Metal-Organic Framework-Derived Synthesis of Cobalt Indium Catalysts for the Hydrogenation of CO2 to Methanol

Article abstract of DOI:10.1021/acscatal.0c00449

Methanol synthesis by means of direct CO₂ hydrogenation has the potential to contribute to climate change mitigation by turning the most important greenhouse gas into a commodity. However, for this process to become industrially relevant, catalytic systems with improved activity, selectivity, and stability are required. Here, we explore the potential of metal-organic frameworks (MOFs) as precursors for synthesis of Co₃O₄-supported In₂O₃ oxide composites for the direct CO₂ hydrogenation to methanol. Stepwise pyrolytic-oxidative decomposition of indium-impregnated ZIF-67(Co) MOFs affords the formation of a nanostructured In₂O₃?Co₃O₄ reticulated shell composite material able to reach a maximum methanol production rate of 0.65 g_MeOH·g_cat ^-1·h^-1 with selectivity as high as 87% over 100 h on stream. Textural characteristics of the sacrificial ZIF-67(Co) matrix and In-loading were found to be important variables for optimizing the catalyst performance such as induction time, methanol productivity, and selectivity. The structural investigation on the catalytic system reveals that the catalyst undergoes reorganization under reaction conditions, transforming from Co₃O₄ with an amorphous In₂O₃ shell into Co₃InC_0.75 covered by a layer consisting of a mixture of amorphous CoO_x and In₂O₃ oxides. Structural reorganization is responsible for the observed induction period, while the amorphous mixed cobalt indium oxide shell is responsible for the high methanol yield and selectivity. Additionally, these results demonstrate the tunable performance of MOF-derived In₂O₃?Co₃O₄ catalysts as a function of the reaction conditions which allows us to establish a reasonable trade-off between high methanol yield and selectivity in a wide temperature and pressure window.

Full text of DOI:10.1021/acscatal.0c00449

Subscriber access provided by University of Wollongong Library
Article
Metal Organic Framework Derived Synthesis of Cobalt
Indium Catalysts for the Hydrogenation of CO2 to methanol
Alexey Pustovarenko, Alla Dikhtiarenko, Anastasiya Bavykina, Lieven E. Gevers, Adrian Ramirez, Artem
Russkikh, Selvedin Telalovic, Antonio Aguilar, Jean Louis Hazemann, Samy Ould-Chikh, and Jorge Gascon
ACS Catal., Just Accepted Manuscript • DOI: 10.1021/acscatal.0c00449 • Publication Date (Web): 03 Apr 2020
Downloaded from pubs.acs.org on April 4, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
Metal Organic Framework Derived Synthesis of
Cobalt Indium Catalysts for the Hydrogenation of
CO2 to methanol
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Alexey Pustovarenko,† Alla Dikhtiarenko,† Anastasiya Bavykina,† Lieven Gevers,† Adrian
Ramírez,† Artem Russkikh, † Selvedin Telalovic,† Antonio Aguilar,§ Jean-Louis Hazemann,§ Samy
Oud-Chikh† and Jorge Gascon†,*
† King Abdullah University of Science and Technology, KAUST Catalysis Center (KCC),
Advanced Catalytic Materials, Thuwal 23955, Saudi Arabia
§ Néel, UPR2940 CNRS, University of Grenoble Alpes, F-38000 Grenoble, France
KEYWORDS: metal-organic framework derived route, direct CO2 hydrogenation, methanol
synthesis, indium oxide, cobalt oxide.
1
ACS Paragon Plus Environment

ACS Catalysis
Page 2 of 39
1
2
3
4
ABSTRACT
5
6
7
8
9
Methanol synthesis by means of direct CO2 hydrogenation has the potential to contribute to climate
change mitigation by turning the most important greenhouse gas into a commodity. However, for
this process to become industrially relevant, catalytic systems with improved activity, selectivity
and stability are required. Here we explore the potential of metal-organic frameworks (MOF) as
precursors for synthesis of Co3O4-supported In2O3 oxide composites for the direct CO2
hydrogenation to methanol. Stepwise pyrolytic-oxidative decomposition of indium-impregnated
ZIF-67(Co) MOF affords the formation of a nanostructured In2O3@Co3O4 reticulated shell
composite material able to reach a maximum methanol production rate of 0.65 gMeOH·gcat-1·h-1 with
selectivity as high as 87% over 100 h on stream. Textural characteristics of the sacrificial ZIF-
67(Co) matrix and In-loading were found to be important variables for optimizing the catalyst
performance such as induction time, methanol productivity and selectivity. The structural
investigation on the catalytic system reveals that the catalyst undergoes reorganization under
reaction conditions, transforming from a Co3O4 with amorphous In2O3 shell into Co3InC0.75 covered
by a layer consisting of a mixture of amorphous CoOx and In2O3 oxides. Structural reorganization
is responsible for the observed induction period, while the amorphous mixed cobalt indium oxide
shell is responsible for the high methanol yield and selectivity. Additionally, these results
demonstrate the tunable performance of MOF-derived In2O3@Co3O4 catalyst as a function of the
reaction conditions which allows to establish a reasonable trade-off between high methanol yield
and selectivity in a wide temperature and pressure window.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
ACS Paragon Plus Environment

Page 3 of 39
ACS Catalysis
1
2
3
4
5
1. Introduction
6
7
8
9
Anthropogenic CO2 emissions increase year by year, with very serious consequences for our
climate. In order to solve these environmental and energy issues, intensive research is being carried
out into technologies that may eventually turn our economy into a low-carbon one.1 In this scenario
of a circular carbon economy, carbon dioxide is viewed as an alternative resource that can be
recycled to produce valuable chemicals and fuels,2-4 turning, in such a way, waste to wealth.
Among the potential products derived from CO2, methanol has attracted particular attention due to
its relevance not only as an alternative fuel 5-6 but also as a convenient chemical feedstock.7-11
Indeed, methanol synthesis through direct hydrogenation of CO2, using pure H2 and CO2 as
starting materials, is considered to be a very promising strategy to utilize CO2 when green
hydrogen (produced via electrolysis using renewable energy) is available.2,11-13 At the current stage,
the Cu-ZnO-Al2O3 catalyst commercially used in the production of methanol from synthesis gas
(CO/CO2/H2) is the benchmark.14-15 In direct CO2 hydrogenation, this catalyst is able to drive, to
some extent, the production of methanol, however exhibiting low selectivity and fast deactivation
caused by competitive reverse water gas shift reaction (RWGS).15-17 To address these selectivity
and stability issues, a number of multi-metallic composite systems have been proposed, i.e.
CuO/ZrO2, CuO/ZnO/ZrO2, CuO/ZnO/Ga2O3, Cu/ZnO/ZrO2/Al2O3/SiO2,18-21 and their doped
variations (i.e., by Ag, Au, Pt, Pd, Rh promotors).22-27 These mix-component systems are able to
convert CO2 to methanol with high selectivities at low conversions per pass, such as in case of
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cu@ZnO core-shell (SMeOH ~ 100 %, XCO2 conversion of 2.3 %)21 or Au-doped ZnO-ZrO2 (SMeOH
~
100%, XCO2 ~ 1.5 %).27 Contrariwise, when higher CO2 conversions are targeted, low methanol
selectivities are usually achieved. This is the case, for instance, for Cu-K/Al2O3 (SMeOH = 2 % at X
= 29)28 and for Cu/Zn/Al/Y (SMeOH = 52 % at X = 27).29
3
ACS Paragon Plus Environment

