DSpace at VNU: Improving the ethanol sensing of ZnO nano-particle thin films-The correlation between the grain size and the sensing mechanism

Sn was doped into the ZnO film in order to reduce the average grain size and vary the surface morphology, which was expected to increase the response.. XRD patterns of the undoped and Sn-

Contents lists available atScienceDirect Sensors and Actuators B: Chemical

j o u r n a l h o m e p a g e :w w w e l s e v i e r c o m / l o c a t e / s n b

Improving the ethanol sensing of ZnO nano-particle thin films—The correlation between the grain size and the sensing mechanism

Thanh Thuy Trinha,c, Ngoc Han Tuc, Huy Hoang Lec, Kyung Yul Ryua, Khac Binh Lec,

Krishnakumar Pillaia, Junsin Yia,b,∗

a Information and Communication Device Laboratory, School of Information and Communication Engineering, Sungkyunkwan University, Republic of Korea

b Department of Energy Science, Sungkyunkwan University, Republic of Korea

c University of Sciences, Vietnam National University, Ho Chi Minh City, Vietnam

a r t i c l e i n f o

Article history:

Received 15 January 2010

Received in revised form 3 September 2010

Accepted 22 September 2010

Available online 27 October 2010

Keywords:

ZnO:Sn thin films

Neck-controlled sensitivity

Ethanol sensing

Sol–gel process

a b s t r a c t

ZnO and Sn doped ZnO (ZnO:Sn) thin films at various doping concentrations from 1 to 10 at.% were prepared by the sol–gel method for an ethanol sensing application The Sn doping significantly influenced the film growth, grain size and response of the films The XRD patterns showed that the hexagonal wurtzite structure of the ZnO film was retained even after the Sn doping The crystallite grain sizes of the ZnO:Sn thin films at 0, 2 and 4 at.% were estimated by using the typical Scherrer’s equation The crystalline quality of the films at 6, 8 and 10 at.% of Sn was degenerated Typical FESEM images demonstrated the different morphologies for the ZnO:Sn thin films at various Sn concentrations; many pores of various dimensions were observed depending on the doping level A TEM analysis of the ZnO:Sn thin films at 0,

2 and 4 at.% was performed to verify the grain size The optimum Sn doping level of ZnO:Sn thin film for ethanol sensing was estimated to be 4 at.% The 4 at.% sample obtained the highest response to ethanol vapor in the 10–400 ppm level range at a low operating temperature of 250◦C The sensing mechanism was explained by a variation in the sensitivity model from a neck–grain-boundary controlled sensitivity

to a neck-controlled sensitivity Our work demonstrates the ability to reduce the working temperature

as well as to increase the response of ZnO thin film based gas sensors to detect ethanol, which would be

of great merit for commercialized applications

© 2010 Elsevier B.V All rights reserved

1 Introduction

Semiconductor oxides such as SnO2, Fe2O3, Ga2O3, and Sb2O3

are widely used for the detection of inflammable gases (CH4, C3H8,

H2, etc.) and toxic gases (CO, H2S, etc.) due to their distinct

advan-tages, such as high response time and low cost[1,2] In particular,

zinc oxide is a potential candidate for toxic and combustible gas

sensing applications The gas sensing mechanism of ZnO type gas

sensors involves the chemisorptions of oxygen onto the oxide

sur-face at high temperatures, creating a space charge layer around

the particles, followed by a charge transfer during the interaction

between the chemisorbed oxygen and the target gas molecules,

thus leading to a change in the surface resistance of the sensor

ele-ment[3,4] The lower edge of the conduction band of ZnO is∼4.3 eV

below the vacuum energy, which is higher than the chemical

poten-tial of oxygen (∼5.7 eV below the vacuum energy)[5] So when ZnO

is exposed to air, O2will be adsorbed; it acts as electron acceptors

∗ Corresponding author at: Tel.: +82 31 290 7139; fax: +82 31 290 7159.

E-mail address: yi@yurim.skku.ac.kr (J Yi).

that generate a space charge layer around the crystallite particles

in the material[6] The gas sensing properties of zinc oxides, which are difficult

to explain, depend naturally on their catalytic or surface chemical properties as well as on their physical or morphological properties Because of the surface reaction, this type of gas sensor shows a lack

of selectivity resulting in an unspecific gas detection mechanism and so many types of reducing gases can be detected simultane-ously Furthermore, most of the reactions involved in the detection

of gases rely on the reaction between the adsorbed oxygen and the test gas, so that the working temperature is usually quite high[7] Zinc oxide thin films doped with various materials such as Fe[8],

Cu[9]and Al[10]have been widely explored for the development

of selective gas sensors

It has been shown that a lower operating temperature may be achieved by the doping effect and a significant resistance change can be obtained in doped ZnO rather than in undoped ZnO, which results in a higher sensor response[11] Consequently one of the techniques used to improve the performance of ZnO thin film type sensor is by doping the films with suitable elements which vary the surface morphology of the films The surface morphology 0925-4005/$ – see front matter © 2010 Elsevier B.V All rights reserved.

