The reactive surface sites and the H2S sensing potential for the SnO2 produced by a mechanochemical milling

Article abstract of DOI:10.1149/1.1570414

SnCl₂ powder was milled with Ca(OH)₂ and K₂CO₃ powder respectively in a ball mill at room temperature and in an air atmosphere. Pure SnO₂ was obtained by removal of the CaCl₂ or KCl by-product by washing the powder. An initial composition of SnCl₂, K₂CO₃, and Cr(NO₃)₃·9H₂O was milled to produce the doped SnO₂ phase with Cr on the surface. Heat-treatment results in the formation of a tetragonal phase. The infrared spectra for the different powders and from thick films of these materials are reported. An important finding is that during milling in the presence of water, OH groups are formed at the surface in great numbers. After heat-treatment at 400°C or higher, the OH bonds of adjacent grains are probably transformed into Sn-O-Sn bridges. Washing the material with pure ethanol, instead of water, can result in the reduction of intensity for the band at 1450 cm^-1 due to the presence of the bridge bonding Sn-O-Sn. On exposure to H₂S and CO gas, the SnO₂ film prepared from anhydrous powder, has higher sensitivity than the SnO₂ film, prepared from hydrated powder. It is shown that the SnO₂ produced by mechanochemical synthesis can be successfully applied to a H₂S gas sensor.

Full text of DOI:10.1149/1.1570414

Journal of The Electrochemical Society, 150 ͑6͒ H129-H134 ͑2003͒
H129
0013-4651/2003/150͑6͒/H129/6/$7.00 © The Electrochemical Society, Inc.
The Reactive Surface Sites and the H2S Sensing Potential for
the SnO2 Produced by a Mechanochemical Milling
a,z
U. Kersen and M. R. Sundbergb
¨
aDepartment of Electrical and Communications Engineering, Laboratory of Electromechanics, Helsinki
University of Technology, Fin-02015 Finland
bDepartment of Chemistry, Laboratory of Inorganic Chemistry, University of Helsinki, Fin-00014 Helsinki,
Finland
SnCl2 powder was milled with Ca(OH)2 and K2CO3 powder respectively in a ball mill at room temperature and in an air
atmosphere. Pure SnO2 was obtained by removal of the CaCl2 or KCl by-product by washing the powder. An initial composition
of SnCl2 , K2CO3 , and Cr(NO3)3•9H2O was milled to produce the doped SnO2 phase with Cr on the surface. Heat-treatment
results in the formation of a tetragonal phase. The infrared spectra for the different powders and from thick films of these materials
are reported. An important finding is that during milling in the presence of water, OH groups are formed at the surface in great
numbers. After heat-treatment at 400°C or higher, the OH bonds of adjacent grains are probably transformed into Sn-O-Sn bridges.
Washing the material with pure ethanol, instead of water, can result in the reduction of intensity for the band at 1450 cmϪ1 due to
the presence of the bridge bonding Sn-O-Sn. On exposure to H2S and CO gas, the SnO2 film prepared from anhydrous powder,
has higher sensitivity than the SnO2 film, prepared from hydrated powder. It is shown that the SnO2 produced by mechanochemi-
cal synthesis can be successfully applied to a H2S gas sensor.
© 2003 The Electrochemical Society. ͓DOI: 10.1149/1.1570414͔ All rights reserved.
Manuscript submitted August 19, 2002; revised manuscript received December 16, 2002. Available electronically April 11, 2003.
inorganic sol-gel route, where tetrachloride is used as a tin precur-
sor. Heating at 350°C causes complete decomposition of organic
compounds, and FTIR shows a broad band in the range 3750-2500
cmϪ1 due to different O-H stretching modes of alcohol. Molecular
water and Sn-OH groups are not present, only the band of SnO2
vibrations at about 600 cmϪ1 and that of a lattice mode at about 690
cmϪ1 is observed.5 The measured FTIR spectra in the range 4000-
800 cmϪ1 revealed different bands, assigned to the unsymmetrical
hydroxyl stretching, water deformation, and surface hydroxyl
stretching modes.2-4
The repeated ball-powder collisions during milling induce struc-
tural changes and continually regenerate the reacting interfaces, al-
lowing chemical reactions to occur.6 Extensive surface treatment by
use of high rotation speeds ͑100-200 rpm͒ and enhanced ball-to-
powder mass ratio ͑10:1͒ during milling give rise to quite different
surfaces, which may lead to differences in surface charge and hence
intergranular barrier heights. The presence of free water causes a
considerable change in the reactivity of the surface. Thus, mecha-
nochemically synthesized SnO2 surfaces may have considerably dif-
ferent chemical compositions and reactivities than those prepared by
other methods.2-4 It was found in the present work that with appro-
priate selection of milling conditions and chemical reaction, it is
possible to reduce the intensity of the antisymmetric Sn-O-Sn modes
of surface-bridging oxide at 1450 cmϪ1. Hydroxyls attached to Sn
atoms are highly electron accepting, causing the resulting complex
to be strongly dipolar.7 Thus, the consumption of surface OH groups
by reacting with gas could lead to an increase in the conductance.