ACS Catalysis
Page 4 of 39
1
2
3
4
5
6
7
8
9
Among recent studies on simpler metal oxide systems featuring higher stability, density
functional theory (DFT) calculations predict that methanol synthesis via CO2 hydrogenation is
favored over In2O3 surfaces.30 The proposed catalytic mechanism involves cyclic generation of
oxygen vacancies on In2O3 by H2 and their replenishment via CO2 activation. Later, Sun et. al.31
experimentally proved the superior activity of In2O3 for CO2 hydrogenation to methanol. When
In2O3 supported is supported on ZrO2, further improvements in catalytic performance have been
reported: at low conversion levels, almost 100% selectivity to methanol with space time yield
(STY) of 0.32 gMeOH·gcat-1·h-1 and stable performance over 1000 h on stream can be achieved.32 In
light of recent advances made in In-based bimetallic formulations33 with improved catalytic merits
and inspired by the drastic importance of carrier supports in such catalytic systems, Bavykina et.
al.34 prepared a pre-catalyst consisting in In2O3 supported on Co3O4 by means of reverse co-
precipitation. The resulting In@Co system catalyzes methanol formation via CO2 hydrogenation
with selectivities above 80% and productivity of 0.86 gMeOH·gcat-1·h-1, at conversion levels close to
thermodynamic equilibrium.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
It is well known that, in heterogeneous catalysis, the preparation method significantly affects
catalytic performance.25 Over the last decade, particular interest has been paid to Metal-Organic
Framework (MOF)-supported and MOF mediated synthetic approaches to prepare highly effective
catalysts for a variety of catalytic transformations.35-37 Being microporous crystalline materials
with a tailored arrangement of metal-centers and bridging linkers,38 MOFs dispose a system of
pores that can be loaded with active components offering an optimal distribution of active sites.
For instance, an encapsulation of Cu and/or ZnO nanoparticles within UiO-66(Zr) 39,40 resulted in
highly active catalyst showing an 8-fold enhanced yield and 100% selectivity towards methanol
(albeit at very low conversions, where the effect of RWGS is minimized) when compared to the
4
ACS Paragon Plus Environment

Page 5 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
benchmark Cu/ZnO/Al2O3. Besides the promising perspectives of MOF-supported catalysts, recent
investigations reveal the advantages in preparation of highly efficient heterogeneous catalytic
materials from metal-organic framework supports/precursors by means of thermal
decomposition.37,41-42 MOF-derived nanocomposite materials, i.e. MOF-derived carbons and
oxides, feature a range of advantages, such as high specific area, porosity and excellent distribution
of the active component within the host matrix, compared to similar materials prepared by co-
precipitation or impregnation methods.37,42
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In light of these results, here we explore the MOF mediated route for the preparation of a highly
efficient Co3O4-supported In2O3 catalyst in line with our previous findings on this inorganic
system.34 The impregnation of ZIF-67(Co) with indium, followed by thermal decomposition,
results in the formation of In@Co oxide composite materials with enhanced catalytic behavior in
CO2 hydrogenation to methanol, revealing higher methanol yields, selectivities and significantly
shortened induction times (10 h vs. 30 h observed for In@Co prepared by co-precipitation).34 The
influence of the crystal size of the parent ZIF-67(Co) and the amount of loaded indium on the
catalytic performance has been thoroughly studied. Moreover, the importance of two-step thermal
treatment (pyrolysis followed by calcination) for In-impregnated ZIF-67(Co) material in order to
obtain highly performing catalysts is also discussed in detail.
2. Experimental section
2.1. Materials
All chemicals were purchased from Sigma-Aldrich and used as received. Commercial Cu-ZnO-
Al2O3 methanol synthesis catalyst was obtained from Alfa Aesar™ (composition: 60-68% CuO,
22-26% ZnO, 8-12% Al2O3, 1-3% MgO).
5
ACS Paragon Plus Environment

ACS Catalysis
Page 6 of 39
1
2
3
4
2.2. Synthetic procedures
5
6
7
8
9
Synthesis of 300 nm ZIF-67(Co). The synthesis of ZIF-67(Co) with a crystallite size around 300
nm was performed in accordance to the previously reported method.43-44 Following the typical
synthetic protocol, 2.933 g (10 mmol) of Co(NO3)2·6H2O and 6.489 g (79 mmol) of 2-
methylimidazole (MeIm) were separately dissolved in 200 mL of methanol. The MeIm-containing
clear solution was rapidly poured into the pink solution of the cobalt salt. The mixture was stirred
for 12 min and then kept standing for 24 h to allow precipitation to take place. The bright purple
product was collected by centrifugation, washed several times with methanol, and dried at 423 K
for 10 h under vacuum.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Synthesis of 22 nm ZIF-67(Co). ZIF-67(Co) with a crystallite size around 22 nm was obtained
through modulated synthesis with triethylamine (TEA). In typical procedure, 0.582 g (2 mmol) of
Co(NO3)2·6H2O and 4.866 g (60 mmol) of 2-methylimidazole (MeIm) were dissolved in 400 mL
methanol. Afterwards, 5 mL of triethylamine was added to the vigorously stirred mixture and
stirring continued for 15 min. When precipitation started, the mixture was kept steady for 24 h.
The product was collected by centrifugation, washed several times with methanol, and dried at 423
K for 10 h under vacuum.
Synthesis of MOF-derived In@Co catalysts by incipient wetness impregnation method. The
catalysts were prepared by incipient wetness impregnation (IWI) in the following manner: 1 g of
activated ZIF-67(Co) was impregnated with 3 mL of aqueous indium nitrate solution of different
indium concentrations as stated in Table S1 and then dried under vacuum at 100 °C overnight. For
thermal treatment, an In-impregnated ZIF-67(Co) precursor sample was transferred into a quartz
reactor (L = 1 m and ID = 5 cm), vertically inserted into tubular oven equipped with temperature
(Nabertherm) and mass-flow controllers (Bronkhorst® High-Tech), and followed by two thermal
6
ACS Paragon Plus Environment

