Direct conversion of CO2 into methanol over promoted indium oxide-based catalysts

Article abstract of DOI:10.1016/j.apcata.2019.117144

Supported indium oxide catalysts are investigated for the CO₂ hydrogenation to methanol at a total pressure of 40 bar (528–573 K) using a laboratory flow reactor. Surface reducibility, optical spectral characteristics, and catalytic rates and selectivity were correlated to catalyst composition. Promoted catalysts, especially Yttrium or Lanthanum-promoted indium oxide, require higher temperatures (H₂-TPR) for surface reduction and display higher CO₂ desorption temperatures (CO₂-TPD). The promoted samples also have ?20% higher methanol selectivity compared to the non-promoted catalyst, while having similar methanol formation rates (0.330–0.420 g_MeOH g_cat.^?1 h^?1 at 573 K). From 528 K to 558 K, methanol selectivity was over 80%, over all the promoted catalysts, and nearly 100% selectivity was observed at the low temperature range (?528 K) investigated. The reaction kinetics of Y-promoted catalyst and the results of CO co-feeding experiments suggest that formate pathway is the likely reaction mechanism for methanol formation.

Full text of DOI:10.1016/j.apcata.2019.117144

Accepted Manuscript
Title: Direct Conversion of CO into Methanol over Promoted
2
Indium Oxide-based Catalysts
Authors: Chen-Yu Chou, Raul F. Lobo
PII:
DOI:
Article Number:
S0926-860X(19)30299-6
https://doi.org/10.1016/j.apcata.2019.117144
117144
Reference:
APCATA 117144
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
21 February 2019
1 July 2019
3 July 2019
Please cite this article as: Chou C-Yu, Lobo RF, Direct Conversion of CO into Methanol
2
over Promoted Indium Oxide-based Catalysts, Applied Catalysis A, General (2019),
https://doi.org/10.1016/j.apcata.2019.117144
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Direct Conversion of CO into Methanol over
2
Promoted Indium Oxide-based Catalysts
Chen-Yu Chou and Raul F. Lobo*
Center for Catalytic Science and Technology, Department of Chemical and Biomolecular
Engineering, University of Delaware, Newark, DE 19716, USA
*Corresponding author: Fax: 302-831-1048, lobo@udel.edu
Graphical abstract
Highlights
1




Yttrium loading in In-based catalysts reduce surface reducibility.
Y and La promoters increase number of CO adsorption sites.
Y and La promoters promoted catalysts have 40-60% higher methanol selectivity than
conventional Cu-based catalyst.
2

The formate pathway is, most likely, the dominant reaction channel for methanol
synthesis on Y-promoted supported In-catalyst.
Abstract
Supported indium oxide catalysts are investigated for the CO
2
hydrogenation to methanol
at a total pressure of 40 bar (528-573K) using a laboratory flow reactor. Surface reducibility,
optical spectral characteristics, and catalytic rates and selectivity were correlated to catalyst
composition. Promoted catalysts, especially Yttrium or Lanthanum-promoted indium oxide,
require higher temperatures (H
temperatures (CO -TPD). The promoted samples also have ~20% higher methanol selectivity
compared to the non-promoted catalyst, while having similar methanol formation rates (0.330-
2
-TPR) for surface reduction and display higher CO
2
desorption
2
0
.420 gMeOH g
cat. h-1 at 573 K). From 528 K to 558 K, methanol selectivity was over 80 %, over
all the promoted catalysts, and nearly 100% selectivity was observed at the low temperature range
~528 K) investigated. The reaction kinetics of Y-promoted catalyst and the results of CO co-
-1
(
feeding experiments suggest that formate pathway is the likely reaction mechanism for methanol
formation.
Keywords: Methanol synthesis, CO hydrogenation, Indium oxide, Yttrium, Lanthanum.
2
Introduction
The direct carbon dioxide (CO ) hydrogenation to methanol is a potentially important route
2
to decrease CO
2
emissions: methanol is a versatile compound that can be used as fuel or as a
2

precursor to the production of many commodity chemicals such as acetic acid, formaldehyde, and
dimethyl ether [1,2]. Liquid methanol is also preferable to hydrogen in terms of energy storage
density and ease of transportation. If a green (CO
2
-free) H
2
source is used, CO
2
hydrogenation to
methanol is sustainable as exemplified by the George Olah CO
2
to Renewable Methanol Plant in
Iceland [3]. This plant uses electricity, generated from hydro and geothermal energy, to make
hydrogen. The process can be driven by a variety of renewable energy sources such as solar, wind,
and waste heat from chemical or nuclear plants, making the process more sustainable. CO
2
, thus,
can be chemically transformed from a greenhouse gas into a valuable and renewable carbon source
ꢁ
1
[
2]. Although methanol synthesis from CO
CO conversion to methanol is kinetically limited at low temperatures and thermodynamically
limited at high temperatures, resulting in a low theoretical methanol yield [4][5].
Cu/ZnO/Al (CZA) ternary catalysts are currently employed for industrial methanol
synthesis from syngas (CO/CO /H ) at elevated pressures (50-100 bar) and temperatures (473-543
2
and H is exothermic (∆퐻°298퐾 = −4ꢀ.5 푘퐽 푚표푙 ),
2
2
2
O
3
2
2
K) [6][7]. However, the CZA catalysts typically have low selectivity (around 40%), because of the
competing reverse water–gas shift (RWGS) reaction, and limited stability, due to sintering of the
active phase under reaction conditions [8]. Among the Cu-based materials investigated, only a few
catalysts such as LaCr0.5Cu0.5
3
O [9], Cu-Ga/ZnO [10], and Cu@ZnO (core-shell) [11] exhibit high
selectivity to methanol (88-100%); the long-term stability and scalability of these materials,
however, have not been evaluated. Many investigations have attempted to identify new catalyst
formulations (such as Cu/CeO
promising catalytic activity for CO
metal alloys, in particular, have been recognized as highly selective to CO
2
[12], Pd/Ga
hydrogenation to methanol. Indium oxide and In-containing
in the methanol and
2
O
3
[13], Pd/CeO
2
[14], and In
2
O [15]) showing
3
2
2
ethanol steam reforming reaction[16–18]; consequently, these materials also have the potential for
3

