Effect of dilution and contaminants on strength and hydraulic conductivity of sand grouted with colloidal silica gel

Colloidal silica (CS) is a low-viscosity liquid that can be made to gel by addition of brine. This property allows it to be injected into, or mixed with, soil, so that after gelling the colloidal silica blocks the pore space in the soil and forms a barrier to the flow of contaminated groundwater or non-aqueous liquids (NAPLs). Gelled-in-place CS was first studied for the petroleum industry and later for protecting groundwater quality. Noll investigated the use of colloidal silica diluted so that its solids content was reduced from 30% (a typical nominal value for material as delivered) to values as low as 5%. The more dilute colloids could still be made to gel, although more slowly, and the resulting gel was weaker. Because the proposed application of colloidal silica grout involves emplacing it in the subsurface by permeation, jet grouting, or soil mixing where its role as a barrier will be to resist flow of contaminants, the effects of these contaminants on the properties of the grouted soil is also of interest. This work comprised four tasks. In Task 1, samples of grouted sand were prepared with a range of CS dilutions, for measurement of hydraulic conductivity and unconfined-compressive strength. In Task 2, these properties were measured on samples of grouted sand that incorporated 5% volumetric saturation of NAPLs. In Task 3, samples, prepared without any contaminants, were immersed in contaminant liquids and tested after 30 and 90 days. Task 4 was added because NAPL contamination in the samples of Tasks 2 and 3 impelled modifications in the test methods, and comparison of the results of Task 2 and Task 1 suggested that these modifications had introduced errors. In Task 4, samples were tested both ways, to confirm that in Tasks 2 and 3 strength was underestimated and hydraulic conductivity was overestimated. Despite the existence of these known systematic errors, the inclusion of control samples in Tasks 2 and 3 permits conclusions to be drawn from these data.