Controllable Conversion of CO2 on Non-Metallic Gold Clusters

Article abstract of DOI:10.1002/anie.201913635

Gold nanoparticles in metallic or plasmonic state have been widely used to catalyze homogeneous and heterogeneous reactions. However, the catalytic behavior of gold catalysts in non-metallic or excitonic state remain elusive. Atomically precise Au_n clusters (n=number of gold atoms) bridge the gap between non-metallic and metallic catalysts and offer new opportunities for unveiling the hidden properties of gold catalysts in the metallic, transition regime, and non-metallic states. Here, we report the controllable conversion of CO₂ over three non-metallic Au_n clusters, including Au₉, Au₁₁, and Au₃₆, towards different target products: methane produced on Au₉, ethanol on Au₁₁, and formic acid on Au₃₆. Structural information encoded in the non-metallic clusters permits a precise correlation of atomic structure with catalytic properties and hence, provides molecular-level insight into distinct reaction channels of CO₂ hydrogenation over the three non-metallic Au catalysts.

Full text of DOI:10.1002/anie.201913635

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Controllable Conversion of CO2 on Non-metallic Gold Clusters
Authors: Dan Yang, Wei Pei, Si Zhou, Jijun Zhao, Weiping Ding, and
Yan Zhu
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201913635
Angew. Chem. 10.1002/ange.201913635
Link to VoR: http://dx.doi.org/10.1002/anie.201913635
http://dx.doi.org/10.1002/ange.201913635

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
Controllable Conversion of CO2 on Non-metallic Gold Clusters
Dan Yang,[a]# Wei Pei,[b]# Si Zhou,*[b] Jijun Zhao,[b] Weiping Ding,[a] and Yan Zhu*[a]
Dedication to the 100th anniversary of the School of Chemistry and Chemical Engineering, Nanjing University
Abstract: Gold nanoparticles in metallic or plasmonic state have been
widely used to catalyze the homogeneous and heterogeneous
reactions. However, the catalytic behaviours of gold catalysts in non-
metallic or excitonic state remain elusive. Atomically precise Aun
clusters (n = number of gold atoms) bridge the gap between non-
metallic and metallic catalysts and offer new opportunities for
unveiling the hidden properties of gold catalysts in the metallic,
transition regime, and non-metallic states, respectively. Here we
report the controllable conversion of CO2 over three non-metallic Aun
clusters including Au9, Au11, and Au36, towards different target
products: methane produced on Au9, ethanol on Au11, and formic acid
on Au36. Structural information encoded in the non-metallic clusters
permits a precise correlation of atomic structure with catalytic
properties and hence provides molecular-level insight into distinct
reaction channels of CO2 hydrogenation over the three non-metallic
Au catalysts.
bacteria to enable photosynthesis of acetic acid from CO2.[18]
These above studies are typically impossible for the traditional
metallic catalysts.
The catalytic hydrogenation of CO2 is selected as a target
reaction in this work, as an effective conversion of CO2 into liquid
fuels and chemicals is of paramount importance for mitigating the
global warming and energy supply problems.[19-27] Especially,
recent successes in the transformation of CO2 into C2 products
have added additional spotlights into this area, mainly because C2
compounds, relative to C1 species, are really high-value-added
products and can be used as entry platform chemicals for existing
value.[28] The controllable reduction of CO2 with renewable
hydrogen into C1 or C2 products is thus highly desirable, which
may become true by modulating the competition of C-C bond
formation with the C-H and C-O bond formation. Several reports
have shown that CO2 can be converted into C2 products by both
homogeneous and heterogeneous catalysis.[29-34] However, it
remains challenging to control the rate-determining route or
intermediates from CO2 conversion that determine different
pathways purposefully towards C1 or C2 products.
Introduction
Herein we utilize the non-metallic gold clusters such as
[Au9(PPh3)8](NO3)3 (denoted as Au9), [Au11(PPh3)8Cl2]Cl (denoted
as Au11), and Au36(TBBT)24 (denoted as Au36; TBBT = 4-tert-
butylbenzenethiol) as catalysts to catalyze the hydrogenation
reaction of CO2, and investigate how the atomicity of gold clusters
dictates the reaction channels of CO2 hydrogenation selectively
towards C1 or C2 products, which shows totally different
performances from the plasmonic/metallic gold nanoparticle
catalysts.