The aim of this work is to present more information about the
surface composition and reactivity of SnO2 powder used in the
preparation of a gas sensors. Neither the study of the surface reac-
tivity of mechanochemically synthesized particles, nor the influence
of the surface reactivity to the gas-sensing response, have been re-
ported before.
The current methods of fabrication of semiconductor SnO2 gas
sensors requires the use of nanocrystalline powder with grain size of
less than 20 nm. After synthesis the powder is mixed with organic
solvent to form a paste. Nanosized SnO2 powders for gas sensing
applications have been produced by several methods, including sol-
gel processing, spray pyrolysis, pulsed laser ablation, chemical va-
por deposition, and sputtering. Often nanocrystalline structures ob-
tained by these methods have a high state of agglomeration.
Mechanochemical processing ͑MCP͒ is an alternative method for
the production of nanosized powders, where separated nanoparticles
can be formed.1 This method is not time consuming and is of low
cost.
Only one previous paper has been published on the uses of SnO2
thin films prepared by mechanochemical processing as a gas sensor.2
Separated tin dioxide nanoparticles, averaging 24 nm, were pro-
duced, and examined by X-ray photoelectron spectroscopy ͑XPS͒,
but in this work, the surface was not characterized by infrared spec-
troscopy.
In this paper, the synthesis of SnO2 nanoparticles by MCP is
reported. It is found that the details of powder synthesis and pro-
cessing have a large influence on the final electrical properties of the
gas-sensitive layer. Since the surface reactions important to the gas
sensing mechanism are complex, involving the transfer of electronic
charges, it is important to analyze the reactive surface sites.
Fourier transform infrared ͑FTIR͒ spectrometry is a reliable tool
for the determination of existing surface species; however, large
differences in sensor fabrication lead to very different surfaces, mak-
ing it difficult to compare results obtained by different authors. The
FTIR spectra of SnO2 powders calcined at 250°C ͑for use in im-
proved sol-gel fabricated thick film gas sensors͒ shows bands at
1632 cmϪ1 indicating the presence of adsorbed water and also hy-
droxyl absorption bands in the range 2000-3600 cmϪ1. The water
content and associated lattice distortions largely disappear after
heat-treatment above 450°C as indicated by the disappearance of the
associated absorption bands.3 Another synthesis method based on
the microwave treatment of a solution of tin chloride produces pow-
der precursors in a few minutes. In the infrared spectrum at 450°C
there are again clear absorption bands from molecular water and
hydroxyl groups associated with lattice distortions and defects.4
SnO2 thin films may also be prepared by using a chloride-based
Experimental
Nanosized SnO2 powder particles were produced by mecha-
nochemical synthesis, where a milling process of duration 4 h was
performed at room temperature in a centrifugal type mill. Two
chemical reactions were used. The first one is
SnCl ϩ Ca OH͒ → SnO ϩ CaCl ϩ H O
͓1͔
͑
2
2
2
2
It is important to notice that this reaction produces water. The start-
ing materials ͑2.33 g of SnCl2 , 0.91 g of Ca(OH)2 , and 15.45 g of
z E-mail: ukersen@hotmail.com
Downloaded on 2014-10-16 to IP 134.71.135.126 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H130
Journal of The Electrochemical Society, 150 ͑6͒ H129-H134 ͑2003͒
CaCl2) were sealed in a nonmetal vial in an air atmosphere and
milled to produce 5g of SnO. Milling was performed using steel
balls with a diam of 12.7 mm. The ball-to-powder mass ratio of 10:1
was used with a rotation speed of 75 rpm ͑low speed͒, 125 rpm
͑moderate speed͒, and 200 rpm ͑high speed͒. The milled powder
was subsequently heat-treated at 400°C for 1 h in air atmosphere to
complete the SnO2 formation. Removal of the CaCl2 by-product
was carried out both, with distilled water and pure ethanol.
