We describe a six-step method for making colloidal clusters of 2, 3, or 4 silica particles with a radius of 1.2 microm. This method, originally described by Manoharan et al. (Manoharan, V. N.; Elsesser, M. T.; Pine, D. J. Science 2003, 301, 483), is based on the encapsulation of silica spheres in emulsion droplets. The originality of our work lies in the preparation of monodisperse emulsions, which allows us to obtain some high yields of small aggregates over a wide range of conditions. Using optical microscopy and disk centrifugation, we show that the relative fractions of 2, 3, and 4 particle aggregates are controlled by the emulsification conditions, particularly the concentration of silica in the dispersed phase. Our best yields are obtained using low to moderate shear rates, a highly viscous continuous phase, and intermediate amounts of silica. The sedimentation of the colloidal solution into a gradient of concentration leads to aqueous suspensions of identical clusters. Since the overall process can easily be scaled up, large quantities of identical clusters may be prepared, which should allow the thermodynamic properties of these new colloidal objects to be measured for the first time. These nonspherical particles could serve as building blocks for more complex assemblies, such as colloidal crystals which could find applications as photonic materials.

Maghemite nanoparticles (γ-Fe 2 O 3 NP), and maghemite/silica nanocomposite microspheres (γ-Fe 2 O 3 /SiO 2 MS), have been evaluated as magnetic heterogeneous Fenton catalysts. The catalysts were fully characterized by electronic microscopies, magnetometry, XRD, UV-Vis-NIR spectroscopy, and sorption volumetry. It was found that the two materials differ in size, morphology, porosity and microstructure, although the maghemite nanoparticles are not modified by their encapsulation into the silica. Both catalysts have a strong magnetic susceptibility, but only the MS catalyst can be easily recovered by magnetic settlement. The mineralization and decolorization of aqueous solutions containing a model pollutant in presence of the catalysts were comparatively studied. Three model pollutants differing in their structure and their electrostatic charge were tested. The obtained reaction rates depend on the nature of the pollutant and catalyst. The results indicate the existence of a correlation between the amount of adsorbed pollutant and the decolorization rate. The free NP are usually more active than the MS catalyst, but larger velocity can also be obtained with the MS catalyst when the pollutant is strongly adsorbed on this material. Moderate mineralization rates were observed for both catalysts illustrating the larger stability toward oxidation of the uncoloured organic intermediates resulting from the primary degradation of the model pollutants. Moreover the efficiency and stability of the MS catalyst were established since this material showed an activity for a pollutant during five consecutive tests. This was also confirmed by characterization of the catalyst after these tests.

No abstract