Recycling of CO2 by Hydrogenation of Carbonate Derivatives to Methanol: Tuning Copper–Oxide Promotion Effects in Supported Catalysts

Article abstract of DOI:10.1002/cssc.202000166

The selective hydrogenation of organic carbonates to methanol is a relevant transformation to realize flexible processes for the recycling of waste CO₂ with renewable H₂ mediated by condensed carbon dioxide surrogates. Oxide-supported copper nanoparticles are promising solid catalysts for this selective hydrogenation. However, essential for their optimization is to rationalize the prominent impact of the oxide support on performance. Herein, the role of Lewis acid centers, exposed on the oxide support at the periphery of the Cu nanoparticles, was systematically assessed. For the hydrogenation of propylene carbonate, as a model cyclic carbonate, the conversion rate, the apparent activation energy, and the selectivity to methanol correlate with the Lewis acidity of the coordinatively unsaturated cationic sites exposed on the oxide support. Lewis sites of markedly low and high electron-withdrawing character promote unselective decarbonylation and decarboxylation reaction pathways, respectively. Supports exposing Lewis sites of intermediate acidity maximize the selectivity to methanol while inhibiting acid-catalyzed secondary reactions of the propanediol product, and thus enable its recovery in cyclic processes of CO₂ hydrogenation mediated by condensed carbonate derivatives. These findings help rationalize metal–support promotion effects that determine the performance of supported metal nanoparticles in this and other selective hydrogenation reactions of significance in the context of sustainable chemistry.

Full text of DOI:10.1002/cssc.202000166

Accepted Article
Title: Recycling of CO2 via hydrogenation of carbonate derivatives to
methanol: tuning copper-oxide promotion effects in supported
catalysts
Authors: Jonglack Kim, Norbert Pfänder, and Gonzalo Prieto
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.202000166
Link to VoR: http://dx.doi.org/10.1002/cssc.202000166
A Journal of
www.chemsuschem.org

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Recycling of CO2 via hydrogenation of carbonate derivatives to
methanol: tuning copper-oxide promotion effects in supported
catalysts
Jonglack Kim[a], Norbert Pfänder[b], and Gonzalo Prieto*[a]
[a]
Dr. J. Kim, Dr. G. Prieto
Heterogeneous Catalysis
Max-Planck-Institut für Kohlenforschung
Kaiser-Wilhelm-Platz 1, 45470 Mülheim an der Ruhr, Germany.
E-mail: prieto@mpi-muelheim.mpg.de
N. Pfänder
[b]
Max-Planck-Institut für Chemische Energiekonversion
Stiftstraße 34-36, 45470 Mülheim an der Ruhr, Germany.
Supporting information for this article is given via a link at the end of the document.
Abstract: The selective hydrogenation of organic carbonates to
methanol is a relevant transformation to realize flexible processes for
the recycling of waste CO2 with renewable H2 mediated by
condensed carbon dioxide surrogates. Oxide-supported copper
nanoparticles (NPs) are promising solid catalysts for this selective
hydrogenation. However, essential for their optimization is to
rationalize the prominent impact of the oxide support on
performance. Herein, the role of Lewis acid centers, exposed on the
oxide support at the periphery of the Cu NPs is systematically
assessed. For the hydrogenation of propylene carbonate, as a
model cyclic carbonate, the conversion rate, the apparent activation
energy as well as the selectivity to methanol correlate with the Lewis
acidity of the coordinatively unsaturared cationic sites (cus) exposed
on the oxide carrier. Lewis sites of markedly low and high electron
withdrawing character promote decarbonylation and decarboxylation
unselective reaction pathways, respectively. Supports exposing
Lewis sites of intermediate acidity maximize the selectivity to
methanol while inhibiting secondary, acid-catalyzed reactions of the
propanediol product, enabling its recovery in cyclic processes of CO2
hydrogenation mediated by condensed carbonate derivatives. These
findings help rationalize metal-support promotion effects which
determine the performance of supported metal NPs in this and other
selective hydrogenation reactions of significance in the context of
sustainable chemistry.
Introduction
In supported metal catalysts, the role of the oxide support goes
often beyond that of a high-surface-area carrier which adds to
the metal dispersion and the mechanical stability of a catalyst.
Oxide species are known to exert "promotion" effects which
modify profoundly the catalytic performance and stability of the
metal species they are in contact to.[1] Metal/oxide promotion
effects are particularly significant for reactions which require the
activation of polarized bonds (C-O, N-O, S-O, etc) on "oxophilic"
oxide surfaces to operate in conjunction with classical metal-
catalyzed elementary steps, e.g. H2 dissociation.[2] Identifying
experimentally accessible physicochemical descriptors, as a
means to reach a quantitative, ideally predictive description of
such metal/oxide promotion effects is currently a major goal.
Such insights are expected to set firmer basis for a deeper
understanding of existing catalytic systems, as well as a more
rational development of innovative catalysts and processes.
Catalysis by nanoparticles of coinage metals, i.e. Au, Ag and Cu,
is particularly dominated by metal/oxide promotion effects. This
is in part associated to the relatively low oxophilicity and
remarkably low carbophilicity displayed by these metals, which
often require the concerted action of oxide species to activate
reactants with polar bonds. Exemplary, copper/oxide interfaces
have been proposed to play a pivotal role for the very rich, but
also strongly support-dependent, reactivity displayed by copper
nanoparticles dispersed on oxide carriers. This includes
reactions of particular significance for a more sustainable
chemical industry, including the selective hydrogenation of
(biogenic) oxygenated compounds[3], alcohol synthesis from
syngas,[4] and CO2,[5] lignin depolymerization,[6] or (bio)ester
hydrogenolysis.[7]
The hydrogenation of CO2 organic derivatives, i.e. carbonates,
carbamates or ureas, into methanol are transformations of
significant interest in the context of the chemical CO2 recycling,
and archetypal catalytic reactions where the activation of highly
polarized bonds is central for the overall selectivity. Under
certain scenarios, the hydrogenation of CO2 to methanol using
H2 of renewable origin is considered a feasible route to
chemically recycle waste CO2 streams into a versatile platform
1
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
chemical, as well as to store renewable energy into a liquid
vector.[8] Surmounting some of the limitations envisaged for
direct CO2 hydrogenation processes, the aforementioned
organic derivatives could play a substantial role as high-density,
therefore cost-effectively transportable, intermediate CO2
"couriers" and thus facilitate the bridging of remote point sources
of waste CO2 and renewable H2, as well as the buffering of
transient fluctuations which are inherent to both supplies
(Scheme S1 in the Supporting Information).[9] These compounds
can be obtained with high selectivity by reaction of CO2 with
different organic compounds such as alcohols, glycols, epoxides
or amines.[10] In
a separate catalytic step, CO2 organic
derivatives can be hydrogenatively converted in the presence of
H2 as co-reactant. If the latter transformation is achieved with
high selectivity, the CO2-derived carbon may act as the
precursor for methanol in a reaction which, unlike the direct CO2
hydrogenation, is not bonded to thermodynamic limitations to the
methanol yield, while the organic residues can be recovered and
recycled for further CO2 "capture", enabling a cyclic overall
process consisting of a net conversion of CO2 and H2 into
methanol.
