Ligand-stabilized and atomically precise gold nanocluster catalysis: A case study for correlating fundamental electronic properties with catalysis

Article abstract of DOI:10.1002/chem.201300600

We present results from our investigations into correlating the styrene-oxidation catalysis of atomically precise mixed-ligand biicosahedral-structure [Au₂₅(PPh₃)₁₀(SC ₁₂H₂₅)₅Cl₂

Full text of DOI:10.1002/chem.201300600

FULL PAPER
DOI: 10.1002/chem.201300600
Ligand-Stabilized and Atomically Precise Gold Nanocluster Catalysis: A
Case Study for Correlating Fundamental Electronic Properties with Catalysis
Jing Liu,[a, b] Katla Sai Krishna,[a, b] Yaroslav B. Losovyj,[a, b] Soma Chattopadhyay,[c]
Natalia Lozova,[a] Jeffrey T. Miller,[d] James J. Spivey,[b] and Challa S. S. R. Kumar*[a, b]
Abstract: We present results from our
investigations into correlating the styr-
ene-oxidation catalysis of atomically
precise mixed-ligand biicosahedral-
Au bond contraction and higher d-
band vacancies in both the ligand-stabi-
lized Au clusters. The ligands were
found not only to act as colloidal sta-
bilizers, but also as d-band electron ac-
ceptor for Au atoms. Au25-bi clusters
have a higher first-shell Au coordina-
tion number than Au25-i, whereas Au25-
bi and Au25-i clusters have the same
number of Au atoms. The UPS re-
vealed a trend of narrower d-band
width, with apparent d-band spin–orbit
splitting and higher binding energy of
d-band center position for Au25-bi and
Au25-i. We propose that the differences
in their d-band unoccupied state popu-
lation are likely to be responsible for
differences in their catalytic activity
and selectivity. The findings reported
herein help to understand the catalysis
of atomically precise ligand-stabilized
metal clusters by correlating their
atomic or electronic properties with
catalytic activity.
structure [Au
(Au25-bi) and thiol-stabilized icosahe-
dral core–shell-structure [Au25-
(SCH2CH2Ph)18]À (Au25-i) clusters with
(SC12H25)5Cl2]2+
25ACHTUNGTRENNUNG(PPh3)10ACHUTNGTRENNUGN
ACHTUNGTRENNUNG
their electronic and atomic structure by
using a combination of synchrotron ra-
diation-based X-ray absorption fine-
structure spectroscopy (XAFS) and ul-
traviolet photoemission spectroscopy
(UPS). Compared to bulk Au, XAFS
Keywords: cluster compounds
·
gold · nanostructures · ultraviolet
photoemission spectra · X-ray ab-
sorption fine structure
À
revealed low Au–Au coordination, Au
Introduction
particle catalysts, there has been significant work on gold
nanoparticles, especially as catalysts for liquid–solid and
gas–solid oxidation reactions.[9,10] Their size- and shape-de-
pendent catalytic properties have also been examined.[11–15]
For instance, Goodman and co-workers have reported an
unusual size-dependent catalytic behavior of chemical-
vapor-deposited gold clusters (1–6 nm) on single-crystalline
titania surface for CO oxidation.[16] Theoretical studies by
Landman and co-workers[17] have shown that the Au8 is the
smallest catalytically active size for the CO oxidation. Sever-
al attempts have been made to investigate their structure–
catalytic-activity relationships.[18] However, there have been
very few attempts to directly correlate their fundamental
electronic structure with their catalytic properties.[19–21] One
of the main reasons for this is the polydispersity issue of
gold nanoparticles. With recent successes in the synthesis of
atomically precise ultra-small gold-cluster catalysts, there is
now an opportunity to obtain new knowledge about their
electronic structure–catalytic-activity relationships.[12,22–24]
There are a number of experimental and computational
studies on the catalytic activity of atomically precise gold
clusters.[2,24–28] The CO oxidation was explored by Spivey
and co-workers[2] for ligand-stabilized atomically precise
Au38 clusters supported on titania. Their results showed that
supported Au clusters of controllable size can be prepared
by thiol-ligated solution-based method. Recently, Jin and
co-workers[24] have studied and compared the catalytic activ-
ity of silica-supported gold clusters Au25, Au38, and Au144 for
styrene oxidation by using both dioxygen and tert-butyl hy-
Colloidal nanoparticle catalysts continue to attract attention
both for solution[1] and for gas-phase catalysis.[2,3] The influ-
ence of their size, shape, and oxidation state on their cataly-
sis was previously investigated.[4–7] Their catalytic activity
and selectivity are generally known to increase when the
cluster size decreases to about 1 nm or below.[8] Of the nano-
[a] Dr. J. Liu, Dr. K. S. Krishna, Dr. Y. B. Losovyj, Dr. N. Lozova,
Dr. C. S. S. R. Kumar
Center for Advanced Microstructures and Devices (CAMD)
Louisiana State University, Baton Rouge, LA 70806 (USA)
[b] Dr. J. Liu, Dr. K. S. Krishna, Dr. Y. B. Losovyj, Prof. J. J. Spivey,
Dr. C. S. S. R. Kumar
Center for Atomic-Level Catalyst Design
324, Cain Department of Chemical Engineering
Louisiana State University, 110 Chemical Engineering
South Stadium Road, Baton Rouge, LA 70803 (USA)
E-mail: ckumar1@lsu.edu
[c] Dr. S. Chattopadhyay
CSRRI-IIT, MRCAT, Sector 10
Bldg 433B, Argonne National Laboratory
Lemont, IL 60439 (USA)
Physics Department, Illinois Institute of Technology
Chicago, IL 60616 (USA)
[d] Prof. J. T. Miller
CSE Division, Argonne National Laboratory
9700 South Cass Avenue
Argonne, Illinois 60439-4837 (USA)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201300600.
