| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Plant and Soil Sciences, Box 9555, 117 Dorman Hall, Mississippi State University, Mississippi State, MS 39762. Mississippi Agriculture and Forestry Experiment Station #J10257
* Corresponding author (csmith{at}pss.msstate.edu)
Received for publication October 7, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: BTC, breakthrough curve • SOM, soil organic matter
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
 | [1] |
Isotherms of nonionic organic compounds are often assumed to be linear (Chiou, 1989; Hamaker and Thompson, 1972; Karickhoff, 1984). However, when sorption is examined across a large range of concentrations, the partitioning coefficient of most herbicides is not linear. As a result, the linear Kd usually underpredicts pesticide sorption on soil at low solution concentrations and overpredicts pesticide sorption at high solution concentrations. To more accurately describe the isotherms of most organic compounds a curvilinear equation, such as the Freundlich equation, should be used.
The sorption isotherm is described by the Freundlich equation, in which:
 | [2] |
Atrazine is a weak-base herbicide with a pKa of 1.7. However, under field conditions, atrazine is nonionized. The partitioning coefficient for atrazine varies widely, depending on soil properties. Atrazine sorption is directly correlated to soil organic matter and clay content. Reported atrazine Kd or Kf values range from 0.1 to 25 (Bouchard, 1999; Ma and Selim, 1994; Pignatello and Huang, 1991; Sigua et al., 1995; Sonon and Schwab, 1995; Wauchope and Myers, 1985). The reported values of N for atrazine range from 0.59 to 1.04 (Clay et al., 1988; Hamaker and Thompson, 1972; Huang et al., 1984; Ma and Selim, 1994; Pignatello and Huang, 1991; Rao and Davidson, 1979; Sonon and Schwab, 1995; Swanson and Dutt, 1973).
Imazaquin is a weakly acidic herbicide with a pKa of 3.8. Imazaquin is moderately sorbed in the nondissociated form but weakly sorbed in the anionic form (Mangels, 1991). Imazaquin sorption is negatively correlated with soil pH (Basham et al., 1987; Goetz et al., 1986; Loux et al., 1989; Loux and Reese, 1992; Renner et al., 1988). At a solution pH 4.5, a Hoytville clay soil (fine, illitic, mesic Mollic Epiaqualf) sorbed 35% of the imazaquin initially in the soil solution (Loux and Reese, 1992). However, when the pH was 5.5 and above, sorption decreased to less than 10%. The reported Kd values for imazaquin range from 0.049 to 2.76 as soil pH decreased from 8.0 to 3.0 (Renner et al., 1988). Organic matter and clay content also greatly influence imazaquin sorption within a moderate pH range. The N value for imazaquin appears to be very close to unity (Loux et al., 1989; Renner et al., 1988; Seifert, 1999).
Pesticide sorption on soil is customarily determined using batch equilibrium assays. For these assays a relatively high solution to soil ratio (i.e., 10:1 to 2:1) is used and a rapid establishment of equilibrium is often assumed. Often, solution–soil mixtures are allowed to equilibrate from 4 to 24 h. Then, a single partitioning coefficient is calculated to describe pesticide–soil interactions that occur across a large time scale (i.e., a few seconds to several days). Extrapolating partitioning coefficients across a range of times may result in poor prediction of herbicide reactivity in either rapid reactions (i.e., leaching through a soil profile) or of long-term reactions (i.e., sorbing to or from aged pesticide residues). Thus, batch studies poorly represent normal field conditions. Under field conditions, a herbicide is exposed to lower solution to soil ratios and can interact with soil across a range of time. Additionally, herbicide dissipation influences equilibrium shifts between solution and soil phases.
In reality, herbicide sorption is probably a multiphasic reaction. Research demonstrates that sorption is composed of a rapid, almost instantaneous phase, followed by a more gradual phase (Johnson et al., 2000; O'Dell et al., 1992). The rapid phase is probably the result of sorption on surface sites of organic matter, clay colloids, and soil organic matter colloid complexes (Johnson et al., 2000; O'Dell et al., 1992). The rapid phase of sorption is more reversible than the gradual phase of sorption. Thus, in rapid reactions, sorption and desorption partitioning coefficients are more similar than those resulting from gradual phase of sorption (De Jonge et al., 2000; Pignatello and Xing, 1996).
