Size dependence of atomically precise gold nanoclusters in chemoselective hydrogenation and active site structure

Article abstract of DOI:10.1021/cs500533h

We investigate the catalytic properties of water-soluble Au _n(SG)_m nanocluster catalysts (H-SG = glutathione) of different sizes, including Au₁₅(SG)₁₃, Au ₁₈(SG)₁₄, Au₂₅(SG)₁₈, Au ₃₈(SG)₂₄, and captopril-capped Au₂₅(Capt) ₁₈ nanoclusters. These Au_n(SR)_m nanoclusters (SR represents thiolate generally) are used as homogeneous catalysts (i.e., without supports) in the chemoselective hydrogenation of 4-nitrobenzaldehyde (4-NO₂PhCHO) to 4-nitrobenzyl alcohol (4-NO₂PhCH ₂OH) with ～100% selectivity in water using H₂ gas (20 bar) as the hydrogen source. These nanocluster catalysts, except Au ₁₈(SG)₁₄, remain intact after the catalytic reaction, evidenced by UV-vis spectra, which are characteristic of nanoclusters of each size and thus serve as spectroscopic "fingerprints". We observe a drastic size dependence and steric effect of protecting ligands on the gold nanocluster catalysts in the hydrogenation reaction. Density functional theory (DFT) modeling of the 4-nitrobenzaldehyde adsorption shows that both the -CHO and -NO₂ groups closely interact with the S-Au-S staples on the gold nanocluster surface. The adsorptions of the 4-nitrobenzaldehyde molecule on the four different sized Au_n(SR)_m nanoclusters are moderately strong and similar in strength. The DFT results suggest that the catalytic activity of the Au_n(SR)_m nanoclusters is primarily determined by the surface area of the Au nanocluster, consistent with the observed trend of the conversion of 4-nitrobenzaldehyde versus the cluster size. Overall, this work offers molecular insight into the hydrogenation of 4-nitrobenzaldehyde and the catalytically active site structure on gold nanocluster catalysts.

Full text of DOI:10.1021/cs500533h

Research Article
pubs.acs.org/acscatalysis
Size Dependence of Atomically Precise Gold Nanoclusters in
Chemoselective Hydrogenation and Active Site Structure
Gao Li,† De-en Jiang,‡ Santosh Kumar,† Yuxiang Chen,† and Rongchao Jin*,†
†Department of Chemistry, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213, United States
‡Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States
S
* Supporting Information
ABSTRACT: We investigate the catalytic properties of water-soluble Aun(SG)m
nanocluster catalysts (H-SG = glutathione) of different sizes, including Au15(SG)13,
Au18(SG)14, Au25(SG)18, Au38(SG)24, and captopril-capped Au25(Capt)18 nanoclusters.
These Aun(SR)m nanoclusters (SR represents thiolate generally) are used as
homogeneous catalysts (i.e., without supports) in the chemoselective hydrogenation of
4-nitrobenzaldehyde (4-NO2PhCHO) to 4-nitrobenzyl alcohol (4-NO2PhCH2OH) with
∼100% selectivity in water using H2 gas (20 bar) as the hydrogen source. These
nanocluster catalysts, except Au18(SG)14, remain intact after the catalytic reaction,
evidenced by UV−vis spectra, which are characteristic of nanoclusters of each size and
thus serve as spectroscopic “fingerprints”. We observe a drastic size dependence and steric
effect of protecting ligands on the gold nanocluster catalysts in the hydrogenation
reaction. Density functional theory (DFT) modeling of the 4-nitrobenzaldehyde adsorption shows that both the -CHO and
-NO2 groups closely interact with the S-Au-S staples on the gold nanocluster surface. The adsorptions of the 4-nitrobenzaldehyde
molecule on the four different sized Aun(SR)m nanoclusters are moderately strong and similar in strength. The DFT results
suggest that the catalytic activity of the Aun(SR)m nanoclusters is primarily determined by the surface area of the Au nanocluster,
consistent with the observed trend of the conversion of 4-nitrobenzaldehyde versus the cluster size. Overall, this work offers
molecular insight into the hydrogenation of 4-nitrobenzaldehyde and the catalytically active site structure on gold nanocluster
catalysts.
KEYWORDS: Au nanocluster, water-soluble, hydrogenation, chemoselective, size dependence
consists of a biicosahedral Au23 core and six dimeric Au2(SR)3
and three monomeric Au(SR)2 (-SR-Au-SR-) staples (Scheme
1f). These well-defined gold nanoclusters with known crystal
structures can be employed as model catalysts to investigate the
correlation between the catalytic performance and the gold
nanocatalyst structure.9 The unique cluster structures provide
an exciting opportunity to gain insight into the mechanism of
nanocatalysis.
