Unraveling the role of low coordination sites in a cu metal nanoparticle: A step toward the selective synthesis of second generation biofuels

Article abstract of DOI:10.1021/cs500581a

The acidity of a prereduced Cu/SiO₂ catalyst was extensively investigated by means of FT-IR of adsorbed pyridine and by titration with 2-phenylethylamine in cyclohexane. Comparison with the parent CuO/SiO ₂ material, which was alread

Full text of DOI:10.1021/cs500581a

Research Article
pubs.acs.org/acscatalysis
Unraveling the Role of Low Coordination Sites in a Cu Metal
Nanoparticle: A Step toward the Selective Synthesis of Second
Generation Biofuels
Nicola Scotti,†,§ Milind Dangate, Antonella Gervasini, Claudio Evangelisti, Nicoletta Ravasio,*
†,§
‡
†
,†
and Federica Zaccheria†
†
‡
CNR, Institute of Molecular Science and Technologies (CNR-ISTM) and Dipartimento di Chimica, Universita
9, 20133 Milano, Italy
̀
di Milano, Via Golgi
1
*
S Supporting Information
ABSTRACT: The acidity of a prereduced Cu/SiO catalyst was extensively
2
investigated by means of FT-IR of adsorbed pyridine and by titration with 2-
phenylethylamine in cyclohexane. Comparison with the parent CuO/SiO2
material, which was already shown to exhibit Lewis acid sites due to the high
dispersion of the CuO phase, provided evidence that reduction of this phase to
the metallic state increases the acidity of the material. This allowed us to set up
a bifunctional catalyst showing acidic and hydrogenation activity, both
ascribable to the presence of the metal particle, without the need of an acidic support. This catalyst was tested in the one-pot
transformation of γ−valerolactone into pentyl valerate and showed comparable activity (91% vs 92% conversion) and improved
selectivity (92% vs 72%) with respect to the previously reported copper catalyst supported on acidic material. The role of Cu in
activating the substrate was also evidenced through FTIR of adsorbed γ-valerolactone.
KEYWORDS: GVL, pentyl valerate, valeric fuels, supported Cu catalysts, solid Lewis acids
INTRODUCTION
obtained from GVL, valeric esters showed excellent features as
gasoline or diesel fuel depending on the alcoholic residue.
■
26
In the search for renewable energy, we have several potential
sources at hand: water, sunlight, wind, geothermal heat, and
tidal flow can all be used to generate power. On the contrary,
the search for renewable transportation fuels has to rely on only
one source: biomass.
Among the different raw materials available from biomass,
vegetable oils are very suitable for the synthesis of diesel
components such as fatty acids methyl esters (biodiesel) and
hydrocarbons obtained through hydrocracking of vegetable oils
Some of us recently reported a process for the production of
ethyl and pentyl valerate directly from GVL and the proper
alcohol by using an heterogeneous copper catalyst supported
27
on silica-zirconia. The reaction with ethanol allowed us to
elucidate the reaction mechanism and to test the stability and
reusability of the catalyst, whereas using pentanol, GVL
conversions higher than 90% and selectivities to pentyl valerate
up to 83% were achieved. The reaction takes place under H2
through nucleophilic addition of the alcohol to the carboxylic
group giving hydroxypentanoate 1, followed by dehydration to
pentenoate 2 and hydrogenation to pentyl valerate 3 (Scheme
1
and fats (green diesel or renewable diesel), but the most
abundant biomass is cellulose, and for this reason its
transformation in platform molecules has received increased
attention in the past decade. These platform molecules were
first pinpointed by the U.S. Department of Energy in 2004 in a
report identifying 15 compounds as those having high potential
for application in future biorefineries, which was revised in 2010
to include more recent advances in biomass conversion
1
).
The acid-catalyzed reactions were ascribed to the acidic
support, whereas hydrogenation activity was linked to the
presence of supported copper.
However, it is well-recognized that the type and strength of
2
technologies. Among the compounds listed in the 2010
acidic sites are very important in designing solid acid catalysts
review, levulinic acid can be considered one of the most
important due to its reactive nature. Moreover, it can be
produced from lignocellulosic wastes at low cost, for example,
through the Biofine process that can be fed with agricultural
28,29
for renewable raw material transformations,
solid Lewis
acids emerging as a very valuable alternative to Brønsted ones
such as zeolites and sulfonated materials. The role of both the
support and the catalyst acidity in the particular case of levulinic
acid transformation to pentenoic acids or to valeric esters in the
3
residues, papermill sludge, or urban waste paper.
Levulinic acid or its esters can be hydrogenated to furan
4
,5
6−14
derivatives or to γ-valerolactone (GVL)
that in turn has
been identified as a platform molecule with many applications
Received: April 30, 2014
Revised: July 15, 2014
Published: July 15, 2014
15
16−18
15
as a precursor for chemicals, as solvent,
fuel additive,
19−25
and fuels precursor.
Among the molecules that can be
©
2014 American Chemical Society
2818
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
Scheme 1. Reaction Mechanism for the Conversion of GVL
into Pentyl Valerate
and dilution with highly deionized water (Milli-Q Academic,
Millipore) to a final weight of 100 g.
