Kinetics of reduction of supported nanoparticles of iron oxide

Article abstract of DOI:10.1007/s10973-007-8427-7

A series of catalysts prepared by dispersing iron oxide on supports of different nature and acidity has been studied. Silica (S) and silica-zirconia mixed oxides (SZ) with different ZrO₂ content (from 5 to 45 mass%) were used as supports for the iron oxide phase which was deposited over them by an equilibrium-adsorption method. The red-ox properties of the Fe-catalysts were studied by temperature programmed reduction (TPR) technique. The two well defined and narrow TPR peaks observed could be associated with the reduction steps: I) Fe₂O₃→FeO and Fe₂O ₃→Fe(0) (at ca. 400°C) and II) FeO→Fe(0) (at ca. 800-900°C). The temperature of the second-step-peak increased with the zirconia content in the support, likely because of the stronger interaction of the iron oxide phase with the support. Activation parameters for the two step-reduction processes were obtained by a simple computation procedure applied to the TPR profiles.

Full text of DOI:10.1007/s10973-007-8427-7

Journal of Thermal Analysis and Calorimetry, Vol. 91 (2008) 1, 93–100
KINETICS OF REDUCTION OF SUPPORTED NANOPARTICLES
OF IRON OXIDE
Claudia Messi1, P. Carniti1 and Antonella Gervasini1,2*
1Dipartimento di Chimica Fisica ed Elettrochimica, Universit´ degli Studi di Milano, via C. Golgi n. 19, 20133 Milano, Italy
2Centro di Eccellenza CIMAINA, Universit´ degli Studi di Milano, via C. Golgi n. 19, 20133 Milano, Italy
A series of catalysts prepared by dispersing iron oxide on supports of different nature and acidity has been studied. Silica (S) and
silica-zirconia mixed oxides (SZ) with different ZrO2 content (from 5 to 45 mass%) were used as supports for the iron oxide phase
which was deposited over them by an equilibrium-adsorption method. The red-ox properties of the Fe-catalysts were studied by
temperature programmed reduction (TPR) technique.
The two well defined and narrow TPR peaks observed could be associated with the reduction steps: I) Fe2O3®FeO and
Fe2O3®Fe(0) (at ca. 400°C) and II) FeO®Fe(0) (at ca. 800–900°C). The temperature of the second-step-peak increased with the
zirconia content in the support, likely because of the stronger interaction of the iron oxide phase with the support. Activation param-
eters for the two step-reduction processes were obtained by a simple computation procedure applied to the TPR profiles.
Keywords: iron oxide, kinetics of reduction, supported catalysts, temperature programmed reduction
Introduction
when an oxide is concerned as substrate. In this case,
the low number of exchangeable sites does not allow
to exploit the same preparation methods. The form-
ation of highly dispersed and nanosized iron oxide
aggregates on oxide supports is then a very difficult
task. For this purpose, different synthetic approaches
and Fe-precursors of the iron oxide phase have been
tested [15–18].
Thermal analysis techniques are widely used in
catalysis for various characterisation analyses of
catalytic solids and surfaces. Among these, temper-
ature programmed reduction (TPR) analysis has
found great success not only for studying the reducing
properties of active metal phases but also for
collecting information on the dispersion and aggre-
gation state of supported phases. Qualitative (shape
and position of TPR peaks) and quantitative (amount
of reducible gas consumed) data provide very useful
information on the metal phase in terms of nature and
dispersion. Therefore, TPR technique has been
recognized as an important tool in the research aimed
at characterizing both the bulk and dispersed metal
phases [1–7].
Recently, supported iron oxide catalysts have
received much attention because their potentiality for
many applications in environmental catalysis (N2O
decomposition and reduction) [8, 9] and in fine
chemical industry for reactions requiring strong
Lewis acid sites (Friedel-Crafts, isomerisations, etc.)
