Surface Engineering of a Supported PdAg Catalyst for Hydrogenation of CO2 to Formic Acid: Elucidating the Active Pd Atoms in Alloy Nanoparticles

Article abstract of DOI:10.1021/jacs.8b04852

The hydrogenation of carbon dioxide (CO₂) to formic acid (FA; HCOOH), a renewable hydrogen storage material, is a promising means of realizing an economical CO₂-mediated hydrogen energy cycle. The development of reliable heterogeneous catalysts is an urgent yet challenging task associated with such systems, although precise catalytic site design protocols are still lacking. In the present study, we demonstrate that PdAg alloy nanoparticles (NPs) supported on TiO₂ promote the efficient selective hydrogenation of CO₂ to give FA even under mild reaction conditions (2.0 MPa, 100 °C). Specimens made using surface engineering with atomic precision reveal a strong correlation between increased catalytic activity and decreased electron density of active Pd atoms resulting from a synergistic effect of alloying with Ag atoms. The isolated and electronically promoted surface-exposed Pd atoms in Pd@Ag alloy NPs exhibit a maximum turnover number of 14 839 based on the quantity of surface Pd atoms, which represents a more than 10-fold increase compared to the activity of monometallic Pd/TiO₂. Kinetic and density functional theory (DFT) calculations show that the attack on the C atom in HCO₃^- by a dissociated H atom over an active Pd site is the rate-determining step during this reaction, and this step is boosted by PdAg alloy NPs having a low Pd/Ag ratio.

Full text of DOI:10.1021/jacs.8b04852

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Article
Surface Engineering of a Supported PdAg Catalyst for Hydrogenation of
2
CO to Formic Acid: Elucidating the Active Pd Atoms in Alloy Nanoparticles
Kohsuke Mori, Taiki Sano, Hisayoshi Kobayashi, and Hiromi Yamashita
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b04852 • Publication Date (Web): 22 Jun 2018
Downloaded from http://pubs.acs.org on June 23, 2018
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 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Surface Engineering of a Supported PdAg Catalyst for Hydrogena-
tion of CO2 to Formic Acid: Elucidating the Active Pd Atoms in Alloy
Nanoparticles
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
Kohsuke Mori*,†, ‡, §, Taiki Sano†, Hisayoshi Kobayashi,ǁ and Hiromi Yamashita*,†, §
† Division of Materials and Manufacturing Science, Graduate School of Engineering, Osaka University, 2ꢀ1 Yamadaꢀoka,
Suita, Osaka 565ꢀ0871, Japan. Tel&FAX: +81ꢀ6ꢀ6879ꢀ7460, +81ꢀ6ꢀ6879ꢀ7457
Eꢀmail: mori@mat.eng.osakaꢀu.ac.jp, yamashita@mat.eng.osakaꢀu.ac.jp
‡ JST, PRESTO, 4ꢀ1ꢀ8 Honcho, Kawaguchi, Saitama, 332ꢀ0012, Japan
§ Elements Strategy Initiative for Catalysts Batteries ESICB, Kyoto University, Katsura, Kyoto 615ꢀ8520, Japan.
ǁ
Kyoto Institute of Technology, Matsugasaki, Sakyoꢀku, Kyoto, 606ꢀ8585, Japan
ABSTRACT: The hydrogenation of carbon dioxide (CO2) to formic acid (FA; HCOOH), a renewable hydrogen storage material, is
a promising means of realizing an economical CO2ꢀmediated hydrogen energy cycle. The development of reliable heterogeneous
catalysts is an urgent yet challenging task associated with such systems, although precise catalytic site design protocols are still
lacking. In the present study, we demonstrate that PdAg alloy nanoparticles (NPs) supported on TiO2 promote the efficient selective
hydrogenation of CO2 to give FA even under mild reaction conditions (2.0 MPa, 100 °C). Specimens made using surface engineerꢀ
ing with atomic precision reveal a strong correlation between increased catalytic activity and decreased electron density of active
Pd atoms resulting from a synergistic effect of alloying with Ag atoms. The isolated and electronically promoted surfaceꢀexposed
Pd atoms in Pd@Ag alloy NPs exhibit a maximum turnover number of 14,839 based on the quantity of surface Pd atoms, which
represents a more than tenꢀfold increase compared to the activity of monometallic Pd/TiO2. Kinetic and density functional theory
–
(DFT) calculations show that the attack on the C atom in HCO3 by a dissociated H atom over an active Pd site is the rateꢀ
determining step during this reaction, and this step is boosted by PdAg alloy NPs having a low Pd/Ag ratio.
Introduction
geneous transition metal complexes as catalysts for FA proꢀ
duction, especially those based on Ir and Ru, and interesting
results have been obtained.8, 11, 20ꢀ23 Unfortunately, the develꢀ
opment of heterogeneous catalysts lags significantly in spite of
the obvious practical advantages of such materials.5, 24ꢀ32 Thus,
heterogeneous catalysts still require the use of high catalyst
concentrations, extremely elevated pressures and organic solꢀ
vents.
The selective transformation of carbon dioxide (CO2) to
useful chemicals or fuels has the potential to alleviate both
climate change and future demands for fossil fuels.1ꢀ4 The hyꢀ
drogenation of CO2 to produce formic acid (FA; HCOOH) is a
promising strategy, since the target product is a valuable
commodity chemical commonly used for preservative and
antibacterial purposes.5ꢀ6 Because FA has a relatively high
hydrogen content (53 g·L–1) and is a lessꢀtoxic, nonflammable
liquid under ambient conditions, it is also regarded as a promꢀ
ising hydrogen storage material. Using this compound, chemiꢀ
cally stored hydrogen could be liberated in a controlled manꢀ
ner in the presence of appropriate catalysts, even at room temꢀ
perature.7ꢀ14 Recently, FA has also been considered as an alterꢀ
native to methanol as a fuel source for direct liquid fuel cell
systems generating electricity.15 The current industrial FA
production methods traditionally include the hydrolysis of
methyl formate and oxidation of biomass.16ꢀ17 Thus, the appliꢀ
cation of CO2 hydrogenation catalysts to generate FA must be
researched in order to ensure efficient CO2 utilization and the
realization of economicallyꢀviable CO2ꢀmediated hydrogen
energy cycles.18
The immobilization of existing molecular organometallic
complexes on insoluble matrixes could combine the tunability
of the wellꢀdefined active centers in homogeneous catalysts
and the operational simplicity of heterogeneous materials.33
Surfaceꢀgrafted silicaꢀbased substances having a variety of
functional groups allow the coordination of analogous metal
complexes.34 Porous organic framework (POF) materials, inꢀ
cluding metal organic frameworks (MOFs), are also potential
supports for singleꢀatom active centers, based on the utilizaꢀ
tion of the framework skeleton to provide suitable coordinaꢀ
tion sites.35ꢀ36 The design of this category of catalysts is relaꢀ
tively simple because the resulting coordination environments
are similar to those of their homogeneous counterparts.
