Applied Catalysis A: General 511 (2016) 141–148

Contents lists available at ScienceDirect

Applied Catalysis A: General

journal homepage: www.elsevier.com/locate/apcata

Ethanol reactivity over La1+x FeO3+ ı perovskites

Guillaume Tesquet a , Jérémy Faye a,b , Fadime Hosoglu a , Anne-Sophie Mamede a,c ,
Franck Dumeignil a,d , Mickaël Capron a,∗
a
Université Lille 1 Sciences et Technologies, UCCS, Unité de Catalyse et Chimie du Solide, UMR CNRS 8181, 59655 Villeneuve d’Ascq Cedex, France
b
TEAMCAT Solutions SAS, Ecole Centrale de Lille, Cité Scientifique-CS 20048, 59655 Villeneuve d’Ascq Cedex, France
c
Ecole Nationale Supérieure de Chimie de Lille, 59650 Villeneuve d’Ascq, France
d
Institut Universitaire de France, Maison des Universités, 103 Boulevard Saint-Michel, 75005 Paris, France

a r t i c l e i n f o a b s t r a c t

Article history: Mixed oxide catalysts composed of mixtures of La1+x FeO3+ı perovskite and La2 O3 were prepared using
Received 23 July 2015 the auto-combustion method with glycine as an ignition promoter. A series of samples was prepared
Received in revised form with x = 0, 0.1, 0.3, 0.5, 0.8 and 1, which were hereafter called La1+x Fe. The as-prepared powders were
23 November 2015
characterized by XRD, XPS, DTA-TGA and N2 -physisorption. Their acid-base properties were evaluated
Accepted 3 December 2015
from their 2-propanol reactivity. Surface enrichment in La2 O3 was detected when the lanthanum con-
Available online 8 December 2015
tent increased. For high lanthanum concentrations, lanthanum hydroxide was also formed. The highest
ethanol conversion was observed over La1.3 Fe (32% at 400 ◦ C), which contained the largest amount of basic
Keywords:
Ethanol conversion
sites. The main reaction products were ethylene, acetone, 2-pentanone and 1-butanol. Other minor prod-
Mixed oxides ucts originating from dehydration, reverse aldolization and hydrogenation reactions were also detected.
Perovskite Such a diversity of products was due to the presence of different kinds and distribution of sites on the
XPS catalysts surface (acid, base, hydrogenation, etc.). The catalysts activity was depending on their acid-base
Basicity properties: an increase in the basic character due to the La2 O3 presence at the surface leaded to higher
selectivities to 1-butanol and 2-pentanone, which were 6 and 31% at 400 ◦ C, respectively, over the sample
with the highest number of basic sites, which were also stronger (La1.3 Fe). On the other hand, over La1.8 Fe,
which contained the highest quantity of La(OH)3, lower selectivities to 1-butanol and 2-pentanone (4.3
and 22%, respectively) were observed due to a decrease in the number of basic sites, together with a high
ethylene selectivity (37%), with thus a predominant action of acid sites.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction
Hundreds of molecules are issued from oil processing. These
chemicals are used in several sectors such as textile industry,
food-industry, health, transport, chemistry, etc. [1]. However, the
progressive depletion of oil resources together with an increasing
demand in countries like China, Brazil and India [2] lead to the
development of new technologies based on alternative resources,
such as biomass.
Biomass can be pre-processed to yield platform molecules such
as furfural, various acids or alcohols such as glycerol or ethanol. The
ethanol global production has reached 89 billion of liters in 2011.
It was mainly produced in the United States and in Brazil, which
represent together about 87% of the global production [3]. Ethanol
is used in various fields: in organic chemistry as a solvent, in the
∗ Corresponding author. Fax: +33 3 20 43 65 61.
E-mail address: mickael.capron@univ-lille1.fr (M. Capron).
cosmetic industry and in paints as a dissolvent, but the main appli-
cation, which is driving the market, consists in its use as a biofuel
[4]. Among the molecules that can be obtained from ethanol, 1-
butanol is of particular interest as a solvent in organic chemistry,
it possesses various applications in the cosmetic industry, and can
be further used as an additive in gasoline and diesel [5,6] thanks
to a high octane index and a low saturation vapor pressure com-
pared to that of ethanol [4]. It can be synthesized from ethanol
via the Guerbet reaction, which was discovered in the 19th cen-
tury by Marcel Guerbet [7] and is recognized as a useful synthetic
tool to obtain dimer alcohol by a self-condensation of primary
alcohol [8]. It is often considered as a multi-step reaction (Fig. 1,
given in the case of ethanol as a raw material): (i) dehydrogenation
of ethanol over metallic or basic sites (weak or medium) leading
to acetaldehyde;(ii) then, aldolization between two molecules of
as-formed acetaldehyde on a strong basic site to obtain the associ-
ated aldol (i.e., 3-hydroxy-butyraldehyde); (iii) Elimination of one
water molecule on acid sites to form an ␣,␤-unsaturated alde-
hyde. Finally, (iv) double hydrogenation of this aldehyde occurs

http://dx.doi.org/10.1016/j.apcata.2015.12.005
0926-860X/© 2015 Elsevier B.V. All rights reserved.
142 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148

