Slowing the Kinetics of Alumina Sol–Gel Chemistry for Controlled Catalyst Overcoating and Improved Catalyst Stability and Selectivity

Article abstract of DOI:10.1002/smll.201801733

Catalyst overcoating is an emerging approach to engineer surface functionalities on supported metal catalyst and improve catalyst selectivity and durability. Alumina deposition on high surface area material by sol–gel chemistry is traditionally difficult to control due to the fast hydrolysis kinetics of aluminum-alkoxide precursors. Here, sol–gel chemistry methods are adapted to slow down these kinetics and deposit nanometer-scale alumina overcoats. The alumina overcoats are comparable in conformality and thickness control to overcoats prepared by atomic layer deposition even on high surface area substrates. The strategy relies on regulating the hydrolysis/condensation kinetics of Al(^sBuO)₃ by either adding a chelating agent or using nonhydrolytic sol–gel chemistry. These two approaches produce overcoats with similar chemical properties but distinct physical textures. With chelation chemistry, a mild method compatible with supported base metal catalysts, a conformal yet porous overcoat leads to a highly sintering-resistant Cu catalyst for liquid-phase furfural hydrogenation. With the nonhydrolytic sol–gel route, a denser Al₂O₃ overcoat can be deposited to create a high density of Lewis acid–metal interface sites over Pt on mesoporous silica. The resulting material has a substantially increased hydrodeoxygenation activity for the conversion of lignin-derived 4-propylguaiacol into propylcyclohexane with up to 87% selectivity.

Full text of DOI:10.1002/smll.201801733

FULL PAPER
Nanostructured Catalysts
www.small-journal.com
Slowing the Kinetics of Alumina Sol–Gel Chemistry
for Controlled Catalyst Overcoating and Improved Catalyst
Stability and Selectivity
Yuan-Peng Du, Florent Héroguel, and Jeremy S. Luterbacher*
and selectivity.[1,2] Catalyst overcoating
is an emerging postsynthesis modifica-
tion strategy for engineering the surface
Catalyst overcoating is an emerging approach to engineer surface
functionalities on supported metal catalyst and improve catalyst selectivity
and durability. Alumina deposition on high surface area material by sol–gel
chemistry is traditionally difficult to control due to the fast hydrolysis kinetics
of aluminum-alkoxide precursors. Here, sol–gel chemistry methods are
adapted to slow down these kinetics and deposit nanometer-scale alumina
overcoats. The alumina overcoats are comparable in conformality and thick-
ness control to overcoats prepared by atomic layer deposition even on high
surface area substrates. The strategy relies on regulating the hydrolysis/con-
densation kinetics of Al(sBuO)3 by either adding a chelating agent or using
nonhydrolytic sol–gel chemistry. These two approaches produce overcoats
with similar chemical properties but distinct physical textures. With chelation
chemistry, a mild method compatible with supported base metal catalysts, a
conformal yet porous overcoat leads to a highly sintering-resistant Cu cata-
lyst for liquid-phase furfural hydrogenation. With the nonhydrolytic sol–gel
route, a denser Al2O3 overcoat can be deposited to create a high density of
Lewis acid–metal interface sites over Pt on mesoporous silica. The resulting
material has a substantially increased hydrodeoxygenation activity for the
conversion of lignin-derived 4-propylguaiacol into propylcyclohexane with up
to 87% selectivity.
nanostructure and functionality of various
catalysts.[3,4] Many studies have reported
improving the selectivities and stabili-
ties of a variety of catalysts by depositing
SiO2 overcoats.[5–8] However, silica is not
hydrothermally stable which has limited
its use in renewable catalytic systems that
generally feature aqueous conditions.[9,10]
Catalytic systems featuring biomass-
derived streams are one such category
of reactions where the water content in
plants and/or hydrolysis-based decon-
struction processes lead to hydrothermal
catalytic conversions. This has increased
the interest in more hydrothermally stable
oxides like alumina (Al2O3). Al2O3 deposi-
tion has also generated interest because the
addition of its intrinsic Lewis acidity has a
potential to create new active sites.[11,12]
Atomic layer deposition (ALD) is well
known as an effective method for depos-
iting alumina on substrate in either
powder or film form and it has been widely
applied to improve the performances of
heterogeneous catalysts.[4,13–15] The greatest advantage of ALD
is the highly controllable growth of the deposited overcoat with
respect to thickness and uniformity.[16] However, ALD is hard
to scale and requires a heavy instrumentation and the use of
expensive gaseous precursors. In contrast, sol–gel methods are
considered simpler and cheaper alternatives.[17] Nonetheless,
depositing Al2O3 over nanoparticles with high surface areas via
sol–gel process remains particularly challenging because of the
high hydrolysis/condensation rate of aluminum alkoxides.[18]
For instance, SiO2 overcoating can be easily performed by
applying the so-called “Stöber method,” which is carried out by
adding tetraethoxysilane (TEOS) into a suspension containing
the substrate, a controlled amount of ethanol, ammonia, and
water.[19] However, the same procedure with classic alumina
precursors such as aluminum sec-butoxide (Al(sBuO)3) leads to
the rapid formation of alumina precipitates.
1. Introduction
Supported metal catalysts are widely used in many commercial
enterprises including petroleum refineries, emission control,
pharmaceutical production and, increasingly, in biomass con-
version. In the past, catalyst research focused on controlling
metal particle size and screening suitable supports to enhance
catalytic activities. As synthetic techniques increasingly allow
us to control the material’s nanoenvironment, the importance
of catalyst structure and surface properties has begun to be rec-
ognized as it is found to be closely related to catalyst stability
Y.-P. Du, Dr. F. Héroguel, Prof. J. S. Luterbacher
Laboratory of Sustainable and Catalytic Processing
Institute of Chemical Sciences and Engineering
École Polytechnique Fédérale de Lausanne (EPFL)
CH-1015, Lausanne, Switzerland
In order to synthetize conformal overcoats with a controlled
thickness and density using liquid phase sol–gel chemistry
techniques, our main strategy was to tailor the condensation
kinetics of an alumina precursor. We had previously explored
nonhydrolytic sol–gel (NHSG) chemistry to deposit titania
E-mail: jeremy.luterbacher@epfl.ch
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/smll.201801733.
DOI: 10.1002/smll.201801733
©
1801733 (1 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2018, 1801733

www.advancedsciencenews.com
www.small-journal.com
(TiO2) that has similar highly reactive precursors in aqueous
media.[20] Currently, no NHSG methods exist for Al2O3 dep-
osition on high surface area materials. On the other hand,
ligands containing β-dicarbonyl groups, such as acetylacetone
(AcAc) and ethyl acetoacetate (EAA) have also been reported
to chelate with metal alkoxides by ligand substitution and
slow down their high condensation rates. Although chelation
chemistry methods have been previously applied for alumina
deposition on thin films,[21,22] it has similarly not been used
for high surface area powder samples such as those used
in heterogeneous catalysis. Our objective was to develop a
toolbox of alumina overcoating methods using these comple-
mentary methods that both slow down the kinetics of alumina
condensation.
Here, we demonstrate that the sol–gel kinetics can be modi-
fied by nonhydrolytic sol–gel and chelation chemistry to achieve
conformal alumina overcoats with controlled thicknesses of
nanometer precision. Specifically, we report two Al2O3 deposi-
tion methods performed in liquid phase with Al(sBuO)3 which
is an inexpensive and common precursor for alumina sol–gel
processing. We applied said methods to overcoat two typical
high surface area supported metal catalysts and demonstrate
the complementary uses of both resulting catalysts in biomass
conversion reactions, which lead to increased stability and
activity, respectively.
was systematically initiated by injecting the precursors into
these respective suspensions. A syringe pump (Figure 1b) was
used to inject the precursors and avoid any undesirable homo-
geneous condensation due to a high precursor concentration
within the catalyst suspension.
