Continuous Gas-Phase Condensation of Bioethanol to 1-Butanol over Bifunctional Pd/Mg and Pd/Mg–Carbon Catalysts

Article abstract of DOI:10.1002/cssc.201801381

The condensation of ethanol to 1-butanol in the presence of different catalyst systems based on a Pd dehydrogenating/hydrogenating component and magnesium hydroxide-derived materials as basic ingredient was studied in a fixed-bed reactor. The metal was incorporated by wetness impregnation, and the resulting material was then reduced in situ with hydrogen at 573 K for 1 h before reaction. The bifunctional catalysts were tested in a fixed-bed reactor operated in the gas phase at 503 K and 50 bar with a stream of helium and ethanol. A bifunctional catalyst supported on a synthetic composite based on Mg and high surface area graphite (HSAG) was also studied. Improved catalytic performance in terms of selectivity towards 1-butanol and stability was shown by the Pd catalyst supported on the Mg–HSAG composite after thermal treatment in helium at 723 K, presumably due to the compromise between two parameters: adequate size of the Pd nanoparticles and the concentration of strongly basic sites. The results indicate that the optimal density of strongly basic sites is a key aspect in designing superior bifunctional heterogeneous catalyst systems for the condensation of ethanol to 1-butanol.

Full text of DOI:10.1002/cssc.201801381

Accepted Article
Title: Continuous gas phase condensation of bioethanol to 1-butanol
over bifunctional Pd/Mg and Pd/Mg-carbon catalysts
Authors: Cristina López-Olmos, Maria Virtudes Morales, Carolina
Ramirez-Barria, Esther Asedegbega-Nieto, Inmaculada
Rodriguez-Ramos, and Antonio Guerrero-Ruiz
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201801381
Link to VoR: http://dx.doi.org/10.1002/cssc.201801381
A Journal of
www.chemsuschem.org

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Continuous gas phase condensation of bioethanol to 1-butanol
over bifunctional Pd/Mg and Pd/Mg-carbon catalysts
C. López-Olmos,[a] M.V. Morales,[b] A. Guerrero-Ruiz,*[b, c] C. Ramirez-Barria,[a,b] E. Asedegbega-
Nieto,[b] I. Rodríguez-Ramos[a, c]
Abstract: The condensation of ethanol into 1-butanol, in the
consumption compare to petrol.[4] Furthermore, apart from its
fuel properties, 1-butanol finds applications in the paint, solvents
and plasticizers and it is also the raw material for acrylic acid
and acrylic esters.
presence of different catalytic systems based on Pd
dehydrogenating/hydrogenating
component
and
magnesium
hydroxide derived materials as basic ingredient, was studied in a
fixed bed reactor. The incorporation of metal was done by wetness
impregnation and the resulting material was then reduced in-situ with
hydrogen at 573 K for 1 h before reaction. The bifunctional catalysts
were tested in a fixed bed reactor operated in gas phase at 503 K
and 50 bar using a stream of helium and ethanol. Bifunctional
catalyst supported over synthetized composite based on Mg and a
carbon material: high surface area graphite, was also studied.
Improved catalytic performance in terms of selectivity towards 1-
butanol and stability was shown by the Pd catalyst supported on Mg-
HSAG composite after thermal treatment in helium at 723K,
presumably due to the compromise between two parameters: the
adequate size of Pd nanoparticles and the concentration of strong
basic sites. Our results indicate the optimal density of strong basic
Traditionally, butanol has been produced through the acetone-
butanol-ethanol (ABE) bacterial fermentation. But in the 1950s
this method was substituted by the petrochemical route known
as the Oxo process, where propylene is hydroformylated using
syn-gas over a homogeneous rhodium catalyst to yield butanal,
which is then hydrogenated to butanol.[5] However, this route
would not endure owing to the rising prices in crude oil, which
has favored the industrial ABE fermentation to emerge again in
many countries. Nevertheless, some studies have pointed out to
the limitations of this fermentation process such as low butanol
yield and by-product formation (acetone and ethanol).[6] On the
contrary, the direct conversion of butanol from ethanol has been
suggested to be a more desirable route, since the reaction
proceeds faster when compared to the fermentation process and
fewer steps are necessary to get the product. Bearing in mind
that ethanol could also be a biomass derivative (bioethanol),[7]
the catalytic conversion of bioethanol into butanol represents a
promising alternative for fuel production from renewable
biomass resources[8].
sites is
a
key aspect in designing superior bifunctional
heterogeneous catalytic systems for the condensation of ethanol to
1-butanol.
Introduction
The synthesis of higher alcohols from light alcohols is known as
Guerbet reaction and has been used for more than 100 years.[9]
Traditionally, it was carried out employing a homogeneous base
(alkali metal hydroxide) in presence of a metal. The chemical
route for the synthesis of 1-butanol from ethanol implies a multi-
step mechanism (see Figure 1): Step (1) is the dehydrogenation
of ethanol to acetaldehyde, step (2) is the base-catalyzed aldol
condensation of acetaldehyde to 2-butenal, and steps (3) and
(4) are hydrogenation reactions of 2-butenal, 2-butenol, and/or
butanal into 1-butanol. The metal carries out the ethanol
dehydrogenation to acetaldehyde, as well as the hydrogenation
of reaction intermediates to 1-butanol, while the homogeneous
base is responsible for the aldol-condensation step in this
Guerbet reaction mechanism.[10] However, despite the fact that
homogeneous catalysts have been widely studied with rather
good performance,[11] it is known that the use of basic
heterogeneous catalysts is more desirable from a sustainable
point of view. In this sense, the solid basic catalysts such as
the non-expensive alkaline earth metal oxides and hydroxides
are receiving more attention in last years as promising
alternatives in base-catalyzed reactions.[12]
The interest in 1-butanol production has grown in recent years
because it is contemplated as a promising alternative to ethanol
as petrol substitute. Butanol offers many advantages when used
as fuel over ethanol due to its properties such as its lower
volatility, immiscibility with water, less corrosive nature (which
leads to improved safety) and mainly its higher heat of
combustion, similar to that of gasoline.[1,2] Also unlike other
alcohols, butanol has the air-fuel ratio closer to that of petrol,
which makes it more suitable when used in current cars as no
modification in combustion engine is required.[3] As a matter of
fact, the application of butanol as biofuel has already been
demonstrated, and at the same time a reduction in emission of
pollutants was attained with a very low increase in fuel
[a]
[b]
C. López-Olmos, C. Ramírez-Barria, Prof. I. Rodríguez-Ramos
Instituto de Catálisis y Petroleoquímica, CSIC
C/ Marie Curie 2, Cantoblanco, 28049 Madrid, Spain
M.V. Morales, Prof. A. Guerrero-Ruiz, Dr. E. Asedegbega-Nieto
Departamento de Química Inorgánica y Técnica
UNED, Facultad de Ciencias
Quite a few recent reviews can be found in literature offering a
detailed discussion on heterogeneous catalysts employed in
this reaction.[6,13-15] From these reviews, it is generally drawn
that, due to the plurality of reactions involved, the use of
catalytic systems simultaneously displaying basic, acidic and
Paseo Senda del Rey 9, 28040 Madrid, Spain
E-mail: aguerrero@ccia.uned.es
[c]
Prof. A. Guerrero-Ruiz, Prof. I. Rodríguez-Ramos
Grupo de Diseño y Aplicación de Catalizadores Heterogéneos,
Unidad Asociada UNED-CSIC (ICP), Spain
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Figure 1. Ethanol condensation mechanism based on the Guerbet reaction. DHG: dehydrogenation, C: aldol-condensation, DH: dehydration, H: hydrogenation.
