You are on page 1of 6

Sensors and Actuators B 98 (2004) 148153

Flame spray synthesis of tin dioxide nanoparticles for gas sensing


T. Sahm a , L. Mdler b , A. Gurlo a, , N. Barsan a , S.E. Pratsinis b , U. Weimar a
b

Institute of Physical and Theoretical Chemistry, University of Tbingen, Auf der Morgenstelle 8, 72076 Tbingen, Germany Particle Technology Laboratory, Swiss Federal Institute of Technology (ETH) Zurich, Sonneggstrasse 3, CH-8092 Zurich, Switzerland Received 22 July 2003; received in revised form 30 September 2003; accepted 1 October 2003

Abstract Flame spray pyrolysis (FSP) has been used to synthesize tin oxide nanoparticles from the ethylhexanoate precursor in ethanol. The particles were highly crystalline having a primary particle and crystallite size of 17 nm. The single crystalline particles were only slightly aggregated and directly used for thick lm sensor deposition by drop coating. The ame made SnO2 nanoparticles showed high and fast response to both reducing (propanal) and oxidizing (NO2 ) gases. 2003 Elsevier B.V. All rights reserved.
Keywords: Tin dioxide; Gas sensors; Flame spray pyrolysis; Aerosol

1. Introduction Metal oxides in general and SnO2 , in particular, have attracted the attention of many users and scientists interested in gas sensing under atmospheric conditions. SnO2 sensors are the best-understood prototype of oxide-based gas sensors [13]. In recent years, nano- and micrometer SnO2 particles for the gas-sensing applications have been produced by different chemical and physical methods, e.g. by solgel [1,4], gas-phase condensation [5], decomposition of organometallic precursor [6], oxidation of metallic tin [7], hydrothermal treatment of colloidal solutions [8], laser ablation [9,10], and mechano-chemical processing [11,12]. All these methods lead to submicrometer tin dioxide particles with different shapes and morphologies, from spherical to rod-like. Also, recently the gas-sensing properties of SnO2 realized as single-crystalline nanobelts [13], opal-like [14] and ordered mesoporous structures [15,16] have been studied. The importance of size control, the required large and easily accessible surface area (large pore size, no micropores), the desired high crystallinity, the ability of noble metal doping and competitive production rates put high demands on the method of nanoparticle production for sensor materials.
Corresponding author. Tel.: +49-7071-2978765; fax: +49-7071-295960. E-mail address: alexander.gurlo@ipc.uni-tuebingen.de (A. Gurlo). URL: http://www.ipc.uni-tuebingen.de.

Dry aerosol synthesis routes are able to fulll these requirements [17] as tin oxide can be made readily in vapor ames from its chloride [18,19] and from tetramethyl tin gaseous precursor with sizes down to 6 nm [20]. However, the present study is an extension of the work on ame made tin oxides, directly linking sensor performance with the more versatile ame spray method. Furthermore, vapor ame methods have limitations when tin oxide has to be doped with materials where only low vapor pressure precursors are available such as platinum or ceria. Flame spray pyrolysis (FSP) overcomes this drawback by spraying liquid precursors, thus forming a spray ame where the precursor evaporates or decomposes and reacts in the gas phase with the subsequent particle formation. In this way, FSP is a very promising technique for sensor material fabrication since it enables primary particle and crystal size control (e.g. [21]), which is important to tailor sensitivity [5], as well as the controlled in situ deposition of noble metal clusters [22,23]. It has been shown that due to the morphology of the FSP-made particles, the mass transfer rates in catalysis are higher compared to microporous materials because of the large external surface area of ame-made materials [23]. Furthermore, FSP bears the advantage to completely manufacture the nanopowder in a single high-temperature step without affecting the microstructure and noble metal particle size in a subsequent annealing process as it is necessary in conventional spray pyrolysis [24] or wet methods in general. In the present study, the possibility of FSP in synthesis of nanoparticles of tin dioxide for gas sensing is presented. The aim was to demonstrate that FSP and ame process in

0925-4005/$ see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.snb.2003.10.003

T. Sahm et al. / Sensors and Actuators B 98 (2004) 148153

149

general can be successfully applied as a new preparation method of sensing materials.