In order to approximately estimate the amount of precursor
needed to form a single atomic monolayer on the surface of
the catalyst, we calculated the surface projection of the Van Der
Waals volumes of the precursor molecules by MarvinSketch
software (Figure 1c), which were estimated to be 0.48 and
1.0 nm2 for Al(sBuO)Br2 and Al(sBuO)2(EAA), respectively.
Our past work with alumina has shown that steric hin-
drances between precursors is the determining factor for cal-
culating maximum precursor coverage on the surface instead
of hydroxyl density.[28] The necessary amount of precursor for
achieving monolayer coverage on a specific material was then
calculated by dividing the specific surface area of the catalyst
by this projected precursor area. Knowing the amount of pre-
cursor needed to form a monolayer allows us to target a tar-
geted number of deposited monolayers by controlling the total
amount of injected precursor, leading to sub-nanometer control
of the overcoat thickness.
2.2. Al2O3 Deposition by Chelation Chemistry
Al2O3 overcoating using the chelating method was first per-
formed on SiO2 spheres with relatively low specific surface area
(as determined by the Brunauer–Emmett–Teller (BET) method
as 15 m2 g−1, SBET) to easily image the overcoat. An amount of
40 monolayers coverage of Al(sBuO)2(EAA) was deposited and
the product was referred as C-Al2O3@SiO2, where C denotes
materials prepared using the chelation chemistry method.
Though high-angle annular dark field scanning transmission
electron microscopy (HAADF STEM) provides low contrast
between the SiO2 core and the Al2O3 shell (Figure 2a), energy
dispersive X-ray (EDX) spectroscopy mapping clearly shows
that a conformal alumina overcoat was formed on the silica
spheres (Figure 2b) with an average thickness of 6.1 2.0 nm.
We subsequently investigated Al2O3 deposition on Pt supported
on high surface area mesoporous silica (SBA-15), which is
more representative of a material used for catalytic applications.
The final injected amount of Al per unit weight of SBA-15 was
3.4 mmol gcatalyst−1, which corresponded to approximately three
monolayers of Al(sBuO)2(EAA). Figure 2c,d shows no modifica-
tion of SBA-15’s morphology after alumina deposition (for com-
parison, a TEM image of the uncoated material is available in
Figure S1, Supporting Information). EDX mapping (Figure 2e)
showed a uniform distribution of Si and Al, which implied
that Al2O3 growth occurred within the pore structure. Forma-
tion of the Al2O3 layer inside the pores was also supported by
the measurement of a reduced average pore size (from 6.8 to
6.3 nm) as determined by nitrogen physisorption (Figure S2a,
Supporting Information). Notably, the periodic hexagonal
structure of SBA-15 was fully preserved after overcoating as
evidenced by powder X-ray diffraction (XRD) (Figure S2b,
Supporting Information). In agreement with previous work
studying the surface alumination of SBA-15 and NHSG TiO2
deposition, the SBET of C-Al2O3@2%Pt/SBA-15 was drastically
2. Results and Discussions
2.1. Design of Al2O3 Precursors with Slower
Hydrolysis/Condensation Kinetics
Our first strategy was chemically modifying Al(sBuO)3 based
on NHSG chemistry of alumina, which was first reported by
Acosta et al.[23] Typically, the precursor of the NHSG reaction
can be prepared by mixing metal halides and metal alkoxides.
Ligand exchange occurs between the metal halide and metal
alkoxide at room temperature while condensation only takes
place at higher temperatures. The gelation is likely initiated
by the ligand exchanged precursors as shown in the previous
studies of TiCl4–Ti(iPrO)4 NHSG chemistry.[24] The actual pre-
cursors involved in those gelation processes have been iden-
tified as Ti(iPrO)3Cl and Al(iPrO)2Cl for Ti(iPrO)4–TiCl4 and
Al(iPrO)3–AlCl3 pairs, respectively.[25] Mixing metal alkoxides
and metal halides with nonoptimal ratios has been shown to
cause long gelation time.[24] Therefore, the ratio of Al(sBuO)3
and AlBr3 was kept to 2 by presuming that Al(sBuO)3–AlBr3
route follows the same chemistry reported in the literature of
analogous Ti and Al systems (Figure 1a).[25] AlBr3 was preferred
over AlCl3 to allow using CH2Br2 as a solvent instead of the
more toxic CCl4.[23] The second strategy was adding EAA to che-
late Al(sBuO)3 by ligand substitution.[18,26,27] The sol–gel kinetics
of the resulting chelated alumina precursor and its associated
gelation rate is dependent on the ratio between chelating agents
and Al(sBuO)3. In this work, we prepared the precursor by
reacting 0.75 equivalent EAA with Al(sBuO)3 (Figure 1a).
After preparing the precursors, the catalysts were either
dispersed in a Stöber solution (chelated precursor) or a heated
anhydrous CH2Br2 (NHSG precursors). Alumina deposition
©
Small 2018, 1801733
1801733 (2 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Figure 1. a) NHSG and chelation chemistry of Al(sBuO)3 for controlled Al2O3 deposition. b) Schematic illustration of the continuous injection syringe
pump setup. c) Van Der Waals structure and the volumes of the two chemically modified Al(sBuO)3.
decreased, from 723 to 392 m2 g−1, which has been determined
as being linked to the filling of micropores.[20,29] Importantly,
this method is not merely restricted to overcoating silica-based
substrates. We successfully overcoated other metal oxides like
Fe2O3 ellipsoid nanoparticles and even hydrophobic substrates
such as carbon spheres (Figure S3, Supporting Information).
Previous attempts to overcoat Al2O3 on powder substrates
by sol–gel based methods had been made by repeatedly
grafting aluminum alkoxides with surface hydroxyl groups
in anhydrous conditions.[29–31] Nonetheless, the overlayers
in those studies did not uniformly grow over the surfaces or
pore structures. We previously reported a new alumina over-
coating method to control the condensation by stochiometri-
cally controlling the H2O amount but this method gave poor
results when coating catalysts with SBET above 150 m2 g−1.[28]
In contrast, the overcoat of C-Al2O3@SiO2 was comparable in
appearance and conformality to the alumina overcoat produced
by ALD.[32] Another disadvantage of those published methods
was that they required time-consuming multistep approaches
to achieve thicker overcoats. However, this presented approach
can be performed in a single step with an automated injection.
Furthermore, it still produced high quality (e.g., uniform and
conformal) overcoats when the amount of alumina loading was
increased by five times (14 monolayer equivalents of precursor,
sample denoted as C-5Al2O3@SBA-15, Figure S1a, Supporting
Information). In this case, the average pore size of C-5Al2O3@
SBA-15 was reduced to 4.1 nm while narrow pore size distri-
butions as well as the characteristic diffraction peak of SBA-15
were still observed (Figure S2, Supporting Information).
According to the pore size determined by physisorption, the
thickness of the alumina overcoat in the pores (1.35 nm) was
close to the predicted thickness based on the calculated thick-
ness of one monolayer of alumina (1.25 nm for 14 monolayers,
or 0.09 nm per monolayer estimated from C-Al2O3@2%Pt/
SBA-15).
Several important parameters govern our chelation method.