dehydrogenating/hydrogenating properties is required. Even so,
alternatively to the two aldehyde self-aldolization mechanism, a
few works employing hydroxyapatite catalysts[16,17] or MgO[18]
addresses the mechanism as a direct bimolecular condensation
of ethanol (proton abstraction mechanism). Note that this
reaction mechanism could only be expected in the case of not
transition metal-oxide catalysts.
Despite the generally accepted fact that the coupling reaction of
alcohols requires a base as a critical component, it is often
claimed that appropriate strength of acid-base pairs is also
essential. For this reason, many heterogeneous catalytic
systems described in the literature and patents for the
condensation of ethanol to 1-butanol are mixed oxides, where
Lewis acidic metals (such as Al) are incorporated in the MgO
structure, often prepared from thermal treatment of
hydrotalcites[19-21] as well as mixed oxides modified with
transition metals.[22-26] Nevertheless, there are other materials
free of magnesia that have proven to be active for ethanol
Guerbet condensation reaction such as zeolites modified with
alkali metals,[27] transition metals supported on slightly basic
alumina[28,29], Cu on high surface area Ce2O[30] and modified
zirconia with Na.[31]
challenging task owing to the uncontrolled formation of side-
products due to the high reactivity of the intermediate product
acetaldehyde.[35] Furthermore, finding a selective and stable
catalyst in condensation reactions, which are quite significant in
the preparation of fuels and chemicals from biomass derivatives,
emerges as a great task in heterogeneous catalysis. Magnesium
oxides have been extensively explored as reference material for
processes involving aldol condensation reactions,[36,37] however,
although there are a few papers describing the use of MgO-
based materials in the condensation of ethanol, they usually
refer to processes at atmospheric pressure and high
temperatures.[14,18,19] Besides, to our knowledge, very little has
been published about the effect of addition of transition metals to
MgO as well as employing carbon supports in the Guerbet
reaction of ethanol in fixed bed reactors. In this work, we have
prepared Pd catalysts supported on Mg-based materials. This
metal was chosen because of the high selectivities reported for
this reaction.[24,25] Thereafter, the influence in the catalytic
activity and selectivity towards 1-butanol has been studied using
Pd over composites of Mg-carbon material (HSAG) and
comparatively studied with the Pd/Mg.
Most of these works employing metal oxides describe a two-
phase process where ethanol is previously evaporated and
passed with a carrier gas through a solid fixed bed catalytic
system, where high reaction temperatures are required (usually
starting from 573 K).[18,21,22] The dehydrogenation is generally
considered the rate determining step explaining the high
reaction temperatures of the above mentioned references. For
this reason and in order to reduce the reaction temperature,
most of the reported catalytic systems combine an alkali metal
Results and Discussion
Structural and textural features. The X-ray diffraction patterns
of the commercial Mg(OH)2, the support MgO and the resulting
Pd/Mg catalyst are represented in Figure 2. The Mg(OH)2
diffractogram can be indexed as the hexagonal brucite phase
(JCPDS card 084-2164) while the thermal dehydration process
at 723 K gave rise to the characteristic periclase structure of
MgO (JCPDS card 45-946). The crystallite size of MgO was
estimated from line broadening of (200) diffraction peak
(2θ=42.9º) using Scherrer formula, while for Mg(OH)2 the peak
at 2 θ =58.8º corresponding to reflection plane (110) was used
to determine the mean size of crystallites. The values are
included in Table 1, together with some other parameters related
to textural and structural properties. As it was expected, the
thermal decomposition of Mg(OH)2 resulted in the formation of
MgO particles of a much smaller size which is in good
agreement with other reported observations.[38,39] This was also
reflected in the increase in the measured surface area (from 17
m2/g to 105 m2/g, Table 1). The subsequent incorporation of
metal NPs to the MgO support caused again some evident
changes in the structure and texture of the material (Figure 2,
Table 1). The bifunctional Pd/Mg catalyst showed the
characteristic peaks of brucite structure of Mg(OH)2, which
confirmed that aqueous impregnation transformed MgO back to
Mg(OH)2 as a consequence of its interaction with water.[40]
compound with
a
reduced transition metal as
a
hydrogenating/dehydrogenation agent. Some reported examples
have been previously mentioned for Cu,[22] Ni[28, 29] or Pd.[24, 25]
Carbon materials, such as high surface area graphites (HSAG)
and graphene derived materials,[32] are widely used as catalyst
supports with the aim of improving dispersion of the active
phase. Moreover, it has been interestingly reported in literature
that the strength of basic sites necessary for aldol condensation
reactions can be improved by dispersing the metal oxides onto
them.[33,34] In this sense, carbon materials represent ideal
supports for that goal thanks to their inert nature contrary to
other inorganic supports that normally have acid/basic sites that
can interfere with the catalytic performance giving rise to
undesired side reactions. Furthermore, the hydrophobic nature
of carbon materials could result advantageous taking into
account that water is a byproduct in this condensation reaction.