and therefore the effective specific surface area of the sensor film

depend on the concentration of the dopants That is, dopants

effec-tively vary the contact and adsorption area between the gas sensing

element and the target gas[12]which would change the response

of the film Thus the sensing mechanism of a sensor film will be

altered depending on the concentration of dopants in the film

It has been reported that the surface area as well as the ratio

between the crystallite size (D) and the space charge layer thickness

(L) play an important role in enhancing the gas response

char-acteristics and in the gas selectivity of the SnO2 gas sensor[13]

However, to date, there are hardly any reports regarding the

corre-lation between the grain size of the particles and the sensitivity of

the ZnO thin film In addition, the exact role of the dopants in the

gas sensing process is not well understood, though there are

var-ious reports discussing the role of dopants based on the catalytic

effect and/or the oxygen transfer effect on the surface of the ZnO

particles[14]

Controlling and monitoring the ethanol concentration is vital

in fields such as the testing of the alcohol levels of drivers and

the monitoring of chemical synthesis[7] Both undoped[4,15,16]

and doped ZnO[3,8]have been investigated widely due to their

high sensitivity to ethanol However, the working temperatures of

zinc oxide sensors are quite high, in the range of 400–500◦C[8,15]

The sensitivity of zinc oxide to ethanol vapor can still be improved

upon, and there is need to further explore new ethanol-sensitive

materials as well

Gas sensing films can be deposited by several methods, such as

thermal evaporation, successive ionic layer adsorption and reaction

(SILAR)[17], pulsed laser deposition (PLD)[18], and sol-gel

pro-cess The solution-based sol–gel process offers a simple, low cost

and large area thin film coating method as an alternative to vacuum

deposition techniques[19] Moreover, it has the advantage of

fab-ricating thin films with a small grain size, porous microstructure,

and a large surface area, useful for gas sensing applications

In this study, ZnO and Sn doped ZnO (ZnO:Sn) thin films at

var-ious Sn concentrations were prepared using the sol-gel process for

an ethanol sensing application The structural and morphological

properties of the ZnO thin films as well as the effect of Sn doping

on the ethanol sensing behavior were investigated Sn was doped

into the ZnO film in order to reduce the average grain size and

vary the surface morphology, which was expected to increase the

response When the average grain size was close to the space charge

layer thickness (L), the sensitivity mechanism of ZnO:Sn thin films

changed from a neck–grain-boundary controlled sensitivity to a

neck-controlled sensitivity model This property was found to

sig-nificantly increase the film’s response

2 Experimental

2.1 The sensing films fabrication and characterization

Fig 1 depicts the details regarding the preparation of the

films Zinc acetate dehydrate (Zn(CH3COO)2·2H2O, 99.99% pure)

was dissolved in a mixture of 2-methoxyethanol (99.99% pure)

and monoethanolamine ((MEA) HOCH2CH2NH2, 99% pure)

solu-tion and tin tetrachloride (SnCl4·5H2O, 99% pure) was dissolved in

2-methoxyethanol to prepare two different types of solutions,

solu-tions A and B MEA acted as a solution stabilizer The concentration

of the Zn ions in all of the ZnO:Sn sols was controlled to 0.75 M; the

Sn/Zn ratio was varied from 0 to 10 at.% in steps of 2 at.% Both the A

and B solutions were stirred for 2 h and then aged at 60◦C for 44 h

until a transparent and homogenous sol was obtained Prior to

dip-ping, the glass substrates were cleaned ultrasonically by a freshly

prepared dilute hydrochloric acid, detergent solution, a sodium

hydroxide solution, acetone, and distilled water Finally, the

sub-strates were dried in a flow of N2gas The ZnO:Sn gel films were coated onto the glass substrates (25 mm× 25 mm × 1 mm) using the dip coating method at a speed of 15 cm/min These as-coated films were pre-heated at 250◦C for 20 min immediately after coat-ing After repeating the coating procedure 5 times the films were annealed in air at 500◦C for 2 h The thickness of the films was estimated to be around 100–150 nm using the alpha-step method The surface morphology of the films was studied using a Hitachi S-4800 model ultra high resolution field emission scanning elec-tron microscope (FESEM) X-ray diffraction (XRD) patterns were obtained by employing a Siemens Kristalloflex using CuK␣ radi-ations ( = 1.54059 ˚A) The measuring range was 20–60◦, at room

temperature The crystalline grain size of the ZnO and the ZnO:Sn thin films at 2 and 4 at.% was measured by transmission electron microscopy (TEM) (Model JEM-3101-JEOL) In order to prepare a specimen for ZnO (Fig 4(a)), the film deposited on the glass sub-strates was scraped and dispersed in a solvent followed by placing

a drop of the solution on a carbon coated copper grid The speci-mens for the TEM observation of the ZnO:Sn thin films with 2 and