2.2.2. N2 Physisorption Measurements. Values of the
sample surface area and porosity were determined (Table 1)
by collecting N adsorption/desorption isotherms at −196 °C
2
using a Sorptomatic 1900 series instrument purchased from CE
Instruments Ltd. Prior to the analysis, the sample (ca. 0.1 g)
was outgassed at 150 °C for 2 h. Specific surface area was
calculated by the BET equation in the relative pressure interval
of P/P = 0.033−0.35, assuming the N molecular area in the
0
2
2
3
−1
adsorbed state as 16.2 Å . The total pore volume (cm ·g ,
STP) was estimated from the total uptake of N adsorbed at a
2
relative pressure P/P = 0.99, converted into liquid volume
0
3
−1
(
0
cm ·g ) using the N density in the normal liquid state (ρ =
2
−
3
.8081 g cm ). Pore size distribution was obtained from the
desorption branch of the collected isotherm using the Barrett−
37
Joyner−Halenda (BJH) model equation.
.2.3. Scanning Electron Micrographs. Scanning electron
2
micrographs (SEM) were obtained by a JEOL JSM-5500LV
coupled with energy dispersive X-ray spectroscopic (EDS)
analyzer working at 20 keV to obtain quantitative information
on the distribution of Cu and Si elements. The samples were
analyzed under moderate vacuum after coating.
2.2.4. HRTEM Analysis. High-resolution transmission
electron microscopy (HRTEM) analysis of copper catalyst
was operated at 200 kV with a LIBRA 200FE analytical
transmission electron microscope, equipped with FEG source
and purchased from Zeiss. Samples were deposited on holey-
carbon-coated grids from alcohol suspensions. Samples, in the
form of powders, were ultrasonically dispersed in isopropyl
alcohol, and a drop of the suspension was deposited on a holey-
carbon film grid (300 mesh). Histograms of the particle size
distribution (Figure S11) were obtained by counting onto the
micrographs at least 500 particles; the mean particle diameter
presence of Ru supported on zeolitic or amorphous systems
was put in evidence by Weckhuisen et al. and by Fu et al.
14
30
We have long been investigating the relationships between
metal oxide dispersion and acidity in supported catalysts and
31−33
recently reported on a new concept of solid Lewis catalyst.
Thus, we could show a relationship between the high
dispersion of a non acidic oxide such as CuO and the acidic
activity of a supported CuO/SiO catalyst in epoxide ring
2
opening, cellulose deconstruction, and Friedel Craft acyla-
3
4−36
tion.
Therefore, we were interested in investigating the
acidity of the same system in the reduced state. If the acidity of
the reduced Cu/SiO catalyst was preserved, this would result
2
in a true bifunctional catalyst where both acidic and reduction
activity are carried out by the same site.
This may allow us to tune the activity and/or selectivity in
the one-pot transformation of GVL into pentyl valerate by
using a more simple catalyst.
Here we wish to report a detailed analysis of Cu/SiO2
catalysts and catalytic tests showing that a more selective
reaction can be obtained by using this bifunctional system
where both the acidity and the hydrogenation activity can be
ascribed to the metallic phase.
(d
) was calculated by using the formula d
was the number of particles of diameter d .
i
= Σd n
i
/Σn where
m
m
i
i
n
i
2.2.5. Surface Composition of the Catalysts  Cuboocta-
hedron Model. Assuming that the catalyst nanoparticles are
cubooctahedral in shape with a face-centered cubic (fcc)
structure, it is possible to calculate the number of different sites
3
8,39
depending on the size of the nanocrystals.
The total
number of atoms of a copper nanoparticle for a fcc crystal (as is
1
/3
EXPERIMENTAL SECTION
.1. Catalyst Preparation. Silica materials used as
the case of copper) can be calculated from d = 1.105d N
T
,
np
at
■
in which dnp is the average diameter of nanoparticles obtained
2
2
by HRTEM and d is the atom diameter of copper (2.56 Å).
supports, SiO A (SSA = 413 m /g, PV = 0.75 mL/g) and
at
2
2
The formula does not depend on particle shape.
Equation 1 allows one to calculate m, the length of an edge
expressed as an atom.
SiO B (SSA = 693 m /g, PV = 0.62 mL/g), were obtained
2
from Sigma-Aldrich and Sumitomo Chemicals, respectively.
Catalysts were prepared by the chemisorption−hydrolysis
3
3
(
[
CH) method by adding the support to an aqueous
3
2
m−3
N = (10m − 15m + 11 )/3
(1)
2+
T
Cu(NH ) ] solution prepared by dropping NH OH to a
3
4
4
Cu(NO ) ·3H O solution until pH 9 had been reached. After
All of the following, including N (bulk atoms), N (surface
B S
3
2
2
2
0 min under stirring, the slurry, held in an ice bath at 0 °C,
was diluted with water. The solid was separated by filtration
with a Buchner funnel, washed with water, dried overnight at
20 °C, and calcined in air at 350 °C. Two series of catalysts
atoms), N (high coordination atoms), and N (low
HS LS
coordination atoms), are given from eqs 2−5:
̈
3
2
N = 16m − 63m + 84m − 38
(2)
(3)
(4)
(5)
B
1
were obtained with a nominal copper loading of 8% and 16%,
respectively.