2. Experimental 2.1. Powder synthesis and characterization Tin(II) 2-ethylhexanoic acid (Aldrich) was diluted in ethanol (J.T. Baker, purity > 99.9%) to obtain a 0.5 M precursor solution. The precursor was fed into a ame spray pyroylsis reactor [22,25] by a syringe pump with a rate of 8 ml/min and was dispersed by 3 l/min of oxygen (Pan Gas, purity > 99%) into ne droplets by a gas-assist nozzle (pressure drop at the nozzle tip: 1.5 bar). The water-cooled system of the reactor avoided any evaporation of the precursor within the liquid feed lines or overheating of the nozzle. The spray ame was maintained by a concentric supporting amelet ring of premixed methane/oxygen (CH4 = 1.5 l/min, O2 =3.2 l/min). In order to assure enough oxidant for complete conversion of the reactants, an additional outer oxygen ow (5 l/min) was supplied. The powder was collected with the aid of a vacuum pump on a glass ber lter (GF/D Whatman, 257 mm in diameter). During the experiment, the lter was placed in a water-cooled holder, 400 mm above the nozzle, keeping the off-gas temperature below 200 C. X-ray diffraction (XRD) patterns were recorded with Bruker AXS D8 Advance (40 kV, 40 mA) and used to obtain the crystallite size (dXRD ) based on the fundamental parameter approach and the Rietveld method [26] with the structural parameters of casserite (ICSD Coll. Code: 084576 [27]). A linear background was used when matching the XRD patterns while microstrain was not considered. The BET powder-specic surface area (SSA), was measured by nitrogen adsorption at 77 K (Micromeritics Gemini 2375) after degassing the sample, at least, for 1 h at 150 C in nitrogen. Assuming within an aggregate, the equivalent

Intensity (scaled), a.u.


20

30

2 , degree

40

50

60

70

Fig. 2. X-ray diffraction pattern of as-prepared tin oxide. The vertical lines represent the diffraction pattern of SnO2 [12]. The average crystal diameter of 17.8 nm was determined by Rietveld analysis.

average primary particle diameter dBET is calculated by dBET = 6/(SSA p ), where p is the density of SnO2 (6.85 g/cm3 ). The product powder was further analyzed by transmission electron microscopy (TEM, Hitachi H600, operated at 100 kV). 2.2. Characterization of the gas-sensing properties by DC electrical tests DC electrical measurements (sensor tests) have been performed to monitor the response to CO, propanal (propionaldehyde, C3 H6 O) and NO2 in dry synthetic air and in synthetic air with 50% relative humidity. The sensing layers were fabricated by drop coating from the suspension on the alumina substrates with interdigitated Pt electrodes on the front side and heater on the back side [2]. After deposition, the sensors were annealed for 10 min in the belt oven at 500 C in air. For the comparison tests,

Fig. 1. TEM micrographs (different magnication) of as-prepared tin oxide made in a ame spray reactor operating at 36 g/h. The particles are polyhedral and show low degree of aggregation.

150

T. Sahm et al. / Sensors and Actuators B 98 (2004) 148153

commercially available SnO2 powder (Aldrich, 325 mesh, d = 330 nm) was also used for the sensor deposition. The measurements were performed with a set of two identical sensors placed symmetrically in a teon-made test chamber and operated in the same conditions. The operating temperature of the sensors was adjusted between 200 and 400 C. The sensor signal is given in the following as the resistance ratio Rgas /Rair for NO2 and as Rair /Rgas for CO and propanal, where Rgas and Rair denote the sensors resistances in the presence and in the absence of NO2 , CO and propanal, respectively. A computer driven gas-mixing system provided the analyte gas. A typical gas mixing bench consists of a combination of computer-controlled mass ow controllers and computer-controlled valves. The sensors were exposed to NO2 (105000 ppb), CO (50010,000 ppm) and propanal (10300 ppm) in dry and humid (50% RH) synthetic air. The humidity was adjusted by bubbling synthetic air through a column of water and subsequently mixing it with dry air in a gas blender. Dened concentration of NO2 , CO and propanal were obtained in the PC-controlled gas-mixing bench by
NO2 Concentration [ppb]

mixing certied NO2 , CO or propanal test gas. Incoming NO2 concentrations were controlled by a ML 9841B NOx Chemiluminescence analyzer [28].