Molecular catalysts based on Ru and Ir have been designed to
achieve high reaction rates and product selectivities for the
selective hydrogenation of organic CO2 derivatives in solution.[11]
However, solid catalysts are highly preferred, as they enable an
effective product recovery and catalyst recycling. Catalysts
based on supported copper nanoparticles have been proposed
for the heterogeneously catalyzed hydrogenation of organic
carbonates into methanol.[12] Stark differences in activity and
selectivity have been reported as a function of the nature of the
oxide support. However, in spite of the significance of these
oxide support effects for a rational development of advanced
catalysts, they remain poorly understood at present.
Figure 1: Representative Cs-HAADF-STEM micrographs for TaOx@Al2O3
(a,b) and SmOx@Al2O3 (c,d) supports consisting of oxide overlays on a
high-surface-area γ-Al2O3 substrate. Arrows point to the higher Z-contrast
oxide overlay observed as a rim around the γ-Al2O3 nanocrystals. In panel
d, samarium atoms within the SmOx overlay are resolved.
the overlay oxide. X-ray diffraction did not detect crystallites of
the MOx species (Figure S1). All materials, including the neat γ-
Al2O3 substrate, showed type-IV N2-physisorption isotherms
characteristic of mesoporous materials (Figure S2), with
essentially identical Al2O3-normalized specific surface areas
(277±21 m2 g-1Al2O3) and mesopore volumes (0.70±0.05 cm3 g-
1
Al2O3), Table 1. These results prove the absence of any
Herein, we address the promotion effects of peripheral Lewis
oxide centers on the Cu-catalyzed hydrogenation of propylene
carbonate –as a model organic carbonate– to methanol. For a
series of model Cu catalysts, the surface Lewis acidity (electron
accepting) character of the peripheral oxide support has been
systematically modified and quantitatively assessed in a broad
study space. This physico-chemical feature is shown to be a
rather general and quantitative descriptor for both activity and
selectivity, and thus a suitable design parameter for optimized
copper-based catalysts.
significant mesopore plugging upon the deposition of the MOx
oxides, which would have blocked part of the porosity of the
alumina carrier to N2 uptake. In addition, the MOx@Al2O3 oxides
showed an average mesopore value of 9.2±0.6 nm, i.e. ca. 2 nm
narrower than for the pristine γ-Al2O3 carrier. This reduction in
mesopore size agrees reasonably well with that expected upon
deposition of a monolayer of the MOx species on the wall of the
Al2O3 mesopores. It is therefore inferred that the MOx species
are in all cases deposited as an amorphous, essentially two-
dimensional overlay on the Al2O3 surface. Indeed, Cs-HAADF-
STEM could provide a direct visual proof for the existence of a
(few-)atom-thick oxide overlay enveloping the γ-Al2O3
nanocrystals for selected materials, e.g. TaOx@Al2O3 and
SmOx@Al2O3, for which the Z-contrast of the M atoms (M=Ta,
Sm) relative to the Al2O3 matrix was maximum among the series
of materials (Figure 1).
Results and Discussion
Characterization of oxide support materials
In order to synthesize a series of oxide support materials
showing similar textural properties (specific surface area, pore
volume and diameter) but different surface Lewis acidities in a
broad study space, the surface of a mesoporous γ-Al2O3 carrier
was overlaid with different transition metal and lanthanide oxides
(MOx) of markedly different intrinsic Lewis acidity. In all cases,
the surface coverage of M, where M=Sm, Y, Sc, Zr and Ta) was
set to ca. 4.5-5.0 Mat nm-2, which has been previously identified
A notable uniformity in textural properties is achieved within the
series of MOx@Al2O3 support materials, which would otherwise
not be possible should bulk oxides be employed. Moreover, the
incorporation of the different oxide overlays on the common
Al2O3 substrate resulted in a series of materials spanning in a
very broad range of surface Lewis acidity. Lewis acidity in oxides
stems from coordinatively unsaturated metal sites (cus) exposed
on their outer surface. The relative Lewis acidity of the cus on
the surface of the oxide support materials was quantified by
to correspond to the oxide monolayer content of different oxides
13]
on γ-Al2O3.[5c,
The resulting series of support oxides is
hereafter denoted as MOx@Al2O3, where MOx corresponds to
2
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
20%H2/N2 was selected to activate the catalysts prior to catalysis
(Figure S5).
Following in situ catalyst reduction, X-ray photoemission
spectroscopy (XPS) showed Cu2p3/2 binding energies (BEs) of
932.2±0.4 eV for the entire series of materials (Supplementary
Table S1), with Cu L3M4,5M4,5 at a kinetic energy of ca. 919 eV.
Jointly, these observations point to the full reduction of copper
species to metallic state, in line with what could be inferred from
the H2-TPR profiles. The overlay MOx oxides showed, however,
no appreciable reduction following the H2 activation step at 523
K (Table S1). Only in the case of TaOx could a contribution from
a
sub-oxide species be detected. This contribution was
nevertheless determined to account for <15% of the tantalum
species on the catalyst surface. These results accredit the low
reducibility of the oxides deposited as an overlay on the surface
of the γ-Al2O3 carrier.
Cs-HAADF-STEM was coupled to Energy Dispersive X-ray
(EDX) spectroscopy to investigate the dispersion of Cu and MOx
species on the surface of reduced catalysts. Figure 3 compiles
representative STEM micrographs and compositional EDX
mappings, collected at both the mesoscopic and nanometric
lengthscales, respectively, for selected catalysts after reduction.
A remarkably uniform mesospatial distribution for the MOx
overlay oxides could be ascertained in all cases, discarding any
appreciable zoning or agglomeration. Owing to the relatively
high Z-contrast contributed by the MOx@Al2O3 support oxides,
the presence of Cu nanoparticles (with sizes of ca. 6-15 nm)
could only be appraised in few catalysts, with the assistance of
EDX compositional maps to identify copper-enriched nanoscale
regions. For other materials, however, direct visualization of the
Cu nanoparticles proved challenging owing to limited Z-contrast
differences and lower Cu nanoparticle sizes (Figure 3c and
Figure S6).
Figure 2: Comparison between the surface acidity determined experimentally
for the overlay MOx oxides supported on γ-Al2O3 with the theoretical (Lewis)
acidity of the corresponding bulk oxides, as defined by Jeong et al.[14], i.e.
represented by NM-2δM, where NM denotes the formal oxidation state and δM is
the Sanderson's partial charge of the cations in the bulk oxides: Sm2O3, Y2O3,
Sc2O3, ZrO2 and Ta2O5, respectively.
means
of
UV-vis
spectroscopy
coupled
to
1,2-
dihydroxyanthraquinone (alizarin) as a surface probe molecule,
as described elsewhere.[14] The lowest intramolecular charge
transfer (IMCT) from the catechol subunit to the polycyclic
system of the probe molecule appears as a band at around 505
nm, in the visible range of the spectra (Figure S3). This energy
is hereafter denoted as the spectroscopic parameter η, and it is
plotted in Figure 2 versus the theoretical Lewis acidity of the
corresponding bulk-type MOx oxide for the entire series of
[14]
MOx@Al2O3 materials.