2
N = 10m − 20m + 12
S
2
.2. Catalyst Characterization. 2.2.1. Inductively
2
NHS = 6(m − 2) + 4(m − 3)(m − 2)
Coupled Plasma (ICP) Analysis. Copper metal loadings were
determined by ICP-OES (ICAP6300 Duo purchased from
Thermo Fisher Scientific) and an external calibration method-
ology, after microwave digestion of the samples (ca. 10 mg of
catalyst, and therefore 0.8−1.5 mg of Cu) in 3 mL of aqua regia
NLS = 24(m − 2) + 12
2.2.6. FT-IR Spectra of Adsorbed Probe Molecules (Pyridine
and GVL). The FT-IR studies of probe molecules (pyridine and
2
819
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
a
Table 1. Morphological Data of the Samples Obtained from N Adsorption−Desorption
2
surface area (m ·g−1)
2
PV (mL/g)
average pore diameter (Å)
b
material
Cu content (%)
SiO A
412.8
384.7
350.9
692.9
228.6
318.6
0.75
0.77
0.52
0.62
0.43
0.42
72
80
58
36
75
53
2
8% CuO/SiO A
9.4
2
16% CuO/SiO A
15.4
2
SiO B
2
8% CuO/SiO B
9.1
2
16% CuO/SiO B
12.1
2
a
b
See Figure S1. Determined by the Gurvitsch rule.
GVL) adsorption and desorption were carried out with a FTS-
0 spectrophotometer equipped with mid-IR MCT detector
purchased from BioRad. The experiments were performed on
was prereduced at 270 °C (20 min in air + 20 min under
6
vacuum + reduction under H ). The reaction mixture was
2
analyzed by GC-MS (GC-MS chromatography, 5%-phenyl-
methyl polysiloxane column) with the following temperature
programm (60 °C for 3 min; 2,5 °C/min until 90 °C, 15 °C/
min until 280 °C, 12 min at 280 °C) and by verifying the linear
response in a concentration range of 0.01−0.1 M.
sample disk (15−20 mg) after a simple calcination treatment
(
270 °C, 20 min air + 20 min under vacuum) for SiO and
2
CuO/SiO samples or after a reduction pretreatment (270 °C,
2
2
0 min air + 20 min under vacuum + reduction under H ) for
2
Cu/SiO samples. One spectrum was collected before probe
2
RESULTS AND DISCUSSION
The textural features of the two series of materials show a
different behavior (See Table 1 and Figure 1). On silica A, we
molecule adsorption as a blank experiment. Therefore, pyridine
■
(Py) or GVL adsorption was carried out at room temperature,
and the following desorption steps were performed from room
temperature to 250 °C. The spectrum of each desorption step
was acquired every 50 °C after cooling the sample. IR spectra of
adsorbed GVL were deconvoluted, by a peak fitting procedure,
using GRAMS/AI software (Thermo Scientific Inc., U.S.A.).
2.2.7. Acidity Determination. Acidity measurements of the
supports and catalysts have been performed in a home-
assembled liquid chromatographic (HPLC) line, consisting of a
L-6200A pump, an AS-2000A auto sampler injector, and a L-
4
250 UV−vis detector all purchased from Merck-Hitachi
coupled to a personal computer for the collection, storage,
40
and processing of the data. The analyses have been performed
by using 2-phenylethylamine (PEA) as probe molecule; PEA
was dissolved in cyclohexane (for the intrinsic acidity) and in 1-
pentanol (for the effective acidity) working at constant
temperature of 30 ± 1 °C. The sample (ca. 0.050 g, 80−200
mesh particles) was first activated under air flowing (the
calcined samples at 150 °C for 2 h) in a stainless steel tube (i.d.
2
mm, length 12 cm). After the transfer of the tube into the
adsorption line, successive injections of PEA solution (50 μL,
.15 M in cyclohexane or 1-pentanol) were put into the line in
0
which the basic solution continuously circulated through the
sample holder, until adsorption equilibrium was achieved.
During the first run, the injections of PEA solution (50 μL)
were done until surface saturation; then pure liquid was flowed
(
250 mL) through the sample to remove the physisorbed PEA,
and the second run adsorption was collected. Subtraction of the
second run from the first gives the chemisorption isotherm,
from which the strong acid sites can be computed.
Figure 1. Pore volume distribution as a function of pore radius for
support (SiO B) and Cu catalysts (see Figure S2 for SiO A).
2
2
2
.2.8. Temperature-Programmed Reductions. Temper-
ature-programmed reduction (TPR) analysis was performed
with a modified version of the Pulse Chemisorb 2700 apparatus
from Micromeritics. Catalysts (samples containing ca. 2 mg of
observed a small decrease of the surface area going from the
support to the 8% Cu catalyst and to the 16% Cu catalyst,
whereas on silica B, we observed a dramatic decrease in the
surface area values going from the support to the 8% Cu
catalyst and then an increase going to the 16% Cu material.
This second behavior can be explained by a very high
dispersion on the support of the CuO phase in the 8% sample,
covering two-thirds of the surface and filling the smallest pores,
as shown by a decrease in pore volume and an increase in pore
radius. When the Cu loading was increased to 16%, the CuO
Cu) were diluted with quartz, calcined at 500 °C under O (40
2
mL/min), and then reduced at 8 °C/min under a flow (15 mL/
min) of a 8% H /Ar mixture. The H consumption was
2
2
detected by a thermal conductivity detector (TCD).
.3. General Procedure for the Synthesis of Alkyl
2
Valerates. In a typical run, the procedure which we followed is
as follows: GVL (99% from Sigma-Aldrich, 4 g, 40 mmol), 1-
pentanol (200 mmol) and catalyst (400 mg) were loaded in a
Parr reactor under 10 atm H at 250 °C for 10 h. The catalyst
2
2
820
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
particles grow outside the pores developing a new surface and
causing the increase in surface area.
higher surface area in order to further increase the metallic
phase dispersion.
