Turning Au Nanoclusters Catalytically Active for Visible-Light-Driven CO2 Reduction through Bridging Ligands

Article abstract of DOI:10.1021/jacs.8b06723

Development of visible-light photocatalytic materials is an ultimate goal for solar-driven CO₂ conversion. Au nanoclusters (NCs) may potentially serve as components for harvesting visible light but can hardly perform solar-driven CO₂ reduction due to the lack of catalytic sites. Herein, we report an effective strategy for turning Au nanoclusters catalytically active for visible-light CO₂ reduction, in which metal cations (Fe²⁺, Co²⁺, Ni²⁺, and Cu²⁺) are grafted to the Au NCs using l-cysteine as a bridging ligand. The metal-S bonding bridge facilitates the electron transfer from Au NCs to metal cations so that the grafted metal cations can receive photoinduced electrons and work as catalytic sites for CO₂ reduction. The varied d-band centers and binding energies with CO₂ for different metal cations allow tuning electron transfer efficiency and CO₂ activation energy. Furthermore, the photostability of Au NCs-based catalyst can be significantly enhanced through the encapsulation with metal-organic frameworks. This work opens a new door for the photocatalyst design based on metal clusters and sheds light on the surface engineering of metal clusters toward specific applications.

Full text of DOI:10.1021/jacs.8b06723

Subscriber access provided by Kaohsiung Medical University
Article
Turning Au Nanoclusters Catalytically Active for Visible-
Light-Driven CO2 Reduction through Bridging Ligands
Xiaofeng Cui, Jin Wang, Bing Liu, Shan Ling, Ran Long, and Yujie Xiong
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.8b06723 • Publication Date (Web): 08 Nov 2018
Downloaded from http://pubs.acs.org on November 8, 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 8
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Turning Au Nanoclusters Catalytically Active for Visible-Light-
Driven CO2 Reduction through Bridging Ligands
Xiaofeng Cui,†, ‡, ¶ Jin Wang,†, ‡ Bing Liu,†, ┴ Shan Ling,¶ Ran Long,‡ and Yujie Xiong*, ‡
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
‡Hefei National Laboratory for Physical Sciences at the Microscale, iChEM (Collaborative Innovation Center of Chemistry
for Energy Materials), School of Chemistry and Materials Science, and National Synchrotron Radiation Laboratory,
University of Science and Technology of China, Hefei, Anhui 230026, P. R. China.
¶Anhui Key Laboratory of PhotoelectricꢀMagnetic Functional Materials, School of Chemistry and Chemical Engineering,
Anqing Normal University, Anqing, Anhui 246011, P. R. China.
┴School of Chemical and Material Engineering, Jiangnan University, Wuxi, Jiangsu 214122, R. P. China.
KEYWORDS. Photocatalysis; gold cluster; catalytic sites; ligand; CO2 reduction.
Supporting Information Placeholder
ABSTRACT: Developing visibleꢀlight photocatalytic materials is an ultimate goal for solarꢀdriven CO2 conversion. Au
nanoclusters may potentially serve as the components for harvesting visible light, but can hardly perform the solarꢀdriven CO2
reduction due to the lack of catalytic sites. Herein, we report an effective strategy for turning Au nanoclusters catalytically active
for visibleꢀlight CO2 reduction, in which metal cations (Fe2+, Co2+, Ni2+ and Cu2+) are grafted to the Au NCs using Lꢀcysteine as a
bridging ligand. The metalꢀS bonding bridge facilitates the electron transfer from Au NCs to metal cations so that the grafted metal
cations can receive photoꢀinduced electrons and work as catalytic sites for CO2 reduction. The varied dꢀband centers and binding
energies with CO2 for different metal cations allow tuning electron transfer efficiency and CO2 activation energy. Furthermore, the
photostability of Au NCsꢀbased catalyst can be significantly enhanced through the encapsulation with metalꢀorganic frameworks.
This work opens a new door for the photocatalyst design based on metal clusters, and sheds light on the surface engineering of
metal clusters toward specific applications.
levels.11ꢀ13 For this reason, metal NCs exhibit analogous
properties to molecular complexes whose light absorption can
INTRODUCTION
Solarꢀdriven conversion of CO2 and H2O into chemical fuels
be readily tuned, and have emerged as visibleꢀlight harvesters
such as CH4, CO/H2 and methanol has been developed as a
in photocatalytic and photovoltaic systems.14ꢀ17 Thus Au NCs
very promising approach to address current energy and
environmental challenges.1ꢀ6 To improve solar utilization, the
are ideal candidates for absorbing visible light alternatively to
semiconductor. Although some metal NCs have demonstrated
visible light, which accounts for about 43% photons in the
their catalytic activity in reactions such as oxidation,
solar spectrum, should be harnessed for CO2 photoreduction.
hydrogenation and CꢀC coupling,9, 10, 18ꢀ21 pristine Au NCs are
Bandgap engineering and doping are the major approaches to
almost inactive in photocatalytic reactions. The tight ligand
tuning the spectral range of light harvesting by semiconductor
capping on metal surface results in the lack of active sites for
− the key material to photocatalysis. Tailoring the parameters
catalyzing reactions, and as such, the photogenerated electrons
of semiconductor often causes the changes in electronic
can hardly participate in reactions. To overcome this
structures, charge separation and transfer, which in turn results
limitation, surface engineering should be performed on Au
in the variation of catalytic activity and selectivity.7, 8 This
NCs to create active sites for CO2 reduction.