catalyzing its reverse reaction, CO
nanoparticles were widely used for photocatalytic hydrogenation of CO
have shown that key intermediates (*HCOO) involved in CH OH synthesis are more stable on a
defective In surface than those on the Cu surface, strongly suppressing the formation of CO
20] and demonstrating great potential for selective catalytic CO hydrogenation. A more recent
DFT and kinetic study from Frei et. al. showed a different energetically favored path with
CH OOH and CH (OH) as main intermediates on the In (1 1 1).surface.[21]. Martin et. al.
investigated ZrO supported indium oxide catalysts and observed significant and stable reaction
rates with high selectivity towards methanol [22]. Other studies using In or In-based materials
15,23,24] have not shown as high methanol selectivity (99.8%) as reported by Martin et al. Sun
et. al., for example, showed 54.9 % methanol selectivity with unsupported In catalyst; Rui et.
al. investigated In and Pd/In catalyst, demonstrating remarkably high methanol formation
rates and ~60-72% methanol selectivity over Pd/In catalyst prepared with peptide templates
also see Table 1). Dou and coworkers [25] have investigated the effect of ZrO on a (110) In
surface (three atomic layers thick) and found clear effects of the ZrO . In particular, they found
that the presence of ZrO under the In overlayer inhibits the dissociation of CO into CO, and
increases the concentration of oxygen vacancies on the In layer. The ZrO also enhances the
stability of the key intermediate H COO on the In surface.
2
hydrogenation to methanol. Among them, defective In
2
O
3
2
[19]. DFT calculations
3
2
O
3
[
2
2
2
2
2
O
3
2
2
O
3
[
2
O
3
2
O
3
2
O
3
2
O
3
(
2
2
O
3
2
2
2
O
3
2
2
O
3
2
2
2
O
3
Inui et. al. have investigated the promotional effect of lanthanum oxide on Cu-Zn-Cr-Al-
Oxides catalysts by increasing CO adsorption and catalytic reaction rates in the CO to methanol
reaction [26]. Addition of Y into CZA catalysts also leads to higher CO conversion and Cu
2
2
2
O
3
2
dispersion [27]. Despite the improvements observed upon addition of La and Y to Cu catalysts, Y
4

or La-promoted indium oxide catalysts have not been investigated for the hydrogenation of CO
to methanol.
2
Herein, Y or La-promoted In
2
O
3
/ZrO catalysts have been prepared by wet impregnation
2
and investigated for the CO to methanol reaction. In particular, Y and La effect on selectivity and
2
reducibility of the supported oxide have been investigated in detail. The catalysts are characterized
using scanning electronic microscopy (SEM), X-ray diffraction (XRD), X-ray photoelectron
spectroscopy (XPS), temperature-programmed reduction (TPR), and temperature-programmed
desorption (TPD). The catalytic properties, including CO conversion, methanol selectivity, and
2
methanol or CO formation rate, were correlated to the catalysts composition and reducibility. We
found that the incorporation of Y and La enhance methanol selectivity significantly by decreasing
the surface reducibility and increasing the CO
catalyst. A mechanistic discussion comparing the initial steps of CO
In /ZrO and Y-promoted catalysts is provided. Furthermore, the catalytic kinetics suggests that
the reaction proceeds via a formate pathway.
2
adsorption capacity on both Y or La-containing
2
hydrogenation over
2
O
3
2
5

Experimental
Catalyst preparation
The ‘unpromoted’ catalyst, In
2
O
3
/ZrO
2
, was prepared by wet impregnation as follows: 0.76
3
g In(NO
3
)
3
·xH
2
O (Alfa Aesar, 99.99%) was first dissolved in a mixture of ethanol (70 cm ) and
3
deionized water (24 cm ). 2.00 g ZrO
2
(NORPRO, SZ 31164 extrudates, monoclinic phase,
crushed prior to use) as the catalyst support was subsequently added to the solution and the
resulting slurry was stirred for 5 h at room temperature. The solvent was removed using a rotary
evaporator (Büchi Rotavap R-114) at 343 K. The sample obtained was dried in a drying oven at
−
1
3
53 K overnight and calcined in static air at 573 K (heating rate 2 K min , hold for 2 h). The
promoted catalysts were prepared in the similar fashion. For 1.5Y9In/ZrO , 2Y8In/ZrO , and
O (Sigma Aldrich, 99.99%), 0.61 g
2
2
3
Y8In/ZrO
2
, 0.61 g In(NO
3
)
3
·xH
2
O and 0.12 g Y(NO
3
)
3
·4H
2
In(NO ·xH
3
)
3
2
O and 0.17 g Y(NO
3
)
3
·4H O, and 0.57 g In(NO
2
3
)
3
·xH
2
O and 0.22 g Y(NO
3
)
3
·4H O,
2
3
3
respectively, were dissolved in a mixture of ethanol (70 cm ) and deionized water (24 cm ). For
3
La10In/ZrO
dissolved in the same solvent as above. The rest of the procedures were the same as the unpromoted
In /ZrO
Catalyst Characterization
The bulk composition in catalysts was determined using a Rigaku wavelength-dispersive
2
, 0.61 g In(NO
3
)
3
·xH
2
O and 0.22 g La(NO
3
)
3
·6H
2
O (Sigma Aldrich, 99.99%) were
2
O
3
2
.
X-ray fluorescence (WDXRF) spectrometer. The sample names of the catalysts are based on the
metal weight percentages obtained from WDXRF. Textural characterization of the samples was
carried out using N adsorption at 77 K. Isotherms were collected on a Micrometrics 3Flex and
2
analyzed using Brunauer–Emmett–Teller (BET) method to calculate surface area, and the Barrett-
Joyner-Halenda (BJH) method was used to calculate pore volume and pore size distribution. X-
6

Ray Diffraction (XRD) patterns of catalyst powders were collected at room temperature on a
Bruker diffractometer using Cu Kα radiation (λ = 1.5418 Å). Measurements were taken over the
range of 20° < 2θ < 70° with a step size of 0.02°. The phase was identified by comparing the
diffraction patterns with the conventional Joint Committee on Powder Diffraction Standards
(JCPDSs). Scanning electronic microscopy (SEM) analyses were carried out on a Zeiss SEM/FIB
Auriga-60. X-ray photoelectron spectroscopy (XPS) measurements were performed on a K-alpha
Thermo Fisher Scientific spectrometer using monochromated Al Kα X-ray source. The
measurements of indium oxides on supported ZrO
2
(In
2
O
3
/ZrO
2
), and other Y or La-promoted
) were done with a spot size
catalysts (1.5Y9In/ZrO , 2Y8In/ZrO , 3Y8In/ZrO , and 3La-10In/ZrO
2
2
2
2
-7
of 400 μm at ambient temperature and a chamber pressure of ~10 mbar. A flood gun was used
for charge compensation. All the measured spectra were calibrated by setting the reference binding
energy of adventitious carbon 1s signal at 284.8 eV. The spectra were analyzed using Avantage®
surface chemical analysis software (v5.986). For the spectral fitting, each component consists of a
linear combination of Gaussian and Lorentzian product functions. SMART background was used
over the region to define the peaks.
Temperature-programmed reduction (TPR) and temperature-programmed desorption
(
TPD) were carried out by using an Altamira AMI-200 catalyst characterization flow system.
Analysis of the gaseous effluent was carried out using a thermal conductivity detector (TCD) to
detect the H or CO concentrations during the experiment. H -TPR was conducted via 2 steps: (i)
the sample (50 mg) was dried at 398 K for 30 min in Ar (60 sccm); (ii) The TPD trace was obtained
by increasing the temperature up to 973 K in a gaseous mixture of 10% H /Ar at the rate of 5 K
-TPD was conducted as the following steps: (i) 50 mg of sample was dried at 398 K for
0 min in He (40 sccm); (ii) the sample was then treated with He (40 sccm) at 573 K for 60 min
2
2
2
2
-1
min . CO
2
3
7