First, the quantity of added H2O is crucial for controlling hydrol-
ysis kinetics. The ratio of water and ethanol used in this work
was much lower (0.04) than that present in the Stöber solution
used for preparing spherical silica (0.15) because we observed
that Al(sBuO)2(EAA) still reacts with H2O faster than tetraeth-
oxysilane does. Nevertheless, an insufficient concentration of
water hindered the gelation of precursor and so an appropriate
balance was required. Second, the concentration of NH3, which
acts as a catalyst to initiate the hydrolysis and condensation of
metal alkoxides, needs to be decreased when overcoating high
surface area substrates with porous structures such as SBA-15.
A high NH3 concentration led to preferential condensation
on the surface of SBA-15 particles before the precursor could
diffuse into the pore structure. Figure S1b in the Supporting
Information shows the image of overcoated SBA-15 prepared
with an excess NH3(aq) (0.04 mL instead of 0.02 mL). The for-
mation of fibrous structures around the particles was indicative
of alumina growth outside of the catalyst pellet. Third, higher
EAA/Al(sBuO)3 ratios are known to decrease gelation times.[21]
Therefore, NH3 and/or H2O concentration had to be adjusted
to obtain suitable gelation times when a higher EAA/Al(sBuO)3
ratio was used. These effects showed how the interplay of EAA
binding, water, and ammonia can be used to control the sol–gel
kinetics of Al(sBuO)3.
©
1801733 (3 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2018, 1801733

www.advancedsciencenews.com
www.small-journal.com
was a similar order of magnitude albeit
below that of N-Al2O3@SiO2. This differ-
ence could either be due to the fact that our
imaging method did not allow us to measure
an overcoat thickness measurement as pre-
cisely as through physisorption, or that the
initial micropore filling in SBA-15 skewed
the pore size measurement by not contrib-
uting to the reduction of pore diameters in
the mesopore range.
Compared to the chelation method, the
overall concentration of the precursor loaded
in the syringe was particularly important for
controlling the Al2O3 growth during NHSG-
based overcoating. Since the condensation
was initiated by the higher temperature in
the reaction flask, we found that the hot vapor
generated during reflux caused the undesired
gelation of precursor droplets before they
dripped into the substrate suspension. When
the total aluminum concentration in the
syringe was set below 0.1 m, this early gela-
tion did not occur anymore. With this adjust-
ment, we were able to avoid any uncontrolled
growth of Al2O3 outside the pore structure of
SBA-15 (Figure S4, Supporting Information).
It is also worth mentioning that the thick-
ness of Al2O3 overcoats on 2%Pt/SBA-15 pro-
duced with the NHSG method were thinner
then that prepared by the chelation method
when the same molar amount of Al was
injected (Figure S2a, Supporting Informa-
tion). Such difference likely resulted from
the less bulky NHSG precursor (Figure 1c),
which could lead to a denser deposited mono-
Figure 2. TEM and STEM images of C-Al2O3@SiO2 a) under HAADF mode and b) with EDX
mapping or c) C-Al2O3@2%Pt/SBA-15 d) under HAADF mode and e) with EDX mapping.
2.3. Deposition of Al2O3 by Nonhydrolytic Sol–Gel Chemistry
layer. We further investigated the physical properties of these
overcoats by analyzing the overcoated silica spheres, which
allowed us to more accurately probe the overcoat texture and
thickness by microscopy. Because the overcoat on C-Al2O3@
SiO2 was thinner than that of N-Al2O3@SiO2, silica spheres with
a thicker alumina coating were prepared for analysis (C9nm-
Al2O3@SiO2, the synthesis conditions are given in Section S1,
Supporting Information). Physisorption results (Figure S5, Sup-
porting Information) suggested that the pore volume of C9nm-
Al2O3@SiO2 was three times higher than that of N-Al2O3@SiO2
(0.11 and 0.03 cm3 g−1, respectively), which confirmed that two
overcoating routes lead to different Al2O3 structural properties.
We suggested that is not only due to the different steric hin-
drances of precursors but also the differences in condensation
chemistry and growing mechanisms. We hypothesized that the
rapid initial grafting reaction between Al(sBuO)Br2 and surface
hydroxyl groups could ensure consecutive surface condensa-
tions, leading to a denser alumina network. On the other hand,
Al(sBuO)2(EAA) may proceed through an oligomer deposition
mechanism. Specifically, the precursors start polymerizing into
sub-nanometer oligomers (sols) when injected into catalyst
suspension. These sols gradually crosslink and heterogeneous
nucleation may only happen when supersaturation is reached
by progressive injection. Our previous coarse grain modeling
After depositing an amount of precursor corresponding to
40 monolayers of Al2O3 on SiO2 spheres, the formation of a rough
surface was observed by STEM (Figure 3a). EDX mapping
revealed a conformal alumina overcoat with an average thick-
ness 12 4 nm, which corresponded to a monolayer thickness of
about 0.3 nm. The coating of 2%Pt/SBA-15 was also performed
−1
and 3.4 mmol gcatalyst of Al was injected (corresponding to
a coverage of 1.3 monolayer for Al(sBuO)Br2), which was the
same molar amount as the aforementioned C-Al2O3@2%Pt/
SBA-15 material (though corresponding to fewer monolayer).
Similar to the materials obtained using our chelating method,
deposition with NHSG chemistry did not alter the SBA-15
morphology (Figure 3c,d) and the characteristic XRD peaks of
SBA-15 (Figure S2b, Supporting Information). The uniform
mixture of SiO2 and Al2O3 observed by EDX (Figure 3e) again
suggested that alumina was deposited inside the pores, which
was further supported by the reduced average pore diameter
from 6.8 to 6.6 nm (Figure S2a, Supporting Information). We
then estimated that the thickness formed by one monolayer
equivalent of Al(sBuO)Br2 to be roughly 0.075 nm of alumina
overcoat, which was comparable but slightly denser than what
was obtained for C-Al2O3@2%Pt/SBA-15. This thickness
©
Small 2018, 1801733
1801733 (4 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
amounts of Al(sBuO)3 and H2O and applied
it to the same reaction.[28] Although that
method also stabilized Cu/Al2O3, it was
only able to stabilize the Cu on Al2O3
with low SBET (45 m2 g−1) as the use of this
method on the high surface area Al2O3 led
to no discernable stabilization (Figure S6,
Supporting Information), which limits the
applicability of our prior method to less indus-
trially relevant supports. Our first attempts to
deposit Al2O3 using NHSG method leached
all the copper out during the coating proce-
dure, presumably due to the presence of the
corrosive HBr byproduct. Therefore, we used
the chelated Al(sBuO)3 precursor for over-
coating Cu/Al2O3 to synthesize C-Al2O3@Cu/
Al2O3 and demonstrate the benefit of having
several available methods to slow down pre-
cursor condensation kinetics. Although in
initial tests, we had found that EAA interacted
with metallic copper and caused leaching
(the color of reaction media turned blue
after overcoating) during the coating process,
this issue was easily resolved by oxidizing
the copper with a mild thermal treatment
before overcoating. With this pretreatment,
the color change was not observed anymore.
The SBET of the resulting C-Al2O3@Cu/Al2O3
slightly decreased to 144 from 154 m2 g−1.