complexity calls for the efforts to develop alternative materials
In this article, we report a facile approach to the surface
modification of glutathioneꢀprotected Au NCs (namely, Auꢀ
GSH NCs) through ligand grafting with Lꢀcysteine (Lꢀcys),
which enables the effective immobilization of various
transition metal cations as catalytic sites on surface. As such,
the catalytic activity and selectivity in CO2 photoreduction can
be modulated by simply altering the grafted metal cations
(Fe2+, Co2+, Ni2+and Cu2+). Density functional theory (DFT)
calculations indicate that different active metals possess
diverse dꢀband centers and adsorption energy with CO2,
altering the activity and selectivity of CO2 reduction. As a
proof of concept, the Au NCs grafted with Co2+ cations
which can harvest visible light through a different mechanism.
Such a material platform would help reveal the fundamental
mechanisms in photocatalytic CO2 reduction.
Metal nanoclusters (NCs) are a class of ultrasmall
nanoparticles as alternative materials to harvest solar light. In
particular, thiolateꢀprotected Au NCs can be obtained in
solution phase through atomically precise synthesis, and
possess high conductivity, strong visibleꢀlight response and
abundant surface binding sites for cocatalysts and/or reaction
species.9, 10 As their particle sizes (∼1 nm) approach to de
Broglie wavelength, the continuous electronic bands in
metallic nanoparticles would evolve to discrete energy
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 8
achieve a rate of 3.54 ꢁmol⋅gcat−1⋅h−1 and a selectivity up to
would be able to serve as additional catalytic sites to Au NCs.
1
2
3
4
5
6
7
8
In this work, we select waterꢀsoluble AuꢀGSH NCs as a model
system owing to its intrinsic unique photochemical property
and synthetic feasibility. The abundant surface ligands on Auꢀ
GSH NCs provide the opportunity for constructing catalytic
sites through ligands modification.
65.2% for CO production.
EXPERIMENTAL SECTION
Functionalization of Au-GSH NCs with L-cysteine
(Auc-C). The synthetic procedures for AuꢀGSH NCs are
included in the Supporting Information. 10 mL of AuꢀGSH
NCs dispersion (1 g⋅Lꢀ1) was added into 10 mL of 2ꢀ(Nꢀ
morpholino) ethane sulfonic acid (MES) buffer (3.9 mg⋅mLꢀ1)
in a 50ꢀmL roundꢀbottom flask, which was then heated to 65
°C. Subsequently, 6 mg of 1ꢀ(3ꢀdimethylaminopropyl)ꢀ3ꢀ
ethylcarbodiimide hydrochloride (EDC) and 1.6 mg of Nꢀ
hydroxysuccinimide (NHS) were added as catalysts to form
amide bond between the carboxylic groups of the AuꢀGSH
NCs and the amino groups of Lꢀcys. After stirring for 15 min,
40 mg of Lꢀcys was added in. The final mixture was stirred at
65 °C for 6 h. The resulting Lꢀcysꢀmodified AuꢀGSH NCs
(AucꢀC) were collected by centrifugation at 8,000 rpm for 5
min and washed with deionized water three times. Finally, 5
mL of deionized water was added into the purified AucꢀC, and
30 ꢁL of NaOH (30% wt) was added timely for better
dispersing AucꢀC in aqueous solution.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Schematic illustration for the process of modifying
Au clusters with Lꢀcys ligands (AucꢀC) and further grafting
them with metal cations (AucꢀCꢀM).
The concept and preparation procedures are illustrated in
Figure 1. AuꢀGSH NCs are facilely prepared through a simple
oneꢀpot method by slightly modifying the reported protocol, in
which aqueous solutions of HAuCl4 and GSH are mixed and
allowed to react under gentle stirring at a temperature of 70 °C
(see details in Supporting Information).26
The lowest
unoccupied molecular orbital (LUMO) and highest occupied
molecular orbital (HOMO) energy levels of AuꢀGSH NCs are
located at ꢀ0.63 V and 1.51 V (vs. RHE), which are
thermodynamically favorable for CO2 reduction and H2O
oxidation, respectively.17 As indicated by matrixꢀassisted laser
desorption/ionization timeꢀofꢀflight (MALDIꢀTOF) mass
spectrometry (Figure S1), this simple oneꢀpot method can
yield AuꢀGSH NCs with a size distribution ranging from Au24
to Au39. The AuꢀGSH NCs are excellent absorbers to harvest
visible light up to 500 nm as well as offer relatively high
stability as compared with monodispersed Au NCs, which can
improve photocatalytic activity and durability. The quantum
yield of AuꢀGSH NCs may reach 15%, orders of magnitude
higher than that of the reported Auꢀthiolate NCs (typically
0.001−0.1%).26 The AuꢀGSH NCs display excellent stability,
including storage in water at room temperature and elevated
temperature (80 °C), in solutions of high salt concentration,
and in common buffer solutions.26 The capping ligands on
AuꢀGSH NCs only provide ꢀNH2 and ꢀCOOH groups for
anchoring transition metal cations, which have lower affinity
as compared with ꢀSH group. To bind catalytic sites more
firmly on the surface of Au NCs, AuꢀGSH NCs are further
functionalized with Lꢀcys (namely, AucꢀC). This step is
achieved through the formation of amide bond between the
carboxylic groups of GSH and the amino groups of Lꢀcys,
catalyzed by 1ꢀ(3ꢀdimethylaminopropyl)ꢀ3ꢀethylcarbodiimide
hydrochloride (EDC) and Nꢀhydroxysuccinimide (NHS).27
The structure of Au NCs is well retained after the
functionalization (see Figure S2), and the uncoordinated thiols
introduced by Lꢀcys provide a platform for anchoring metal
cations onto the nanoclusters (namely, AucꢀCꢀM; M = Fe, Co,
Ni and Cu).
