Materials Research Bulletin 47 (2012) 557–561

Enhancement in ethanol sensing response by surface activation of ZnO with SnO2

Onkar Singh, Ravi Chand Singh *

Department of Physics, Guru Nanak Dev University, Amritsar 143005, Punjab, India

A R T I C L E I N F O

Article history:

Received 25 August 2011

Received in revised form 29 October 2011

Accepted 28 December 2011

Available online 5 January 2012

Keywords:

A. Nanostructures

A. Oxides

B. Chemical synthesis

C. Electron microscopy

C. X-ray diffraction

A B S T R A C T

A surface functionalized gas sensing material convincingly giving enhanced response to ethanol is

demonstrated by SnO2 activated ZnO. Zinc oxide was synthesized by a chemical route, deposited on an

alumina substrate and activated by tin dioxide obtained by on-site oxidation of tin chloride. The XRD

study of samples confirmed wurtzite hexagonal structure of zinc oxide and FESEM investigation revealed

that surface of activated ZnO microrods was covered by nanoparticles of tin dioxide. Sensing response of

sensing elements activated with different concentrations of tin chloride solution has been investigated.

It was found that response to ethanol vapor significantly enhanced (eight times) by surface activation

with tin dioxide, which optimized at a concentration of 3 wt.%.

� 2012 Elsevier Ltd. All rights reserved.

Contents lists available at SciVerse ScienceDirect

Materials Research Bulletin

jo u rn al h om ep age: ww w.els evier .c o m/lo c ate /mat res b u

1. Introduction

Metal-oxide gas sensors have attracted attention of manyworkers since their first commercial release in 1968 [1]. Simpleconstruction and vast applications have put them in a class of themost investigated group of gas sensors [2–6]. Resistance orconductance of poly-granular metal oxides change upon exposureto reducing and oxidizing gases [7]. It is imperative to operatemetal-oxide gas sensors at elevated temperatures to activatechemical reaction, improve their selectivity and reduce responsetime [8]. In addition, humidity effects decrease at temperatureswell above 100 8C [8]. Hazardous and toxic gases from vehicles andindustries are polluting the environment. Therefore, for controllingpollution, vehicles and industries should have an alarm systemdetecting and warning for dangerous gas concentration levels.Thus, the need to monitor and control these gases has led toresearch and development of a wide variety of sensors usingdifferent materials and technologies. The n-type materials such asSnO2, ZnO and WO3 are commonly used base materials for gassensing. In these materials oxidizing gases are adsorbed on thesurface of the grains as ions, causing the electrical conductivity ofthe material to decrease. Reducing species are generally sensedbased on their reaction with the pre adsorbed oxygen ions, leadingto an increase in material conductivity as the ions are consumed. Inthe present study, zinc oxide has been utilized as base material andhas many advantages over the other metal oxides because of highmobility of conduction electrons and good chemical and thermal

* Corresponding author. Tel.: +91 991 412 9939.

E-mail address: ravichand.singh@gmail.com (R.C. Singh).

0025-5408/$ – see front matter � 2012 Elsevier Ltd. All rights reserved.

doi:10.1016/j.materresbull.2011.12.049

stability under the operating conditions of sensors. Nanostructureslike rods and particles have become the most promising researchmaterial because of their wide range of applications such astransparent electrodes in solar cells, varistors, piezoelectricdevices, gas sensor, etc. [9]. Different techniques namely sol–gel[10], spray pyrolysis [11], hydrothermal method [12], thermalevaporation [13], etc. are prevalent for the synthesis of zinc oxidenanostructures. In recent years research has been concentrated onimproving the gas sensing by activating zinc oxide with noblemetal catalyst, by incorporating additives into it or by surfaceactivation of thick films [14–18].

In the present investigation, our aim was to develop the sensorby modifying ZnO thick films, which could enhance the sensingresponse. In this work, we synthesized zinc oxide and the surface ofZnO sensor has been activated by SnO2 and its effect on sensorresponse to ethanol vapor has been studied.

2. Experimental details

2.1. Preparation of zinc oxide powder

The chemicals used in the present work were of analytical gradeobtained from Loba Chemie, Mumbai, India. Zinc oxide powder wasprepared by following a chemical route, starting with a 0.2 Msolution of ZnCl2 prepared in distilled water, adding an ammoniumhydroxide drop wise at room temperature with continuous stirringto yield precipitates. The precipitates thus obtained wereseparated from the rest of the liquid by filtering and dried intopowder at 120 8C temperature. The powder thereby obtained wascrushed and calcined at 450 8C for three hours.

