Effect of Drying on Viscoelasticity and Permeability of Gel
- PDF / 361,423 Bytes
- 7 Pages / 414.72 x 648 pts Page_size
- 74 Downloads / 240 Views
EXPERIMENTAL PROCEDURE The silica gels used in this study are of the type known as B2 [12], made from tetraethox-
ysilane in a two-step process: initial hydrolysis under acidic conditions with a substoichiometric amount of water, followed gelation under basic conditions with a final water/alkoxide ratio of -3.7. Details of the preparation and properties of this gel are available in refs. [6,8,12]. The sol was cast into polystyrene pipette tubes (with inner diameter 7.77 mm) and left to gel at room temperature overnight; then the rods of gel were removed from the tubes and given two 24-hour washings in a large excess of pure ethanol to remove partially hydrolyzed alkoxide [6]. The rods were then given successive 24-hour washings in ethanol/water solutions with volume ratios of 50/50, 25/75, and 5/95. The samples were stored in the 5/95 solution, rather than pure water, to avoid microbial growth during long aging. Based on the amount of alkoxide, and assuming a skeletal density for the gel of 2.0 g/cm3 , the solids content (or relative density, p) is estimated to be 0.065 at the time of gelation [8]. For controlled drying, rods of gel were transferred onto a plastic mesh inside a glass tube with loose-fitting endcaps, and held at room temperature. Partially dried samples were transferred into a bath of pure water for 3-point bending measurements. The apparatus and procedure are described in ref. [13]. Briefly, the gel rod is supported at the ends on teflon®rollers, and a pushrod is dropped onto the gel to produce a sudden predetermined deflection; the load is measured as a function of time at constant deflection (typically -1% strain). Each run lasted 20 h, during which time the baseline drift could be as much as 0.5 g; the initial load was chosen to be > 20 g, of which >15 g relaxed away, so the baseline drift amounts to < 3% of the measured 209 Mat. Res. Soc. Symp. Proc. Vol. 346. 01994 Materials Research Society
change in load. In these experiments, the span between end supports was set at 5.4 cm, giving a span/diameter ratio of 7-10. The sample diameter was measured with an optical microscope at the end of each run. Water (neutral pH) was chosen as the medium for the experiments, because it provides fast viscoelastic relaxation for silica gels [13]. Early tests were performed in ethanol, but it was found that the partially dried samples expanded when immersed in the bath of the 3point bending apparatus, returning almost to their initial dimensions. In alcohol, the gels show no viscoelasticity [14], so the strain caused by capillary forces is largely reversible. There is also a slight recovery in water as the elastic part of the strain recovers, but most of the strain (both during drying and bending) is irreversible. The load on the gel rod relaxes by two independent mechanisms: hydrodynamic relaxation caused by flow of liquid within the gel network, and viscoelastic relaxation attributed to hydrolysis of siloxane bonds by the pore liquid [13]. Although the rigorous result is quite complex [15], the kinetics of
Data Loading...