The atomically precise metal clusters ideally composed of an
exact number of metal atoms recently have attracted significant
interests due to the breakthrough in determination of total
structures (surface plus core) and distinctive electronic
structures.[1-6] Different from metallic catalysts with the collective
excitation behavior, the atomically precise metal clusters can be
classified into three distinct states: metallic, transition regime, and
non-metallic.[7] Non-metallic or molecular-like clusters possess
discrete electron energy levels.[8] With the evolution of electronic
structures from metallic to non-metallic state, these clusters can
reveal a hidden mystery in the catalytic field, that is, the non-
metallic cluster catalysts may have similar or even superior
catalytic abilities compared to the metallic ones. More than that,
they are opening up a new era for catalytic science, e.g. these
clusters provide exciting opportunities to map out the explicit
structure-property relation and allow for identification of the active
sites with atomic precision.[9-15] Specific examples are that they
can be used to explore how the catalytic performance can be
tailored by only one atom alteration[16,17] and also can circumvent
the sluggish kinetics of electron transfer for non-photosynthetic
Results and Discussion
Electrospray ionization mass (ESI-MS) spectra showed true
mono-dispersity of the three clusters, confirming the exact
formulas: [Au9(PPh3)8]3+, [Au11(PPh3)8Cl2]+, and Au36(TBBT)24. As
shown in Figure 1a, the 3208.12, 3470.44 and 3732.71 Da peaks
- +
- +
were assigned to [Au9(PPh3)53++2NO3 ] , [Au9(PPh3)63++2NO3 ] ,
and [Au9(PPh3)73++2NO3 ] , respectively. In Figure 1b, only one
- +
sharp peak at 4336.3 Da was observed, verifying the formula of
[Au11(PPh3)8Cl2]+. The m/z 5528.77 peak was 2+ charged (Figure
1c), as revealed by the isotope peak spacing, which corresponded
to [Au36(TBBT)24-2e]2+. A small peak at 5446.16 Da arose from a
fragment of [Au36(TBBT)23-2e]2+.
The three gold clusters have individually precise atom-
packing structures. The total structure of [Au9(PPh3)8]3+ resembles
a butterfly with eight gold atoms surrounding the central gold atom
and the peripheral gold atoms are bonded to PPh3 (Figure 1d).[35]
[Au11(PPh3)8Cl2]+ can be viewed as an incomplete icosahedron
with an approximate C3v symmetry and the ten surface gold atoms
are linked to eight PPh3 and two halogen atoms (Figure 1e).[36]
Au36(TBBT)24 has a 28-gold-atom kernel with a truncated fcc
[a]
[b]
D. Yang, Prof. W. Ding, Prof. Y. Zhu
Key Lab of Mesoscopic Chemistry, School of Chemistry and
Chemical Engineering, Nanjing University, Nanjing 210093, China
E-mail: zhuyan@nju.edu.cn
W. Pei, Dr. S. Zhou, Prof. J. Zhao
Key Laboratory of Materials Modification by Laser, Ion and Electron
Beams, Dalian University of Technology, Dalian 116024, China
E-mail: sizhou@dlut.edu.cn
# These authors contributed equally.
Supporting information for this article is given via a link at the end of the
document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
(face-centered cubic) tetrahedron exposing 4 of {111} facets and
12 of {100} facets, where each {111} facet is protected by one
dimeric staple, and each {100} square is capped by one bridging
thiolate (Figure 1f).[37]
X-ray absorption near edge structure (XANES) experiments
were carried out to monitor the charge states of the gold clusters.
As shown in Figure 3a, the white line intensities of the Au9, Au11,
and Au36 clusters were higher than that of Au foil, which were
generally caused by the ligands interaction with surface Au
atoms.[38] The Au L3-edge white line results from electronic
transition from occupied 2p to unoccupied 5d states, where higher
5d electronic density is due to more metal-ligand charge
transfer.[39] By the aids of liner combination fits, the charge states
of the three gold clusters were positive: the average charge state
of Au9 was the highest, that of Au36 was the lowest, and that of
Au11 was in between. Furthermore, local coordination
environments in the three clusters were probed by the extended
X-ray absorption fine structure (EXAFS) studies. The Au-X (Au-P
for Au9, Au-P/Cl for Au11, and Au-S for Au36) bonds and the Au-
Au bonds can be observed for each gold cluster in Figure 3b and
Table S1, indicating that the intensities of Au-Au bonds in the
three clusters were shown in the descending order: Au36 > Au11
Au9.
>
Figure 1. ESI-MS profiles of (a) [Au9(PPh3)8]3+, (b) [Au11(PPh3)8Cl2]+, and (c)
Au36(TBBT)24. Atomic structures of (d) [Au9(PPh3)8]3+, (e) [Au11(PPh3)8Cl2]+, and
(f) Au36(TBBT)24. Color codes: green = Au, pink = P, yellow = S; orange = Cl.
