True
thermodynamic stability of a solid colloidal dispersion is
generally unexpected, so much so that a thorough experimental validation
of proposed stable systems remains incomplete. Such dispersions are
underinvestigated and would be of interest because of their long-term
stability and insensitivity to the preparation pathway. We apply classical
nucleation theory (CNT) to such colloidal systems, providing a relationship
which links the size-dependent interfacial free energy density of
the particles to their size distribution, and use this expression
in the fitting of previously reported size distributions for putatively
thermodynamically stable nanoparticles. Experimental data from a gold–thiol
system exhibiting inverse coarsening or “digestive ripening”
can be well described in terms of a power-law dependence of the interfacial
free energy γ on radius r based on capacitive
charging of the nanoparticles, going as r
–3, as suggested by prior authors. Data from magnetite nanoparticles
in highly basic solutions also can be well fit using the CNT relation
but with γ going as r
–2.
Slightly better fits are possible if the power of the radius is nonintegral,
but we stress that more complex models of γ will require richer
data sets to avoid the problem of overfitting. Some parameters of
the fits are still robustly at odds with earlier models that implicitly
assumed absolute thermodynamic stability: first,
the extrapolated free-energy density of the flat surface in these
systems is small and positive rather than strongly negative; second,
the shape of the distributions indicates the solution phase to be
supersaturated in monomer relative to the bulk and thus that these
two systems may only be metastable at best. For future work, we derive
expressions for the important statistical thermodynamic and chemical
parameters of the interface energy in terms of (1) the surfactant
concentration, (2) the temperature dependence of the distribution,
and (3) the concentrations of particles in the tails of the distribution.