Although the Guerbet reaction with longer-chain alcohols has
been broadly studied, the ethanol coupling reaction is still a
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Although it can be observed that the main peak at ~43º
corresponding to residual MgO was also present in the sample.
It should be noted that the conversion of MgO into Mg(OH)2
during the impregnation procedure to incorporate the metal
caused a decrease in the surface area as well as in the pore
volume with respect to the starting MgO and this is in agreement
with other authors’ findings[39,40] (Table 1). The distinct peaks of
the Pd NPs (JCPDS card 001-1201) were detected at 2ϴ of 40.4,
46.8, 68.4 and 82.4º.
metallic Pd were detected and the diffractogram confirmed that
incorporation of Pd NPs during the synthesis procedure
converted the MgO in Mg(OH)2 as was previously observed with
the massive Pd/Mg catalyst.
Figure 3. XRD pattern of Pd/Mg-HSAG catalyst. For comparison purposes the
patterns of the supports Mg-HSAG and HSAG are also represented.
In Figure 4 the XRD patterns of Pd/Mg and Pd/Mg-HSAG
treated and not in helium at 723 K prior to hydrogen reduction at
573 K are shown. As can be observed for both catalysts, the
Figure 2. XRD pattern of Pd/Mg catalyst. For comparison purposes the
patterns of the starting Mg(OH)2 and MgO are also represented.
thermal treatment at 723
K after deposition of Pd NPs,
decomposed Mg(OH)2 to MgO as the reflections corresponding
to Mg(OH)2 disappear. The diffraction patterns show sharper
palladium peaks after thermal treatment, implying an increase of
particle size, which is more evident for the Pd/Mg-HSAG catalyst.
In Figure 3 the XRD pattern of the Pd/Mg-HSAG catalyst is
represented. As can be observed, reflections corresponding to
Table 1. Textural and structural parameters of supports and catalysts.
SUPPORTS
CATALYSTS
Mg(OH)2
SBET
Pore volume[a]
(cm3/g)
MgO loading[b]
(wt %)
Crystal size
(nm)
SBET
Pore volume
(cm3/g)
Sample
Sample
Crystal size[c]
(nm)
(m2/g)
(m2/g)
Mg(OH)2
MgO
17
0.11
0.39
0.56
-
-
44[c]
9.9[d]
20[d]
Pd/Mg
Pd/Mg-HSAG
Pd/Mg*
41
0.23
0.54
-
17
22
-
105
234
224
166
Mg-HSAG
45
Pd/Mg-HSAG*
215
-
-
[a] BJH desorption pore volume. [b] Determined from thermogravimetric analysis in air. Crystallite size was determined applying Scherrer formula to [c] peak at
2ϴ=58.8º corresponding to Mg(OH)2, [d] peak at 2ϴ=42.9º corresponding to MgO, * After thermal treatment in helium at 723K prior reduction with hydrogen.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Figure 4. XRD patterns corresponding to A) Pd/Mg and B) Pd/Mg-HSAG catalysts a) thermally treated at 723 K and reduced at 573 K and b) directly reduced at
573 K.
In Figure 5 are represented the TEM images of the Pd catalysts
and the corresponding histograms representing the particle size
distribution are depicted in Figure 6. TEM micrographs
confirmed the formation of metal nanoparticles during the
activation process with hydrogen. The Pd catalysts supported
over Mg have a Gaussian particle size distribution in which the
maximum of the peak coincides with the average Pd particle
size. The combination of metal and Mg gave rise to bigger
average Pd particle sizes (d=3.5-3.8, Fig.5b and 5d) than that of
Pd/HSAG (d=2.6 nm, Fig.5a). It is interesting to note that
Pd/HSAG shows non Gaussian particle size distribution and that
approximately 45% of the palladium particles has a diameter
smaller than 2 nm (Fig.6). Additionally, the TEM study confirms
the conclusions drawn from the X-ray study in which palladium
peaks of larger intensity were observed for catalysts treated in
helium at 723K. This thermal treatment has a remarkable effect
on the particle size and its distribution, considering that both
catalysts studied lose their Gaussian distribution and suffered an
increase in the average particle size, noticing that approximately
15% of the particles have a diameter superior than 9 nm.
types of basic sites have been established: strong basic sites
(Qdiff>90 KJ/mol), medium-strength basic sites (90>Qdiff>60
KJ/mol) and weak-strength basic sites (Qdiff<60 KJ/mol). The
total amount of basic sites and the corresponding percentage of
each type for every sample are summarized in Table 2. As
reported in Table 2, Mg(OH)2 and MgO have identical total
quantity of basic sites and quite similar strength distribution and
heats of adsorption, presumably due to Mg(OH)2 being
degasified at 623K for 10 hours prior to calorimetric
measurements, conditions that lead to the formation of MgO as
reveled in the XRD patterns (do not shown for the sake of
brevity). While exhibiting quite comparable evolution of the
differential heat of adsorption to MgO, a decrease in the total
amount of basic sites was observed for samples containing Pd
(Fig.7b), which may be due to the hydroxylation of MgO surface
that takes place during wetness impregnation with the acidic
aqueous palladium nitrate solution.[42] However, the calorimetry
study of Pd/Mg and Pd/Mg-HSAG after thermal treatment at
723K (Fig.7c) showed higher heats of adsorption at initial values
of surface CO2 coverage, which implies
a considerable
Figure 7a) represents the evolution of the differential heat of
adsorption of CO2 with the surface coverage of Mg(OH)2 and
MgO. The increase in the CO2 uptake is accompanied by a
continuous decrease in the differential heat from initial values
around 125 kJ/mol to ~40 kJ/mol, this last value is attributed to
the limit between chemical and physical for CO2 adsorption.[41]
The differential heats of adsorption indicate the basic strength of
strengthening of strong basic sites compared to the other
samples and even MgO. Probably this effect is due to metal-
support interaction at high temperatures, as it was previously
reported.[43-47] Particularly, Ueda et al[43-45] observed that the
addition of metals has a remarkable effect in the basic surface of
MgO, revealing that the incorporation of metal atoms causes a
distortion in the lattice, the expansion of Mg-O bond length and
the localization of electron on oxygen atom. As a result, the
combination of M-MgO exhibits higher basic properties.
the adsorption sites; thus this figure points out to
a
heterogeneous basic sites strength distribution. Attending to the
differences in the strength of the adsorption sites of CO2, three
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
a)
b)
c)
e)
d)
Figure 5. TEM images of a) Pd/HSAG, b) Pd/Mg, c) Pd/Mg-HSAG, d) Pd/Mg*, e) Pd/Mg-HSAG* catalysts. *Thermally treated in helium at 723K prior reduction
with hydrogen.