B- (Ca10 (PO4 )6 (OH)2 ). Authors have proposed a mechanism showing

- H2
OH O O-
- H2 (ii)
O O-
(i)
OH O
(ii) B- H+
+ 2H2 - H2O
(iv) (iii) O OH
OH O an optimal acid/base ratio equal to 5, which enabled to reach a
unsaturated Aldol
aldehyde
Fig. 1. Guerbet reaction pathway with ethanol as a reactant.
to produce 1-butanol, requiring a hydrogenation or Lewis acid site
function on the catalyst.
Due to the complexity of the reaction pathway and to the
diversity of catalytic sites needed to realize this reaction, many
byproducts, such as acetone, ethylene, ethane, ester, 2-pentanone
can be formed from ethanol. However, the Guerbet reaction is
not the only reaction pathway to obtain butanol from ethanol.
Scalbert et al. [9] have also evidenced 2 other different possible
routes. After the first dehydrogenation step, which is the same for
both alternative reaction pathways, one molecule of ethanol and
one molecule of acetaldehyde react to form butenol or butane-
diol. In the case of the butanediol formation, this molecule looses
a water molecule to obtain the butenol. The last step of both alter-
native routes is a reaction between ethanol and butenol to yield
butanol and acetaldehyde via a hydride transfer according to a
Meerwein–Pondorf–Verley mechanism.
The Guerbet reaction has been realized over heterogeneous
catalysts [10], in the liquid phase [11–14] and in the gas phase
[15] using alcohol mixtures or alcohols heavier than ethanol as
reactants. Only a few studies specifically report the conversion
of ethanol, mainly over hydrotalcite [16] or over hydroxyapatites
(HAP) [17–20].
Hosoglu et al. [16] have synthesized hydrotalcite-type Mg6-x Cux
Al2 (OH)16 CO3 ·4H2 O catalysts (x = 0; 1; 3; 5), in order to evaluate
the influence of the copper introduction in the hydrotalcite struc-
ture. Such substitution led to an important increase in the ethanol
conversion at 300 ◦ C, which reached 80% for x of 5. However, this
type of catalysts was selective to acetaldehyde, with a selectivity of
almost 60%, but not in 1-butanol, which was detected in only very
low quantities.
Tushida et al. [18] have reached a maximum yield to
butanol with an ethanol conversion equal to 20% and a selec-
tivity to 1-butanol of 70% at 370 ◦ C using stoichiometric HAP
interactions between ethanol and the various active sites of HAP
(acidic and basic sites of various strength and nature) during the
Guerbet reaction. Then, Ogo et al. [19] have improved the selectiv-
ity in 1-butanol (i.e., 75% for 20% of ethanol conversion) substituting
partially calcium by strontium (Ca5 Sr5 (PO4 )6 (OH)2 ). This small 1-
butanol selectivity increase has been attributed to the presence of
residual sodium, which was included in the strontium precursor,
highlighting the benefit effect on butanol formation when the num-
ber of basic sites is well tune. Finally, Silvester et al. [17,20] have
prepared several HAP, which present various acid/base ratio. The
goal was to understand finely the influence of acidic and basic site
on ethanol reactivity and butanol production. They have evidenced
maximum butanol and other heavier alcohols selectivity via Guer-
bet reaction, and minimize the others side reactions. MgO has been
also tested to convert ethanol into 1-butanol through the Guerbet
reaction [21]. Irrespective of the catalytic system, the best yield
reported so far does not exceed 25%.
The guideline followed in this work was to design multi-
functional catalytic materials to carry out a one-pot reaction from
ethanol. To reach this goal, we have chosen to use perovskite-type
materials. Indeed, perovskites present a wide range of modularity
owing to their general formula LnMO3 , where Ln corresponds to an
element of the lanthanide group, and M a transition metal [22,23].
The elementary lattice of the perovskite structure is composed
of MO6 octahedrons linked by theirs corners. These octahedrons
form a tri-dimensional framework and the lanthanide element is
located in the cavity induced by the framework. The perovskite
structure is obtained if the electroneutrality and the tolerance fac-
tor are respected [24]. Perovskites have already been often used in
catalytic reactions for their oxidant character [25–31], but also for
their reductive properties [32–34].
In this study, non-stoichiometric lanthanum-iron based per-
ovskites were synthesized to investigate the influence of the
lanthanum excess, in particular to increase the basic properties of
the material, which seem to be the key parameter for the Guerbet
reaction.
2. Experimental
2.1. Catalysts synthesis
A series of mixed oxides of the La1+x FeO3+ı general formula
(with x = 0, 0.1, 0.3, 0.5, 0.8 and 1) was prepared using the so-called
auto-combustion method [25,35–37]. This preparation procedure
was selected due to its quickness compared with common methods
(e.g., co-precipitation or freeze drying [33,38–40]), which further
enables obtaining larger specific surface areas.
Lanthanum and iron nitrates precursors [La(NO3 )3 ·6H2 O, purity
99.99%; Fe(NO3 )3 ·9H2 O, purity 99.99%; both from Sigma–Aldrich]
were dissolved in a minimum quantity of water. Then, glycine

Table 1
Elemental composition and physical properties of the La1-x Fe catalysts, Binding energies and relative abundances of the elements constituting the La1+x FeO3+ı samples. and
atomic distribution in oxygen-containing surface species obtained from the O1s spectral fitting.

Sample SSA (m2/g) Ratio La/Fea Surface composition O1s spectral fitting (at.%)b

Th ICP XPS C1s O1s La3d5/2 Fe2p3/2 OLaFeO3 OLa2 O3 OLa(OH) 2− OHz O
3 +CO3

BE (eV) at.% BE (eV) at.% BE (eV) at.% BE (eV) at.%

LaFe 6 1 1.02 2.04 285.0 30.6 529.3 46.7 834.0 15.0 710.0 7.3 60.0 – 34.2 5.8
La1.1 Fe 10 1.1 1.11 3.26 285.0 28.2 529.1 49.3 834.4 17.3 710.1 5.3 42.2 3.5 50.8 3.5
La1.3 Fe 9 1.3 1.31 4.36 285.0 23.7 529.0 52.8 834.6 19.2 710.4 4.4 30.3 9.1 59.3 1.3
La1.5Fe 9 1.5 1.50 5.70 285.0 23.5 528.9 53.8 834.5 19.4 709.9 3.4 22.7 7.6 68.5 1.3
La1.8Fe 12 1.8 1.80 6.30 285.0 23.9 528.9 54.2 834.1 18.9 709.9 3.0 17.2 6.8 74.8 1.3
La2Fe 21 2 2.01 6.97 285.0 21.2 529.0 55.7 834.6 20.2 710.0 2.9 17.4 12.1 69.6 1.0
 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148 143

(NH2 CH2 COOH, pure, Sigma–Aldrich) was added in a stoichiomet-

ric amount relative to the nitrate precursors quantities. Once the
dissolution process was complete, the solution was heated to 120 ◦ C
in order to evaporate water until obtaining a viscous solution.
Then, the temperature was increased to 370 ◦ C to provide the ther-
mal energy necessary to initiate the auto-combustion. The glycine
enabled the decomposition of the reactants owing to the exother-
micity of its combustion, and leading to the formation of a light
brown powder. In order to completely remove the carbonaceous
residues originating from incomplete combustion of glycine, a final
calcination step was applied at 600 ◦ C (heating rate = 2 ◦ C/min) dur-
ing 8 h under air (flow = 1.8 L/h). Finally, the catalysts were pelleted
and grinded to obtain particles sizes between 80 and 100 ␮m.
All the as-obtained solids were labeled La1+x Fe according to the
theoretical lanthanum stoichiometry.
Fig. 2. X-ray diffractograms of the La1+x Fe catalysts.
2.2. Characterization