[10, 11]. For most catalytic applications, high disper-
sion of the metal centers is desirable to enhance the
activity-selectivity pattern of the catalyst. Conven-
tional methods used to disperse metal centers in silica
structures (zeolites, MCM, SBA, etc.) exploit the
ionic exchange properties of the support substrates
[12–14]. A completely different situation is present
In this work, we present the study performed on
a series of catalysts constituted of a supported iron
oxide phase over oxide supports of varying nature and
acidity. The supported iron phase was deposited by an
adsorption equilibrium method, already use for the
preparation of copper-based catalysts [19, 20]. It is
known that the properties of the dispersed phases are
dependent on the interaction with the relevant
support. The supported iron oxide phase can then
have some properties, in particular the red-ox pro-
perties and metal dispersion, influenced by the
support nature. TPR was chosen as thermal analysis
technique to investigate on the reducibility of the
variously supported iron oxide. The reducing profiles
of the iron-catalysts have been evaluated from quanti-
tative point of view and kinetically interpreted to
obtain the activation parameters of the reaction of
iron oxide reduction. Strong influence of the support
nature and properties have been observed on the path
followed by the dispersed iron oxide for the reduction
to metal iron.
*
Author for correspondence: antonella.gervasini@unimi.it
1388–6150/$20.00
Akadémiai Kiadó, Budapest, Hungary
Springer, Dordrecht, The Netherlands
© 2008 Akadémiai Kiadó, Budapest

MESSI et al.
Table 1 Catalyst presentation with their main properties
Composition/mass%
Sample code
Surface area/m2 g–1
Pore sizea/nm
SiO2
ZrO2
Fe2O3
Fe/S
92.74
89.44
80.21
65.56
52.62
84.51
0
7.26
5.85
5.63
6.35
4.33
11.04
239.2
237.4
231.5
298.1
237.5
224.0
7
4
4
4
4
4
Fe/SZ-5
4.71
Fe/SZ-15
Fe/SZ-30
Fe/SZ-45
Fe/SZ-5(w.i.)b
14.16
28.09
43.05
4.45
amain population of pore size calculated from B.J.H. [33], bcatalyst prepared by wetness impregnation (w.i.)
flowing, 45 mL min–1). A TPDRO–1100 (Thermo
Electron Corporation) equipped with a quartz reactor
with porous septum (ca. 8 mm i.d.) and a filter filled
with soda lime for trapping acid gases and water was
employed. TPR measurement was carried out using
H2/Ar (5.03% v/v) as reducing gas mixture. The gas
flow rate was adjusted by mass flow controllers to
15 cm3 min–1. A thermal conductivity detector (TCD)
measured the amount of H2 uptake. The obtained TPR
peaks were integrated for quantitative determination
of H2 consumed. Calibrations for the determination of
H2-consumption were performed by both CuO bulk
standard per analysis as reference material and pure
H2 (Sapio, Italy; purity >99.9999%) injections.
Conventional TPR runs forecast a first heating
step from 50 to 1000°C (typical heating rate,
8°C min–1) followed by an isothermal step at 1000°C
for 2 h. Sample size used (ca. 0.07 g sieved between
45–60 mesh) corresponded to about 30 mmol of
Fe2O3; this allowed to maintain K and P values
[21, 22] around 70 s and 10°C, respectively, for all the
analyses. Pure iron oxides (Fe2O3, Fe3O4 and FeO),
used as reference materials, were mixed with silica
sand to obtain mixture containing 4–6 mg of
reducible oxide in 100 mg of sand.
Experimental
Materials
Iron catalysts were prepared from Fe(III)-acetyl-
acetonate precursor, adsorbing it on the supports
suspended in a water/propanol solution. As supports,
a pure silica (S) and a series of silica-zirconia oxides
(SZ) (containing from 5 to 45 mass% of ZrO2)
prepared by sol-gel route were used. After filtration,
drying, and calcination (500°C), orange solids
containing Fe in an amount of 4–5 mass% were
obtained (Table 1). For comparative purpose, a
conventional Fe-catalyst was prepared by wetness
impregnation (w.i.) starting from Fe(NO3)3 precursor
and SZ-5 as support.
The iron content of the samples were measured
by ionic exchange chromatographic analysis after
solid dissolution in acid mixtures and reaction in a
microwave digestor at 220°C. The composition of
each sample is reported in Table 1, expressed as
mass% of Fe2O3, SiO2 and ZrO2 oxides.
