Little Barriers, Big Effects: Repulsion Yields Profound Impacts

The forces and torques in particle trajectory simulations underlying classic CFT prevent colloid attachment in porous media under unfavorable conditions, where colloid and collector surfaces are like-charged. This prediction runs in direct contradiction to experimental observations demonstrating attachment under unfavorable conditions, albeit reduced relative to favorable conditions No functional, easily applied theory previously existed to quantitatively predict colloid retention in porous media in the presence of energy barriers (unfavorable colloid attachment conditions), which is thought to be the prevalent condition in environmental systems.

As described before, the mean-field approach (use of a single zeta potential to represent each surface) predicts zero colloid attachment in the presence of significant repulsion. However, above a threshold IS, net repulsion becomes sufficiently reduced by compression of the electric double-layer interaction closer to the surface, and attachment is predicted. So, mean-field DLVO parameters can yield attachment, but they predict an extremely stiff dependence of simulated attachment on IS, one that does not match the gradual experimentally observed trends with IS.

In sediment-packed column transport experiments, a profound change occurs between favorable and unfavorable conditions as indicated not only by lesser retention, but by the extended tailing of low colloid concentrations during elution, which is absent or present when bulk repulsion is absent or present, respectively. The resulting retention profiles from experiments using unfavorable conditions (Figure 5) have shown non-log-linear (e.g. nonmonotonic or hyperexponential) decreases in concentration as function of distance.

Figure 5. Experimental data (symbols) for the breakthrough-elution behavior (a) and retained profiles (b) of polystyrene latex microspheres in quartz sand in unfavorable and favorable conditions described in Li and Johnson. Lines are continuum model descriptions with a single deposition rate coefficient. The single deposition rate coefficient was used to generate a probabilistic deposition distribution. Adapted from Li and Johnson (2005)

The hyperexponential profiles (initially observed for bacteria and protozoa) were attributed to heterogeneity among the bacterial population such that “stickier" individuals were retained upgradient of “less sticky" individuals. Distributions in colloid size, surface charge, coatings, and hydrophobicity have been inferred to yield heterogeneity among colloid populations and hyperexponential retention profiles as illustrated in Figure 5.

Other studies have attributed the inferred “fast" and “slow" dual-deposition behavior to differences in interaction energies between colloids and collectors emanating partially from localized soil heterogeneities as well as deep secondary energy minima, the latter yielding “fast" retention, and the former yielding “slow" retention on the basis that the former mechanism requires overcoming the repulsive energy barrier prior to retention. Additional studies have also suggested that soil heterogeneities, such as heterogeneities of attractive iron oxyhydroxide situated within bulk repulsive silica, can create “fast" and “slow" deposition rates in the favorable and unfavorable regions.

While the above studies implicate soil heterogeneity in generating fast versus slow retention rates, they do not articulate how this would generate preferential upgradient versus downgradient retention (hyperexponential profiles). Assuming that inferred heterogeneity is distributed throughout the column (not predominantly located near the column inlet), then the kinetic retention coefficient across the column would be uniformly increased or decreased across the column by the presence of heterogeneity. Based on similar reasoning, a number of studies have concluded that soil heterogeneity alone is likely not the primary cause of this behavior.