Structural analysis was performed by X-ray Diffraction using a 100 18

BRUKER AXS D8 Advanced diffractometer (CuK␣1 = 1.5406 Å). The 90 16

% wt La2O3 and La(OH)3

diffractograms were recorded between 20◦ and 90◦ , with a step
80 LaFeO3 La2O3 14
of 0.04◦ and an acquisition time of 0.04 s. The phase’s identifica-

tion was performed by comparison with powder diffraction profiles
provided by the ICDD.
The specific surface areas (SSAs) were determined by nitrogen
adsorption–desorption using a Micromeritics FlowSorb III appara-
tus. 200 mg of pelleted and sieved catalyst (size comprised between
80 and 100 ␮m) were introduced in the cell in which they were out-
gassed at 150 ◦ C during 30 min under a gas mixture composed of
30% N2 in helium. Then, the cell was immersed in liquid nitrogen to
perform the adsorption, and then in a water bath at room temper-
ature to desorb N2 . The quantification was realized with a thermal
conductivity detector,TCD, calibrated beforehand.
X-ray Photoelectron Spectroscopy (XPS) experiments were car-
ried out on an AXIS Ultra DLD Kratos spectrometer equipped
with a monochromatized aluminum source with a wavelength at
1486.6 eV. All the spectra were collected with a constant analy-
sis energy mode (CAE = 40 eV) and corrected using the C1s binding
energy at 285 eV as a reference. The C1s, O1s, Fe2p and La3d binding
energy regions were analyzed and the background was subtracted
using a Shirley type curve [41]. The spectra were fitted using the
Casa XPS software.
In order to determine the acidity and the basicity of the catalytic
materials, a 2-propanol test was performed. A mixture composed of
100 mg of catalyst and 100 mg of carborundum was introduced in
a fixed-bed reactor. The 2-propanol concentration was controlled
with an evaporator/saturator device and tuned to obtain a ratio
2-propanol/helium = 4.5/95.5 (vol%) with a helium flow fixed to
25 ml/min. The tested temperatures were 250, 300 and 350 ◦ C to
keep a low 2-propanol conversion and a reaction order equal to
1.The products were analyzed using an Alpha MOS PR2100 gas
chromatograph equipped with an HP-INNOWAX column (L = 30 m;
ID = 0.25 mm; film thickness = 0.5 ␮m) connected to a FID.
2.3. Catalytic tests
The activity of the catalysts was evaluated using a Multi-
R® device from Teamcat Solutions. Each fixed-bed reactor of the
parallel equipment was filled with a mixture of 200 mg of catalyst
at 100 ␮m and 200 mg of SiC (particles size of 125 ␮m). 500 mg of
SiC was added above the catalytic bed to favor the homogeniza-
tion of the flow composition (downflow reactor). A high pressure
pump injected the desired flow of ethanol in an evaporator main-
tained at 120 ◦ C along with helium (40 ml /min). The composition
of the mixture was fixed to obtain an ethanol/helium ratio equal to
13/87 (vol%), corresponding to a GHSV of 1125 h−1 . The unreacted
% wt LaFeO3
70 La(OH)3 12
60
10
50
8
40
6
30
20 4
10 2
0 0
LaFe La1.1Fe La1.3Fe La1.5Fe La1.8Fe La2Fe
Fig. 3. Semi-quantitative analysis of the different phases obtained from XRD results.
ethanol and products were analyzed on-line using an Agilent 7890A
gas chromatograph. The organics were separated on a Zebron Phe-
nomenex column (L = 30 m, ID = 0.25 mm, film thickness = 1 ␮m)
and quantified using a FID, whereas a HP-PLOT/Q Agilent column
(L = 30 m, ID = 0.32 mm, film thickness = 20 ␮m) connected to a TCD
was used for light molecules quantification. The tested tempera-
tures were 300, 350 and 400 ◦ C, and the results presented in this
paper correspond to the average of five analyses once the steady
state has been reached. The steady state was typically reached after
3 h of reaction at a given temperature.
3. Results and discussion
3.1. Structural analysis and chemical composition
The diffractograms of the catalysts after calcination are reported
in Fig. 2. For x = 0, the major crystallized phase was the LaFeO3
perovskite (JCPDS # 01-070-7777) with an orthorhombic sym-
metry, but the presence of La2 O3 cannot be excluded even if its
quantity cannot be evaluated. The characteristic peaks of LaFeO3
were present in all the samples. Increasing the quantity of La led
to the appearance of a second crystalline phase, La2 O3 (JCPDS #
01-074-2430). The intensity of the main peak of La2 O3 , located
at 29.9◦ increased until x = 0.3.When the value of x was superior
or equal to 0.5, a third phase appeared, identified as lanthanum
hydroxide La(OH)3 (JCPDS # 00-036-1841). This phenomenon was
then amplified with the increase of x. Thanks to the X-ray diffrac-
tograms (semi-quantitative analysis), we quantified the proportion
of the different crystalline phases (Fig. 3).
144 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148
These results are in agreement with the literature [42] in
which the crystallization of the oxide in the perovskite structure
is observed when x = 0, with further appearance of the La2 O3 and
La(OH)3 phases upon introduction of a lanthanum excess.
The bulk chemical composition of the catalysts was determined
by ICP analysis (Table 1). We observed a good correlation between
the ICP La/FeICP ratio and the theoretical La/FeTh ratio.
3.2. Textural properties
The SSAs are reported in Table 1. They were comprised between
6 and 21 m2 /g. The SSA remained stable at ≈9 m2 /g except for the
last sample of the series. For x equal to 0.8, the SSA was significantly
higher (i.e., twice larger). This increase may be correlated with an
increase in the quantity of the La2 O3 phase, which is well known to
develop a high SSA [43,44]. These values are in agreement with the
values reported by Wu et al. [42] for non-stoichiometric LaCo1-x Fex
O3 perovskites.
3.3. Surface properties
The nature and the relative abundances of the surface species
were determined by X-ray photoelectron spectroscopy (XPS).
Table 1 gathers the binding energies (BEs) of the C1s, O1s, Fe2p3/2
and La3d5/2 levels, as well as the calculated relative atomic abun-
