CO2 Hydrogenation on Cu/Al2O3: Role of the Metal/Support Interface in Driving Activity and Selectivity of a Bifunctional Catalyst

Article abstract of DOI:10.1002/anie.201908060

Selective hydrogenation of CO₂ into methanol is a key sustainable technology, where Cu/Al₂O₃ prepared by surface organometallic chemistry displays high activity towards CO₂ hydrogenation compared to Cu/SiO₂, yielding CH₃OH, dimethyl ether (DME), and CO. CH₃OH formation rate increases due to the metal–oxide interface and involves formate intermediates according to advanced spectroscopy and DFT calculations. Al₂O₃ promotes the subsequent conversion of CH₃OH to DME, showing bifunctional catalysis, but also increases the rate of CO formation. The latter takes place 1) directly by activation of CO₂ at the metal–oxide interface, and 2) indirectly by the conversion of formate surface species and CH₃OH to methyl formate, which is further decomposed into CH₃OH and CO. This study shows how Al₂O₃, a Lewis acidic and non-reducible support, can promote CO₂ hydrogenation by enabling multiple competitive reaction pathways on the oxide and metal–oxide interface.

Full text of DOI:10.1002/anie.201908060

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: CO2 Hydrogenation on Cu/Al2O3: Role of Metal/Support
Interface in Driving Activity and Selectivity of a Bifunctional
Catalyst
Authors: Christophe Copéret, Erwin Lam, Juan José Corral-Pérez,
Kim Larmier, Gina Noh, Patrick Wolf, Aleix Comas-Vives, and
Atsushi Urakawa
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: Angew. Chem. Int. Ed. 10.1002/anie.201908060
Angew. Chem. 10.1002/ange.201908060
Link to VoR: http://dx.doi.org/10.1002/anie.201908060
http://dx.doi.org/10.1002/ange.201908060

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
CO2 Hydrogenation on Cu/Al2O3: Role of Metal/Support Interface
in Driving Activity and Selectivity of a Bifunctional Catalyst
Erwin Lam[a], Juan José Corral-Pérez[b], Kim Larmier[a], Gina Noh[a], Patrick Wolf[a], Aleix Comas-
Vives[a,d], Atsushi Urakawa[b,c], Christophe Copéret*,[a]
Abstract: Selective hydrogenation of CO2 into methanol is a key
sustainable technology. Here, we show that Cu/Al2O3 prepared via
Surface Organometallic Chemistry, displays particularly high activity
towards CO2 hydrogenation compared to Cu/SiO2, yielding CH3OH,
dimethyl ether (DME) as well as CO. Then, by combining advanced
operando spectroscopy and DFT calculation, we show that CH3OH
formation involves formate intermediates and that the increase of rate
originates from the metal-oxide interface. Al2O3 also promotes the
subsequent conversion of CH3OH to DME, evidencing bifunctional
catalysis, but also increases the rate of CO formation. The latter takes
place i) directly by activation of CO2 at the metal-oxide interface, and
ii) indirectly via the conversion of formate surface species and CH3OH
to methyl formate that is further decomposed into CH3OH and CO.
This study showcases how Al2O3, a Lewis acidic and non-reducible
support, can promote CO2 hydrogenation by enabling multiple
competitive reaction pathways on the oxide and metal-oxide interface.
competitive reverse-water-gas-shift (RWGS) reaction that
transforms CO2 into CO in addition to the thermodynamic limits of
the reaction (~20% conversion at 230 °C and 30 bar). Among Cu-
based catalysts, Cu/ZnO/Al2O3 and Cu/ZrO2 have been shown to
be effective in the selective hydrogenation of CO2 to CH3OH.[1, 4-
11]
For Cu/ZrO2, the selective CO2 hydrogenation to CH3OH is
ascribed to Lewis acidic Zr surface sites, located at the periphery
of Cu nanoparticles that facilitate the formation of formate and
methoxy reaction intermediates.[12-13] For comparison, Cu/SiO2
that does not contain Lewis acidic sites shows significantly lower
activity and selectivity favoring CO formation. In fact, introducing
surface Zr(IV) or Ti(IV) sites on SiO2 at the interface with Cu
nanoparticles via Surface Organometallic Chemistry (SOMC)
increases the CH3OH activity and selectivity, further supporting
the importance of Lewis acid sites in this reaction.[13-14] Regarding
CO, it has mainly been proposed to be formed on copper-based
catalysts via two different mechanisms: The so-called redox and
formate pathways. The former consists of the direct conversion of
CO2 to CO (that may involve M–COOH intermediates) and the
subsequent reduction of Cu2O surface species to Cu(0) metal
forming H2O, while the latter involves the decomposition of
formate into CO.[9, 15-27] Additionally, the formation of CO but also
CH3OH is often proposed to be related to the presence of oxygen
defect sites originating from the oxide support.[28-36]
For Cu/ZnO/Al2O3, the origin of the promotional effect towards
selective CH3OH formation is still a matter of debate. For instance,
CuZn alloy or the Cu/ZnO interface has been proposed to
generate the active site responsible for the high CH3OH
selectivity.[37-44] In contrast, Al2O3, a minor component of the
catalyst, is typically viewed as a stabilizing oxide that maintains
the dispersion of Cu and Zn. However, considering the high
reactivity of alumina toward many small molecules, including
CH3OH (vide infra),[45-48] e.g. the dehydration of CH3OH to
dimethyl ether (DME),[46] one may wonder about the involvement