The linear correlation observed
demonstrates that the ranking of Lewis acidities established
experimentally for those cus exposed on the surface of the
Al2O3-supported oxide overlays matches well that expected for
the corresponding bulk-type oxides, covering a very broad range
of acidity.
Given the limitations associated to gas chemisorption methods
to reliably determine Cu dispersion from series of catalysts
supported on a wide range of oxides,[16] we have herein resorted
to a quantitative analysis of the XPS spectra, registered for the
in situ reduced catalysts, as a means to evaluate Cu dispersion
and average particle size in the series of Cu/MOx@Al2O3
catalysts. This bulk-sampling method provides a quantitative
assessment which is complementary to the local electron
microscopy analysis. As summarized in Table 1 and Figure S7,
the average Cu nanoparticle sizes determined for the series of
catalysts synthesized with a copper content of 1.5 Cu nm-2 were
in the range of 4-16 nm. At this constant Cu surface content, the
average Cu nanoparticle size increased for catalysts supported
on increasingly Lewis basic oxides, notably YOx@Al2O3 and
SmOx@Al2O3. This trend may be associated to a partial
hydrolysis of the copper nitrate precursor during impregnation
and drying, which is known to result in a higher extent of metal
agglomeration[15c] and it is expected to be facilitated on oxide
supports of basic character via local pH increments of the
mesopore-confined impregnation solution during catalyst
synthesis. Catalysts with analogous average Cu nanoparticle
sizes could also be synthesized on the most Lewis acidic
TaOx@Al2O3 support, by increasing the surface-specific Cu
loading (Figure S7).
Characterization of copper-based catalysts
Copper was incorporated on the series of the MOx@Al2O3
support materials by incipient wetness impregnation. The
nominal surface copper content was generally set to 1.5 Cuat
nm-2. On selected supports, catalysts with higher surface Cu
contents of 3.0 and 4.5 Cuat nm-2 were also synthesized. The
thermal treatment following impregnation was designed to
ensure full dehydration of the copper nitrate precursor under
mild temperature (343 K) in flow of N2, prior to its decomposition
at higher temperatures. Previous studies have shown that this
route leads to highly dispersed Cu(II) species on the surface of
the oxide carrier. [15] X-ray diffraction of the as-calcined catalysts
showed no signs of CuO crystals for catalysts with a Cu content
of 1.5 Cu nm-2 (2.5-4 wt%) indicative of the very high copper
dispersion achieved (Figure S4). Hydrogen temperature-
programmed (H2-TPR) experiments showed H2 consumptions,
corresponding to the reduction of Cu(II) species to their metallic
state, peaking in the temperature range of 425-510 K. On the
basis of this reducibility study, a treatment at 523 K under flow of
3
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Cu
Al
Sm
a)
100 nm
100 nm
100 nm
100 nm
10 nm
10 nm
10 nm
10 nm
Y
Cu
Al
b)
100 nm
100 nm
100 nm
100 nm
10 nm
10 nm
10 nm
10 nm
Ta
Cu
Al
c)
100 nm
100 nm
100 nm
100 nm
10 nm
10 nm
10 nm
10 nm
Figure 3: Representative Cs-HAADF-STEM micrographs (left panel) and the corresponding Energy Dispersive X-ray Spectroscopy (EDX) compositional maps
for nanometer-thick cross-sections of a) Cu/SmOx@Al2O3, b) Cu/YOx@Al2O3, and c) Cu/TaOx@Al2O3 catalysts, with a copper content of 1.5 Cuat nm-2, after
reduction.
Propylene carbonate hydrogenation
the Lewis acidity of the cus on the oxide support. The highest
-1
The catalytic performance of the set of Cu/MOx@Al2O3 model
catalysts was evaluated in the hydrogenation of propylene
carbonate in the liquid phase. Methanol and propane(di)ols, CO
and CO2 were the major reaction products. Light (C1-C3)
hydrocarbons were detected under the highest tested reaction
temperatures, albeit only in minor amounts (<5%). Under
identical reaction settings, blank experiments with the Cu-free
MOx@Al2O3 oxide support materials showed insignificant
propylene carbonate conversion (<5% after 24 h), evidencing
that the presence of Cu is essential for catalysis.
reaction rates (0.29-0.31 mol gCu L-1 min-1) were registered for
Cu nanoparticles deposited on oxide supports exposing surface
cus of the lowest Lewis acidity, i.e. SmOx@Al2O3 and
YOx@Al2O3. Upon progressively increasing the surface Lewis
acidity of the support, the reaction rate decreased progressively,
by a factor of ca. 3 over the entire study space.
In order to assess whether a secondary reaction pathway, via
hydrogenation of CO2, contributed to the observed methanol
formation rates, a set of control experiments was conducted
feeding CO2 and H2 as reactants. In these tests, the CO2 partial
pressure in the gas feed was set to 10 bar, i.e. simulating the
case in which propylene carbonate had initially been fully
Remarkably, as shown in Figure 4a, the initial propylene
carbonate conversion rate showed a strong dependence with
4
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Figure 4: Impact of the Lewis acidity of cus exposed on the surface of the oxide support on the catalytic activity. Evolution of a) the initial propylene carbonate
(PC) conversion rate; and b) the apparent activation energy with the spectroscopic parameter η, in the hydrogenation of propylene carbonate for the series of
Cu/MOx@Al2O3 catalysts. Reaction conditions: T= 453 K, P(H2)=40 bar, initial propylene carbonate concentration= 0.25
M in 1,4-dioxane, catalyst
concentration= 0.15 mM Cu. Error bars represent the standard error from 3 independent tests. Dashed lines are included as guides to the eye.
decarboxylated under reaction conditions. However, the
methanol formation rates registered from CO2 were ca. one to
two orders of magnitude lower than those observed from
propylene carbonate (see Table S2 and the accompanying
discussion in the Supporting Information). These results
confirmed the prevalence of a direct carbonate hydrogenation
pathway as the major methanol production route.
cus sites exposed on the surface of the support turn either more
electron donating or more electron accepting.
The copper particle size is another parameter which might
potentially have an impact on the catalytic performance.
Structure sensitivity phenomena, i.e. particle size-dependent
turnover frequencies, have been reported for hydrogenation
reactions with supported copper nanoparticles in the sub-10 nm
size regime.[17] However, an analysis of the catalytic
performance for catalysts displaying different average Cu
nanoparticle sizes in the range of 4.3-15.2 nm supported on a
common oxide carrier, i.e. TaOx@Al2O3, showed very similar
apparent activation energy (Figure 4a) as well as product
selectivities at a constant carbonate conversion level of 80%
(Figure 5). These results further confirm that the Lewis acidity of
the oxide support is a dominant factor for reactivity. In addition,
these findings suggest that coordinatively unsaturated cationic
centers on the surface of the oxide species, at the periphery of
the copper nanoparticles, are involved in a kinetically relevant
reaction step.
