Chemical Engineering Journal 485 (2024) 150093

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Local In-O-Pd Lewis acid-base pair boosting CO2 selective hydrogenation

to methanol
Yujie Shi a, b, 1, Qingqing Gu c, 1, Yi Zhao a, b, Yuehong Ren d, *, Bing Yang c, Jing Xu e,
Ying Zhang a, b, Chengsi Pan a, b, Yongfa Zhu d, Yang Lou a, b, *
a
Key Laboratory of Synthetic and Biological Colloids, Ministry of Education, School of Chemical and Material Engineering, Jiangnan University, Wuxi, Jiangsu 214122,
China
b
International Joint Research Center for Photoresponsive Molecules and Materials, Jiangnan University, Wuxi, Jiangsu 214122, China
c
Dalian National Laboratory for Clean Energy, Dalian Institute of Chemical Physics, 457 Zhongshan Road, Dalian 116023, China
d
Department of Chemistry, Tsinghua University, Beijing 100084, China
e
School of Food Science and Technology, Jiangnan University, Wuxi, Jiangsu 214122, China

A R T I C L E I N F O A B S T R A C T

Keywords: CO2 hydrogenation into valuable chemicals is industrially essential but very challenging since CO2 is chemically
CO2 hydrogenation inert and the selectivity is difficult to control. Here, we design and construct In2O3 supported Pd single atoms
Methanol (Pd1/In2O3 SAC) to boost the methanol yield (254.8 gmethanol⋅g−Pd1⋅h− 1) and selectivity (99.4 %) via the Lewis
Single-palladium site
acid-base pair, which activates the CO2 and forms the intermediate species of formate (HCOO*). The compre­
Lewis acid-base pair
hensive experimental and theoretical results confirm that the local Lewis acid-base pair is composed of the
isolated Pd atom coupling with the adjacent O and In atoms. The space–time yield (STY) of methanol over Pd1/
In2O3 SAC is around 1.7 times higher than the best reported catalysts under comparable conditions. The strategy
of constructing local Lewis acid-base pairs with single-atom site and metal oxides reported in this work opens
new avenues to develop highly selective catalysts for CO2 utilization.

1. Introduction coordination with two sites (C atom coordination with Lewis basic sites
while one O atom coordination with Lewis acidic sites) [9–14], thereby
Hydrogenation of CO2 to methanol is crucial, which not only pro­ promoting the formation of key intermediates like formate species
vides a significant route to alleviate the greenhouse effect caused by (HCOO*) for better harvesting value-added products. However, homo­
CO2, but also converts CO2 into highly valuable chemicals and geneous Lewis acid-base pair catalyst systems are difficult to separate
environmental-friendly engine fuel [1–6]. However, implementation of from the products and/or recycle [15,16]. In contrast, extensive efforts,
such a highly selective and desirable reaction is a challenging task due to for example, transition metal ions-containing zeolites [17], metal­
the thermodynamic stability and chemical inertness of CO2 [7,8]. –organic framework (MOF)-based catalysts [18,19], metal-embedded
Furthermore, the reverse water–gas shift (RWGS) or over-hydrogenation nitrogen-doped carbon materials [20] and tuning active sites on the
side reaction is difficult to avoid, forming undesired products such as metal-oxide interfaces [21,22], have been devoted to develop hetero­
carbon monoxide and methane. Hence, it is highly desired to develop geneous Lewis acid-base pair catalyst systems. When the metal sites are
robust catalysts capable of simultaneously modulating the activation downsized to atomic level, the atomic Lewis acid-base pairs formed with
process of CO2 and H2 for harvesting high selectivity and activity for CO2 isolated metal atom and active sites on support not only manipulate the
hydrogenation. electronic state of active metal centers via the electronic interaction, but
Lewis acid-base pairs, composed of the proximal Lewis acidic sites also provide isolated and uniform adsorption sites, thereby improving
and the Lewis basic sites, can efficiently and synergistically activate the catalytic ability of simultaneously modulating the activation process
small molecules such as H2 via heterolytic dissociation and CO2 via of CO2 and key intermediates.

* Corresponding authors at: Key Laboratory of Synthetic and Biological Colloids, Ministry of Education, School of Chemical and Material Engineering, Jiangnan
University, Wuxi, Jiangsu 214122, China(Y. Lou).
E-mail addresses: yhrenbitipe@mail.tsinghua.edu.cn (Y. Ren), yang.lou@jiangnan.edu.cn (Y. Lou).
1
Y. S. and Q. G. made equal contribution to this work.

https://doi.org/10.1016/j.cej.2024.150093
Received 1 January 2024; Received in revised form 12 February 2024; Accepted 28 February 2024
Available online 1 March 2024
1385-8947/© 2024 Elsevier B.V. All rights reserved.
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

In this work, we designed and constructed In2O3 supported single Pd and flushed three times with 10 bar H2 to remove the inside air, followed
atoms (denoted as Pd1/In2O3 SAC) with unique Pd1-O4 configuration. by filling with CO2 (0.55 MPa) and H2 (2.75 MPa) at room temperature.
The synergistic effects of monoatomic Pd as well as In and O sites Afterwards, it is heated to reaction temperature with stirring at a speed
contribute to the excellent catalytic performance by generating local In- of 700 rpm for 1 h to initiate the reaction. After reaction, the reactor is
O-Pd Lewis acid-base pairs, which favors CO2 adsorption and activation cooled down in the ice water bath. The gas phase including CO2, H2, CO,
and the formation of intermediate species of formate (HCOO*). As a CH4 is determined by Ruineng GC-3900II gas chromatography (GC)
result, the Pd1/In2O3 SAC exhibits superior performance for CO2 hy­ equipped with a TDX-01 column connected to a thermal conductivity
drogenation to methanol under mild reaction conditions (3.3 MPa @ detector (TCD) and a flame ionization detector (FID). The liquid mixture
200 ◦ C). Under comparable conditions, the TOF of Pd1 atoms in Pd1/ is collected by filtration and is analyzed using a Ruimin GC-2060 GC
In2O3 SAC (799.2 h− 1) is around 5.2 times higher than that of Pd NPs/ equipped with SE-54 column and flame ionization detector (FID), in
In2O3 and higher than most of the catalysts for this reaction in the open which a known amount of 1-Butanol is introduced into the liquid
literature. Fundamental origins triggering the excellent activity and mixture as an internal standard. The peaks are calibrated by the stan­
selectivity of Pd1/In2O3 SAC were discussed in depth via a series of dard chemicals from Sigma-Aldrich.
combined experimental and computational analyses, covering the The conversion of CO2 (XCO2) was calculated using Eq. (1).
optimal mechanism, reaction kinetics, and so on.
[CH3OH]n + [CO]n + [CH4]n
XCO2(%) = × 100% (1)
[CO2]n
2. Experimental
where [CH3OH]n represent the molar amount of CH3OH in the liquid
2.1. Materials
mixture, [CO]n and [CH4]n represent the molar amount of CO and CH4 in
the outlet gas, respectively. [CO2]n represents molar amount of CO2 in
All chemicals were used without further purification. Indium chlo­
the inlet gas.
ride, sodium carbonate anhydrous and ethylene glycol were purchased
CO selectivity (SCO), CH4 selectivity (SCH4) and CH3OH selectivity
from innochem. Palladium(II) nitrate dihydrate was purchased from
(Smethanol) were calculated using Eqs. (2)–(4), respectively.
Sigma-Aldrich.
[CO]n
SCO (%) = × 100% (2)
2.2. Preparation of In2O3 samples [CH3OH]n + [CO]n + [CH4]n

