Published on Web 08/25/2006
Generalized Route to Metal Nanoparticles with Liquid Behavior
Scott C. Warren,†,‡ Matthew J. Banholzer,† Liane S. Slaughter,† Emmanuel P. Giannelis,‡
Francis J. DiSalvo,† and Ulrich B. Wiesner*,‡
Department of Chemistry & Chemical Biology and Department of Materials Science & Engineering,
Cornell UniVersity, Ithaca, New York 14853
Received June 23, 2006; E-mail: ubw1@cornell.edu
Scheme 1. Synthesis of Liquid Metal Nanoparticles
Most metals, especially transition metals, must be heated above
1000 °C to flow as a liquid.1 This basic physical property of metals
has strongly influenced the range of applications for this broad class
of materials. At the same time, the high melting point has severely
limited the ways that many metals can be handled and used. For
example, it is unlikely that molten platinum will be cast in a rubber
mold or used as a solvent for organic reactions. Here, we
demonstrate that metals can be made to flow like liquids at room
temperature by attaching an appropriate ligand to a metal nano-
particle. This new class of materials allows metals to be handled,
assembled, and used in ways not previously possible. Furthermore,
since the dynamics of nanoparticle flow contribute favorably to
the thermal conductivity2 and the fully accessible phonon modes
of the metal increase the heat capacity of the liquid,3 such materials
may be of interest in heat-transfer fluids.
Numerous strategies for metal and ceramic nanoparticle func-
tionalization have been explored. Although thiol-containing poly-
mers,4,5 ionic liquids,6 and amphiphilic molecules7 have been
attached to metal nanoparticles previously, liquid-like properties
have not been reported. In the case of ceramic nanoparticles,
Giannelis et al. have demonstrated liquid-like properties when a
silicon alkoxide with a quaternized ammonium was covalently
attached to the nanoparticle surface.8-10
by centrifugation in methanol and then 90:10 water/methanol,
afforded the pure product 3, a powdery solid. This synthetic
approach was generalized to gold, palladium, and rhodium nano-
particles (Supporting Information).
We synthesized the thiol N,N-dioctyl-N-(3-mercaptopropyl)-N-
methylammonium bromide (2) in a simple, two-step process
(Scheme 1). Reaction of excess 1,3-dibromopropane with N,N-
dioctyl-N-methylamine, followed by vacuum distillation of the
excess 1,3-dibromopropane, afforded 1 in 92% yield. The remaining
8% comprised the â-hydrogen rearrangement products N,N-dioctyl-
N-methylammonium bromide and allyl bromide; the allyl bromide
was removed during vacuum distillation. Upon dissolving the
product in methanol, we added sodium hydrosulfide incrementally
and stirred the solution at room temperature. Removal of methanol
and dissolution in chloroform precipitated the sodium bromide and
excess sodium hydrosulfide, which were removed by filtration.
Evaporation of the chloroform in vacuo afforded a product mixture
that included the thiol 2 (43%), the disulfide of 2 (35%), the sulfide
of 2 (14%), and the side product from the previous reaction, N,N-
dioctyl-N-methylammonium bromide (8%). This mixture required
no further purification, as the thiol (and possibly the disulfide)11
preferentially ligated to the nanoparticle surface and the side
products were removed during nanoparticle purification.
We examined 3 and 5 by TEM (Figure 1), PXRD (Figure 2),
DSC, NMR, and TGA (Supporting Information). TEM and PXRD
revealed that the gold and platinum particles had a high degree of
crystallinity, while the palladium and rhodium nanoparticles were
predominantly amorphous. Figure 1C,D demonstrates the ability
to tune nanoparticle size by varying synthesis conditions, e.g., the
reaction temperature or the injection rate of superhydride (Sup-
porting Information). NMR revealed successful removal of all
unwanted organic compounds, as the only organic material present
1
is 2 bound to the nanoparticle surface. In both H and 13C NMR,
the peak width increased upon nanoparticle ligation, and the
methylene resonances closest to the thiol disappeared completely,
as expected.13
To generate nanoparticles with liquid-like properties, sulfonate
4 was added to 3 to achieve a 1:1 bromide:sulfonate molar ratio
(Supporting Information). After the solution was stirred in chlo-
roform for several hours, water was added and stirring was
continued for several more hours. Upon removal of the water by
pipet, the chloroform was distilled off under vacuum to afford 5.
Vacuum reached <0.05 mbar, and analysis of 5 by NMR and TGA
showed no residual solvent. These nanoparticles exhibited liquid-
like flow at room temperature, as shown in Figures 3 and 4. The
nanoparticle liquid was viscous, and the 50 mg drop moved 1.3
cm along an inclined glass plane over 30 min. DSC of the 2.0 nm
platinum nanoparticles 5 (Figure 3) revealed that the ligands
crystallize at -11 °C and melt at 16 °C (using a scan rate of 10
°C/min). A comparison between 4 and 5 revealed that the enthalpy
To synthesize 2.0 nm platinum nanocrystals,12 0.300 g of H2-
PtCl6‚6H2O, 0.227 g of 2, and 18 g of anhydrous THF were stirred
for 30 min at 20 °C. We injected 10 mL of 1.0 M superhydride in
THF at a constant rate over 10 s. After the solution was stirred for
30 min, a small amount of 2-propanol was added to quench any
excess superhydride. Rotary evaporation to remove THF, followed
† Department of Chemistry & Chemical Biology.
‡ Department of Materials Science & Engineering.
9
12074
J. AM. CHEM. SOC. 2006, 128, 12074-12075
10.1021/ja064469r CCC: $33.50 © 2006 American Chemical Society