Thiol-functionalized undecagold clusters by ligand exchange: Synthesis, mechanism, and properties

Article abstract of DOI:10.1021/ic048686+

Ligand exchange of phosphine-stabilized undecagold precursor particles, Au₁₁(PPh₃)₈Cl₃, with ω-functionalized thiols provides a convenient and general approach for the rapid preparation of large families of thiol-stabilized, subnanometer (d _CORE ～ 0.8 nm) particles. The approach permits rapid incorporation of specific functionality into the stabilizing ligand shell, is tolerant of a wide range of functional groups, and provides convenient access to new materials inaccessible by other methods. Mechanistic studies and trapping experiments give insight into the progression of the ligand exchange, providing evidence that the core size of the phosphine-stabilized undecagold precursor particles is preserved during ligand exchange. The optical properties of the thiol-stabilized nanoparticles depend strongly on the composition of the ligand shell, and a series of studies suggests that this dependence is a result of the ligand shell's influence on the electronic structure of the particle core, as opposed to a structural change within the nanoparticle core.

Full text of DOI:10.1021/ic048686+

Inorg. Chem. 2005, 44, 6149−6158
Thiol-Functionalized Undecagold Clusters by Ligand Exchange:
Synthesis, Mechanism, and Properties
Gerd H. Woehrle and James E. Hutchison*
Department of Chemistry and Materials Science Institute, 1253 UniVersity of Oregon,
Eugene, Oregon 97403
Received September 19, 2004
Ligand exchange of phosphine-stabilized undecagold precursor particles, Au11(PPh
provides a convenient and general approach for the rapid preparation of large families of thiol-stabilized, subnanometer
0.8 nm) particles. The approach permits rapid incorporation of specific functionality into the stabilizing
3 8 3
) Cl , with ω-functionalized thiols
(
d
CORE
∼
ligand shell, is tolerant of a wide range of functional groups, and provides convenient access to new materials
inaccessible by other methods. Mechanistic studies and trapping experiments give insight into the progression of
the ligand exchange, providing evidence that the core size of the phosphine-stabilized undecagold precursor particles
is preserved during ligand exchange. The optical properties of the thiol-stabilized nanoparticles depend strongly on
the composition of the ligand shell, and a series of studies suggests that this dependence is a result of the ligand
shell’s influence on the electronic structure of the particle core, as opposed to a structural change within the
nanoparticle core.
I. Introduction
metal clusters with less than 20 core atoms have been studied
7
for use as coordination complexes, catalysts, and tagging
Metal nanoparticles with subnanometer core dimensions
are of interest for fundamental studies and may be useful as
building blocks for nanoscale devices because they are small
enough to possess discrete electronic states1 and can exhibit
semiconductor-like electronic properties.3 To study the
fundamental properties of these nanoparticles and explore
their use in specific applications, convenient access to ligand-
stabilized nanoparticles possessing terminal functional groups
is essential. Until recently, the lack of general synthetic
strategies to introduce specific functionalities into the
terminal positions of the ligand shell has limited detailed
experimental investigations of the properties and reactivities
8
,9
reagents for biomolecules, the vast majority of these
clusters are stabilized by ligands (e.g., CO, Cp, H O, halides,
2
etc.) that lack the chemical functionality needed to specifi-
cally bind the clusters to molecules, substrates, and surfaces.
,2
,4
Most examples of the introduction of functional groups
into the ligand shell of subnanometer particles involve the
1
0-13
use of functionalized triarylphosphines.
The synthesis
of functionalized triarylphosphine-stabilized undecagold
1
4
(
Au11) clusters typically requires an involved series of
synthesis and separation steps, including phosphine-for-
1
2
phosphine ligand exchange reactions, postsynthetic trans-
1
3
4
-6
formations of functional groups, and ion exchange chro-
of these materials. Although a number of small transition
*
Author to whom correspondence should be addressed. E-mail: hutch@
(7) (a) Hill, C. L.; Prosser-McCartha, C. M. Coord. Chem. ReV. 1995,
143, 407-455. (b) Pignolet, L. H.; Aubart, M. A.; Craighead, K. L.;
Gould, R. A. T.; Krogstad, D. A.; Wiley, J. S. Coord. Chem. ReV.
1995, 143, 219-263. (c) Thimmappa, B. H. S. Coord. Chem. ReV.
1995, 143, 1-34. (d) Braunstein, P.; Rose, J. Met. Clusters Chem.
1999, 2, 616-677. (e) Pignolet, L. H. Catal. Di- Polynucl. Met. Cluster
Complexes 1998, 95-126.
uoregon.edu. Phone: 1-541-346-4228. Fax: 1-541-346-0487.
(
1) Collings, B. A.; Athanassenas, K.; Lacombe, D.; Rayner, D. M.;
Hackett, P. A. J. Chem. Phys. 1994, 101, 3506-3513.
(
2) Schaaff, T. G.; Shafigullin, M. N.; Khoury, J. T.; Vezmar, I.; Whetten,
R. L.; Cullen, W. G.; First, P. N.; Wing, C.; Ascensio, J.; Yacaman,
M. J. J. Phys. Chem. B 1997, 101, 7885-7891.
(
3) Chen, S.; Ingram, R. S.; Hostetler, M. J.; Pietron, J. J.; Murray, R.
W.; Schaaff, T. G.; Khoury, J. T.; Alvarez, M. M.; Whetten, R. L.
Science 1998, 280, 2098-2101.
(8) Hainfeld, J. F. Science 1987, 236, 450-453.
(9) Jahn, W. J. Struct. Biol. 1999, 127, 106-112.
(10) Cariati, F.; Naldini, L. Inorg. Chim. Acta 1971, 5, 172-174.
(11) Bartlett, P. A.; Bauer, B.; Singer, S. J. J. Am. Chem. Soc. 1978, 100,
5085-5089.
(
(
4) Yang, Y.; Chen, S. Nano Lett. 2003, 3, 75-79.
5) Woehrle, G. H.; Warner, M. G.; Hutchison, J. E. J. Phys. Chem. B
2002, 106, 9979-9981.
(12) Jahn, W. Z. Naturforsch., B: Chem. Sci. 1989, 44, 1313-1322.
(13) Reardon, J. E.; Frey, P. A. Biochemistry 1984, 23, 3849-3856.
(14) Frey, P. A.; Frey, T. G. J. Struct. Biol. 1999, 127, 94-100.
(
6) Grant, C. D.; Schwartzberg, A. M.; Yang, Y.; Chen, S.; Zhang, J. Z.
Chem. Phys. Lett. 2004, 383, 31-34.
1
0.1021/ic048686+ CCC: $30.25
© 2005 American Chemical Society
Inorganic Chemistry, Vol. 44, No. 18, 2005 6149
Published on Web 07/06/2005

Woehrle and Hutchison
matography.9,12,13 The limited availability of functionalized
phosphines and need for multistep transformations and
separations limit the utility of these approaches as convenient
and general routes to a diverse family of functionalized
clusters. In addition, triarylphosphine-stabilized clusters
exhibit only limited solution stability, especially in acidic
solution, and are often prone to oxidative decomposition
under ambient conditions.9 A potential solution to enhanc-
ing the stability of these clusters involves stabilization
through thiol ligands. Thiol-stabilized nanoparticles are
known to exhibit higher stability than the phosphine-
stabilized nanoparticles;1 however, only a few examples
of thiol-stabilized nanoparticles with subnanometer core
dimensions are available.2
we show that the optical properties of the thiol-stabilized
exchange products depend on the nature of the ligand shell
and discuss possible causes for this dependence.
