Research article Received: 4 February 2015,
Revised: 17 April 2015,
Accepted: 23 April 2015
Published online in Wiley Online Library: 26 May 2015
(wileyonlinelibrary.com) DOI 10.1002/bio.2943
Development of highly sensitive sensor system for methane utilizing cataluminescence Gu. Gonga,b* and Hua Zhua ABSTRACT: A gaseous sensor system was developed for the detection of methane based on its cataluminescence emission. Cataluminescence characteristics and optimal conditions were studied in detail under optimized experimental conditions. Results showed that the methane cataluminescence sensor system could cover a linear detection range from 10 to 5800 ppm (R = 0.9963, n = 7) and the detection limit was about 7 ppm (S/N = 3), which was below the standard permitted concentration. Moreover, a linear discriminant analysis method was used to test the recognizable performance of the methane sensor. It was found that methane, ethane, propane and pentane could be distinguished clearly. Its methane sensing properties, including improved sensitivity, selectivity, stability and recognition demonstrated the TiO2/SnO2 materials to be promising candidates for constructing a cataluminescence-based gas sensor that could be used for detecting explosive gas contaminants. Copyright © 2015 John Wiley & Sons, Ltd. Keywords: cataluminescence; methane; intensity; gas sensor; CO
Introduction
Luminescence 2016; 31: 183–189
* Correspondence to: G. Gong, School of Computer Science and Technology, Jiangsu Normal University, Jiangsu, Xuzhou, 221116, People’s Republic of China. E-mail: [email protected] a
School of Mechatronic Engineering, China University of Mining and Technology, Xuzhou, 221116, People’s Republic of China
b
School of Computer Science and Technology, Jiangsu Normal University, Xuzhou, 221116, People’s Republic of China Abbreviations: ICP, inductively coupled plasma; LDA, linear discriminant analysis; PCA, principal component analysis; RSD, relative standard deviation; RF, radio frequency.
Copyright © 2015 John Wiley & Sons, Ltd.
183
In recent years, as the concentrations of explosive gases such as methane propane, and n-butane have increased, so has the frequency of accidental explosions. The requirement for detection of explosive gases is pressing in order to prevent the occurrence of such accidents. Accordingly, the development of sensor systems that can selectively detect and recognize these types of explosive gases when below the gas concentration limit, in situ, is urgently required (1). Methane, one of the explosive gases, is widely used both in the laboratory and in industry. However, due to its colorless, flammable and explosive nature, any increase in the concentration of methane may lead to an explosion. Therefore, an excellent selective determination method for methane has been eagerly sought (2). Robot ’noses’ have been made with automatic gas analyzers. Recently, researchers have successfully developed a kind of olfactory acuity meter, in which the gas analyzer not only can smell acetone and chloroform, but is also able to sniff out carbon monoxide, which is not detectable by the human olfactory system. In recent years, this gas automatic analyzer has been widely used in detecting gas, for gas analysis in the spacecraft cabin and as an environmental monitor. The principle and display of the gas analyzer is related to electronics, so it was called an ’electronic nose’ (3). The so-called olfactory system of the robot was made by integrating the electronic nose and the computer. Gas sensors systems have become increasingly necessary with the rapid development of modern science and technology (4–8). An excellent system with high sensitivity, plus fast response and recovery speed to gas analytes is needed (9). For example, a sensor array with 16 oxide semiconductor gas-sensing materials, along with an artificial neural network has been developed for detecting and recognizing methane, propane, butane and so on (10), but this sensor system is not appropriate for real time and continuous detection on site. Since the concept of ’cataluminescence’ was first put forward by Breysse and his co-workers for catalytic oxidation of CO on the thorium surface (7), cataluminescence-based sensor
systems have attracted great interest due to their high intensity, signal/noise (S/N) ratio and rapid response. Furthermore, there was almost no consumption of sensor materials during the test process (11). Even though cataluminescence phenomena have been studied on a variety of catalytic materials, more effort is still needed to develop novel excellent catalysts with higher intensities and lower experimental temperatures (12). It is still a great challenge to prepare a novel material to achieve a remarkably enhanced performance for the cataluminescencebased sensor system. In our study, an individual cataluminescence (CTL) sensor system was selected to test on an explosive gas. SnO2 nanomaterials with dense TiO2 particles that could improve the contact opportunity with explosive gas were synthesized successfully. The high emission generated on the surface product by the catalyzing oxidization of methane was detected, and was the so-called cataluminescence phenomenon. The methane sensor device was designed based on a new CTL, and sensitivity, selectivity and response are discussed in detail. The results showed that the melt spun CTL sensors had a fast response and recovery speed toward methane. In addition, linear discriminant analysis (LDA) and principal component analysis (PCA) were used to detect the sensor identification ability.
G. Gong and H. Zhu
Experimental Chemical reagents and materials All chemicals used in the experiments were of high purity (more than 99%) and high analytical grade; they were purchased from the Shanghai Chemical Reagent Company, China and could be used in the experiments without further purification.
tube. When gas flows through the tube, catalysis occurs on the surface of the material. The CTL intensity from the photomultiplier tube (PMT) was collected by an ultraweak chemiluminescence analyzer. The wavelength selected could be changed through the optical filter in the range of 400–600 nm.
Results and discussion Morphology, structure and composition of nanostructures
Synthesis of materials A SnO2/carbon precursor was synthesized by a hydrothermal method, as reported previously (13). The SnO2/carbonaceous precursor was prepared by an inductively coupled plasma (ICP) treatment device, including one three-neck glass bottle and several cycles of copper wire wound round the glass bottle. The diameter and length of the bottle were 5 and 20 cm, respectively. A radio frequency (RF) of 13.5 MHz was fed into the wire. The system was flushed with argon at flow rate, then pulled back to 10 Pa. After washing and discarding three times, and continuous stirring for 30 min under oxygen, the SnO2/carbon precursor liquidity ratio was 5:1. After that, the plasma-treated samples were dried in a vacuum oven at 60 °C for further use. The nano TiO2-modified SnO2 nanomaterial preparation and adsorption process was followed by the annealing treatment. For TiO2/SnO2 nanomaterial preparation, firstly the SnO2/carbon precursor plasma was processed (0.2 g) by ultrasonic dispersion 10 mL (0.1 M) in aqueous solution and by dissolving certain amounts of titanium in ethanol. Then, the mixed solution was left to stand at room temperature. Secondly, after several washes and centrifugations, the samples were dried in an oven at 70 °C. Finally, the samples were heated in an annealing furnace process in a controlled air environment.