3. Results and discussion 3.1. Properties of ame-made tin oxide A stable red/orange spray ame was formed when igniting the precursor spray having a visible ame height of 190 mm. The tin oxide was produced at a rate of 36 g/h while one lter had a capacity of about 1.2 g. Fig. 1 shows two TEM images of the as-prepared powder at two different magnications. The aggregated powder consists of polyhedral primary particles of about 20 nm. The powder morphology is typical for particles formed in the gas phase [29]. The measured average primary particle size and the average crystallite size are dBET = 17.1 and dXRD = 17.8 nm, respectively. The coincidence of the average primary particle and the average

10000 2503 1000 166 100 27 10 10 58 975 435

4794

2654 1034 540 176 68 34

NO2 [ppb]

15

10

R []
10
7

1200 1320 1440 1560 1680 1800 1920 2040 2160 2280 2400 2520

(A)

Time [min]

Sensor Signal

10

1 0
(B)

1000

2000

3000

4000

5000

6000

NO2 Concentration [ppb]

Fig. 3. Change of the resistance of the ame-pyrolized SnO2 sensors under exposure to NO2 (A) and sensor signal dependence on NO2 concentration (B) in humid (50% RH) air at 220 C. In the upper part of (A) incoming NO2 concentrations measured on-line by NOx analyzer are shown.

T. Sahm et al. / Sensors and Actuators B 98 (2004) 148153

151

crystal sizes give strong evidence that the primary particles are single crystals with a low degree of aggregation. This was observed also for FSP-made bismuth oxide [30]. Fig. 2 shows the XRD pattern of the as-prepared tin oxide powder. The measured structure matches the reference structure [27] quite well and no other crystalline species than SnO2 were detected. The diffraction pattern shows further a homogeneous well-crystalline powder, while the presence of larger particles (above 100 nm) or an amorphous product was not detectable compared to other FSP studies [21,30]. The FSP-derived tin oxide powder is similar to the powder produced in a hydrogen diffusion-type burner reacting tetramethyl tin (TMT) at an estimated production rate of 33 g/h [31]. The particle properties derived from the very reactive gaseous TMT in the diffusion ame are similar concerning the average primary particle diameter of 15 nm and the estimated crystal size of about 15 nm. The authors also report single crystalline SnO2 particles having only a small degree of aggregation. Similar results were obtained [20] in a low pressure at ame burner using the TMT but at a much lower concentrations. The authors showed good control of the primary particle size from about 6 to 15 nm by varying the precursor concentration in the ame. Therefore, the tin oxide particles produced with the ame spray process are similar to the particles synthesized in the vapor ame reactor corroborating the fact that the precursor reaction and particle formation takes place within the gas phase and that all liquid precursor left the droplet environment before the formation of the tin oxide. However, the vapor ame-made tin oxide powders were not tested for their sensor performance and therefore the versatile FSP technique of SnO2 was explored. 3.2. Gas-sensing properties As mentioned above, the aim of the present study was to demonstrate that FSP and ame process in general can be successfully applied as a new preparation method of sensing materials. The main results of the sensor tests can be summarized as follows: exposing the sensor with oxidizing gases (NO2 ) increases its resistance (Fig. 3A) while reducing gases (propanal and CO) decreases the resistance of the sensors (Fig. 4A). This behavior is typical for the tin dioxide as an n-type semiconductor; the sensors show low signals to CO even at high concentrations (>500 ppm, both in humid and dry air) and much higher signals to propanal, that is a typical behavior for the undoped SnO2 [32]; the calibration curves and the corresponding sensor signals can be approximated by power laws (Figs. 3B and 4B); at low operating temperatures (<300 C), the sensor signals to propanal and CO are higher in dry air in comparison with the humid one. This difference disappears at higher operating temperatures (>300 C) (Fig. 5).

