Ethanol to 1,3-Butadiene Conversion by using ZrZn-Containing MgO/SiO2 Systems Prepared by Co-precipitation and Effect of Catalyst Acidity Modification

Article abstract of DOI:10.1002/cctc.201600331

The conversion of ethanol to 1,3-butadiene (1,3-BD) has been investigated over ZrO₂- and ZnO-containing magnesia silica oxides prepared by a co-precipitation method at different Mg-to-Si molar ratios. The effect of reaction temperature and ethanol flow rate was investigated. The catalyst acidity was modified through the addition of alkali metals (Na, K and Li) to the final materials. Catalysts were characterised by nitrogen physisorption analysis, X-ray diffraction, scanning electron microscopy with energy dispersive X-ray, temperature-programmed desorption of ammonia, infrared spectroscopy and ²⁹Si/(⁷Li) NMR spectroscopy. The catalytic results showed that the controlled reduction of catalyst acidity allows ethanol dehydration to be suppressed, whilst increasing 1,3-BD selectivity. The best catalytic performance achieved 72 mol % for the combined 1,3-BD and acetaldehyde selectivity.

Full text of DOI:10.1002/cctc.201600331

DOI: 10.1002/cctc.201600331
Full Papers
Ethanol to 1,3-Butadiene Conversion by using ZrZn-
Containing MgO/SiO2 Systems Prepared by Co-
precipitation and Effect of Catalyst Acidity Modification
Simonꢀ Da Ros,[a, b] Matthew D. Jones,*[b] Davide Mattia,[c] Jose C. Pinto,*[a] Marcio Schwaab,[d]
Fabio B. Noronha,[e] Simon A. Kondrat,[f] Tomos C. Clarke,[f] and Stuart H. Taylor[f]
The conversion of ethanol to 1,3-butadiene (1,3-BD) has been
investigated over ZrO2- and ZnO-containing magnesia silica
oxides prepared by a co-precipitation method at different Mg-
to-Si molar ratios. The effect of reaction temperature and etha-
nol flow rate was investigated. The catalyst acidity was modi-
fied through the addition of alkali metals (Na, K and Li) to the
final materials. Catalysts were characterised by nitrogen physi-
sorption analysis, X-ray diffraction, scanning electron microsco-
py with energy dispersive X-ray, temperature-programmed de-
sorption of ammonia, infrared spectroscopy and 29Si/(7Li) NMR
spectroscopy. The catalytic results showed that the controlled
reduction of catalyst acidity allows ethanol dehydration to be
suppressed, whilst increasing 1,3-BD selectivity. The best cata-
lytic performance achieved 72 mol% for the combined 1,3-BD
and acetaldehyde selectivity.
Introduction
1,3-Butadiene (1,3-BD) is an important starting material used as
a monomer for the production of a variety of synthetic poly-
mers.[1] The polymerisation of 1,3-BD with itself and with other
olefin monomers represents its largest commercial use; exam-
ples include the production of styrene-butadiene-rubber (SBR),
polybutadiene (PB), styrene-butadiene latex (SBL), acrylonitrile-
butadiene-styrene (ABS), adiponitrile, nitrile rubber (NBR),
chloroprene and styrene-butadiene block copolymers (SBS and
SEBS).[2] 1,3-BD is currently produced from naphtha steam
crackers as a co-product of ethene manufacturing.[3,4] Besides
environmental concerns as a result of the use of petroleum-de-
rived hydrocarbons, the need for a new route to 1,3-BD is fur-
ther exacerbated owing to the possible future shortfall in
supply as a result of the changes of feedstock from naphtha to
ethane in the U.S.[5,6] The catalytic conversion of ethanol into
1,3-BD is an attractive alternative owing to the availability of
bioethanol, which is expected to significantly increase over the
next few years from the fermentation of sugar-rich waste ma-
terials (second generation bioethanol).[1,5,7,8] For example, in
Brazil alone, 23.4 billion litres of bioethanol were produced in
2014.[9]
[a] S. Da Ros, Prof. Dr. J. C. Pinto
Programa de Engenharia Quꢀmica/COPPE
Universidade Federal do Rio de Janeiro
Cidade Universitꢁria-CP: 68502, 21941-972, Rio de Janeiro (Brazil)
E-mail: pinto@peq.coppe.ufrj.br
The route most widely accepted to account for 1,3-BD pro-
duction from ethanol involves five consecutive reactions.[10–16]
Initially, ethanol is dehydrogenated to acetaldehyde. Then,
3-hydroxybutanal is formed from acetaldehyde self-aldolisa-
tion. Next, 3-hydroxybutanal dehydrates to crotonaldehyde,
which is reduced (Meerwein–Ponndorf–Verley (MPV) reduction)
with ethanol to produce crotyl alcohol and acetaldehyde. Final-
ly, crotyl alcohol is dehydrated to afford 1,3-BD. This mecha-
nism was initially suggested and experimentally supported by
Quattlebaum et al.,[17] who verified that higher 1,3-BD yields
were achieved when crotonaldehyde and ethanol were used
compared with a feed comprising acetaldehyde and ethanol,
with either a SiO2-based system or Ta2O5/SiO2 catalyst. With the
latter catalyst, the authors also observed that a feed compris-
ing of only acetaldehyde was converted into crotonaldehyde,
which was not observed in significant quantities when ethanol
and acetaldehyde were passed over the catalyst. This suggests
that crotonaldehyde was rapidly converted into crotyl alcohol
(presumably by an MPV process) and subsequently to 1,3-BD.
However, some reports have recently ruled out the aldol con-
[b] S. Da Ros, Dr. M. D. Jones
Department of Chemistry
University of Bath
Claverton Down, Bath BA2 7AY (UK)
E-mail: mj205@bath.ac.uk
[c] Dr. D. Mattia
Department of Chemical Engineering
University of Bath
Claverton Down, Bath BA2 7AY (UK)
[d] Dr. M. Schwaab
Departamento de Engenharia Quꢀmica
Universidade Federal do Rio Grande do Sul
90040040, Porto Alegre (Brazil)
[e] Dr. F. B. Noronha
Catalysis Division
National Institute of Technology
Av. Venezuela 82, 20081-312, Rio de Janeiro (Brazil)
[f] Dr. S. A. Kondrat, Dr. T. C. Clarke, Prof. Dr. S. H. Taylor
School of Chemistry, Cardiff Catalysis Institute
Cardiff University
CF10 3AT, Cardiff (UK)
Supporting information for this article can be found under http://
dx.doi.org/10.1002/cctc.201600331.
ChemCatChem 2016, 8, 1 – 12
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
densation as the main path, suggesting instead that crotyl al-
cohol is produced through the reaction between an activated
form of ethanol and acetaldehyde.[18,19]
comparison between wet-kneading and a co-precipitation
method, at the same Mg/Si ratio, has pointed to the former as
more suitable for ethanol conversion into 1,3-BD.[22,27] However,
the co-precipitation method may be an efficient preparation
procedure to produce catalysts with homogeneous properties
throughout their surface, therefore increasing the potential of
controlling the physical and chemical properties of the catalyst
and facilitating the determination of structure–activity relation-
ships. Despite that, a rigorous study using the co-precipitation
method has not been reported.