Page 7 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
treatment steps in sequential order: first, pyrolysis at 600 °C (1 °C·min-1, 4 h) in nitrogen flow (25
mL·min-1), then calcination at 400 °C (1 °C·min-1, 2 h) in air flow (25 mL·min-1). The obtained
In@Co samples are denoted as aIn:bCo(x) where x is the crystal size of the ZIF-67(Co) precursor,
a and b indicate the desired In:Co molar ratio.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2.3. Characterization methods
Powder X-ray diffraction (PXRD) patterns were acquired on a Bruker D8 Advance operated at
30 kV and 30 mA using monochromatic Cu-Ka (l = 1.5418 Å) radiation, a scan speed of 15 s per
step and a step size of 0.05° on 10 – 80° 2q range. For phase analysis of reacted catalysts, the X-
ray diffraction data were measured under inert atmosphere using an airtight sample holder which
was loaded in a glove box under Ar atmosphere. The crystalline phase was identified with the help
of the PDF-4+ (2019) crystal database. The profile matching of all PXRDs with expected
crystalline phases was confirmed by Pawley fitting 45 and quantitative estimation of the phases was
obtained by Rietveld refinement using TOPAS V5 (academic version) software.46 The results of
refinements are provided in Fig. S28.
Nitrogen adsorption experiments were carried out at 77 K using a Micromeritics ASAP 2040
instrument. Before the measurements, the ZIF-67(Co) samples were degassed at 150 °C for 10
hours and the calcined products were degassed at 100 °C for 10 hours.
Thermogravimetric data (TG) were collected in nitrogen and air atmospheres using a Mettler
Toledo thermal analyzer at a heating rate of 5 °C·min-1 in the 25 – 950 °C temperature range and
at a dynamic gas flow rate of 20 mL·min-1.
Scanning electron microscopy (SEM) imaging was performed with secondary electrons on a FEI
TENEO VS microscope using 5 kV acceleration voltage and 5 mm working distance. The crystal
7
ACS Paragon Plus Environment

ACS Catalysis
Page 8 of 39
1
2
3
4
5
6
size distribution for each sample was obtained analyzing several images recorded at different
magnifications using ImageJ software.47
7
8
9
The annular Dark-Field scanning transmission electron microscopy (ADF-STEM) in
conjunction with Electron Energy Loss Spectroscopy (EELS) study was carried out with a Cs-
Probe Corrected Titan microscope (Thermo-Fisher Scientific) which was also equipped with a GIF
Quantum of model 966 from Gatan Inc. (Pleasanton, CA). STEM-EELS analysis was performed
by operating the microscope at the accelerating voltage of 300 kV, using a convergence angle a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
of 17 mrad and a collection angle b of 38 mrad. Spectrum-imaging dataset includes the
simultaneous acquisition of zero-loss and core-loss spectra (DualEELS) using a dispersion of 0.5
eV/channel and were recorded using a beam current of 0.2 nA and a dwell time of 20 ms/pixel.
The Co L2,3-edge, In M4,5-edge, and O K-edge were selected to build the chemical maps. Carbon
K-edge was acquired by continuously scanning a Co3InC0.75 nanoparticle while simultaneously
acquiring 10 Dual-EELS spectra (frame time = 10 s) using a dispersion of 0.1 eV/channel. The
energy scale of all high loss EELS spectra was referenced to the zero-loss peak. Carbon K-edge
spectra were normalized using ATHENA software.48 Typically, the specimens were prepared by
dispersing the catalyst in methanol under ultrasound irradiation during 30 min and placing several
drops of the resulting dispersion on the copper grids having a carbon layer. For the particular study
of carbon K-edge, a nickel grid with a lacey-carbon layer was used to record EELS spectra without
the interference of the supporting carbon. The QSTEM software49 was used to simulate high-
resolution lattice images of the Co3InC0.75 nanoparticle along the <101> zone axis using a slab
thickness of 4.8 nm. For the microscope parameters, we used an accelerating voltage of 300 kV, a
spherical aberration coefficient (Cs) of 1.0 mm, a defocus (df) of -1.7 nm, a chromatic aberration
coefficient (Cc) of 1 mm, an energy spread (dE) of 1.0 eV, a convergence angle a = 17 mrad, and
8
ACS Paragon Plus Environment

Page 9 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
a detector collection angle of [63, 200] mrad. Ab-initio simulations of the carbon K-edge were
performed using the FDMNES package. The FDMNES code features mono-electronic
calculations, which are carried out in real space with Hedin-Lundqvist exchange-correlation
potential and using clusters built around each non-equivalent absorbing atom.50-51 The finite-
difference method (FDM) implemented in the FDMNES code was used since the latter is a full
potential method that introduces no approximation on the shape of the potential. To take into
account the core-hole lifetime and other multielectronic phenomena occurring in the absorption
process, a convolution procedure was applied to the calculated spectra. At the Fermi level, the
Lorentzian (FWHM) width was set to 0.16 eV. The instrumental energy resolution was taken into
account by further convoluting the calculated spectra with a Gaussian broadening of 1.5 eV width.
The XAS experiments were performed at CRG-FAME beamline (BM30B) at the European
Synchrotron Radiation Facility (ESRF). The ring is operated at 6 GeV with a nominal current of
200 mA in 7/8+1 mode. The beamline is equipped with a liquid-nitrogen-cooled doubled crystal
Si(220) monochromator surrounded by two Rh-coated mirrors for harmonic rejection. The beam
size on the sample is 240×110 μm (H×V, FWHM). Cobalt K-edge spectra were collected in
fluorescence mode using a CANBERRA 30-elements Ge solid state detector. Indium K-edge
spectra were collected in transmission mode using silicon diodes to measure the incident and
transmitted beam intensities. The monochromator was energy calibrated measuring the cobalt and
indium K absorption edges using a metallic cobalt and indium foil respectively. First maximum of
the 1st derivative of the absorption cobalt K-edge was set at 7709 eV and 27940 eV for indium K-
edge. The data acquisitions of reacted catalysts were performed ex-situ either in fluorescence or
transmission mode with the samples retrieved inside a glove box and loaded into a XAS cell
dedicated for air-sensitive compounds. Non-air sensitive fresh catalyst and reference compounds
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
ACS Paragon Plus Environment

ACS Catalysis
Page 10 of 39
1
2
3
4
5
6
7
8
9
were handled outside of glove box and measured in fluorescence mode using conventional sample
holder. All samples were diluted with boron nitride and pressed into pellets to optimize the
effective XAFS edge-step. On average four scans were acquired for each sample to improve the
signal to noise level of the data. All XAS data were analyzed using the HORAE package, a
graphical interface to the AUTOBK and IFEFFIT code.48 The EXAFS spectra were obtained after
performing standard procedures for pre-edge subtraction, normalization, polynomial removal, and
wavevector conversion. The amplitude factors (S02) were fitted to the EXAFS spectra recorded for
a Co metallic foil and an In2O3 powder (Table S3 and Figure S18). S02 were determined to be 0.73
± 0.04 and 0.99 ± 0.03 for Co and In, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
CHN elemental analyses on selected samples were performed using a Thermo Flash 2000
Organic Elemental Analyzer with the detection limits of 0.2 % (w/w) for carbon, 0.1% for nitrogen
and 0.08% for hydrogen. For the analysis of the used catalysts, the corresponding samples were
transferred unopened from the reactor to the glovebox and were kept under Ar atmosphere. To
avoid undesired reaction of the used catalysts with air, all samples were sealed in tin sample
crucibles in a glove box under Ar atmosphere. For each sample three individual measurements
were performed and the mean was calculated as a final result.
X-ray fluorescence measurements (XRF) were performed on a HORIBA XGT-700 XRF
analyzer. For every measurement, three different spots were analyzed for each sample with a total
acquisition time of 1500 s per sample.
2.4. Catalytic testing in CO2 hydrogenation to methanol
Catalytic tests were performed using a parallel reactor system Flowrence® from Avantium
consisting of 16 tubular fix-bed reactors installed in a furnace and operated at one mixed feed gas
flow which is distributed over 16 channels with a relative standard deviation of 2%. 50 mg of the
10
ACS Paragon Plus Environment

