Influence of Fluid Velocity and Cell Concentration on the Transport of Motile and Nonmotile Bacteria in Porous Media TERRI A. CAMESANO AND BRUCE E. LOGAN* Department of Civil and Environmental Engineering, The Pennsylvania State University, 212 Sackett Building, University Park, Pennsylvania 16802 The effect of fluid velocity on the transport of motile and nonmotile bacteria was studied in saturated soil columns using radiolabeled cells. According to colloid filtration theory, decreasing the bulk fluid velocity in a porous medium increases the number of collisions of passive colloids with particles and, therefore, should result in increased colloid retention in porous media. However, for motile cells, there was a variation in cell retention significantly different from that predicted by filtration theory at low fluid velocities, leading to the conclusion that filtration theory is not applicable for this motile bacterial strain at low fluid velocities. As the pore velocity was decreased from 120 to 0.56 m/day, the fractional retention of motile cells (Pseudomonas florescens P17) decreased by 65%, and the collision efficiency (R) defined as the ratio of particles that attach to soil grains to particles that collide with the soil (calculated using a filtration equation) decreased from 0.37 (120 m/day) to 0.003 (0.56 m/day). For passive colloids, the fractional retention (if R is a constant equal to 0.01) would increase by more than 800% over this same velocity range. To support our conclusion that cell motility was the factor producing this change from filtration theory, we rendered P17 cells nonmotile and tested this strain and a second nonmotile strain [ Burkholderia (Pseudomonas) cepacia G4] under the same conditions. Collision efficiencies for both nonmotile suspensions were constant. For nonmotile P17, R was equal to 0.018 ( 0.003 (0.56-590 m/day). Over a wide velocity range for nonmotile G4, R was equal to 0.22 ( 0.067 (11-560m/day). Swimming cells were presumably able to avoid sticking to soil grains at low fluid velocities, but at high fluid velocities, cell motility did not reduce attachment. Two additional factors known to affect cell transport (solution ionic strength and cell concentration) were also examined with these two strains in porous media. Decreasing the ionic strength from 4.14 to 0.0011 mM (at a constant pH) decreased cell retention for motile P17 by 39 ( 12%, but this is less of a reduction than is typically observed for nonmotile strains. Increasing the cell concentrations of motile P17 increased the overall retention of cells, suggesting that previously deposited cells provided a more favorable surface for adhesion than the native soil (ripening). In contrast, increasing the cell concentrations of G4 resulted in lower retention, suggesting that deposited cells provided a less favorable collector surface (blocking). These results need to be further investigated with other motile and nonmotile species. However, our results do suggest that wider dispersal of cells during bioaugmentation than previously thought possible may be achieved by using a combination of motile cells, low pumping velocities, and low ionic strength solutions. Optimal cell concentrations to use for in situ bioaug- mentation of contaminated soil will depend on the adhesion of the bacterial strains for soil grains and with each other, but in general blocking-type cells are capable of greater dispersal at higher concentration than ripening-type cells. Introduction A variety of microorganisms have been isolated that can degrade common chemical pollutants such as BTEX com- pounds and chlorinated aliphatics (1, 2). Some of these microorganisms show promise for subsurface bioremedia- tion. However, to widely introduce these laboratory-grown microbesat high concentrationsinto contaminated aquifers, it may be necessary to facilitate their transport over larger distances (3). Large losses of cells due to attachment to soil grains have been observed in bioaugmentation tests (4, 5). Most laboratory strains investigated have high collision efficiencies (R values) at ionic strengths comparable to groundwater (0.1 <R< 1; 6, 7),where R is defined in filtration models as the fraction of colliding bacteria that attaches to the soil. Orderofmagnitude reductionsin cellconcentrations in a few centimeters can occur if R values are high (8). Previous experiments on the transport of inorganic colloids, viruses, and bacteria in porous media suggest that bacterial transport in soil can be enhanced by altering physicalfactors such as water velocityand cellconcentration (9-12). Column testson nonmotile speciessupport the trend of decreased colloid retention with increased velocities predicted by filtration models (13-15). However, previous column studies were designed to examine attachment at velocities produced by natural aquifer gradients, typically 0.1-10 m/day. During pump-and-treat soil remediation, high pumping rates may increase flow velocities by orders ofmagnitude (5), resulting in a flow range that has not been tested using filtration theory in subsurface applications, although high flow rates are common for water treatment filters (16). Clean-bed filtration theory predicts no effect of particle concentration on colloid retention (17), although high colloid concentrations are known to affect colloid retention in porous media (11, 18). Due to the large range of experimental conditions reported in the literature, it has not been satisfactorily explained whether the discrepancies between theory and experiments for the effects of colloid concentration and velocity on transport are due to violation ofthe clean-bed assumption or ifin fact there are phenomena, such as blocking, ripening, or the orientation of cells and collectors during deposition with respect to velocity, that are not included in the theory. Rijnaarts et al.(11),usingcell concentrationsof10 7 or 10 8 mL -1 and high surface coverages (g20%),proposed that some bacteria exhibit blockingwhere attached cellsprevent further cellsfrom attachingbyblocking a portion ofthe collector surface (17). Ifcolloidalinteractions are favorable, increasing cell concentrations results in multilayer coverage (18), defined here as ripening. The effect offluid velocityon colloid transport is correctly predicted for passive particles, but it is not known whether *Correspondingauthor phone: (814)863-7908;fax: (814)863-7304; e-mail: blogan@psu.edu. Environ. Sci. Technol. 1998, 32, 1699-1708 S0013-936X(97)00996-6 CCC: $15.00  1998 American Chemical Society VOL. 32, NO. 11, 1998 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1699 Published on Web 04/29/1998