Selectivity controlled transformation of carbon dioxide into a versatile bi-functional multi-carbon oxygenate using a physically mixed ruthenium-iridium catalyst

Article abstract of DOI:10.1039/d1cy00149c

To mitigate environmental concern and energy issues, the conversion of carbon dioxide (CO₂) to valuable and useful carbon chemicals offers a promising strategy for the development of a carbon neutral economy. The utilization of inert CO₂as a building block in the synthesis of multi-carbon (>2) oxygenated compounds, specifically propylene glycol methyl ether (PGME; C₄H₁₀O₂), produced annually at a multi-million-ton scale from petroleum-based propylene oxide is of particular interest because of its multifaceted industrial and commercial applications. Herein, we present a simple and straightforward system that uses CO₂in compressed form as the C1 feedstock for the synthesis of PGME by direct hydrogenation and subsequent C-C coupling without any other sacrificial reagents. In addition, we combine experimental results with DFT calculations to elucidate the synergistic contributions of two catalytic metals (Ru and Ir) to activation of CO₂for hydrogenation and consequent mediation of the C-C bond formation leading to the generation of PGME. Taken together, our findings suggest that the strategy presented herewith may serve as a starting point for the development of a sustainable chemical synthesis platform for multi-carbon oxygenates by utilizing CO₂as the starting material.

Full text of DOI:10.1039/d1cy00149c

Volume 11
Number 14
21 July 2021
Pages 4661–5016
Catalysis
Science &
Technology
rsc.li/catalysis
ISSN 2044-4761
PAPER
Maya Chatterjee, Hajime Kawanami et al.
Selectivity controlled transformation of carbon dioxide into
a versatile bi-functional multi-carbon oxygenate using a
physically mixed ruthenium–iridium catalyst

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Selectivity controlled transformation of carbon
dioxide into a versatile bi-functional multi-carbon
oxygenate using a physically mixed ruthenium–
iridium catalyst†
Cite this: Catal. Sci. Technol., 2021,
11, 4719
Maya Chatterjee, * Abhijit Chatterjee, Mitsunori Kitta and Hajime Kawanami *be
a
c
d
To mitigate environmental concern and energy issues, the conversion of carbon dioxide (CO
and useful carbon chemicals offers a promising strategy for the development of a carbon neutral economy.
The utilization of inert CO as a building block in the synthesis of multi-carbon (>2) oxygenated
compounds, specifically propylene glycol methyl ether (PGME; C ), produced annually at a multi-
2
) to valuable
2
4 10 2
H O
million-ton scale from petroleum-based propylene oxide is of particular interest because of its multifaceted
industrial and commercial applications. Herein, we present a simple and straightforward system that uses
2
CO in compressed form as the C1 feedstock for the synthesis of PGME by direct hydrogenation and
subsequent C–C coupling without any other sacrificial reagents. In addition, we combine experimental
Received 27th January 2021,
Accepted 28th April 2021
results with DFT calculations to elucidate the synergistic contributions of two catalytic metals (Ru and Ir) to
2
activation of CO for hydrogenation and consequent mediation of the C–C bond formation leading to the
generation of PGME. Taken together, our findings suggest that the strategy presented herewith may serve
as a starting point for the development of a sustainable chemical synthesis platform for multi-carbon
DOI: 10.1039/d1cy00149c
rsc.li/catalysis
2
oxygenates by utilizing CO as the starting material.
compounds starting from CO , research efforts are mainly
2
Introduction
concentrated on the formation of C1 (formic acid, formate,
2–12
The development of an effective method to produce
petroleum-derived compounds using a renewable feedstock is
emerging as an increasingly important strategy for the
methanol) and C2 (ethanol)
products. Although, multi-
carbon oxygenates (>C ) are more valuable and display
2
excellent compatibility with gasoline as the blending
1
3
synthesis of value-added chemicals. In this context, CO is an
fraction, the catalytic conversion of CO2 to multi-carbon
2
attractive renewable C1 feedstock because of its abundance,
oxygenates (>C ) is still in its infancy due to the involvement
2
nontoxicity
and
nonflammability.
However,
the
of a complex network of hydrogenation and C–C coupling
under reducing conditions in a controlled manner.
Traditionally, such studies have relied on electro-reduction or
high energy reagents (organometallic and strong organic bases)
by utilizing additional sacrificial carbon sources, hence
producing mixtures of products or forming by-products,
1
4,15
thermodynamically stable and kinetically inert nature of CO2
has been a major hurdle to the transformation of this C1
feedstock, which has required highly energetic reactants,
severe reaction conditions (200 °C or above) and complex
1–12
designer catalysts.
If we focus on the oxygenated
9–12,16,17
limiting their subsequent applicability.
remains persistent need for the development of
straightforward methodologies to selectively synthesize multi-
carbon oxygenates (>C ) that are of industrial and commercial
As such, there
a
Microflow Chemistry Group, Research Institute for Chemical Process Technology,
AIST Tohoku, 4-2-1, Nigatake, Miyagino-ku, Sendai, 983-8551, Japan.
E-mail: c-maya@aist.go.jp; Fax: +81 22 237 5388; Tel: +81 22 237 5213
a
b
2
Interdisciplinary Research Center for Catalytic Chemistry, AIST, Central 5, 1-1-1
interest, utilizing CO as a starting reactant material.
2
Higashi Tsukuba, Ibaraki, 305-8565, Japan. E-mail: h-kawanami@aist.go.jp
c
An example of an industrially relevant and versatile multi-
carbon oxygenate is propylene glycol methyl ether (PGME),
with a global market value of >$1500 million (USD) in 2017,
which is expected to grow further to >$2000 million (USD) by
2026. Some of the established uses of PGME include: fuel
additives to improve the distillation curve by reducing the
Dassault Systemes K.K., BIOVIA, Think Park Tower, 2-1-1 Osaki Shinagawa-ku,
Tokyo, 141-6020, Japan
d
Research Institute of Electrochemical Energy, Department of Energy and
Environment, National Institute of Advanced Industrial Science and Technology, 1-
8
-31 Midori-gaoka, Ikeda, Osaka 563-8703, Japan
e
CREST, Japan Science and Technology (JST), 4-1-8, Honcho, Kawaguchi, Saitama,
32-0012, Japan
3
1
8
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/d1cy00149c
emission of particulate matter and NOx,
brake fluid
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4719

View Article Online
Paper
Catalysis Science & Technology
components to improve the responsiveness of brakes, and
paint, varnish and lacquer components, as well as an
intermediate for PGME–acetate (which is a photoresist
mixed by different mixing processes such as mechanical,
chemical and physical mixing (details are in the Materials
and methods section). Structural characterization of
individual Ru/MCM-41 and Ir/MCM-41 using XRD confirmed
the retention of the mesoporous structure even after the
incorporation of metals (ESI;† Fig. S1). Furthermore, the
mesoporous structure was also preserved in the catalysts
obtained from various mixing processes (mechanical,
chemical and physical) (ESI;† Fig. S2a–c). Focusing, on the
physically mixed Ru/MCM-41 and Ir/MCM-41 (Ru–Ir-PM)
catalysts, the existence of a mesoporous moiety as well as
Ru and Ir was also evident even from the XRD patterns of
spent catalysts (ESI;† Fig. S2d and S2 inset). In addition, N2
adsorption–desorption also reflects the typical mesoporosity
of the studied Ru–Ir-PM catalysts (ESI†).
1
9
solvent in the semiconductor industry). Traditionally, the
industrial production of PGME involves propylene oxide (a
petroleum-based highly toxic and hazardous compound) and
alcohol; however, the epoxide ring of propylene oxide may
open through either of its C–O bonds, resulting in low
2
0
activity and PGME selectivity. Although, the application of
acidic or basic catalysts has improved the selectivity towards
PGME, the process still suffers from several drawbacks
including the necessity of liquid waste treatment as well as
2
1
corrosion problems. As a result, developing an alternative
sustainable process that utilizes CO as the starting material
2
is particularly prudent at this time.