Apparatus and procedures A diagram of the CTL detection system is shown in Fig. 1. Materials for testing were placed on a heating rod in a quartz tube inner with diameter of about 10 mm. The air pump could not only control flow but could also disperse test gas. The gas filter, which is a filtering net, was installed in the gas input port. Material dispersed to form a suspension in ethanol and then to form a layer of material, of 0.5 mm thick in the heating rod. The working temperature could be adjusted by controlling the voltage of the quartz
Field-emission scanning electron microscopy (FESEM; FEI Sirion-180), high-resolution transmission electron microscopy (HRTEM; JEM2010) and X-ray diffractometer (XRD; Philips X’ PERT PRO) analyses were used to study the morphology, structure and composition of the experimental samples. The XRD spectra of pure samples modified by TiO2 are shown in Fig. 2. As shown in Fig. 2(A), all the diffraction peaks could be easily identified as belonging to a SnO2 lattice consistent with a rutile structure of A = 0.48 nm and C = 0.32 nm, and are in good agreement with values reported in the literature ( JCPDS no.41-1445). From Fig. 2(B), the TiO2 phase could be observed ( JCPDS no. 84-1286). Figure 3 shows the Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) of the plasma-treated pure SnO2 and TiO2/SnO2 materials. As can be seen from the pure SnO2 (Fig. 3a), the shape retains a coral-like outline, without destroying the structure. Figure 3(b) gives the TiO2/SnO2 nanostructure, a porous and hollow structure can be clearly observed. In contrast, as shown in Fig. 3(c), the surface of the TiO2/SnO2 nanostructures was rougher compared with that of pure SnO2. It was modified by TiO2 with a high density on the surface of SnO2. This modification may be due to overlap in such a high density of nanoparticles. The composition of the as-prepared samples was investigated by X-ray Photoelectron Spectrometer (XPS) analysis, as shown in Fig. 4. As indicated in Fig. 5, the Ti component could be detected in TiO2/SnO2. A valley-type shape is clearly presented, indicating that the nanostructure is hollow and porous. Moreover, the quantitative results of elements corresponding to the XPS spectra are shown in Table 1. It was found that the molar ratio of Ti to Sn in the TiO2/SnO2 was ca. 0.45:1. Cataluminescence application Enhanced CTL intensity on TiO2/SnO2 compared with pure SnO2. In our study, with a controlled optimized flow rate of
184
Figure 1. Schematic diagram of the cataluminescence-based sensor system.
wileyonlinelibrary.com/journal/luminescence
Copyright © 2015 John Wiley & Sons, Ltd.
Luminescence 2016; 31: 183–189
Development of highly sensitive sensor system
Figure 2. XRD patterns of pure SnO2 materials and TiO2/SnO2 composites.
230 mL/min, it was shown that this suitable flow resulted in faster cycle determination (14,15). Molecular volatile organic compounds could rapidly and fully contact with oxygen at 230 mL/min. From Fig. 6, an enhanced cataluminescence performance of methane using chemiluminescence on nano-sized TiO2/SnO2 was observed and compared with pure SnO2 at 270 °C. The TiO2/SnO2 nanostructures exhibited a strong reaction for methane, which confirmed it was an excellent methane sensor using CTL. It was clear that the TiO2/SnO2 composite exhibited strong and steady CTL responses to methane. In addition, both the response time (5 sec) and the recovery time (10 sec) were very short. The decorated nanocomposites showed superior optical and catalytic properties compared with their individual single-component materials. The introduction of TiO2 into the composite was thought to play an important role in the enhanced CTL sensing performance for methane. Therefore, further research on the detection characteristics of methane CTL was carried out. Wavelength selection. The cataluminescence emission wavelength of methane on the surface of materials was first tested with optical filters ranging from 400 to 590 nm (400, 420, 455, 480, 500, 520, 548 and 590 nm). Figure 7 showed the cataluminescence intensity of methane at different wavelengths under a flow rate of 230 mL/min and temperature 270 °C. The noise signal increased with the longer wavelength radiation from incandescent lamps from a ceramic heater. The S/N ratio was used to calculate the real luminescence intensity. The radiated noise signal from the incandescent lamps was higher at longer wavelengths as the ceramic heater increased in incandescence at higher temperatures. Figure 7 shows that the maximal emission of methane was at 455 nm with the maximum signal to noise (S/N) ratio. Therefore, this wavelength (455 nm) was selected as the optimal test wavelength for the current cataluminescence-based methane sensor.
Luminescence 2016; 31: 183–189
Figure 3. FESEM images of (a) pure SnO2 materials; (b) TiO2/SnO2 structures; and (c) TEM images of TiO2/SnO2 particles.
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/luminescence
185
Optimization of working temperature. The influence of experimental temperature on the cataluminescence sensor system is still a very important challenge for system design. As can be seen from Fig. 8, the cataluminescence intensity of methane increased with test temperature, a finding that could be attributed to the larger conversion of methane on TiO2/SnO2 particles at higher temperatures. The high performance at very high temperatures was not suitable because of the unfavourable signal to noise (S/N) ratio. As can be seen from Fig. 8, due to the noise increase, a significant decrease in signal to noise (S/N) ratio was
G. Gong and H. Zhu
Figure 6. Response curves of the cataluminescence methane sensor based on TiO2/SnO2 compared with pure SnO2. Concentration: 180 ppm. Flow rate of methane: 230 mL/min, wavelength: 420 nm, and the temperature of cataluminescence sensor: 270 °C.
Figure 4. XPS survey spectra of (a) pureSnO2; and (b) TiO2/SnO2.
Figure 7. The cataluminescence intensity of methane changed at different wavelengths. Temperature of cataluminescence sensor: 286 °C. Air flow rate: 230 mL/min. Concentration: 180 ppm.
Figure 5. The compositional lines of Sn and Ti.
Table 1. Atomic percentages of the pure SnO2 nanostructures and TiO2/SnO2 Atomic percentage (at%)
Pure SnO2 TiO2/SnO2
Sn 3d
Ti 2p
O 1s
C 1s
22.1 20.6
– 9.2
61.7 52.9
16.2 17.3
found at about 286 °C. In view of these results, the test temperature of 286 °C was finally chosen for our further study.