Figure 6. Particle size distribution of Pd determined from TEM micrographs of the different catalysts.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Figure 7. Differential heats of CO2 adsorption vs. coverage at 323 K. a) Mg(OH)2 and MgO b) Pd bifunctional catalysts c) Pd bifunctional catalysts*
Thermally treated in helium at 723K prior reduction with hydrogen.
Catalytic tests. First of all, a blank experiment without catalyst
showed no ethanol conversion after 6h at 503 K, which indicated
the reaction can barely take place in absence of a catalyst.
Table 3 summarizes the catalytic parameters obtained with the
Table 2. CO2 chemisorption capacities at 323 K and type of basic sites .
Total
µmol CO2/g
(Qdiff> 40
kJ/mol)
Basic sites distribution
support Mg-HSAG, as well as with the studied bifunctional
catalysts reduced at 573 K and those samples previously treated
in helium at 723K. In general, the main products detected were
1-butanol (ButOH), acetaldehyde (Ac), diethyl ether (DEE),
carbon monoxide and methane. Other products such as acetone
and 1,1-diethoxy ethane, ethyl acetate, 2-butanone, diethoxy
butane, butanal, 2-butenal, 2-butanol, 2-ethyl-1-butanol, 1-
hexanol, 1-octanol, 2-ethyl-1-hexanol were observed in low
quantities. The detection of 1,1-diethoxy ethane, ethyl acetate
and other higher alcohols (>C4), is in agreement with other
author reports for the continuous condensation of ethanol.[17,48-51]
% Strong
sites
% Medium-
strength
% Weak-
Sample
strength
(60>Qdiff>40
kJ/mol)
(Qdiff>90
kJ/mol)
(90>Qdiff>60
kJ/mol)
Mg(OH)2
MgO
Pd/Mg
Pd/Mg-HSAG
Pd/Mg*
Pd/Mg-HSAG*
487
488
374
295
411
214
51
58
47
47
61
68
38
32
41
38
29
22
11
10
12
15
10
10
Table 3. Catalytic activity and product selectivities obtained at 503 K and 50 bar after 24 h on stream.
Conv. (%)
Selectivity (%)
Average particle
Sample
ButOH
0
0
34
41
Ac
8
2
13
15
13
13
DEE
73
37
27
25
1
CO
-
10
4
CH4
Acetone
Others[a]
size (nm)
Mg-HSAG
Pd/HSAG
Pd/Mg
Pd/Mg-HSAG
Pd/Mg*
-
0.5
4.5
17
19
22
17
-
14
5
0
19
7
17
7
29
22
2.6
3.5
3.8
5.9
6.4
30
0.0
0.9
1
5
6
43
54
6
7
Pd/Mg-HSAG*
< 1
4
6
<1
Ac – Acetaldehyde, DEE – Diethyl ether, ButOH – 1-Butanol. [a] Other products also detected in low quantities: 1,1-diethoxy ethane, ethyl acetate, 1-Hexanol, 2-
butanona, diethoxy butane, 1-octanol, butanal, 2-butenal, 2-butenol, 2-butanol, 2-ethyl-1-butanol, 1-octanol, 2-ethyl-1-hexanol. * After thermal treatment in helium
at 723K prior reduction with hydrogen.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Figure 8. Conversion and 1-butanol selectivity through 24 hours of reaction. * After thermal treatment in helium at 723K prior reduction with hydrogen.
The bifuncional catalyst proved to be very stable during 24 hours
of reaction in terms of conversion and selectivity (Fig. 8) So,
after the maximum in conversion is reached at hour 2-3 the
catalysts suffer a slight deactivation but after 10 h the reaction
achieves the steady state conditions.
product can be obtained via i) decarbonylation and
dehydrogenation of 3-hydroxybutanal (R9), which also leads to
carbon monoxide and hydrogen formation, and ii) ketonization of
acetic acid (R10), product of ethyl acetate hydrolysis.[53] The
former of these two routes is unlikely taking place, as
dehydration of 3-hydroxybutanal should predominate. The
second route is plausible since carbon dioxide (the other product
of acetic acid ketonization) is detected, while acetic acid could
not be seen due to the low sensibility of the FID detector. The
acetone formation is favored in the Pd-HSAG catalyst, whereas
the presence of Mg in Pd/Mg-HSAG reduces considerably this
by-product (Table 3).
The first step in the Guerbet route mechanism is
dehydrogenation. According to literature,[19,20,56] dehydrogenation
of ethanol to acetaldehyde over metal oxides takes place over
basic medium-strength Mg2+-O2- pairs, which are predominant in
MgO with respect to Mg(OH)2. Aldol condensation of adsorbed
acetaldehyde is faster than dehydrogenation[31,57] and also
implies acid-base pairs as well as high quantity of basic sites.[19]
The presence of trace amounts of 2-butenal, crotyl alcohol and
butanal among the reaction products suggests that with our
catalysts the reaction pathway follows the Guerbet route.
The presence of the main reaction products can be explained by
the different reaction pathways represented in Fig. 9. Briefly,
acetaldehyde is the primary dehydrogenation (R1) product which
can undergo subsequent condensation reactions: reacting with
another two ethanol molecules giving rise to 1,1-diethoxy ethane
(R2) or with another acetaldehyde molecule yielding 3-
hydroxybutanal (R3), which readily dehydrates to 2-butenal (R4).
