Heats of adsorption and activation energies of surface processes measured by infrared spectroscopy

Article abstract of DOI:10.1016/j.molcata.2007.08.013

The temperature-programmed adsorption equilibrium of CO at 4 kPa on Pd/SiO₂ and of CO₂ at 101.3 kPa on γ-Ga₂O₃ was studied by infrared spectroscopy between 298 and 723 K. The relative coverage (θ) evolution of each adsorbed species as a function of the adsorption temperature was measured by normalizing the IR absorption bands. The Temkin's model was used to fit the experimental data of θ versus temperature of each surface species and to obtain the initial (θ = 0) and final (θ = 1) isobaric heats of adsorption of linear and bridged CO adsorbed on palladium, and of HCO₃^- (bicarbonate), b-CO₃^2- (bidentate carbonate), br-CO₃^2- (bridged carbonate) and CO₂^- (carboxylate) on gallium oxide. The reactive adsorption of CO (101.3 kPa) on gallium oxide was also studied. CO reacts with surface OH groups to give formate species from 448 K. These formate species are further decomposed to CO₂(g) and Ga-H species from 573 K onwards. The whole process was modeled by considering the rate of reaction for each elementary step. Thus, the activation energy of the reactive adsorption of CO and of the formate decomposition on Ga₂O₃ was calculated. The calculated kinetic parameters are discussed and compared to those reported in the literature.

Full text of DOI:10.1016/j.molcata.2007.08.013

Available online at www.sciencedirect.com
Journal of Molecular Catalysis A: Chemical 281 (2008) 73–78
Heats of adsorption and activation energies of surface
processes measured by infrared spectroscopy
∗
´
´
Sebastian E. Collins, Miguel A. Baltanas, Adrian L. Bonivardi
Instituto de Desarrollo Tecnolo´gico para la Industria Qu´ımica (INTEC), Gu¨emes 3450, S3000GLN Santa Fe, Argentina
Available online 19 August 2007
Abstract
The temperature-programmed adsorption equilibrium of CO at 4 kPa on Pd/SiO2 and of CO2 at 101.3 kPa on ␥-Ga2O3 was studied by infrared
spectroscopy between 298 and 723 K. The relative coverage (θ) evolution of each adsorbed species as a function of the adsorption temperature
was measured by normalizing the IR absorption bands. The Temkin’s model was used to fit the experimental data of θ versus temperature of each
surface species and to obtain the initial (θ = 0) and final (θ = 1) isobaric heats of adsorption of line−ar and bridged CO adsorbed on palladium, and
−
2−
2−
of HCO3 (bicarbonate), b-CO3 (bidentate carbonate), br-CO3 (bridged carbonate) and CO2 (carboxylate) on gallium oxide. The reactive
adsorption of CO (101.3 kPa) on gallium oxide was also studied. CO reacts with surface OH groups to give formate species from 448 K. These
formate species are further decomposed to CO2(g) and Ga–H species from 573 K onwards. The whole process was modeled by considering the
rate of reaction for each elementary step. Thus, the activation energy of the reactive adsorption of CO and of the formate decomposition on Ga2O3
was calculated. The calculated kinetic parameters are discussed and compared to those reported in the literature.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Carbon monoxide; Carbon dioxide; Activation energies; Gallium oxide; Palladium
1. Introduction
instance, temperature-programmed desorption (TPD) methods
permit the determination of the desorption activation energies,
which can be equal to the heat of adsorption whenever the
adsorption process is not activated.
However, these techniques present experimental difficulties
such as: (i) the presence of several co-adsorbed species on the
surface, (ii) possible parallel reactions, (iii) mass and heat trans-
fer phenomena and (iv) readsorption of the products. In this
respect, Gorte [3] and Gorte and Demmin [4] established a series
of criteria allowing to set the experimental conditions, particu-
larly in regard to TPD, where the influence of factors such as
diffusion, lag time and readsorption over the measurement could
be minimized.