II. Experimental Section
Materials. Hydrogen tetrachloroaurate was purchased from
Strem and used as received. Dichloromethane was dried over
calcium hydride and distilled prior to use. Chloroform was filtered
through a plug of basic alumina prior to use to remove acidic
impurities. Triphenylphosphine-stabilized undecagold particles,
,15
Au11(PPh
3 8 3
) Cl (1), were synthesized as described previously and
5
19
had a core diameter of 0.8 ( 0.2 nm. 1-Azido-2,4-dinitrobenzene,
6,17
20
2
-(2-mercaptoethoxy)ethanol, and2-[2-(2-mercaptoethoxy)ethoxy]-
20
ethanol were synthesized according to known procedures. ((2,4-
Dinitrophenyl)imino)(triphenyl)phosphorane was prepared as de-
,4,5,18
We recently investigated the ligand exchange of tri-
phenylphosphine-stabilized undecagold (Au11) precursor
particles (dCORE ∼ 0.8 nm) with ω-functionalized thiols as a
reliable and convenient approach for producing functional-
17
scribed previously. All other compounds were purchased from
Aldrich and used as received.
Analytical Procedures. Nuclear magnetic resonance (NMR)
spectra were collected on a Varian Unity Inova 300 MHz instrument
5
13
31
ized, subnanometer gold particles. Initial studies demon-
equipped with a 4-channel probe ( C, 75.42 MHz; P, 121.43
1
13
MHz). For H and C NMR, chemical shifts were referenced to
the residual proton resonance of the solvent. For P NMR
spectroscopy, the spectra were referenced to H PO (external
3 8 3
strated successful ligand exchange of Au11(PPh ) Cl with a
3
1
limited set (three) of different thiols but did not investigate
the broader scope of this approach. In that study, we also
noted a surprising dependence of the optical properties of
the thiol-stabilized product particles on the thiol ligand used
during ligand exchange but could not determine whether this
was due to electronic effects of the ligand or to core size
changes taking place during the exchange. Recent mecha-
nistic work in our group showed that the ligand exchange
of 1.5-nm phosphine-stabilized gold nanoparticles with thiols
results in the loss of gold core atoms.17 Although core size
analysis by transmission electron microscopy (TEM) sug-
gested that Au11 core is preserved during ligand exchange,
small changes in the number of core atoms are undetectable
by TEM. Thus, we sought to investigate further whether core
size changes occur during the ligand exchange of phosphine-
stabilized Au11 clusters with thiols and determine whether
such changes could be responsible for the observed differ-
ences in the optical properties.
3
4
standard). X-ray photoelectron spectroscopy (XPS) was performed
on a Kratos Axis HSi instrument operating at a base pressure of
-
9
∼
5 × 10 mmHg using monochromatic Al KR radiation at 15
mA and 13.5 kV. Nanoparticle samples were drop-cast from
solution onto clean glass slides. Samples were charge compensated,
and binding energies were referenced to carbon 1s at 284.4 eV.
UV-visible spectra were obtained on a Hewlett-Packard 8453 diode
array spectrometer with a fixed slit width of 1 nm using 1-cm quartz
cuvettes. Thermal gravimetric analysis (TGA) was performed on a
TA Instruments Hi-Res TGA 2950 thermogravimetric analyzer
under nitrogen atmosphere (flow rate 100 mL/min). Samples (1-2
mg) were deposited onto Al pans as powders or by drop-casting
from dichloromethane and placed in the instrument until a stable
weight was obtained prior to analysis. The samples were heated at
a rate of 2 °C/min up to 100 °C, held at that temperature for 20
min to ensure complete solvent evaporation, and then heated to
5
(
00 °C at a rate of 1 °C/min. Transmission electron microscopy
TEM) was performed on a Philips CM-12 operating at 120 kV
accelerating voltage. Samples were prepared by aerosol deposition
of aliquots onto SiO -coated 400-mesh Cu TEM grids (Ted Pella).
Here, we report the general features and scope of the ligand
exchange reaction of Au11(PPh
3
)
8
Cl
3
with a wide range of
x
ω-functionalized thiols. The approach is convenient and
tolerates organic- and water-soluble thiols with neutral or
charged headgroups. Mechanistic studies of the exchange
reaction provide strong evidence that the core size of the
precursor particle remains unchanged during the ligand
exchange. These studies show that, during the exchange of
thiols for phosphines, the phosphine ligands are lost as free
triphenylphosphine, demonstrating that undecagold particles
follow a different mechanism for exchange reactions with
The samples were dried under ambient conditions prior to inspection
by TEM. Images were recorded and processed as described
previously.21
Synthetic Procedures. General Procedure for the Preparation
of Organic-Soluble Undecagold Nanoparticles. To a solution of
-
3
5
.0 × 10 mmol 1 in dichloromethane/1-chlorobutane (1:3; 30
mL) was added 0.1 mmol of the organic-soluble thiol. The mixture
was stirred rapidly at 55 °C until completion of the ligand exchange
reaction. The reaction time depended on the incoming ligand and
varied from 6 h for propanethiol up to 24 h for long-chain
alkanethiols. Upon completion of the exchange reaction the solvent
was removed in vacuo. To remove excess free ligand and
byproducts, the crude material was dissolved in the minimum
amount of ethanol or 2-propanol22 and purified by gel filtration
17
thiols than their larger (dCORE ∼ 1.5 nm) analogues. Finally,
(
15) Schoenauer, D.; Lauer, H.; Kreibig, U. Z. Phys. D 1991, 20, 301-
04.
16) (a) Brown, L. O.; Hutchison, J. E. J. Am. Chem. Soc. 1997, 119,
3
(
1
2384-12385. (b) Warner, M. G.; Reed, S. M.; Hutchison, J. E. Chem.
Mater. 2000, 12, 3316-3320.
(19) Bailey, A. S.; Case, J. R. Tetrahedron 1958, 3, 113-131.
(20) Woehrle, G. H.; Warner, M. G.; Hutchison, J. E. Langmuir 2004, 20,
5982-5988.
(21) (a) Brown, L. O.; Hutchison, J. E. J. Phys. Chem. B 2001, 105, 8911-
8916. (b) Woehrle, G. H.; Hutchison, J. E.; O¨ zkar, S.; Finke, R. G.
Submitted for publication.
(
17) Woehrle, G. H.; Brown, L. O.; Hutchison, J. E. J. Am. Chem. Soc.
2005, 127, 2172-2183.
(
18) (a) Schaaff, T. G.; Whetten, R. L. J. Phys. Chem. B 2000, 104, 2630-
2
4
641. (b) Negishi, Y.; Tsukuda, T. J. Am. Chem. Soc. 2003, 125,
046-4047.
6150 Inorganic Chemistry, Vol. 44, No. 18, 2005

Thiol-Functionalized Undecagold Clusters
-
2
chromatography using Sephadex LH-20 (eluent, ethanol or 2-pro-
panol; collect colored fraction).
