| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Earth Sciences, Utrecht University, P.O. Box 80021, 3508 TA Utrecht, The Netherlands
b Microbiological Laboratory for Health Protection, National Institute of Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands
* Corresponding author (Torkzaban{at}geo.uu.nl)
Received 15 July 2005.
 |
ABSTRACT |
|---|
 TOPEXECUTIVE SUMMARY ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
X174, used as surrogate pathogenic viruses, at various water contents and solution chemistries in terms of pH and ionic strength (IS). The objective was to explore the interaction of viruses with the solid–water interfaces (SWI) and air–water interfaces (AWI) for a range of conditions. The experimental data were fitted with a transport model to determine the adsorption (attachment and detachment rate) parameters. Our results show that the retention of viruses in the soil column increases as water saturation decreases when the chemical conditions are favorable for adsorption (pH 7 and relatively high IS). Our analysis indicates that the enhanced retention of X174 viruses at lower water contents is caused by increased attachment to the SWI and that retention by the AWI is not significant. Results obtained from a first series of experiments (pH 9 and low IS) showed insignificant attachment of MS2 viruses to both the SWI and the AWI. The MS2 breakthrough data for a second series of experiments (pH 7 and high IS) did not allow us to separate out the role of the AWI. Although attachment of MS2 viruses to the AWI cannot be ruled out in our experiments, we suspect that the increased retention of this phage under unsaturated condition was also due to enhanced attachment to the SWI. Increased attachment to the SWI under unsaturated conditions is attributed to increased mass transfer of viruses to the SWI due to a reduced diffusion length at lower water contents. Our results demonstrate that if there is any attachment to the AWI, it is reversible. When unfavorable conditions occur for attachment to the SWI, the attached viruses may be detached by moving solid–water–air contact lines (SWA).
Abbreviations: AWI, air–water interfaces • EC, electrical conductivity • IS, ionic strength • pfu, plaque-forming units • SWA, solid–water–air contact lines • SWI, solid–water interfaces
|
INTRODUCTION |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The enhanced removal of viruses in the vadose zone, as compared with the saturated zone, has been attributed to increased adsorption onto SWI, irreversible attachment to AWI, and/or attachment to SWA (e.g., Chu et al., 2001; Lance and Gerba, 1984; Bitton et al., 1984; Powelson et al., 1990; Thompson et al., 1998). The hypothesis of irreversible attachment of viruses to the AWI was supported by visualization studies involving colloid transport in micromodels (Wan and Wilson, 1994a; Sirivithayapakorn and Keller, 2003). Wan and Wilson (1994a) suggested that a wide variety of colloids, including hydrophobic and hydrophilic latex microspheres, clay particles, and bacteria, can irreversibly accumulate at the AWI and that this accumulation increases with increasing particle hydrophobicity and solution IS. Colloids that are transported to the AWI are believed to be retained by either capillary or electrostatic forces. Hence, interfacial attachment will depend on pH, IS, and colloid surface properties. An increase in IS reduces the magnitude of the repulsive energy barrier between the negatively charged air–water interface and similarly charged mineral colloids. This produces more favorable conditions for attachment and faster rates of AWI capture (Wan and Wilson, 1994a; Saiers and Lenhart, 2003). In contrast, Crist et al. (2004) and (2005) observed that hydrophilic latex microsphere colloids did not adsorb to the AWI, and that colloid retention occurred at the SWA. Wan and Tokunaga (2002) demonstrated in bubble column experiments that only positively charged particles attached to the negatively charged AWI. This suggests that colloid immobilization at the AWI would be limited to a small subset of environmental colloids.