Turning CO2 into a feedstock requires large amounts of
concentrated CO2 with high purity. Nevertheless, direct
recovery of the gas from the atmosphere (e.g., atmosphere; 0.03
vol%) or exhaust gases from transportation is unfavourable,
has high energy penalties, and possesses economic constraints
or sometimes is infeasible because of very low concentrations.
Morphological analysis of individual Ru/MCM-41 and Ir/
MCM-41 is presented in Fig. S3 (ESI†). The high-angle
annular
dark-field
scanning
transmission
electron
microscopy (HAADF-STEM) image shows two distinct
contrasts. The high contrast part represents the presence of
heavy particles inside the silica channels (existence of
smaller particles) and the larger bright spots in the darker
background represent the amorphous silica matrix (larger
particles). An average particle size of 5.1 and 4.8 nm was
determined for Ru/MCM-41 and Ir/MCM-41, respectively.
The representative HAADF-STEM images of mechanically
and chemically mixed catalysts are also shown in Fig. S4
and S5 (ESI†), respectively. For the mechanically mixed
catalysts, an irregular distribution of metal particles with a
mean diameter of 9.5 nm was observed (ESI;† Fig. S4a).
Elemental mapping and the corresponding EDS analysis
(ESI;† Fig. S4b–f) revealed the isolated existence of both
metals in a single location. In contrast, the chemically
mixed catalyst displayed a uniform distribution of bimetallic
nanoparticles of an average size of 9.8 nm and featured Ru
and Ir adjacent to each other in a selected region (Fig. S5a–
f†). Furthermore, Fig. 1 presents a montage of HAADF-
STEM images, electron mapping images and corresponding
EDS spectra of Ru–Ir-PM. A representative HAADF-STEM
image of the fresh catalyst shows dispersion of Ru and Ir
throughout the surface depending on locations. Both of the
metals largely preserved their crystallinity (Fig. 1a inset) and
maintained their individuality even after mixing, which can
be best illustrated by EDS of the tested regions (Fig. 1b).
The spent catalyst also maintained the dispersion and
individuality of both metals (Fig. 1c); however, an
overlapping signal originates in the selected location
designated as region 2 (Fig. 1d and e), directed towards the
possibility of single particle formation. High resolution
HAADF-STEM images of region 2 (Fig. 1f and g) labelled
with squares (green) (Fig. 1g) show the coexistence of both
Ru and Ir, which was again confirmed from the EDS profile
Alternative opportunities exist to capture CO in a concentrated
2
form, for example from the exhaust gas of thermal power
plants, oil refineries, and industrial chemical plants before it
2
2
enters the atmosphere. Supercritical CO2 (scCO ) is more
2
reliable in terms of its availability in pure and concentrated
forms, with tuneable hybrid properties of either gaseous or
liquid state, which allows the formation of products with
higher yields under comparatively milder conditions; thus,
catalytic deactivation and destruction of the reactor system can
be prevented. Nonetheless, even after the exemplary work of
2
3
24,25
2
Jessop et al., scCO was rarely explored as a reactant.
Herein, we have designed an approach to produce PGME
multi-carbon oxygenate) utilizing scCO2 as a reactant. To
(
2
develop a strategy for utilizing scCO as a reactant, laboratory
scale verification and validation of predicted performance is
critical for securing access to the CO economy. To this end, we
2
utilized scCO
(
(
2
and hydrogen as the reactants, and ruthenium
Ru) and iridium (Ir) as the catalysts because both of the metals
Ru and Ir) were highly efficient towards hydrogenation and
2
3,26,27
C–C coupling reactions.
Furthermore, mesoporous silica
(MCM-41), which possesses a large surface area that increases
the effectiveness of the active atoms (surface area is
proportional to utilization efficiency), was used as a support
material for Ru and Ir. Having developed this system, we began
by optimizing the catalyst and the reaction conditions. Finally,
we integrated our experimental findings with theoretical
calculations to validate the reaction pathways. Altogether, we
were able to successfully synthesize PGME with high-selectivity
using a simple and efficient strategy.
Results and discussion
Characterization of catalysts
(
Fig. 1h). The formation of single particles might be due to
the segregation of Ru and Ir that occurred during the
reaction. No alloy particles were detected in any of the
cases.
MCM-41 supported Ru and Ir catalysts were prepared by a
hydrothermal method and then both of the catalysts were
4720 | Catal. Sci. Technol., 2021, 11, 4719–4731
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
Fig. 1 HAADF-STEM images, electron mapping and the corresponding EDS spectra of Ru–Ir-PM. Before reaction: (a) HAADF-STEM images and
elemental mapping of two selected regions of Ru–Ir-PM and (b) the corresponding EDS. After reaction: (c) HAADF-STEM images and elemental
mapping; (d) identification of single particles; (e) EDS of selected regions; (f and g) high resolution HAADF-STEM images and elemental mapping of
single particles; (h) EDS analysis. Elemental mapping of Ru (red), Ir (turquoise) and Si (green).
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4721

View Article Online
Paper
Catalysis Science & Technology
−
1
−1
−1 −1
−1
Catalytic performance of single-metal Ru/MCM-41 and Ir/
MCM-41 catalysts
h ; 79 h ) > PGME (0.3 mmol g
h ; 34 h ) > methanol
cat
−
1
−1
−1
−1
(0.2 mmol g
h ; 22 h ) > 2-propanol (0.02 mmol g
cat
cat
−
1
−1
h ; 2 h ), whereas, on the Ir catalyst, the efficiency order
Our initial experimental approach was to assess the viability
of the single-metal Ru/MCM-41 and Ir/MCM-41 catalysts
individually to design a process for synthesizing multi-carbon
oxygenated compounds using scCO2 as a reactant with
selectivity control. We found that both of the individual
−
1
−1
−1
was: 2-propanol (1.16 mmol g
h ; 130 h ) > PGME (0.14
cat
−
1
−1
−1
−1 −1
mmol g
h ; 15 h ) > ethanol (0.12 mmol g
h ; 13
cat
cat
−
1
h ) (Fig. 2b). All the results reflected the astounding ability
of both the catalysts to generate multi-carbon oxygenated
products through hydrogenation and C–C coupling. Analyses
of the resultant gaseous products confirmed the formation of
methane on both of the single-metal catalysts (Ru = 0.001
mol%; Ir = 0.004 mol%) but CO was below the detection limit
catalysts successfully converted scCO
2
into methanol,
ethanol, 2-propanol and PGME (Fig. 2a and b). All of the
products were confirmed from the analysis of the
condensable liquid phase after reaching steady state
(
<6 ppm).
(
definition in the Materials and methods section). The
product selectivity on Ru/MCM-41 [ethanol (65.5%) > PGME
22.0%) > methanol (11.0%) > 2-propanol (1.5%)] varied
(
Mixture of Ru/MCM-41 and Ir/MCM-41 catalysts
considerably from that on Ir/MCM-41 [2-propanol (92.9%) >
PGME (5.9%) > methanol (0.8%) > ethanol (0.4%)] under the
optimized set of screened conditions (total pressure = 16
MPa, PH2/PCO2 = 0.33, reaction time = 16 h) (Fig. 2a). To better
understand the catalytic efficiency, the performance of each
catalyst was evaluated in terms of molar productivity
The implementation of the individual single-metal Ru/MCM-
41 and Ir/MCM-41 catalysts resulted in a mixture of
compounds rather than a single targeted product; thus, we
applied a physical mixture of the Ru/MCM-41 and Ir/MCM-41
(
Ru–Ir-PM) as a catalyst for the conversion of scCO . It is
2
−
1
−1
−1
important to note that a physical mixture of solid catalysts
can have a significant impact on the product distribution
compared to their individual counterparts due to the additive
(
mmol gcat h ) and turnover frequency (TOF; h ). On the
−
1
Ru catalyst, the efficiency order was: ethanol (0.7 mmol gcat
2
8,29
effect (i.e., synergy between the two metals).