Grafting with metal cations (Auc-C-M). The grafting of
metal cations was carried out right before photocatalytic
reactions. A certain amount of metal cations (Fe2+, Co2+,
Ni2+or Cu2+) were added into 10 mL of AucꢀC dispersion (1
g⋅Lꢀ1) under vigorous stirring. Metal cations were immediately
coordinated with the thiol of Lꢀcys to serve as catalytic sites in
photocatalytic reactions. The Au NCs grafted with metal
cations were denoted as AucꢀCꢀM.
Photocatalytic CO2 conversion measurements. The
photocatalytic CO2conversion measurement was performed in
a homeꢀmade quartz reactor with a total volume of 50 mL. 1
mL of triethanolamine (TEOA) was added into 10 mL of
pristine AucꢀCꢀCo solution (1 g⋅mLꢀ1) to sacrifice
photogenerated holes. The mixture was bubbled with CO2
about 30 min to eliminate air and form saturated CO2
atmosphere. A 300ꢀW Xe lamp (Solar edge 700, China) with
a 420ꢀnm longꢀwaveꢀpass cutꢀoff filter (λ≥420 nm) was used
as a visibleꢀlight source. The power density was measured to
be 100 mW cmꢀ2. The photocatalytic reaction was typically
performed for 3h at room temperature and under ambient
pressure. The amounts of CH4, CO and H2 evolved were
determined using gas chromatograph (GC, 7890A, Ar carrier,
Agilent). H2 was detected using a thermal conductivity
detector (TCD). CH4 was measured by a flame ionization
detector (FID). CO was converted to CH4 by a methanation
reactor and then analyzed by the FID. The isotopeꢀlabeled
experiments were performed using 13CO2 instead of 12CO2, and
the products were analyzed using gas chromatographyꢀmass
spectrometry (GCꢀMS, 7890A and 5975C, Agilent).
RESULTS AND DISCUSSION
To investigate whether the amide bond has been formed
between AuꢀGSH NCs and Lꢀcys, synchrotron radiationꢀbased
Fourierꢀtransform infrared spectroscopy (FTIR) spectroscopy
is employed to examine the samples. As shown in Figure 2a,
pure GSH display three peaks of amide bond at 1655 cmꢀ1
(C=O stretching), 1545 cmꢀ1 (N–H bending) and 1397 cmꢀ1
The central concept of this work is to bridge lightꢀ
harvesting Au NCs and catalytic metal sites with ligands. As a
matter of fact, we can receive some inspirations from the
applications of Au NCs as fluorescent probes in the detection
of metal cations, in which the fluorescence of Au NCs is
quenched by the efficient electron transfer from Au NCs to
metal cations.22ꢀ25 If the metal cations that receive the
photogenerated electrons can catalyze CO2 reduction, they
(C–N stretching).
However, these characteristic peaks
disappear after binding GSH to Au NCs (i.e., AuꢀGSH NCs),
ACS Paragon Plus Environment

Page 3 of 8
Journal of the American Chemical Society
most likely as the amide bonds on AuꢀGSH NCs are not
so that more Lꢀcys molecules can be adsorbed onto the surface
of AuꢀGSH NCs, which favors the reaction of GSH with the
amine group of adsorbed Lꢀcys.
1
2
3
4
5
6
7
8
infrared active. This feature has been reported for the amide
bonds in GSHꢀprotected nanoclusters.28, 29 After grafting Lꢀ
cys onto the AuꢀGSH NCs (i.e., AucꢀC), the peaks at 1660 cmꢀ
The next question would be whether metal cations have
been bound to the thiols of modified Au NCs. Highꢀ
resolution XPS has been employed to provide the evidence.
Here we take the earth abundant Co2+ as an example to
demonstrate this approach, which can be coordinated with the
thiols of AucꢀC to form AucꢀCꢀCo. As charge transfer should
take place upon the coordination of thiols and Co2+ cations,
we pay more attention to the XPS peak shift of S and grafted
Co. The S2p XPS spectrum (Figure 2c) shows the S2p1/3 and
S2p2/3 peaks with binding energies at 162.8 eV and 163.9 eV,
respectively, which can be assigned to the thiols of GSH
and/or Lꢀcys.31 After grafting Co2+, the S2p1/3 and S2p2/3
peaks are shifted about 0.2 eV toward higher binding energy,
which indicates the reduction in electron density of the S in
AucꢀCꢀCo. Meanwhile, the peak of grafted Co2p (779.6 eV)
is shifted toward opposite direction with respect to the Co2+
cations mixed with unmodified AuꢀGSH NCs (AucꢀCo, 780.4
eV) as shown in Figure 2d, revealing that the electron density
of the Co atoms increases after grafting. These XPS peak
shifts can be attributed to the coordination of Co2+ with thiol
which causes the migration of electrons from S atoms to Co
atoms.31 The XPS characterization thus confirms the effective
binding of Co cations to Au NCs though Lꢀcys ligands.
