Sensors and Actuators B 48 (1998) 277 – 284

In-situ EXAFS analysis of the local environment of Pt particles

incorporated in thin films of SnO2 semi-conductor oxide used as
gas-sensors
M. Gaidi a,*, M. Labeau a, B. Chenevier a, J.L. Hazemann b
a
Laboratoire des Matériaux et du Génie Physique, Institut National Polytechnique de Grenoble, UMR 5628 CNRS, BP 46,
F-38402 Saint Martin d’Hères Cedex, France
b
Laboratoire de Géophysique Interne et Tectonophysique, URA 733, BP 53, F-38041 Grenoble Cedex, France

Abstract

A very fine and homogeneous dispersion of metallic nanoparticles (3 – 5 nm) inside polycrystalline SnO2 films was carried out
by modified CVD co-deposition. It improves the sensitivity and selectivity for reducing gas detection. In order to verify and refine
the models which can explain the role of metallic particles in the detection process, the environment of metallic nanoparticles were
studied in Pt doped SnO2 thin film by in-situ EXAFS analysis. XANES region of EXAFS spectra were used to study the variation
of the local environment and oxidation state as a function of the temperature and atmosphere. © 1998 Elsevier Science S.A. All
rights reserved.

Keywords: Gas sensors; Thin film; Tin oxide; Conductance; X-ray; Absorption; Metallic aggregate

1. Introduction This work is an in-situ analysis of the role of the Pt

SnO2 is often used as an active element for gas
detection. It interacts with reducing gases by surface
adsorption leading to a modification of the free electron
density on the surface. By dispersing a low concentra-
tion of small metallic particles (Pd, Pt, Cu…) on the
surface of the SnO2 grains, the sensitivity and selectivity
can be improved [1 – 3]. The interaction phenomena
between CO and the composite (SnO2 +Pt or Pd (or
PdO) clusters) can be described by a simple model
where the desorption of reducing species leads to a
modification of local environment of the metallic ele-
ment and of its oxidation state as well, leading to an
electronic charge transfer between the clusters and the
SnO2 support. Another model describes the metallic
particles as catalysts which activate the interaction be-
tween gas and SnO2 surface (spill-over effect) [4]. These
models nevertheless need to be verified and refined to
allow accurate manipulation of the sensitivities to vari-
ous gases and temperatures.
* Corresponding author. Fax: +33 476826394;
gaidi@inpg.fr
particles in the CO detection process using a very fine
and homogeneous dispersion of nanoparticles included
in polycristalline SnO2 films. To limit the number of
parameters having an influence on the CO detection,
this study was carried out using well controlled atmo-
spheres. It is to notice that these conditions do not
correspond exactly to a realistic surrounding atmo-
sphere in which the humidity, for instance, acts as an
interfering gas for the CO detection.
2. Experiments
Thin films of Pt-doped tin oxide (3 atomic %, 6 and
12% Pt in the solution) have been deposited on oxidised
silicon substrates by submicronic aerosol pyrolysis (Py-
rosol process) [5]. A very fine and homogeneous disper-
sion of metallic aggregates (3–5 nm) within the film is
obtained by co-deposition of the two species (tin oxide
and metal). Metallic size inside the film increases with
the amount of metal.
Electrical measurements were carried out on pure
e-mail: and Pt-doped SnO2 overcoated with evaporated gold
electrodes. Two kinds of measurement have been done.

0925-4005/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.

PII S0925-4005(98)00056-2
278 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284

Fig. 1. Effects of the Pt additive concentration on SnO2 film conductance as a function of temperature under pure air and air + CO(300 ppm):
– – – pure air —— air+CO (300 ppm).

Fig. 2. XANES region of 6%Pt/SnO2 sample at 350°C: height (hwl) of the white line at the Pt LIII edge under various working conditions: pure
air, air+ CO(300 ppm) and N2 + CO (900 ppm).

