Laboratory scale investigations into the influence of particle diameter on colloid transport in highly weathered and fractured shale saprolite

A series of colloid transport experiments were performed in an undisturbed column of fractured shale saprolite using fluorescent latex microspheres of uniform shape and surface coating as analogs for colloidal contaminants or microorganisms. The objectives of this study were to determine if there was an optimum size range for transport of latex microspheres in a highly weathered and fractured shale, and to investigate possible causes of colloid retention and size preferential transport. The column, excavated from a depth of 1.5 meters in highly weathered and fractured shales of the Conasauga group in eastern Tennessee, had a bulk hydraulic conductivity value of 2 x 10^-5 m/s and a porosity of 42 percent. The experiments were performed at a constant flow rate of 3.9 ml/min, with a hydraulic gradient value of 0.08, which is within the range of gradients typically observed in these deposits under field conditions. For the first tracer experiment, four different sizes of latex microspheres (.05, 0.1, 0.5, and 1.0 μm) were added to a solution of 0.006 M CaBr₂, and two pore volumes (10.7 liters) of the solution were pumped through the column. A second tracer experiment was performed using 0.5, 1.7 and 4.25 μm diameter microspheres in a .005 M CaBr₂ solution. Microsphere breakthrough curves for both experiments exhibited a common pattern of rapid initial breakthrough, followed by a concentration "plateau". which was up to four orders of magnitude below the injection concentration. These experiments clearly show that the 0.5 and 1.0 μm microspheres were near the optimum size for transport. Microsphere losses were greater for the smaller sizes (.05 and 0.1 μm diameter) and for the larger sizes (1.7 and 4.25 μm). After each injection the column was flushed with at least four pore volumes of 0.005 M CaCl₂, producing an abrupt decrease in colloid concentration, followed by a long “tail” of slowly declining tracer concentrations. The long "tail" indicates that the processes causing microsphere losses were at least partly reversible. Upon completion of the tracer experiments, the saprolite column was dissected under UV light to examine the distribution of microspheres along the fractures and in the adjoining matrix. Comparisons of ambient and UV light observations and analysis of microcore soil samples taken from the column during dissection indicate that microspheres traveled along some, but not all of the fractures, and that they often tended to travel through flow channels along fracture surfaces. Size segregation of microspheres along flow channels was observed and the smaller microspheres were found in the fine grained matrix at distances of up to 4 mm from the fractures. This was likely due to diffusion of the smaller microspheres into the fine-grained matrix.. Quantitative evaluation of microsphere loss mechanisms indicate that gravitational settling could be an important factor in losses of the 4.25 μm diameter microspheres. Calculations of Brownian diffusion rates within the fractures show that this process could cause significant numbers of the smaller (.05 to 0.5 μm) microspheres to collide with fracture walls where electrostatic attachment is likely. Calculated "hydraulic" fracture apertures were large enough for transport of all microsphere sizes. however, the cubic law ignores irregularities in fracture surfaces and variations in fracture apertures which may have caused straining of larger microspheres.