CO2 Hydrogenation to Ethanol over Cu@Na-Beta

Article abstract of DOI:10.1016/j.chempr.2020.07.001

Here, we report a high-performance catalyst Cu@Na-Beta, prepared via a unique method to embed 2～5 nm Cu nanoparticles in crystalline particles of Na-Beta zeolite, for CO₂ hydrogenation to ethanol as the only organic product in a traditional fixed-bed reactor. The ethanol yield in a single pass can reach ～14% at 300°C, ～12,000 mL·g_cat^?1·h^?1, and 2.1 MPa, corresponding to a space-time yield of ～398 mg·g_cat^?1·h^?1. The key step of the reaction is considered as the rapid bonding of CO₂? with surface methyl species at step sites of Cu nanoparticles to CH₃COO? that converts to ethanol in following hydrogenation steps. The points of the catalyst seemed to be that the irregular copper nanoparticles stuck in zeolitic frameworks offer high density of step sites and the intimate surrounding of zeolitic frameworks strongly constrain the CO₂ reactions at the copper surface and block by-products, such as methanol, formic acid, and acetyl acid. The high-performance catalyst Cu@Na-Beta, prepared via a unique method to embed 2～5 nm Cu nanoparticles in crystalline particles of Na-Beta zeolite, is reported for CO₂ hydrogenation to ethanol as the only organic product in a traditional fixed-bed reactor. The ethanol yield in a single pass can reach ～14% at 300°C, ～12,000 mL·g_cat^?1·h^?1, and 2.1 MPa, corresponding to space-time yield of ～398 mg·g_cat^?1·h^?1. The key step of the reaction is the rapid bonding of CO₂? with surface methyl species at step sites of Cu nanoparticles to CH₃COO?, which converts to ethanol in the following hydrogenation steps. The points of the catalyst seem to be that the irregular copper nanoparticles stuck in zeolitic frameworks offer a high density of step sites and that the intimate surrounding of zeolitic frameworks strongly constrains the CO₂ reactions at the copper surface and blocks byproducts such as methanol, formic acid, and acetyl acid. CO₂ direct reduction to ethanol is a much-anticipated research topic worldwide. A big progress has been made in the current investigation toward industry application. A high-performance catalyst Cu@Na-Beta, prepared via a unique method to embed 2～5 nm Cu nanoparticles in crystalline particles of Na-Beta zeolite, is reported for CO₂ hydrogenation to ethanol in a traditional fixed-bed reactor, with ethanol space-time yield of ～398 mg·g_cat^?1·h^?1. Peripherals-surrounded catalysts, which may be called mesocatalysts, appear to be one focus of future investigations on catalysis.

Full text of DOI:10.1016/j.chempr.2020.07.001

ll
Article
CO2 Hydrogenation to Ethanol over Cu@Na-
Beta
Liping Ding, Taotao Shi, Jing
Gu, ..., Yan Zhu, Zhaoxu Chen,
Weiping Ding
zxchen@nju.edu.cn (Z.C.)
dingwp@nju.edu.cn (W.D.)
HIGHLIGHTS
Ethanol as the only organic
product obtained from high-
performance CO2 hydrogenation
Key step for C–C bond is achieved
between CO2 and surface methyl
at Cu step sites
Current ethanol synthesis from
CO2 hydrogenation should be
useful in industry
Assemblies of surroundings and
reactive centers are important for
novel catalysts
CO2 direct reduction to ethanol is a much-anticipated research topic worldwide. A
big progress has been made in the current investigation toward industry
application. A high-performance catalyst Cu@Na-Beta, prepared via a unique
method to embed 2ꢀ5 nm Cu nanoparticles in crystalline particles of Na-Beta
zeolite, is reported for CO2 hydrogenation to ethanol in a traditional fixed-bed
reactor, with ethanol space-time yield of ꢀ398 mg$gcatÀ1$hÀ1. Peripherals-
surrounded catalysts, which may be called mesocatalysts, appear to be one focus
of future investigations on catalysis.
Q8
Ding et al., Chem 6, 1–17
October 8, 2020 ª 2020 Elsevier Inc.
https://doi.org/10.1016/j.chempr.2020.07.001

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
CO2 Hydrogenation to Ethanol
over Cu@Na-Beta
Liping Ding,1,4 Taotao Shi,1,4 Jing Gu,1 Yun Cui,1 Zhiyang Zhang,1 Changju Yang,1 Teng Chen,1
Ming Lin,2 Peng Wang,3 Nianhua Xue,1 Luming Peng,1 Xuefeng Guo,1 Yan Zhu,1 Zhaoxu Chen,1,
*
and Weiping Ding1,5,
*
SUMMARY
The Bigger Picture
The high-performance catalyst
Cu@Na-Beta, prepared via a
Here, wereportahigh-performancecatalystCu@Na-Beta,preparedvia
a unique method to embed 2ꢀ5 nm Cu nanoparticles in crystalline par-
ticles of Na-Beta zeolite, for CO2 hydrogenation to ethanol as the only
organic product in a traditional fixed-bed reactor. The ethanol yield in
a single pass can reach ꢀ14% at 300ꢁC, ꢀ12,000 mL$gcatꢂ1$hꢂ1, and
unique method to embed 2ꢀ5 nm
Cu nanoparticles in crystalline
particles of Na-Beta zeolite, is
reported for CO2 hydrogenation
to ethanol as the only organic
product in a traditional fixed-bed
reactor. The ethanol yield in a
single pass can reach ꢀ14% at
2.1 MPa, corresponding to a space-time yield of ꢀ398 mg$gcatꢂ1$hꢂ1
.
The key step of the reaction is considered as the rapid bonding of
CO2* with surface methyl species at step sites of Cu nanoparticles to
CH3COO* that converts to ethanol in following hydrogenation steps.
The points of the catalyst seemed to be that the irregular copper nano-
particles stuckinzeoliticframeworks offerhigh density of step sites and
the intimate surrounding of zeolitic frameworks strongly constrain the
CO2 reactions at the copper surface and block by-products, such as
methanol, formic acid, and acetyl acid.
300ꢁC, ꢀ12,000 mL$gcatꢂ1$hꢂ1
,
and 2.1 MPa, corresponding to
space-time yield of
ꢀ398 mg$gcatꢂ1$hꢂ1. The key step
of the reaction is the rapid
bonding of CO2* with surface
methyl species at step sites of Cu
nanoparticles to CH3COO*, which
converts to ethanol in the
INTRODUCTION
The utilization of carbon dioxide has a massive impact on carbon cycling for the
development of the recycling economy1,2 and has attracted much research attention
worldwide. Among the products from CO2 conversion, methanol as the main product
has been investigated intensively in recent years.3–5 Comparatively, few results have
been reported about the synthesis of ethanol or C2+OH from CO2 hydrogenation,
even though ethanol is not only a nontoxic but also a more valuable product that
can be easily converted to high-value chemicals such as ethylene. Considering the
synthetic routes of ethanol from feed gases of CO2+3H2 or CO+2H2, the relevant re-
actions are shown as Equations 1, 2, and 3. This means that the hydrogenation of CO2
to ethanol highly relates to the syngas conversion and the difference is just the water-
gas shift reaction, a well-established industrial reaction.6 As to the utilization of chem-
ical energy included in the feed gases, one C2H5OH (g) reserves ꢀ94% or ꢀ89%
chemical energy included in 6H2 or 2CO+4H2, respectively, reflecting to some extent
the advantages of CO2 hydrogenation to ethanol over the syngas conversion.
following hydrogenation steps.
The points of the catalyst seem to
be that the irregular copper
nanoparticles stuck in zeolitic
frameworks offer a high density of
step sites and that the intimate
surrounding of zeolitic
frameworks strongly constrains
the CO2 reactions at the copper
surface and blocks byproducts
such as methanol, formic acid, and
acetyl acid.