Fig. 3a shows
Fig. 2 Catalytic performance of the individual Ru/MCM-41 and Ir/
MCM-41catalysts for conversion of CO (a) Selectivity and (b)
productivity for each product formed. Reaction conditions: catalysts =
.1 g, temperature = 150 °C, reaction time = 16 h and total pressure =
6 MPa (P 2 = 4 MPa and the rest is CO ). Each experimental result is
the average value from three measurements with a reproducibility of
10%.
2
.
Fig. 3 Comparison of (a) product distribution and (b) productivity of
each product on Ru–Ir-PM as a function of mixing ratio. Reaction
conditions: Ru–Ir-PM = 0.1 g, temperature = 150 °C, reaction time = 16
0
1
H
2
2
h and total pressure = 16 MPa (PH2 = 4 MPa and the rest is CO ). Each
experimental result is the average value from three measurements with
a reproducibility of ±10%.
±
4722 | Catal. Sci. Technol., 2021, 11, 4719–4731
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
−
1
−1 −1
the catalytic performance of Ru–Ir-PM as a function of
composition. Starting from a ratio of 10% Ru/MCM-41: 90%
135 h ) compared to that of Ru–Ir-PM (3.1 mmol g
h ;
cat
−
1
348 h ). Taken together, the results suggested that the
mixing process had a significant impact on the catalytic
activity. Despite the same compositions, morphological
differences (distribution pattern of Ru and Ir) arise due to
the variation of the mixing processes (as mentioned in the
Characterization of catalysts section) and have a substantial
influence on particle size, shape, interaction with support,
and adsorption–desorption behavior of reactant and
Ir/MCM-41 (weight basis), PGME was formed with
a
selectivity of 30.3%, which reached a maximum of >99% at a
catalytic composition of 80% Ru/MCM-41: 20% Ir/MCM-41.
The formation of PGME as a single product was confirmed
1
by GCMS (ESI;† Fig. S6a and b), and H (ESI;† Fig. S7) and
1
3
C NMR (ESI;† Fig. S8) analysis. In addition, the catalytic
efficiency (molar productivity and TOF) towards PGME
formation was also maximized at a catalyst composition of
3
1
products, which might govern the catalytic efficiency and
selectivity of the reactions.
8
0% Ru/MCM-41: 20% Ir/MCM-41, with a value of 3.1
−
1
−1
−1
−1 −1
mmol gcat
h
(348 h ) as compared to 0.2 mmol gcat
h
−
1
−1 −1
−1
(22 h ) and 0.14 mmol g
h
(15 h ) for 100% Ru/MCM-
cat
Optimization of reaction conditions
4
1 and 100% Ir/MCM-41, respectively (Fig. 3b). These results
emphasized (i) the beneficial synergistic effects between the
two metals and (ii) the necessity of controlling the catalytic
composition efficiently to obtain PGME as the single
product.
Based on the maximum catalytic efficiency, we continued our
studies with the Ru–Ir-PM catalyst of the composition that
maximized PGME selectivity (i.e., 80% Ru/MCM-41: 20% Ir/
MCM-41) for optimization of different reaction parameters
(e.g., total pressure, PH2/PCO2 ratio and reaction time) related
to catalytic performances. Our target was to develop a system
under mild reaction conditions so we have chosen a
temperature of 150 °C for parametric optimization required
to obtain measurable catalytic activity for the studied
reaction.
Different processes of mixing. For solid catalysts, the
mixing process can greatly influence the catalytic outcome
due to possible changes in the concentration of metal
components or arrangements of the active constituents
3
0
(
ensemble effects). As such, after optimizing the ratio of the
individual metal components, we also tested the catalysts
prepared via mechanical (ball-milling) and chemical (bi-
metallic) mixing (Fig. 4). Under the optimized reaction
conditions (80% Ru/MCM-41: 20% Ir/MCM-41; total pressure
Variation of the total reactant pressure. First, the effect of
total reactant pressure on the product distribution was
investigated at a fixed temperature of 150 °C (Fig. 5a). PGME
remained the predominant product throughout the entire
=
16 MPa, PH2/PCO2 = 0.33), the mechanically mixed catalyst
produced methanol, ethanol and PGME with a selectivity of
4.4%, 14.4% and 51.2%, respectively (Fig. 4). The catalytic
efficiency (molar productivity and TOF) to produce PGME
3
−
1
−1
−1
(
0.05 mmol gcat h ; 5.6 h ) was strikingly low compared to
−
1
−1
−1
that of Ru–Ir-PM (3.1 mmol gcat h ; 348 h ). It is likely
that the implementation of mechanical forces modulated the
electronic and geometric structure of metal catalyst. In
comparison, the chemically mixed catalyst produced
2
-propanol and PGME with a selectivity of 17.6% and 82.4%,
respectively. Although, the PGME selectivity (82.4%) was
−
1
−1
relatively high, the productivity was low (1.2 mmol gcat h ;
Fig. 4 Effect of the mixing process on product selectivity and PGME
productivity. Physical (Ru–Ir-PM), chemical (bimetallic) and mechanical
Fig. 5 Evaluation of the effect of total pressure on (a) product
distribution; (b) PGME productivity; corresponding phases of the
reactant are also shown on the right side of the graph. Reaction
conditions: Ru–Ir-PM = 0.1 g, temperature = 150 °C, reaction time = 16
h. Each experimental result is the average value from three
measurements with a reproducibility of ±10%.
(
°
ball-milling). Reaction conditions: catalysts = 0.1 g, temperature = 150
C, reaction time = 16 h and total pressure = 16 MPa (PH2 = 4 MPa and
the rest is CO ). Each experimental result is the average value from
2
three measurements with a reproducibility of ±10%.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4723

View Article Online
Paper
Catalysis Science & Technology
pressure range studied (8–20 MPa). The selectivity to PGME
was changed from 88.1% (8 MPa) to >99% (16 MPa), and
then decreased to 73.7% at 20 MPa due to the unexpected
formation of ethanol. Alteration of pressure invoked a strong
influence on PGME productivity which increased from 0.46
performances, varying from 0.041 to 0.6. Fig. 6a shows PGME
as the only detectable product throughout the entire range
studied. However, we noticed a significant role of PH2/PCO2
ratio in PGME productivity, which switched from 0.26
−
1
−1
−1
mmol gcat
h ; 29.2
h
(PH2/PCO2
= 0.041) to 0.03
−
1
−1
−1
−1 −1
−1
−1
−1
mmol gcat h ; 52.0 h (8 MPa) to 3.1 mmol gcat h ; 348
mmol gcat h ; 3.4 h (PH2/PCO2 = 0.6) after a attaining
−
−
1
−1 −1
−1 −1
−1
h
h
(16 MPa) and then dropped to 0.37 mmol gcat h ; 41.6
maximum of 3.1 mmol gcat h ; 348 h (PH2/PCO2 = 0.33).
This significant lowering of the PGME productivity at a
higher PH2/PCO2 ratio could be attributed to the hydrogenation
of surface C1 species prior to coupling. Fig. 6b compares the
PGME productivity and methane concentration (via gaseous
product analysis) as a function of the PH2/PCO2 ratio. At the
higher PH2/PCO2 ratio of 0.6, the simultaneous reduction of
1
(20 MPa) (Fig. 5a). This pressure-based performance
2
emphasized the density inhomogeneity of scCO (P = 8 MPa,
−
1
−1
d = 2.332 mol l ; P = 16 MPa, d = 5.0772 mol l ; calculated
using the Peng–Robinson EOS) that is, with decreasing
2
pressure, CO preferred to exist in a low density gas-like state,
but with increasing pressure, CO2 preferred to exist in a
densified liquid-like state as verified by video monitoring
performed under the conditions relevant to all reactions
−
1
−1
−1
PGME productivity (3.1 mmol g
h
to 0.03 mmol gcat
cat
−
1
h ) and enhanced methane concentration (0.002 mol% to
∼4 mol%) (ESI;† Fig. S10a and b) reflected the preference for
hydrogenation of C1 species due to the higher availability of
(ESI;† Fig. S9a–e). The fluctuation of PGME selectivity and
productivity above 16 MPa was attributed to the liquid-like
state, which enabled the desorption of intermediates other
than that required for PGME formation. Intriguingly, the
highest PGME productivity was achieved at a total reactant
pressure of 16 MPa corresponding to the liquid-like density
3
2
hydrogen. In particular, although a more detailed account
of the effect of the PH2/PCO2 ratio on the reaction could be
obtained by constructing a solubility model (as reported
previously for the hydrogenation reaction using CO
2
as a co-
3
3,34
of CO , which causes a possible change in the adsorption–
solvent
), such a model cannot be extrapolated to the high
2
desorption behavior of reactants and products crucial for
heterogeneous catalysis (Fig. 5b).