The second reaction used is
SnCl2 ϩ K2CO3 → SnO ϩ 2KCl ϩ CO2
͓2͔
This reaction does not produce water. The starting materials ͑4.23 g
of SnCl2 and 3.09 g of K2CO3) were sealed in a nonmetal vial in an
air atmosphere and milled to produce 3 g of SnO. The ball-to-
powder mass ratio of 22:1 was used with a rotation speed of 125
rpm. The milled powder was washed by distilled water.
The starting materials ͑2.33 g of SnCl2 , 0.91 g of Ca(OH)2 ,
0.098 g of Cr(NO3)•9H2O and 15.45 g of CaCl2) were sealed and
milled to produce the doped SnO phase with Cr distributed on the
surface. The ball-to-powder mass ratio of 10:1 was used with a
rotation speed of 125 rpm. The milled powder was washed by deion-
ized water. After washing all powders were dried at 65°C for 8 h.
A Philips PW 1830 X-ray powder diffractometer with Cu K␣
radiation was used to follow the crystallization of the powder as a
function of heat-treatment temperature. The crystallite size, D, is
calculated according to the Scherrer equation
Figure 1. X-ray diffraction pattern of the milled ͑for 4 h͒, heat treated ͑at
400°C for 1 h͒, and washed powder with a typical distribution of ͑ϩ͒ tetrag-
onal and ͑Ϫ͒ orthorhombic SnO2 phases. The sharp peak ͑x͒ at 29.4 ͑2⌰°͒
indicates residual CaCl2 .
An X-ray diffraction ͑XRD͒ pattern of the synthesized powder, after
removal of the CaCl2 mass and drying, is presented in Fig. 1. The
tetragonal phase is clearly dominant with the orthorhombic phase,
giving the rise to only two very weak peaks. It may be concluded
that the crystallization has succeeded at this low temperature of
400°C. The main three peaks are located at 26.6, 33.9, and 51.8
͑2⌰°͒ degrees, respectively, and show a high degree of crystallinity.
L 2␪͒ ϭ K␣/ D cos ␪͒
͓3͔
͑
͑
*
The sharp peak ͑ ͒ at 29.4 ͑2␪°͒ indicates residual CaCl2 .
X-ray analysis is important for the confirmation of the micro-
structure of the synthesized powder obtained after washing and re-
moval of all CaCl2 . The SnO2 crystallite sizes determined using
Scherrer’s equation were in the range 10-25 nm dependent on the
different milling conditions. The sample of powder for Brunauer,
Emmett, Teller analysis was milled for 5 h at a speed of 125 rpm.
The total pore volume comprised 0.25 mL/g with a measured spe-
cific surface area of 79 m2/g.
where L(2␪) is the width at half-maximum intensity ͑in radians͒, ␪
is the Bragg angle of the ͑110͒ plane in the diffraction pattern of the
peak, K is a constant depending on the line shape profile ͑currently
K ϭ 1), ␣ is the wavelength of the X-ray source ͑in the case of Cu
radiation ␣ ϭ 1.54059 Å), and D is the crystallite size.8 The spe-
cific surface area and porosity of the synthesized powders are mea-
sured with a Coulter Omnisorp 100 CX, using a N2 gas adsorption
method at 77 K. The FTIR spectra were taken with a Bio-Rad FTS
6000 spectrometer with photoacoustic output in the range 4000-400
cmϪ1. All powders synthesized by MCP were washed both in water
and in ethanol, dried, additionally heat-treated at 750°C for 1 h, and
cooled down to room temperature before making of the FTIR mea-
surements.