2CO2 + 6H2 / C2H5OH (g) + 3H2O (g) OH= ꢂ173.7 kJ/mol
2CO + 4H2 / C2H5OH (g) + H2O (g) OH= ꢂ255.9 kJ/mol
CO + H2O (g) / H2 + CO2 OH= ꢂ41.1 kJ/mol
(Equation 1)
(Equation 2)
(Equation 3)
Chem 6, 1–17, October 8, 2020 ª 2020 Elsevier Inc.
1

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
For the practical use of CO2 hydrogenation to ethanol, a significant promotion on
catalytic efficiency and ethanol selectivity is highly desired.7–15 At the present
time, it is a great challenge to obtain high or ideally exclusive selectivity to ethanol
by using traditional heterogeneous catalysts at valuably high CO2 conversion. New
ideas for the catalyst design are called to achieve high ethanol selectivity as well as a
high activity for CO2 conversion.
Copper-based catalysts have been widely applied to catalyze the hydrogenation of
COx and the production of organic oxygenates from syngas,16–18 of which the ac-
tivity and selectivity strongly depend on the support and/or promoters. Especially,
Liao et al. have revealed the importance of morphology-dependent interactions of
ZnO with Cu nanoparticles at the interface in between in selective hydrogenation
of CO2 to CH3OH.16 On the other hand, zeolites with well-organized 3D porous
structure have strong confining or modulating effects on entering molecules and
the clusters enclosed, as reported in the recent years.19–22 Zeolite beta is a disor-
dered intergrown hybrid of tetragonal and clinorhombic system, which are two
distinct but closely related structures and possesses a three-dimensional pore sys-
tem containing 12-membered ring apertures. The inner field potential of Beta
zeolite has fascinating characteristics, different from the surrounding effects of
the ZSM-5 and X zeolites, which we have reported earlier.19,22 In current work,
we combined the Beta zeolite and the active Cu nanoparticles to prepare a highly
efficient catalyst, in an intimately cooperated structure with copper enclosed in
crystalline particles of zeolite Beta, which showed the superior performance for
CO2 hydrogenation to ethanol. The CuO nanoparticles were embedded in crystal
particles of Na-Beta zeolite (abbreviated as CuO@Na-Beta) by a two-step prepara-
tion procedure using dry gel conversion and crystallization methods (Preparation
of Catalysts), similar to what we previously described.19 Then the sample was
reduced by hydrogen to Cu@Na-Beta before the catalytic reaction. The obtained
material is considered as ꢀ2–5 nm Cu nanoparticles embedded in the zeolite crys-
tal particles, as characterized by various techniques (see Experimental Procedures).
RESULTS AND DISCUSSION
The equilibrium conversions of CO2 with the mixture of CO2+3H2 as reactants at
different temperatures and pressures were calculated according to thermodynamic
parameters of the related compounds (Supplemental Information; Figure S1),
implying the reaction conditions for ethanol synthesis from CO2 hydrogenation
can be quite mild. Figures 1A–1D show the catalytic performance of the catalyst,
ꢀ8.24 wt % Cu@Na-Beta (the Cu content was measured by X-ray fluorescence
[XRF] spectroscopy) for CO2 hydrogenation. With the reaction temperature
increasing from 200ꢁC to 350ꢁC, the conversion of CO2 increases from 0.85% to
12.2% at a rather high space velocity 12,000 mL$gcatꢂ1$hꢂ1 and a pressure of
mere 1.3 MPa. When the temperature reaches 250ꢁC, CO appears as a by-product
and the selectivity of CO increases to ꢀ 45.2% as the temperature increases to 350ꢁC
(Figure 1A), while the ethanol is the only organic product. The pressure has a great
effect on the performance of the catalyst. When the pressure changes from 0.5 to 2.1
MPa (300ꢁC), the conversion of CO2 increases significantly from ꢀ2.0% to ꢀ18% and
the selectivity of CO decreases from 94.6% to 21% (Figure 1B), and ethanol is still the
only organic product observed in the tests of the reaction. When the space velocity
increases from 6,000 to 18,000 mL$gcatꢂ1$hꢂ1, the CO2 conversion decreases
from 13.5% to 3.7%, and the CO selectivity increases from ꢀ26.9% to ꢀ35.9% (Fig-
ure 1C). The ethanol yield of ꢀ14 % is obtained at 300ꢁC, 2.1 Mpa, and
1Key Laboratory of Mesoscopic Chemistry,
School of Chemistry and Chemical Engineering,
Nanjing University, Nanjing 210093, China
2Institute of Materials Research and Engineering,
Agency for Science, Technology and Research
(A*STAR), 3 Research Link, Singapore 117602,
Singapore
3National Laboratory of Solid State
Microstructures, College of Engineering and
Applied Sciences and Collaborative Innovation
Center of Advanced Microstructures, Nanjing
University, Nanjing 210093, China
4These authors contributed equally
5Lead Contact
*Correspondence: zxchen@nju.edu.cn (Z.C.),
dingwp@nju.edu.cn (W.D.)
12,000 mL$gcatꢂ1$hꢂ1
.
https://doi.org/10.1016/j.chempr.2020.07.001
2
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Figure 1. Catalytic Performance of the Cu@Na-Beta Catalyst
(A) Effect of reaction temperature (12,000 mL$gcatꢂ1$hꢂ1; 1.3 MPa; CO2+3H2).
(B) Effect of reaction pressure (12,000 mL$gcatꢂ1$hꢂ1; 300ꢁC; CO2+3H2).
(C) Effect of space velocity (1.3MPa; 300ꢁC; CO2+3H2).
(D) Prolonged test of reaction (300ꢁC; 1.3 MPa; 12,000 mL$gcatꢂ1$hꢂ1; CO2+3H2. Inset: Products
collected by a condenser during the first 24 h of reaction).
(E) The product selectivities for syngas conversion (1.5 MPa; 12,000 mL$gcatꢂ1$hꢂ1; 300ꢁC;
CO+2H2).
(F) Ethanol selectivity in organic products using the mixture of CO, CO2, and H2 as reactants
(1.5 MPa; 12,000 mL$gcatꢂ1$hꢂ1; 300ꢁC).
More importantly, the catalytic performance of the catalyst Cu@Na-Beta remains
stable during the 100-h reaction on stream (Figure 1D), and ethanol is the only
organic product formed under the reaction conditions (H2/CO2 = 3/1, 300ꢁC,
1.3 MPa, 12,000 mL$gcatꢂ1$hꢂ1). The products of the first 24-h reaction during the
prolonged test have been collected with a condenser cooled by liquid nitrogen.
The weight of the products obtained is ꢀ0.71 g, in the composition of ꢀ43.5 wt %
ethanol, coincidently according to the stoichiometry of the reaction 2CO2 + 6H2 =
C2H5OH + 3H2O, as analyzed by gas chromatography (GC) using acetone as inner
standard. Referring to the collection result, the isolated yield of ethanol is calculated
as ꢀ5.1%, which is in accordance with the analysis results by online GC. The NMR
Chem 6, 1–17, October 8, 2020
3

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Table 1. Catalytic Performance of Some Related Catalysts for CO2 Hydrogenation
Catalyst
P (MPa)
T (ꢁC)
Conv. (%)
Sel. (%)b
CH4
N/D
5.6
STYc
Ref
CH3OH
ꢀ100
30.1
8.2
C2H5OH
N/D
CO
41.6
66.4
25.4
30.5
19.7
57.5
6.9
ꢂ
a
Cu/SiO2
1.3
300
0.47
3.1
1.8
7.9
26.7
32
0 (0)
This work
Cu/Na-Betaa
Cu@Na-ZSM5a
Cu@Na-Betaa
Ru-Fe/SiO2
64.3
46 (3.0)
59 (3.9)
258 (17.1)
ꢂ
6.8
85.0
N/D
34.7
36.7
56.4
ꢂ
N/D
29.4
47.1
4.7
ꢀ100
16
5.0
260
340
300
270
200
200
200
140
Kusama et al.7
Nieskens et al.8
Li et al.9
CoMoS
10.3
6.0
12.9
ꢂ
CuZnFe0.5K0.15
42.3
ꢂ
32 (C2+
)
148(C2+)
d
K-CuCo/MoOx
4.0
ꢂ
ꢂ
27 (C2+
0.29
ꢂ
)
Prieto et al.10
He et al.11
Pt/Co3O4
8.0
ꢂ
ꢂ
12.5
57
ꢂ
Co/Mo2C
4.0
ꢀ10
ꢂ
9.5
46
25
9.5
Chen et al.12
Bai et al.13
Pd2Cu/P25
CoAlOx
3.2
ꢂ
ꢂ
92
ꢂ
41.5
0.444
ꢂ
4.0
ꢂ
ꢂ
ꢂ
ꢂ
92.1
100
ꢂ
ꢂ
Wang et al.14
Song et al.15
Meso Carbone
ꢂ0.56 V
aCO is not included to calculate the organic product selectivity.