However, to date, there have been few insights with regard
to the most promising design strategies for active catalyst sites
in supported metal nanoparticles (NPs). Pdꢀ, Auꢀ and Ruꢀ
based catalysts have all been shown to be active for the target
reaction and their activity can be enhanced by the appropriate
selection of support materials, by increasing the NP dispersion
The gas phase hydrogenation of CO2 to FA involves a posiꢀ
tive free energy change (∆G = 33 kJ molꢀ1), while the same
reaction in aqueous solution proceeds more readily because of
the relatively low (negative) free energy (∆G = –4 kJ molꢀ1).19
Significant efforts have been devoted to the research of homoꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
and by altering the electronic structures of the active metals.
Page 2 of 9
PdAg/LDH (entry 2). However, the other heterogeneous PdAg
catalysts, such as PdAg/MgO, PdAg/CeO2 and PdAg/Al2O3,
were found to be less active (entries 3ꢀ5). Interestingly,
PdAg/TiO2 exhibited considerably high activity compared
with the corresponding PdCu, PdZn, PdAu and PdNi catalysts
supported on TiO2 (entries 6ꢀ9). In the Pd Kꢀedge Xꢀray abꢀ
sorption nearꢀedge structure (XANES) spectra, the edge posiꢀ
tion of the PdAg/Al2O3 was lower than that of the PdAg/TiO2
(Figure S1), which presumably ascribed to the differences in
the work functions between Al2O3 (4.70 eV) and TiO2 (6.21
eV). This result also suggests that the moderate interaction
between PdAg NPs and TiO2 is suitable for attaining active
catalysts for this target reaction.
1
2
3
4
5
6
7
8
The electronic and geometric effects associated with bimetalꢀ
lic NPs, whose architectural configuration involves two metꢀ
als, can be surface engineered to obtain random alloys, segreꢀ
gated or core–shell structures. This flexibility offers signifiꢀ
cant opportunities to maximize catalytic activity and selectiviꢀ
ty compared with monometallic counterparts, owing to the
unique interactions between neighbouring metals.37ꢀ39 As a
result, there has been remarkable progress in the development
of supported bimetallic NP catalysts, with the creation of novꢀ
el catalysts based on a wellꢀunderstood design strategy. Howꢀ
ever, although it is well known that catalytic performance can
be modified as a result of alloying, additional elucidation of
the catalytically active species during the hydrogenation of
CO2 to FA is required.
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
Table 1. Screening results for CO2 hydrogenation to formic
acid over various catalysts.a
In this paper, we present a new catalyst based on PdAg alloy
NPs supported on TiO2, intended to promote heterogeneous
CO2 hydrogenation to produce FA. This material was used
because surface engineering with atomic precision can provide
PdAg alloys with different surface compositions. When using
this catalyst, isolated and electronꢀrich Pd atoms created with
the aid of neighbouring Ag atoms at a low Pd/Ag ratio boost
the electronegativity of the dissociated hydride species. This
Molar
ratio
(Pd:M)
Entry
Metal NPs
Support
TON@24h
1
2
PdAg
PdAg
PdAg
PdAg
PdAg
PdCu
PdZn
PdAu
PdNi
Pd
1:1
1:1
1:1
1:1
1:1
1:1
1:1
1:1
1:1
1:0
TiO2
LDH
MgO
CeO2
Al2O3
TiO2
TiO2
TiO2
TiO2
TiO2
748
342
200
142
40
3
4
ꢀ
ultimately enhances the rate determining HCO3 hydrogenation
step, as demonstrated by kinetic and density functional theory
(DFT) calculations.
5
6
353
344
251
120
281
7
8
9
Results and Discussion
10
Supported PdAg catalysts were prepared using a convenꢀ
tional method. Briefly, TiO2 (JRCꢀTiOꢀ6, BET surface area
(SBET) = 100 m2·g–1, rutile) was impregnated with an aqueous
solution of [Pd(NH3)4]Cl2 and AgNO3. The Pd/Ag ratio in the
resulting material was controlled by varying the concentration
of Ag while maintaining a constant Pd loading of 1.0 wt%.
The samples obtained in this manner were subsequently reꢀ
duced with NaBH4 without calcination, affording PdAg alloy
NPs supported on TiO2 having different atomic ratios. For
comparison purposes, various other supports, including layꢀ
ered double hydroxide (LDH), MgO, CeO2 and Al2O3, were
also employed for the deposition of PdAg alloy NPs. In addiꢀ
tion, Pdꢀbased alloy NPs supported on TiO2 were synthesized
along with Cu, Zn, Au or Ni precursors as the second metal by
the same impregnation method, followed by reduction with
NaBH4.
a Conditions: catalyst (50 mg), 1.0 M aqueous NaHCO3 solution (10 mL), H2:CO2
(1:1, total 2.0 MPa), 100 °C.
Figure 1. (A) TON values (based on Pd atoms) versus the mole
fraction of Ag in PdAg/TiO2 catalysts having different Pd and Ag
concentrations. (B) TEM image of a single PdAg/TiO2 NP and the
elemental distributions as determined by crossꢀsectional EDX line
profiling.