Several column-scale experiments were previously conducted to study the effect of unsaturated conditions on virus removal using bacteriophages MS2 and/or X174 (Chu et al., 2001, 2003; Keller and Sirivithayapakorn, 2004). Chu et al. (2001) obtained different retention results for different sand surface properties. For clean sand (treated to remove all metal oxides), they found that the slight attachment of viruses to the AWI was irreversible, possibly due to strong interfacial forces. Results for untreated sand suggested that the presence of iron oxides created favorable conditions for attachment to the SWI and dominated the removal of viruses under unsaturated conditions. Keller and Sirivithayapakorn (2004) observed that the removal of MS2 increased significantly with decreasing water content. The enhanced removal was ascribed to attachment to the AWI and straining in thin water films. Hydrophobic interaction (Schijven and Hassanizadeh, 2000) has also been suggested to contribute to the adsorption of viruses to the SWI and the AWI. However, there is insufficient evidence to show that this process is a significant contributor to virus attachment.
Although earlier studies have clearly shown that lower water contents lead to more virus retention in soil columns, the relative importance of the various mechanisms involved is still unclear. The objective of this study was to examine the effect of water content on the fate and transport of viruses in porous media, in particular their adsorption (attachment and detachment) behavior. Two sets of column experiments involving different chemical conditions and saturation levels were performed. Two bacteriophages were employed with different hydrophobicity and surface charge properties. In the first set of experiments, a solution of high pH (9) and low IS (0.6 mM) was used. For these conditions, one would expect negligible adsorption of both bacteriophages to the SWI; however, they may attach to the AWI. The second set of experiments was conducted using a solution of neutral pH (7) and higher IS (19 mM). Under these conditions, adsorption to the SWI may become significant. The various experiments also allowed us to investigate the effect of water content on SWI adsorption. A mathematical model (HYDRUS-1D) that takes into account virus attachment and detachment to two types of kinetic adsorption sites was used to study the various retention mechanisms.
|
MATERIALS AND METHODS |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
X174 were selected as model viruses. MS2 is an icosahedral phage with a diameter of 27 nm and a low isoelectric point of 3.5 (Penrod et al., 1996); conversely, X174 is less hydrophobic than MS2 (Shields and Farrah, 1987) and has an isoelectric point of about 6.7 and a size of 23 nm (Dowd et al., 1998).
Highly concentrated suspensions of MS2 and X174 were prepared as described in ISO 10705–1 and ISO10705–2 (ISO, 2000a, 2000b), respectively. These concentrated bacteriophage suspensions were used to prepare stock suspensions, diluted with 1 g L–1 peptone-saline to a concentration of 1010 to 1011 phages L–1. The concentrated stock suspensions were subsequently used to prepare seeding suspensions of approximately 109 phages L–1. MS2 was assayed as described in ISO 10705–1 (2000a) using host strain WG49 (ATCC 15597). Bacteriophage X174 was assayed according to ISO 10705–2 (2000b) using WG5 (ATCC 70078) as the host.
Tris (pKa of 8.55 at 0°c) and deionized water were used to prepare the buffer solution. In the first series of experiments, the pH of the solution was adjusted to 9 by addition of NaOH (1 mM), while the ionic strength was kept low (
0.6 mM). In the second series of experiments, the pH and IS of the buffer solution were adjusted to 7 and 19 mM by addition of HCl (1 mM) and NaCl, respectively. The pH and IS of each experiment are listed in Table 1.
|
X174 at about the same concentration as the input concentrations of the column experiments. The suspensions were regularly analyzed for a period of 1 mo to yield inactivation rates for the two different chemical conditions.
Column Set-Up and Experiments
A schematic of the column setup is shown in Fig. 1
. A cylindrical PVC column with an internal diameter of 10 cm and length of 23 cm was used. A hydrophilic nylon membrane with 10-µm pore size (SaatiTech, Veniano, Italy), supported by a stainless-steel plate, was used as a capillary barrier at the bottom of the column. Results of preliminary experiments (performed without sand) demonstrated that neither the column nor the end-fitting assemblies removed viruses from the suspension. Sand with a median grain size (d50) of 140 µm and uniformity (d60/d10) of 1.6 was used as the porous medium. The sand was treated with HCl (1 mM) for about 20 min to remove any attached colloids, and then rinsed thoroughly with distilled water followed by oven drying at 110°C. The free iron oxide content on the sand surface was measured to be 45 mg Fe kg–1 sand using the procedure modified by Mehra and Jackson (1960). The sand was wet-packed in the column following the procedure of Robinson and Friedman (2001) to produce a homogeneous packing. A shaker was constructed and used to uniformly apply the buffer or the virus suspension to the inlet of the column.
