Analyzing the kinetic response of tin oxide carbon and tin oxide CNT composites gas sensors for alcohols detection Vinayak Kamble and Arun Umarji Citation AIP Advances 5, 037138 (2015); doi 10 1063/1[.]
Analyzing the kinetic response of tin oxide-carbon and tin oxide-CNT composites gas sensors for alcohols detection Vinayak Kamble and Arun Umarji Citation: AIP Advances 5, 037138 (2015); doi: 10.1063/1.4916755 View online: http://dx.doi.org/10.1063/1.4916755 View Table of Contents: http://aip.scitation.org/toc/adv/5/3 Published by the American Institute of Physics AIP ADVANCES 5, 037138 (2015) Analyzing the kinetic response of tin oxide-carbon and tin oxide-CNT composites gas sensors for alcohols detection Vinayak Kamblea and Arun Umarji Materials Research Centre, Indian Institute of Science, Bangalore, 560012, India (Received 28 January 2015; accepted 19 March 2015; published online 30 March 2015) Tin oxide nanoparticles are synthesized using solution combustion technique and tin oxide – carbon composite thick films are fabricated with amorphous carbon as well as carbon nanotubes (CNTs) The x-ray diffraction, Raman spectroscopy and porosity measurements show that the as-synthesized nanoparticles are having rutile phase with average crystallite size ∼7 nm and ∼95 m2/g surface area The difference between morphologies of the carbon doped and CNT doped SnO2 thick films, are characterized using scanning electron microscopy and transmission electron microscopy The adsorption-desorption kinetics and transient response curves are analyzed using Langmuir isotherm curve fittings and modeled using power law of semiconductor gas sensors C 2015 Author(s) All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported License [http://dx.doi.org/10.1063/1.4916755] I INTRODUCTION Metal oxide semiconductors (MOS) are one of the most widely used materials in commercial gas sensor devices The basic principle of chemoresistive gas sensor operation stems on the high sensitivity of resistance to ambient gaseous conditions SnO2 is one such prototype oxide materials which posses’ qualities required for ideal sensor device and is being utilized in many of commercial gas sensing devices.1 SnO2, ZnO and other binary oxides offer great properties including high sensitivity, better thermal stability, however they lack selectivity.1–5 Hence often the sensor performance are improved by doping6 or using nanocomposites of metal oxide with polymer or carbon nanotubes (CNTs).7 The components of nanocomposites synergistically show enhanced sensing characteristics The properties of nanocomposites are largely affected by the type of components as well as their morphology and interracial contacts Carbon and its allotropes, when doped or added in metal oxides, it exhibits very good gas sensing properties.8–12 These MOS based sensors are used in detections of several hazardous gases including CO, NOX, SOX, NH3 and VOCs etc.13 Volatile Organic Compounds (VOCs) are relatively mild and very common atmospheric pollutants yet cause prolonged adverse effects on humans as well as rest of the ecosystem.14,15 VOC sensors find application in food, perfume, medicine and chemical industries.16 An empirical model usually used to explain the fundamental gas sensing mechanism is as follows According to this model, a depletion layer is created at the interfaces of the sensor (oxide) as a result of adsorbed oxygen from air Further reducing gases are adsorbed by trapping some of the charge carriers (conduction electron) The charge transfer between a surface adsorbed gaseous species and the oxide surface atoms leads to change in carrier concentration of the oxide and widening or narrowing of the depletion layer This phenomenon affects the current flow through the oxide matrix and further results in change in resistance of the oxide.5,17,18 However other than temperature and concentration of the test gas, largely the change in resistance is determined by the interaction between the surface adsorbing species and the oxide surface e g the nature of adsorbing gas molecules (oxidizing or reducing), type of functional group present, the nature of a Author to whom correspondence should be addressed; E-Mail: vinbk@mrc.iisc.ernet.in; Tel.: +91-080-2293-2944; Fax: +91-080-2360-7316 2158-3226/2015/5(3)/037138/9 5, 037138-1 © Author(s) 2015 037138-2 V Kamble and A Umarji AIP Advances 5, 037138 (2015) preadsorbed oxygen species I e molecular (O2−) or atomic oxide (O−) ions Here it is important to note that interaction between the MO surface and the test gas involves exchange of charge