General Procedure for the Preparation of Water-Soluble
Undecagold Nanoparticles. An aqueous solution (15 mL) of 0.1
mmol of the water-soluble thiol ligand was added to a solution of
Phosphines). A solution of hexanethiol (1.5 mg, 1.0 × 10 mmol)
-3
in DMSO-d
6
(0.6 mL) was added to 1 (5 mg, 1.2 × 10 mmol),
1
and the reaction was monitored by H NMR spectroscopy. The first
spectrum was recorded at t
min for a total time of 12 h.
Ligand Exchange between 1 and Hexanethiol in the Presence
0
) 1 min followed by spectra every 15
-
3
5
.0 × 10 mmol 1 in CHCl
3
(15 mL). The biphasic mixture was
-
2
stirred rapidly at 55 °C until completion of the ligand exchange
reaction which was monitored by the complete transfer of the
colored nanoparticles from the organic to the aqueous phase. The
reaction time depended on the incoming ligand and varied from 4
to 15 h. Upon completion of the exchange reaction the layers were
of Excess PPh
mmol) and PPh
3
. A solution of hexanethiol (1.5 mg, 1.0 × 10
-2
3
(10 mg, 4.0 × 10 mmol) in DMSO-d
6
(0.6 mL)
-
3
was added to 1 (5 mg, 1.2 × 10 mmol), and the reaction was
monitored by 1H NMR spectroscopy. The first spectrum was
recorded at t
time of 12 h.
Trapping Free PPh
solution of hexanethiol (2.8 mg, 2.4 × 10 mmol) and 1-azido-
0
) 1 min followed by spectra every 15 min for a total
separated, and the aqueous layer was washed with CH
mL) and evaporated in vacuo. The crude material was redispersed
in the minimum amount of water and purified by gel filtration
2
Cl
2
(3 × 10
3
Using 1-Azido-2,4-Dinitrobenzene. A
-
2
-
2
chromatography using Sephadex LH-20 (eluent, H
colored fraction) to remove excess free ligand and byproducts.
Alternatively, the crude products were dissolved in H O and purified
2
O; collect
2,4-dinitrobenzene (10 mg, 4.7 × 10 mmol) in DMSO-d
6
(0.6
-
3
mL) was added to 1 (5 mg, 1.2 × 10 mmol), and the reaction
1
was monitored by H NMR spectroscopy. The first spectrum was
2
by ultracentrifugation for 18 h at 340 000g. The nanoparticles
formed a highly concentrated solution at the bottom of the centrifuge
tube which could be separated from the supernatant.
Synthesis of Carboxylic Acid-Functionalized, Thiol-Stabilized
Undecagold Nanoparticles. An aqueous solution (15 mL) of 0.1
mmol of the ω-carboxyalkanethiol which was buffered to pH 8
0
recorded at t ) 1 min followed by spectra every 1 min for a total
time of 625 min.
III. Results and Discussion
In the following section, we first describe the scope of
the ligand exchange method to prepare functionalized, thiol-
stabilized Au11 clusters from a common precursor particle.
We will discuss ligand exchange reactions of Au (PPh ) -
using a 0.1 mM KH
2
PO
4
/K
2
HPO
4
buffer was added to a solution
(15 mL). The biphasic mixture
-
3
of 5.0 × 10 mmol 1 in CHCl
3
11
3 8
was stirred rapidly at 55 °C until the transfer of the colored
nanoparticles from the organic to the aqueous phase was complete.
The reaction time depended on the incoming ligand and varied from
3
Cl with a representative family of 22 methyl-terminated or
ω-functionalized thiols which include organic- and water-
soluble alkane- or arenethiols with neutral or charged
headgroups. Our results demonstrate the general nature of
this approach and its tolerance to a wide range of important
functional groups. The ease of preparation and convenient
purification of the new materials synthesized by this method
show its broad utility as a general route to functionalized
Au11 particles.
A mechanistic investigation of the ligand exchange reac-
tion by NMR spectroscopy is also described that gives
fundamental insight into the progression of the exchange
reaction and provides evidence that the number of core atoms
is preserved during ligand exchange. These results are based
upon product formation studies and trapping experiments.
3
to 15 h. The layers were separated, and the aqueous layer was
washed with CH Cl
2
2
(3 × 5 mL) and evaporated in vacuo. The
residue was dissolved in Nanopure water (10 mL) and precipitated
by acidifying with 10% HCl to about pH 2. After the mixture was
filtered, the residue was washed with 20 mL of 10% HCl, 20 mL
2
of H O, and 5 mL of methanol. If higher purity is desired, this
material can be dissolved in water and further purified by
ultracentrifugation for 18 h at 340 000g. The nanoparticles form a
highly concentrated solution at the bottom of the centrifugation tube
which could be separated from the supernatant.
NMR Monitoring of the Ligand Exchange between 1 and
Thiols. General Procedure of Monitoring the Ligand Exchange
between 1 and Hexanethiol by NMR. All NMR experiments to
monitor the ligand exchange reaction were conducted at 55 °C.
Hexanethiol was placed in an NMR tube and dissolved in DMSO-
Unlike their 1.5-nm analogues, ligand exchange of Au11(PPh
Cl with thiols does not produce AuCl(PPh ) as a byproduct,
and all phosphine ligands are liberated from the nanoparticle
as free PPh . Further evidence for the replacement of all
phosphine ligands in form of PPh is demonstrated by
experiments in which blocking of the ligand exchange was
3 8
) -
3
3
6
d , and the solution was equilibrated in the NMR instrument to 55
1
°
C. A single H NMR spectrum was obtained as a starting point
3
for the reaction. The contents of the NMR tube were then added to
a scintillation vial charged with 1. The resulting mixture was quickly
agitated until everything had dissolved and placed back into the
NMR tube. The NMR tube was returned to the instrument and
reshimmed, and spectra were collected at preset time intervals.
Ligand Exchange between 1 and Excess Hexanethiol. A
3
attempted through the addition of excess PPh
mixture.
3
to the reaction
In the final part of this section, we discuss the optical
properties of the thiol-stabilized exchange products. This
study is based upon UV-visible spectroscopy and suggests
a strong dependence of the optical properties on the nature
of the ligand shell.
-
2
solution of hexanethiol (2.8 mg, 2.4 × 10 mmol) in DMSO-d
6
-3
(
0.6 mL) was added to 1 (5 mg, 1.2 × 10 mmol), and the reaction
1 31
was monitored by either H NMR or P spectroscopy. The first
spectrum was recorded at t ) 1 min followed by spectra every 15
min for a total time of 12 h.
0
Scope of the Ligand Exchange Approach. The phos-
Ligand Exchange between 1 and a Stoichiometric Amount
of Hexanethiol (To Replace All Nanoparticle-Bound
phine-stabilized undecagold precursor particle, Au11(PPh
3 8
) -
Cl (1), was synthesized using a modification of a procedure
3
1
1
developed by Bartlett et al. and had an average core size
(
22) Most organic-soluble Au11 clusters can be dissolved in ethanol.
Exceptions are clusters stabilized with long-chain alkanethiols such
octadecanethiol and hexadecanethiol. In these cases, less polar solvents
such as 2-propanol must be used for chromatography. Alternatively,
dichloromethane or chloroform can be used.
