BACTERIAL TRANSPORT
IN
NAPL-CONTAMINATED POROUS MEDIA
By Brock Rogers1 and Bruce E. Logan,2 Member, ASCE ABSTRACT: To understand the effect of a non-aqueous-phase liquid (NAPL) on bacterial transport, column experiments in the presence and absence of a NAPL were conducted using two bacterial strains and two different porous media. The presence of a NAPL (tetrachloroethene) decreased the retention of Pseudomonas fluorescens P17 in quartz and soil columns (factor of 2 for quartz and 1.6 for the soil). In contrast, there was little change in the overall transport of P. putida KT2442 in soil columns in the presence of a NAPL. To understand how a NAPL phase might affect bacterial transport, several different mechanisms of particle removal by a NAPL were hypothesized, and a filtration model was modified to test each hypothesis. Only one of the five models was consistent with the increased transport of P17 in the presence of the NAPL suggesting that the NAPL produced large, immobile zones of water. The presence of immobile water zones would decrease overall porosity, reduce the number of packing grains available for particle filtration, and increase water velocity, resulting in increased bacterial transport in the presence of a NAPL.
INTRODUCTION Research on the transport of biocolloids through the subsurface originally focused on the travel of pathogenic bacteria and viruses, with the goal of inhibiting biocolloid transport (Gerba and Bitton 1984; Yates 1988). In more recent years, bacterial transport studies have centered on enhancing the transport of bacteria through porous media to enhance bioremediation via bioaugmentation (Johnson et al. 1996; Camesano and Logan 1998; Li and Logan 1999). Bioaugmentation is a process by which contaminant-degrading microbes are injected into the subsurface for remediation purposes. For bioaugmentation to be successful, bacteria must be able to be transported through the soil over reasonably large distances (several meters) to avoid clogging the wellhead (Marlow et al. 1991). The transport of bacteria through porous media is often described by filtration theory (Yao et al. 1971; Rajagopalan and Tien 1976). Particle removal is divided into two steps: collision of the particle with the collector and attachment to the surface (Yao et al. 1971). These two steps are incorporated into filtration models that can account for gravitational settling, interception, Brownian diffusion, and London-van der Waals forces (Logan et al. 1995). The collision efficiency a, defined as the ratio of the number of particles that stick to a collector to the number of particles that strike the collector, is used to quantify the attachment of bacteria to porous media. Bacterial attachment is affected by factors such as solution ionic strength (IS) (Gannon et al. 1991; Shonnard et al. 1994), pH (Scholl and Harvey 1992; McCaulou et al. 1995), hydrophobicity and electropotential (Sharma et al. 1985; van Loosdrecht et al. 1987a,b; Stenstro¨m 1989), and the presence or absence of extracellular polymers (Fletcher and Floodgate 1973; Jucker et al. 1998). The widespread use of chlorinated solvents and their improper disposal has resulted in extensive ground-water contamination. Frequently, these contaminants are in the form of non-aqueous-phase liquid (NAPL), which can persist for long periods of time in aquifers (Hunt et al. 1988). In situ bio1
Res. Assoc., Dept. of Chem. and Envir. Engrg., Univ. of Arizona, Tucson, AZ 85721; currently, URS Greiner Woodward Clyde, 7878 North 16th St., Ste. 200, Phoenix, AZ 85020. 2 Prof., Dept. of Civ. and Envir. Engrg., Pennsylvania State Univ., University Park, PA 16802; Corresponding author E-mail:
[email protected] Note. Associate Editor: Susan E. Powers. Discussion open until December 1, 2000. To extend the closing date one month, a written request must be filed with the ASCE Manager of Journals. The manuscript for this paper was submitted for review and possible publication on July 23, 1999. This paper is part of the Journal of Environmental Engineering, Vol. 126, No. 7, July, 2000. qASCE, ISSN 0733-9372/00/0007-0657– 0666/$8.00 1 $.50 per page. Paper No. 21504.