4 at.% Sn concentration (Fig 4(d) and (e)) were prepared by a car-bon and metal coating method A thin layer of platinum (∼10 nm) above which a thick layer of carbon was coated by the ion beam method (Model Gatan 682 PECS) on top of the deposited ZnO:Sn film in order to provide conductivity and beam damage protection 2.2 The sensing measurements

The characteristics of the sensors were studied using a home-made heated gas flow chamber (6.7 dm3 in volume) The sensor film with the CrNi electrode was placed on a heater which was kept inside the gas flow chamber; the temperature of the heater was controlled from room temperature to 500◦C by a heat con-troller The current to the heater was controlled with a variable voltage transformer The temperature of the sensor was detected using a copper-constantan thermocouple The temperature inside the chamber was always maintained at over 80◦C, which is over the boiling point of the ethanol solution The response S, of the film was defined as Ra/Rgwhere Raand Rgare the electrical resistances

in the air and in the ethanol–air mixed gas, respectively[3,13] The ethanol was injected into the hot-chamber by a micro pipette in the range of 0.2–2␮l The ethanol is then converted to its vapor phase

in approximately 2 min The sensing characteristics were studied

at a temperature range of 200–300◦C The volume of the ethanol injected into the chamber was estimated by the following equation [20,21]:

C (ppm) =ı × VM × Pr× R × T

where ı is the ethanol density, Vris the volume of the ethanol injected, R is the universal gas constant, T is the absolute tem-perature, M is the molecular weight, Pbis the pressure after the ethanol vaporization inside the chamber, and Vbis the volume of the chamber

3 Results and discussions

3.1 The structural characteristics Fig 2depicts the XRD patterns of the ZnO and ZnO:Sn films

at various Sn concentrations As shown in the figure, the undoped and low percentage Sn doped films have (0 0 2) as the preferred orientation This (0 0 2) preferred orientation is due to the minimal surface energy in which the hexagonal structure, c-plane to the ZnO crystallites, corresponds to the densest packed plane No phases corresponding to tin or the related tin compounds were detected

in the XRD pattern due to the low doping concentration, implying

Magnetic stirring,

1 hr

SnCl4.5H2O 2MEsolvent

SolutionA (Snsource)

Additive MEA

2MEsolvent

Solution 2ME+MEA

Magnetic stirring, 30 minutes

Znacetate

stirring,

SolutionB (ZnOsource)

SolutionC

Magnetic stirring, 2 hrs,

Transparent sol

Wetfilms

Dip coated

5 times

XRD, FESEM, TEM, ethanol sensing test

Fig 1 Experimental process used to fabricate the sensing films.

that the Sn dopant did not alter the typical ZnO hexagonal wurtzite

structure The figure shows that the (0 0 2) peak was slightly shifted

to higher diffraction angles, especially for the 6 at.% Sn doped

sam-ple when compared to the undoped one For the undoped samsam-ple,

the (0 0 2) peak was at 2 around 33.7◦whereas for the 6 at.% doped

sample the peak was at around 34.1◦ This is in agreement with

other report[22] When the ZnO film was doped with Sn, Sn4+

sub-stituted into the Zn2+site in the crystal structure The difference in

the ion radius between Sn4+(0.069 nm)[23]and Zn2+(0.074 nm)

[24]might have resulted in a small lattice distortion and so

there-fore reduced the XRD Bragg peak intensity as well as the grain size

[22]

Additional peaks corresponding to the (1 0 1) and (1 0 0) planes

of the ZnO were also observed inFig 2, but with low relative

intensities As the Sn concentration increased, the intensity of the

Bragg peaks decreased and the full width at half maximum (FWHM)

increased The average crystallite size D, of the films at various Sn

Fig 2 XRD patterns of the undoped and Sn-doped ZnO thin films at various doping

doping was estimated by Scherrer’s formula using the equation:

where and ˇ are the wavelength of CuK␣ radiation and the FWHM

of the strongest peak, respectively.Table 1presents the average grain sizes of the films which show that the grain size decreases with an increase in the Sn concentration The grain sizes of the ZnO:Sn thin films with 0, 2 and 4 at.% Sn concentration were esti-mated to be∼23, 16.2 and 11.5 nm, respectively As the grain size

in the film decreases, the total surface area of the grains in the film

is expected to increase, which could enhance the response of the film when used as a sensor The crystalline quality of the films with

6, 8 and 10 at.% of Sn doping is degenerated The disappearance of the (0 0 2), (1 0 0) and (1 0 1) peaks in the samples with the 8 at.% and 10 at.% Sn doping imply a deterioration of the crystalline qual-ity with an increase in the Sn doping level The degeneration of the crystallinity with an increase of the Sn concentration was also observed earlier[22]