bSpace-time yield of ethanol (mg$gcatꢂ1$hꢂ1). Data in parentheses are turnover frequencies (TOF in hꢂ1).
c0.1 g Cat; GHSV: 12,000 mL$gcatꢂ1$hꢂ1
dCO hydrogenation to ethanol.
.
eElectroreduction of CO2 to ethanol.
spectrum of the liquid product collected is also listed in Figure S2 and only the sig-
nals from ethanol and water are observed.
It is very interesting that the catalyst gives totally different results with syngas
CO+2H2 used as the reactants, of which products similar to Fischer-Tropsch synthe-
sis are obtained, even though some methanol and ethanol are also produced, as
shown in Figure 1E. If the mixture of CO, CO2, and H2 was used as the feed gas,
the ethanol became the only organic product with a ratio of CO2/(CO + CO2) in
the feed gas larger than 0.65, indicating the CO2 is important for the formation of
ethanol (Figure 1F).
For insight into the catalytic mechanism, three control samples were prepared with
the same copper content. Two of them were prepared by traditionally impregnated
copper over Na-Beta zeolite (Cu/Na-Beta) and silica (Cu/SiO2). The third control
sample was Cu@Na-ZSM-5, which was prepared using the same method as
Cu@Na-Beta with Na-ZSM-5 zeolite. The typical catalytic performances of the cata-
lysts for the title reaction are listed in Table 1. Under the same reaction conditions,
the catalyst Cu/SiO2 shows poor activity for CO2 conversion, and no ethanol is de-
tected in products and the results seem consistent with that reported by Wang
et al.,23 in which the Cu/SiO2 needed higher pressure to promote its activity.
Comparatively, the Cu/Na-Beta is active for CO2 conversion and ethanol formation
but much inferior to the Cu@Na-Beta, and the yield of ethanol over the former is only
ꢀ16% of the latter. The catalyst Cu@Na-ZSM-5 shows fairly good catalytic perfor-
mance but the selectivity to ethanol is not exclusive and methanol and methane
are detected. The results of the papers published on conversion of CO2 to ethanol
are listed in Table 1. Among traditional catalysts, the results reported by this work
are much more excellent for the selectivity of ethanol. Almost 100% selectivity to
ethanol can be achieved by electroreduction over nitrogen-doped mesocarbon;
however, the conversion of CO2 is low. These experimental results suggest that
4
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Chem 6, 1–17, October 8, 2020
5

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Figure 2. The Characterization of the Catalyst Cu@Na-Beta at Different Status
(A) TEM image of the Na-Beta zeolite and the pore-size distribution obtained from N2 sorption (inset) of the zeolite treated by alkaline solution for
formation of irregular mesopores.
(B) HRTEM image of Cu@Na-Beta reduced from CuO@Na-Beta and the size distribution of Cu nanoparticles (inset).
(C) HRTEM of copper nanoparticles in Cu@Na-Beta.
(D) HRTEM of a Cu nanoparticle in Cu@Na-Beta (the same sample with C).
(E) TEM image of spent Cu@Na-Beta after a long-term reaction (100 h) and the size distribution of Cu nanoparticles (inset).
(F) H2-TPR of samples of CuO@Na-Beta and CuO/Na-Beta.
(G) Operando X-ray absorption spectra (XAS) of CuO@Na-Beta recorded every 40 min from its contact with reaction gases (0.5 MPa, H2/CO2 = 3:1,
250ꢁC).
(H) Radial distribution functions of Cu from Fourier transformed k-edge of Cu EXAFS of the samples Cu@Na-Beta, Cu/Na-Beta, and Cu/SiO2.
the zeolites and the cooperation manner between the zeolites and copper are
important to obtain the excellent catalytic performance and the enclosure of copper
nanoparticles in the crystal particles of Beta zeolite is the key factor. The intimate
cooperation or synergism between them, however, appears different from previ-
ously reported remarkable results on syngas conversion using Ox-Zel catalysts, in
which the bifunctional catalytic mechanism is established, i.e., CO hydrogenation
on oxides to intermediates that are further converted to hydrocarbons on acidic
SAPO-34 or H-ZSM-5 at distance, as reported by Bao et al. and Wang et al.24,25
For current Cu@Na-Beta catalyst, the surrounding Na-Beta framework constrains
the copper nanoparticles at special shapes and electronic status suitable for the
CO2 hydrogenation to ethanol occurr at the nanoparticles, and the enveloping of
zeolitic frameworks on the nanoparticle surface is also highly influential, which re-
stricts side reactions for the production of methanol, formic acid, acetic acid, etc.
Figure 2 shows the characterization results of the related samples. The alkali-treat-
ment on raw Na-Beta zeolite led to the formation of mesopores in their crystal par-
ticles, and the sizes of the mesopores were measured by N2 sorption as ꢀ4 nm (Fig-
ure 2A and inset). CuO nanoparticles were introduced into the mesopores and then
enclosed in the interior of the zeolite crystal particles after the two-step dry gel con-
version (DGC) synthesis. After a reduction in 5% H2/N2 at 350ꢁC, the CuO@Na-Beta
was reduced to Cu@Na-Beta and the copper species remain as highly dispersed
nanoparticles enclosed in the Beta zeolite crystal particles, as shown in Figure 2B.
HRTEM of irregular copper nanoparticles in Cu@Na-Beta is shown in Figure 2C.
The irregular shape of Cu nanoparticles was embedded in Na-Beta zeolites. The
high-resolution image of a copper particle in a gourd shape enclosed in the zeolite
is shown in Figures 2D and S3A. By the difficultly obtained image, the copper nano-
particle enclosed in the zeolite is constructed as accurately as possible and shown in
Figure S3B. It appears that step sites show up on the particle surface, due to the
constraint of the irregular mesopores of the zeolite etched by alkali during the prep-
aration. After a long-term reaction of 100 h, the copper nanoparticles retain their
original sizes, as revealed by the TEM image of spent Cu@Na-Beta (Figure 2E), in
agreement with its stable catalytic performance, as shown in Figure 1D. The XRD
(X-ray diffraction) patterns of related samples are shown in Figure S4, indicating
the reduction of CuO to metallic copper during the treatment in 5% H2/N2 at
350ꢁC. The crystallinities of the zeolite in CuO@Na-Beta, Cu@Na-Beta, and spent
Cu@Na-Beta appear to be similar to the original Beta zeolite. The stability of
Na-Beta framework surrounding the copper nanoparticles ensures the stable activity
of the catalyst. For the control catalyst Cu/Na-Beta, however, Cu nanoparticles
significantly grow up after reaction (Figure S5) and, accordingly, the catalyst, though
with poor activity, deactivates in a much shorter time on stream. It should be pointed
out that the copper species loaded on or enclosed in the zeolite Beta are signifi-
cantly different, as shown by the measurement of H2 temperature-programed
6
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Table 2. Cu k-edge EXAFS Data for the Cu-Based Catalysts
a
2
2 c
CN (Cu-Cu)b
8.59 G 0.86
7.74 G 0.87
10.29 G 0.63
s (A )
0.0138
0.0136
0.0135
˚
R(A)
˚
Sample
Scattered
Cu@Na-Beta
Cu/Na-Beta
Cu/SiO2
Cu
Cu
Cu
2.57
2.54
2.55
All the parameters are obtained with the catalysts in reaction of CO2+3H2 in the XAS cell.
aBond distances
bCoordination number
cDebye-Waller factor
reduction. The copper species enclosed are reduced at temperatures about 100ꢁC
higher than those loaded on the zeolite (Figure 2F). The results show that, in Cu@Na-
Beta, Cu particles are tightly surrounded by a molecular sieve and interact with each
other.
