A better understanding of the risk of water contamination from viruses, bacteria or contaminants adsorbed on colloids can be achieved by modeling the retention of colloidal particles in soils. Colloids transported by fluid remain in the soil when they arrive at constrictions in pore space too small to admit them. Pore throats have been typically considered the place where colloids are trapped. However, numerous column experiments offer unequivocal evidence of straining of colloids that are smaller than the smallest pore throat in the column.

We propose to explain this observation in terms of a geometric mechanism, independent of physicochemical factors. Our hypothesis is that colloids can be trapped not just in throats between three grains, but also in gaps between pairs of grains. The occurrence of appropriately sized gaps in model soils, represented by computer generated dense random packs of mono-dispersed spheres, is large enough to trap a considerable number of colloids. A geometric analysis of throats and gaps in the model soils was combined with a new methodology to compute flow rates in gaps. The results were input to an existent straining theory to study the dependence of straining rate on colloid size. The analysis suggests that straining in a gap cannot be treated correctly without reference to the throat associated with the gap. A purely geometric argument gives a reasonably accurate prediction of the scaling of straining rate with particle size. Refining this argument to account for flow distribution within a throat and associated gaps improves the prediction.