186
Analytical characteristics of the sensor system. Under the discussed selected test conditions, the concentration of methane in the range 10 to 5800 ppm was selected for further calibration. As can be seen from Fig. 9, the cataluminescence sensing performance of the TiO2/SnO2 materials was tested by the response of the cataluminescence sensor system to methane at different concentrations (50, 850 and 1200 ppm). The flow injection mode was applied in our research by repeatedly injecting different volumes of methane gas. The calibration curve with a dynamic range for methane gas was acquired. It was obvious that these TiO2/SnO2
wileyonlinelibrary.com/journal/luminescence
Figure 8. The cataluminescence intensity and signal to noise ratio for methane on TiO2/SnO2 materials at different temperatures. Concentration of methane: 180 ppm. Air flow rate: 230 mL/min. Selected wavelength: 455 nm.
materials showed a strong and steady cataluminescence response intensity to methane, the relative cataluminescence intensity increased with the methane concentration.
Copyright © 2015 John Wiley & Sons, Ltd.
Luminescence 2016; 31: 183–189
Development of highly sensitive sensor system
Figure 9. Cataluminescence curve between CTL intensity and methane concentration. Air flow rate: 230 mL/min. Temperature: 286 °C. Selected wavelength: 455 nm.
Moreover, the regression equation of the cataluminescence intensity of methane concentration was linear in the range 10 to 5800 ppm. The linear regression equation was Y = 67.2X + 1318.4 (R = 0.9963, n = 7), where Y was the average relative cataluminescence responses intensity of replicate teats at the same concentration value, X was the concentration of methane gas, and R was the regression coefficient. The good relative standard deviation (RSD) of the responses (Fig. 9) showed excellent repeatability and linearity for this cataluminescence determination. Selectivity and stability of the CTL-based methane sensor. In our study, we assessed if the CTL-based methane sensor showed an excellent selectivity towards methane by testing another nine common possible gases, including ethane, propane, pentane, hydrogen, kerosene, diesel oil, ethylene, toluene, xylene under the same experimental conditions at a concentration of 180 ppm at 286 °C and with a flow rate of 230 mL/min. As shown in Fig. 10, there was hardly any intensity from the surface of CTL sensor for hydrogen, kerosene, diesel oil, ethylene, toluene and xylene, while for the other gases cataluminescence intermediates were produced and the detector was sensitive enough to test specifically for methane.
Luminescence 2016; 31: 183–189
Possible mechanism. Although many studies have been conducted on TiO2- or SnO2-based sensors, their gas-sensing mechanisms remain controversial (16–19). The mechanisms for improved performance were discussed based on experimental results. In
Figure 11. Linear discriminant analysis results acquired from the methane, ethane, propane and pentane patterns.
Figure 12. Principal component analysis results acquired from the methane, ethane, propane and pentane patterns.
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/luminescence
187
Figure 10. Selectivity towards methane of the cataluminescence-based sensor. Flow rate of carrier gas: 230 mL/min. Temperature of cataluminescence sensor: 286 °C. Concentration of eight kinds VOCs: 180 ppm. Selected wavelength: 455 nm.
In our study, the stability and durability of this cataluminescence sensor system based on TiO2/SnO2 nanostructure were also discussed. The CTL intensity of the TiO2/SnO2 material to methane vapors did not change markedly after several days under the optimal detecting conditions, therefore it can be concluded that this enhanced CTL performance on TiO2/SnO2 material was stable. In Figs. 11 and 12, the LDA and PCA results showed that different point sets acquired from the ethane, propane and pentane patterns were well recognized. The canonical patterns clustered into four different groups to achieve good discrimination. The first three canonical factors contained 64.88, 27.54 and 5.81% of the variation, occupying 98.23% of total variation. As can be seen from Fig. 11, each chemical was tested five times and the variances plotted as PC1, PC2 and PC3 were 84.82, 8.52 and 5.97%, respectively. The three-dimensional LDA and PCA scoreboard showed clear coding of the five injections, and the classification rate was excellent. The pattern recognition results of the methane gas illustrated the good selectivity and repeatability of this CTL-based sensor system.
G. Gong and H. Zhu Table 2. Analysis of methane in the artificial samples Sample no.
Composition
1
Methane Ethanol Methane Formaldehyde Methane Acetaldehyde Methane Diethyl ether
2 3 4
Standard values (ppm)
Average measured values (ppm, n = 7)
1000 800 1000 500 1000 300 1000 300
954 782 977 506 1046 298 1036 298
Recovery (%)a
RSD (%)
95.4 97.8 97.7 101.2 104.6 99.3 103.6 99.3
1.1 1.4 1.7 1.2 1.3 0.8 1.3 0.8
a
Recovery (%) is defined as a percentage of measured value/standard value.
present study, it is difficult for the gas molecules to diffuse into the interior of the sensing layer, resulting in a poor utilization rate of the sensing materials located internally. Only the materials located at the surface of the sensing layer can contact the gas analytes effectively. If the small gas molecules diffuse into the deep location of the sensing layer, the gas sensor would exhibit excellent performance. As reported in a previous study, reagents immobilization onto proper substrates play an important role in the development of high-quality CTL-based sensors. A widely accepted CTL mechanism was the formation of an intermediate of a highly reactive endoperoxide. The acquired high selectivity of methane was possibly attributed to the formation of cataluminescence intermediates. Moreover, it is known that gases could be detected not only when they were catalytically oxidized on the catalyst but also when they produced excited intermediates that emitted light during the transition to the electronic ground state. The CTL intermediate quantity that ether produced was larger than that of other materials, although many gases could be catalytically oxidized on nano-sized materials (20,21). When VOC gases were introduced, the oxidization reaction took place on the sensor surface in the following way: VOC þ O →VOC O þ e VOC þ O2 →VOC O þ 2e when the CTL-based gas sensor is exposed to reducing gases such as methane, some chemical reactions between oxygen species and gas analytes will occur, resulting in the release of electrons back to TiO2/SnO2 and thereby increasing the response. The results presented in this paper demonstrate that TiO2–SnO2 had a significantly higher sensor response. Moreover, it is known that gases can be detected not only when they are catalytically oxidized on the catalyst but also when they produce excited intermediates emitting light during the transition to the electronic ground state (20). In air surroundings, O2 is adsorbed onto the surface of the TiO2/SnO2 microstructures. It will capture and react with electrons from TiO2/SnO2 to produce negative oxygen species. When the CTL-based gas sensor is exposed to methane, some chemical reactions between oxygen species and gas analytes will occur, resulting in the release of electrons back to TiO2/SnO2 and thereby increasing the current of the circuit.