dances.
The energy of the Fe2p3/2 level was located around
710 eV ± 0.4 eV. The corresponding photopeak was accompa-
nied with a satellite peak at 8.4 eV ± 0.2 eV for all the samples.
These values were characteristic of Fe3+ species [45]. The La3d
level with a spin-orbital coupling of 5/2 exhibited a binding energy
of 834.3 eV ± 0.3 eV. A satellite peak was also present at an energy
of 3.8 eV larger than that of the central transition. The binding
energies of these two peaks were characteristic of the presence of
La species with an oxidation degree of +III [46].
The surface XPS La/Fe ratios (Table 1) are quite different from
the bulk ratios determined by ICP analysis. For the stoichiometric
LaFe sample, a two-fold lanthanum surface enrichment was already
detected, in agreement with the presence of a low quantity of La2 O3
at the sample surface as detected by XRD. Then, a linear evolution in
function of x was observed between the La/FeICP and LaFeXPS ratios
over the non-stoichiometric samples. Indeed, when x increased,
there was an enrichment in lanthanum on the surface by a factor
3.5 (slope of the line). This enrichment over all the samples came
from the formation of both secondary phases La2 O3 and La(OH)3 at
the surface of samples [25].
Unfortunately, fitting the La3d5/2 spectrum is not an easy and
straight forward task, as reported else were [46,47]. In order to go
deeper in the understanding of the surface La species speciation,
we focused our attention on the C1s and O1s contributions. Fig. 4
gives an example of O1s spectrum fitting, which needs 4 compo-
nents of which the assignment was performed thanks to the studies
of Faye et al. [25] and Van Der Heide et al. [48]. The first peak
was then assigned to oxygen atoms present in the LaFeO3 struc-
ture at a BE of 529.1 eV [49,50], the second peak was associated
to La2 O3 (BE = 530.4 eV) [48], the third one located at 533.7 eV was
assigned to oxygen atoms present in water molecules present at
the surface of the materials [51]. Finally, the last contribution at
531.3 eV was attributed to oxygen in La(OH)3 , but also to oxygen
present in surface carbonate species [52]. The presence of carbon-
ate species at the surface was confirmed by the analysis of the C1s
core level, which showed a characteristic additional contribution
at 289.3 eV. Surface carbonates were indeed easily formed by reac-
tion between surface lanthanum species and CO2 from the ambient
atmosphere [48,53]. In addition, it could not be excluded that these
species have been created during the auto-combustion of glycine,
539 537 535 533 531 529 527 525
Binding Energy (eV)
Fig. 4. Example of the O1s core level spectrum convolution obtained for the La1.3 Fe
sample. OLaDeO3 at 529.1 eV; OLa2 O3 at 530.4 eV; OLa(OH)3 + CO2− 3
at
531.3 eV; OHz O at 533.7 eV.
which has generated CO2 in situ, as underlined by Valange et al.
[54]. Such carbonated species seems to be present in an amor-
phous phase or are very dispersed, since no traces of lanthanum
carbonate were identified by XRD analysis. The distribution of
the different species as a function of the catalyst formulation is
given in Table 1. The quantity of LaFeO3 at the surface of the cat-
alyst decreased with the increase of x. In contrast, the amounts of
La2 O3 and La(OH)3 + carbonate increased with x. These results have
confirmed the semi-quantitative analysis of the secondary phases
obtained by the bulk analysis XRD (Fig. 3), i.e., that the increase in
the peak intensity was due to an increase of the quantity of this
phase and no from a better cristallinity.
3.4. Acid-base properties
The acid and base properties of the samples were determined
using a 2-propanol catalytic test. The reaction mechanism and the
phenomena taking place could become very complicated, but keep-
ing a low 2-propanol concentration in the feed limits the formation
of secondary products, thus considerably simplifying the system.
The main expected products were acetone and propylene accord-
ing to the presence of basic sites (Eq. (1)) or acid sites (Eq. (2)),
respectively:
CH3 CH(OH)CH3 → CH3 COCH3 + H2 (1)
CH3 CH(OH)CH3 → CH2 CHCH3 + H2 O (2)
The selectivity to acetone (Sa ), to propylene (Sp ) and the Sa /Sp
ratio are plotted in Fig. 5 as a function of La2 O3 phase percentage
determined thanks to the O1S contribution (cf XPS analysis in the
Table 1), for a 2-propanol conversion of 30%.
The curves assigned to the acetone and propylene selectivities
roughly follow an inverse trend, whereas the curves correspond-
ing to the Sa /Sp ratio and the selectivity to acetone are very similar.
It means that the basic character was the predominant property of
our samples, as suggested by the Sa /Sp ratio, which was always larg-
erthan1. A maximum Sa /Sp value was reached for x = 0.3 (3.8), and
could be associated to the high quantity of La2 O3 (9.1 at% deter-
mined by XPS) at the surface of the sample. La2 O3 is considered as
the most basic phase among the rare earth simple oxide [43,55]. As
a consequence, the sample presenting the higher La2 O3 quantity on
its surface and no lanthanum hydroxide was also that presenting
the highest basic character (i.e., in term of basic site number and/or
basic site strength). A further increase in the lanthanum content led
to a decrease in the Sa /Sp ratio to 2.4 and 2.2 for La1.5 Fe and La1.8 Fe,
respectively. This result can be correlated with the appearance of
lanthanum hydroxide observed by XRD. The presence of this phase
seems to limit the basic character of the catalysts by neutraliza-
 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148 145

50
La1.3Fe
4,5 to the temperature increase. The increase is moderate because our
Selectivitiy to acetone (A) and
A catalysts exhibited quite low specific surface areas (Table 1). At
45 4
40 the higher test temperature, a quite linear evolution was observed
3,5
propylene (P) (%)