Some of the above-mentioned problems can be overcome
by carrying out adsorption experiments under equilibrium con-
ditions because, among other considerations, the theoretical
modeling of the recorded experimental data becomes simpli-
fied [5]. Consequently, some authors have proposed to use
variable-temperature infrared spectroscopy to measure heats of
adsorption for different surface species [5–8]. The infrared spec-
tra are used to follow the evolution of the coverage of a particular
surface species as a function of the adsorption temperature at a
constant pressure of the adsorbed gas, and a suitable model is
then used to calculate the heat of adsorption.
Heats of adsorption and activation energies of chemisorbed
species participating in surface reactions on solid catalysts are
required parameters to perform a microkinetic modeling, i.e., to
correlate the elementary reaction steps with the performance of
a catalyst [1].
Classically, adsorption microcalorimetry and adsorption
isotherms over a temperature range are the methods used for
determining heats of adsorption [2]. Microcalorimetry allows
to measure differential heats of adsorption as a function of the
surface species coverage and, thus, average integral heats can be
derived. Then, the Claussius–Clapeyron equation can be applied
to a series of isotherms recorded at different temperatures to give
the isosteric heat (enthalpy) of adsorption.
In addition, elementary kinetic parameters are frequently
measured by temperature-programmed experiments, that is,
measures taken under transient conditions, which consist in a
controlled perturbation to the gas/solid system and the measure-
ment of a signal generated from the gas phase or the surface. For
∗
Corresponding author. Tel.: +54 342 4559175; fax: +54 342 4550944.
E-mail address: abonivar@ceride.gov.ar (A.L. Bonivardi).
1381-1169/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2007.08.013

74
S.E. Collins et al. / Journal of Molecular Catalysis A: Chemical 281 (2008) 73–78
Bianchi et al. have been successful in using variable-
To study the adsorption of CO over Pd/SiO2, this catalyst
was exposed to flowing CO (4%)/N2 (100 cm3/min) at 298 K
and heated to 723 K (5 K/min). The adsorption of pure CO2 or
CO on calcined ␥-Ga2O3 was performed under flow of each gas
(100 cm3/min) from 298 to 723 K (5 K/min) at 101.3 kPa.
A Shimadzu 8210 FTIR spectrometer with a DGTS detec-
tor was used in the transmission mode; spectra were acquired
every 25 K at 4 cm−1, averaging 25–100 scans. When needed,
background correction of the spectra was achieved by subtract-
ing the spectra of the pretreated wafers at each temperature. A
Lorentzian sum function was used to fit the overlapping bands
and to measure peak areas and/or intensities [25].
temperature infrared spectroscopy to measure the isobaric
heats of adsorption of individual, linear and bridged CO
species on several metal supported catalysts, such as Pt/Al2O3
[9–12], Cu/Al2O3 [5,13,14], Ru/Al2O3 [15], Pd/Al2O3 [16,17],
Pd/CeO2/Al2O3 and Pd/La2O3/Al2O3 [17], Pt/Ru/CeO2/Al2O3
[8,17], Ir/Al2O3 [18], Pt/ZrO2 [19] and Au/TiO2 [20]. How-
ever, as far as we know, this approach has been seldom applied
to the study of the adsorption and/or reactive adsorption over the
surface of metal oxides [7,19].
Here we explore the use of this approach to study two
cases: (i) the temperature-programmed equilibrium adsorption
of CO and CO2 on Pd/SiO2 and ␥-Ga2O3, respectively and
(ii) the temperature-programmed reactive adsorption of CO on
␥-Ga2O3.
All the gases used in this study were high purity grade
and were further purified as follows: H2 (AGA, 99.999%),
N2 (AGA, 99.998%) and CO2 (AGA, 99.996%) were passed
˚
through MnO/Al2O3 and molecular sieve (3 A, Fisher Co.) traps
2. Experimental
to eliminate oxygen and water impurities, respectively; CO
(AGA Research grade, 99.99%) and O2 (AGA, 99.999%) were
passed through molecular sieve (3 A, Fisher Co.) and Ascarite
˚
2.1. Materials
traps to remove water and carbon dioxide, respectively.