We initially attempted to identify the optimal catalyst for
the hydrogenation of CO2 to FA, and generated the data sumꢀ
marized in Table 1. A typical reaction was performed in a 25
mL stainless steel reactor in a basic aqueous solution containꢀ
ing 1.0 M NaHCO3 under a total pressure of 2.0 MPa (H2:CO2
= 1:1) at 100 °C over 24 h. FA was obtained with >99 % seꢀ
lectivity for all catalysts and any other products cannot be
observed. The support used for the alloy NPs was regarded as
a crucial parameter because the support can provide additional
heterogeneous catalytic sites, and may also modify particle
sizes and electronic properties. The PdAg/TiO2 (Pd:Ag=50:50)
specimen was found to be an efficient heterogeneous catalyst,
with a high turnover number (TON) of 748 (entry 1) after 24
h. In addition, trials without a catalyst or using only the TiO2
support material showed no activity. We recently reported that
strongly basic LDH can act as an efficient support for a singleꢀ
site Ru catalyst during the hydrogenation of CO2, and the preꢀ
sent data show moderate activity was obtained when using
Additionally, the catalytic activity was greatly affected by
the composition of the PdAg alloy NPs. Figure 1A plots TON
values versus the Ag content in various specimens. The monꢀ
ometallic Pd/TiO2 gave poor results, while the reaction using
monometallic Ag/TiO2 was also extremely sluggish (TON =
7). The TON increased with increases in the Ag proportion up
to 70%, after which it decreased. This “volcano” type variaꢀ
tion in activity based on the Pd/Ag ratio suggests the forꢀ
mation of a uniform PdAg alloy structure on the surface of the
TiO2 as well as a synergic effect originating from the integraꢀ
tion of the Pd and Ag.40 Highꢀangle annular darkꢀfield scanꢀ
ning transmission electron microscopy (HAADFꢀSTEM) imꢀ
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
ages demonstrated highly dispersed PdAg alloy NPs with a atoms had the opposite distribution for the Ag@Pd/TiO2 (Fig-
1
2
3
4
5
6
7
8
narrow size distribution on the TiO2 support in the case of the
PdAg/TiO2 (Pd:Ag = 30:70), with a mean particle diameter
(dave) of 3.3 nm (Figure S2). This value is similar to that obꢀ
tained for the Pd/TiO2 (dave = 3.2 nm, see Figure S3). Energyꢀ
dispersive Xꢀray spectroscopy (EDX) line analysis revealed
that both Pd and Ag were situated on the same particles, conꢀ
firming the formation of a random PdAg alloy (Figure 1B). In
contrast, the mean particle size of the PdAg NPs on the Al2O3,
which exhibited activity substantially inferior to that observed
using TiO2, was 9.5 nm with a wide size distribution (Figure
S4). The formation of small PdAg NPs with a narrow size
distribution could therefore be one important factor contribꢀ
uting to the high catalytic activity of the PdAg/TiO2.
ure 2B).
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
As discussed, the maximum activity was obtained with a
low Pd/Ag ratio (Pd:Ag = 30:70) in trials using random PdAg
alloy NPs. This result indicates that Pd atoms surrounded by
excess Ag atoms can efficiently facilitate the hydrogenation of
CO2. Based on these data, we investigated the effect of surꢀ
faceꢀexposed active Pd atoms in alloy NPs on the catalytic
activity, applying a surface engineering approach.
Figure 3. (A) Pd Kꢀedge and (B) Ag Kꢀedge FTꢀEXAFS spectra
of PdAg/TiO2, Pd@Ag/TiO2, Ag@Pd/TiO2, Pd/TiO2 and referꢀ
ence materials.
Table 2. Curve fitting results for Pd and Ag Kꢀedge EXAFS data.
sample
edge
Pd K
shell
CN CNtotal R/Å
ꢁσ2/Å2
Pd–Pd
Pd–Ag
Ag–Ag
Ag–Pd
Pd–Pd
Pd–Ag
Ag–Ag
Ag–Pd
Pd–Pd
Pd–Ag
Ag–Ag
Ag–Pd
3.1
4.0
5.1
4.4
3.3
4.8
4.5
4.3
1.8
3.8
7.0
3.7
2.77
2.85
2.86
2.80
2.75
2.80
2.83
2.78
2.75
2.79
2.84
2.79
0.086
0.018
0.095
0.035
0.098
0.079
0.011
0.036
0.047
0.010
0.011
0.036
7.1
9.5
PdAg/TiO2
Pd/Ag = 30/ 70
Ag K
Pd K
Ag K
Pd K
Ag K
8.1
Pd@Ag/TiO2
Pd/Ag = 30/ 70
Figure 2. TEM images of and elemental distributions along single
NPs of (A) Pd@Ag/TiO2 and (B) Ag@Pd/TiO2 as determined by
crossꢀsectional EDX line profiling.
8.8
5.6
Ag@Pd/TiO2
Pd/Ag = 30/ 70
Because of the complete solid solubility and similar reducꢀ
tion potentials of Pd and Ag ions (E0(Pd2+/Pd0) = +0.99 V,
E0(Ag+/Ag0) = +0.80 V), random PdAg alloy NPs can be readꢀ
ily formed via the coꢀreduction of Pd and Ag precursors. Thus,
we were able to tune the surface composition of PdAg NPs to
form PdcoreAgshell structures via the successive reduction of
metal precursors. This was accomplished by first impregnating
the TiO2 with [Pd(NH3)4]Cl2 followed by reduction with
NaBH4 to form the Pd core. Next, AgNO3 impregnation was
carried out and the material again subjected to reduction with
NaBH4 to form a Ag shell, affording Pd@Ag/TiO2. To proꢀ
duce AgcorePdshell catalysts, Ag@Pd/TiO2 was also synthesized
by the same successive reduction method but in the opposite
order. Pd:Ag molar ratio was unified to 30:70 for fair comparꢀ
ison. In this paper, the term of alloy was used for both ranꢀ
domly distributed PdAg NPs coreꢀshell type nanoparticles.
10.7
Xꢀray absorption measurements were carried out to preciseꢀ
ly determine average local structural information for the Pd
and Ag atoms. The shapes of the normalized Xꢀray absorption
nearꢀedge structure (XANES) spectra at the Pd Kꢀedge and the
edge positions of three PdAg samples resembled those of Pd
foil and monometallic Pd/TiO2 but differed from those of PdO.
These results suggest that the oxidation state of the Pd atoms
was close to Pd0 (Figure S7). More detailed inspection found
two distinct peaks at approximately 24,367 and 24,385 eV,
corresponding to the allowed 1s → 5p transition in the PdAg
samples. These peaks were slightly shifted to lower energy
values compared to the Pd/TiO2, indicating that the symmetry
of the Pd metal face centered cubic (fcc) structure was slightly
disordered by integration with the Ag.41 The Pd Kꢀedge Fouriꢀ
er transformꢀextended Xꢀray absorption fine structure (FTꢀ
EXAFS) spectra contained a single sharp peak associated with
Pd−Pd bonds at approximately 2.6 Å that was also suggestive
of metallic Pd (Figure 3A). However, the Pd–Pd distance in
the PdAg samples was found to be slightly longer compared to
the values for Pd foil and Pd/TiO2, indicating the presence of
heteroatom Pd−Ag bonds. No peaks due to Pd−O or Pd−O−Pd
Elemental analysis confirmed that the average composition
of the NPs in both types of materials was Pd30Ag70. HAADFꢀ
STEM images also showed highly dispersed PdAg NPs with
average diameters of 3.1 and 3.2 nm for the Pd@Ag/TiO2 and
Ag@Pd/TiO2, respectively (Figures S5 and S6). EDX line
scans confirmed the successful surface engineering of the NPs,
such that Pd atoms were preferentially located in the core reꢀ
gion, while the Ag atoms were situated in the shell region in
the Pd@Ag/TiO2 (Figure 2A). Conversely, the Pd and Ag
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
bonds (expected at approximately 1.6 and 3.0 Å in the case of
observed to decrease in the order of Pd@Ag/TiO2 >
PdAg/TiO2 > Ag@Pd/TiO2 (Figure 4B).