La2Fe between the ethanol conversion and the samples basicity repre-
35

30
La1.8Fe
25 LaFe
La1.1Fe La1.5Fe
20
15
10
5 0,5
0 0
0 5
O La2O3
Fig. 5. Selectivity to acetone (A), to propylene (P) and Sa /Sp ratio in the reaction
of iso propanol conversion as a function of the lanthanum oxide phase percent-
age determined thanks to the oxygen O1s spectrum deconvolution determined by
XPS. Catalytic test parameters: catalytic bed is composed of a mixture 100 mg cat-
alyst/100 mg SiC; the ratio ethanol/helium is fixed to 4.5/95.5 (vol%) with helium
flow of 25 mL/min; GHSV: 1125 h−1 .
tion or covering of the surface basic sites. Finally, according to the
XPS data interpretation, the La2 Fe sample presented the highest
quantity of La2 O3 , i.e., 12.1% (against 9.1% for La1.3 Fe sample) and
by consequence, this sample should have the highest Sa/Sp ratio,
which was not the case (3.8 and 3.6 for the La1.3 Fe and La2 Fe sam-
ples, respectively). In this sample, the presence (not evidenced for
La1,3 Fe) of La(OH)3 leads to a decrease in the basic character.
3.5. Catalytic performances
Below 300 ◦ C, the catalysts exhibited no activity, while above
400 ◦ C deactivation by carbon deposition became too prominent.
All the obtained catalytic results, including conversion and selectiv-
ities to all the products (including the minor ones), are gathered in
the Electronic Supplementary Information (ESI) file 1. Fig. 6A shows
the conversion of the catalysts as a function of the selected temper-
ature and Fig. 6B illustrates the evolution of the ethanol conversion
at 400 ◦ C as a function of the basicity of the samples represented by
the Sa /Sp ratio.
In Fig. 6A, we observed that the conversion increased with the
temperature overall the samples. At 300 ◦ C, the ethanol conversion
was inferior or equal to 10%, irrespective of the sample. An increase
in the reaction temperature to 350 ◦ C had only a moderate influence
on the ethanol conversion, while working at 400 ◦ C leads to a slight
improvement of the catalysts conversion of about 15 points. This
ethanol conversion increase was due to the energetic input owing
Sa / Sp Ratio
3 sented by the Sa /Sp ratio (Fig. 6B). As previously mentioned (Section
2,5 2.7), this ratio was strongly linked to the presence of the La2 O3
and La(OH)3 secondary phases. LaFe and La1.1 Fe exhibited a sim-
2
ilar conversion of about 20%, while a maximum conversion was
1,5
P observed overLa1,3 Fe (32%). The ethanol conversion can be corre-
1 lated with the evolution of the Sa /Sp ratio. For La1,5 Fe and La1,8 Fe,
the lanthanum excess responsible for the formation of La(OH)3 to
the detriment of La2 O3 led to a decrease in the ethanol conversion.
10 15 Finally, the conversion observed over La2 Fe, which had a higher
La2 O3 concentration than La1.5 Fe and La1.8 Fe, was substantially
identical to that of La1,3 Fe (31%).
In brief, the highest conversion was obtained for catalyst with
the highest Sa /Sp ratio, i.e., La1.3 Fe and La2 Fe, and the conversion
then seemed to be governed by the quantity of additional basic
sites provided by La2 O3 .
Among the reaction products, no traces of butenol and/or
butanediol [9], but the presence of acetone coming from the 4-
hydroxybutan-2-al, which is the tautomeric specie of the aldol
formed in the Guerbet reaction, were detected during the catalytic
tests. This suggests that the main reaction pathway consists on the
Guerbet reaction mechanism [8]. Due to the high level of complex-
ity of this reaction, which requires basic and acid sites of different
strength and/or nature, and of samples used in this study, com-
peting reactions can occur. Among the different products formed
at low and high temperature, we found Guerbet intermediates
and final products, namely acetaldehyde, butyraldehyde and 1-
butanol. However, other compounds were detected like alcohols
such as 2-propanol and 2-pentanol, as well as ketones and aldehy-
des like acetone, 2-pentanone, 2-heptanone and 4-heptanone and
also others such as CO, CO2 , ethylene and ethane, all these products
being issued from competing reactions. Hereafter, a general view
of the catalytic results will be presented, and, after a more detailed
study on the correlation between the physico-chemical properties
of catalysts and their catalytic performances, we will accordingly
propose a general reaction scheme.
Fig. 7A and B present the ethanol conversion, the products selec-
tivity and the carbon balance (CB) observed over all the catalysts
at 300 and 400 ◦ C, respectively. The selectivity values correspond
to an average calculated on 5 analyses. These five analyses are per-
formed once the steady state reached, i.e., after 3 h of reaction at
each temperature. The CB decreased with the increase in x, which
was mainly due to the increase in carbon species deposition on the

35 A 35
B La1.3Fe
300 °C 30 La1.8Fe
30 La2Fe
Conversion (%)

350 °C
Conversion (%)

25 400 °C 25

20 La1.1Fe
20 La1.5Fe
15 LaFe
15
10
10
5
5
0
0 0 1 2 3 4
LaFe La1.1Fe La1.3Fe La1.5Fe La1.8Fe La2Fe Sa / Sp

Fig. 6. Ethanol conversion as a function of the lanthanum content at 300, 350 and 400 ◦ C; (B) ethanol conversion at 400 ◦ C as a function of the Sa/Sp ratio. Catalytic test
parameters: catalytic bed is composed of a mixture 200 mg catalyst/200 mg SiC; the ratio ethanol/helium is fixed to 13/87 (vol%) with helium flow of 40 mL/min; GHSV:
1125 h−1 .
146 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148

A A 40 █ SEthylène █ % wt. La2O3 18

% weight in La2O3 and La(OH)3

35 █ % wt. La(OH)3 16

Selectivity to ethylene (%)

100%
Conversion, selectivities and Carbon

90% 30 14
80% 12

from XRD
25
balance (%) at 300°C

70% 10
Ethyl butyrate
60% 2-pentanone
20
8
50% Ethyl acetate
15
Acetone 6
40% Acetaldehyde
10 4
30% Conversion
Carbon balance 5
20% 2
10% 0 0
La1.1Fe La1.5Fe La2Fe
0% LaFe La1.3Fe La1.8Fe
La1.5Fe La Fe La2Fe
LaFe La1.1Fe La1.3Fe 1.8
B 16 4,5
B 4
14