The non-metallic properties of the three gold clusters are
indeed reflected in the discrete electron energy levels and multiple
absorption bands in optical spectra. UV-vis absorption spectrum
of the [Au9(PPh3)8]3+ cluster showed two prominent peaks at 313
and 442 nm and two weak peaks at 348 and 378 nm (Figure 2a).
The [Au11(PPh3)8Cl2]+ cluster showed the optical absorption
bands centered at 312, 375 and 415 nm (Figure 2b). The
Au36(TBBT)24 cluster exhibited two obvious absorption bands at
374 and 580 nm and one broad band at 426 nm (Figure 2c). From
DFT (density functional theory) calculations, the Kohn−Sham (KS)
molecular orbital (MO) energy levels and atomic orbital
components in each KS MO of the three clusters indicate that the
absorption peaks in the three clusters mainly involve the Au(sp)
→ Au(sp) transitions (Figure 2d-f). Accordingly the optical energy
gaps were determined to be 1.49, 1.73 and 1.84 eV between the
highest occupied molecular orbital (HOMO) and lowest
unoccupied molecular orbital (LUMO) by our DFT calculations.
Figure 3. (a) XANES profiles and (b) EXAFS data with fits of [Au9(PPh3)8]3+
,
[Au11(PPh3)8Cl2]+, Au36(TBBT)24 and Au foil.
We next turned our attention to the catalytic properties of the
non-metallic gold clusters. The three gold clusters were exploited
as heterogeneous catalysts to explore if the non-metallic gold can
be indeed operative in determining the reaction pathways for the
hydrogenation of CO2. Regardless of the supports being SiO2,
TiO2, CeO2, ZnO, Al2O3 or ZrO2, a similar trend was found (Figure
S1 and Table S2; the supplemental discussion of the effect of
supports on the catalytic conversion of CO2 can be found in
Supporting Information). The Au9 catalysts gave rise to the
selectivity of methane with ~90% and other products like
methanol, ethanol and formic acid et al as minor products. The
Au36 catalysts can give the high selectivity (~85%) for formic acid
with a small amount of methanol, methane and ethanol et al.
Interestingly, the Au11 catalysts can improve C-C coupling so as
to form C2 chemical (ethanol) with >80% selectivity. Furthermore,
the temperature effect on the catalytic performances of the three
cluster catalysts for CO2 hydrogenation was shown in Figure 4a.
The activities in terms of TOF (turnover frequency) of the three
catalysts were increased with the increasing temperature (note
that the atomic structures of the clusters can be destroyed at
higher temperatures). The results in Figure 4b revealed that the
increase of reaction pressure was also beneficial to the
conversion of CO2, but the selectivity of products was irrelevant
to the reaction pressure. Catalytic behaviors of the three clusters
were further studied by exploring the activity and product
selectivity as a function of the reaction time. The conversion of
CO2 (where TON is turnover number) increased with the reaction
Figure 2. UV-vis spectra of (a) [Au9(PPh3)8]3+, (b) [Au11(PPh3)8Cl2]+ and (c)
Au36(TBBT)24. Kohn−Sham orbitals of (d) Au9(PPh3)8, (e) Au11(PPh3)8Cl2 and (f)
Au36(SCH3)24 from DFT calculations and the corresponding HOMO and LUMO
charge density distributions. The H, C, O, S and Au atoms are shown in white,
gray, red, brown and yellow colors, respectively.
This article is protected by copyright. All rights reserved.

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
time, and the selectivity of products almost remained unchanged
with the reaction time (Figure 4c). Moreover, the distinct selectivity
towards C1 or C2 products were also embodied in the recyclability
of the three catalysts (Figure S2). Transmission electron
microscopy studies showed that the clusters had no drastic
aggregation throughout the reactions (Figure S3). As much, the
typical UV-vis spectra of the spent catalysts showed that a certain
quantity of the clusters was still intact after reactions (Figure S4).
approach this goal, the changes of ligands in the three clusters
under reaction conditions were detected by in-situ diffuse
reflectance infrared Fourier transform spectroscopy (DRIFTS). In
comparison with the FT-IR features of TBBT and PPh3 ligands
(Figure S9), the IR features marked with purple rectangular
frames in Figure S10 and S11 were assigned to PPh3, and the
characteristic IR bands in Figure S12 were assigned to TBBT.
Clearly, these bands gradually decreased in intensity with the
treatment time, and finally held steady. It indicates that the ligands
could be partially removed under reaction conditions and hence
the gold sites of the clusters can be exposed and activate
reagents.