Methods
Catalyst characterization for the determination of the
structure, by XRD, morphology, by N2 adsorp-
tion-desorption isotherms and SEM, electronic prop-
erties by DR-UV, EPR, and XPS spectroscopies, acid
properties, by TPD of basic probe, was performed and
results will be presented elsewhere [20].
Pure Fe2O3, Fe3O4, and FeO (Aldrich,
purity >99.9%) powders were used as reference
samples for the analyses of reduction under prog-
rammed heating rate.
Reduction experiments in isothermal and
temperature programmed heating conditions
Scheme 1 Thermal programme used for the reduction
experiments in isothermal conditions (from 300 to
600°C) followed by temperature programmed
analyses (from 50 to 1000°C)
Reduction analyses in programmed temperature
(TPR) and/or isothermal conditions were realized on
freshly oxidized samples (350°C for 1 h under air
94
J. Therm. Anal. Cal., 91, 2008

IRON OXIDE NANOPARTICLES
On selected samples, a series of isothermal
Temperature programmed reduction (TPR) analyses
reductions under H2/Ar flowing (5.03% v/v)
performed at defined temperature, starting from the
lowest at 200°C to the higher ones (300, 400, 500 and
600°C) each one maintained for 0.5 h, were carried
out. After each isothermal pre-reduction, the samples,
cooled to 50°C in inert flowing atmosphere (Ar,
20 mL min–1), were further reduced under program-
med temperature (TPR) from 50 to 1000°C, maintain-
ing the final temperature for 1 h only (Scheme 1). The
sample mass, and TPR conditions were analogous to
those above reported.
Much work has been performed on the reducibility of
bulk iron oxides under linear heating rate (TPR) and
isothermal conditions [24–26]. Analysis by TPR
furnishes the fingerprint of the reduction of each
oxide showing one or more peaks with/without
shoulders with well defined positions of the peak
maxima of reduction (Tmax). Moreover, quantitative
results on the amount of reductant (e.g., H2)
consumed during the reduction complement the infor-
mation for the determination of the iron oxide nature.
Reduction of pure hematite by hydrogen is a
complex event that can proceed in different steps via
intermediate oxides (i.e., magnetite, Fe3O4, and
wüstite, FeOx). The possible reduction reactions, in-
volving hematite and the other derived oxides, have
been written here below. The reduction to metal iron
can occur by a reaction sequence or by a one single re-
duction step; each step has a defined molar ratio
among the oxide, reductant, and product:
Kinetic computation
The conventional TPR profiles (TCD signal, mV g–1,
vs. analysis time, s) were divided into very small con-
stant intervals (Dt =1 s) which were integrated obtain-
ing the H2 consumption value for each interval. The
obtained H2 consumption values were summed each
other to obtain a curve describing the total H2 con-
sumed as a function of analysis time or temperature.
On each curve, the zones of exponential increase of the
H2 consumption vs. analysis time were numerically
treated to obtain the Arrhenius activation parameters.
(i)
Fe2O3 + 3H2®2Fe + 3H2O
3Fe2O3 + H2®2Fe3O4 + H2O
3Fe2O3 + 3H2®6FeO + 3H2O
Fe3O4 + 4H2®3Fe + 4H2O
Fe3O4 + H2®3FeO + H2O
(ii)
(iii)
(iv)
(v)
Results and discussion
The studied samples comprised an oxide support
(silica and mixed silica-zirconias with variable
amount of zirconia) and an iron oxide supported
phase. Composition and main morphologic properties
are listed in Table 1. The adsorption equilibrium
method used to deposit the iron phase on the supports
guaranteed the obtainment of highly dispersed iron
aggregates that, after calcination, gave dispersed iron
oxide nanophases. The diffractometric analysis
(XRD) did not show any line, indicating the total
amorphous character of both the support and iron
oxide phases and the presence of aggregates of iron
oxides of nanometric dimensions below the limit of
detection of XRD.
The spectrophotometric (XPS) and spectro-
scopic (DRS-UV) analyses [23] converged on the
identification of the nature of the iron phase. Over all
the surfaces, binding energy in the 710.9–711.7 eV
interval (Fe 2p3/2) was diagnostic of the Fe2O3
presence. UV-bands with maxima at ca. 400–500 and
330–350 nm, typical of nanosized Fe2O3 particles
with dimension around 5 nm and of lower nuclearity
Fe-oxide particles, were observed, respectively.