degradation via bioaugmentation can reduce off-site migration of the chemical. However, the effect of the NAPLs on biocolloid transport has not been previously investigated. The presence of an immiscible liquid in porous media could affect filtration rates by changing biosorption to soil surfaces as a result of chemical sorption, increasing removal by providing a pure chemical phase for partitioning, and increasing pore water porosity and creating immobile water zones. To understand and predict the effect of a residual NAPL on the transport of bacteria through porous media, filtration models based on these factors were developed and compared to experimental results. METHODS The effect of a pure-phase NAPL residual on the transport of bacteria through porous media was experimentally tested using two strains of bacteria and two types of porous media. It is well known that by suspending cells in low ionic strength (LIS) solutions the transport of bacteria can be substantially increased (Gannon et al. 1991; Jewett et al. 1995; Li and Logan 1999). Therefore, the effect of NAPLs on bacterial transport were investigated using both an artificial ground water (AGW) and a LIS water. Bacteria and Culture Conditions Pseudomonas putida KT2442, obtained from Daryl F. Dwyer at the Department of Civil Engineering, University of Minnesota, is the parent strain of a genetically engineered microbe capable of degrading toluene and 4-ethyl benzoate (Nu¨ßlein et al. 1992). KT2442 is an aerobic, rod-shaped, gram-negative, motile bacterium and is resistant to the antibiotic rifampicin. Pseudomonas fluorescens P17 (ATCC 49642) is an aerobic, rod-shaped, gram-negative, motile bacterium previously used in transport experiments (Kinoshita et al. 1993; Jewett et al. 1995). P17 was grown on a morpholino-propane-sulfonic acid/mineral salts media amended with glucose (250 mg/L) as described in Jewett et al. (1995). Cells were harvested during late log phase (A600 = 0 .1), or at ;108 cells/mL as determined using an acridine orange staining procedure (Hobbie et al. 1977), and diluted to ;107 cells/mL with 200 mL of a test solution in a 500-mL flask containing either an artificial ground water (AGW) (Jewett et al. 1995) or LIS water (MilliQ, Millipore Corp., New Bedford, Mass.; 1025 M). After dilution, 80 mL of 3H-leucine (L-leucine, 1 mCi/mL, Amersham, Arlington Heights, Ill.) was added to the solution, and this suspension was incubated on a shaker table at room temperature for 8 h to allow for radiolabeled uptake by the microbes. Unassimilated radiolabel was removed by syringe filtration JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000 / 657
(Camesano and Logan 1998). Total suspension radioactivity (1 mL) was measured in 20-mL vials (in triplicate), containing 10 mL of scintillation cocktail (CytoScint ES, ICN, ICN Biomedicals Inc., Irvine, Calif.), using a scintillation counter (LS3801, Beckman Instruments, Palo Alto, Calif.). Background (noncell-associated) radioactivity was measured by filtering a portion (15 mL) of the influent solution through a 0.2-mm polycarbonate membrane (Poretics Corp., Livermore, Calif.) and by analyzing 1-mL samples (in triplicate) for radioactivity. Background counts were always low ( 1 (Johnson et al. 1996; Li and Logan 1999). Filtration Models Based on Presence of NAPL Five different models were developed to explain the results of bacterial transport in the presence of a NAPL residual. These models account for the effects that a NAPL residual may have on soil properties and flow conditions through the porous media and include porosity, PV, and the number and size of collectors. The conceptual development of the models and their overall prediction on bacterial transport in the presence of a NAPL residual are summarized below and presented in detail by Rogers (1997). To compare predictions of the different models, fractional removals calculated using the different models were normalized by the PV model using 660 / JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000
(10)
If it is assumed that the NAPL phase decreases the soil porosity in proportion to the occupied pore space by the NAPL, this decrease in porosity will increase the PV (Fig. 2). It is assumed for Model 1 that Nc does not change due to the presence of the NAPL and that total porosity for flow is the water porosity fw. The collector efficiency for Model 1, h1, is calculated using (5) except that the water velocity in the presence of NAPL is based on the water mobile phase. Therefore, u* [ U/fw = U/(ft 2 fn), where fn is the porosity occupied by the immobile NAPL. For Model 1, (3) becomes CzuAfwDt = Cz1DzuAfwDt 1
3 ? a ? h 1(1 2 ft)CuADzDt 2dc
(11)
producing a filter coefficient l1 defined as 3 (1 2 ft) l1 = ? h1 ? 2 dcfw
The PV filtration model therefore consists of (6) and (7), and in the absence of a NAPL, ft = fw. All variables except a in the PV filtration equation can be calculated from the experimental conditions; therefore, a must be calculated from the experimental results. A detailed discussion of the PV model is presented in Logan et al. (1995). Because C/C0 is the relative concentration of the column effluent at steady state, the total fraction of particles retained within the column is Fr = 1 2 (C/C0), and a can be calculated as a=2
1 2 exp(2lmaL) 1 2 exp(2lpvaL)
Model 1
where lpv = PV model filter coefficient, defined as lpv =
(9)
(12)
Overall, Model 1 predicts an increase in the fraction of bacteria retained in the presence of a pure-phase NAPL versus water alone (PV model) because the change in filtration rate due to fw < ft [compare (7) and (12)] is larger than that due to changes due to h1(u* = U/fw) < hpv(u* = U/ft). Model 2 Model 2 is derived in the same manner as for Model 1 except that the number of collectors is assumed to decrease due to the obstruction of flow by the NAPL (Fig. 2). The number of collectors available for collision in the presence of a NAPL residual can be determined by reducing Nc in proportion to the reduction in porosity produced by the NAPL, using the factor b defined as b=
Vpv 2 Vn (ftVt 2 fnVt) ft 2 fn fw = = = Vpv ftVt ft ft
(13)
where Vpv = total pore volume (cm3); Vn = volume of NAPL residual (cm3); and Vt = total volume (cm3). Including b in the calculation of the number of collectors yields Nc =
A(1 2 ft)Dz ft 2 fn A(1 2 ft)Dz fw ? = ? p 3 ft p 3 ft dc dc 6 6
(14)
Notice that if there is no NAPL in the system, fn = 0, and Nc is identical to the PV model calculation. If all the pores are filled with NAPL, fn = ft, and Nc = 0 (i.e., all collectors are blocked by NAPL). A NAPL residual will cause a change in the collector efficiency hpv as a result of the increase in PV. The resulting mass balance yields a filter coefficient for Model 2 of 3 (1 2 ft) l2 = ? h2 ? 2 ft ? dc
(15)
where the collector efficiency in the presence of a NAPL for Model 2 is calculated as above for Model 1 (i.e., h2 = h1). The fraction retained predicted using Model 2 is less than the retention predicted by the PV filtration model because h2(u* = U/fw) < hpv(u* = U/ft).