In recent studies, conventional gold nanoparticles (poly-
dispersed ones, usually between 2 and 20 nm) have been widely
applied in the selective hydrogenation reactions.15−20 For
example, Corma and co-workers studied hydrogenation of
nitroaromatics with an aldehyde substituent by H2 and found
that the Au/TiO2 catalyst can selectively reduce nitro
compounds to amines instead of aldehyde to alcohol.21
Ultrasmall gold nanoparticles (1−2 nm) also exhibit good
catalytic activity in the hydrogenation processes.22−26 In our
recent work, we investigated the catalytic performance of the
unsupported and oxide-supported Aun(SPh)m nanoclusters (n =
25, 36, and 99, and m = 18, 24, and 42, respectively) in the
chemoselective hydrogenation of the reactants with both nitro
INTRODUCTION
■
Well-defined gold nanoclusters with ∼10 to a few hundred gold
atoms (i.e., core size ranging from 1 to 2 nm) constitute a new
and promising class of material and are attracting significant
research interest.1−8 These atomically precise gold nanoclusters
with an organic shell (e.g., thiolate), formulated as Aun(SR)m (n
and m represent the numbers of gold atoms and thiolate
molecules, respectively), have been applied in a range of
catalytic reactions, such as selective oxidation and hydro-
genation as well as carbon−carbon cross-coupling reac-
tions.9−12 It is well-known that the catalytic properties of
nanocatalysts are largely affected by the surface area of the
particle. Unlike conventional gold nanocatalysts (e.g., naked
nanoparticles) whose exact surface structures are largely
unknown, more and more atomically precise nanoclusters are
being crystallographically characterized.2 The structures of
Aun(SR)m generally exhibit a gold kernel protected by staplelike
surface motifs, and the surface area of these well-defined gold
nanoclusters is mainly determined by the staple shell
comprising both surface gold atoms and the protecting ligands.
Gold nanoclusters have unique Au core packing modes and
different protecting staple motifs; for example, the Au25(SR)18
nanocluster13 comprises an Au13 icosahedral core (or kernel)
and a staple shell comprising six Au2(SR)3 (-SR-Au-SR-Au-SR-)
dimeric staples (Scheme 1d), and the Au38(SR)24 nanocluster14
Received: April 21, 2014
Revised: June 11, 2014
© XXXX American Chemical Society
2463
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469

ACS Catalysis
Research Article
of H-SG were dissolved in 50 mL of methanol. The solution
was then cooled to 0 °C in an ice bath for 30 min. Then, 5 mL
of an aqueous solution of NaBH4 (99 mg, 2.6 mmol) was
poured rapidly into the mixture under vigorous stirring at 0 °C.
The mixture was allowed to react for an additional 2 h. The
reaction mixture was centrifuged to remove the insoluble
Au(I)-SG polymers. The supernatant was collected and
concentrated by rotary evaporation. The concentrated solution
was then precipitated by adding acetone. The precipitate was
collected and then etched with excess glutathione at 55 °C in
water for 3 h. The etched product was then centrifuged to
remove insoluble Au(I)-SG polymers as a precipitate. The
supernatant was precipitated with ethanol. The precipitate was
washed three times with ultrasonication and centrifugation to
obtain the clean product. This product was finally run through
the polyacrylamide gel electrophoresis (PAGE) gel to obtain
pure Au15(SG)13, Au18(SG)14, Au25(SG)18, and Au38(SG)24
nanoclusters, which showed up as four distinct bands in gel
analysis.
Purification by Polyacrylamide Gel Electrophoresis.
The PAGE experiment was conducted by using an OWL
P10DS-2 vertical gel electrophoresis system. The gel size was
20 cm × 20 cm. The separating and stacking gels were prepared
with monomer (acrylamide) concentrations of 30 and 4%,
respectively, with a cross-linker bis-acrylamide concentration of
4% for the best separation. The gel was eluted with 25 mM Tris
and 192 mM glycine buffer for 16 h at 300 V. The cluster was
dissolved in a 10% (v/v) glycerol/water mixture and loaded in
the gel. The well-resolved bands in the gel were then cut,
crushed, and soaked into water for 2 h. The nanoclusters diffuse
into water because of their water solubility. The nanocluster
solution with the gel matrix was then filtered with a 0.2 μm
filter and then acidified with 100 μL of 10% (v/v) glacial acetic
acid. This acidified solution was then concentrated by a cutoff
filter of 3 kDa. The concentrated solution was then precipitated
with ethanol and dried under vacuum to obtain pure
nanoclusters.
Scheme 1. Molecular Structures of (a) Glutathione and (b)
Captopril and Cluster Structures of (c) Au15(SR)13,27 (d)
14
Au25(SR)18,13 (e) Au18(SR)14,28 and (f) Au38(SR)24
a
Nanoclusters
a
Carbon and hydrogen atoms have been omitted for the sake of
clarity. Color code: green for core Au, yellow for surface Au, and blue
for S.
and aldehyde functional groups (e.g., nitrobenzaldehyde) to
alcohols (e.g., nitrobenzyl alcohol) using H2 as a hydrogen
source.25 Remarkably, a complete chemoselectivity for alcohol
(∼100%) was obtained, and oxide-supported Aun(SPh)m
nanoclusters offer a high level of conversion (80−99%) in
water (solvent); however, the unsupported nanoclusters were
less efficient and gave rise to only ∼10% conversion of
nitrobenzaldehyde in a homogeneous solution using a toluene/
ethanol mixture as the solvent.25 All of the works mentioned
above mainly focus on the organic soluble Aun(SR)m nano-
cluster catalysts supported on various oxides, but how the
water-soluble Aun(SR)m nanoclusters perform in aqueous-phase
homogeneous catalysis remains unclear.
Herein, we are motivated to explore the catalytic activity of
aqueous soluble Aun(SR)m nanocluster catalysts as homoge-
neous catalysts for chemoselective hydrogenation of 4-nitro-
benzaldehyde (4-NO2PhCHO). The nanoclusters studied in
this work include Au15(SG)13, Au18(SG)14, Au25(SG)18,
Au38(SG)24, and Au25(Capt)18 nanoclusters, where SG and
Capt stand for glutathione and captopril, respectively; note that
the Au25(SG)18 nanocluster throughout this work is anionic
{i.e., [Au25(SG)18]− in the native state}, whereas others are
charge neutral. The reaction is conducted in water (the
cheapest and environmentally friendly solvent) under relatively
mild conditions (80 °C, 20 bar of H2 gas). A complete
chemoselectivity to the 4-nitrobenzyl alcohol (4-
NO2PhCH2OH) product is obtained, and the catalytic activity
exhibits a distinct size dependence on the gold core (from Au15
to Au38, size range from ∼0.8 to ∼1.3 nm) and also a distinct
ligand effect [i.e., glutathione vs captopril (Scheme 1a,b)].