The catalyst prepared by impregnation gave well
visible, even if not much intense, diffraction peaks,
typical of the presence of Fe2O3 crystalline aggregates
(35.6, 44.2, 63.2° 2q) of about 130  of dimension.
(vi) (1–x)Fe3O4 + (1–4x)H2®3Fe(1–x)O + (1–4x) H2O
(vii)
FeO + H2®Fe + H2O
(viii)
Fe(1–x)O + H2®(1–x)Fe + H2O
Generally, the first step in the hematite reduction
is the formation of Fe3O4. Accordingly, observation
of a TPR peak at low temperature (the exact position
diverges to a large extent dependent on impurities)
corresponds to the Fe3O4 formation. Our experimental
results confirmed those of the literature; TPR of pure
Fe2O3 powder showed one first peak at low temper-
ature (370°C) followed by a principal peak with
maximum at 550°C. Reduction (TPR spectrum
integration gave 100% reduction) likely passed
through the (ii) and (iv) reduction steps. The Fe3O4 re-
duction profile gave one symmetrical peak with Tmax
at 550°C, while the reduction of FeO was observed at
very higher temperature (797°C). Quantitative
reductions could be calculated for Fe3O4 (step iv) and
FeO (step vii), too.
A very different and complex situation occurs
when supported iron phases were concerned. Comp-
lex reduction patterns are usually observed because of
the presence of mixtures of iron oxides, promoters,
different sizes of oxide aggregate, and development
of oxide-support interaction [27–30]. On such convo-
J. Therm. Anal. Cal., 91, 2008
95

MESSI et al.
Table 2 Quantitative results obtained from TPR analyses of the Fe-catalysts
H2 consumed/mmol g–1
I-TPR peak (Tmax,1)b/%
II-TPR peak
(Tmax,2)c/%
T
max,1/°C
T
max,2/°C
Sample code
Total
Calc.a
1364.8 1386.8
979.4 1080.2
Partial
Path A
Path B
b
c
Exp.
Tmax,1
Tmax,2
Fe2O3®FeO
Fe2O3® Fe
FeO®Fe
Fe/S
420
420
420
430
450
800
780
900
900
900
590.0
512.6
638.5
569.5
481.3
n.d.
819.2
550.8
432.9
534.1
353.8
n.d.
29.6
25.5
20.2
22.4
17.6
n.d.
9.8
59.1
51.1
40.4
44.8
35.4
n.d.
Fe/SZ-5
Fe/SZ-15
Fe/SZ-30
Fe/SZ-45
14.1
24.4
32.8
46.9
n.d.
1058.1 1071.2
1192.5 1103.4
1001.9
824.4
Fe/SZ-5(w.i.) 371–510 671–785 2073.5 2397.8
a Fe2O3 + 3H2®2Fe + 3H2O; b Fe2O3+H2®2FeO (path A) plus Fe2O3+3H2®Fe+3H2O (path B); c FeO+H2®Fe+H2O
luted profiles, the identification of the reduction path
followed by iron oxide is not an easy task. The TPR
profile of the catalyst prepared by impregnation,
Fe/SZ-5(w.i.), gave a convoluted curve (Fig. 1) that
resembled those reported in the literature for the
shape of the curve and temperatures of the reduction
peaks [27, 28]. Comparing the reduction peak
temperatures (Table 2) with those observed for the
iron oxide standard samples, it is possible to guess the
reduction path followed by the iron oxide phase. The
reduction of Fe2O3 to Fe(0) proceeded through Fe3O4
(Tmax at 371 and 510°C) and FeO (Tmax at 785°C)
intermediates. Because FeO should not be present on
the fresh surface of Fe/SZ-5(w.i.), it had to be formed
from a more oxidized Fe-phase. Reduction steps (iii)
and (v) are relevant to Fe2O3 and Fe3O4 reduction to
FeO; one of these reactions could be responsible of
the TPR peak observed at 671°C. The complexity of
the TPR profile of Fe/SZ-5(w.i.) reveals the high
heterogeneity of the iron surface in terms of nature of
the oxides, dimension of oxide aggregates, and
oxide-support interaction.
and high (780<Tmax,2/°C<900°C) temperatures. The
position of the low temperature peak was lightly
affected by the nature of the support; only a light shift
at higher temperature can be noticed by increasing the
zirconia content in the support (Tmax,1 of Fe/SZ-30 and
Fe/SZ-45, Table 2). The position of the high
temperature peak (Tmax,2) was comprised in a large
interval and the value depended on the support nature.