Model 5 is therefore essentially identical to Model 2, except that all calculations of Model 5 are made with respect to the measured (mobile) water velocity.
Model 3
Effect of NAPL on Removal of P17 in AGW
The same assumptions made above for Model 2 are made with respect to the fluid velocity, except for Model 3 it is assumed that the residual NAPL acts as an individual collector, increasing the total number of collectors available for collision (Fig. 2). Because the NAPL ganglia function as collectors, Model 3 is constructed with separate terms for the collector efficiency of the soil grain collectors and the NAPL residual collectors as
The presence of a pure phase of PCE (14.8% by volume) decreased the fraction of retained cells of P17 in quartz media from 0.56 to 0.33, an overall change in Fr of 41% (Table 1). Bacteria were more highly retained in columns packed with the Arizona soil than with the quartz, with Fr = 0.78 for the Arizona soil. When a pure phase of PCE was placed in the soil column, there was a reduction in P17 retention in columns packed with the Arizona soil, except the decrease was smaller than observed with quartz-packed columns. The fraction retained decreased from 0.78 to 0.61, or by 22%. The PV filtration model was used to estimate the overall and intracolumn sticking efficiencies of P17 for the two media by assuming that the presence of the pure phase did not alter the flow hydrodynamics in the columns. The overall collision efficiency was calculated to be slightly higher for the Arizona soil (a = 0.25) than for the quartz media (a = 0.20) in the absence of the PCE phase (Table 1). In both cases the collision efficiency declined following the addition of the PCE, a result expected based on the corresponding decreases in Fr following PCE addition. The collision efficiency was reduced by 50% to a = 0.10 for the quartz media and by 36% to a = 0.16 for the soil. Intracolumn collision efficiencies were not constant over the length of the column. According to filtration theory, bacterial removal is first-order with respect to cell concentration. When the filtration equation is applied to removals measured along the column, the intracolumn collision efficiencies should be constant over the column length. However, removal rates were apparently much higher near the column inlet than at other distances in the column. As a result, the intracolumn collision efficiencies ai for P17 decreased by two orders of magnitude over the length of the column for both types of porous media (Fig. 3). At the quartz-media column entrance ai = 0.14, and at the effluent end of the column ai = 0.002. The decrease in ai was not due to the NAPL, as these decreases were measured in the absence of the NAPL. Decreases in ai have been previously documented in a number of studies and are not fully understood (Albinger et al. 1994; Martin et al. 1996; Camesano and Logan 1998).
3 1 l3 = ? 2 fw
F
h s3 ?
G
(1 2 ft) fn 1 h n3 ? ds dn
(16)
where h s3 and h n3 = sediment and NAPL collector efficiencies calculated using the sediment and NAPL collector diameters, respectively. The prediction of the magnitude of bacterial retention for Model 3 depends on the size chosen for the NAPL collector. However, regardless of the NAPL size, the fraction retained predicted by Model 3 is always larger than the fraction retained predicted by the PV filtration equation. Therefore, for Model 3 calculations it was assumed that the NAPL and sediment collector diameters were the same. Model 4 For this model it is assumed that the residual NAPL encompasses other sediment collectors, decreasing the total number of collectors Nc by forming fewer and larger NAPL collectors. For example, in Fig. 2, one NAPL collector covers and replaces four sediment collectors. The total number of collectors must be equal to the sum of the number of packing or sediment collectors Nsc and the number of NAPL collectors Nnc, or Nc = Nsc 1 Nnc. If the NAPL covers only a few soil grains, then the number of sediment collectors must be much larger than the number of NAPL collectors, so that Nc ' Nsc. If all the NAPL formed a single collector, it still would not greatly affect the number of collectors. Because it would be difficult to determine how many NAPL collectors there were, we instead assumed that there was not enough NAPL to appreciably change the total number of collectors. Therefore, the filtration equations for Models 2 and 4 are mathematically identical.