Chem. Eur. J. 2013, 19, 10201 – 10208
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
10201

drogen peroxide (TBHP) as oxidants. In a different study,
they also compared the catalytic activity of ceria-supported
Au25 nanorods and nanospheres[29] for styrene oxidation and
benzalacetone hydrogenation. However, their study was
mainly focused on elucidating the effect of calcination on
the product yields and a comparison of their activity and se-
lectivity. Tsukuda and co-workers reported that selectivity
for styrene epoxide reached 92% for Au25 clusters support-
ed on hydroxyapatite when TBHP was used as oxidant.[25] In
a similar approach, selective oxidation of styrene in the
presence of dioxygen by Au55 clusters was reported by Lam-
bert and co-workers.[26] They found a sharp size threshold
for activity, with particles larger than 2 nm being completely
inactive. Their observations suggest that catalytic activity
arises from the altered electronic structure intrinsic to small
gold nanoparticles. Further progress has been made by Jiang
and co-workers[30] by using DFT calculations to show that
bare Au38 (face-centered cubic (fcc) structure) is more cata-
lytically active than Au55 (fcc structure) for styrene oxida-
tion due to the high electropositive nature of Au atoms on
the Au38 surface.
What is clearly lacking is a concerted effort to correlate
the fundamental electronic structure of these atomically pre-
cise gold clusters with their catalysis by using a combination
of spectroscopic methods, such as X-ray absorption fine
structure (XAFS) and ultraviolet photoemission spectrosco-
py (UPS). Such a correlation is of paramount importance
and a first step towards meeting one of the grand science
challenges to understand how remarkable catalytic proper-
ties emerge from complex correlations of atomic or elec-
tronic constituents of the catalysts and how one can control
their activity and selectivity.[31] We, therefore, not only stud-
ied the styrene-oxidation reaction on atomically precise
Results and Discussion
The UV/Vis absorption spectra obtained for both Au25-i and
Au25-bi clusters are in agreement with the literature (Fig-
ure S1 in the Supporting Information).[51,52] The spectrum
for Au25-bi clusters shows discrete peaks typical of mole-
cule-like electronic levels, whereas the spectrum for the
Au25-i clusters shows broad absorption bands at l=400, 450,
and 670 nm. The matrix-assisted laser-desorption ionization
(MALDI) of the clusters was also obtained to ascertain
their size (Figure S2 in the Supporting Information). To in-
vestigate the catalytic activity of clusters, the as-prepared
Au25-bi and Au25-i clusters were separately immobilized onto
silica supports (fumed silica) by using conventional wet im-
pregnation method and were utilized as catalysts for oxida-
tion of styrene. The catalysis reaction was performed using
uncalcined cluster samples to ensure comparison of the re-
sults was made on clusters of the same size. The styrene-oxi-
dation procedure was modified from previously reported
work by Jin and co-workers.[24] The oxidation reaction was
carried out by using acetonitrile as a solvent and TBHP as
oxidant for both the clusters. Prior to performing the oxida-
tion, a blank reaction without the catalyst was carried out to
confirm that the catalytic activity was solely due to the clus-
ters. Thermogravimetric analysis (TGA) was also performed
on the catalyst samples before and after the catalysis reac-
tion to ensure there is no leaching of the ligands during the
process. The TGA results showed nearly equal weight loss
for the catalyst samples before and after reaction confirming
that the ligands were intact during the catalysis reaction
(Figure S3 in the Supporting Information). The oxidation of
styrene catalyzed by Au25-i resulted in 66% overall conver-
sion with decrease in selectivity from benzaldehyde (48%)
to styrene oxide (46%), and benzene acetaldehyde (6%).
However, the catalysis of styrene by using Au25-bi led to de-
creased conversion (43%) albeit with higher selectivity for
benzaldehyde formation (75%) compared to styrene oxide
(16%) and benzene acetaldehyde (8%).
The normalized Au L3-edge XANES spectra of Au25-bi
and Au25-i clusters along with Au foil are shown in Figure 1.
The features in the Au25-bi and Au25-i spectra, including res-
onance peak position and shape, were similar to the bulk
gold foil. It indicates that the local environment of Au is
still similar to bulk gold. The oscillations beyond the absorp-
tion edge in Au25-bi and Au25-i XANES spectra were broad-
ened, which were related to lower coordination number,
higher structural disorder, and change of bonding distance
in comparison with the bulk gold foil, which is typical for
nanoclusters. The small shift in absorption-edge energy E0 in
Au25-bi (0.7 eV) and Au-i (0.4 eV) relative to the Au foil
was consistent with the observation of d-electron depletion.
It also suggests that the Au25-i and Au25-bi clusters were
mostly in metallic Au0 state and possibly partially oxidized
AuI. There is an increase of intensity of the shoulder near
the absorption edge (white line) was observed in Au25-bi
and Au25-i spectra relative to the bulk Au. The white line in-
tensity of Au L3-edge is associated with 2p3/2 to 5d transition.
[Au
(PPh3)
E
(Au25-bi)
and
[Au25-
ACHTUNGTRENNUNG
to understand the electronic structure of the catalysts by
using a combination of synchrotron radiation-based XAFS
and UPS, an approach rarely used previously.[32–34] XAFS
(including X-ray absorption near edge structure (XANES)
and extended X-ray absorption fine structure (EXAFS)) is a
powerful tool to explore the local atomic and electronic
structure of different kinds of gold nanoclusters.[20,35–46] The
UPS is particularly sensitive in the surface region (ca. 1 nm),
providing information about the valence-band structure
near the Fermi level, which has been employed to investi-
gate bare Au, as well as ligand-stabilized Au nanoparti-
cles.[47–50] Our own recent investigation utilizing UPS con-
firms s–d hybridization in partially “exposed” Au38(SR)24
clusters deposited on the native SiO2 oxide in addition to
pointing towards the importance of ligand effect.[50] Our hy-
pothesis was that a combination of these two techniques en-
ables us to more completely understand the relationship be-
tween the local atomic environment, electronic structure
and catalytic selectivity and activity of Au25-bi and Au25-i
nanoclusters.