For the catalysts prepared on silica and on
silica-zirconia at lowest zirconia content (Fe/S and
Fe/SZ-5), Tmax,2 was around 800°C. On the contrary,
for the catalysts prepared on the supports with higher
zirconia content (Fe/SZ-15, Fe/SZ-30 and Fe/SZ-45),
Tmax,2 was around 900°C. Integration of the TPR
spectra indicated complete reduction of iron oxide
(Fe2O3 to Fe) in any case, as shown by the comparison
between the calculated and experimental H2
consumed (Table 2), but Fe/SZ-45 for which a
reduction of 82% could be calculated. The two well
defined and sharp reduction peaks of the Fe-catalysts
prepared by the adsorption method suggest that iron
oxide aggregates were of very smaller dimension than
those of impregnated catalyst. Moreover, the shift at
higher temperatures of the FeO reduction peak
suggests that a strong metal oxide-support interaction
was active when zirconia was concerned.
A completely different situation emerges when
TPR profiles of the Fe-catalysts prepared by
adsorption method are examined (Fig. 2). All the
spectra are dominated by two well defined and
intense peaks positioned at low (420<Tmax,1/°C<450)
Fig. 1 TPR profile of Fe/SZ-5(w.i.) sample collected at
8°C min–1 from 50 to 1000°C under H2/Ar (5% v/v)
flowing
Scheme 2 Proposed reduction path for the supported iron
species prepared by equilibrium adsorption method
according to TPR results
96
J. Therm. Anal. Cal., 91, 2008

IRON OXIDE NANOPARTICLES
Table 3 Quantitative results of the reduction experiments in isothermal conditions followed by temperature programmed (TPR)
analyses of the Fe/S and Fe/SZ-30 samples
Isothermal reduction
TPR
S(isothermal+ TPR)
Sample
Fe/S
T/°C
Red./%
T
max1/°C
T
max2/°C Red.Tmax,1/% Red.Tmax,2/% Red./%
Red./%
200
300
400
500
600
<2
420
420
420
440
460
780
780
780
780
780
52.0
47.5
37.5
22.7
9.1
48.0
53.4
62.5
77.3
90.9
93.7
97.3
73.4
57.2
50.8
93.7
102.2
95.5
97.6
96.0
3.9
22.0
40.4
45.2
300
400
500
4.2
420
420
420
900
900
920
53.2
41.4
19.3
46.8
58.6
80.7
97.6
74.4
57.5
101.4
96.7
22.3
45.5
Fe/SZ-30
102.9
600
51.7
410
920
18.9
81.1
56.0
107.7
Separate quantification of the H2-consumption
of the two TPR peaks gave information on the reduc-
tion path of the supported iron oxide, under the as-
sumption that Fe2O3 was the sole phase present in the
catalyst samples. A simple-two-step-reduction path
(Fe2O3®FeO®Fe, reduction steps iii and vii) did not
hold, on the basis of the quantitative data of H2 con-
sumption and the amount of iron present in the cata-
lysts. The high temperature reduction peak (II-TPR
peak) could be doubtless assigned to the FeO reduc-
tion (step vii), due to the Tmax,2 value. From the H2
consumption of II-TPR peak, it was possible to know
the amount of FeO reduced and that of the parent
Fe2O3 phase (step iii). From the stoichiometric com-
putations, it was found that the amount of Fe2O3 pres-
ent in each sample was superior than that necessary
for the FeO formation. This signified that a part of
Fe2O3 should be reduced by a different path, likely
leading directly to metallic iron from hematite phase
(step i). Scheme 2 reports our proposed reduction path
for the supported iron oxide phase on the basis of the
quantitative evaluation of H2-consumption. In this
case, Fe3O4 formation has not been clearly detected,
as in the case of the impregnated Fe-catalyst (pres-
ence of the peak at 370°C). An inspection of Table 2
reveals that the II-TPR peak regularly decreased with
the increase of zirconia content in the support. In par-
allel, the direct formation of metallic iron from hema-
tite (path B of the I-TPR peak) regularly increased
Fig. 3 Reduction experiments in isothermal conditions (from
300 to 600°C) (left side) followed by temperature
programmed analyses (from 50 to 1000°C) (right side)
for Fe/S and Fe/SZ-30 samples
Fig. 2 TPR profiles of A – Fe/S; B – Fe/SZ-5; C – Fe/SZ-15;
D – Fe/SZ-30; E – Fe/SZ-45 samples collected at 8°C
min–1 from 50 to 1000°C under H2/Ar (5% v/v) flowing
J. Therm. Anal. Cal., 91, 2008
97

MESSI et al.