According to literature, the first condensation is the acid-
catalyzed acetylation reaction[52] while the second one is the
base-catalyzed
aldol
condensation.[1,40,53]
Subsequent
hydrogenations of 2-butenal give rise to the desired product 1-
butanol (R5-6).
1-butanol can also react with unreacted ethanol to form C6
alcohols like 1-hexanol[15] whose presence in our product stream
is higher as the higher the selectivity of 1-butanol is. So Pd/Mg,
Pd/Mg-HSAG, Pd/Mg* and Pd/Mg-HSAG* samples shows
correspondently 2, 4,
4
and 5% hexanol selectivities,
respectively, whereas the corresponding 1-butanol selectivities
are 34, 41, 43 and 54% (Table 3).
As can be seen in Table 3, Mg-HSAG showed negligible
conversion meanwhile Pd-HSAG exhibited no selectivity to 1-
butanol. Based on the catalytic tests, the metal/Mg combination
is compulsory to obtain good 1-butanol yields. The incorporation
of Pd to Mg highly improved the catalytic performance in the
condensation of ethanol, because transition metals have better
dehydrogenation properties than metal oxides.[58] Comparison of
Pd/Mg catalyst with that supported over HSAG (Pd/Mg-HSAG)
reveals that for a very similar conversion (17% and 19%,
respectively) the latter gives a higher selectivity towards 1-
butanol production 34% and 41%, respectively (Table 3). The
differences in the catalytic performance between Pd/Mg and
Pd/Mg-HSAG can not be ascribed to different basicity properties
determined by CO2 chemisorption as both catalysts have similar
percentage of strong surface basic sites (47%). Moreover, the
average Pd particle size of the two samples is very close, 3.5
Diethyl ether is the product of the acid-catalyzed direct ethanol
dehydration (R7)[28,53,54] while methane and carbon monoxide are
obtained by metal-catalyzed acetaldehyde decarbonylation
(R8).[55] The presence of Mg in the Pd catalysts reduces
considerably the formation of diethyl ether, methane and carbon
monoxide, as can be seen in Table 3, presumably because the
acid sites of the HSAG support are covered by Mg. Moreover,
the selectivity of diethyl ether drops remarkably with Pd catalysts
thermally treated in helium at 723 K prior to hydrogen reduction
(Table 3), which is in agreement with the XRD and calorimetry
studies where it is observed that the interaction of Pd and MgO
at this temperature involve a strengthening of the basic sites.
Thus the MgO covering the HSAG acid sites has an important
role in the decrease of acetone formation. This undesired
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
nm and 3.8 nm for Pd/Mg and Pd/Mg-HSAG, respectively.
Therefore, we can speculate about some additional effect of the
graphite support which has not been identify by the
characterization techniques used but is observed in the
increased catalytic selectivity towards 1-butanol. Likely the
graphite support makes possible to maximize the interaction
between Pd atoms and the strong basic sites of the MgO
originating cooperative sites for the tandem reaction. In addition,
the Pd/Mg-HSAG catalyst shows a 2-butenal selectivity of 3%
while the selectivity with Pd/Mg is 1%, confirming the greater
tendency of Pd/Mg-HSAG to catalyze reactions that require
coupled Pd atoms and stronger basic surface sites, since 2-
butenal is the product of the base catalyzed aldol condensation
of acetaldehyde, and an intermediate in the formation of 1-
butanol.
Figure 9. Reaction scheme proposed for ethanol condensation reaction in the liquid phase.[23]
Catalyst treated in helium at 723 K, in particular Pd/Mg-HSAG,
exhibited the best catalytic behavior in terms of 1-butanol
selectivity, which coincides with the tendency observed in Table
2 where these catalysts showed higher amount of strong basic
sites than their counterparts. The selectivity towards 1-butanol
increased from 34% with Pd/Mg up to 43% after thermal
treatment with helium (Table 3). This effect is even more
remarkable with Pd/Mg-HSAG, in which 1-butanol selectivity
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
increases from 41% up to 54%, and it is in accordance with
other results reported in bibliography using Ni–MgAlO
catalysts[59] that exhibit conversion and selectivity of 19 and 55%, this parameter is observed when catalysts are supported over
Figure 10 shows STY (%) for Pd-Mg catalysts based on the
number of strong basic sites per gram. Considerable increase of
respectively. In order to rationalize these findings, the type and
HSAG and for those treated in helium at 723K in comparison
quantity of active sites involved in the reaction must be analyzed. with their counterparts. These tendencies are in consonance
Concerning basic sites, it is worth mentioning that quite similar
basic strength distribution was found in both catalysts after
thermal treatment as revealed the Qdiff profiles (Fig.5c),
suggesting that both catalysts preserve the nature of active sites
(same strength of basic sites) for aldol-condensation step.
However, the total quantity of strong basic sites measured as
CO2 chemisorbed was 42% lower in Pd/Mg-HSAG than in Pd/Mg
after treatment with helium (Table 2) at 723 K.
Based on the microscopy results and catalytic tests, Pd particle
size is not critic to achieve Pd-Mg bifunctionality, but it possibly
has an effect in ethanol conversion. As it has already been
mentioned, it is often claimed dehydrogenation of ethanol to
acetaldehyde is the limit step in Guerbet reactions, so this would
explain the differences between the Pd catalysts. TEM images
(Fig. 5) show quite comparable average particle size for Pd/Mg
and Pd/Mg-HSAG catalysts as well as fairly similar particle size
distributions (Fig. 6) and ethanol conversion results (Table 3).
Nevertheless, the catalyst treated at 723 K showed higher
average particle size than their counterparts and considerably
different conversion values: Pd/Mg* has lower particle size (5.9
nm) than Pd/Mg-HSAG* (6.4 nm), whereas ethanol conversion
is higher for the former (22%) than the latter (19%). On the other
hand, it should be remarked that our Pd catalysts are able to
dehydrogenated ethanol at relativekly lower reaction
temperatures, 503 K, than in the case of non-metallic catalysts,
for instance hydroxyapatite, which work minimum at 573K.[60]
with 1-butanol selectivity results, which have great dependence
with the catalyst strong basic sites distribution.