Owing to the specific features of the catalyst that are re-
quired for this cascade reaction, materials with multifunctional
properties have been studied, especially MgO/SiO2 systems.[2,5]
This is related to the fact that Mg-O pairs might act as Lewis
acid/Brønsted basic sites and the silanol moieties as a Brønsted
acid, which are necessary for ethanol dehydrogenation,[20] acet-
aldehyde condensation,[21] crotonaldehyde reduction and the
further crotyl alcohol dehydration to 1,3-BD. However, owing
to the presence of acid sites in these systems, ethanol dehy-
dration to ethene and diethyl ether are significant competitive
reactions. Efforts have been dedicated to the design of cata-
lysts able to suppress these undesirable parallel reactions. For
instance, the addition of metals and/or metal oxides based on
Cu, Zr, Zn and Ag to the MgO/SiO2 system has been shown to
be beneficial to the 1,3-BD yield.[13,21–23,28] In particular, a syner-
gic effect between ZrO2 and ZnO has been demonstrated.[12,23]
ZnO may support ethanol dehydrogenation and ZrO2 is ex-
pected to assist aldol condensation and crotonaldehyde reduc-
tion.[12,14,23,28–30] Conversely, in addition to catalyst composition,
the catalyst preparation method is of paramount importance
for 1,3-BD formation, as different acid/basic features may be
obtained depending on the synthesis conditions.[18,22,23,31]
Owing to this, different optimum Mg-to-Si molar ratios for 1,3-
BD formation have been reported, depending on the synthesis
procedure employed.[13,18,23]
Herein, we report for the first time an extensive investigation
using the co-precipitation method for the synthesis of magne-
sium silica oxide catalysts with different Mg-to-Si (Mg/Si) molar
ratios for ethanol conversion into 1,3-BD. The materials were
used as supports for ZnO and ZrO2 doping and the Mg/Si
molar ratio effect of materials containing ZrO2 and ZnO was
also evaluated. The effect of reaction temperature and ethanol
flow rate was also studied. In particular, the ethanol flow rate
was investigated to assess catalyst performance regarding 1,3-
BD productivity (in gBD gcatꢀ1 hꢀ1), a variable usually neglect-
ed[18,22,32,33] but that is important for industrial applications. As
it is well-known that the acidic/basic features play a key role
for this process, we have modified the catalyst acidity through
the addition of alkali metals (Na, K and Li) to the final materials.
The catalyst preparation method was further optimised by
modification of the number of calcination steps. As a result, it
is shown that the co-precipitation method is suitable to pre-
pare MgO/SiO2 systems that, after ZrO2 and ZnO addition, pro-
duce 1,3-BD productivities as high as those obtained by using
expensive Ag-containing systems.[2,14] Besides this, the catalyst
acidity modification through the alkali metal doping was suc-
cessful for the suppression of ethanol dehydration, allowing
the combined 1,3-BD and acetaldehyde selectivity to be in-
creased up to 72%. Catalysts were characterised by nitrogen
physisorption, X-ray diffraction, scanning electron microscopy
with energy dispersive X-ray, temperature-programmed de-
sorption of ammonia, infrared spectroscopy and 29Si/(7Li) NMR
spectroscopy.
Among the catalyst preparation procedures, different meth-
ods have been investigated such as physical mixtures of MgO
and SiO2,[11,13] wet-kneading,[10–13,22,23] sol-gel,[18,24] impregna-
tion[13] and co-precipitation.[22,25,26] Whereas it has been proven
that a physical mixture between precursor oxides is not suita-
ble for 1,3-BD formation, as the resultant catalysts show similar
features to single MgO and SiO2 phases, wet-kneading meth-
ods have been the most widely discussed in the literature with
less attention being dedicated to sol-gel and co-precipitation
methods.[5]
In the wet-kneading preparation procedure, MgO and SiO2
are usually mixed at the desired molar ratio in the presence of
water, then dried and calcined. Besides the specific features of Results and Discussion
the MgO and SiO2 precursors employed and their molar ratio,
The effect of the Mg/Si molar ratio
the amount of water, time, ageing temperature, drying and
calcination procedure represent some of the preparation steps
that might alter the catalyst’s behaviour.[5] As observed by
energy-dispersive X-ray (EDX) and TEM analyses, materials pre-
pared by wet-kneading are usually inhomogeneous in their
composition and morphology.[11,22,27] They are generally consti-
Magnesium silica oxide catalysts prepared at different Mg/Si
molar ratios were evaluated at 3258C, by using an ethanol
weight hourly space velocity (WHSV) of 0.62 hꢀ1. At this condi-
tion, ethanol conversion was typically lower than 20%, allow-
ing clear observation of the different catalytic properties
among the catalysts samples.
¨
¨
tuted of islands of MgO and SiO2, and a limited amount of
amorphous magnesia hydrosilicate phase.[22,27] These unique
characteristics are described as the key factor for the conver-
sion of ethanol into 1,3-BD, as an intrinsic basic/acid sites dis-
tribution is obtained on these materials.[22,27]
Figure 1 shows the selectivity profile of the main carbon-
containing products, obtained after 3 h of time on stream
(TOS), as a function of the Mg/Si molar ratio for (a) MgO/SiO2
systems and (b) MgO/SiO2 systems containing a fixed amount
of ZrO2 and ZnO, which has previously shown to be optimum
for SiO2 and acts as a comparison with our previous work.[12,23]
The profiles observed for ethene, diethyl ether (DEE) and acet-
aldehyde (AcH) show the same general trend as a function of
By using a sol-gel method, Ochoa et al.[18] observed that the
Mg/Si molar ratio affected the number, strength and distribu-
tion of the basic/acid sites, the surface area and the crystalline
structure of the catalysts, impacting the product distribution. A
&
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
with those reported here, in Figure 1, emphasising that the co-
precipitation preparation procedure has dramatically changed
the physicochemical properties of the catalyst compared with
the wet-kneading procedure.
Regarding the textural properties, samples with higher
Mg/Si molar ratios showed lower surface areas compared with
the 1:1 and 25:75 ratios, which may indicate the formation of
MgO particles in the pores of the samples, Table S1 in the Sup-
porting Information.[13,23] Diffraction patterns indicated samples
with amorphous features, the MgO periclase phase being ob-
served only in the MgO/SiO2 (95:5) sample (peaks at Bragg
angles of 37.08, 43.08, 62.48),[2] Figure S1. In particular, the
ZrZn/MgO-SiO2-1 system presented broad peaks (at 25–308,
33–398 and 58–628), characteristic of magnesium silicate hy-
drates.[34,35] As samples containing ZrO2 and ZnO have shown
similar PXRD patterns compared to the initial MgO/SiO2 sup-
port, ZrO2 and ZnO should be dispersed into the -Mg-O-Si- net-
work or their small loading was undetectable by PXRD. These
results highlight the effect of the preparation method on the
catalyst properties, as the MgO phase was observed even for
a Mg/Si molar ratio equal to 1 by using a sol-gel technique.[18]
A clear modification of silicon environments as a function of
the Mg/Si molar ratio was suggested by 29Si{1H} CP MAS NMR
experiments, Figure S2 (in the Supporting Information). Cata-
lysts with higher amounts of MgO, that is, the MgO/SiO2 (95:5)
and MgO/SiO2 (75:25) samples, presented a single resonance
with maxima around ꢀ71 ppm, indicating a high concentration
of Q1 species. Conversely, as the Mg/Si molar ratio was de-
creased, MgO/SiO2 (1:1) and MgO/SiO2 (25:75) samples, reso-
nance maxima were shifted down field, to approximately ꢀ87
and ꢀ94 ppm, indicating an increase in Q2 and Q3 spe-
cies.[2,34,36] In comparison with the wet-kneading method,[23] Q2
and Q3 species were observed at a Mg/Si molar ratio equal to
75:25, and Q2, Q3 and Q4 species were observed at a Mg/Si
molar ratio equal to 1. Thus, the co-precipitation method used
in this work seems to be more efficient in the formation of
Mg-O-Si linkages. These results are in agreement with a general
uniform distribution of elements inside catalyst particles as
verified for the ZrZn/MgO-SiO2-1 sample by scanning electron
microscopy with energy dispersive X-ray, Figure S3 (in the Sup-
porting Information). Even though a residual amount of Na
was observed from catalyst preparation, as will be more de-
tailed in the catalyst characterisation section, these Na traces
did not significantly affect catalyst activity.