The differences observed in the overall carbonate conversion
rate among the series of catalysts might arise from either active
sites of different intrinsic reactivity and/or differences in the
relative abundance of identical active sites. However, an
analysis of the initial carbonate conversion rates in the
temperature range of 413-473 K (Figure S8) revealed a stark
dependence of the apparent activation energy (Ea) for carbonate
conversion with the oxide cus Lewis acidity. As shown in Figure
4b, Ea increased rather linearly with the spectroscopic parameter
η, from 28 kJ mol-1 for Cu/SmOx@Al2O3 to 52.5 kJ mol-1 for
Cu/TaOx@Al2O3. It is hence inferred that the differences in
reactivity are indeed connected to reaction pathways with
different overall energetic barriers operating on different
catalysts, rather than to simply differences in the surface density
of active sites.
Not only the overall selectivity to carbon oxides but also the
composition of these gas-phase by-products depended
markedly on the Lewis character of the peripheral cus sites. As
shown in Figure 5b the CO/CO2 molar ratio in the products, at a
given carbonate conversion level of 80%, decreased linearly
upon increasing the spectroscopic parameter η. Different side-
reaction pathways might be responsible for the production of CO
and CO2 under the reaction conditions applied for propylene
carbonate hydrogenation. Figure 6a summarizes these possible
reaction pathways (pathways (ii) through (iv)), together with the
desired selective hydrogenation to methanol and propanediol
(pathway (i)). Carbon monoxide formation might be the result of
decarbonylation of the carbonate substrate (pathway (ii)). It is
therefore inferred from all these catalytic results that oxide
supports exposing surface cus of an increasingly stronger Lewis
In order to assess the influence of the support Lewis acidity on
product selectivity, the reaction product pattern was examined
as a function of the carbonate conversion for all catalysts at a
reaction temperature of 453 K. Relatively steady product
selectivity patterns were obtained in all cases after a propylene
carbonate conversion level of 70% had been reached (Figure
S9). Figure 5a depicts the evolution of the methanol selectivity,
at a constant propylene carbonate conversion of 80%, as a
function of η. A volcano dependence is observed, according to
which the selectivity to methanol is maximized (>60%) for Cu
NPs deposited on oxides exposing cus of intermediate Lewis
acidity, namely ZrOx@Al2O3, and decreases remarkably as the
5
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Figure 5: Impact of the Lewis acidity of cus exposed on the surface of the oxide support on the reaction selectivity. Evolution of a) the selectivity to methanol;
and b) the CO/CO2 molar ratio within the products with the spectroscopic parameter η, in the hydrogenation of propylene carbonate with the series of
Cu/MOx@Al2O3 catalysts. Reaction conditions: T= 453 K, P(H2)=40 bar, initial propylene carbonate concentration= 0.25
M in 1,4-dioxane, catalyst
concentration= 0.15 mM Cu, propylene carbonate conversion: 80%. Error bars represent the standard error from 3 independent tests. Dashed lines are included
as guides to the eye.
basic character promote
a
fast propylene carbonate
In order to assess the effect of water on the conversion of
propylene carbonate, separate experiments were conducted on
the most Lewis acidic Cu/TaOx@Al2O3 catalyst in which different
amounts of water were intentionally added at the beginning of
the reaction. As shown in Figure 7b, the addition of increasing
amounts of exogenous water resulted, expectedly, in a steep
decrease in methanol selectivity, from ca. 30% to <5% for
H2O/carbonate initial molar ratios ≥0.1, in favor of CO2 as the
decarbonylation. This reaction pathway is by and large
responsible for both the higher carbonate conversion rates and
the higher selectivity to CO observed as the spectroscopic
parameter η decreases (Figures 4 and 5, respectively).
Carbon dioxide may form as the product of
a direct
decarboxylation of propylene carbonate (pathway (iii) in Figure
6a) or via a carbonate hydrolysis in the presence of H2O
(pathway (iv)). In turn, water might be in situ sourced to the
reaction medium as the product of intermolecular dehydration
(etherification with methanol and/or oligomerization) secondary
reactions of the 1,2-propanediol primary product, as summarized
schematically in Figure 6b. Both routes undesirably contribute to
a depletion of the glycol product and thus disable its recycling as
"CO2 carrier" in an overall process of CO2 utilization. These
secondary reactions are known to be acid catalyzed. As a matter
of fact, a detailed analysis of the reaction products arousing from
the propylene organic moiety in the carbonate reactant revealed
that the extent to which secondary dehydration reactions take
place is higher for catalysts supported on oxides exposing cus of
marked Lewis acid nature. As shown in Figure 7a, the lumped
selectivity to dehydration products, including products of
methanol cross-etherification with propane(di)ols as well as
propaneglycol oligomers, increased upon increasing η beyond
2.48 eV.
product of carbonate hydrolysis. In parallel,
a significant
increase in the initial carbonate consumption rate of about a
factor of 4 was also observed. These results evidence that,
reaction conditions favoring propylene carbonate hydrolysis lead
to notably higher carbonate conversion rates alongside
remarkably lower methanol selectivity than those registered
under standard reaction settings, i.e. without the addition of
exogenous water into the reaction medium. It is therefore
deduced that carbonate hydrolysis pathways driven by the
presence of endogenous water are not the major reaction
pathway underlying the production of CO2 under propylene
carbonate hydrogenation conditions. Rather, our results suggest
that the direct carbonate decarboxylation pathway dominates
and it is by and large responsible for the formation of CO2 on Cu
NPs supported on the most Lewis acidic oxides. The carbonate
decarboxylation pathway being facilitated on oxides of marked
Lewis acid character is in agreement with previous predictions
based on quantum-mechanical calculations.[18]
In view of these results, a higher water concentration in the
reaction medium, and thus a higher driving force for propylene
carbonate hydrolysis, is inferred for catalysts supported on
oxides bearing strongly Lewis acidic cus. However, the
selectivity to 1-propanol, the product of the carbonate direct
decarboxylation route (pathway (iii) in Figure 6a), increased also
remarkably with increasing η beyond 2.45 eV, and dominated
The fact that the catalytic performance correlates strongly with
the Lewis acidity of those cus exposed on the oxide surface, at
the periphery of the Cu nanoparticles, suggests strongly the
involvement of these Lewis centers in the activation of the
carbonate reactant. Lewis centers of marked electron
withdrawing (acidic) character promote the cleavage of C-O
bonds between the carbonate functional group and the propane
over dehydration side-products for the most Lewis acidic catalyst, backbone of the substrate molecule, thus its unselective
i.e. Cu/TaOx@ Al2O3.
decarboxylation. On the contrary, centers of less electron
withdrawing character (more Lewis basic) seem to facilitate the
6
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
a)
b)
Figure 7: Contribution of dehydration secondary reactions and water byproducts to the catalytic performance. a) Evolution of the selectivity to n-propanol, 1,2-
propandiol and products of 1,2-propanediol dehydration reactions as a function of the Lewis acidity of the cus on the oxide support at a propylene carbonate
conversion of 80%. b) Influence of the addition of increasing amounts of exogenous water on the initial reaction rate (●) and methanol selectivity (▲) for the
hydrogenation of propylene carbonate catalyzed by Cu/TaOx@γ-Al2O3. Reaction conditions: PH2 =40 bar, initial propylene carbonate concentration= 0.25M in 1,4-
dioxane, catalyst concentration= 0.15 mM Cu. Lines in panel b are added as guide to the eye.