Operando X-ray Afine structure (XAFS) measurements were performed to determine
the status of copper species during the reaction under 250ꢁC and 0.5 MPa pressure
and the results are shown in Figures 2G and 2H. The XANES spectra in Figure 2G
show that as soon as reaction gas is introduced, the intensity of the peak at 8,997
eV starts to decrease, indicating the reduction of CuO. Simultaneously, the peak
at 8,981.3 eV due to the Cu species appears and increases with time on stream.
By the edge fitting, CuO species are reduced to Cu after 40 min on stream, indi-
cating that metallic copper is mainly the active species for the reaction. The radial
distribution functions of the copper obtained by Fourier transform of the EXAFS
are shown in Figure 2H and those of the control catalysts, i.e., Cu/Na-Beta and
Cu/SiO2, are also listed. The coordination parameters of Cu in the three samples
are given in Table 2. The Cu-Cu shell coordination number and distances in the three
samples are given in Table 2, which shows the copper nanoparticles in Cu@Na-Beta
and Cu/Na-Beta are similar in size of around ꢀ3 nm26,27 in accordance with the TEM
results and the copper nanoparticles in Cu/SiO2 are slightly larger in size. This means
that the crystal structure of the copper nanoparticles in the three samples are in fact
similar, which cannot explain their significant differences in the catalytic property for
CO2 hydrogenation. We think the true cause lay in the different shapes of the copper
nanoparticles in the samples and the intimate interaction between the copper and
the support. The Cu@Na-Beta has irregular copper nanoparticles and the most inti-
mate interaction between copper and zeolite. The intimacy of copper and zeolite for
Cu/Na-Beta is inferior to that of Cu@Na-Beta and only the bottoms of the copper
particles contact with the zeolite and it was prepared with traditional impregnation
method and without encapsulation (Figure S5). A small amount of ethanol observed
on the Cu/Na-Beta may form at the boundary of copper and Na-Beta zeolite. The
results of X-ray photoelectron spectroscopy (XPS) and Cu LMM Auger spectra (Fig-
ure S6) recorded with the samples reduced in the instrument show that a small
amount of electropositive copper species is detected in the Cu@Na-Beta but not de-
tected in other two samples, which implies copper particles, or the surface copper
atoms in the particles, interact strongly with the framework oxygen of the Na-Beta
surrounding, and, a relatively more intimate cooperation between Cu site and the
zeolite of Cu@Na-Beta can be deduced.
As to the formation mechanism of CH3CH2OH from CO2 hydrogenation on Cu@Na-
Beta, it was explored by a transient measurement between the switching of
H2/12CO2 mixture (3/1) to H2/13CO2 mixture (3/1). Under steady-state reaction con-
ditions (300ꢁC and 0.5 MPa), the stream of H2/12CO2 mixture (3/1) was switched to
Chem 6, 1–17, October 8, 2020
7

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Figure 3. Investigation on the Reaction Intermediates of CO2 to CH3CH2OH on Cu@Na-Beta
Catalyst
(A) The relative concentrations, monitored at the outlet of the reactor, of 13CO2 or 12CO2 during
switching the reactants between 13CO2+3H2 and 12CO2+3H2 (up panel); and concentrations of
products (12CH312CH2OH, black; 12CH313CH2OH, red; 13CH312CH2OH, blue; and 13CH313CH2OH,
pink) during the switching (lower panel), over the catalyst Cu@Na-Beta in reaction.
(B) In situ FT-IR spectrum during exposure of Cu@Na-Beta to 0.7 MPa reaction gas (CO2/H2 = 1:3) at
300ꢁC. CH3COOꢂ:3,010 cmꢂ1 d(CH), 1,680 cmꢂ1 nas(OCO), 1,511 cmꢂ1 ns(OCO); CH3CH2OH:
ns(OH) 3,100 cmꢂ1, 1,300 cmꢂ1d(OH), 2,850 cmꢂ1 ns(CH2), and 2,950 cmꢂ1 nas(CH2).
(C) The 1Ha` 13C CP MAS NMR and 1H MAS NMR recorded on the catalyst quenched from steady-
state reaction on stream of H2/13CO2 (3/1).
(D) TPD profiles detected using a mass spectrometer from the catalyst Cu@Na-Beta with
adsorption of HCOOH, CH3COOH, CH3OH, and CH3CH2OH, respectively.
H2/13CO2 mixture (3/1) and then switched back to the original H2/12CO2 mixture
(3/1) gases. The products at the exit of the reactor were stored using a multiposition
micro-electric valve actuators with 12 sample loops of 500 ml and analyzed afterward
by a GC-MS (Figure S7), and the results are shown in Figure 3. The products of
12CH312CH2OH, 12CH313CH2OH, 13CH312CH2OH, and 13CH313CH2OH are detected
by their characteristic m/z peaks listed in Table S1. The changes in the relative con-
centration of 13CO2 and 12CO2 respond to the transient switching accordingly. The
most important result is the synchronous variation of 12CH313CH2OH signal with the
13CO2 along the switching time (Figure 3), suggesting that the gaseous 13CO2 reacts
with surface 12CH3* readily to give 12CH313CH2OH. The results also imply that the
8
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
formation of C–C bond over this catalyst is in fact faster than the formation of C1 spe-
cies, such as methane, coincident with the results of the catalytic reaction listed in
Figure 1. The surface 12CH3* species can be inferred on the copper catalyst surface
because methane is a common product of CO2 hydrogenation on copper-based
catalysts.23
For further detection of the surface intermediates, the measurements of in situ FT-IR
on Cu@Na-Beta were applied. The wavenumbers of 3,010 cmꢂ1 d(CH), 1,680 cmꢂ1
nas(OCO), and 1,511 cmꢂ1 ns(OCO) can be assigned to adsorbed CH3COO* and
the wavenumbers of 3,100 cmꢂ1 ns(OH), 1,300 cmꢂ1 d(OH), 2,850 cmꢂ1 ns(CH2),
[9a]
and 2,950 cmꢂ1 nas(CH2) can be assigned to CH3CH2OH
(Figure 3B). For com-
parison, the results of similar measurements on Cu/Na-Beta are shown in Figure S8.
In general, the differences are obvious, such as the much weaker signals of
CH3COO* and different distribution of surface species, in accordance with its lower
activity for CO2 hydrogenation to produce a small amount of CH3OH, CH4, and
CH3CH2OH.
The catalyst was rapidly quenched to room temperature from steady-state reaction
at 250ꢁC and the surface species were then detected by NMR for 13C-NMR measure-
ment, the Cu@Na-Beta was quenched from steady-state reaction and then loaded in
a glove box to an solid state NMR rotor. The results of NMR measurement are de-
picted in Figure 3C. The 13C NMR signal at 160.0 ppm (peak 2) and the 1H NMR
peak at 8.6 ppm (trace hydrogen attached to a carboxyl group) are consistent with
the presence of carboxyl species. The 13C NMR resonance at 60.7 ppm (peak 3) sug-
gests the presence of surface methoxy species while the other signals observed
at 20.9 (peak 4),39.7 ppm (peak 5), and 1ꢀ3 ppm (1H NMR) are in agreement
with the formation of different surface methyl species.4 The 13C NMR signal at
170.0 ppm (peak 1) is assigned to carbonate, which is not so important in the reac-
tion. It is suggested by the results of IR and NMR that the intermediate for ethanol is
surface-adsorbed acetic acid, i.e., *CH3 + O=C=O* a` CH3COO* + *. The energy bar-
rier for desorption of the intermediate acetate radical is too high,28 which can only
be hydrogenated to ethanol in further steps.
