Transport and retention of silver nanoparticles in granular media filtration

Kim, Ijung
Journal Title
Journal ISSN
Volume Title

The increasing use of engineered nanoparticles such as silver nanoparticles (AgNPs) has focused more attention on the transport of nanoparticles in natural and engineered systems. Despite a substantial number of studies on the transport of nanoparticles in groundwater flow conditions, other conditions such as those in granular media filtration in water treatment plant have not been fully explored. This study was designed to investigate the transport of AgNPs in granular media filtration with a relatively high filtration velocity (~2 m/hr) and a low influent AgNP concentration (~100 [mu]g/L). Effects of several physical and chemical parameters on the transport and attachment of AgNPs were examined, focusing on the colloidal filtration theory and particle-particle interaction, respectively. Regarding the transport of AgNPs, four physical parameters (filter depth, filtration velocity, filter media size, and AgNP size) were varied at a fixed chemical condition. Positively charged branched polyethylenimine (BPEI) capped AgNPs were chosen to examine the transport of AgNPs under electrostatically favorable attachment conditions. The effects of filter depth, filtration velocity, and filter media size on transport of AgNPs were adequately described by the well-known colloidal filtration model. However, deviation from the model prediction was apparent as the AgNP size became smaller, implying a possible variation of nanoparticle properties in the smaller size such as 10 nm. In the AgNP attachment study, negatively charged citrate- and polyvinylpyrrolidone (PVP)-capped AgNPs were employed to examine the chemical effects on particle (AgNP)-particle (filter media) interaction. When the ionic strength and ion type in the background water were varied, the attachment of citrate AgNPs followed the DLVO theory. Ca- or Mg-citrate complexation was found to lead to charge neutralization, resulting in a greater AgNP deposition onto the filter media. However, PVP AgNPs were only marginally affected by the electrostatic effect, demonstrating a stronger stabilizing effect by PVP than citrate. When natural organic matter (NOM) was introduced in the background water, the deviation from the DLVO theory was considered primarily due to the steric interaction by NOM coating onto particles. Different amounts of AgNP deposition for different types of NOM suggest the variation of steric effects according to the molecular weight of NOM. The deposition of humic acid-coated AgNPs was similar regardless of the capping agent, indicating the possible displacement of the capping agent by NOM. The electrostatic and steric interactions affected the detachment of AgNPs as well as the attachment of AgNPs. The amount of detachment depended on the depth and width of the secondary energy minimum. Also, the detachment was enhanced with NOM coating, probably due to a weak attachment by the steric effect. However, the hydrodynamic force employed in this study was insufficient to yield a remarkable detachment. Overall, the retention profile was a relatively vertical line (i.e., equal deposition with depth) when the AgNP aggregation was prevented by the electrostatic or steric repulsion, implying homogeneous AgNP capture throughout the filter bed. On the other hand, ripening (the capture of particles by attraction to previously retained particles) was favored at the top of the filter bed when the AgNP aggregation was allowable.