Figure 6: a) Schematic representation of the possible reaction pathways for the conversion of propylene carbonate under the applied reaction conditions: (i)
desired selective hydrogenation to methanol and 1,2-propanediol; (ii) direct decarbonylation pathway; (iii) direct decarboxylation pathway; and (iv) hydrolysis to
CO2 and 1,2-propanediol. b) Illustrative acid-catalyzed (i) esterification with methanol, and (ii) oligomerization secondary reactions for the 1,2-propanediol
primary product.
activation of the electrophilic carbonate functional group in the
reactant molecule, resulting in significantly reduced overall
activation energies and notably higher conversion rates.
However, the enhanced reactivity is in this case largely due to
the promotion of C-O bond cleavage within the carbonate group,
leading to an undesired decarbonylation of the substrate. Oxides
bearing cus Lewis centers of intermediate acidity, exemplary
those exposed on ZrOx, are essential to favour the selective
hydrogenation of the carbonate functional group in the reactant
to methanol. Moreover, mild Lewis acidity of the oxide support
effectively suppresses undesired secondary reactions of the
glycol side-product, such as dehydration and polymerization,
which are catalyzed by stronger Lewis acid sites. Such
preservation of the propanediol backbone of the carbonate
molecule enables its recycling as CO2 "carrier" in an overall
cyclic process of CO2 hydrogenation to methanol. These findings
underscore the central role of the oxide support at the periphery
of Cu nanoparticles in determining the catalytic performance.
Moreover, they serve to clearly identify the relative Lewis acidity
(electron donating/withdrawing character) of coordinatively
unsaturated cationic sites on the oxide surface as a central
physicochemical parameter to rationalize and optimize such
metal-oxide promotion effects, which play a pivotal role in a
number of catalytic reactions of environmental significance, such
as CO2 hydrogenation to alcohols[19] and CO[20] chemical vectors,
or the selective conversion of biogenic oxygenates,[3b, 3d] among
other.
identical porosity, whilst exposing coordinatively unsaturared
surface metal sites (cus) which span in a very wide range of
surface Lewis acidity. The dispersion of copper nanoparticles on
this battery of oxide supports enables a systematic assessment
of metal-support promotion effects in the selective
hydrogenation of propylene carbonate to methanol. Under
relevant reaction conditions, the overall (apparent) activation
energy for the reaction scales with the relative electron
withdrawing character of the cus on the oxide at the periphery of
the Cu nanoparticles – as quantified by UV-vis spectroscopy
coupled to 1,2-dihydroxyanthraquinone as probe molecule –
indicating the involvement of oxide Lewis centers in the
activation of the carbonate reactant. Moreover, the selectivity of
the reaction correlates also with the Lewis acidity of the cus on
the oxide support. Lewis sites of lower acidity, i.e. lower electron
withdrawing power promote
a low-Ea, thus faster, albeit
unselective decarbonylation of the carbonate reactant. On the
contrary, Cu nanoparticles interfaced with oxides bearing strong
Lewis acid sites favour a decarboxylation route. Oxide supports
exposing sites of intermediate Lewis acidity are optimal for the
hydrogenation of the carbonate functional group and maximize
the selectivity to methanol. In addition, a mild Lewis acidity of the
oxide carrier is also essential to inhibit secondary, acid-
catalyzed reactions of the propanediol primary product, enabling
its effective recovery and reuse, as required to realize a cyclic
process of hydrogenation of waste CO2 to methanol mediated by
the condensed carbonate derivative. The insights achieved on
this set of model catalysts are significant for the design of
optimized copper catalysts for the selective hydrogenation of
CO2-derived carbonates to methanol. More broadly, they
contribute towards an unifying and quantitative description of
copper-oxide promotion effects which play a crucial role in a
large number of selective hydrogenation reactions of
significance for sustainable chemistry.
Conclusion
The deposition of monolayer-content overlays of various
transition metal and lanthanide oxides on a high-surface-area γ-
Al2O3 carrier results in oxide support materials with essentially
7
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Figure 7: Contribution of dehydration secondary reactions and water byproducts to the catalytic performance. a) Evolution of the selectivity to n-propanol, 1,2-
propanediol and soluble products of 1,2-propanediol dehydration reactions as a function of the Lewis acidity of the cus on the oxide support at a propylene
carbonate conversion of 80%. b) Influence of the addition of increasing amounts of exogenous water on the initial reaction rate (●) and methanol selectivity (▲)
for the hydrogenation of propylene carbonate catalyzed by Cu/TaOx@γ-Al2O3. Reaction conditions: PH2 =40 bar, initial propylene carbonate concentration= 0.25
M in 1,4-dioxane, catalyst concentration= 0.15 mM Cu. Lines in panel b are added as guide to the eye.
Cu(NO3)3·3H2O (99%, Sigma-Aldrich) dissolved in 0.25 M HNO3(aq). The
Cu concentration of the stock solution was adjusted to attain preset
surface-specific Cu contents of 1.5, 3.0 or 4.5 Cuat nm-2. The as-
Experimental Section
impregnated solids were transferred to a quartz packed-bed reactor,
dried at 343 K for 10 hours and calcined at 623 K (3 K min-1) for 4 hours
Catalyst synthesis
under N2 flow. Then, the catalysts were transferred to a U-shaped
packed-bed glass reactor and activated by reduction at 523 K (1 K min-1)
A high-surface-area γ-Al2O3 carrier was synthesized from a nanosized
pseudo-boehmite precursor (Disperal-P2, Sasol Materials). A synthesis
gel containing the alumina precursor, the polyethyleneglycolether non-
ionic surfactant Tergitol 15-S-7 (Sigma-Aldrich) as porogen agent and
water, with a molar composition of Al:EO:H2O 1:1.51:43 (where EO
represents the ethylenoxide building units in the polymer) was
hydrothermally treated at 383 K in an oven for 48 h. Then, the gel was
dried at 353 K for 48 hours and at 393 K for 5 hours, and finally air-
calcined in a muffle oven at 873 K (3 K min-1) and sieved in the 100-200
μm size range for further catalyst synthesis steps.
in flow of 10 vol% H2/N2. Last, the reactor was let cool down to RT and
the solid recovered, transferred under exclusion of air into a glove-box
(O2 < 0.1 ppm, H2O < 0.1 ppm), and ground into a fine powder.