Therefore, both the gold core size and the protecting organic
ligands are important factors in the catalytic process. Further,
the available nanocluster structures13,14,27,28 (Scheme 1c−f)
permit density functional theory (DFT) modeling, and insights
into the reactant adsorption mode and catalytically active sites
are obtained for the selective hydrogenation reaction.
To analyze the degradation of the nanoclusters after catalysis,
the solutions of pure nanoclusters and the nanoclusters after
catalysis were run side by side in the gel and the resultant
nanoclusters were isolated from the gel.
Synthesis of Au25(Capt)18 Nanoclusters.30 HAuCl4·
3H2O (0.20 mmol, 78.7 mg) and TOABr (0.23 mmol, 126.8
mg) were first dissolved in 10 mL of methanol and vigorously
stirred for 15 min. After that, captopril (1 mmol, 217.2 mg) was
dissolved in 5 mL of methanol and that solution was rapidly
poured into the reaction mixture, and the mixture was stirred
for a further 30 min. The color of the solution quickly changed
to white. After 30 min, NaBH4 (2 mmol, 75.6 mg) was
dissolved in 5 mL of ice-cold water and rapidly added to the
solution in the flask under vigorous stirring. The color of the
solution immediately changed to brown-black. The reaction
was conducted further overnight. Then, the reaction mixture
was centrifuged to remove the insoluble Au(I)-Capt polymers.
The supernatant was collected and concentrated by rotary
evaporation. The concentrated solution was then precipitated
by adding acetone. The precipitate was extracted several times
with minimal amounts of methanol and then finally precipitated
with acetone and dried under vacuum.
EXPERIMENTAL METHODS
Characterization of Gold Nanoclusters. UV−vis spectra
of the clusters (dissolved in water) were acquired on a Hewlett-
Packard (HP) Agilent 8453 diode array spectrophotometer at
room temperature. Matrix-assisted laser desorption ionization
■
Synthesis of Aun(SG)m Nanoclusters. The Aun(SG)m
nanoclusters were prepared by a modified literature method.29
Typically, HAuCl4·3H2O (102 mg, 0.26 mmol) and 1.1 mmol
2464
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469

ACS Catalysis
Research Article
Figure 1. (a) Full gel image of the synthesized Aun(SG)m nanoclusters [including Au15(SG)13, Au18(SG)14, Au25(SG)18, and Au38(SG)24 nanoclusters
and another unidentified nanocluster] after PAGE purification. (b) UV−vis spectra and (c) Matrix-assisted laser desorption ionization mass spectra
of the Aun(SG)m nanoclusters.
(MALDI) mass spectrometry was performed with a PerSepti-
veBiosystems Voyager DE super-STR time-of-flight (TOF)
mass spectrometer. 2,4-Dihydroxybenzoic acid (DHB) was
used as the matrix during MALDI analysis. The isolated
nanoclusters and the matrix were dissolved in a 50% water/
methanol mixture to prepare a 1 μM solution (the same
concentration for both clusters and DHB). Equal volumes of
the matrix and the nanocluster solutions were mixed, and then
5 μL of this mixture was deposited on a MALDI steel plate and
dried in a vacumm oven prior to MALDI analysis.
oxidation of styrene: Au25(SR)18 > Au38(SR)24 > Au144(SR)60.39
Inspired by the results described above, we chose five different
aqueous soluble gold nanoclusters protected by functional
thiolate ligands, i.e., Au15(SG)13, Au18(SG)14, Au25(SG)18,
Au38(SG)24, and Au25(Capt)18, to investigate the size depend-
ence of the gold nanoclusters and the steric effect of the
protecting organic ligands in the chemoselective hydrogenation
of the reactants with nitro and aldehyde groups using water as
the solvent.
Nanocluster Synthesis and Characterization. The
atomically precise Aun(SG)m nanoclusters were synthesized
by a literature approach29 but with the following modifications.
Briefly, the Aun(SG)m nanocluster mixture was first obtained via
NaBH4 reduction of HAuCl4 in the presence of excess
glutathione. Then the reaction mixture was etched by excess
glutathione at 55 °C for 3 h, followed by purification via PAGE.