Figure 1. Mg/Si molar ratio effect on the main carbon-containing reaction
products for a) MgO/SiO2 systems and b) MgO/SiO2 systems containing ZrO2
and ZnO (T=3258C, TOS=3 h, WHSV=0.62 hꢀ1, contact time=4 s).
the Mg/Si molar ratio. Whereas ethene and DEE selectivities
decreased as the Mg/Si molar ratio was increased, AcH selectiv-
ity showed a sharp increase, from less than 10% with the 1:1
species to over 40% with higher amounts of MgO. The de-
crease in the ethanol dehydration along with the increase in
the ethanol dehydrogenation as the Mg/Si molar ratio was in-
creased is in agreement with a reduction in the catalyst acidity,
as expected when the Mg/Si increases.[10,18]
On the other hand, interestingly, the effect of the Mg/Si
molar ratio on 1,3-BD selectivity was different between MgO/
SiO2 systems and MgO/SiO2 systems containing ZrO2 and ZnO.
Although a smooth rise in 1,3-BD selectivity was verified as the
Mg/Si molar ratio increased for pure MgO/SiO2 samples, Fig-
ure 1(a), the same trend was not observed for catalysts con-
taining ZrO2 and ZnO, Figure 1(b). Instead, a maximum in the
1,3-BD selectivity was obtained for the catalyst with the Mg/Si
molar ratio equal to one.
As the synergistic effect of ZrO2 and ZnO on the MgO/SiO2
system was more beneficial to 1,3-BD formation at the Mg/Si
molar ratio equal to one, this catalytic system, labelled as
ZrZn/MgO-SiO2-1, was selected for further investigation.
We have previously studied the effect of the Mg/Si molar
ratio for ZnII- and ZrIV-containing MgO/SiO2 systems, using the
wet-kneading method for the MgO/SiO2 preparation.[12] In line
with results presented in Figure 1, MgO/SiO2 systems prepared
by the wet-kneading method have shown a different 1,3-BD
selectivity behaviour upon the addition of ZrO2/ZnO to the cat-
alyst. However, although a maximum in the 1,3-BD selectivity
was observed at a Mg/Si molar ratio equal to 3:1 for the pure
MgO/SiO2 systems, an increase in 1,3-BD selectivity was ob-
served as the Mg/Si molar ratio was increased for the systems
prepared by incipient wetness.[12] These results are in contrast
Reaction temperature and WHSV effect
Initially, the reaction temperature and the WHSV effect were in-
vestigated to evaluate catalyst performance and afford more
insights into the kinetic mechanism. Mass transfer limitations
were excluded by the apparent activation energy estimation,[37]
Figure S4 (in the Supporting Information).
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 2. WHSV and reaction temperature effect on a) ethene and b) DEE se-
lectivities (ZrZn/MgO-SiO2-1, TOS=3 h).
Figure 3. WHSV and reaction temperature effect on a) AcH and b) 1,3-BD se-
lectivities (ZrZn/MgO-SiO2-1, TOS=3 h).
The effect of reaction temperature and WHSV on ethanol de-
hydration products is shown in Figure 2. It should be noted
that, in this work, higher WHSV conditions correspond to
a higher ethanol molar fraction in the gas phase and lower
contact times.
Figure 3(a). As ethanol dehydrogenation to AcH should involve
Brønsted basic sites and Lewis acid sites,[20,43] verified tenden-
cies suggest a high concentration of active sites for ethanol
dehydrogenation on the catalyst surface.
The further AcH transformation involves 3-hydroxybutanal
formation and its subsequent dehydration to crotonalde-
hyde.[5,12] Whereas the aldol coupling is an endergonic reaction
in the temperature range studied, which becomes more ender-
gonic as reaction temperature increases,[5] 3-hydroxybutanal
dehydration to crotonaldehyde is favourable in this tempera-
ture range and becomes more favourable as the temperature
increases.
As expected, an increase in reaction temperature increased
ethene selectivity, Figure 2(a), and decreased DEE selectivity,
Figure 2(b), as ethanol dehydration to ethene is endothermic
and to DEE it is exothermic.[7,38] Even though the mechanism
of DEE formation is still being discussed in the literature, for in-
stance, regarding whether it involves acid/base pairs,[39,40]
Brønsted acid sites and/or Lewis acid sites,[40] it is understood
that DEE formation should involve the reaction of the two
nearest chemisorbed ethanol moieties.[41] On the other hand,
ethene formation should occur through a concerted mecha-
nism, where the methyl hydrogen of the ethoxide species,
chemisorbed on a Lewis[41] or Brønsted acid site,[42] is abstract-
ed by the adjacent Brønsted basic site. Indeed, by using an
alumina catalyst, Arai et al.[41] verified a rise in DEE formation as
the concentration of surface ethoxide was increased, while
ethene formation was suppressed. Therefore, the effect of
WHSV on ethene and DEE selectivities observed in this work
suggests an increase in the concentration of chemisorbed eth-
anol species on the catalyst surface with a concurrent increase
in WHSV, as higher ethanol molar fractions were fed at higher
WHSV conditions.
Thus, as ethanol dehydrogenation to AcH is favoured ther-
modynamically as the reaction temperature increases,[5] an
excess of acetaldehyde in the system might contribute to fur-
ther AcH condensation, explaining its lower selectivities at
3758C.
The 1,3-BD selectivity behaviour presents different tenden-
cies as a function of reaction temperature and WHSV, Fig-
ure 3(b). At the higher temperature, the increase in 1,3-BD se-
lectivity as WHSV increased might be related to an excess of
AcH in the system, whereas at the lower temperature, the con-
tact time appears to affect 1,3-BD formation more dramatically.
This same effect of residence time was observed by Sushkevich
et al.,[14] who evaluated the WHSV effect by using an Ag/ZrO2/
SiO2 system at 3208C. In addition, additional experiments vary-
ing the WHSV at 3758C and using the same ethanol molar frac-
tion (the amount of catalyst remained constant and the gas
flow and ethanol feed rate were both varied), indicated the
Figure 3 shows the effect of reaction temperature and WHSV
on AcH and 1,3-BD selectivities. AcH shows a selectivity in-
crease for both reaction temperatures as the WHSV was raised,
&
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
same general tendency observed at 3258C, Figure S4(c). This
implies that the extra ethanol present in the higher WHSV pro-
cesses facilitates the full conversion to 1,3-BD, as observed
previously.[12]
Butene selectivity (1-butene, cis- and trans-2-butene) de-
creased smoothly as the WHSV was increased at both reaction
temperatures, Figure 4. The formation of butene from ethanol
Figure 5. WHSV and reaction temperature effect on ethanol conversion
(ZrZn/MgO-SiO2-1, TOS=3 h).
Figure 4. WHSV and reaction temperature effect on butene selectivity (ZrZn/
MgO-SiO2-1, TOS=3 h).
is thought to occur through deoxygenation of butanal pro-
duced from crotyl alcohol isomerisation.[5] Other studies sug-
gest butene as a butanol dehydration product, butanol being
produced from the hydrogenation of butanal, which, in turn,
might be obtained from the hydrogenation of the C=C double
bond of crotonaldehyde.[16,25] However, no traces of butanal or
butanol were observed in this work and butene may be a prod-
uct of 1,3-BD hydrogenation. Thus, the reduction in butene se-
lectivity with WHSV can be explained by a reduction in contact
time of 1,3-BD in the reactor.
Other minor products observed were ethane, acetone, pro-
pene and propane, their combined selectivities did not ach-
ieved more than 7% in all experiments. Traces of ethyl acetate,
C5 (pentadienes and its isomers) and C6 (hexadienes) com-
pounds were also identified in the output stream.
Regarding ethanol conversion, an increase was observed as
reaction temperature increased, whereas an ethanol conver-
sion drop was observed as WHSV was raised, Figure 5. The
higher WHSV increases the ethanol molar fraction in the gas
stream, at the same time shortening the contact time, factors
that might help to explain the reduced conversion.
Figure 6. Effect of WHSV on the yield of the main carbon-containing prod-
ucts at a) 3258C and b) 3758C. (ZrZn/MgO-SiO2-1, TOS=3 h).
Figure 6 shows the effect of reaction temperature and WHSV
on the yield of the main carbon-containing reaction products.
Firstly, it is worth noting that the ethanol conversion increase
resulting from the temperature rise has boosted the 1,3-BD
yield from 9–16% at 3258C to 27–30% at 3758C, within the
same range of WHSV between 0.3 and 1.2 hꢀ1
.
