Covalently Binding Atomically Designed Au9 Clusters to Chemically Modified Graphene

Article abstract of DOI:10.1002/anie.201504334

Atomic-resolution transmission electron microscopy was used to identify individual Au₉ clusters on a sulfur-functionalized graphene surface. The clusters were preformed in solution and covalently attached to the surface without any dispersion or aggregation. Comparison of the experimental images with simulations allowed the rotational motion, without lateral displacement, of individual clusters to be discerned, thereby demonstrating a robust covalent attachment of intact clusters to the graphene surface.

Full text of DOI:10.1002/anie.201504334

.
Angewandte
Communications
International Edition: DOI: 10.1002/anie.201504334
German Edition: DOI: 10.1002/ange.201504334
Surface Chemistry
Covalently Binding Atomically Designed Au9 Clusters to Chemically
Modified Graphene**
Concha Bosch-Navarro,* Zachary P. L. Laker, Helen R. Thomas, Alexander J. Marsden,
Jeremy Sloan, Neil R. Wilson,* and Jonathan P. Rourke*
Abstract: Atomic-resolution transmission electron microscopy
was used to identify individual Au9 clusters on a sulfur-
functionalized graphene surface. The clusters were preformed
in solution and covalently attached to the surface without any
dispersion or aggregation. Comparison of the experimental
images with simulations allowed the rotational motion, without
lateral displacement, of individual clusters to be discerned,
thereby demonstrating a robust covalent attachment of intact
clusters to the graphene surface.
allowing direct correlation between property and structural
features.
Graphene (G), as a two-dimensional system with out-
standing electronic properties and a high surface area,[13]
offers the ideal platform for the deposition of NPs.[14] In
addition, the diversity of carbon chemistry offers many routes
to producing chemically modified graphene (CMG).[13,15,16]
Nanoparticles have been stably attached to both G and CMGs
for a plethora of different applications.[17] In particular,
different routes for the hybridization of Au NPs with G
have been studied,[18–23] but surprisingly, the fabrication of
atomically precise Au NCs supported on G remains unex-
plored. Recently, we have described an easy way to
chemically modify graphene with sulfur functionalities,[24]
and now, by taking advantage of the affinity between gold
and sulfur, we describe the stable attachment of preformed
[Au9(PPh3)8](NO3) clusters[25] to our CMG. Aberration-cor-
rected transmission electron microscopy (ac-TEM) has been
employed to directly identify individual covalently attached
Au9 clusters, and to track their relative orientation.
Chemically modified graphene with sulfur functionalities
was synthesized by treatment of graphene oxide (GO) with
potassium thioacetate, followed by an aqueous work-up.[24]
This route is simple and scalable, and gives a single-layer
material with reactive thiol groups that offer anchoring points
for further functionalization; this material is referred to as
GOSH. Moreover, the synthetic route results in a partial
reduction of the GO when the sulfur functionalities are
introduced, thereby giving a more graphene-like substance;
this is particularly relevant for applications in which a semi-
conducting/conducting behavior is required, as the reduction
of GO results in a partial restoration of the sp2 structure of
G.[26]
The [Au9(PPh3)8](NO3) cluster (abbreviated as Au9) was
selected as the target cluster.[10,25] The D2h-symmetric cluster
is composed of nine gold atoms arranged such that one central
gold atom is surrounded by the remaining eight gold atoms,
each of which is coordinated by a monodentate phosphine
ligand (see Figures S1 and S2 in the Supporting Information).
The average metal–metal distance is around 0.27 nm, which
results in a cluster diameter between 0.45 nm and 0.54 nm,[10]
far smaller than that typically exhibited by Au NPs (particle
size > 3 nm).[11] The binding between Au9 and GOSH was
achieved by simply stirring Au9 with a dispersion of GOSH
(Scheme 1). A covalent bond is formed between sulfur and
gold, which is accompanied by displacement of a phosphine
ligand. As a result, a neutral GOSH@Au9 hybrid is formed.