5
of 0.8 ( 0.2 nm as determined by TEM. Ligand exchange
between 1 and alkane- or arenethiols was achieved by
combining an excess of the thiol (∼20 molar equiv) with a
Inorganic Chemistry, Vol. 44, No. 18, 2005 6151

Woehrle and Hutchison
Scheme 1. Schematic Representation of the Ligand Exchange
Reaction of Au11(PPh3)8Cl3 and ω-Functionalized Thiols
essentially no decomposition and the ligand exchange
resulted in completely exchanged nanoparticles indistin-
guishable from those obtained in chloroform; however, the
high boiling point of dimethyl sulfoxide makes its removal
from the nanoparticle samples inconvenient. The best results
were achieved using a mixture of 1-chlorobutane/dichloro-
methane (3:1) as the solvent. In this solvent system, ligand
exchange resulted in completely exchanged thiol-stabilized
Au11 clusters with no decomposition of the precursor
2
5
clusters.
Water-soluble Au11 clusters can also be prepared using
this ligand exchange approach if a biphasic solvent system
3
(water/CHCl ) is used in place of the organic solvent. As in
the case of monophasic exchanges, the organic layer of the
biphasic system can be replaced with a 3:1 mixture of
1
-chlorobutane/dichloromethane; however, we did not find
nanoparticle solution in an appropriate solvent such as
chloroform or water at 55 °C. To investigate the full scope
of the ligand exchange approach, exchanges were carried
out with a representative family of ω-functionalized thiols,
including organic- and water-soluble alkane- or arenethiols
with either neutral or charged headgroups (Scheme 1). All
thiols tested were compatible with this approach, and only a
few ligands (e.g., mercaptophenol, mercaptoethanesulfonic
acid) required special reaction conditions to achieve complete
exchange. In each case, the approach is generally applicable,
leading to complete exchange of the original phosphine
ligand shell.
significant differences in yield using either alumina-filtered
chloroform or 1-chlorobutane/dichloromethane as the sol-
2
6
vent. The pH of the aqueous layer was kept between 5 and
to avoid decomposition of the precursor particle under
8
acidic conditions and prevent significant disulfide formation
at higher pH. Typical reaction times for the biphasic ligand
exchange reactions ranged from 4 h for ligand exchanges
with (2-mercaptoethyl)trimethylammonium chloride up to 15
h using 2-[2-(2-mercaptoethoxy)ethoxy]ethanol. In exchange
reactions involving thiols that contain ionizable headgroups
(e.g., carboxylic acids), the pH must be raised to ensure
complete deprotonation of the headgroups to avoid formation
of insoluble material that can prevent complete exchange.
In nearly all cases, the use of a phosphate buffer (pH ) 8)
as the aqueous layer yielded the best results.
As in the case of exchange reactions of 1.5 nm gold
1
7,23
nanoparticles,
the reaction time depends on the thiol
ligand used. The time increases as the chain length and
bulkiness of the incoming ligand increase. However, the
ligand exchange of Au11 particles with thiols occurs only at
elevated temperatures (∼55 °C) and typically requires longer
reactions times (approximately twice as long as in the case
The thiol-stabilized ligand exchange products can be
purified either by solvent washes or, more conveniently, by
gel filtration chromatography using Sephadex LH-20 resin.
In practically every case, chromatography is more reliable,
less wasteful, and faster than solvent washes. The use of the
LH-20 resin allows for rapid purification of both organic-
and water-soluble exchange products with a wide range of
solvents, including alcohols, chlorinated organics, and water.
The recovery of nanoparticle material from the column is
nearly quantitative, and the support can be reused after
sufficient rinsing.
of 1.5 nm gold nanoparticles) due to the lower reactivity of
Au11 clusters.24
Ligand exchange reactions with organic-soluble thiols were
carried out in a single organic phase and have been applied
to a wide variety of alkane- and arenethiols (Scheme 1).
Reaction times depend on the incoming ligand and vary from
6
h for propanethiol to 24 h for long-chain alkanethiols.
Ligand exchange reactions with organic-soluble thiols were
initially carried out in chloroform; however, this solvent leads
to the rapid decomposition of the precursor cluster (1) and,
thus, reduced yield. Filtering the chloroform through basic
alumina immediately prior to use improved the yields, but
partial decomposition of 1 still occurred.
To determine the chemical composition of the thiol-
stabilized exchange products and ensure sufficient purity,
we analyzed each exchange product using a combination of
NMR, UV-visible spectroscopy, TEM, TGA, and XPS
1
(
Table 1). We used H NMR spectroscopy to confirm the
purity of each sample. The resonances associated with ligands
bound to the gold core show significant line broadening.27
The absence of sharp resonances indicates that excess free
ligand and byproducts of the ligand exchange have been
To avoid the difficulties arising from particle decomposi-
tion in chloroform, a number of alternative reaction media
were investigated, including alcohols, ethyl acetate, acetone,
and dimethyl sulfoxide. Use of alcohols, ethyl acetate, or
acetone eliminated decomposition of 1 but resulted in
incomplete exchange due to precipitation of the particles
during the exchange. In dimethyl sulfoxide, 1 showed
2
5
removed. When nonaromatic thiols are used for the
(
25) See Supporting Information.
(26) This is probably due to the rapid transfer of the exchanging particles
from the organic to the aqueous phase.
(
23) Hostetler, M. J.; Templeton, A. C.; Murray, R. W. Langmuir 1999,
(27) Terrill, R. H.; Postlethwaite, T. A.; Chen, C.-h.; Poon, C.-D.; Terzis,
A.; Chen, A.; Hutchison, J. E.; Clark, M. R.; Wignall, G.; Londono,
J. D.; Superfine, R.; Falvo, M.; Johnson, C. S. H., Jr.; Samulski, E.
T.; Murray, R. W. J. Am. Chem. Soc. 1995, 117, 12537-12548.
15, 3782-3789.
(24) At temperatures below 40 °C, no ligand exchange is observed over a
period of several days.