carriers and hence is highly affected by polarity and electron affinities of the analyte Though in the present study all the test gas contain same functional group (i e OH−), it is well known that in alcohols as the number of carbon atom increases the polarity of the molecule reduces Concurrently, other factors such as the bulkiness of the molecule and hence charges screening also increases Considering above circumstances, it will be justified to analyze the sensing characteristics and time constant studies to probe the variation and to understand the kinetics involved in gas sensing of alcohols In this paper, comparisons of alcohols sensing properties of SnO2 - Carbon / CNT composite films are demonstrated Usually the practical application of alcohol sensors demands a good sensitivity over a wide range from 100 ppm to 1% (which is 10,000 ppm) for several domestic as well as industrial applications like food quality control, breath analyzers, industrial reaction monitoring and environmental pollution control etc In the present study, we have tried to imitate the practical situations by keeping the alcohol liquids at room temperature and depending on the volatility of the compound the vapors are generatred, which are calculated using the vapor pressure data of the respective analyte Further, the gas sensing properties and kinetics of gas response of carbon doped and CNT doped SnO2 are compared using Langmuir adsorption isotherms The response and recovery time constants are calculated and they have been modeled the power law of MOS sensors.19 II EXPERIMENTAL Ultrafine SnO2 nanoparticles are prepared by solution combustion synthesis using tin oxalate (SnC2O4, from Sigma Aldrich) The details of preparation method can be found elsewhere.20,21 Multiwalled carbon nanotubes (CNT) were synthesized by simple pyrolysis method using Benzene and ferrocene at 750 oC.22 The as prepared CNTs were functionalized by treating with conc HNO3 SnO2 nanoparticles (with and without 10% CNT loading ) were mixed with a dispersing agent Triton-X 100 in order to make 10 mg/ml of SnO2 in Triton-X 100 Here Triton-X is used as a surfactant as well as carbon source A uniform dispersion of oxide into triton-X is made The SnO2-C composite films were deposited by drop casting this dispersion on glass substrates followed by annealing in air at 250 oC for hr Whereas CNT loaded SnO2 films were annealed at 450 oC to remove the amorphous carbon residues The SnO2 films with and without CNT loaded are coded as TOCN and TOC respectively The as synthesized SnO2 nanoparticles were characterized by XRD using Bruker D8 advance diffractometer at a scan rate at o/min and step size of 0.02, Cu Kα radiation (1.5418 Ao) Further, the Brunauer–Emmett–Teller (BET) surface area and porosity were evaluated by adsorptiondesorption of N2 gas at 77 K on Belsorp-Aqua Porosimeter The thermal analysis of the SnO2 Triton-X dispersion was carried out using a simultaneous TG-DTA apparatus (TA instruments SDT Q600) in air atmosphere, at 10 oC / heating rate The thick films deposited using drop casting are characterized by raman spectra using Jobin-yvon LabRam HRUVraman system with a 515 nm wavelength laser in the range from 3000 cm−1 to 100 cm−1 The morphology and Energy dispersive spectroscopy analysis of the thick films was carried out using high resolution scanning electron microscope (SEM) (FEI Quanta ) and the transmission electron microspy (TEM) images are recorded using Technai T20 200eV microscope The gas sensing measurements were done on an in-house built gas sensing apparatus capable of accommodating two samples simultaneously The details of the experimental setup can be found elsewhere.23 The alcohol vapors were generated by bubbling air gas through respective liquids and the different concentrations were calculated using vapor pressure data obtained from Antoine’s constants Zero air was used as diluting gas Total gas flow was maintained at 300 sccm The temperature was controlled using a heater placed at the bottom of the sensor films The data were measured and collected using Keithley 2700 multimeter interfaced data acquisition system III RESULTS AND DISCUSSION The thermal analysis of the starting mixture i e Dispersion (SnO2 + CNTs + Triton) was done in order to study the decomposition temperature of surfactant and quantification of carbon residue 037138-3 V Kamble and A Umarji AIP Advances 5, 037138 (2015) FIG The TG-DTA of the dispersion used for depositing thick films of TO-C and TO-CN FIG X-ray diffraction patterns of TO-C and TO-CN films TABLE I Comparison of full width