Variation of the PH2/PCO2 ratio. Having selected a total
pressure of 16 MPa as the optimized total pressure for PGME
formation at a fixed temperature of 150 °C, we next
investigated the effect of the PH2/PCO2 ratio on the catalytic
temperatures used in this reaction system. Regardless,
control over the PH2/PCO2 ratio was critical to increase the
productivity of PGME formation.
Time profiles of mixed catalysts. After the confirmation of
reactant pressure and PH2/PCO2 ratio, we evaluated the time
profile for the Ru–Ir-PM catalyst (Fig. 7). Within 1 h, the
resultant product mixture contained methanol, ethanol and
2-propanol with a selectivity of 4.2%, 43.3% and 52.5%,
respectively. In 2 h, methanol disappeared completely, and
the selectivity to ethanol and 2-propanol was also reduced to
33.6% and 39.9%, respectively, whereas, PGME was detected
with a selectivity of 26.5%. Over the next 4 h, PGME became
the only detectable product which reached the highest
selectivity of >99% and then remained constant. Initially
(
2
after 1 h), the catalytic efficiency followed the order:
−
1
−1
−1
-propanol (0.7 mmol gcat h ; 78.6 h ) > ethanol (0.27
−
1
−1
−1
−1 −1
mmol gcat h ; 30.0 h ) > methanol (0.13 mmol gcat h ;
1
−
1
5.0 h ) suggesting the preference for C–C coupling over
Fig. 6 Plots of the PH2/PCO2 ratio dependent (a) PGME productivity
and selectivity and (b) change of PGME productivity with CH
formation. Reaction conditions: Ru–Ir-PM = 0.1 g, temperature = 150
C, total pressure 16 MPa and reaction time 16 h. Each
4
Fig. 7 Time dependent product distribution and PGME productivity
using Ru–Ir-PM. Reaction conditions: catalysts = 0.1 g, temperature =
150 °C, total pressure = 16 MPa and PH2/PCO2 = 0.33. Each
°
=
=
experimental result is the average value from three measurements with
a reproducibility of ±10%.
experimental result is the average value from three measurements
with a reproducibility of ±10%.
4724 | Catal. Sci. Technol., 2021, 11, 4719–4731
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
hydrogenation on Ru–Ir-PM. Furthermore, as the time
studied reaction conditions. Notably, there was no reaction
right after the addition of reactants in any of the cases.
increased from 2 h to 16 h, a substantial enhancement of
−
1
−1
−1
PGME productivity (TOF) from 0.09 mmol gcat
h
(10.0 h )
−
1
−1
−1
to 3.1 mmol gcat
h
(348 h ) was attained. The robustness
Proposed reaction pathways of PGME formation
of the catalyst was evidenced from its efficiency after 72 h of
reaction with a trivial change of PGME productivity from 3.1
The time profile of Ru–Ir-PM (Fig. 7) evidences the (i)
formation of methanol, ethanol and 2-propanol within 1 h as
initial products and also highlights (ii) the gradual
disappearance of the initial products with the formation of
PGME. For the first case, simultaneous detection of
methanol, ethanol and 2-propanol indicated that the reaction
proceeded via parallel routes rather than in a sequential
pathway (methanol → ethanol → 2-propanol), which was also
confirmed from the control experiments (ESI;† Table S1;
entries 1–2). In the second case, we expected the possible
involvement of methanol, ethanol and 2-propanol in the
PGME formation. Thus, control experiments were performed
with a defined amount (10 mmol) of methanol, ethanol and
2-propanol as probe compounds along with scCO2 and
hydrogen (ESI† Table S1; entries 1–3). The result shows that
−
1
−1
−1 −1
mmol gcat
h
to 2.8 mmol gcat
h
(Fig. 7) due to the
slight increase in methane concentration as observed from
the analysis of the gaseous product (ESI;† Fig. S11). Due to
the differences in catalytic performance based on the mixing
process, we proceeded to study the time-profile of the
reaction on both mechanically and chemically mixed
catalysts. Our kinetic investigations demonstrated two
different product distribution trends related to the
mechanical (ESI† Fig. S12a) and chemical mixing processes
(
ESI† Fig. S12b). For the mechanically mixed catalysts, in 1 h,
methanol and ethanol were formed with a selectivity of
2.5% and 27.5%, respectively (Fig. S11a†). PGME was
7
detected with a selectivity of 18.2% after 2 h, which then
reached up to 51.2% with prolonging the reaction time to 16
h along with methanol (34.4%) and ethanol (14.4%). No
propanol was detected throughout the studies. Although, the
molar productivity and TOF of PGME formation increased
2
scCO was successfully converted into PGME with >99%
selectivity in the presence of methanol, which remained
inactive (ESI† Table S1; entry 1). Considering ethanol as a
probe compound, diethoxyethane was the major product with
a selectivity of 79.4% along with PGME (20.6%) (ESI† Table
S1; entry 2). On the other hand, incorporation of 2-propanol
resulted in no PGME and isopropyl acetate was the sole
product (selectivity >99%) (ESI† Table S1; entry 3). Notably,
diethoxyethane and 2-propyl acetate were reported to form
−
3
−1
−1
−1
from 4 × 10
mmol gcat
h
and 0.44 h
to 0.05
−
1
−1
−1
mmol gcat
h
and 5.6 h , the results suggested a relatively
slow rate of product formation after mechanical mixing of
the Ru and Ir catalysts. On the other hand, the chemically
mixed catalyst (ESI† Fig. S11b) showed no detectable
products during
2
2
1
h
of reaction. After
-propanol and PGME were evident with a selectivity of
2.9%, 36.2% and 40.9%, respectively. Upon extending the
2 h, ethanol,
3
5
from the reaction of CH
3
CHO with ethanol
and
3
6
2-propanol, respectively. Additional experiments with CH3-
CHO, which produced ethanol (59.3%), 2-propanol (21.3%),
ethyl acetate (10.0%) and diethoxyethane (9.4%) (ESI;† Table
reaction time to 16 h, ethanol disappeared and the selectivity
to 2-propanol also dropped to 17.6% but PGME reached a
maximum selectivity of 82.4%. Moreover, the molar
productivity and TOF of PGME were also increased from 0.07
3
7
S1; entry 6), confirmed its intermediacy. Hence, the results
of control experiments suggested the existence of CH CHO in
3
−
1
−1
−1
−1 −1
mmol gcat
h
h
and 8.0 h to 1.2 mmol gcat
as the reaction time was changed from 2 h to 16 h. Thus,
compared to the catalytic efficiency of the mechanically
h
and 135
the system. However, we did not observe this compound
−
1
during our scCO hydrogenation experiments possibly due to
2
the instrumental limitations and high reactivity of CH CHO.
3
−
3
−1
mixed catalyst towards PGME formation (4 × 10 mmol gcat
We collated our experimental findings to derive a reaction
−
1
−1
0
.46 h ; h ), a significant enhancement was evident on the
pathway for the transformation of scCO to PGME on Ru–Ir-
2
−
1
−1
−1
chemically mixed catalyst (1.2 mmol gcat h ; 135 h ).
Different product distributions on the catalysts from
various mixing processes are indicative of different reaction
pathways (hydrogenation and C–C coupling) occurring at
different rates. Despite the same metal content, differences
in the mixing process result in significant changes in the
catalytic performance and determine the product
distribution. The catalytic structure and its relation with
activity are a critical, as well as challenging, issue, which rely
on the concentration of metal components or arrangements
of the active constituents (ensemble effects) related to the
electronic and geometric structure of key elements. Thus,
understanding such a correlation needs separate attention.