10202
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 10201 – 10208

Ligand-Stabilized Gold Nanocluster Catalysis
FULL PAPER
Figure 1. Normalized Au L3-edge XANES spectra of Au25-bi, Au25-i clus-
ters, and gold foil.
Figure 2. Fourier transformed EXAFS data (k2 weighting, k range: 2–
11.5 ꢁÀ1) of Au25-i, Au25-bi clusters, and gold foil.
À
The intensity of the first resonance in Au L3-edge XANES
is usually used to evaluate the unoccupied d-band states of
Au clusters.[39,41,53] The white line of bulk Au is contributed
by s–p–d hybridization.[44,54,55] For bare Au and polymer-sta-
bilized Au nanoclusters,[20,44–46] the d–d interaction is very
strong and s–d hybridization is relatively weak. Thus, the
corresponding white line intensity is lower than bulk Au.
These observations have been previously attributed to the
strong ligand effect on small gold nanoclusters.[39,41] The
white line of Au25-bi spectra was larger than that of Au25-i,
which in turn was slightly larger than bulk Au. It suggests
the number of unoccupied d-band states of Au25-bi is larger
than that of Au25-I, which is in turn larger than bulk Au.
The Fourier transform of Au L3-edge EXAFS with k2
weight is presented in Figure 2. The radial positions of
peaks represent the average distance between the Au atom
and its near neighbors within 5 ꢁ. The fitting results for
Au25-bi and Au25-i are shown in the Table 1 (fitted data of
the Au L3-edge k2-weighted of EXAFS c(k) spectra shown
in Figure S4 in the Supporting Information). The first promi-
nent peak at approximately 2.3 ꢁ in both Au25-bi and Au25-i
nanoclusters was absent in the Au foil, which confirmed
S(P) bonding to Au atoms leading to the change in the local
environment of the Au clusters. The nearest neighbor at
Au S(P) single scattering, because the contributions from
À
À
Au S and Au P scattering were indistinguishable. The high
À
Au S(P) coordination numbers in samples, especially for
Au25-i, indicated the presence of sulfur and/or phosphorous.
À
The first Au Au single scattering in Au foil was at
2.88 ꢁ.[29,56] It was 2.80 ꢁ for Au25-i and 2.82 ꢁ for Au25-bi,
0.08 and 0.06 ꢁ shorter than that in Au foil, respectively,
which was consistent with the previous observations.[36,38,42]
À
The first-shell Au Au bonding distance was reported at
2.78 ꢁ in thiolate-protected Au38
(PPh3)12Cl6
tude of the first-shell Au Au in Au25-bi and Au25-i was much
smaller than that in bulk Au, implying the low coordination
numbers and higher structural disorder present in small Au
[2]
at 2.80 ꢁ in Au55-
[40,43]
[41]
A
and at 2.83 ꢁ in Au144(SR)60. The ampli-
À
À
particles. The first-shell Au Au coordination number for
Au25-bi is 6.9, which is similar to the experimental result and
the calculated value for the previously assigned Au13 cluster
in icosahedron structure.[38,42] It is in agreement with our
proposal of the molecular structure of biicosahedral Au25
cluster: two icosahedral Au13 cages adjoined through one
sharing Au atom and the Au25 core, stabilized by phospho-
rous and sulfur ligands. This value is also very close to the
1–2 nm gold particles.[21,57] The coordination number of the
À
first Au Au shell in Au25-i is 4.8, which is much smaller than
in the case of Au25-bi. It is likely to be due to differences in
the local atomic structures of the two Au25 clusters, as the
À
2.3 ꢁ in Au25-i was contributed by Au S single scattering.
For Au25-bi, the peak at 2.3 ꢁ was likely associated with
À
average Au Au coordination number is sensitive to
the local atomic arrangement of the clusters.[38,58]
The presence of -S-Au-S-Au-S- staples on the Au13
core in the Au25-i cluster results in decreasing first
Table 1. EXAFS fit results of Au25-bi and Au25-i clusters.
À
Sample 1st shell
1st shell
Au S(P)
Coordination d-Band
number (S/P) depletion energy
Binding
À
shell Au Au coordination number and increasing
À
Au Au
coordination bond
number (Au) length
[ꢁ]
[53,57]
À
bond
(relative) (relative) Au S coordination number at the same time.
length [ꢁ]
The ligands significantly affected the electronic
Au25-bi 2.82Æ0.03 6.9Æ1.3
2.30Æ0.05 0.7Æ0.4
2.30Æ0.04 1.2Æ0.4
high
low
lower
high
low
lower
structure and local structure of the Au25-bi and
Au25-i nanoclusters. The XANES study at the K-
edges of sulfur and phosphorous was another way
Au25-i
2.80Æ0.04 4.8Æ1.3
2.782 2.52
Au38
2.302
1.12
Chem. Eur. J. 2013, 19, 10201 – 10208
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10203

C. S. S. R. Kumar et al.
gands. By considering the Au L3-edge XANES results and
difference in electronegativity between P and Au, it is rea-
sonable to assume that electrons are transferred from Au to
P. Similar result has been reported on mixed ligand stabi-
lized Au13 clusters.[42] The charge transfer from Au to P in
Au13ACHTNUTRGNE(NUG PH3)6 clusters was also previously supported by the
DFT analysis.[59]
Upon calcination of the catalysts, there was a distinct
change observed in their catalytic activity. The calcined
Au25-i clusters showed a better conversion of 77% with se-
lectivity of benzaldehyde (70%), styrene oxide (22%), ben-
zene acetaldehyde (3%), and acetophenone (5%), whereas
the calcined Au25-bi clusters had a conversion of 67% with
selectivity of the products being benzaldehyde (47%) and
styrene oxide (53%). These catalysis results are in agree-
ment with those earlier published by Jin et al.,[24] wherein a
6–10% increase in catalytic activity was observed for cal-
cined samples over uncalcined samples (Table 2).