The behavior discussed above concerning the material
microstructure is also consistent with the SEM images collected
on the silica support and the copper catalysts. In particular,
Figure S3 in the Supporting Information section shows the
comparison between the smooth polyhedral grains of silica B
and the wrinkled surface of the relevant highest copper catalyst,
whose silica grains are covered by a new phase grown on it.
In spite of this evidence, TPR profiles of all the catalytic
systems analyzed show a single and sharp reduction peak with
the maximum centered around 250 °C (Figure 2). These data
demonstrate that although the metal loading is quite high, the
very high dispersion of the copper oxide phase allows its easy
reduction to the metallic state.
In order to directly compare the acidity of supported Cu
systems, we used Py adsorption spectroscopy. Pyridine is a
common basic probe molecule used to investigate the surface
acidity of solid materials and gives also the chance to
distinguish between Lewis and Brønsted acid sites.
FT-IR spectra of adsorbed pyridine are known to show many
−1
peaks ranging from 1400 to 1700 cm . Pyridine bound to
Lewis acid sites (L) is usually assigned to a first band around
1450 cm− and a second around 1610 cm , whereas an
1
−1
−1
−1
absorption at 1550 cm followed by a peak near 1640 cm is
related to the presence of Brønsted acid sites (Br). On the
other hand, the band at around 1490 cm− is not clearly
assigned and should be the result of a combination of pyridine
1
lying on both Lewis and Brønsted acid sites, whereas a band at
1
1
620 cm− was identified as the result of either kind of
interactions. Finally, a weak interaction between the pyridine
and the solid surface (physisorption or hydrogen bonding, P),
resulting from very weak or no acidity, accounts for a two bands
−
1
−1
in the region of 1440−1450 cm and at 1580−1600 cm ,
which almost completely disappear by outgassing at 100
3
5,42−45
°
C.
The analyzed catalysts, both reduced and unreduced ones,
show the presence of the bands ascribable to Lewis acid sites at
1
1
450 and 1610 cm− that are resistant to the outgassing at high
temperatures, at least up to 250 °C, whereas physisorbed
pyridine completely disappears at 100 °C (Figure 3, S5−S10).
On the other hand, spectra of both silica supports show only
the two bands of physisorbed pyridine.
From the spectra here reported, it is evident that the reduced
samples exposing metallic copper on the surface not only keep
the properties of the precursor CuO phase (the same bands in
the spectra are even more evident) but also show an even
increased Lewis acid character with respect to the unreduced
materials.
Thus, the amount of adsorbed pyridine (mmolPy/gcat),
calculated on the basis of the relationship reported by
46
−1
Emeis from the integration of the band around 1450 cm ,
clearly show this trend, both in the case of Cu/SiO A and of
2
Cu/SiO B (see Table 2). This amplification results to be
2
almost 5 times in the case of Cu/SiO B versus CuO/SiO B
2
2
(
0.342 mmol /g vs 0.073 mmol /g ).
Py cat
Py cat
Moreover, in the spectra of adsorbed pyridine on reduced
Figure 2. TPR reduction profiles for CuO/SiO A (8% and 16% Cu)
2
−1
copper catalysts, also a significant band at 1575 cm and a
and CuO/SiO B (8% and 16% Cu).
2
−1
shoulder at 1441 cm , respectively, are visible. These two
bands are ascribable to the 8b and 19b modes of adsorbed
4
7
Some of us recently reported that copper catalysts supported
on non acidic silica show acidic properties because of the high
dispersion obtained with the preparation method, namely, the
pyridine corresponding to the vibrational modes as those of
48
benzene ring. The presence of these bands is therefore even
more diagnostic of the Lewis-type interaction of these catalytic
systems with pyridine (Figure 3).
3
3
CH one. Thus, pyridine desorption spectra of the fresh
catalyst CuO/SiO prepared by CH were compared with the
To have a deeper insight into the acidic properties of Cu/
2
spectrum of the bare SiO2 support, and this comparison
provides evidence for the presence of Lewis acid sites on the
CuO containing material.
SiO B, liquid-phase acid−base titrations by a strong base
2
probe, 2-phenylethylamine (PEA) was also performed. For the
titrations, cyclohexane was chosen to dissolve PEA. As a result
of the apolar aprotic character of cyclohexane, the amount of
the acid sites determined in this liquid corresponds to the
On the other hand, the very great advantage offered by a
dispersed metal-containing phase would be the opportunity to
also exploit other properties of the metal such as hydrogenation
4
0,45
intrinsic acidity of the samples.
4
1
activity, providing that the reduced phase preserve the acidic
properties. In the present paper, we therefore compared the
prereduced Cu/SiO material with its parent CuO/SiO in
Figure 4 shows the PEA adsorption isotherms of silica B and
of the two copper catalysts at different loadings collected in
cyclohexane. All the isotherms, both those related to the first
adsorption run and those of the second adsorption run, have a
more or less Langmuirian shape. The computed chemisorption
2
2
order to investigate its character. Moreover, we moved to
another silica support with similar pore volume and much
2
821
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
Figure 3. IR spectra of adsorbed pyridine on different catalysts. (A) Comparison of bare silicas (SiO A and SiO B), unreduced copper supported
2
2
catalysts (16% CuO/SiO A and 16% CuO/SiO B), and reduced ones (16% Cu/SiO A and 16% Cu/SiO B) after adsorption of Py and degassing
2
2
2
2
at 150 °C. (B) Py adsorption patterns at different desorption temperature for 16% Cu/SiO B.
