Effect of Preparation Method and CuO Promotion in the Conversion of Ethanol into 1,3-Butadiene over SiO2–MgO Catalysts

Article abstract of DOI:10.1002/cssc.201402361

Silica–magnesia (Si/Mg=1:1) catalysts were studied in the one-pot conversion of ethanol to butadiene. The catalyst synthesis method was found to greatly influence morphology and performance, with materials prepared through wet-kneading performing best both in terms of ethanol conversion and butadiene yield. Detailed characterization of the catalysts synthesized through co-precipitation or wet-kneading allowed correlation of activity and selectivity with morphology, textural properties, crystallinity, and acidity/basicity. The higher yields achieved with the wet-kneaded catalysts were attributed to a morphology consisting of SiO₂spheres embedded in a thin layer of MgO. The particle size of the SiO₂catalysts also influenced performance, with catalysts with smaller SiO₂spheres showing higher activity. Temperature-programmed desorption (TPD) measurements showed that best butadiene yields were obtained with SiO₂–MgO catalysts characterized by an intermediate amount of acidic and basic sites. A Hammett indicator study showed the catalysts’ pK_avalue to be inversely correlated with the amount of dehydration by-products formed. Butadiene yields could be further improved by the addition of 1 wt % of CuO as promoter to give butadiene yields and selectivities as high as 40 % and 53 %, respectively. The copper promoter boosts the production of the acetaldehyde intermediate changing the rate-determining step of the process. TEM-energy-dispersive X-ray (EDX) analyses showed CuO to be present on both the SiO₂and MgO components. UV/Vis spectra of promoted catalysts in turn pointed at the presence of cluster-like CuO species, which are proposed to be responsible for the increased butadiene production.

Full text of DOI:10.1002/cssc.201402361

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201402361
Effect of Preparation Method and CuO Promotion in the
Conversion of Ethanol into 1,3-Butadiene over SiO2–MgO
Catalysts
Carlo Angelici, Marjolein E. Z. Velthoen, Bert M. Weckhuysen,* and Pieter C. A. Bruijnincx*[a]
Silica–magnesia (Si/Mg=1:1) catalysts were studied in the one-
pot conversion of ethanol to butadiene. The catalyst synthesis
method was found to greatly influence morphology and per-
formance, with materials prepared through wet-kneading per-
forming best both in terms of ethanol conversion and buta-
diene yield. Detailed characterization of the catalysts synthe-
sized through co-precipitation or wet-kneading allowed corre-
lation of activity and selectivity with morphology, textural
properties, crystallinity, and acidity/basicity. The higher yields
achieved with the wet-kneaded catalysts were attributed to
a morphology consisting of SiO2 spheres embedded in a thin
layer of MgO. The particle size of the SiO2 catalysts also influ-
enced performance, with catalysts with smaller SiO2 spheres
showing higher activity. Temperature-programmed desorption
(TPD) measurements showed that best butadiene yields were
obtained with SiO2–MgO catalysts characterized by an inter-
mediate amount of acidic and basic sites. A Hammett indicator
study showed the catalysts’ pKa value to be inversely correlat-
ed with the amount of dehydration by-products formed. Buta-
diene yields could be further improved by the addition of
1 wt% of CuO as promoter to give butadiene yields and selec-
tivities as high as 40% and 53%, respectively. The copper pro-
moter boosts the production of the acetaldehyde intermediate
changing the rate-determining step of the process. TEM-
energy-dispersive X-ray (EDX) analyses showed CuO to be pres-
ent on both the SiO2 and MgO components. UV/Vis spectra of
promoted catalysts in turn pointed at the presence of cluster-
like CuO species, which are proposed to be responsible for the
increased butadiene production.
Introduction
It has been advocated that future biorefineries need to pro-
duce high-value bio-based chemicals in addition to the lower-
value biofuels to be economically competitive.[1] The produc-
tion of ‘drop-in’ renewable chemicals, that is, direct, molecular-
ly identical yet sustainably produced replacements for current
petrochemical chemicals, is particularly attractive as efficient
production routes can serve existing markets and make use of
existing infrastructure.
(hereon simply called butadiene), which finds major applica-
tion in the polymer industry, can for instance be produced
from ethanol.
Butadiene is currently obtained as by-product of ethylene
production in hydrocarbon steam cracking, with its isolation
requiring a number of expensive extractive distillation steps.[5]
The predicted shift towards lighter feeds for steam cracking, as
a result of the recent shale gas boom, threatens butadiene
production and is expected to significantly affect its price; effi-
cient on-purpose production technologies for butadiene are,
therefore, highly desirable.[6]
Due to its (mandated) use as fuel and fuel additive, bioetha-
nol is currently produced in increasingly large volumes, leading
to decrease in ethanol market prices. Importantly, this future
growth in bioethanol production is expected to mainly result
from the utilization of 2nd or 3rd generation biomass.[2,3] The ex-
pected increased availability and shift to inedible feedstocks
will ensure that low-cost ethanol will not only be available as
fuel but also for the production of renewable bulk chemicals.
Indeed, ethanol can serve as an excellent platform chemical, as
many bulk chemicals currently are or can in principle be pro-
duced from ethanol.[1,4] The bulk chemical 1,3-butadiene
During the first decades of the twentieth century, two close-
ly related processes were developed for the synthesis of buta-
diene from ethanol, that is, the Ostromisslenski[7] and Lebedev
processes.[8,9] The first consists of two consecutive steps, that
is, dehydrogenation of ethanol to acetaldehyde followed by
condensation of acetaldehyde and ethanol to give butadiene;
the Lebedev process, on the other hand, is a one-pot, one-step
process used for direct butadiene synthesis from ethanol. The
mechanism of this gas-phase conversion is complex and has
not yet been unambiguously demonstrated. Most commonly
a mechanistic pathway is proposed that involves consecutive
dehydrogenation, aldol condensation, reduction and dehydra-
tion steps (see Scheme 1).[10] The ideal catalyst for this process
thus requires a subtle balance of different active sites (acidic,
basic, and redox) to be struck, both in terms of strength as
well as number.
[a] C. Angelici, M. E. Z. Velthoen, Prof. B. M. Weckhuysen, Dr. P. C. A. Bruijnincx
Inorganic Chemistry and Catalysis
Debye Institute for Nanomaterials Science
Utrecht University
Universiteitsweg 99, 3584 CG Utrecht (The Netherlands)
E-mail: b.m.weckhuysen@uu.nl
p.c.a.bruijnincx@uu.nl
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2505

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Herein, we report on the activity of various SiO2–MgO mate-
rials as catalysts for the Lebedev process. Different preparation
methods were employed for the SiO2–MgO catalysts with the
aim of establishing a structure–activity relationship. The signifi-
cant differences in terms of butadiene yield observed can
indeed be correlated to structural differences, for example, in
morphology, dispersion, and acid/base properties. It is also
shown that CuO promotion on the SiO2–MgO systems im-
proves butadiene yield substantially in all cases. The nature of
the supported CuO is also shown to have an important effect
on catalysis.