of alumina during the hydrogenation of CO2. In fact, studies on
CO2 hydrogenation catalysts based on alumina-supported Cu
nanoparticles are rare, and the influence of CH3OH or the effect
of the Lewis acidic support on activity/product selectivity are often
overlooked.[15, 49-55]
Introduction
The hydrogenation of CO2 into CH3OH has been identified as a
very promising route to mitigate the increasing anthropogenic CO2
content in the atmosphere by incorporating CO2 into a carbon-
cycle, in the so-called methanol economy, provided that i)
hydrogen is derived from green technology, e.g. water electrolysis
using excess intermittent energy, and ii) efficient CO2 capture
technologies are available.[1-3] In that regard, copper-based
catalysts are good candidates in terms of activity and selectivity
for transforming CO2 to CH3OH.[1] However, the efficiency of the
CO2-to-CH3OH synthesis can be significantly decreased by the
[a]
E. Lam, Dr. K. Larmier, Dr. G. Noh, Dr. P. Wolf, Dr. A. Comas-Vives,
Prof. Dr. C. Copéret
Department of Chemistry and Applied Biosciences
ETH Zürich, Vladimir Prelog Weg 1-5, 8093 Zürich, Switzerland
E-mail: ccoperet@ethz.ch
[b]
[c]
J.J. Corral-Pérez, Prof. Dr. A. Urakawa
Institute of Chemical Research of Catalonia (ICIQ), The Barcelona
Institute of Science and Technology, 43007, Tarragona, Spain
Current address:
Catalysis Engineering, Department of Chemical Engineering, Delft
University of Technology, Van der Maasweg 9, 2629 HZ Delft, The
Netherlands
In order to understand the role of alumina, we reasoned that
studying alumina-supported Cu nanoparticles in the absence of
zinc oxide would be ideal. We thus prepared small and narrowly
dispersed Cu nanoparticles on Al2O3 via SOMC.[56-57] Using
detailed catalytic studies augmented by ex-situ solid-state NMR,
operando transient diffuse reflectance infrared spectroscopy
[d]
Current address:
Department of Chemistry, Universitat Autonoma de Barcelona,
08193 Cerdanyola del Vallèes, Catalonia, Spain
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
Cu2.4wt%/Al2O3
(DRIFTS) and DFT modelling, we demonstrate that Cu/Al2O3 is a
highly effective CO2 hydrogenation catalyst in forming CH3OH that
is further converted into dimethyl ether (DME). We also show
however that Cu/Al2O3 catalyzes the conversion of CO2 to CO,
and the formation and decomposition of methyl formate into CO,
thus providing key information about the role of alumina in
determining the selectivity in CO2 hydrogenation.
3.1 ± 0.6 nm
Results and Discussion
Cu3.8wt%/Al2O3
Catalyst Synthesis and Characterization
3.1 ± 0.6 nm
Copper nanoparticles supported on alumina are prepared via
SOMC[56] that involves grafting [Cun(Mesityl)n](n=2,4,5) in toluene
on -alumina dehydroxylated at 500 °C (Al2O3-500, 2.1 OH nm-2),
followed by a reduction step under H2 at 500 °C, yielding
CuXwt/Al2O3 where X specifies the Cu weight loading (Figure 1a).
The copper loading is controlled by grafting a specific amount of
[Cun(Mesityl)n] (n=2,4,5) per nm2 yielding catalysts with 2.4 and
3.8 wt% copper, referred to as Cu2.4wt/Al2O3 and Cu3.8wt/Al2O3,
respectively (Table 1). After the grafting step, the IR spectrum
shows the consumption of the surface Al-OH groups (bands
above 3500 cm-1) and the appearance of C-H stretching bands at
around 3000 cm-1 (Figure 1b and Figure S1 for Cu3.8wt/Al2O3).
Subsequent treatment under H2 at 500 °C restores the Al-OH
groups while all C-H stretching bands disappear (Figure 1b),
consistent with the removal of all organic ligands.
Figure 2. HAADF-STEM and particle size distribution (particle counts > 150) of
CuXwt/Al2O3 catalysts.
Dark-field scanning transmission electron microscopy indicates
the formation of Cu nanoparticles in both samples with sizes of
3.1 ± 0.6 nm for Cu2.4wt%/Al2O3 and Cu3.8wt%/Al2O3, respectively
(Figure 2). The dispersal of Cu particles takes place without
change in specific surface area that is ca. 140 m2 g-1 according to
N2 adsorption/desorption isotherms (77 K) for both samples. In
addition, powder X-ray diffraction shows only the crystalline phase
of -Al2O3 and the absence of crystalline copper, consistent with
the presence of small nanoparticles (Figure S2). N2O
chemisorption shows that the copper surface site concentration is
40 mol gcat-1 and 67 mol gcat-1 for Cu2.4wt/Al2O3 and Cu3.8wt/Al2O3,
respectively (Table 1), consistent with a similar particle size
distribution by microscopy and dispersion by H2 chemisorption (62
(a)
Mesitylene
CuyMesy-1
O
+
OH
OH
O
O
O
O
O
O
[Cun(Mes)n
(n=2,4,5)
]
H2
Al2O3
Al2O3
Al2O3
Toluene, 4h
500 °C, 12h
Al2O3-500
CuyMesy-1/Al2O3-500
Cu2.4wt%/Al2O3-500
(b)
-1
-1
mol gcat and 86 mol gcat H2 uptake for Cu2.4wt%/Al2O3 and
Cu3.8wt%/Al2O3, respectively – Figure S3), where a higher H2
uptake is observed compared to N2O. Overall using the
methodology of grafting an organometallic precursor followed by
reduction under H2 leads to the formation of Cu nanoparticles with
similar dispersions, for both Cu loadings.
Al2O3-500
CuyMesy-1/Al2O3-500
Cu2.4wt%/Al2O3-500
Figure 1. (a) Reaction scheme for the synthesis of copper nanoparticles on
Al2O3-500 with a Cu loading of 2.4 wt% with Mes representing the mesitylene
ligand. (b) IR spectra of grafting and reduction of [Cun(Mesityl)n] (n=2,4,5) on
Al2O3-500
.