RESULTS
Effect of Sorption on Collision Efficiencies
Model 5 In experiments using a conservative tracer (described above) it was observed that in columns containing NAPL the tracer eluted sooner than predicted based on the reduced water porosity produced by the NAPL. Based on the mean breakthrough time of the tracer in columns containing the PCE residual, an effective porosity was calculated as fe, and it replaced the water porosity fw in (11) that was calculated from the volume displaced by the NAPL. It was also assumed that the number of collectors decreased by a ratio be based on replacing fw with fe in (13), or be = fe /ft. Defining the experimentally measured velocity as ue, (11) becomes CzueAfeDt = Cz1DzueAfeDt 1
3 ? a ? h1(1 2 ft)CUeAfeDzDt 2dcft
(17)
and the collector efficiency for Model 5 h5 is 3 (1 2 ft) l5 = ? h5 ? 2 ft ? dc
(18)
The presence of a residual PCE phase in the column can change column porosity and fluid flow, but it is also possible that the PCE alters the relative adhesion of the cells to the column media. To determine if the reduction in collision efficiencies resulted from the presence of sorbed PCE, a column TABLE 1. AGW
Bacterium (1)
Effect of PCE on Transport of P17 Suspended in
Porous medium (2)
P17
Quartz
P17
Arizona soil
KT2442
Arizona soil
PCE phase (3)
Fraction retained, Fr (4)
Overall collision efficiency a (5)
None Sorbed NAPL None NAPL None NAPL
0.56 0.51 0.33 0.78 0.61 0.56 0.63
0.020 0.020 0.010 0.025 0.016 0.016 0.020
JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000 / 661
that reductions in ai along the column were not due to some unusual aspect of P17 (duplicate experiments shown in Fig. 4). However, the presence of a NAPL did not appear to alter the intracolumn collision efficiencies (Fig. 4). The presence of the PCE ganglia slightly increased the overall retention of species KT2442 from 0.56 to 0.63, resulting in calculated values of a of 0.16 (no NAPL) and 0.20 (NAPL) (Table 1). However, the intracolumn profiles of ai in the presence and absence of the NAPL were essentially identical. The larger a for KT2442 in the presence of the NAPL resulted from the slightly higher fractional retention (and therefore higher ai) measured only at the column entrance and not at other distances along the column (Fig. 4). Therefore, it was concluded that there was no measurable change in a as a result of NAPL for species KT2442. Effect of Low IS Solutions on Bacterial Transport
FIG. 3. Intracolumn Collision Efficiencies of Pseudomonas fluorescens P17 in AGW in Presence and Absence of PCE Residual Phase in: (a) Quartz Columns; (b) Arizona Soil Columns
experiment was performed using a carrying solution saturated with PCE and media pre-equilibrated with the PCE solution. The overall collision efficiency of P17 was unchanged by the presence of sorbed PCE (Table 1). The collision efficiency was a = 0.020 in the presence and absence of sorbed PCE. Thus, we concluded that the reduction in a in the presence of PCE ganglia was due to changes in column porosity and not the presence of PCE on the media surface. Effect of NAPL on Transport of KT2442 To determine if the effects of the NAPL were specific to P17, we conducted bacterial injection experiments using strain KT2442 and the Arizona soil. Intracolumn collision efficiencies of KT2442 decreased with distance along the column (Fig. 4) in the same manner as observed for P17, indicating
Decreasing the carrying solution IS below that typical of ground water (1022 M) increased bacterial transport through quartz and Arizona soil. A reduction in IS from ;1022 M (AGW) to ;1025 M (LIS water) resulted in an order-of-magnitude reduction in the fraction of P17 retained on quartz sediment and KT2442 on the Arizona soil (Table 2). The overall collision efficiencies a at the two different ISs were similarly decreased by an order of magnitude (Table 2). Intracolumn collision efficiencies measured using LIS water were not only lower than those measured using AGW, they were also much more constant over the column length (Fig. 5), although there was substantial scatter in ai in replicate experiments (particularly under LIS conditions). Values of ai for low IS water varied only by a factor of 2 for P17 and by 2.5 for KT2442 along the length of the 15-cm-long column versus a nearly two orders-of-magnitude variation under AGW conditions (Fig. 5). Effect of NAPL on Bacterial Transport in Low IS Solutions The presence of a NAPL did not produce a measurable change in bacterial retention when bacteria were suspended in LIS solutions (Table 2). The overall a for P17 in quartz-packed columns in the presence and absence of the NAPL were the same (a = 0.015). The effect (if any) of the presence of NAPL on intracolumn ai could not be determined due to the large scatter of ai measured in LIS experiments as was shown in Fig. 5(a). The overall a for KT2442 in low IS experiments was slightly larger in the presence than in the absence of the NAPL (Table 2), but again based on the scatter of the intracolumn ai at LIS (Fig. 6), this change was not considered to be measurable in these experiments. Tracer Tests Tracer tests confirmed the presence of a large fraction of immobile water in the columns produced by the NAPL. The TABLE 2.