and then cooled down to 313 K under the same gas stream; (iii) CO
13 K with 50% CO /He gaseous mixture (40sccm) for 2 h; (iv) the desorption process was
programmed from 313 K to 873 K with 40 sccm He. The ramping rates in all the CO -TPD
2
adsorption was carried out at
3
2
2
-1
experiments was 5 K min .
Reactor setup and catalytic activity
The reaction rates and other kinetic parameters were measured using a packed-bed
microreactor operated in a down-flow mode under high pressure. 40 bar is the standard operating
®
pressure controlled by a back pressure regulator (Swagelok , KPB series, 0-2000 PSIG). The setup
was equipped with multiple mass flow controllers (Brooks SLA5850 and Bronkhorst EL-Flow
F211CV) to feed H
2
(Keen, 99.999%), CO
2
(Keen, 99.999%), CO (Matheson, 99.99%), and He
(Keen, 99.999%). A fixed-bed and down-stream reactor setup was used in this study. 0.15 g
catalyst pellets (particle size: 250-420 μm) were placed between two plugs of quartz wool within
a 316 SS tube (4.6 mm ID). The reactor bed temperature was monitored by the insertion of a
thermocouple through the top of the reactor (Omega, K-type, 1/8 in. diameter). The thermocouples
also helped maintain the position of the bed within the reactor tube. Prior to the reaction, the
catalyst was activated at 573 K and 5 bar He for 1 h. Activity tests were conducted under Ptot = 40
3
-1
bar, T = 528K, 543K, 558K, or 573 K, a total flow rate Ftot = 130 cm STP min (sccm), and the
volumetric feed composition of H :CO :He = 80:20:30. The standard gas hourly space velocity
2
2
GHSV) that is used was 52000 cm h gcat.-1. Each condition was held for 3 h or when the effluent
composition was in steady state. Helium was used as internal standard gas for concentration
calibration due to the change of the total volumetric flow rate in the effluent after the CO
3
-1
(
2
hydrogenation reaction. Gas lines after the reactor were kept at 433 K using heating tape to avoid
condensation of products. The composition of the effluent stream was analyzed online by a gas
8

chromatograph (490 MicroGC, Agilent) during continuous flow experiments. The MicroGC was
equipped with a thermal conductivity detector (TCD) to quantify H , He, CO , CO, and CH OH.
Gas hourly space velocity (GHSV) in cm h gcat. was calculated according to the
equation below:
2
2
3
3
-1
-1
퐹푡ꢂ푡
퐺퐻푆푉 = 푤
(1)
푐푎푡.
where ꢃꢄꢅꢆ. is the weight of the catalyst.
The CO
2
conversion (XCO2), CH OH selectivity (SMeOH), CO formation rate (rCO), and
3
methanol formation rate (rMeOH) were calculated according to the following expressions:
푁
̇
ꢁ푁
̇
ꢇꢈ ,푖푛
ꢇꢈꢉ,ꢂ푢푡
ꢇꢈ ,푖푛
ꢉ
푋퐶푂2
=
(2)
푁
̇
ꢉ
̇
where ꢊ stands for the molar flowrate of species i.
ꢋ
푁
̇ ꢍꢎꢈꢏ,ꢂ푢푡
̇ ꢍꢎꢈꢏ,ꢂ푢푡+푁
푆푀푒푂ꢌ = 푁
(3)
(4)
̇ ꢇꢈ,ꢂ푢푡
푁
̇ 푖,푖푛
∙ꢐ
∙ꢑ ∙푀
ꢇꢈꢉ
푖 푖
푤푐푎푡.
푟ꢋ =
where 푆 and ꢒ are the selectivity and the molar mass of species i, respectively.
ꢋ
ꢋ
Apparent kinetic parameters (activation energy and reaction orders) were determined near
the region of CO /H = 4:1 molar ratio over 0.15 g indium oxide-based catalysts. The apparent
2
2
activation energy was obtained by measuring the methanol formation rate with respect to the
temperature of the reaction from 573 K to 528 K in intervals of 15 K. The reaction orders of
2
Y8In/ZrO
or CO
pressure, which corresponded to the change of H
2
catalyst at 573 K were calculated from the dependence of methanol formation rate on
partial pressure. The H reaction order was obtained from the change of H partial
flow rate (FH2). FH2 was 60, 70, 80, 85, and 90
H
2
2
2
2
2
9

sccm, while FCO2 was 20 sccm. FHe was adjusted accordingly to keep constant total flow rate. In
CO reaction order experiment, FCO2 was 12 to 28 sccm with 4 sccm increment. FHe was also
adjusted stepwise to keep Ftot = 130 sccm while a constant FH2 = 80 sccm was fed into the reactor.
In the CO co-feeding experiments over 2Y8In/ZrO at 573K and 543K, FH2 and FCO2
remained constant at 80 and 20 sccm, respectively. The amount of He fed into the system, FHe
2
2
3
−1
(
initially at 110 cm STP min ) was reduced stepwise while feeding progressively higher amounts
of CO resulting in a molar carbon ratio of R = CO/(CO +CO) = 0, 0.20, 0.33, 0.43, 0.50. The
2
conversion at R = 0 was acquired again after the measurement at R = 0.50 to control for any change
of activity due to deactivation. An additional CO-TPD experiment was conducted in the same
reactor setup as measuring the catalytic reaction rates and the CO signal was monitored using a
mass spectrometer (MS, Pfeiffer, GSD320). Before the TPD process, 0.1 g 2Y8In/ZrO
2
was
/He
50 sccm) to partially reduce the surface. 20% CO/He (50 sccm) was then fed into the reactor at
pretreated in 100 sccm He at 458 K for 1 h to clean the sample surface, followed by 20% H
2
(
3
13 K for 2h. The TPD trace was obtained by increasing the temperature from 313 K to 873 K
under a constant flow of 50 sccm He.
10

Results and Discussion
The SEM images of In
particles are formed of aggregates of smaller particles. The aggregates have a broad range of sizes
~100 nm to ~5µm, see Figure 1). The physical and textural properties of In /ZrO and the
promoted catalysts were determined using N adsorption, XRF, and XPS (Table 2). The measured
2
O
3
/ZrO
2
, 1.5Y9In/ZrO
2
, and 3La10In/ZrO show that the catalysts
2
(
2
O
3
2
2
surface area and the total pore volume (V
P
) correspond closely to the properties of the ZrO
2
carrier
2
2
(
~100 m /g) in all cases. The BET surface area (SBET) of the un-promoted In
2
O
3
/ZrO
2
was 87 m /g,
2
2
while for the promoted catalysts the surface area ranged from 84 m /g to 88 m /g. Total pore
3
3
volume (V ) of the catalysts was in the narrow range of 0.21 cm /g to 0.23 cm /g. The incorporation
P
of Y and La slightly decreased the relative fraction of the smaller pores and larger pores but
increased the amount of medium sized pores as seen in the pore size distribution (see Error!
Reference source not found.). Clearly, the effect of yttrium and lanthanum oxides on surface area
and porosity is small (vide infra). The bulk weight percentage, in metallic base, of Y, La and In of
the catalyst (W
close to the nominal weight percentages calculated from the amount of the precursors used during
catalyst preparation. The surface density of In atoms in the (1 1 1) surface of crystalline In was
Y
, WLa, and WIn) were also determined from XRF. All of observed amounts were
2
O
3
1
8
2
estimated to be 8.67×10 In atoms/m [28,29]. Based on the BET surface area, molecular weight
of indium (M.W. of In) and the loading of In atoms (wIn) measured from XRF, the un-promoted
In
2
O
3
/ZrO
had approximately 6.1 In atoms/nm2.
2
ꢃ (%)
ꢚꢛ
퐼ꢓ ꢔꢕ표푚 푠ꢖ푟푓ꢔꢗꢘ 푑ꢘꢓ푠ꢙꢕ푦 =
× 퐴푣표푔ꢔ푑푟표 ꢓꢖ푚푏ꢘ푟
(
퐵퐸푇 푠ꢖ푟푓ꢔꢗꢘ ꢔ푟ꢘꢔ) × (ꢒ. 푊. 표푓 퐼ꢓ)
Assuming a uniform distribution, the thickness of the deposited indium oxide layer can be
estimated by the ratio of the measured In atom density to the theoretical In atom surface density.
11