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
Table 1. Physicochemical properties of copper-based catalysts.
Cu surface area
Cu surface area
Cu content[a]
[wt%]
Specific surface area[b]
Particle size
distribution[e] [nm]
Entry
Support
based on N2O[c] [mol
based on H2[d] [mol
[m2 g-1]
-1
-1
gcat
]
gcat
]
[f]
1
2
3
Cu/SiO2
4.19
2.40
3.81
197
141
139
50/34
45/40
66/67
NA
2.9 ± 1.3
3.1 ± 0.6
3.1 ± 0.6
Cu2.4wt%/Al2O3
Cu3.8wt%/Al2O3
62
86
[a] Determined by ICP-OES; [b] Determined from N2 physisorption applying the BET theory; [c] Measured on fresh/spent catalyst, assuming
a 1:2 stoichiometry between N2O and Cu surface sites; [d] Measured on fresh catalyst, assuming a 1:2 stoichiometry between H2 and Cu surface
sites [e] Determined by TEM on the fresh catalyst. [f] Reference [13].
Catalytic Studies
involvement of Lewis acidic sites and their inhibition by the
coordination of reaction intermediates/products (Figure S10-
S14). To further investigate the role (promotional effect) of
Al2O3 on the observed selectivity patterns, in particular with
respect to the increase of CO selectivity, not observed for
Cu/ZrO2, CO2 hydrogenation was also investigated with
physical mixtures of Cu/SiO2 with varying amounts of Al2O3,
while keeping the amount of copper constant. We note that
Al2O3 in the absence of Cu does not shows any catalytic activity
in CO2 hydrogenation at 230 °C. The CO2 hydrogenation
carried out with a physical mixture consisting of a 5:1 (w/w)
ratio of Cu/SiO2:Al2O3 shows a similar formation rate for
hydrogenated products (CH3OH) as observed for Cu/SiO2 but
includes the formation of DME. In addition, the formation rate
The thus-prepared catalysts were then evaluated in CO2
hydrogenation in a fixed-bed flow reactor at 230 °C and 25 bar
producing CH3OH, CO and DME as the main products (Figure
2). Since DME is likely formed from the secondary reaction of
CH3OH on Al2O3 (vide infra),[46] we compare catalysts based
on the combined formation rate of CH3OH and DME (1:2
DME:CH3OH stoichiometry), expressed in g CH3OH gCu h-1.
-1
The catalysts are compared to the Cu/SiO2 benchmark
catalysts prepared via SOMC that also contain similar particles
sizes according to TEM (Table 1).[13] The formation rates are
given by extrapolation to 0 s gcat ml-1 contact time (Figure S4-
8) from catalytic results reaching up to 7% conversion (Figure
S9). The equilibrium conversion under the given reaction
conditions is around 17% including the dehydration of CH3OH
to DME and 15% without. For Cu2.4wt%/Al2O3 the intrinsic
-1
of CO is ca. doubled (0.49 g gCu h-1) by comparison with
Cu/SiO2, leading to an overall CO selectivity increase.
Increasing the ratio of Al2O3 with Cu/SiO2 to 1:1 (w/w) further
-
CH3OH formation rate is around three times greater (0.89 g gCu
1
-1
-1
h-1) than with Cu/SiO2 (0.27 g gCu h-1), a non-Lewis acidic
increases the CO formation rate marginally (0.53 g gCu h-1),
support. This increase in CH3OH formation rate is paralleled by
an increased intrinsic CO formation rate, of ca. four times (1.04
implying that a larger amount of Al2O3 does not further affect
the rate determining step for CO formation. The overall CH3OH
formation rate remains similar as found for Cu/SiO2 with a rate
-1
g gCu h-1) for Cu2.4wt%/Al2O3 compared to Cu/SiO2. Increasing
-1
the Cu loading to 3.8 % (Cu3.8wt%/Al2O3) has hardly any effect
on the intrinsic CH3OH formation rate, while the CO formation
of ca. 0.25 g gCu h-1 and a higher ratio for DME vs. CH3OH
(Figure 2). The same trends are observed when formation
rates at higher conversions (1% and 2%) are considered, with
the only observation being the decrease in formation rates for
all the products due to product inhibition (Figure S15). All the
data are also recorded at conversions that are far away from
equilibrium, with the catalytic reaction being kinetically
controlled. Thus, Al2O3 is responsible for the increased
formation rate of CO in the physical mixtures between Cu/SiO2
and Al2O3, while it promotes both CO and CH3OH formation
rates in Cu/Al2O3 pointing out the role of interfacial sites in this
reaction. In fact, for CuXwt/Al2O3 (X = 2.4 and 3.8) samples
(Figure 2), the CH3OH formation rates are similar when
normalized by the amount of Cu due to having similar Cu
dispersions for both samples and the formation rate of CO
decreases with higher Cu loading but is proportional to the
mass of the catalyst (Figure S16).
-1
rate decreases (0.73 g gCu h-1) (Figure 2). Overall, Cu/Al2O3
catalysts show higher formation rates for both CH3OH and CO
compared to Cu/SiO2 while generating large amounts of DME
that is not observed for Cu/SiO2 as a result from the
subsequent dehydration of CH3OH on alumina.[46] While DME
appears as a primary product as indicated by an intrinsic
formation rate greater than zero; this is most likely due to strong
adsorption of CH3OH allowing the further dehydration to DME
prior to desorption (vide infra). As proposed for Cu/ZrO2 and
related systems,[12-13] Lewis acidic supports likely promote the
formation of CH3OH, but stronger Lewis acids such as those
found on Al2O3 are also able to further catalyze the dehydration
of CH3OH to DME.[46] Interestingly, Cu/Al2O3 also promotes CO
formation, which is not observed for Cu/ZrO2.[12] The CO and
CH3OH (including DME) formation rate decreases at higher
conversions, i.e., longer contact time, consistent with the
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
CO
CH3OH
DME
CH3OH + DME
Cu/SiO2
Al2O3
Cu/SiO2
Al2O3
Cu/SiO2 Cu2.4wt%/Al2O3 Cu3.8wt%/Al2O3
(5:1 w/w)
(1:1 w/w)
Figure 2. Intrinsic formation rates for Cu/SiO2 and Cu/Al2O3 as well as physical mixtures of Cu/SiO2 and Al2O3 for the formation of CO (blue), CH3OH (red), DME
(black) and overall CH3OH formation (CH3OH + DME red-dashed).