Thick-film gas sensors were prepared by conventional hand coat-
ing of a paste, the SnO2 powder mixed with an ␣-terpineol-based
solvent, over the 0.2 mm gap of Au sensing electrodes on a 7
ϫ 8 mm Al2O3 substrate. Subsequently these substrates were sin-
tered at 700°C for 1 h in air. Pt wires were attached to the pads of
the heater and sensing electrodes by using an Au paste and heat-
treatment. The H2S and CO gas response tests were carried out with
a PC-controlled measuring system. Sensors to be tested were placed
in a measuring chamber of about 0.51 in volume. Optimum detec-
tion temperatures were determined and thick-film sensors main-
tained at this fixed temperature were exposed to varying test gas
concentrations of H2S ͑3 ppm and 10 ppm͒ and CO ͑50 ppm and
100 ppm͒, using a gas flow controller unit with an air flow rate of
0.51 minϪ1. Sensitivity, S, to gas is defined
It should be noted that water vapor present inside the milling
vessel or produced by the chemical reaction accompanying milling,
has a considerable influence on the milling mechanism. The current
work shows that molecular water and hydroxyls are remained on the
surface at much higher temperature than 500°C. The hydroxyl
groups appear as a result of the acid-base reaction with the OHϪ
sharing its electronic pair with the Lewis acid ͑Sn͒ leaving weakly
bonded protons. The correlation between the behavior of the con-
ductance and the intensity of the surface hydroxyl deformation band
͑Sn-OH͒ in the infrared spectrum shows that the decrease in con-
ductivity in the temperature range 150-450°C is due to loss of hy-
droxyl groups from the oxide surface.9
Two types of mechanism, proposed by Heiland and Kohl,10 to
explain the increase in surface conductivity in the presence of water
vapor can also result in the formation of two types of OH groups
during MCP, one including lattice oxygen Olat the other bound to
lattice tin Snlat
H2O ϩ Snlat ϩ Olat
→
HO Ϫ Sn ͒ ϩ O H ϩ e
͓6͔
͑
lat
lat
Dissociation of water in the gas phase results in the reaction
H ϩ OϪlat O H͒Ϫ ϩ eϪ
S ϭ R air͒ Ϫ R g͒/R air͒ 100%
͓4͔
͓
͑
͑
͑
͔
ϭ
͓7͔
͑
lat
where R(air) is the sensor resistance in clean air and R(g) is the
resistance in the test gas.
The second mechanism is related to the reaction between the proton
and the lattice oxygen with subsequent binding of the resulting OH
group to the Snlat atom
Results and Discussion
The mixture, produced after 4 h ballmilling with a rotation speed
of 125 rpm, was heat treated at 400°C for 1 h, resulting in the
exothermic reaction
H2O ϩ 2Snlat ϩ Olat → 2 HO Ϫ Sn ͒ ϩ V
͓8͔
͑
lat
o
The oxygen vacancy created produces electrons in the conduction
band. The OH group bound to Snlat has been detected spectroscopi-
SnO ϩ 1/2 O͒ g͒ → SnO
͓5͔
͑
͑
2
2
Downloaded on 2014-10-16 to IP 134.71.135.126 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑6͒ H129-H134 ͑2003͒
H131
Figure 2. FTIR spectra of SnO2 powder milled with moderate speed. Curve
1 shows results for a powder heat-treated at 400°C and washed in pure
ethanol, and curve 2 the results for a powder washed in distilled water.
Figure 3. FTIR spectra of SnO2 powder milled with high speed. For curve 1
the powder was heat-treated at 400°C and washed in distilled water, while
for curve 2 the powder was washed in pure ethanol.
Both adsorption of water and hydroxylation are unfavourable at
high temperatures. It is possible that at a sufficiently high tempera-
ture, the surface may be free from molecular water, but may be
partially covered by OH groups. In Fig. 5 there are spectra for five
different powders obtained by mechanochemical synthesis and heat-
treated at 750°C for 1 h. It can be noticed that the intensities of all
the adsorption bands are reduced to some extent after high tempera-
ture heat-treatment compared to those in Fig. 4. The powders
washed with water and milled with speeds of 75-125 rpm have the
highest intensities for the asymmetric Sn-O-Sn stretching and hy-
droxyl deformation modes in Fig. 5. Treatment of the milled com-
position with ethanol causes the band at 1450 cmϪ1 to decrease
͑curves 3, 4͒. The milled composition produced by the reaction
given by ͑Eq. 2͒ contains no water, and therefore the deformation
bands are of lowest intensity ͑curve 5͒, resulting only from the ad-
sorption of water from the gas phase during the milling.
cally. Probably during milling the reaction shown by Eq. 8 is initi-
ated and a certain number of OH groups are bound to Snlat . The
reaction shown by Eq. 6 may be partially started by milling, but it is
of evident that reaction requires some activation of the composition
by heat-treatment at 400°C.