The more gradual phase of sorption probably results from diffusion of herbicides into three-dimensional soil structures. Herbicide sorption experiments conducted across several days have demonstrated that sorption coefficients are not constants but rather increase with time (De Jonge et al., 2000; Johnson et al., 2000; Lehmann et al., 1990; O'Dell et al., 1992; Pignatello and Huang, 1991; Scribner et al., 1992). Additionally, diffusion-driven reactions result in more tightly sorbed herbicide molecules that are thought to be less available for desorption than the herbicide molecules simply sorbed to surface sites (De Jonge et al., 2000; Johnson et al., 2000; O'Dell et al., 1992).
Experiments conducted across several days also suggest pesticide–soil equilibrium may be a misnomer. True equilibrium may never be reached within a soil, because dissipation processes such as leaching or chemical, physical, or biological transformations can occur before equilibrium is reached (Brusseau and Rao, 1989). Thus, it is appropriate to discard the idea of an equilibrium-partitioning coefficient and define the reactions of interest as either rapid or gradual and examine the degree of sorption as it relates to the contribution of each phase of reaction. The temporal division between rapid and gradual phases of sorption is arbitrary (Pignatello and Xing, 1996). Weber and Huang (1996) examined changes in sorption isotherms at incubation times ranging from 1 min to 14 d. However, Xing and Pignatello (1996) examined rapid reactions after 1 d of incubation and gradual reactions after 30 and 180 d of incubation.
Nonequilibrium sorption can be the result of diffusion-limited transport of pesticide molecules through soil structures (Brusseau and Rao, 1989). These structures include, but are not limited to, interstitial micropores of soil aggregates, interstitial colloid spaces, and the three-dimensional matrices of natural organic matter (Bouchard et al., 1988; Nkedi-Kizza et al., 1989; Pignatello, 1990; Wu and Gschwend, 1986, 1988).
Given the high solution to soil ratio and the difficulty resolving rapid reactions with the traditional batch method, and that pesticide–soil interactions may never reach a true equilibrium in soil, a nonequilibrium sorption method may be more appropriate for studying partitioning isotherms and sorption kinetics than traditional equilibrium batch techniques (Hinz and Selim, 1999; Miller et al., 1989; Skopp and McCallister, 1986). This is mainly supported by the fact that pesticide sorption under saturated flow conditions may only have a few seconds to a few minutes to interact with the soil matrix. Additionally, soil aggregates, micropores, macropores, and solution to solid ratios are vastly different in natural conditions when compared with soil in the batch method. Nonequilibrium sorption techniques include the miscible-displacement method (Brusseau et al., 1991; Gamerdinger et al., 1991), stirred slurry filtration method (Li et al., 1996; Lindstrom et al., 1970; Wahid and Sethunathan, 1978), and thin-disc flow method (Miller et al., 1989; Skopp and McCallister, 1986; Thabet and Selim, 1996). Because of the constraints of the batch method, a thin-disc flow or some other nonequilibrium method may better resolve herbicide soil interaction than traditional batch equilibrium methods. Second, a thin-disc flow method may allow simultaneous examination of pesticide partitioning in a dynamic state of equilibrium and the kinetics of pesticide sorption and desorption.
The objective of this research was to explore the rapid and gradual sorption isotherms of atrazine and imazaquin by comparing partitioning coefficients derived with traditional batch techniques to partitioning coefficients determined in a nonequilibrium thin-disc flow apparatus.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Atrazine and imazaquin soil interactions were examined by using uniformly ring-labeled 14C-herbicides. Herbicide purity and specific activity was confirmed by high performance liquid chromatography (HPLC) analysis and liquid scintillation analysis. Atrazine was confirmed to be 960 g kg-1 radiochemically pure with a specific activity of 1528 kBq mg-1. Imazaquin was confirmed to be 990 g kg-1 radiochemically pure with a specific activity of 1343 kBq mg-1. The relative amount of atrazine and imazaquin parent material present after incubation and herbicide degradation products were also determined by HPLC analysis after each experiment.