A comparison of the thermogravimetric analysis (TGA)
results for GOSH and GOSH@Au9 gives the first evidence of
G
old nanoparticles (Au NPs) have structure- and size-
dependent optical and electronic properties,[1–4] and their
catalytic activity increases when the particle size drops down
to around 1 nm.[5] To achieve this size, an accurate design of
ligand-protected gold nanoclusters (Au NCs) is required. A
family of phosphine-coordinated Au NCs ([AunLm]z+; n = 1–
11, z = 1–4; L = PPh3 or PPh2(CH2)3PPh2) having well-defined
nuclearity and geometrical structures have been synthe-
sized[6,7] and shown to exhibit clearly distinguishable optical
and electronic properties.[7–10] However, the immobilization of
Au NCs onto a support is an important step towards their
implementation in practical devices, which has so far not been
realized satisfactorily.[1,2,11,12] Therefore, the challenge is to
develop a strategy to chemically bind predesigned Au NCs
onto the surface of solid conductors/semiconductors, thus
[*] Dr. C. Bosch-Navarro, Z. P. L. Laker, A. J. Marsden, Dr. J. Sloan,
Dr. N. R. Wilson
Department of Physics, University of Warwick
Coventry, CV4 7AL (UK)
E-mail: concepcion.bosch@uv.es
Neil.Wilson@warwick.ac.uk
Dr. C. Bosch-Navarro, H. R. Thomas, Dr. J. P. Rourke
Department of Chemistry, University of Warwick
Coventry, CV4 7AL (UK)
E-mail: j.rourke@warwick.ac.uk
[**] C.B.-N. acknowledges support for her fellowship from the Vali+D
program of the Generalitat Valenciana (Spain). Z.P.L.L. thanks the
EPSRC for support through a studentship (EP/M506679/1). We
thank Marc Walker for assistance with XPS measurements.
Supporting information for this article (full experimental details,
together with spectroscopic data and a complete set of image
simulations and a movie of experimental images, complementary
image simulations, and corresponding molecular models) is
available on the WWW under http://dx.doi.org/10.1002/anie.
201504334.
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co.
KGaA. This is an open access article under the terms of the Creative
Commons Attribution License, which permits use, distribution and
reproduction in any medium, provided the original work is properly
cited.
9560
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 9560 –9563

Angewandte
Chemie
out the possibility of an inter-
action between Au9 and any
remaining oxygen functionali-
ties on the GOSH, thus con-
firming that Au9 attachment is
through the sulfur functionali-
ties.
Determining where the
Au9 is attached and whether
the clusters remain intact
Scheme 1. Synthesis of GOSH@Au9.
requires direct imaging of the
GOSH@Au9 hybrid at atomic
hybrid formation (Figure S7). In both cases, a first weight loss
appears around 1308C, which corresponds to desorption of
water, while the major mass loss is above 5008C and comes
from decomposition of graphene-like sheets. The main differ-
ence appears at higher temperatures, where a mass of 10.5%
remains for GOSH@Au9, thus pointing to the presence of
metallic centers at a level of order of 1 atomic%. The
presence of gold in the hybrid was confirmed by energy-
dispersive X-ray (EDX) elemental analysis, which showed
atomic gold and phosphorus contents of around 1 atomic%
for each element. Additional, and more accurate, corrobo-
ration came from X-ray photoelectron spectroscopy (XPS;
Figures S4–S6), in which, in addition to the signals corre-
sponding to C, S, and O that are typical of GOSH,[24] signals
for Au and P were found, consistent with the presence of
Au9.[6] Moreover, the ratio between the atomic content
determined by XPS (1.6% Au:1.3% P) is consistent with
the presence of Au9P7 clusters.
resolution. This was achieved by ac-TEM. Atomically thin 2D
materials such as graphene or graphene oxide have been
employed as supports for imaging molecular species;[27,28] at
one atom thickness of carbon, they are almost transparent
under the electron beam so that individual molecules can be
resolved at atomic resolution, whilst their well-defined crystal
lattice enables accurate calibration for quantitative measure-
ments. TEM grids were prepared by adding one drop of
a well-dispersed 0.08 mgmLÀ1 solution of GOSH@Au9 in
DMSO to a lacey carbon support and allowing it to dry under
ambient conditions.
An ac-TEM image of a typical area of GOSH@Au9 is
shown in Figure 1a. As is characteristic for samples derived
from graphene oxide,[29] the image shows ordered regions
where the graphene-like lattice is visible, regions of higher
contrast that are apparently disordered and can be attributed
to oxidation debris[30,31] or other carbonaceous material
adhered to the surface, defects, and small holes. A single
hexagon of spots can be seen in the inset diffraction pattern,
which shows that the long-range crystalline order of graphene
is retained, and that this region consists of a single monolayer
A control experiment was performed whereby, instead of
GOSH, GO was treated with Au9. With GO, no gold was
detected by either EDX or by XPS (Figure S5), thus ruling
Figure 1. a) General ac-TEM image of GOSH@Au9. Inset: the typical hexagonal pattern of a graphene monolayer. b) Three different ac-TEM
images with high resolution for Au9 imaged over GOSH in three different orientations. b1 corresponds to the molecular cluster highlighted with
a square in (a). c) ac-TEM image simulation produced from the crystal-structure model depicted in (d).