Bacterium (1) P17 P17
FIG. 4. Intracolumn Collision Efficiencies of Pseudomonas putida KT2442 in AGW in Presence and Absence of PCE Residual Phase in Arizona Soil Columns 662 / JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000
Effect of Solution IS and PCE on Bacterial Transport
Chemical solution (2)
IS (M) (3)
AGW
1022
LIS water
25
10
22
Porous medium (4) Quartz Quartz
P17 KT2442
AGW AGW
10 1022
Arizona soil Arizona soil
KT2442
LIS water
1025
Arizona soil
Overall Fraction collision PCE retained, efficiency a phase Fr (7) (5) (6) None NAPL None NAPL None None NAPL None NAPL
0.56 0.33 0.046 0.049 0.78 0.56 0.63 0.049 0.073
0.020 0.010 0.0015 0.0015 0.025 0.016 0.020 0.0010 0.0014
presence of a PCE residual (15.2%) decreased the breakthrough time of a salt tracer from 54 to 31.5 minutes, or by a factor of 1.7. In addition to the more rapid breakthrough time, the PCE residual produced an 8.5-fold increase in the dispersion coefficient from EL (water only) = 1.3 3 1024 cm2/s to EL (PCE) = 1.1 3 1023 cm2/s. Similar changes in the PV occurred in tests using the Arizona soil. The breakthrough of the salt tracer in the presence of a residual NAPL occurred at 0.52 pv, an increase in velocity by a factor of 1.9 compared to breakthrough in the absence of the NAPL. The dispersion coefficient for the Arizona soil increased by a factor of 1.5 (5.2 3 1024 to 7.7 3 1024 cm2/s) in the presence of a 19.6% PCE residual. Modeling Effect of NAPL on Bacterial Transport
FIG. 5. Intracolumn Collision Efficiencies of: (a) Pseudomonas fluorescens P17 in Quartz-Packed Columns; (b) Pseudomonas putida KT2442 in Arizona Soil in LIS Water and AGW
FIG. 6. Intracolumn Collision Efficiencies of Pseudomonas putida KT2442 in Arizona Soil in LIS Water and AGW in Presence and Absence of PCE Residual Phase TABLE 3. Column packing (1) Quartz
Arizona soil
Five different filtration models were proposed to explain the decreases in retention of P17 in the presence of a NAPL phase. Models 2 and 4, although based on different concepts of how the NAPL was distributed in the porous media, produced identical filtration equations, thus resulting in a total of four model predictions of the effect of NAPL on particle transport. The model parameters used to calculate fractional retention of the bacteria are summarized in Table 3 for the quartz and Arizona soil. Models 2, 4, and 5 predicted decreased particle retention, whereas Models 1 and 3 predicted increased particle retention. The three models (Models 2, 4, and 5) that predicted a decrease in bacterial retention in the presence of a NAPL were compared with experimental data for P17 for quartz and Arizona soil (Fig. 7). Model 5 predicted larger decreases than Model 2 in the fractional retention of the bacteria in both the soil and quartz-media columns. This comparison of P17 data with the models indicated that Model 5, based on a decrease in the number of collectors and an increase in the PV (due to obstruction of flow paths) provided the best agreement with our P17 data. The two remaining models (Models 1 and 3) that predicted a slight increase in the fractional retention of bacteria are shown in Fig. 8. The magnitudes of the predicted increases in the fractional retention of bacteria using Models 1 and 3 are so slight that it is unlikely that these changes would be detectable. From this comparison of P17 data with the five models, Model 5 most accurately predicts the effect of the ganglia on P17 transport. Unfortunately, this model does not account for the lack of an effect of the PCE phase on the retention of KT2442. It was demonstrated (Fig. 3) that the introduction of the PCE phase did not alter KT2442 retention. A greater attraction of KT2442 to the PCE ganglia than to the soil grains, however, could partially or completely offset the decrease in retention produced by the NAPL phase. Because we observed a reproducible reduction in retention in the column of P17, but not KT2442, we believed we could support our hypothesis of
Values Used in Model Calculations for Quartz- and Soil-Packed Columns
Model (2)
u* (m/day) (3)
0 (PV) 1 2 3 4 5 0 (PV) 1 2 3 4 5
3.85 4.50 4.50 4.50 4.50 7.66 3.37 4.29 4.29 4.29 4.29 8.25
h sm (4) 7.37 6.63 6.63 6.63 6.63 4.60 8.09 6.84 6.84 6.84 6.84 4.36
3 3 3 3 3 3 3 3 3 3 3 3
1023 1023 1023 1023 1023 1023 1023 1023 1023 1023 1023 1023
h nm (dn = dc) (5)
lm (m21) (6)
lm /l0 (7)
— — — 6.63 3 1023 — — — — — 6.84 3 1023 — —
95.27 99.89 85.62 108.22 85.62 59.39 88.83 90.28 70.90 104.38 70.90 45.23
1.00 1.05 0.90 1.14 0.90 0.59 1.00 1.08 0.85 1.25 0.85 0.54
DFr (%) (8) 0.0 2.1 24.9 5.6 24.9 222.5 0.0 3.7 28.5 10.5 28.5 231.2
JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000 / 663
TABLE 4. Hydrophobic chemical (1) Hexadecane PCE
Hydrophobicity of P17 and KT2442
Solution (2)
IS (M) (3)
AGW LIS water AGW LIS water
1022 1025 1022 1025
Hydrophobicity (%) P17 (4) 79 29 97.5 95.2
6 6 6 6
KT2442 (5) 2 3 0.1 0.5
58 34 91.7 68
6 6 6 6
6 9 0.4 7