3.2 The surface morphology Fig 3depicts the representative FESEM images of the undoped ZnO thin films and the ZnO:Sn thin films at 2, 4, 6 and 8 at.% Sn dop-ing levels As demonstrated in the figure, variations in the surface morphology of the ZnO:Sn films with an increase in the Sn dop-ing concentration were observed The average particle size on the surface of the 4 at.% film decreased significantly when compared to the undoped ZnO and 2 at.% ZnO:Sn thin film Further Sn doping to

Table 1

The average grain sizes of the ZnO:Sn thin films with 0, 2 and 4 at.% Sn concentrations.

Fig 3 Representative FESEM images of (a) undoped ZnO, (b) 2 at.%, (c) 4 at.%, (d) 6 at.%, and (e) 8 at.%, Sn doped ZnO thin films.

6 at.% appears to slightly increase the overall grain size, as shown in

Fig 3(d) Many pores of various dimensions can be observed in the

films The figure shows that the number of pores increases when

the Sn concentration increases from 0 to 4 at.% (Fig 3(a)–(c)) The

increase of the Sn concentration to 6 and 8 at.% (Fig 3(d) and (e))

reduced the number of pores; and crystallite grains in the films

seem to be agglomerated The pores on the surface of the films are

likely to improve the surface area, in the sense that the real

spe-cific surface that the sensing gas can enter into and contact with

effectively increases This could play a significant role in the

sur-face reactions resulting in an increase in the sensitivity of the films

Thus, from the above observations, ZnO:Sn thin films with a 4 at.%

Sn concentration are expected to show the highest response when

compared to the other films

The HRTEM method was used to reaffirm the grain sizes in the

ZnO:Sn thin films The results of the undoped, 2 and 4 at.% Sn doped

films are shown inFig 4.Fig 4(a) shows that the crystallite size of

the ZnO film is in the∼20–25 nm range (marked by white lines)

Fig 4(b) and (c) shows that the crystallite sizes of the ZnO:Sn thin

films with the 2 and 4 at.% Sn concentrations are∼10 to 20 nm and

∼8 to 10 nm, respectively These size variations of the crystallites

to the Sn concentrations are almost in accordance with the sizes

calculated by Scherrer’s equation using XRD results.Fig 4(d) and (e)

shows cross-sectional views of the TEM images of the ZnO:Sn thin films with the 2 and 4 at.% Sn concentrations on the glass substrates The white background indicates the pores in the film The figures demonstrate that the film with the 4 at.% Sn concentration has more

of pores when compared to the 2 at.% Sn concentration film This is

in accordance with our earlier FESEM observations

In accordance to the structure and morphology observations above, the sample with the 4 at.% Sn doping level was expected

to have the highest response to ethanol vapor compared to the other samples Section 3.3will present and discuss the sensing characteristics of the undoped and the Sn doped ZnO films 3.3 The sensing characteristics

3.3.1 The influence of doping concentration on the ethanol sensitivity of films

Fig 5shows the response of the films with the different dop-ing concentrations as a function of the ethanol concentration at a working temperature of 300◦C As demonstrated in the figure, the undoped sample showed a poor response with a value of around 7.7 In contrast, the response of the ZnO:Sn films increased with

an increase in the Sn doping concentration up to 4 at.% Above the

4 at.% doping concentration, the response of the films tended to

Fig 4 Representative TEM images of (a) undoped ZnO, (b) 2 at.%, and (c) 4 at.%, Sn doped ZnO thin films.

400 300

200 100

0

0

10

20

30

40

50

60

(R a

Gas concentration (ppm)

Undoped

1at%

2at%

3at%

4at%

5at%

6at%

Fig 5 Response of the undoped and Sn doped ZnO thin films at various doping

levels as a function of the ethanol concentration at 300 ◦ C.

decrease, as we can see for the 5 at.% and 6 at.% samples from the

figure The response of the 8 at.% and 10 at.% doping samples (not

shown here) decrease further, when compared to the 6 at.%

sam-ple The highest response, at over 50 (S = Ra/Rg), was obtained from

the 4 at.% doping sample in ethanol vapor concentrations ranging

from 100 to 400 ppm Repeated experiments were also performed

in order to determine the reliability of the films

The sensing mechanism of the pure and the ZnO:Sn sensor films

can be explained as follows: at an elevated temperature, the

reac-tive oxygen species, such as O2 −, O2−and O−, are adsorbed on the

ZnO:Sn film surface and the concentration of these oxygen species

is changed by the chemisorptions due to surface reactions[4] The

doping of the ZnO by Sn creates electronic defects in the same way

that Al doped ZnO does[25,26], and also changes the surface

mor-phology of the films (as noted from the FESEM image inFig 3),

which causes the variations in the adsorbed oxygen This develops a

potential barrier which enhances the resistance of the material[4]