The FTIR difference spectrum of powder, synthesized at moder-
ate speed, using the reaction shown by Eq. 1 is illustrated in Fig. 2.
A broad absorption band in the range 3750-2500 cmϪ1 is due to the
asymmetric hydroxyl stretching mode and the band between 820
and 1300 cmϪ1 is assigned to the hydroxyl deformation mode. The
molecular water deformation mode is located at 1640 cmϪ1. The
band at 1450 cmϪ1 is most probably an overtone of the asymmetric
Sn-O-Sn stretching mode of a surface-bridging oxide formed by
condensation of adjacent surface hydroxyl groups. The SnO2 vibra-
tions are observed at 625 cmϪ1 and lattice modes of SnO2 appear at
690 cmϪ1. Washing of the heat-treated powders with pure ethanol
leads to lower band intensities compared to washing with water. Of
especial importance is the intensity of the band at 1450 cmϪ1. It has
been concluded by Morimoto that dehydration of the surface of
SnO2 occurs principally via condensation of adjacent surface hy-
droxyl groups on the ͑100͒ plane with secondary condensation on
the ͑101͒ plane.11 Hydroxyl group condensation on these two planes
will lead to the formation of surface Sn-O-Sn bridges. It was con-
cluded in previous work that a possible factor, affecting the sensi-
tivity to H2S gas, is the formation of strong Sn-O-Sn stretching
mode deformation and its influence on the H2S adsorption.12 It was
suggested that strong oxygen bonds, formed during the dehydration,
remain on the surface and thus reduce the number of active sites for
gas adsorption. At high speed milling ͑200 rpm͒ there is no essential
difference between OH band intensities, whether the composition
was washed by water or ethanol, as shown in Fig. 3. The differences
can be found in the intensity of the band at 1450 cmϪ1. ͑The sharp
peak at 890 cmϪ1 can be attributed to a surface hydroxyl deforma-
tion band.͒
The response and recovery times of sensors made from the pow-
ders depend on the operating temperature, both being slow at low
The FTIR spectra of powders synthesized with different milling
conditions and chemical reactions are presented in Fig. 4. The pow-
der produced by the chemical reaction ͑Eq. 2͒ and milled with mod-
erate speed ͑curve 1͒ is characterized by the lowest intensities of
adsorption bands. The curve 2 is the infrared spectrum of SnO2
powder with Cr2O3 distributed on the surface, where the band at
1450 cmϪ1 is increased in intensity. Curves 3 and 4 show the FTIR
spectra of two undoped SnO2 thick film sensors, sintered at 700 and
600°C, respectively.
Figure 4. FTIR spectra of SnO2 powder milled with high speed. Curve 1
shows the spectrum of the anhydrous powder and curve 2 that of the hydrous
powder. Curves 3 and 4 show the spectra of the SnO2 thick films prepared
from these powders.
Downloaded on 2014-10-16 to IP 134.71.135.126 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H132
Journal of The Electrochemical Society, 150 ͑6͒ H129-H134 ͑2003͒
Figure 7. Response of an SnO2 thick films to 10 ppm of H2S in air. The
powder was ͑curve 1͒ milled with high speed, heat-treated and washed in
pure ethanol, or ͑curve 2͒ milled with moderate speed and washed in pure
ethanol.
Figure 5. FTIR spectra of SnO2 powders milled with different speeds and
heat-treated at 750°C.
temperatures, and fast at higher temperature. An operating tempera-
ture of about 225-250°C was found to give the maximum sensitivity
to H2S gas. The increase of the conductivity in the presence of H2S
may be explained by the reaction between H2S and oxygen ions on
the surface of SnO2
operating temperatures of 200-250°C have been observed
previously.14,15 The response time depends considerably on the sen-
sor operating temperature. At a higher temperature of 300°C the
sensor current in air is below 2 mA, which corresponds to a sensor
resistance of 2.5 k⍀. The current increases to 20 mA in gas ͑curve b
in Fig. 7͒, and there is a faster gas response of the surface, but the
sensitivity to H2S is decreased to 12.