As to the reason for that the methanol, formic acid, or acetic acid is not detected in
the product during the steady-state reaction of CO2 hydrogenation, the constraint of
the zeolite Na-Beta framework surrounding the copper nanoclusters must take ef-
fect. As shown in Figure 3D, basically no CH3OH, HCOOH, or CH3COOH can be de-
tected by a mass spectrometer, during heating the catalyst, Cu@Na-Beta, with the
adsorption of CH3OH, HCOOH, or CH3COOH, respectively, which all decomposed
to COx during the heating of temperature-programmed desorption (TPD), indicative
of the strong interaction of the compounds with the catalyst. The ethanol, however,
can desorb successfully from the catalyst around 200ꢁC ꢀ300ꢁC. Interestingly, no
CH3OH, HCOOH, or CH3COOH can be detected by the mass spectrometer, during
heating the catalyst Cu/Na-Beta with the adsorption of CH3OH, HCOOH, or
CH3COOH, respectively, which all decomposed to COx during the heating of
TPD. Even most of ethanol decomposed to COx (Figure S8B). The results indicate
the strong adsorption of the zeolite to these species.
To shed insightful light on the mechanism of ethanol formation, density functional
and slab model calculations were performed on the two key and crucial steps for
this process: the formation of surface methyl and the C–C bond. Many documented
results on copper-based catalysts catalyzed CO2 hydrogenation indicate that
methyl is likely formed from CH3OH and/or CH3O whose formation is quite facile
Chem 6, 1–17, October 8, 2020
9

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
as demonstrated by Kattel et al.18 Hence, we considered the pathways (Equation 4)
and (Equation 5) for CH3 formation on the model surfaces. Reactions (Equation 6)
and (Equation 7) were designed for C–C bond formation.
CH3O* / CH3* + O*
(Equation 4)
(Equation 5)
(Equation 6)
(Equation 7)
CH3OH* / CH3* + OH*
CO2* + CH3* + 2H2 / CH3CH2OH* + OH*
CO2* + CH3O* + 2H2 / CH3CH2OH* + O* + OH*
Test calculations show that the (100), (110), and (111) facets of copper were not
active enough for the reactions. The (221) facet of copper, which was observed on
the copper nanoparticles (Figure 2), was found to be effective and, thus, was
selected to mimick the catalyst. To better reflect the real catalyst surface, three
(221) surface models were constructed, i.e., the perfect, with Cu vacancy or O-doped
at the edge (Figure 4A,). The obtained binding energies for CO2 on the three sur-
faces are 0.31, 0.39, and 0.47 eV, respectively.
Figures 4B–4D illustrate the initial states (IS), transition states (TS), and final states
(FS) of the kinetically favorable pathways for reactions (Equation 4) and (Equation 5).
Other pathways are depicted in Figure S9. Surface defects, vacancy, and doped
atoms, reduce the barriers by 0.2–0.3 eV, as shown in Table S2. The barrier/reaction
heat calculated for a kinetically favorable route of CH3O* / CH3* + O* is 0.99/0.57
eV for the O-doped Cu (221). Methyl formation via the C–O bond breaking of
CH3OH is easier than through CH3O decomposition. The two lowest barriers are
0.78 eV with exothermicity of ꢂ0.35 on the O-doped surface (Table S2). A similar
route that has 0.80 eV for the barrier and ꢂ0.36 for the reaction heat is calculated
on the Cu vacant Cu (221). These results indicate that the formed methanol can be
easily converted into methyl, consistent with undetectable methanol in the product.
When zero-point energy (ZPE) correction is included, the barriers are only 0.68 eV. A
similar route that has 0.80 (0.70 with ZPE correction) eV for the barrier and ꢂ0.36 for
the reaction heat is calculated on the Cu vacant Cu(221). These results indicate that
the formed methanol can be easily converted into methyl, consistent with undetect-
able methanol in the product.
For the reactions (Equation 6) and (Equation 7), the initial states, transition states, and
final states of CO2* + CH3* + 2H2 reaction paths on the three (221) surfaces are de-
picted in Figures 4E–4G, respectively. The reaction CO2 + CH3O + 2H2 on the three
(221) surfaces are depicted in Figure S9 of Supplemental Information. Figures 4H and
4I show the calculated energy barriers/reaction energies without zero-point energy
correction for CO2* + CH3* / CH3COO* on the perfect, Cu vacancy, and O-doped
defect surfaces are ④ 0.32/ꢂ1.43, ⑤ 0.19/ꢂ1.49, and ⑥ 1.49/ꢂ0.05 eV, respec-
tively. The reaction energies ⑦ for CH3COO* reduction to CH3CH2OH* are
ꢂ1.17,ꢂ0.67, and ꢂ0.92 eV on the corresponding (221) surfaces, respectively. The
reactions are exothermic. Figure 4J shows the energy barriers/reaction energies for
the reaction CO2* + CH3O* + 2H2 on the three model (221) surfaces.
10
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Figure 4. Reaction Pathway of CO2 to CH3CH2OH and Reaction Energetics Calculated by DFT
(A) Illustrations of perfect, Cu vacancy, and O-doped surfaces on the edges. The differently colored atoms represent different sites on first layer
(yellowish-brown for site type 1, yellow for site type 2, green for site type 3, and blue for site type 4. Dark brown sphere is for Cu at lower layers).
(B–D) Illustrations of the initial state, transition state, and final state of methyl formation from CH3O* on (B) O-doped Cu(221) and from CH3OH on (C) Cu
vacancy defect surface and (D) O-doped surface.
(E–J) Illustrations of the initial state, transition state, and final state of CO2+CH3 reaction paths on (E) Cu vacancy defect surface, (F) perfect
surface, (G) O-doped surface and potential energy surfaces for the reaction of (H) CO2 + CH3OH / CO2 + CH3 / CH3CH2OH, (I) CO2 + CH3O / CO2
+
CH3 / CH3CH2OH, and (J) CO2 + CH3O / CH3CH2OH. Dark brown sphere, Cu; red sphere, O; gray sphere, C; white sphere, H; yellowish-brown, Cu on
edge.
As can be seen from the results, all the steps are exothermic except the reaction
CO2*+CH3O* to CH3COO* on the O-doped defect (221) surface, which is slightly
endothermic. For CO2* + CH3* to CH3COO*, the Cu vacancy surface is very favor-
able for this process with a barrier of 0.19 eV that is smaller than the adsorption en-
ergy of CO2 (0.39 eV), followed by the perfect surface with the barrier of 0.32 eV and
CO2 adsorption energy of 0.31 eV. The high barrier of 1.49 eV on the O-doped sur-
face indicates that the reaction hardly takes place here. For CO2* + CH3O* to
CH3COO*, the barrier heights on the three models are larger than the calculated
adsorption energy of CO2. Hence, CO2 would desorb before reacting with
CH3O*. These results suggest that the C–C bond forms mainly from the reaction
of CO2 with methyl (not methoxide) and the step surface with Cu vacancy is the
most effective sites for the rapid bonding of CO2* to surface methyl species to
Chem 6, 1–17, October 8, 2020
11

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
Scheme 1. Summary of the CO2 Hydrogenation to Ethanol over the Catalyst Cu@Na-Beta
The active copper nanoparticles were enclosed in the zeolite particles in irregular shapes
constrained by the zeolite mesopores purposely prepared by alkali etching. At such nanoparticles
of copper, step surfaces can be stabilized by the intimate constraint of the zeolite, which are active
for the bonding of CO2 with adsorbed methyl to form intermediate CH3COO*, which is converted
to ethanol in following hydrogenation.
form surface acetyl acid which converts to the final product of ethanol in following
exothermic hydrogenations.