When exposed to ethanol vapor, the chemisorbed oxygen will react

with the ethanol vapor due to the sensing reaction and re-inject

the free carriers, thereby reducing the resistance of the ZnO and

the ZnO:Sn The observed variations in the response of the ZnO:Sn

films at various Sn doping concentrations can be attributed to the

variations in the electronic defects created due to the Sn doping,

the surface morphology, and to the variations in the adsorbed

oxy-gen quantity As shown in the FESEM images fromFig 3, there are

many pores in the films that allow the gas to quickly diffuse into

the films from the outside This means that the oxygen as well as

the ethanol vapors can diffuse into the film and come into contact

with the inner surface thereby increasing the effective surface area

during sensing[12] It should be noted fromFig 3that the number

of pores increases when the Sn concentration increases from 0 to

4 at.% (supported by the TEM images inFig 4), so that the response

also increases Above the 4 at.% Sn concentration, the response of

the film decreases due to the reduced number of pores in the film

Furthermore, it should be noted that the average crystalline size of

the ZnO:Sn films estimated from the XRD pattern decreased when the Sn concentration in the film increased from 0 to 4 at.% The ZnO:Sn film with the 4 at.% Sn concentration showed the lowest crystalline size of∼11.5 nm As the grain size in the film decreases, the total surface area of the film is expected to increase, which enhances the response of the film In fact, oxygen adsorption plays

an important role in the electrical properties of the Sn-doped ZnO nano material with multi microstructures and depends strongly on temperature At low temperatures, O2 − is chemisorbed, while at

high temperatures O2 −and O−are chemisorbed, and the O2 −

dis-appears rapidly The complete process of the oxygen adsorption can

be described by the following equations[12]:

O2(adsorbed)+ e−↔ O2(adsorbed) − (4)

O2(adsorbed)−+ e−↔ 2O(adsorbed) − (5)

When exposed to a reduction gas, such as ethanol vapor, the reac-tion between the ethanol vapor and the oxygen that is adsorbed onto the surface of the film can be expressed by:

CH3CH2OH(adsorbed)+ 6O(adsorbed) −→ 2CO2+ 3H2O+ 6e− (7)

Due to the reaction, a number of free electrons are re-injected into the film, so that the resistance of the films decreases as the ethanol gas flows into the test chamber and is subsequently adsorbed onto the surface of the ZnO:Sn thin film

3.3.2 The effect of working temperature on the ethanol sensitivity of films

In order to estimate the efficacy of the films in reduced work-ing temperature for senswork-ing ethanol, experiments were performed using the ZnO:Sn thin films at various doping concentrations rang-ing from 0 to 5 at.% at the temperatures of 200, 250 and 300◦C The results are shown inFig 6

Fig 6(a) presents the response of the films as a function of the various doping concentrations for 300 ppm of ethanol at different temperatures The figure shows that at 200◦C the response of the ZnO:Sn film improved dramatically when compared to the pure ZnO film The response of the undoped film was around 3, whereas for the 1, 2, 3, 4 and 5 at.% ZnO:Sn films, the response was around

30, 64, 72, 55 and 42, respectively When the working temperature was increased to 250◦C, all the doped-films exhibited an extremely high response A higher response of around 150 was observed for the 4 at.% sample Above 4 at.% doping value, the response of the film tended to decrease, however this value was still higher than the undoped films The experiments were repeated to verify their reproducibility; the trend was observed to be same

We also estimated the film’s ability to detect ethanol vapor at

a low concentration of 50 ppm The results are shown inFig 6(b) The figure demonstrates that the 4 at.% sample could detect the ethanol vapor even at the very low concentration of 50 ppm with a high response when compared to all the other doped samples This result suggests that ZnO:Sn thin films with an appropriate doping concentration will be able to detect ethanol at low concentrations FromFig 6(a) and (b), it is obvious that the working temperature plays a crucial role in determining the response of the ZnO:Sn film Typically, the optimum working temperature of a sensor depends

on the target gas, specifically the dissociation mechanism and the reactions between the gas and the oxygen chemisorption on the sensor surface [27] Furthermore, the chemisorptions of atmo-spheric oxygen depend on the sensor surface morphology In our case, the optimum operating temperature was found to be∼250◦C

for the 4 at.% sample As described earlier, the sensing mechanism

is controlled by surface reactions, due to which a potential barrier

5 4 3 2 1 0

0

20

40

60

80

100

120

140

160

Doping concentration (at%)

200 o C

250 o C

300 o C (a)

5 4 3 2 1 0

0

10

20

30

40

50

Doping concentration (at%)

(b)

Fig 6 Block diagram representing the response of undoped and Sn doped ZnO films

at different operating temperatures as a function of the doping concentration: (a)