Page 11 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
catalyst with particle fraction between 150 µm and 250 µm was loaded onto the stainless-steel tube
reactor (30 cm long with 2 mm of internal diameter) preliminary filled with 9.5 cm bed of coarse
SiC (particle grit 40, 300 µL) in order to ensure the isothermal zone for the catalytic bed. Before
the catalytic test, the catalysts were pretreated in-situ under N2 at 300 °C for 1 hour. Then, the
reactors were pressurized with a mixed feed containing 20 vol.% of CO2 and 80 vol.% of H2 to 50
bar using a membrane-based pressure controller. In addition, 0.5 mL/min of He was mixed with
the feed and used as internal standard. The products were analyzed with an Agilent 7890B
chromatograph equipped with two loops, where one was connected to the Column 5 HaysepQ6 Ft
G3591-80013 and TCD and the second Gaspro 30M, 0.32 mm OD column followed by FID.
Conversion (X, %), space time yield (STY, gi·gcat-1·h-1), and selectivity (S, %) are defined as
follows:
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ꢅꢆꢇ,ꢈꢉꢊ ∙ ꢅꢁꢂ ,ꢋ
ꢃ
ꢀꢁꢂ = ꢄ1 −
ꢌ ∙ 100
ꢃ
ꢅꢆꢇ,ꢋ ∙ ꢅꢁꢂ ,ꢈꢉꢊ
ꢃ
ꢅꢎ,ꢋ
ꢅꢆꢇ,ꢋ
ꢏ
ꢐ
ꢍꢎ =
∙ 100
ꢅꢁꢂ ,ꢈꢉꢊ
ꢅꢆꢇ,ꢈꢉꢊ
ꢅ
ꢅꢆꢇ,ꢋ
ꢃ
ꢃ
ꢏ
−
ꢁꢂ ,ꢋ ꢐ
ꢀꢁꢂ
100 100
ꢍꢎ
ꢃ
∙
∙ ꢓꢎ ∙ ꢔꢕꢍꢖ
ꢍꢑꢒ =
ꢎ
22.4
where CHe,blk, CHe,R, CCO2,blk, CCO2,R are the concentrations determined by GC analysis of He in the
blank, He in the reactor effluent, CO2 in the blank, and CO2 in the reactor effluent, respectively,
11
ACS Paragon Plus Environment

ACS Catalysis
Page 12 of 39
1
2
3
4
5
6
Ci,R is the concentration of the product determined by GC, Mi is the molecular weight of product i
and GHSV is CO2 gas hourly space velocity in L·gcat-1·h-1.
7
8
9
Carbon balance in all experiments accounted for more than 97.8% of the total carbon input whereas
a small amount of carbon trapped in form of mixed metal carbide was comparatively modest.
Reproducibility of the catalytic experiment was carried out for representative sample and was
found to be deviating within 1.7% relative experimental error.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. Results and Discussion
In our study we explored ZIF-67(Co) framework as a precursor material for the preparation of
indium loaded MOF-derived cobalt oxide supported catalysts (In@Co). ZIF-67(Co) is a zeolitic
imidazolate framework formed by cobalt and 2-methylimidazole (MeIm) featuring a porous
architecture with sodalite (sod) topology (Fig. 1A-B), and has been chosen as a sacrificial material
due to its high cobalt content (26 wt.% Co). We synthesized two types of ZIF-67(Co) samples with
different crystal sizes in order to evaluate the effect they may induce on the performance of the
derived In@Co catalysts. Thus, applying a modulated synthesis protocol, ZIF-67(Co) with average
crystal sizes of 20 nm and 300 nm were obtained. These samples are referred as ZIF-67(Co)-20
and ZIF-67(Co)-300, respectively. As it can be observed from Figures 1D-E (insets) and S1, both
samples possess narrow size distribution profiles. To confirm matching of ZIF-67(Co)-20 and -
300 with the ZIF-67 structure, powder X-ray diffraction analyses have been performed and
compared in Figure 1C. Both materials reveal the typical reflections attributed to the cubic ZIF-67
structure.52-53 However, the notable broadening of the peaks in case of ZIF-67(Co)-20 indicates
that the sample has a smaller crystal size compared to the ZIF-67(Co)-300.
12
ACS Paragon Plus Environment

Page 13 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1 (double column). (A) Cage structure of ZIF-67(Co) and (B) its topological view reveling
sod fragment. (C) Powder X-ray diffraction patterns compared to theoretically expected ZIF-
67(Co) (cubic I-43m);53 (D-E) SEM micrographs along with size distribution histograms (insets)
and (F) N2 adsorption isotherms acquired at 77 K for ZIF-67(Co) precursor samples of 300 ± 85nm
(blue symbols) and 20 ± 5 nm (violet symbols) crystal sizes, respectively. On the isotherm graph
the filled markers correspond to the adsorption curve and the hollow ones to desorption branch.
In order to evaluate the textural properties of the ZIF-67(Co)-20 and -300 materials with
different crystal sizes, nitrogen adsorption experiments have been carried out. Figure 1F shows
corresponding nitrogen sorption isotherms for both 20 nm and 300 nm ZIF-67(Co) precursors. The
BET analyses reveal very close microporous surface areas of 1579 and 1566 m2·g-1 for ZIF-67(Co)-
300 and -20, respectively (Table S2). Notably, the N2 adsorption isotherm for ZIF-67(Co)-300
13
ACS Paragon Plus Environment