Figure 4. Catalytic performances of CO2 hydrogenation over the Au9, Au11 and
Au36 supported on ZrO2 under different conditions: (a) reaction temperature, (b)
reaction pressure, and (c) reaction time. (d) Catalytic results of CO2
hydrogenation over the three gold clusters pretreated to remove all ligands.
Reaction conditions: catalyst (100 mg, ～ 1% Au), H2O (15 mL), 3 MPa
(CO2+3H2), 120 oC, 10 hr. Stars denote as TOF or TON. Error bars correspond
to the deviations from three independent experiments.
Notably, supports had no activity (Table S3, entries 1-5).
Considering the clusters themselves containing carbon sources
(ligands), additional experiments were performed: either CO2 or
H2 was introduced and no product was detected (Table S3,
entries 6-11), which ruled out the possibility of the products from
ligands. A question naturally is whether the ligands can have an
influence on the catalytic activity and selectivity. To address this,
the clusters protected by different ligands (Figure S5) were
synthesized to catalyze the hydrogenation of CO2. As shown in
Figure S6, methane was still the major product over the Au9
protected by diphenyl-methylphosphine (MePPh2); formic acid
was the major product over the Au36 protected by p-isopropyl
benzenethiol (TPBT); ethanol was the major product over the Au11
capped by 2-diphenylphosphino pyridine (PyPPh2). These
findings support that the catalytic selectivity is mainly determined
by the gold sites rather than the surface ligands. Of note, when
the ligands were completely removed by the thermal treatment
(Figure S7), the clusters became gold nanoparticles with metallic
state (namely nanocrystals) (Figure S8). The metallic Au
nanoparticles with 5-10 nm size had very low activity with
methane as major product (Figure 4d), implying that underlying
reaction path has been altered.
Figure 5. Time-resolved in-situ IR spectra of the reaction intermediates formed
on the (a) Au9, (b) Au11 and (c) Au36 supported on ZrO2 for the hydrogenation of
CO2 at 120 °C and 0.8 MPa (CO2+3H2).
The three gold clusters are capable of driving reaction
pathways of CO2 conversion towards different products, which
implies the different reaction mechanisms happened on the three
catalysts. Therefore, the reaction intermediates occurred in the
three catalyst systems were monitored by time-resolved in-situ
FT-IR spectroscopy. As shown in Figure 5a, a series of bands
were observed: the band at 1457 cm-1 assigned to OCO*;[31,39,40]
the bands at 1630, 1868 and 2072 cm-1 assigned to HCO*, OH*
and CO* species.[41,42] Thus, the reaction mechanism of CO2
hydrogenation on the Au9 catalyst was proposed: methane was
produced by H-assisted pathway possibly via formyl (HCO*) and
formaldehyde (H2CO*, 2921 cm-1) species to form surface CHx. In
terms of the Au11 reaction system (Figure 5b), two bands at 1457
Considering that the three clusters were protected by organic
ligands, the active site-blocking impact of ligands might lead to no
gold sites available for adsorption and conversion of reagents.
How can the three clusters catalyze the conversion of CO2? To
This article is protected by copyright. All rights reserved.

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
and 1396 cm-1 were assigned to the adsorption OCO*
species,[31,39,40] and the band at 2060 cm-1 was related to CO*
species adsorbed on gold.[43-45] IR bands at 2977 and 2892 cm-1
were also observed, which were assigned to (CH).[29-31,39] The
band at 1687 cm-1 was assigned to CH2O*.[39] It was speculated
that C2H5OH on the Au11 catalyst was likely to be formed via CO*
insertion in CH2* species. In addition, the band at 1747 cm-1
emerged, corresponding to (C=O) in the adsorbed HCOOH
species,[46] indicating that HCOOH was also formed on the Au11
catalyst. For the Au36 catalyst system (Figure 5c), the bands at
2877, 1627 and 1385 cm-1 assigned to (CH) and δ(CH) modes
in HCOO* species were found[47,48] and the bands at 2955 and
1747 cm-1 corresponding to (CH) and (C=O) in the adsorbed
HCOOH were also observed.[46,48] Therefore, this meant the
formation of HCOOH on the Au36 catalyst via the adsorbed
formate intermediates. IR spectra of standard samples were
shown in Figure S13 and S14 for clear comparison. Overall, the
above observations confirm the COOH*, CO*, and CHx* species
formed on the Au9 and Au11 catalysts by CO2 hydrogenation,
which are regarded as important intermediates for the formation
of methane and ethanol. Also, the HCOO* species is an essential
intermediate for formic acid formation on the Au36 catalyst.
molecules and reaction intermediates involved in CO2
hydrogenation. With increasing number of shedding ligands, the
binding strength of reaction intermediates on the exposed Au
atoms enhances and quickly achieves convergence with about
half of the ligands being removed (Figure 6b).