Scheme 1. Commonly proposed mechanism for the Lebedev conversion of
ethanol into 1,3-butadiene.[10]
Different classes of catalysts (metal nanoparticles supported
on clay materials, on oxides, or on mixed oxides) have been
employed in the Lebedev process, and the catalytic aspects Results and Discussion
have been recently reviewed by us.[4] (Mixed) metal oxides
Catalytic activity
have been studied most, with SiO2–MgO systems, of different
composition and obtained through different preparation pro-
cedures, often performing best in terms of butadiene
yield.[11,12] It was previously shown that redox-active promoters
can improve performance, considering that butadiene yields
are typically lower than 20% with unpromoted SiO2–MgO sys-
tems,[12,13] the only exception being the results reported by Oh-
nishi et al. (butadiene yield 42% at 623 K).[14] The same study
showed a remarkable sensitivity of catalyst performance to im-
pregnation of the SiO2–MgO materials with 0.1 wt% of Na2O
and K2O (leading to yields of 87 and 70%, respectively), thus
implying that the balance of acidic and basic active sites is
subtle yet critical for high activity and selectivity.[14] A recent
publication by Jones et al. investigated for the first time the
use of bi- and trimetallic systems supported on various metal
oxides again pointing at the importance of using SiO2.[15] Fur-
thermore, a number of transition metals and metal oxides,
such as CuO,[13,16] ZnO,[17] Cr2O3,[17] and NiO[18] were employed
as promoters to improve the dehydrogenation activity of the
catalyst and thus the final butadiene yield. In most cases,
scarce details on catalyst structure, product distribution, and
long-term activity were provided for the various catalysts
tested; as a result the salient features of these (promoted)
SiO2–MgO catalysts are not yet fully understood.
Combinations of SiO2 and MgO of different molar ratios have
often been reported as promising catalysts for the Lebedev
process.[11,12] Here, we study a series of six differently prepared
SiO2–MgO catalysts, having a nominal 1:1 molar ratio (see the
Experimental Section).
The influence of preparation method on catalytic per-
formance was studied using a physical mixture [PM, SiO2–MgO
(I)], three wet-kneaded catalysts [WK, SiO2–MgO (II)–(IV)], and
two co-precipitated ones [CP, SiO2–MgO (V) and (VI)] (see
Table 2 for details on the preparation methods). These samples
show distinct differences in morphology (vide infra) and, as
a result, high variability in catalytic activity. For these catalyst
materials, activity and selectivity with 24 h on stream are re-
ported (Figure 1), providing the first data on long-term catalyst
performance in the Lebedev process. The physical mixture
SiO2–MgO (I) showed low ethanol conversion and butadiene
yield. Catalytic testing of the two single oxides under identical
conditions (data not shown) showed that SiO2 was quite inert
and gave negligible ethanol conversion and no butadiene,
whereas the use of MgO resulted in low ethanol conversion
and butadiene yield (1.7% after 4 h time on stream). Notably,
the physical mixture SiO2–MgO (I) showed a butadiene yield
similar to pure MgO, again illustrating the inertness of isolated
Figure 1. Ethanol conversion (left) and butadiene yield (right) as a function of time-on-stream over the different SiO2–MgO catalysts tested. Conditions: 0.2 g
catalyst, reaction temperature 698 K, ethanol (gas phase) and nitrogen flow 2 and 98 mLminꢀ1, respectively.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2506

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
SiO2. As expected, all chemically mixed SiO2–MgO catalysts
showed higher ethanol conversions and butadiene yields. The
time on stream data furthermore shows that conversions and
butadiene yields are stable in time (Figure 1) and no apprecia-
ble deactivation can be observed.
varied and, taken as a measure of acidity of the catalyst, could
be correlated with SiO2 particle size (see below).
Acetaldehyde, a key intermediate that can be formed by
ethanol dehydrogenation on materials containing redox
sites[20] but also on pure MgO,[21] is formed in too low amounts
to allow any correlations to be made with catalyst structure.
Although the mechanism depicted in Scheme 1 is generally ac-
cepted, different rate-determining steps have been proposed
for butadiene production over SiO2–MgO catalysts.[11,12,22] The
fact that quite a low amount of acetaldehyde is observed for
the SiO2–MgO catalysts employed in this study might suggest
that acetaldehyde formation is rate determining, as previously
proposed by Niiyama et al.[11] Finally, various amounts of other
components, identified to be mainly butenes, 1-butanol, and
acetone, are formed over the different catalysts. SiO2–MgO (II),
for instance, produces a relatively large amount of others at
the expense of ethylene and butadiene for which the yields
are the lowest among the WK catalysts. Notably, SiO2–MgO (IV)
shows a remarkably low amount of other components, with
~90% of the products due to unconverted ethanol, ethylene,
and butadiene. Finally, it should be noted that the combined
yield of the different butenes (i.e., 1-butene, cis- and trans-2-
butene), already discussed to be critical for purification of bu-
tadiene, is well below 5% in all cases.
At similar conversion levels (50–65%), the catalysts showed
significant differences in butadiene yield with performance in-
creasing in the order: PM<CP<WK (Figure 1). The catalysts
prepared by co-precipitation performed poorly, with almost
negligible amounts of butadiene produced with SiO2–MgO (VI)
and a low, yet stable butadiene yield of approximately 6% for
SiO2–MgO (V). The catalysts prepared through wet-kneading
[SiO2–MgO (II)–(IV)] gave significantly higher butadiene yields.
While SiO2–MgO (II) showed the highest ethanol conversion, it
gave the lowest butadiene yield of the wet-kneaded catalysts.
SiO2–MgO (III) performed best, yielding ~17% butadiene with
33% selectivity, even after 24 h time on stream. These differen-
ces in catalytic activity can be clearly related to differences in
morphology and acidity/basicity of the materials, as shown
below.
Not only the butadiene yield, but also the identity and
number of by-products formed have a large influence on the
ability of the Lebedev process to compete with traditional bu-
tadiene routes.[19] Indeed, as product separation is often costly,
an ethanol conversion process that yields a limited number of
products, which are in turn easily separable and individually
can find a value-added application, is highly desired. Any ethyl-
ene produced can, for instance, be easily separated by distilla-
tion, whereas co-produced butenes would make separation of
butadiene from the reaction mixture much more difficult.[5]
Other compounds that can be present in the product stream
are diethyl ether, higher alcohols and, of course, process inter-
mediates such as acetaldehyde.