The coverage for In
amount of In deposited on the catalysts was slightly less than a monolayer if the distribution of
In is uniform on the ZrO surface.
2
O
3
/ZrO
2
was ~0.7, whereas for the Y-promoted catalysts was ~0.6. The
2
O
3
2
In contrast to XRF, XPS is a surface sensitive technique and can reveal information about
the surface composition of the promoted catalysts. WIn measured from XPS were also similar to
the nominal concentration of the catalysts; however, W
than the nominal concentration. W or WLa from XPS of 1.5Y9In/ZrO
and 3La10In/ZrO were 3.1%, 4.0%, 5.3%, and 7.0%, respectively. It appears then that Y or La
are preferably located on the surface of the sample after the synthesis process. Further analyses of
the XPS spectra show that the atomic In/Zr ratios of In /ZrO and 3La10In/ZrO were both 0.17,
while the In/Zr ratios of 1.5Y9In/ZrO , 2Y8In/ZrO , 3Y8In/ZrO were 0.11-0.12 (see Table 2).
The In/Zr ratios of Y-modified samples decreased after the incorporation of yttrium oxide. This
result implies that the surface of Y-modified samples contained less In that were deposited per
ZrO particle on the surface, resulting in lower methanol formation rate with respect to the catalyst
weight. (0.465 gMeOH
K. See Figure 5 (c)). This may occur if there are In
XPS) in the final form of the catalyst. On the other hand, 3La10In/ZrO
ratio than In /ZrO , which is consistent to its similar methanol formation rate compared to
In /ZrO catalyst (see below).
Y
, WLa on the surface were much higher
Y
2
, 2Y8In/ZrO , 3Y8In/ZrO
2
2
,
2
2
O
3
2
2
2
2
2
2
O
3
2
cat. h-1 for In
-
1
O
3
/ZrO
and 0.420 gMeOH
g
cat. h-1 for 1.5Y9In/ZrO
oxide particles (with In inaccessible to
had a similar In/Zr atomic
-1
at 573
g
2
2
2
2
O
3
2
2
O
3
2
2
O
3
2
The XRD patterns of ZrO
2
, In
2
O
3
, In
2
O
3
/ZrO
2
, 1.5Y9In/ZrO
2
, and 3La10In/ZrO (
2
Figure 2) show that the support is monoclinic ZrO
structure did not change appreciably by the wet impregnation or the calcination. Peaks of pure
In were identified at 2θ = 21.49°, 30.58°, 35.45°, and 51.02° and were assigned to the (2 1 1),
2
(PDF#00-001-0750): the carrier
2
O
3
12

(
2 2 2), (4 0 0), and (4 4 0) reflections of the cubic In
found in In /ZrO , 1.5Y9In/ZrO , and 3La10In/ZrO
In
and 3La10In/ZrO
2
O
3
(PDF#00-044-1087): these peaks can be
as well, indicating that at least some of
2
O
3
2
2
2
2
O
3
in the supported materials exist in crystalline form after calcination in air. For 1.5Y9In/ZrO
2
,
2
catalysts, no characteristic diffraction peaks of Y or La related compound can
be observed. However, Y3+ or La3+ could still be identified on the surface by XPS (see Error!
Reference source not found.). The absence of reflections of Y and La in the XRD patterns indicates
that yttrium oxide and lanthanum oxide particles are too small or the loading of these two materials
is too low to be detected by XRD. XRD patterns and the XPS spectra together suggest that Y and
La oxides exist in a highly dispersed form such as nano-crystallites or well-mixed with the oxide
in thin layers on the surface of the material after the calcination in air. It should be noted that
indium, yttrium, and lanthanum were in their oxides form (In
2
O
3,
Y
2
O
3
, La
2
3
O ) in the samples of
Y and La-promoted catalysts after the calcination in air. The nomenclature of sample names for
those promoted catalysts was simplified as element name only (e.g. 1.5Y9In/ZrO
etc.) for simplicity.
2
, 3La10In/ZrO
2
,
The reducibility of the catalysts was examined by H
2
-TPR (Figure 3). For In
2
O /ZrO
3
2
,
evidence for reduction starts at around 423 K, consistent with past reports [23][30]. The main peak
at 470 K was assigned to the surface reduction of In , and the later small and broad signal near
47 K is assigned to the bulk reduction of In . H consumption near 647 K is small due to the
low loading of the In in In /ZrO . A control TPR experiment (Error! Reference source not
found.) performed on unsupported In showed much higher H consumption in this region, as
expected from bulk reduction studies described in other reports [23]. All the surface In
2
O
3
6
2
O
3
2
2
O
3
2
O
3
2
2
O
3
2
2
O
3
reduction peaks of our samples had similar reduction temperature windows (~50 K), but the
modified catalysts showed shifted TPR profiles: the reduction peaks of Y-modified samples,
13

1
.5Y9In/ZrO
2
, 2Y8In/ZrO
2
, 3Y8In/ZrO
2
, were at 481 K, 496 K, and 505 K, respectively. That is,
also showed a
higher yttrium concentration led to higher reduction temperatures. 3La10In/ZrO
2
higher reduction temperature for surface In
. These data show that, as expected, incorporation of the less reducible Y3+
and La3+ cations onto the surface of In
2
O
3
at 503 K, which is 33 K higher than the un-
promoted In
2
O
3
/ZrO
2
2
3
O decrease the reducibility of In-based catalysts.
Furthermore, the reduction window remains nearly the same as the amount of Y and La is increased.
This is an indication of excellent mixing between the promoter and the In phase, since
2
O
3
otherwise a separate reduction peak (or clear broadening) would be observed in the reduction peak.
As can be seen in Figure 3, this reduction window remains remarkably similar.
The CO
after heat treatment at 573 K in helium for 1 h. The CO
exhibited one peak at T< 400 K and one broad peak in the region of 570 K < T < 800 K. The first
peak is assigned to the desorption of physically adsorbed CO . The later peaks, which can be
further deconvoluted into Gaussian peaks α and β for 2Y8In/ZrO and 3La10In/ZrO , are assigned
to the desorption of chemically adsorbed CO . There are two common ways to create oxygen
vacancies on In clusters for CO hydrogenation [23]. One is thermal-induced oxygen vacancies
denoted here as Ov1) and the other is H -induced oxygen vacancies (denoted here as Ov2). HCOO
has been suggested as a key intermediate that is both thermodynamically and kinetically favorable,
based on DFT investigations on In catalyst [20,31]. Oxygen vacancy sites on the In surface
facilitate both CO adsorption, hydrogenation, and stabilize intermediates (HCOO, H COO, and
CO) involved in methanol formation.
2
-TPD profile (Figure 4) for samples with different compositions were recorded
2
desorption profiles of all catalysts
2
2
2
2
2
O
3
2
(
2
2
O
3
2
O
3
2
2
H
2
Since the samples for CO
2
adsorption measurements were treated under inert gas at 573 K
-TPD profiles are
(
without any reduction steps), the high-temperature peaks (α and β) in the CO
2
14