Based on the highest PGME productivity, Ru–Ir-PM was
considered as a superior catalyst to its mechanical and
chemical analogs with the same composition under the
PM (Scheme 1). PGME was formed from scCO and hydrogen
2
through the hydrogenation and C–C coupling reactions.
Catalytic hydrogenation of CO2 generally proceeds via a
formate or CO pathway. We hypothesized the formate
2
pathway for CO hydrogenation under the present conditions
as CO was below the detection level during the studies. The
hypothesis was further validated through two separate
reactions using HCHO (possible intermediate of the formate
pathway) and CO (0.2 MPa) as substrates. Under the same
conditions applied for scCO2 hydrogenation, HCHO was
transformed into methanol (ESI;† Table S1; entry 4), whereas,
no products were detected from CO (ESI;† Table S1; entry 5).
Altogether, these results were consistent with the notion that
the hydrogenation of scCO
2
proceeded via the formate
3
8
pathway on Ru–Ir-PM. Detection of methanol was notable
−
1
−1
only on Ru (0.14 mmol gcat h ) (Fig. 1); hence, preference
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4725

View Article Online
Paper
Catalysis Science & Technology
Scheme 1 Possible reaction pathways of PGME formation from CO
2
and H
2
based on experimental results obtained under the present reaction
conditions and validated by calculating the energies of each step using DFT.
for hydrogenation of HCHO on the Ru surface can be
suggested, which corresponds well with previous reports.
The extent of HCHO hydrogenation strongly depends on the
on the Ir surface. The negative enolate ion was then stabilized
on positively charged catalytic surface species (RuO ) and
consequently transformed into PGME by the attack of
3
9
2
0
40
0
reduction of RuO
=
2
to Ru . The existence of RuO
85% and Ru = 15%) in Ru–Ir-PM was confirmed from the
XPS analysis (details are in the ESI;† Fig. S13a–d), as a low
2
/Ru (RuO
2
methoxy species via an anti-Markovnikov fashion to the less
0
50
substituted carbon.
Alternatively, depending on the Ir
+
content in Ru–Ir-PM, the enolate may take up H and be
4
1
loading of Ru has a tendency for partial reduction; hence,
hydrogenation of HCHO to methanol was restricted.
Therefore, considering the experimental findings, it is
reasonable to suggest that instead of further hydrogenation,
the HCHO species generated from the hydrogenation of
scCO2 on the catalyst surface coupled to form structure I
preferably on Ir sites of Ru–Ir-PM. The propensity for C–C
coupling on Ir sites was reported earlier.
further hydrogenated into 2-propanol (Fig. 3).
DFT calculations
To gain further insight into the role of each of the catalytic
components (Ru, Ir) and to validate our experimental
findings, we performed density functional theory (DFT)
4
2
4
3,44
In addition, the
calculations. To develop a comprehensive model, the
rapidity of 2-propanol (C3) formation on Ir/MCM-41 (1.16
mmol gcat h ; 130 h ) compared to that on Ru/MCM-41
optimized adsorption geometries of the concerned molecules
on the individual catalysts with different oxidation states
were computed (ESI;† Fig. S14). In addition, the activation
energies of concerned reaction pathways and the
corresponding transition states were also calculated and are
presented in Table S2 and Fig. S15, respectively (ESI†).
Furthermore, Fig. 8a compares the reaction energy profiles of
different steps involved in PGME formation on RuO2 and
−
1
−1
−1
−
1
−1
−1
(
0.02 mmol gcat
inclination for C–C coupling on Ir sites. Thus, in Ru–Ir-PM, Ir,
which exists as IrO (confirmed from the XPS studies; details
h ; 2 h Fig. 2b) also displayed an
2
4
5
are in the ESI†), drove the coupling reaction to form
structure I, which was then transformed into structure II after
the shifts of hydrogen and consequent rearrangement,
which explains the generation of the CH CHO intermediate
Scheme 1). Furthermore, the results from our control
4
6
3
2
IrO surfaces. It is important to note that the reaction
energies of each pathway vary from one metal to another. For
(
experiments also revealed that CH
3
CHO is an active
instance, after adsorption, CO was hydrogenated to produce
CH O species. The calculated reaction energy and activation
2
2
4
7,48
intermediate generated from CO
2
and hydrogen
on Ru–Ir-
PM. In the following steps, the chain propagation was
proposed to involve the nucleophilic attack of an adsorbed
formyl on structure II yielding Cn+1 species (structure III),
which then transformed into an enolate (structure IV) after
energy (E ) (ESI;† Table S2) for this reaction were −1.30 eV
a
and E = 0.39 eV and −1.18 eV and E = 0.59 eV on RuO and
a
a
2
IrO
2
, respectively, which confirmed that the RuO
2
surface was
energetically preferable for CH O formation. The generated
2
4
9
the shifts of hydrogen atoms and subsequent hydrogenation
CH O species, rather than being hydrogenated to methanol
2
4726 | Catal. Sci. Technol., 2021, 11, 4719–4731
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
compared to their oxide counterparts (ESI;† Table S2).
5
3
Interestingly, enolate, as a non-stabilized species that had a
+
high propensity to take up H to form 2-propanol, was
0
0
energetically preferred on Ir (Ir = −2.34 eV; E = 0.44 eV),
a
which explained the rapidity of 2-propanol formation on Ir/
−
1
−1
MCM-41 (130 h ) compared to Ru/MCM-41 (2 h ) as well as
on Ru–Ir-PM depending on the Ir content (Fig. 2b). Although,
it is too early to predict precisely the active sites for the
2
involved reaction steps in the conversion of CO to PGME,
the combined approach (theoretical and experimental)
highlighted the importance of oxidic species rather than the
reduced metal. Overall, based on the energetics, the DFT
analyses confirmed that the formation and stabilization of
reaction intermediates involved in the conversion of CO to
2
PGME require the synergy of both the metals, which provide
multiple active sites for hydrogenation and coupling
reactions to promote the activation of the reactants and the
ability for chain propagation to achieve PGME formation with
excellent selectivity and productivity.
Conclusion
Fig. 8 Calculated energy profiles of the reaction steps are given
0
0
Here, we have shown that utilization of scCO as a prospective
2
relative to (a) RuO
2
and IrO
2
and (b) Ru and Ir of the Ru–Ir-PM
+ H → PGME. Each surface
= black; Ru = pattern and Ir =
reactant for the synthesis of organic compounds can be
drastically improved and likely goes beyond the conventional
hydrogenation of CO2 to form C1 and C2 compounds. Our
findings revealed PGME as a single product, which was
catalysts to simulate the reaction of CO
2
2
0
0
is represented by RuO
gray.
2
= white; IrO
2
obtained from the combination of scCO , hydrogen and
2
on the RuO
to form a CH
reaction energy of −2.05 eV (E = 0.33 eV). This CH CHO
2
surface (E
a
= 0.49 eV), can undergo C–C coupling
highly robust catalysts based on physically mixed Ru/MCM-41
and Ir/MCM-41 at a comparatively low temperature of 150 °C
required for the conversion of CO2 to multi-carbon
compounds. The process seems particularly promising
without the requirement of any additional reagents or
solvents, which have typically flattened the environmental
benefits of CO2 as a feedstock. This method facilitated the
energetically constrained reaction sequence required to
3
CHO intermediate on the IrO
2
surface with a
a
3
intermediate can be further hydrogenated into ethanol on
RuO (ref. 51) with a reaction energy of −2.83 eV (E = 0.46
eV) which explained the high TOF of ethanol on Ru/MCM-41
2
a
−
1
−1
(
0
79 h ) compared to that on Ir (13 h ) [IrO
2
= −1.71 eV (E
a
=
5
2
.82 eV)], also in agreement with a previous report. Prior to
the generation of CH CHO, structure II underwent chain
propagation through a formyl attack and formed structure
produce
hydrocarbon generation. By combining our experimental
results with DFT calculations, we propose plausible
a
multi-carbon oxygenate and also avoids
3
III, where the formyl species was generated on the IrO
2
a
surface with a reaction energy of −2.13 eV. Subsequently,
structure III formed a negatively charged enolate species
mechanism for the conversion of CO2 to PGME. This
alternative strategy of PGME formation may open up many
(structure IV). In the final step, stabilization of the enolate
2
opportunities to utilize CO to mitigate the energy and
ion on the positively charged RuO2 followed by PGME
formation was corroborated by the lower reaction energy on
environmental issues involved in the traditional synthesis of
multi-carbon compounds. This platform may enable further
RuO
2
[−2.19 eV; E
a
= 0.39 eV] compared to that on IrO
2
[−1.03
2
exploration of multi-carbon syntheses with CO as a feedstock
upon additional investigation of similar catalytic systems and
mechanistic pathways. Altogether, this work posits a novel
method to utilize scCO as a feedstock for fuels and chemicals
2
considering environmental perspectives by adopting new
strategies and efficient catalysts to make desired products.