10

1 2

10

R []

10

10

150 10 20 2100 2160 100 50 50 20 10 Propanal [ppm] 2220 2280 2340 2400 100

10

(A)

Time [min]

100

Sensor Signal

10 1 2 3 4 0 20 40 60 80 100 120 140 160

(B)

Propanal Concentration [ppm]

Fig. 4. (A) Change of the resistance of the ame-pyrolized SnO2 sensors under exposure to propanal in humid (50% RH) air at 220 C and (B) sensor signal dependence on propanal concentration in dry air (1 and 2) and at 50% relative humidity (3 and 4).

Sensor Signal

1 2 3 4 5 6

100

10 220 240 260 280 300 320 340 360

Temperature [C]
Fig. 5. Sensor signal (to 150 ppm propanal) dependence on operating temperature of the sensors in dry air (1, 2 and 5) and at 50% relative humidity (3, 4 and 6), 14: FSP SnO2 ; 5 and 6: commercially available SnO2 .

152

T. Sahm et al. / Sensors and Actuators B 98 (2004) 148153

One interesting nding is that the sensor signals to propanal of the ame-pyrolized SnO2 are completely different in humid (50% RH) and dry air at low operating temperature (220 C) and namely the signals in dry air are much higher in comparison with the humid air (see Fig. 5). However, this phenomenon is observed only at relatively low operating temperatures (less than 300 C). One possible explanation can be the different reactivity of the surface of the ame-pyrolized SnO2 in dry and humid air at low temperatures, that can be related to the different hydroxylation of the surface of SnO2 . Inuence of surface hydroxyl groups on sensor response and related effects are usually observed at relatively low operating temperatures. For example, as recently shown [2,33,34], surface hydroxyl groups play an important role in the detection of CO especially at low temperatures. The reason why this happens here is to be understood. Compared to state-of-the-art SnO2 thick lm sensors [35] the response to NO2 and propanal of the reported sensors is fast and show the expected power law behavior. Although for the low NO2 concentrations (below 200 ppb) the sensor response is fast and stable, small oscillations are observed for higher NO2 concentrations (more than 400 ppb) (see Fig. 3A). In our opinion, this problem can be related to the way in which the resistance of the sensors were measured, i.e. especially for the high sensor resistance (more than 1 M for high NO2 concentrations). For the reducing gases (CO and propanal) and therefore for the lower sensor resistance (see Fig. 4A) this phenomenon was not observed. The observed small sensor signal when exposing the sensor to CO is expected from undoped tin dioxide. This was underlined by the direct comparison with commercially available powders (Fig. 5). This is probably due to the lack of additives generally employed for increasing the sensitivity to CO by catalytic effects. While a direct comparison of the sensor performance to other SnO2 preparation methods is rather difcult, the obtained results of the FSP made pure tin dioxide were satisfying. Although the variation of the resistance of the sensitive layer upon time (drift) is also one of the important parameters to characterize the sensor performance, these data are not available at this moment.

Acknowledgements We gratefully acknowledge the technical support of Dr. M. Mller (ETH) providing the TEM facilities and stimulating discussions with Dr. A. Rssler (ETH).