H2S ϩ 3O2Ϫ → SO2 ϩ H2O ϩ 6eϪ
͓9͔
An attempt was made to Cr dope the SnO2 particles by ballmill-
ing using chromium nitrate nonhydrate Cr(NO3)3•9H2O. The de-
composition of the main starting materials during milling could be
followed according to the Eq. 1. By heat-treatment at 400°C, chro-
mium nitrate particles decompose and Sn and Cr ions change their
oxidation states as below
According to this reaction, the interaction of H2S with previously
adsorbed O2Ϫ ions on the surface, is resulting in injection of elec-
trons into the depletion layer of SnO2 grains.13 This reaction reduces
the band bending of the intergranular depletion layer and leads to an
enhanced conduction of the sensor. Figure 6 shows the gas response
to 3 and 10 ppm of H2S gas. The powder for the sensor was milled
at high speed and washed with water. Since the thick film, as pre-
pared, has a low porosity, the reaction in the bulk takes a consider-
able time. Removal of the salt matrix with ethanol increases, to
some extent, the sensitivity to the gas, as shown by curve a in Fig. 7.
Comparison with Fig. 6 shows that gas is adsorbed much more
quickly. Similar response times, of the order 20-40 min, for H2S at
2 Cr NO ͒ •9H O → Cr O ϩ 6NO ϩ 1/2O ϩ 18H O
͓
͑
͔
3
3
2
2
3
2
2
2
͓10͔
Coordination bonds between Cr3ϩ ion and water in the starting ma-
terial will be transformed into Cr-O-Cr bonds upon heating to
Figure 8. Dynamic response of a Cr doped SnO2 thick film to 50 and 100
ppm of H2S in air. The powder used in the preparation was milled with high
speed, heat-treated and washed in distilled water.
Figure 6. Dynamic response of an SnO2 thick film to 3 and 10 ppm of H2S
in air. The powder was milled with high speed, heat-treated, and washed in
distilled water.
Downloaded on 2014-10-16 to IP 134.71.135.126 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 150 ͑6͒ H129-H134 ͑2003͒
H133
Figure 9. Dynamic response of an SnO2 thick film prepared from anhydrous
powder to 3 and 10 ppm of H2S in air.
Figure 11. Response of an SnO2 thick film prepared from anhydrous pow-
der to 50 and 100 ppm of CO in air.
400°C. The occupation of Cr6ϩ ion sites by H2O or OH groups may
be one explanation for the reduced sensitivity of this Cr doped SnO2
thick film as presented in Fig. 8. It is noticeable that there is a
considerably higher intensity for Sn-O-Sn stretching mode of sur-
face bridging oxide for this sample ͑curve 2 in Fig. 4͒.
The main sensing mechanism may be explained by the reaction
between H2S and oxygen ions adsorbed on the surface of SnO2 . It
can be seen in Fig. 9 that enhanced gas sensitivity (S ϭ 44) was
achieved upon exposure to 10 ppm of H2S.
A comparison of the sensitivity to H2S and CO gas for the dif-
ferent types of SnO2 sensors has been performed in order to get
confirmation of the proposed sensing mechanism. The two chemical
reactions that underly the chemical sensing mechanism are
heat-treatment, is shown in Fig. 10. Due to the very small concen-
trations of CO ͑50 and 100 ppm͒, the responses are not high. When
the powder was obtained using milling at moderate speed with the
chemical reaction given by Eq. 2, the resulting thick film has a more
distinct response, shown in Fig. 11, as well as a higher response to
CO in comparison to Fig. 10. A possible explanation for this could
be that powders manufactured with high and very high milling
speeds have greater surface charge due to a large number of surface
OH groups. During heat-treatment, however, most of OH groups are
transformed into Sn-O-Sn bonds, which are not involved in gas
adsorption.