6152 Inorganic Chemistry, Vol. 44, No. 18, 2005

Thiol-Functionalized Undecagold Clusters
Table 1. Analytical Data for the Ligand Exchange Products
1H NMR chem shift (ppm)
and NMR solvent
ligand
reaction time, h
6
dCOREa
S:Au ratiob
0.96
% ligand by massc
23.9
HS(CH2)2CH3d
HS(CH2)5CH3d
0.8 ( 0.2 (N ) 1103)
1.15 (b), CDCl3
1.79 (b)
10
0.8 ( 0.2 (N ) 623)
0.90 (b), CDCl3
0.93
36.9
1
1
.32 (b)
.80 (b)
HS(CH2)7CH3d
HS(CH2)11CH3d
HS(CH2)15CH3d
HS(CH2)17CH3d
14
16
20
24
0.7 ( 0.2 (N ) 920)
0.8 ( 0.3 (N ) 1032)
0.8 ( 0.2 (N ) 810)
0.8 ( 0.2 (N ) 623)
0.88 (b), CDCl3
.27 (b)
0.90 (b), CDCl3
.25 (b)
0.89 (b), CDCl3
.27 (b)
0.89 (b), CDCl3
.26 (b)
0.86
0.91
0.87
0.89
41.6
47.1
53.8
56.1
1
1
1
1
4
4
-mercaptophenold
-methylbenzenethiol
24h
20
0.8 ( 0.2 (N ) 902)
0.7 ( 0.2 (N ) 718)
7.23 (b), CD3OD
1.80 (b), CDCl3
0.92
0.89
36.7
35.7
d
7.11 (b)
4
-mercaptobiphenyld
20
18
0.8 ( 0.2 (N ) 512)
0.8 ( 0.3 (N ) 627)
7.32 (b), CD2Cl2
3.40 (b), CD3OD
0.86
1.01
44.8
46.1
d
HS(CH2)3Si(OCH3)3
3.91 (b)
HS(CH2)COOHe
HS(CH2)2COOH
HS(CH2)5COOH
3
4
8
0.9 ( 0.2 (N ) 651)
0.8 ( 0.2 (N ) 564)
0.9 ( 0.3 (N ) 812)
3.10 (b), CD3OD
3.30 (b), CD3OD
1.21 (b), CD3OD
0.81
0.76
0.80
28.8
31.1
38.9
e
e
2.10 (b)
HS(CH2)11COOHe
12
0.8 ( 0.2 (N ) 712)
1.31 (b), CD3OD
0.74
46.8
1
2
.55 (b)
.14 (b)
HS(CH2)2O(CH2)2OHf
HS(CH2)2O(CH2)2O(CH2)2OH
HS(CH2)2PO(OH)2
10
15
7
0.8 ( 0.2 (N ) 1034)
0.8 ( 0.3 (N ) 832)
0.9 ( 0.2 (N ) 591)
3.73 (b), D2O
3.71 (b), D2O
3.51 (b), D2O
0.90
0.83
1.02
35.0
41.5
39.7
f
f
3
2
.27
.26 (b)
HS(CH2)2NMe2‚HClf
9
0.9 ( 0.2 (N ) 704)
2.93 (b), D2O
0.74
35.1
3
3
.55 (b)
.80 (b)
+
- f
HS(CH2)2NMe3 Cl
HS(CH2)2O(CH2)2NMe3 Cl
HS(CH2)2O(CH2)2O(CH2)2NMe3 Cl
4
8
0.8 ( 0.3 (N ) 923)
0.8 ( 0.2 (N ) 689)
0.8 ( 0.3 (N ) 834)
0.8 ( 0.2 (N ) 783)
3.13 (b), D2O
.52 (b)
3.11 (b), D2O
.62 (b)
1.65 (b), D2O
.18 (b)
3.42 (b), D2O
.75 (b)
0.84
0.79
0.92
0.94
38.0
44.2
49.7
39.1
3
+
- f
3
+
- f
9
3
-
+ f,g
HS(CH2)2SO3 Na
10
3
a
b
Core diameter in nm (mean ( std dev) from analysis of representative TEM images. N refers to the number of particles measured. Ratio obtained from
quantification of the areas of the XPS signals. Obtained from TGA analysis. d Synthesized according to the general procedure for the preparation of organic-
soluble nanoparticles. e Synthesized according to the general procedure for the preparation of water-soluble nanoparticles but using a 3:1 KH2PO4/K2HPO4
buffer (pH 8) as the aqueous phase. f Synthesized according to the general procedure for the preparation of water-soluble nanoparticles. This material is
c
g
5
h
obtained in a two-step synthesis as described elsewhere.
The reaction is carried out in CH3Cl:MeOH (1:1).
1
exchange reaction, H NMR spectroscopy can also be used
to confirm the completion of the ligand exchange (through
the disappearance of the triphenylphosphine signals).
The core sizes of the exchange products were determined
by analyzing representative TEM images for each nano-
particle sample.21 The typical average core size of 0.8 (
considered. However, UV-visible spectroscopy was not
useful for probing the nanoparticle core size of the product
particles because the optical properties of the nanoparticle
are dependent on nature of the ligand shell, as described
below.
The chemical composition of each exchange product was
determined using a combination of XPS and TGA (Table
0
.2 nm (N > 500) is the same as the core size of the
phosphine-stabilized precursor particle 1 (see Table 1).
Because the optical properties of subnanometer gold clusters
are highly sensitive to the number of core atoms,
determining the core size by UV-visible spectroscopy was
2
5,28
1
). The absence of phosphorus (with the exception of
mercaptoethanephosphonic acid-stabilized particles) in the
XPS analysis confirmed complete ligand exchange and
removal of all phosphine-containing byproducts within the
detection limits of the instrument. For organic-soluble
exchange products, XPS analysis also revealed the absence
of chloride. Quantitative XPS analyses gave sulfur:gold ratios
of organic-soluble exchange products ranging from 0.86:1.00
to 1.01:1.00, which correlate to ∼9-11 thiols/nanoparticle.
Water-soluble exchange products showed slightly lower
sulfur:gold ratios of 0.74:1.00 to 1.02:1.00, corresponding
1
,29,30
(
28) The reported size dispersity of ∼30% is due to the low contrast of
the samples, the presence of touching particles, and irregularities in
the support film which limits the accuracy of the analysis. Even in
the case of the monodisperse, phosphine-stabilized undecagold precur-
sor, size analysis gives 30% size dispersity.
(
29) Hall, K. P.; Mingos, D. M. P. Prog. Inorg. Chem. 1984, 32, 237-
325.
(
30) Haeberlen, O. D.; Chung, S.-C.; Stener, M.; Roesch, N. J. Chem. Phys.
997, 106, 5189-5201.
1
Inorganic Chemistry, Vol. 44, No. 18, 2005 6153

Woehrle and Hutchison
to 8-11 thiols/nanoparticle. The sulfur:gold ratios of the
exchange products obtained by XPS and TGA analyses are
consistent with one another. The slightly lower thiol coverage
of certain water-soluble exchange products could be due to
some unexchanged chloride ligands or the presence of oxo
ligands that could be produced by oxidation of some of the
uncoordinated surface Au atoms during ligand exchange.
The ligand exchange approach is general and tolerates a
surprisingly wide range of functional groups, using either a
mono- or biphasic system. The approach produces function-
alized Au11 particles with technologically relevant headgroups
exchange between 1 and hexanethiol was carried out in an
NMR tube at 55 °C using a 20-fold molar excess of the thiol.
Deuterated DMSO was chosen as solvent to eliminate the
decomposition of 1 to AuCl(PPh
3
) that occurs rapidly in
34
3
CDCl . The exchange reaction was monitored until no more
changes in the spectrum were observed (∼12 h). At the end
31
of the reaction, the P NMR spectrum showed a major peak
at 26.7 ppm, corresponding to triphenylphosphine oxide, and
a small peak at 32.6 ppm, which was attributed to a trace
3
amount of AuCl(PPh ), produced either by partial decom-
position of the phosphine-stabilized precursor under the
exchange conditions or as byproduct of the exchange
31
such as alcohols for biological applications, phosphonic
acids and silanes for surface applications, and quaternary
ammonium headgroups for self-assembly on DNA tem-
3
4
reaction. Signals for phosphines bound to the Au11 core
(δ ) 53.2 ppm) or free PPh (δ ) -5.5 ppm) were not
3
2
0,32
plates,
making this method potentially useful in many
detected.
different research areas (e.g., nanoelectronics, biotagging,
catalysis, etc.). In addition, the general nature of the approach,
in combination with the ease of preparation and convenient
purification, allows for the rapid preparation of large families
of thiol-stabilized Au11 particles.33
Triphenylphosphine oxide can either be formed by oxida-
tion of the phosphine ligands on the nanoparticle surface (i.e.,
triphenylphosphine oxide is the leaving group of the ligand
exchange) or is produced by the subsequent oxidation of
35
liberated PPh
3
in solution. To determine the origin of
triphenylphosphine oxide, a trapping experiment was used
to selectively probe for the presence of free PPh in solution.