There is a clear effect of calcination on the catalytic activ-
ity and product selectivity of the catalysts. Although the un-
calcined Au25-i sample showed almost equal selectivity to-
wards benzaldehyde and styrene oxide, the calcined sample
favored the formation of benzaldehyde over styrene oxide.
Conversely, for Au25-bi clusters, the uncalcined sample fa-
vored the formation of benzaldehyde over styrene oxide,
whereas the calcined sample showed comparable selectivity
for both. Not surprisingly, in either case, the calcined sam-
ples showed higher catalytic activity (conversion percentage)
than the uncalcined samples. The increased catalytic activity
of the calcined samples can be attributed to the increased
accessibility of the catalyst surface due to partial removal of
ligands. The nature of the oxidant also played a significant
role in determining the selectivity of the products. For ex-
ample, TBHP is known to be activated even on larger parti-
cles,[17,24] and hence, the observed selectivity for calcined
samples was different from the uncalcined samples.
Figure 3. a) Normalized sulfur K-edge XANES spectra of Au25-bi and
Au25-i clusters; b) normalized phosphorous K-edge XANES Au25-bi and
PPh3.
to probe the Au-ligand interaction from the perspective of S
and P. As seen in Figure 3a, the normalized sulfur K-edge
XANES spectra of Au25-bi and Au25-i showed different fea-
Taking Au25-i as an example, Au L3-edge EXAFS was per-
formed on silica supported Au25-i nanoclusters after calcina-
tion at 2008C for 2 h in vacuum. After calcination treat-
tures. The white line in Au25-i XANES at 2472.7 eV was as-
[41]
À
À
signed to S 1s to S C transition. The white line splitting
ment, the Au S scattering and Au-low Z-scattering contri-
in the Au25-bi cluster spectrum was observed at 2471.8 and
bution in Au25-i was negligible, because the thiolate ligands
À
À
2473 eV. It was probably the result of differences in Au S
were mostly removed. The first shell Au Au coordination
bonds in Au25-bi and Au25-i
clusters because the XANES is
quite sensitive to the chemical
Table 2. Catalytic-activity comparison for uncalcined and calcined Au25-bi and Au25-i clusters.
environment of the absorbing
atom. The Au25-bi contained
both sulfur and phosphine li-
gands, so the phosphorus K-
edge XANES of Au25-bi (Fig-
ure 3b) was collected along
with that of PPh3 as a reference.
The first strong shift in reso-
nance to 2147.4 and 2150.2 eV
by 1.8 and 2.6 eV in comparison
to PPh3 was due to the bonding
between Au and phosphine li-
Reaction conditions
Selectivity for different products [%]
Total
conversion
[%]
Benzaldehyde Styrene Benzene
AcetoACTHNGUTRENNUpG henone
oxide
acetaldehyde
Styrene, TBHP, acetonitrile, 758C,
Au25-i@SiO2 (uncalcined)
Styrene, TBHP, acetonitrile, 758C,
Au25-bi @SiO2 (uncalcined)
Styrene, TBHP, acetonitrile, 758C,
Au25-i@SiO2 (calcined)
48
75
70
47
46
6
–
–
5
–
66
43
77
67
16
22
53
8
3
–
Styrene, TBHP, acetonitrile, 758C,
Au25-bi @SiO2 (calcined)
10204
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 10201 – 10208

Ligand-Stabilized Gold Nanocluster Catalysis
FULL PAPER
Table 3. EXAFS fit results of Au25-i/SiO2 clusters calcined in the air and He.
Sample
Paths
Bond length R [ꢁ]
Coordination number
Debye–Waller factor [ꢁÀ1
]
Energy shift DE [eV]
5.3Æ1.9
À
Au Au
2.84Æ0.02
2.28Æ0.02
2.83Æ0.01
2.25Æ0.04
2.83Æ0.01
2.83Æ0.01
2.29Æ0.01
2.82Æ0.01
2.24Æ0.02
2.82Æ0.02
2.22Æ0.05
4.9Æ0.9
1.5Æ0.2
9.7Æ1.2
0.4Æ0.2
10.8Æ0.7
6.4Æ1.5
1.2Æ0.3
8.6Æ1.4
0.5Æ0.2
9.2Æ1.1
0.2Æ0.1
0.014
0.008
0.013
0.009
0.015
0.014
0.008
0.014
0.009
0.015
0.003
Air 1008C
À
Au S
À
Au Au
3.0Æ0.7
Air 2008C
Air 3008C
He 1008C
À
Au S
À
Au Au
2.8Æ0.9
4.2Æ1.1
À
Au Au
À
Au S
À
Au Au
2.5Æ0.9
2.3Æ1.5
He 2008C
He 3008C
À
Au S
À
Au Au
À
Au S
number increased to about 9.5 in calcined Au25-i/SiO2, which
slightly changed to 8.5 after catalysis (Table S1 in the Sup-
cycle and began to be more pronounced at about 3.4 eV for
5d5/2 orbital.[50] As can be seen in Figure 4a, the valence-
band photoemission experiments of Au25-bi and Au25-i nano-
clusters revealed higher binding energies for gold d-band
center (Ed) of Au25 clusters (Au25-bi (4.56 eV), Au25-i
(4.36 eV)) relative to bulk Au (4.25 eV). The d-band center
of Au25-bi and Au25-i clusters with a higher binding energy
À
porting Information). The increase of the first shell Au Au
coordination number indicated the growth of Au particle
size on SiO2, which was not surprising, based on previously
published results, after thermal treatment of ligand protect-
ed Au nanoclusters.[24,60]
The electronic and geometric changes in silica-supported
Au25-i nanoclusters during the calcination in air and helium
(He) was probed using in situ XANES and EXAFS at Au
L3-edge to understand how the local environment of the
atomically precise Au catalyst changes during the calcina-
tion treatment. The white line intensity in XANES spectrum
of Au25-i/SiO2 calcined at 3008C was smaller than that of the
bulk Au, which was associated with the d-band electron
counts suggesting that the calcination treatment at 3008C re-
moved most of the ligands, and thus the quantum size effect
dominated d-band electron character (Figure S5a in the
Supporting Information). It is similar to the case of bare Au
clusters.[21,44] The fitting results of Au L3-edge EXAFS are
À
summarized in Table 3. It was found that the first shell Au
Au bond remained at about 2.83 ꢁ, whereas the coordina-
tion number increased with the increase in temperature. The
significant change of coordination number happened around
2008C, especially for sulfur. Air appeared to be more effec-