1
(C=O stretching), 1530 cmꢀ1 (N–H bending) and 1390 cmꢀ1
(C–N stretching) appear again, which can be ascribed to the
characteristics for bending modes of newly formed amide
bond between GSH and Lꢀcys.30 Moreover, the stretching
vibration peak of N−H bond at 3000−3200 cmꢀ1 is shifted
toward higher frequency in AucꢀC with respect to bare Lꢀcys,
which is attributed to the inductive effect of electronꢀ
withdrawing group − carbonyl in amide. This verifies the
formation of amide bond during the functionalization on Auꢀ
GSH NCs. This conclusion is further confirmed by Xꢀray
photoelectron spectroscopy (XPS). As shown in Figure 2b,
the Au4f XPS peaks for AucꢀC are shifted about 0.4 eV
toward lower binding energy after grafting Lꢀcys. This
evidences the existence of Lꢀcys ligands, as the electron
donation by ꢀNH2 in the Lꢀcys can increase the electron
density of Au atoms.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. (a) Average production rates of CH4, CO and H2 in
lightꢀdriven CO2 reduction with H2O in the presence of
TEOA, catalyzed by 10ꢀmg AucꢀC grafted with different
amounts of Co2+. (b) Mass spectra of 13CO (m/z = 29)
produced over AucꢀCꢀCo in lightꢀdriven reduction of 13CO2.
(c) Average production rates of CH4, CO and H2 in lightꢀ
driven CO2 reduction with H2O in the presence of TEOA,
catalyzed by 10ꢀmg AucꢀC grafted with 4ꢀꢁmol Fe2+, Co2+,
Ni2+ or Cu2+. The irradiation is performed with visible light (λ
≥ 420 nm) with a power density of 100 mW·cm−2 for 3 h.
Figure 2. (a) FTIR spectra of AuꢀGSH NCs and AucꢀC in
reference to Lꢀcys and GSH. Highꢀresolution XPS spectra of
(b) Au4f in AuꢀGSH NCs and AucꢀC, (c) S2p in AucꢀC and
AucꢀCꢀCo, and (d) Co2p in AucꢀCo and AucꢀCꢀCo.
To further examine whether this functionalization process
1
results in ligand exchange, we employ H nuclear magnetic
resonance (NMR) spectroscopy to characterize the solution
after the reaction of modifying AuꢀGSH NCs with Lꢀcys for 6
h. As shown in Figure S3, no free GSH can be detected in the
solution, which indicates that the ligand exchange between the
Lꢀcys and AuꢀGSH NCs should be negligible in this system.
The AucꢀC is fully purified prior to the further coordination
with metal cations so that no free Lꢀcys or GSH is involved in
the system (Figure S4). As determined by highꢀperformance
liquid chromatography (HPLC, Figure S5), the molar ratio of
GSH to grafted Lꢀcys is about 1:0.81 after the
functionalization. It is also worth pointing out that AuꢀGSH
NCs have not aggregated through cross coupling after the
ligand modification as indicated by their light absorption
We are now in a position to assess the performance of
grafted metal cations as catalytic sites. The evaluation is
performed for our catalysts with CO2 and H2O under visibleꢀ
light irradiation, using triethanolamine (TEOA) to sacrifice
photogenerated holes. As shown in Table S1, the AuꢀGSH
NCs and AucꢀC without metal cations anchored do not exhibit
catalytic activity for CO2 reduction. As Au NCs are grafted
with Co2+ through Lꢀcys (AucꢀCꢀCo, Figure 3a), they can be
turned catalytically active for CO2 reduction. In contrast,
simply mixing AuꢀGSH NCs with metal cations does not
generate the catalytic activity, indicating that the Lꢀcys
bridging is the key to electron transfer from lightꢀharvesting
center to catalytic sites. To trace the carbon source of CO,
isotopic 13CO2 is used as the reactant to perform the
(Figure S6).
The undesired crosslinking is effectively
suppressed by adding excess Lꢀcys into AuꢀGSH NCs solution
photocatalytic reaction under identical conditions.
As
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 8
indicated by mass spectrometry (Figure 3b), the peak at m/z=
quenching effect, and nearly fully quenches the PL at the
amount of 4 ꢁmol. Co2+ and Ni2+ quench the PL 52.1% and
49.7% at the same amount of metal cations, respectively,
while no apparent quenching can be observed for the addition
of Fe2+. The electron transfer between clusters and metal
cations can induce such a quenching effect on the PL of metal
NCs. This argument is supported by the PL measurements for
the samples in the absence of bridging ligands. The PL of the
unmodified AuꢀGSH NCs (Auc) whose strong photoemission
exhibits a maxima around 630 nm can hardly be quenched by
the addition of Fe2+, Co2+ and Ni2+ cations (Figure S11). The
only exception is the distinct quenching observed for the
addition of Cu2+ cations. A previous report indicates a similar
case caused by aggregationꢀinduced fluorescence quenching
through the complexation of Cu2+ with GSH rather than
electron transfer.32
29 can be ascribed to 13CO. It confirms that the CO indeed
originates from photocatalytic CO2 reduction.