ACS Catalysis
Page 14 of 39
1
2
3
4
5
6
7
8
exhibits a typical type I shape, corroborating the fully microporous nature of the material, whereas
ZIF-67(Co)-20 features IV-type with a hysteresis loop at high relative pressures, most likely
derived from interparticle condensation of N2.54-55
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Considering that in our study the catalysts are prepared by thermal decomposition of ZIF-
67(Co), thermogravimetric analysis is a necessary tool to determine optimal conditions for the
treatment. ZIF-67(Co) decomposes in nitrogen atmosphere at temperatures above 550 °C (Fig. S2)
yielding Co nanoparticles of ~ 4.7 nm size highly dispersed within a carbon matrix, which mainly
consists of graphite nanotubes (Fig. S3). The PXRD data for the Co@C material derived from
pristine ZIF-67(Co) show (Fig. S4) broad peaks originated from graphite (hcp C allotrope) and
metallic Co (fcc form, a-Co allotrope) phases. Notably, the pyrolysis of ZIF-67(Co) at 600 °C
results in the formation of a-Co phase,56 allotrope which is expected to be thermodynamically
favored at temperatures above 417 °C and exists as a metastable phase at room temperature.56 In
contrast, the decomposition of ZIF-67(Co) in air flow proceeds in two steps (Fig. S2): i) oxidative
destruction of the framework with a weight loss of ca. 63 wt.% yielding Co3O4 as main
decomposition product at 400 °C; and ii) transformation of Co3O4 to CoO 57 occurring at 900 °C
with additional 2.7 wt.% mass loss. In a similar way, Co@C decomposes during the calcination
step (Fig. S5), losing 43 wt.% due to carbon oxidation resulting in Co3O4 at 400 °C. The product
obtained by two-step decomposition comprises uniform nanoparticles of 5.8 nm average size (Fig.
S6) with the spinel Co3O4 structure (Fig. S7).
Typically, we prepared In@Co catalysts in three steps: i) incipient wetness impregnation of ZIF-
67(Co) with an aqueous In(NO3)3 solution; ii) pyrolysis of In@ZIF-67(Co) at 600 °C in nitrogen;
iii) calcination of the obtained product at 400 °C in flow of synthetic air. Upon IWI step, ZIF-
67(Co) crystals undergo agglomeration with visible sealing at the edges of the particles (Fig. S8).
14
ACS Paragon Plus Environment

Page 15 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
Thereafter, during the pyrolysis step, In@ZIF-67(Co) transforms into mixture of freestanding
spherical particles of about 12 nm size, which tend to sinter under the electron beam, and
agglomerations of smaller ones with a mean size of about 10 nm (Fig. S9). Qualitative analysis
performed on the XRD pattern indicates a mixture of nanosized a-Co and mixed-metal carbide,
Co3InC0.75 phases, where the carbide phase is dominant (Fig. S10). This observation supports the
effectiveness of IWI method and suggests high dispersion of the loaded In component. At the last
step of catalyst preparation, the pyrolyzed product has been calcined in a stream of air at the
temperature suitable for conversion to oxide form (i.e. 400 °C).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The resulting catalyst, having an In:Co molar ratio of 3:8 and referred to as 3In@8Co(300),
consists of 5-10 nm sized oxide nanoparticles (Fig. S11) developing a considerable high surface
area of 107 m2·g-1. Notably, the XRD pattern reveals broad diffraction lines attributed only to the
Co3O4 phase. This is also in accordance with the EXAFS spectroscopy applied at Co K-edge, which
reveals the characteristic fingerprint of the Co3O4 crystal structure and is further supported by
subsequent fitting of the EXAFS data (details in supplementary information, Table S3, Figs. 18A-
B). An interesting observation is the absence of any In-derived compounds in the PXRD pattern,
although the sample was calcined at 400 °C (Fig. S12). The state of indium in the 3In@8Co(300)
pre-catalyst was thus refined by EXAFS spectroscopy applied at the indium K-edge. The spectrum
of 3In@8Co(300) pre-catalyst is shown on Figs. 2(A)-(B) together with the spectrum of a
crystalline In2O3.
15
ACS Paragon Plus Environment

ACS Catalysis
Page 16 of 39
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2 (double column). (A) Magnitude of the Fourier transform for the EXAFS k3·c(k)
functions acquired at In K-edge for In2O3, Co3InC0.75, 3In@8Co(300) catalyst before and after CO2
hydrogenation reaction and, (B) their related back Fourier transforms (real part) applied within
their respective R-range for EXAFS fitting. Solid line represents experimental data whereas dotted
line corresponds to fit results conducted with [1.1–2.5] Å and [1.2–3.6] Å R-range together with a
[2.9–12.8] Å-1 and [3.8–10.4] Å-1 k-range, respectively for 3In@8Co(300) catalyst before and after
reaction. (C) ADF-STEM image and elemental mappings for 3In@8Co(300): (D) Co, (E) O, (F)
In maps and superimposed (G) Co/In maps and (H) Co/In/O maps.
16
ACS Paragon Plus Environment

Page 17 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
Qualitative assessment of the FT-EXAFS spectra for the In2O3 reference (without phase
correction), indicate a main peak at 1.7 Å attributed to In-O scattering and several peaks between
2.5 and 4.3 Å assigned mainly to In-In path with some minor contributions from In-O scattering
and some multiple scattering processes.58 In contrast, the FT-EXAFS of the 3In@8Co(300) pre-
catalyst show only one prominent peak centered around 1.6 Å which is similar to the low-R peak
of the In2O3 FT-EXAFS spectrum. The lack of contribution from more distant coordination shells
(e.g. In-In scattering path, Table S3) is associated with deficiency of the long-range order and
proves the amorphous nature of the indium component in the 3In@8Co(300) pre-catalyst.
Quantitative EXAFS fitting of the first shell with a single In-O scattering path provides a
coordination number of 6.3 ± 0.7 at a bonding distance of 2.14 ± 0.01Å, consistent with an
octahedral coordination (Table S3). The distribution of the indium oxide within the sample is
further highlighted by ADF-STEM imaging and EELS elemental mapping (Figs. 2(C)-(H): the
amorphous indium oxide covers homogeneously the 5 nm Co3O4 nanoparticles as a thin layer.
The catalytic performance of ZIF-67(Co)-derived indium containing catalysts in CO2
hydrogenation has been evaluated at 300 °C and 50 bar with a constant 20%CO2/80%H2 feed flow.
The kinetic profile of the reaction on 3In@8Co(300) catalyst (Fig. 3A) shows an induction period
of 20 h until it reaches 0.4 gMeOH·gcat-1·h-1 methanol productivity. As the reaction proceeds, the
methanol selectivity stabilizes at 63% with CO and CH4 by-product selectivities of 27% and 9%,
respectively (Fig. 3B, Fig. S29). Overall, the CO2 conversion is stable over 100 h on steady state
performance, yielding an STY of 0.5 gMeOH·gcat-1·h-1 for methanol.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17
ACS Paragon Plus Environment