1
2
3
4
5
6
7
8
PdO) were observed. The Ag Kꢀedge XANES spectra of these
three PdAg samples were also found to resemble those of Ag
foil and monometallic Ag/TiO2, confirming the presence of
metallic Ag (Figure S8). Their FTꢀEXAFS spectra contained
a single intense peak ascribed to contiguous Ag−Ag bonds
with a length of approximately 2.7 Å (Figure 3B). This peak
was shifted to a slightly shorter interatomic distance compared
with those of pure Ag foil and Ag/TiO2, again suggesting
Pd−Ag bonds.
9
Coordination numbers (CNs) and bond lengths (R) obtained
from EXAFS curve fitting can provide critical information
regarding the composition of the bimetallic NPs, and these
data are presented in Table 2. The Pd–Pd bond distances were
evidently shorter than those of Pd–Ag bonds, which in turn
were shorter than Ag–Ag bonds. These results are consistent
with the shifts of the main peaks due to metallic bonding in the
FTꢀEXAFS spectra. It is widely accepted that the atoms inside
an fcc lattice are completely coordinated with a CN of 12,
while the surface atoms are coordinately unsaturated with a
CN of 7, 8 or 9 for (110), (100) and (111) facets,
respectively.42 As a result, the CN of the shell atoms will typiꢀ
cally be less than that of the core atoms, since atoms at the
surface have fewer neighbors than those in the core.43 In the
present case, the dave for the three PdAg NP samples were all
approximately 3 nm, meaning that the total CN was consideraꢀ
bly less than 12. The CNtotal (CNPd–Pd + CNPd–Ag) at the Pd Kꢀ
edge for PdAg/TiO2 made with a random alloy was found to
be 7.1, which is lower than the value of 9.5 (CNAg–Ag + CNAg–
Pd) at the Ag Kꢀedge because the Pd concentration was subꢀ
stantially lower than that of Ag (Pd:Ag = 30:70). As expected,
the CNtotal at the Pd Kꢀedge for the Pd@Ag/TiO2 was inꢀ
creased, while the Ag Kꢀedge was decreased, confirming a
PdcoreAgshell structure. The opposite trend was observed in the
case of the Ag@Pd/TiO2 with a AgcorePdshell structure.
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. (A) FTꢀIR spectra of CO chemisorbed on Pd and PdAg
samples. (B) XPS spectra of Pd and PdAg samples. (C) Relationꢀ
ship between the TON for CO2 hydrogenation based on surfaceꢀ
exposed Pd atoms (as determined by CO pulse adsorption) and the
Pd 3d binding energy (as determined by XPS).
The catalytic activities of a series of supported PdAg cataꢀ
lysts having different surface compositions, and of Pd/TiO2,
were assessed during CO2 hydrogenation, with the results proꢀ
vided in Figure 5. Comparing the TON values based on the
total quantities of Pd employed, the Pd@Ag/TiO2 exhibited an
elevated TON (2,496) despite the low density of surface exꢀ
posed Pd atoms. In an effort to elucidate the cause of these
activities, the dispersion of Pd species was determined by
pulsed CO adsorption measurements (Table S2). As a result, a
maximum TON value of 14839 was obtained from the
Pd@Ag/TiO2 based on the quantity of surface Pd atoms,
which represents a more than tenꢀfold increase compared with
the Pd/TiO2. These TON values obtained under relatively mild
conditions are superior to those reported previously for other
heterogeneous catalysts, including Au/Al2O3 (580 at 4 MPa),29
Ru/γꢀAl2O3 (139 at 13.5 MPa),26Pd/mpgꢀC3N4 (85 (24 h) at
100 °C, 4 MPa),30 Ru/LDH (698 (24 h) at 100 °C, 2 MPa),18
We also characterized the surface PdAg NPs by Fourier
transform infrared (FTꢀIR) experiments following CO adsorpꢀ
tion (Figure 4A). In a preliminary DFT calculation, the adꢀ
sorption energies (Ead) of CO adsorbed at a Pd site on the Pd
(111) and PdAg (111) surfaces were determined to be 31.7 and
33.9 kcal/mol, which are noticeably larger than 12.9 and 12.5
kcal/mol at an Ag site on the Ag (111) and PdAg (111) surfacꢀ
es, respectively (Table S1). In addition, the amount of CO in
the pulsed CO adsorption measurement of monometallic
Ag/TiO2 was 0.009 cm3/g, which was ca.0.2% of the theoretiꢀ
cal value, by assuming monodentate type adsorption onto Ag
atoms (Table S2). These results suggest that the CO molecules
are predominantly adsorbed on the Pd atoms rather than the
Ag atoms. Monometallic Pd/TiO2 generated two distinct peaks
assignable to the linear and bridging stretching vibrations of
adsorbed CO at 2076 and 1943 cm–1, respectively.44 The conꢀ
tribution of bridgingꢀtype CO decreased as the Pd/Ag ratio
was lowered, and was completely absent in the case of the
Pd@Ag/TiO2, suggesting the isolation of Pd atoms. Additionꢀ
ally, linearꢀtype CO was predominantly observed for all samꢀ
ples, and the peaks were gradually shifted to lower waveꢀ
numbers with decreases in the Pd/Ag ratio. Thus, the Pd atoms
in the PdAg NPs were evidently electron enriched by charge
transfer from Ag atoms owing to the net difference in ionizaꢀ
tion potential between the two metals (Pd: 8.34 eV, Ag: 7.57
eV). A similar tendency was observed in the XPS data. The Pd
3d peaks generated by the PdAg samples were shifted to lower
binding energies than those of the Pd/TiO2, and this shift was
Au/TiO2 (215 (20 h) at 70 °C,
4
MPa),29 IrꢀPNꢀ
MPa),45
PEI@TNT(Na+) (1012 (20 h) at 140 °C,
2
PdAg/SBAꢀ15ꢀphenylamine (874 (24 h) at 100 °C, 2 MPa)38
and PdAg/amineꢀmodified mesoporous carbon (839 (24 h) at
100 °C, 2 MPa).37 Moreover, a TEM image of the recovered
catalyst confirmed that the particle sizes remained essentially
unchanged without significant agglomeration of the NPs and
the average diameter was determined to be 3.7 nm (Figure
S9A). EXAFS analysis also demonstrated no significant variaꢀ
tions in the alloy structure (Figure S9B-E). Furthermore, the
spent catalyst keeps its original activity even second reaction.