Pd/SiO2 (2 wt.% Pd) was obtained by ion exchange of
Pd(II) acetate (Sigma, 99.97%) onto silica (Davison 59; BET
surface area = 301 m2/g) at pH 11 in NH4OH(aq) [21]. The
exchanged material was washed at the same pH, and dried at
423 K. The resulting diammine palladium complex was then
calcined in flowing air (200 cm3/min), by heating from 298
to 673 K (2 h) at 2 K/min to obtain dispersed PdO on the sur-
face. The PdO/SiO2 was reduced in a 5% (v/v) H2/Ar mixture
(200 cm3/min), heating from 298 to 723 K (2 h) at 2 K/min.
Finally, the catalyst was passivated at 298 K by flowing O2/Ar
mixtures (200 cm3/min) with increasing O2 content, from 0.1
to 5% (v/v). The Pd metal dispersion, measured by the double
isotherm of hydrogen chemisorption at 293 K, was 58%. ␥-
Ga2O3 was obtained from Ga(NO3)3·xH2O (Strem, 99.99 wt.%
Ga) as previously described [22–24]. The presence of a single
crystallographic phase was verified by XRD (Shimadzu model
XD-D1 diffractometer; Cu K␣ radiation). The specific surface
area of the ␥-gallia polymorph was 105 m2/g, measured by the
BET isotherm (N2, 77 K) in a Micromeritics Accusorb 2001E
unit.
3. Results and discussion
3.1. Surface coverage and heats of adsorption of CO
species on Pd/SiO2
Fig. 1 shows the temperature-variable infrared spectra taken
in the 298–723 K temperature range, under flow of CO (4%)/N2.
The IR band initially centered at 2090 cm−1 is assigned to linear
oron-topCO(COL), whichdownshiftsduetocoveragedecrease
at increasing temperature. The broad and overlapped bands at
1975 and 1925–1900 cm−1 correspond to bridged-bonded CO
[COB on (1 0 0) planes and particle edges] and hollow-bonded
CO [COH on (1 1 1) planes], respectively [26–29]. As it can be
seen in the spectra, the absorbance of the COB species decreases
monotonically and down shifts, while the COH band slightly
2.2. In situ FTIR spectroscopy
Temperature-programmed experiments were performed by
employing self-supported wafers of 13 mm in diameter, pressing
10 mg of Pd/SiO2 or 30 mg of ␥-Ga2O3 at 5 t/cm2. A heated IR
cellmadeinPyrexglasswithwatercooledNaClwindows, which
was attached to a conventional high-vacuum system (base pres-
sure = 1.33 × 10−4 Pa) equipped with a manifold for gas flow,
was used for the in situ experiments.
Samples were pretrated in situ with O2 (100 cm3/min,
5 K/min)at723 Kandcooledundervacuumtoroomtemperature
(RT). Additionally, the Pd/SiO2 catalyst was further reduced in
H2 (100 cm3/min, 5 K/min) at 723 K (30 min) and again cooled
down under vacuum to 298 K. It was verified that this activa-
tion protocol does not change the morphology of the employed
solids.
Fig. 1. Infrared spectra during the temperature-programmed adsorption of CO
(4 kPa) on Pd/SiO2, between 298 and 723 K (the arrow indicates spectra at
increasing temperature).

S.E. Collins et al. / Journal of Molecular Catalysis A: Chemical 281 (2008) 73–78
75
comparing the coverage of the CO adsorbed species obtained
using FTIR with mass spectroscopy data. To this end, in an
additional experiment we measured the CO coverage by mass
spectrometry, following the procedure described by Bourane et
al. [31], and a fairly good agreement was achieved as compared
to infrared relative coverages determined from integrated peak
areas. Moreover, it can be theoretically shown that the transition
probability between two vibrational states is independent of the
temperature, as long as the re-population of the vibrational states
of higher energy does not occur upon increasing the temperature
[32]. Since CO adsorption on Pd is considered a non-activated
process (the activation energy of adsorption is close to zero), a
new equilibrium can be reached in ∼10−2 s for a temperature
perturbation equal to 10 K [5]. Then, under our experimental
conditions (that is, for a heating rate of 5 K/min) the adsorption
equilibrium is reached almost instantaneously, which satisfies
point (c).