CuO was added as promoter to the catalysts prepared
through wet-kneading and co-precipitation. Different loadings
(0.5, 1 and 2 wt%) as well as various thermal treatments (calci-
nation in stagnant air, calcination in N2 flow, and calcination in
N2 flow followed by reduction) were employed for CuO/SiO2–
MgO (II). The CuO loading did not have a strong effect on cata-
lytic activity and selectivity; thermal treatment did matter and
calcination in stagnant air provided the best catalyst (data not
To judge the efficiency of a pro-
cess, it is, therefore, important
that the whole product composi-
tion is reported, rather than just
the butadiene yield. The product
distributions for the six SiO2–
MgO catalysts at t=0.5 and 4 h
show considerable differences
(Figure 2). Diethyl ether produc-
tion, which is undesired as two
ethanol molecules are consumed
upon formation, is ꢁ5% for all
catalysts. On the other hand,
substantial amounts of ethylene
were observed, in particular over
the two CP catalysts.
All WK catalysts, but particu-
larly SiO2–MgO (II), showed
a much lower selectivity to eth-
ylene and diethyl ether, suggest-
ing that the CP catalysts are
much more acidic than the WK
ones. The amount of ethylene
Figure 2. Product distribution for the different SiO2–MgO catalysts after 0.5 and 4 h on stream. Conditions: 0.2 g
formed over the WK catalysts catalyst, reaction temperature 698 K, ethanol (gas phase) and nitrogen flow 2 and 98 mLminꢀ1, respectively.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2507

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 3. Ethanol conversion (left) and butadiene yield (right) as a function of time over the different CuO (1 wt%)/SiO2–MgO catalysts tested. Conditions:
0.2 g catalyst, reaction temperature 698 K, ethanol (gas phase) and nitrogen flow 2 and 98 mLminꢀ1, respectively.
shown). Promotion of the WK SiO2–MgO catalysts with CuO re-
sulted in a strong increase in ethanol conversion, whereas the
CuO-containing CP catalysts either showed a similar [CuO/
SiO2–MgO (V)] or reduced [CuO/SiO2–MgO (VI)] conversion
level (cf. Figure 1 and Figure 3). Most importantly, butadiene
yield was strongly boosted upon addition of CuO: SiO2–MgO
catalysts show butadiene yields in the range 0.3–17% (selectiv-
ities 2–36%), whereas CuO-containing catalysts show 6–36%
yields (selectivities 19–50%) after 4 h. Again, WK catalysts
showed much higher yields than CP ones.
by considering that, for these catalysts, the rate-determining
step is not the formation but rather the transformation of acet-
aldehyde, as proposed by Kvisle et al. for SiO2–MgO systems.[12]
Furthermore, the selectivity towards ethylene and diethyl ether
is highly reduced in all cases. This suggests that CuO not only
introduces redox-active sites, but also poisons the most acidic
sites of the SiO2–MgO samples. Alternatively, ethanol is prefer-
entially dehydrogenated to acetaldehyde in the presence of
CuO and, as a consequence, does not get a chance to undergo
dehydration; given the incomplete ethanol conversion, the
second hypothesis seems less likely. Finally, a substantial in-
crease in the amount of other compounds formed was ob-
served for the CuO-containing catalysts, especially with the WK
samples. In addition to the previously observed butenes, 1-bu-
Wet-kneaded catalyst CuO/SiO2–MgO (II) gave moderate but
fairly stable butadiene yields of 23–28%. As with the non-CuO-
containing catalysts, CuO/SiO2–MgO (III) performed best with
an initial yield of approximately 40%, corresponding to a buta-
diene
selectivity
of
53%
(Table 1). In contrast to the un-
promoted catalysts, the CuO-
containing ones did show some
deactivation, as evidenced by
the gradual drop in ethanol con-
version and butadiene yield with
time on stream; the gradual de-
activation can at least partially
be ascribed to the increased
amount of carbonaceous depos-
its that was detected for the pro-
moted catalysts using UV/Vis
analysis of the spent catalysts
(data not shown).
Not only butadiene yield but
also
product
distribution
changed dramatically upon in-
troduction of CuO (Figure 4). No-
tably, an increase in the amount
of acetaldehyde produced was
observed for all catalysts. The
fact that butadiene and acetal-
dehyde yields increased simulta-
neously for the copper-promot-
Figure 4. Product distribution for the CuO/SiO2–MgO catalysts after 0.5 and 4 h on stream. Conditions: 0.2 g cata-
ed catalysts might be explained lyst, reaction temperature 698 K, ethanol (gas phase) and nitrogen flow 2 and 98 mLminꢀ1, respectively.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2508

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
The basic sites on the different materials were ini-
tially probed by using CO2-TPD, which showed the
amount of basic sites to vary in the following order:
SiO2–MgO (II)>SiO2–MgO (III)>SiO2–MgO (IV)>SiO2–
MgO (V)>SiO2–MgO (VI). This order follows inversely
the amount of acidic sites as determined by NH3-
TPD, again emphasizing that different preparation
methods lead to the presence of different amounts
of acidic/basic sites on the catalysts. However, the
calculated number of basic sites from the TPD data
Table 1. Butadiene yield, ethanol conversion, and butadiene selectivity observed for
the best SiO2–MgO and CuO (1 wt%)/SiO2–MgO catalysts at 698 K tested at the start
of the reaction (0.5 h) and after 4 h on stream.
Entry Catalyst
BD yield [%]
EtOH conversion [%]
0.5 h 4 h
BD selectivity [%]
0.5 h 4 h
0.5 h
4 h
1
2
SiO2–MgO(III)
CuO/SiO2–MgO(III)
17
38
16
37
51
80
52
74
34
32
48 49
tanol and acetone, C6-oxygenated compounds were now ob-
served among the main byproducts.
was unexpectedly low in all cases, that is, about tenfold lower
than the number of acidic sites calculated using NH3-TPD.
Given the overall basicity of the materials (see below), this sug-
gests that not all (catalytically) relevant basic sites are properly
probed by CO2 under the applied TPD conditions. Such a dis-
crepancy has been observed before, for instance by Liu and
coworkers for MgO–ZrO2 materials,[23] and requires further
study.