Table 4 Kinetic results obtained from the TPR analyses of the Fe-catalysts (heating rate, 8°C min–1)
I-TPR peak
Ea/kJ mol–1
II-TPR peak
Sample
code
ln A/s
r
k(450°C)/s
ln A/s–1
Ea/kJ mol–1
r
k(850°C)/s–1
Fe/S
1.853
2.932
4.494
5.249
3.195
47.7
52.3
62.0
65.4
57.4
0.999
0.999
0.999
0.999
0.999
0.0023
0.0032
0.0030
0.0035
0.0017
27.542
44.148
42.671
89.943
95.612
292.7
428.0
466.6
913.7
977.2
0.993
0.999
0.996
0.998
0.996
0.0223
0.1851
Fe/SZ-5
Fe/SZ-15a
Fe/SZ-30
Fe/SZ-45
0.000682
0.000372
0.000119
aEa values of 60.6 (lnA of 4.138) and 65.5 (lnA of 5.173) kJ mol–1 for heating rate of 5 and 12°C min–1, respectively for the I-TPR peak
with the zirconia content in the support. It can be in-
ferred that support surfaces richer in zirconia pro-
moted the Fe2O3 reduction by a different path that sil-
ica without the intermediate formation of the
metastable FeO3 phase.
For an in depth study of the reducibility of the
supported iron phase, other reduction experiments
were carried out on selected samples. Fe/S and
Fe/SZ-30 were chosen as representative samples due
to the very large difference in the support compo-
sition. After complete oxidation of their surfaces, they
were first isothermally reduced in the temperature
range from 200 to 600°C, then cooled to 50°C and
reduced again with programmed heating rate (TPR) to
1000°C (Scheme 1). When the isothermal temper-
ature of reduction was low (200 and 300°C), the TPR
spectra obtained were very similar to those of Fig. 2.
For isothermal reductions at higher temperatures
(400–600°C), besides a well evident pre-reduction
(Fig. 3 (left side) and Table 3), modified TPR spectra
were obtained. The I-TPR peak regularly decreased
with increasing isothermal reduction temperature, as
expected (Table 3). It is interesting to note that the
calculated amount of FeO reduced in the II-TPR peak,
deriving from the parent Fe2O3 phase, was constant in
all the experiments (around 47% and 45% for Fe/S
and Fe/SZ-30, respectively). This indicated that the
formation of the metastable FeO phase was not
affected by the thermal history of the sample but it
depended on the nature of the support on which the
parent Fe2O3 was stabilized. This conclusion is in
contrast with the literature results indicating that the
reduction path and intermediate formed during bulk
Fe2O3 reduction are dependent on reduction
temperature [24]. In the case of the supported Fe2O3
phase, the development of metal-support interaction
predominates over the massive oxide interaction.
Kinetic analysis
The kinetics of reduction of supported iron oxides is
an important point that deserves investigation. There
are numerous models in the literature for the estima-
tion of the kinetic parameters for the reduction of
oxidic species based on nucleation model or contract-
ing sphere model [1, 2, 27, 30–32]. Some of them
make use of the well-known correlation describing
the shift in Tmax with the heating rate while other are
based on peak-shape analysis.
In this study, the kinetic analysis to obtain the
activation parameters was performed applying a
classical Arrhenius plot by using a peak-shape
analysis of the data reported in Fig. 2.