tive in assisting the removal of the ligands than He gas at
the same temperature. At 3008C, the k2-weighted EXAFS
c(k) of Au25-i clusters with fewer ligands still showed higher
disorder and low coordination number relative to bulk Au
(Figure S5b in the Supporting Information). The average
particle size of the Au25-i/SiO2 after calcination to 3008C
was approximately 3–4 nm.[20]
The UPS was employed to study the surface electronic
structure of ligand-stabilized Au25-bi and Au25-i clusters, par-
ticularly how chemical nature of ligands and size effect can
be correlated with their electronic structure. The valence-
band spectra of gold clusters gave the characteristics of the
6s and 5d bands density of states. In a previous study of
Figure 4. Valence-band photoemission spectra of Au25-bi (red), Au25-i
(black; without sputtering), and Au38 clusters (blue; after sputtering to
remove the thiolate ligands) deposited from solution on SiO2 native
thiolACHTUNGTRENNUNGate-stabilized Au38 nanoclusters supported on SiACHTUNGTNER(NUGN 111)
oxide surfaces prepared on SiACTHNURTGNEU(GN 111), compared to a micron thick (111)
wafer, it was shown that the valence-band spectra of Au38
barely resembled the gold-valence band features before
sputtering.[50] The intensity of 5d-band features at about 3
and 6 eV starts to increase after first gentle Ar sputtering
textured gold film (magenta). All spectra were taken at a photon energy
of 84 eV; b) d-band width (Wd), d-band apparent spin–orbit splitting (Eso)
and d-band center position (Ed) of Au25-bi and Au25-i nanoclusters ob-
tained from (a). The region close to Fermi level is shown as an inset.
Chem. Eur. J. 2013, 19, 10201 – 10208
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10205

C. S. S. R. Kumar et al.
may be due to the presence of partially oxidized Au atoms
in them, an observation also supported by the analysis of
XANES spectra. The d-band width (Wd) and apparent spin–
orbit splitting (Eso) were narrower in smaller size nanoclus-
ters (Au25-bi<Au25-i<Au38 <bulk Au). This behavior was
also seen previously from valence-band spectra studies on
thiolate-capped Au clusters,[29,41] as well as the observations
on bare Au nanoclusters (fcc structure) to some extent.[47,48]
Both valence-band spectra for Au25-bi and Au25-i (synchro-
tron-radiation flux normalization was done during compari-
son of photoemission intensities; see Figure 4) were collect-
ed without any sample treatment. It means that any differ-
ence in the intensity of the experimental bands can be at-
tributed to chemical nature of ligand assuming different
bond strengths or just simple presence/absence of ligands on
the surface of the clusters. The intensity of 5d band for
Au25-bi was slightly higher than Au25-i clusters, implying that
the contribution from surface Au atoms in Au25-bi is higher
than for Au25-i. Another possible reason for such electronic
structure difference may be simply different net charges on
Au25Àbi and Au25Ài (positive vs. negative). The suppression
of the intensity of 5d5/2 band in Au25Ài spectrum may imply
that charge transfer from Au d band to sulfur is mainly con-
tributed by d5/2 electrons in Au25-i nanoclusters. Such change
in electronic structure relative to Au25-bi could be related to
the different local structure of Au25-i, for example, the pres-
ence of -S-Au-S-Au-S- staples. A close look at the spectra
indicated higher 6s-band density of states at the Fermi level
region for Au38 cluster than Au25-bi and Au25-i nanoclusters.
Despite the fact that the photoemission intensity in that
region was quite lower than that for the gold film, it still
suggested that Au38 could be considered as more metallic
than Au25-bi and Au25-i in accordance with the observed
trend for d-band binding energies.
flects that the contribution from the surface Au atoms in
Au25-bi was larger than that in Au25-i. One of the reasonable
explanations seems to be the easier removal of ligands in
Au25-bi clusters compared to thiol ligands in staple motif
layer of Au25-i under ultrahigh vacuum (UHV) condition.
This is consistent with the hypothesis that the phosphine li-
gands in Au25-bi are more weakly bound to the Au clusters
than the thiol ligands in Au25-i, which was also seen in our
previous investigations on Au38.[50] The observed correlation
was consistent with the increasing d-band intensity, although
thiol ligands in Au38 were removed after Ar sputtering
cycle.[50] The 4f band for Au25-bi showed a positive shift rela-
tive to Au25-i, which was consistent with the observation of
d-band shift.
The correlation between the catalytic activity and elec-
tronic structure of bare Au catalysts shows that differences
in CO oxidation activity of Au clusters can be explained by
the change in the s–p–d hybridization and increased d-elec-
tron density in smaller particles.[21] The more negatively
charged Au core in polymer-stabilized Au clusters was more
active in aerobic oxidation of alcohols, because O2 is more
readily activated by the higher negative charge on the Au
core to promote the oxidation reaction. Furthermore, with
Ag doping (<10%), the catalytic activity of the polymer-
stabilized Au clusters was enhanced due to the anionic
charge transferred from Ag to Au sites.[46]
The XAFS and UPS results in the present study revealed
differences in the d-band structure of the Au25-bi and Au25-i
clusters. Higher d-band depletion of Au sites in Au25-bi in
comparison with the Au25-i clusters, due to more d-band
electron transferred to ligands, is likely to make the Au sites
in Au25-bi clusters more relatively electropositive (Table 1).