2
Table 2. Amount of Adsorbed Pyridine (mmol /g ) on the
Results obtained from the titrations have been collected in
Table 3. It can be observed that silica shows the highest amount
of total acid sites and of strong acid sites (expressed in terms of
Py cat
Different Catalysts (Reduced and Not Reduced) Evaluated
−
1
from the Band at 1450 cm
equivalent of sites per unit mass, m /g) among the samples
because of its higher surface area. The percentage of strong
acidic sites increases from silica (52%) to the 8% Cu sample
eq
pyridine on L sites (mmol/g)
sample
150 °C
200 °C
250 °C
1
1
1
1
8
6% CuO/SiO A
0.096
0.156
0.073
0.342
0.236
0.067
0.121
0.049
0.240
0.155
0.047
0.089
0.036
0.159
0.101
2
(
60%) and the 16% Cu sample (71%); then, concerning the
6% Cu/SiO2A
acid strength, the following ranking can be observed: 16% Cu/
SiO B > 8% Cu/SiO B > SiO B (Table 3). This clearly shows
6% CuO/SiO B
2
2
2
2
6% Cu/SiO B
2
the change in the nature of acidic sites going from silica to the
Cu catalysts, which are also associated with Lewis Cu centers.
% Cu/SiO B
2
The Lewis acidity expressed by CuO/SiO is quite unusual as
2
CuO in its bulk form shows, on the contrary, a basic behavior.
This character was previously ascribed to the surface defectivity
of the highly dispersed CuO phase, shown to be in distorted
isotherms, related to the strong acid sites of the samples, have a
more pronounced Langmuirian shape, as expected.
2
822
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
higher concentration of Lewis sites. To get a deeper insight into
the dispersion and morphology of the Cu metallic phase high
resolution micrographs of the 16% Cu/SiO B sample were
2
recorded. The lattice-fringe image/FFT of a representative Cu
particle on the surface (Figure 5B,C) shows spots at 2.1 and 1.8
Figure 5. (A) HRTEM micrographs of 16% Cu/SiO B catalyst; (B)
2
magnified view of the Cu nanoparticle squared in A; (C) inverted FFT
pattern taken from image B.
Å, corresponding to Cu (111) and Cu (200) lattice planes of
metallic copper, respectively. High resolution micrographs
together with the observed spots in the FT pattern are in good
agreement with the presence of cuboctahedron copper particles
50,51
in a face cubic centered (fcc) crystal structure.
Moreover,
transmission electron micrographs (Figure 5A) show well-
dispersed spherical Cu nanoparticles over the SiO support
2
with particle size centered at about 4.7 nm (Figure S11).
Therefore, HRTEM analysis clearly shows the presence of well
dispersed metallic phase with well formed, highly defective
metal nanoparticle.
These findings prompted us to test the activity of these
materials in the one-pot conversion of GVL into pentyl valerate
(
PV) (Scheme 1). Previous results obtained in the same
transformation with Cu/SiO −ZrO clearly showed the
importance of both acidic and hydrogenation activity for this
transformation. The use of an acidic mixed oxide such as silica
2
2
Figure 4. First (blue full circle line) and second (red full square line)
runs of PEA adsorption isotherms collected at 30 °C in cyclohexane
27
zirconia as the support granted the presence of acidic sites,
whereas the well-dispersed, reduced copper phase allowed the
reduction of alkyl pentenoates into alkyl valerates.
on SiO B and relevant Cu catalysts; the chemisorption isotherm
2
(
black solid line) has been computed from subtraction of the second
from the first isotherm.
We therefore found it compelling to investigate whether the
catalyst presenting the higher density of Lewis sites, namely,
Table 3. Results of the Intrinsic Acidity Determination
Collected in Cyclohexane
1
6% Cu/SiO B, was able to behave as a true polyfunctional
2
catalyst, where nucleophilic addition to the carbonyl group,
dehydration of the intermediate alcohol 1, and hydrogenation
of pentenoate 2 are all promoted by the metallic phase.
Table 4 reports turnover frequencies calculated after a 1 h
reaction for 8% Cu/SiO −ZrO and 16% Cu/SiO catalysts
with respect to total surface atoms, high coordination sites, and
low coordination sites according to a cuboctrahedron model, as
well as Lewis sites as evaluated from pyridine adsorption
spectra. The ratio between the two values calculated for low
coordination sites and for Lewis sites are close to 1, clearly
showing that Lewis behavior is linked with low coordination
sites on the metal particle.
code
intrinsic acid sites
strong
total
mmol·g−
1
2
2
2
SiO B
1.236
0.627
0.561
0.641 (52%)
0.376 (60%)
0.400 (71%)
2
8
% Cu/SiO B
2
1
6% Cu/SiO B
2
octahedral configuration by means of EXAFS and DRUV
33
spectroscopic analysis, that is a peculiarity deriving from the
CH preparation technique. The reduction of copper oxide into
Cu(0) enhances even more this defectivity, the cuboctahedron
or truncated octahedron being the best approximations for the
Further insights into the role of copper can be drawn by
analyzing the IR spectra of adsorbed GVL (see Figure 6). The
3
9,49
geometry of a metallic nanoparticle
thus resulting in a
spectrum of GVL adsorbed onto SiO B shows two bands, one
2
2
823
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
Table 4. Turnover Frequencies Evaluated after a 1 h reaction and Calculated on Both Surface Atoms and Acidic Sites
TOF (h−1)
c
TOF (h−1
d
)
a
b
D
(nm)
4.7
3.7
L
(mmol /g )
conv (%)
NS
NHS
NLS
L
Py cat
1
8
6% Cu/SiO B
0.159
0.183
49
54
104.6
125.3
133.3
172.3
486.3
459.5
308.2
2
% Cu/SiO −ZrO
295.4
2
2
a
b
c
Mean particle diameter evaluated by HRTEM. Lewis acid sites density evaluated by adsorbed pyridine titration after degassing at 250 °C. TOF
39 d
(
(
after 1 h reaction) evaluated by using the surface atoms composition according to the cuboctahedral shape model reported by Kaneda.
after 1 h reaction) evaluated on Lewis acidic sites L.