References
[1] N. Barsan, U. Weimar, Conduction model of metal oxide gas sensors, J. Electroceram. 7 (2001) 143167. [2] N. Barsan, U. Weimar, Understanding the fundamental principles of metal oxide based gas sensors the example of CO sensing with SnO2 sensors in the presence of humidity, J. Phys. Condens. Matter 15 (2003) R813R839. [3] N. Barsan, M. Schweizer-Berberich, W. Gpel, Fundamental and practical aspects in the design of nanoscaled SnO2 gas sensors: a status report, Fresenius J. Anal. Chem. 365 (1999) 287304. [4] S. Capone, P. Siciliano, N. Barsan, U. Weimar, L. Vasanelli, Analysis of CO and CH4 gas mixtures by using a micromachined sensor array, Sens. Actuators B 78 (2001) 4048. [5] M. Kennedy, F.E. Kruis, H. Fissan, B.R. Mehta, S. Stappert, G. Dumpich, Tailored nanoparticle lms from monosized tin oxide nanocrystals: particle synthesis, lm formation, and size-dependent gas-sensing properties, J. Appl. Phys. 93 (2003) 551560. [6] C. Nayral, E. Viala, P. Fau, F. Senocq, J.C. Jumas, A. Maisonnat, B. Chaudret, Synthesis of tin and tin oxide nanoparticles of low size dispersity for application in gas sensing, Chem. Eur. J. 6 (2000) 40824090. [7] N. Sergent, P. Glin, L. Prier-Camby, H. Praliaud, G. Thomas, Preparation and characterisation of high surface area stannic oxides: structural, Sens. Actuators B 84 (2002) 176188. [8] N. Baik, G. Sakai, N. Miura, N. Yamazoe, Hydrothermally treated sol solution of thin oxide for thin-lm gas sensor, Sens. Actuators B 63 (2000) 7479. [9] K.I. Gnanasekar, B. Rambabu, K.C. Langry, Role of grain boundaries in exceptionally H2 sensitive highly oriented laser ablated thin lms of SnO2 , J. Electrochem. Soc. 149 (2002) H19H27. [10] S. Nicoletti, L. Dori, G. Cardinali, A. Parisini, Gas sensors for air quality monitoring: realisation and characterisation of undoped and noble metal-doped SnO2 . Thin sensing lms deposited by the pulsed laser ablation, Sens. Actuators B 60 (1999) 9096. [11] U. Kersen, M.R. Sundberg, The reactive surface sites and the H2 S sensing potential for the SnO2 produced by a mechanochemical milling, J. Electrochem. Soc. 150 (2003) H129H134. [12] L. Cukrov, L. McCormick, K. Galatsis, W. Wlodarski, Gas sensing properties of nanosized tin oxide synthesised by mechanochemical processing, Sens. Actuators B 77 (2001) 491495. [13] E. Comini, G. Faglia, G. Sberveglieri, Z.W. Pan, Z.L. Wang, Stable and highly sensitive gas sensors based on semiconducting oxide nanobelts, Appl. Phys. Lett. 81 (2002) 18691871. [14] R.W.J. Scott, S.M. Yang, G. Chabanis, N. Coombs, D.E. Williams, G.A. Ozin, Tin dioxide opals and inverted opals: near-ideal microstructures for gas sensors, Adv. Mater. 13 (2001) 1468 1472. [15] T. Hyodo, N. Nishida, Y. Shimizu, M. Egashira, Preparation and gas-sensing properties of thermally stable mesoporous SnO2 , Sens. Actuators B 83 (2002) 209215. [16] R.W.J. Scott, M. Mamak, K. Kwong, N. Coombs, G.A. Ozin, Making sense out of sulfated tin dioxide mesostructures, J. Mater. Chem. 13 (2003) 14061412. [17] W.J. Stark, S.E. Pratsinis, Aerosol ame reactors for manufacture of nanoparticles, Powder Technol. 126 (2002) 103108. [18] W.H. Zhu, S.E. Pratsinis, Synthesis of SiO2 and SnO2 particles in diffusion ame reactors, AIChE J. 43 (1997) 26572664.

4. Conclusions In summary, we demonstrated that FSP can be successfully used for the preparation of nanoparticles of SnO2 for the application in gas sensing. Single crystalline tin oxide particles of about 20 nm were produced using the versatile FSP technique. The fabricated sensors show high sensitivity and fast responses to NO2 and propanal. The direct control of particles sizes with the FSP parameters [21] and the deposition of Pt- and Pd-doped SnO2 nanoparticles [23], which is known to be a good method for promoting the detection of CO [36], are the next steps towards tailor made sensors from gas phase-derived powders.