Characterization methods
Nitrogen physisorption isotherms were recorded in a Micromeritics ASAP
instrument (3Flex) unit after degassing the sample (ca. 100 mg) at 423 K
under vacuum for 5 h. Specific surface areas were derived using the
Brunauer-Emmett-Teller (B.E.T) method applied in the relative pressure
(P/P0) regime of 0.05-0.30. Total mesopore volumes were obtained from
the amount of N2 taken up at a relative pressure P/P0=0.95. Pore size
distributions and average mesopore diameters were determined applying
the Barrett-Joyner-Halenda (B.J.H) method to the desorption branch of
the isotherms.
A series of oxide support materials was synthesized via the deposition of
monolayer amounts of different transition metal and lanthanide oxides
(MOx, M=Sm, Y, Sc, Zr and Ta) on the γ-Al2O3 carrier. Metal precursors
were selected on the basis of previous studies [13a, 13b] to achieve a strong
interaction with the γ-Al2O3 surface. Sm(NO3)3·6H2O (99.9%, Sigma-
Aldrich), Y(NO3)3·6H2O (99.8%, Sigma-Aldrich), Sc(NO3)3·xH2O
(99.9%,Sigma-Aldrich), Zr(OC3H7)4 (70 wt. % in 1-Propanol, Sigma-
Aldrich) and Ta(OC2H5)5 (99.98%, Sigma-Aldrich) were used as received.
Stock solutions were prepared by dissolving nitrate precursors in Milli-Q
water, Zr(OC3H7)4 in dry 1-Propanol and Ta(OC2H5)5 in dry ethanol,
respectively. The γ-Al2O3 carrier was dried at 523 K for 3 hours under
dynamic vacuum (3 mbar). Then, metal/lanthanide incorporation was
achieved by incipient wetness impregnation. In all cases, the metal
Powder X-ray diffraction patterns were collected in a STOE Theta/Theta
diffractometer using Cu-Kα radiation monochromatized by a graphite
crystal (λ = 1.5406 Å) using a step size of 0.02° and a dwell time of 3 s
step-1.
UV-vis spectroscopy coupled to alizarin (1,2-dihydroxyanthraquinone) as
surface probe molecule was applied to assess the Lewis acidity of those
coordinatively unsaturated metal sites (cus) exposed on the surface of
the oxide supports, as reported previously.[14] The solids were dried and
soaked in a stock solution of the probe molecule in dry ethanol (0.15 mM)
under the exclusion of air. After solid recovery by filtration, excess
(unbound) alizarin was washed off with dry ethanol and the solid was
dried at RT under vacuum (3 mbar). Diffuse reflectance (DR)-UV-vis
spectra were recorded in a Pelkin Elmer Lamda 365 spectrometer using
BaSO4 as reflectance standard and converted into absorption using the
Kubelka-Munk (K/M) formalism. The peaking energy for the
intramolecular charge transfer (IMCT) band (EIMCT) of the adsorbed probe
was determined after subtraction of the spectra for the support material
prior to alizarin uptake.
content was adjusted to achieve
a surface-specific metal content
corresponding to the experimentally determined monolayer, i.e. 4.5-5.0
Mat nm-2.[13a] The as-impregnated solid was transferred into a quartz
packed-bed reactor, dried at 343 K for 10 hours and calcined at 773 K for
4 hours under flow of synthetic air (heating rates of 3 K min-1). The
calcined solid was transferred into a glove-box under exclusion of air.
The series of thus synthesized support materials was denoted as
MOx@Al2O3, where MOx stands for the overlay oxide.
Copper was incorporated on the surface of the series of MOx@Al2O3
support oxides by incipient wetness impregnation of copper nitrate
8
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
H2-Temperature programmed reduction (H2-TPR) experiments were
performed in a Micromeritics Autochem 2910 device. First, the sample
was flushed with Ar at 393 K for 2 hours (10 K min-1). Materials
containing SmOx and YOx oxide overlays, of marked basic character,
were additionally calcined in situ at 773 K (10 K min-1) for 4 hours under
flow of 5 % O2/Ar, in order to decompose surface carbonate species
which might have developed via uptake of atmospheric CO2 during
specimen manipulation and experiment setup. After being cooled down
to RT, the temperature was ramped to 1073 K (5 K min-1) under flow of
10% H2/Ar. Water evolved was removed from the outlet stream by a cold
trap (195 K) and the H2 consumption profiles were recorded using a
thermal conductivity detector (TCD).
column (60 m) and a flame ionization detector (FID), and a second one
connected to two consecutive packed bed columns (Agilent HS-Q 80/120,
1 and 3 m respectively) and a 13X molecular sieve column and two TCD
detectors for the analysis of permanent gases. Ar was used as internal
standard for quantification of gaseous products. Liquid-phase samples
were collected and analyzed in an Agilent 6890 gas chromatograph
equipped with a DB-Waxetr capillary column (15 m) and a TCD detector
using 1,3-propanediol and 2-pentanol as standards. Initial reaction rates
were determined by linear regression of analyses taken at reaction times
in the range from 0 to 60 minutes. Product selectivities are reported on a
molar basis at preset propylene carbonate conversion levels, and
independently for products arising from either the O-C-O (methanol, DME,
and carbon oxides) or the propane backbone (C3 alcohols, oligomers
thereof and hydrocarbons) "synthons" in the propylene carbonate
reactant.
High-Angle Annular Dark-Field Scanning-Transmission Electron
Microscopy (HAADF-STEM) and Energy-Dispersive X-ray Spectroscopy
(EDX) studies were performed in an spherical aberration-corrected (Cs)
dedicated STEM microscope (Hitachi HD-2700) equipped with a cold-
field emission gun and two EDAX Octane T Ultra W EDX detectors and
operated at 200 kV. Prior to observation, the reduced catalysts were
embedded in a low-viscosity resin (Spurr, Sigma-Aldrich) in a glove box.
The resin was then cured in an oven at 333 K overnight. Specimen
cross-sections with a nominal thickness of 50 nm were obtained with a
DIATOME diamond knife mounted on a Reichert Ultracut ultramicrotome
and collected on a nickel TEM grid (400 mesh) coated with a Lacey
carbon film (PLANO).
Acknowledgements
Sasol Materials (Germany) is acknowledged for the supply of the
high-purity pseudo-boehmite alumina precursor. E. Andrés (MPI-
KOFO) is thanked for contributions to the synthesis of
SmOx@Al2O3. S. Ruthe and N. Duyckaerts (MPI-KOFO) are
acknowledged for assistance with chromatographic product
quantification.
J.P.
Holgado
(ICMS-CSIC,
Spain)
is
X-ray photoelectron spectra (XPS) were collected in a customized
spectrometer equipped with a hemispherical SPECS PHOIBOS 100
analyzer using fixed transmission mode at 20 eV pass energy.