Moreover, even though the increase in WHSV has reduced
ethanol conversion, Figure 5, 1,3-BD yield has only reduced
slightly. Thus,
BD gcatꢀ1 hꢀ1) was obtained with WHSV, Figure 7.
As discussed by Makshina et al.,[13] catalytic data is usually
reported at low ethanol concentrations and 1,3-BD productivi-
a linear increase in BD productivity (in
g
Figure 7. Effect of WHSV and reaction temperature on 1,3-BD productivity
(gBD gcatꢀ1 hꢀ1) on ZrZn/MgO-SiO2-1 catalyst (TOS=3 h).
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
ties are usually too low to be industrially relevant. The 1,3-BD
productivity values achieved in this work suggest that the
ZrZn/MgO-SiO2-1 material prepared by co-precipitation is
a promising catalyst for 1,3-BD production, as it presents high
productivity with reasonable 1,3-BD selectivity compared with
other catalytic systems.[13,30] For instance, Janssens et al.[2] re-
ported productivity equal to 0.15 gBD gcatꢀ1 hꢀ1 over a catalyst
based on Ag/MgO-SiO2, at 4008C and WHSV=1.2 hꢀ1. Using
and AcH. Conversely, no significant change was observed to
the butene selectivity, which fluctuated around 10%.
A positive linear relation was verified when 1,3-BD and AcH
selectivities were considered as a function of Na2O content,
Figure 8(b), achieving 66% for the combined 1,3-BD and AcH
selectivities for the sample with the highest Na content. Also,
a negative linear relation was obtained when the ethene and
DEE selectivities were considered as a function of catalyst Na2O
content, Figure 8(b).
a
lower temperature (3208C), but much higher WHSV
(10.3 hꢀ1), Sushchevich et al.[14] observed 0.23 gBD gcatꢀ1 hꢀ1 over
a Ag/ZrO2/SiO2 system.
Moreover, a reduction in ethanol conversion was observed
as a function of Na content, Figure 8(b), resulting in lower 1,3-
BD yields and productivities, Table 1, entries 1–4. These ten-
dencies were confirmed by using a lower WHSV equal to
0.3 hꢀ1, entries 5–7, Table 1. However, it should be emphasised
that avoiding ethanol dehydration is the most important step
to attain high 1,3-BD yields, as ethene is the most thermody-
namically stable product.[5] The thermodynamics of the reac-
tion mixture in the case where ethanol dehydration occurs to-
gether with ethanol to 1,3-BD conversion was discussed by
Makshina et al.,[5] who have shown that the thermodynamic
yield of 1,3-BD is considerably lower when ethanol dehydration
is present.
Based on these results, we conclude that our ZrZn/MgO-
SiO2-1 catalyst has shown a suitable performance for ethanol
to 1,3-BD conversion. However, ethanol dehydration occurred
in significant quantities and, thus, we aimed to modify the
acidic features of this system through the addition of the alkali
metals Na, K and Li.
Catalyst acidity modification
The effect of added Na2O content on selectivity for the main
carbon-containing products is shown in Figure 8(a). The in-
crease in Na2O content decreased the selectivities to ethene
and diethyl ether, while increasing the selectivities to 1,3-BD
The suppression of acid sites through Na doping was recent-
ly studied on ZnxZryOz mixed metal oxides in the ethanol to
1,3-BD conversion.[30] An increase in the AcH and 1,3-BD selec-
tivity and a decrease in ethene selectivity were observed for
Na2O-containing samples. The catalytic results were rational-
ised through a reduction in the number of strong acid sites as
a result of Na doping, which was verified by temperature-pro-
grammed desorption of ammonia (NH3-TPD).
In this work, the acidity of the catalyst samples was investi-
gated through ammonia adsorption followed by TPD experi-
ments and IR measurements. Gases released during TPD ex-
periments were monitored by MS, Figure S5 (in the Supporting
Information). A dramatic reduction in the m/z signal attributed
to NH3, on going from the ZrZn/MgO-SiO2-1 system to samples
impregnated with sodium, was verified, indicating a large re-
duction in the number of acidic sites. A minor reduction in the
number of acidic sites as the quantity of sodium was increased
in the samples was also indicated by NH3-TPD experiments,
Table S2 (in the Supporting Information). Furthermore, sodium-
containing samples presented peaks with maximum intensity
at temperatures around 3808C, Figure S5(b), whilst the ZrZn/
MgO-SiO2-1 system exhibited peak maximum intensity around
5208C. Therefore, the addition of sodium resulted in a reduc-
tion in the concentration and strength of the acid sites.
IR measurements after NH3 adsorption supported the NH3-
TPD conclusions, presenting a clear reduction in the NH3
stretching intensity as the Na content was increased in the
samples, Figure S6 (in the Supporting Information). In particu-
lar, a weaker NH3 signal was observed for the precursor MgO-
SiO2-1 material, indicating that the ZrO2 and ZnO addition in-
creased the acidity on going from the MgO-SiO2-1 to the ZrZn/
MgO-SiO2-1 system. Thus, the sodium addition should have de-
activated the Lewis acid sites associated with ZrO2 and ZnO, as
will be discussed more later, contributing to the reduction of
the ethanol dehydrated fraction and ethanol conversion.
Figure 8. a) Effect of catalyst Na content on the selectivity of the main car-
bonaceous products and b) selectivities comparison with ethanol conver-
sion. (T=3758C, WHSV=0.62 hꢀ1, TOS=3 h).
&
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Table 1. Catalytic results for 3 h of time on stream and reaction temperature equal to 3758C.
Entry
Catalyst
WHSV
X
[%]
Selectivity [mol%]
1,3-BD yield[d]
[mol%]
1,3-BD productivity
[gBD gcatꢀ1 hꢀ1
]
1,3-BD
AcH
Ethene
DEE
Butene
1
2
3
4
5
6
7
ZrZn/MgO-SiO2-1
0.62
0.62
0.62
0.62
0.31
0.31
0.31
0.62
0.62
0.62
0.62
0.62
0.62
0.62
0.62
0.62
0.62
0.31*
0.62*
1.24*
40
26
24
24
52
35
36
46
28
32[c]
35.9
39.1
46.5
48.5
31.8
49.3
50.6
32.5
52.2
44.6
44.7
51.6
44.3
26.2
49.5
55.1
47.7
57.8
55.9
44.6
8.3
14.2
13.1
17.3
7.0
11.8
16.9
6.6
20.0
8.2
34.1
13.7
20.5
6.3
22.8
17.1
17.0
12.9
19.2
30.9
32.2
22.9
18.7
15.5
34.7
16.3
13.8
34.9
10.5
24.2
5.8
17.4
16.2
43.6
11.3
12.2
13.4
10.3
10.2
7.5
9.8
5.1
4.6
3.6
5.6
4.4
3.4
10.4
2.7
6.6
1.2
5.2
4.4
12.7
3.2
3.5
2.6
2.5
9.2
11.5
10.9
9.8
14.6
12.4
9.9
10.6
10.2
10.8
9.9
30.4
18.1
17.3
13.7
27.6
25.4
17.4
26.8
19.6
41.4
25.4
21.1
11.4
24.6
22.2
27.1
15.9
37.2
20.1
13.1
0.13
0.08
0.07
0.06
0.06
0.05
0.03
0.11
0.09
0.17
0.10
0.09
0.05
0.11
0.10
0.12
0.07
0.07
0.09
0.12
0.8-Na/ZrZn/MgO-SiO2-1
1.2-Na/ZrZn/MgO-SiO2-1
2-Na/ZrZn/MgO-SiO2-1
ZrZn/MgO-SiO2-1
1.2-Na/ZrZn/MgO-SiO2-1
2-Na/ZrZn/MgO-SiO2-1
Water/ZrZn/MgO-SiO2-1
1.2-Na/ZrZn/MgO-SiO2-1
ZrZn/MgO-SiO2-1
1.2-Na/ZrZn/MgO-SiO2-1
1.2-K/ZrZn/MgO-SiO2-1
1.2-Li/ZrZn/MgO-SiO2-1
ZrZn/MgO-SiO2-1
1.2-Na/ZrZn/MgO-SiO2-1
1.2-K/ZrZn/MgO-SiO2-1
1.2-Li/ZrZn/MgO-SiO2-1
1.2-K/ZrZn/MgO-SiO2-1
1.2-K/ZrZn/MgO-SiO2-1
1.2-K/ZrZn/MgO-SiO2-1
8
9[a]
10[b]
11[a,b]
12
6.9[c]
22
18
44
19
26
19
44
35
26
7.5
8.7
7.5
9.0
13
14[e]
15[a,e]
16[a,e]
17[a,e]
18[a,e,f]
19[a,e,f]
20[a,e,f]
7.9
13.7
10.3
7.5
2.78
2.36
5.0
[a] No calcination after alkali metal doping. [b] Acetaldehyde in the feed (8:2 ethanol to AcH feed ratio). [c] Acetaldehyde conversion was not calculated.