T. Sahm et al. / Sensors and Actuators B 98 (2004) 148153 [19] S. Vemury, S.E. Pratsinis, L. Kibbey, Electrically controlled ame synthesis of nanophase TiO2 , SiO2 , and SnO2 powders, J. Mater. Res. 12 (1997) 10311042. [20] D. Lindackers, C. Janzen, B. Rellinghaus, E.F. Wassermann, P. Roth, Synthesis of Al2 O3 and SnO2 particles by oxidation of metalorganic precursors in premixed H2 O2 Ar low pressure ames, Nanostruct. Mater. 10 (1998) 12471270. [21] L. Mdler, W.J. Stark, S.E. Pratsinis, Flame-made ceria nanoparticles, J. Mater. Res. 17 (2002) 13561362. [22] L. Mdler, W.J. Stark, S.E. Pratsinis, Simultaneous deposition of Au nanoparticles during ame synthesis of TiO2 and SiO2 , J. Mater. Res. 18 (2003) 115120. [23] R. Strobel, W.J. Stark, L. Mdler, S.E. Pratsinis, A. Baiker, Flamemade platinum/alumina: structural properties and catalytic behaviour in enantioselective hydrogenation, J. Catal. 213 (2003) 296304. [24] M.V. Cabanas, G. Delabouglise, M. Labeau, M. Vallet-Regi, Application of a modied ultrasonic aerosol device to the synthesis of SnO2 and Pt/SnO2 for gas sensors, J. Solid State Chem. 144 (1999) 8690. [25] L. Mdler, H.K. Kammler, R. Mueller, S.E. Pratsinis, Controlled synthesis of nanostructured particles by ame spray pyrolysis, J. Aerosol. Sci. 33 (2002) 369389. [26] R.W. Cheary, A.A. Coelho, Axial divergence in a conventional X-ray powder diffractometer. I. Theoretical foundations, J. Appl. Crystallogr. 31 (1998) 851861. [27] A.A. Bolzan, C. Fong, B.J. Kennedy, C.J. Howard, Structural studies of rutile-type metal dioxides, Acta Crystallogr. Sect. B Struct. Commun. 53 (1997) 373380.

153

[28] A. Gurlo, N. Barsan, U. Weimar, Mechanism of NO2 sensing on SnO2 and In2 O3 thick lm sensors as revealed by simultaneous consumption and resistivity measurements, in: Proceedings of the XIV European Conference on Solid State Transducers Eurosensors XVI, 2002, Prague, Czech Republic, WP18. [29] S.E. Pratsinis, Flame aerosol synthesis of ceramic powders, Prog. Energy Combust. Sci. 24 (1998) 197219. [30] L. Mdler, S.E. Pratsinis, Bismuth oxide nanoparticles by ame spray pyrolysis, J. Am. Ceram. Soc. 85 (2002) 17131718. [31] D.L. Hall, P.V. Torek, C.R. Schrock, T.R. Palmer, M.S. Wooldridge, Gas-phase combustion synthesis of tin oxide nanoparticles, Mater. Sci. Forum 383386 (2002) 347352. [32] K. Ihokura, J. Watson, The Stannic Oxide Gas Sensor, CRS Press, Boca Raton, 1994. [33] S. Emiroglu, N. B rsan, U. Weimar, V. Hoffmann, In situ diffuse a reectance infrared spectroscopy study of CO adsorption on SnO2 , Thin Solid Films 391 (2001) 176185. [34] S. Harbeck, A. Szatvanyi, N. Barsan, U. Weimar, V. Hoffmann, DRIFT studies of thick lm un-doped and Pd-doped SnO2 sensors: temperature changes effect and CO detection mechanism in the presence of water vapor, Thin Solid Films 436 (2003) 7683. [35] http://www.garosensor.com/; http://www.sinc.co.jp/. [36] N. Barsan, J.R. Stetter, M. Findlay, W. Gpel, High performance gas sensing of CO: comparative tests for semiconducting (SnO2 -based) and for amperometric gas sensors, Anal. Chem. 71 (1999) 2512 2517.

You might also like