It is worth noting that we modified the previous approach29 by
incorporating the etching step, as this process is very useful for
size focusing of the polydispered gold nanoclusters into only
several robust sizes. Unlike the reported approach (e.g., the size
ranged from Au10 to Au39),29 in our modified synthesis, only
Au15(SG)13, Au18(SG)14, Au25(SG)18, and Au38(SG)24 nano-
clusters were present after size focusing, and they are readily
separated by PAGE (Figure 1a). The purified Aun(SG)m
nanoclusters were characterized by UV−vis spectra (Figure
1b) and matrix-assisted laser desorption ionization (MALDI)
mass spectra (Figure 1c) using 2,4-dihydroxybenzoic acid
(DHB) as the matrix. Aun(SG)m nanoclusters of each size
exhibit their characteristic UV−vis spectra with multiple peaks,
which can conveniently serve as the fingerprints for the purpose
of identification. Their sizes are assigned according to the
previous work by comparing the UV−vis spectra with standard
spectra.29 The MADLI mass spectra of Aun(SG)m nanoclusters
gave rise to broadened peaks due to MALDI-caused ligand loss
and other processes, but their peak masses roughly match the
expected values. Specifically, Au15(SG)13 nanoclusters show
saddlelike double peaks at 375/485 nm in the optical spectrum
(Figure 1a) and a mass peak centered at m/z ∼6.2k (intact
formula mass of 6.9k, where k = 1000) (Figure 1b). In the case
of Au18(SG)14 nanoclusters, a sharp optical peak at 570 nm and
a mass peak centered at m/z ∼6.8k are seen (intact formula
mass of 7.8k). The Au25(SG)18 nanoclusters give rise to a
distinct peak at 670 nm in the UV−vis spectrum and a mass
peak at m/z ∼9.4k (intact formula mass of 10.4k). A sharp peak
at 630 nm is shown in the UV−vis spectrum of Au38(SG)24, and
its mass peak is centered at m/z ∼13.6k (intact formula mass of
14.8k). The UV−vis spectra of the four gold nanoclusters [i.e.,
Typical Procedure for the Hydrogenation of Nitro-
benzaldehyde. In a typical selective hydrogenation reaction,
4-nitrobenzaldehyde (0.05 mmol), pyridine (0.1 mmol),
Aun(SG)m nanoclusters (∼0.1 μmol), and 1 mL of water
were added to a reactor (Parr Instrument Co., 22 mL capacity,
series 4700) under 20 bar of H2. The reaction mixture was kept
at 80 °C. After the reaction, the mixture was extracted with
ethyl acetate. The crude product was obtained after the removal
of solvent. The conversion of 4-nitrobenzaldehyde and the
1
selectivity for 4-nitrobenzyl alcohol were determined by H
nuclear magnetic resonance (NMR) (300 MHz) spectroscopic
analysis. After the hydrogenation reaction, the gold nanocluster
catalyst was analyzed by the UV−vis spectrum and PAGE.
Computational Details. Parallel, resolution-of-identity
density functional theory (DFT) calculations with the
Perdew−Burke−Ernzerhof form of the generalized gradient
approximation for electron exchange and correlation,31 the
def2-SV(P) basis sets, and empirical dispersion correction32
were performed with the quantum chemistry program
Turbomole version 6.5.33 Effective core potentials that have
19 valence electrons and include scalar relativistic corrections
were used for Au.34 To calculate the surface area of a cluster in
a solvent such as water, one constructs a cavity by uniting atom-
centered spheres with radii of ri and rsolv for all atoms i in the
cluster.35 Default values for ri and rsolv from Turbomole version
6.5 are used. The surface area of the cavity is defined as the
surface area of the cluster.
RESULTS AND DISCUSSION
■
The catalytic reactivity of metal nanoclusters mainly depends
on the metal core, such as the surface-to-volume ratio, size,
morphology, stability, etc.36−40 Meanwhile, the organic func-
tional shell can affect the electronic properties of the metal core
and further influence the catalytic properties of the metal
clusters, leading to peculiar selectivity in catalytic reactions. For
example, it was found that the catalytic activity of different sized
nanoclusters decreases in the following order in the selective
2465
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469

ACS Catalysis
Research Article
Au15(SG)13, Au18(SG)14, Au25(SG)18, and Au38(SG)24] match
well with the literature.29 The as-obtained pure Aun(SG)m
nanoclusters are then utilized in catalytic hydrogenation.
that the chemoselectivity for the product of 4-aminobenzyl
alcohol is ∼100% under the catalysis by the Aun(SG)m
nanocluster catalysts. This complete chemoselectivity is the
same as in our previous work involving Aun(SPh)m nano-
clusters,25 indicating that the surface ligands (SG vs SPh) do
not affect the catalytic selectivity (i.e., ∼100% in both cases).
Size Dependence of Nanoclusters. With respect to the
size dependence of nanoclusters, the conversion of 4-nitro-
benzaldehyde when catalyzed by Au15(SG)13 nanoclusters is
relatively low, ∼7.8% (Table 1, entry 1). The catalytic activity
(hereafter termed the conversion of 4-nitrobenzaldehyde) of
the gold nanoclusters increases to 10.2% [Au18(SG)14], 16.0%
[Au25(SG)18], and 20.6% [Au38(SG)24] when equal numbers of
moles of nanoclusters are used (Table 1, entries 2−4). Thus,
the catalytic performance of the gold nanoclusters decreases in
Table 1. Hydrogenation of 4-Nitrobenzaldehyde Catalyzed
by Unsupported Aqueous Soluble Aun(SG)m Nanoclusters as
a
Catalysts
b
selectivity (%)
b
entry
catalyst
conversion (%)
1
2
3
d
d
d
d
d
d
d
d
1
2
3
4
5
Au15(SG)13
Au18(SG)14
Au25(SG)18
Au38(SG)24
glutathione
7.8
10.2
16.0
20.6
100
100
100
100
−
nd
nd
nd
nd
−
nd
nd
nd
nd
−
the following order: Au38(SG)24 > Au25(SG)18 > Au18(SG)14
>
Au15(SG)13. The observed trend from 7.8 to 20.6% (a nearly 3-
fold increase) with increasing cluster size is distinct and
interesting. It is worth noting that the selectivity for 4-
nitrobenzyl alcohol is constantly 100% in the four hydro-
genation reactions described above. No reaction was observed
in the blank or control experiments [i.e., using glutathione and
captopril ligands as the catalysts under otherwise identical
reaction conditions (Table 1, entries 6 and 7, respectively)],
indicating that the gold nanoclusters are essential and
constitute the catalytically active sites in the hydrogenation
reaction. We also note that the catalytic performance of
aqueous soluble Au25(SG)18 is much better than that of the
hydrophobic Au25(SPh)18 catalyst; for instance, the turnover
frequency (TOF) of Au25(SG)18 is ∼20 h−1 (i.e., moles of
alcohol product per hour per mole of gold nanocluster) versus
5.6 h−1 [for unsupported Au25(SPh)18 clusters25], which is due
to the enhanced activation of H2 in a water solution compared
to an organic solution, as evidenced by the higher hydro-
genation activity of the gold nanoclusters [e.g., CeO2-supported
Au99(SPh)42 catalyst] in water than in the organic solution
(93.1% vs 78.4%).