Selectivity to acetone (%)

100%
12 3,5
Carbon balance (%) at 400°C

90%
Conversion, Selectivities and

Sa / Sp ratio
Others 3
80% Ethane 10
CO2 2,5
70% 8
Ethylene
60% Ethyl butyrate
2
2-pentanone
6
50% 1,5
Butanol
40% Ethyl acetate
4 1
Acetone
30% 2 0,5
Acetaldehyde
20% Conversion
0 0
10% Carbon balance La1.1Fe La1.5Fe La2Fe
LaFe La1.3Fe La1.8Fe
0% La1.5Fe La2Fe
La1.1Fe
LaFe La1.3Fe La1.8Fe C 35 █ Sn-butanol 4,5
█ S2-pentanone
Selectivity to n-butanol and

Fig. 7. Ethanol conversion, carbon balance and selectivities to products at (A) 300 ◦ C 30 4
and (B) 400 ◦ C for all the samples of the La1+x Fe series. Catalytic test parame-
2-pentanone (%)

ters: catalytic bed is composed of a mixture 200 mg catalyst/200 mg SiC; the ratio 25 3,5

Sa / Sp ratio
ethanol/helium is fixed to 13/87 (vol%) with helium flow of 40 mL/min; GHSV:
1125 h−1 . 20 3