The high catalytic activity of the Pd@Ag/TiO2 can be asꢀ
cribed to the generation of isolated and electronꢀrich Pd atoms
with the aid of the surrounding Ag atoms, since the average
diameters estimated by TEM revealed no significant differꢀ
ences in NP size. There is also evidently a correlation between
the TON based on surface Pd atoms and the Pd 3d5/2 binding
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
energy determined by XPS analysis (Figure 4C). The imꢀ
significantly decreased. These experiments provide firm eviꢀ
dence that the FA originates from the CO2.
1
2
3
4
5
6
7
8
portance of the electronic state of the active center to the hyꢀ
drogenation of CO2 to FA has been reported previously. As an
example, the strong electronꢀdonating ability of PNP pincerꢀ
type ligands or Nꢀheterocyclic carbenes in Ir(III) complexes
are responsible for promoting the reaction.46ꢀ47 In prior work,
our own group also identified a correlation between catalytic
activity and the Ru 3p binding energy for a series of Ruꢀbased
LDHꢀsupported catalysts, in which the TON based on Ru conꢀ
tent increased as the binding energy decreased.18
In the case of a reaction under a flow of H2 and D2 through
the catalyst, the TOF for HD formation was almost independꢀ
ent of the surface composition, as shown in Figure 7A.51 On
the contrary, the effect of the HCO3– concentration was greatly
dependent on the surface composition (Figure 7B). The reacꢀ
tion order decreased as follows; Pd/TiO2 (0.67)
>
Ag@Pd/TiO2 (0.36) > PdAg/TiO2 (0.33) > Pd@Ag/TiO2
(0.22), which is consistent with the TON values based on the
quantity of surface exposed Pd atoms for the CO2 hydrogenaꢀ
tion. These kinetic investigations demonstrate that the elemenꢀ
tary steps involving HCO3– species (step 2 or 3) is the rateꢀ
determining step, rather than the dissociation of H2 (step 1).
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. Comparison of the catalytic activities of a series of
supported PdAg catalysts with different surface compositions and
Pd/TiO2 during CO2 hydrogenation. Conditions: catalyst (10 mg),
1.0 M aqueous NaHCO3 solution (10 mL), H2:CO2 (1:1, total 2.0
MPa), 100 °C.
Considering previous experimental and theoretical studies
regarding the reaction mechanisms of metal NPs,29, 37ꢀ38, 48 we
propose a possible catalytic cycle for CO2 hydrogenation over
supported PdAg NPs (Figure 6A). The reaction is initiated by
the dissociation of H2 to afford a metalꢀhydride species (step
–
1). This is followed by the adsorption of HCO3 generated
under basic conditions (step 2), which undergoes hydrogenaꢀ
tion to give a formate intermediate (step 3). The reduction of
Figure 6 (A) Possible reaction mechanism and (B) potential
energy profiles for CO2 hydrogenation to formic acid.
–
HCO3 is energetically more likely to proceed if active H atꢀ
–
tacks the C atom of HCO3 rather than the O atoms. Finally,
the production of formate accompanied by H2O regenerates
the initial active species (step 4). NaHCO3 and NaOH gave
almost similar activity as an additive, while the use of trimeꢀ
thylamine showed moderate result (Figure S10). The activity
without additives is quite low because of the limited solubility
of CO2 in aqueous solution. Kinetic studies showed almost
zeroꢀorder dependence in CO2 pressure (Figure S11). The FTꢀ
IR spectrum of the Pd@Ag/TiO2, upon treatment in an aqueꢀ
ous NaHCO3 solution at 1.0 MPa of CO2 showed the peak due
To better understand the positive effect of alloying with Ag,
potential energy profiles were obtained using DFT calculaꢀ
tions, employing Pd22, Pd11Ag11 and Pd6Ag16 clusters as modꢀ
els for monometallic Pd and alloy NPs (Figure 6B). In the
case of Pd22, the dissociation of H2 (step 1) occurs at Pd sites
via TSI/II with a barrier of 13.9 kcal/mol. Following this,
–
HCO3 is adsorbed on the Pd to produce intermediate III (step
2), followed by reduction through the attack of a hydride via
TSIII/IV, with a barrier of 77.4 kcal/mol (step 3). The energy
barrier to step 4, in which formate is spontaneously generated
ꢀ
to ν (CO3) band of the adsorbed HCO3 was observed at
around 1454 cmꢀ1 (Figure S12).49 These results suggest that
–
along with H2O when the OH of HCO3 is attacked by another
the present catalytic system mainly proceeds via the reduction
–
hydride, is negligibly small. These results show that the rateꢀ
determining step in the present catalytic cycle is step 3. In the
cases of the Pd11Ag11 and Pd6Ag16 clusters, the dissociation of
H2 occurs with a barrier of 11.9 and 11.0 kcal/mol, respectiveꢀ
ly. These activation energies are similar to that obtained with
the Pd22, but differ from the 39.0 kcal/mol calculated for an
Ag22 cluster model. These results agree with the kinetic data
obtained from the HD formation reaction, and show that alloyꢀ
ing with Ag does not affect the H2 dissociation step. In sharp
of HCO3 transformed from gaseous CO2 under basic aqueous
solutions, while the involvement of the direct CO2 activation is
the minor pathway. The origin of the FA during this hydroꢀ
genation process was investigated using 13CO2 in D2O as a
solvent (with 0.5 M NaOH) followed by analysis of the prodꢀ
ucts by 13C NMR spectroscopy. Deprotonated H13COO– genꢀ
erated a sole peak at 170. 8 ppm without the formation of byꢀ
ꢀ
products (Figure S13).50 Additionally, HCO3 was used as a
sole carbon source without gaseous CO2, the attained TON
–
contrast, the activation energy for the reduction of HCO3 via
ACS Paragon Plus Environment

Journal of the American Chemical Society
TSIII/IV was decreased with increases in the proportion of Ag
Page 6 of 9
participate in the dissociation of H2. Thus, these charges deꢀ
crease in the order of –0.115 (Pd22) > –0.168 (Pd11Ag11) > –
0.216 (Pd6Ag16). These results are supported experimentally
by the FTꢀIR spectra following CO adsorption as well as the
XPS analysis. Such changes accordingly decrease the electroꢀ
negativity of the dissociated hydride species on the Pd atoms.