Based on the previous assumptions, the thermal evolutions
of θL and θB and at a constant CO pressure (4 kPa) were used
to determine the heats of adsorption of COL and COB species,
respectively. Temkin’s model produced the best fitting, in agree-
ment with previous results on palladium catalysts [16,17]. This
model assumes that the heat of adsorption decreases linearly
with the increase of the coverage. So:
ꢀ
ꢁ
RT
1 + L0P
1 + L1P
θ =
ln
(1)
ꢀE
where T and P stand for the adsorption temperature and pressure,
respectively, ꢀE stands for the difference between the heat of
adsorption at θ = 0 (E0) and the one at θ = 1 (E1); and L0 and L1
stand for the adsorption coefficients at θ = 0 and 1, respectively.
The adsorption coefficients (Lθ) were estimated by statistical
thermodynamicsassumingthat theCOadsorbedspeciesis local-
ized (loss of three translational degrees of freedom) [5,15–17]:
Fig. 2. Thermal evolution of the: (a) COL, and (b) COB, COH and COB+H
relative coverages on Pd/SiO2 at increasing temperature. Full lines correspond
to the simulation using the Temkin’s model.
ꢀ
ꢁ
h3
1
T5/2
Eθ
reduces its intensity between 550 and 673 K due to its higher
heat of adsorption as compared to the other CO adsorbed species
[26]. This last observation is in agreement with the detection of
COH species on Pd monocrystals at temperatures higher than
700 K, by IRAS [26,30].
The evolution of the relative surface coverage (θ) of linear-,
bridged- and hollow-bonded CO on Pd/SiO2 (θL, θB and θH,
respectively) as a function of the adsorption temperature is
Lθ =
exp
(2)
k(2πmk)3/2
RT
where h is the Planck’s constant, k the Bolztmann’s constant, m
the weight of the adsorbate and Eθ is the heat of adsorption at
θ = 0 or θ = 1.
The best fits of the experimental CO coverages as a func-
tion of the adsorption temperature are shown in Fig. 2 (full
lines) for the following heats of adsorption: E0 = 125 kJ/mol and
E1 = 49 kJ/mol for COL and E0 = 98 kJ/mol and E1 = 51 kJ/mol
for COB. We also calculated the overall heat of adsorption (E0
and E1) for COB plus COH (that is, COB+H) in order to compare
our results with those reported in the literature (it is clear from
Fig. 2b, though, that thermodinamic data cannot – and should
not – be computed for COH due to the narrow range of θH vari-
ation). Table 1 shows a fair agreement between the different
Eθ values for the COL and COB+H adsorptions over supported
palladium crystallites. The differences can be adscribed to the
different supports and metal particle sizes.
shown in Fig. 2. For comparison purposes (see below) the θB+H
,
which corresponds to COB+H ( COB + COH,) is also included.
The relative coverage of CO was determined as the ratio
between the integrated absorbance of each IR band (or pair
of bands for the case of COB+H) at each adsorption tempera-
ture (Fig. 1) and the highest IR band area at 298 K. That is, we
assumed that: (a) the intensity of the IR bands are proportional to
the amount of the CO species on the surface, (b) the IR absorp-
tion coefficients are independent of the adsorption temperature
and, also, that (c) the system is in equilibrium at each tempera-
ture. The first two assumptions, (a) and (b), can be sustained by

76
S.E. Collins et al. / Journal of Molecular Catalysis A: Chemical 281 (2008) 73–78
Table 1
Heats of adsorption of bonded CO on palladium particles supported on different metal oxides, calculated by the variable-temperature infrared adsorption approach
Catalyst
COL
COB
COB+H
Reference
E0 (kJ/mol)
E1 (kJ/mol)
E0 (kJ/mol)
E1 (kJ/mol)
E0 (kJ/mol)
E1 (kJ/mol)
Pd/Al2O3
92
92
92
92
125
54
54
54
54
49
–
–
–
–
98
–
–
–
–
51
168
168
168
168
168
92
75
92
92
62
[16,17]
[17]
[17]
[17]
This work
Pd/(Cl)Al2O3
Pd/CeO2/Al2O3
Pd/La2O3/CeO2/Al2O3
Pd/SiO2
δ(OH) = 1225 cm−1]; bidentate carbonate, b-CO32− [νas(CO) =
3.2. Coverage and heats of adsorption of CO2 adsorbed on
γ-Ga2O3
1587 cm−1
,
νs(CO) = 1325 cm−1]; polydentate carbonate,
2−
p-CO3
[νas(CO) = 1460 cm−1
,
νs(CO) = 1410 cm−1];
2−
bridged carbonate, br-CO3 [νas(CO) = 1750 cm−1, νs(CO) =
Upon exposure of the ␥-gallia polymorph at 323 K to increas-
ing CO2 pressure, up to atmospheric pressure (101.3 kPa),
several signals readily evolved in the 2000–1000 cm−1 region.