This could be correlated to their activity and selectivity for
the styrene oxidation reaction, in which the uncalcined Au25-
bi clusters favored the formation of more oxidized product
benzaldehyde over styrene oxide. For uncalcined samples
containing ligands, the interaction between the Au core and
TBHP is weak. Thus, selectivity is mainly determined by the
charge of Au core. However, for the calcined samples the in-
teraction between the Au clusters and oxidant TBHP has to
be considered. The calcined Au25-i/SiO2 displayed higher se-
lectivity to benzaldehyde than uncalcined Au25-i/SiO2,
though the Au core became more electronegative after cal-
cination. The hybridization between TBHP molecular states
and Au d-band states in calcined Au25-i/SiO2 interaction may
be responsible for promoting the formation of benzaldehyde
in styrene oxidation.
The differences seen in the intensity of gold d band in the
valence-band spectra for Au25-bi and Au25-i while using
photon energies in the range of 60–80 eV almost vanished,
although photon energy was increased to over 100 eV (as
shown in Figure 5 spectra recorded at 125 eV). In spite of
close similarities in valence band at this photon energy, an
enhanced intensity of gold 4f band (at about 84 and 88 eV
with respect to the Fermi level) for Au25-bi clusters was
clearly seen, whereas it was barely visible for Au25-i. This re-
The styrene oxidation can be understood to proceed in
different pathways on each of these two types of uncalcined
Au clusters. Uncalcined Au25-i produced benzaldehyde and
styrene oxide as the major products at comparable selectivi-
ties, whereas uncalcined Au25-bi was more selective to benz-
AHCTUNGTRENGUaNN lAHCTNUGRTNEGdNU eCAHTNUGTRENhGUNN yde than styrene oxide. We propose a styrene-oxida-
tion mechanism that is consistent with that of Jin and co-
workers.[24] Initially, TBHP is bound to the clusters forming
a hydroperoxy species, which later becomes peroxyformate
intermediate upon losing a water molecule. The styrene is
Figure 5. Wide-range ultraviolet photoemission spectra of ligand-stabi-
lized Au25-bi and Au25-i nanoclusters recorded at 125 eV photon energy
at room temperature.
10206
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 10201 – 10208

Ligand-Stabilized Gold Nanocluster Catalysis
FULL PAPER
cient Chemical Transformations (IACT), an Energy Frontier Re-search
Center funded by the U.S. Department of Energy, Office of Science,
Office of Basic Energy Sciences. We are grateful to Dr. Tianpin Wu and
Dr. Tomohiro Shibata of MRCAT-10ID beam line for their help during
EXAFS measurements of Au L3-edge. Acknowledgements are also due
to Dr. Vladislav Zyryanov for designing the sample cells used for loading
the samples.
likely adsorbing on to this intermediate through the C=C ac-
tivation. The C=C bond interacts with adjacent oxygen
À
atoms on the peroxyformate through p p bonding forming
a secondary intermediate that later rearranges to form dif-
ferent product depending on the selectivity of the cluster.
Control experiments in the absence of catalyst and also
TBHP were carried out to confirm that the catalytic effect
was only due to clusters.
[1] D. Li, C. Wang, D. Tripkovic, S. Sun, N. M. Markovic, V. R. Stamen-
kovic, ACS Catal. 2012, 2, 1358–1362.
[2] S. Gaur, J. T. Miller, D. Stellwagen, A. Sanampudi, C. S. S. R.
Kumar, J. Spivey, Phys. Chem. Chem. Phys. 2012, 14, 1627–1634.
[3] M. Shekhar, J. Wang, W.-S. Lee, W. D. Williams, S. M. Kim, E. A.
Stach, J. T. Miller, W. N. Delgass, F. H. Ribeiro, J. Am. Chem. Soc.
2012, 134, 4700–4708.
[4] R. M. Rioux, H. Song, P. Yang, G. A. Samorjai, Catalysis and Mate-
rials Science: The Issue of Size Control 2008, 149–166.
[5] Y. Pꢂrez, M. L. Ruiz-Gonzꢃlez, J. M. Gonzꢃlez-Calbet, P. Concep-
ciꢄn, M. Boronat, A. Corma, Catal. Today 2012, 180, 59–67.
[6] I. Lee, F. Delbecq, R. Morales, M. A. Albiter, F. Zaera, Nat. Mater.
2009, 8, 132–138.
[7] M. P. Casaletto, A. Longo, A. Martorana, A. Prestianni, A. M. Vene-
zia, Surf. Interface Anal. 2006, 38, 215–218.
[8] E. Gross, J. M. Krier, L. Heinke, G. A. Somorjai, Top Catal 2012, 55,
13–23.
[9] Z. J. Li, D. S. Huang, W. Liu, R. M. Xiao, J. Liu, C. Xu, Y. Jiang,
L. D. Sun, Gold Bull. 2004, 37, 3–11.
[10] C. D. Pina, E. Falletta, M. Rossi, Chem. Soc. Rev. 2012, 41, 350–369.
[11] L. Guczia, A. Becka, Z. Pꢃsztib, Catal. Today 2012, 181, 26–32.
[12] R. Jin, Nanotechnol. Rev. 2012, 1, 31–56.
[13] Y. Zhu, R. Jin, Y. Sun, Catalyst 2011, 1, 3–17.
[14] J. Hernꢃndez, J. Solla-Gullꢄn, E. Herrero, A. Aldaz, J. M. Feliu, J.
Phys. Chem. C 2007, 111, 14078–14083.
[15] S. H. Brodersen, U. Grønbjerg, B. Hvolbæk, J. Schiøtz, J. Catal.