ACS Catalysis
Page 18 of 39
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. (double column). (A) Kinetic profile of evolution MeOH space time yield with time and
(B) MeOH, CO and CH4 selectivities at time-on-stream (TOS) of 100 h over 3In@8Co(300)
(orange line), 3In@8Co(20) (navy line) and 3In@8Co(300)-dc (turquoise line) catalysts at XCO2 of
18%, 19% and 13%, respectively. (B) MeOH, CO and CH4 selectivities on stable plateau of the
kinetic profile (100 h) on studied catalysts. Reaction conditions: 80%H2-20%CO2 feed; T = 300
°C, P = 50 bar, mcat = 50 mg, GHSV = 15600 h-1. (C) ADF-STEM imaging and elemental mapping
for 3In@8Co(300) after the reaction: (from left to right) ADF-STEM image, Co, O, In maps and
superimposed Co/In maps and Co/In/O maps. (D) XRD patterns for fresh (top) and spent (bottom)
3In@8Co(300) (orange) and 3In@8Co(20) (navy) catalysts compared to identified phases: Co3O4
(■), Co3InC0.75 (*), a-Co (fcc,◆), c-In2O3 (cubic,●) and rh-In2O3 (rhombohedral,▼).
As previously reported by our group 34 and characterized below (vide infra), the induction period
is associated with the formation of the active phase by a solid reaction between the indium oxide
layer with its supporting cobaltite phase. Having a higher external surface area in the starting
18
ACS Paragon Plus Environment

Page 19 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
sacrificial ZIF-67(Co) is expected to enhance the contact between indium and cobalt oxides and
to increase substantially the formation rate of the active phase. To this end, another ZIF-67(Co)
support with smaller crystal size - ZIF-67(Co)-20nm – was used to improve the dispersion of the
In-component. The impregnation at the same In:Co molar ratio (3:8) followed by a two steps
thermal decomposition resulted in a material possessing 119 m2·g-1 of total surface area,
3In@8Co(20) (Table S4). The kinetic profile of reaction expressed in STYMeOH on 3In@8Co(20)
catalyst is shown on Figure 3A and compared to 3In@8Co(300) behavior at the same reaction
conditions. As anticipated, 3In@8Co(20) catalyst obtained from MOF with smaller particle size
shows considerably reduced induction period of 10 h versus 19 h (STYMeOH of 0.38 gMeOH·gcat-1·h-1)
observed in case of 3In@8Co(300). However, both catalysts demonstrate similar selectivities
towards MeOH (~63%) and by-products (CO, ~27% and CH4, ~9-11%) reaching stable STY
values of 0.52 gMeOH·gcat-1·h-1 for methanol production.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To gain some insights on the structural composition of 3In@8Co after the catalytic run, PXRD
analyses were performed for reacted catalysts which were expressly protected under inert
atmosphere in order to exclude any possible phase transformation caused by exposure to air.
Comparison of the diffraction patterns for fresh and spent catalysts along with the d-spacing
positions for identified crystalline phases is shown in Figure 3D. Evidently, both catalysts undergo
a phase transformation – from crystalline spinel with amorphous In2O3 to a complex mixture
consisting of mixed-metal carbide Co3InC0.75, a-Co (fcc form), cubic allotrope of In2O3 (c-In2O3),
traces of rhombohedral In2O3 phase (rh-In2O3) and both CoO polymorphs (Fig. S13). Besides the
evidences gained from PXRD, these structural transformations were confirmed by In and Co K-
edge EXAFS analyses performed ex-situ for the catalyst before and after catalytic process (Figs.
2(A)-(B), S17(A)-(B), Table S3). Interestingly, the hexagonally packed rh-In2O3 is a metastable
19
ACS Paragon Plus Environment

ACS Catalysis
Page 20 of 39
1
2
3
4
5
6
7
8
9
polymorphic form of indium oxide which irreversibly transforms, under elevated temperatures and
pressures of 0.1 MPa-3GPa, to the more stable cubic form.59-60 Additionally, it is known that in
strongly reductive atmospheres, such as CO and H2, the allotrope transformation proceeds
following the same pathway - rh-In2O3 → c-In2O3 meanwhile the reverse transformation happens
only at high energy input.61 Extrapolating the observation, one of the processes occurred during
the catalytic run is recrystallization of amorphous In2O3: firstly, to rh-In2O3 and then to more stable
c-In2O3. However, a minor part of rh-In2O3 is still present after catalysis. Simultaneously with the
formation of crystalline In2O3, the Co3O4 phase undergoes reductive transformation following the
sequence Co3O4 → CoO → Co which results in nanoparticles of a-Co (fcc) polymorph and traces
of CoO (hcp/fcc) phases (Fig. 3D, Fig. S13B). Finally, the third and more peculiar component
found to be formed is a mixed metal carbide, Co3InC0.75 (Fig. 3D). Since the reductive atmosphere
of reaction should drive the transformation of oxides to their corresponding metallic phases, the
occurrence of this carbide is attributed to the reductive transformation under carbon-rich
atmospheres better described as carburization process.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Imaging operated by ADF-STEM microscopy and performed on the 3In@8Co(300) material
after the catalytic run shows also that the catalyst undergoes textural changes along with phase
reorganization. The vast majority of the sample is a composite formed from aggregated
nanoparticles, all displaying a core-shell morphology. The average diameter of the core-shell
nanoparticle is about 10 nm with a shell thickness of 2 nm (Fig 3.C). Additional EELS elemental
mapping reveals that elemental composition of the shell includes cobalt, indium and oxygen atoms
which agrees with the formation of cobalt-indium oxides.34 No lattice fringes could be observed in
the shell which means that the latter cobalt-indium oxides are also amorphous. On the other hand,
the core is clearly polycrystalline as evidenced from the various orientation of lattice fringes
20
ACS Paragon Plus Environment

Page 21 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
imaged by high-resolution HAADF-STEM (Fig. 4A-C). The high-resolution image of the lattice
(Fig. 4C-D) is reproduced by a multi-slice STEM image simulation of a Co3InC0.75 slab viewed
along the <110> zone axis (Fig. 4F); the alternance of bright atom columns containing indium and
cobalt atoms with fainter atom columns containing only cobalt atoms is especially well reflected.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. (double column). (A) High-resolution HAADF-STEM image of 3In@8Co(300) after
reaction. (B) Fourier transform (d* = 2.6 nm-1, d = 3.83 Å) of (C) zoomed region (blue dots) of
interest in (A) representing the atomic arrangement of the Co3InC0.75 phase, viewed along <110>
zone axis. (D) Zoomed region in (C, yellow dots) processed with a Wiener and Gaussian band-
pass filters (E), and (F) corresponding multi-slice STEM image simulation of a Co3InC0.75 slab.
(G) Comparison of experiment and theory (FDMNES calculations) for the carbon K-edge
spectrum of Co3InC0.75 recorded by EELS spectroscopy.
21
ACS Paragon Plus Environment