The absorbance of these bands reached a saturation plateau at
approximately 600 Torr of CO2 (not shown) [24]. The infrared
spectra collected during the temperature-programmed adsorp-
tion of CO2 (at 101.3 kPa) on ␥-Ga2O3 are reported in Fig. 3. All
these bands followed the same, reverse evolution after decreas-
ing the temperature from 723 to 300 K under flowing pure
CO2, which suggests that all carbonaceous species are reversibly
adsorbed on Ga2O3.
−
1170 cm−1]; and carboxylate, CO2 [νas(CO) = 1680 cm−1
,
νs(CO) = 1280 cm−1]. More details about these assignments
can be found in Refs. [23,24].
The relative coverage of the carbonate species during the
temperature-programmed adsorption experiment (298–723 K)
of CO2 are plotted in Fig. 4, taking into account the same
assumptions made regarding the CO adsorption on Pd [former
items (a–c)], but for CO2 now. The coverage of HCO3−, b-
2−
−
CO32−, br-CO3 and CO2 decreased monotonically during
the adsorption of CO2 at increasing temperature, while the con-
centration of p-CO32− remained almost constant (not shown in
Fig. 4). It is worth noticing that the same evolution of the cov-
erage of the carbonate groups was obtained during the cooling
process of the sample in the IR cell from 723 to 298 K under
flowing CO2 at 101.3 kPa, which confirms that the adsorption of
CO2 ongalliaalsosatisfiedtheadsorptionequilibriumcondition.
The experimental data were first adjusted using the Lang-
muir’s adsorption model, which assumes that the heat of
adsorption is independent of the coverage. However, the Lang-
muir’s model did not provide a successful fitting for the CO2
adsorption data on gallia. Then, the thermal evolution of the
coverage of each surface carbonate species was predicted by
For the assignment of the IR signals of the carbonaceous
surface species chemisorbed onto Ga2O3, it is appropriate to
examine the following features [23,24]: (i) their frequency posi-
tion and its comparison with reported IR signals of carbonates
adsorbed over other metal oxides, (ii) the width of the ꢀν3-band
2−
splitting of the CO3 anion (that is, the ꢀν3 = νas − νs of the
CO3 stretching modes) due to the loss of its D3h symmetry by
chemisorption and (iii) the thermal stability of each signal.
Accordingly, the IR signals in the 2000–1000 cm−1 region
were assign−ed to the following surface species: bicarbon-
ate, HCO3
[νas(CO) = 1630 cm−1
,
νs(CO) = 1430 cm−1
,
Fig. 3. Infrared spectra during the temperature-programmed adsorption of CO2
(101.3 kPa) on ␥-Ga2O3, between 298 and 723 K (the arrow indicates spectra at
increasing temperature).
Fig. 4. Thermal evolution of the relative coverages of carbonate and bicarbonate
species on ␥-Ga2O3 (except for p-CO32−) at increasing temperature. Full lines
correspond to the simulation using the Temkin’s model.

S.E. Collins et al. / Journal of Molecular Catalysis A: Chemical 281 (2008) 73–78
77
using the Temkin’s model, and the best fit is depicted in
Fig. 4.