assigned to Ov1 vacancies. The β peaks are assigned to stronger basic sites induced by the
incorporation of Y and La. The α peak maxima associated with Ov1,α were 680 K, 682 K, and 678
K for In
with Ov1,β were 728 K and 732 K for 2Y8In/ZrO
CO desorption temperature in the β peaks indicates an increase of the basicity for Ov1 sites in Y
or La-promoted samples. The strength of the adsorption sites in terms of CO adsorption follows
the order 3La10In/ZrO > 2Y8In/ZrO > In /ZrO . Note that 3La10In/ZrO and 2Y8In/ZrO
desorbed more CO than In /ZrO according to the integrated peak area from their TPD profiles.
The total peak area associated with Ov1 was normalized by the amount of indium loading for each
catalyst, and the ratio of the area per indium loading was In /ZrO :2Y8In/ZrO :3La10In/ZrO
1:1.6:1.7, that is, the density of surface Ov1 sites was 60 or 70% higher for Y or La-promoted
2
O
3
/ZrO
2
, 2Y8In/ZrO
2
, and 3La10In/ZrO
2
, respectively. The β peak maxima associated
2
and 3La10In/ZrO
2
, respectively. The increasing
2
2
2
2
2
O
3
2
2
2
2
2
O
3
2
2
O
3
2
2
2
=
catalysts compared to In
2
O
3
/ZrO
2
. The XPS spectra of the concentration of oxygen near defects
are consistent with the higher density of oxygen
for In /ZrO , 2Y8In/ZrO
2
O
3
2
2
, and 3La10In/ZrO
2
vacancies in the promoted samples, as shown above (see also Error! Reference source not
found.). Together, the results from H -TPR and CO -TPD indicate that Y and La promoters
2
2
suppress the creation of Ov2 by making the catalysts less reducible, strengthen the CO
of Ov1, and consequently, enhance the methanol selectivity (see below).
2
adsorption
The catalytic properties of In
La10In/ZrO (Figure 5) can be compared in terms of (a) CO
c) methanol formation rate, and (d) CO formation rate. The methanol selectivity versus CO
conversion for all catalysts is also shown in Figure 5 (e). For all catalysts, CO conversion,
2
O
3
/ZrO
2
, 1.5Y9In/ZrO
2
, 2Y8In/ZrO
2
, 3Y8In/ZrO
2
, and
3
2
2
conversion, (b) methanol selectivity,
(
2
2
methanol and CO formation rate increased with increasing reaction temperature from 528 K to 573
K. In addition, for all samples the conversion ranged between 0.9% to 10.5%: at 573 K, 558 K,
15

and 543 K. The relative magnitude of CO
.5Y9In/ZrO > 2Y8In/ZrO > 3Y8In/ZrO . At 528 K, CO
from 0.9% to 1.3%. Methanol and CO were the only two carbon-containing products that were
found in this CO hydrogenation process (Figure 5 (b)), thus confirming that the In -based
2
conversion was: In
2
O
3
/ZrO
2
> 3La10In/ZrO >
2
1
2
2
2
2
conversion was similar for all catalysts
2
2
O
3
catalysts have high methanol selectivity. The selectivity to methanol decreased with increasing
temperature and CO production is more favorable at higher temperature due, in part, to the
endothermicity of the RWGS reaction (Figure 5 (d)). This has been also found in the CO
hydrogenation over the classical Cu-based catalysts [32–34]. The CO conversion and methanol
formation rate over In /ZrO in this work was higher (At 573 K, 10.5% and 0.465 gMeOH
h , respectively) than most other reports in the CO to methanol reaction [15,22,35–39] (also see
Table 1). The properties of the In /ZrO catalyst investigated here were very sensitive to
2
2
-1
2
O
3
2
g
cat.
-1
2
2
O
3
2
temperature (53-81% in methanol selectivity from 573 K to 528 K). Given the relatively low
pressure (40 bar) used in this investigation, the selectivity of this catalyst is expected to be lower
than the values reported by Martin et. al., that indicated 99.8% methanol selectivity over
In
2
O
3
/ZrO
2
catalyst at 50 bar [22]. In a more recent report [21] from the same group, a methanol
:CO ratios near 5 (see the supporting information of Ref. [21])
selectivity of 45%-65% with H
2
2
was observed under similar reaction conditions used in the previous publication [22]. Other reports
have also showed moderate selectivity, between 40-70% to methanol [15] [23] (Table 1),
suggesting that the high selectivity shown in Martin et. al.’s report is difficult to reproduce.
Rare earth elements, such as La, Ce, and Y were found to enhance the Cu surface area and
CO
improve the catalytic activity in CO
introduction of Y was seen as a method to increase the Cu dispersion and surface area for
2
adsorption [26,40–42]. Y
2
O
3
or La
2
O
3
has been used in Cu-based or Fe/Al
2
O
3
catalysts to
2
hydrogenation reaction [26,27,34,40,43–45]. Specifically,
2
O
3
16

Cu/Zn/ZrO
the density of basic sites and enhanced the methanol selectivity for 5-7% compared to the un-
promoted Cu-based catalysts. To further improve the selectivity, Y and La were
investigated as promoters on In-based catalysts by making the In /ZrO catalyst less reducible
and more basic. The methanol selectivity of YIn/ZrO and 3La10In/ZrO catalysts remained
around 70% even when the reaction temperature reached 573 K, while the un-promoted
In /ZrO had ~53% selectivity at the same reaction conditions. The increased selectivity occurs
in parallel with a reduction in catalytic rates: CO conversion at 573 K decreased from 10.5 % (on
In /ZrO ) to ~7.7 % (on 1.5Y9In/ZrO and 2Y9In/ZrO catalysts). The decrease of CO
conversion rate resulted in part from a suppression in the CO formation: at 573 K, the CO
2
[34], while Gao et. al. [40] found that Mn, La, Ce, Zr and Y modifiers could increase
2
O
3
2
O
3
2
O
3
2
2
2
2
O
3
2
2
2
O
3
2
2
2
2
-
1
formation rate of 1.5Y-9In/ZrO
formation rate over non-promoted In
decrease with yttrium addition, but the decrease is much less than it is with CO. At 573K for
2
was 0.164 gCO
g
cat. h-1, which is a reduction of 57% of the CO
cat. h-1). Methanol formation rates also
-1
2
O
3
/ZrO (0.380 gCO g
2
-1
example, 1.5Y-9In/ZrO
with In /ZrO . Note too that this trend is consistent with the changes in reducibility of the
samples as determined by the temperature of the reduction peak in the hydrogen TPR experiment
Figure 3). Figures 5(c) and 5(d) also show a change in the effective activation energy for the
formation of methanol and CO. For methanol, the higher the temperature, the closer the formation
rates of the promoted catalysts than the In /ZrO catalyst. In contrast, with CO, the higher the
reaction temperature, the higher the disparity between the CO formation rate of the promoted
catalysts versus In /ZrO
2
was 0.4 gMeOH g
cat. h-1, only 9% less than the formation rate observed
2
O
3
2
(
2
O
3
2
2
O
3
2
.
The promotion effects of Y and La are observed under all conditions investigated. For
example, an increase in methanol selectivity by about 20% is observed at all the temperatures (528
17