300 ppm ethanol concentration and (b) 50 ppm ethanol concentration.

is created to further charge transfer When doping with Sn, due to

the low concentration, Sn may substitute the Zn site in the lattice

of the ZnO, forming more trap states on the surface of the films

These surface trap states make the oxygen chemisorption process

occur more easily, even at low temperatures (lower than 250◦C),

and enhances the O−concentration on the surface On a further

increase of the working temperature over 250◦C, the adsorbed

oxy-gen species available on the film surface may not be enough to react

with the ethanol vapors, which in turn may decrease the response of

the film More studies are needed to arrive at a specific conclusion

The response and the recovery times are vital parameters in the

design of sensors for desired applications The response time is

usu-ally defined as the time taken to achieve 90% of the final change in

the current value following a step change in the gas concentration

at the sensor The recovery time can be defined as the time needed

1500 1000

500 0

10 0

10 1

10 2

C 2 H 5 OH out

C 2 H 5 OH in

C 2 H 5 OH out

C 2 H 5 OH in

C 2 H 5 OH out

4at% Sn doped Undoped

Time (sec)

C 2 H 5 OH in

Fig 7 Response and recovery time of the undoped and the 4 at.% Sn doped ZnO thin

films at 250 ◦ C.

to return to 90% of the initial current value after recovering to a dry air flow as opposed to the operating temperature[28].Fig 7depicts the response and recovery time of the undoped and the 4 at.% Sn doped samples at 250◦C for 300 ppm of ethanol As shown in the figure, a high response (∼150), a short response time (∼40 s) and recovery time (∼60 s) to 300 ppm ethanol were observed for the

4 at.% sample at 250◦C Repeated experiments showed the same trend The short response and recovery times will be a merit for the ZnO:Sn thin films for use in sensor applications

Overall, our studies illustrate that the ZnO:Sn thin films with the 4 at.% doping have a high response to ethanol vapor This high response characteristic can be explained by the surface reaction mechanism of the crystallite grains or nano particles associated with the thin films As explained earlier, at elevated temperatures, the oxygen adsorption from the gas phase results in the forma-tion of an acceptor surface state in n-type semiconductors[29] The oxygen vacancies govern the position of the Fermi level in the ZnO films A near-surface portion of the vacancies can capture oxygen from the surrounding atmosphere As a result, the concentration of free vacancies in the subsurface layer and, hence, the concentration

of free charge carriers tend to decrease Thus, under oxygen adsorp-tion, the crystal surface of n-type semiconductors has an electron depleted layer in which the concentration of electrons is lower than that in the bulk[30] The size of this layer, which is created due to the oxygen chemisorptions onto the film surface, strongly depends

on the ratio of the crystallite size (D) to the space charge layer thick-ness (L) on the crystallite grains of the film[13] The space charge layer thickness can be calculated by:

L =εK

BT

q2No

1 /2

(8) whereε is the static dielectric constant, KBis Boltzmann’s constant,

T is the absolute temperature, q is the electrical charge of the carrier, and No is the carrier concentration The estimated value of L for ZnO was around 7.5 nm, obtained by substituting the values of the physical parameters as T = 573 K,ε = 7.9 × 8.85 × 10−12F m−1, and

No= 4.0× 1017cm−3[31]

It is apparent from the above discussions that as the quantity of adsorbed oxygen increases, the region where the movement of the