The loading of Cu was determined induc-
tively coupled plasma optical emission spec-
trometry (ICP-OES) and it was reduced from
5.0 to 2.4 wt%, as a result of the additional
weight of Al2O3 in the overcoat. Based on N2O
chemisorption, the original number of Cu
surface sites on Cu/Al2O3 was 53.6 µmol g−1
Figure 3. TEM and STEM images of N-Al2O3@SiO2 a) under HAADF mode and b) with EDX
mapping or c) N-Al2O3@2%Pt/SBA-15 d) under HAADF mode and e) with EDX mapping.
studies showed that the preferred mechanism of Al2O3 deposi-
tion in liquid phase using Al(sBuO)3 occurs via an oligomer depo-
sition mechanism, where the alumina overcoat formed by the
polymerization of alumina oligomers instead of monomeric pre-
cursors.[33] If the formation of alumina by the chelation method
did indeed proceed through this oligomer-deposition mecha-
nism, this would likely lead to a much more porous framework.
whereas no surface Cu sites were detected in the as-synthesized
C-Al2O3@Cu/Al2O3. However, Cu sites could be re-exposed by
calcination, after which 9.9 µmol g−1 of Cu surface sites was
measured. This re-exposure of copper could be attributed to
either pore formation by overcoat crystallization or the removal
of residual precursor ligands strongly bound on the metal par-
ticles.[8,14] If all the active sites were re-exposed, the number of
Cu surface sites should be roughly 25.7 µmol g−1 because of the
reduced Cu loading. Therefore, ≈38% of active sites were recov-
erable after overcoating.
The uncoated and overcoated catalysts were then used to
catalyze furfural hydrogenation in a fixed-bed flow reactor
(Figure 4). In both cases, coke formation led to rapid deacti-
vation of the catalyst, which could be partially reversed. How-
ever, the sintering of Cu particles during calcination and the
liquid phase reaction could cause irreversible deactivation. The
activity of Cu/Al2O3 was not fully recovered after calcination
(Figure 4a), whereas the deactivation of C-Al2O3@Cu/Al2O3
could be completely reversed (Figure 4b). The spent catalysts
were analyzed by TEM (Figure S7, Supporting Information) and
significant particle sintering was observed on Cu/Al2O3. In con-
trast, most of the Cu particles in C-Al2O3@Cu/Al2O3 were still
well dispersed after five catalytic cycles, which can be ascribed
2.4. Improving the Stability of Cu/Al2O3 for Furfural
Hydrogenation
As we and others have shown, deposition of a metal oxide over-
coat onto a supported heterogeneous catalyst can stabilize the
metal nanoparticles against sintering.[6,14,34–36] However, stabili-
zation is only successful if the overcoat is conformal and success-
fully covers all of the metal nanoparticles. Herein, we use furfural
hydrogenation, which is a copper-based industrially relevant
reaction for producing furfural alcohol (FFA), to investigate the
stability of an overcoated copper catalyst and, indirectly, measure
the quality and conformality of the coating. Previously, we had
also developed a layer-by-layer alumina overcoating method
achieved by alternately reacting substrate with stoichiometric
©
1801733 (5 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2018, 1801733

www.advancedsciencenews.com
www.small-journal.com
Figure 4. Results of continuous furfural hydrogenation over a) Cu/Al2O3 and b) C-Al2O3@Cu/Al2O3.
to the physical barrier formed by the protective Al2O3 overcoat.
These results suggested that the chelation chemistry-based
approach works on a broader range of supports compared to
previous methods, which underlines the importance of slowing
condensation kinetics. Accordingly, we consider chelation-based
overcoating to be an easily performed but powerful method for
overcoating a large array of supported metal catalysts, including
those containing less leaching-resistant base metals.
The HDO of 4-propylguaiacol was studied over a series of over-
coated and uncoated Pt-based catalysts. Our HDO reaction tem-
perature was somewhat lower compared to most (though not all)
reported HDO conditions of lignin model compounds (typically
above 200 °C)[48–50] to ensure that the pore structure of SBA-15
remains stable (TEM images of spent catalysts were shown in
Figure S8, Supporting Information). Among all catalysts we syn-
thesized, uncoated 2%Pt/SBA-15 had the highest rate in con-
verting 4-propylguaiacol (Figure 5a), likely because of its greater
number of accessible Pt sites (Table 1). Full conversion was
reached after 3 h of reaction. Nevertheless, it had the lowest selec-
tivity to fully deoxygenated products and the main product was
2-methoxy-4-propylcyclohexanol, which resulted only from ring
hydrogenation. The selectivity of this product was 61% at a conver-
sion of 59% and decreased with increasing conversion. However,
increasing conversion only yielded 4-propylcyclohexanol and very
small amount of fully deoxygenated C9 alkane (propylcyclohexane),
which indicated its limited HDO activity. Replacing the SBA-15
support with Al2O3 only slightly improved the selectivity to pro-
pylcyclohexane (≈40%) because the Lewis acidity facilitated HDO.
Interestingly, the two differently overcoated catalysts displayed
very different selectivities. C-Al2O3@2%Pt/SBA-15 led to a similar
propylcyclohexane selectivity to Pt/Al2O3 and other products like
1-propylcyclohexene and 4-propylcyclohexanone were also detected
(the yields of these minor products are shown in Table S1, entry 7,
Supporting Information), indicating that the overcoat prepared
using the chelation method did not improve deoxygenation selec-
tivity. In contrast, N-Al2O3@2%Pt/SBA-15 showed a significantly
enhanced propylcyclohexane selectivity (80%). N-Al2O3@2%Pt/
SBA-15 did display a slightly lower conversion (77%) at the
same reaction time compared to C-Al2O3@2%Pt/SBA-15, which
was attributed to its reduced number of accessible Pt sites as
determined by CO chemisorption (Table 1). Notably, using
N-Al2O3@2%Pt/SBA-15 led to a yield of about 10% aromatic
products (4-propylphenol and propylbenzene). The selectivities
toward these aromatic products were decreased when full conver-
sion was reached, where the selectivity toward propylcyclohexane
was close to 90% (Table S1, entry 6, Supporting Information).
The production of aromatics suggests that the reaction proceeds
through aromatic intermediates when N-Al2O3 overcoat is pre-
sent. In comparison, selectivities to aromatic products were always
less than 2.5% at the same reaction time when the other three
2.5. Improving the Selectivity of Pt/SBA-15
for 4-Propylguaiacol Hydrodeoxygenation
To test the potential of our overocoats for improving catalyst
activity, we tested our materials for the hydrodeoxygenation
(HDO) of lignin-derived propyl guaiacol. Lignin is a plant frac-
tion that is a polymer of phenylpropanoid subunits. Because
of its aromatic-rich and energy dense structure, lignin is one
of the most promising potential sources of chemicals and
drop-in fuels to be derived from biomass.[37,38] However, the
molecules that can be directly produced from lignin contain
too much oxygen to be used directly as a fuel.[39] A typical
example of such a molecule is propyl guaiacol, which we
and others have produced as a majority product from soft-
wood and, as the second most abundant product from hard-
wood lignin.[40–43] Hydrodeoxygenation of this molecule (and
closely related ones) is a typical route for obtaining high
quality fuels from lignin. In these reactions, the synergistic
effects between metal and acid sites play an important role
and a variety of bifunctional catalysts for catalytic upgrading
of oxygenated phenolic molecules have been reported.[20,44–47]
Although zeolites are an industrially relevant support, having
abundant acid sites and high surface area, their hydrothermal
stability is an issue for widespread use in biomass upgrading
due to the amount of water that is generally present in such
streams. Furthermore, their small pore size limits the diffu-
sion of several molecules including lignin-derived oligomers
and even some monomers.[45] Hence, creating acid sites on
mesoporous SBA-15 supported metal catalysts would be an
interesting application of Al2O3 overcoating method for syn-
thetizing a stable bifunctional catalyst with a larger pore size
for lignin valorization.
©
Small 2018, 1801733
1801733 (6 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
Figure 5. a) Selectivities and conversions of 4-propylguaiacol HDO over Pt based catalysts. The reaction time was 3 h and the amounts of catalysts
were all 20 mg except for 1%Pt/Al2O3 (40 mg, due to its metal loading being only half of 2%Pt/SBA-15). The first bar on the left represents the low
conversion result for 2%Pt/SBA-15. The conditions for these data were a reaction time of 1.5 h and a fivefold loading of 4-propylguaiacol substrate.