to 573 K). In fact, selectivity as high as 100% at low temperatures (528 K) could be reached for
both Y and La-promoted indium oxide catalysts. Moreover, under isoconversion conditions (see
Figure 5 (e)), e.g. 4%, methanol selectivity for the catalysts decreases in the following order:
3
La10In/ZrO
2
> 1.5Y9In/ZrO
2
> 2Y8In/ZrO
2
> 3Y8In/ZrO
2
> In
2
O
3
/ZrO , demonstrating that the
2
modification of the catalysts is a feasible way to boost catalytic performance.
Figure 6 shows an Arrhenius plot of the methanol formation rate from CO
2
hydrogenation
conversion was
less than 10% at all temperatures to ensure that the reactor operates under a differential regime.
The apparent activation energy for CO to methanol over In /ZrO , 1.5Y9In/ZrO , 2Y8In/ZrO
Y8In/ZrO , and 3La10In/ZrO were 66.3 kJ/mol, 91.4 kJ/mol, 92.0 kJ/mol, 95.7 kJ/mol, and 59.1
over In-based catalysts within the temperature range of 528 K and 573 K. The CO
2
2
2
O
3
2
2
2
,
3
2
2
kJ/mol, respectively. The La-promoted indium oxide catalyst showed the lowest activation energy
compared to the other catalysts. The activation energy of supported Cu-based catalysts is between
3
8 kJ/mol and 58 kJ/mol [46–50]; and the activation energy of bulk In
kJ/mol [21]. A higher activation energy (91.4-95.7 kJ/mol) was found in Y-promoted indium oxide
catalysts compared to In /ZrO . This implies that Y is directly affecting the active site of
In /ZrO catalyst since otherwise the kinetically relevant reaction would occur on the original
active sites. From the molecular stand point, the ionic radius of In , Zr , Y , and La are 92 pm,
7 pm, 106 pm, and 122 pm, respectively suggesting that the ionic radius of the ions, with similar
2
O
3
was found to be 103
2
O
3
2
2
O
3
2
3
+
4+
3+
3+
8
3
+
size with In , is more likely to affect the reaction on In
2
O
3
/ZrO catalysts, resulting in a different
2
activation energy.
The catalyst can also be compared with respect to methanol formation rate normalized by
mole percent of surface In (n%In, surface) (see Error! Reference source not found.). The normalized
methanol formation rate is almost the same at every temperature for In
2
O
3
/ZrO
2
and 3La10In/ZrO ,
2
18

but very different for the series of Y-incorporated samples, which is consistent with the apparent
activation energies in the reactions over these catalysts. The effect of reaction rates over
2
Y8In/ZrO
Reference source not found.). The methanol formation rate has a high dependence on H
pressure with the apparent reaction order of 1.07. CO , on the other hand, showed a negative
reaction order of -0.13 near H /CO = 4:1 relative compositions. This result is similar to values
recent reported by Frei et. al. [21] which showed that for bulk In , the reaction orders were
estimate at -0.1 for CO and 0.5 for H above the stoichiometric ratio of H /CO = 3:1. The negative
reaction order of CO on methanol formation rate was mainly due to the fact that higher CO
concentration would also favor the RWGS reaction and thus produce more CO. To reach high
methanol formation rates, an excess of H to CO ratio is beneficial.
2
catalyst on partial pressure of the reactants was investigated at 573 K (Error!
2
partial
2
2
2
2
O
3
2
2
2
2
2
2
2
2
From the above kinetic results, a power law rate expression and an Arrhenius form of the
rate constant with respect to temperature can be obtained:
풓푴풆푶푯 = 풌 [푯ퟐ]ퟏ.ퟎퟕ [푪푶ퟐ]ꢁퟎ.ퟏퟑ
ꢁ
ퟗퟔퟐퟕퟒ
풌 (푻) = ퟏ. ퟔퟔ × ퟏퟎ 풆( 푹푻
ퟕ
)
-1
-1
-1
where 풓푴풆푶푯 is the methanol formation rate in mol s kgcat. , 푘 is the rate constant in mol s kgcat.
-1
mol/L)-0.94, R is the gas constant (8.314 J mol-1 K-1).
(
The effect of co-feeding CO in the methanol formation rate was evaluated over 2Y8In/ZrO
2
at 543 K and 573 K (Error! Reference source not found.) to assess the kinetic effects of CO and
determine any potential recycling of the product stream, since CO is the main byproduct of the
process at high temperature. CO has been used earlier as a strategy to generate more active oxygen
vacancies [22]. However, a recent report showed that upon CO addition into the feed over bulk
19

In
more than 40 % loss in methanol reactivity at higher pressure (5.5 MPa) was observed [21]. In our
case, 2Y8In/ZrO appears to be unaffected by CO and displays a fixed methanol formation rate
within a ratio of CO/(CO+CO ) from 0 to 0.5. This implies that CO is the main carbon source of
methanol synthesis over this sample. In a CO-TPD experiment over 2Y8In/ZrO catalyst (Error!
2
O
3
, a slightly higher methanol reactivity at lower pressure (3.5 MPa) could be obtained, but
2
2
2
2
Reference source not found.), we found no evidence of stable CO adsorption complexes on the
surface after the catalyst was treated under reductive conditions.
The possible reaction channels for CO
extensively[12,21,31,51–54]. In general, the first step of the hydrogenation of adsorbed CO
lead to *HCOO (a formate intermediate, see Figure 7 (a)) or *HOCO (a carboxyl intermediate,
see Figure 7 (b) and (c)). Alternatively, adsorbed CO can dissociate into CO + O via the direct
2
hydrogenation to methanol have been discussed
2
can
2
C−O bond cleavage pathway, and the formed *CO can desorb as a product CO in gas phase [55],
but this path was not considered since CO is not easily hydrogenated to methanol on the indium
catalysts. On the formate pathway, *HCOO can be hydrogenated to *HCOOH, and then to
*
H
2
COOH, followed by bond cleavage to *H
COH, sequentially. For the carboxyl pathway, HOCO was likely to go through the RWGS
CO, *H CO,
COH, sequentially. Frei and co-workers [21] have modeled the methanol formation
mechanism over the 111 surface of In , while Ye et. al. [20] have modeled the same process
over 110 surface. Over these two surfaces, the preferred reaction path could involve H COOH* on
the 111 oxide surface, but involve H CO* over the 110 surface. Based on our own results, we
cannot clearly differentiate between these two alternatives.
2
CO+*OH, *H
2
CO+*H
2
O, *H COH, and finally
2
*
H
3
reaction and form *CO+ *OH. Further hydrogenation on *CO will form *HCO, *H
2
3
*
H
3
2
O
3
2
3
20