188
Sample analysis. In our study, we collected methane air samples from a coal mine, methane was not detected using the gas chromatography (GC) method. Methane, ethanol, formaldehyde,
wileyonlinelibrary.com/journal/luminescence
acetaldehyde and diethyl ether were chosen as test samples. Detection results for the CTL sensor are shown in Table 2, the recovery of methane (n = 7) ranged from 95.4 to 104.6%.
Conclusion In recent years, the application of gas sensors in the chemiluminescence analysis field has made great progress, and has benefitted from the development of nanoscience techniques. A fast response catalyzed by nanometer luminescent materials in the gas sensor was found. With low operation cost, easy miniaturization, and the in-depth study of material, the sensor also has significant development prospects, and follows the current direction for sensor detector research detection and multicomponent development. This assay has important significance for the detection of explosive gases. The successful preparation here of TiO2/SnO2 materials is meaningful for practical application and fundamental research. In our study, we describe a synthesized material that is a novel structure as a chemiluminescence sensor due to its excellent sensitivity and selectively for methane gas. The chemiluminescence-based sensor showed excellent analytical indices such as rapid intensity, high sensitivity and outstanding selectivity, and promises a wide application in explosive gas detection and environmental monitoring. Acknowledgments This project is supported by the National Basic Research Program of China (863 Program grant no. 2012AA041504). The authors are also grateful to the editor and the anonymous reviewers for their useful comments and advice, which were vital for improving the quality of this paper.
Conflicts of interest The authors declare no conflicts of interest.
References 1. Zhang YM, Wu YL, Wu XZ, Xi X, Wang JQ. A novel vibration mode testing method for cylindrical resonators based on microphones. Sensors 2015;15:1954–63. 2. Cao XA, Zhang XR. A research on determination of explosive gases utilizing cataluminescence sensor array. Luminescence 2011;20:243–50. 3. Li B, Sauve G, Iovu MC, Zhang R, Cooper J, Santhanam S, et al. Volatile organic compound detection using nanostructured copolymers. Nano Lett 2012;6:1598–602.
Copyright © 2015 John Wiley & Sons, Ltd.
Luminescence 2016; 31: 183–189
Development of highly sensitive sensor system 4. Lin HB, Shih JS. Fullerene C60-cryptand coated surface acoustic wave quartz crystal sensor for organic vapors. Sens Actuators B Chem 2010;92:243–54. 5. Lee DS, Jung HY, Lim JW, et al. Explosive gas recognition system using thick film sensor array and neural network. Sens Actuators B Chem 2008;71:89–98. 6. Hu J, Xu KL, Jia YZ, Lv Y, Li YB, Hou XD. Oxidation of ethyl ether on borate glass: chemiluminescence, mechanism, and development of a sensitive gas sensor. Anal Chem 2009;80:7964–9. 7. Potyrailo RA, Mirsky VM. Combinatorial and high-throughput development of sensing materials: the first 10 years. Chem Rev 2008;108:770–813. 8. Breysse M, Claudel B, Faure L, Guenin M, Roberto J, Williams J. Chemiluminescence during the catalysis of carbon monoxide oxidation on a thoria surface. J Catal 1976;45:137–44. 9. Wang QH, Xie LJ, Zhu B, Zheng Y, Cao SH. Harmful gas recognition exploiting a CTL sensor array. Sensors 2013;13:13509–20. 10. Shi GL, Sun B, Jin Z, Liu JH, Li MQ. Synthesis of SiO2/Fe3O4 nanomaterial and its application as cataluminescence gas sensor material for ether. Sens Actuators B Chem 2012;171–172:699–704. 11. Cao XA, Zhang ZY, Zhang XR. A novel gaseous acetaldehyde sensor utilizing cataluminescence on nanosized BaCO3. Sens Actuators B Chem 2004;99:30–5. 12. Rao ZM, Liu LJ, Xie JY, Zeng YY. Development of a benzene vapour sensor utilizing chemiluminescence on nanosized Y2O3. Luminescence 2008;23:163–8.
13. Zhou KW, Ji XL, Zhang N, Zhang XR. On-line monitoring of formaldehyde in air by cataluminescence-based gas sensor. Sens Actuators B Chem 2006;119:392–7. 14. Zhu YF, Shi JJ, Zhang ZY, Zhang C, Zhang XR. Development of a gas sensor utilizing chemiluminescence on nanosized titanium dioxide. Anal Chem 2002;74:120–4. 15. Wu YY, Zhang SC, Na N, Wang X, Zhang XR. A novel gaseous ester sensor utilizing chemiluminescence on nano-sized SiO2. Sens Actuators B Chem 2007;126:461–6. 16. Kawi S, Tang YP, Hidajat K, Yu LE. Synthesis and characterization of nanoscale CeO2 catalyst for deNOx. J Metastab Nanocryst Mater 2005;23:95–8. 17. Zhao Y, Wei W, Meiping H, Shen P, Minli Z, Wang W. An effective secure routing way to reduce energy consumption in wireless sensor networks. AISS 2012;4:277–83. 18. Llinas J, Hall DL. An introduction to multi-sensor data fusion. In: Proceedings of the 1998 IEEE International Symposium on Circuits and Systems (ISCAS 1998), 31 May–3 June 1998; Monterey, CA, USA 1998:537–40. 19. Mao SY, Zhao W. Comments on multisensor image fusion techniques. J Beijing Univ Aeronaut Astronaut 2002;28:512–8. 20. Zhao Y, Wei W, Tang X.J, Zhao ML. Choice of optimized scheme parameters of neural networks based on error handling. AISS 2012;4:76–82. 21. Zhao Y, Wei W, Song HB, Li W. The hardware design of intelligent circuit breaker. Energ Educ Sci Technol A 2014;33:695–702.