Products listed as Others included 1-propylcyclohexene and 4-propylcyclohexanone. b) Reaction pathways of 4-propylguaiacol HDO.
catalysts were used. As has been previously reported, when using
noble metal based catalyst, the benzene ring of 4-propylguaiacol
is rapidly hydrogenated as the first step and subsequent deoxy-
genation occurs (Figure 5b, pathway 1), which is in agreement to
our observations.[47,48] On the other hand, transition metal cata-
activity did not follow the same trend as N-Al2O3@2%Pt/
SBA-15 showed the highest selectivity toward fully deoxy-
genated alkanes compared to other catalysts. Powder XRD
analyses confirmed the amorphous nature of alumina over-
coats produced from both methods (Figure S9, Supporting


lysts can catalyze the hydrogenolysis of C O bonds first and then
Information). According to the shift of the Si OH band at
−1
the hydrogenation of aromatic ring. However, even when this
occurs, this pathway is much less favored than the alternate.[50]
The majority of aromatic intermediates that were produced in the
3745 cm−1 to the Al OH band at 3739 cm in the fourier-

transform infrared (FTIR) spectra and the shifts as well as the
broadenings of the proton signals in the H magic angle spin-
1

case of N-Al2O3@2%Pt/SBA-15 indicates that the C O cleavage
ning solid-state nuclear magnetic resonance (ssNMR) spectra
(Figures S10 and S11, Supporting Information), we con-
cluded that the surfaces of both N-Al2O3@2%Pt/SBA-15 and
C-Al2O3@2%Pt/SBA-15 were fully covered by Al2O3.[52] Addi-
tionally, diffuse reflectance infrared Fourier transform spec-
troscopy with a pyridine probe (pyridine-DRIFT) suggested
that both overcoated catalysts had stronger Lewis acid sites
compared to 1%Pt/Al2O3 (Figure S12, Supporting Informa-
tion). Therefore, the different selectivities of C-Al2O3@2%Pt/
SBA-15 and N-Al2O3@2%Pt/SBA-15 seem to be unrelated to
the crystal structure of alumina, degree of overcoat coverage
and the strength of acidity. We performed 27Al solid-state NMR
to further elucidate the chemical compositions of two Al2O3
overcoats (Figure 6a). The octahedrally coordinated Al sites
predominated in 1%Pt/Al2O3, which corresponds to the char-
acteristic structure of γ-Al2O3. Interestingly, penta-coordinated
mechanism might be the primary one over this particular catalyst.
However, no conversion was observed when the reaction was run
under N2 with this catalyst (Table S1, Supporting Information),
which means that any HDO still required an H2 environment to
proceed.
2.6. The Effects of Physical and Chemical Properties
of Overcoats on Hydrodeoxygenation
Pt particles and surface acid sites have previously been reported
to have a synergistic effect during hydrodeoxygenation.[45,47,51]
The catalyst with the most to the fewest acid sites per unit
surface area were 2%Pt/Al2O3 > N-Al2O3@2%Pt/SBA-15 >
C-Al2O3@2%Pt/SBA-15 (Table 1). However, the deoxygenation
Table 1. Summary of the characterization data of Pt catalysts.
Catalyst
Al/Si ratioa)
N/A
dPt [nm]d)
3.5 2.4
3.8 1.5
4.3 2.6
5.0 1.9
SBET [m2 g−1
723
]
Vmicropore [cm3 g−1 b)
]
Lewis acid sites [µmol m−2 c)
]
Ptsurface [µmol g−1
]
2%Pt/SBA-15
0.03
0
58.5
9.4
≈0
1.3
0.9
1.5
N-Al2O3@2%Pt/SBA-15
C-Al2O3@2%Pt/SBA-15
1%Pt/Al2O3
0.50
449
0.47
392
0.01
0
30.2
35.2
N/A
164
a)Determined by inductively coupled plasma optical emission spectrometry (XPS), using the signals of Si 2s and Al 2s; b)Estimated using the t-plot method; c)Determined
by NH3 temperature programmed desorption and the value of SBET. Pyridine-DRIFT analysis (Figure S12, Supporting Information) suggested the absence of Brønsted acid
site; d)Determined by TEM analysis.
©
1801733 (7 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2018, 1801733

www.advancedsciencenews.com
www.small-journal.com
Figure 6. a) 27Al solid-state NMR spectra of C-Al2O3@2%Pt/SBA-15, N-Al2O3@2%Pt/SBA-15, and 1%Pt/SBA-15. Spinning sidebands were labeled by
asterisks. b) Schematic illustration of the effect of overcoat texture on the hydrodeoxygenation of 4-propylguaiacol.
coordinated Al sites were found in both overcoated catalysts.
This Al (V) site is considered electron-deficient thus leading to
a strong Lewis acidity,[53] which confirmed the pyridine-DRIFT
results. Although the ratio of Al(V) and Al(IV) may differ in the
two overcoated catalysts, we assumed and results confirmed
that the alumina overcoats synthesized from Al(sBuO)3–AlBr3
and Al(sBuO)3–EAA were chemically similar.
on the C-Al2O3@2%Pt/SBA-15, 4-propylguaicaol likely had less
interactions with the Pt-SiO2–Al2O3 interfaces due to the higher
porosity of the overcoat. Therefore, the reaction proceeded mostly
through molecules bound directly to the metal (Figure 6b). In
such cases, the dominant reaction is known to be ring hydro-
genation. These HDO results exemplify the significant control
that overcoat nanostructures can have on the selectivities of the
reactions catalyzed by the corresponding heterogeneous catalyst.
Many studies have suggested that the metal-support interface
plays an important role in HDO.[20,44] A recent investigation on
phenol HDO revealed a mechanism showing that direct deoxy-
genation can proceed by the first interaction between the phenol
tautomer and an surface oxophilic site, followed by the hydro-
genation of the carbonyl group and a final dehydration step to
form benzene.[48] The presence of electron-deficient Al(V) sites
in the alumina overcoat and aromatic products suggests that
4-propylguaiacol was converted in a similar direct deoxygenation
pathway over N-Al2O3@2%Pt/SBA-15. Because Al(V) sites were
present in both overcoated Pt/SBA-15, we attributed the different
selectivities to the distinct structural (rather than chemical) prop-
erties of two overcoats. As discussed in 3.3, the Al2O3 overcoat
prepared from chelation chemistry method is more porous. Such
physical differences of overlayers can also explain the fact that
the accessible Pt sites of C-Al2O3@2%Pt/SBA-15 outnumbered
those of N-Al2O3@2%Pt/SBA-15 by a factor of 3 after overcoating
despite their similar metal loadings (as determined by ICP-OES:
1.7% and 1.9%). Compared to C-Al2O3@2%Pt/SBA-15, the
denser overlayer on N-Al2O3@2%Pt/SBA-15 led to increased
intimate contact between Pt particles and Al2O3, which creates
more interfacial metal-support active sites. In summary, with
the more active overcoat, the 4-propyluaiacol was likely mainly
anchored on the abundant Pt-Al(V) sites of N-Al2O3@2%Pt/SBA-
15, which promoted initial deoxygenation/hydrogenolysis and
subsequent ring hydrogenation to form cyclohexane. In contrast,
3. Conclusions
We studied the formation of alumina overcoats over high sur-
face area supported catalysts by slowing down the kinetics
of alumina precursor condensation and developed two dis-
tinct approaches to achieve this kinetic control. In one case,
we used nonhydrolytic sol–gel chemistry and in the other we
used a chelation agent to manipulate the hydrolysis/condensa-
tion kinetics of Al(sBuO)3. Both methods were easy to conduct,
yet highly versatile, allowing us to deposit conformal alumina
overcoats onto various high surface area substrates with sub-
nanometer precision. However, both methods offered distinct
features that were advantageous in different cases. The non-
hydrolytic route produced a denser Al2O3 overcoat and could
drastically improve the catalyst’s selectivity during lignin hydro-
deoxygenation—a reaction for which, the metal–metal oxide
interface plays an important role. However, the nonhydrolytic
sol–gel method could not be used with a supported base metal
catalyst like Cu because it lead to metal leaching during the
coating procedure. In this case, the Al2O3 overcoat prepared
with chelation chemistry was the only one that could produce a
conformal overcoat and protect the metal nanoparticles against
sintering during liquid phase hydrogenation. Moreover, both
©
Small 2018, 1801733
1801733 (8 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
into the reaction flask containing catalyst suspension refluxing at 100 °C
and an automatic syringe pump (KDS 100 legacy syringe pump) was
used to control the injection rate. SiO2 spheres were dispersed in 10 mL
CH2Br2 and then overcoated by injecting a precursor solution comprising
0.08 mL Al(sBuO)3 and 0.17 g AlBr3. The injection rate was adjusted to
0.5 mL h −1, which correspond to one monolayer amount of precursor
per hour. The solution was left to react for 24 h after the injection was
completed. 2%Pt/SBA-15 was dispersed in 20 mL CH2Br2 and coated
by injecting a precursor solution containing 0.14 mL Al(sBuO)3 and
0.3 g AlBr3 at a rate of 1 mL h−1. The solution was left to react for 6 h
after the injection was completed. In both cases the following two
subsequent postsynthesis treatments followed: first, 20 mL of diethyl
ether were added into the suspension to quench highly reactive residual
Al–Br groups. The solution was then kept stirring at 80 °C overnight.