Firstly conductance was measured in steady state in a
50 – 500°C temperature range using two kinds of atmo-
sphere: (1) pure air; and (2) air charged with 300 ppm
CO. A detailed description of this kind of measure-
ments has been reported in a previous paper [6]. Sec-
ondly, the electrical conductance has been measured at
a fixed temperature as a function of time and gas. In
that case pure air and 900 ppm CO in nitrogen were
used.
As EXAFS is a very powerful technique to detect
and analyse the surroundings of metallic elements un-
der controlled atmosphere, the authors have used X-ray
 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284
Fig. 3. Observation of the EXAFS oscillations as a function of the reduction process in 6%Pt/SnO2 sample and comparison with Pt and PtO2
references.
absorption to study the local environment and the
oxidation state (XANES) of the metallic elements
added into the films. To confirm that the method could
be used, the authors decided to analyse at first a
favourable case, films with platinum as metallic addi-
tive. An EXAFS in-situ experiment has been carried
out as the material (SnO2 films with various amounts of
Pt particles—12, 6 and 3% — deposited on an oxidised
2× 2 cm2 silicon substrate) was inserted in a specific
reaction cell, heated to maximum temperatures of
400°C and put in contact with air or with a polluting
gaseous mixture containing small amounts of reducing
gas (Air+300 ppm CO, N2 +900 ppm CO). The thick-
ness of the films was in the range of 1 mm. The
Fig. 4. XANES region of the sample (white line) under pure air at
350°C, as a function of Pt content. Pt and PtO2 references are also
reported.
279
experiment has been done at ESRF on the beam line
BM32. The X-ray absorption has been measured in
fluorescence mode within an energy range close to Pt
LIII edge.
3. Results
The conductance of pure and Pt-doped SnO2 films
was measured in steady state successively under pure
air and 300 ppm of CO in air. As shown in Fig. 1
Pt-doped SnO2 films have a very good sensitivity to CO
in air when compared with pure tin oxide films which
have a bad sensitivity and a high temperature reply. In
the polluted gaseous atmosphere, the increase of the
conductance of undoped SnO2 does not show any well
defined maximum as a function of temperature. In
contrast, Pt-doped SnO2 films show a maximum of
conductance at low temperature (100°C).
Thus, the authors use the well-known correlation
between the X-ray absorption at the LIII edge and the
occupation of the metal 5d band. Many works have
shown the potential of the technique to follow the
modification of the electronic state of platinum during
reduction [7–10]. Schematically, the greater the white
line intensity, the less the d band is occupied. In other
words, the more oxidised platinum is, the higher the
white line becomes.
To study the effect of CO on the oxidation state of
Pt, the samples were put in contact with three kinds of
gas (900 ppm CO in nitrogen, air + 300 ppm CO, air).
Fig. 2 shows the evolution of white line under various
working conditions. It can be seen that the platinum
particles exhibit two states: either oxidised or reduced,
depending on the reducing ability of the gas. In the case
of pure air (respectively N2 + CO) Pt shows an oxida-
tion state (respectively reduction state) where qualita-
tively, the XANES and EXAFS spectrum can be
280 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284

Fig. 5. 6%Pt/SnO2 sample, evolution of the height (hwl) of the white line as a function of the gas, temperature and time.

Fig. 6. Kinetic measurement of the electrical conductance under N2 +900 ppm CO at T= 300°C (semi-log scale).

Fig. 7. Kinetic measurement of the electrical conductance under N2 +900 ppm CO at different temperatures.
 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284 281

Fig. 8. Evolution of the duration (in min) of (i) the electrical reply (transition between part A and part B in Fig. 6) and (ii) Pt reduction observed
on XANES region (Fig. 5). (i) = white dots, (ii) =black dots.

Fig. 9. Effect of the gas flow on the electrical conductance and on the time reply (inset curve).