The details are similar to those in our previous publication. In a SCH4 (%) =
[CH4]n
× 100% (3)
typical synthesis of In2O3 samples, 4.42 g of Indium chloride (InCl3) is [CH3OH]n + [CO]n + [CH4]n
dissolved into 80 mL of ethylene glycol and the system is degassed by a
mechanical pump. Then, the mixture is heated to 160 ◦ C and kept in Ar Smethanol (%) =
[CH3OH]n
× 100% (4)
flow for 40 min under vigorous stirring. Na2CO3 aqueous solution (200 [CH3OH]n + [CO]n + [CH4]n
mL, 0.22 M) heated at 80 ◦ C is dropwise added into the solution, and the
mixture is further aged at 160 ◦ C for 1 h under vigorous stirring. After where [CH3OH]n represent the molar amount of CH3OH in the liquid
washing and filtering, the product obtained is dried at 60 ◦ C overnight mixture, [CO]n and [CH4]n represent the molar amount of CO and CH4 in
and calcined at 350 ◦ C for 4 h in air. the outlet gas.
The space–time yield of CH3OH (STYmethanol, gmethanol⋅g−Pd1 h− 1) was
calculated using Eq. (5).
2.3. Preparation of Pd1/In2O3
[CH3 OH]n × Mmethanol
The Pd1/In2O3 SACs are prepared via an adsorption method as STYmethanol = (5)
gcat × [Pd] × t
shown in previous publication [23]. In order to synthesize Pd1/In2O3
SAC, the In2O3 nanorods are firstly dispersed into the aqueous solution where [CH3OH]n represent the molar amount of CH3OH in the liquid
at room temperature and then the palladium (II) nitrate dihydrate with mixture; Mmethanol is the relative atomic mass of CH3OH (32.04 g/mol);
corresponding amounts is slowly dropped in above mixture solution. gcat is the mass of catalyst loaded (in gram); [Pd] content is the content
The mixtures are filtered out after stirring at room temperature for 2 h, of Pd in catalyst measured by ICP-OES (wt %), and t is the reaction time.
and then the obtained sample is dried at 60 ◦ C for 12 h in air. The The turnover frequency (TOF) of Pd was calculated using Eq. (6).
samples are calcined in air at 400 ◦ C for 4 h. The actual Pd loading is
[CO2 ]n × XCO2 × MPd ( − 1 )
0.80 wt% as measured by ICP-OES (AGILENT ICP-OES 730). TOF = h (6)
gcat × [Pd] × t
2.4. Preparation of Pd NPs/In2O3
where [CO2]n is the molar amount of CO2 in the inlet gas; XCO2 is the
conversion of CO2; MPd is the relative atomic mass of Pd (106.42 g/mol);
The In2O3 supported nano-Pd particles (Pd NPs/In2O3) catalysts are
gcat is the mass of catalyst loaded (in gram); [Pd] content is the content
prepared via the NaBH4 reduction method as the control catalyst. In
of Pd in catalyst measured by ICP-OES (wt %), and t is the reaction time.
order to synthesize Pd NPs/In2O3 catalysts, the In2O3 and the palladium
(II) nitrate dihydrate with corresponding amounts are firstly dispersed
into the aqueous solution at room temperature. Subsequently, 6 mL 0.2 2.6. Catalyst characterization
M NaBH4 was slowly dropped into the above solution under vigorous
stirring. After stirring for 6 h, the mixture was filtered out and then the HAADF-STEM images were acquired on a JEM-ARM200F TEM/
obtained catalyst was dried at 60 ◦ C for 12 h in air. The actual Pd loading STEM with a resolution of 0.08 nm. Before collecting data, the catalysts
is 0.82 wt% as measured by ICP-OES (AGILENT ICP-OES 730). were dispersed by ultrasonic machine in ethanol. Then a droplet of
dispersed catalyst solution was dropped onto a copper grid coated with
2.5. Evaluation of the catalytic performance lacey carbon film.
Diffuse Reflectance Infrared Fourier Transform Spectroscopy
The hydrogenation of CO2 is performed in a stainless-steel autoclave (DRIFTS) of NO adsorption on Pd/In2O3 catalysts was conducted on a
(TKA 50 mL). Without the pre-reduction treatment, 50 mg of catalyst Nicolet 6700 spectrometer with a MCT detector. The in-situ cell was
and 5 mL of DMF are mixed into the reactor. Then the reactor is sealed equipped with ZnSe windows. The DRIFTS spectra were acquired with a