Angew. Chem. Int. Ed. 2015, 54, 9560 –9563
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9561

.
Angewandte
Communications
Figure 2. Selected ac-TEM image frames taken from the dynamics of a single Au9 with their corresponding molecular models (the full ac-TEM
sequence representing the motion of a single Au9 cluster over a period of 11.4 s is shown in Figure S12).
of chemically modified graphene (see also Figure S8).[27]
Additional features are seen, not observed on GO or
GOSH. Clusters of dark spots indicative of atoms with
a high atomic number are apparent and dispersed across the
GOSH surface. There is clear structure within the clusters
and, as discussed below, this can be used to unambiguously
identify these features as Au9 clusters. No aggregation of the
clusters was observed, with isolated clusters distributed
uniformly across the sheets (Figure S9b), thus suggesting
that they are attached to the GOSH and hence unable to
diffuse and coalesce.
Counting the Au9 clusters gives an estimate of the extent
of functionalization. From analysis of images such as Fig-
ure 1a, the estimated concentration of Au9 clusters is 7 Æ 2 per
100 nm2, which corresponds to just under 0.2% of the carbon
atoms in GOSH being functionalized through the sulfur to
Au9 linkage. Note that this is only an estimate of the Au9
cluster density, as the image contrast of the Au9 clusters varies
considerably depending on their orientation (as discussed
below), which complicates their identification. However, this
functionalization density would correspond to an atomic
content of 1.6 Æ 0.4% Au, which is in close agreement with
the TGA, XPS, and EDX measurements.
More detailed analysis requires comparison between the
experimental images and image simulations of the Au9
clusters on GOSH. Using the known crystal structure of
Au9, a tableau of multislice image simulations was constructed
by rotating the molecule about two orthogonal symmetry axes
(Figure S10). For Au9, the contrast is dominated by the gold
atoms (Z = 79), which scatter the electrons to a greater extent
than do the phosphorus (Z = 15) or carbon (Z = 6) atoms.
However, for accurate comparison all atoms were contained
in the image simulation, including the coordination sphere of
ligands and a section of graphene lattice.[32] Figure 1b shows
three different regions on the GOSH@Au9, each from an ac-
TEM image acquired with 0.3 s exposure. The region in
Figure 1b1 is marked by the box in Figure 1a. A comparison
with the simulations (Figure 1c) facilitates identification of
the orientation of the Au9 clusters (Figure 1d).
a close match for the contrast arising from the gold atoms and
also subtler variations in contrast that are consistent with the
ligands still being present. Further confirmation comes from
the measurement of the spacings of the atomic columns within
the clusters, which shows that the Au–Au distances in the
experimental images are consistent with those expected for
Au9 (Figure S11). This proves that intact, undamaged, Au9
NCs are present on the GOSH surface.
A study of the dynamics of these Au9 clusters demon-
strates that they are covalently bound to GOSH and not
simply adsorbed to the surface. The ac-TEM image shown in
Figure 1a was taken with a 0.3 s exposure; the match between
the experimentally observed contrast and image simulation
shows not only that Au9 is present and intact, but also that
those clusters are stationary on the GOSH surface for the
period of that exposure. However, inspection of subsequent
images shows that the clusters are not permanently fixed and
their contrast changes over time. Figure 2 shows a sequence of
ac-TEM images of a single Au9 cluster on GOSH; images
were acquired at 0.3 s intervals over a period of more than
10 s (the full image sequence is shown in Figure S12 in the
Supporting Information, with only selected images shown in
Figure 2). The images are from the same region and, through
comparison with fixed points in the larger image (Figure S9a),
show no apparent lateral displacement of the cluster relative
to the underlying GOSH. The change in contrast is indicative
of rotations of the Au9 cluster relative to the GOSH surface,
and comparison with the image simulation tableau enables
each of the images shown in Figure 2 to be identified as
specific Au9 orientations. From this it is apparent that the Au9
cluster is rotating, but without lateral displacement. This is
consistent with covalent attachment of Au9 to GOSH through
À À
the S Au bond.