2011, 284, 34–41.
[16] M. Valden, X. Lai, D. W. Goodman, Science 1998, 281, 1647–1650.
[17] A. Sanchez, S. Abbet, U. Heiz, W. D. Schneider, H. Hakkinen, R. N.
Barnett, U. Landman, J. Phys. Chem. A 1999, 103, 9573–9578.
[18] Y. Zhang, X. Cui, F. Shi, Y. Deng, Chem. Rev. 2012, 112, 2467–2505.
[19] T. V. W. Janssens, A. Carlsson, A. Puig-Molina, B. S. Clausen, J.
Catal. 2006, 240, 108–113.
[20] M. S. Chen, D. W. Goodman, Catal. Today 2006, 111, 22–33.
[21] J. T. Miller, A. J. Kropf, Y. Zha, J. R. Regalbuto, L. Delannoy, C.
Louis, E. Bus, J. A. van Bokhoven, J. Catal. 2006, 240, 222–234.
[22] M. Boronat, A. Corma, J. Catal. 2011, 284, 138–147.
[23] R. Jin, Nanoscale 2010, 2, 343–362.
Conclusion
Our results provide an insight into correlation between the
local structure, electronic structure, and catalytic selectivity
and activity of ligand-stabilized Au25-bi and Au25-i nanoclus-
ter catalysis. Compared to bulk Au, there is a d-band deple-
À
tion and Au Au bond contraction in Au25-bi and Au25-i asso-
ciated with strong ligand binding. The d-band narrowing, ap-
parent spin–orbit splitting reduction, and d-band center shift
in ligand-stabilized Au25-bi, Au25-i and Au38 nanoclusters was
observed in UPS. The higher electropositivity of uncalcined
Au25-bi compared to Au25-i could be associated with its
higher oxidizing capability leading to enhanced selectivity to
benzaldehyde.
On the other hand, calcination led to effective removal of
the ligands that resulted in increase in the total conversion
of styrene for both Au25-bi and Au25-i. However, calcination
also led to the increase in the size of Au core and caused
the particle size to increase to 4 nm for Au25-i. The number
of d-band hole population decreased compared to the uncal-
cined Au25-i clusters, implying decrease in the relative elec-
tropositivity after the calcination treatment. Such a decrease
in electropositivity still led to increase in selectivity for ben-
zaldehyde; which is in contradiction to the case of catalysis
by uncalcined samples. Thus, the electropositivity of Au
atoms may not be the dominant factor for catalytic selectivi-
ty in calcined Au25-bi/SiO2 and Au25-i/SiO2 nanoclusters. In-
stead, the interaction between the catalyst and oxidant
TBHP appear to play a key role in the selectivity. While in-
vestigations to date attempt to explain nanocluster catalysis
based on their electronic and geometric structure,[68] our
findings point out the importance of atomic purity of cata-
lysts during the catalysis to arrive at meaningful conclusions
with respect to correlation of their electronic structure with
catalytic properties.
[24] Y. Zhu, H. Qian, R. Jin, Chem. Eur. J. 2010, 16, 11455–11462.
[25] Y. Liu, H. Tsunoyama, T. Akita, T. Tsukuda, Chem. Commun. 2010,
46, 550–552.
[26] M. Turner, V. B. Golovko, O. P. H. Vaughan, P. Abdulkin, A. B.
Murcia, M. S. Tikhov, B. F. G. Johnson, R. M. Lambert, Nature 2008,
454, 981–983.
[27] Y. Gao, N. Shao, Y. Pei, Z. Chen, X. C. Zeng, ACS Nano 2011, 5,
7818–7829.
[28] Y. Zhu, H. Qian, R. Jin, J. Mater. Chem. 2011, 21, 6793–6799.
[29] Y. Zhu, H. Qian, A. Das, R. Jin, Chin. J. Catal. 2011, 32, 1145–1150.
[30] W. Gao, X. F. Chen, J. C. Li, Q. Jiang, J. Phys. Chem. C 2010, 114,
1148–1153.
[31] Directing Matter and Energy: Five Challenges for Science and the
Imagination, A report from the basic energy sciences advisory com-
mittee, DOE, USA 2007.
[32] H. Hoffmann, F. Zaera, R. M. Ormerod, R. M. Lambert, J. M. Yao,
D. K. Saldin, L. P. Wang, D. W. Bennett, W. T. Tysoe, Surf. Sci. 1992,
268, 1–10.
[33] A. K. Chakraborty, K. S. Coleman, V. R. Dhanak, Nanotechnology
2009, 20, 155704–155709.
Acknowledgements
This material is based upon work supported as part of the Center for
Atomic Level Catalyst Design, an Energy Frontier Research Center
funded by the U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences under Award Number DE-SC0001058. MRCAT
operations are supported by the Department of Energy and the MRCAT
member institutions. The use of the advanced photon source at ANL was
supported by the U.S. Department of Energy, Office of Science, Office of
Basic Energy Sciences under Contract No. DE-AC02-06CH11357. Finan-
cial support for J.T.M. was provided as part of the Institute for Atom-effi-
[34] Y. Zheng, D. Qi, N. Chandrasekhar, X. Gao, C. Troadec, A. T. Wee,
Langmuir 2007, 23, 8336–8342.
Chem. Eur. J. 2013, 19, 10201 – 10208
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10207

C. S. S. R. Kumar et al.
[35] C. Lꢄpez-Cartes, T. C. Rojas, R. Litrꢃn, D. Martꢅnez-Martꢅnez, J. M.
de La Fuente, S. Penadꢂs, A. Fernꢃndez, J. Phys. Chem. B 2005, 109,
8761–8766.
[51] Y. Shichibu, Y. Negishi, T. Watanabe, K. N. Chaki, H. Kawaguchi, T.