c
nr
a
Reaction conditions: 0.1 μmol of unsupported Aun(SG)m nanocluster
catalyst in 1 mL of water, 0.05 mmol of 4-nitrobenzaldehyde (molar
ratio of nitrobenzaldehyde to Au clusters of 500), and 0.1 mmol of
b
pyridine under 20 bar of H2 at 80 °C for 4 h. The conversion of 4-
nitrobenzaldehyde and selectivity for 4-nitrobenzyl alcohol (1) were
c
d
1
determined by the H NMR spectrum. No reaction. Not detected.
Chemoselective Hydrogenation. Homogeneous hydro-
genation was conducted under 20 bar of H2 gas at 80 °C for 4 h
in water (for details, see Experimental Methods). After the
1
catalytic reaction, the crude product was analyzed by H NMR
spectroscopy; only two species, i.e., the residual reactant (4-
nitrobenzaldehyde) and only one product [4-nitrobenzyl
alcohol (1)], are observed in the NMR spectrum (Figure 2).
The Aun(SG)m nanocluster catalysts [i.e., Au15(SG)13,
Au18(SG)14, Au25(SG)18, and Au38(SG)24], except Au18(SG)14,
are found to be stable and remain intact after the selective
hydrogenation, which is supported by the unchanged UV−vis
spectra after the hydrogenation reactions (Figure S1 of the
Supporting Information). Further analysis of the gold nano-
cluster catalysts [i.e., Aun(SG)m after the hydrogenation
reactions] by PAGE purification is shown in Figure 3. Only
one band appears in the PAGE gel, implying the purity of the
nanoclusters. To further confirm the purity of the gel band, we
measured the UV−vis spectra at the top and bottom positions
of the gel band and found that the Au15(SG)13, Au25(SG)18, and
Au38(SG)24 nanoclusters remain intact, judging by the optical
spectra being identical to those of the fresh nanoclusters
(Figure 3a,c,d vs Figure 1b), while Au18(SG)14 is less stable and
changes to different sized nanoclusters after the catalytic
hydrogenation process (Figure 3b vs Figure 1b, blue line).
Ligand Steric Effect. To investigate the potential steric
effect of the protecting thiolate ligand, we further chose the
captopril-protected Au25(Capt)18 nanocluster and compared its
performance with that of Au25(SG)18. The captopril ligand is
less bulky than glutathione (Scheme 1a). The synthesis of the
Au25(Capt)18 nanocluster follows a reported protocol30 (see
Experimental Methods). The as-obtained Au25(Capt)18 nano-
cluster shows a UV−vis spectrum (Figure 4a) identical to that
of Au25(SG)18; note that the slight change in the UV region is
1
Figure 2. H NMR spectrum of the crude product after the catalytic
reaction catalyzed by the Au25(SG)18 nanocluster catalyst. Only
residual reactant 4-nitrobenzaldehyde (-CHO at 10.18, the proton on
the phenyl ring at 8.44 and 8.41, 8.11, and 8.08 ppm) and the exclusive
4-nitrobenzyl alcohol product (-CH2 at 4.87, the proton on phenyl
ring at 8.26 and 8.23, 7.58, and 7.55 ppm) are found in the spectrum.
Other possible products [e.g., 4-aminobenzaldehyde (2) and 4-
aminobenzyl alcohol (3)] are not detected. As shown in Figure
2, the residual 4-nitrobenzaldehyde shows signals at δ 10.18
(-CHO, 1H, s), 8.42 (2H, d), and 8.10 (2H, d), and the
chemoselective hydrogenation product (4-nitrobenzyl alcohol)
shows signals at δ 8.25 (2H, d), 7.56 (2H, d), and 4.87 ppm
(-CH2OH, 2H, s). The extra peak at 7.28 ppm belongs to the
residual solvent peak of CDCl3. The NMR analysis indicates
2466
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469

ACS Catalysis
Research Article
Table 2. Hydrogenation of 4-Nitrobenzaldehyde Catalyzed
by Unsupported Aqueous Soluble Au25(SR)18 (SR = SG or
a
Capt) Nanoclusters and the Polymeric Au(I)-SG Complex
b
selectivity (%)
b
entry
catalyst
conversion (%)
1
2
3
d
d
d
d
d
1
2
3
4
Au25(Capt)18
Au25(SG)18
Au(I)-SG
23.3
16.0
65.6
100
100
24
nd
nd
76
−
nd
nd
nd
−
c
captopril
nr
−
a
Reaction conditions: 0.1 μmol of unsupported Au25(SR)18 nano-
cluster or Au(I)-SG complex catalyst in 1 mL of water, 0.05 mmol of
4-nitrobenzaldehyde, and 0.1 mmol of pyridine under 20 bar a H2 at
b
80 °C for 4 h. The conversion of 4-nitrobenzaldehyde and selectivity
for 4-nitrobenzyl alcohol (1) and 4-aminobenzaldehyde (2) were
c
d
1
determined by the H NMR spectrum. No reaction. Not detected.