1
2
3
4
5
6
7
8
Furthermore, a volcano tendency (Figure 3a) has been
observed for the photocatalytic activity by continuously
increasing the amount of Co2+ cations from 1 to 16 ꢁmol. At
the volcano maximum (4ꢀꢁmol Co2+), the reduction of CO2 to
CO can be achieved with an evolution rate of 3.45
ꢁmol⋅gcat−1⋅h−1 and a selectivity of 65.2% (Table S1). In
addition to a small amount of produced CH4, the major side
reaction is the reduction of protons in H2O to H2. The
synthesis gas with a CO/H2 ratio between 0.5 and 1 is the
desired gas source for the thermochemical synthesis of
methanol and FischerꢀTropsch reactions in the existing
industrial processes. In general, more catalytic sites can offer
higher activity; however, the photocatalytic activity is sharply
declined together with the reduced CO selectivity as the
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
To look into the origin of PL quenching, normalized timeꢀ
resolved PL decay is conducted with a pulsed laser excitation
at λ = 365 nm (Figure 4a). The decay traces can be fitted
using triexponential decay kinetics with the parameters listed
in Table S4. Similarly to the decay tendency in steadyꢀstate
PL spectra, AucꢀCꢀCu, AucꢀCꢀCo and AucꢀCꢀNi show fast
decay with average PL decay time (Taverage) of 0.08 ns, 0.51 ns
and 0.89 ns, respectively, while AucꢀCꢀFe exhibits comparable
amount of Co2+ is increased beyond 4 ꢁmol.
This
deteriorative performance may result from the redundantly
free Co2+ cations in solution (Figure S7), as we have not
purified the AucꢀCꢀCo NCs after the coordination with metal
cations. The free Co2+ cations may act as a shuttle redox
mediator for the recombination of photogenerated electrons
and holes to reduce photocatalytic activity.16 The activity for
CO2 reduction is more affected by free Co2+ than H2
generation as proton reduction can more easily take place than
CO2 reduction (the rateꢀlimiting step for CO2 reduction CO2 +
e−→CO2•−, −1.85V versus SHE; water reduction 2H2O(I) +
2e−→ H2(g) + 2OH−(aq), −0.414 V versus SHE). As the Co2+
loading reaches 24 µmol, both H2 production and CO2
reduction are totally shut off. After removing the free Co2+,
the photocatalytic activity can be improved as shown in Figure
S8.
Taverage to AucꢀC. Nevertheless, electron transfer is not the only
factor responsible for PL decay. It is known that the
contributions of τ1 and τ2 are more related to charge transfer.33,
34
Although the τ1 and τ2 for AucꢀCꢀCu are smaller than those
for AucꢀCꢀCo, its PL decay is more governed by τ3.
Another merit of our synthetic scheme is the variety of
metal cations that can be readily grafted to Au NCs, which
allows correlating catalytic performance with metal sites.
Figure 3c shows the average production rates of CH4, CO and
H2 by AucꢀC grafted with different metal cations (see the
actual loading amounts in Table S2). Apparently, the other
three catalytic centers (Fe2+, Ni2+ and Cu2+) are more favorable
to produce H2, whose H2 selectivity even reach 81.11%,
94.51% and 100%, respectively. Notably, AucꢀCꢀNi exhibits
the highest activity with the average rate of electron
consumption (Relectron) reaching 31.1 ꢁmol⋅gcat−1⋅h−1. Table S3
summarizes the performance comparison of our samples with
the reported cases.
Figure 4. (a) Timeꢀresolved transient fluorescence spectra of
AuꢀGSH NCs modified with Lꢀcys (AucꢀC) and AucꢀC grafted
with different metal cations (Fe2+, Co2+, Ni2+ and Cu2+, 4
ꢁmol) in aqueous solution, excited at 455 nm. (b) ESR spectra
of AucꢀCo (lower) and AucꢀCꢀCo (upper) in dark or under light
irradiation for 2 min.
To appreciate the roles of surface modification and metal
cation grafting, we collect spectroscopic information for light
absorption and charge behavior. The absorption spectrum in
Figure S6 indicates that AucꢀC is an excellent absorber to
harvest visible light up to 500 nm, whose absorption profile
has not been altered by grafting various metal cations. Given
the well maintained light absorption, we further look into the
transfer of photogenerated electrons from AuꢀGSH NCs to the
grafted metal sites, whose efficiency should be a crucial factor
for photocatalytic reactions according to the Marcus theory.
To correlate the electronꢀtransfer process with metal sites, we
employ steady and transient photoluminescence (PL)
spectroscopy to examine the samples. The steadyꢀstate PL of
Au NCs is gradually quenched by increasing the loading
amounts of Co2+ cations (Figure S9). Furthermore, various
metal cations display different quenching effects on the PL of
Au NCs (Figure S10). Cu2+ exhibits an exceptionally strong
We also employ electron spin resonance (ESR)
spectroscopy to probe the information for trapped electrons at
metal sites. As shown in Figure 4b, strong signals are
observed at g=1.99, which can be ascribed to highꢀspin state
of Co2+.35 The peak shift for AucꢀCꢀCo is caused by the
interaction between AucꢀC and Co, which further confirms the
strong coordination ability of functionalized AucꢀC.
Furthermore, the ESR signal intensity for AucꢀCꢀCo is
significantly weakened after light irradiation, while no distinct
change is observed for AucꢀCo in the absence of Lꢀcys bridge.