189
Luminescence 2016; 31: 183–189
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/luminescence
Revised: 17 April 2015,
Accepted: 23 April 2015
Published online in Wiley Online Library: 26 May 2015
(wileyonlinelibrary.com) DOI 10.1002/bio.2943
Development of highly sensitive sensor system for methane utilizing cataluminescence Gu. Gonga,b* and Hua Zhua ABSTRACT: A gaseous sensor system was developed for the detection of methane based on its cataluminescence emission. Cataluminescence characteristics and optimal conditions were studied in detail under optimized experimental conditions. Results showed that the methane cataluminescence sensor system could cover a linear detection range from 10 to 5800 ppm (R = 0.9963, n = 7) and the detection limit was about 7 ppm (S/N = 3), which was below the standard permitted concentration. Moreover, a linear discriminant analysis method was used to test the recognizable performance of the methane sensor. It was found that methane, ethane, propane and pentane could be distinguished clearly. Its methane sensing properties, including improved sensitivity, selectivity, stability and recognition demonstrated the TiO2/SnO2 materials to be promising candidates for constructing a cataluminescence-based gas sensor that could be used for detecting explosive gas contaminants. Copyright © 2015 John Wiley & Sons, Ltd. Keywords: cataluminescence; methane; intensity; gas sensor; CO
Introduction
Luminescence 2016; 31: 183–189
* Correspondence to: G. Gong, School of Computer Science and Technology, Jiangsu Normal University, Jiangsu, Xuzhou, 221116, People’s Republic of China. E-mail: [email protected] a
School of Mechatronic Engineering, China University of Mining and Technology, Xuzhou, 221116, People’s Republic of China
b
School of Computer Science and Technology, Jiangsu Normal University, Xuzhou, 221116, People’s Republic of China Abbreviations: ICP, inductively coupled plasma; LDA, linear discriminant analysis; PCA, principal component analysis; RSD, relative standard deviation; RF, radio frequency.
Copyright © 2015 John Wiley & Sons, Ltd.
183
In recent years, as the concentrations of explosive gases such as methane propane, and n-butane have increased, so has the frequency of accidental explosions. The requirement for detection of explosive gases is pressing in order to prevent the occurrence of such accidents. Accordingly, the development of sensor systems that can selectively detect and recognize these types of explosive gases when below the gas concentration limit, in situ, is urgently required (1). Methane, one of the explosive gases, is widely used both in the laboratory and in industry. However, due to its colorless, flammable and explosive nature, any increase in the concentration of methane may lead to an explosion. Therefore, an excellent selective determination method for methane has been eagerly sought (2). Robot ’noses’ have been made with automatic gas analyzers. Recently, researchers have successfully developed a kind of olfactory acuity meter, in which the gas analyzer not only can smell acetone and chloroform, but is also able to sniff out carbon monoxide, which is not detectable by the human olfactory system. In recent years, this gas automatic analyzer has been widely used in detecting gas, for gas analysis in the spacecraft cabin and as an environmental monitor. The principle and display of the gas analyzer is related to electronics, so it was called an ’electronic nose’ (3). The so-called olfactory system of the robot was made by integrating the electronic nose and the computer. Gas sensors systems have become increasingly necessary with the rapid development of modern science and technology (4–8). An excellent system with high sensitivity, plus fast response and recovery speed to gas analytes is needed (9). For example, a sensor array with 16 oxide semiconductor gas-sensing materials, along with an artificial neural network has been developed for detecting and recognizing methane, propane, butane and so on (10), but this sensor system is not appropriate for real time and continuous detection on site. Since the concept of ’cataluminescence’ was first put forward by Breysse and his co-workers for catalytic oxidation of CO on the thorium surface (7), cataluminescence-based sensor
systems have attracted great interest due to their high intensity, signal/noise (S/N) ratio and rapid response. Furthermore, there was almost no consumption of sensor materials during the test process (11). Even though cataluminescence phenomena have been studied on a variety of catalytic materials, more effort is still needed to develop novel excellent catalysts with higher intensities and lower experimental temperatures (12). It is still a great challenge to prepare a novel material to achieve a remarkably enhanced performance for the cataluminescencebased sensor system. In our study, an individual cataluminescence (CTL) sensor system was selected to test on an explosive gas. SnO2 nanomaterials with dense TiO2 particles that could improve the contact opportunity with explosive gas were synthesized successfully. The high emission generated on the surface product by the catalyzing oxidization of methane was detected, and was the so-called cataluminescence phenomenon. The methane sensor device was designed based on a new CTL, and sensitivity, selectivity and response are discussed in detail. The results showed that the melt spun CTL sensors had a fast response and recovery speed toward methane. In addition, linear discriminant analysis (LDA) and principal component analysis (PCA) were used to detect the sensor identification ability.
G. Gong and H. Zhu
Experimental Chemical reagents and materials All chemicals used in the experiments were of high purity (more than 99%) and high analytical grade; they were purchased from the Shanghai Chemical Reagent Company, China and could be used in the experiments without further purification.
tube. When gas flows through the tube, catalysis occurs on the surface of the material. The CTL intensity from the photomultiplier tube (PMT) was collected by an ultraweak chemiluminescence analyzer. The wavelength selected could be changed through the optical filter in the range of 400–600 nm.
Results and discussion Morphology, structure and composition of nanostructures
Synthesis of materials A SnO2/carbon precursor was synthesized by a hydrothermal method, as reported previously (13). The SnO2/carbonaceous precursor was prepared by an inductively coupled plasma (ICP) treatment device, including one three-neck glass bottle and several cycles of copper wire wound round the glass bottle. The diameter and length of the bottle were 5 and 20 cm, respectively. A radio frequency (RF) of 13.5 MHz was fed into the wire. The system was flushed with argon at flow rate, then pulled back to 10 Pa. After washing and discarding three times, and continuous stirring for 30 min under oxygen, the SnO2/carbon precursor liquidity ratio was 5:1. After that, the plasma-treated samples were dried in a vacuum oven at 60 °C for further use. The nano TiO2-modified SnO2 nanomaterial preparation and adsorption process was followed by the annealing treatment. For TiO2/SnO2 nanomaterial preparation, firstly the SnO2/carbon precursor plasma was processed (0.2 g) by ultrasonic dispersion 10 mL (0.1 M) in aqueous solution and by dissolving certain amounts of titanium in ethanol. Then, the mixed solution was left to stand at room temperature. Secondly, after several washes and centrifugations, the samples were dried in an oven at 70 °C. Finally, the samples were heated in an annealing furnace process in a controlled air environment.