Similar to a study reported previously, we chose 1-azido-
,4-dinitrobenzene as the trapping reagent because it selec-
3 8 3
Mechanism of the Ligand Exchange of Au11(PPh ) Cl
and Thiols. Only a few mechanistic investigations have been
reported that provide information about the course and
outcome of ligand exchange reactions of ligand-stabilized
3
1
7
2
1
7,23
tively undergoes a fast Staudinger reaction with PPh to
nanoparticles.
In this study, we investigated the mech-
3
irreversibly form an iminophosphorane which is easily
anism for the replacement of the phosphine ligands by thiols
to determine whether the core size is preserved during ligand
exchange. Recent studies of the ligand exchange reaction of
3
6
observable by NMR. Triphenylphosphine oxide, produced
as the leaving group of the ligand exchange reaction, would
not result in the formation of iminophosphorane because the
trapping reagent does not react with triphenylphosphine
oxide. If triphenylphosphine oxide is formed by the oxidation
1
“
.5 nm triphenylphosphine-stabilized gold nanoparticles,
”, with alkanethiols showed that part of the
phosphine ligand shell is replaced as AuCl(PPh ), leading
) produced
3 5
Au101(PPh )21Cl
3
17
of liberated PPh , rapid formation of iminophosphorane
to loss of core atoms. The amount of AuCl(PPh
3
3
would be observed. In addition, the amount of triphenylphos-
phine oxide will be greatly reduced because PPh is rapidly
3
removed from the mixture by the irreversible reaction with
the trap.
When the ligand exchange reaction of 1 with an excess
of hexanethiol is performed in the presence of 1-azido-2,4-
dinitrobenzene, the formation of iminophosphorane is ob-
served throughout the course of the ligand exchange reaction.
Only trace amounts of triphenylphosphine oxide are produced
in the presence of the trapping reagent. Since the formation
is limited by the number of chlorides available in the original
ligand shell which means that about 5% of the core atoms
are lost as result of the exchange reaction.17
In the case of Au11(PPh
of the phosphine ligands as AuCl(PPh
loss of three Au atoms and produce nanoparticles with a Au
3
)
8
Cl
3
, a similar, partial replacement
3
) would result in the
8
core according to the number of chlorides in the original
ligand shell. Since the optical properties of gold clusters with
6
to 13 core atoms are highly sensitive to the number of
1,29
8
core atoms, a change from a Au11 core to a Au during
of iminophosphorane only occurs in the presence of free PPh
in solution, this observation indicates that the phosphine
3
ligand exchange should lead to significant changes in the
optical properties of the exchange products. These concerns
led us to investigate the course of the ligand exchange
reactions of the Au11 core.
ligands dissociate from the clusters as free PPh
3
and not
triphenylphosphine oxide.
To investigate if the trace amount of AuCl(PPh
3
) observed
The progression of the ligand exchange reaction between
and hexanethiol was followed by in situ P NMR
spectroscopy to monitor the formation of phosphine-contain-
ing byproducts. To ensure complete exchange, the ligand
31
during our initial NMR studies was formed as leaving group
of the ligand exchange or was a result of the slow
decomposition of 1 under the reaction conditions, we set up
an experiment to block the ligand exchange by addition of
1
(
(
(
31) Mrksich, M. Chem. Soc. ReV. 2000, 29, 267-273.
32) Warner, M. G.; Hutchison, J. E. Nat. Mater. 2003, 2, 272-277.
33) Thiol-stabilized exchange products show an increased stability towards
heat, pH changes, and salt addition with respect to the phosphine-
(34) In a control experiment, we found that Au11(PPh3)8Cl3 produces only
trace amounts of AuCl(PPh3) in DMSO-d6 at 55 °C over the course
of 12 h.
(35) In a control experiment, we found that PPh3 is oxidized to tri-
phenylphosphine oxide in DMSO-d6 at 55 °C over the course of 2 h.
(36) (a) Triphenylphosphine oxide and gold-bound PPh3 ligands do not
react with the azide under identical conditions. (b) Staudinger, H.;
Meyer, J. HelV. Chim. Acta 1919, 2, 635-646.
5
stabilized precursor particle 1. In addition, while water-soluble,
phosphine-stabilized undecagold derivatives are being slowly oxidized
when stored in solution for a few days or at low pH,9 a similar
sensitivity to oxidation has not been observed for 1 and the thiol-
stabilized exchange products.
6154 Inorganic Chemistry, Vol. 44, No. 18, 2005

Thiol-Functionalized Undecagold Clusters
Figure 1. Evolution of the R-methylene (R-CH2) resonance of hexanethiol as a function of time for the ligand exchange of 1 with hexanethiol monitored
1
by H NMR spectroscopy. (A) The reaction was carried out in the presence a stoichiometric amount of hexanethiol to replace all of the ligands of 1. The
intensity of the R-CH2 resonance decreases over time as the thiols are exchanged onto the nanoparticle until it is broadened into the baseline after completion
of the exchange. (B) The reaction was carried out under identical conditions as the experiment in (A) but in the presence of a 4-fold molar excess of PPh3
over hexanethiol added prior to ligand exchange. The intensity of the R-CH2 resonance remains unchanged throughout the measurement, indicating complete
blocking of the ligand exchange.
an excess of free PPh
Addition of excess of PPh
only inhibit the loss of phosphine ligands as free PPh
not their replacement as AuCl(PPh ). If AuCl(PPh ) was a
leaving group of the ligand exchange, addition of excess PPh
should only lead to partial blocking of the ligand exchange.
Figure 1 shows the effect of addition of excess PPh to
3
to the ligand exchange mixture.
to the reaction mixture should
but
liberated from the nanoparticle as free PPh
the phosphine ligands as AuCl(PPh ) would lead to partial
ligand exchange. Therefore, the observed AuCl(PPh ) is
most likely caused by partial decomposition of 1 under the
reaction conditions and not a result of the ligand exchange.
The experimental evidence from both the trapping studies
and the blocking experiment suggests that the ligand
exchange of Au11 clusters with thiols follows a reaction
mechanism different from the analogous exchange involving
3
because loss of
3
3
1
7
3
3
3
3
3
3
the ligand exchange mixture on the progression of ligand
exchange. The extent of ligand exchange can be monitored
1
3 21 5 101 3 21
“Au101(PPh ) Cl ”. The ligand exchange of “Au (PPh ) -
by the change in intensity of the H NMR resonance for the
Cl
which about 25% of the phosphine ligands are replaced as
AuCl(PPh
) during the initial stage of the ligand exchange.17
The remaining phosphine ligands are removed in a subse-
quent stage by transfer to closely associated AuCl(PPh ) to
5
” with thiols proceeds in a three-stage mechanism in
R-methylene protons (CH protons closest to the sulfur) as
2
the incoming thiol is exchanged onto the nanoparticle surface.