Unfortunately, neither E0 nor E1 of individual surface car-
bonate groups have been reported, as far as we know, in the
open literature. Petre et al. [33] reported an initial global heat of
adsorption of CO2 over Ga2O3 equal to 115 kJ/mol measured by
microcalorimetry. This value is higher than the E0 determined
in this work (E0 = 55–72 kJ/mol, see Fig. 4). Nevertheless, it is
clear that initial heat of adsorption reported by Petre et al. [33] is
an average, overall value, i.e., the result of the adsorption of sev-
eral surface species, including the p-CO32− group. Fig. 3 shows
2−
that the p-CO3 coverage slightly changed along the adsorp-
tion ramp (see also Fig. 9 in Ref. [24]). Then, it is not possible
to calculate any heat of adsorption for this species. Yet, for a
surface species whose coverage does not change appreciably
throughout the temperature range used here, we can still make
anestimationofthelowerexpectedvalueoftheinitialadsorption
Fig. 6. Thermal evolution of the relative coverages of OH, HCOO and Ga–H
species on ␥-Ga2O3 at increasing temperature. Full lines correspond to the
simulation (see text).
2−
heat. Hence, E0 > 200 kJ/mol is predicted for p-CO3 species.
Therefore, the average value for the initial heat of adsorption
of all carbonate species over Ga2O3 might be closed to the one
reported by Petre et al. [33].
species (HCOO) [νas(COO) = 1580 cm−1, δ(CH) = 1386 cm−1
νs(COO) = 1372 cm−1] [23]. From 573 K onwards HCOO
decomposes into Ga–H surface species [ν(Ga–H) = 1990 cm−1
,
3.3. Reactive adsorption of CO on γ-Ga2O3
]
[22] and CO2(g). Considering that the adsorption of CO pro-
ceeds according to an Eley–Rideal mechanism, the following
pathway for the reactive adsorption of CO on gallia can be
proposed:
We have shown in the previous section that variable-
temperature infrared spectroscopy can be used to measure heats
of adsorption of carbonate species on gallium oxide, because
the adsorption process is reversible and can be considered
under equilibrium condition at each temperature. However, this
approach can be further employed when the adsorbate reacts
irreversibly with a solid surface, as is the case of CO adsorption
on gallium oxide.
Ga–OH + CO(g) → HCOO–Ga
HCOO–Ga → Ga–H + CO2(g)
(3)
(4)
CO chemisorbs on ␥-Ga2O3 from 448 K (Fig. 5) by reacting
In the same direction, Haneda et al. proposed the HCOO
formation and the subsequent release of CO2(g) for the CO
adsorption on Ga2O3/Al2O3, using FTIR and mass spectrom-
etry [34]. Furthermore, the complete removal of OH surface
group was verified over the surface of our sample in additional
CO2 and/or CO adsorption experiments and after the absence of
any bicarbonate and/or formate group formation.
with surface OH (3800–3000 cm−1) groups to produce formate
Fig. 6 shows the normalized coverage of OH, HCOO and
Ga–H surface groups. These relative coverages were calcu-
lated assuming that the total concentration of said surface
species was equal to the initial concentration of OH groups [i.e.,
θ(OH) + θ(HCOO) + θ(Ga–H) = 1].
According to the above reaction mechanism, the reactive
adsorption of CO on ␥-Ga2O3 can be described by the following
differential equations:
ꢀ
ꢁ
dθ(OH)
dT
A1
β
Ea1
RT
= −
T
−3/2θ(OH)PCO exp
−
(5)
(6)
ꢀ
ꢁ
dθ(HCOO)
dT
A1
Ea1
=
T
−3/2θ(OH)PCO exp
−
β
RT
ꢁ
ꢀ
Fig. 5. Infrared spectra during the temperature-programmed reactive adsorption
of CO (101.3 kPa) on ␥-Ga2O3, between 298 and 723 K (the arrow indicates
spectra at increasing temperature).