O. Singh, R.C. Singh / Materials Research Bulletin 47 (2012) 557–561558

2.2. Sensor preparation and testing method

To fabricate thick film sensors, a paste was prepared by mixing asmall amount of sintered powder with distilled water. The paste wasthen painted onto an alumina substrate (12 mm � 5 mm), having apre-deposited pair of gold electrical contacts 2 mm apart, to obtain athick film of around 25 mm. To obtain sensors of identical geometryalumina substrates were appropriately masked with polymer filmand after painting, extra wet material was removed. No material as abinder was used since, fine particles of zinc oxide were self-bindingvery well. The sensor design is shown in Fig. 1. Samples prepared tobe heated for curing at 350 8C for one hour. To activate the sensors,4 ml of SnCl4 solution of different concentrations (1.0 wt.%, 3.0 wt.%,5.0 wt.%) was added to the surfaces of batch of sensors followed byheating at 350 8C for one hour.

Crystal’s structure of prepared samples was characterized bypowder X-ray diffraction (XRD) using Cu-Ka radiation withShimadzu 7000 systems. Morphologies and grain sizes of thesamples were analyzed by field emission scanning electronmicroscope (FESEM) with JEOL, model-JSM 6700F. To preventcharge build up during the FESEM observations, we coated thesamples with gold by using sputtering system.

Fig. 2. Schematic of sensing unit. (a) Testing chamber and (b) data acquisition

system.

Fig. 1. Schematic of gas sensor.

The measurements of gas sensor response were carried outwith a home-built apparatus consisting of a potentiometerarrangement, a 40 l test chamber in which a sample holder, asmall temperature controlled oven, and a mixing fan wasinstalled. The schematic of sensing unit is shown in Fig. 2. Thefabricated sensor was placed in the test chamber oven at somefixed temperature, and a known quantity of ethanol was injectedusing Hamilton’s syringe into the test chamber. Variation of realtime voltage signal across a resistance connected in series withsensor was recorded with Keithley Data Acquisition ModuleKUSB-3100 connected to a computer. The sensor responsemagnitude was determined as Ra/Rg ratio, where Ra and Rg werethe resistances of sensor in air and air-gas ambience respectively.The response time is defined as the time required for theconductance to reach 90% of the equilibrium value after test gas isinjected. The recovery time is the time necessary for the sensor toattain a conductance 10% above the original value in the air. All thesensors were tested following same procedure at temperatures,250, 300, 350, 400 and 450 8C.

3. Results and discussion

3.1. Structural analysis

Fig. 3 shows the X-ray diffraction pattern of pure and SnO2

activated ZnO. The presence of well resolved and sharp lines wascompared with standard data at the investigation lab, whichconfirmed that the synthesized samples were of ZnO of hexagonalwurtzite phase. The additional peaks (marked with + and *)appearing in the image are of alumina substrate and goldelectrodes. Due to smaller wt.% of SnO2 in the sample there areno prominent peaks of SnO2 associated in the XRD pattern.

Fig. 4 represents the FESEM images of pure and SnO2 activatedZnO. Microrods of pure ZnO of various sizes are visible in Fig. 4(a).The sample activated with 1 wt.% Fig. 4(b) where one finds verysmall grains, most likely of SnO2, sticking on the surface of ZnOmicrorods. The population of SnO2 grains on the ZnO surfacegradually increased as we increased the concentration of tinchloride solution, which is clearly visible in Fig. 4(c and d).Another observation from Fig. 4(d) is that with an increase inconcentration of activating solution, ZnO surface has beencompletely masked by SnO2 grains and SnO2 grains have grownin size as well.

6050403020

+ Alumina* Au

5%

3%

1%

0%

** ++*

(110

)

(102

)

(101

)

(002

)

(100

)

Inte

nsity

(a.u

.)

2θ (deg.)

Fig. 3. XRD patterns of pure ZnO (0 wt.%), 1 wt.%, 3 wt.%, and 5 wt.% SnO2 activated

ZnO.

Fig. 4. FESEM of (a) pure ZnO (0 wt.%), (b) 1 wt.%, (c) 3 wt.%, and (d) 5 wt.% SnO2 activated ZnO.