Noticeably, the three gold clusters exhibit distinct chemical
reactivity. The kinetic barriers for dissociation of H2 molecule on
Au9(P(CH3)3)4, Au11(P(CH3)3)6 and Au36(SCH3)14 are 0.31, 0.55
and 0.75 eV, respectively, leading to the formation of two H*
adatoms on the Au sites. The binding capabilities of the clusters
reduce as the atomicity increases (Figure 6c). Specially, the
adsorption energies of CO* (a critical intermediate toward various
products) relative to free CO molecule are -1.24, -1.15 and -0.73
eV for Au9(P(CH3)3)4, Au11(P(CH3)3)6 and Au36(SCH3)14,
respectively. For Au36(SCH3)14 with weak binding strength,
HCOOH molecules can desorb easily from the cluster and thus
may be the main products of CO2 hydrogenation. Au9(P(CH3)3)4
and Au11(P(CH3)3)6 having strong binding with CO* and other
intermediates would allow further hydrogenation of CO* or C-C
coupling, such that C1 products beyond CO and HCOOH and
even C2 products are possible. The binding capabilities of gold
clusters are correlated to the d-orbital center of Au atoms, as
demonstrated by their electronic structures in Figure S16. The
smaller-size cluster has a higher d-orbital center relative to HOMO,
which results in less occupancy of the antibonding states formed
with reaction intermediates, and thus provides stronger binding
strength with reaction intermediates, consistent with the d band
theory.[50]
To further evaluate the activity and selectivity of the three
gold clusters, we explored the full reaction pathways and energy
diagrams of CO2 hydrogenation, as presented by Figure 7. On all
these gold clusters, CO2 molecule can only be physisorbed, and
thus has to be activated by direct reaction with a H* adatom to
form either formate (HCOO*) or carboxyl (COOH*) species. Then,
HCOOH molecule and CO* intermediate are produced by
hydrogenation of HCOO* and COOH*, respectively. Further
hydrogenation of CO* and C-C coupling can yield C1 and C2
products, such as CH3OH, CH4 and C2H5OH. As expected,
Au36(SCH3)14 with weak binding capability favors the HCOO*
pathway and produces HCOOH easily with the maximum barrier
of 0.51 eV. The COOH* pathway toward the other products
involves a large barrier of 0.93 eV during the formation of COOH*,
and thus is kinetically hindered. For Au9(P(CH3)3)4 and
Au11(P(CH3)3)6, the COOH* pathway is kinetically more favorable
than the HCOO* one with barriers of 0.49 and 0.18 eV,
respectively. Formation of HCOOH requires a large barrier of
about 0.88 eV due to the strong adsorption of HCOO* and
unfavorable desorption of HCOOH. In particular, Au9(P(CH3)3)4
catalyzes CO* hydrogenation by going through HCO*, H2CO* and
H2COH*, to form CHx and finally produce CH4 with the maximum
barrier of 0.49 eV. In comparison, hydrogenation of H2CO* to
CH3OH or C-C coupling toward C2 products involves larger
barriers of 0.60 and 0.93 eV, respectively. For Au11(P(CH3)3)6, C-
C coupling by chemical bonding between CO* and CH2* is much
easier than that on Au9(P(CH3)3)4, involving a small barrier of 0.24
eV mainly due to the steric effect. Further hydrogenation of CH2*‒
CO* toward C2H5OH is kinetically favorable, and the rate-limited
step occurs at CH2OH* formation with a barrier of 0.60 eV. CH3OH
Figure 6. (a) Energies required for removing ligands from the Au9(P(CH3)3)8,
Au11(P(CH3)3)8Cl2 and Au36(SCH3)24 clusters by DFT calculations. (b) Binding
energies of COOH* intermediate on the three clusters as a function of number
of removed ligands. (c) Binding energies of key reaction intermediates and (d)
kinetic barriers of the rate-limited steps toward various products on
Au9(P(CH3)3)4, Au11(P(CH3)3)6 and Au36(SCH3)14 clusters.
To understand the distinct selectivity of the three non-metallic
gold clusters, we performed DFT calculations to explore the CO2
hydrogenation processes on the clusters. The phenyl group in -
PPh3 and -TBBT ligands were replaced by the -methane thiolates
(-CH3) to simplify our models and reduce the computational
cost.[49] According to the in-situ FTIR measurement (Figure S10-
S12), the ligands were partially shed from the gold clusters.