at half maxima (FWHM) of (110) planes and crystallite size calculated using Scherrer equation Sensor Films (codes) Carbon doped SnO2 (TOC) CNT doped SnO2 (TOCN) 2θ (110) FWHM t (nm) 26.617 26.657 1.101 0.761 7.41 10.73 Fig shows the TG-DTA of the dispersion in air atmospheres The TGA of dispersion shows two successive weight losses in the range 200 - 325 oC accompanied by corresponding exothermic peaks in DTA The first weight loss occurs at the ∼260 oC and marks the decomposition of the surfactant I e Triton-X While the second weight loss at 300 – 350 oC with strong exothermic peak corresponds to the burning of carbon residue left after the decomposition, which continues up to 450 oC beyond which, the weight remains nearly constant Fig depicts the XRD patterns of TO-C and TO-CN thick films respectively Since the TO-CN films are annealed in air atmosphere at higher temperature (400 oC), it exhibits high crystallinity i e the intensities are refined as compared to TO-C films besides the peaks became narrower The crystallite sizes calculated from Scherrer formula are nearly and 10 nm for TO-C and TO-CN respectively The corresponding parameters are summarized in Table I 037138-4 V Kamble and A Umarji AIP Advances 5, 037138 (2015) FIG (a) Comparison of Raman spectra of TO-C and TO-CN films showing distinctly different carbon peaks in both the films, and (b) deconvolution of the enlarged view of lower range Raman spectra of TO-CN film showing characteristic SnO2 Raman modes (colored peaks) During combustion synthesis, a large amount of gases are evolved (40 moles of gases per moles of SnO2 formed), hence the product is highly porous with high surface area The BET surface area (as) of the as synthesized SnO2 nanoparticles is estimated to be 95.8 m2/g The average grain size from BET (DBET) is calculated using equation (1) and is estimated to be 9.01 nm, which is in fair accordance with the value obtained by XRD DBET = /ρas (1) Where, ρ is the theoretical density of SnO2 (6.95 g/cm3) Raman spectroscopy is a widely used and very efficient tool in probing the nature of carbon species The comparison of Raman spectra of TO-C and TOCN films, is shown in Fig Both the films show the characteristic raman modes of SnO2 (A2g, B1g and Eg) at 485, 630 and 750 cm-,21 while the D band and G band peaks are clearly visible in raman spectrum of only TO-CN films as can be seen in Fig Nevertheless the D band and G band are absent in the Raman spectrum of TO-C films However, raman spectrum of TO-C films show a peak at 2450 cm−1, which is attributed to the presence of amorphous carbon in the composites The scanning electron micrographs of the TO-C and TO-CN composite thick films are shown in Fig As can be seen in Fig 4(a), the TO-C films show significant agglomeration due to the efficient binding of oxide with the surfactant i e Triton-X On the other hand, TO-CN films show significantly lesser agglomeration and high porosity, due to evaporation of carbonaceous matter leading to formation of pores Nevertheless, at higher magnifications TOCN films show several elongated structures, which are found to consist of carbon along with Sn and O in energy dispersive spectroscopy as shown in Fig 4(f) The films were sputtered with a layer of Au for grounding the charge during SEM image acquisition, hence Au is also seen in EDS spectrum (Fig 4(f)) The morphology was further examined by transmission electron microscopy (TEM) and the corresponding TEM images of the samples are shown in Fig Here, the stronger adhesion between the particles is also evident from TEM images of TOC film shown in Figures 5(a) and 5(b) The HRTEm image shown in Fig 5(c) shows spacing corresponding to (110) and (211) reflections while the Selected area electron diffraction (SAED) pattern (shown inset in Fig 5(a)) is also indexed for rutile SnO2, which confirms the crystalline nature of the sample Further, the nature of morphology of TOCN films seen from Figure 5(d) confirms the presence of dimensional CNTs entangled with SnO2 crystallites The SAED pattern shows reflections pertaining tto Rutile SnO2 The bare CNT with a few SnO2 nanocrystallites is shown in Figure 5(e), which in HRTEM (Fig 5(f)) shows lattice spacing corresponding to rutile SnO2 (enlarged in inset of Fig 5(f)) The gas sensing studies of the composite films were carried out towards various concentrations of different alcohols, and the sensing characteristics are depicted in Fig 6(a through d) A wide range of 037138-5 V Kamble and A Umarji AIP Advances 5, 037138 (2015) FIG Scanning electron micrographs of [(a) and (b)] TOC