Since Al2O3 itself is not active for the RWGS reaction, these
data also suggest that Al2O3 is able to mediate the formation of
CO from products resulting from the hydrogenation of CO2 on
Cu/SiO2. Plotting the ratio of CH3OH with DME formation rates
shows a clear correlation between the products, expected from
DME originating from the dehydration of CH3OH (Figure S17).
The ratio between CH3OH and CO formation rates normalized
by both the Cu loading and mass of catalyst does not show a
clear correlation amongst different catalysts (Figure S18-S19),
thus hinting to multiple reaction pathways for CO formation.
transmission IR, where upon adsorption of 60 mbar of CH3OH
on Al2O3 and heating to 200 °C; the presence of H2O and
CH3OH can be observed, but not adsorbed DME as evidenced
by the absence of signals at around 1160 cm-1 (Figure S21)
which is only observed in the gas phase by NMR upon
condensation of the gas phase in C6D6 (Figure S22), showing
that DME is more weakly adsorbed on the catalyst surface than
other reaction intermediates and products such as CH3OH or
H2O.[46] This is consistent with the strong adsorption of CH3OH
and the appearance of DME as a primary product generated
by subsequent dehydration of CH3OH.
Mechanism for the Formation of Products and Byproducts
– DME and CO – on Al2O3
In order to identify possible reaction intermediates formed on
Cu2.4wt%/Al2O3, we used solid state NMR (Magic Angle Spinning
(MAS) 1H- and 13C-ssNMR as well as 1H-13C HETCOR ssNMR
after reaction with 5 bar of 13CO2:1H2 (1:3 ratio) at 230 °C for
12 hours in a batch reactor (for details see ESI), followed by
evacuation of the gas phase at room temperature under high
vacuum (10-5 mbar). The 1H-13C HETCOR spectrum shows the
presence of two correlation peaks at 8/170 ppm and 2.9/48
ppm (Figure 3a), consistent with the formation of formate and
methoxy species on the oxide support as previously observed
on Cu/ZrO2.[12-13] No adsorbed DME (expected at around 60
ppm in 13C-ssNMR) is observed (Figure S20), while both DME
and CH3OH are observed in the gas phase by 13C solution
NMR following the condensation of the gas phase after
reaction in deuterated benzene (C6D6) (Figure 3b). The
formation and desorption of DME is in fact also evidenced by
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
(a)
reaction of CH3OH with surface formate species (denoted as
HCOO*) generating a surface hydroxyl group (OH*), could
subsequently decompose to form CH3OH and CO (eq. 5). In
fact, surface formate species are observed by solid-state NMR
upon reaction of the Cu/Al2O3 or Al2O3 with 5 bar of 13CO2:H2
(1:3) at 230 °C for 12 hours (Figure 3a and Figure S23-S24).
The formation of methyl formate is known to occur during CO2
hydrogenation at high pressures (300 bar) and can be further
promoted by cofeeding CH3OH,[59] but its decomposition is
likely hindered at high pressures (300 bars).[49]
2
H
13C
H
13C
1
H
O
O
O
H
Al
Al
Al
Al
O
O
Scheme 1. Possible reaction pathways for the formation of CO.
1
2
CH3OCH3
CH3OH
CO + CH4 + H2
CO + 2 H2
(1)
13C Chemical Shift [ppm]
(2)
(b)
13CO2
C6D6
2 CH3OH
HCOOCH3
HCOOCH3 + 2 H2
2 CO + 2 H2
CO + H2O
(3a)
(3b)
(4)
13CH3O13CH3
13CH3OH
HCOOH
CH3OH
+ HCOO*
- OH*
HCOOCH3
(5a)
13C Chemical Shift [ppm]
HCOOCH3
CH3OH + CO
(5b)
Figure 3. (a) Ex-situ MAS-NMR 1H -13C HETCOR spectra of Cu2.4wt%/Al2O3
Reaction of methyl formate with Al2O3 at 230 °C also yields
CH3OH, DME and CO, consistent with the decomposition of
methyl formate into CH3OH and CO at 230 °C with the
concomitant dehydration of CH3OH into DME under reaction
conditions (Figure S25). Furthermore, methyl formate is also
shown to decompose to CO and CH3OH on Cu/Al2O3 (Figure
S26).
reacted with CO2:H2 (1:3) at 5 bars for 12 hours and 230 °C. (b) Solution 13
NMR of the gas phase of Cu2.4wt%/Al2O3 after reaction recorded in C6D6.
C
CO can originate from several reaction pathways mediated by
Al2O3. Among possible explanations, Al2O3 contains highly
reactive surface sites able to activate small molecules that can
mediate the formation of CO through the decomposition of
DME/CH3OH or the formation and the subsequent
decomposition of formic acid/methyl formate (Scheme 1).[47, 58]
The decomposition of DME into CO requires concomitant
formation of CH4 (eq. 1),[47] a product that is only observed in
trace amounts under the investigated reaction conditions.
Alternatively, the direct decomposition of CH3OH to CO (eq. 2)
or the formation of methyl formate from two equivalents of
CH3OH followed by decomposition to CO (eq. 3) are not likely
either, due to a higher CO2 conversion for the Cu/SiO2 and
Al2O3 physical mixtures. Similarly, the proposed decomposition
of CH3OH and methyl formate according to eq. 2 and 3 would
lead to the formation of H2, which is thermodynamically less
likely considering the low conversion and since the use of high
pressure of H2 (15 bar) disfavor a reaction that generates H2.