Conclusions
1/2O g͒ ϩ eϪ ↔ OadϪ
͓11͔
͓12͔
͑
2
Mechanochemical processing of known stoichiometric amounts
of SnCl2 with Ca(OH)2 or with K2CO3 in a centrifugal action re-
sulted in the formation of the desired mass of SnO. Use of the
former reactant results in the production of water, while use of the
latter does not. The infrared spectra of the powders produced sup-
port the evidence that two mechanisms proposed by Heiland and
Kohl to describe the increase of sensor conductance by exposure to
water vapor, can be applied to explain the formation of surface
hydroxyl groups during the milling. The FTIR spectra of the pow-
ders and of the SnO2 thick film sensors fabricated from them are
influenced to a large degree by the energy input in the milling and
the type of chemical reaction to be used during of milling. Using
K2CO3 reactant, the chemical reaction initiated by ball milling does
not produce water, therefore the FTIR spectrum of the synthesized
powder has no surface hydroxyl deformations, only a small amount
of molecular water shown by the band at 1640 cmϪ1. For this case
most of adsorbed oxygen ions are distributed near the Sn ions on the
surface of SnO2 grains. The interpretation of gas sensing mechanism
is that H2S or CO gas reacts with these oxygen ions, releasing elec-
trons into the conduction band, with simultaneous decrease in sur-
face charge and an increase in conductance.
CO g͒ ϩ OadϪ → CO g͒ ϩ eϪ
͑
͑
2
where OadϪ is the negatively charged surface oxygen that can react
with the CO to produce electrons to the conduction band. The re-
sponse of the SnO2 thick film, where the powder for the preparation
was obtained with high speed milling and washed with water after
The response of SnO2 thick film sensors to H2S gas may be
improved by ͑i͒ selection of the proper chemical reaction that does
not produce water during mechanochemical synthesis; ͑ii͒ producing
powder grains smaller than 10 nm and ͑iii͒ the using FTIR spectros-
copy to control the conditions of manufacture of the SnO2 surface.
Acknowledgments
The authors would like to express their sincere thanks to Dr.
George Darling, Surface Science Research Center, University of
Liverpool, for his help on preparing the manuscript.
Figure 10. Response of an SnO2 thick film to 50 and 100 ppm of CO in air.
The powder used in the preparation was milled with high speed, heat-treated,
and washed in distilled water.
Downloaded on 2014-10-16 to IP 134.71.135.126 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

H134
Journal of The Electrochemical Society, 150 ͑6͒ H129-H134 ͑2003͒
8. B. D. Cullity, Elements of X-Ray Diffraction, 2nd ed., p. 102, Addison-Wesley
Publishing, Reading, MA, ͑1978͒.
References
1. T. Tsuzuki and P. G. McCormick, Physica B, 239, 378 ͑1997͒.
2. L. M. Cukrov, P. G. McCormick, K. Galatsis, and W. Wlodarski, Sens. Actuators B,
77, 491 ͑2001͒.
9. P. G. Harrison and M. J. Willett, J. Chem. Soc., Faraday Trans. 1, 85, 1928 ͑1989͒.
10. G. Heiland and D. Kohl, in Chemical Sensor Technology, Vol. 1, T. Seiyama,
Editor, p. 15, Elsevier, Amsterdam ͑1988͒.
3. A. Dieguez, A. Romano-Rodriguez, J. R. Morante, J. Kappler, N. Barsan, and W.
Gopel, Sens. Actuators B, 60, 125 ͑1999͒.
11. K. Morishige, S. Kittaka, and T. Morimoto, Bull. Chem. Soc. Jpn., 53, 2128 ͑1980͒.
12. U. Kersen, Mater. Sci. Forum, 386-388, 637 ͑2002͒.
¨
¨
4. A. Cirera, A. Vila, A. Dieguez, A. Cabot, A. Cornet, and J. R. Morante, Sens.
Actuators B, 64, 65 ͑2000͒.
5. P. Siciliano, Sens. Actuators B, 70, 153 ͑2000͒.
13. J. N. Zemel, Thin Solid Films, 163, 189 ͑1988͒.
14. M. Rumyantseva, M. Labeau, G. Delabouglise, L. Ryabova, I. Kutsenok, and A.
Gaskov, J. Mater. Chem., 7, 1785 ͑1997͒.
6. P. G. McCormick and F. H. Froes, JOM, 50, 61 ͑1998͒.
7. K. D. Schierbaum, U. Weimar, and W. Gopel, Sens. Actuators B, 3, 205 ͑1991͒.
15. L. Jianping, W. Yue, G. Xiaoguang, M. Qing, W. Li, and H. Jinghong, Sens. Ac-
tuators B, 65, 111 ͑2000͒.
¨
Downloaded on 2014-10-16 to IP 134.71.135.126 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).