Conclusions
The synthesis of 1-butanol from ethanol through the
acetaldehyde condensation has been studied in a fixed-bed gas
continuous reactor using bifunctional heterogeneous systems
based on Pd nanoparticles and magnesium oxide, which
incorporated basic properties, supported or not on high surface
area graphite. Pd/Mg catalysts supported on graphite were the
most selective to 1-butanol, while the thermal treatment in
helium at 723 K played an important role on the selectivity of the
catalysts for the Guerbet condensation reaction of ethanol.
Characterization of the active sites by microcalotrimetry of CO2
chemisorption (depicting basic sites provided by MgO) and TEM
analysis (reveling dehydrogenating/hydrogenating properties
related to Pd particle size) allowed us to explain the differences
displayed in the catalytic performance. Catalysts treated in
helium at 723 K exhibited the best catalytic behavior in terms of
1-butanol selectivity, in particular Pd/Mg-HSAG* material, whose
amount of strong basic sites is the highest among the catalysts
studied, whereas its larger Pd particle sizes have a slightly
negative impact in ethanol conversion. As the condensation of
ethanol to 1-butanol is a tandem reaction, ethanol conversion is
not only dependent on the size of Pd particles, but also on the
efficiency of aldol condensation and hydrogenation over stronger
MgO basic sites. Thus the compromise between both
parameters in the catalysts, after adequate thermal treatment,
gives rise to higher 1-butanol yields.
Experimental Section
Synthesis of the bifunctional catalysts
Preparation of the supports. The support MgO was obtained
by calcination at 723 K of a commercial Mg(OH)2 (Fluka,
SBET=17 m2/g). The Mg-carbon materials were prepared by a
deposition-precipitation method. The carbon support (HSAG)
was suspended by constant stirring in a solution containing the
necessary
quantity
of
nitrate
magnesium
precursor
(Mg(NO3)2·6H2O, Sigma Aldrich) to incorporate a 50 wt% of final
MgO loading to the support. The pH was adjusted to 11 with
NaOH and the final solution was kept stirring for 1 h. After that,
the resulting material was centrifuged several times, and dried at
383 K. Thereafter it was treated in a horizontal furnace using a
N2 flow with a heating rate of 5 K/min until 873 K, giving rise to
the support Mg-HSAG.
Figure 10. Site time yield (%) for bifunctional catalyst after 24 hours of
reaction. *After thermal treatment in helium at 723K prior reduction with
hydrogen.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Deposition of metallic Pd nanoparticles on the supports.
The incorporation of metallic palladium nanoparticles (NPs) was
done by the wetness impregnation technique using an aqueous
solution of Pd(NO3)2·2H2O (Sigma Aldrich) with the adequate
concentration to incorporate a metal loading of 5 wt% on the
support (MgO or Mg-HSAG). The resulting materials were dried
in air at 383 K overnight. Finally, the samples were treated at
573K for 1 h under flowing hydrogen (20 mLSTP/min).
mL/min with a HPLC-pump. Condensed products were collected
and analyzed by gas chromatography (GC) while gas products
were analyzed by on-line GC. The analysis of the collected
reaction products was carried out with a gas chromatograph
(Bruker Scion 436) equipped with a FID detector and configured
with a capilar column BR-Swax. The temperature program for
GC separation was hold at 50ºC for 1 minute, ramp to 150ºC at
15ºC/min and hold at 150ºC for 3.61 min (total=12 min). The
analysis of gaseous products was carried out with two gas
cromatographs (Varian CP 3380) equipped with FID and TCD
detectors and configured with a capilar column SupelQ Plot and
60/80 Carboxen-1000 column, respectively. The temperature
program for GC-FID separation was ramp from 35 to 51ºC at
2ºC/min and hold at 51ºC for 1 min, then ramp to 200ºC at
16ºC/min and hold at 200ºC for 7 min (total = 25 min). For the
GC-TCD, the temperature program was hold at 50ºC for 3
minutes, ramp to 150 at 5ºC/min and then to 225 at 20 ºC/min
(total = 26.75 min). Calibration of the possible reaction products
was done with commercial standards. After 24 hours of reaction
time, the ethanol conversion XEtOH is calculated using the
following equation:
Catalysts characterization
Structural properties of the supports and the catalysts were
determined by X-ray diffraction (XRD), using a Polycristal X’Pert
Pro PANalytical diffractometer with Ni-filtered Cu/K radiation (λ =
1.54 Å) operating at 45 kV and 40 mA. For each sample,
Bragg’s angles between 4◦ and 90◦ were scanned at a rate of
0.04º/s.
Values of specific surface area (SBET) and average volume
diameter of the supports as well as those of the catalysts were
determined by N2 adsorption-desorption isotherms at 77K in an
automatic volumetric adsorption apparatus (Micromeritics ASAP
2010). Prior to nitrogen adsorption, the samples were outgassed
for 5 h at 423 K.
Σ푖푛푖푚표푙푖
푋퐸푡푂퐻 (%) =
· 100
2푚표푙퐸′ 푡푂퐻 + Σ푖푛푖푚표푙푖
To obtain information on the particle size and distribution of
metal NPs on the different supports, the catalysts were
subjected to a detailed transmission electron microscopy (TEM)
study. TEM micrographs were obtained on a JEOL JEM-2100F
microscope at 200 kV. The samples were ground and
ultrasonically suspended in ethanol before deposition over a
copper grid with carbon coated layers. The mean diameter (d) of
Pd particles was calculated based on a minimum of 300
particles, using the following equation where ni is the number of
particles with diameter di:
Where ni is the number of carbon atoms of the product i, moli is
the number of mol of the product i, and mol´EtOH is the number of
mol of unreacted ethanol.