O. Singh, R.C. Singh / Materials Research Bulletin 47 (2012) 557–561 559

3.2. Gas sensing behavior

The sensing response of fabricated sensors was investigated atvarious temperatures, which optimized at 400 8C (Fig. 5). Gasspecies to be detected requires a certain amount of thermal energyto cross the potential barrier to combine with the adsorbed oxygen.At an optimum operable temperature, a large number of gasmolecules possess required energy, which reacts with adsorbedoxygen resulting in enormous change in conductance of sensingelement. Above 400 8C, the amount of adsorbed oxygen on the

450400350300250

0

5

10

15

20

25

30

35

40

45

50

55

60

0%

5%

3%

1%

Sen

sing

Res

pons

e

Operating Temperature(oC)

Fig. 5. Ethanol sensing response for pure and SnO2 activated ZnO thick film at

different operating temperatures.

sensor surface decreases and as a consequence a lesser amount ofgas species is being consumed. Thus sensing response falls above400 8C. For further investigations, all the sensors were operated atthe optimum operating temperature of 400 8C. To find out theoptimum concentration of tin chloride solution required for thebest sensor response, the sensors activated with various wt.%concentrations of tin chloride solution were investigated. Fig. 6shows sensor response to 250 ppm ethanol at 400 8C versus timefor pure and SnO2 activated ZnO gas sensors. It is clear from Fig. 6that the response of the sensor to ethanol vapor increased

100806040200-505

1015202530354045505560

Sen

sing

Res

pons

e

Time(s)

0% 1% 3% 5%

Fig. 6. Sensing response versus time for pure and SnO2 activated ZnO gas sensor to

250 ppm ethanol at 400 8C.

600550500450400350300250200150100500

0

20

40

60

80

100

120

140

Sen

sing

Res

pons

e

Time (s)

10ppm

50ppm

100ppm

250ppm

500ppm

1000ppm

Fig. 7. Sensing response of optimum activated ZnO gas sensor to different

concentrations of ethanol at 400 8C.

O. Singh, R.C. Singh / Materials Research Bulletin 47 (2012) 557–561560

gradually with the increase in concentration of tin chloridesolution and optimized at 3 wt.%. Further increase of concentrationof tin chloride, sensor response decreased. There is a significantimprovement in the response of sensor with surface activation. Thesensor response increased eight times when activated with 3 wt.%tin chloride solution. Experiment was repeated on a batch of threesamples to affirmatively check the reproducibility of results.Another observation is that response time (�5–10 s) and recoverytime (�15–20 s) of all the sensors in the batch is very short, and ithas not been altered by activation. Fig. 7 shows the variation ofsensing response of activated sensor versus concentration ofethanol, which is linear up to 1000 ppm concentration.

The adsorption–desorption of molecules at the sensor surface isthe principal mechanism for gas detection. It is well known thatwhen an n-type semiconductor particle is exposed to air, oxygenmolecules can adsorb on the surface of the particle and form O2�,O2

2�, O2� ions by capturing electrons from the conductance band,which in turn produces an electron-depleted space-charge layer inthe surface region of the particle. Molecular-type adsorbates (O2,O2�), dissociative type one (O2

2�) and surface (lattice) oxygen(O2

�) are confirmed to exist on the surface of an n-typesemiconductor particle. In addition, all of these adsorbed oxygenspecies are discerned to desorb depending on the adsorptionconditions. Since gas sensors are usually operated at 250 8C andabove, the O2

2� species is more important than another oxygenadsorbates. The target gas (ethanol) may undergo differentreactions, and then can take two routes of decomposition reaction,i.e., dehydration and dehydrogenation:

C2H5OH ! C2H4þ H2OðacidicoxideÞ (1)

2C2H5OH ! 2CH3CHO þ H2ðbasicoxideÞ (2)

These primary products thus formed are consecutively oxidized toCO, CO2 and H2O.

C2H4þ 3O22�ðadÞ ! 2CO2þ 2H2O þ 6e� (3)

2CH3CHOðadÞ þ 5O22�ðadÞ ! 4CO2þ 4H2O þ 10e� (4)

Since ZnO is a basic oxide, dehydrogenation is favored. Thecatalytic oxidation of alcohol over ZnO agrees with the aboveresults because only one intermediate product (CH3CHO) can be

detected. As both the alcohol response and the conversion ratio ofC2H5OH increased with an increase in working temperature, theresponse of the ZnO sensor toward C2H5OH is dependent on theconversion ratio of C2H5OH or formation of CH3CHO. The formationof intermediate product (CH3CHO) increases the response toalcohol. Namely, formation of acetaldehyde plays a key role in thegas sensing process of alcohol.

2C2H5OHðadÞ þ O22�ðadÞ ! 2C2H4O�ðadÞ þ 2H2O (5)

C2H4O�ðadÞ ! CH3CHOðadÞ þ e� (6)

Thus, the gas sensing mechanism of ZnO toward C2H5OH is themode controlled by chemisorption of negatively charged oxygen[19].