Now, in phenomenology, the hydrogenation of CO2 to CH3CH2OH occurs at the
reactive centers of copper nanoparticles in irregular shapes, which are shaped and
confined by the 3D Na-Beta zeolitic frameworks from all sides. The reaction process
can be summarized and shown in Scheme 1. The unique copper nanoparticles in
irregular shapes with step surface and the intimate cooperation between Cu nano-
particles and Na-Beta play the key role in the high performance of reaction over
the catalyst Cu@Na-Beta. CO2 is firstly hydrogenated to CH3* at the surface of cop-
per nanoparticles and the subsequent CO2* adsorbed reacts with CH3* to form
CH3COO* as the most important surface intermediate, which cannot desorb directly
due to the high thermodynamics barrier.29 The followed hydrogenation of
CH3COO* to ethanol is exotic and easy to take place on the catalyst.30 The 3D
Na-Beta zeolitic framework functions at two aspects, i.e., to confine and modulate
the copper nanoparticals in unique shapes with surface sites for the reactions and
to constrain the reactants at the layer nearby the nanoparticle surface to restrict
the formation of by-products such as methanol, formic acid, or acetic acid.
Conclusion
In conclusion, CO2 can be converted to ethanol in high efficiency over the complex
catalyst Cu@Na-Beta. Combining the experimental and theoretical calculation
12
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
results, it is considered that the synergies between the irregular copper nanopar-
ticles and the surrounding zeolitic framework are responsible for the high-perfor-
mance catalyst Cu@Na-Beta. The current findings reflect the importance of the
catalytic assembly of ‘‘reactive centers plus their surroundings’’ that would be ex-
tracted as a new concept of mesocatalyst, and the appropriate use of such combina-
tory structures will be great helpful to design more highly efficient catalysts.
EXPERIMENTAL PROCEDURES
Resource Availability
Lead Contact
Weiping Ding is the Lead Contact for this work and can be reached here: dingwp@
nju.edu.cn.
Materials Availability
All materials generated in this study are being made available.
Data and Code Availability
The published article includes all Data and Code generated or analyzed during this
study.
Preparation of Catalysts
In a typical synthesis of Cu@Na-Beta, 6.00 g commercial Na-Beta zeolites (Si/Al ꢀ 25)
in a size of ꢀ0.1 mm were introduced intracrystalline mesopores by alkaline treat-
ment with 200ml 0.200mol/L NaOH solution at 65ꢁC for 30 min, then filtered,
washed, and dried (denoted as alk-Beta). Then CuO was introduced in the pores
of alk-Beta by impregnating Cu(NO3)2 ethanol solution and then calcined at
450ꢁC for 3 h (denoted as CuO/alk-Beta). Then in the next step, 1.00 g CuO/alk-
Beta was thoroughly mixed with a kind of Na-Beta zeolite synthesis gel, which was
prepared by 0.125 g Al2(SO4)3$18H2O, 3.16 g 20% Tetraethylammonium hydroxide
(TEAOH), 0.540g SiO2, and 0.0275g NaOH. The slurry was treated at 180ꢁC for
2 days for crystallization using the dry gel conversion method as we previously re-
ported19 and then thoroughly rinsed with water and calcined at 550ꢁC for 4 h in
air to remove the template, and the sample was denoted as CuO@Na-Beta. The
CuO@Na-Beta was reduced by 5% H2/N2 at 350ꢁC for 1.5 h before the catalytic
test (denoted as Cu@Na-Beta). CuO@Na-ZSM-5 and Cu@Na-ZSM-5 were prepared
similarly with Na-ZSM-5. The CuO/Na-Beta and CuO/SiO2 were prepared by tradi-
tional impregnation method. Cu(NO3)2 ethanol solution was impregnated to Na-
Beta zeolite or SiO2 (Cu: Na-Beta or SiO2 = 82.4 mg: 917.6 mg) and then calcined
at 450ꢁC for 3 h to be CuO/Na-Beta or CuO/SiO2, which were reduced to Cu/Na-
Beta or Cu/SiO2 by 5% H2/N2 at 350ꢁC for 1.5 h before the catalytic test. The copper
contents in all the samples were controlled as the same for the convenience of
comparison.
Characterization of Catalysts
The crystalline structure and morphology of the catalysts were characterization by
XRD, TEM, and HRTEM. XRD measurements were performed on an X’pert PAN
analytical diffractometer (Cu Ka radiation, 40 kV, 40 mA). TEM and HAADF-STEM im-
ages of the samples were obtained using JEM-2010 UHR and FEI Tecnai G2F20
(200 kV). The status of Cu in the samples during the reduction and reaction in the
mixed gases of CO2 + 3H2 at 250ꢁC and 0.5 MPa was determined by in situ measure-
ment of Cu K-edge X-ray absorption spectra (XANES and EXAFS). These data
were acquired in transmission mode at beamline BL14W1 (Shanghai Synchrotron Ra-
diation Facility). The station was operated with
a Si (111) double crystal
Chem 6, 1–17, October 8, 2020
13

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
monochromator. Around 100 mg sample was pressed into 1cm wafer and placed in
an in situ cell attached to a controlled heating equipment. Gas flows were controlled
by mass flow controllers. The valence states of Cu in Cu@Na-Beta were also carefully
analyzed by XPS and AES (Auger electron spectroscopy). For the measurement, the
catalyst was treated at 250ꢁC in flowing CO2 + 3H2 at 0.5 MPa for 40 min and then
transferred into XPS chamber without contacting air.
The H2 temperature-programmed reduction was carried out at apparatus TP5080
(Xianquan, Tianjin). The sample, placed in a quartz tube, was pretreated at 200ꢁC
in Ar for 1 h and cooled down to room temperature. Then the sample was heated
in flowing 5% H2/Ar at 10ꢁC/min, and the consumption of H2 was monitored using
a thermal conductivity detector. CH3OH, CH3CH2OH, HCOOH, or CH3COOH tem-
perature-programmed desorption were measured using the same apparatus. The
catalyst was firstly pretreated and reduced at 350ꢁC. After cooled down to room
temperature, the compounds of CH3OH, CH3CH2OH, HCOOH, and CH3COOH
were introduced into the carrier gas at 70ꢁC to the catalyst for adsorption, respec-
tively. Then, the desorbed species from the catalyst were monitored by a mass spec-
trometer during linear heating in flowing He at 10ꢁC/min.
Switching Experiment between H2/12CO2 Mixture (3:1) and H2/13CO2 Mixture
(3:1)
Figure S6 shows the schematic drawing of equipment with a computer-controlled
twelve-position valve with sample loops specifically designed for collecting multiple
compositions at short time intervals of the product gas at transient state for later
GC-MS analysis. Under steady-state conditions (250ꢁC and 0.5 MPa), the stream
of H2/12CO2 mixture (3:1) was switched to H2/13CO2 mixture (3:1) and then switched
back to the original H2/12CO2 mixture (3:1) gases. The products at various time of
transient at the exit of the reactor were stored in the 12 sample loops of 500 ml
and analyzed afterward by GC-MS.
NMR and IR Detection on Quenched Catalyst
The Cu@Na-Beta during the steady-state reaction with H2/13CO2 mixture (3:1) at
250ꢁC for 0.5 h, 0.5 MPa was quenched rapidly to room temperature and loaded
in a glove box to a solid-state NMR rotor for 1H MAS NMR or 1Ha`13C CP MASNMR
measurements. For IR detection, the Cu@Na-Beta was similarly quenched from
steady-state reaction to room temperature with H2/12CO2 (3:1) as reaction gas.
Then the sample was placed in a quartz cell for IR measurement.
Catalytic Tests
CO2 hydrogenation was carried out in a fixed-bed reactor with 6 mm inner diameter.