[d] Calculated with Equation (3). [e] No calcination before Zr and Zn addition. [f] The ethanol molar fraction was kept equal to 0.4.
As expected, however, besides acidity, the sodium addition
modified the samples’ textural properties as well. Although no
significant change in pore structure was indicated by N2 ad-
sorption–desorption isotherms, Figure S7 (in the Supporting In-
formation), the surface area was reduced with Na content,
Table S3 (in the Supporting Information), probably owing to
Na2O particle formation in the catalyst pores.[13,23] Thus, the
ethanol conversion reduction should be associated, mainly,
with the deactivation of acid sites and to the smaller access to
the active sites as a result of surface area reduction.
over, ethanol dehydration to ethene and DEE was further sup-
pressed upon AcH addition, this suppression being more pro-
nounced for the catalyst containing Na2O.
The acidity modification of the ZrZn/MgO-SiO2-1 system was
also investigated by using two other alkali metals, K and Li, en-
tries 12 and 13, Table 1. Both systems were effective in the sup-
pression of ethanol dehydration, presenting lower selectivities
to ethene and DEE compared with the starting ZrZn/MgO-
SiO2-1 material. A similar catalytic performance was observed
between samples doped with Na and K (compare entries 3 and
12). Conversely, the 1.2-Li/ZrZn/MgO-SiO2-1 sample presented
a lower 1,3-BD yield and productivity, which is in line with its
lower Brunauer–Emmett–Teller (BET) surface area, equal to
81 m2 gꢀ1, compared with the surface area of 1.2-Na/ZrZn/
MgO-SiO2-1 and 1.2-K/ZrZn/MgO-SiO2-1 samples, which are
equal to 219 and 243 m2 gꢀ1, respectively.
The effect of Na2O on the catalytic properties was confirmed
by an additional experiment in which the process of Na addi-
tion was imitated by using only water, entry 8, Table 1. The cat-
alytic performance was similar to that obtained for the ZrZn/
MgO-SiO2-1 (entry 1, Table 1) system. Thus, the additional calci-
nation step involved in the Na addition process, which is illus-
trated in the Figure S8 (in the Supporting Information) as Calci-
nation 3, does not appear to have affected the catalyst activity,
and therefore, the effect of the removal of this calcination step
was investigated, entry 9, Table 1. A slight increase in the 1,3-
BD selectivity, yield and productivity (compare entries 3 and 9)
was obtained. However, there is a difference in the surface
area, (219 versus 333 m2 gꢀ1) of the two materials, which may
contribute to the difference in performance.
Finally, the effect of calcination step removal before Zr and
Zn addition was investigated, entries 14–17. The removed cal-
cination step is illustrated in Figure S8 as Calcination 1. Regard-
ing the ZrZn/MgO-SiO2-1 system, a slight enhancement of the
ethene and diethyl ether selectivities was observed upon the
removal of the calcination step (compare entries 1 and 14,
Table 1), suggesting an increase in the acidity of the catalyst
and/or an improvement in the access to active acid sites, as
the BET surface area changed from 323 to 416 m2 gꢀ1 for the
ZrZn/MgO-SiO2-1 system prepared with and without the calci-
nation step before Zr and Zn addition, respectively. For sam-
ples containing Na2O, K2O and Li2O, the removal of the calcina-
tion step was beneficial to the 1,3-BD yield and productivity
(compare entries 3 and 15 for Na2O, entries 12 and 16 for K2O
and entries 13 and 17 for Li2O containing samples). As ob-
served for the ZrZn/MgO-SiO2-1 system, 1.2-K/ZrZn/MgO-SiO2-
In an attempt to further improve the catalytic data, the
effect of acetaldehyde in the feed was evaluated by using an
ethanol-to-acetaldehyde feed ratio equal to 8:2, as this condi-
tion has resulted in better catalytic performances in the litera-
ture.[12] Catalytic data is shown in Table 1, entries 10–11. As ex-
pected, the addition of AcH in the feed increased the amount
of 1,3-BD produced (compare yields and productivities of en-
tries 1 and 10, and between entries 9 and 11, Table 1). More-
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
1 and 1.2-Li/ZrZn/MgO-SiO2-1 samples also presented an in-
crease in the surface area as a result of the initial calcination
step removal, Table S4 (in the Supporting Information).
These results suggest a clear relationship among 1,3-BD for-
mation, acidic/basic concentration and its distribution on cata-
lyst surface, as 1,3-BD yield and productivity were strongly cor-
related with catalyst surface area for samples containing alkali
metals, Figure 9.
Catalyst characterisation
The elemental dispersion of selected samples at specific loca-
tions on the catalyst particles was investigated through scan-
ning electron microscopy with energy dispersive X-ray analysis.
Table 2 shows the average values determined. A general uni-
form distribution of Mg, Si, Zr, Zn and Na was observed.
However, a higher measurement scattering was observed for
Zr, Zn and Na (compare standard deviations, Table 2), which
may be due to a less uniform distribution of these compounds
and/or related to a lower analysis sensitivity associated with
the smaller concentration of these species. Moreover, the Na
doping was confirmed by analyses, which highlight the differ-
ent Na2O content among samples. The residual amount of Na
observed in the ZrZn/MgO-SiO2-1 catalyst might be a result of
an incomplete removal of Na during catalyst washing.
Element distribution on the catalysts surface was further
evaluated through EDX mapping analysis. Figure 10 shows the
element distribution for the 1.2-Na/ZrZn/MgO-SiO2-1 catalyst,
emphasising the homogeneous chemical distribution of
elements.
Element distribution for samples 0.8-Na/ZrZn/MgO-SiO2-
1 and 2-Na/ZrZn/MgO-SiO2-1 is shown in Figures S10–S11 in
the Supporting Information. The dispersion homogeneity was
further verified by an additional analysis by using a higher
magnification, Figure S12 (in the Supporting Information).
Figure 9. Relationship between surface area, 1,3-BD yield (left) and 1,3-BD
productivity (right) for 1.2-Na/ZrZn/MgO-SiO2-1, 1.2-K/ZrZn/MgO-SiO2-1 and
1.2-Li/ZrZn/MgO-SiO2-1 samples. Reactions performed as entries 3, 12, 13,
15–17, Table 1.