ACS Catalysis
Page 22 of 39
1
2
3
4
5
6
7
8
9
With high-angle annular dark field (HAADF) detection, electrons arising from Rutherford
scattering are collected, meanwhile electrons deviated by coherent elastic scattering are mostly
excluded (removing the phase contrast). The intensity I in the resulting images is then, in a first
approximation, given by I a t ∙ Za (a = 1.5 - 2) with a thickness t and an average atomic number
Z.62 As a direct consequence, the difference between indium and cobalt atoms is indeed observable
but the carbon atoms located in-between the columns of cobalt atoms are not visualized easily. To
definitely support the formation of a mixed carbide phase in the particle core, the carbon K-edge
was recorded by EELS spectroscopy (Fig. 4G). The C K-edge originates principally from dipolar
transitions from the 1s state to the unoccupied 2p state. Unlike X-ray Raman scattering
spectroscopy, EELS spectroscopy is highly forward scattered: with a microscope tension of 300
kV, 90% of the intensity is collected within ≈ 2.6 mrad scattering angle corresponding to a small
momentum transfer q equal to 0.9 Å-1.63 With those experimental conditions, the monopole
transitions (s→s) are thus neglected.64 Self-consistent field simulations for Co3InC0.75 clusters with
radii of up to 10 Å were performed to ensure the convergence of the calculation. Using the full-
potential FDM approach, the three main dipolar transitions observed in the experimental spectra
are relatively well reproduced. The agreement is not perfect, especially for the energy position of
the third transition, but nevertheless the main features are there.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To summarize, the phase composition of the spent catalyst is dependent of several processes
taking place during CO2 hydrogenation: i) the principal process is the formation of Co3InC0.75
constituting the core of catalytically active nanoparticles surrounded by a Co-In oxides shell, ii)
crystallization of In2O3 from amorphous phase as separated nanoparticles, iii) reduction of Co3O4
to metallic Co.
22
ACS Paragon Plus Environment

Page 23 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
In order to address the question concerning the relevance of two-step thermal treatment, i.e.
sequential pyrolysis-calcination process, the catalytic behavior of 3In@8Co(300)-dc sample
prepared through direct calcination of In-impregnated ZIF-67(Co)-300 at 400 °C without the
intermediate pyrolytic step has been evaluated. Kinetic profile of 3In@8Co(300)-dc performance
in CO2 hydrogenation towards methanol (Fig. 3A) reveals considerably slower induction period
that takes approx. 55 h to reach a plateau of 0.28 gMeOH·gcat-1·h-1 STY for the methanol product.
Additionally, the catalyst prepared by direct calcination route concedes 3In@8Co(300) material in
methanol productivity (0.28 gMeOH·gcat-1·h-1 vs 0.52 gMeOH·gcat-1·h-1) albeit with similar SMeOH and
higher SCO (Fig. 3A-B). Given the poorer catalytic performance of 3In@8Co(300)-dc, one may
assume the significance of intermediate pyrolysis step for stabilization of indium and cobalt
domains in close vicinity to each other, evidently in form of cobalt indium carbide phase,
Co3InC0.75. We presume that such mixed metal carbide redounds the homogeneity of In@Co
system after calcination rendering more effective distribution of Co3O4 and In2O3 oxides within the
composite catalyst which, consequently, diminishes the induction period for hydrogenation
process.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Taking into account the above-mentioned findings and considering the higher performance of
catalyst derived from ZIF-67(Co) sacrificial support with 20 nm crystal size, a set of experiments
was carried out in order to study the effect of indium loading. Thereby, ZIF-67(Co)-20 nm has
been impregnated with different amounts of In(NO3)3 in accordance with Table S1 following the
same IWI procedure as before and for simplicity denoted as xIn@yCo(20) representing the
nominal In:Co molar ratios (x:y) in each sample. As it can be observed from nitrogen absorption
measurements (Fig. S22), the shape of isotherms of In@ZIF-67(Co)-20 nm materials changes from
type IV- to II- and show the total (BET) surface area descending as the indium content increases
23
ACS Paragon Plus Environment

ACS Catalysis
Page 24 of 39
1
2
3
4
5
6
7
8
(Table S2, S4). Although indium concentrations were increased gradually, the micropore area
deviates from a linear trend; this indicates that the indium species settle not only on external
surfaces of ZIF-67(Co) but also occupy the micropore area of the framework.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Afterwards, In@ZIF-67(Co)-20 nm materials with different amounts of loaded indium
underwent the sequential pyrolysis-calcination treatment in order to render In@Co composite
catalysts. The resulting oxide-containing materials exhibit total surface area ranging from 119 to
87 m2·g-1 (Table S4) and show II-type adsorption isotherms typical for macroporous adsorbents
(Fig. S23). Notably, the total surface area decreases as the indium content increases; that
successfully agrees with the findings, above discussed, concerning the embedment of Co3O4
nanoparticles possessing the large interparticle surface which is occupied by amorphous In2O3
phase. The content of loaded indium in MOF-derived catalysts was confirmed by XRF
measurements and calculated from TG curves considering the temperature region of 870-930 °C,
in which Co3O4 → CoO transition is taking place (Figs. S24, S25), in order to compare to the
nominal In:Co molar ratios (Table S4). The results summarized in Table S4 show a fair agreement
with the expected In:Co ratios aimed to be attained in the xIn@yCo(20) catalysts.
A series of MOF-derived catalysts aIn@bCo(20) with variable In:Co molar ratios was tested in
direct CO2 hydrogenation to methanol under standard conditions (80%H2-20%CO2 feed; P = 50
bar and GHSV = 15600 h-1, Fig. S30). Figure 5A shows kinetic profiles of STY evolution for
MeOH product on the catalysts with variable In:Co molar ratios. As can be observed, the
aIn@bCo(20) catalysts with lower In content, i.e. 1In@23Co(20) and 1In@12Co(20), exhibit the
-
largest induction periods of 40 and 90 h and a low MeOH productivity of 0.01 and 0.04 gMeOH·gcat
1·h-1, respectively. Further increase of indium content up to 1In:6Co value reduces the induction
24
ACS Paragon Plus Environment

Page 25 of 39
ACS Catalysis
1
2
3
4
5
6
period down to 30 h yet reaching maximum STYMeOH of 0.21 gMeOH·gcat-1·h-1 over a short period of
time (5 h) after which the catalyst slowly deactivates.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. (double column). (A) Kinetic profile of evolution MeOH space time yield vs time and
(B) selectivities to MeOH product, CO and CH4 by-products at time-on-stream (TOS) of 100 h
over catalysts with different In loadings: 3In@4Co(20) (XCO2 = 20.5%, red), 3In@8Co(20) (XCO2
=
19.2%, orange), 1In@6Co(20) (XCO2 = 13.9%, blue), 1In@12Co(20) (XCO2 = 11.1%, purple) and
1In@23Co(20) (XCO2 = 10.1%, green). Reaction conditions: 80%H2-20%CO2 feed; T = 300 °C, P
= 50 bar, mcat = 50 mg, GHSV = 15600 h-1. XRD patterns for (C) fresh and (D) spent catalysts and
comparison with the identified phases: Co3O4 (■), CoO (｜), e-Co (hcp,★), Co3InC0.75 (*), a-Co
(fcc,◆), c-In2O3 (cubic,●) and rh-In2O3 (rhombohedral,▼). ADF-STEM imaging (E) and
25
ACS Paragon Plus Environment