Therefore, we considered Au9(P(CH3)3)4, Au11(P(CH3)3)6 and
Au36(SCH3)14 as model structures (displayed by Figure S15), such
that some of the Au atoms are exposed and can serve as the
active sites. According to our test calculations, the energy
required for removing ligands has little dependence on the
coverage of ligands on the clusters (Figure 6a). The gold clusters
fully protected by ligands have no activity for chemisorption of
This article is protected by copyright. All rights reserved.

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
and CH4 products require larger barriers of 1.26 and 0.78 eV,
respectively, and thus have lower yield.
metallic metal clusters for extensively catalytic applications and
thus they will open an avenue for a new type of metal catalysts
with high efficiency for a series of challenging chemical reactions.
Based on the DFT calculations, the three gold clusters
possess different binding capabilities with intermediate species
governed by their electronic structures, which in turn determine
the reaction pathway and product selectivity of CO2
hydrogenation. As revealed by Figure 6d, Au9(P(CH3)3)4,
Au11(P(CH3)3)6 and Au36(SCH3)14 exhibit high selectivity for CH4,
C2H5OH and HCOOH with the limited kinetic barriers of 0.49, 0.60
and 0.51 eV, respectively. The product selectivity and emerging
intermediates during the reaction agree well with our experimental
results.
Acknowledgements
We acknowledge financial supports from National Natural
Science Foundation of China (21773109, 91845104
).
Keywords: cluster • non-metallic • gold • CO2 • selectivity
[1]
T. Higaki, Q. Li, M. Zhou, S. Zhao, Y. W. Li, S. Li, R. C. Jin, Acc. Chem.
Res. 2018, 51, 2764-2773.
[2]
[3]
[4]
[5]
I. Chakraborty, T. Pradeep, Chem. Rev. 2017, 117, 8208-8271.
Z. B. Gan, N. Xia, Z. K. Wu, Acc. Chem. Res. 2018, 51, 2774-2783.
X. Kang, M. Z. Zhu, Coordin. Chem. Rev. 2019, 394, 1-38.
A. Ghosh, O. F. Mohammed, O. M. Bakr, Acc. Chem. Res. 2018, 51,
3094-3103.
[6]
[7]
[8]
[9]
Z. Lei, J. J. Li, X. K. Wan, W. H. Zhang, Q. M. Wang, Angew. Chem. Int.
Ed. 2018, 57, 8639-8643.
M. Zhou, C. J. Zeng, Y. X. Chen, S. Zhao, M. Y. Sfeir, M. Z. Zhu, R. C.
Jin, Nat. Commun. 2016, 7, 13240.
C. J. Zeng, Y. X. Chen, K. Kirschbaum, K. Appavoo, M. Y. Sfeir, R. C. Jin,
Sci. Adv. 2015, 1, e1500045.
P. X. Liu, R. X. Qin, G. Fu, N. F. Zheng, J. Am. Chem. Soc. 2017, 139,
2122-2131.
[10] G. Li, R. C. Jin, J. Am. Chem. Soc. 2014, 136, 11347-11354.
[11] X. K. Wan, J. Q. Wang, Z. A. Nan, Q. M. Wang, Sci. Adv. 2017, 3,
e1701823.
[12] R. R. Nasaruddin, T. K. Chen, N. Yan, J. P. Xie, Coordin. Chem. Rev.
2018, 368, 60-79.
[13] W. Choi, G. X. Hu, K. Kwak, M. Kim, D. E. Jiang, J. P. Choi, D. Lee, ACS
Appl. Mater. Inter. 2018, 10, 44645-44653.
[14] C. Liu, H. Abroshan, C. Y. Yan, G. Li, M. Haruta, ACS Catal. 2016, 6, 92-
99.
[15] D. R. Kauffman, J. Thakkar, R. Siva, C. Matranga, P. R. Ohodnicki, C. J.
Zeng, R. C. Jin, ACS Appl. Mater. Inter. 2015, 7, 15626-15632.
[16] Y. Y. Liu. X. Q. Chai, X. Cai, M. Y. Chen, R. C. Jin, W. P. Ding, Y. Zhu,
Angew. Chem. Int. Ed. 2018, 57, 9775-9779.
Figure 7. Energy diagrams of CO2 hydrogenation on Au9(P(CH3)3)4,
Au11(P(CH3)3)6 and Au36(SCH3)14 clusters by DFT calculations. Kinetic barriers
for each elementary step are presented, and the limiting barriers for various
products are highlighted. The insets show the atomic structures of reaction
intermediates during the formation of CH4, C2H5OH and HCOOH for
Au9(P(CH3)3)4, Au11(P(CH3)3)6 and Au36(SCH3)14, respectively. The H, C, O, P,
S and Au atoms are shown in white, grey, red, green, magenta and yellow colors,
respectively.