Apparatus and procedures A diagram of the CTL detection system is shown in Fig. 1. Materials for testing were placed on a heating rod in a quartz tube inner with diameter of about 10 mm. The air pump could not only control flow but could also disperse test gas. The gas filter, which is a filtering net, was installed in the gas input port. Material dispersed to form a suspension in ethanol and then to form a layer of material, of 0.5 mm thick in the heating rod. The working temperature could be adjusted by controlling the voltage of the quartz
Field-emission scanning electron microscopy (FESEM; FEI Sirion-180), high-resolution transmission electron microscopy (HRTEM; JEM2010) and X-ray diffractometer (XRD; Philips X’ PERT PRO) analyses were used to study the morphology, structure and composition of the experimental samples. The XRD spectra of pure samples modified by TiO2 are shown in Fig. 2. As shown in Fig. 2(A), all the diffraction peaks could be easily identified as belonging to a SnO2 lattice consistent with a rutile structure of A = 0.48 nm and C = 0.32 nm, and are in good agreement with values reported in the literature ( JCPDS no.41-1445). From Fig. 2(B), the TiO2 phase could be observed ( JCPDS no. 84-1286). Figure 3 shows the Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) of the plasma-treated pure SnO2 and TiO2/SnO2 materials. As can be seen from the pure SnO2 (Fig. 3a), the shape retains a coral-like outline, without destroying the structure. Figure 3(b) gives the TiO2/SnO2 nanostructure, a porous and hollow structure can be clearly observed. In contrast, as shown in Fig. 3(c), the surface of the TiO2/SnO2 nanostructures was rougher compared with that of pure SnO2. It was modified by TiO2 with a high density on the surface of SnO2. This modification may be due to overlap in such a high density of nanoparticles. The composition of the as-prepared samples was investigated by X-ray Photoelectron Spectrometer (XPS) analysis, as shown in Fig. 4. As indicated in Fig. 5, the Ti component could be detected in TiO2/SnO2. A valley-type shape is clearly presented, indicating that the nanostructure is hollow and porous. Moreover, the quantitative results of elements corresponding to the XPS spectra are shown in Table 1. It was found that the molar ratio of Ti to Sn in the TiO2/SnO2 was ca. 0.45:1. Cataluminescence application Enhanced CTL intensity on TiO2/SnO2 compared with pure SnO2. In our study, with a controlled optimized flow rate of
184
Figure 1. Schematic diagram of the cataluminescence-based sensor system.
wileyonlinelibrary.com/journal/luminescence
Copyright © 2015 John Wiley & Sons, Ltd.
Luminescence 2016; 31: 183–189
Development of highly sensitive sensor system
Figure 2. XRD patterns of pure SnO2 materials and TiO2/SnO2 composites.
230 mL/min, it was shown that this suitable flow resulted in faster cycle determination (14,15). Molecular volatile organic compounds could rapidly and fully contact with oxygen at 230 mL/min. From Fig. 6, an enhanced cataluminescence performance of methane using chemiluminescence on nano-sized TiO2/SnO2 was observed and compared with pure SnO2 at 270 °C. The TiO2/SnO2 nanostructures exhibited a strong reaction for methane, which confirmed it was an excellent methane sensor using CTL. It was clear that the TiO2/SnO2 composite exhibited strong and steady CTL responses to methane. In addition, both the response time (5 sec) and the recovery time (10 sec) were very short. The decorated nanocomposites showed superior optical and catalytic properties compared with their individual single-component materials. The introduction of TiO2 into the composite was thought to play an important role in the enhanced CTL sensing performance for methane. Therefore, further research on the detection characteristics of methane CTL was carried out. Wavelength selection. The cataluminescence emission wavelength of methane on the surface of materials was first tested with optical filters ranging from 400 to 590 nm (400, 420, 455, 480, 500, 520, 548 and 590 nm). Figure 7 showed the cataluminescence intensity of methane at different wavelengths under a flow rate of 230 mL/min and temperature 270 °C. The noise signal increased with the longer wavelength radiation from incandescent lamps from a ceramic heater. The S/N ratio was used to calculate the real luminescence intensity. The radiated noise signal from the incandescent lamps was higher at longer wavelengths as the ceramic heater increased in incandescence at higher temperatures. Figure 7 shows that the maximal emission of methane was at 455 nm with the maximum signal to noise (S/N) ratio. Therefore, this wavelength (455 nm) was selected as the optimal test wavelength for the current cataluminescence-based methane sensor.
Luminescence 2016; 31: 183–189
Figure 3. FESEM images of (a) pure SnO2 materials; (b) TiO2/SnO2 structures; and (c) TEM images of TiO2/SnO2 particles.
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/luminescence
185
Optimization of working temperature. The influence of experimental temperature on the cataluminescence sensor system is still a very important challenge for system design. As can be seen from Fig. 8, the cataluminescence intensity of methane increased with test temperature, a finding that could be attributed to the larger conversion of methane on TiO2/SnO2 particles at higher temperatures. The high performance at very high temperatures was not suitable because of the unfavourable signal to noise (S/N) ratio. As can be seen from Fig. 8, due to the noise increase, a significant decrease in signal to noise (S/N) ratio was
G. Gong and H. Zhu
Figure 6. Response curves of the cataluminescence methane sensor based on TiO2/SnO2 compared with pure SnO2. Concentration: 180 ppm. Flow rate of methane: 230 mL/min, wavelength: 420 nm, and the temperature of cataluminescence sensor: 270 °C.
Figure 4. XPS survey spectra of (a) pureSnO2; and (b) TiO2/SnO2.
Figure 7. The cataluminescence intensity of methane changed at different wavelengths. Temperature of cataluminescence sensor: 286 °C. Air flow rate: 230 mL/min. Concentration: 180 ppm.
Figure 5. The compositional lines of Sn and Ti.
Table 1. Atomic percentages of the pure SnO2 nanostructures and TiO2/SnO2 Atomic percentage (at%)
Pure SnO2 TiO2/SnO2
Sn 3d
Ti 2p
O 1s
C 1s
22.1 20.6
– 9.2
61.7 52.9
16.2 17.3
found at about 286 °C. In view of these results, the test temperature of 286 °C was finally chosen for our further study.