|
The column was flushed under saturated conditions with about seven pore volumes of the buffered solution to standardize the ionic strength and pH of the system. The chemical conditions were verified by measuring the pH and IS of both the influent and effluent. Unsaturated experiments were performed at various saturation levels. Uniform saturation and capillary pressure along the column were established in each experiment; deviations in water content between the upper and lower ports were <7%. Uniform saturation conditions were achieved at a given flow rate by gradually increasing the suction at the bottom of the column by lowering the outflow level. Readings from the three minitensiometers confirmed that a unit hydraulic gradient was established along the column. Once uniform and steady-state flow was established at a given water content, the only driving force for water flow was gravity. In this case Darcy flux equals the hydraulic conductivity at that saturation level. To minimize the effect of the flow velocity on our results, we applied a hydraulic gradient in the saturated experiments such that the flow velocity was approximately the same as in one of the unsaturated experiments. Table 1 shows a summary of the various column experiments.
Before phage application, tracer experiments were performed to determine the dispersion coefficient of each experiment. A solution containing 1 g L–1 of NaCl was applied to the column at a known water saturation. The breakthrough curve was then obtained by measuring the electrical conductivity (EC) of the outflow as a function of time. Following the salt tracer experiment and after the salt was thoroughly flushed from the column, a suspension of bacteriophages MS2 and X174 containing about 106 plaque-forming units (pfu) per milliliter of each phage was introduced onto the column. The virus suspension was applied for five to ten pore volumes (seeding duration) in attempts to obtain a steady-state breakthrough curves. Virus-free solution was subsequently applied to study the elution portion of the effluent curve. In most cases, the column was drained to investigate the role of a moving SWA on virus remobilizing. Several studies (Gomez-Suarez et al., 2001; Saiers et al., 2003) previously reported the detachment of bacteria and colloids from solid surfaces on the passage of AWI, as a result of the SWA.
Effluent samples were collected from the bottom of the column at regular intervals using a fraction collector, and analyzed for virus concentrations. All experiments were conducted in a cold room (5 ± 3°C) to minimize inactivation of viruses. Five experiments were performed at different saturation and solution chemistry levels, as shown in Table 1. They were labeled by the corresponding conditions. For example, Exp. HpLi100 was performed at high pH, low IS, and 100% saturation. Experiments HpLi100 and HpLi65 were conducted for conditions in which adsorption to the SWI was expected to be minimal. This was achieved by increasing the pH of the solution to 9 and maintaining a very low IS (0.6 mM). For this pH condition, positively charged sites on the surface of sand grains originated from iron oxides should have reversed to become negatively charged (Elimelech et al., 2000). Experiments LpHi100, LpHi66 and LpHi52 were performed at pH 7 and IS of 19 mM, where more extensive attachment to the SWI was expected due to presence of iron oxides on the surface grains (45 mg Fe Kg–1).
Modeling of Transport and Fate of Viruses under Saturated and Unsaturated Conditions
After a certain travel time and travel distance through the porous media, viruses are removed from the soil water. Processes of major importance for the removal of viruses during this passage are adsorption and inactivation. The governing equation of one-dimensional unsaturated virus transport, including terms accounting for adsorption to the SWI and the AWI is:
 | [1] |
is the water content; D is the dispersion coefficient (L2 T–1); q is the Darcy water flux density (L T–1); µl is the inactivation rate coefficient for free viruses (T–1); and rs and ra are adsorption rates to the SWI and AWI, respectively (pfu L–3 T–1). Note that inactivation of free viruses is assumed to be of a first- order process (Schijven and Hassanizadeh, 2000).