the possibility that the high selectivity (∼100%) might be due
to the relatively low conversion in the reaction, we adjusted the
reaction conditions to improve the conversion by increasing the
amount of Au25(Capt)18 catalyst from 0.1 to 0.2 μmol and
decreasing the amount of nitrobenzaldehyde from 0.05 to 0.025
mmol (with other reaction conditions kept the same as
described in footnote a of Table 1). Under such conditions, the
conversion was largely increased to 76.9% at 4 h and 83.2% at 9
h (Figure 5), whereas the chemoselectivity for the alcohol
Figure 3. (a−d) Gel images and corresponding UV−vis spectra (at
different positions of the gel) of the gold nanocluster catalysts after the
hydrogenation: (a) Au15(SG)13, (b) Au18(SG)14, (c) Au25(SG)18, and
(d) Au38(SG)24.
caused by the trace organic residues (e.g., pyridine and the
catalytic product). MALDI mass spectra show a sharp peak
around 8.4 kDa (Figure 4b), in which some ligands are lost due
to MALDI process (intact mass of 8.8k). The hydrogenation
reaction catalyzed by Au25(Capt)18 nanoclusters was conducted
under conditions identical to those of the Aun(SG)m cases. A
high conversion [23.3% (Table 2, entry 1)] of 4-nitro-
benzaldehyde with 100% selectivity for the 4-nitrobenzyl
alcohol product was obtained, which is indeed better than
those of the other four glutathione-protected Aun(SG)m
nanoclusters. The TOF value of the Au25(Capt)18 catalyst
was increased to ∼30 h−1. The UV−vis spectra and the MALDI
mass spectra (Figure 4) indicate that the Au25(Capt)18
nanocluster catalyst is intact after the hydrogenation reaction.
The higher catalytic activity of the Au25(Capt)18 nanocluster
catalyst can be attributed to the steric effect [i.e., less bulky
captopril compared to the -SG ligand on the surface of the
Aun(SG)m nanoclusters]. We also found that the Au(I)-SG
complex can catalyze the reaction (Figure S2 of the Supporting
Information) but the ligand itself cannot. The Au(I)-SG
complex gave rise to a 65.6% conversion and a 76% selectivity
for the 4-aminobenzaldehyde product and 24% for 4-nitro-
benzyl alcohol (Table 2, entry 3).
Figure 5. Time-dependent conversion of 4-nitrobenzaldehyde as a
function of reaction time over unsupported Au25(Capt)18. Reaction
conditions: 0.2 μmol of Au25(Capt)18 nanocluster catalyst in 2 mL of
water, 0.025 mmol of 4-nitrobenzaldehyde, and 0.1 mmol of pyridine
under 20 bar of H2 at 80 °C for 9 h.
product is constantly at ∼100%. Thus, the high chemo-
selectivity is not due to the relatively low conversion in the
reaction; instead, it is the intrinsic characteristics of the catalyst.
It is worth noting that the Au25(Capt)18 catalyst still remains
intact after the catalytic reactions.
Further, the selective hydrogenation kinetics was investigated
by monitoring the conversion of 4-nitrobenzaldehyde using
Au25(Capt)18 as a catalyst at different time intervals. To rule out
Figure 4. (a) UV−vis and (b) MALDI mass spectrometry spectra of the Au25(Capt)18 nanocluster catalyst before and after the hydrogenation
reaction. (c) Gel images and corresponding UV−vis spectra (at different positions) of the gold nanocluster catalysts after the hydrogenation.
2467
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469

ACS Catalysis
Research Article
Structure−Property Correlation. Au25(SCH2CH2Ph)18
and Au38(SCH2CH2Ph)24 nanoclusters have been crystallo-
graphically characterized previously,13,14 and the Au15(SCH3)13
and Au18(SCH3)14 structures have been predicted by DFT
calculation.27,28 Thus, we chose these four cluster structures as
models for the DFT simulations on the 4-nitrobenzaldehyde
adsorption mode and further correlated the catalytic activity
with the cluster structures. It is worth noting that the capping
thiolate ligand (-SG) was replaced by -SCH3 only for the ease
of DFT calculations. Calculations from dispersion-corrected
DFT suggest that the adsorption energies of the 4-nitro-
benzaldehyde (4-NO2PhCHO) molecule on the four
Aun(SCH3)m nanoclusters are moderately strong and similar.
The 4-nitrobenzaldehyde molecule has the strongest adsorption
on the Au25(SCH3)18 nanoclusters, probably because of the
Table 3. Adsorption Energy of the 4-Nitrobenzaldehyde
Molecule on Various Aun(SCH3)m Nanoclusters
a
entry
Au nanocluster
Ead‑max (eV)
Ead‑average (eV)
1
2
3
4
Au15(SCH3)13
Au18(SCH3)14
Au25(SCH3)18
Au38(SCH3)24
−0.99
−0.94
−1.05
−0.86
−0.88
−0.92
−1.03
−0.86
a
Both the adsorption energy for the maximal interaction (Ead‑max) and
the average interaction over several adsorption sites (Ead‑average) are
listed. They were computed at the dispersion-corrected DFT-PBE
level (DFT-PBE-D3) with the def2-SV(P) basis sets in Turbomole. A
negative value means a favorable interaction.
hyde molecule on the basis of the same number of moles of the
Aun(SCH3)m nanoclusters against the surface area of the cluster
(Figure 7) and found that indeed the conversion increases with
the surface area of the gold cluster (see Computational Details
for how the surface area is calculated).
extra negative charge on the native Au25(SCH3)18 cluster13
−
that provides higher polarizability to a strongly polar group
such as -NO2 on the molecule. Figure 6 shows an adsorption
Figure 7. Conversion of 4-nitrobenzaldehyde vs the surface area of the
Aun(SCH3)m nanoclusters.