The inhibited rotation is caused by interaction with the
electron beam and is indicative of a set of metastable
orientations. Prior work has observed similar electron beam
induced molecular motion on graphene oxide,[32] on carbon
nanotubes,[33,34] and for molecules attached to carbon nano-
horns, where it was shown that lower acceleration voltages in
the TEM resulted in higher frequency of molecular motion as
a result of a larger scattering cross-section.[35,36] An accelerat-
Comparison between the image simulations and exper-
imental images clearly shows that the clusters are Au9, with
9562 www.angewandte.org
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2015, 54, 9560 –9563

Angewandte
Chemie
[11] B. H. Kim, M. J. Hackett, J. Park, T. Hyeon, Chem. Mater. 2014,
ing voltage of 80 kV was used here to minimize damage to the
chemically modified graphene by the electron beam. At this
acceleration voltage, the clusters are fixed in each orientation
for timescales on the order of seconds before switching to
another orientation. This indicates that each observed
orientation is metastable, and corresponds to a local energy
minimum. As each image shows well-defined spots rather
than blurred streaks, it is also clear that the transition between
orientations must be relatively rapid.
In conclusion, we have proven that atomically designed
clusters can be covalently attached to chemically modified
graphene by taking advantage of the affinity between gold
and the sulfur functionalities present on the surface. We have
demonstrated that the Au9 clusters are intact and well-
dispersed over GOSH, and show no apparent aggregation.
Dynamic ac-TEM measurements show how a single molec-
ular cluster rotates as a result of the effect of the electron
beam, but without lateral diffusion, which is indicative of
a strong covalent interaction between Au9 and GOSH.
Moreover, the results of the dynamic study suggest the
presence of metastable orientations that may appear as
a consequence of the steric demands of the ligands.
26, 59 – 71.
[12] J. J. Pietron, R. M. Stroud, D. R. Rolison, Nano Lett. 2002, 2,
545 – 549.
[13] Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts, R. S. Ruoff,
Adv. Mater. 2010, 22, 3906 – 3924.
[14] N. M. Julkapli, S. Bagheri, Int. J. Hydrogen Energy 2014, 39,
948 – 979.
[15] D. R. Dreyer, S. Park, C. W. Bielawski, R. S. Ruoff, Chem. Soc.
Rev. 2010, 39, 228 – 240.
[16] S. Park, R. S. Ruoff, Nat. Nanotechnol. 2009, 4, 217 – 224.
[17] X. Huang, X. Qi, F. Boey, H. Zhang, Chem. Soc. Rev. 2012, 41,
666 – 686.
[18] C. Bosch-Navarro, E. Coronado, C. Martí-Gastaldo, Carbon
2013, 54, 201 – 207.
[19] P. Pandey, G. R. Bell, J. P. Rourke, A. M. Sanchez, M. D. Elkin,
B. J. Hickey, N. R. Wilson, Small 2011, 7, 3202 – 3210.
[20] G. Goncalves, P. Marques, C. M. Granadeiro, H. I. S. Nogueira,
M. K. Singh, J. Grµcio, Chem. Mater. 2009, 21, 4796 – 4802.
[21] Z. Xiong, L. L. Zhang, J. Ma, X. S. Zhao, Chem. Commun. 2010,
46, 6099 – 6101.
[22] M. Quintana, K. Spyrou, M. Grzelczak, W. R. Browne, P. Rudolf,
M. Prato, ACS Nano 2010, 4, 3527 – 3533.
[23] M. Quintana, A. Montellano, A. E. Del Rio Castillo, G. Van
Tendeloo, C. Bittencourt, M. Prato, Chem. Commun. 2011, 47,
9330 – 9332.
[24] H. R. Thomas, A. J. Marsden, M. Walker, N. R. Wilson, J. P.
Rourke, Angew. Chem. Int. Ed. 2014, 53, 7613 – 7618; Angew.
Chem. 2014, 126, 7743 – 7748.
Our approach is generally applicable to the whole family
of gold nanoclusters, and may be extended to other atomically
designed clusters. This will allow fine-tuning of the graphene–
nanocluster properties (e.g. optical or charge-transfer proper-
ties), thus permitting exploration of the effect that the size
and morphology of clusters have in applications ranging from
biosensors or biomedicine, to energy storage, or heteroge-
neous catalysis.
[25] F. Cariati, L. Naldini, J. Chem. Soc. Dalton 1972, 2286 – 2287.