TOF
is worth noting that the interaction with silica catalysts is
generally more labile with degassing at increasing temperature
with respect to the one with silica zirconia.
Scheme 2. Schematic Representation of the GVL Ring
Opening Reaction Mechanism Activated by the Cu/SiO2
Catalyst
The presence of a unique catalytic site and in particular the
absence of catalytically relevant Brønsted acid sites on the
support may in fact have a positive effect on product selectivity.
Results obtained in the catalytic tests for longer reaction times
are reported in Table 5.
Table 5. Reaction of GVL with 1-Pentanol To Form Pentyl
Valerate
Figure 6. GVL adsorption IR spectra after outgassing at 150 and 100
C for SiO2 B, Cu/SiO2 B, and SiO −ZrO2 compared with the
a
°
2
spectrum of pure GVL.
conv
(%)
Sel PV Sel 1
Sel 2
(%)
Sel 4
(%)
entry
catalyst
no catalyst
(%)
(%)
52
1
2
3
4
5
6
7
8
30
7
70
9
12
57
40
16
4
corresponding to the physisorbed molecule centered at 1759
−
1
−1
SiO B
14
51
2
cm and one at 1726 cm , ascribable to a weak chemical
interaction with GVL. This band is indeed still evident,
although very weak, also after degassing at 150 °C, whereas the
SiO −ZrO2
58
92
81
75
91
83
1
2
8% Cu/SiO −ZrO2
72
70
77
92
65
2
−
1
16% Cu/SiO A
10
11
2
band at 1759 cm
completely disappears under these
8% Cu/SiO B
5
2
conditions. When Cu is present on the same SiO , a shoulder
2
−1
16% Cu/SiO
2
B
7
lying at wavenumber lower than 1700 cm is also present
witnessing a strong chemisorption of GVL CO double bond
with the catalyst.
b
16% CuO/SiO B
8
3
2
a
Reaction conditions: 0.1 wt % catalyst, 250 °C, 10 bar H , 690 rpm,
2
b
1
0h, GVL/pentanol = 1:5. Unreduced catalyst.
Indeed, deconvolution of the spectrum shows that three
bands are present in the case of GVL adsorbed on SiO (1759,
2
−1
1
726, and 1705 cm ), whereas on Cu/SiO a new band,
centered at 1683 cm , is clearly detectable.
These data agree well with pyridine characterization, the
catalyst showing the highest number of Lewis acid sites,
2
,
−1
It is interesting to note that this same signal at 1683 cm− is
1
namely, the reduced 16% Cu/SiO B being the most selective
2
present also in the bare SiO −ZrO spectrum (with a very small
material. On the other hand, the unreduced 16% CuO/SiO B,
2
2
2
−1
shift to 1685 cm ), showing a very similar behavior of the two
materials, at least with respect to GVL.
the reduced 8% Cu/SiO B, and catalysts prepared on SiO A
2
2
are less active.
These spectra strongly confirm the hypothesis of the relevant
role of reduced copper not only in the hydrogenation reaction
but also in the activation of the substrate toward nucleophilic
addition by virtue of its Lewis acidic properties (Scheme 2). It
The use of naked silica as a catalyst, even with very low
conversion, gave selectively product 4 coming from dehy-
dration between the intermediate hydroxyl compound 1 and
the solvent alcohol, which is also favored on the naked silica
2
824
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
zirconia catalyst (entry 3), suggesting a determining role of
Bronsted-like sites in this pathway.
AUTHOR INFORMATION
Corresponding Author
E-mail: n.ravasio@istm.cnr.it. Fax: +39 0250314405. Tel.: +39
0250314382.
Author Contributions
N.S. and M.D. contributed equally.
Notes
The authors declare no competing financial interest.
■
When Cu is deposited on the surface the decrease in this
kind of acidic sites, witnessed by intrinsic acidity measurements
carried out with PEA, results in a very low selectivity toward the
ether in favor of the olefinic compound 2, which is then
hydrogenated to pentyl valerate (PV) and never accumulates in
the reaction mixture. The high selectivity observed (entry 7) is
therefore due to both the contribution to acidity of the metallic
phase and its high hydrogenation activity, thereby driving the
reaction toward the upper reaction pathway (Scheme 1). This
effect is even more pronounced in the case of copper deposited
*
§
ACKNOWLEDGMENTS
■
The authors acknowledge financial support by the Italian
Ministry for Education, University and Scientific Research
(MIUR) through the Project PRIN 2010A2FSS9_001.
on SiO with respect to Cu/SiO −ZrO , because residual
2
2
2
Bronsted-like sites on the surface of silica-zirconia make the
lower pathway still possible.
REFERENCES
■
This is in agreement with the very low number of acid sites
observed when titrating with PEA dissolved in 1-pentanol (i.e.,
the actual reaction solvent). The acidity determined in 1-
pentanol corresponds to the effective acidity shown by the
(
1) Perego, C.; Ricci, M. Catal. Sci. Technol. 2012, 2, 1776−1786.