Catalyst characterization
The large variation in catalytic activity of the (CuO-impregnat-
ed) SiO2–MgO catalysts could be correlated with differences in
catalyst structure. Considering the commonly proposed mech-
anisms, and, in particular, the different elementary steps re-
ported, it seems evident that a specific balance between acidic
and basic sites is required for selective butadiene production;
quite surprisingly, not much is reported in the open literature
with respect to the type, strength, and abundance of acidic/
basic sites required for a good catalyst material for the Lebe-
dev process. The strength and number of acidic and basic sites
on the different catalysts were measured by temperature-pro-
grammed desorption (TPD) of NH3 and CO2, respectively,
whereas overall acidity/basicity was determined using Ham-
mett indicators. Broad desorption peaks are observed by NH3-
TPD for all SiO2–MgO catalysts, especially for the catalysts pre-
pared through co-precipitation. The number of acid sites varies
in the order: SiO2–MgO (VI)>SiO2–MgO (V)ꢂSiO2–MgO (IV)>
SiO2–MgO (III)>SiO2–MgO (II). This order is in good agreement
with the combined amounts of ethylene and diethyl ether by-
products, which are produced through ethanol dehydration on
the acidic sites of the catalysts. The significantly increased di-
ethyl ether formation over the two co-precipitated catalysts
[and partially over SiO2–MgO (IV) prepared through wet-knead-
ing] also points at a larger fraction of strongly acidic sites, as
also evidenced by a broadening of the NH3 desorption signal
at higher temperatures (not shown).
The overall acidity/basicity of the different catalysts was as-
sessed using a set of Hammett indicators of increasing pKa
value. If these indicators are protonated or deprotonated by
a solid acid/base a color change is observed, which allows for
a pKa range to be determined for the solid materials.[24,25] More
specifically, when testing indicators in order of increasing pKa
value, the lower limit of the pKa value of the catalyst is given
by the pKa value of the last indicator that showed color
change (i.e., the indicator for which the initially neutral form
was still mainly converted into the deprotonated one) and the
upper limit of the catalyst’s pKa value is defined by the first in-
dicator that does not show color change. As expected, the dif-
ferent preparation methods used led to a different overall ba-
sicity/acidity for the various SiO2–MgO catalysts (Table 2). The
materials are found to be predominantly basic, with some ma-
terials capable of inducing color change in indicators having
pKa values as high as 15.0 (2,4-dinitroaniline). This shows that
the total amount of basic sites is actually underestimated by
the CO2-TPD measurements. The order of overall basicity of
the catalysts was found to be: SiO2–MgO (II)ꢂSiO2–MgO (III)>
SiO2–MgO (IV)>SiO2–MgO (VI), with the co-precipitated SiO2–
MgO (VI) thus having less acidity than all of the wet-kneaded
catalysts. SiO2–MgO (IV) is the least basic of the wet-kneaded
Table 2. Overview of the different preparations for the SiO2–MgO catalyst materials, their corresponding BET surface areas, total acidity, and overall basici-
ty.
Sample
Preparation
method[a]
Preformed
(hydr)oxides
SiO2 particle
size [nm]
Sg
[m2 gꢀ1
Total acidity[b]
[mmolgꢀ1
]
pKa range[c]
Comments
]
SiO2–MgO (I)
SiO2–MgO (II)
SiO2–MgO (III)
SiO2–MgO (IV)
SiO2–MgO (V)
SiO2–MgO (VI)
PM
WK
WK
WK
CP
Y
Y
Y
Y
N
N
~425
~425
30–100
20–40
~425
124
139
358
156
225
0.145
0.219
0.234
0.242
0.268
15.0–17.2
15.0–17.2
~15.0[d]
SiO2: adding EtOH/NH3 to TEOS
SiO2: adding TEOS to EtOH/NH3
SiO2: commercial Aerosil 300
Mg(NO3)2 added to TEOS after 20 min
Mg(NO3)2 and TEOS added simultaneously
n.d.[e]
[f]
CP
–
9.3–15.0
[a] PM=physical mixture; WK=wet-kneading; CP=co-precipitation. [b] Determined by means of NH3-TPD. [c] Determined using Hammett indicators.
[e] The color observed is between that of the neutral (protonated) and deprotonated forms of 2,4-dinitroaniline (pKa =15.0). [f] pKa could not be deter-
mined due to intrinsic coloration of the material. [d] No SiO2 phase was observed for this catalyst material.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2509

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
catalysts, showing the influence of the SiO2 particle size on the
acid–base properties.
ly 425 nm could be clearly distinguished. The lack of intimate
contact between the two components is believed to be the
main reason of its poor activity. Indeed, both acidic and basic
sites are required for a series of elementary steps in the Lebe-
dev process (Scheme 1) and lack of proximity between the dif-
ferent active sites results in the very limited selectivity to buta-
diene.
The acidity/basicity studies suggest that the co-precipitated
catalysts [SiO2–MgO (V) and (VI)] combine an excessive amount
of acidic sites with an insufficient amount of basic ones, which
results in the high amounts of by-products produced through
ethanol dehydration (i.e., ethylene and diethyl ether). This ulti-
mately leads to a poor butadiene yield. Conversely, the large
number of basic sites in SiO2–MgO (II) also leads to a butadiene
yield lower than in the case of the other two wet-kneaded cat-
alysts, SiO2–MgO (III) and SiO2–MgO (IV). The latter catalysts
contain an intermediate amount of both acidic and basic sites,
which results in a different product distribution and, eventual-
ly, a higher butadiene yield.
The TEM results for samples obtained by WK or CP were
quite different. The WK catalysts (II)–(IV) show spherical SiO2
particles (as expected for the Stçber-like preparation) of varia-
ble size that are to different extents covered by or embedded
in MgO. SiO2–MgO (II), for instance, was found to consist of
425 nm SiO2 spheres and islands of thick platelets of MgO (Fig-
ure 6b). However, closer inspection revealed the intimate con-
tact between the phases as a roughening of the surface of the
SiO2 spheres, which are decorated with a thin layer of MgO
(Figure 6b1 and b2). The morphologies of SiO2–MgO (III) and
SiO2–MgO (IV) again showed thick platelets of MgO and MgO-
embedded SiO2 spheres, but the latter are now much smaller
(30–100 nm for III and 20–40 nm for IV). As a result of the
smaller SiO2 particles in these two catalysts, the contrast with
the MgO component is insufficient to clearly observe MgO
layers embedding the SiO2 spheres. Nevertheless, given the
similar preparation method, such layers are expected to play
an important role in these two samples as well. The TEM
image of CP SiO2–MgO (V) shows SiO2 spheres completely cov-
ered in a thick MgO layer. The other CP sample SiO2–MgO (VI)
displayed an ill-defined morphology in which neither SiO2
spheres nor MgO platelets could be identified.