favorable KT2442 attachment to PCE by demonstrating that KT2442 partitioned more favorably into a PCE phase than P17. To determine if KT2442 would adhere more favorably to PCE than P17, we performed BATH tests using both PCE and hexadecane as the immiscible phases. Surprisingly, KT2442 was found to partition less (not more) favorably into a PCE phase than P17. In AGW, 91.7% of KT2442 bacteria partitioned into the PCE versus 97.5% of the P17 (Table 4). Additional BATH experiments using LIS water and hexadecane provided similar partitioning trends between the two bacteria. From the BATH results, we concluded that the lower retention predicted for KT2442 by Model 5 simulations was likely not due to greater attraction of KT2442 than P17 to the PCE phase. The reason for this lack of an effect of the PCE on KT2442 transport was therefore not resolved by these experiments, but based on the small changes in ai predicted for the Arizona soil in the presence of a NAPL (Fig. 7), it may be that changes in ai were too small to be observable in our experiments. FIG. 7. Comparison of Model Predictions with Intracolumn Collision Efficiencies of Pseudomonas fluorescens P17 in AGW in Presence and Absence of PCE Residual Phase in: (a) Quartz Columns; (b) Arizona Soil Columns; Data Points Are Not Shown for Clarity; Model Calculations Assume a = 0.1
FIG. 8. Comparison of Model Predictions with Intracolumn Collision Efficiencies of Pseudomonas fluorescens P17 in AGW in Presence and Absence of PCE Residual Phase in: (a) Quartz Columns; (b) Arizona Soil Columns; Data Points Are Not Shown for Clarity; Model Calculations Assume a = 0.1 664 / JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000
DISCUSSION The presence of a NAPL in soil columns did not produce increases in the retention of two bacterial strains in the porous media despite hydrophobicity measurements that indicated that both strains favorably partitioned from water into PCE or hexadecane phases. The overall effect the NAPL had on bacterial retention varied from the two strains and depended on the column packing and solution ionic strength. In AGW solutions, P17 retention decreased by 41% in quartz media and by 22% in soil columns as a result of the presence of a residual (15–21%) PCE phase in the column. Based on overall retention and intracolumn data, however, the presence of a PCE phase appeared to have little effect on the overall transport of species KT2442 in soil column experiments. The reasons for the changes in bacterial filtration rates in the presence of NAPL were explored using filtration models adapted to include different effects of the NAPL. The presence of an immiscible liquid in porous media can result in many factors that affect filtration rates including chemical sorption to soil and mineral surfaces that modifies the sticking coefficient, reduction in water porosity, increases or reduction in the number of soil grains that can function as particle collectors, and creation of immobile water in the column. The presence of a sorbed PCE phase did not measurably change the retention of cells in the quartz-media column indicating that PCE sorption by itself was not responsible for observed changes in cell transport. By including each of these factors into separate models, we were able to evaluate the relative effects of these factors. The main effect of the NAPL was apparently the creation of immobile water in the column. If the NAPL had either not altered the number of soil grains participating in filtration (Model 1), or had acted to form new collectors (Model 3), then bacterial retention would have increased in the presence of the NAPL. The large decreases measured for P17 retention in quartz-media columns could not be accounted for solely by a reduction in porosity or the number of soil grain collectors (Models 2 and 4). However, decreased retention of P17 in quartz-media columns containing a NAPL was consistent with
the presence of a large fraction of immobile water in the porous media (Model 5). The presence of immobile water was verified using a salt tracer. The presence of a 15% residual phase in the quartzmedia experiments, for example, should have increased the pore velocity from 3.85 to 4.3 m/day. Instead, a PV of 7.66 m/day was measured in the presence of the NAPL, an increase that was only possible if a substantial fraction of the void volume was now immobile (Table 3). Immobile water was also evidenced by a large increase in the dispersivity in the quartzmedia column (from 0.028 to 0.165 cm) due to the presence of the NAPL. Increases in chemical dispersivities measured here in the presence of a NAPL are also consistent with values reported by others (Pennell et al. 1994; Jin et al. 1995). The magnitude of the change in the dispersion coefficient indicates that the amount of immobile water can vary depending on the media grain size and homogeneity. Much smaller increases in chemical dispersivity were measured