In unsaturated porous media, sorption of colloids at the air–water (AW) interface is accepted as a mechanism for controlling colloid retention and mobilization. However, limited actual pore‐scale observations of colloid attachment to the AW interface have been made. To further investigate these processes, a real‐time pore‐scale visualization method was developed. The method builds on the light transmission technique for fingered flow studies in packed‐sand infiltration chambers and combines it with high‐resolution, electro‐optical hardware and public domain imaging software. Infiltration and drainage of suspensions of hydrophilic negatively charged carboxylated latex microspheres provides compelling visual evidence that colloid retention in sandy porous media occurs via trapping in the thin film of water where the AW interface and the solid interface meet, the air–water–solid (AWS) interface. With this modified theory of trapped colloids at the AWS interface, we are able to explain the apparent discrepancy between previous experimental evidence of hydrophilic colloids seemingly partitioning to the AW interface and more recent findings that suggest this type of colloid does not adsorb at the AW interface.

tion and deposition has improved in the past 10 yr, scientific reviews emphasize the need for more research onThe transport, retention, and release of hydrophobic and hydrothe mechanisms controlling transport in the unsaturated philic polystyrene latex microsphere colloids were examined in 0.5cm-thick, 26-cm-long slab chambers filled with either regular (hydro-

Colloids have long been known to facilitate the transport of contaminants in soils, but few direct observations have been made of transport and retention in unsaturated porous media. Studies have typically been limited to evaluation of column breakthrough curves, resulting in differing and sometimes conflicting proposed retention mechanisms. We carried out pore scale visualization studies of colloid transport in unsaturated quartz sand to directly observe and characterize colloid retention phenomena. Synthetic hydrophilic (0.8, 2.6, and 4.8 microm carboxylated polystyrene latex) and relatively hydrophobic (5.2 microm polystyrene latex) colloidal microspheres were added to steady-state water flow (0.15 mm min(-1)) applied to an inclined infiltration chamber. Bright field microscopy was used to determine the positions and movement of water and colloids. Confocal laser scanning microscopy was used to determine water film geometry in an unsaturated horizontal chamber. We determined mechanisms of hydrophilic colloid retention at what is generally termed the air/water/solid (AWS) interface. Based on our observations, the AWS interface is here more accuratelytermed the air/water meniscus/solid (AWmS) interface, denoting the region where between-grain water meniscii diminish to thin water films on the grain surfaces. Colloids were retained at the AWmS interface where the film thickness approximately equaled colloid diameters. The greater retention for hydrophilic colloids at this interface (compared to elsewhere in the solid/water interface) can be explained by the additional surface tension capillary potentials exerted on colloids at the AWmS interface. While some 0.8-microm colloids were observed in thin water films, film straining played no significant role in the retention of larger colloids. Mechanisms for slightly hydrophobic colloids differed slightly. In addition to primary retention at the AWmS interface, hydrophobic colloids attached to others already present atthat interface resulting in apparent retention at the air/water (AW) interface. Attachment of hydrophobic colloids was also observed at water-solid interfaces, as hydrophobicity impelled the colloids to avoid water. Factors contributing to retention of slightly hydrophobic colloids were sand grain roughness and possibly a tendency for these colloids to flow near surfaces and interfaces, consonant with the enhanced retention of hydrophobic colloids (relative to hydrophilic colloids) observed in the literature.

Wan and Wilson (1994a), colloid transport experiments with porous media have been based on mass balance In unsaturated porous media, sorption of colloids at the air-water of breakthrough colloid concentrations in packed-sand (AW) interface is accepted as a mechanism for controlling colloid retention and mobilization. However, limited actual pore-scale obser-columns (Wan and Wilson, 1994b;Schafer et al., 1998; vations of colloid attachment to the AW interface have been made. To Jewett et al., 1999; Jin et al., 2000; Lenhart and Saiers, further investigate these processes, a real-time pore-scale visualization 2002). On the basis of analyses of outflow concentrations method was developed. The method builds on the light transmission of colloid particles, these authors found that more partitechnique for fingered flow studies in packed-sand infiltration chamcles were retained in the column at lower water contents bers and combines it with high-resolution, electro-optical hardware (or under conditions where there were higher percentand public domain imaging software. Infiltration and drainage of ages of trapped air). Retention, measured as reduced suspensions of hydrophilic negatively charged carboxylated latex micolloid concentrations in the column outflow, was attribcrospheres provides compelling visual evidence that colloid retentionuted to sorption at the AW interface and film straining. in sandy porous media occurs via trapping in the thin film of waterIn one case retention of bacteriophages in a batch syswhere the AW interface and the solid interface meet, the air-watersolid (AWS) interface. With this modified theory of trapped colloids tem was ascribed to the presence of a dynamic AWS at the AWS interface, we are able to explain the apparent discrepancy interface (Thompson et al., 1998; between previous experimental evidence of hydrophilic colloids seem-1999). New data suggest, however, that conceptual modingly partitioning to the AW interface and more recent findings that els of colloid transport in unsaturated media need to be suggest this type of colloid does not adsorb at the AW interface.reexamined. It has been generally assumed, based on visualization studies using etched-glass micromodels (Wan and Wilson, 1994a), that retention at the AW interface sity,