The selectivity to a specific product was defined as follows:
푛푖푚표푙푖
푆푖 (%) =
· 100
Σ푖푛푖푚표푙푖
The carbon balance (C%), which resulted higher than 90%, was
defined as:
Σ푖푛푖푚표푙푖 + 2푚표푙퐸′ 푡푂퐻
퐶 (%) =
· 100
2푚표푙퐸푡푂퐻
∑푛푖푑푖3
푑 =
In addition, site time yield STY (%) of butanol was defined as:
∑푛푖푑푖2
휇푚표푙 퐵푢푡푂퐻
⁄
ℎ · ꢀ
The amount and strength of the basic sites were determined by
CO2 adsorption using a volumetric equipment described in detail
in a previous work.[61] Successive doses of CO2 were introduced
into the system at 323 K to titrate the surface of the catalysts
until a final pressure of 7 Torr was reached. Prior to the
measurements, aliquots (200 mg) of the materials were
pretreated for 10 hours under vacuum at 623 K and outgassed
overnight. The differential heats of adsorption (Qdiff) were
obtained as the ratio between the exothermic integrated values
of each pulse and the adsorbed amount of CO2.
(
)
푆푇푌 % =
. 100
퐾퐽
푚표푙
휇푚표푙 퐶푂2
⁄
(> 90
)
ꢀ
Where µmol ButOH/h·g is the number of 1-butanol produced per
hour and per gram of catalyst, and µmol CO2/g (> 90 KJ/mol) is
the number of µmol CO2 absorbed per gram of catalyst with heat
of adsorption superior than 90 KJ/mol, which correspond to the
strong basic sites.
Acknowledgements
Catalytic reaction
We acknowledge financial support from the Spanish
Government (projects CTQ2017-89443-C3-2-R and CTQ2017-
89443-C3-3-R). Also CLO thanks the MECD of Spain for a FPU
predoctoral grant FPU16/05131.
The reactions were performed in a fixed bed reactor operated in
gas phase at 503 K and 50 bar using a stream of helium and
ethanol. Catalysts (0.5 g) were reduced in-situ with hydrogen at
573 K for 1 h before reaction. In order to study the influence of
calcination temperature, two additional experiments were carried
out with catalysts Pd/Mg and Pd/Mg-HSAG, therefore they were
also treated in helium at 723 K for 1 h prior reduction with
hydrogen. For all experiments, helium flow rate during reaction
was 50 mLSTP/min and ethanol flow rate was adjusted to 0.02
Keywords: bioethanol • biobutanol • basic sites • bifunctional
catalyst • metal nanoparticle
[1]
[2]
P. Dürre, Biotechnol. J. 2007, 2, 1525-1534.
P.S. Nigam. A. Singh. Prog. Energy Combust. Sci. 2011, 37 52-68.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
[3]
[4]
[5]
K. Brekke. Ethan. Today 2007, 36-39.
[36] J.I. Di Cosimo, V.K. Diez, C.R. Apestegula. App. Catal. A: Gen. 1996,
137 149-166.
P. Durre. Ann. N. Y. Acad. Sci. 2008, 1125 353-362.
E. Billing and D. R. Bryant, in Kirk-Othmer Encyclopedia of Chemical
Technology, John Wiley & Sons, Inc., 2000.
[37] L. Zhang, T.N. Pham, J. Faria, D. Santhanaraj, T. Sooknoi, Q. Tan, Z.
Zhao, D.E. Resasco. ChemSusChem 2016, 9, 736-748.
[38] S. E. Wankeki, R.M. J. Characterization of Porous Solids, Elsevier
Science Publishers B.V., Amsterdam - Printed in The Netherlands,
1998.
[6]
[7]
B. Ndaba, I. Chiyanzu, S. Marx. Biotechnol. Reports 2015, 8 1-9.
J.A. Posada, A.D. Patel, A. Roes, K. Blok, A.P.C. Faaij, M.K. Patel.
Bioresour. Technol. 2013, 135 490-499.
[8]
C. Angelici, B. M. Weckhuysen, P.C.A. Bruijnincx, ChemSusChem.
2013, 6 1595-1614.
[39] M. Estrada, V.V. Costa, S. Beloshapkin, S. Fuentes, E. Stoyanov, E.V.
Gusevskaya, Appl. Catal. A: Gen. 2014, 473 96-103.
[40] S.A.C. Carabinero, N. Bogdanchikova, A. Pestryakov, P.B. Tavares,
L.S.G. Fernandes, J.L. Figueiredo. Nanoscale Res. Let. 2011, 6 435-
440.
[9]
M. Guerbet, C. R. Acad. Sci. Paris 1899, 128, 1002-1004.
[10] R.E. Miller, G.E. Bennett, Ind. Eng. Chem. 1961, 53, 33–36.
[11] H. Aitchison, R.L. Wingad, D. F. Wass. ACS Catal. 2016, 6, 7125–7132.
[12] A. Corma, S. Iborra. Adv. Catal. 2006, 49, 239–302.
[13] D. Gabriëls, W.Y. Hernández, B. Sels, P. Van Der Voort, A.
Verberckmoes. Catal. Science Technol. 2015, 5 3876-3902.
[14] J.T. Kozlowski, R.J. Davis. ACS Catal. 2013, 3, 1588-1600.
[15] A. Galadima, O. Muraza. Ind. Eng. Chem. Res. 2015, 54, 7181-7194.
[16] T. Tsuchida, S. Sakuma, T. Takeguchi, W. Ueda. Ind. Eng. Chem. Res.
2006, 45 8634–8642.
[41] M. León, E. Díaz, S. Bennici, A. Vega, S. Ordónez, A. Auroux. Ind. Eng.
Chem. Res. 2010, 49 (8) 3663-3671.
[42] K. Refson, R. A. Wogelius, D. G. Fraser. Physical Review B, 1995, 52
(15) 10823-10827
[43] W, Ueda, H. Kurokawa, Y. Morooka, T. Ikawa. Chem. Lett., 1985, 6
819-820
[44] W. Ueda, T. Yokoyama, Y. Morooka, T. Ikawa. Chem. Lett., 1985, 14C
1059-1062
[17] J. Scalbert, F. Thibault-Starzyk, R. Jacquo, D. Morvan, F. Meunier. J.
Catal. 2014, 311 28–32.
[45] W. Ueda, T. Yokoyama, Y. Morooka, T. Ikawa. Chem. Lett., 1984, 1 39-
40
[18] A.S. Ndou, N. Plint, N.J. Coville. Appl. Catal. A: Gen. 2003, 251 337–
345.