Batch Isotherm Experiments
Atrazine and imazaquin partitioning isotherms were determined in separate experiments with incubation times of 5 min and 24 h. The 5-min isotherms were conducted to measure the rapid and reversible binding of herbicide to soil. The 24-h isotherms were intended to approximate short-term sorption equilibrium.
In the rapid phase (5 min) isotherm experiments, 5.0 g soil and 9.0 mL of 5.0 mM CaCl2 were combined in 50-mL polypropylene, screw-cap centrifuge tubes. All tubes were placed on a shaker with 12 cm horizontal travel and 260 oscillations per minute. Tubes were shaken for approximately 24 h ± 15 min, at 21°C. This initial exposure allowed soil aggregates and solution to equilibrate before fortification. Thus the physical, chemical, and biological properties of the soil slurry were more comparable with a 24-h batch study than if the entire 10 mL of herbicide solution was combined with dry soil and incubated for 5 min. Additionally, preliminary research had shown that disintegration of soil aggregates during the first 2 h of incubation greatly influenced the binding of weakly sorbed imazaquin. This resulted in unacceptably high experimental error and confounded herbicide-binding results (data not shown). Weber and Huang (1996) also reported greater experimental error using this method in batch studies with incubation times less than 100 min when compared with longer incubation times.
After the initial equilibration, tubes were fortified with 1.0 mL of concentrated atrazine or imazaquin solutions in 5.0 mM CaCl2. Each milliliter of fortified solution contained 0.42 kBq 14C-labeled herbicide and the appropriate amount of nonradiolabeled herbicide to achieve the desired solution concentrations. The final soil slurry had a 2:1 solution to soil ratio. After fortifying the soil slurries, the initial atrazine solution concentrations were 0.120, 0.454, 1.39, 2.78, 4.55, and 6.95 µM. A 4.64 µM solution would represent a labeled rate of atrazine diluted in the top-8-cm soil layer at the designated solution to soil ratio. The initial imazaquin concentrations were 0.134, 0.405, 1.67, 3.30, 6.51, and 9.72 µM. A 0.2 µM solution would represent a labeled rate of imazaquin diluted in the top-8-cm soil layer at the designated solution to soil ratio. Concurrently, 1.0 mL of each 14C-herbicide solution was added to 9.0 mL of 5.0 mM CaCl2 without soil. This provided a no-soil control to determine the initial herbicide concentration for each herbicide at all concentrations, experimental variance, and sorption to the laboratory apparatus.
The fortified soil slurries were immediately placed back on the shaker for 5 min. Tubes were then centrifuged for 10 min at 4500 rpm. One milliliter of supernatant from each tube was transferred to 20-mL liquid scintillation vials and combined with 10 mL of water-accepting liquid scintillation cocktail.
The solution concentration of each sample was determined directly by counting the disintegrations per minute (dpm) for 20 min and comparing the results to the radioactivity in the control samples. All samples were corrected for quench using an internal standard and external quench curves. The soil concentration of each herbicide was determined indirectly by assuming all pesticide removed from solution was sorbed to soil. This assumption was confirmed by combustion of soil in a biological oxidizer for 4 min. The evolved 14CO2 was trapped in 15 mL of scintillation cocktail and counted using liquid scintillation counting.
The gradual phase (24 h) isotherms were determined using the technique described above; however, 10 mL of 0.042 kBq mL-1 radiolabeled herbicide solution in 5.0 mM CaCl2 were added directly to centrifuge tubes containing 5.0 g soil. Thus, the amount of time the soil remained in slurry was 24 h for both the 5-min and 24-h batch experiments, but the amount of time for sorption to occur varied. The tubes containing the herbicide–soil slurry were then placed on a horizontal shaker for 24 h. The initial atrazine concentrations were 0.120, 0.464, 2.32, 4.64, 9.27, 23.2, 46.4, and 69.54 µM. The 4.64 µM solution represents a labeled rate of atrazine diluted in the top-8-cm soil layer at the designated solution to soil ratio. The initial imazaquin concentrations were 0.073, 0.738, 1.20, 1.27, 2.72, 4.83, 6.42, 7.22, 12.0, 41.3, 81.4, and 162 µM. A 0.2 µM solution would represent a labeled rate of imazaquin for an 8-cm ha furrow-slice at the designated solution to soil ratio.