The reduction reaction of the supported Fe2O3
phase can be described by the rate equation of the type:
Fig. 4 Examples of TPR peak integration; the obtained curve
of H2 consumption as a function of analysis time (right
axis) is drawn
rate =k [Fe2O3]
(1)
98
J. Therm. Anal. Cal., 91, 2008

MESSI et al.
very large interval passing from 293 (Fe/S) to 977
(Fe/SZ-45) kJ mol–1. Once again, the influence of the
zirconia content in the support affects the reducibility
of the iron oxide-phase (FeO). Likely on zirconia, the
FeO interaction is stronger than on SiO2 and higher
activation energy is necessary for its reduction.
Conclusions
Temperature programmed reduction was successfully
used for the study of the reducing properties of
supported iron oxide phases. The reducing path of
hematite was found to be dependent on the support;
acidic oxide (zirconia) favours complete reduction of
hematite to metallic iron while on more neutral oxide
(silica) hematite preferentially reduced through
wüstite formation. The strong interaction between
iron oxide and zirconia led to a shift to high temper-
ature of the FeO reduction, very high activation
energy values for this reduction were observed.
Fig. 5 Arrhenius plot for the I-TPR peak reduction of the
Fe-catalysts
where [Fe2O3] is the actual concentration at any reac-
tion time and k is the rate constant. Equation (1)
makes use of first order with respect to reducible solid
phase and zero order in H2 concentration, in agree-
ment with some authors [2].
The classical Arrhenius dependence on tempera-
ture for k can hold,
ln k = ln A–Ea/RT
(2)
References
where Ea is the activation energy and A the Arrhenius
pre-exponential constant.
1 N. Hurst, S. J. Gentry, A. Jones and B. D. McNicol, Catal.
Rev.-Sci. Eng., 24 (1982) 233.
Firstly, a graphical procedure was made, as
shown in Fig. 4, as an example. The TPR spectrum
was divided in small portions (Dt =1 s) that were
integrated to obtain the amount of H2 consumed in
each time interval. Summing all the integrated
portions, the curve of total H2 consumption as a
function of reaction time was obtained. In the time
interval correspondent to the ascending I-TPR and
II-TPR peaks, section of exponential curves could be
observed (Fig. 4). For each point of these portions of
curve, the rate was evaluated and the rate constant
was calculated by Eq. (1). Because each calculated
rate constant was related to a defined temperature, the
classical Eq. (2) was applied to obtain the activation
parameters. Arrhenius plot for the I-TPR peaks of all
the catalysts is shown in Fig. 5. The obtained kinetic
results are collected in Table 4. In the Table are also
reported the rate constants calculated at 450°C (for
the I-TPR peak) and at 850°C (for the II-TPR peak).
The obtained kinetic parameters are not affected by
the condition used in the TPR analysis. On Fe/SZ-15,
TPR data were collected at 5, 8 and 12°C min–1 of
heating rate and the values of the calculated acti-
vation parameters did not vary significantly.
2 S. Bhatia, J. Beltramini and D. D. Do, Catal. Today,
7 (1990) 309.
3 M. Fadoni and L. Lucarelli, Stud. Surf. Sci. Catal., Vol.
120A (Adsorption and its Applications in Industry and
Environmental Protection, Vol. 1), Ed. A. Dabrowski,
Elsevier, Amsterdam, 1999, pp. 177–225.
4 J. W. Jenkins, B. D. McNicol and S. D. Robertson,
Chemtech, (1977) 316.
5 S. Subramanian, Platinum Metal Rev., 36 (1992) 98.
6 F. Kapteijn, J. A. Moulijn and A. Tarfaoui, Stud. Surf. Sci.
Catal., Vol. 123 (Catalysis: an integrated approach),
Elsevier, Amsterdam, pp. 525–541.
7 B. Jouguet, A. Gervasini and A. Auroux, Chem. Eng.
Technol., 18 (1995) 243.
8 V. Boissel, S. Tahir and C. A. Koh, Appl. Catal. B,
64 (2006) 234.
9 F. J. Perez-Alonso, I. Melián-Cabrera, M. López
Granados, F. Kapteijn and J. L. G. Fierro, J. Catal.,
239 (2006) 340.