The crude products were centrifuged and sequentially washed twice
with diethyl ether, ethanol, and water. After washing, the samples were
stirred in 0.1 m NH3(aq) at 60 °C overnight. The products were washed
with water and dried in a vacuum oven at 60 °C. Subsequently, the
crude was reacted with 0.1 m NH3(aq), washed thoroughly and calcined
at 500 °C to completely remove the byproduct HBr. Finally, Overcoated
Pt/SBA-15 was reduced at 300 °C for 3 h prior to the catalytic reaction.
The resulting materials are referred to as N-Al2O3@2%Pt/SBA-15 and
N-Al2O3@SiO2 below.
Catalyst Overcoating—Chelation Chemistry-Based Approach: C-Al2O3@
SiO2: SiO2 spheres (0.5 g) were dispersed in a mixture containing 25 mL
ethanol, 0.04 mL of 25 wt% NH3(aq) and 1 mL H2O. The precursor
solution was prepared by stirring 0.13 mL Al(sBuO)3 and 0.05 mL EAA
in 9.8 mL of 2-butanol for 1 h. This precursor solution was then injected
into the mixture containing silica spheres using the aforementioned
syringe setup at a rate of 1 mL h−1. The suspension was stirred at room
temperature for 8 h after the injection.
C-Al2O3@Cu/Al2O3: Cu/Al2O3 was preoxidized at 250 °C for 2 h
by flowing synthetic air before overcoating. Then, 0.5 g of CuO/Al2O3
was dispersed in the same solution as described above. The precursor
solution was prepared by stirring 2 mL Al(sBuO)3 and 0.8 mL EAA in
27.2 mL 2-butanol for 1 h. This precursor solution was then injected
into the mixture containing the substrate at a rate of 1 mL h−1. The
suspension was heated to 40 °C and stirred for 6 h after the injection.
C-Al2O3@2%Pt/SBA-15: The uncoated catalyst (0.5 g of 2%Pt/
SBA-15) was dispersed in a mixture containing 25 mL ethanol, 0.02 mL
25 wt% NH3, and 1 mL H2O. The precursor solution was prepared by
stirring 0.44 mL Al(sBuO)3 and 0.165 mL EAA in 9.4 mL 2-butanol for
1 h. The precursor was then injected into the mixture containing the
substrate at a rate of 1 mL h−1. The solution was left to react at room
temperature and stirred for 8 h after the injection.
methods create stronger Lewis acid sites compared to γ-Al2O3
according to pyridine-DRIFT and SSNMR analyses. These dif-
fering characteristics provide an especially versatile toolbox for
renewable chemistry where stability and selectivity is especially
important in the presence of water in and numerous oxygen
functionalities in the reactants. At the same time, the high
degree of control over the overcoat thickness and texture could
allow the tailoring of the surface nanostructures to several
materials that are used in applications beyond heterogeneous
catalysis.
4. Experimental Section
Chemicals and Materials: All chemicals were analytical grade and
obtained from commercial suppliers. They were used without further
purification unless stated otherwise. Air and moisture-sensitive
reagents were handled using a nitrogen filled glove box and a standard
Schlenk line apparatus. TEOS, tetramethyl orthosilicate (TMOS), AlBr3,
CaH2, H2PtCl6 ·6H2O, furfural, and 1-butanol were obtained from
Acros. Benzophenone, magnesium, sodium, Pluronic 123 (P123),
3-propylguaiacol, 2,2,4-trimethylpentane (isooctane), CH2Br2, EAA,
2-butanol and HNO3(aq) (≈68%) were purchased from Sigma-Aldrich.
Molecular sieve (4 Å), Al(sBuO)3, and HCl(aq) (≈37%) were obtained from
Merck. Ethanol and diethyl ether were obtained from Fisher Scientific.
Silicon carbide (100 mesh) was obtained from Strem. Cu(NO3)2 ·3H2O
was obtained from ABCR. NH3(aq) (≈25%) was purchased from VWR
Chemicals. γ-Al2O3 (Pural SB) was obtained from Sasol. Diethyl
ether was dried over Na-benzophenone, distilled and stored over
molecular sieves. 2-butanol was dried over Mg, distilled, and stored
over molecular sieves. CH2Br2 was dried over CaH2, distilled, and stored
over molecular sieves. Furfural and EAA were purified by distillation under
reduced pressure. The water used in this study was purified by a Millipore
Milli-Q Advantage A10 water purification system resulting in a resistivity
higher than 18 MΩ cm. All gases were purchased from Carbagas.
Preparation of Supports and Catalysts: The silica spheres were
synthesized by the Stöber method.[54] Typically, 3.8 mL of TEOS, 7.7 mL
NH3 (≈25%), and 18 mL of deionized water were mixed with 120 mL
ethanol. The mixture was stirred and reacted at room temperature for
24 h. The product was centrifuged and washed three times first with
ethanol then water. SBA-15 was synthesized based on a previously
published method that was slightly modified.[55] Briefly, 2.8 g of Pluronic
P-123 was dissolved in 104 g of 1.6 m HCl(aq) by vigorous stirring. Then,
4 mL of TMOS was slowly added and the resulting mixture was stirred
for 24 h. The suspension was transferred into a Teflon lined stainless
steel autoclave and heated at 110 °C for 24 h. After the hydrothermal
treatment, the remaining surfactant was removed by ethanol using
a standard Soxhlet apparatus for 24 h. All synthetized supports were
calcined at 500 °C for 5 h under flow of synthetic air (ramping rate:
2 °C min−1) and dried under reduced pressure (below 10−2 bar) at 120 °C
prior to impregnation.