ˇ
´
´
[26] P. Simek, Z. Sofer, O. Jankovsky, D. Sedmidubsky, M. Pumera,
Adv. Funct. Mater. 2014, 24, 4878 – 4885.
[27] N. R. Wilson, P. Pandey, R. Beanland, R. J. Young, I. Kinloch, L.
Gong, Z. Liu, K. Suenaga, J. P. Rourke, S. J. York, J. Sloan, ACS
Nano 2009, 3, 2547 – 2556.
[28] F. Schäffel, M. Wilson, J. H. Warner, ACS Nano 2011, 5, 9428 –
9441.
[29] C. Gómez-Navarro, J. C. Meyer, R. S. Sundaram, A. Chuvilin, S.
Kurasch, M. Burghard, K. Kern, U. Kaiser, Nano Lett. 2010, 10,
1144 – 1148.
[30] H. R. Thomas, S. P. Day, W. E. Woodru, C. Valle, R. J. Young,
I. A. Kinloch, G. W. Morley, J. V. Hanna, N. R. Wilson, J. P.
Rourke, Chem. Mater. 2013, 25, 3580 – 3588.
[31] J. P. Rourke, P. Pandey, J. J. Moore, M. Bates, I. Kinloch, R. J.
Young, N. R. Wilson, Angew. Chem. Int. Ed. 2011, 50, 3173 –
3177; Angew. Chem. 2011, 123, 3231 – 3235.
[32] J. Sloan, Z. Liu, K. Suenaga, N. R. Wilson, P. Pandey, L. M.
Perkins, J. P. Rourke, I. J. Shannon, Nano Lett. 2010, 10, 4600 –
4606.
[33] M. Koshino, N. Solin, T. Tanaka, H. Isobe, E. Nakamura, Nat.
Nanotechnol. 2008, 3, 595 – 597.
Keywords: gold nanoparticles · graphene · graphene oxide ·
nanoclusters · thiols
How to cite: Angew. Chem. Int. Ed. 2015, 54, 9560–9563
Angew. Chem. 2015, 127, 9696–9699
[1] M. Stratakis, H. Garcia, Chem. Rev. 2012, 112, 4469 – 4506.
[2] D. T. Thompson, Nano Today 2007, 2, 40 – 43.
[3] L. Dykman, N. Khlebtsov, Chem. Soc. Rev. 2012, 41, 2256.
[4] A. GutØs, C. Carraro, R. Maboudian, Biosens. Bioelectron. 2012,
33, 56 – 59.
[5] A. A. Herzing, C. J. Kiely, A. F. Carley, P. Landon, G. J.
Hutchings, Science 2008, 321, 1331 – 1335.
[6] C. Battistoni, G. Mattogno, R. Zanoni, L. Naldini, J. Electron
Spectrosc. Relat. Phenom. 1982, 28, 23 – 31.
[34] N. Solin, M. Koshino, T. Tanaka, S. Takenaga, H. Kataura, H.
Isobe, E. Nakamura, Chem. Lett. 2007, 36, 1208 – 1209.
[35] E. Nakamura, M. Koshino, T. Tanaka, Y. Niimi, K. Harano, Y.
Nakamura, H. Isobe, J. Am. Chem. Soc. 2008, 130, 7808 – 7809.
[36] R. M. Gorgoll, E. Yücelen, A. Kumamoto, N. Shibata, K.
Harano, E. Nakamura, J. Am. Chem. Soc. 2015, 137, 3474– 3477.
[7] Y. Shichibu, M. Zhang, Y. Kamei, K. Konishi, J. Am. Chem. Soc.
2014, 136, 12892 – 12895.
[8] N. R. Jana, L. Gearheart, C. J. Murphy, S. Carolina, Langmuir
2001, 17, 6782 – 6786.
[9] D. P. Anderson, J. F. Alvino, A. Gentleman, H. Qahtani, L.
Thomsen, M. I. J. Polson, G. F. Metha, V. B. Golovko, G. G.
Andersson, Phys. Chem. Chem. Phys. 2013, 15, 3917 – 3929.
[10] F. Wen, U. Englert, B. Gutrath, U. Simon, Eur. J. Inorg. Chem.
2008, 106 – 111.
Received: May 12, 2015
Published online: July 6, 2015
Angew. Chem. Int. Ed. 2015, 54, 9560 –9563
ꢀ 2015 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
9563