(2) Bozell, J. J. Green Chem. 2010, 12, 525−728.
(3) Climent, M. J.; Corma, A.; Iborra, S. Green Chem. 2014, 16, 516−
5
47.
53
(4) Bermudez, J. M.; Menen
Garcia-Martinez, J.; Luque, R. Green Chem. 2013, 15, 2786−2792.
5) Upare, P. P.; Lee, J.-M.; Hwang, Y. K.; Hwang, D. W.; Lee, J.-H.;
́
dez, J. A.; Romero, A. A.; Serrano, E.;
sample surface in the reaction liquid. Acid−base titration
analysis carried out in the same or similar reaction solvent
allows one to evaluate acidity under conditions similar to those
used during reaction. The amount of effective sites is relevant
for silica and increases in the presence of the supported metal.
All these observations are consistent with the reaction
mechanisms proposed. Thus, the reaction of GVL in the
presence of strong Bronsted sites such as those present on the
(
Shiva, B.; Halligudi, S. B.; Hwang, J.-S.; Chang, J.-S. ChemSusChem
2011, 4, 1749−1752.
(6) Yang, Y.; Gao, G.; Zhang, X.; Li, F. ACS Catal. 2014, 4, 1419−
1
(
2
(
425.
7) Yan, K.; Lafleur, T.; Jarvis, C.; Wu, G. J. Cleaner Prod. 2014, 72,
30−232.
8) Abdelrahman, O. A.; Heyden, A.; Bond, J. Q. ACS Catal. 2014, 4,
surface of Ru/H-β and Ru/H-ZSM5 or Ru/SBA-SO H gives
3
1171−1181.
14,30
pentenoic and pentanoic acids
that are then esterified. On
(9) Mohan, V.; Venkateshwarlu, V.; Pramod, C. V.; Raju, B. D.; Rao,
K. S. R. Catal. Sci. Technol. 2014, 4, 1253−1259.
the contrary, in the presence of Cu/SiO , the lactone ring is
2
(
1
(
10) Fabos, V.; Mika, L. T.; Horvath, I. T. Organometallics 2014, 33,
opened by nucleophilic attack of 1-pentanol to the carbonyl
group activated through interaction with the Lewis acid sites on
the catalyst surface.
81−187.
11) Tang, X.; Chen, H.; Hu, L.; Hao, W.; Sun, Y.; Zeng, X.; Lin, L.;
Liu, S. Appl. Catal., B 2014, 147, 827−834.
12) Yang, Z.; Huang, Y.-B.; Guo, Q.-X.; Fu, Y. Chem. Commun.
2013, 49, 5328−5330.
13) Yuan, J.; Li, S.-S.; Yu, L.; Liu, Y.-M.; Cao, Y.; He, H.-Y.; Fan, K.-
N. Energy Environ. Sci. 2013, 6, 3308−3313.
14) Luo, W.; Upakul Deka, U.; Andrew, M.; Beale, A. M.; Ernst, R.
H.; van Eck, E. R. H.; Pieter, C. A.; Bruijnincx, P. C. A.; Weckhuysen,
B. M. J. Catal. 2013, 301, 175−186.
15) Horvath, I. T.; Mehdi, H.; Fabos, V.; Boda, L.; Mika, L. T. Green
Chem. 2008, 10, 238−242.
16) Fegyverneki, D.; Orha, L.; Lang, G.; Horvath, I. T. Tetrahedron
010, 66, 1078−1081.
17) Luterbacher, J. S.; Rand, J. M.; Martin Alonso, D.; Han, J.; Tyler
Youngquist, J.; Maravelias, C. T.; Pfleger, B. F.; Dumesic, J. A. Science
2014, 343, 277−280.
18) Qi, L.; Fung Mui, Y.; Wing Lo, S.; Lui, M. Y.; Akien, G. R.;
Horvath, I. T. ACS Catal. 2014, 4, 1470−1477.
19) Martin Alonso, D.; Wettstein, S. G.; Dumesic, J. A. Green Chem.
013, 15, 584−595.
20) Serrano-Ruiz, J. C.; Luque, R.; Sepulveda-Escribano, A. Chem.
Soc. Rev. 2011, 40, 5266−5281.
21) Martin Alonso, D.; Wettstein, S. G.; Bond, J. Q.; Root, T. W.;
Dumesic, J. A. ChemSusChem 2011, 4, 1078−1081.
(22) Bond, J. Q.; Martin Alonso, D.; Wang, D.; West, R. M.;
Dumesic, J. A. Science 2010, 327, 1110−1114.
23) Bozell, J. J. Science 2010, 329, 522−523.
24) Serrano-Ruiz, J. C.; Braden, D. J.; West, R. M.; Dumesic, J. A.
Appl. Catal., B 2010, 100, 184−189.
25) Bond, J. Q.; Martin Alonso, D.; West, R. M.; Dumesic, J. A.
Langmuir 2010, 26, 16291−16298.
(26) Lange, J.-P.; Price, R.; Ayoub, P. M.; Louis, J.; Petrus, L.; Clarke,
L.; Gosselink, H. Angew. Chem., Int. Ed. 2010, 49, 4479−4483.