The following kinetic formulation for adsorption of viruses to the SWI was employed (Bales et al., 1991, 1993; Schijven and Hassanizadeh, 2000):
 | [2] |
b is the bulk density of the soil (M L–3), and katts and kdets are the attachment and detachment rate coefficients, respectively (T–1). Note that the attachment, detachment, and inactivation rates are assumed to be linear. In the case of heterogeneities in the soil properties or between virus particles, more than one kinetic adsorption site may be present (Schijven and Simunek, 2002). We note that, even in the case of a homogeneous SWI, katts and kdets are not necessarily constant since they may be affected by such factors as pH, ionic strength, organic matter content, temperature, grain size, and flow velocity (Schijven and Hassanizadeh, 2000). Under unsaturated conditions, the adsorption coefficients may also be a function of water content.
The attachment of viruses to the AWI can be described by the following mass balance equation:
 | [3] |
Parameter Estimation
The governing equations (Eq. [1]
–[3]) were solved numerically subject to appropriate initial and boundary conditions using the HYDRUS-1D software package (Simunek et al., 1998). HYDRUS-1D simulates water, heat, and multiple solute movement in one-dimensional variably saturated porous media. We used a modified version of HYDRUS-1D that permits consideration of two distinct sites for reversible kinetic adsorption (Schijven and Simunek, 2002). The code is coupled to a nonlinear least square optimization routine based on the Marquardt–Levenberg algorithm (Marquardt, 1963) to facilitate the estimation of solute transport parameters from experimental data.
As can be seen from Eq. [1], [2], and [3], the model contains the following parameters: the water content (), the average pore water velocity (v = q/), the dispersivity (
= D/v), three inactivation rate coefficients (µl, µs, µa), and the attachment and the detachment rate coefficients. Several of these parameters were estimated from independent experiments. The water content () was measured directly; this parameter, as explained earlier, was reasonably constant in time and space. The average pore water velocity (v) was calculated from direct measurement of the water flux and the water content for each experiment. The dispersivity () of the saturated and unsaturated column was estimated by inverse analysis of the breakthrough data of NaCl. As expected, the dispersivity was found to increase with decreasing saturation (Toride et al. 2003). We assumed that the NaCl-based dispersivity was the same as for the viruses. Small variations in the dispersivity were judged not to be important because advection was always the dominant transport process in our experiments, as reflected by relatively large values of the column Peclet number, Pe = L, where L is column length. Values of Pe in our experiments were always >40.
The inactivation rate coefficient of free viruses (µl) as measured with the batch experiments was assumed to be applicable to the column experiments as well. The inactivation rate coefficient for bacteriophages attached to the solid surface (µs) was furthermore assumed to be the same as µl. Enhanced or decreased inactivation of attached viruses has been reported (e.g., Sakoda et al., 1997; Gerba, 1984). However, the ratio of µl/µs is generally close to one (Schijven and Hassanizadeh, 2000). In any case, we fitted the saturated breakthrough curves with and without µs and found no significant effect on the values of attachment and detachment coefficients. The fitted value of µs was found to be very small and of the same order as µl. This is because of the short duration of our experiments (5–6 h) and low temperature that we were able to assume solid phase inactivation to be insignificant. Similar results were obtained by fitting µa to unsaturated breakthrough curves. Therefore, µa was also assumed to have the same value as µl.
The attachment and detachment rate coefficients (katts and kdets) for saturated column experiments were obtained by fitting the solution of Eq. [1] and [2] to the virus breakthrough data. The very small concentration values at the tail of the breakthrough curve were found to have an insignificant effect on the fitted parameter values, although they can be given more weight by using log-transformed values of the breakthrough concentrations in the fitting process.