On the basis of the experimental results discussed above and
DFT simulations, here we briefly discuss the plausible
mechanism for selective hydrogenation. The 4-nitrobenzalde-
hyde (4-NO2PhCHO) molecules are first adsorbed onto the
Aux(SR)x+1 staples of the Aun(SR)m nanocluster; both the -NO2
and -CHO groups strongly interact with the S-Au-S staple
motifs on the cluster surface. At the same time, H2 is activated
by the gold nanoclusters under the assistance of pyridine. Then
the -CHO group is selectively hydrogenated via the surface Au
atoms. Finally, the 4-nitrobenzyl alcohol molecule is formed
and desorbed from the Au nanocluster catalyst.
Figure 6. 4-Nitrobenzaldehyde molecule adsorbed on
a
−
Au25(SCH3)18 nanocluster [top view (a) and side view (b)].
Au38(SCH3)24 nanocluster [top view (c) and side view (d)]. The
color code is as follows: yellow for Au, blue for S, gray for C, white for
H, red for O, and pink for N.
geometry of the 4-nitrobenzaldehyde molecule on
−
Au25(SCH3)18 and Au38(SCH3)24. One can see that the
CONCLUSIONS
■
-CHO and -NO2 groups of the molecule are in close contact
with the S-Au-S groups of the cluster surface; this interaction is
mostly likely due to the polar organic groups inducing dipoles
in the S-Au-S motifs. In addition, there are multiple sites on
each of the clusters that can adsorb the molecule. The variation
in adsorption energy is not large, as indicated by the small
difference between the maximal adsorption energy and the
average adsorption energy (Table 3); the energy difference is
only 0.02 eV for Au18(SCH3)14, Au25(SCH3)18, and
Au38(SCH3)24 nanoclusters and 0.11 eV for Au15(SCH3)13.
The fact that the Au nanoclusters have multiple sites with
similar adsorption energies suggests that the catalytic
conversion of the 4-nitrobenzaldehyde molecule on a cluster
should correlate with the surface area of the Au nanocluster.
We therefore plotted the conversion of the 4-nitrobenzalde-
In summary, we have investigated the catalytic properties of
water-solube Aun(SR)m nanocluster catalysts [including
Au15(SG)13, Au18(SG)14, Au25(SG)18, Au38(SG)24, and
Au25(Capt)18] in the homogeneous chemoselective hydro-
genation of 4-nitrobenzaldehyde in water. The well-defined
Aun(SR)m nanocluster catalysts are essentially intact after the
chemoselective hydrogenation, except for the Au18(SG)14
nanocluster. Overall, a drastic size dependence and the ligand’s
steric effects are observed in gold nanocluster-catalyzed
hydrogenation reactions. The catalytic activity (based on the
conversion of 4-nitrobenzaldehyde) of the gold nanoclusters
increases with an increasing core size: Au15(SG)13 < Au18(SG)14
< Au25(SG)18 < Au38(SG)24. On the other hand, the catalytic
properties of the gold nanoclusters with the less bulky ligand
are found to be much better [e.g., Au25(Capt)18 > Au25(SG)18].
2468
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469

ACS Catalysis
Research Article
(15) Mitsudome, T.; Kaneda, K. Green Chem. 2013, 15, 2636−2654.
(16) Pan, M.; Brush, A. J.; Pozun, Z. D.; Ham, H. C.; Yu, W.;
Henkelman, G.; Hwang, G. S.; Mullins, C. B. Chem. Soc. Rev. 2013, 42,
5002−5013.
DFT calculations show that both -CHO and -NO2 groups of
the 4-nitrobenzaldehyde molecule closely interact with the
cluster surface staples (S-Au-S); the adsorption energies of the
4-nitrobenzaldehyde molecule on the four different sized
Aun(SR)m nanoclusters are moderately strong (−0.86 to
−1.05 eV). The increased catalytic activity of larger gold
nanoclusters is consistent with the surface area of the gold
nanoclusters. Overall, this work reveals the molecular level
insight into the hydrogenation reaction of nitrobenzaldehyde
and the catalytically active site structure on gold nanocluster
catalysts.
(17) Yamane, Y.; Liu, X.; Hamasaki, A.; Ishida, T.; Haruta, M.;
Yokoyama, T.; Tokunaga, M. Org. Lett. 2009, 11, 5162−5165.
(18) Corma, A.; Serna, P. Science 2006, 313, 332−334.
(19) Cano, I.; Chapman, A. M.; Urakawa, A.; van Leeuwen, M. J. Am.
Chem. Soc. 2014, 136, 2520−2528.
(20) Ren, D.; He, L.; Yu, L.; Ding, R.; Liu, Y.; Cao, Y.; He, H.; Fan,
K. J. Am. Chem. Soc. 2012, 134, 17592−17598.
(21) Santos, L. L.; Serna, P.; Corma, A. Chem.Eur. J. 2009, 15,
8196−8203.
(22) Zhu, Y.; Qian, H.; Drake, B. A.; Jin, R. Angew. Chem., Int. Ed.
2010, 49, 1295−1298.
ASSOCIATED CONTENT
* Supporting Information
■
S
(23) Yamamoto, H.; Yano, H.; Kouchi, H.; Obora, Y.; Arakawa, R.;
Kawasaki, H. Nanoscale 2012, 4, 4148−4154.