correlated with an oxidised state characterised by Pt–O
bonding (PtO2 type) and reduced state characterised by
Pt – X bonding (X =CO or Pt) (see Fig. 3.). The oxida-
tion state is determined by making a comparison be-
tween experimental spectra and a standard reference (Pt
and PtO2). Table inside Fig. 2 shows the evolution of the
white line intensity (labeled as hwl—height of the white
line—in the following) when the atmosphere changes
from air to air + 300 ppm CO or N2 + 900 ppm CO, the
drop in hwl shows that Pt passes from an oxidation state
to a reduction state. It must be mentioned that platinum
aggregates keep their electronic state unchanged for a
282 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284
certain time before the transformation to the reduction
state. A detailed description of the kinetic behaviour of
the platinum reduction is given in the following para-
graph.
An important metallic particle size effect has been
detected, the change in the white line is particularly
pronounced for the smallest particles concentration, i.e.
the finest particles, whereas it is very weak for 12% Pt
concentration. The XANES study shows that the finest
particles have the highest oxidation state (see Fig. 4).
This size effect can be understood in terms of specific
area. For the small particles all the bulk of the Pt
particles is oxidised, whereas for the big ones only the
surface is oxidised. In that case the metallic phase
widely prevails.
3.1. Kinetics study of the reduction under CO of
incorporated platinum
A study of the kinetics of the oxidation/reduction
process was also carried out on 6% Pt sample by using
the evolution of the white line as a function of gas,
temperature and time. The 6% Pt sample was thermally
cycled between 25 and 400°C, alternating the atmo-
sphere at each temperature (pure air or 900 ppm CO in
nitrogen) (see Fig. 5).
The hwl evolution has been studied as a function of
temperature, gas and time (Fig. 5). The main conclu-
sion of this study is that the reduction of platinum
under CO has a strongly temperature dependent kinet-
ics. In the higher temperature range (T ] 200°C) the
reaction has a slow kinetic and the time necessary for
the total reduction of platinum increases with increas-
ing temperature. This time is obtained by measuring the
time for the drop in hwl ($50 min at T= 300°C).
Three steps in this range of temperature can be
distinguished.
“ A first regime where the platinum aggregates keep
their electronic configuration unchanged (the white
line intensity remains constant).
“ A second regime follows, when the reduction of
platinum begins and the white line intensity drops
rapidly.
“ Thirdly a change in the rate of the platinum reduc-
tion which goes on but at a slower speed.
In the lower temperature range (T= 100°C) a fast
kinetic reaction occurs and only the second and the
third regime can be observed.
3.2. Correlation with kinetic electrical measurement
After introducing the gas (N2 +900 ppm CO), first a
weak and afterwards a strong and fast electrical reply is
observed (see Fig. 6). At first the conductance increases
slowly and remains constant for certain time (region A)
after which it shows a maximum (in order of few
millisiemens: region B). Then the conductance decreases
slowly, tending towards a constant value (region C).
To study the effect of temperature on the fast electri-
cal reply, many series of electrical conductance mea-
surement at a fixed temperature under air and N2 +900
ppm CO were carried out as a function of time (see Fig.
7). A fast reply of tin oxide (great peak of conductance)
at lower temperature and very slow reply at the higher
one is observed. With decreasing temperature the swift-
ness of the second part of the electrical reply increases
and shows a tendency to rejoin the first part (at about
100°C)
The authors have reported in Fig. 8 the time evolu-
tion of platinum reduction and the time of tin oxide
reply. It is to be noticed that the difference in both
EXAFS and electrical measurement times is due to a
difference in the experiment condition (in the rate of
the flow of gas). The authors have demonstrated in Fig.
9 the effect of the gas flow on the electrical conductance
(inside: on the time of reply). The time of reply de-
creases with increasing the gas flow.
4. Discussion
This study shows that the platinum reduction under
CO occurs in all temperature ranges but with various
kinetics. XANES and electrical measurements show
that both platinum reduction and tin oxide reply ex-
hibit three regimes of evolution as a function of time.
From these results a reaction mechanism can be pro-
posed as follows: Platinum oxide reacts with CO and
changes to a state of metallic Pt:
PtO2 + 2CO“ Pt+ 2CO2 (1)
At elevated temperature Pt is oxidised by the
chemisorbed oxygen:
Pt+ 2Oabs “ PtO2 (2)
A relatively moderate electrical reply is then observed
(region A in Fig. 6). The chemisorbed oxygen dimin-
ishes and leads to a change of the electrical conduc-
tance. A second phenomena which can be also
considered is the difference between Eq. (1) and Eq. (2):
Eq. (1) is probably faster than Eq. (2).
After a certain time tin oxide will be unable to give
more oxygen and Pt becomes in the reduced state, at
this moment the great peak in the electrical conduc-
tance appears (Fig. 7) and the white line drops. Here
only Eq. (1) is valid and Pt acts as an electron donor.
At low temperature the Eq. (2) is very slow and only a
fast reduction of platinum occurs (Eq. (1)) and induces
a fast electrical reply (Fig. 7).
In the case of air+ CO, oxygen in the gas state
chemisorbs on the surface of SnO2 and partially com-
pensates the chemisorbed oxygen consumption, leading