This Co ESR signal decay can be assigned to the lowered spin
state during the reduction of Co2+ by photoexcited electrons.36
Given the role of Co2+ as an electron acceptor, we propose that
the grafted Co2+ should be the active site for CO2 reduction. A
similar ESR signal change with light irradiation is also
observed for the Fe2+, Ni2+ and Cu2+ cations that are grafted to
ACS Paragon Plus Environment

Page 5 of 8
Journal of the American Chemical Society
Au NCs through the Lꢀcys bridging ligand (Figure S12). This
characterization also indicates that the Lꢀcys bridging ligand is
the key to electron transfer.
physisorption. The chemisorption of CO2 on AucꢀCꢀCo favors
the reduction of CO2 and improves the selectivity of CO2
reduction over proton reduction.
1
2
3
4
5
6
7
8
9
Upon acquiring the information of electron transfer, we
further examine the electronic structures of metal sites through
DFT calculations as the electron transfer is not the only factor
determining photocatalytic activity and selectivity. Based on
the optimized structural models in Figure S13, we obtain the
dꢀband centers (εd) of 4 different metal sites. The dꢀband
center is defined as the average energy of dꢀband, which
characterizes the ability to inject electrons into the adsorbate
from the dꢀband of metal.37 In general, electrons can be more
easily injected into the antibonding orbitals when the dꢀband
center is closer to the Fermi level, making catalyst surface
more reactive.37ꢀ39 Figure 5a shows that the calculated dꢀband
centers (εd) for the AucꢀC grafted with Fe2+, Co2+, Ni2+, and
Cu2+, whose distance to the Fermi level (Ef) follows the order:
AucꢀCꢀNi (ꢀ1.61) < AucꢀCꢀCo (ꢀ2.13) < AucꢀCꢀFe (ꢀ2.63) <
AucꢀCꢀCu (ꢀ3.18). From the viewpoint of active sites, Ni2+
and Co2+ should offer higher catalytic activity in reactions.
Taken together, the PL characterization and DFT calculation
show that the photocatalytic performance of AucꢀCꢀFe and
AucꢀCꢀCu is largely limited by their low efficiency of electron
transfer and/or low catalytic activity of active sites,
respectively.
To depict the reaction pathway on AucꢀCꢀCo, we calculate
the energies for CO formation, desorption, dissociation and
hydrogenation on AucꢀCꢀCo surface. As displayed in Figure
5c and 5d, the desorption of CO from surface requires a lower
energy (0.94 eV) than the dissociation (3.14 eV) or
hydrogenation (1.22 eV) of CO molecule. For this reason, the
CO formed on AucꢀCꢀCo would preferentially be desorbed
from surface and produce CO gas rather than undergo further
reactions. This improves the selectivity of CO production by
AucꢀCꢀCo catalysts as observed in our experiments. In
particular, the formation of CH4, which requires the cleavage
of CꢀO bond and the formation of CꢀH bond, is limited by the
high energy barriers for CO dissociation and hydrogenation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
To examine the durability of our catalysts, we have
performed a test in 3 successive cycles each of which takes 3
hours.
Unfortunately, the NCꢀbased catalysts exhibit
frustrating photostability in the cycling test, and become
nearly inactive after the 3 cycles (Figure S14). Transmission
electron microscopy (TEM) characterization (Figure S15)
reveals that Au NCs have been aggregated and grown from
1.53 nm up to Au nanoparticles with an average diameter of
3.34 nm after the 3 cycles, together with emerging surface
plasmon resonance (SPR) peaks for Au nanoparticles (Figure
S16). Similar situation has been found for the Au NCs used
as a photosensitizer.17, 40
To prevent NC agglomeration and enhance photostability,
we employ Zn(2ꢀmethylimidazole)2 (ZIFꢀ8) framework as an
encapsulation shell to protect AucꢀCꢀCo (Figure S17 and S18).
Cycling test indicates that the photostability is remarkably
improved by the ZIFꢀ8 encapsulation. No obvious activity
decay is observed during four 3ꢀhour reaction cycles (Figure
S19), as both ZIFꢀ8 coating and AucꢀCꢀCo can be well
maintained (Figure S20 and S21). Thus the encapsulation of
Au NCs with metalꢀorganic frameworks (MOFs) is an
effective strategy for enhancing their photostability. However,
in the case of ZIFꢀ8, the activity and CO selectivity are
lowered after the encapsulation. The cladding shell of ZIFꢀ8
hinders the mass transfer of CO2 and TEOA to lower the
catalytic activity. Meanwhile, the CO selectivity is most likely
reduced by the presence of Zn in ZIFꢀ8, which demonstrates
its role as active sites to produce H2 (Figure S19). It is
anticipated that these deficiencies in our proofꢀofꢀconcept
experiments can be overcome by coating a porous shell with
optimal channels and appropriate metal atoms.
Figure 5. (a) Projected density of states (PDOS) of the dꢀ
orbitals of surface atoms for AucꢀCꢀFe, AucꢀCꢀCo, AucꢀCꢀNi
and AucꢀCꢀCu. The red dashed lines represent the Fermi level,
and the solid lines indicate the dꢀband centers (εd). (b) Models
for CO2 adsorption on AucꢀCꢀCo surface (left) and AucꢀCꢀNi
surface (right). (c) Energy profiles for CO formation and
desorption, CO dissociation, and CO hydrogenation on AucꢀCꢀ
Co surface. (d) Calculated structures of CO formation and
desorption, CO dissociation, and CO hydrogenation on AucꢀCꢀ
Co surface.
To demonstrate the efficacy of our approach to Au NCs,
we also employ a monodispersed Au25(SG)18 system to
perform photocatalytic CO2 reduction. After modified with Lꢀ
cys and grafted with Co2+, the Au25ꢀCꢀCo shows good
photocatalytic activity (Figure S22), proving that our reported
approach should be applicable to various NCs.