A2
β
−Ea2
−
Tθ(HCOO) exp
RT

78
S.E. Collins et al. / Journal of Molecular Catalysis A: Chemical 281 (2008) 73–78
ꢀ
ꢁ
Agency for the Promotion of Science and Technology
(ANPCyT) of Argentina.
dθ(GaH)
dT
A2
β
Ea2
RT
=
T θ(HCOO) exp
−
(7)
where Ai stands for the Arrhenius pre-exponential factor (not a
function of the temperature for localized surface species [2,35]),
Eai for activation energy of reaction i (i = 1 or 2), and β stands
for the heating rate.
References
[1] J.A. Dumesic, ACS Professional Reference Book, American Chemical
Society, Washington, DC, 1993.
[2] M. Boudart, G. Dje´ga-Mariadassou, Kinetics of Heterogeneous Catalytic
Reactions, 1st ed., Princeton University Press, New York, 1984.
[3] R.J. Gorte, J. Catal. 75 (1982) 164.
[4] R.A. Demmin, R.J. Gorte, J. Catal. 90 (1984) 32.
[5] S. Derrouiche, D. Bianchi, Langmuir 20 (2004) 4489, and references
therein.
[6] E. Garrone, C. Otero Area´n, Chem. Soc. Rev. 34 (2005) 846–857.
[7] C. Otero Area´n, O.V. Manoilova, G. Turnes Palomino, M. Rodr´ıguez Del-
gado, A.A. Tsyganenko, B. Bonelli, E. Garrone, Phys. Chem. Chem. Phys.
4 (2002) 5713.
[8] T. Chafik, O. Dulaurent, J.L. Gass, D. Bianchi, J. Catal. 179 (1998) 503.
[9] O. Dulaurent, D. Bianchi, Appl. Catal. A 196 (2000) 196.
[10] A. Bourane, O. Dulaurent, D. Bianchi, J. Catal. 196 (2000) 115.
[11] A. Bourane, O. Dulaurent, D. Bianchi, Langmuir 17 (2001) 5496.
[12] A. Bourane, O. Dulaurent, K. Chandes, D. Bianchi, Appl. Catal. A 214
(2001) 193.
[13] S. Zeradine, A. Bourane, D. Bianchi, J. Phys. Chem. B 105 (2001) 7254.
[14] O. Dulaurent, X. Courtois, V. Perrichon, D. Bianchi, J. Phys. Chem. B 104
(2000) 6001.
[15] O. Dulaurent, M. Nawdali, A. Bourane, D. Bianchi, Appl. Catal. A 201
(2000) 71.
[16] O. Dulaurent, K. Chandes, C. Bouly, D. Bianchi, J. Catal. 188 (1999)
237.
This system of differential equations was numerically inte-
grated (fourth order Runge–Kutta) and the error between the
model and the experimental data was minimized with Ai and Eai
as parameters (Down-Simplex). Full lines in Fig. 6 represent the
best fits of the experimental data with A1 = 1.95 × 105 K3/2/s Pa,
Ea1 = 35 kJ/mol and A2 = 3.27 × 104 (K s)−1, Ea2 = 103 kJ/mol.
The calculated pre-exponential factors are in the same order
of magnitude than the expected ones for adsorption and decom-
position/desorption processes of gases on metal surfaces [2]. A
comparison of the activation energy for formate formation is not
possible at the moment, because the results reported in the litera-
ture usually deal with the synthesis of formate from CO2 and H,
instead of CO and OH [36–38]. Nevertheless, the kinetic of the
synthesis and decomposition of formate species have been stud-
ied on copper single crystals [36,37] and on copper supported
catalysts [37,38] due to the importance of this intermediate on
the water gas shift reaction: Yatsu et al. [37] reported an activa-
tion energy for the formate decomposition on copper equal to
115.7 kJ/mol, which is in reasonable agreement with the value
of Ea2 calculated in this work.
[17] O. Dulaurent, K. Chandes, C. Bouly, D. Bianchi, J. Catal. 192 (2000)
262.
[18] A. Bourane, M. Nawdali, D. Bianchi, J. Phys. Chem. B 106 (2002) 2665.
[19] O. Dulaurent, D. Bianchi, Appl. Catal. A 207 (2001) 211.
[20] S. Derrouiche, P. Gravejat, D. Bianchi, J. Am. Chem. Soc. 126 (2004)
13010.