ACS Catalysis
Page 26 of 39
1
2
3
4
elemental mappings for 3In@4Co(20) after reaction: Co map (F) and O map (H); superimposed
5
6
Co/In maps (G) and Co/In/O maps (I).
7
8
9
In a contrast, 3In@8Co(20) and 3In@4Co(20) catalysts with higher In2O3 loading show
considerably reduced induction time (19 h and 7 h vs 30 h) and outstanding stable performance
over 100 h of reaction with methanol productivity of ca. 0.5 and 0.65 gMeOH·gcat-1·h-1, respectively.
Therefore, the activity towards methanol product on MOF-derived mixed oxide catalysts is
enhancing in order 3In@4Co > 3In@8Co > 1In@6Co > 1In@12Co > 1In@23Co. Regarding the
MeOH selectivity data compared at 100 h of time-on-stream (TOS) shown on Figure 5B, the
methanol selectivity develops following the same trend as STYMeOH whereas the selectivities
toward by-products (CO and CH4) evolve in opposite manner. However, a higher MeOH
selectivity (62%) is reached by 3In@8Co(20) and 3In@4Co(20) materials along with reduced
selectivities toward CO (27-28%) and CH4 (9-11%) by-products. In order to understand the
differences in reactivity, structural analysis was carried out on the experimental XRD data for fresh
and reacted catalysts with different In loadings (Fig. 5C, D). Although the aIn@bCo(20) catalysts
before reaction showed identical XRD pattern with diffraction lines associated solely to the Co3O4
phase (Fig. 5C), the composition of reacted solids enriches by diversity of phases, depending on
the indium content (Fig. 5C, Table S5). Consequently, the difference in reactivities of the MOF-
derived catalysts also relies on the compositional characteristics of the spent phases manifested
during the reaction. As shown in Figure 5D, the spent 1In@23Co(20) catalyst mainly contains fcc-
Co phase and traces of CoO and could be associated with low MeOH selectivity (12 %) and
productivity (0.02 gMeOH·gcat-1·h-1) in contrast to high CO selectivity (70 %). The appearance of
Co3InC0.75 phase in mixture with fcc-Co in 1In@12Co(20) improves the selectivity towards MeOH
twofold (SMeOH = 23%, SCO = 59%, SCH4 = 18%), whereas the dominance of the mixed-metal carbide
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
26
ACS Paragon Plus Environment

Page 27 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
over a-Co in 1In@6Co(20) leads to maximum SMeOH of 40% (SCO = 41%, SCH4 = 19%), after which
the activity starts to decline. Finally, the mixture of Co3InC0.75 and In2O3 (both rh- and c-
polymorphs) more likely reflects the true composition responsible for selective MeOH synthesis
inasmuch as both actively performing 3In@4Co(20) and 3In@8Co(20) catalysts possess similar
composition based on these two components (SMeOH = 62 %, SCO = 27-28%, SCH4 = 9-11% and
STYMeOH = 0.5–0.65 gMeOH·gcat-1·h-1). To visualize relations between the phase contents found by
PXRD studies, a correlation matrix has been added (Fig. S31). The correlation coefficients
between individual phase contents and selectivities towards the catalytic products were scaled
from yellow (strong positive) to navy (strong negative correlation). It demonstrates strong
correlation between the carbide and indium oxides content and methanol selectivity whereas cobalt
and cobalt oxide presence negatively correlated with it. Although carbide content has the strongest
correlation with methanol production, its catalytic activity was excluded by performing an
individual catalytic study (Table S7) and appeared to be a side product of catalysts workout and
carburization (Figs. S35-S38). To support the above-mentioned results, CHN elemental analyses
on the fresh and spent catalysts were carried out and compared in Table S6. As expected, the
carbon content in reacted catalysts increases compared to the fresh ones and can be associated with
the formation of Co3InC0.75 phase; that agrees well with PXRD quantification data (Table S5).
Notably, spent 3In@4Co(20) and 3In@8Co(20) catalysts differ by relative amount of crystalline
In2O3 and Co3InC0.75 phases (Table S5) that might have influence on the induction period
considering previous finding that both components are formed along the sequentially-conjugated
process. Similarly to 3In@8Co(300), the spent version of the best performing 3In@4Co(20)
catalyst shows also core-shell morphological features as suggested by EELS elemental maps (Figs.
5E-I, Figs. S26-S27). In conclusion to all above mentioned, the optimal catalyst composition is
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
27
ACS Paragon Plus Environment

ACS Catalysis
Page 28 of 39
1
2
3
4
5
6
determined by relative molar ratio of 3In:8Co and 3In:4Co possessing enough indium to be
quantitatively converted to Co3InC0.75 phase.
7
8
9
The most active catalysts, 3In@4Co(20) and 3In@8Co(20), were chosen to study the effect of
the reaction temperature with the aim of finding optimal conditions for CO2 to methanol
hydrogenation process. The kinetic profiles for MeOH productivity on 3In@4Co(20) and
3In@8Co(20) catalysts at three temperatures – 270 °C, 285 °C and 300 °C, are shown in Figure
6A. As can be observed, the induction period for both catalysts increases from 7-19 h to 60-70 h
while the reaction temperature decreases from 300 °C to 270°C. This observation can be
rationalized by the fact that both Co3InC0.75 and In2O3 phases formed through conjugated red-ox
process become kinetically unfavorable at lower temperatures. Despite the discrepancy of initial
reaction rates for both catalysts, the MeOH productivity is deviating in a range of 0.5–0.65
gMeOH·gcat-1·h-1 regardless the applied temperature. However, the alternate comparison of
selectivities for MeOH product and CO/CH4 by-products in the steady-state region (Fig. 6B, Fig.
S32) of the catalytic process shows that lower temperature induces high methanol selectivities
(SMeOH ~ 79-80% at 270 °C vs 62 % at 300 °C) yet suppresses side processes (SCO ~ 15 % and SCH4
~ 5 % at 270 °C vs SCO ~ 27-28 % and SCH4 ~ 9-11 % at 300 °C), i.e. methanation and reverse
water-gas shift (RWGS) reactions.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28
ACS Paragon Plus Environment

Page 29 of 39
ACS Catalysis
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. (single column). (A) Kinetic profile of evolution MeOH space time yield vs time and
(B) selectivities to MeOH product, CO and CH4 by-products at time-on-stream (TOS) of 100 h
over catalysts 3In@4Co(20) (red) and 3In@8Co(20) (orange) at different temperatures: 270 ºC
((XCO2 = 14.2% and 11.6%, □), 285 ºC (XCO2 = 17.3% and 15.5%, ○) and 300 ºC (XCO2 = 20.5% and
19.2%, ▽). Reaction conditions: 80%H2-20%CO2 feed; P = 50 bar, mcat = 50 mg, GHSV = 15600
h-1.
Thus, an optimal temperature needs to be considered in order to inhibit undesired side reactions
and favor higher methanol selectivity and productivity yield.
29
ACS Paragon Plus Environment