15 2,5
catalysts surface (coking). The reaction products had between 1 and
10 2
7 carbon atoms.
At 300 ◦ C (Figs. 7A), the main products were acetaldehyde, ethyl 5 1,5
acetate (EA) and ethyl butyrate (EBut). When the temperature
increased at 350 ◦ C (results presented in ESI) and 400 ◦ C, the selec- 0 La1.1Fe La1.5Fe La2Fe
1
tivity to acetaldehyde decreased in favor of the formation of new LaFe La1.3Fe La1.8Fe
products issued from secondary reactions (2-propanol, 1-butanol,
Fig. 8. (A) Selectivity to ethylene at 400 ◦ C and weight concentration in La2 O3 and
2-pentanol, acetone, butyraldehyde, 2-pentanone, 2-heptanone La(OH)3 in the samples bulk (data from XRD); (B) Selectivity to acetone and frac-
and 4-heptanone, CO and CO2 ). This acetaldehyde transformation tion Sa /Sp at 400 ◦ C for each sample; (C) Selectivities to n-butanol, 2-pentanone and
was already visible in Fig. 7A at 300 ◦ C for La1.8 Fe and La2 Fe and lead fraction Sa /Sp at 400 ◦ C for each sample. Catalytic test parameters: catalytic bed is
to minor products such as acetone and 2-pentanone. Acetaldehyde composed of a mixture 200 mg catalyst/200 mg SiC; the ratio ethanol/helium is fixed
to 13/87 (vol%) with helium flow of 40 mL/min; GHSV: 1125 h−1 .
was produced by dehydrogenation, which occurred on basic sites.
In their work, Silvester et al. [17,20] have shown that the first step
of Guerbet reaction, to form acetaldehyde, occurred on weak or
medium basic sites as CaO. In our study, we could suppose that this observe a decrease in ethylene selectivity between LaFe and La1.3 Fe
reaction taken place on La O and/or O La O bond according to from 28 to 17%, which can be correlated to the appearance of La2 O3
mechanism proposed by Diez et al. [56] and Gakkai [57]. since the presence of this oxide in the catalyst induced an increase
In the following discussion, we will mainly focus on the main in the amount of the basic sites while hiding a part of the acidity of
products, i.e., ethylene, acetone, 1-butanol and 2-pentanone. the perovskite most probably by coverage. Indeed, the appearance
The discussion will be performed on the results obtained at of secondary phase at the surface of catalysts decrease the availabil-
400 ◦ C and not at full conversion, because to reach it, the tem- ity of Lewis acidic sites Fe3+ , as it was shown by the XPS analysis in
perature had to be strongly increased. Unfortunately, when the Table 1. La1.5 Fe and La1.3 Fe had nearly the same La2 O3 concentra-
temperature was higher than 400 ◦ C, the ethanol conversion was tion, but the surface of the first one also comprised La(OH)3 , leading
higher due to overoxidation (i.e., mainly production of CO2 ), which to an increase in the ethylene selectivity to 25%. This increase of the
did not allowed finding any correlation. ethylene selectivity can result of the neutralization or coverage of
In our system, ethylene was formed by intramolecular dehydra- basic sites by the La(OH)3 phase. La1.8 Fe was the sample with the
tion of ethanol, which was favored at high temperature over acid highest concentration in La(OH)3 and with the highest selectivity
sites (Bronsted and Lewis). Fig. 8A presents the ethylene selectivity to ethylene (about 37%). Over La2 Fe, the presence of an important
at 400 ◦ C and the weight fraction of the two secondary crystalline quantity of La2 O3 together with a decrease in the La(OH)3 content
phases (i.e., La2 O3 and La(OH)3 ) over the various catalysts. We induced a drop in ethylene selectivity.
 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148
Fig. 9. General reaction pathway.
A similar evolution between the acetone selectivity and the Sa /Sp
ratio as function of x was observed in Fig. 8B. The selectivity to ace-
tone was maximal for La1.3 Fe (14%), suggesting a formation linked
with the presence of basic properties. As previously described by
Gines et al. [58] and Bussi et al. [59], acetone could be formed
through a reverse aldolization (ESI 2). The aldol (3-hydroxybutanal)
is in tautomeric equilibrium with 4-hydroxybutan-2-al (so-called
‘keto’ form). This latter is more stable than the aldol of 25 kJ/mol.
The keto can then react through a reverse aldolization to form
acetone and formaldehyde. The latter, never observed in our
experiments, can be directly converted to CO and H2 . The dif-
ference instability between both tautomeric species was also
reflected in the distribution of products resulting from their sub-
sequent reactions. We could notice that the sum of the derived
products selectivities from 4-hydroxybutan-2-al, “keto” form (i.e.,
acetone, 2-propanol, 2-pentanone, 2-pentanol, 2-heptanone and
4-heptanone, Fig. 8B) was equal to 50% whereas the sum of the
derived products selectivities from the “aldol” form (i.e., butyralde-
hyde and 1-butanol) almost reached 9%.
The selectivities to 1-butanol and 2-pentanone at 400 ◦ C as well
as the Sa /Sp ratio have been plotted as a function of each sam-
147
ple (Fig. 8C). These two products followed the same trend as the
Sa /Sp ratio, with a maximum of selectivity for x = 0.3. As a matter
of fact, these products were both formed by aldolization, between
two acetaldehyde molecules to 1-butanol, and between acetone
and ethanol to 2-pentanone, meaning that the same active sites,
i.e., basic ones (supposedly strong basic sites), were involved for
the formation of these products. The evolution of the selectivity in
1-butanol as a function of the basicity is in according to the previ-
ous work of Ogo et al. [19] on hydroxyapatite. However, as already
mentioned, the Guerbet reaction needs acid/base properties and
Silvester et al. [20] have determined an optimal ratio of acidity on
basicity of 5 for the production of 1-butanol, and generally alcohols.
The results presented in this study seemed to be characteristics of
an acidity/basicity ratio widely superior to 5, i.e., a good conversion,
a high selectivity in ethylene and a low selectivity in 1-butanol and
alcohols.
The same meticulous methodology was applied to explain the
formation of each observed product. This allowed the development
of a general reaction scheme, which is the conclusion of this work
(Fig. 9)
148 G. Tesquet et al. / Applied Catalysis A: General 511 (2016) 141–148
4. Conclusion
The aim of this work was to study the influence of the stoi-
chiometry, and more particularly, of the lanthanum excess in mixed
oxides, based on the LaFeO3 perovskite, on their ethanol catalytic
activity. Catalysts of the general formula La1+x FeO3+ı were pre-
pared by the so-called auto-combustion method and characterized
using several techniques. XRD analysis allowed the identification
of the different crystalline phases. For 0 < x ≤ 0.3, the perovskite
phase (LaFeO3 ), together with La2 O3 , were identified, while a third
phase La(OH)3 was also detected for the samples with 0.3 < x ≤ 1.
The secondary phases [LaFeO3 , La2 O3 and La(OH)3 ] were located
at the surface of the solids deduced from XPS analysis. Adding La
in excess in the perovskite structure leads to a change in term of