in the soil column, for example, than in the quartz media. This lower immobile water fraction in the soil columns than that in the quartz columns, implied by the reduced dispersivity, would account for the relatively smaller reduction in bacterial transport in the soil columns than measured in the quartz columns. Changes in bacterial retention due to replacement of the AGW with LIS water were much larger than those produced by the presence of a NAPL phase. For example, reducing the IS from 1022 M (AGW) to 1025 M resulted in a 92% decrease in overall retention of P17 in the quartz-media column and of KT2442 in the soil column. In contrast, the maximum decrease in overall retention due to the NAPL was 41% for P17 and was not detectable for KT2442. The reduction in bacterial retention due to LIS water measured here is consistent with previous studies demonstrating that LIS solutions can enhance bacterial transport in many different types of porous media (Gross and Logan 1995; Camesano and Logan 1998; Li and Logan 1999). The effect of the NAPL on bacterial transport using LIS solutions, however, has not been previously examined. However, based on the intracolumn results obtained here, it was not possible to detect an effect of the NAPL on bacterial transport under LIS conditions. Under LIS conditions, the scatter of the data was larger than the magnitude of the expected effect of the NAPL on intracolumn retention. Overall retention of the bacteria in the column under LIS conditions was also very low, thus making precise measurements more difficult under LIS conditions than in AGW. Therefore, the lack of an effect of the NAPL under LIS conditions is probably due to the relatively large scatter in the intracolumn data under LIS conditions and not a lack of an effect of the NAPL under LIS conditions. CONCLUSIONS The presence of a NAPL such as PCE in a porous medium created immobile-water zones that were larger than those attributable to just water displacement by the NAPL. The decrease in flow paths through the porous medium produced by the large immobile water phase resulted in an increase in the PV of water and, therefore, a decrease in the number of collisions possible between bacteria and soil particles. A large decrease in bacterial retention was measured for P17 injected into quartz-media columns, but smaller (P17) to nonmeasurable (KT2442) changes were measured for bacteria injected into columns packed with a sandy Arizona soil. It is concluded that the presence of the NAPL will be dependent on the specific bacterial strain and type of porous medium. ACKNOWLEDGMENTS The writers thank K. Unice for writing the computer program used to calculate the dispersion coefficients and T. Camesano and Y. Fang for
their assistance in the laboratory. This publication was made possible by Grant No. ES-04940 from the National Institute of Environmental Health Science, Research Triangle Park, N.C. Its contents are solely the responsibility of the writers and do not necessarily represent official views of the funding agency. This research was performed while the writers were at the University of Arizona.
APPENDIX.
REFERENCES
Albinger, O., Biesemeyer, B. K., Arnold, R. G., and Logan, B. E. (1994). ‘‘Effect of bacterial heterogeneity on adhesion to uniform collectors by monoclonal populations.’’ FEMS Microbiol. Lett., 124, 321–326. Camesano, T. A., and Logan, B. E. (1998). ‘‘Influence of fluid velocity and cell concentration on the transport of motile and non-motile bacteria in porous media.’’ Envir. Sci. and Technol., 32, 1699–1708. Fletcher, M., and Floodgate, G. D. (1973). ‘‘An electron-microscopic demonstration of an acid polysaccharide involved in the adhesion of a marine bacterium to solid surfaces.’’ J. Gen. Microbiol., 74, 325– 334. Gannon, J., Tan, Y., Baveye, P., and Alexander, M. (1991). ‘‘Effect of sodium chloride on transport of bacteria in a saturated aquifer material.’’ Appl. Envir. Microbiology, 57, 2497–2501. Gerba, C. P., and Bitton, G. (1984). ‘‘Microbial pollutants, their survival and transport pattern to groundwater.’’ Groundwater pollution microbiology, C. P. Gerba, ed., Wiley, New York, 53–67. Gross, M. J., and Logan, B. E. (1995). ‘‘Influence of different chemical treatments on transport of Alcaligenes paradoxus in porous media.’’ Appl. Envir. Microbiology, 61, 1750–1756. Hobbie, J. E., Daley, R. J., and Jasper, S. (1977). ‘‘Use of nucleopore filters for counting bacteria by fluorescence microscopy.’’ Appl. Envir. Microbiology, 33, 1225–1228. Hunt, J. R., Sitar, N., and Udell, K. S. (1988). ‘‘Nonaqueous phase liquid transport and cleanup. 1. Analysis of mechanisms.’’ Water Resour. Res., 24, 1247–1258. Jewett, D. G., Hilbert, T. A., Logan, B. E., Arnold, R. G., and Bales, R. C. (1995). ‘‘Bacterial transport in columns and filters: Influence of ionic strength and pH on collision efficiency.’’ Water Res., 29, 1673–1680. Jin, M., et al. (1995). ‘‘Partitioning tracer test for the detection, estimation, and remediation performance assessment of subsurface nonaqueous phase liquids.’’ Water Resour. Res., 31, 1201–1211. Johnson, W. P., Martin, M. J., Gross, M. J., and Logan, B. E. (1996). ‘‘Facilitation of bacterial transport through porous media by changes in solution and surface properties.’’ Colloids and Surfaces A: Physiochem. Engrg. Aspects, 107, 263–271. Jucker, B. A., Zehnder, A. J. B., and Harms, H. (1998).‘‘Quantification of polymer interactions in bacterial adhesion.’’ Envir. Sci. and Technol., 32, 2909–2915. Kinoshita, T., Bales, R. C., Yahya, M. T., and Gerba, C. P. (1993). ‘‘Bacterial transport in a porous medium: Retention of Bacillus and Pseudomonas on silica surfaces.’’ Water Res., 27, 1295–1301. Li, Q., and Logan, B. E. (1999). ‘‘Enhancing bacterial transport for bioaugmentation of aquifers using low ionic strength solutions and surfactants.’’ Water Res., 33, 1090–1100. Logan, B. E. (1999). Environmental transport processes, Wiley, New York. Logan, B. E., Jewett, D. G., Arnold, R. G., Bouwer, E. J., and O’Melia, C. R. (1995). ‘‘Clarification of clean-bed filtration models.’’ J. Envir. Engrg., ASCE, 121(12), 869–873. McCaulou, D. R., Bales, R. C., and Arnold, R. G. (1995). ‘‘Effect of temperature-controlled motility on transport of bacteria and microspheres through saturated sediment.’’ Water Resour. Res., 31, 271– 280. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). Molecular cloning: A laboratory manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Marlow, H. J., Duston, K. L., Wiesner, M. R., Thompson, M. T., Wilson, J. T., and Ward, C. H. (1991). ‘‘Microbial transport through porous media: The effects of hydraulic conductivity and injection velocity.’’ J. Haz. Mat., 28, 65–74. Martin, M. J., Logan, B. E., Johnson, W. P., Jewett, D. G., and Arnold, R. G. (1996). ‘‘Scaling bacterial filtration rates in different sized porous media.’’ J. Envir. Engrg., ASCE, 122(5), 407–415. Nu¨ßlein, K., Maris, D., Timmis, K., and Dwyer, D. F. (1992). ‘‘Expression and transfer of engineered catabolic pathways harbored by Pseudomonas spp. introduced into activated sludge microcosms.’’ Appl. Envir. Microbiology, 58, 3380–3386. Pennell, K. D., Jin, M., Abriola, L. M., and Pope, G. A. (1994). ‘‘Surfactant enhanced remediation of soil columns contaminated by residual tetrachloroethylene.’’ J. Contam. Hydrol., 16, 35–53. Rajagopalan, R., and Tien, C. (1976). ‘‘Trajectory analysis of deep-bed JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000 / 665
filtration with the sphere-in-cell porous media model.’’ AIChE J., 22, 523–533. Rogers, B. (1997). ‘‘Bacterial transport through a NAPL contaminated porous media.’’ MS thesis, Dept. of Chemical and Envir. Engrg., University of Arizona, Tucson, Ariz. Rosenberg, M. (1984). ‘‘Bacterial adherence to hydrocarbons: A useful technique for studying cell surface hydrophobicity.’’ FEMS Microbiol. Lett., 22, 289–295. Runkel, R. L. (1996). ‘‘Solution of the advection-dispersion equation: Continuous load of finite duration.’’ J. Envir. Engrg., ASCE, 122(9), 830–832. Scholl, M. A., and Harvey, R. W. (1992). ‘‘Laboratory investigations on the roles of sediment surface and groundwater chemistry on the attachment of bacteria to representative aquifer material.’’ J. Contam. Hydrol., 6, 321–336. Sharma, M. M., Chang, Y. I., and Yen, T. F. (1985). ‘‘Reversible and irreversible surface charge modification of bacteria for facilitating transport through porous media.’’ Colloids Surf., 16, 193–206. Shonnard, D. R., Taylor, R. T., Hanna, M. L., Boro, C. O., and Duba, A. G. (1994). ‘‘Injection-attachment of Methylosinus trichosporium OB3b
666 / JOURNAL OF ENVIRONMENTAL ENGINEERING / JULY 2000
in a two-dimensional miniature sand-filled aquifer simulator.’’ Water Resour. Res., 30, 25–35. Stenstro¨m, T. A. (1989). ‘‘Bacterial hydrophobicity, an overall parameter for the measurement of adhesion potential to soil particles.’’ Appl. Envir. Microbiology, 55, 142–147. van Loosdrecht, M. C. M., Lyklema, J., Norde, W., Schraa, G., and Zehnder, A. J. B. (1987a). ‘‘The role of bacterial cell wall hydrophobicity in adhesion.’’ Appl. Envir. Microbiology, 53, 1893–1897. van Loosdrecht, M. C. M., Lyklema, J., Norde, W., Schraa, G., and Zehnder, A. J. B. (1987b). ‘‘Electrophoretic mobility and hydrophobicity as a measure to predict the initial steps of bacterial adhesion.’’ Appl. Envir. Microbiology, 53, 1898–1901. Wagner-Dobler, I., Pipke, R., Timmis, K. N., and Dwyer, D. F. (1992). ‘‘Evaluation of aquatic sediment microcosms and their use in assessing possible effects on introduced microorganisms on ecosystem parameters.’’ Appl. Envir. Microbiology, 58, 1249–1258. Yao, K. M., Habibibian, T., and O’Melia, C. R. (1971). ‘‘Water and waste water filtration: Concepts and applications.’’ Envir. Sci. and Technol., 5, 1105–1112. Yates, M. V. (1988). ‘‘Modeling microbial fate in the subsurface environment.’’ Crit. Rev. Envir. Cont., 17, 307–344.