eV; E = 0.79 eV]. We also calculated the reaction energy on
a
0
0
Ru and Ir for the same reaction pathways of PGME
formation from CO (Fig. 8b), logical for the system
containing hydrogen. The hydrogenation of CO2 to CH O
2
2
species is exothermic with reaction energies of −0.57 eV (E
a
=
0
0
0
.46 eV) and −1.22 eV (E
a
= 0.52 eV) on Ru and Ir ,
respectively. Further coupling of CH O species leads to the
2
Materials and methods
Materials
0
3
formation of CH CHO on Ir which corresponded to a
0
reaction energy of −2.05 eV (E
a
= 0.38 eV), whereas, on Ru , it
was −1.89 eV (E = 0.51 eV) (Fig. 8b). It has been observed
Tetraethyl orthosilicate, sodium hydroxide and ruthenium(III)
chloride n-hydrate were purchased from Wako Pure
a
that almost each step needs to bypass a high energetic hurdle
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4727

View Article Online
Paper
Catalysis Science & Technology
Chemicals.
Cetyltrimethylammonium
bromide
and
30 mA). In general, the diffraction data were collected using
continuous scan mode with a scan speed of 2 deg. min over
the scan range of 2θ = 1.5–15° and 2θ = 20–90° for low and
−
1
dihydrogen hexachloroiridate(IV) hydrate were purchased
from Merck and Alfa-Aesar, respectively. CO (>99.99%) and
2
hydrogen have been provided by Nippon Sanso Co. Ltd.
wide-angle, respectively. For textural analysis, N -adsorption
2
measurements were carried out using a Micromeritics ASAP
2
2400 analyser. The volume of adsorbed N was normalized to
Catalyst preparation
standard temperature and pressure. Before the experiment,
samples were degassed at 300 °C for one night. The specific
surface area was determined from the linear part of the BET
The Ru and Ir catalysts supported on mesoporous MCM-41
were synthesized by an in situ method of metal
5
4
introduction. This protocol involves tetraethyl orthosilicate
TEOS) and cetyltrimethylammonium bromide (CTAB) as a
silica source and templating agent, respectively.
plot (P/P = 0.05–0.30).
0
(
One of the most important technique to visualize the
mixtures of supported metal nanoparticles is scanning
transmission microscopy (STEM) in combination with the
bright field (BF) and high-angle annular dark field (HAADF-
STEM) techniques. Images acquired with Z-contrast and
high-resolution stand out very clearly on a light background
of Si and oxygen. As Z contrast imaging is used in HAADF-
STEM, the change in contrast depending on the atomic
number of the elements (Ru = 44 and Ir = 77) created notable
differences of the images to determine the distribution of
each element. TEM, HAADF-STEM, and STEM-EDS images
a
Incorporation of metals was conducted with aqueous
solutions of the corresponding metal salts such as
ruthenium(III)
chloride
n-hydrate
and
dihydrogen
hexachloroiridate(IV) hydrate. Hydrothermal synthesis of Ru
and Ir containing MCM-41 was accomplished from a gel
mixture of molar composition: SiO
.56Na O : 120H O, where x is the desired amount of metal to
be introduced. The gel was prepared through the addition of
.30 g of CTAB in 52 g of deionized water containing 0.54 g
2
: xRu/Ir salt : 0.14 CTAB :
0
2
2
1
of NaOH under stirring at ambient temperature for 1 h to
obtain a clear solution. To this solution, 5 g of TEOS was
added and then stirring was continued for another 1 h. After
that, the required amount of metal salt solution was added
slowly to the surfactant and silica mixture with continuous
stirring until homogeneity was achieved (2 h). The pH of the
gel was maintained at 11.3 using sodium hydroxide (NaOH).
The resultant gel was autoclaved for 48 h at 140 °C. After
that, the autoclave was quenched with water at room
temperature. The resulting solid was separated by filtration
and washed thoroughly with deionized water followed by
drying at 50 °C for overnight. Prior to use, the template
3
were collected using a TITAN ™ G2 60-300 (FEI) operated at
300 kV equipped with STEM-mode with Super-X detector.
Energy dispersive spectroscopy (EDS) analysis was performed
with the ESPRIT imaging software (Bruker).
XPS spectra were recorded on an Ulvac PHI 5601ci
spectrometer. The pressure in the sample chamber was kept
−
7
under 1 × 10 MPa. The binding energy scales were adjusted
to the highest C (1s) peak position equal to 284.9 eV.
Monochromatized Al Kα X-rays (14 kV, 200 W) were used in
place of a conventional X-ray source to obtain high-quality
spectra to avoid overlapping of the satellite peaks; thus,
uncertainty in the determinations of the peak positions and
peak areas was nullified. An electron flood gun was used as a
neutralizer, and the number of acquisitions was maintained
at 12 for each analysis. To ensure the accuracy of the data,
(CTAB) was removed from the synthesized material after
calcining in air at 550 °C for 8 h. The physically mixed
catalyst was prepared by blending of calcined Ru/MCM-41
and Ir/MCM-41 at the desired ratio (wt basis) at room
temperature by hand without adding any external forces. For
the ball-milled catalyst, milling was performed in a planetary
ball mill at a rotation speed of 150 rpm for 2 h using 10 mm
diameter stainless balls at a ball-to-sample weight ratio of
the XPS system was calibrated using the peaks of Cu (2P3/2
)
and Cu (3P) whose binding energies are 932.67 and 75.14 eV,
respectively. All of the powder mineral samples were fixed
onto a copper sample holder using double-sided sticky tape.
Catalytic performance. The reaction was performed in a
1
0 : 1. On the other hand, bimetallic Ru–Ir catalysts were
50 ml stainless steel batch reactor placed over a magnetic
synthesized by the same approach adopted for monometallic
catalysts. Instead of a single metal salt solution, a mixture
with the desired ratio of both metal precursor solutions was
added following the same sequence as described for the
individual metals. Catalysts were obtained in ten different
batches with a reproducibility of ±5%. In each case, the metal
content was controlled to ∼1 wt% based on the gel
composition and then confirmed by elemental analysis.
stirrer in an oven fitted with a fan heater to maintain a
constant temperature of 150 °C (ESI† Fig. S15a–c). Typically,
0.1 g of solid catalyst (Ru–Ir-PM) was incorporated inside the
reactor, which contains a stirrer bar for continuous stirring
of the contents during the reaction. After the stabilization of
the required temperature (generally takes 90 min), the reactor
was pressurized with 4 MPa of hydrogen and then the
2
required total pressure of 16 MPa was completed with CO . A
back-pressure regulator was used to maintain the reaction
pressure inside the reactor and a pressure monitor was used
for monitoring the pressure inside the reactor. Meanwhile,
we prepared a collecting vessel in a cold trap (ice–acetone
mixture; −5.5 °C) to collect condensable products, which was
attached to the outlet of the backpressure regulator.