CO could also arise from the decomposition of formic acid that
can potentially be formed on Cu/SiO2 and subsequently
decomposes on Al2O3 to CO and H2O (eq. 4) .[21] It is for
example known that formic acid can decompose to CO and
H2O on Al2O3 or SiO2.[21] Finally, methyl formate formed upon
In order to understand additional reaction pathways besides
the decomposition of formic acid forming CO and H2O on Al2O3
based catalyst,[15-16, 21] the effect of CH3OH and its interaction
with surface species for the formation of CO is further
investigated. Thus, Cu3.8wt%/Al2O3 was studied by transient
operando diffuse reflectance infrared spectroscopy (DRIFTS)
with online activity monitoring to identify correlation between
the gas phase composition with the concentrations of adsorbed
intermediates. The changes in the IR spectrum and the gas
phase composition are monitored at 230 °C and 5 bar under a
transient condition, i.e. switching the atmosphere between a
mixture of CO2 and H2 and a mixture of CO2, H2 and CH3OH
(vapor) (Figure 4a). In the latter atmosphere, concentrations of
CH3OH, DME, CO and methyl formate increased (Figure 4b
and S27), confirming that the formation of DME, CO and methyl
formate are largely influenced by the concentration of CH3OH.
The IR spectrum in the region of 1000-1900 cm-1 shows two
formate species that are behaving kinetically different (at ca.
1640 and 1670 cm-1) based on multivariate curve resolution
analysis.[60] The two formate species are proposed to be Al2O3
bound formates[16] and are quickly decreasing in intensity when
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
CH3OH is present (Figure 4c and S28), in line with the rapid
formation of methyl formate. The evolution of adsorbed
methoxy species between 1000 and 1900 cm-1 is also
observed (Figure S28). In addition, when switching between
CO2/H2 and Ar, the concentration of the formate species at ca.
1640 cm-1 is decreasing with a slower rate (Figure 4d blue),
indicating that CH3OH can indeed facilitate the
removal/reaction of this surface formate species. This is further
supported by a similar profile of the IR signals observed
between 2500 and 3100 cm-1 corresponding to formate species
(Figure S29-S30). In addition the evolution of methoxy species
are shown in the region between 2500 and 3100 cm-1: with a
decrease in concentration of methoxy surface species when
switching between CO2/H2/CH3OH and CO2/H2 (Figure S29) vs.
an increase in concentration of methoxy surface species when
switching between Ar and CO2/H2 (Figure S30). Overall, it
shows the possible formation of methyl formate from CH3OH
that could then further decompose to CO on Cu3.8wt%/Al2O3.
(a)
(b)
(c)
(d)
CO2/H2/Ar vs. CO2/H2/Ar/CH3OH
CO2/H2/Ar vs. Ar
Figure 4. (a) Evolution of surface species between 1000 and 1900 cm-1 during transient operando DRIFTS experiments at 5 bar and 230 °C switching between
CO2/H2/Ar and CO2/H2/Ar/CH3OH for Cu3.8%/A2O3. (b) Gas phase concentration of CH3OH, methyl formate and CO between gas mixtures of CO2/H2/Ar and
CO2/H2/Ar/CH3OH. (c) (left) Concentration profile of two surface species (wavenumber 1000-1900 cm-1) corresponding to (right) component spectra in blue and
black obtained by MCR analysis when switching between gas mixtures of CO2/H2/Ar (green) and CO2/H2/Ar/CH3OH (blue). (d) (left) Concentration profile of two
surface species (wavenumber 1000-1900 cm-1) corresponding to (right) component spectra in blue and black obtained by multivariate spectra analysis when
switching between gas mixtures of CO2/H2/Ar (green) and Ar (blue).
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
Role of the Metal/Oxide Interface on CH3OH and CO
Formation
(a)
In order to obtain further information about the reaction
mechanism, possible active site structures, and the role of
perimeter sites around the Cu nanoparticle, we investigated
potential reaction pathways via a computational approach. The
Cu/Al2O3 catalyst is modeled by placing a Cu67 particle
(corresponding to ca. 1 nm diameter) on the most abundant
(110)-facet of -Al2O3 (Figure 5a). This facet exposes different
aluminum sites with different coordination number that can be
found at the periphery (interface) between Cu and Al2O3. Due
to the Cu nanoparticle size, the interface between the metal
and the support can have various interfacial sites that are
further investigated by DFT calculations. They are denoted
Al(III), Al(IVA) and Al(IVB), the number indicating the
coordination number and the letter describing two types of four-
coordinated sites (Figure 5a).[45, 61] First, CO2 can adsorb on
these interfacial sites to generate bent monodentate (Figure 5b
– Al(III)) or bridging (μ2) CO2 species by involving two
neighboring aluminum sites (Figure 5b - Al(III)-Al(IVB) and
Al(IVA)-Al(IVA)) with CO2 adsorption free energy (calculated
at 500 K) ranging from rG = –59 to 63 kJ mol-1 (Figure 5b and
Table S1), while CO2 does not interact with the Cu(111)
surface and remains linear (rG = +46 kJ mol-1). This adsorbed
CO2 at the interface between Cu and Al2O3 can then further
react with surface Cu hydrides generated by splitting H2 on the
Cu particle forming CH3OH or CO (Figure 5c, Figure S31-S32
and Table S2), similar to the pathway proposed by DFT
calculation for Cu/ZrO2.[12] CO2 is hydrogenated to formate
(HCOO*), acetal (CH2OO*) and (CH2OOH*) upon which the C-
(b)
Cu(111)
46 kJ mol-1
Al(III)
-59 kJ mol-1
Al(III)-Al(IVB)
-24 kJ mol-1
Al(IVA)-Al(IVA)
49 kJ mol-1
DrG =
Al(III)
Al(IVA)-Al(IVA)
TS(CO2* -> CO* + O*)
DrG‡ = 7 kJ mol-1 DrG = -104 kJ mol-1 DrG‡ =138 kJ mol-1
TS(CO2* + H* -> HCOO*)
CO* + O*
HCOO*
DrG = 18 kJ mol-1
(c)
H2 split on Cu
H2
2 H*
MeOH synthesis (formate pathway)
H*
CO2*
2 H*
H*
2 H*
CH3OH + H2O
HCOO*
H2CO* + OH*
CH3O* + OH*
CO2
2 H*
CO + H2O
CO* + O*
Reverse water-gas shift
O
bond is broken forming CH2O* and OH*. Further
Figure 5. (a) Top (left) and side (right) view of the computational model
consisting of the (110) facet of -Al2O3 and a Cu67 particle with the aluminum
interfacial sites circled for Al(III) (black), Al(III)-Al(IVB) (green) and Al(IVA)-
Al(IAV) (blue). (b) Structure and adsorption free energy (rG) calculated at
500 K in kJ mol-1 of CO2 interacting with the Cu(111) facet and the interfacial
sites with respect to gas phase CO2. Al(III) site (black circle), an Al(III) site
in close proximity to an Al(IVB) site (green circle), and two Al(IVA) sites in
close proximity (blue circle). Transition state for the formation of CO* + O*
and HCOO* on Al(III) and Al(IV)-Al(IV), respectively. (c) Reaction
mechanism for the formation of Cu-H, CH3OH via the formate pathway, and
CO.