2
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

resolution of 4 cm− 1 and 64 scans, and the data was recorded in approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) func­
Kubelka-Munk units. When collecting the data of NO-DRIFTS on Pd1/ tional [25] was used to calculate the exchange–correlation interactions,
In2O3 and Pd NPs/ In2O3 samples, the samples were firstly heated at and the DFT-D method by Grimme [26] was applied to describe the vdW
120 ◦ C for 40 min to remove the possible H2O and other impurities on interaction. The core-electron interaction was described by the DFT
the surface. And then the samples were cooled down to 25 ◦ C and the semi-core pseudopotential (DSPP) core treatment [27], which replaced
background spectrum was collected before the background was stable. core electrons by a single effective potential. The double numerical plus
Then NO (0.5 vol% NO/N2) molecules were introduced to adsorb on the polarization (DNP) basis set with global orbital cutoff of 4.4 Å was
catalyst surfaces until the gaseous NO peak intensity stopped increasing employed. The geometry optimization was achieved until the conver­
(~15 min), suggesting the saturation of NO adsorption on the catalyst gence criteria of energy and force on each atom were below 2.0 × 10− 5
surfaces. Then pure Ar was used to purge out the gaseous NO from the Ha and 0.004 Ha/Å, respectively. The minimum-energy reaction path­
sample cell so that the chemically adsorbed/bound NO species could be ways and corresponding transition states (TS) were obtained by the
detected. After collecting the background signal, the whole procedure complete linear synchronous transit and quadratic synchronous transit
was recorded by FTIR equipped with MCT detector (cooled by liquid (LST/QST) method [28,29], and which was verified by vibrational fre­
nitrogen) including the NO adsorption and Ar purging process. For CO2 quencies. Moreover, the zero-point energy and the entropy contribu­
hydrogenation, the similar processes were followed, but the temperature tions at 475 K and 1 atm to obtain the Gibbs free energies of elementary
was set to 200 ◦ C and the ratio of H2/CO2 was set to 5. reaction pathways was considered in the calculation process.
The Pd K-edge extended X-ray absorption fine structure (EXAFS) of The In2O3 (4 3 1) surface was constructed from the optimized In2O3
the Pd1/In2O3, Pd NPs/ In2O3 and Pd foil experiments were performed bulk unit cell with the lattice parameters of 10.28 Å. The In2O3 (4 3 1)
on the 20-BM-B beamline of the Advanced Photon Source at Argonne surface was modeled in a (1 × 1) supercell with three periodic atomic
National Laboratory, operated at 7.0 GeV and an average current of 100 layers containing 72 O atoms and 48 In atoms and was separated by a
mA. The XAFS spectra were collected in fluorescence mode due to low vacuum of 15 Å. Based on our experimental results and model screening,
Pd concentration. The radiation was monochromatized by a Si (1 1 1) the Pd1/In2O3 SAC catalyst was modelled by substituting a defective
double-crystal monochromator. Data reduction, data analysis, and surface In atom. Moreover, the bottom one layer of the Pd1/In2O3 SAC
EXAFS fitting were performed with the Athena and Artemis software was fixed in their equilibrium bulk positions, and the rest were allowed
packages. The energy calibration of the catalysts was conducted through to relax in the geometry optimization.
a standard Pd foil, which as a reference was simultaneously measured.
Raman spectra were collected in the range of 100–800 cm− 1 on an 3. Results and discussion
inVia reflex Renishaw Raman Spectrometer and excited with a 532 nm
laser source, and the resolution was within 1 cm− 1. The laser power was 3.1. Construction and identification of Pd single atoms in Pd1/In2O3 SAC
set at 7 mW, and the integration time was 5 s.
X-ray photoelectron spectroscopy (XPS) analysis was carried out on a The Pd1/In2O3 SAC was synthesized via an adsorption method
Thermo Scientific K-Alpha Nexsa XPS spectrometer equipped with Al Kα [23,30] and the details of synthesis are listed in experimental proced­
X-ray under ultrahigh vacuum conditions. The spectrometer was oper­ ures. High-angle annular dark-field scanning transmission electron mi­
ated at 12 kV and 6 mA. All measured binding energies were calibrated croscopy (HAADF-STEM) images confirm that the fabricated In2O3
with the C 1 s peak (284.8 eV). mainly consists of (4 3 1) crystal planes. When Pd single atoms were
X-ray diffraction (XRD) patterns were recorded on a Bruker D8 focus deposited on In2O3 (Pd1/In2O3 SAC), the transmission electron micro­
diffraction spectrometer Cu Kα radiation with a scanning angle (2 theta) scopy (TEM) and HAADF-STEM images acquired from different areas of
of 5◦ to 90◦ at a speed of 2◦ /min, operated at 40 kV and 40 mA at room fresh Pd1/In2O3 SAC reveal that no Pd clusters and particles can be
temperature. observed (Fig. 1a and S1). It indicates that Pd species are atomically
Hydrogen temperature programed reduction (H2-TPR) was con­ dispersed on the surface of In2O3. Because of the limitation on Z-contrast
ducted on the Micromeritics AutoChem 2920 analyzer. The H2-TPR between Pd atoms and In2O3, it is difficult to directly image the atom­
technique was employed to analyze the reducibility of the catalysts. 100 ically dispersed Pd species by HAADF-STEM. The energy dispersive
mg of the as prepared (calcined) catalyst was placed in a quartz reactor spectroscopy (EDS) elemental maps further confirm that the Pd species
and pretreated in an N2 flow (20 mL min− 1) at 150 ◦ C for 30 min to are atomically dispersed in selected regions of the Pd1/In2O3 SAC
remove the adsorbed water and carbonates, and then the sample was (Fig. 1b).
cooled down to 50 ◦ C in a N2 flow (20 mL min− 1). After that, the sample The crystalline structure of In2O3 support and Pd/In2O3 catalysts was
was heated from 50 to 500 ◦ C with a heating rate of 10 ◦ C min− 1 under a investigated by XRD analyses. As shown in Fig. 1c, the characteristic
mixture of 5 % H2-95 % N2 (40 mL min− 1). The amount of H2 con­ peaks of In2O3 that appear at 21.6◦ , 30.7◦ , 35.6◦ , 51.1◦ , and 60.9◦ can be
sumption was determined by a gas chromatograph with a thermal ascribed to (2 1 1), (2 2 2), (4 0 0), (4 4 0), and (6 2 2) crystalline planes of
conductivity detector (TCD). cubic In2O3 (PDF#06-0416). The XRD patterns of Pd1/In2O3 SAC and Pd
Temperature-programmed desorption of carbon dioxide (CO2-TPD) NPs/In2O3 suggest that the original crystal structure of In2O3 is main­
was conducted on a Micromeritics AutoChem II 2920 chemisorption tained, indicating that the metal species do not change the lattice
analyzer. Typically, 100 mg of catalyst was placed in a U-shaped quartz structure of In2O3. In addition, no characteristic diffraction peaks of Pd
tube. A 10 % H2/Ar gas mixture with a flow rate of 30 mL min− 1 was particles are observed from XRD, which reveals that palladium species
used as the reductant. After reduction at 300 ◦ C for 1 h, the tube was are highly dispersed in In2O3.
cooled to 50 ◦ C under flowing argon, followed by CO2 adsorption (30 As shown in Raman spectra (Fig. S2), the peak at around 130 cm− 1 is
mL min− 1) at 50 ◦ C for 0.5 h. Afterward, the flowing Ar (30 mL min− 1) ascribed to the In-O vibration of the InO6 structure. The two peaks at 306
was injected to remove residual and physically adsorbed CO2. The whole and 493 cm− 1 are assigned to the bending vibration and stretching vi­
system was then purged with Ar and heated from 50 to 600 ◦ C at a rate of bration of the octahedral InO6. The peak centered at 364 cm− 1 belongs
5 ◦ C min− 1. The effluent gas was monitored by a TCD detector. The TCD to the stretching vibration of In-O-In (ν (In-O-In)). The peak at 630 cm− 1
signal was normalized for further analysis. is ascribed to the stretching vibration of In-O bonds. Moreover, the ratio
of two peaks (I2/I1) at 306 cm− 1 (I1) and 368 cm− 1 (I2) is adopted to
2.7. Computational method intuitively compare the amount of oxygen vacancies [31,32]. The I2/I1
values for fresh In2O3, Pd NPs/In2O3 and Pd1/In2O3 SAC are 0.13, 0.43
Spin-polarized density functional theory (DFT) calculations were and 0.70, respectively, which indicates more oxygen vacancies on Pd1/
implemented in the Dmol3 Module [24]. The generalized gradient In2O3 SAC compared with that on In2O3 and Pd NPs/In2O3. More oxygen