0.10 g catalyst (pelletized and sieved into 0.38–0.83 mm particles) was packed in
the reactor and rested in the middle of the furnace heating zone. After reduction
of the catalyst at 350ꢁC in the reactor by 5% H2/N2 gas flow (20 ml/min) for 1.5 h,
the reactor was charged with a H2/CO2 mixed gas (3:1, and 8% N2 as internal stan-
dard) at various flows and pressures. The products were analyzed by online GC with
dual column and dual detector. CO2, H2, N2, CO, and light alkane were separated by
TDX-01 and detected by Thermal conductivity detector (TCD), while alcohols or hy-
drocarbons were separated by supelco Petrocol DH 50.2 chromatographic column
and detected by FID.
CO2 conversion was calculated by equation:
CO2 conversion = CO2 conversion was calculated by equation:
14
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
CO2in ꢂ CO2out
CO2 conversion =
3 100%
CO2in
where CO2 in and CO2 out represent the molar fraction of CO2 at the inlet and outlet,
respectively.
Product selectivities were calculated according to equations:
CO2out
SðCOÞ =
3 100%
CO2in ꢂ CO2out
f
OHA
CH3
CH3OH
SðCH3OHÞ =
3 100%
f
OHA OH + f
A
+ 2f
OHA
CH3
CH3
CH4 CH4
CH3CH2 CH3CH2OH
f
A
CH4 CH4
SðCH4Þ =
3 100%
f
OHA OH + f
A
+ 2f
OHA
CH3
CH3
CH4 CH4
CH3CH2 CH3CH2OH
2f
OHA
CH3CH2OH
CH3CH2
SðCH3CH2OHÞ =
3 100%
f
OHA OH + f
A
+ 2f
OHA
CH3CH2OH
CH3
CH3
CH4 CH4
CH3CH2
S(CH3OH), S(CH3CH2OH), S(CH4) were calculated in organic products exclude CO.
where CO2 in and CO2 out represent the molar fraction of CO2 at the inlet and outlet,
respectively.
Theoretical Calculations
All the calculations were performed using the Vienna ab initio simulation package
(VASP)31 in the generalized gradient approximation (GGA-PBE). A plane-wave ba-
sis set with a cutoff energy of 450 eV and a (3 3 5 3 1) k-point grid generated with
the Monkhorst-Pack scheme32 were used. The atomic positions were relaxed until
the forces on all unconstrained atoms were %0.02 eV/A. DFT-D3 scheme33 for cor-
˚
rections of dispersion interactions and the DFT+U that better describes the local-
ized d electrons with U = 5.0 eV for Cu were employed. The copper catalyst was
modeled with Cu (221) surface which in the perfect surface model contains 64
Cu atoms built from the theoretically optimized cubic bulk Cu structure with a lat-
˚
tice parameter a = 3.61 A using (15 3 15 3 15) k-point grid. This optimized bulk
34
˚
parameter agrees with the experimental value of 3.61 A very well. Two defect
˚
surface models were constructed (Figure 4). A vacuum space of at least 10 A thick-
ness was applied to avoid artificial interactions between the slab and its periodic
neighbors. In all calculations, the bottom two stepped layers having 32 Cu atoms
were fixed at their equilibrium bulk positions, whereas the atoms in the top two
layers together with the adsorbates were allowed to fully relax. Transition states
were located using the nudged elastic band (NEB) method35 and verified by the
presence of one and only one imaginary frequency corresponding to the C–C
bond formation/breaking while there were no imaginary frequencies at all for
the initial and final states.
The relative stabilities of defect surfaces were first studied. Two defect types, Cu va-
cancy and O-doped, were considered. Figure 4 depicts the perfect surface (B),
defect surfaces with Cu vacancy (C), and O-doped (D) defects. The first layer Cu
atoms were colored differently to represent four unequal surface sites. The positions
of defects (vacancy and doped site) were determined by comparing the stability of
the defect locating at row 1 to row 4 (see Figure 4). Both Cu vacancy and doped ox-
ygen atom prefer the edge sites (Figure 4).
Chem 6, 1–17, October 8, 2020
15

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
All bond lengths and O–C–O angles of CO2* in Figures 4 and S9 were listed in
Table S3.
SUPPLEMENTAL INFORMATION
Supplemental Information can be found online at https://doi.org/10.1016/j.chempr.
2020.07.001.
ACKNOWLEDGMENTS
Funding: the authors thank the financial support from the National Science Founda-
tion of China (21932004, 91745108, and 21773105), the Ministry of Science and
Technology of China (2017YFB0702800), and the Graduate Scientific Research
Foundation of Nanjing University (2016CL05). We are grateful to the High-Perfor-
mance Computing Center (HPCC) of Nanjing University for doing the numerical cal-
culations in this paper on its blade cluster system. Shanghai Synchrotron Radiation
Facility of China is appreciated for help on XAS measurements (beamline BL14W1).
The authors thank the help from Dr. Yu Deng, Kewen Huang, Dr. Xiangke Guo, Dr.
Pei Sheng, Dr. Fanfei Sun, and Dr. Jianguo Wang.
AUTHOR CONTRIBUTIONS
W.D. conceived and organized the work. L.D., J.G., Y.C., and Z.Z. designed and
joined the experiments. L.D. performed the main experiments. Z.C. and T.S. per-
formed the theoretical calculations. M.L. and P.W. participated in the TEM experi-
ments. C.Y. and L.P. carried out the NMR analysis. T.C. and Y.Z. helped the EXAFS
measurements and data processing. N.X. and X.G. joined discussions. L.D. and W.D.
wrote the manuscript, and W.D. finalized the work. All authors discussed the results.
DECLARATION OF INTERESTS
The authors declare no competing interests.
Received: November 29, 2019
Revised: January 30, 2020
Accepted: July 1, 2020
Published: July 23, 2020
REFERENCES
1. Enthaler, S., von Langermann, J., and
Schmidt, T. (2010). Carbon dioxide and
formic acid—the couple for environmental-
friendly hydrogen storage? Energy Environ.
Sci. 3, 1207–1217.
5. An, B., Zhang, J., Cheng, K., Ji, P., Wang, C.,
and Lin, W. (2017). Confinement of ultrasmall
Cu/ZnOx nanoparticles in metal–organic
frameworks for selective methanol synthesis
from catalytic hydrogenation of CO2. J. Am.
Chem. Soc. 139, 3834–3840.
promoter on structure and performance of K/
Cu–Zn catalyst for higher alcohols synthesis
from CO2 hydrogenation. Catal. Lett. 143,
345–355.
10. Prieto, G., Beijer, S., Smith, M.L., He, M., Au, Y.,
Wang, Z., Bruce, D.A., de Jong, K.P., Spivey,
J.J., and de Jongh, P.E. (2014). Design and
synthesis of copper-cobalt catalysts for the
selective conversion of synthesis gas to ethanol
and higher alcohols. Angew. Chem. Int. Ed. 53,
6397–6401.
2. Ghoniem, A.F. (2011). Needs, resources and
climate change: clean and efficient conversion
technologies. Prog. Energy Combust. Sci. 37,
15–51.
6. Spivey, J.J., and Egbebi, A. (2007).
Heterogeneous catalytic synthesis of ethanol
from biomass-derived syngas. Chem. Soc. Rev.
36, 1514–1528.
3. Moret, S., Dyson, P.J., and Laurenczy, G. (2014).
Direct synthesis of formic acid from carbon
dioxide by hydrogenation in acidic media. Nat.
Commun. 5, 4017.
7. Kusama, H., Okabe, K., Sayama, K., and
Arakawa, H. (1997). Ethanol synthesis by
catalytic hydrogenation of CO2 over Rh-
FeSiO2 catalysts. Energy 22, 343–348.
11. He, Z., Qian, Q., Ma, J., Meng, Q., Zhou, H.,
Song, J., Liu, Z., and Han, B. (2016). Water-
enhanced synthesis of higher alcohols from
CO2 hydrogenation over a Pt/Co3O4 catalyst
under milder conditions. Angew. Chem. Int.
Ed. 55, 737–741.