Catalyst characterization techniques
Different spectroscopic techniques were employed for
structural characterization of the developed catalysts. X-ray
diffraction (XRD) measurements were conducted with a
Rigaku Smart lab X-ray diffractometer (Cu-Kα source, 40 kV,
4728 | Catal. Sci. Technol., 2021, 11, 4719–4731
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
1
Depending on the required reaction time, the reactor was
allowed to attain room temperature and the gasses from the
reactor started to pass to the collecting vessel slowly through
the backpressure regulator until the pressure had the same
value in both places (reactor and collecting vessel) (steady
state). In each case, products were detected from the analysis
of condensable liquid products after reaching steady state
using a headspace-GCMS. For gaseous products, samples were
collected after attaining steady state using a Tedlar sampling
gas bag and then analysed by GC-TCD (Shimadzu GC-8A
equipped with one molecular sieve column and one Porapak
Q column). The solid catalyst was then ready for recycling.
Each experimental result is the average of values from three
AVANCE III spectrometer. H NMR (400 MHz, D O): δ 5.52 (s,
2
1
3
–OH), 4.1–4.3 (m, –OCH CH CH), 1.94, 1.93 (d, –CH ).
NMR (100 MHz, D O): δ 71.35, 65.92, 58.32, 18.33.
2
Phase observation. The phase behavior of scCO2 and
hydrogen at 150 °C was studied separately in a 10 ml high
pressure view cell (ESI† Fig. S16d). After the stabilization of
temperature, hydrogen at a fixed pressure was introduced into
C
3
2
3
the cell followed by the introduction of CO to reach the desired
2
pressure. For phase observation during the reaction, the
contents were stirred continuously. All images were recorded
during stirring to observe the change in the state of reactants.
Calculation of density. Gas densities were calculated using
a
chemical engineering process simulator COCO
measurements with
balance was calculated using the formula moles of carbon in
products (gaseous condensation)/moles of carbon in
reactants. The carbon balance for this reaction floated from
3% to 98%. In order to avoid contamination and cross-
a
reproducibility of ±10%. Carbon
(AmsterCHEM version 3.2) using the Peng–Robinson equation
of state. Graboski and Daubert have modified the coefficients
in the SRK EOS and provided a special relation for hydrogen.
This modification of the Soave–Redlich–Kwong equation of
state (SRK EOS) has been recommended by the American
Petroleum Institute (API). The uncertainty was estimated as
0.5% from the difference between experiment and calculation.
Computational methodology. The calculations were
performed on a (100) surface model with a 4 layer thickness
+
9
correlation between experimental reactants and products, we
performed a blank test before each experimental trial. In
addition, we performed a blank run prior to product analysis
to ensure the identities of the products formed. In this way,
we were able to avoid contamination in the reaction products.
GCMS analysis for liquid products. The analyses were
carried out with an Agilent 6890/5973 inert MSD (mass
selective detector) system. Headspace analysis was performed
with an Agilent G1888MS auto-sampler equipped with a 1 ml
sample loop. A GC (HP 6890) was equipped with a VF-
WAXms column (J&W, 60 m × 0.25 mm, 0.50 μm) and a flame
ionization detector. The GC method used was as follows: an
initial oven temperature of 40 °C was held for 2 minutes. In
created from the bulk structure of RuO
2 2
and IrO . The initial
geometry of the adsorbed molecules was calculated over (100)
0
0
RuO , IrO and their hydrogenated (Ru , Ir ) surfaces using
2
2
the Adsorption Locator module (Dassault Systemes BIOVIA)
that allows finding out the low energy adsorption sites on
both periodic and non-periodic substrates as well as
investigations on the preferential adsorption of mixtures of
adsorbate components. The Adsorption Locator module
identifies possible adsorption sites by carrying out Monte
Carlo searches of the configurational space of the substrate–
adsorbate system as the temperature is slowly decreased
within the simulated annealing process of a molecular
−
1
the next step, the temperature was ramped at 10 °C min
until it reached 230 °C and held for 5 minutes. Helium was
−
1
used as a carrier gas with a flow rate of 0.7 ml min in a
constant flow mode. For MS analysis, the temperature of the
ion source was maintained at 230 °C. The loop and transfer
line temperatures were set at 65 °C and 75 °C, respectively.
Before sampling, vial equilibration was set at 20 min and the
vial pressurization was set at 10 psi for 0.15 min. The
5
5,56
dynamics run.
The minimum energy structures obtained
from the Monte Carlo method were then fully optimized with
the DFT method as described below.
All the reaction energy calculations were performed using
5
7
3
density functional theory (DFT) using DMol (ref. 58 and
59) of DASSAULT SYSTEMES BIOVIA. Gradient corrected
selectivity to each product was calculated by the following
P
6
0,61
62
equation S
product ‘i’ as monitored by GC and
concentration of the product (moles). Quantification of the
products was obtained using a multi-point calibration curve
for each product. Productivity and turnover frequency were
calculated as follows:
i
= C
i
/
C
p
, where C
i
is the concentration of the
functional BLYP
and the DNP basis set with the spin
P
C
p
is the total
unrestricted option were used throughout the calculations.
The initial spin states were decided according to the spin
states of the metal cations present in the model. The Kohn–
5
7
Sham equation was expanded in a double numeric quality
basis set with polarization functions (DNP). The orbital
cutoff range and Fermi smearing were selected as 5.0 Å and
0.001 Ha, respectively. The self-consistent-field (SCF)
procedures were performed to obtain well-converged
Productivity of PGME
=
amount of PGME (mmol)/g of catalyst used/h
geometrical and electronic structures at
a convergence
criterion of 10–6 a. u. The energy, maximum force, and
maximum displacement convergence were set to 10–6 Ha,
Turnover frequency (TOF) = amount of product produced/total
−
1
amount of metal used × time.
0.002 Ha Å , and 0.005 Å, respectively. The activation
energy was calculated using the transition state search
3
NMR analysis. Solution-state spectra of the resultant
algorithm as implemented in DMol with the nudged elastic
liquid product were acquired using a Bruker 400 MHz
band method (NEB).
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4729

View Article Online
Paper
Catalysis Science & Technology
Conflicts of interest
21 C. De, Q. Cai, X. Wang, J. Zhao and B. Lu, J. Chem. Technol.
Biotechnol., 2011, 86, 105–108.
There are no conflicts to declare.
2
2 P. G. Jessop and R. Noyori, in Green Chemistry using liquid
and supercritical carbon dioxide, ed. J. M. DeSimone and W.
Tumas, Oxford press, 2003, ch. 3, pp. 48–61.
Acknowledgements
2
2
2
2
3 P. G. Jessop, T. Ikariya and R. Noyori, Nature, 1994, 368,
H. K. is funded by the Japan–U.S. cooperation project for
research and standardization of Clean Energy Technologies,
The Ministry of Economy Trade and Industry (METI), CREST
and Japan Science and Technology (JST).
2
31–233.
4 S. Wesselbaum, U. Hintermair and W. Leitner, Angew. Chem.,
Int. Ed., 2012, 51, 8585–8588.
5 Z. He, Q. Qian, Z. Zhang, Q. Meng, H. Zhou, Z. Jiang and B.
Han, Philos. Trans. R. Soc., A, 2015, 373, 20150006.
6 W. Reeve and A. Sadle, J. Am. Chem. Soc., 1950, 72, 1251–1254.
References
1
2
3
4
5
Z. Li, J. Wang, Y. Qu, H. Liu, C. Tang, S. Miao, Z. Feng, H.
An and C. Li, ACS Catal., 2017, 7, 8544–8548.
G. H. Gunasekar, K. Park, K.-D. Jung and S. Yoon, Inorg.
Chem. Front., 2016, 3, 882–895.
G. Centi, E. A. Quadrelli and S. Perathoner, Energy Environ.
Sci., 2013, 6, 1711–1731.
27 M. Inoue, T. Miyake, T. Inui and Y. Takegami, J. Chem. Soc.,
Chem. Commun., 1983, 70–72.
28 G. C. Bond, Platinum Met. Rev., 1983, 21, 16–18.
29 Y. Wang, S.-L. Chen, Y.-L. Gao, Y.-Q. Cao, Q. Zhang, W.-K.
Chang and J. B. Benziger, ACS Catal., 2017, 7, 5572–5584.
30 J. Wang, G. Li, Z. Li, C. Tang, Z. Feng, H. An, H. Liu, T. Liu
and C. Li, Sci. Adv., 2017, 3, e1701290.