[21] A.L. Bonivardi, D.L. Chiavassa, C.A. Querini, M.A. Baltana´s, Stud. Surf.
Sci. Catal. 130D (2000) 3747.
[22] S.E. Collins, M.A. Baltana´s, A.L. Bonivardi, Langmuir 21 (2005) 962.
[23] S.E. Collins, M.A. Baltana´s, A.L. Bonivardi, J. Catal. 226 (2004) 410.
[24] S.E. Collins, M.A. Baltana´s, A.L. Bonivardi, J. Phys. Chem. B 110 (2006)
5498.
[25] J.D. Ingle, S.R. Crouch, Spectrochemical Analysis, 1st ed., Prentice Hall,
Upper Saddle River, NJ, 1988, p. 211.
4. Conclusions
In situ variable-temperature infrared spectroscopy was
employed to follow the relative coverage of different surface
species generated by adsorption of CO over a Pd/SiO2 catalyst
and of CO2 and CO over ␥-Ga2O3. Under equilibrium reversible
adsorption, the individual heats of adsorption of CO on Pd and
CO2 on ␥-Ga2O3, leading to the formation of different sur-
faces species in each case, were calculated using the Temkin’s
adsorption model.
The variable-temperature infrared spectroscopy strategy was
also used to study the reaction of CO with OH groups on
the surface of ␥-Ga2O3. The thermal evolution of OH, HCOO
and Ga–H species was successfully simulated considering the
system of differential equations that describe the elementary
reaction steps, by using the pre-exponential factors and the acti-
vation energies as fitting parameters.
[26] E. Ozensoy, D.W. Goodman, Phys. Chem. Chem. Phys. 6 (2004) 3765.
[27] T. Lear, R. Marshall, J.A. Lo´pez-Sanchez, S.D. Jackson, T.M. Klapo¨tke,
M. Bra¨umer, G. Rupprechter, H.J. Freund, D. Lennon, J. Chem. Phys. 123
(2005) 174706.
[28] F.M. Hoffmann, Surf. Sci. Rep. 3 (1983) 1.
[29] G.C. Cabilla, A.L. Bonivardi, M.A. Baltana´s, Catal. Lett. 55 (1998) 147.
[30] W.K. Khun, J. Szanyi, D.W. Goodman, Surf. Sci. 303 (1994) 377.
[31] A. Bourane, O. Dulaurent, B. Bianchi, J. Catal. 195 (2000) 406.
[32] G. Herzberg, Molecular Spectra and Molecular Structure. I. Spectra of
Diatomic Molecules, 2nd ed., Van Nostrand Reinhold Company, 1950.
[33] A.L. Petre, A. Auroux, P. Ge´lin, M. Caldararu, N.I. Ionescu, Thermochim.
Acta 379 (2001) 177.
[34] M. Haneda, E. Joubert, J.C. Me´ne´zo, D. Duprez, J. Barbier, N. Bion, M.
Daturi, J. Saussey, J.C. Lavalley, H. Hamada, Phys. Chem. Chem. Phys. 3
(2001) 1371.
In general, the calculated values of different kinetic param-
eters for both reversible (CO and CO2 on Pd/SiO2 and Ga2O3,
respectively) and reactive adsorption (CO on Ga2O3) following
the variable-temperature infrared spectroscopy approach were in
good agreement with those reported employing other classical
methods.
[35] J.I. Steinfeld, J.S. Francisco, W.L. Hase, Chemical Kinetics and Dynamics,
Prentice-Hall, New Jersey, 1989.
[36] H. Nakano, I. Nakamura, T. Fujitani, J. Nakamura, J. Phys. Chem. B 105
(2001) 1355.
Acknowledgements
[37] T. Yatsu, H. Nishimura, T. Fujitani, J. Nakamura, J. Catal. 191 (2000) 423.
[38] D.B. Clarke, B.-K. Lee, M.J. Sandoval, A.T. Bell, J. Catal. 150 (1994) 81.
This work was supported by the National Council for Sci-
entific and Technical Research (CONICET) and the National