Powder X-ray diffraction (XRD) patterns also show clear
structural differences as a result of preparation method. As
a result of the low copper loading, the XRD patterns for the
CuO-containing catalysts (data not shown) are identical to the
unpromoted ones (Figure 5). For all catalysts, a halo in the
region 2q=208–408 is observed due to the presence of amor-
phous SiO2.[13,26] Only the WK samples show peaks that could
be attributed to the periclase phase of MgO, at 2q=42.98,
50.28, 73.88, 89.48, and 94.68,[27] with the intensity of the peri-
clase peaks varying for the different samples.
Good butadiene selectivity can thus be correlated to proper
intimate contact between the two components; too intimate
a contact, as in the CP material SiO2–MgO (VI), is detrimental
for butadiene production, but a (crystalline) MgO layer that
embeds the SiO2 particles, as observed for the WK samples, on
the other hand, greatly improves activity and selectivity to bu-
tadiene. A thin layer appears to be more advantageous, imply-
ing that the best performance is achieved when SiO2 can still
interact with the substrate, but in a way that is mediated by
MgO. Large MgO-only areas, such as those observed in SiO2–
MgO (II), probably contribute to ethanol conversion but with
lower selectivity to butadiene.
Figure 5. XRD diffractograms of the various SiO2–MgO catalysts.
TEM analysis of the CuO-containing catalysts showed that
the typical morphologies shown in Figure 6 for the non-im-
pregnated samples are retained upon addition of copper-con-
taining (nano)particles could be seen on any of the samples
through bright field or dark field TEM nor by high angle annu-
lar dark field scanning transmission electron microscopy
(HAADF-STEM). Energy dispersive X-ray (EDX) analysis did show
that Cu-containing species can be found at both SiO2-rich and
MgO-rich areas (Figure 6C’1 and C’2). This lack of preference is
repeatedly seen on different catalysts and at different areas of
the same catalyst.
The morphology of the various SiO2–MgO materials that
have been reported as active catalysts for the Lebedev process
has been little studied. Only Kvisle et al. used TEM to study the
morphology of SiO2–MgO catalysts prepared through wet-
kneading.[12] In their case, the single oxides were observed to
consist of leaflets (5–10 nm) of SiO2 and platelets (30–40 nm)
of MgO. The wet-kneaded catalysts were shown to contain
substantially bigger SiO2 particles, with the overall morphology
consisting of inhomogeneous leaflets and platelets containing
both components.
The fact that no CuO nanoparticles are observed by TEM
can either point at an intrinsic lack of contrast with the MgO
phase, depending on crystallinity and specific morphology of
stacked MgO platelets, at the absence of nanometer-sized CuO
particles or at particles or clusters too small to be resolved.
In our case, TEM analysis proved extremely useful as the ob-
served morphological differences could be correlated to cata-
lytic performance. In the PM catalyst (Figure 6a), the MgO pla-
telets and monodisperse, smooth SiO2 spheres of approximate-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2510

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 6. TEM images of A) SiO2–MgO (I), B) SiO2–MgO (II), C) SiO2–MgO (III), D) SiO2–MgO (IV), E) SiO2–MgO (V), F) SiO2–MgO (VI), and C’) CuO (1 wt%)/SiO2–
MgO (III). EDX scan of a C’1) SiO2-rich region of CuO (1 wt%)/SiO2–MgO (III) and C’2) of a MgO-rich region of CuO (1 wt%)/SiO2–MgO (III).
The lack of crystallinity (Figure 5) in the CP samples V and VI
seem to discard the former explanation, but an intrinsic impos-
sibility of observing CuO on MgO cannot be completely ex-
cluded. Although metallic Cu particles on MgO have been
often studied using TEM, we are not aware of any other stud-
ies in which the particle size of CuO on MgO was studied
using TEM.
by UV/Vis spectroscopy before and after addition of CuO. The
unpromoted SiO2–MgO catalysts show three characteristic
bands at 210, 250, and 290 nm (Figure 7). Coluccia et al.[29,30]
previously assigned two similar bands at 230 and 274 nm of
MgO to charge transfer between Mg2+ and four- and three-co-
ordinated O2ꢀ ions, respectively. The latter absorption band at
274 nm was thus thought to correspond to the more unsatu-
rated and more reactive sites on MgO. In our case, the two
bands are slightly shifted, yet match the position of the bands
measured for pure MgO (data not shown). The 290 nm band is
similar for all catalysts, which, therefore, all seem to contain
only a small amount of these unsaturated sites; on the other
hand, the four-coordinated oxygen CT bands vary both in in-
tensity and position. It is particularly intense in SiO2–MgO (V)
and SiO2–MgO(III), whereas the intensity is much lower for
SiO2–MgO (VI); this sample is prepared through a pure co-pre-
cipitation route and as a result has a lower amount of oxygen
atoms coordinated with Mg2+ only; this again points at the ab-
Identification of the nature and size of the CuO species is
further complicated by the possibility that chemical structures
other than CuO nanoparticles or clusters can be formed. El-
Shobaky et al.[28] suggested that CuO/MgO solid solutions are
formed in CuO-doped MgO catalysts, based on the absence of
CuO diffraction peaks in samples containing up to 23 wt%
CuO. Solid solution formation was attributed to a thermal
treatment performed at 673 K, a temperature that is lower
than the one used for calcination of our samples.
To gain further insight into the nature of the CuO species
supported on SiO2–MgO, the samples were also characterized
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2511

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
hedral coper species and as a weak band at ~1380 nm (not
shown) for d–d transitions in tetrahedral copper species.
Neither of these d–d transitions, however, provided further
insight into the differences in catalytic performance and nature
of CuO. On the contrary, the LMCT bands associated with the
CuO species proved insightful for correlating structure to activ-
ity. Multiple bands are observed in the 200–300 nm region,
corresponding to O2ꢀ to Cu2+ CT of isolated copper oxide spe-
cies.[31–33] The presence of such CuO species on both compo-
nents (as shown by EDX) of the catalyst accounts for the multi-
ple observed features. The position of these CT bands is
known to depend on the type and morphology of the sup-
ports and the procedure employed to support CuO.[33–35] Kong
et al., for instance, observed variations in the position and in-
tensity of CT bands of isolated CuO species for CuO/SBA-15
and CuO/KIT-6.[36] Bravo-Suꢁrez et al. recently reported two dif-
ferent bands centered at 260 and 360 nm for CuO/MgAlOx
mixed oxides.[37] The 260 nm band was assigned to an LMCT of
isolated octahedral Cu2+ꢀO2ꢀ species, whereas the latter band
was attributed to the presence of oligomers of the type Cuꢀ
OꢀCu. Derrien and co-workers[38] reported a band at 235 nm
for CuO supported on mesoporous SiO2 spheres, whereas Shi-
mokawabe et al.[32] reported the same band to be at 270 nm
for CuO supported on SiO2, which illustrates the additional de-
pendence on morphology.