186
Analytical characteristics of the sensor system. Under the discussed selected test conditions, the concentration of methane in the range 10 to 5800 ppm was selected for further calibration. As can be seen from Fig. 9, the cataluminescence sensing performance of the TiO2/SnO2 materials was tested by the response of the cataluminescence sensor system to methane at different concentrations (50, 850 and 1200 ppm). The flow injection mode was applied in our research by repeatedly injecting different volumes of methane gas. The calibration curve with a dynamic range for methane gas was acquired. It was obvious that these TiO2/SnO2
wileyonlinelibrary.com/journal/luminescence
Figure 8. The cataluminescence intensity and signal to noise ratio for methane on TiO2/SnO2 materials at different temperatures. Concentration of methane: 180 ppm. Air flow rate: 230 mL/min. Selected wavelength: 455 nm.
materials showed a strong and steady cataluminescence response intensity to methane, the relative cataluminescence intensity increased with the methane concentration.
Copyright © 2015 John Wiley & Sons, Ltd.
Luminescence 2016; 31: 183–189
Development of highly sensitive sensor system
Figure 9. Cataluminescence curve between CTL intensity and methane concentration. Air flow rate: 230 mL/min. Temperature: 286 °C. Selected wavelength: 455 nm.
Moreover, the regression equation of the cataluminescence intensity of methane concentration was linear in the range 10 to 5800 ppm. The linear regression equation was Y = 67.2X + 1318.4 (R = 0.9963, n = 7), where Y was the average relative cataluminescence responses intensity of replicate teats at the same concentration value, X was the concentration of methane gas, and R was the regression coefficient. The good relative standard deviation (RSD) of the responses (Fig. 9) showed excellent repeatability and linearity for this cataluminescence determination. Selectivity and stability of the CTL-based methane sensor. In our study, we assessed if the CTL-based methane sensor showed an excellent selectivity towards methane by testing another nine common possible gases, including ethane, propane, pentane, hydrogen, kerosene, diesel oil, ethylene, toluene, xylene under the same experimental conditions at a concentration of 180 ppm at 286 °C and with a flow rate of 230 mL/min. As shown in Fig. 10, there was hardly any intensity from the surface of CTL sensor for hydrogen, kerosene, diesel oil, ethylene, toluene and xylene, while for the other gases cataluminescence intermediates were produced and the detector was sensitive enough to test specifically for methane.
Luminescence 2016; 31: 183–189
Possible mechanism. Although many studies have been conducted on TiO2- or SnO2-based sensors, their gas-sensing mechanisms remain controversial (16–19). The mechanisms for improved performance were discussed based on experimental results. In
Figure 11. Linear discriminant analysis results acquired from the methane, ethane, propane and pentane patterns.
Figure 12. Principal component analysis results acquired from the methane, ethane, propane and pentane patterns.
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/luminescence
187
Figure 10. Selectivity towards methane of the cataluminescence-based sensor. Flow rate of carrier gas: 230 mL/min. Temperature of cataluminescence sensor: 286 °C. Concentration of eight kinds VOCs: 180 ppm. Selected wavelength: 455 nm.
In our study, the stability and durability of this cataluminescence sensor system based on TiO2/SnO2 nanostructure were also discussed. The CTL intensity of the TiO2/SnO2 material to methane vapors did not change markedly after several days under the optimal detecting conditions, therefore it can be concluded that this enhanced CTL performance on TiO2/SnO2 material was stable. In Figs. 11 and 12, the LDA and PCA results showed that different point sets acquired from the ethane, propane and pentane patterns were well recognized. The canonical patterns clustered into four different groups to achieve good discrimination. The first three canonical factors contained 64.88, 27.54 and 5.81% of the variation, occupying 98.23% of total variation. As can be seen from Fig. 11, each chemical was tested five times and the variances plotted as PC1, PC2 and PC3 were 84.82, 8.52 and 5.97%, respectively. The three-dimensional LDA and PCA scoreboard showed clear coding of the five injections, and the classification rate was excellent. The pattern recognition results of the methane gas illustrated the good selectivity and repeatability of this CTL-based sensor system.
G. Gong and H. Zhu Table 2. Analysis of methane in the artificial samples Sample no.
Composition
1
Methane Ethanol Methane Formaldehyde Methane Acetaldehyde Methane Diethyl ether
2 3 4
Standard values (ppm)
Average measured values (ppm, n = 7)
1000 800 1000 500 1000 300 1000 300
954 782 977 506 1046 298 1036 298
Recovery (%)a
RSD (%)
95.4 97.8 97.7 101.2 104.6 99.3 103.6 99.3
1.1 1.4 1.7 1.2 1.3 0.8 1.3 0.8
a
Recovery (%) is defined as a percentage of measured value/standard value.
present study, it is difficult for the gas molecules to diffuse into the interior of the sensing layer, resulting in a poor utilization rate of the sensing materials located internally. Only the materials located at the surface of the sensing layer can contact the gas analytes effectively. If the small gas molecules diffuse into the deep location of the sensing layer, the gas sensor would exhibit excellent performance. As reported in a previous study, reagents immobilization onto proper substrates play an important role in the development of high-quality CTL-based sensors. A widely accepted CTL mechanism was the formation of an intermediate of a highly reactive endoperoxide. The acquired high selectivity of methane was possibly attributed to the formation of cataluminescence intermediates. Moreover, it is known that gases could be detected not only when they were catalytically oxidized on the catalyst but also when they produced excited intermediates that emitted light during the transition to the electronic ground state. The CTL intermediate quantity that ether produced was larger than that of other materials, although many gases could be catalytically oxidized on nano-sized materials (20,21). When VOC gases were introduced, the oxidization reaction took place on the sensor surface in the following way: VOC þ O →VOC O þ e VOC þ O2 →VOC O þ 2e when the CTL-based gas sensor is exposed to reducing gases such as methane, some chemical reactions between oxygen species and gas analytes will occur, resulting in the release of electrons back to TiO2/SnO2 and thereby increasing the response. The results presented in this paper demonstrate that TiO2–SnO2 had a significantly higher sensor response. Moreover, it is known that gases can be detected not only when they are catalytically oxidized on the catalyst but also when they produce excited intermediates emitting light during the transition to the electronic ground state (20). In air surroundings, O2 is adsorbed onto the surface of the TiO2/SnO2 microstructures. It will capture and react with electrons from TiO2/SnO2 to produce negative oxygen species. When the CTL-based gas sensor is exposed to methane, some chemical reactions between oxygen species and gas analytes will occur, resulting in the release of electrons back to TiO2/SnO2 and thereby increasing the current of the circuit.