In Figure 1A, the evolution of the R-methylene proton peaks
is shown as a function of time for a ligand exchange reaction
between 1 and a stoichiometric amount of hexanethiol to
completely replace all of the ligands of 1 (∼11 molar equiv
3
3
form polyphosphine Au complexes. During the final stage
of the exchange reaction the thiol ligand shell is being
completed and reorganizes into a more crystalline state
3 8 3
of thiol to Au11(PPh ) Cl ). As the ligand exchange progresses,
1
the intensity of the H NMR resonance for the R-methylene
protons decreases until it broadens into the baseline toward
the end of the exchange reaction. Approximately 200 min
after the start of the ligand exchange, the resonance for the
R-methylene protons of the free thiol is not observable by
NMR, indicating that all thiol is removed from solution and
exchanged completely onto the nanoparticle.
3 8 3 3
(Scheme 2A). In the case of Au11(PPh ) Cl , AuCl(PPh ) is
not produced as a leaving group ligand exchange with thiols.
Instead, the complete ligand shell is removed in form of free
PPh (Scheme 2B).
3
The results from our mechanistic studies provide strong
evidence that the Au11 core remains intact during the ligand
exchange reaction, demonstrating the ability to control the
core size of the product particles during the synthesis of the
phosphine-stabilized precursor nanoparticles. Thus, it is
possible to predict the core size of the product particles from
the core size of the precursor.
2
Figure 1B shows the evolution of the R-CH protons over
time for a ligand exchange reaction between 1 and stoichio-
metric hexanethiol in the presence of a 4-fold molar excess
of PPh
3
over the amount of hexanethiol. In contrast to the
shown
exchange reaction without the addition of excess PPh
3
The fate of the protons during the ligand exchange reaction
remains unknown. The absence of a signal for the HS protons
in Figure 1A, the intensity of the resonance the R-methylene
protons remains unchanged over the complete course of the
experiment (∼420 min) as indicated by the constant integra-
1
37
in the H NMR spectra of the thiol-stabilized exchange
products suggest that the HS protons are being lost as a result
2
tion ratio of the R-CH protons and the terminal methyl
of ligand exchange. It is unclear whether the HS protons
protons. These results show that no thiol is removed from
solution in the presence of an excess PPh , indicating that
the ligand exchange reaction is blocked completely. Com-
plete inhibition of the ligand exchange by an excess of PPh
provides additional evidence that all phosphine ligands are
+
are lost in form of H or as H
2
. The NMR experiments did
3
not provide any evidence for either one of the two species
3
(
37) Hasan, M.; Bethell, D.; Brust, M. J. Am. Chem. Soc. 2002, 124, 1132-
1133.
Inorganic Chemistry, Vol. 44, No. 18, 2005 6155

Woehrle and Hutchison
Scheme 2. Differences in the Mechanism for the Ligand Exchange of Phosphine-Stabilized Gold Nanoparticles with ω-Functionalized Thiols
possibly due to the insufficient sensitivity. Another possibility
is that the sulfur-bound hydrogen is transferred to an exiting
ligand, e.g., chloride or triphenylphosphine, although these
species were not observed in our experiments. Such a
mechanism was recently shown for the exchange of thiols
onto sulfide-stabilized Au nanoparticles in which intact thiols
adsorb onto the nanoparticle surface and subsequently
transfer hydrogen to a leaving ligand.37
similar to those of the phosphine-stabilized precursor 1 with
maxima at 318 nm, 330, 375, and 416 nm (due to interband
transitions) but show an additional, broad peak at 690 nm.
The latter peak was attributed to the excitonic transition of
4
the subnanometer core and is believed to reflect the
38
semiconductor character of the Au11 cluster. In contrast to
the spectrum of ODT-stabilized Au11, the spectra for water-
Optical Properties of the Thiol-Stabilized Product
Particles. Access to a large family of functionalized Au11
particles provided by the ligand exchange method offers the
opportunity to study the influence of the ligand shell on the
optical properties. Although a number of Au11 clusters have
been prepared, their optical properties have not been studied
in detail until recently.4 Phosphine-stabilized Au11 clusters
have a characteristic UV-vis spectrum with maxima at 415,
-6
3
85, and 301 nm which are only slightly sensitive to changes
9,29
in the phosphine and halide ligands.
Storage of water-
soluble Au11 clusters at low pH or addition of oxidizing
agents leads to noticeable shifts of the maxima to 425, 366,
and 324 nm. These changes are completely reversible on
addition of a small amount of a reducing agent (e.g., NaBH
4
)
and are believed to reflect some rearrangement in the gold
2
9
core.
The optical properties of several thiol-stabilized Au11
cluster also have been studied recently. Chen et al. investi-
gated the electronic structure of dodecanethiol-stabilized Au11
(
Au11SC12) by voltammetric and spectroscopic measure-
4
ments. By comparing the voltammograms of Au11SC12 with
those of Au11(PPh Cl , they found that the band gap of the
3
)
7
3
thiol-stabilized cluster is about 0.35 eV larger than for the
phosphine-stabilized cluster. Under the assumption that the
core dimensions remain unchanged during the exchange
reactions, this observation was attributed to the stronger
bonding of Au-S compared to that of Au-P(Cl). Further-
more, the authors found that thiol-stabilized Au11 clusters
(in contrast to the phosphine-stabilized precursor particles)
exhibit photoluminescence and behave similarly to indirect
band-gap semiconductors.
Figure 2. Dependence of the UV-visible spectra of thiol-stabilized Au1
1
particles on the nature of the ligand shell. The spectra were normalized at
600 nm. (A) Comparison of the optical properties of organic-soluble and
In the course of our studies, we observed distinct differ-
ences in the optical spectra of organic-soluble and water-
soluble thiol-stabilized exchange products (Figure 2A).
Undecagold particles stabilized with long-chain alkanethiols,
such as octadecanethiol (ODT), have absorbance spectra
water-soluble Au11 particles stabilized by (a) octadecanethiol, (b) mercap-
toethanesulfonate, (c) mercaptopropionic acid, (d) mercaptoethanephos-
phonic acid, and (e) N,N-dimethylaminoethanethiol. (B) Dependence of the
UV-visible spectra of organic-soluble Au11 particles stabilized with straight-
chain alkanethiols on the chain length. The stabilizing ligands were (a)
octadecanethiol, (b) dodecanethiol, (c) octanethiol, and (d) hexanethiol.
6156 Inorganic Chemistry, Vol. 44, No. 18, 2005

Thiol-Functionalized Undecagold Clusters
soluble exchange products do not show defined peaks (Figure
2A).
A surprising trend was observed for the optical properties
of organic-soluble, thiol-stabilized Au11 particles when the
chain length of the stabilizing alkanethiol ligand was
decreased. Figure 2B shows the optical spectra of thiol-
stabilized Au11 particles with ligand shells composed of
octadecanethiol, dodecanethiol, octanethiol, and hexanethiol.
In this series of straight-chain alkanethiol-stabilized Au11
particles, the defined maxima at 318, 330, and 416 nm
broadened more and more with decreasing chain length of
the thiol ligands. The Au11SC18 particles showed the signature
peaks of Au11 particles observed in the absorbance spectrum
of 1; however, the optical spectra the Au11 particles stabilized
with the shortest thiol in the series, hexanethiol, resembled
the optical spectra of water-soluble, thiol-stabilized Au11
particles. This trend was unexpected considering the fact that
the only known difference between the Au11 clusters was
the chain length of the stabilizing thiol ligand.