All the overcoated samples were washed three times with ethanol
followed by water and then dried in a vacuum oven at 60 °C. Both
overcoated catalysts (C-Al2O3@Cu/Al2O3 and C-Al2O3@Pt/SBA-15) were
further calcined at 400 °C for 3 h and reduced at 300 °C for 5 h. The
additional materials synthesis details and description of characterization
techniques are shown in Sections S1 and S2 in the Supporting
Information, respectively.
Catalytic Activity Tests—Hydrodeoxygenation of
a Model Lignin
Monomer: The catalyst was mixed with 10.5 mL isooctane and 0.1 mL
4-propylguaiacol in a 25 mL stainless-steel reactor (Parr). The reactions
were stirred at 600 rpm under 15 bar H2 at 200 °C. The actual pressure
after the temperature rising to 200 °C was ≈25 bar. The reaction
time and the amounts of reactants were varied to reach different
conversions. The temperature was ramped up with a heating plate and
heating tape controlled by an Omega Proportional-Integral-Derivative
controller. The product was analyzed by an Agilent Technologies 7890 A
gas chromatography apparatus equipped with a flame ionization
detector (GC-FID) and an HP-5 column (50 m, 0.32 mm). The products
were quantified using calibration curves obtained prepared with
authenticated standards except for 2-methoxy-4-propylcyclohexanol and
4-propylcyclohexanol. These two products were quantified using the
effective carbon number method with n-decane as an internal standard,
which has proven to be accurate for lignin-derived monomers.[40]
Cu/Al2O3 (5%) was prepared by incipient wetness impregnation
using Cu(NO3)2 ·3H2O dissolved in 0.1 m HNO3(aq) as the precursor
and calcined γ-Al2O3 as the support. The impregnated powder was
subsequently dried in an oven at 105 °C overnight and reduced at
300 °C for 5 h by flowing H2 (ramping rate: 1 °C min−1). 2%Pt/SBA-15
was prepared by incipient wetness impregnation using H2PtCl6 ·6H2O
dissolved in 1 m HCl(aq) as the precursor. The impregnated powder was
dried overnight at 105 °C, calcined at 400 °C for 3 h and then reduced at
300 °C for 3 h. Both 1% Pt/Al2O3 and 2% Pt/SBA-15 were prepared with
this procedure.
Catalyst Overcoating—NHSG-Based Approach: All catalysts (0.5 g)
were dried overnight under reduced pressure prior to overcoating.
The precursor was prepared by dissolving AlBr3 and Al(sBuO)3 in
20 mL CH2Br2. The precursor solution was loaded into a glass syringe
(Hamilton US) with a Teflon plunger. The syringe needle was positioned
©
1801733 (9 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2018, 1801733

www.advancedsciencenews.com
www.small-journal.com
External diffusion limitations were assumed to be negligible based on
experiments where the stirring speed was varied and internal diffusion
limitations were assumed to be negligible based on our calculation of
the Weisz–Prater criterion (see Section S4, Supporting Information).
Catalytic Activity Tests—Hydrogenation of Furfural: Furfural
hydrogenation was performed using a stainless steel fixed-bed down flow
tubular reactor with an inner diameter 4.5 mm, which was previously
described.[28] The catalytic bed was prepared by mixing the catalysts
with SiC and immobilized in the metal tube between two quartz wool
plugs. The remainder of the tube was filled with SiC and both ends were
plugged with quartz wool. The heated zone corresponding to the area
containing the catalytic bed was delimited by two external conductive
stainless-steel blocks enclosed in a furnace. Before the first run, the
catalyst was reduced under H2 flow (100 mL min−1) for 5 h at 300 °C.
After cooling down to the reaction temperature (130 °C), hydrogen flow
was adjusted to 35 mL min−1 using a Brooks mass flow controller and
pressure was set to 23 bar using a back pressure regulator (Tescom). A
furfural solution (70 g kg−1 in 1-butanol) was fed into the reactor using
SSI Series II high-performance liquid chromatography (HPLC) pump at
a rate of 0.10 mL min−1. Liquid samples were collected using a Jerguson
gage equipped with a needle valve and analyzed by GC-FID. The products
were quantified by the calibration curves obtained using standard
chemicals. Periodically, furfural flow was stopped and the catalyst was
regenerated by calcination at 400 °C for 1 h (under a flow of synthetic air
at 100 mL min−1) and reduction at 300 °C for 5 h (under a flow of H2 at
100 mL min−1). Catalytic activities were evaluated by the yield of furfural
alcohol per unit time and mol of Cu. The internal diffusion limitation
was assumed to be negligible based on our calculation of the Weisz–
Prater criterion (see Section S4, Supporting Information).
[1] Q. Zhang, I. Lee, J. B. Joo, F. Zaera, Y. Yin, Acc. Chem. Res. 2013, 46,
1816.
[2] F. Héroguel, B. Rozmysłowicz, J. S. Luterbacher, CHIMIA
Int. J. Chem. 2015, 69, 582.
[3] H. N. Pham, A. E. Anderson, R. L. Johnson, T. J. Schwartz,
B. J. O’Neill, P. Duan, K. Schmidt-Rohr, J. A. Dumesic, A. K. Datye,
ACS Catal. 2015, 5, 4546.
[4] B. J. O’Neill, D. H. K. Jackson, J. Lee, C. Canlas, P. C. Stair,
C. L. Marshall, J. W. Elam, T. F. Kuech, J. A. Dumesic, G. W. Huber,
ACS Catal. 2015, 5, 1804.
[5] H. Yang, Y. Chong, X. Li, H. Ge, W. Fan, J. Wang, J. Mater. Chem.
2012, 22, 9069.
[6] J. Lu, B. Fu, M. C. Kung, G. Xiao, J. W. Elam, H. H. Kung, P. C. Stair,
Science 2012, 335, 1205.
[7] S. H. Joo, J. Y. Park, C.-K. Tsung, Y. Yamada, P. Yang, G. A. Somorjai,
Nat. Mater. 2009, 8, 126.
[8] H. Zhu, Z. Ma, S. H. Overbury, S. Dai, Catal. Lett. 2007, 116, 128.
[9] S.-C. Shen, S. Kawi, J. Phys. Chem. B 1999, 103, 8870.
[10] N. Pham, A. E. Anderson, R. L. Johnson, K. Schmidt-Rohr,
A. K. Datye, Angew. Chem., Int. Ed. 2012, 51, 13163.
[11] X. She, J. H. Kwak, J. Sun, J. Hu, M. Y. Hu, C. Wang, C. H. F. Peden,
Y. Wang, ACS Catal. 2012, 2, 1020.
[12] A. C. Alba-Rubio, B. J. O’Neill, F. Shi, C. Akatay, C. Canlas, T. Li,
R. Winans, J. W. Elam, E. A. Stach, P. M. Voyles, J. A. Dumesic, ACS
Catal. 2014, 4, 1554.
[13] J. Lu, B. Liu, J. P. Greeley, Z. Feng, J. A. Libera, Y. Lei, M. J. Bedzyk,
P. C. Stair, J. W. Elam, Chem. Mater. 2012, 24, 2047.
[14] B. J. O’Neill, D. H. K. Jackson, A. J. Crisci, C. A. Farberow, F. Shi,
A. C. Alba-Rubio, J. Lu, P. J. Dietrich, X. Gu, C. L. Marshall,
P. C. Stair, J. W. Elam, J. T. Miller, F. H. Ribeiro, P. M. Voyles,
J. Greeley, M. Mavrikakis, S. L. Scott, T. F. Kuech, J. A. Dumesic,
Angew. Chem., Int. Ed. 2013, 52, 13808.
[15] J. Lu, J. W. Elam, P. C. Stair, Acc. Chem.Res. 2013, 46, 1806.