Values of katts and kdets determined from the saturated experiments cannot be transferred directly to the unsaturated experiments since they likely depend on water content. Therefore, unsaturated breakthrough curves were fitted with Eq. [1], [2], and [3] to determine the rate coefficients katts, kdets, katta, and kdeta at a given water content. Results of the fitting process were at first very much affected by initial estimates for the parameters. This problem was avoided by using the following protocol: values of katts and kdets obtained from the saturated experiments were used as initial estimates for the corresponding parameters under unsaturated conditions. With these initial values, the breakthrough curves were fitted without log-transformation, so data from the breakthrough part of the curve and the plateau were dominant. The resulting fitted parameters were then used as initial guesses for fitting the log-transformed effluent curves. This resulted in a final set of parameter values, which were still compared with the data to make sure that they fitted the entire effluent curve. A similar procedure was used for fitting saturated breakthrough data with the two-site kinetic model (katt1, kdet1, katt2, and kdet2) when a single site model was found to be unable to produce a reasonable fit.
|
RESULTS |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
X174 was 1.3 x 10–5 min–1 (0.014 d–1). As mentioned, these values were also used for µs and µa.
Description of the Breakthrough Curves
Measured and fitted breakthrough curves for all experiments (conducted at different saturation levels and solution chemistry) are shown in Fig. 3
and 4.
Breakthrough curves (left) are plotted with the normalized concentration (C/C0) on a linear scale to emphasize differences in the maximum values of C/C0 between experiments. Breakthrough curves (right) are also plotted with C/C0 on a logarithmic scale to focus on the tails of the breakthrough curves.
|
|
|
X174, a one-site kinetic model still provided a satisfactory fit of the breakthrough curves. In the case of MS2, however, a more complex transport behavior was observed. The tails of the breakthrough curves of MS2 plotted in log scale showed a relatively smooth transition of the declining limb to a long straight tail, even in the case of 100% saturation (Exp. LpHi100). It was therefore necessary to employ a reversible two-site kinetic adsorption model to fit the breakthrough curves of MS2.
Effect of Solution Chemistry
As can be seen from Table 2, katt increases and kdet decreases as the solution condition becomes favorable for adsorption. To compare the retention capacity of the bacteriophages at different solution chemistry and saturation levels, the distribution coefficient KD for each site was determined as the ratio of katt/kdet. Calculated values are given in Table 3. Moreover, the mass balance of viruses was determined by comparing the number of viruses coming out of the column, calculated from integrating the area under the breakthrough curves, with the total number of seeded viruses (Table 4).
|
|
Under saturated conditions, the value of katt for X174 is about 20 times higher at pH 7 and higher IS (Exp. LpHi100) than at pH 9 and low IS (Exp. HpLi100), whereas the value of kdet is about 20 times lower. A similar effect is observed for unsaturated conditions (Exp. HpLi65 vs. LpHi66 and LpHi52).
It should be noted that we cannot evaluate individually the relative roles of pH and IS. These data suggest only the combined influence of the two parameters on katt and kdet.
Effect of Water Content
From Table 3, we can clearly see that when the condition is favorable for attachment, the retention of both X174 and MS2 tends to increase as the saturation decreases. At high pH and low IS, the effect of water content was negligible for both bacteriophages, as adsorption was small. The reverse seemed to be true in experiments conducted at pH 7 and higher IS. This is also reflected in the values of katt presented in Table 2. The value of katt for both phages increased with decreasing water content. However, there was no clear effect of water content on the value of kdet.
Effect of Drainage
To investigate whether the bacteriophages retained in the column were attached to the SWI or to the AWI, the column was fully drained at the end of Exp. HpLi65, LpHi100, and LpHi66. This was done by stopping the water influx and lowering the effluent hanging tube to gradually drain the column under gravity.
A sharp increase of outflow concentration for X174 was observed after draining the column (Fig. 3b, 4a, and 4b). At 100% saturation (Exp. LpHi100), X174 particles could only have been retained at the SWI and thus were subsequently remobilized by moving the SWA contact lines during drainage. Remobilization of bacterial and clay particles attached to the SWI by moving SWA contact lines has also been reported by other researchers (Powelson and Mills, 2001; Saiers et al., 2003; Sirivithayapakorn and Keller, 2003).Under unsaturated conditions (Exp. HpLi65 and LpHi66), no distinction could be made between release of X174 from the SWI or AWI.