The effect of calcination step removal was further
7
investigated through Li MAS NMR spectroscopy. Fig-
Table 2. Elemental dispersion of catalyst samples in wt%.[a]
ure S9(a) (in the Supporting Information) shows the
spectra for samples 1.2-Li/ZrZn/MgO-SiO2-1 and 1.2-
Li/ZrZn/MgO-SiO2-1 prepared with only one calcina-
tion step after ZrO2 and ZnO addition. Even though
a similar chemical shift was observed between the
samples, the different lineshape, broader in the half
width for the sample with higher number of calcina-
tion steps, Figure S9(b), and the different intensities
of the spinning sideband distribution indicated
a modification in the local environment of the lithium
Sample
Mg
Si
Zr
Zn
Na
ZrZn/MgO-SiO2-1
32.1ꢁ0.5
31.8ꢁ1.0
31.5ꢁ0.9
31.8ꢁ0.4
59.6ꢁ2.2
56.7ꢁ2.0
55.8ꢁ0.5
56.7ꢁ0.5
6.3ꢁ2.0
7.6ꢁ2.5
7.2ꢁ0.3
6.5ꢁ0.7
1.4ꢁ0.4
1.7ꢁ0.4
1.9ꢁ0.6
1.9ꢁ0.3
0.6ꢁ0.2
2.2ꢁ0.2
3.7ꢁ0.5
3.2ꢁ0.1
0.8-Na/ZrZn/MgO-SiO2-1
1.2-Na/ZrZn/MgO-SiO2-1
2-Na/ZrZn/MgO-SiO2-1
[a] Values of Mg, Si, Zr, Zn and Na were normalised to 100 and represent a dispersion
measure only.
nuclei. Thus, the reduction of the number of calcination steps,
besides largely affecting surface area, in agreement with other
studies,[44] also produces different structural ordering between
samples.
The metal loadings were confirmed by inductively coupled
plasma optical emission spectroscopy (ICP-OES). Na loadings of
0.98 and 1.96, Zn loadings of 0.39 and 0.37, and Zr loadings of
1.13 and 1.07 wt% were observed for samples 1.2-Na/ZrZn/
MgO-SiO2-1 and 2-Na/ZrZn/MgO-SiO2-1, respectively.
The catalyst 1.2-K/ZrZn/MgO-SiO2-1 was further investigated
regarding WHSV effects, entries 18–20, Table 1. A similar kinetic
behaviour to the ZrZn/MgO-SiO2-1 system at 3258C was ob-
served, as higher contact times (obtained at lower WHSVs)
were beneficial to 1,3-BD selectivity and yield.
Figure S13 (in the Supporting Information) shows the PXRD
patterns for MgO/SiO2 and the metal-doped variants. Similar
PXRD patterns were observed for K2O- and Li2O-containing
samples, Figure S14 (in the Supporting Information), for the
1.2-Na/ZrZn/MgO-SiO2-1 sample prepared with no calcination
after Na doping, Figure S15 (in the Supporting Information),
and for samples prepared with no calcination step before Zr
and Zn addition, Figure S16 (in the Supporting Information).
Thus, samples presented a common amorphous structure with
the three broad peaks (at 25–308, 33–398 and 58–628) charac-
teristic of magnesium silicate hydrates.[34,35]
Therefore, the acidity modification of the ZrZn/MgO-SiO2-
1 system prepared by co-precipitation, through the addition
the alkali metals, specially by using K2O, seems to be a promis-
ing—and cheap—catalyst preparation method to maximize
1,3-BD formation from ethanol conversion. Firstly, ethanol de-
hydration might be avoided, which is a thermodynamic re-
quirement to achieve higher 1,3-BD yields. In addition, uncon-
verted ethanol and the acetaldehyde produced might be recy-
cled in the process, overcoming the lower ethanol conversion
obtained with these systems.
The local environment of the silicon atoms on the catalyst
surface was investigated through 29Si MAS NMR spectroscopy,
Figure S17 (in the Supporting Information). Figure S17(a)
&
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Table 3. Summary of 29Si MAS NMR data. Numbers represent the relative
proportion of each silicon environment.
Catalyst sample
Q2 (ꢀ87 ppm)
Q3 (ꢀ95 ppm)
i) MgO-SiO2-1
ii) ZrZn/MgO-SiO2-1
iii) 0.8-Na/ZrZn/MgO-SiO2-1
iv) 1.2-Na/ZrZn/MgO-SiO2-1
v) 2-Na/ZrZn/MgO-SiO2-1
21.9
27.4
28.6
27.0
26.3
78.1
72.6
71.4
73.0
73.7
tion,[45,46] Figure S18 (in the Supporting Information). Subtle dif-
ferences in the CHCl3 stretching intensities were observed
among samples. A comparison with MgO suggested that sam-
ples containing higher alkali metal content (1.2 and 2 wt%)
may have a slightly higher basicity compared with the precur-
sor MgO/SiO2 and the ZrZn/MgO-SiO2 system.
Therefore, the main effect of alkali metal doping may be re-
lated to their interaction with Lewis acid sites associated with
ZrO2 and ZnO species, as also suggested by NH3-TPD experi-
ments and IR measurements from NH3 adsorption.
Conclusions
In this work, the effect of the Mg-to-Si molar ratio was investi-
gated in the synthesis of magnesia silicate oxides prepared by
co-precipitation for the ethanol to 1,3-butadiene conversion.
Catalysts were used as support for ZrO2 and ZnO and the etha-
nol conversion was studied in a wide range of WHSVs by using
two reaction temperatures. 29Si MAS NMR data suggested that
the co-precipitation method was more efficient in the forma-
tion of Mg-O-Si linkages compared with the traditional wet-
kneading. The Mg/Si molar ratio equal to 1 was more suitable
to 1,3-BD formation, whereas higher Mg/Si molar ratios pro-
duced more acetaldehyde. High 1,3-BD productivities (in
Figure 10. Elemental mapping of the 1.2-Na/ZrZn/MgO-SiO2-1 catalyst.
shows the spectra for catalysts i) MgO-SiO2-1, ii) ZrZn/MgO-
SiO2-1, iii) 0.8-Na/ZrZn/MgO-SiO2-1, iv) 1.2-Na/ZrZn/MgO-SiO2-
1 and v) 2-Na/ZrZn/MgO-SiO2-1. Spectra are in agreement with
the PXRD patterns, the two broad resonances with maxima
around ꢀ87 and ꢀ95 ppm suggest the presence of -Mg-O-Si-
linkages. Moreover, the lack of signal at ꢀ110 indicates the ab-
sence of silica in the catalyst structure, or the presence of an
amount too small to be detected.[34] Similar signals with chemi-
cal shift between ꢀ85 and ꢀ89 ppm and between ꢀ92 and
ꢀ99 ppm were already reported for magnesium silicate sys-
tems and they were attributed to Q2 and Q3 species, respec-
tively, as Si*(OMg)(OSi)2(OH) and Si*(OMg)(OSi)3.[2,34,36] 29Si MAS
spectra were confirmed by cross-polarization (CP) experiments,
Figure S17(b). A shoulder at ca. ꢀ80 ppm was observed for all
samples, which might be related to Q1 species.
29Si MAS spectra were fitted by a function containing two
Gaussian distributions to compute spectra areas and obtain
the relative proportion of each silicon environment, the data
being summarised in Table 3. Only subtle differences were ob-
served between the relative proportion of Q2 and Q3 species,
which are probably due to the intrinsic experimental analysis
fluctuation. This data indicates that the catalytic results are not
explained through the interaction of Na species with surface
Brønsted acidic silanol moieties.
g
BD gcatꢀ1 hꢀ1) were obtained with the ZrZn/MgO-SiO2-1 system,
results that might be associated with the homogeneity of the
catalyst properties. Catalytic results supported the usual kinetic
route of ethanol to 1,3-BD conversion involving acetaldehyde
condensation. The catalyst acidity was modified through the
addition of alkali metals (M=Na, Li and K) to the final materi-
als. This process resulted in a decrease in the fraction of etha-
nol dehydrated, boosting 1,3-BD selectivity. A positive linear re-
lation was obtained for the combined 1,3-BD and acetaldehyde
selectivities as a function of the catalyst Na content. Further
catalyst optimisation was performed through the reduction of
calcination steps in the catalyst preparation, resulting in higher
surface areas, 1,3-BD yields and productivities. In particular,
a strong correlation between surface area, 1,3-BD yield and
productivity was observed for 1.2-M/ZrZn/MgO-SiO2-1 samples.