Figure 7. UV/Vis spectra of the SiO2–MgO samples: i) MgꢀO-based charge
transfer involving tetra-coordinated oxygen, ii) MgꢀOꢀSi-based charge trans-
fer links, and iii) MgO-based charge transfer involving tris-coordinated
oxygen.
sence of an appropriate MgO-like structure in this catalyst, as
also observed using XRD and TEM. For SiO2–MgO (II) and SiO2–
MgO (IV) a shoulder can be observed at approximately
210 nm, whereas an additional band seems to be centered at
a wavelength lower than 200 nm.
The band located at around 260 nm was also seen by Kvisle
et al.[12] for a wet-kneaded SiO2–MgO catalyst. It was assigned
to MgꢀOꢀSi links based on the observation that the mineral
antigorite (Mg3(OH)3Si2O5) has a band at the same position and
is known to be active in the Lebedev process. In our case, the
intensity of the 250 nm band varies consistent with this assign-
ment, that is, the wet-kneaded samples show only a weak
shoulder at ~250 nm, whereas co-precipitated materials, ex-
pected to contain a much larger amount of MgꢀOꢀSi bonds,
show a much higher absorption at this wavelength. These re-
sults again indicate that an optimal catalyst for the Lebedev
process requires an intimate contact between the two compo-
nents but not full mixing of the two (see Figure 3).
Moreover, the weak shoulder in the region 300–550 nm
(Figure 8), assigned to LMCT in cluster-like species (CuꢀOꢀCu),
is found to vary strongly for our CuO/SiO2–MgO materials,
being highest in intensity for the WK catalysts, SiO2–MgO(III)
and SiO2–MgO(IV) in particular. Only in SiO2–MgO(VI), which
gives the lowest butadiene yield, it is completely absent. The
fact that the two WK catalysts SiO2–MgO(III) and SiO2–MgO(IV)
are more active and more selective towards butadiene forma-
tion suggests that these cluster-like CuO species promote bu-
tadiene formation. This is tentatively explained considering
that the formation of acetaldehyde molecules on different sites
in close proximity with each other could make their condensa-
tion (proposed by some authors to be the rate-limiting step of
the complete process)[12,22] easier than on isolated CuO species,
thus leading to cascade reactions for the formation of 3-hy-
droxybutanal, crotonaldehyde, and crotyl alcohol, all known to
be key intermediates in the Lebedev process.
The UV/Vis spectra of the CuO-containing samples show two
additional transitions (Figure 8): i) a ligand to metal charge
transfer band (LMCT) in the region 200–550 nm and ii) d–d
transitions seen as a broad band around 700–750 nm for octa-
Conclusions
SiO2–MgO mixed oxides have often been reported as most effi-
cient catalyst materials for the one-step conversion of ethanol
to butadiene. However, the reason as to why this combination
of oxides is so selective towards the production of butadiene
remains elusive. A better understanding of the influence of
preparation method and promotion on activity and selectivity
is needed for the design of better catalysts for the Lebedev
process.
We observed a clear effect of the synthesis method of SiO2–
MgO on morphology, acidity, and basicity. More specifically, all
catalyst materials prepared by wet-kneading proved to be
more active and selective towards butadiene, which points at
Figure 8. UV/Vis spectra of the CuO/SiO2–MgO samples before reaction:
i) bands assigned to LMCT transitions and ii) bands assigned to d–d transi-
tions.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2512

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
ized water and recovered through centrifugation, dried at 393 K,
and finally calcined in stagnant air at 773 K for 5 h. A modified
Stçber route was used for the preparation of SiO2. TEOS was hy-
drolyzed using an ethanol–NH3 solution (5:1 v/v) at room tempera-
ture for 24 h. After washing with ethanol, SiO2 was dried and cal-
cined as described for the other oxides.
the importance of the thin layer of MgO embedding the SiO2
spheres and the specific local structure obtained through this
preparation method. Decreasing the size of these SiO2 spheres
further improved the selectivity towards butadiene production.
In addition, the introduction of a small amount of CuO as
promoter is shown to bring about a significant improvement
in both total ethanol conversion and butadiene yield, resulting
in a butadiene selectivity of around 53%; the added CuO in-
creases acetaldehyde production, effectively shifting the rate-
determining step of the process.
The samples SiO2–MgO (II-IV) were synthesized by a wet-kneading
(WK) technique similar to the one reported by Kvisle et al.[12] The
two uncalcined components (molar ratio of 1:1) were mixed at
room temperature for 4 h in water; in all cases Mg(NO3)2 ꢂ6H2O
was employed as precursor and the corresponding hydroxide was
obtained as described before. For catalyst SiO2–MgO (II), the SiO2
component was prepared using the modified Stçber route de-
scribed above. For catalyst SiO2–MgO (III), the SiO2 component
(final concentration is 0.34 M) was prepared by addition of the de-
sired amount of TEOS (17.3 g) at once to an ethanol–NH3 solution
in a closed vessel, followed by aging at 308 K overnight; the mate-
rial was subsequently dried in a rotary evaporator at 328 K. SiO2–
MgO (IV) was prepared by WK using Aerosil 300 silica (Degussa).
The catalysts SiO2–MgO (V) and (VI) were prepared by co-precipita-
tion (CP) methods: in the case of SiO2–MgO (V) an EtOH/NH3 =5:1
(v/v) solution (240 mL) was added at once to the desired amount
of TEOS (approximately 2.6 g). After 20 min, Mg(NO3)2 (approxi-
mately 3.2 g) dissolved in 200 mL of ethanol was added to the pre-
vious solution. SiO2–MgO (VI) was prepared through addition of
240 mL of ethanol–NH3 solution (5:1 v/v) to the same amounts of
the two precursors dissolved in 200 mL of ethanol.
For all the SiO2–MgO samples, the precipitate was washed several
times with deionized water and recovered through centrifugation,
dried at 393 K, and then calcined at 773 K for 5 h.
CuO (1 wt%) was supported on SiO2–MgO (II)–(VI) by IWI (incipient
wetness impregnation). 0.1 mL of a 0.63 M solution of Cu(NO3)2 ꢂ
3H2O in water was added to the support material (previously dried
for 1 h at 353 K) and, once the impregnation was completed, the
sample was left for 1 h to equilibrate, then dried for 12 h under
vacuum at RT, and finally, calcined at 773 K for 5 h.