For MS2, an increased concentration following drainage was only observed at pH 9 and low IS (Exp. HpLi65, Fig. 3b), but not at pH 7 and higher IS (Exp. LpHi100 and LpHi66, Fig. 4a and 4b). We believe that at pH 9 and low IS, remobilization of MS2 by the moving SWA contact lines was facilitated by the strong negative charge of MS2. In Exp. LpHi100, despite the fact that more MS2 was retained, no increased concentration was observed after draining the column. This may suggest that the attached MS2 particles were strongly bound to the SWI such that passing of the SWA contact lines was not able to remobilize them. This hypothesis is in agreement with the high value of katt and the low value of kdet to the first site of adsorption. In Exp. LpHi66, no increased concentration in breakthrough curve was also observed after draining the column.
|
DISCUSSION |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Adsorption of viruses to the AWI is believed to be controlled by solution chemistry, particle surface charge, and hydrophobicity (Wan and Tokunaga, 2002; Wan and Wilson, 1994a). Studies have shown that air–water and oil–water interfaces are negatively charged for pH values larger than 2 (Li and Somasundaran, 1991; Graciaa et al., 1995; Marinova et al., 1996). Wan and Tokunaga (2002) demonstrated in bubble column experiments that only positively charged particles attached to the negatively charged AWI. Under our experimental conditions, where both MS2 and X174 are negatively charged, we may expect little attachment to the AWI. Indeed, we believe that virus retention by the solid–water interface was the dominant effect and that retention by the air–water interface was insignificant.
Although it may seem difficult to differentiate between attachment to the SWI and the AWI, we believe that a careful analysis of the breakthrough curves, especially the tail part, may provide clues to the relative significance of these two attachment mechanisms. If there is attachment to both the AWI and SWI, a two-site kinetic model would likely be required to simulate the breakthrough curves under unsaturated conditions. However, in the case of X174, for both high and low pH experiments and at various saturations, a one-site kinetic model could simulate breakthrough curves satisfactorily. Similar results were obtained for MS2, but only for high pH and low IS conditions. Therefore, we conclude that attachment to the AWI was negligible in these cases.
For low pH, high IS experiments, however, a two-site kinetic model was needed for simulating MS2 breakthrough curves in both saturated and unsaturated experiments (see parameters for LpHI100 in Tables 2 and 3). This suggests that there are two different adsorption sites on the SWI: a weak-adsorbing site and a strongly adsorbing site. If there was significant adsorption to the AWI, we believe that a three-site kinetic model would be required to simulate the breakthrough curves of MS2 in the LpHi66 and LpHi52 experiments. This was, however, not the case. We are therefore led to the conclusion that adsorption to the AWI was not significant, or that it was of the same order of magnitude as one of the adsorption sites on the SWI, so that the two effects could be lumped. Although adsorption of MS2 to the AWI cannot be ruled out completely, we speculate that increased attachment under unsaturated experiments occurs because of enhanced attachment to the SWI instead of to the AWI. Further research is needed to prove this hypothesis.
Additional evidence on the insignificance of AWI adsorption comes from the drainage of the columns at the end of the HpLi65, LpHi100, and LpHi66 experiments. The drainage process creates AWI in the column. If there was adsorption to the AWI, drainage should create additional adsorption sites, and thus, it should produce a drop in the breakthrough concentration. In the case of X174, the opposite trend was observed and resulted in a sharp increase in the breakthrough concentration. A similar result was obtained for MS2 in Exp. HpLi65, where adsorption to SWI was weak. No effect of drainage on the breakthrough concentration of MS2 was observed in Exp. LpHi100 and LpHi66, where there was strong adsorption to the SWI.
Our results are in contrast to some earlier studies that suggest that irreversible attachment to the AWI is the main mechanism for enhanced removal of viruses and colloidal particles with decreasing water content (e.g., Wan and Wilson, 1994a; Powelson et al., 1990; Jin et al., 2000; Powelson and Mills, 2001; Keller and Sirivithayapakorn, 2004).Our results indicate that if there was any attachment to AWI, it certainly was not irreversible. In all cases, we needed to include relatively large detachment rate coefficients in our model (Table 2) to obtain a satisfactory fit to the breakthrough curves, especially in the tail portion of the curves. Moreover, C/C0 values of both bacteriophages reached a plateau with the value of unity, and the mass recovery was almost 100% in saturated and unsaturated experiments conducted at a high pH. This means that for the duration of our experiments, inactivation and irreversible attachment to both the SWI and AWI were negligible.
