Centrifugation: concentrating colloidal silica

The 3M Company manufactures a continuous polycrystalline alumina—silica—boria fiber (Nextel) by a sol process (17). Aluminum acetate is dissolved in water and mixed with an aqueous dispersion of colloidal silica and dimethylformamide. This mixture is concentrated in a Rotavapor flask and centrifuged. The viscous mixture is then extruded through spinnerettes at 100 kPa (1 atm) the filaments are collected on a conveyor and heat-treated at 870°C to convert them to metallic oxides. Further heating at 1000°C produces the 10-pm diameter aluminum borosilicate fibers, which are suitable for... 

For the coating reactions with APS (Table I), a procedure similar to that described by Philipse and Vrij (4) was applied. Philipse and Vrij used the coupling agent 3-methacryloxypropyltrimethoxysilane to coat colloidal silica. Because an alcosol was used, the actual concentrations of water as listed in Table I have to be corrected for the (not exactly known) amount consumed in the hydrolysis-condensation reactions. After addition of the APS to 400 mL of alcosol, the solution was stirred slowly for an hour. After 2 h of refluxing, 250 mL was distilled slowly for an hour. The remaining unreacted APS was removed in four centrifugation redispersion steps. The ultracentrifuge used was from Beckmann (L5-50B), and the sediment was redispersed each time in absolute ethanol. 

Laurent, et al. examine ten biomacromolecules and colloidal silica in 1.7 MDa hyaluronic acid(5). For all probe particles, s had a stretched-exponential dependence on matrix c, with v = 0.5. 5 was independent of the centrifugal acceleration, i.e., experiments were in a linear domain in which solution viscoelastic properties were not evident. Contrast is to be made with the electrophoretic mobility, as discussed in the next chapter, in which one can enter a nonlinear domain in which particle mobility depends on the applied field E. The constant B was found, to good accuracy, to be linear in the particle radius, except for the very largest spheres. Laurent, et al. also measured the diffusion coefficients D of four smaller probes in the same polymer solutions. Within experimental error, s/D was independent of matrix concentration. 

The amounts oi adsorption of the polymer on latex and silica particles were measured as follows. Three milliliters of the polymer solution containing a known concentration was introduced into an adsorption tube(lO ml volume) which contained 2 ml of latex (C = l+.O wt %) and silica(C = 2.0 wt %) suspensions. After being rotated(l0 rpm) end-over-end for 1 hr in a water bath at a constant temperature, the colloid particles were separated from the solution by centrifugation(25000 G, 30 min.) under a controlled temperature. The polymer concentration that remained in the supernatant was measured colorimetrically, using sulfuric acid and phenol for the cellulose derivatives(12), and potassium iodide, iodine and boric acid for PVA(13). From these measurements, the number of milligrams of adsorbed polymer per square meter of the adsorbent surface was calculated using a calibration curve. 

PS colloids at all acrylic acid concentrations tested. However, at low acrylic acid concentrations the particles detached easily from the surface. Only at concentrations above % wt/wt the silica particles were sufficiently strongly bonded to the surface that they remained attached even after several centrifugation steps. The size of the silica particles did not depend on the size of the PS colloids but is determined by the amount of tetraethoxysilane. 

---