UV−vis spectra of the Aun(SG)m nanocluster catalysts before
and after hydrogenation, and NMR spectrum of the reaction
product when catalyzed by the Au(I):SG polymer. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(24) Shivhare, A.; Ambrose, S. J.; Zhang, H.; Purves, R. W.; Scot, R.
W. J. Chem. Commun. 2013, 49, 276−278.
(25) Li, G.; Zeng, C.; Jin, R. J. Am. Chem. Soc. 2014, 136, 3673−3679.
(26) Zhu, Y.; Wu, Z.; Gayathri, G. C.; Qian, H.; Gil, R. R.; Jin, R. J.
Catal. 2010, 271, 155−160.
(27) Jiang, D.; Overbury, S. H.; Dai, S. J. Am. Chem. Soc. 2013, 135,
8786−8789.
AUTHOR INFORMATION
Corresponding Author
*E-mail: rongchao@andrew.cmu.edu.
■
(28) Tlahuice, A.; Garzon
3737−3740.
́
, I. L. Phys. Chem. Chem. Phys. 2012, 14,
Notes
(29) Negishi, Y.; Nobusada, K.; Tsukuda, T. J. Am. Chem. Soc. 2005,
127, 5261−5270.
The authors declare no competing financial interest.
(30) Kumar, S.; Jin, R. Nanoscale 2012, 4, 4222−4227.
(31) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77,
3865−3868.
ACKNOWLEDGMENTS
■
This work is financially supported by the U.S. Department of
Energy Office of Basic Energy Sciences (Grant DE-FG02-
12ER16354). The DFT calculations were supported by the U.S.
Department of Energy, Office of Science, Basic Energy
Sciences, Chemical Sciences, Geosciences, and Biosciences
Division, and used resources of the National Energy Research
Scientific Computing Center, which is supported by the Office
of Science of the U.S. Department of Energy (Grant DE-AC02-
05CH11231).
(32) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. J. Chem. Phys.
2010, 132, 154104.
(33) Ahlrichs, R.; Bar, M.; Haser, M.; Horn, H.; Kolmel, C. Chem.
Phys. Lett. 1989, 162, 165−169.
(34) Andrae, D.; Haussermann, U.; Dolg, M.; Stoll, H.; Preuss, H.
Theor. Chim. Acta 1990, 77, 123−141.
(35) Klamt, A. WIREs Computational Molecular Science 2011, 1, 699−
709.
(36) Haruta, M. Catal. Today 1997, 36, 153−166.
(37) Fenger, R.; Fertitta, E.; Kirmse, H.; Thunemann, A. F.;
̈
Rademann, K. Phys. Chem. Chem. Phys. 2012, 14, 9343−9349.
REFERENCES
■
(38) Nikolaev, S. A.; Smirnov, V. V. Catal. Today 2009, 147, S336−
(1) Yamazoe, S.; Koyasu, K.; Tsukuda, T. Acc. Chem. Res. 2014, 47,
816−824.
(2) Qian, H.; Zhu, M.; Wu, Z.; Jin, R. Acc. Chem. Res. 2012, 45,
1470−1479.
S341.
(39) Zhu, Y.; Qian, H.; Jin, R. Chem.Eur. J. 2010, 16, 11455−
11462.
(40) Liu, Y.; Tsunoyama, H.; Akita, T.; Xie, S.; Tsukuda, T. ACS
Catal. 2011, 1, 2−6.
(3) Yu, Y.; Luo, Z.; Chevrier, D. M.; Leong, D. T.; Zhang, P.; Jiang,
D.; Xie, J. J. Am. Chem. Soc. 2014, 136, 1246−1249.
(4) Yi, C.; Tofanelli, M. A.; Ackerson, C. J.; Knappenberger, K. L., Jr.
J. Am. Chem. Soc. 2013, 135, 18222−18228.
(5) Kwak, K.; Kumar, S. S.; Pyo, K.; Lee, D. ACS Nano 2014, 8, 671−
679.
(6) Mathew, A.; Natarajan, G.; Lehtovaara, L.; Hakkinen, H.; Kumar,
̈
R. M.; Subramanian, V.; Jaleel, A.; Pradeep, T. ACS Nano 2014, 8,
139−152.
(7) Kwak, K.; Kumar, S. S.; Lee, D. Nanoscale 2012, 4, 4240−4246.
(8) Sakai, N.; Tatsuma, T. Adv. Mater. 2010, 22, 3185−3188.
(9) Li, G.; Jin, R. Acc. Chem. Res. 2013, 46, 1749−1758.
(10) Li, G.; Jiang, D.; Liu, C.; Yu, C.; Jin, R. J. Catal. 2013, 306, 177−
183.
(11) Li, G.; Liu, C.; Lei, Y.; Jin, R. Chem. Commun. 2012, 48, 12005−
12007.
(12) Liu, J.; Krishna, K. S.; Losovyj, Y. B.; Chattopadhyay, S.; Lozova,
N.; Miller, J. T.; Spivey, J. J.; Kumar, C. Chem.Eur. J. 2013, 19,
10201−10208.
(13) Zhu, M.; Aikens, C. M.; Hollander, F. J.; Schatz, G. C.; Jin, R. J.
Am. Chem. Soc. 2008, 130, 5883−5885.
(14) Qian, H.; Eckenhoff, W. T.; Zhu, Y.; Pintauer, T.; Jin, R. J. Am.
Chem. Soc. 2010, 132, 8280−8281.
2469
dx.doi.org/10.1021/cs500533h | ACS Catal. 2014, 4, 2463−2469