hydrogenation yields methoxy (CH3O*) and hydroxide (OH*)
(Figure 5c) and finally CH3OH and H2O that can desorb from
the catalyst, closing the catalytic cycle. Alternatively, the CO2
adsorbed at the interface of the Cu nanoparticle and Al2O3 can
directly split into CO* and O* (Figure 5b). The O* is located at
the interface between Cu and Al2O3 while CO coordinated on
Cu that desorbs. Further hydrogenation yields H2O that
desorbs from the catalyst, similar to what was shown on
Ni/Al2O3, where the interfacial sites are playing a key role
during the RWGS reaction.[62-63]
Overall, sites where CO2 is coordinated in a monodentate
geometry as on Al(III) appears to favor CO formation, whereas
CO2 adopting a bridging (μ2) coordination favors the formation
of HCOO* and ultimately CH3OH on Al(IVA)-Al(IVA). Thus,
depending on the nature of the aluminum site it can either favor
CH3OH or CO. This can be explained by the geometry of CO2
(Figure 5b). Bonding in a monodentate geometry allows to
easily break the C-O bond leading to the RWGS reaction being
favored, while having a bridging (μ2) coordination CO2* consist
of a geometry that resembles this of formate species (HCOO*),
hence the low activation barrier for its formation. It shows that
CO2 hydrogenation at the Cu/Al2O3 interface lowers the
activation energy for both CH3OH and CO in comparison to the
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
Cu(111) surface (for further details see ESI). This is consistent
with the increase in CH3OH and CO formation rate found
experimentally and in line with the inhibition for both the CH3OH
and CO formation at longer contact times due to coordination
on Lewis acid sites by reaction products (H2O/CH3OH) that
have a stronger adsorption energy than CO2. The formation of
DME is shown not to require the presence of Cu and can solely
take place on the Al2O3 surface, where by DFT calculation,
DME is preferentially formed on more strongly Lewis acidic
sites (Scheme S1). In addition, the most strongly adsorbed
reaction intermediates determined by DFT calculation (HCOO*
and CH3O*) are the same intermediates observed with ex situ
solid state NMR experiments (vide supra).
NMR measurements, Dr. Peter Chen (ETH Zurich) for HAADF-
STEM and Dr. Lucas Foppa (ETH Zurich) for helpful
discussions.
Keywords: Computational Chemistry • Hydrogenation •
Metal/Support Interface • Nanoparticles • Operando
Spectroscopy
[1]
A. Álvarez, A. Bansode, A. Urakawa, A. V. Bavykina, T. A.
Wezendonk, M. Makkee, J. Gascon, F. Kapteijn, Chem. Rev. 2017,
117, 9804-9838.
[2]
[3]
G. A. Olah, A. Goeppert, G. K. S. Prakash, Beyond Oil and Gas: The
Methanol Economy, 2nd ed., Wiley-VCH: Weinheim, Germany, 2011.
G. A. Olah, Angew. Chem., Int. Ed. 2013, 52, 104-107, Angew. Chem.
2013, 125, 112-116.
[4]
[5]
K. T. Jung, A. T. Bell, Catal. Lett. 2002, 80, 63-68.
J. Weigel, R. A. Koeppel, A. Baiker, A. Wokaun, Langmuir 1996, 12,
5319-5329.
Conclusion
Cu nanoparticles (2-4 nm) supported on Al2O3 is a bifunctional
catalyst that is particularly active toward the catalytic
hydrogenation of CO2 into CH3OH, DME and CO, in contrast to
what is observed with SiO2 or ZrO2 supports that show low
activity and selectivity or promote only CH3OH formation rate,
respectively. The formation of CH3OH and CO as well as the
further dehydration of CH3OH to DME is due to the strongly
Lewis acidic Al2O3 support that promotes both of these
reactions. The observed promotional effect is traced back to
specific interfacial sites present between Cu and Al2O3, that
favors CH3OH formation via formate intermediates, but also the
direct conversion of CO2 to CO. In addition, these Lewis acid
sites open additional reaction pathways, converting surface
formate intermediates upon reaction with CH3OH to methyl
formate that can then further decompose into CO. This study
thus showcases how the Al2O3 support can participate in the
formation of a broad range of products, increasing both CH3OH
and CO production rates and also helps rationalizing the effect
of Lewis acidic oxide supports on the catalytic properties of Cu
nanoparticles in CO2 hydrogenation. Such information provides
guidelines on ways to tune the catalytic performances of CO2
hydrogenation catalysts that we are currently exploring.