Some authors have suggested that hydrophobic interactions are responsible for attachment of viruses to the AWI (Jin et al., 2000). Here again, if this were the case, we would have needed two different sites for modeling unsaturated breakthrough curves. Moreover, hydrophobic interactions should have been more pronounced at a high pH and low IS, where conditions for adsorption to the solid phase are unfavorable. This behavior, however, was not observed. We therefore believe that hydrophobic interactions do not play a significant role in adsorption of these viruses to the AWI.
A two-site kinetic model was needed to simulate MS2 breakthrough curves in both saturated and unsaturated experiments that were performed at pH 7 and IS of 19 mM. Table 3 shows that there was a strong adsorption site and a weak adsorption site, with saturated KD values equal to 123 and 0.005, respectively. The only plausible explanation for such a strong adsorbing site under saturated conditions is the presence of iron oxide on the surface of sand (45 mg Fe kg–1; see Materials and Methods). At a neutral pH, iron oxides are positively charged, whereas MS2 is still negatively charged. This results in strong adsorption of MS2 to the SWI under both saturated and unsaturated conditions. The same result does not hold for X174 because it has an almost neutral surface charge at a pH of 7. Indeed, breakthrough curves for X174 were satisfactorily modeled using a one-site kinetic model with a relatively small distribution coefficient, even at a pH of 7 (Table 3).
There is a clear increase in the retention of both phages as the saturation decreases. This enhanced adsorption can be explained as follows. Kinetic adsorption is the result of two mechanisms: diffusion of viruses from within the pores to the soil grain surfaces and then attachment to the surfaces. The reverse occurs in the case of desorption. As the saturation decreases, the water moves into smaller pores or in narrow wedge-shaped corners. As a result, the diffusion length decreases and this leads to faster adsorption kinetics. Also, at lower saturations, zones of immobile water will be formed within the soil. Diffusion into these zones will manifest itself as enhanced adsorption. Some researchers have suggested that increased attachment of viruses to the SWI is due to an increase in electrostatic and hydrophobic interactions under unsaturated conditions, as larger pores are no longer available for transport and viruses are closer to the SWI (Lance and Gerba, 1984; Chu et al., 2001). However, this is unlikely to be the case, as was also shown by Powelson et al. (1990), because the effective range of electrostatic and hydrophobic forces are on the order of nanometers and the sizes of water-filled pores are much larger.
We emphasize that our results do not exclude the possibility of significant adsorption of viruses to the AWI under other chemical conditions. More studies are needed to determine the range of chemical conditions where significant attachment to AWI may occur.
|
CONCLUSIONS |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
X174), two different solution chemistries, in terms of pH and ionic strength, and various saturation levels. The following conclusions may be drawn:X174 and MS2 were retained significantly by attachment to the SWI due to electrostatic interactions. X174 and MS2 increased, whereas detachment was little affected. The increased attachment was attributed to enhanced attachment to the SWI. |
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPEXECUTIVE SUMMARYABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. A. Bradford, E. Segal, W. Zheng, Q. Wang, and S. R. Hutchins Reuse of Concentrated Animal Feeding Operation Wastewater on Agricultural Lands J. Environ. Qual., September 2, 2008; 37(5_Supplement): S-97 - S-115. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. A. Bradford and S. Torkzaban Colloid Transport and Retention in Unsaturated Porous Media: A Review of Interface-, Collector-, and Pore-Scale Processes and Models Vadose Zone J., May 27, 2008; 7(2): 667 - 681. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Pang, M. McLeod, J. Aislabie, J. Simunek, M. Close, and R. Hector Modeling Transport of Microbes in Ten Undisturbed Soils under Effluent Irrigation Vadose Zone J., January 23, 2008; 7(1): 97 - 111. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