films and [(c) through (e)] of TO-CN films at low magnification and high magnification respectively and (f) EDS of TO-CN films concentrations of different alcohols is used, ranging from ppm level to sub-percentage level Further the gas sensing of Ethanol is done at two different temperatures The TO-CN films show higher conductivity as compared to that of TO-C films The increased in conductivity in TO-CN film is due to excess conducting channels created as a result of CNT addition The sensitivities of both the films towards all the alcohols are compared and shown in Fig 6(d) The TO-C films show higher sensitivity toward most of alcohols except ethanol and IPA Further the gas sensing properties are analyzed using Langmuir adsorption kinetics Essentially, the basic sensing mechanism of the gas sensing is governed by the type of functional groups present in the test gas Hence, analyzing the adsorption-desorption kinetics and transient responses are of immense help to understand the surface phenomenon The gas sensing behavior is explained using 037138-6 V Kamble and A Umarji AIP Advances 5, 037138 (2015) FIG (a) and (b) TEM image of TOC films showing strong adhesion between the particles (c) HRETEM image of TOC films showing SnO2 crystallites (d) , (e) and (f) The TEM and HRTEM images of TOCN films showing CNTs along with SnO2 crystallites While the inset shows the SAED pattern of the respective films the depletion layer model of the sensors, where the adsorption and desorption (A/D) of the gases causes change in carrier concentration at the interfaces of crystallites This change in carrier concentration is reflected in change in resistance (hence equally in conductance) of the oxide material Here the operating temperature, porosity and size of the crystallite play a crucial role along with the type and concentration of the analyte.18,24 Apart from strong dependency on thermal conditions, the A/D phenomenon is kinetically controlled If a monolayer coverage of the reducing gas is assumed on the MO surface at constant temperature T, assuming Langmuir adsorption kinetics for a single adsorption site, the conductance (G(t)) at any given instant of time t is given by G(t) = G0 + G1(1 − exp(−t/τ)) (2) Where G0 is the base conductance value and τ is the characteristic time constant Hence the individual equation for the response and recovery process of transient can be written as25 G(t) = G0 + G1(1 − exp(−t/τresponse)) G(t) = G′0 + G′1(exp(−t/τrecovery)) (2a) (2b) Where, τresponse and τrecovery are the time constants for response and recovery process respectively Thus, the sensor transients are analyzed using equations (2a) and (2b), and the experimental data and its fit, are shown in 7(a) and 7(b) for two analytes Here, it can be seen in Fig 7, a very good fit is obtained (R2 values ∼0.97) and the values of time constants for response (τresponse) and recovery (τrecovery) are extracted The number of terms and coefficients in equation (2) represents the effective number of energetically different adsorption sites and their individual contribution respectively Here, in this case the equation used for fitting contains only one exponential term, this signifies that there exists only one effective adsorption site in the composite films The values of time constants are not only determined by the energetics of adsorption site, but also depend on the concentration of the test gas being analyzed Hence, the variation of time constant values against concentration is further studied The time constants exhibit an exponential dependence 037138-7 V Kamble and A Umarji AIP Advances 5, 037138 (2015) FIG (a) to (c) The simultaneous gas sensing response curves of TO-C and TO-CN films for various concentration of alcohols obtained using in-house built gas sensing set up and (d) comparison of sensitivities calculated towards various gases FIG The typical gas sensing response obtained and response and recovery curve fittings for (a) 800 ppm of t-Butanol and (b) 6500 ppm of Methanol vapors at 300 oC on the concentration Hence, mathematically expressed as shown in eq (3) τ = τ0C β (3) log τ = log τ0 + β log C (3a) Where, τ is the time constant, C is concentration of gas and β is the power factor The value of β depends upon gas and surface reaction So, log τ Vs logC is essentially a straight line whose slope 037138-8 V Kamble and A Umarji AIP Advances 5, 037138 (2015) FIG The variation of logτ with logC for (a) response and (b) recovery of Ethanol gas at 250 oC and 300 oC and comparison of logτ Vs logC plots for (c) response and (d) recovery of different