3
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

Fig. 1. Resolved structures of the Pd1/In2O3 SAC. (a) HAADF-STEM images of the synthesized Pd1/In2O3 SAC. (b) Energy dispersive spectroscopy (EDS) mapping
images of the Pd1/In2O3 SAC. (c) XRD patterns of In2O3, Pd NPs/ In2O3, and Pd1/In2O3 SAC. (d) The k3-weighted Pd K-edge Fourier transform EXAFS spectra of the
Pd1/In2O3 SAC as well as Pd foil and PdO references. (e) Experimental and fitting Pd K-edge EXAFS curves of the Pd1/In2O3 SAC. (f) DFT calculated stable Pd1-O4
configuration of the Pd1 active site. The WT-EXAFS plot of the Pd1/In2O3 SAC (g), Pd foil (h) and PdO (i).

vacancies on Pd1/In2O3 SAC may originate from the Pd single atoms Furthermore, the wavelet transform (WT) of the Pd K-edge EXAFS
distributed on In2O3. oscillations was analyzed to better reveal the coordination environment
Next, three procedures were conducted to realize the local coordi­ of Pd atoms in Pd1/In2O3 SAC as shown in Fig. 1g–i. The WT contour
nation environment of the Pd single-atom sites in the Pd1/In2O3 SAC. plots of Pd1/In2O3 SAC present one intensity maximum at approxi­
First, the extended X-ray absorption fine structure (EXAFS) was carried mately 8.4 Å− 1 that can be assigned to the Pd-O coordination and no
out. The Pd K-edge EXAFS spectra of Pd1/In2O3 SAC (Fig. 1d) show a intensity maximum related to Pd-Pd or Pd-O-Pd coordination can be
sharp peak located at 1.5 Å (phase uncorrected), which can be assigned observed compared with that of the reference Pd foil and PdO samples.
to Pd-O coordination [33,34]. Furthermore, the Pd K-edge EXAFS Meanwhile, there is a weak intensity at approximately 12.8 Å− 1 that can
spectra show no significant peak associated with the Pd-Pd path be assigned to the Pd-O-In scattering. The results further confirm that
(centered at 2.5 Å, phase uncorrected) [35], which indicates that the Pd the Pd species are atomically dispersed on the Pd1/In2O3 SAC.
species are atomically dispersed on In2O3 instead of nanoparticles or Thirdly, the crystal model built for DFT simulation (Fig. S3) was used
multi-core clusters, and further verifies the EDS and XRD results. The to fit the EXAFS data (Fig. 1d). The fitting data fits the original EXAFS
weak peak located at 2.90 Å can be assigned to the Pd-O-In coordination data of Pd species well as shown in Fig. 1e and Table S1 and S2, which
path (phase uncorrected) [36,37]. secures the reliability of the fitting results on the local structure of Pd

4
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

single-atom species in Pd1/In2O3 SAC. In details, each Pd atom in the The unique electronic states of Pd1/In2O3 SAC are further verified by
Pd1/In2O3 SAC is averagely coordinated with four O atoms (CN = 3.9 ± X-ray photoelectron microscopy (XPS, Fig. 2b, S5, S6 and S7). The
0.3) in the Pd-O shell and averagely coordinated with one In atom (CN binding energies (BEs) of Pd 3d5/2 in Pd1/In2O3 SAC are solely located at
= 1.1 ± 0.4) in the Pd-O-In shell. Combined with the DFT simulation 336.7 eV is higher than that of Pd0 (BEs = 335.5 eV) [44] but lower than
model and EXAFS results, it can be concluded that the Pd1 atoms are that of Pd2+ (BEs = 337.2 eV) [4], which can be assignable to the Pdδ+
anchored at the In vacancy sites of In2O3, which forms a unique Pd1-O4 species (0 < δ < 2). Alternatively, the Pd species in Pd NPs/In2O3 sample
configuration (Fig. 1f and S4). show two oxidation states with Pd 3d5/2 binding energy at 335.5 eV and
336.7 eV reveals the existence of Pd0 and Pdδ+ (0 < δ < 2) species,
3.2. Identifying the electron-deficiency states of Pd single atoms and respectively. By calculating the peak area of these two Pd species, the
interfacial Lewis acid-base pairs over Pd1/In2O3 SAC content of Pd0 is estimated to be 76 %, while Pdδ+ (0 < δ < 2) is esti­
mated to be 24 %. Such feature of oxidation states over Pd1/In2O3 SAC
To probe the electronic states and geometric structure of Pd species can be further confirmed via the X-ray absorption near edge structure
in Pd1/In2O3 SAC, in-situ diffuse reflectance infrared Fourier transform (XANES) data. The white line intensity of the Pd species in the Pd1/In2O3
spectroscopy of NO (NO-DRIFTS) was performed on the Pd1/In2O3 SAC SAC is higher than that of Pd foil whereas lower than that of PdO
and Pd NPs/In2O3 (Fig. 2a). For NO adsorption over Pd NPs/In2O3, the (Fig. 2c), corroborating the in-situ NO-DRIFTs (Fig. 2a) and XPS results
adsorption peak centered at 1558 cm− 1 is associated with the adsorption (Fig. 2b). Those results confirm that the Pd species in Pd1/In2O3 SAC
of NO on the surface of Pd NPs/In2O3 in the form of nitrate ions [38,39]. possess unique electron-deficiency state, which indicates that the single-
The absence of the adsorbed NO species on Pd species in the region of atom Pd species in Pd1/In2O3 SAC possess Lewis acid properties.
1700 to 1900 cm− 1 confirms the absence of cationic Pd species in Pd The presence of Lewis acid-base pairs is further verified both theo­
NPs/In2O3. For NO adsorption over Pd1/In2O3 SAC, the adsorption retically and experimentally. Firstly, the electrostatic potential figure
peaks centered at 1540, 1556, 1563 and 1650 cm− 1 are associated with (Fig. 2d) of the surface model of Pd1/In2O3 SAC reveals that the presence
the adsorption of NO on the surface of Pd1/In2O3 SAC in the form of of monoatomic Pd creates a negative potential relative to In atom, while
nitrate ions [39–41]. While the peak centered at 1748 cm− 1 can be the Pd atom remains positively charged relative to O atom. Moreover,
assigned to the linearly adsorbed NO species on isolated Pd atom on the deformation electron density (Fig. S8) confirms the electron-
Pd1/In2O3 SAC (Pd1-NO nitrosyl species) [40,42,43], which corrobo­ deficiency state of Pd1 atoms. The Mulliken charge profile (Fig. S9)
rates the results that Pd species in Pd1/In2O3 SAC are atomically demonstrates the relative charge density of adjacent In, O and Pd atoms.
dispersed on In2O3. As such, a local In-O-Pd Lewis acid-base pair is formed between the Pd

Fig. 2. The electronic structure and chemical properties of the Pd/In2O3 catalysts. (a) In situ NO-DRIFTS adsorbed on Pd1/In2O3 SAC and Pd NPs/In2O3 at
30 ◦ C. (b) Pd 3d XPS peak of the fresh Pd1/In2O3 SAC and Pd NPs/In2O3. (c) Pd K-edge X-ray absorption near-edge structure (XANES) spectra. (d) Electrostatic
potential maps of perfect In2O3 (upper) and Pd1/In2O3 SAC (lower) with an isosurface of 0.04 electron/Å3.