188
Sample analysis. In our study, we collected methane air samples from a coal mine, methane was not detected using the gas chromatography (GC) method. Methane, ethanol, formaldehyde,
wileyonlinelibrary.com/journal/luminescence
acetaldehyde and diethyl ether were chosen as test samples. Detection results for the CTL sensor are shown in Table 2, the recovery of methane (n = 7) ranged from 95.4 to 104.6%.
Conclusion In recent years, the application of gas sensors in the chemiluminescence analysis field has made great progress, and has benefitted from the development of nanoscience techniques. A fast response catalyzed by nanometer luminescent materials in the gas sensor was found. With low operation cost, easy miniaturization, and the in-depth study of material, the sensor also has significant development prospects, and follows the current direction for sensor detector research detection and multicomponent development. This assay has important significance for the detection of explosive gases. The successful preparation here of TiO2/SnO2 materials is meaningful for practical application and fundamental research. In our study, we describe a synthesized material that is a novel structure as a chemiluminescence sensor due to its excellent sensitivity and selectively for methane gas. The chemiluminescence-based sensor showed excellent analytical indices such as rapid intensity, high sensitivity and outstanding selectivity, and promises a wide application in explosive gas detection and environmental monitoring. Acknowledgments This project is supported by the National Basic Research Program of China (863 Program grant no. 2012AA041504). The authors are also grateful to the editor and the anonymous reviewers for their useful comments and advice, which were vital for improving the quality of this paper.
Conflicts of interest The authors declare no conflicts of interest.
References 1. Zhang YM, Wu YL, Wu XZ, Xi X, Wang JQ. A novel vibration mode testing method for cylindrical resonators based on microphones. Sensors 2015;15:1954–63. 2. Cao XA, Zhang XR. A research on determination of explosive gases utilizing cataluminescence sensor array. Luminescence 2011;20:243–50. 3. Li B, Sauve G, Iovu MC, Zhang R, Cooper J, Santhanam S, et al. Volatile organic compound detection using nanostructured copolymers. Nano Lett 2012;6:1598–602.
Copyright © 2015 John Wiley & Sons, Ltd.
Luminescence 2016; 31: 183–189
Development of highly sensitive sensor system 4. Lin HB, Shih JS. Fullerene C60-cryptand coated surface acoustic wave quartz crystal sensor for organic vapors. Sens Actuators B Chem 2010;92:243–54. 5. Lee DS, Jung HY, Lim JW, et al. Explosive gas recognition system using thick film sensor array and neural network. Sens Actuators B Chem 2008;71:89–98. 6. Hu J, Xu KL, Jia YZ, Lv Y, Li YB, Hou XD. Oxidation of ethyl ether on borate glass: chemiluminescence, mechanism, and development of a sensitive gas sensor. Anal Chem 2009;80:7964–9. 7. Potyrailo RA, Mirsky VM. Combinatorial and high-throughput development of sensing materials: the first 10 years. Chem Rev 2008;108:770–813. 8. Breysse M, Claudel B, Faure L, Guenin M, Roberto J, Williams J. Chemiluminescence during the catalysis of carbon monoxide oxidation on a thoria surface. J Catal 1976;45:137–44. 9. Wang QH, Xie LJ, Zhu B, Zheng Y, Cao SH. Harmful gas recognition exploiting a CTL sensor array. Sensors 2013;13:13509–20. 10. Shi GL, Sun B, Jin Z, Liu JH, Li MQ. Synthesis of SiO2/Fe3O4 nanomaterial and its application as cataluminescence gas sensor material for ether. Sens Actuators B Chem 2012;171–172:699–704. 11. Cao XA, Zhang ZY, Zhang XR. A novel gaseous acetaldehyde sensor utilizing cataluminescence on nanosized BaCO3. Sens Actuators B Chem 2004;99:30–5. 12. Rao ZM, Liu LJ, Xie JY, Zeng YY. Development of a benzene vapour sensor utilizing chemiluminescence on nanosized Y2O3. Luminescence 2008;23:163–8.
13. Zhou KW, Ji XL, Zhang N, Zhang XR. On-line monitoring of formaldehyde in air by cataluminescence-based gas sensor. Sens Actuators B Chem 2006;119:392–7. 14. Zhu YF, Shi JJ, Zhang ZY, Zhang C, Zhang XR. Development of a gas sensor utilizing chemiluminescence on nanosized titanium dioxide. Anal Chem 2002;74:120–4. 15. Wu YY, Zhang SC, Na N, Wang X, Zhang XR. A novel gaseous ester sensor utilizing chemiluminescence on nano-sized SiO2. Sens Actuators B Chem 2007;126:461–6. 16. Kawi S, Tang YP, Hidajat K, Yu LE. Synthesis and characterization of nanoscale CeO2 catalyst for deNOx. J Metastab Nanocryst Mater 2005;23:95–8. 17. Zhao Y, Wei W, Meiping H, Shen P, Minli Z, Wang W. An effective secure routing way to reduce energy consumption in wireless sensor networks. AISS 2012;4:277–83. 18. Llinas J, Hall DL. An introduction to multi-sensor data fusion. In: Proceedings of the 1998 IEEE International Symposium on Circuits and Systems (ISCAS 1998), 31 May–3 June 1998; Monterey, CA, USA 1998:537–40. 19. Mao SY, Zhao W. Comments on multisensor image fusion techniques. J Beijing Univ Aeronaut Astronaut 2002;28:512–8. 20. Zhao Y, Wei W, Tang X.J, Zhao ML. Choice of optimized scheme parameters of neural networks based on error handling. AISS 2012;4:76–82. 21. Zhao Y, Wei W, Song HB, Li W. The hardware design of intelligent circuit breaker. Energ Educ Sci Technol A 2014;33:695–702.
189
Luminescence 2016; 31: 183–189
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/luminescence