In contrast, the electronic charges of the C atoms of the adꢀ
1
2
3
4
5
6
7
8
atoms, with calculated values of 58.7 and 46.2 kcal/mol for
Pd11Ag11 and Pd6Ag16, respectively. These results are also in
good agreement with the kinetic investigation into the effect of
–
HCO3 concentration, and further demonstrate that the PdAg
alloy NPs with a low Pd/Ag ratio plays a pivotal role in boostꢀ
ing the rateꢀdetermining step. In a preliminary calculation, the
energy barrier in the dissociation of H2 on Ag atom of Pd6Ag16
cluster was determined to be 38.9 kcal/mol, which is almost
similar to that on the Ag22 cluster, but is substantially larger
than those on the Pd atoms of Pd22, Pd11Ag11, and Pd6Ag16,
respectively. This result definitely excludes the participation
of Ag atoms on the catalytic cycle as an active center even
after the formation of Pd@Ag coreꢀshell structure.
–
sorbed HCO3 are almost constant for all cluster models, and
maintain their positive charges. Therefore, electronically proꢀ
moted hydride species on the Pd6Ag16 will tend to attack the C
atoms of the adsorbed HCO3 , whereas electronic repulsion
between the less negative hydride species on the Pd22 and the
positively charged C atoms results in a greater activation enerꢀ
gy. It can be concluded that the electronic ligand effect resultꢀ
ing from the interplay of the neighboring Ag atoms explains
the enhanced activity for CO2 hydrogenation.
–
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
Conclusion
In summary, we have elucidated the effects of highly active
Pd sites surrounded by Ag atoms during CO2 hydrogenation to
FA. By tuning the surfaceꢀexposed Pd atoms in the alloy NPs,
the optimized Pd atoms in Pd@Ag/TiO2 performed as an effiꢀ
cient catalyst even under low pressure conditions. Excellent
selectivity was obtained owing to the electronic ligand effect
caused by the Ag atoms, with activity greater than that of
monometallic Pd/TiO2 by a factor of more than ten. The enꢀ
hanced electronegativity was found to facilitate the rate deterꢀ
–
mining reduction step of the adsorbed HCO3 species, as eviꢀ
Figure 7. (A) Comparison of activity during the HD exchange
reaction using Pd and PdAg catalysts. (B) Effect of HCO3 conꢀ
centration on TOF values using Pd and PdAg catalysts
denced by kinetic and DFT calculations. This study provides
advanced insights into the architecture of catalytically active
sites for CO2 hydrogenation to FA. This work also emphasizes
the importance of the surface engineering of alloy NPs with
atomic precision for optimization of the target reaction. Furꢀ
ther improvement of the catalytic performance by the enꢀ
hancement of CO2 adsorption properties may lead to the deꢀ
velopment of environmentallyꢀbenign CO2ꢀmediated hydrogen
storage/release systems.
ꢀ
Experimental Section
Materials: TiO2 (JRCꢀTiO6) was kindly supplied by the
Catalysis Society of Japan. [Pd(NH3)4]Cl2 was obtained from
the Aldrich Chemical Co., AgNO3 was purchased from
Nacalai Tesque and NaHCO3 was purchased from Wako Pure
Chemical Industries, Ltd. All commerciallyꢀavailable comꢀ
pounds were used asꢀreceived. Distilled water was employed
as the reaction solvent.
Figure 8. Representative Mulliken atomic charges as determined
by DFT calculations in the reaction intermediate C for (A) Pd22,
(B) Pd11Ag11 and (C) Pd6Ag11 clusters.
As discussed above, the electronic state of the active Pd atꢀ
oms is important to the hydrogenation of CO2, and there is a
strong correlation between the TON based on surface Pd atꢀ
oms and the extent to which the Pd atoms are electron rich.
The significant enhancement of the rateꢀdetermining step by
the PdAg alloy NPs having a low Pd/Ag ratio can be explained
by considering the electronic state in reaction intermediate III.
Figure 8 depicts the highest occupied molecular orbital
(HOMO) level of reaction intermediate III and the representaꢀ
tive Mulliken atomic charges of selected atoms as determined
by DFT calculations for Pd22, Pd11Ag11 and Pd6Ag16 clusters.
The HOMO can be tuned through coupling of the electronic
structure as the NP composition is varied, and is elevated
(from –1.65 to –1.51 eV) when Pd atoms are surrounded by
more Ag atoms. This higher HOMO level would be expected
to increase the extent to which the Pd atoms are electron rich
and so to decrease the electronic charge of the Pd atoms that
Catalyst synthesis: TiO2 (0.5 g) was dispersed in 100 mL
of an aqueous solution containing [Pd(NH3)4]Cl2 (0.047 mmol)
and AgNO3 (0.047 mmol), followed by stirring at room temꢀ
perature for 1 h. The suspension was evaporated under vacuꢀ
um and the obtained powder was dried overnight. Subsequentꢀ
ly, the sample was reduced with NaBH4 (0.28 mmol) and
washed with distilled water several times to yield PdAg/TiO2
(Pd: 1.0 wt%, Pd/Ag molar ratio = 1/1). Using the same coꢀ
reduction procedure, the support, the second metal in the alloy
and the Pd/Ag molar ratio were all changed. During the synꢀ
thesis of the Pd@Ag/TiO2 catalyst (Pd: 1.0 wt%, Pd/Ag molar
ratio = 3/7), TiO2 was first impregnated with Pd(NH3)4]Cl2,
followed by reduction using NaBH4, after which AgNO3 was
added, followed by
a second reduction with NaBH4.
Ag@Pd/TiO2 was also synthesized by the same successive
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
reduction method but in the opposite order. Analysis by inducꢀ
corresponding bulk position and the top layer allowed to relax
during geometry optimizations.
1
2
3
tively couple plasma atomic emission analysis (ICPꢀAES)
clearly indicated that the desired amounts of metal species
were successfully loaded onto each of the catalysts.
4
5
6
7
8
9
ASSOCIATED CONTENT
General procedure for the catalytic hydrogenation of
CO2: CO2 hydrogenation to FA was conducted with a batch
reactor system using a stainless steel autoclave (60 mL). In
each trial, the catalyst (10 mg) and a 1.0 M aqueous NaHCO3
solution (10 mL) were transferred into the reactor and the
pressure was set to 2.0 MPa by adding H2:CO2 (1:1). The sysꢀ
tem was subsequently heated to 100 °C and stirred for 24 h.