electrons is disturbed increases When ethanol is introduced, the

adsorbed oxygen ions are removed (by Eq.(7)) and therefore the

potential barrier decreases so that the mobility of the charge

car-riers increases Typically, a semiconductor sensor film like ZnO:Sn

consists of crystallites which are connected to each other by necks,

forming aggregates of large particles which in turn are connected

to the neighboring particles by the grain boundary contacts The

response of an oxide semiconductor such as ZnO:Sn thin film is

independent of the crystallite size if the crystallite size is

signif-icantly large compared to the space charge layer (D 2L) At this

condition, the resistance of the sensor film is mostly determined by

the resistance offered by the potential barrier at the grain boundary

contacts, which is independent of the crystallite size If the

crystal-lite size approaches the space charge layer thickness (D≥ 2L), the

space charge layer penetrates deeper into each of the crystallites

and forms channels at each neck within a particle or crystallite

grain[13,32] At this stage the conduction electrons must move

through these channels and so experience an extra potential barrier

in addition to that found at the grain boundaries Due to the large

number of crystallite grains and hence necks in the film; the

resis-tance of the sensor film is determined predominantly by the neck

resistance Since the neck size is observed to be proportional to the

crystallite size, the response of the gas sensor is dependent on the

crystallite size[13,32] Thus, the actual crystal size (D) relative to

the space charge depth is one of the most important factors

affect-ing the sensaffect-ing properties of the ZnO-like semiconductor oxide gas

sensor

Compared to the L value (7.5 nm) of ZnO, typically, a high

response can be expected for the ZnO:Sn based gas sensor if

the crystallite size is below ∼15 nm In our case, the highest

response was obtained for the ZnO:Sn film with the 4 at.% doping,

which showed a surface morphology having small crystallite grains

(∼11.5 nm) and many pores, as shown inFig 3 This makes the

sensitivity model vary from the neck–grain-boundary controlled

sensitivity to a neck-controlled sensitivity In addition, the 4 at.%

sample exhibited the best surface area due to the association of

the small grain size and the many surface pores The number of

pores increases when the Sn concentration increases to 4 at.% At

the 6 and 8 at.% Sn doping, the number of pores reduced,

per-haps due to the agglomeration of the crystallite grains (Fig 3)

As discussed earlier, the gas sensing mechanism of undoped and

ZnO:Sn thin film is based on the surface reactions Therefore, the

samples with higher surface area are expected to show a better

response

4 Conclusions

ZnO and ZnO:Sn thin films were prepared using a simple sol–gel

method The XRD patterns of the as-prepared samples showed that

the hexagonal wurtzite structure of the ZnO thin film was retained

even after Sn doping No traces of tin or related tin compounds were

detected A TEM analysis of the ZnO:Sn thin films at 0, 2 and 4 at.%

was performed to verify the grain size The FESEM images showed

that the ZnO:Sn thin film consists of nano particles associated with

small grains and many pores The ethanol sensing mechanisms of

the ZnO:Sn thin films at various doping levels were analyzed using

a custom built home-made device It was found that the proper

Sn-doping of the ZnO film greatly improved the response of the gas

sensor to ethanol The best response (∼150), the shortest response

time (∼40 s) and recovery time (∼60 s) to 300 ppm of ethanol was

observed for the sample with the 4 at.% Sn concentration at a

tem-perature of 250◦C Our work demonstrates the ability to reduce the

working temperature and to increase the response of ZnO:Sn thin

film based gas sensors, which would have great merit in

commer-cialized applications

Acknowledgments

This work is financially supported by KRCF (Korea Research Council of Fundamental Science & Technology) and SKKU (Sungkyunkwan University) for National Agenda Project program

References

[1] Y Liu, E Koep, M.L Liu, A highly sensitive and fast-responding SnO 2 sensor fabricated by combustion chemical vapor deposition, Chem Mater 17 (2005) 3997–4000.

[2] C Bose, P Thangadurai, S Ramasamy, Grain size dependent electrical studies

on nanocrystalline SnO 2 , Mater Chem Phys 95 (2006) 72–78.

[3] F Paraguay, M Yoshida, J Morales, J Solis, W Estrada, Influence of Al, In, Cu, Fe and Sn dopants on the response of thin film ZnO gas sensor to ethanol vapour, Thin Solid films 373 (2000) 137–140.

[4] P Feng, Q Wan, T.H Wang, Contact controlled sensing properties of flower like ZnO nanostructures, Appl Phys Lett 87 (2005) 213111.

[5] M Gratzel, Photoelectrochemical cells, Science 414 (2001) 338– 344.

[6] Q.H Li, Y.X Liang, Q Wan, T.H Wang, Oxygen sensing characteristics of individual ZnO nanowire transistors, Appl Phys Lett 85 (2004) 6389– 6391.

[7] R.D Delgado, Tin oxide gas sensors: an electrochemical approach, PhD Thesis, University of Barcelona, 2002.

[8] H.H Hsu, H Paul Wang, C.Y Chen, C.J.G Jou, Y.-L Wei, Chemical structure of zinc in the Fe/ZnO thin films during sensing of ethanol, J Electron Spectrosc Relat Phenom 156–158 (2007) 344–346.

[9] H Gong, J.Q Hu, J.H Wang, C.H Ong, F.R Zhu, Nano-crystalline Cu-doped ZnO thin film gas sensor for CO, Sens Actuators B 115 (2006) 247– 251.

[10] S.C Navale, V Ravi, I.S Mulla, S.W Gosavi, S.K Kulkarni, Low temperature synthesis and NO x sensing properties of nanostructured Al-doped ZnO, Sens Actuators B 126 (2007) 382–386.

[11] H Nanto, T Minami, S Takata, Zinc oxide thin film ammonia gas sensor with high sensitivity and excellent selectivity, J Appl Phys 60 (1986) 482– 484.

[12] Z Yang, Y Huang, G Chen, Z Guo, S Cheng, S Huang, Ethanol gas sensor based on Al-doped ZnO nanomaterial with many gas diffusing channels, Sens Actuators B 140 (2009) 549–556.

[13] S Seal, S Shukla, Nanocrystalline SnO 2 gas sensors in view of surface reactions and modifications, JOM J Min Met Mater Soc 54 (2002) 35–38.