[6]
A. Baiker, M. Kilo, M. Maciejewski, S. Menzi, A. Wokaun, Stud. Surf.
Sci. Catal., 1993, 75, 1257-1272.
[7]
[8]
C. Schild, A. Wokaun, A. Baiker, J. Mol. Catal. 1990, 63, 243-254.
S. Tada, K. Larmier, R. Buchel, C. Coperet, Catal. Sci. Technol. 2018,
8, 2056-2060.
[9]
Y. Amenomiya, J. Catal. 1979, 57, 64-71.
[10] E. Ramaroson, R. Kieffer, A. Kiennemenn, Appl. Catal. 1982, 4, 281-
286.
[11] J. Schumann, T. Lunkenbein, A. Tarasov, N. Thomas, R. Schlögl, M.
Behrens, ChemCatChem 2014, 6, 2889-2897.
[12] K. Larmier, W. C. Liao, S. Tada, E. Lam, R. Verel, A. Bansode, A.
Urakawa, A. Comas‐Vives, C. Copéret, Angew. Chem., Int. Ed.
2017, 56, 2318-2323, Angew. Chem. 2017, 129, 2358-2363.
[13] E. Lam, K. Larmier, P. Wolf, S. Tada, O. V. Safonova, C. Copéret, J.
Am. Chem. Soc. 2018, 140, 10530-10535.
[14] G. Noh, E. Lam, J. L. Alfke, K. Larmier, K. Searles, P. Wolf, C.
Copéret, ChemSusChem 2019, 12, 968-972.
[15] C.-S. Chen, W.-H. Cheng, S.-S. Lin, Catal. Lett. 2000, 68, 45-48.
[16] N. Ishito, K. Hara, K. Nakajima, A. Fukuoka, J. Energy Chem. 2016,
25, 306-310.
[17] G. Jacobs, A. C. Crawford, B. H. Davis, Catal. Lett. 2005, 100, 147-
152.
[18] T. van Herwijnen, W. A. de Jong, J. Catal. 1980, 63, 83-93.
[19] T. van Herwijnen, R. T. Guczalski, W. A. de Jong, J. Catal. 1980, 63,
94-101.
[20] J. F. Edwards, G. L. Schrader, J. Phys. Chem. 1984, 88, 5620-5624.
[21] Y. Noto, K. Fukuda, T. Onishi, K. Tamaru, Trans. Faraday Soc. 1967,
63, 2300-2308.
Acknowledgements
[22] A. A. Gokhale, J. A. Dumesic, M. Mavrikakis, J. Am. Chem. Soc. 2008,
130, 1402-1414.
E.L., K.L., G.N., and P.W. were supported by the SCCER-Heat
and Energy Storage program of InnoSuisse. J.J.C.-P. and A.U.
thank the ICIQ foundation, the Generalitat de Catalunya
through the CERCA Programme and recognition (2017 SGR
1633), MCIU (CTQ2016-75499-R (AEI/FEDER, UE)) and
Severo Ochoa Excellence Accreditation 2014-2018 (SEV-
2013-0319) for financial support. A.C.-V. acknowledges the
financial support from the “Ramon y Cajal” Fellowship, funded
by Spanish MEC and the European Social Fund (RyC-2016-
19930). We acknowledge Dr. Wei-Chih Liao (ETH Zurich) and
Christopher Gordon (ETH Zurich) for assistance in solid-state
[23] D. Tibiletti, A. Goguet, F. C. Meunier, J. P. Breen, R. Burch, Chem.
Commun. 2004, 1636-1637.
[24] G. Jacobs, B. H. Davis, Appl. Catal., A 2005, 284, 31-38.
[25] R. Zhang, B. Wang, H. Liu, L. Ling, J. Phys. Chem. C 2011, 115,
19811-19818.
[26] D. C. Grenoble, M. M. Estadt, D. F. Ollis, J. Catal. 1981, 67, 90-102.
[27] E. L. Kunkes, F. Studt, F. Abild-Pedersen, R. Schlögl, M. Behrens, J.
Catal. 2015, 328, 43-48.
[28] D. Ferri, T. Bürgi, A. Baiker, Phys. Chem. Chem. Phys. 2002, 4, 2667-
2672.
[29] Y. A. Daza, J. N. Kuhn, RSC Adv. 2016, 6, 49675-49691.
This article is protected by copyright. All rights reserved.

10.1002/anie.201908060
Angewandte Chemie International Edition
RESEARCH ARTICLE
[30] S. S. Kim, H. H. Lee, S. C. Hong, Appl. Catal., A 2012, 423-424, 100-
107.
[47] A. Comas-Vives, M. Valla, C. Copéret, P. Sautet, ACS Cent. Sci.
2015, 1, 313-319.
[31] L. F. Bobadilla, J. L. Santos, S. Ivanova, J. A. Odriozola, A. Urakawa,
ACS Catal. 2018, 8, 7455-7467.
[48] G. Busca, P. F. Rossi, V. Lorenzelli, M. Benaissa, J. Travert, J. C.
Lavalley, J. Phys. Chem. 1985, 89, 5433-5439.
[32] T. Jin, Y. Zhou, G. J. Mains, J. M. White, J. Phys. Chem. 1987, 91,
5931-5937.
[49] A. Bansode, B. Tidona, P. R. von Rohr, A. Urakawa, Catal. Sci.
Technol. 2013, 3, 767-778.
[33] C. M. Kalamaras, S. Americanou, A. M. Efstathiou, J. Catal. 2011,
279, 287-300.