Assuming bidentate CO
2
adsorption, which is likely with indium oxide and Y-modified
zirconia surface according to the DFT calculations reported in [20][31] and the FTIR absorption
peaks of adsorbed CO
2
[56], the first step of hydrogenation of the adsorbed CO
2
can be either to
3
+
3+
the oxygen that is bonded to In or Y or to the carbon. Since Y is less reducible than In , the
activated hydrogen will be more likely to attack the O next to In rather than the O next to Y (see
Figure 7 (b) and (c)), leading to slower formation of *HOCO intermediate for the promoted
catalysts, and therefore, a lower formation rate of unwanted CO. The trade-off is lower CO
2
consumption rates for the promoted catalysts since the CO adsorbed near Y will be less likely to
2
react. The results of the CO co-feeding experiment (Error! Reference source not found.) and
CO-TPD (Error! Reference source not found.) also suggest that the reaction likely follows the
formate pathway (*HCOO) rather than the RWGS + CO Hydro pathway (*HOCO carboxyl
intermediate) or direct C-O bond cleavage pathway (*CO + *O) since CO was not a stable
adsorbate to react on the surface [55][57].
Conclusions
Y and La were used to promote the catalytic properties of supported In
2
O
3
(In
2
O
3
/ZrO )
2
for methanol synthesis by firstly mixing yttrium nitrate and lanthanum nitrate during the wet
impregnation process. Characterization studies show that the reducibility of the indium oxide
catalysts in a hydrogen atmosphere is correlated to methanol production rates and selectivity. The
Y and La-promoted catalysts had 11-35 K higher surface reduction temperatures compared to the
supported indium oxide catalyst. Flow microreactor studies indicated that the incorporation of Y
or La into the In
TPD traces show 60% to 70% more heat-induced oxygen vacancies (Ov1) in 2Y8In/ZrO
La10In/ZrO
2
O
3
/ZrO
2
catalysts increased the number of surface CO
2
adsorption sites. The CO
2
2
or
3
2
. The Y or La-promoted catalysts increased the selectivity of methanol significantly
21

(
about 20% more selective, from 528 K to 573 K). A selectivity of nearly 100% can be achieved
at 528K and 40 bar, which is a mild reaction condition compared to the commercial process (513-
33K, 50-100 bar). CO co-feeding experiment suggest a likely reaction route (formate pathway)
for methanol formation over 2Y8In/ZrO within the reaction conditions we have examined. These
5
2
results also show that the incorporation of Y and La into indium oxide catalysts is a feasible method
for the design of selective catalysts for sustainable methanol economy.
Acknowledgements
The authors are grateful to the U.S. Army for the funding under grant GTS-S-17-013. The authors
would also like to acknowledge Dr. Terry Dubois, Dr. Young Jim Kim, and Dr. Ting Jiang for the
help and suggestions.
References
[
1] A. Goeppert, M. Czaun, J.P. Jones, G.K. Surya Prakash, G.A. Olah, Chemical Society
Reviews 43 (2014) 7995–8048.
[
2] G.A. Olah, A. Goeppert, G.K.S. Prakash, Journal of Organic Chemistry 74 (2009) 487–
4
98.
[
[
[
3] G.A. Olah, Angewandte Chemie - International Edition 52 (2013) 104–107.
4] M.D. Porosoff, B. Yan, J.G. Chen, Energy Environ. Sci. 9 (2016) 62–73.
5] A. Goeppert, M. Czaun, J.-P. Jones, G.K. Surya Prakash, G.A. Olah, Chem. Soc. Rev. 43
(2014) 7995–8048.
[
6] M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Hävecker, F. Abild-Pedersen, S. Zander, F.
22

Girgsdies, P. Kurr, B.-L. Kniep, M. Tovar, R.W. Fischer, J.K. Nørskov, R. Schlögl,
Science 336 (2012) 893–7.
[
[
7] G.C. Chinchen, P.J. Denny, D.G. Parker, M.S. Spencer, D.A. Whan, Applied Catalysis 30
(1987) 333–338.
8] M.B. Fichtl, D. Schlereth, N. Jacobsen, I. Kasatkin, J. Schumann, M. Behrens, R. Schlögl,
O. Hinrichsen, Applied Catalysis A: General 502 (2015) 262–270.
[
[
9] L. Jia, J. Gao, W. Fang, Q. Li, Catalysis Communications 10 (2009) 2000–2003.
10] J. Toyir, P.R. de la Piscina, J.L.G. Fierro, N. Homs, Applied Catalysis B: Environmental
2
9 (2001) 207–215.
[
[
[
11] A. Le Valant, C. Comminges, C. Tisseraud, C. Canaff, L. Pinard, Y. Pouilloux, Journal of
Catalysis 324 (2015) 41–49.
12] J. Graciani, K. Mudiyanselage, F. Xu, a. E. Baber, J. Evans, S.D. Senanayake, D.J.
Stacchiola, P. Liu, J. Hrbek, J.F. Sanz, J. a. Rodriguez, Science 345 (2014) 546–550.
13] T. Fujitani, M. Saito, Y. Kanai, T. Watanabe, J. Nakamura, T. Uchijima, Applied
Catalysis A: General 125 (1995) 199–202.
[
[
14] L. Fan, K. Fujimoto, Applied Catalysis A: General 106 (1993) L1–L7.
15] K. Sun, Z. Fan, J. Ye, J. Yan, Q. Ge, Y. Li, W. He, W. Yang, C.J. Liu, Journal of CO2
Utilization 12 (2015) 1–6.
[
16] H. Lorenz, W. Jochum, B. Klötzer, M. Stöger-Pollach, S. Schwarz, K. Pfaller, S. Penner,
Applied Catalysis A: General 347 (2008) 34–42.
23

[
[
[
17] N. Iwasa, T. Mayanagi, N. Ogawa, K. Sakata, N. Takezawa, Catalysis Letters 54 (1998)
19–123.
1
18] T. Umegaki, K. Kuratani, Y. Yamada, A. Ueda, N. Kuriyama, T. Kobayashi, Q. Xu,
Journal of Power Sources 179 (2008) 566–570.
19] L. Wang, M. Ghoussoub, H. Wang, Y. Shao, W. Sun, A.A. Tountas, T.E. Wood, H. Li,
J.Y.Y. Loh, Y. Dong, M. Xia, Y. Li, S. Wang, J. Jia, C. Qiu, C. Qian, N.P. Kherani, L. He,
X. Zhang, G.A. Ozin, Joule 2 (2018) 1369–1381.
[
[
20] J. Ye, C. Liu, D. Mei, Q. Ge, ACS Catalysis 3 (2013) 1296–1306.
21] M.S. Frei, M. Capdevila-Cortada, R. García-Muelas, C. Mondelli, N. López, J.A. Stewart,
D. Curulla Ferré, J. Pérez-Ramírez, Journal of Catalysis 361 (2018) 313–321.
[
22] O. Martin, A.J. Martín, C. Mondelli, S. Mitchell, T.F. Segawa, R. Hauert, C. Drouilly, D.
Curulla-Ferré, J. Pérez-Ramírez, Angewandte Chemie International Edition 55 (2016) 1–
6
.
[
[
23] N. Rui, Z. Wang, K. Sun, J. Ye, Q. Ge, C. jun Liu, Applied Catalysis B: Environmental
18 (2017) 488–497.
2
24] A. García-Trenco, A. Regoutz, E.R. White, D.J. Payne, M.S.P. Shaffer, C.K. Williams,
Applied Catalysis B: Environmental 220 (2018) 9–18.
[
[
25] M. Dou, M. Zhang, Y. Chen, Y. Yu, Surface Science 672–673 (2018) 7–12.
26] T. Inui, T. Takeguchi, A. Kohama, K. Tanida, Energy Conversion and Management 33
(1992) 513–520.
[
27] P. Gao, L. Zhong, L. Zhang, H. Wang, N. Zhao, W. Wei, Y. Sun, Catal. Sci. Technol. 5
24