Thus, efforts should be dedicated to increasing the surface
area of these systems, keeping the acid/basic site distribution
constant. The best catalytic results were obtained with the
1.2-K/ZrZn/MgO-SiO2-1 material, achieving 72 mol% for the
combined selectivity of 1,3-BD and acetaldehyde, at a reasona-
ble 1,3-BD yield and productivity level. Therefore, as uncon-
verted ethanol and acetaldehyde may easily be recycled in the
Furthermore, basicity features of the catalyst samples were
investigated through IR measurements from CHCl3 adsorp-
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
state magic angle scattering (MAS) NMR spectroscopy was per-
formed by using a Varian VNMRS 400 MHz spectrometer, operating
at a resonance frequency of 79.44 MHz with a spinning rate of
6 kHz. 1000 scans were accumulated with a recycle time of 60 s,
the pulse length being 4.5 ms. The 29Si{1H} CP MAS NMR spectra
were recorded with the same spectrometer. 4000 scans were accu-
mulated with a recycle time of 1 s. The CP contact time was
3.0 ms. The 29Si chemical shifts are referenced to tetramethylsilane.
The acidity of the samples were determined by temperature-pro-
grammed desorption of ammonia (NH3-TPD) with a Setsys Evolu-
tion TGA Setaram system coupled with an in-line mass spectrome-
ter, OmniStarꢃ Pfeiffer Vacuum Quadrupole, for the measurement
of the outgas composition. The release of ammonia (m/z=15) was
monitored. The signal m/z equal to 15 was used to avoid interfer-
ence by the fragmentation of water molecules. Samples were ex-
posed to NH3 for 48 h at room temperature before the TPD experi-
ments. Pure argon, 100 mLminꢀ1, was used as the sweep gas.
Before starting the analysis, the analytical chamber was purged of
ambient air by using an argon flow at 200 mLminꢀ1 for 40 min.
The NH3-TPD analyses were started by heating the sample at
108Cminꢀ1 from room temperature to 7008C and maintaining that
temperature for 0.5 h, under argon. In situ IR spectra were record-
ed with a PerkinElmer Frontier spectrometer. Measurements were
process, the x-K/ZrZn/MgO-SiO2-1 system prepared by co-pre-
cipitation is a promising material that deserves more investiga-
tion to maximize 1,3-BD production. Catalyst acidity modifica-
tion was further confirmed by NH3-TPD and IR measurements
from NH3 adsorption, but 29Si MAS NMR data indicated that
the role of the alkali metal in the catalyst structure was not re-
lated to its interaction with Brønsted acidic silanol moieties. As
IR measurements from CHCl3 adsorption indicated only subtle
differences between catalyst basicity, the main effect of alkali
metal doping should be associated with a selective deactiva-
tion of Lewis acid sites related to ZrO2 and ZnO.
Experimental Section
Catalyst preparation
In a typical synthesis, catalysts with the Mg/Si molar ratios equal to
25:75, 50:50, 75:25 and 95:5 were prepared by co-precipitation.
For the 50:50 material, SiO2 (9.01 g, Sigma–Aldrich, 99.8%) was dis-
solved in NaOH solution (100 mL of 1.2m, Sigma–Aldrich, 99%).
The mixture was heated to between 60 and 808C under vigorous
stirring until complete SiO2 dissolution, the solution was cooled
performed by accumulating 15–30 scans at a resolution of 4 cmꢀ1
.
and Na2CO3 (42.4 g, Sigma–Aldrich, 99.9%) added.
A
CHCl3 was used as a molecular probe for the basicity evaluation.[45]
Samples were exposed to CHCl3 at 208C. MgO (Sigma–Aldrich) was
evaluated for comparison.
Mg(NO3)2·6H2O solution (Sigma–Aldrich, 99%) was added drop-
wise into this mixture whilst stirring at 258C (38.85 g of
Mg(NO3)2·6H2O in 200 mL). The pH was maintained at 10.5 by
adding appropriate quantities of 1.2m NaOH solution and, at the
end of the process, the solution volume was adjusted to 600 mL
with deionized water. The resultant mixture was stirred for 2 h
before ageing for 22 h at 258C. Finally, the mixture was filtered
and washed with 7.5 L of hot water. The precipitate was dried in
static air at 808C for 24 h before grinding. Materials were calcined
Catalytic reactions
The catalytic tests were carried out in a flow quartz packed-bed re-
actor at atmospheric pressure. Argon was used as the carrier gas
(8 mLminꢀ1). The ethanol weight hourly space velocity (WHSV) was
varied within 0.3–2.4 hꢀ1 through the modification of the ethanol
flow rate, keeping the catalyst mass and carrier gas flow rate fixed.
The WHSV range investigated corresponded to ethanol molar frac-
tions between 0.41 and 0.85. The contact time (calculated as the
ratio between the catalyst volume and the total gas flow at the re-
action temperature) ranged from 1.3 to 5.5 s. The reaction temper-
ature was within 325–3758C. The exhaust gases were analysed
after 3 h of time on stream (TOS) by GC-MS with an Agilent 7890A
instrument with a HP-PLOT/Q, 30 m long, 0.53 mm diameter
column equipped with FID/MS detectors. The GC was calibrated as
detailed elsewhere.[13] Carbon balances were typically better than
80%. Carbon balances and reaction conditions are summarized in
the Table S5 (in the Supporting Information).
Ethanol conversion (X), selectivity (S) and yield were computed as
in Equations (1), (2) and (3), respectively, where NEtOH,in and NEtOH,out
represent the number of mols of ethanol that were added and col-
lected, respectively, Ni denotes the number of mols of the product
i, NP is the total number of products and r is the ratio between the
number of carbons of the product i and of ethanol. For 1,3-BD, for
instance, r is equal to 2.
in air at 5008C for 4 h, using a heating rate of 58Cminꢀ1
.
To produce materials with 1.5 wt% and 0.5 wt% of ZrIV and ZnII, re-
spectively, ZrO(NO3)2·H2O (0.57 g, Sigma–Aldrich, 99%) and
Zn(NO3)2·6H2O (0.24 g, Sigma–Aldrich, 98%) were dissolved in
water (50 mL), and the solution was added to 10 g of the MgO/
SiO2 system. This was stirred until the mixture was completely dry.
Finally, the solid was calcined in air at 5008C for 5 h. For the Na
doping, the appropriate volume of 1.2m NaOH solution was added
to the final catalyst drop-wise. The mixture was kept under stirring
for 1 h at 258C before drying at 808C for 5.5 h and finally calcined
at the same previous condition. KOH (Sigma–Aldrich, 90%) and
LiOH·H2O (Alfa Aesar, 99%) were used instead of NaOH for compar-
ison. The catalyst preparation procedure is illustrated in the
Scheme 1 in the Supporting Information. Samples were labelled as
y-M/ZrZn/MgO-SiO2-x, where y denotes the content of the alkali
metal M in wt% and x represents the Mg/Si molar ratio.
Catalyst characterisation
SEM images and SEM/EDX mapping were carried out with
a JEOL6480LV at 5–25 kV. Energy-dispersive X-ray spectroscopy
(EDX) was carried out in situ during SEM analysis. At least five dif-
ferent spots were selected throughout the images, to evaluate the
homogeneity of elements within catalyst particles. Measurements
of static adsorption of N2 at ꢀ1968C were obtained by using a Mi-
cromeritics 3Flex instrument. Samples were degassed at 1508C
under vacuum for 2 h prior to analysis. Powder X-ray diffraction
(PXRD) was performed with a BRUKER D8-Advance diffractometer
using CuKa (l=1.5406 ꢂ) radiation. Intensities were measured with
a 0.028 step size and a measuring time of 0.3 s per point. 29Si solid-
NEtOH;in ꢀ NEtOH;out
Xð%Þ ¼
ꢂ 100
ð1Þ
ð2Þ
ð3Þ
NEtOH;in
Ni
Sið%Þ ¼ P
yieldð%Þ ¼
ꢂ 100
NP
i¼1 Ni
r ꢃ Ni
ꢂ 100
NEtOH;in
&
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
10
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[19] A. Chieregato, J. Velasquez Ochoa, C. Bandinelli, G. Fornasari, F. Cavani,
M. Mella, ChemSusChem 2015, 8, 377–388.
Acknowledgments
[20] J. I. Di Cosimo, V. K. Diez, M. Xu, E. Iglesia, C. R. Apesteguꢀa, J. Catal.