5
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

single atom and In2O3. Furthermore, the formation of oxygen vacancy 84.7 % respectively under the same reaction condition (Fig. S18). These
promotes the linked In atom to produce empty electronic orbit, which results reveal that the Lewis acid-base pairs of Pd1/In2O3 SAC possess
enables to accept the electrons from oxygen atom of CO2 molecules and remarkable catalytic activity, selectivity and stability for CO2 hydroge­
further promotes the activation of CO2 molecules. nation to methanol.
Secondly, the CO2-TPD experimental results (Fig. S10) further verify Furthermore, Pd1/In2O3 SAC shows excellent intrinsic activity for
the strong basicity of Pd1/In2O3 SAC. The CO2-TPD profile is divided CO2 hydrogenation (Fig. 3b) with the TOF up to 799.2 h− 1 at 200 ◦ C,
into two categories, below 200 ◦ C and 200–650 ◦ C, matching weak basic which is higher than most of the catalyst reported in the literature
sites and moderate basic sites, respectively [45]. Pd1/In2O3 SAC and (Table S3) under comparable condition. The TOF of Pd1 atoms in Pd1/
pure In2O3 have almost no CO2 desorption peak below 150 ◦ C while Pd In2O3 SAC (799.2 h− 1) is also around 5.2 times higher than that of Pd
NPs/In2O3 has a low-temperature (around 80 ◦ C) peak, which is NPs/In2O3 (154.8/h) under the same reaction conditions. Besides, the
assigned to the physical interaction with CO2 [46]. Compared with pure STY of methanol over Pd1/In2O3 SAC is up to 254.8 gmethanol⋅g−Pd1 h− 1 at
In2O3 and Pd NPs/In2O3, Pd1/In2O3 SAC exhibits a higher temperature 200 ◦ C, which is around 1.7 times higher than the best catalyst reported
CO2 desorption peak at around 582 ◦ C, indicating the stronger basic sites in the literature (0.2 %Pt/MoS2 catalyst with a STY of 152.2
of Pd1/In2O3 SAC, which has a stronger interaction of CO2, promoting gmethanol⋅g−Pt1 h− 1 at 210 ◦ C [51]; Fig. 3d, S19 and Table S3) under
the adsorption and activation of CO2 [47,48]. comparable condition. On the other hand, the specific reaction rate of
1
Thirdly, the Pyridine FT-IR experimental results (Fig. S11) further Pd1/In2O3 SAC (2.1 gmethanol⋅g−catalyst h− 1 at 200 ◦ C) is around 2.3 times
1
confirm the strong acidity of Pd1/In2O3 SAC. The peaks at 1608, 1589, and 4.2 times higher than that of Pd NPs/In2O3 (0.9 gmethanol⋅g−catalyst h− 1
1441 cm− 1 are assigned to Lewis acid sites [49]. The peaks at 1540 cm− 1 1
at 200 ◦ C) and pure In2O3 (0.5 gmethanol⋅g−catalyst h− 1 at 200 ◦ C), respec­
are assigned to Brønsted acid sites [50]. The peaks at 1577 and 1489 cm- tively. Moreover, to better understand the kinetic behavior of the Pd1/
1
are assigned to pyridine species interacting with these two kinds of acid In2O3 SAC, the apparent activation energy (Ea) has been investigated
sites [49]. Obviously, the strength of Lewis acid sites over Pd1/In2O3 (Fig. S20). Under the conditions of kinetic control, the Ea of the Pd1/
SAC is stronger than that of Pd NPs/In2O3. The CO2-TPD and Pyridine In2O3 SAC (71.2 kJ mol− 1) is much lower than that of Pd NPs/In2O3
FT-IR experimental results confirm that the co-presence of surface acidic (82.4 kJ mol− 1) and pure In2O3 (112.3 kJ mol− 1), which indicates that
sites and basic sites. the catalytic process of hydrogenation of CO2 is intrinsically boosted on
the Pd1/In2O3 SAC compared with that of Pd NPs/In2O3 and In2O3.
These results reveal that the Lewis acid-base pairs of Pd1/In2O3 SAC
3.3. Enhanced catalytic selectivity and activity for methanol on Pd single
possess remarkable catalytic performance for CO2 hydrogenation to
atoms
methanol.
H2 dissociation and CO2 activation is considered as the crucial factor
To investigate the catalytic performance of the synthesized Pd1/
in determining the overall catalytic activity for CO2 hydrogenation. The
In2O3 SAC, the CO2 hydrogenation was conducted and compared with
H2-TPR experimental results indicate that H2 can be easily dissociated
the cases of Pd NPs/In2O3 and pure In2O3 in a batch reactor. Under the
on Pd1/In2O3 SAC. As shown in Fig. S21, the low-temperature H2
same reaction condition (50 mg catalysts, 200 ◦ C, 3.3 MPa, H2/CO2 =
reduction peak on Pd1/In2O3 SAC (centered at 59, 73 and 133 ◦ C) is
5:1, 1 h), the CO2 conversion on Pd1/In2O3 SAC is 33 %, which is around
much lower than that on Pd NPs/In2O3 (centered at 179 ◦ C) and In2O3
2.4 times and 4.1 times higher than that of Pd NPs/In2O3 and pure In2O3,
(centered at 196 and 224 ◦ C), which indicates that the Pd single atoms
which indicates that the Pd1/In2O3 SAC possesses much higher catalytic
anchored to In2O3 possess better capability for H2 dissociation and
activity for CO2 hydrogenation. The selectivity to methanol over Pd1/
activation.
In2O3 SAC is 99.4 %, with CO and CH4 as minor product under mild
condition (Fig. S12 and S13). In contrast, Pd NPs/In2O3 and pure In2O3
exhibit lower methanol selectivity of 98.0 % and 96.7 % respectively 3.4. Activation mechanism for methanol formation over Pd1/In2O3 SAC
under the same reaction condition (Fig. 3a). Moreover, the Pd1/In2O3
SAC can sustain the high methanol selectivity (>98 %) for at least five DFT calculations were performed to clarify the key role of Lewis acid-
cycles (Fig. S14, S15, S16 and S17). Even when the reaction lasts for 5 h, base pair in achieving excellent CH3OH selectivity over the Pd1/In2O3
the selectivity of Pd1/In2O3 SAC to methanol is slightly decreased to SAC (Fig. 4). Our DFT calculation results reveal that the direct dissoci­
97.4 %, with CO and CH4 as minor product under mild condition (50 mg ation of CO2 to CO needs to overcome a barrier of 5.58 eV while the
catalysts, 200 ◦ C, 3.3 MPa, H2/CO2 = 5:1). In contrast, Pd NPs/In2O3 formation of carboxyl (COOH) from CO2 needs to overcome activation
and pure In2O3 exhibit much lower methanol selectivity of 87.1 % and barriers of 2.98 eV. But both of them are higher than that of the

Fig. 3. The catalytic performance of pure In2O3, Pd NPs/In2O3 and Pd1/In2O3 SAC for CO2 hydrogenation. (a) Methanol selectivity. (b) TOF of CO2 hydrogenation at
200 ◦ C. (c) Comparison of the selectivity and productivity of methanol over Pd1/In2O3 SAC with the most efficient catalysts reported in the open literature. Reaction
conditions: 50 mg catalyst in 5 mL DMF with 3.3 MPa reactant gas (H2/CO2 = 5:1). The catalysts mentioned in Fig. 3d are listed in Table S3 and the references cited
in supporting information.