This
monodispersed system also provides a wellꢀdefined platform
for examining the grafting of Lꢀcys to Au NCs through
MALDIꢀTOF mass spectrometry. As shown in Figure S23,
this characterization suggests that Au25(SG)18 is modified with
Lꢀcys and well maintained during the functionalization.
In addition to photocatalytic activity, product selectivity is
a key parameter to CO2 reduction. AucꢀCꢀCo and AucꢀCꢀNi
show very different selectivity in CO2 reduction, which can be
attributed to the varied adsorption strength of CO2 to metal
sites. As shown in Figure 5b, the CO2 adsorption energy on
AucꢀCꢀCo is calculated to be ꢀ0.61 eV with the formation of
CꢀCo bond. In contrast, the adsorption of CO2 on AucꢀCꢀNi is
relatively weak with an adsorption energy of ꢀ0.16 eV. In
particular, the absence of CꢀNi bonding indicates a weak
Another key point to the universal feature of our method is
the selection of bridging ligands. We modify AuꢀGSH NCs
with another ligand − 3ꢀmercaptopropionic acid (MPA) which
has a similar molecular structure to Lꢀcys except the absence
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 8
of amine in MPA, and then graft Co2+ onto the NCs through
MPA. The obtained AucꢀMPAꢀCo shows good activity for
lightꢀdriven CO2 reduction (Figure 24), indicating that MPA
also can work as an electron transporter. In our approach, the
bridging ligand should be a bifunctional linker molecule that
possesses one functional group (e.g., carboxylate and amine)
for binding with Au NCs and one thiol as a terminal group for
coordination with metal cations. We further compare the
photocatalytic activity of AucꢀCꢀCo and AucꢀMPAꢀCo, and
identify that the activity of AucꢀMPAꢀCo is slightly higher
than that of AucꢀCꢀCo. The nitrogen atom possesses a
stronger negative induction effect owing to its high
electronegativity than sulfur, which makes the charge density
of thiol in Lꢀcys lower than that in MPA. DFT calculation
reveals that the binding energy of S−Co in the coordination of
Lꢀcys with Co (ꢀ1.13 eV) is lower than that for MPA (ꢀ1.51
eV). We thus assume that the stronger S−metal cation
interaction promotes the photocatalytic activity.
Science Foundation (1708085QB43), Anhui Province Natural
Science Key Foundation (KJ2016A861), CAS Key Research
Program of Frontier Sciences (QYZDBꢀSSWꢀSLH018), CAS
Interdisciplinary Innovation Team, and Innovative Program of
Development Foundation of Hefei Center for Physical Science
and Technology (2016FXCX003). FTIR characterization was
performed at the Infrared Spectroscopy and Microspectroscopy
Endstation (BL01B) in the National Synchrotron Radiation
Laboratory (NSRL) in Hefei, China.
1
2
3
4
5
6
7
8
REFERENCES
1. Halmann, M. Nature 1978, 275, 115.
2. Inoue, T.; Fujishima, A.; Konishi, S.; Honda, K. Nature 1979,
277, 637.
3. Aresta, M.; Dibenedetto, A.; Angelini, A. Chem. Rev. 2014, 114,
1709.
4. White, J. L.; Baruch, M. F.; Pander, J. E.; Hu, Y.; Fortmeyer, I.
C.; Park, J. E.; Zhang, T.; Liao, K.; Gu, J.; Yan, Y.; Shaw, T. W.;
Abelev, E.; Bocarsly, A. B. Chem. Rev. 2015, 115, 12888.
5. Gao, C.; Wang, J.; Xu, H. X.; Xiong Y. J. Chem. Soc. Rev. 2017,
46, 2799.
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
6. Lingampalli, S. R.; Ayyub, M. M.; Rao, C. N. R. ACS Omega
2017, 2, 2740.
7. Bai, S.; Jiang, J.; Zhang Q.; Xiong, Y. J. Chem. Soc. Rev. 2015,
44, 2893.
8. Jiao, Y.; Zheng, Y.; Chen, P.; Jaroniec, M.; Qiao, S. Z. J. Am.
Chem. Soc. 2017, 139, 18093.
9. Li, G.; Jin, R. C. Acc. Chem. Res. 2013, 46, 1749.
10. Jin, R. C.; Zeng, C. J.; Zhou, M.; Chen, Y. X. Chem. Rev. 2016,
116, 10346.
11. Philip, R.; Chantharasupawong, P.; Qian, H. F.; Jin, R. C.;
Thomas, J. Nano Lett. 2012, 12, 4661.
12. Zhu, M. Z.; Aikens, C. M.; Hollander, F. J.; Schatz, G. C. Jin, R.
C. J. Am. Chem. Soc. 2008, 130, 5883.
13. Kogo, A.; Sakai, N.; Tatsuma, T. Nanoscale 2012, 4, 4217.
14. Wang, H. K.; Chen, F.Y.; Li, W.Y.; Tian, T. J. Power Sources
2015, 287, 150.
15. Sakai, N.; Tatsuma, T. Adv. Mater. 2010, 22, 3185.
16. Chen, Y.S.; Choi, H. B.; Kamat, P.V. J. Am. Chem. Soc. 2013,
135, 8822.