to a moderate change in electrical properties:
 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284
Ogas “Oabs (3)
In that case the same behaviour than in region
(A) under N2 + 900 ppm CO occurs. The proposed
reaction mechanism is valid only in the case of the
finest Pt particles (3 and 6% Pt doped films) which
are able to change in all the bulk its oxidation state
in contact with pure air or CO in correlation with
the higher sensitivity to CO (Fig. 1).The effect of the
particle size on the interaction between CO and
other metal particles has already been reported. C.
Henry et al. [11,12] showed that very fine Pd parti-
cles are more active than the big ones. This size ef-
fect can be understood in term of the presence of
additional adsorption sites for oxygen and CO at the
particle surface which make easier the interaction be-
tween the two molecules.
5. Conclusion
This study shows the wide potentiality of EXAFS
(XANES) to make in-situ reaction and their kinetic
mechanism. Combined with electrical measurement,
the XANES study is able to provide a description of
the interaction between Pt doped tin oxide and re-
ducing gases. The major results of this study can be
summarised as follows:
1. The oxidation state of the platinum depends on
the particles size. The finest particles show the
more oxidised state.
2. The oxidation (reduction) of platinum occurs at
all investigated temperatures but with different ki-
netics.
3. The electrical reply of tin oxide is mainly due to
the change of Pt from the oxidation state to the
reduced one.
4. An analysis of the kinetic reactions provides a
simple explanation for the presence of the con-
ductance peak at lower temperature (fast reduc-
tion of Pt).
These results and the proposed mechanisms are re-
lated to laboratory conditions which do not include
the role of interfering gases for the CO detection, for
instance humidity. They form a good basis for an
understanding of the mechanisms in the surrounding
atmosphere.
References
[1] S. Semancik, J.B. Fryberger, Model studies of SnO2-based gas
sensors: vacancy defects and Pd additive effects, Sens. Actua-
tors B 1 (1990) 97 –102.
[2] S. Matsushima, J. Maekawa, J. Tamaki, N. Miura, N. Yama-
zoe, New methods for supporting palladium on a tin oxide
gas sensor, Sens. Actuators B 9 (1992) 71–78.
283
[3] D. Kohl, The role of noble metals in the chemistry of solid
state gas sensors, Sens. Actuators B 1 (1990) 158 – 165.
[4] N. Yamazoe, New approaches for improving semiconductor
gas sensors, Sens. Actuators B 5 (1991) 7 – 19.
[5] M. Labeau, B. Gautheron, F. Cellier, M. Vallet-Regi, E. Gar-
cia, J.M. Gonzalez Calbet, Pt nanoparticles dispersed on
SnO2 thin films: a microstructural study, J. Solid State Chem.
102 (1993) 434 – 439.
[6] B. Gautheron, M. Labeau, G. Delabouglise, U. Schmatz, Un-
doped and Pd-doped SnO2 thin films for gas sensors, Sens.
Actuators B 15 – 16 (1993) 357 – 362.
[7] D.E. Sayers, D. Bazin, H. Dexpert, A. Jucha, E. Dartyge, A.
Fontaine, P. Lagarde, EXAFS and near edge structure III, in:
K.O. Hodgson, B. Hedmans, J.E. Penner-Hahn (Eds.), Dy-
namical Studies of Catalytic Systems Using Dispersive X-ray
Absorption Spectroscopy, Springer, Berlin, 1984, p. 209.
[8] D. Bazin, H. Dexpert, P. Lagarde, The EXAFS tool, Topic
in Current Chemistry, 154, Springer, Berlin, 1988, p. 70.
[9] D. Bazin, H. Dexpert, N.S. Guyot-Sionnest, J.P.
Bournonville, J. Lynch, EXAFS characterization of reforming
catalysts: Examples of recent applications, J. Chim. Phys. 86
(1989) 1707 – 1716.
[10] A. Caballero, A.R. Gonzalez, P. Malet, G. Munuera, J. Gar-
cı́a, J.C. Conesa, E. Burattini, EXAFS/XANES studies of the
influence of the drying pretreatements on the reducibility of
Pt/Al2O3 and Pt-Re/Al2O3 catalysts, Phys. B 158 (1989) 158–
159.
[11] C.R. Henry, On the effect of the diffusion of carbon monox-
ide on the substrate during CO oxidation on supported palla-
dium clusters, Surf. Sci. 223 (1989) 519.
[12] C. Becker, C.R. Henry, A second adsorption state on palla-
dium clusters on MgO (100), Catal. Lett. 43 (1997) 55.
Biographies
Mounir Gaidi, graduated at Monastir University of
Science (Tunisia) in 1995 (Master in Chemistry and
Physics), started in 1996 a PhD thesis work in Mate-
rial Physics at the LMGP on the ‘Structure and
Electronic State of Metallic Aggregates Incorporated
in Semi-Conducting SnO2 Thin Films Used as Pollu-
ant Gas Sensors’. He has acquired a good expertise
in the optimisation of sensor properties by manipu-
lating the concentrations of metallic nanoparticles
added in SnO2 films.
Michel Labeau, Professor in Material Sciences,
Doctor of Physics Science, 23 years experience of re-
search in Material Sciences, including thin films syn-
thesis of semiconductor for gas applications. Head of
the ‘Films for components and Sensors’ Group of
LMGP. Responsible of several national and interna-
tional cooperation programs.
Bernard Chene6ier, Senior Scientist, Doctor of
Physics Science, he has strong expertise in mi-
crostructural and physical properties analysis of thin
films. He has been responsible for LMGP partner-
ship in several national ant international programs.
284 M. Gaidi et al. / Sensors and Actuators B 48 (1998) 277–284

Jean-Louis Hazemann, head of the EXAFS Station
on the Beam line project ‘IF’ at the European Synchro-
tron Radiation Facility (ESRF), Doctor in Material
Science, he has been involved for 7 years in synchrotron
instrumentation. Chief of the EXAFS Station, he has in
charge the development of the scientific activity on
various uses of the X-ray Absorption Spectroscopy
(XAS) in the frame of the program of structural investi-
gation on disordered materials in environmental prob-
lems.

.
.