The different preparation methods lead to significant differ-
ences in the amount of acidic and basic sites. The large
amounts of by-products formed by ethanol dehydration over
the co-precipitated catalysts can be directly related to the high
acidity of these materials. The wet-kneaded SiO2–MgO catalysts
are more basic, with the exact acidity/basicity depending on
the size of the SiO2 spheres, with smaller SiO2 spheres provid-
ing the best balance of basic and acidic sites.
UV/Vis studies provided further insights into the nature of
the catalyst materials, that is, in the extent of intimate mixing
of the SiO2 and MgO phases and into the type of CuO species
deposited and responsible for the increased butadiene yields.
Although CuO nanoparticles could not be seen by our TEM
analysis, Cu is present on both phases of the catalyst and small
cluster-like CuO species are proposed to have a positive
impact on butadiene formation. The results reported here thus
provide new insights into the structural characteristics required
for a good catalyst for the Lebedev process.
Further studies are now required to disclose more details on
the exact nature of acidic/basic active sites required for buta-
diene production and on the nature, location, and temporal
behavior of the CuO promoter. Together with the results pre-
sented herein, this will result in a more complete structure–ac-
tivity relationship and ultimately in improved methods for the
synthesis of efficient Lebedev catalysts.
Catalyst testing
Experimental Section
For all catalytic tests, quartz wool was placed in a U-shaped quartz
reactor, before addition of the catalyst (0.2 g; sieved to 425–90 mm
particle size). The desired amount of ethanol was fed through
a Bronkhorst CEM system consisting of three parts: a liquid flow
controller to check the amount of ethanol fed, a gas flow control-
ler for the nitrogen used as carrier gas, and finally, a mixing cham-
ber kept at 303 K where the gaseous mixture was formed and fed
downstream into the reactor. The total flow used was
100 mLminꢀ1, of which 2 mLminꢀ1 consisted of ethanol in the gas
phase. Reactions were run at 698 K. The analysis of the reaction
mixture was performed by means of GC-FID (gas chromatography
with flame ionization detector) using a CP poraplot Q-HT column;
quantification of the main components (ethanol, ethylene, acetal-
dehyde, butadiene, and diethyl ether) was based on calibration
curves obtained by feeding known amounts of the various com-
pounds. The following definitions were used (mol: moles of ob-
served substance:
Materials
Davicat Si 1404 silica from Grace and Aerosil 300 silica by Degussa
were used as purchased and without further treatment. Mg(NO3)2 ꢂ
6H2O (99+%, Acros), Cu(NO3)2 ꢂ3H2O (99%, Acros), and tetraethyl
orthosilicate (TEOS, 98%, Aldrich) were employed for the prepara-
tion of the different oxides. NH3 (25%, Merck) and ethanol (100%,
Interchema) were used during synthesis. Benzene (Sigma–Aldrich,
ACS reagentꢃ99.0%) was used for the Hammett indicator study.
Catalyst preparation
Six SiO2–MgO [denoted (I)–VI] catalysts (all with molar ratio 1:1)
were prepared in different ways to investigate the effect of prepa-
ration method on catalytic performance. SiO2–MgO (I) was a physi-
cal mixture (PM) prepared by mixing the two calcined individual
components, prepared as detailed below, for 10 min in a mortar.
MgO was prepared by dissolution of the nitrate precursor in water
(0.5 M), followed by dropwise addition of a 1m aqueous NH3 solu-
tion to precipitate the corresponding hydroxide (Mg(OH)2). The
precipitate was aged overnight, washed several times with deion-
Ethanol conversion:
molEtOH converted
XEtOHð%Þ ¼ 100 ꢄ
molEtOH initial
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2513

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Yields:
tral form) were added to the suspension of the catalyst in benzene.
After a few minutes, a color change was observed on the surface
of the solid catalyst, indicating that the majority of the indicator
was converted into its deprotonated form. The flask was then
stored for 24 h to confirm the color change (or its absence).
molproduct i
Yið%Þ ¼ 100 ꢄ
molEtOH initial
Selectivity of the i-main component:
molproduct i
Sið%Þ ¼ 100 ꢄ
Acknowledgements
Smolproduct i
This project has been performed in the framework of the Catch-
Bio program. The authors gratefully acknowledge the support of
the Smart Mix Program of the Netherlands Ministry of Economic
Affairs and the Netherlands Ministry of Education, Culture and
Science.
Catalyst characterization
TEM images of the SiO2–MgO samples were obtained on
a Tecnai 12 apparatus operated at 120 keV. The catalyst particles
were deposited on a TEM grid and analyzed as received. The CuO-
containing catalysts were imaged on a Tecnai F20 apparatus oper-
ating at 200 keV and equipped with an EDX detector. Nickel grids
were used for the EDX study of CuO-containing samples.
Keywords: butadiene · copper · ethanol · heterogeneous
catalysis · silica–magnesia
Textural properties of the materials were studied by nitrogen physi-
sorption measurements at 77 K using a Micromeritics Tristar 3000.
Prior to the adsorption measurements the samples were dried at
573 K for 12 h. The Brunauer–Emmett–Teller (BET) method was ap-
plied to calculate the total surface area. The t-plot method was ap-
plied to obtain the micropore volume and external surface area.
The Barrett–Joyner–Halenda (BJH) model was used to determine
the size of the mesopores.
[1] J. J. Bozell, G. R. Petersen, Green Chem. 2010, 12, 539–554.
[2] S. Gonzꢁlez-Garcꢄa, L. Luo, M. T. Moreira, G. Feijoo, G. Huppes, Biomass
Bioenergy 2012, 36, 268–279.
[3] R. P. John, G. S. Anisha, K. M. Nampoothiri, A. Pandey, Bioresour. Technol.
2011, 102, 186–193.
[4] C. Angelici, B. M. Weckhuysen, P. C. a Bruijnincx, ChemSusChem 2013, 6,
1595–1614.
[5] W. C. White, Chem.-Biol. Interact. 2007, 166, 10–14.
[6] P. C. A. Bruijnincx, B. M. Weckhuysen, Angew. Chem. Int. Ed. 2013, 52,
11980–11987; Angew. Chem. 2013, 125, 12198–12206.
[7] J. Ostromisslenski, J. Russ. Phys. Chem. Soc. 1915, 47, 1472–1506.
[8] S. V. Lebedev, Br. Pat. 331,402, 1929.
[9] S. V. Lebedev, J. Gen. Chim. 1933, 3, 698–708.
[10] S. K. Bhattacharyya, S. K. Sanyal, J. Catal. 1967, 7, 152–158.