6
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

Fig. 4. Activation mechanism for methanol formation over Pd1/In2O3 SAC. (a) Relative Gibbs free energy (black) and the corresponding activation energy
barriers (red) calculated at 475 K for the hydrogenation of CO2 to methanol on Pd1/In2O3 SAC. The calculated energies are referenced to the surface slab and CO2(g)
+ 3H2(g). The “g” and “*” represent gaseous and adsorbed states, respectively. (b) Optimized geometries of the corresponding transition states for CO2 hydrogenation
toward methanol over Pd1/In2O3 SAC. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

formation of formate (HCOO) from CO2 (2.81 eV). Therefore, the CO2 In-O-Pd sites in the Pd1/In2O3 SAC are the key active sites for promoting
activation and conversion proceeding the HCOO pathway will be more the transformation of CH3OH products.
energy favored compared with the CO and/or COOH pathway. To further verify the above reaction mechanism of methanol for­
Fig. 4a shows the relative Gibbs free energy and corresponding mation over Pd1/In2O3 SAC, the surface species evolved during the CO2
activation barriers of each elementary reaction, with the geometry of the hydrogenation reaction were monitored using in-situ DRIFTS. Fig. S23
key intermediates in the catalytic process. The corresponding transition shows the infrared vibration peaks obtained from the intermediate
states are presented in Fig. 4b. The structural details of the initial re­ species when Pd1/In2O3 SAC was exposed to CO2/H2 gas at 200 ◦ C. The
actants, intermediates and final products are provided in the Supporting peaks observed at 1363, 1375, 1577 cm− 1 were attributed to the pres­
Information (Fig. S22). As shown in Fig. 4a, CO2 hydrogenation follows ence of formate (HCOO*) species. [52] The peaks observed at 1458
the HCOO* intermediate pathway, which is first triggered by the acti­ cm− 1 were attributed to the H3CO* species. At the same time, the bands
vation of CO2 adsorption on the In-O-Pd site, and then the adsorbed related to HCO3* (3600–3800 cm− 1) and CO3* (1389 cm− 1) were also
CO2* is generated into CO3* intermediate by overcoming the lower observed due to CO2 adsorbed on the surface hydroxyl group. [47,53]
energy barrier of 0.49 eV, and then generated into HCOO* intermediate These results coincide well with the reaction mechanism identified by
through protonation (O-H) and hydrogenation (C-H). Subsequently, the DFT where the above key intermediates are formed during CO2 hydro­
formed HCOO* is further hydrogenated into H2COO* intermediate with genation to methanol.
an energy of 0.69 eV, which needs to overcome the higher energy barrier
of 3.14 eV. Then, H2COO* is further hydrogenated into H2COOH* and 4. Conclusions
H2CO* intermediates through protonation (O-H), with the correspond­
ing energy barriers of 0.64 eV and 1.11 eV, respectively. The formed To conclude, constructing Lewis acid-base pairs centered by isolated
H2CO* further interacts with H on adjacent O atoms and is transformed Pd species on Pd1/In2O3 SAC enables to efficiently convert CO2 to
into H3CO* intermediate through the energy barrier of 1.01 eV. Finally, methanol, which significantly enhances the CO2 conversion (33 %) and
the H atoms adsorbed on Pd protonate H3CO* to generate the final methanol selectivity (99.4 %). Detailed experiments and characteriza­
product CH3OH*, releasing 0.67 eV. Further comparison of the activa­ tions combined with DFT calculations fully demonstrate the unique local
tion energy barriers of each elementary reaction shows that TS4 should In-O-Pd Lewis acid-base pairs contribute to the facile adsorption and
be the rate-determining step of the reaction. In addition, according to activation of CO2, which consequently boosts CO2 selective hydroge­
Fig. S22, all intermediates are nearly adsorbed on the In-O-Pd sites, nation to methanol. Although further refinement on this Pd1/In2O3 SAC
where intermediates tightly bind to the In sites, and H2 molecules are may enhance the catalytic performance for CO2 hydrogenation, the
more easily activated by the Pd sites. Based on the above discussion, the tactics of designing atom-precision active sites reported here is believed