FA yields were determined by high performance liquid chroꢀ
matography (HPLC), using a Shimadzu instrument equipped
with a BioꢀRad Aminex HPXꢀ87H ion exclusion column (300
× 7.8 mm) and 5 mM H2SO4 (0.5 mL/min) as the mobile
phase, at 40 °C. TON values were determined by dividing the
quantity of FA produced after 24 h by the moles of either total
Pd or surfaceꢀexposed Pd.
Supporting Information is available free of charge via the Internet
at http://pubs.acs.org. TEM, CO adsorption, XANES, 13CNMR.
AUTHOR INFORMATION
Corresponding Author
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
*K. M. mori@mat.eng.osakaꢀu.ac.jp
*H. H. yamashita@mat.eng.osakaꢀu.ac.jp
ACKNOWLEDGMENT
The present work was supported by JSTꢀPRESTO
(JPMJPR1544). The part of this work was supported by Grantsꢀ
inꢀAid for Scientific Research (Nos. 18H020740 and 26220911)
from the Japan Society for the Promotion of Science (JSPS) and
MEXT.
Characterization: Powder Xꢀray diffraction (XRD) patꢀ
terns were recorded using a Rigaku Ultima IV diffractometer
with Cu Kα radiation (λ=1.54056 Å). TEM images were obꢀ
tained with a Hitachi HFꢀ2000 FEꢀTEM instrument operating
at 200 kV. Nitrogen adsorptionꢀdesorption isotherms were
acquired at ꢀ196 °C using a BELSORPꢀmax system (Microꢀ
tracBEL Corp.). Samples were degassed at 150 °C for 3 h
under vacuum to vaporize physisorbed water prior to each
trial. Specific surface areas were calculated by the Brunauerꢀ
EmmettꢀTeller (BET) method using nitrogen adsorption data
over the range from p/p0 = 0.05 to 0.35. XPS was performed
with a Shimadzu ESCAꢀ3400 system, using Mg Kα radiation
(hν = 1253.6 eV) as the excitation source. The binding energy
values of the spectra were calibrated using the contaminant C
1s core level at 284.5 eV. Metal concentrations in the samples
were determined by ICPꢀAES analysis with a Nippon Jarrellꢀ
Ash ICAPꢀ575 Mark II instrument. STEM images and eleꢀ
mental maps were obtained using a JEOLꢀARM 200F appaꢀ
ratus equipped with a Kvex energyꢀdispersive Xꢀray detector
(JEDꢀ2300T) operating at 200 kV. CO pulse adsorption was
performed by using a BELꢀMETALꢀ1 instrument (BEL Japan,
Inc.) to measure the amount of surface exposed Pd. The samꢀ
ples were pretreated under a helium flow at 323 K for 15 min,
and subsequently under a H2 flow at 323 K for 30 min. The
CO adsorption was measured at 323 K at a CO flow rate of 20
cm3 min–1. Pd and Ag Kꢀedge XAFS spectra were recorded
using a fluorescence yield collection technique at the 01B1
beamline station at the SPringꢀ8 facility, JASRI, Harima, Jaꢀ
pan (proposal nos. 2017B1081 and 2017B1084), with a
Si(111) monochromator. Data reduction was performed using
the REX2000 software program (Rigaku). Fourier transforꢀ
mation of the k3ꢀweighted normalized EXAFS data was carꢀ
ried over the range of 3.0 Å < k/Åꢀ1 < 12 Å to obtain radial
structure functions, and backscattering amplitudes and phase
shift parameters for curve fitting analyses were calculated with
the FEFF8.40 code.
REFERENCES
1.
Wang, W.; Wang, S.; Ma, X.; Gong, J., Chem. Soc. Rev.
2011, 40, 3703ꢀ3727.
2.
Kondratenko, E. V.; Mul, G.; Baltrusaitis, J.;
Larrazabal, G. O.; PerezꢀRamirez, J., Energy Environ. Sci. 2013,
6, 3112ꢀ3135.
3.
Álvarez, A.; Bansode, A.; Urakawa, A.; Bavykina, A.
V.; Wezendonk, T. A.; Makkee, M.; Gascon, J.; Kapteijn, F.,
Chem. Rev. 2017, 117, 9804ꢀ9838.
4.
Jessop, P. G.; Ikariya, T.; Noyori, R., Chem. Rev. 1995,
95, 259ꢀ272.
5.
Gunasekar, G. H.; Park, K.; Jung, K.ꢀD.; Yoon, S.,
Inorg. Chem. Front. 2016, 3, 882ꢀ895.
6.
Reutemann, W.; Kieczka, H., Formic Acid. In
Ullmann's Encyclopedia of Industrial Chemistry, WileyꢀVCH
Verlag GmbH & Co. KGaA: 2000.
7.
Singh, A. K.; Singh, S.; Kumar, A., Catal. Sci. Tech.
2016, 6, 12ꢀ40.
8.
Mellmann, D.; Sponholz, P.; Junge, H.; Beller, M.,
Chem. Soc. Rev. 2016, 45, 3954ꢀ3988.
9.
Eppinger, J.; Huang, K.ꢀW., ACS Energy Lett. 2017, 2,
188ꢀ195.
10.
Grasemann, M.; Laurenczy, G., Energy Environ. Sci.
2012, 5, 8171ꢀ8181.
11.
12.
Joó, F., ChemSusChem 2008, 1, 805ꢀ808.
Mori, K.; Dojo, M.; Yamashita, H., ACS Catal. 2013, 3,
1114ꢀ1119.
13.
Mori, K.; Tanaka, H.; Dojo, M.; Yoshizawa, K.;
Yamashita, H., Chem. Eur. J. 2015, 21, 12085ꢀ12092.
14.
Yamashita, H., ChemCatChem 2017, 9, 3456ꢀ3462.
15. Uhm, S.; Lee, H. J.; Lee, J., Phys. Chem. Chem. Phys.
2009, 11, 9326ꢀ9336.
Mori, K.; Naka, K.; Masuda, S.; Miyawaki, K.;
DFT calculations. All DFT calculations were performed
with the DMol3 program in the Materials Studio 17.2 software
package.52ꢀ53 The generalized gradient approximation (GGA)
exchangeꢀcorrelation functional proposed by Perdew, Burke
and Ernzerhof (PBE) was combined with the double numerical
basis set plus polarization functions (DNP). Pd22, Pd11Ag11 and
Pd6Ag16 clusters were chosen as models of monometallic Pd
and PdAg alloy NPs, with the lower two layers fixed at the
16.
Wang, W.; Niu, M.; Hou, Y.; Wu, W.; Liu, Z.; Liu, Q.;
Ren, S.; Marsh, K. N., Green Chem. 2014, 16, 2614ꢀ2618.
17.