[11] H. Niiyama, S. Morii, E. Echigoya, Bull. Chem. Soc. Jpn. 1972, 45, 655–
659.
[12] S. Kvisle, A. Aguero, R. P. A. Sneeden, Appl. Catal. 1988, 43, 117–131.
[13] E. V. Makshina, W. Janssens, B. F. Sels, P. A. Jacobs, Catal. Today 2012,
198, 338–344.
[14] R. Ohnishi, T. Akimoto, K. Tanabe, J. Chem. Soc. Chem. Commun. 1985,
1613–1614.
[15] 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.
[16] W. J. Toussaint, J. T. Dunn, D. R. Jackson, Ind. Eng. Chem. 1947, 39, 120–
125.
[17] G. Natta, R. Rigamonti, Chim. Ind. 1947, 29, 195–201.
[18] Y. Kitayama, M. Satoh, T. Kodama, Catal. Lett. 1996, 36, 95–97.
[19] A. D. Patel, K. Meesters, H. den Uil, E. de Jong, K. Blok, M. K. Patel,
Energy Environ. Sci. 2012, 5, 8430–8444.
[20] F.-W. Chang, H.-C. Yang, L. S. Roselin, W.-Y. Kuo, Appl. Catal. A 2006, 304,
30–39.
[21] N. Takezawa, C. Hanamaki, H. Kobayashi, J. Catal. 1975, 38, 101–109.
[22] W. M. Quattlebaum, W. J. Toussaint, J. T. Dunn, J. Am. Chem. Soc. 1947,
69, 593–599.
[23] S. Liu, X. Zhang, J. Li, N. Zhao, W. Wei, Y. Sun, Catal. Commun. 2008, 9,
1527–1532.
[24] Y. Ono, H. Hattori, Solid Base Catalysis, Springer, 2011.
[25] J. Take, N. Kikychi, Y. Yoneda, J. Catal. 1971, 21, 164–170.
[26] R. Brambilla, C. Radtke, J. H. Z. dos Santos, M. S. L. Miranda, Powder Tech-
nol. 2010, 198, 337–346.
[27] J. C. A. A. Roelofs, J. A. van Bokhoven, A. J. van Dillen, J. W. Geus, K. P.
de Jong, Chem. Eur. J. 2002, 8, 5571–5579.
[28] G. A. El-Shobaky, M. N. Hamed, F. F. Abdalla, S. A. El-Molla, Colloids Surf.
A 2002, 207, 293–301.
[29] S. Coluccia, A. J. Tench, R. L. Segall, J. Chem. Soc. Faraday Trans. 1979,
75, 1769–1779.
[30] S. Coluccia, A. Barton, A. J. Tench, J. Chem. Soc. Faraday Trans. 1981, 77,
2203–2207.
XRD patterns were obtained by using a Bruker-AXS D2 Phaser
powder X-ray diffractometer using CoKa1,2 with l=1.79026 ꢃ, oper-
ated at 30 kV. Measurements were carried out between 2q=108
and 1008 using a step size of 0.058 and a scan speed of 0.58 sꢀ1
.
UV/Vis spectra were collected using a Varian Cary 500 UV/Vis/NIR
spectrometer equipped with a diffuse reflectance spectroscopy
(DRS) accessory. The spectra were collected between 200 and
800 nm with a data interval of 1 nm and at a rate of 600 nmminꢀ1
.
NH3-TPD measurements were performed on
a Micromeritics
ASAP2920 apparatus. Typically, the sample (0.2 g) was placed in
a quartz reactor. The samples were dried in the apparatus in
a helium flow by heating with a temperature ramp of 5 Kminꢀ1 to
a maximum temperature of 873 K. Subsequently, the sample was
cooled to 373 K; at this point, NH3 pulses of 25.31 cm3 minꢀ1 were
applied. The sample was then heated to 873 K with a ramp of
5 Kminꢀ1 to induce desorption of NH3. A procedure similar to the
one described for NH3-TPD was employed in the case of CO2-TPD,
the main difference being the lower temperature (313 vs 373 K) at
which CO2 pulses were fed to the sample. For the calculation of
the number of acidic or basic sites, it was assumed that only one
molecule of NH3 or CO2 could adsorb on a single site.
For the Hammett indicator study, benzene was dried over molecu-
lar sieves and stored under an argon atmosphere. All experiments
were performed using standard Schlenk techniques. A solution of
each indicator was prepared by dissolving a specific indicator
(25 mg) in dry benzene (25 mL). The six indicators used were (in
order of increasing pKa): bromothymolblue (ABCR, pKa =7.2), phe-
nolphtalein (Sigma–Aldrich, ACS reagent, pKa =9.3), 2,4-dinitroani-
line (Acros Organics, 99%, pKa =15.0), 4-chloro-2-nitroaniline
(Sigma–Aldrich purum, ꢃ98% (HPLC), pKa =17.2), 4-nitroaniline
(Acros Organics, 99%, pKa =18.4), and 4-chloroaniline (Sigma–Al-
drich, 98%, pKa =26.5). Before each experiment, the catalyst (0.1 g)
was dried at 473 K in vacuum. The flask containing the catalyst
was then put under argon atmosphere, after which 2 mL of dry
benzene were added to the sample under investigation; at this
point a few drops of the indicator solution (in the protonated, neu-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2514

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[31] K. V. R. Chary, K. K. Seela, G. V. Sagar, B. Sreedhar, J. Phys. Chem. B 2004,
108, 658–663.
[36] A. Kong, H. Wang, X. Yang, Y. Hou, Y. Shan, Micropor. Mesopor. Mat.
2009, 118, 348–353
[32] M. Shimokawabe, N. Takezawa, H. Kobayashi, Appl. Catal. 1982, 2, 379–
387.
[37] J. J. Bravo-Suꢁrez, B. Subramaniam, R. V. Chaudhari, J. Phys. Chem. C
2012, 116, 18207–18221;
[33] S. Velu, K. Suzuki, M. Okazaki, M. P. Kapoor, T. Osaki, F. Ohashi, J. Catal.
2000, 194, 373–384.
[38] G. Derrien, C. Charnay, J. Zajac, D. J. Jones, J. Roziꢅre, Chem. Commun.
2008, 3118–3120.
[34] H. Praliaud, S. Mikhailenko, Z. Chajar, M. Primet, Appl. Catal. B 1998, 16,
359–374.
Received: April 29, 2014
[35] Z. Wang, H. Wan, B. Liu, X. Zhao, X. Li, H. Zhu, X. Xu, F. Ji, K. Sun, L.
Dong, Y. Chen, J. Colloid Interface Sci. 2008, 320, 520–526.
Published online on July 14, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2505 – 2515 2515