Acquisition of the spectra was performed using a non-monochromatic
dual X-ray source (MgK or AlK radiation) with an anode current of 20 mA
and a potential acceleration of 12 kV. As-calcined samples were pressed
into small discs and evacuated in a pre-chamber at 423 K and <10-7
mbar. Then, catalyst reduction was performed in a high-temperature
SPECS HPC-20 reaction cell with IR heating. In this case, the samples
were treated in a flow of 20% H2/Ar (v/v), by heating from RT to 543 K ( 3
K min-1) and holding the final temperature for 2 hours at 1 bar. After
cooling down to room temperature, the samples evacuated at <10-7 mbar
and transferred to the chamber of the spectrometer. Given the low
surface carbon content after the in situ reduction treatment, binding
energies (BE) were referred to the Al2p signal from γ-Al2O3 at 74.10 eV.
To derive surface relative atomic ratios, peak intensities were determined
after nonlinear Shirley-type background subtraction and corrected by
sensitivity factors (Scofield). Average Cu particle sizes were derived from
the experimental Cu/Al surface ratios using the Kerkhof-Moulijn model[21]
acknowledged for XPS experiments. This research received
funding from the Max Planck Institute für Kohlenforschung, and
the Fonds der Chemischen Industrie (Germany) and the
Bundesministerium für Bildung und Forschung (CAT2BIOL,
01DG17019).
Keywords: CO2 recycling • metal nanoparticles • selective
hydrogenation • metal-oxide promotion • oxide Lewis sites
References
[1]
[2]
[3]
a) S. J. Tauster, S. C. Fung, R. T. K. Baker, J. A. Horsley, Science
1981, 211, 1121-1125; b) Y. Wang, S. Van de Vyver, K. K. Sharma, Y.
Román-Leshkov, Green Chem. 2014, 16, 719-726; c) G. Prieto, F.
Schüth, J. Catal. 2015, 328, 59-71; d) G. Kumar, E. Nikolla, S. Linic, J.
W. Medlin, M. J. Janik, ACS Catal. 2018, 8, 3202-3208; e) I. Ro, J.
Resasco, P. Christopher, ACS Catal. 2018, 8, 7368-7387.
a) K.-D. Jung, A. T. Bell, J. Catal. 2000, 193, 207-223; b) M. Boronat,
P. Concepción, A. Corma, S. González, F. Illas, P. Serna, J. Am. Chem.
Soc. 2007, 129, 16230-16237; c) K.-i. Shimizu, Y. Miyamoto, A.
Satsuma, J. Catal. 2010, 270, 86-94; d) R. Albilali, M. Douthwaite, Q.
He, S. H. Taylor, Catal. Sci. & Technol. 2018, 8, 252-267; e) S.
Janampelli, S. Darbha, Energy & Fuels 2018, 32, 12630-12643.
a) Y. Zhu, Y. Zhu, G. Ding, S. Zhu, H. Zheng, Y. Li, Appl. Catal. A:
General 2013, 468, 296-304; b) F. Dong, Y. Zhu, H. Zheng, Y. Zhu, X.
Li, Y. Li, J. Mol. Catal. A: Chem.2015, 398, 140-148; c) J. Ding, T.
Popa, J. Tang, K. A. M. Gasem, M. Fan, Q. Zhong, Appl. Catal. B:
Environ. 2017, 209, 530-542; d) Q. Xu, X. Li, T. Pan, C. Yu, J. Deng, Q.
Guo, Y. Fu, Green Chem. 2016, 18, 1287-1294.
for high-surface-area supported catalysts, modified to consider
a
monolayer of the corresponding MOx oxide on the surface of the Al2O3
carrier.
Catalytic experiments
Catalytic experiments were performed in a PTFE-lined 36 ml autoclave
reactor. The pre-reduced catalyst powder (< 20 μm) was loaded into the
reactor under exclusion of air in a glove box, where the Ar overpressure
was kept constant to 3 mbar. In a typical experiment, 0.210 g (2 mmol) of
[4]
[5]
a) X. Li, J. Zhang, M. Zhang, W. Zhang, M. Zhang, H. Xie, Y. Wu, Y.
Tan, Catalysts 2019, 9; b) S. Schimpf, A. Rittermeier, X. Zhang, Z.-A. Li,
M. Spasova, M. W. E. van den Berg, M. Farle, Y. Wang, R. A. Fischer,
M. Muhler, ChemCatChem 2010, 2, 214-222.
a) L. Angelo, M. Girleanu, O. Ersen, C. Serra, K. Parkhomenko, A.-C.
Roger, Catal.Today 2016, 270, 59-67; b) G. Noh, E. Lam, J. L. Alfke, K.
Larmier, K. Searles, P. Wolf, C. Copéret, ChemSusChem 2019, 12,
968-972; c) J. Kim, B. B. Sarma, E. Andrés, N. Pfänder, P. Concepción,
G. Prieto, ACS Catal. 2019, 9, 10409-10417; d) M. Behrens, Angew.
Chem. Int. Ed. 2014, 53, 12022-12024.
anhydrous propylene carbonate (99.7%, Sigma-Aldrich),
8 ml of
anhydrous 1,4-dioxane as solvent (99.8%, Alfa-Aesar) and an amount of
catalyst corresponding to 75 μmol of Cu were added into the PTFE liner
of the autoclave. The reactor was then sealed, taken out from the glove
box and dosed with H2 (Air Liquide, 99.999%) up to a total pressure of 40
bar at room temperature. All catalytic tests were performed under
vigorous magnetic stirring (1000 rpm). The temperature in the reactor
was increased to the preset reaction temperature using an aluminum
heating jacket coupled to a heating plate (8 K min-1). After selected
reaction times, the reaction was quenched by immersing the autoclave in
an ice bath. The gas phase was then recovered into a gas bag and
analyzed in an Agilent 7890B gas chromatograph equipped with two
sampling loops: the first one connected to a Restek RTX-1 capillary
[6]
[7]
a) Z. Strassberger, A. H. Alberts, M. J. Louwerse, S. Tanase, G.
Rothenberg, Green Chem. 2013, 15, 768-774; b) X. Huang, C. Atay, T.
I. Korányi, M. D. Boot, E. J. M. Hensen, ACS Catal. 2015, 5, 7359-
7370.
a) P. P. Upare, Y. K. Hwang, J.-M. Lee, D. W. Hwang, J.-S. Chang,
ChemSusChem 2015, 8, 2345-2357; b) V. Pospelova, J. Aubrecht, O.
9
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Kikhtyanin, K. Pacultová, D. Kubička, ChemCatChem 2019, 11, 2169-
2178.
[8]
a) G. A. Olah, Angew. Chem. Int. Ed. 2005, 44, 2636-2639; b) A.
Goeppert, M. Czaun, J.-P. Jones, G. K. Surya Prakash, G. A. Olah,
Chem. Soc. Rev. 2014, 43, 7995-8048; c) G. Prieto, ChemSusChem
2017, 10, 1056-1070; d) A. Álvarez, A. Bansode, A. Urakawa, A. V.
Bavykina, T. A. Wezendonk, M. Makkee, J. Gascon, F. Kapteijn, Chem.
Rev. 2017, 117, 9804-9838.