Tsukuda, J. Phys. Chem. C 2007, 111, 7845–7847 (reference for
Au25-bi structure adapted in TOC figure).
[52] M. Zhu, E. Lanni, N. Garg, M. E. Bier, R. Jin, J. Am. Chem. Soc.
2008, 130, 1138–1139.
[36] M. A. MacDonald, D. M. Chevrier, H. Qian, R. Jin, J. Phys. Chem.
C 2011, 115, 15282–15287.
[53] G. Simms, J. D. Padmos, P. Zhang, J. Chem. Phys. 2009, 131,
214703–1–214703–9.
[54] L. F. Mattheiss, R. E. Dietz, Phys. Rev. B 1980, 22, 1663–1676.
[55] M. G. Mason, Phys. Rev. B 1983, 27, 748–762.
[56] CRC Handbook of Chemistry and Physics, 75th ed. (Ed.: D. R.
Lide), CRC Press, Boca Raton, 1994.
[57] M. Zhu, C. M. Aikens, F. J. Hollander, G. C. Schatz, R. Jin, J. Am.
Chem. Soc. 2008, 130, 5883–5885.
[58] J. Kilmartin, R. Sarip, R. Grau-Crespo, D. Di Tommaso, G. Hogarth,
C. Prestipino, G. Sankar, ACS Catal. 2012, 2, 957–963.
[59] G. Shafai, S. Hong, M. Bertino, T. S. Rahman, J. Phys. Chem. C
2009, 113, 12072–12078.
[60] L. D. Menard, F. Xu, R. G. Nuzzo, J. C. Yang, J. Catal. 2006, 243,
64–73.
[61] M. Lemonnier, O. Collet, C. Depautex, J. M. Esteva, D. Raoux,
Nucl. Instrum. Methods Phys. Res. Sect. A 1978, 152, 109–111.
[62] C. U. Segre, N. E. Leyarovska, L. D. Chapman, W. M. Lavender,
P. W. Plag, A. S. King, A. J. Kropf, B. A. Bunker, K. M. Kemner, P.
Dutta, R. S. Duran, J. Kaduk, Synchrotron Radiation Instrumenta-
tion: Eleventh U.S. National Conference (Ed.: P. Pianetta), American
Institute of Physics, New York, 2000, Chapter 512, pp. 419–422.
[63] M. Newville, J. Synchrotron. Rad. 2001, 8, 322–324.
[64] J. J. Rehr, R. C. Albers, Rev. Mod. Phys. 2000, 72, 621–654.
[65] B. Ravel, M. Newville, J. Synchrotron. Rad. 2005, 12, 537–541.
[66] P. A. Dowben, D. LaGraffe, M. Onellion, J. Phys. Condens. Matter
1989, 1, 6571–6587.
[37] D. Zanchet, H. Tolentino, M. C. Martins Alves, O. L. Alves, D.
Ugarte, Chem. Phys. Lett. 2000, 323, 167–172.
[38] L. D. Menard, S. Gao, H. Xu, R. D. Twesten, A. S. Harper, Y. Song,
G. Wang, A. D. Douglas, J. C. Yang, A. I. Frenkel, R. G. Nuzzo,
R. W. Murray, J. Phys. Chem. B 2006, 110, 12874–12883.
[39] P. Zhang, T. K. Sham, Phys. Rev. Lett. 2003, 90, 245502–1–245502–
4.
[40] P. D. Cluskey, R. J. Newport, R. E. Benfield, S. J. Gurman, G.
Schmid, Zeitschrift fꢀr Physik D: Atoms Molecules and Clusters
1993, 26, 8–11.
[41] M. A. MacDonald, P. Zhang, H. Qian, R. Jin, J. Phys. Chem. Lett.
2010, 1, 1821–1825.
[42] A. I. Frenkel, L. D. Menard, P. Northrup, J. A. Rodriguez, F.
Zypman, D. Glasner, S. P. Gao, H. Xu, J. C. Yang, R. G. Nuzzo, AIP
Conf. Proc. 2007, 882, 749–751.
[43] M. M. Marcus, M. P. Andrews, J. Zegenhagen, A. S. Bommannavar,
P. Montano, Phys. Rev. B 1990, 42, 3312–3316.
[44] J. A. Van Bokhoven, J. T. Miller, J. Phys. Chem. C. 2007, 111, 9245–
9249.
[45] H. Tsunoyama, N. Ichikuni, H. Sakurai, T. Tsukuda, J. Am. Chem.
Soc. 2009, 131, 7086–7093.
[46] N. K. Chaki, H. Tsunoyama, Y. Negishi, H. Sakurai, T. Tsukuda, J.
Phys. Chem. C 2007, 111, 4885–4888.
[47] A. Visikovskiy, H. Matsumoto, K. Mitsuhara, T. Nakada, T. Akita,
Y. Kido, Phys. Rev. B 2011, 83, 165428–1–165428–9.
[48] D. C. Lim, I. Lopez-Salido, R. Dietsche, M. Bubek, Y. D. Kim,
Chem. Phys. 2006, 330, 441–448.
[49] M. Quinten, I. Sander, P. Steiner, U. Kreibig, K. Fauth, G. Schmid,
Z. Phys. D 1991, 20, 377–379.
[67] M. W. Heaven, A. Dass, P. S. White, K. M. Holt, R. W. Murray, J.
Am. Chem. Soc. 2008, 130, 3754–3755 (reference for Au25-i structure
adapted in TOC figure).
[68] H. Li, L. Li, Y. Li, Nanotech. Rev. DOI: 10.1515/ntrev-2012-0069.
[50] Y. B. Losovyj, S.-C. Li, N. Lozova, K. Katsiev, D. Stellwagen, U. Die-
bold, L. Kong, C. S. S. R. Kumar, J. Phys. Chem. C 2012, 116, 5857–
5861.
Received: February 15, 2013
Revised: May 14, 2013
Published online: June 20, 2013
10208
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 10201 – 10208