[17] X. Cai, G. Saranya, K. Shen, M. Y. Chen, R. Si, W. P. Ding, Y. Zhu, Angew.
Chem. Int. Ed. 2019, 58, 9964-9968.
[18] H. Zhang, H. Liu, Z. Q. Tian, D. Lu, Y. Yu, S. Cestellos-Blanco, K. K.
Sakimoto, P. D. Yang, Nat. Nanotech. 2018, 13, 900-905.
[19] Q. G. Liu, X. F. Yang, L. Li, S. Miao, Y. Li, Y. Q. Li, X. K. Wang, Y. Q.
Huang, T. Zhang, Nat. Commun. 2017, 8,1407.
[20] J. J. Wang, G. N. Li, Z. L. Li, C. Z. Tang, Z. C. Feng, H. Y. An, H. L. Liu,
T. F. Liu, C. Li, Sci. Adv. 2017, 3, e1701290.
[21] J. Graciani, K. Mudiyanselage, F. Xu, A. E. Baber, J. Evans, S. D.
Senanayake, D. J. Stacchiola, P. Liu, J. Hrbek, J. F. Sanz, J. A. Rodriguez,
Science 2014, 345, 546-550.
Conclusion
Our results explicitly indicate that the non-metallic gold
clusters, Au9, Au11 and Au36, exhibit atomicity-dependent catalytic
performances in the hydrogenation processes of CO2, where they
can selectively determine the reaction pathways towards C1 or C2
products. Specially, Au11 can give >80% selectivity for ethanol as
high-added value chemicals and Au36 can achieve >85%
selectivity for formic acid. In contrast, the activity of the non-
metallic gold clusters are 10-70 times higher than that of metallic
nanoparticles. DFT calculations reveal the molecular-level
mechanisms of CO2 reaction with H2 over the three clusters,
which are governed by their distinct binding capabilities and
electronic structures. The exceptional properties of the non-
metallic gold clusters highlight the importance of pursuing non-
[22] D. Preti, C. Resta, S. Squarcialupi, G. Fachinetti, Angew. Chem. Int. Ed.
2011, 50, 12551-12554.
[23] B. An, J. Z. Zhang, K. Cheng, P. F. Ji, C. Wang, W. B. Lin, J. Am. Chem.
Soc. 2017, 139, 3834-3840.
[24] Z. H. He, M. Cui, Q. L. Qian, J. J. Zhang, H. Z. Liu, B. X. Han, Proc. Natl.
Acad. Sci. 2019, 116, 12654-12659.
[25] C. S. Yang, S. H. Liu, Y. N. Wang, J. M. Song, G. H. Wang, S. Wang, Z.
J. Zhao, R. T. Mu, J. L. Gong, Angew. Chem. Int. Ed. 2019, 58, 1-6.
[26] C. T. Wang, E. Guan, L. Wang, X. F. Chu, Z. Y. Wu, J. Zhang, Z. Y. Yang,
Y. W. Jiang, L. Zhang, X. J. Meng, B. C. Gates, F. S. Xiao, J. Am. Chem.
Soc. 2019, 141, 8482-8488.
[27] S. Kattel, W. T. Yu, X. F. Yang, B. H. Yan, A. Q. Huang, W. M. Wan, P.
Liu, J. G. Chen, Angew. Chem. Int. Ed. 2016, 55, 7968 -7973.
This article is protected by copyright. All rights reserved.

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
[28] Y. N. Gao, S. Z. Liu, Z. Q. Zhao, H. C. Tao, Z. Y. Sun, Acta Phys. Chim.
Sin. 2018, 34, 858-872.
[29] L. X. Wang, L. Wang, J. Zhang, X. L. Liu, H. Wang, W. Zhang, Q. Yang,
J. Y. Ma, X. Dong, S. J. Yoo, J. G. Kim, X. G. Meng, F. S. Xiao, Angew.
Chem. Int. Ed. 2018, 57, 6104-6108.
[30] S. Bai, Q. Shao, P. T. Wang, Q. G. Dai, X. Y. Wang, X. Q. Huang, J. Am.
Chem. Soc. 2017, 139, 6827-6830.
[31] B. An, Z. Li, Y. Song, J. Z. Zhang, L. Z. Zeng, C. Wang, W. B. Lin, Nat.
Catal. 2019, 2, 709-717.
[32] X. L. Pan, Z. L. Fan, W. Chen, Y. J. Ding, H. Y. Luo, X. H. Bao, Nat. Mater.
2007, 6, 507-511.