[9]
X.-L. Du, Z. Jiang, D. S. Su, J.-Q. Wang, ChemSusChem 2016, 9, 322-
332.
[10] a) C. Bruneau, P. H. Dixneuf, J. Mol. Catal. 1992, 74, 97-107; b) M.
Honda, M. Tamura, K. Nakao, K. Suzuki, Y. Nakagawa, K. Tomishige,
ACS Catal. 2014, 4, 1893-1896; c) M. Tamura, K. Ito, M. Honda, Y.
Nakagawa, H. Sugimoto, K. Tomishige, Sci. Reports 2016, 6, 24038; d)
A. Ion, V. Parvulescu, P. Jacobs, D. de Vos, Appl. Catal. A: Gen. 2009,
363, 40-44; e) P. P. Pescarmona, M. Taherimehr, Catal. Sci. & Technol.
2012, 2, 2169-2187.
[11] E. Balaraman, C. Gunanathan, J. Zhang, L. J. W. Shimon, D. Milstein,
Nature Chem. 2011, 3, 609.
[12] a) H. Liu, Z. Huang, Z. Han, K. Ding, H. Liu, C. Xia, J. Chen, Green
Chem. 2015, 17, 4281-4290; b) C. Lian, F. Ren, Y. Liu, G. Zhao, Y. Ji,
H. Rong, W. Jia, L. Ma, H. Lu, D. Wang, Y. Li, Chem. Commun. 2015,
51, 1252-1254; c) Y. Cui, X. Chen, W.-L. Dai, RSC Adv. 2016, 6,
69530-69539; d) M. Tamura, T. Kitanaka, Y. Nakagawa, K. Tomishige,
ACS Catal. 2016, 6, 376-380; e) H. Li, Y. Cui, Q. Liu, W.-L. Dai,
ChemCatChem 2018, 10, 619-624.
[13] a) G. Prieto, P. Concepción, A. Martínez, E. Mendoza, J. Catal. 2011,
280, 274-288; b) G. Prieto, M. I. S. De Mello, P. Concepción, R.
Murciano, S. B. C. Pergher, A. Mart nez, ACS Catal. 2015, 5, 3323-
3335; c) A. M. Turek, I. E. Wachs, E. DeCanio, J. Phys. Chem. 1992,
96, 5000-5007.
[14] N. C. Jeong, J. S. Lee, E. L. Tae, Y. J. Lee, K. B. Yoon, Angew. Chem.
Int. Ed. 2008, 47, 10128-10132.
[15] a) P. Munnik, M. Wolters, A. Gabrielsson, S. D. Pollington, G.
Headdock, J. H. Bitter, P. E. de Jongh, K. P. de Jong, J. Phys. Chem. C
2011, 115, 14698-14706; b) G. Prieto, J. Zečević, H. Friedrich, K. P. de
Jong, P. E. de Jongh, Nature Mater. 2012, 12, 34; c) G. Prieto, M.
Shakeri, K. P. de Jong, P. E. de Jongh, ACS Nano 2014, 8, 2522-2531.
[16] R. Chatterjee, S. Kuld, R. van den Berg, A. Chen, W. Shen, J. M.
Christensen, A. D. Jensen, J. Sehested, Top.Catal. 2019.
[17] a) R. S. Rao, A. B. Walters, M. A. Vannice, J. Phys. Chem. B 2005, 109,
2086-2092; b) R. van den Berg, G. Prieto, G. Korpershoek, L. I. van der
Wal, A. J. van Bunningen, S. Lægsgaard-Jørgensen, P. E. de Jongh, K.
P. de Jong, Nature Commun. 2016, 7, 13057; c) A. Karelovic, G.
Galdames, J. C. Medina, C. Yévenes, Y. Barra, R. Jiménez, J. Catal.
2019, 369, 415-426.
[18] D. J. Darensbourg, A. D. Yeung, Green Chem. 2014, 16, 247-252.
[19] S. Kattel, B. Yan, Y. Yang, J. G. Chen, P. Liu, J. Am. Chem. Soc. 2016,
138, 12440-12450.
[20] D. L. Jurković, A. Pohar, V. D. B. C. Dasireddy, B. Likozar, Chem.
Eng.& Technol.2017, 40, 973-980.
[21] F. P. J. M. Kerkhof, J. A. Moulijn, J. Phys. Chem. 1979, 83, 1612-1619.
10
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Table 1. Chemical composition, textural and surface electronic properties of MOx@Al2O3 oxide supports and metal dispersion in the corresponding copper-based
catalysts.
Composition
Textural properties[a]
cus Lewis acidity[b]
d(Cu)[i]
Sample
(nm)
[e]
[f]
[g]
M[c]
(wt%)
δ[d]
SBET
(m2 gAl2O3
Vp
dp
η[h]
(eV)
(at nm-2)
)
(cm3 gAl2O3
)
(nm)
-1
-1
Al2O3
---
---
265
0.80
11.5
2.50
2.45
2.38
2.40
2.48
2.56
6.7
ScOx@Al2O3
SmOx@Al2O3
YOx@Al2O3
ZrOx@Al2O3
TaOx@Al2O3
9.4
4.5
4.9
4.6
4.7
4.5
255
265
270
285
310
0.65
0.71
0.78
0.67
0.71
9.4
9.6
9.9
8.6
8.4
4.6
23.8
14.5
15.4
24.1
16.0
10.5
8.9
4.3
[a] As determined by N2 physisorption. [b] Relative Lewis acidity of surface-exposed coordinatively unsaturated metal sites (cus), determined by UV-vis
spectroscopy with 1,2-dihydroxyanthraquinone as surface probe molecule. [c] Metal loading corresponding to the MOx overlay oxide. [d] Surface-specific M
loading. [e] B.E.T. specific surface area, normalized per unit mass of Al2O3. [f] Total pore volume, normalized per unit mass of Al2O3. [g] B.J.H. average pore
diameter. [h] Peaking energy for the lowest-energy intramolecular charge-transfer (IMCT) of adsorbed 1,2-dihydroxyanthraquinone as determined by UV-vis
spectroscopy after surface-saturation with the probe molecule at room temperature. The experimental error for this parameter was determined to be <0.01 eV
from independent repetitions. [i] Average copper nanoparticle size, as determined by XPS after catalyst reduction, for Cu/MOx@Al2O3 with a copper content of 1.5
Cu nm-2.
11
This article is protected by copyright. All rights reserved.

10.1002/cssc.202000166
ChemSusChem
FULL PAPER
Entry for the Table of Contents
Oxide is a pillar of support. The relative electron-withdrawing character of coordinatively unsaturated Lewis centers on the oxide
support dials promotion effects in oxide-supported copper nanoparticles as solid catalysts for the selective hydrogenation of
propylene carbonate. An intermediate oxide Lewis acidity maximizes the selectivity to methanol while preserving the propanediol
molecular backbone, as required for cyclic processes of CO2 recycling to methanol with green hydrogen mediated by condensed
carbonate intermediates.
12
This article is protected by copyright. All rights reserved.