Figure 3. UV-visible spectrum of (a) organic-soluble Au11 particles
obtained by the ligand exchange of 1 with hexanethiol, (b) the thiol-stabilized
Au11 particles from (a) after a second ligand exchange with ODT, and (c)
Au11SC18 prepared directly by ligand exchange of 1 with ODT. All spectra
were normalized at 600 nm.
Possible explanations for the observed differences in the
optical spectra are changes in core size and size dispersity.
It is known that the optical and electronic properties of
subnanometer gold clusters are strongly dependent on the
number of core atoms.1,29,30 Although our mechanistic studies
suggested that no core atoms are lost during ligand exchange,
it was possible that ligand exchange lead to ripening and
rearrangement of nanoparticle core, resulting in polydisperse
undecagold core. The absorbances in the spectrum of Au11
SC18 by ligand exchange of Au11SC and ODT were not as
defined as in the spectrum of Au11SC18 synthesized directly
from Au11(PPh Cl . A possible explanation for this obser-
vation is the presence of nanoparticles with mixed hex-
anethiol/ODT ligand shells. Thiol-for-thiol ligand exchange
reactions of larger gold nanoparticles are typically incomplete
and result in particles with mixed ligand shells.40 Nano-
particles with mixed hexanethiol/ODT ligand shells might
exhibit optical properties that lie between those of Au11SC18
-
6
)
3 8
3
3
9
samples.
To probe if the differences in core size and dispersity cause
the different optical spectra of short-chain and long-chain
thiol-stabilized Au11 particles, we prepared Au11SC18 particles
by ligand exchange of hexanethiol-stabilized Au11 particles,
Au11SC , with ODT and compared the absorbance spectrum
6
with the one of Au11SC18 prepared directly by ligand
6
and Au11SC particles.
Because the differences in the optical spectra appeared to
be due to effects of the ligand shell and not to differences in
core size and dispersity, we investigated if the different ligand
shells of the nanoparticle samples affected the optical
properties by influencing interparticle interactions and ag-
gregation. Since these phenomena are concentration-depend-
ent, the optical properties of long- and short-chain thiol-
stabilized Au11 particles were studied over a concentration
exchange of 1 with ODT. Thiol-for-thiol ligand exchanges
have been show to preserve the core size;2 thus, the Au11
3
-
SC18 particles prepared from Au11SC
possess the same core dimensions as the Au11SC
Figure 3 shows the absorbance spectra of a sample of Au11
SC particles before and after ligand exchange with ODT.
6
particles should
6
precursor.
-
4
1
range of 0.2-3.0 mM. We found a linear relationship
between the absorbance and the concentration, indicating that
interparticle interactions or aggregation do not have notice-
6
Before ligand exchange, the spectrum was featureless, as is
typical for undecagold particles stabilized by short-chain
alkanethiols (Figure 3, trace a). After ligand exchange of
this sample with ODT (Figure 3, trace b), the spectrum was
similar to the spectrum of Au11SC18 synthesized directly from
2
5
able effects on the optical spectra.
We also investigated if solvation effects cause the observed
difference in the optical properties. The optical properties
of thiol-stabilized Au11 particles were studied in various
solvents including chloroform, ethanol, toluene, THF, and
3 8 3
Au11(PPh ) Cl with two distinct maxima at 322 and 416
nm (Figure 3, trace c).
1
-chlorobutane, but we did not observe a noticeable depen-
The similarity between the spectra of the two Au11SC18
samples prepared by the two methods suggests that both
materials possess the same core size, suggesting that the
initial thiol-for-phosphine ligand exchange of 1 as well as
the subsequent thiol-for-thiol exchange reaction preserve the
dence on the nature of the solvent in the investigated
2
5
concentration range.
Our optical studies indicate that during ligand exchange
reactions of 1 with thiols no core size changes occur,
confirming the results from the mechanistic investigation.
We could further rule out that interparticle interactions and
(
38) (a) Alivisatos, A. P. J. Phys. Chem. 1996, 100, 13226-13239. (b)
Zhang, J. Z. Acc. Chem. Res. 1997, 30, 423-429.
(
39) It is known that ligand exchange reactions can result in growth of the
nanoparticles and changes in the size dispersity. For example, ligand
exchange of 1.5-nm phosphine-stabilized Au nanoparticles with
alkylamines leads to nanoparticle growth and ripening (see: Brown,
L. O.; Hutchison, J. E. J. Am. Chem. Soc. 1999, 121, 882-883).
(40) Hostetler, M. J.; Green, S. J.; Stokes, J. J.; Murray, R. W. J. Am.
Chem. Soc. 1996, 118, 4212-4213.
(41) The concentration range was chosen to include the concentration
typically used for UV-visible spectroscopy of the nanoparticle
samples.
Inorganic Chemistry, Vol. 44, No. 18, 2005 6157

Woehrle and Hutchison
solvation effects are responsible for the pronounced depen-
dence of the optical properties of alkanethiol-stabilized Au11
particles on the chain length of the thiol ligands. These results
suggest that the nature of the thiol ligand has a direct
influence on the electronic structure of the Au11 core.
Although it is known that thiols can act as electron acceptors
showed that the ligand exchange of these particles follows
a pathway different from that for ligand exchanges of larger
gold nanoparticles such as “Au (PPh ) Cl ”. Optical studies
of the products show a strong dependence on the nature of
the stabilizing thiol ligands. The ligand exchange approach
offers the opportunity to study and tune systematically the
optical and electronic properties of the nanoparticles through
manipulation of the ligand shell.
101
3 21
5
4
decreasing the electron density of the gold core, and possibly
influencing the electronic structure of the gold core, the thiols
used in this study have only minute differences in the electron
acceptor properties. An improved understanding of these
ligand effects would offer the opportunity to tune the optical
properties of nanoparticles with the same core size through
choice of the ligand shell.
Acknowledgment. Support for this work was provided
by the National Science Foundation (Grant DMR-9705343).
We thank Shaun Swartz for his assistance in synthesizing
the propionic acid-stabilized undecagold particles and Lallie
McKenzie and Darren Johnson for their assistance in
designing and preparing the models used in generating the
cover art.
IV. Conclusions
The ligand exchange chemistry of phosphine-stabilized
Au11 clusters with ω-functionalized thiols is shown to be a
powerful synthetic method that provides convenient access
to a diverse family of functionalized Au11 clusters. The
general nature of this ligand exchange approach, in combina-
tion with the ease of preparation, makes it of broad utility.
The approach is general and shows the high tolerance for a
wide variety of functional groups. Mechanistic studies
provided conclusive evidence that the Au11 core of the
precursor particle remains intact during ligand exchange and
Supporting Information Available: TEM images and NMR
spectra of representative nanoparticle samples, thermal stability of
1
in 1:3 dichloromethane/1-chlorobutane, effects of the concentra-
tion on the optical properties of organic-soluble undecagold
particles, and effects of the solvent on the optical properties of
organic-soluble undecagold particles. This material is available free
of charge via the Internet at http://pubs.acs.org.
IC048686+
6158 Inorganic Chemistry, Vol. 44, No. 18, 2005