(2015) 4365–4377.
[
28] M. Wagner, P. Lackner, S. Seiler, A. Brunsch, R. Bliem, S. Gerhold, Z. Wang, J. Osiecki,
K. Schulte, L.A. Boatner, M. Schmid, B. Meyer, U. Diebold, ACS Nano 11 (2017)
1
1531–11541.
[
[
29] M. Wagner, S. Seiler, B. Meyer, L.A. Boatner, M. Schmid, U. Diebold, Advanced
Materials Interfaces 1 (2014) 1–6.
30] T. Bielz, H. Lorenz, W. Jochum, R. Kaindl, F. Klauser, B. Kl??tzer, S. Penner, Journal of
Physical Chemistry C 114 (2010) 9022–9029.
[
[
31] J. Ye, C. Liu, Q. Ge, The Journal of Physical Chemistry C 116 (2012) 7817–7825.
32] K. Kobl, S. Thomas, Y. Zimmermann, K. Parkhomenko, A.C. Roger, Catalysis Today 270
(2016) 31–42.
[
[
33] J.F. Portha, K. Parkhomenko, K. Kobl, A.C. Roger, S. Arab, J.M. Commenge, L. Falk,
Industrial and Engineering Chemistry Research 56 (2017) 13133–13145.
34] J. Słoczyński, R. Grabowski, P. Olszewski, A. Kozłowska, J. Stoch, M. Lachowska, J.
Skrzypek, Applied Catalysis A: General 310 (2006) 127–137.
[
[
35] X.M. Liu, G.Q. Lu, Z.F. Yan, Applied Catalysis A: General 279 (2005) 241–245.
36] J. Toyir, P.R. De la Piscina, J.L.G. Fierro, N. Homs, Applied Catalysis B: Environmental
3
4 (2001) 255–266.
[
37] A. Ota, E.L. Kunkes, I. Kasatkin, E. Groppo, D. Ferri, B. Poceiro, R.M. Navarro Yerga,
M. Behrens, Journal of Catalysis 293 (2012) 27–38.
25

[
38] T. Witoon, J. Chalorngtham, P. Dumrongbunditkul, M. Chareonpanich, J. Limtrakul,
Chemical Engineering Journal 293 (2016) 327–336.
[
[
39] Y. Hartadi, D. Widmann, R.J. Behm, Journal of Catalysis 333 (2016) 238–250.
40] P. Gao, F. Li, N. Zhao, F. Xiao, W. Wei, L. Zhong, Y. Sun, Applied Catalysis A: General
4
68 (2013) 442–452.
41] X. Guo, D. Mao, G. Lu, S. Wang, G. Wu, Journal of Molecular Catalysis A: Chemical 345
2011) 60–68.
42] G. Avgouropoulos, T. Ioannides, H. Matralis, Applied Catalysis B: Environmental 56
2005) 87–93.
[
[
[
[
(
(
43] S.-S. Nam, G. Kishan, M.-W. Lee, M.-J. Choi, K.-W. Lee, Applied Organometallic
Chemistry 14 (2000) 794–798.
44] S. Natesakhawat, J.W. Lekse, J.P. Baltrus, P.R. Ohodnicki, B.H. Howard, X. Deng, C.
Matranga, ACS Catalysis 2 (2012) 1667–1676.
[
[
45] L. Znak, K. Stołecki, J. Zieliński, Catalysis Today 101 (2005) 65–71.
46] J. Graciani, K. Mudiyanselage, F. Xu, A.E. Baber, J. Evans, S.D. Senanayake, D.J.
Stacchiola, P. Liu, J. Hrbek, J. Fernández Sanz, J.A. Rodriguez, Science 345 (2014) 546–
5
50.
[
[
47] Y. Yang, J. Evans, J.A. Rodriguez, M.G. White, P. Liu, Physical Chemistry Chemical
Physics 12 (2010) 9909.
48] S. Du, W. Tang, X. Lu, S. Wang, Y. Guo, P. Gao, Advanced Materials Interfaces 5 (2018)
1
700730.
26

[
[
[
[
[
49] K. Chang, T. Wang, J.G. Chen, Applied Catalysis B: Environmental 206 (2017) 704–711.
50] A. Karelovic, P. Ruiz, Catal. Sci. Technol. 5 (2015) 869–881.
51] L.C. Grabow, M. Mavrikakis, ACS Catalysis 1 (2011) 365–384.
52] M. Behrens, Angewandte Chemie International Edition (2014) 12022–12024.
53] E.L. Kunkes, F. Studt, F. Abild-Pedersen, R. Schlögl, M. Behrens, Journal of Catalysis
3
28 (2015) 43–48.
[
[
[
[
54] K. Larmier, W.-C. Liao, S. Tada, E. Lam, R. Verel, A. Bansode, A. Urakawa, A. Comas-
Vives, C. Copéret, Angewandte Chemie International Edition (2017) 2358–2363.
55] S. Kattel, P. Liu, J.G. Chen, Journal of the American Chemical Society 139 (2017) 9739–
9
754.
56] E.-M. Köck, M. Kogler, T. Bielz, B. Klötzer, S. Penner, The Journal of Physical
Chemistry C 117 (2013) 17666–17673.
57] S. Kattel, B. Yan, Y. Yang, J.G. Chen, P. Liu, Journal of the American Chemical Society
1
38 (2016) 12440–12450.
[
[
58] R. Gaikwad, A. Bansode, A. Urakawa, Journal of Catalysis 343 (2016) 127–132.
59] J. Słoczyński, R. Grabowski, A. Kozłowska, P. Olszewski, J. Stoch, J. Skrzypek, M.
Madej-Lachowska, Catalytic Activity of the M/(3ZnO·ZrO2) System (M=Cu, Ag, Au) in
the Hydrogenation of CO2 to Methanol, 2004.
[
60] J. Toyir, P. Ramírez De La Piscina, J.L.G. Fierro, N. Homs, Applied Catalysis B:
Environmental 29 (2001) 207–215.
27

[
61] X.L. Liang, X. Dong, G.D. Lin, H. Bin Zhang, Applied Catalysis B: Environmental 88
2009) 315–322.
(
28

Figure 1. SEM images of (a) In
2
O
3
/ZrO
2
, (b) 1.5Y9In/ZrO
2
, and (c) 3La10In/ZrO .
2
29