10 G. Neri, G. Rizzo, S. Galvagno, G. Loiacono, A. Donato,
M. G. Mugolino, R. Pietropaolo and E. Rombi, Appl.
Catal. A, 274 (2004) 243.
11 Y. Sun, S. Walspurger, J. P. Tessonnier, B. Louis and
J. Sommer, Appl. Catal. A, 300 (2006) 1.
12 J. C. Groen, A. Brückner, E. Berrier, L. Maldonado,
J. A. Moulijn and J. Pérez-Ramírez, J. Catal.,
243 (2006) 212.
The Ea values for the I-TPR peak are comprised
in a narrow interval (48–65 kJ mol–1) with mean value
of 57 kJ mol–1. This suggests that there is not a great
difference in the dimension and aggregation state of
the Fe2O3 phase among the different samples. On the
contrary, the Ea values of the II-TPR peak vary in a
13 T. Nobukawa, M. Yoshida, K. Okumura, K. Tomishige
and K. Kunimori, J. Catal., 229 (2005) 374.
14 M. Yoshida, T. Nobukawa, S. Ito, K. Tomishige and
K. Kunimori, J. Catal., 223 (2004) 454.
J. Therm. Anal. Cal., 91, 2008
99

MESSI et al.
15 F. Arena, G. Gatti, G. Martra, S. Coluccia, L. Stievano,
L. Spadaro, P. Famulari and A. Parmaliana, J. Catal.,
231 (2005) 365.
25 E. E. Unmuth, L. H. Schwartz and J. B. Butt, J. Catal.,
63 (1980) 404.
26 H.-Y. Lin, Y.-W. Chen and C. Li, Thermochim. Acta,
400 (2003) 61.
16 R. S. Prakasham, G. Sarala Devi, K. Raiya Laxmi and
Ch. Subba Rao, J. Phys. Chem. C, 111 (2007) 3842.
17 A. Br²ileanu, M. R²ileanu, M. Criêan, D. Criêan,
R. Bîrjega, V. E. Marinescu, J. Madarász and G. Pokol,
J. Therm. Anal. Cal., 88 (2007) 163.
27 T. Herranz, S. Rojas, F. J. Pérez-Alonso, M. Ojeda,
P. Terreros and J. L. G. Fierro, Appl. Catal. A,
308 (2006) 19.
28 K. Chen and Q. Yan, Appl. Catal. A., 158 (1997) 215.
29 H. Hayashi, L. Z. Chen, T. Tago, M. Kishida and
K. Wakabayashi, Appl. Catal. A, 231 (2002) 81.
30 K. D. Chen, Y. N. Fan, Z. Hu and Q. J. Yan, Catal. Lett.,
36 (1996) 139.
18 M. Ÿtef²nescu, O. Ÿtef²nescu, M. Stoia and C. Lazau,
J. Therm. Anal. Cal., 88 (2007) 27.
19 S. Bennici, A. Auroux, C. Guimon and A. Gervasini,
Chem. Mater., 18 (2006) 3641.
20 S. Bennici, A. Gervasini and V. Ragaini, Ultrason.
Sonochem., 10 (2003) 61.
31 S. J. Gentry, N. W. Hurst, and A. Jones, J. Chem. Soc.
Faraday Trans. I, 77 (1981) 603.
21 D. A. Monti and A. Baiker, J. Catal., 83 (1983) 323.
22 P. Malet and A. Caballero, J. Chem. Soc., Faraday Trans.
I, 84 (1988) 2369.
32 A. L. Boyce, S. R. Granville, P. A. Sermon and
M. S. W. Vong, React. Kinet. Catal. Lett., 44 (1991) 1.
33 E. P. Barrett, L. G. Joyner and P. Halenda, J. Am. Chem.
Soc., 73 (1951) 373.
23 A. Gervasini, C. Messi, A. Ponti and N. Ravasio, Nano
Lett., submitted.
24 A. Pineau, N. Kanari and I. Gaballah, Thermochim. Acta,
447 (2006) 89.
DOI: 10.1007/s10973-007-8427-7
100
J. Therm. Anal. Cal., 91, 2008