4. Larmier, K., Liao, W.C., Tada, S., Lam, E., Verel,
R., Bansode, A., Urakawa, A., Comas-Vives, A.,
and Cope´ ret, C. (2017). CO2-to-methanol
hydrogenation on zirconia-supported copper
nanoparticles: reaction intermediates and the
role of the metal-support interface. Angew.
Chem. Int. Ed. 56, 2318–2323.
8. Nieskens, D.L.S., Ferrari, D., Liu, Y., and
Kolonko, R. (2011). The conversion of carbon
dioxide and hydrogen into methanol and
higher alcohols. Catal. Commun. 14, 111–113.
12. Chen, Y., Choi, S., and Thompson, L.T. (2016).
Low temperature CO2 hydrogenation to
alcohols and hydrocarbons over Mo2C
9. Li, S., Guo, H., Luo, C., Zhang, H., Xiong, L.,
Chen, X., and Ma, L. (2013). Effect of iron
16
Chem 6, 1–17, October 8, 2020

Please cite this article in press as: Ding et al., CO2 Hydrogenation to Ethanol over Cu@Na-Beta, Chem (2020), https://doi.org/10.1016/
j.chempr.2020.07.001
ll
Article
supported metal catalysts. J. Catal. 343,
147–156.
Functionalized iron oxide nanoparticles: mild
oxidation of benzyl C–H bonds exclusive to
carbonyls by molecular oxygen. ACS Catal. 4,
2746–2752.
characterisation of copper in Cu-Beta zeolite.
Solid State Sci. 3, 637–640.
13. Bai, S., Shao, Q., Wang, P., Dai, Q., Wang, X.,
and Huang, X. (2017). Highly active and
selective hydrogenation of CO2 to ethanol by
ordered Pd–Cu nanoparticles. J. Am. Chem.
Soc. 139, 6827–6830.
28. Graciani, J., Mudiyanselage, K., Xu, F., Baber,
A.E., Evans, J., Senanayake, S.D., Stacchiola,
D.J., Liu, P., Hrbek, J., Ferna´ndez Sanz, J., and
Rodriguez, J.A. (2014). Catalysis. Highly active
copper-ceria and copper-ceria-titania catalysts
for methanol synthesis from CO₂. Science 345,
546–550.
21. Liu, L., Dıaz, U., Arenal, R., Agostini, G.,
´
Concepcio´ n, P., and Corma, A. (2017).
Generation of subnanometric platinum with
high stability during transformation of a 2D
zeolite into 3D. Nat. Mater. 16, 132–138.
14. Wang, L., Wang, L., Zhang, J., Liu, X., Wang, H.,
Zhang, W., Yang, Q., Ma, J., Dong, X., Yoo, S.J.,
et al. (2018). Selective hydrogenation of CO2 to
ethanol over cobalt catalysts. Angew. Chem.
Int. Ed. 57, 6104–6108.
29. Zhao, Y.T., Cui, C.N., Han, J.Y., Wang, H., Zhu,
X.L., and Ge, Q.F. (2016). Direct C–C coupling
of CO2 and the Methyl Group from CH4
activation through facile insertion of CO2 into
Zn–CH3s-bond. J. Am. Chem. Soc. 138, 10191–
10198.
22. Zhang, Z., Li, Y., Gu, J., Ding, L., Xue, N., Peng,
L., Guo, X., Zhu, Y., Ma, J., and Ding, W. (2018).
The effect of electrostatic field on the catalytic
properties of platinum clusters confined in
zeolite for hydrogenation. Catal. Sci. Technol.
8, 6384–6395.
15. Song, Y., Chen, W., Zhao, C., Li, S., Wei, W., and
Sun, Y. (2017). Metal-free nitrogen-doped
mesoporous carbon for electroreduction of
CO2 to ethanol. Angew. Chem. Int. Ed. 56,
10840–10844.
30. Zhang, M., Yao, R., Jiang, H., Li, G., and Chen,
Y. (2017). Insights into the mechanism of acetic
acid hydrogenation to ethanol on Cu (111)
surface. Appl. Surf. Sci. 412, 342–349.
23. Wang, Z., Xu, Z.N., Peng, S.Y., Zhang, M.J., Lu,
G., Chen, Q.S., Chen, Y., and Guo, G.C. (2015).
High-performance and long-lived Cu/SiO2
nanocatalyst for CO2 hydrogenation. ACS
Catal. 5, 4255–4259.
16. Liao, F., Huang, Y., Ge, J., Zheng, W., Tedsree,
K., Collier, P., Hong, X., and Tsang, S.C. (2011).
Morphology-dependent interactions of ZnO
with Cu nanoparticles at the materials’
31. Kresse, G., and Hafner, J. (1993). Ab
initiomolecular dynamics for liquid metals.
Phys. Rev. B 47, 558–561.
interface in selective hydrogenation of CO2 to
CH3OH. Angew. Chem. Int. Ed. 50, 2162–2165.
24. Jiao, F., Li, J., Pan, X., Xiao, J., Li, H., Ma, H.,
Wei, M., Pan, Y., Zhou, Z., Li, M., et al. (2016).
Selective conversion of syngas to light olefins.
Science 351, 1065–1068.
32. Monkhorst, H.J., and Pack, J.D. (1976). Special
points for Brillouin-zone integrations. Phys.
Rev. B 13, 5188–5192.
17. Su, J., Zhang, Z., Fu, D., Liu, D., Xu, X.-C., Shi, B.,
Wang, X., Si, R., Jiang, Z., Xu, J., and Han, Y.-F.
(2016). Higher alcohols synthesis from syngas
over CoCu/SiO2 catalysts: dynamic structure
and the role of Cu. J. Catal. 336, 94–106.
25. Cheng, K., Gu, B., Liu, X., Kang, J., Zhang, Q.,
and Wang, Y. (2016). Direct and highly selective
conversion of synthesis gas into lower olefins:
design of a bifunctional catalyst combining
methanol synthesis and carbon–carbon
coupling. Angew. Chem. Int. Ed. 55, 4725–
4728.
33. Grimme, S., Antony, J., Ehrlich, S., and Krieg, H.
(2010). A consistent and accurate ab initio
parametrization of density functional
18. Kattel, S., Ramı´rez, P.J., Chen, J.G., Rodriguez,
J.A., and Liu, P. (2017). Active sites for CO2
hydrogenation to methanol on Cu/ZnO
catalysts. Science 355, 1296–1299.
dispersion correction (DFT-D) for the 94
elements H-Pu. J. Chem. Phys. 132, 154104.
34. Straumanis, M.E., and Yu, L.S. (1969). Lattice
parameters, densities, expansion coefficients
and perfection of structure of Cu and of Cu-in a
phase. Acta Crystallogr. A 25, 676–682.
26. Bersani, M., Gupta, K., Mishra, A.K., Lanza, R.,
Taylor, S.F.R., Islam, H.-U., Hollingsworth, N.,
Hardacre, C., de Leeuw, N.H., and Darr, J.A.
(2016). Combined EXAFS, XRD, DRIFTS, and
DFT study of Nano copper-based catalysts for
CO2 hydrogenation. ACS Catal. 6, 5823–5833.
19. Gu, J., Zhang, Z., Hu, P., Ding, L., Xue, N., Peng,
L., Guo, X., Lin, M., and Ding, W. (2015).
Platinum nanoparticles encapsulated in MFI
zeolite crystals by a two-step dry gel conversion
method as a highly selective hydrogenation
catalyst. ACS Catal. 5, 6893–6901.
35. Henkelman, G., and Jo´ nsson, H. (2000).
Improved tangent estimate in the nudged
elastic band method for finding minimum
energy paths and saddle points. J. Chem. Phys.
113, 9978–9985.
20. Li, L., Lv, J., Shen, Y., Guo, X., Peng, L., Xie, Z.,
and Ding, W. (2014). Hexadecylphosphate-
27. Ma´rquez, F., and Palomares, A. (2001). EXFAS
electron spectroscopy as a new tool of local
Chem 6, 1–17, October 8, 2020
17