[33] Y. Z. Chen, H. L. Li, W. H. Zhao, W. B. Zhang, J. W. Li, W. Li, X. S. Zheng,
W. S. Yan, W. H. Zhang, J. F. Zhu, R. Si, J. Zeng, Nat. Commun. 2019,
10, 1885.
[34] C. Y. Wu, P. Zhang, Z. F. Zhang, L. J. Zhang, G. Y. Yang, B. X. Han,
ChemCatChem 2017, 9, 3691-3696.
[35] S. J. Yamazoe, S. Matsuo, S. Muramatsu, S. Takano, K. Nitta, T. Tsukuda,
Inorg. Chem. 2017, 56, 8319-8325.
[36] L. C. Mckenzie, T. O. Zaikova, J. E. Hutchison, J. Am. Chem. Soc. 2014,
136, 13426-13435.
[37] C. J. Zeng, Y. X. Chen, K. Iida, K. Nobusada, K. Kirschbaum, K. J.
Lambright, R. C. Jin, J. Am. Chem. Soc. 2016, 138, 3950-3953.
[38] Y. N. Sun, E. D. Wang, Y. J. Ren, K. Xiao, X. Liu, D. Yang, Y. Gao, W. P.
Ding, Y. Zhu, Adv. Funct. Mater. 2019, 27, 1904242.
[39] G. A. Simms, J. D. Padmos, P. Zhang, J. Chem. Phys. 2009, 131, 214703.
[40] L. B. Wang, W. B. Zhang, X. E. Zheng, Y. Z. Chen, W. L. Wu, J. X. Qiu,
X. C. Zhao, X. Zhao, Y. Z. Dai, J. Zeng, Nat. Energy. 2017, 2, 869-876.
[41] Y. M. Wang, C. Wöll, Chem. Soc. Rev. 2017, 46, 1875-1932.
[42] V. Hody, T. Belmonte, C. D. Pintassilgo, F. Poncin-Epaillard, T. Czerwiec,
G. Henrion, Y. Segui, J. Loureiro, Plasma. Chem. Plasm. 2006, 26, 251-
266.
[43] F. Somodi, I. Borbath, M. Hegedus, A. Tompos, I. E. Sajo, A. Szegedi, S.
Rojas, J. L. G. Fierro, J. L. Margitfalvi, Appl. Catal. A: Gen. 2008, 347,
216-222.
[44] F. Boccuzzi, A. Chiorino, M. Manzoli, Surf. Sci. 2000, 454-456, 942-946.
[45] S. H. Pang, C. A. Schoenbaum, D. K. Schwartz, J. W. Medlin, Nat.
Commun. 2013, 4, 2448.
[46] J. R. R. García, J. C. F. Gonzalez, B. E. Handy, L. H. Reyes, D. A. D. H.
D. Río, C. J. L.O. S. V. Cervantes, G. A. F. Escamilla, ChemistrySelect.
2008, 4, 4206-4216.
[47] K. L. Miller, C. W. Lee, J. L. Falconer, J. W. Medlin, J. Catal. 2010, 275,
294-299.
[48] T. Chen, G. P. Wu, Z. C. Feng, G. S. Hu, W. G. Su, P. L. Ying, C. Li, Chin
J. Catal. 2008, 29, 105-107.
[49] S. Zhao, N. Austin, M. Li, Y. Song, S. D. House, S. Bernhard, J. C. Yang,
G. Mpourmpakis, R. C. Jin, ACS Catal. 2018, 8, 4996-5001.
[50] J. K. Nørskov, F. Abild-Pedersen, F. Studt, T. Bligaard, Proc. Natl. Acad.
Sci. 2011, 108, 937-943.
This article is protected by copyright. All rights reserved.

10.1002/anie.201913635
Angewandte Chemie International Edition
RESEARCH ARTICLE
Entry for the Table of Contents (Please choose one layout)
Layout 1:
RESEARCH ARTICLE
Non-metallic gold clusters (Au9, Au11
and Au36) exhibit atomicity-dependent
catalytic performances in the
Dan Yang, Wei Pei, Si Zhou*, Jijun
Zhao, Weiping Ding, and Yan Zhu*
Page No.1 – Page No.5
hydrogenation processes of CO2,
where they can selectively determine
the reaction pathways towards C1 or
C2 products. The exceptional
Controllable Conversion of CO2 on
Non-metallic Gold Clusters
properties of the non-metallic gold
clusters highlight the importance of
pursuing non-metallic metal clusters
for extensively catalytic applications.
Layout 2:
RESEARCH ARTICLE
Author(s), Corresponding Author(s)*
((Insert TOC Graphic here))
Page No. – Page No.
Title
Text for Table of Contents
This article is protected by copyright. All rights reserved.