17. Chen, Y. S.; Kamat, P. V. J. Am. Chem. Soc. 2014, 136, 6075.
18. Gaur, S.; Wu, H.; Stanley, G. G.; More, K.; Kumar, C. S.;
Spivey, J. J. Catal. Today 2013, 208, 72.
19. Zhu, Y.; Qian, H.; Drake, B. A.; Jin, R. Angew. Chem. Int. Ed.
2010, 49, 1295.
20. Li, G.; Jiang, D. E.; Liu, C.; Yu, C.; Jin, R. J. Catal. 2013, 306,
177.
21. Chen, W.; Chen, S. Angew. Chem. Int. Ed. 2009, 48, 4386.
22. Shang , L.; Dong, S. J.; Nienhaus, G. U. Nano Today 2011, 6,
401.
23. Zhang, L. B.; Wang, E. K. Nano Today 2014, 9, 132.
24. Shang, L.; Dong, S. J. J. Mater. Chem. 2008, 18, 4636.
25. Durgadas, C. V.; Sharmaa C. P.; Sreenivasan, K. Analyst 2011,
136, 933.
26. Luo, Z. T.; Yuan, X.; Yu, Y.; Zhang, Q. B.; Leong, D. T.; Lee, J.
Y.; Xie, J. P. J. Am. Chem. Soc. 2012, 134, 16662.
27. Zhang, K.; Zhou, H. B.; Mei, Q. S.; Wang, S. H.; Guan, G. J.;
Liu, R. Y.; Zhang, J.; Zhang, Z. P. J. Am. Chem. Soc. 2011, 133,
8424.
CONCLUSION
In conclusion, we have developed
a
facile surface
modification method to construct catalytic sites on inert Au
NCs for photocatalytic CO2 reduction, in which various metal
cations can be grafted to Au NCs through bridging ligands.
As demonstrated by our characterizations, the metal cations
can accept photogenerated electrons from Au NCs through the
bridging ligands and serve as catalytic sites for CO2 reduction.
Given these functions, the selection of metal cations largely
impacts on the efficiency of electron transfer from lightꢀ
harvesting centers to catalytic sites, the ability of donating
electrons to reaction species, and the adsorption of reaction
species. The impact in turn alters the activity and selectivity
of photocatalytic CO2 reduction. This work demonstrates an
effective surface engineering strategy for bridging catalytic
sites with lightꢀharvesting centers, and offers a model system
for investigating the complex effects from catalytic sites
toward photocatalyst design. From the viewpoint of practical
applications, the agglomerationꢀinduced instability of
nanoclusters has to be overcome, to which the integration with
MOFs may provide a solution.
ASSOCIATED CONTENT
Supporting Information. Detailed experimental section,
characterization
characterizations. This material is available free of charge via the
Internet at http://pubs.acs.org.
methods,
and
additional
material
AUTHOR INFORMATION
Corresponding Author
*yjxiong@ustc.edu.cn
Author Contributions
‡These authors contributed equally.
The manuscript was written through contributions of all authors.
28. Le Guével, X; Spies, C.; Daum, N.; Jung, G.; Schneider, M.
Nano Res. 2012, 5, 379.
29. Li, H.; Cui, Z.; Han, C. Sensor. Actuat. B 2009, 143, 87.
30. Tamami, B.; Yeganeh, H. Eur. Polym. J. 2002, 38, 933.
31. Berner, S.; Lidbaum, H.; Ledung, G.; Åhlund, J.; Nilson, K.;
Schiessling, J.; Gelius, U.; Bäckvall, J.E.; Puglia, C.; Oscarsson,
S. Appl. Surf. Sci. 2007, 253, 7540.
32. Chen, W. B.; Tu, X. J.; Guo, X. Q. Chem. Commun. 2009, 13,
1736.
33. Xu Y. F.; Yang M. Z.; Chen B. X.; Wang X. D., Chen H. Y.; D.
B.; Su C. Y. J. Am. Chem. Soc. 2017, 139, 5660.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENT
This work was financially supported in part by National Key
R&D Program of China (2017YFA0207301), NSFC (21725102,
21603003, 21471141, U1532135), Anhui Provincial Natural
ACS Paragon Plus Environment

Page 7 of 8
Journal of the American Chemical Society
34. Lightcap I. V.; Kamat P. V. J. Am. Chem. Soc. 2012, 134, 7109.
35. Ishihara, T.; Akbay, T.; Furutani, H.; Takita Y. Solid State Ion.
1998, 113–115, 585.
1
2
3
4
5
6
7
8
36. Zhang, H. B.; Wei, J.; Dong, J. C.; Liu, G. G.; Shi, L.; An, P. F.;
Zhao, G. X.; Kong, J. T.; Wang, X. J.; Meng, X. G.; Zhang, J.;
Ye, J. H. Angew. Chem. Int. Ed. 2016, 55, 14310.
37. Miller, S. D.; Kitchin, J. R. Surf. Sci. 2009, 603, 794.
38. Li, J.; Croiset, E.; RicardezꢀSandoval, L. J. Mol. Catal. A: Chem.
2012, 365, 103.
39. Huo, C. F.; Li, Y. W.; Wang, J.; Jiao, H. J. Am. Chem. Soc.
2009, 131, 14721.
9
40. Xiao, F. X.; Zeng, Z.; Hsu, S. H.; Hung, S. F.; Chen, H. M.; Liu,
B. ACS Appl. Mater. Interf. 2015, 51, 28105.
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
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 8
1
2
SYNOPSIS TOC
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
ACS Paragon Plus Environment