7
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

to shed the light on developing highly selective catalysts for CO2 [12] A.E. Ashley, A.L. Thompson, D. O’Hare, Non-metal-mediated homogeneous
hydrogenation of CO2 to CH3OH, Angew. Chem. Int. Ed. 48 (2009) 9839–9843.
hydrogenation.
[13] M.A. Courtemanche, M.A. Legare, L. Maron, F.G. Fontaine, Reducing CO2 to
methanol using frustrated Lewis pairs: on the mechanism of phosphine-borane-
CRediT authorship contribution statement mediated hydroboration of CO2, J. Am. Chem. Soc. 136 (2014) 10708–10717.
[14] C.M. Momming, E. Otten, G. Kehr, R. Frohlich, S. Grimme, D.W. Stephan, G. Erker,
Reversible metal-free carbon dioxide binding by frustrated Lewis pairs, Angew.
Yujie Shi: Writing – original draft, Formal analysis. Qingqing Gu: Chem. Int. Ed. 48 (2009) 6643–6646.
Resources, Formal analysis, Data curation. Yi Zhao: Writing – original [15] S. Shyshkanov, T.N. Nguyen, F.M. Ebrahim, K.C. Stylianou, P.J. Dyson, In situ
formation of frustrated Lewis pairs in a water-tolerant metal-organic framework for
draft, Investigation, Formal analysis. Yuehong Ren: Writing – original the transformation of CO2, Angew. Chem. Int. Ed. 58 (2019) 5371–5375.
draft, Software, Formal analysis. Bing Yang: Software, Investigation. [16] Y.-B. Huang, F. Gu, B. Hu, K.-M. Li, Z.-M. Xu, J. Luo, Q. Lu, Flexible
Jing Xu: Formal analysis. Ying Zhang: Formal analysis. Chengsi Pan: borane–nitrogen frustrated Lewis pair organic microsphere for selective alkyne
hydrogenation, Chem. Mater. 35 (2023) 5752–5763.
Formal analysis. Yongfa Zhu: Resources, Formal analysis. Yang Lou:
[17] Y. Chai, B. Qin, B. Li, W. Dai, G. Wu, N. Guan, L. Li, Zeolite-encaged mononuclear
Writing – review & editing, Writing – original draft, Funding acquisition, copper centers catalyze CO2 selective hydrogenation to methanol, Nat. Sci. Rev. 10
Conceptualization. (2023) nwad043.
[18] K. Yang, J. Jiang, Computational design of a metal-based frustrated Lewis pair on
defective UiO-66 for CO2 hydrogenation to methanol, J. Mater. Chemi. A 8 (2020)
22802–22815.
Declaration of competing interest [19] J. Ye, J.K. Johnson, Catalytic hydrogenation of CO2 to methanol in a Lewis pair
functionalized MOF, Cata. Sci. Technol. 6 (2016) 8392–8405.
The authors declare that they have no known competing financial [20] Y. Zhang, Y. Mo, Z. Cao, Rational design of main group metal-embedded nitrogen-
doped carbon materials as frustrated Lewis pair catalysts for CO2 hydrogenation to
interests or personal relationships that could have appeared to influence
formic acid, ACS Appl. Mater. Interfaces 14 (2022) 1002–1014.
the work reported in this paper. [21] J. Zhao, H. Zhang, H. Wang, J. Wang, Tuning Lewis acid/base on the TiO2-
supported Pd-CoOx interfaces to control the CO2 selective hydrogenation, Mol.
Catal. 518 (2022) 112076.
Data availability
[22] K. Zheng, Y. Li, B. Liu, F. Jiang, Y. Xu, X. Liu, Ti-doped CeO2 stabilized single-atom
rhodium catalyst for selective and stable CO2 hydrogenation to ethanol, Angew.
Data will be made available on request. Chem. Int. Ed. 61 (2022) e202210991.
[23] B. Yu, L. Cheng, S. Dai, Y. Jiang, B. Yang, H. Li, Y. Zhao, J. Xu, Y. Zhang, C. Pan, X.
M. Cao, Y. Zhu, Y. Lou, Silver and copper dual single atoms boosting direct
Acknowledgments oxidation of methane to methanol via synergistic catalysis, Adv. Sci. 10 (2023)
2302143.
[24] B. Delley, From molecules to solids with the DMol3 approach, J. Chem. Phys. 113
This project was supported financially by the National Key R&D (2000) 7756–7764.
Program of China (2021YFB3501900), National Natural Science Foun­ [25] J.P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approximation made
dation of China (21908079, 22273021, U21A20326, 22302200), simple, Phys. Rev. Lett. 77 (1996) 3865–3868.
[26] S. Grimme, Semiempirical GGA-type density functional constructed with a long-
Jiangsu Specially-Appointed Professor (1046010241211400), Natural
range dispersion correction, J. Comput. Chem. 27 (2006) 1787–1799.
Science Foundation of Jiangsu Province (BK20211239, BK20201345, [27] B. Delley, Hardness conserving semilocal pseudopotentials, Phys. Rev. B 66 (2002)
BK20221541), Postgraduate Research & Practice Innovation Program of 155125.
Jiangsu Province (Nos. KYCX23_2461), Special Fund Project of Jiangsu [28] T.A. Halgren, W.N. Lipscomb, The synchronous-transit method for determining
reaction pathways and locating molecular transition states, Chem. Phys. Lett. 49
Province for Scientific and Technological Innovation in Carbon Peaking (1977) 225–232.
and Carbon Neutrality (BK20220023) and Dalian Institute of Chemical [29] N. Govind, M. Petersen, G. Fitzgerald, D. King-Smith, J. Andzelm, A generalized
Physics (DICP I201943). We also thank the Central Laboratory, School of synchronous transit method for transition state location, Comput. Mater. Sci. 28
(2003) 250–258.
Chemical and Material Engineering, Jiangnan University. [30] X. Sun, Y. Zhao, K. Chang, B. Peng, Q.Q. Gu, B. Yang, B.Y. Yu, J. Xu, F.D. Liu,
Y. Zhang, C.S. Pan, Y. Lou, 1T-phase MoS2 edge-anchored Pt1-S3 active site
boosting selective hydrogenation of biomass-derived maleic anhydride, Rare Met.
Appendix A. Supplementary data
42 (2023) 2658–2669.
[31] C. Shen, K. Sun, Z. Zhang, N. Rui, X. Jia, D. Mei, C. Liu, Highly active Ir/In2O3
Supplementary data to this article can be found online at https://doi. catalysts for selective hydrogenation of CO2 to methanol: experimental and
org/10.1016/j.cej.2024.150093. theoretical studies, ACS Catal. 11 (2021) 4036–4046.
[32] T. Shi, Y. Men, S. Liu, J. Wang, Z. Li, K. Qin, D. Tian, W. An, X. Pan, L. Li,
Engineering the crystal facets of Pt/In2O3 catalysts for high-efficiency methanol
References synthesis from CO2 hydrogenation, Colloids Surf. A-Physicochem. Eng. Asp. 651
(2022) 129782.
[33] L. Luo, L. Fu, H. Liu, Y. Xu, J. Xing, C.R. Chang, D.Y. Yang, J. Tang, Synergy of Pd
[1] Z. Li, Y. Men, S. Liu, J. Wang, K. Qin, D. Tian, T. Shi, L. Zhang, W. An, Boosting CO2
atoms and oxygen vacancies on In2O3 for methane conversion under visible light,
hydrogenation efficiency for methanol synthesis over Pd/In2O3/ZrO2 catalysts by
Nat. Commun. 13 (2022) 2930.
crystalline phase effect, Appl. Surf. Sci. 603 (2022) 154420.
[34] X. Tao, R. Long, D. Wu, Y. Hu, G. Qiu, Z. Qi, B. Li, R. Jiang, Y. Xiong, Anchoring
[2] C. Yao, H. Fan, A. Adogwa, H. Xiong, M. Yang, F. Liu, Z. Chen, Y. Lou, Recent
positively charged Pd single atoms in ordered porous ceria to boost catalytic
advances in carbon dioxide selective hydrogenation and biomass valorization via
activity and stability in Suzuki coupling reactions, Small 16 (2020) e2001782.
single-atom catalysts, Res. Chem. Mater. 2 (2023) 189–207.
[35] F. Hu, L. Leng, M. Zhang, W. Chen, Y. Yu, J. Wang, J.H. Horton, Z. Li, Direct
[3] G. Tian, Y. Wu, S. Wu, S. Huang, J. Gao, CO2 hydrogenation to methanol over Pd/
synthesis of atomically dispersed palladium atoms supported on graphitic carbon
MnO/In2O3 catalyst, J. Environ. Chem. Eng. 10 (2022) 106965.
nitride for efficient selective hydrogenation reactions, ACS Appl. Mater. Interfaces
[4] N. Rui, Z. Wang, K. Sun, J. Ye, Q. Ge, C. Liu, CO2 hydrogenation to methanol over
12 (2020) 54146–54154.
Pd/In2O3: effects of Pd and oxygen vacancy, Appl. Catal. B 218 (2017) 488–497.
[36] X. Zhou, K. Han, K. Li, J. Pan, X. Wang, W. Shi, S. Song, H. Zhang, Dual-site single-
[5] X. Jiang, Y. Jiao, C. Moran, X. Nie, Y. Gong, X. Guo, K.S. Walton, C. Song, CO2
atom catalysts with high performance for three-way catalysis, Adv. Mater. 34
hydrogenation to methanol on Pd Cu bimetallic catalysts with lower metal
(2022) e2201859.
loadings, Catal. Commun. 118 (2019) 10–14.
[37] M.S. Frei, C. Mondelli, R. Garcia-Muelas, K.S. Kley, B. Puertolas, N. Lopez, O.
[6] Y. Shi, Y. Zhou, Y. Lou, Z. Chen, H. Xiong, Y. Zhu, Homogeneity of supported
V. Safonova, J.A. Stewart, D. Curulla Ferre, J. Perez-Ramirez, Atomic-scale
single-atom active sites boosting the selective catalytic transformations, Adv. Sci. 9
engineering of indium oxide promotion by palladium for methanol production via
(2022) 2201520.
CO2 hydrogenation, Nat. Commun. 10 (2019) 3377.
[7] L. Song, H. Wang, S. Wang, Z. Qu, Dual-site activation of H2 over Cu/ZnAl2O4
[38] A. Wang, K. Lindgren, M. Di, D. Bernin, P.-A. Carlsson, M. Thuvander, L. Olsson,
boosting CO2 hydrogenation to methanol, Appl. Catal. B 322 (2023) 122137.
Insight into hydrothermal aging effect on Pd sites over Pd/LTA and Pd/SSZ-13 as
[8] Y. Yan, R.J. Wong, Z. Ma, F. Donat, S. Xi, S. Saqline, Q. Fan, Y. Du, A. Borgna,
PNA and CO oxidation monolith catalysts, Appl. Catal. B 278 (2020) 119315.
Q. He, C.R. Müller, W. Chen, A.A. Lapkin, W. Liu, CO2 hydrogenation to methanol
[39] P.H. Ho, J. Woo, R.F. Ilmasani, M.A. Salam, D. Creaser, L. Olsson, The effect of Si/
on tungsten-doped Cu/CeO2 catalysts, Appl. Catal. B 306 (2022) 121098.
Al ratio on the oxidation and sulfur resistance of beta zeolite-supported Pt and Pd
[9] D.W. Stephan, Frustrated Lewis pairs: a new strategy to small molecule activation
as diesel oxidation catalysts, ACS Eng. Au 2 (2021) 27–45.
and hydrogenation catalysis, Dalton Trans. 3129–3136 (2009).
[40] I. Friberg, N. Sadokhina, L. Olsson, The effect of Si/Al ratio of zeolite supported Pd
[10] D.W. Stephan, G. Erker, Frustrated Lewis Pair chemistry: development and
for complete CH4 oxidation in the presence of water vapor and SO2, Appl. Catal. B
perspectives, Angew. Chem. Int. Ed. 54 (2015) 6400–6441.
250 (2019) 117–131.
[11] D.W. Stephan, Frustrated Lewis Pairs, J. Am. Chem. Soc. 137 (2015) 10018–10032.