(
CONCLUSIONS
■
(
The role of metal dispersion in determining the catalytic
activity in hydrogenation and oxidation reactions has long been
recognized. On the contrary, the influence of the supported
metal particle size on the acidic properties of the catalyst has
very rarely been investigated. The present paper shows that
very small metal particles can exhibit catalytically relevant Lewis
acidity. This property, joined with a significant hydrogenation
activity of the metal, allows one to design new bifunctional
(
(
(
2
(
catalysts, such as Cu/SiO , without the need of an acidic
2
support. In the case of GVL to pentyl valerate reaction, this
translates not only into a more simple catalytic system but also
into a more selective one as a result of the presence of a unique
catalytic site that prevents secondary reactions to take place.
(
(
2
(
ASSOCIATED CONTENT
■
(
*
S
Supporting Information
N adsorption/desorption isotherms; pore volume distribution
2
as a function of pore radius for SiO A and Cu/SiO A; SEM
2
2
(
(
images of silica B surface and 16% Cu catalyst surfaces; XRD of
6% CuO/SiO B; py adsorption spectra at different desorption
1
2
temperature for all reduced and unreduced Cu catalysts;
HRTEM micrograph and particle size distribution of 16% Cu/
(
SiO B catalyst. This material is available free of charge via the
2
Internet at http://pubs.acs.org.
2
825
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826

ACS Catalysis
Research Article
(
27) Chan-Thaw, C. E.; Marelli, M.; Psaro, R.; Ravasio, N.;
Zaccheria, F. RSC Adv. 2013, 3, 1302−1306.
28) Shimizu, K.; Satsuma, A. Energy Environ. Sci. 2011, 4, 3140−
153.
29) Shimizu, K.; Furukawa, H.; Kobayashi, N.; Itaya, Y.; Satsuma, A.
Green Chem. 2009, 11, 1627−1632.
30) Pan, T.; Deng, J.; Xu, Q.; Xu, Y.; Guo, Q.-X.; Fu, Y. Green Chem.
013, 15, 2967−2974.
31) Gervasini, A.; Messi, C.; Ponti, A.; Cenedese, S.; Ravasio, N. J.
Phys. Chem. C 2008, 112, 4635−4642.
32) Ravasio, N.; Zaccheria, F.; Gervasini, A.; Messi, C. Catal.
Commun. 2008, 9, 1125−1127.
33) Zaccheria, F.; Scotti, N.; Marelli, M.; Psaro, R.; Ravasio, N.
Dalton Trans. 2013, 42, 1319−1327.
34) Zaccheria, F.; Santoro, F.; Psaro, R.; Ravasio, N. Green Chem.
011, 13, 545−548.
35) Mariani, M.; Zaccheria, F.; Psaro, R.; Ravasio, N. Catal.
Commun. 2014, 44, 19−23.
36) Zaccheria, F.; Shaikh, N. I.; Scotti, N.; Psaro, R.; Ravasio, N.
Top. Catal. 2014, 57, 1085−1093.
37) Barrett, E. P.; Joyner, L. G.; Halenda, P. P. J. Am. Chem. Soc.
(
3
(
(
2
(
(
(
(
2
(
(
(
1
(
951, 73, 373−380.
38) Benfield, R. E. J. Chem. Soc., Faraday Trans. 1992, 88, 1107.
(
39) Mori, K.; Hara, T.; Mizugaki, T.; Ebitani, K.; Kaneda, K. J. Am.
Chem. Soc. 2004, 126, 10657−10666.
40) Carniti, P.; Gervasini, A.; Marzo, M. Catal. Today 2010, 152,
2−47.
41) Zaccheria, F.; Psaro, R.; Ravasio, N. Green Chem. 2009, 11,
62−465.
42) Hensen, E. J. M.; Poduval, D. G.; Degirmenci, V.; Michel
Ligthart, D. A. J.; Chen, W. F.; Mauge, F.; Rigutto, M. S.; van Veen, J.
(
4
(
4
(
́
A. R. J. Phys. Chem. C 2012, 116, 21416−21429.
(
(
43) Parry, E. P. J. Catal. 1963, 2, 371−379.
44) Busca, G.; Martra, G.; Zecchina, A. Catal. Today 2000, 56, 361−
370.
(
45) Gervasini, A.; Bennici, S.; Auroux, A.; Guimon, C. Appl. Catal., A
007, 331, 129−137.
46) Emeis, C. A. J. Catal. 1993, 141, 347−354.
47) Sanchez Escribano, G. V.; del Hoyo Martínez, C.; Fernan
Lopez, E.; Gallardo Amores, J. M.; Busca, G. Catal. Commun. 2009, 10,
61−864.
48) Kumar, M.; Srivastava, M.; Yadav, R. A. Spectrochim. Acta, Part A
013, 111, 242−251.
49) Zhang, Q.; Deng, W.; Wang, Y. Chem. Commun. 2011, 47,
275−9292.
50) J. Urban, J.; H. Sack-Kongehl, H.; K.Weiss, K. Catal. Lett. 1997,
9, 101−108.
51) Lignier, P.; Bellabarba, R.; Tooze, R. P. Chem. Soc. Rev. 2012, 41,
708−1720.
52) Tukacs, J. M.; Kiral
Dibo, G.; Kegl, T.; Mika, L. T. Green Chem. 2012, 14, 2057−2065.
53) Carniti, P.; Gervasini, A.; Biella, S.; Auroux, A. Chem. Mater.
005, 17, 6128−6136.
2
(
(
́
́
dez
́
8
(
2
(
9
(
4
(
1
(
́ ́
y, D.; Stradi, A.; Novodarszki, G.; Eke, Z.;
́
́
(
2
2
826
dx.doi.org/10.1021/cs500581a | ACS Catal. 2014, 4, 2818−2826