You might use the same procedure as the one we used for nanocarbons :
The resin (EPIKOTE Resin 828, a medium viscosity liquid epoxy resin produced from bisphenol A resin and epichlorhydrin, containing no diluent) was degassed under vacuum (1–3 mbar) for 12–14 h, then it was put into an oven at 65°C. In the meantime, the carbon nanoparticles were dispersed in propanol through ultrasonic bath for 1.5 h. Afterwards the solution of alcohol and carbon fillers was mixed with the resin. The obtained mixture was inserted inside the oven at 130–150°C for the evaporation of alcohol. The curing agent (A1, a modified Tetraethylenepentamine (TEPA)) was added to the mixture of resin and filler through slow manual mixing for about 7min. The mix was then poured into 1 cm x 1 cm x 7 cm molds, and left as such for 20 h for the curing process at room temperature and then 4 h in oven at 80°C. When the process was completed, the samples were removed from the molds. The uniformity of distribution of the carbon fillers in the polymer samples was studied by scanning electron microscopy; all fabricated composites demonstrated good homogeneity, carbon inclusions were reasonably well dispersed.
It should work for nanosilica.
Alain
You know, the dispersion and distribution of the nanoparticle should be improved to reach the uniform and used maximum capacity of nanoparticles. If the initial viscosity [before curing] of your resin is not high you can used high shear mixer [Kourki, Powder Technology, 2012] to achieve a good composite. Modification of particles with non-polar groups improve your quality of mixing but reduced you product mechanical and rheological properties.
Recommendations of Jun MA is absolutely reasonable. Moreover if a viscosity of selected epoxy resin is quite low and quantities of nanosilica are low (no substantial effect on viscosity of the resin) lets try to use ultrasound treatment. It gives good results.
Nano-silica dispersed in low boiling solvent could be mixed with the resin in a planetary mixer and the slowly removing the solvent under vacuum will help in better mixing of the silica in the matrix. The viscosity increase depends on the loading of silica in the matrix.