[46] B. Wang, R. Yan, H. Liu. Applied surface science, 2012, 258 (22) 8831-
8836
[19] J.I. Di Cosimo, V.K. Díez, M. Xu, E. Iglesia, C.R. Apesteguía. J. Catal.
1998, 178 499-510.
[47] R. Ferrando, A. Fortunelli. Journal of Physics Condensed Matter, 2009,
21 (26)
[20] K. Kourtakis, A. Michael, L. E. Manzer. Patent WO2009026510A1,
2009.
[48] T. Riittonen, K. Eränen, P. Mäki-Arvela, A. Shchukarev, AR. Rautio, K.
Kordas, N. Kumar, T. Salmi, J. Mikkola. Renewable Energy. 2015, 74
369-378
[21] M. León, E. Díaz, S. Ordoñez. Catal. Today 2011, 164 436-442.
[22] M. León, E. Díaz, A. Vega, S. Ordoñez, A. Auroux. Appl. Catal. B:
Environ. 2011, 102 590-599.
[49] Pi. Dziugan, K.G. Jastrzabek, M. Binczarski, S. Karski, A.Witonska, B.
Kolesinska, Z.J.Kaminski. Fuel. 2015, 158 81-90
[23] I. C. Marcu, D. Tichit, F. Fajula, N. Tanchoux. Catal. Today 2009, 147
231-238.
[50] D.L.Carvalho, R.R.de Avillez, M.T.Rodrigues, L.E.P.Borges, L.G.Appel.
Applied Catalysis A: General. 2012, 415–416 96-100
[51] S. Zhuohua, V. Anaís Couto, B. Giovanni, S. Marc C.A., B. Giuseppe, C.
Catia, F.Francesco, B. Katalin. ACS Sustainable Chemistry and
Engineering, 2017, 5 (2) 1738-1746.
[24] I. C. Marcu, N. Tanchoux, F. Fajula, D. Tichit. Catal Lett. 2012, 143 23-
30.
[25] R. Arjona, J.L. Sanz, A. Corma, M. E. Domine. WO2014001595, 2014.
[26] G. Onyestyák. Periodica Polytechnica. Chemical Engineering, 2018, 62
(1) 91-96
[52] M.C. Abelló, M.F. Gómez, L.A. Arima. React. Kinet. Catal. Lett. 2001,
73 143-149.
[27] C. Yang, Z. Y. Meng. J. Catal. 1993, 142 37-44.
[28] T. Riittonen, E. Toukoniitty, D. K. Madnani, A.R. Leino, K. Kordas, M.
Zabo, A. Sapi, K. Arve, J. Wärna, J.P. Mikkola. Catalysts, 2012, 2 68-
84.
[53] E.V. Makshina, M. Dusselier, W. Janssens, J. Degreve, PA. Jacobs,
B.F. Sels. Chem. Soc. Rev. 2014, 43 7917-7953
[54] S. Hanspal, Z.D. Young, J. Tyler, P. Robert, J.Davis. J.Catal. 2017, 352
182-190
[29] T. Riittonen, K. Eränen, P. Mäki-Arvela, A. Shchukarev, A.R. Rautio, K.
Kordas, N. Kumar, T. Salmi, J.P. Mikkola. Renew. Energy, 2015, 74
369-378.
[55] M. Almohalla, E. Gallegos-Suarez, A. Arcoya, I. Rodríguez-Ramos, A.
Guerrero-Ruiz. Catalysis Science & Technology. 2016, 6 521-529
[56] T. Tsuchida, J. Kuboa, T. Yoshioka, S. Sakuma, T. Takeguchi, W.
Ueda. J. Catal. 2008, 259 183–189.
[30] J.H. Earley, R.A. Bourne, M.J. Watson, M. Poliakoff. Green Chem.
2015, 17 3018-3025.
[31] J.T. Kozlowhski, R.J. Davis. J. Energy Chem. 2013, 22 58-64.
[32] N.M. Julkapli, S. Bagher. Int. J. Hydrogen Energ. 2015, 40 (2) 948-979.
[33] L. Faba, E. Díaz, S. Ordoñez. Appl. Catal. B: Environ. 2013 142-143
387-395.
[57] C.L. Kibby, W. K. Hall. J. Catal. 1973, 29(1) 144-159.
[58] M.A. Landau, V.V. Shchekin. Bull. Acad. Sci. USSR, 1960 9(5) 885-887.
[59] P. Jifeng, Z. Mingyuan, H. Lei, L. Lin, P. Xiaoli, W. Aiqin, W. Xiaodong,
Z. Tao. J. Catal. 2016, 344 184-193
[34] L. Faba, E. Díaz, S. Ordoñez. ChemSusChem, 2013, 6 463-473.
[35] S. Chakraborty, P.E. Piszel, C. E. Hayes, R. T. Baker, W. D. Jones. J.
Am. Chem. Soc. 2015, 137 14264-14267.
[60] C.R. Ho, S. Shylesh, A.T. Bell. ACS Catal, 2016 6, 939−948.
[61] B. Bachiller-Baeza, I. Rodriguez-Ramos, A. Guerrero-Ruiz. Langmuir,
1998 14 (13) 3556–3564
This article is protected by copyright. All rights reserved.

10.1002/cssc.201801381
ChemSusChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
A high performance Pd/MgO
bifunctional catalyst for the
C. López-Olmos,[a] M.V. Morales,[b] A.
Guerrero-Ruiz,*[b, c] C. Ramirez-
Barria,[a,b] E. Asedegbega-Nieto,[b] I.
Rodríguez-Ramos[a, c]
continuous ethanol condensation to
1-butanol was designed. An optimum
compromise between the number of
metal atoms and the amount of
strong basic sites exposed on the
surface of the catalyst was found to
give the highest 1-butanol yield.
Page No. – Page No.
Continuous gas phase
condensation of bioethanol to 1-
butanol over bifunctional Pd/Mg and
Pd/Mg-carbon catalysts
Layout 2:
FULL PAPER
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
Text for Table of Contents
This article is protected by copyright. All rights reserved.