8
Y. Shi et al. Chemical Engineering Journal 485 (2024) 150093

[41] F. Lónyi, H.E. Solt, J. Valyon, H. Decolatti, L.B. Gutierrez, E. Miró, An operando heterostructure for enhancing hydrogenation of CO2 to methanol, Small 19 (2023)
DRIFTS study of the active sites and the active intermediates of the NO-SCR 2204914.
reaction by methane over in, H- and in, Pd, H-zeolite catalysts, Appl. Catal. B 100 [48] M.M. Zain, M. Mohammadi, N. Kamiuchi, A.R. Mohamed, Development of highly
(2010) 133–142. selective In2O3/ZrO2 catalyst for hydrogenation of CO2 to methanol: an insight into
[42] Y. Ryou, J. Lee, S.J. Cho, H. Lee, C.H. Kim, D.H. Kim, Activation of Pd/SSZ-13 the catalyst preparation method, Korean J. Chem. Eng. 37 (2020) 1680–1689.
catalyst by hydrothermal aging treatment in passive NO adsorption performance at [49] M. Baca, A. Pigamo, J.L. Dubois, J.M.M. Millet, Fourier transform infrared
low temperature for cold start application, Appl. Catal. B 212 (2017) 140–149. spectroscopic study of surface acidity by pyridine adsorption on the M1 active
[43] J. Chang, M.J. Hulsey, S. Wang, M. Li, X. Ma, N. Yan, Electrothermal water-gas phase of the MoVTe(Sb)NbO catalysts used in propane oxidation, Catal. Commun.
shift reaction at room temperature with a silicomolybdate-based palladium single- 6 (2005) 215–220.
atom catalyst, Angew. Chem. Int. Ed. 62 (2023) e202218265. [50] Y. Wang, G. Wang, L.I. van der Wal, K. Cheng, Q. Zhang, K.P. de Jong, Y. Wang,
[44] G. Tian, Y. Wu, S. Wu, S. Huang, J. Gao, Solid-state synthesis of Pd/In2O3 catalysts Visualizing element migration over bifunctional metal-zeolite catalysts and its
for CO2 hydrogenation to methanol, Catal. Lett. 153 (2022) 903–910. impact on catalysis, Angew. Chem. Int. Ed. 17735–17743 (2021).
[45] S. Zhang, Z. Wu, X. Liu, Z. Shao, L. Xia, L. Zhong, H. Wang, Y. Sun, Tuning the [51] H. Li, L. Wang, Y. Dai, Z. Pu, Z. Lao, Y. Chen, M. Wang, X. Zheng, J. Zhu, W. Zhang,
interaction between Na and Co2C to promote selective CO2 hydrogenation to R. Si, C. Ma, J. Zeng, Synergetic interaction between neighbouring platinum
ethanol, Appl. Catal. B 293 (2021) 120207. monomers in CO2 hydrogenation, Nat. Nanotechnol. 13 (2018) 411–417.
[46] Z. Cai, J. Dai, W. Li, K.B. Tan, Z. Huang, G. Zhan, J. Huang, Q. Li, Pd supported on [52] H. Zhao, R. Yu, S. Ma, K. Xu, Y. Chen, K. Jiang, Y. Fang, C. Zhu, X. Liu, Y. Tang,
MIL-68(In)-derived In2O3 nanotubes as superior catalysts to boost CO2 L. Wu, Y. Wu, Q. Jiang, P. He, Z. Liu, L. Tan, The role of Cu1-O3 species in single-
hydrogenation to methanol, ACS Catal. 10 (2020) 13275–13289. atom Cu/ZrO2 catalyst for CO2 hydrogenation, Nat. Catal. 5 (2022) 818–831.
[47] W.G. Cui, Q. Zhang, L. Zhou, Z.C. Wei, L. Yu, J.J. Dai, H. Zhang, T.L. Hu, Hybrid [53] W.G. Cui, Y.T. Li, L. Yu, H. Zhang, T.L. Hu, Zeolite-encapsulated ultrasmall Cu/
MOF template-directed construction of hollow-structured In2O3@ZrO2 ZnOx nanoparticles for the hydrogenation of CO2 to methanol, ACS Appl. Mater.
Interfaces 13 (2021) 18693–18703.