1998, 178, 499–510.
[21] V. V. Ordomsky, V. L. Sushkevich, I. I. Ivanova, J. Mol. Catal. A 2010, 333,
85–93.
[22] C. Angelici, M. E. Z. Velthoen, B. M. Weckhuysen, P. C. A. Bruijnincx,
ChemSusChem 2014, 7, 2505–2515.
[23] M. Lewandowski, G. S. Babu, M. Vezzoli, M. D. Jones, R. E. Owen, D.
Mattia, P. Plucinski, E. Mikolajska, A. Ochenduszko, D. C. Apperley, Catal.
Commun. 2014, 49, 25–28.
The authors thank CNPq (Conselho Nacional de Desenvolvimento
Cientꢀfico e Tecnolꢂgico, Brazil) for supporting this research and
providing scholarships. We wish to thank the EPSRC Solid-state
NMR facility in Durham (Dr David Apperley). We thank the Uni-
versities of Bath and Cardiff for involvement in the GW4 partner-
ship.
[24] Z. Han, X. Li, M. Zhang, Z. Liu, M. Gao, RSC Adv. 2015, 5, 103982–
103988.
Keywords: 1,3-butadiene
· ethanol · green chemistry ·
[25] M. Leꢅn, E. Dꢀaz, S. OrdꢅÇez, Catal. Today 2011, 164, 436–442.
[26] M. Leꢅn, E. Dꢀaz, A. Vega, S. OrdꢅÇez, A. Auroux, Appl. Catal. B 2011,
102, 590–599.
heterogeneous catalysis · silica magnesia catalyst
[27] C. Angelici, M. E. Z. Velthoen, B. M. Weckhuysen, P. C. A. Bruijnincx,
Catal. Sci. Technol. 2015, 5, 2869–2879.
[28] O. V. Larina, P. I. Kyriienko, S. O. Soloviev, Catal. Lett. 2015, 145, 1162–
1168.
[29] Y. Sekiguchi, S. Akiyama, W. Urakawa, T. Koyama, A. Miyaji, K. Motokura,
T. Baba, Catal. Commun. 2015, 68, 20–24.
[1] D. Cespi, F. Passarini, I. Vassura, F. Cavani, Green Chem. 2016, 18, 1625–
1638.
[2] W. Janssens, E. V. Makshina, P. Vanelderen, F. De Clippel, K. Houthoofd,
S. Kerkhofs, J. A. Martens, P. A. Jacobs, B. F. Sels, ChemSusChem 2015, 8,
994–1008.
[3] W. C. White, Chem.-Biol. Interact. 2007, 166, 10–14.
[4] G. Odian, Principles of Polymerization, 4th ed., Wiley, New York, 2004,
pp. 529–533.
[5] E. V. Makshina, M. Dusselier, W. Janssens, J. Degrꢄve, P. A. Jacobs, B. F.
Sels, Chem. Soc. Rev. 2014, 43, 7917–7953.
[6] P. C. A. Bruijnincx, B. M. Weckhuysen, Angew. Chem. Int. Ed. 2013, 52,
11980–11987; Angew. Chem. 2013, 125, 12198–12206.
[7] C. Angelici, B. M. Weckhuysen, P. C. A. Bruijnincx, ChemSusChem 2013, 6,
1595–1614.
[8] J. A. Posada, A. D. Patel, A. Roes, K. Blok, A. P. C. Faaij, M. K. Patel, Biore-
sour. Technol. 2013, 135, 490–499.
[9] http://www.afdc.energy.gov/data/10331, 2016.
[10] H. Niiyama, S. Morii, E. Echigoya, Bull. Chem. Soc. Jpn. 1972, 45, 655–
659.
[11] S. Kvisle, A. Aguero, R. P. A. Sneeden, Appl. Catal. 1988, 43, 117–131.
[12] M. D. Jones, C. G. Keir, C. Di Iulio, R. A. M. Robertson, C. V. Williams, D. C.
Apperley, Catal. Sci. Technol. 2011, 1, 267–272.
[30] R. A. L. Baylon, J. Sun, Y. Wang, Catal. Today 2016, 259, 446–452.
[31] T. De Baerdemaeker, M. Feyen, U. Mꢆller, B. Yilmaz, F.-S. Xiao, W. Zhang,
T. Yokoi, X. Bao, H. Gies, D. E. De Vos, ACS Catal. 2015, 5, 3392–3397.
[32] A. Klein, K. Keisers, R. Palkovits, Appl. Catal. A 2016, 514, 192–202.
[33] H.-J. Chae, T.-W. Kim, Y.-K. Moon, H.-K. Kim, K.-E. Jeong, C.-U. Kim, S.-Y.
Jeong, Appl. Catal. A 2014, 150–151, 596–604.
[34] D. R. M. Brew, F. P. Glasser, Cem. Concr. Res. 2005, 35, 85–98.
[35] Z. Li, T. Zhang, J. Hu, Y. Tang, Y. Niu, J. Wei, Q. Yu, Constr. Build. Mater.
2014, 61, 252–259.
[36] J. S. Hartman, R. L. Millard, Phys. Chem. Miner. 1990, 17, 1–8.
[37] C. Perego, S. Peratello, Catal. Today 1999, 52, 133–145.
[38] H. Knçzinger, R. Kçhne, J. Catal. 1966, 5, 264–270.
[39] J. R. Jain, C. N. Pillai, J. Catal. 1967, 9, 322–330.
[40] B. Shi, B. H. Davis, J. Catal. 1995, 157, 359–367.
[41] H. Arai, J.-I. Take, Y. Saito, Y. Yoneda, J. Catal. 1967, 9, 146–153.
[42] H. Knçzinger, H. Bꢆhl, K. Kochloefl, J. Catal. 1972, 24, 57–68.
[43] V. Gruver, A. Sun, J. J. Fripiat, Catal. Lett. 1995, 34, 359–364.
[44] S. Da Ros, E. Barbosa-Coutinho, M. Schwaab, V. Calsavara, N. R. C. Fer-
nandes-Machado, Mater. Charact. 2013, 80, 50–61.
[13] E. V. Makshina, W. Janssens, B. F. Sels, P. A. Jacobs, Catal. Today 2012,
198, 338–344.
[14] V. L. Sushkevich, I. I. Ivanova, V. V. Ordomsky, E. Taarning, ChemSusChem
2014, 7, 2527–2536.
[15] M. Gao, Z. Liu, M. Zhang, L. Tong, Catal. Lett. 2014, 144, 2071–2079.
[16] V. L. Sushkevich, I. I. Ivanova, E. Taarning, Green Chem. 2015, 17, 2552–
2559.
[45] M. Tamura, K.-I. Shimizu, A. Satsuma, Appl. Catal. A 2012, 433–434,
135–145.
[46] S. Huber, H. Knçzinger, J. Mol. Catal. A 1999, 141, 117–127.
[17] W. M. Quattlebaum, W. J. Toussaint, J. T. Dunn, J. Am. Chem. Soc. 1947,
69, 593–599.
Received: March 23, 2016
Revised: May 9, 2016
[18] J. V. Ochoa, C. Bandinelli, O. Vozniuk, Green Chem. 2016, 18, 1653–1663.
Published online on && &&, 0000
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
11
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
S. Da Ros, M. D. Jones,* D. Mattia,
J. C. Pinto,* M. Schwaab, F. B. Noronha,
S. A. Kondrat, T. C. Clarke, S. H. Taylor
Ethanol to 1,3-butadiene: The role of
catalyst acidity of ZrZn-containing silica
magnesia oxides prepared by a co-pre-
cipitation method has been investigat-
ed. Catalyst acidity control is found to
be the key step to increase butadiene
selectivity.
&& – &&
Ethanol to 1,3-Butadiene Conversion
by using ZrZn-Containing MgO/SiO2
Systems Prepared by Co-precipitation
and Effect of Catalyst Acidity
Modification
&
ChemCatChem 2016, 8, 1 – 12
www.chemcatchem.org
12
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!