acido/basicity property. These properties play a major role in the
reactivity of the solids, which is largely governed by a Guerbet reac-
tion pathway interweaved with parallel competitive reactions. The
various characterizations of the La1+x FeO3-ı mixed oxides allowed
us establishing of a correlation between the catalytic activity and
the physico-chemical properties of the samples. The most basic cat-
alyst (La1.3 Fe), as determined by the iso propanol test, was although
that exhibiting the highest ethanol conversion. La1.3 Fe is also the
sample that exhibited the most important La2 O3 concentration
at its surface. This catalyst gave the highest yield to n-butanol
(i.e., 2%) and the highest selectivity to 2-pentanone (12%), this lat-
ter compound being formed by an aldolization reaction between
acetaldehyde and acetone, itself formed by reverse aldolization of
the “keto” molecule (tautomeric species of the aldol formed during
Guerbet reaction pathway).
Acknowledgments
The authors would like to thanks the French National Agency
for Research (ANR) for its financial contribution within the VAL-
BIOGAL project (ANR-09-RPDOC-025-01). We would also like to
thank the engineers and technicians who worked on this study,
namely Mrs Laurence BURYLO for the XRD analyses, Mrs Martine
TRENTESAUX for the XPS analyses, Mr. Olivier GARDOLL for the
B.E.T measurements, as well as the SOLAIZE laboratory for the ICP
analyses.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.apcata.2015.12.
005.
References
[1] Top Value Added Chemicals From Biomass, Volume I: Results of Screening for
Potential Candidates from Sugars and Synthesis Gas Produced by Staff at the
Pacific Northwest National Laboratory (PNNL) and the National Renewable
Energy Laboratory (NREL), August 2004.
[2] BP Statistical Review of World Energy June 2012 bp.com/statisticalreview
6–18.
[3] http://www.planetoscope.com/biocarburants/524-production-mondiale-d-
ethanol.html.
[4] F. Lujaji, L. Kristof, A. Bereczky, M. Mbarawa, Fuel 90 (2011) 505–510.
[5] O. Dogan, Fuel 90 (2011) 2467–2472.
[6] T. Imanaka, T. Tanemoto, S. Teranishi, 7th International Congress on Catalysis,
Tokyo, 1980, Preprints of Papers, No.4.
[7] M. Guerbet, C.R. Aced, Science 128 (1899) 1002–1004.
[8] A.J. O’Lennick, J.K. Parkinson, J. Soc. Cosmet. Chem. 45 (1994) 247–256.
[9] J. Scalbert, F. Thibault-Starzyk, R. Jacquot, D. Morvan, F. Meunier, J. Catal. 311
(2014) 28–32.
[10] J.T. Kozlowski, R.J. Davis, ACS Catal. 3 (2013) 1588–1600.
[11] C. Carlini, M. Di Girolamo, A. Macinai, M. Marchionna, M. Noviello, A.M.
Raspolli Galletti, G. Sbrana, J. Mol. Catal. A 204–205 (2003) 721–728.
[12] C. Carlini, M. Di Girolamo, A. Macinai, M. Marchionna, M. Noviello, A.M.
Raspolli Galletti, G. Sbrana, J. Mol. Catal. A 200 (2003) 137–146.
[13] C. Carlini, A. Macinai, M. Marchionna, M. Noviello, A.M. Raspolli Galletti, G.
Sbrana, J. Mol. Catal. A 206 (2003) 409–418.
[14] C. Carlini, C. Flego, M. Marchionna, M. Noviello, A.M. Raspolli Galletti, G.
Sbrana, F. Basile, A. Vaccari, J. Mol. Catal. A 220 (2004) 215–220.
[15] C. Carlini, C. Flego, M. Marchionna, M. Noviello, A.M. Raspolli Galletti, G.
Sbrana, F. Basile, A. Vaccari, J. Mol. Catal. A 232 (2005) 13–20.
[16] F. Hosoglu, J. Faye, K. Mareseanu, G. Tesquet, P. Miquel, M. Capron, O. Gardoll,
J.F. Lamonier, C. Lamonier, F. Dumeignil, Appl. Catal. A 504 (2015) 533–541.
[17] L. Silvester, J.F. Lamonier, R.N. Vannier, C. Lamonier, M. Capron, A.S. Mamede,
F. Pourpoint, A. Gervasini, F. Dumeignil, J. Mater. Chem. 2 (29) (2014)
11073–11090.
[18] T. Tsuchida, J. Kubo, T. Yoshioka, T. Takeguchi, W. Ueda, J. Catal. 259 (2008)
183–189.
[19] S. Ogo, A. Onda, K. Yanagisawa, Appl. Catal. A 402 (2011) 188–195.
[20] L. Silvester, J.F. Lamonier, J. Faye, M. Capron, R.N. Vannier, C. Lamonier, J.L.
Dubois, J.L. Couturier, C. Calais, F. Dumeignil, Catal. Sci. Technol. 5 (2015)
2994–3006.
[21] A.S. Ndou, N.J. Coville, Appl. Catal. A 251 (2003) 337–345.
[22] N.L. Allan, M.J. Dayer, D.T. Kulp, W.C. Mackrodt, J. Mater. Chem. 1 (1991)
1035–1039.
[23] http://www.chemexplore.net/ReO3-NiO.htm.
[24] S. Royer, H. Alamdari, D. Duprez, S. Kaliaguine, Appl. Catal. B 58 (2005)
273–288.
[25] J. Faye, E. Guelou, J. Barrault, J.M. Tatibouët, S. Valange, Top. Catal. 52 (2009)
1211–1219.
[26] A. Perdersen, W.F. Libby, Science 176 (1972) 1355–1356.
[27] H. Wang, C. Tablet, T. Schiestel, J. Caro, Catal. Today 118 (2006) 98–103.
[28] R. Spinicci, M. Faticanti, P. Marini, S. De Rossi, P. Porta, J. Mol. Catal. A 197
(2003) 147–155.
[29] B. Levasseur, S. Kaliaguine, Appl. Catal. A 343 (2008) 29–38.
[30] T. Arakawa, S.I. Tsuchi-Ya, J. Shiokawa, J. Catal. 74 (1982) 317–322.
[31] T. Arakawa, N. Ohara, H. Kurachi, J. Shiokawa, J. Colloid Interface Sci. 108
(1985) 407–410.
[32] Y. Yokoi, H. Uchida, Catal. Today 42 (1998) 167–174.
[33] R. Zhang, A. Villanueva, H. Alamdari, S. Kaliaguine, J. Mol. Catal. A 258 (2006)
22–34.
[34] R. Zhang, H. Alamdari, S. Kaliaguine, J. Catal. 242 (2006) 241–253.
[35] L.A. Chick, L.R. Pederson, G.D. Maupin, J.L. Bates, L.E. Thomas, G.L. Exarhos,
Mater. Lett. 10 (1990) 6–12.
[36] S. Specchia, A. Civera, G. Sarracco, Chem. Eng. Sci. 59 (2004) 5091–5098.
[37] K.C. Patil, S.T. Aruma, T. Mimani, Curr. Opin. Solid State Mater. 6 (2002)
507–512.
[38] L. Shue, S. Kaliaguine, Appl. Catal. B 16 (1998) L303–L308.
[39] S. Lee, J. Lee, Y. Park, J. Wee, Catal. Today 117 (2006) 376–381.
[40] P. Ciambelli, S. Cimino, L. Lisi, P. Porta, Appl. Catal. B 33 (2001) 193–203.
[41] D.A. Shirley, Phys. Rev. B 5 (1972) 4709–4714.
[42] Y. Wu, X. Ni, A. Beaurain, C. Dujardin, P. Granger, Appl. Catal. B 125 (2012)
149–157.
[43] S. Sato, R. Takahashi, M. Kobune, H. Gotoh, Appl. Catal. A 356 (2009) 57–63.
[44] Q. Li, J. Ni, Y. Wu, Y. Du, W. Ding, S. Deng, J. Rare Earths 29 (2011) 416–419.
[45] S. Ponce, M.A. Peña, J.L.G. Fierro, Appl. Catal. B 24 (2000) 193–205.
[46] T. Nitadori, S. Kurihara, M. Misono, J. Catal. 98 (1986) 221–228.
[47] R. Leanza, I. Rossetti, L. Fabbrini, C. Oliva, L. Forni, Appl. Catal. B 28 (2000)
55–64.
[48] P.A.W. Van der Heide, Surf. Interface Anal. 33 (2002) 414–425.
[49] G. Vovk, X. Chen, C.A. Mims, J. Phys. Chem. B 109 (2005) 2445–2454.
[50] J. Faye, A. Baylet, M. Trentesaux, S. Royer, F. Dumeignil, D. Duprez, S. Valange,
J.M. Tatibouët, Appl. Catal. B 126 (2012) 134–143.
[51] C.R. Clayton, Y.C. Lu, J. Electrochem. Soc. 133 (1986) 2465–2473.
[52] N.S. McIntyre, D.G. Zetaruk, Anal. Chem. 49 (1977) 1521–1529.
[53] M.P. Rosynek, D.T. Magnuson, J. Catal. 48 (1977) 417–421.
[54] S. Valange, A. Beauchaud, J. Barrault, Z. Gabelica, M. Daturi, F. Can, J. Catal. 251
(2007) 113–122.
[55] O.V. Manoilova, S.G. Podkolzin, B. Tope, J. Lercher, E.E. Stangland, J.M. Goupil,
B.M. Weckhuysen, J. Phys. Chem. B 108 (2004) 15770–15781.
[56] V.K. Diez, C.R. Apestuguia, J.I. Di Cosimo, Catal. Today 63 (2000) 53–62.
[57] S. Gakkai, « Gensobetu ShokubaiBenran », Chijn Shokan, Tokio, 1978, 788–790.
[58] M.J.L. Gines, E. Iglesia, J. Catal. 176 (1998) 155–172.
[59] J. Bussi, S. Parodi, B. Irigaray, R. Kieffer, Appl Catal. A 172 (1998) 117–129.