ZnO sensor exhibited an optimum response when it wasactivated with 3 wt.% tin chloride solution. In this sample, smallgrains of SnO2 on the ZnO surface are arranged in such a way thatits effective surface area and porosity have become largest amongall the samples. It appears that morphology created by thecombination of microrods, and nanoparticles is the most favorablecondition for gas sensing behavior of a material. As samples areonly surface activated therefore, it is very unlikely that tin has beensubstituted in the ZnO lattice. Tin dioxide, which has a tetragonalstructure, sits on hexagonal zinc oxide and because of differentcrystal geometry, it leads to the generation of surface states atdifferent energy levels. Thus, the number of active sites isenormously increased. As a result, the number of oxygen speciesadsorbed on the activated surface would be more. The larger thenumber of oxygen species adsorbed, the faster would be theoxidation of ethanol. This would increase the conductance of thefilm crucially, enhancing gas response. The sensors were periodi-cally tested to check their long-term stability and results are quitereproducible.

4. Conclusions

Zinc oxide powder can be easily synthesized by a chemicalroute. Activation of ZnO surface with SnO2 enhances its effectivesurface area and porosity. The modified surface of ZnO has shownexceptionally enhanced sensing response to ethanol vapor. Thesurface activation optimized with an addition of 3 wt.% tin chloridesolution on the sensor surface with eight-time improvements insensor response. The optimum operating temperature for all thesamples for ethanol gas was 400 8C. Surface properties of the ZnOcan be modified by surface activation without affecting the bulkproperties. The sensor showed a very quick response (�5–10 s) andrecovery (�15–20 s) to ethanol. The results reveal that tin dioxideis an appropriate catalyst to improve sensing response of ZnO toethanol. The sensors were found to be reproducible as these werechecked with various samples at different times.

Acknowledgments

Authors would like to thank the following: UGC for financialsupport; Prof. (Dr.) A. Ghosh, Indian Association for the Cultivationof Science, Kolkata for FESEM investigations.

References

[1] K. Ihokura, J. Watson, The Stannic Oxide Gas Sensors: Principles and Applications,4th edition, CRC Press, Boca Raton, USA, 1994.

[2] B.K. Miremadi, R.C. Singh, S.R. Morrison, K. Colbow, Appl. Phys. A 63 (1996)271–275.

[3] P. Mitra, A.P. Chatterjee, H.S. Maiti, Mater. Lett. 35 (1998) 33–38.[4] R.C. Singh, O. singh, M.P. Singh, P.S. Chandi, R. Thangaraj, Mater. Res. Bull. 45

(2010) 1162–1164.[5] R.C. Singh, O. Singh, M.P. Singh, P.S. Chandi, Sens. Actuators B 135 (2008) 352–357.

O. Singh, R.C. Singh / Materials Research Bulletin 47 (2012) 557–561 561

[6] C.S. Rout, S.H. Krishna, S.R.C. Vivekchand, A. Govindaraj, C.N.R. Rao, Chem. Phys.Lett. 418 (2006) 586–590.

[7] G. Eranna, B.C. Joshi, D.P. Runthala, R.P. Gupta, Crit. Rev. Solid State Mater. Sci. 29(2004) 111–188.

[8] J.W. Gardner, V.K. Varadan, O.O. Awadel karim, Microsensors, MEMS, and SmartDevices, John Wiley & Sons Ltd, New York, USA, 2001.

[9] H. Nian, S.H. Hahn, K.K. Koo, J.S. Kim, S. Kim, E.W. Shin, E.J. Kim, Mater. Lett. 64(2010) 157–160.

[10] S.M. Rozati, S. Moradi, S. Golshahi, R. Martins, E. Fortunato, Thin Solid Films 518(2009) 1279–1282.

[11] J.Q Xu, Y.P. Chen, Y.D. Li, J.N Shen, J. Mater. Sci. 40 (2005) 2919–2921.[12] T. Gao, T.H. Wang, Appl. Phys. A 80 (2005) 1451–1454.[13] A. Umar, E.K. Suh, Y.B. Hahn, J. Phys. D: Appl. Phys. 40 (2007) 3478–3484.[14] K. Yu, Y. Zhang, L. Luo, W. Wang, Z. Zhu, J. Wang, Y. Cui, H. Ma, W. Lu, Appl. Phys. A

79 (2004) 443–446.[15] P.P. Sahay, R.K. Nath, Sens. Actuators B 133 (2008) 222–227.[16] D.N. Suryawanshi, D.R. Patil, L.A. Patil, Sens. Actuators B 134 (2008) 579.[17] K. Wada, M. Egashira, Sens. Actuators B 53 (1998) 147–154.[18] D.R. Patil, L.A. Patil, Sens. Actuators B 77 (2009) 1409–1414.[19] J.Q. Xu, J.J. han, Y. Zhang, Y.A. Sun, B. Xie, Sens. Actuators B 132 (2008) 334–339.