Liu, Z., Green Chem. 2014, 16, 4931ꢀ4935.
18. Mori, K.; Taga, T.; Yamashita, H., ACS Catal. 2017, 7,
3147ꢀ3151.
Xu, J.; Zhang, H.; Zhao, Y.; Yang, Z.; Yu, B.; Xu, H.;
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
19.
Moret, S.; Dyson, P. J.; Laurenczy, G., Nat. Commun.
36.
An, B.; Zeng, L.; Jia, M.; Li, Z.; Lin, Z.; Song, Y.;
2014, 5, 4017.
Zhou, Y.; Cheng, J.; Wang, C.; Lin, W., J. Am. Chem. Soc. 2017,
139, 17747ꢀ17750.
1
2
3
4
5
6
20.
Enthaler, S.; von Langermann, J.; Schmidt, T., Energy
Environ. Sci. 2010, 3, 1207ꢀ1217.
37.
Masuda, S.; Mori, K.; Futamura, Y.; Yamashita, H.,
21.
Azua, A.; Sanz, S.; Peris, E., Chem. Eur. J. 2011, 17,
ACS Catalysis 2018, 8, 2277ꢀ2285.
3963ꢀ3967.
38.
Mori, K.; Masuda, S.; Tanaka, H.; Yoshizawa, K.; Che,
22.
Schmeier, T. J.; Dobereiner, G. E.; Crabtree, R. H.;
M.; Yamashita, H., Chem. Commun. 2017, 53, 4677ꢀ4680.
Hazari, N., J. Am. Chem. Soc. 2011, 133, 9274ꢀ9277.
39.
Liao, F.; Lo, T. W. B.; Tsang, S. C. E., ChemCatChem
7
8
9
23.
Himeda, Y., Energy Environ. Sci. 2012, 5, 7923ꢀ7926.
24. Wiener, H.; Blum, J.; Feilchenfeld, H.; Sasson, Y.;
Zalmanov, N., J. Catal. 1988, 110, 184ꢀ190.
25. Zhang, Y.; Fei, J.; Yu, Y.; Zheng, X., Catal. Commun.
2004, 5, 643ꢀ646.
Wang, W.ꢀH.; Hull, J. F.; Muckerman, J. T.; Fujita, E.;
2015, 7, 1998ꢀ2014.
40.
Zhang, S.; Metin, Ö.; Su, D.; Sun, S., Angew. Chem. Int.
Ed. 2013, 52, 3681ꢀ3684.
41.
Mori, K.; Kondo, Y.; Yamashita, H., Phys. Chem.
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
Chem. Phys. 2009, 11, 8949ꢀ8954.
42.
Mori, K.; Hara, T.; Mizugaki, T.; Ebitani, K.; Kaneda,
26.
Hao, C.; Wang, S.; Li, M.; Kang, L.; Ma, X., Catal.
K., J. Am. Chem. Soc. 2004, 126, 10657ꢀ10666.
Today 2011, 160, 184ꢀ190.
27.
Angew. Chem. Int. Ed. 2011, 50, 12551ꢀ12554.
43.
44.
Frenkel, A. I., Chem. Soc. Rev. 2012, 41, 8163ꢀ8178.
Rainer, D. R.; Wu, M. C.; Mahon, D. I.; Goodman, D.
Preti, D.; Resta, C.; Squarcialupi, S.; Fachinetti, G.,
W., J. Vac. Sci. Technol., A 1996, 14, 1184ꢀ1188.
45. Kuwahara, Y.; Fujie, Y.; Yamashita, H., ChemCatChem
2017, 9, 1867ꢀ1867.
28.
W. F.; Hicks, J. C., ChemCatChem 2013, 5, 1769ꢀ1771.
29. Filonenko, G. A.; Vrijburg, W. L.; Hensen, E. J. M.;
Pidko, E. A., J. Catal. 2016, 343, 97ꢀ105.
30. Lee, J. H.; Ryu, J.; Kim, J. Y.; Nam, S.ꢀW.; Han, J. H.;
Lim, T.ꢀH.; Gautam, S.; Chae, K. H.; Yoon, C. W., J. Mater.
Chem. A 2014, 2, 9490ꢀ9495.
31.
Wang, X.; Huang, Y.; Zhang, T., Nat. Commun. 2017, 8, 1407.
32. Nguyen, L. T. M.; Park, H.; Banu, M.; Kim, J. Y.;
Youn, D. H.; Magesh, G.; Kim, W. Y.; Lee, J. S., RSC Adv. 2015,
5, 105560ꢀ105566.
33.
3589ꢀ3614.
34.
Xu, Z.; McNamara, N. D.; Neumann, G. T.; Schneider,
46.
Tanaka, R.; Yamashita, M.; Nozaki, K., J. Am. Chem.
Soc. 2009, 131, 14168ꢀ14169.
47.
Jantke, D.; Pardatscher, L.; Drees, M.; Cokoja, M.;
Herrmann, W. A.; Kühn, F. E., ChemSusChem 2016, 9, 2849ꢀ
2854.
Liu, Q.; Yang, X.; Li, L.; Miao, S.; Li, Y.; Li, Y.;
48.
He, C.ꢀS.; Gong, L.; Zhang, J.; He, P.ꢀP.; Mu, Y., J.
CO2 Utilization 2017, 19, 157ꢀ164.
49.
Collins, S. E.; Baltanás, M. A.; Bonivardi, A. L., J.
Catal. 2004, 226, 410ꢀ421.
50.
Lee, Y.; Kim, S.; Kang, J. K.; Cohen, S. M., Chem.
Wight, A. P.; Davis, M. E., Chem. Rev. 2002, 102,
Commun. 2015, 51, 5735ꢀ5738.
51.
Bai, L.; Wang, X.; Chen, Q.; Ye, Y.; Zheng, H.; Guo, J.;
De Vos, D. E.; Dams, M.; Sels, B. F.; Jacobs, P. A.,
Yin, Y.; Gao, C., Angew. Chem. Int. Ed. 2016, 55, 15656ꢀ15661.
Chem. Rev. 2002, 102, 3615ꢀ3640.
35.
Wezendonk, T.; Seoane, B.; Kapteijn, F.; Makkee, M.; Gascon, J.,
ChemCatChem 2016, 8, 2217ꢀ2221.
52.
53.
Delley, B., J. Chem. Phys. 1990, 92, 508ꢀ517.
Delley, B., J. Chem. Phys. 2000, 113, 7756ꢀ7764.
Bavykina, A. V.; Rozhko, E.; Goesten, M. G.;
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
Insert Table of Contents artwork here
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
9
ACS Paragon Plus Environment