[16] M. Leskelä, M. Ritala, Angew. Chem., Int. Ed. 2003, 42, 5548.
[17] R. A. Caruso, M. Antonietti, Chem. Mater. 2001, 13, 3272.
[18] F. Babonneau, L. Coury, J. Livage, J. Non-Cryst. Solids 1990, 121,
153.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
[19] Z. Teng, G. Zheng, Y. Dou, W. Li, C.-Y. Mou, X. Zhang, A. M. Asiri,
D. Zhao, Angew. Chem., Int. Ed. 2012, 51, 2173.
[20] F. Héroguel, L. Silvioli, Y.-P. Du, J. S. Luterbacher, J. Catal. 2018,
358, 50.
[21] R. Nass, H. Schmidt, J. Non-Cryst. Solids 1990, 121, 329.
[22] A. Mitra, G. De, Langmuir 2014, 30, 15292.
[23] S. Acosta, R. J. P. Corriu, D. Leclercq, P. Lefèvre, P. H. Mutin,
A. Vioux, J. Non-Cryst. Solids 1994, 170, 234.
[24] P. Arnal, R. J. P. Corriu, D. Leclercq, P. H. Mutin, A. Vioux,
Chem. Mater. 1997, 9, 694.
[25] A. Vioux, Chem. Mater. 1997, 9, 2292.
[26] L. Bonhomme-Coury, F. Babonneau, J. Livage, J. Sol-Gel Sci. Technol.
1994, 3, 157.
This work was supported by the European Research Council (ERC)
under the European Union’s Horizon 2020 research and innovation
program (Starting grant: CATACOAT, No. 758653), the Swiss National
Science Foundation through grant PYAPP2_154281 and by EPFL.
This work was also accomplished within the framework of the Swiss
Competence Center for Bioenergy Research (SCCER-BIOSWEET). The
authors thank Benjamin Le Monnier for purifying ethyl acetoacetate and
aluminum sec-butoxide. The authors thank the EPFL interdisciplinary
center for electron microscopy for support during electron microscopy
measurements. The authors thank Emilie Baudat and Pierre Mettraux
for their help in performing measurements with ssNMR and XPS,
respectively. The authors thank EPFL’s central environmental laboratory
for the ICP-OES measurements.
[27] S. Kurajica, J. Popovi , T. G. Kraljevi , E. Tkalcˇec, I. Simcˇic´,
V. Mandic´, A. Altomare, A. Moliterni, X. Rocquefelte, J. Sol-Gel
Sci. Technol. 2014, 71, 217.
[28] F. Héroguel, B. P. Le Monnier, K. S. Brown, J. C. Siu,
J. S. Luterbacher, Appl. Catal., B 2017, 218, 643.
Conflict of Interest
The authors declare no conflict of interest.
[29] M. Baca, E. de la Rochefoucauld, E. Ambroise, J.-M. Krafft, R. Hajjar,
P. P. Man, X. Carrier, J. Blanchard, Microporous Mesoporous Mater.
2008, 110, 232.
[30] C. M. A. Parlett, L. J. Durndell, A. Machado, G. Cibin, D. W. Bruce,
N. S. Hondow, K. Wilson, A. F. Lee, Catal. Today 2014, 229, 46.
[31] P. Iengo, M. Di Serio, A. Sorrentino, V. Solinas, E. Santacesaria,
Appl. Catal,, A 1998, 167, 85.
Keywords
alumina, catalyst stability, heterogeneous catalysis, hydrodeoxygenation,
nanostructured catalysts
Received: May 4, 2018
Revised: June 21, 2018
Published online:
[32] C. P. Canlas, J. Lu, N. A. Ray, N. A. Grosso-Giordano,
S. Lee, J. W. Elam, R. E. Winans, R. P. Van Duyne, P. C. Stair,
J. M. Notestein, Nat. Chem. 2012, 4, 1030.
©
Small 2018, 1801733
1801733 (10 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.small-journal.com
[33] K. S. Brown, C. Saggese, B. P. Le Monnier, F. Héroguel,
J. S. Luterbacher, J. Phys. Chem. C 2018, 122, 6713.
[34] X. Liu, Q. Zhu, Y. Lang, K. Cao, S. Chu, B. Shan, R. Chen,
Angew. Chem. 2017, 129, 1670.
[45] H. Lee, H. Kim, M. J. Yu, C. H. Ko, J.-K. Jeon, J. Jae, S. H. Park,
S.-C. Jung, Y.-K. Park, Sci. Rep. 2016, 6, 28765.
[46] T. Nimmanwudipong, R. C. Runnebaum, D. E. Block, B. C. Gates,
Energy Fuels 2011, 25, 3417.
[35] J. Lee, D. H. K. Jackson, T. Li, R. E. Winans, J. A. Dumesic,
T. F. Kuech, G. W. Huber, Energy Environ. Sci. 2014, 7, 1657.
[36] E. D. Goodman, J. A. Schwalbe, M. Cargnello, ACS Catal. 2017, 7,
7156.
[37] J. S. Luterbacher, D. M. Alonso, J. A. Dumesic, Green Chem. 2014,
16, 4816.
[38] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044.
[39] Q. Lu, W.-Z. Li, X.-F. Zhu, Energy Convers. Manage. 2009, 50, 1376.
[40] L. Shuai, M. T. Amiri, Y. M. Questell-Santiago, F. Héroguel, Y. Li,
H. Kim, R. Meilan, C. Chapple, J. Ralph, J. S. Luterbacher, Science
2016, 354, 329.
[47] M. Hellinger, H. W. P. Carvalho, S. Baier, D. Wang, W. Kleist,
J.-D. Grunwaldt, Appl. Catal., A 2015, 490, 181.
[48] P. M. de Souza, R. C. Rabelo-Neto, L. E. P. Borges, G. Jacobs,
B. H. Davis, D. E. Resasco, F. B. Noronha, ACS Catal. 2017, 7,
2058.
[49] C. R. Lee, J. S. Yoon, Y.-W. Suh, J.-W. Choi, J.-M. Ha, D. J. Suh,
Y.-K. Park, Catal. Commun. 2012, 17, 54.
[50] M. V. Bykova, D. Y. Ermakov, V. V. Kaichev, O. A. Bulavchenko,
A. A. Saraev, M. Y. Lebedev, V. A. Yakovlev, Appl. Catal. B 2012,
113–114, 296.
[51] J. Horácˇek, G. Št’ávová, V. Kelbichová, D. Kubicˇka, Catal. Today
2013, 204, 38.
[52] M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat, J. Catal.
2002, 211, 1.
[53] F. R. Chen, J. G. Davis, J. J. Fripiat, J. Catal. 1992, 133,
263.
[54] W. Stöber, A. Fink, E. Bohn, J. Colloid Interface Sci. 1968, 26, 62.
[55] A. Galarneau, H. Cambon, F. Di Renzo, F. Fajula, Langmuir 2001,
17, 8328.
[41] W. Lan, M. T. Amiri, C. M. Hunston, J. S. Luterbacher,
Angew. Chem. Int. Ed. 2018, 57, 1356.
[42] I. Klein, B. Saha, M. M. Abu-Omar, Catal. Sci. Technol. 2015, 5, 3242.
[43] S.-F. Koelewijn, S. V. den Bosch, T. Renders, W. Schutyser,
B. Lagrain, M. Smet, J. Thomas, W. Dehaen, P. V. Puyvelde,
H. Witters, B. F. Sels, Green Chem. 2017, 19, 2561.
[44] A. M. Robinson, J. E. Hensley, J. W. Medlin, ACS Catal. 2016, 6,
5026.
©
1801733 (11 of 11)
2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Small 2018, 1801733