alcohols at 300 oC TABLE II Comparison of values of β obtained from fitting the straight line in logτ Vs logC plots for various analytes TOC Test gas Ethanol (250 oC) Ethanol (300 oC) IPA Methanol n-Butanol t-Butanol TOCN β Response β Recovery β Response β Recovery 1.12 0.9 0.69 1.26 1.08 1.21 0.41 0.68 0.4 0.58 0.401 0.36 0.535 0.929 1.153 1.381 1.028 1.15 0.539 0.866 1.14 1.007 0.98 0.571 gives the value of β and intercept is τ0 Hence the log τ is plotted against log C It exhibits a straight line nature as shown in Fig The slopes of the fitted straight lines (β) are calculated The values of β obtained for various alcohols and for ethanol at two different temperatures (250 oC and 300 oC) are listed in Table II, which clearly demonstrate the power law behavior of metal oxide semiconductor gas sensors IV CONCLUSION In conclusion, the SnO2-carbon and SnO2-CNT composites based thick film gas sensors are fabricated successfully and its gas sensing response are compared for various members of alcohol family, using Langmuir isotherms, transient responses and time constant analysis The thick film sensors are fabricated using simple chemical methods followed by drop casting technique on glass substrates The results clearly demonstrate the power law behavior of the metal oxide semiconductor composites 037138-9 V Kamble and A Umarji AIP Advances 5, 037138 (2015) based gas sensors and critical temperature dependence of the adsorption- desorption kinetics in metal oxide – carbon composite based gas sensors ACKNOWLEDGMENTS The authors are thankful to the Centre for excellence in Nano Science and Engineering(CeNSE), IISc for Raman spectroscopy and Advanced Facility for Microscopy and Microanalysis (AFMM) for TEM facility Wolfgang Gapel and K D Schierbaum, Sens Actuators, B 26-27, 1-12 (2011) V B Kamble and A M Umarji, Applied physics letters 104(25), - (2014) M.-W Ahn, K.-S Park, J.-H Heo, J.-G Park, D.-W Kim, K J Choi, J.-H Lee, and S.-H Hong, Applied Physics Letters 93(26), 263103 (2008) R G Pavelko, A A Vasiliev, E Llobet, V G Sevastyanov, and N T Kuznetsov, Sens Actuators, B 170(0), 51-59 (2012) G Korotcenkov, Mater Sci Eng., B 139(1), 1-23 (2007) N Yamazoe, Sens Actuators, B 5, 7-19 (1991) C Wongchoosuk, A Wisitsoraat, D Phokharatkul, A Tuantranont, and T Kerdcharoen, Sensors 10(8), 7705-7715 (2010) T Zhang, S Mubeen, N V Myung, and M A Deshusses, Nanotechnology 19(33), 332001-332014 (2008) N D Hoa, N V Quy, and D Kim, Sens Actuators, B 142(1), 253-259 (2009) 10 A Wisitsoraat, A Tuantranont, C Thanachayanont, V Patthanasettakul, and P Singjai, J Electroceram 17(1), 45-49 (2006) 11 N V Carreño, M R Nunes, I T S Garcia, M O Orlandi, H V Fajardo, and E Longo, J Nanopart Res 11(4), 955-963 (2009) 12 J Lanˇ cok, A Santoni, M Penza, S Loreti, I Menicucci, C Minarini, and M Jelinek, Surf Coat Technol 200(1–4), 1057-1060 (2005) 13 L Francioso, A Forleo, A M Taurino, P Siciliano, L Lorenzelli, V Guarnieri, A Adami, and G Agnusdei, Sens Actuators, B 134(2), 660-665 (2008) 14 M L Boeglin, D Wessels, and D Henshel, Environ Res 100(2), 242-254 (2006) 15 M Mori, Y Itagaki, and Y Sadaoka, Sens Actuators, B 163(1), 44-50 (2012) 16 J.-B Sanchez, F Berger, M Fromm, and M.-H Nadal, Sens Actuators, B 119(1), 227-233 (2006) 17 N Barsan, D Koziej, and U Weimar, Sens Actuators, B 121(1), 18-35 (2007) 18 C Wang, L Yin, L Zhang, D Xiang, and R Gao, Sensors 10(3), 2088-2106 (2010) 19 N Yamazoe and K Shimanoe, Sens Actuators, B 128(2), 566-573 (2008) 20 V B Kamble and A M Umarji, AIP Advances 3(8), 082120-082125 (2013) 21 V B Kamble, S V Bhat, and A M Umarji, J Appl Phys 113(24), 244307-244307 (2013) 22 P Mahanandia, P N Vishwakarma, K K Nanda, V Prasad, K Barai, A K Mondal, S Sarangi, G K Dey, and S V Subramanyam, Solid State Commun 145(3), 143-148 (2008) 23 V B Kamble and A M Umarji, Journal of Materials Chemistry C (2013) 24 N Barsan and U Weimar, J Electroceram 7(3), 143-167 (2001) 25 K Mukherjee and S B Majumder, J Appl Phys 106(6), 064912-064910 (2009) ...AIP ADVANCES 5, 037138 (2015) Analyzing the kinetic response of tin oxide- carbon and tin oxide- CNT composites gas sensors for alcohols detection Vinayak Kamblea and Arun Umarji Materials Research... data of the respective analyte Further, the gas sensing properties and kinetics of gas response of carbon doped and CNT doped SnO2 are compared using Langmuir adsorption isotherms The response and. .. FIG The variation of logτ with logC for (a) response and (b) recovery of Ethanol gas at 250 oC and 300 oC and comparison of logτ Vs logC plots for (c) response and (d) recovery of different alcohols