Chemistry Reference
In-Depth Information
volume. If these separation and purification techniques do not result in a chemically
pure sample, then the resultant sample remains a mixture of cluster molecules.
If these mixtures have many components, this leads to a Gaussian distribution of
cluster sizes and compositions. The dispersity of the sample indicates the breadth of
the distribution function. The dispersity depends primarily on the variation in the
number of metal atoms which occur in the nanoparticles, but may also be influenced
by the number of ligands attached to the metal surface. These polydispersed
samples are no longer directly analogous to pure molecular clusters and begin to
resemble classical colloidal species. Their dispersities are estimated on the basis of
a statistical analysis of the sample dimensions using electron microscopy. In this
review these alternative possibilities are defined as follows:
nanocluster
for ana-
lytically pure and monodispersed clusters and
nanoparticles
or
nanocolloids
for
polydispersed samples containing several closely related cluster species.
As the average size of the particles increases and the stabilising ligands become
more labile, thenmore facile equilibria involving the exchange of ligands and solvent
on the surface of the metal particle become more important. Faraday's colloids
belong to this category since they have wide distribution of dispersities and in
solution complex equilibria involving the metal atoms, solvent and ligands. The
complexity of colloids is represented pictorially at the bottom of Fig.
1
. An important
“holy grail” of nanochemistry is the preparation of atomically precise nanoclusters
which are as well defined as the giant molecules found in molecular clusters. As Jin
has noted, “the atomic precision” is a stricter and more accurate criterion than the
conventional term “monodispersity” used for regular nanoparticles, and correspond-
ingly, mass spectrometry is a more definitive characterisation technique than TEM
and is indeed indispensible in nanocluster characterisation [
43
].
Recent X-ray crystallographic studies [
44
] have established that the long held
view that clusters and nanoparticles of gold contain a close-packed central core of
gold atoms, resembling that found in the bulk metal, surrounded by soft ligands
bonded to the surface gold atoms is perhaps an oversimplification. In particular
these studies have demonstrated that surface gold atoms may combine with
organothiolato ligands (RS) to form mono-anionic metallo-thiolato ligands [(RS)
Au(SR)]
and [(RS)
3
Au
2
]
. These ligands are able to co-ordinate to core metal
atoms using conventional RS-Au dative bonds, which may be supplemented by
aurophilic gold-gold interactions [
44
,
45
]. These metallo-thiolato ligands are
represented schematically in Fig.
2
. The cyclic [Au
4
(SR)
4
] units shown in Fig.
2
are commonly observed in the fragmentation patterns observed in the mass spectra
of gold thiolato-colloids [
46
,
47
]. These observations suggest that those clusters and
nanoparticles are actually mixed-valency compounds and indeed their characteris-
tic colours may arise from mixed valence electronic transitions from the Fermi level
of the core atoms to the empty valence orbitals on the gold(I) surface atoms.
Although the section above has tried to define more clearly the distinctions
between clusters, colloids and nanoparticles, there are also grey areas in-between
where the differences remain blurred. The original syntheses of gold cluster
compounds stabilised by phosphine ligands originated in the late 1960s from
Malatesta's group in Milan [
48
-
51
] and were extended primarily by Mingos