[14] N Han, L Chai, Q Wang, Y Tian, P deng, Y Chen, Evaluating the doping effect

of Fe, Ti and Sn on gas sensing property of ZnO, Sens Actuators B 147 (2010) 525–530.

[15] J Xu, Q Pan, Y Shun, Z Tian, Grain size control and gas sensing properties of ZnO gas sensor, Sens Actuators B 66 (2000) 277–279.

[16] J Xu, J Han, Y Zhang, Y Sun, B Xie, Studies on alcohol sensing mechanism of ZnO based gas sensors, Sens Actuators B 132 (2008) 334–339.

[17] R.R Salunkhe, C.D Lokhande, Effect of film thickness on liquefied petroleum gas (LPG) sensing properties of SILAR deposited CdO thin films, Sens Actuators

B 129 (2008) 345–351.

[18] C Ristoscu, D Caiteanu, G Prodan, G Socol, S Grigorescu, E Axente, N Stefan,

V Ciupina, G Aldica, I.N Mihailescu, Structural and optical characterization of undoped, doped, and clustered ZnO thin films obtained by PLD for gas sensing applications, Appl Surf Sci 253 (2007) 6499–6503.

[19] Z.A.E.P Vroon, C.I.M.A Spee, Sol–gel coatings on large area glass sheets for electrochromic devices, J Non-Cryst Solids 218 (1997) 189–195.

[20] P Sengupta, S Panigrahi, P Mohapatra, J Amamcharla, Y Chang (2007), ASABE Paper No RRV-07140, St Joseph, MI ASABE/CSBE Intersectional, Conference, Fargo, ND.

[21] T Nakamoto, M Yosihioka, Y Tanaka, K Kobayashi, T Moriizumi, S Ueyama, W.S Yerazunis, Colorimetric method for odor discrimination using dye coated plate and multi LED sensor, Sens Actuators B 116 (2006) 202–206.

[22] C.Y Tsay, H.C Cheng, Y.T Tung, W.H Tuan, C.K Lin, Effect of Sn-doped on microstructural and optical properties of ZnO thin films deposited by sol–gel method, Thin Solid Films 517 (2008) 1032–1036.

[23] G.C Matei, L.R.C Van, J Schoonman, M Lumbreras, Preparation and char-acterization of SnO 2 and Cu-doped SnO 2 thin films using electrostatic spray deposition (ESD), J Eur Ceram Soc 27 (2007) 207–213.

[24] C Xie, L Xiao, M Hu, Z Bai, X Xia, D Zeng, Fabrication and formaldehyde gas-sensing property of ZnO–MnO 2 coplanar gas sensor arrays, Sens Actuators B

145 (2010) 457–463.

[25] K.F Cai, E Müller, C Draˇsar, A Mrotzek, Preparation and thermoelectric properties of Al-doped ZnO ceramics, Mater Sci Eng B 104 (2003) 45– 48.

[26] R.K Nath, S.K Nath, Sn-doped zinc oxide thin films for methanol, Sens Trans-ducers 108 (2009) 168–179.

[27] P.P Sahay, R.K Nath, Al-doped ZnO thin films as methanol sensors, Sens Actu-ators B 134 (2008) 654–659.

[28] P.T Ivanov, Design, Fabrication and Characterization of Thick-Film Gas Sensors, Tarragona, Spain, 2004.

[30] M.N Rumyantseva, E.A Makeeva, A.M Gas’kov, Influence of the microstructure

of semiconductor sensor materials on oxygen chemisorption on their surface,

Russ J Gen Chem 78 (12) (2008) 2556–2565.

[31] C.C Li, Z.F Du, H.C Yu, T.H Wang, Low-temperature sensing and high

sensi-tivity of ZnO nanoneedles due to small size effect, Thin Solid Films 517 (2009)

5931–5934.

[32] H Ogawa, M Nishikawa, A Abe, Hall measurement studies and electrical

con-duction model of tin oxide ultrafine particle films, J Appl Phys 53 (1982)

4448–4455.

Biographies

Thanh Thuy Trinh completed her undergraduate studies at the University of

Sci-ence, Ho Chi Minh City, Vietnam and now studying M.Sc course in Sungkyunkwan

University, South Korea in 2009.

Huy Hoang Le finished his undergraduate studies at the University of Science, Ho

Chi Minh City, Vietnam in 2008.

Ngoc Han Tu finished her undergraduate studies at the University of Science, Ho

Chi Minh City, Vietnam in 2006 and now is a M.Sc student in College of Technology, Hanoi, Vietnam.

Khac Binh Le currently he is a Professor of Physics at the Faculty of Material Science,

University of Science, Ho Chi Minh City, Vietnam.

Junsin Yi is a Professor at the School of Information and Communication

Engineer-ing, Sungkyunkwan University, Korea.