31 L. Liu and A. Corma, Chem. Rev., 2018, 118, 4981–5079.
32 K. P. Kuhl, E. R. Cave, D. N. Abram and T. F. Jaramillo,
Energy Environ. Sci., 2012, 5, 7050–7059.
33 E. Bogel-Łukasik, R. Bogel-Łukasik, K. Kriaa, I. Fonseca, Y.
Tarasenko, M. Nunes da Ponte, A. Paiva and G. Brunner,
Green Chem., 2007, 9, 427–430.
34 E. Bogel-Łukasik, R. Bogel-Łukasik, K. Kriaa, I. Fonseca, Y.
Tarasenko and M. Nunes da Ponte, J. Supercrit. Fluids,
2008, 45, 225–230.
35 M. R. Capeletti, L. Balzano, G. de la Puente, M. Laborde and
U. Sedran, Appl. Catal., A, 2000, 198, L1–L4.
36 T. Kunugi and T. Kono, Kogyo Kagaku Zasshi, 1969, 72,
1282–1289.
37 F. Hayashi, M. Tanaka, D. Lin and M. Iwamoto, J. Catal.,
2014, 316, 112–120.
M. D. Porosoff, B. Yan and J. G. Chen, Energy Environ. Sci.,
2
016, 9, 62–73.
A. Álvarez, A. Bansode, A. Urakawa, A. V. Bavykina, T. A.
Wezendonk, M. Makkee, J. Gascon and F. Kapteijn, Chem.
Rev., 2017, 117, 9804–9838.
S. Li, Y. Xu, Y. Chen, W. Li, L. Lin, M. Li, Y. Deng, X. Wang,
B. Ge, C. Yang, S. Yao, J. Xie, Y. Li, X. Liu and D. Ma, Angew.
Chem., Int. Ed., 2017, 56, 10761–10765.
S. Kuld, M. Thorhauge, H. Falsig, C. F. Elkjaer, S. Helveg, I.
Chorkendorff and J. Sehested, Science, 2016, 352, 969–973.
M. Aresta, A. Dibenedetto and E. Quaranta, J. Catal.,
6
7
8
9
2
016, 343, 2–45.
A. Banerjee, G. R. Dick, T. Yoshino and M. W. Kanan, Nature,
016, 531, 215–220.
0 L. Yu, J. Huang and Y. Wang, ChemCatChem, 2018, 10,
853–4867.
2
1
1
4
1 J. Wu, S. Ma, J. Sun, J. I. Gold, C. S. Tiwary, B. Kim, L. Zhu,
N. Chopra, I. N. Odeh, R. Vajtai, A. Z. Yu, R. Luo, J. Lou, G.
Ding, P. J. A. Kenis and P. M. Ajayan, Nat. Commun., 2016, 7,
38 M. Hus, D. Kopac, N. S. Stefancic, D. L. Jurkovic, V. D. B. C.
Dasireddy and B. Likozar, Catal. Sci. Technol., 2017, 7,
5900–5913.
1
3869.
39 M. Karamad, H. A. Hansen, J. Rossmeisl and J. K. Nørskov,
ACS Catal., 2015, 57, 4075–4081.
1
1
1
1
1
1
1
1
2
2 E. L. Clark, C. Hahn, T. F. Jaramillo and A. T. Bell, J. Am.
Chem. Soc., 2017, 139, 15848–15857.
3 E. Christensen, J. Yanowitz, M. Ratcliff and R. L.
McCormick, Energy Fuels, 2017, 25, 4723–4733.
4 Y. Gao, S. Shizhen, Z. Zhao, H. Tao and Z. Sun, Acta Phys.-
Chim. Sin., 2018, 34, 858–872.
5 A. Banerjee and M. W. Kanan, ACS Cent. Sci., 2018, 4,
40 S. Wang, E. T. Schrunk, H. Mahajan and R. J. Farrauto,
Catalysts, 2017, 7, 88–101.
41 S. Coman, M. Florea, F. Cocu, V. I. Parvulescu, P. A. Jacobs,
C. Danumah and S. Kaliaguine, Chem. Commun.,
1999, 2175–2176.
42 S. Kattel, P. Liu and J. G. Chen, J. Am. Chem. Soc., 2017, 139,
9739–9754.
43 M.-Y. Ngai, A. Barchuk and M. J. Krische, J. Am. Chem. Soc.,
2007, 129, 280–281, and references therein.
6
06–613.
6 Q. Qian, M. Cui, Z. He, C. Wu, Q. Zhu, Z. Zhang, J. Ma, G.
Yang, J. Zhang and B. Han, Chem. Sci., 2015, 6, 5685–5689.
7 M. Cui, Q. Qian, Z. He, Z. Zhang, J. Ma, T. Wu, G. Yanga and
B. Han, Chem. Sci., 2016, 7, 5200–5205.
44 K. Koda, T. Matsu-ura, Y. Obora and Y. Ishii, Chem. Lett.,
2009, 38, 838–839, and references therein.
8 B. L. Smith, L. S. Ott and T. J. Bruno, Environ. Sci. Technol.,
45 L. Yang, J. Huang, J. Cen, D.-L. Chen, H. Zeng, Z. Rui, R.
Luque and P. Duane, J. Mater. Chem. A, 2021, 9, 7094–7101.
46 A. Kiennemann, H. Idriss, R. Kieffer, P. Chaumette and D.
Durand, Ind. Eng. Chem. Res., 1991, 30(6), 1130–1138.
47 R. Kortlever, J. Shen, K. Jan, P. Schouten, F. Calle-Vallejo and
M. T. M. Koper, J. Phys. Chem. Lett., 2015, 6, 4073–4082.
2
008, 42, 7682–7689.
9 Y. D. Chaniago, G. R. Harvianto, A. Bahadori and M. Lee,
Process Saf. Environ. Prot., 2016, 103, 413–423.
0 H. A. Pecorini and J. T. Banchero, Ind. Eng. Chem., 1956, 48,
1
287–1297.
4730 | Catal. Sci. Technol., 2021, 11, 4719–4731
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
4
4
5
8 M. E. Fakley and R. A. Head, Appl. Catal., 1983, 5, 3–18.
9 T. J. Mazanec, J. Catal., 1986, 98, 115–125.
0 W. Zhang, H. Wang, Q. Li, Q. Dong, N. Zhao, W. Wei and Y.
Sun, Appl. Catal., A, 2005, 294, 188–196.
55 N. Metropolis, A. W. Rosenbluth, M. N. Rosenbluth, A. H.
Teller and E. Teller, J. Chem. Phys., 1953, 21, 1087–1092.
56 S. Kirkpatrick, C. D. Gelatt and M. P. Vecchi, Science,
1983, 220, 671–680.
5
5
1 J. Jae, W. Zheng, A. M. Karim, W. Guo, R. F. Lobo and D. G.
Vlachos, ChemCatChem, 2014, 6, 848–856.
2 S. Li, Y. Xu, Y. Chen, W. Li, L. Lin, M. Li, Y. Deng, X. Wang,
B. Ge, C. Yang, S. Yao, J. Xie, Y. Li, X. Liu and D. Ma, Angew.
Chem., Int. Ed., 2017, 56, 10761–10765.
57 W. Kohn and L. J. Sham, Phys. Rev., 1965, 140,
A1133–A1138.
58 B. Delley, J. Chem. Phys., 1990, 92, 508–517.
59 B. Delley, J. Chem. Phys., 2000, 113, 7756–7764.
60 A. Becke, J. Chem. Phys., 1988, 88, 2547–2553.
61 C. Lee, W. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter
Mater. Phys., 1988, 37, 785–789.
5
5
3 T. Graening and J. F. Hartwig, J. Am. Chem. Soc., 2005, 127,
1
7192–17193, and references therein.
4 M. Chatterjee, T. Iwasaki, Y. Onodera and T. Nagase, Catal.
Lett., 1999, 61, 199–202.
62 B. Delley, Phys. Rev. B: Condens. Matter Mater. Phys.,
2002, 66, 155125–155134.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 4719–4731 | 4731