[50] T. Hyakutake, W. van Beek, A. Urakawa, J. Mater. Chem. A 2016, 4,
6878-6885.
[34] A. Goguet, F. C. Meunier, D. Tibiletti, J. P. Breen, R. Burch, J. Phys.
Chem. B 2004, 108, 20240-20246.
[51] K. K. Bando, K. Sayama, H. Kusama, K. Okabe, H. Arakawa, Appl.
Catal., A 1997, 165, 391-409.
[35] K. Sun, M. Kohyama, S. Tanaka, S. Takeda, J. Phys. Chem. C 2017,
121, 12178-12187.
[52] H. Ren, C.-H. Xu, H.-Y. Zhao, Y.-X. Wang, J. Liu, J.-Y. Liu, J. Ind.
Eng. Chem. 2015, 28, 261-267.
[36] K. Samson, M. Śliwa, R. P. Socha, K. Góra-Marek, D. Mucha, D.
Rutkowska-Zbik, J. F. Paul, M. Ruggiero-Mikołajczyk, R. Grabowski,
J. Słoczyński, ACS Catal. 2014, 4, 3730-3741.
[53] J. B. Wang, H.-K. Lee, T.-J. Huang, Catal. Lett. 2002, 83, 79-86.
[54] V. N. Ipatieff, G. S. Monroe, J. Am. Chem. Soc. 1945, 67, 2168-2171.
[55] T. Tagawab, G. Pleizier, Y. Amenomiya, Appl. Catal. 1985, 18, 285-
293.
[37] M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Hävecker, F. Abild-
Pedersen, S. Zander, F. Girgsdies, P. Kurr, B.-L. Kniep, M. Tovar, R.
W. Fischer, J. K. Nørskov, R. Schlögl, Science 2012, 336, 893.
[38] T. Fujitani, I. Nakamura, T. Uchijima, J. Nakamura, Surf. Sci. 1997,
383, 285-298.
[56] C. Copéret, A. Comas-Vives, M. P. Conley, D. P. Estes, A. Fedorov,
V. Mougel, H. Nagae, F. Núñez-Zarur, P. A. Zhizhko, Chem. Rev.
2016, 116, 323-421.
[57] C. Copéret, F. Allouche, K. W. Chan, M. P. Conley, M. F. Delley, A.
Fedorov, I. B. Moroz, V. Mougel, M. Pucino, K. Searles, K. Yamamoto,
P. A. Zhizhko, Angew. Chem., Int. Ed. 2018, 57, 6398-6440, Angew.
Chem. 2018, 130, 6506-6551.
[39] C. Tisseraud, C. Comminges, T. Belin, H. Ahouari, A. Soualah, Y.
Pouilloux, A. Le Valant, J. Catal. 2015, 330, 533-544.
[40] A. Le Valant, C. Comminges, C. Tisseraud, C. Canaff, L. Pinard, Y.
Pouilloux, J. Catal. 2015, 324, 41-49.
[58] A. Y. Rozovskii, G. I. Lin, Top. Catal. 2003, 22, 137-150.
[59] J. J. Corral-Pérez, A. Bansode, C. S. Praveen, A. Kokalj, H. Reymond,
A. Comas-Vives, J. VandeVondele, C. Copéret, P. R. von Rohr, A.
Urakawa, J. Am. Chem. Soc. 2018, 140, 13884-13891.
[60] A. de Juan, J. Jaumot, R. Tauler, Anal. Methods 2014, 6, 4964-4976.
[61] M. Digne, P. Sautet, P. Raybaud, P. Euzen, H. Toulhoat, J. Catal.
2002, 211, 1-5.
[41] J. D. Grunwaldt, A. M. Molenbroek, N. Y. Topsøe, H. Topsøe, B. S.
Clausen, J. Catal. 2000, 194, 452-460.
[42] S. Kuld, M. Thorhauge, H. Falsig, C. F. Elkjær, S. Helveg, I.
Chorkendorff, J. Sehested, Science 2016, 352, 969.
[43] T. Lunkenbein, J. Schumann, M. Behrens, R. Schlögl, M. G. Willinger,
Angew. Chem., Int. Ed. 2015, 54, 4544-4548, Angew. Chem. 2015,
127, 4627-4631.
[62] L. Foppa, T. Margossian, S. M. Kim, C. Müller, C. Copéret, K. Larmier,
A. Comas-Vives, J. Am. Chem. Soc. 2017, 139, 17128-17139.
[63] M.-C. Silaghi, A. Comas-Vives, C. Copéret, ACS Catal. 2016, 6,
4501-4505.
[44] S. Kuld, C. Conradsen, P. G. Moses, I. Chorkendorff, J. Sehested,
Angew. Chem., Int. Ed. 2014, 53, 5941-5945, Angew. Chem., Int. Ed.
2014, 126, 6051-6055.
[45] R. Wischert, P. Laurent, C. Copéret, F. Delbecq, P. Sautet, J. Am.
Chem. Soc. 2012, 134, 14430-14449.
[46] R. S. Schiffino, R. P. Merrill, J. Phys. Chem. 1993, 97, 6425-6435.
RESEARCH ARTICLE
CO2 + x H2
CH3OH + CO + CH3OCH3
More than one way: CO2
Author(s), Corresponding
Hydrogenation on Cu/Al2O3 is
studied by combining state-
of-the art synthetic methods
in combination with operando
spectroscopy and
Author(s)*
Page No. – Page No.
Title
computational chemistry to
CO2 + H2
CO or CH3OH
CH3OCH3
O
CH3OH
CH3OH + CO
elucidate possible reaction
mechanisms forming CH3OH,
CO and CH3OCH3.
O
H
H
Cu
Cu
OCH3
O
O
Al2O3
This article is protected by copyright. All rights reserved.