The smaller range of initial herbicide solution concentrations in the 5-min experiments compared with the 24-h experiments was the result of limiting the concentration of the fortifying solution to no more than one half the aqueous solubility of each herbicide. Because the concentrated fortifying solution in the 5-min batch studies was diluted by a factor of 10 when added to the equilibrated soil slurry, the initial solution concentration range was roughly 10 times lower in the 5-min than the 24-h batch experiments.
All concentrations of both herbicides at both incubation times were replicated four times, and all experiments were duplicated. Data were subjected to analysis of variance. Means were separated using Fisher's protected least significant difference test at the 0.05 significance level.
Thin-Disc Flow Experiments
Figure 1
is a schematic of the dual radiolabeled thin-disc flow experimental apparatus. An inline filter apparatus (330 In-Line Filter Holder; Nalge Nunc International, Rochester, NY) was used in the thin-disc flow experiments. A 0.45-µm hydrophilic mixed cellulose esters filter was placed in the apparatus and secured with the manufacturer-supplied silicon O ring. The filter was evenly covered with 3.0 g air-dried, 1.7-mm-sieved soil. This created a thin disc of soil approximately 3.0 mm deep with a 44-mm diameter and 2.0-mL pore volume.
|
After pretreatment, the system flow was decreased to 1.0 mL min-1. One hundred microliters of a 5.0 mM CaCl2 solution containing 18.5 kBq 3H-water plus 1.85 kBq 14C-herbicide was injected above the soil disc. The injected herbicide solution contained either 6.5 µM atrazine or 5.0 µM imazaquin. Preliminary research suggested that these concentrations of herbicides closely approximated the amount of each herbicide desorbing from soil after a surface application to field moist soil (data not shown). The pulsed herbicide solution was then rinsed through the soil disc in 5.0 mM CaCl2 at 1.0 mL min-1. Effluent from the soil disc was collected as 1.0-mL fractions directly into 20-mL glass liquid scintillation vials. Ten milliliters of water-accepting liquid scintillation cocktail were added to each vial. The amount of 3H-water and 14C-herbicide in each fraction were determined by dual-label (Full Spectrum Dual Label DPM; Packard Instrument Co., Meriden, CT) counting on liquid scintillation analyzer. All samples were corrected for quench using an internal standard and external quench curves.
In preliminary studies, 100 µL Brilliant Blue (BB) was substituted for the radiolabeled herbicide solution. This allowed for visualization of water flow through the soil disc contained within the transparent filter apparatus. Brilliant Blue initially eluted in a fairly concentrated solution within a random zone of the filter face. Gradually, BB elution encompassed the entire filter face. The breakthrough concentration curve (BTC) with BB was determined with UV–VIS spectrophotometry, and followed the same pattern as 3H-water, regardless of the location of the initial breakthrough event at the filter surface (data not shown). Even though this suggests the initial flow was unevenly distributed across the soil disc, the experimental system was deemed satisfactory because subsequent flow included the entire filter face and the resulting BTC curves were unaffected by the location of the initial breakthrough.
The amount of radiolabeled herbicide observed in the effluent was determined directly by 14C-dpm dual label counting. The amount of injected solution remaining in the system was calculated by determining the amount of conservative tracer in the effluent with 3H-dpm dual label counting and subtracting this value from the amount of 3H-water injected into the apparatus. This value allowed the amount of herbicide remaining in the system, but not sorbed to soil, to be calculated. The amount of 14C-herbicide sorbed to the soil was calculated by subtracting the sum of herbicide contained in the effluent and the amount remaining in the apparatus but not sorbed to the soil from the amount of herbicide originally injected.
Control experiments were conducted in an identical manner; however, soil was omitted from the apparatus to test for herbicide retention by the experimental device. Each herbicide was pulsed through three separate soil discs resulting in a total of three replicate injections for each herbicide. The no-soil controls were also repeated in triplicate.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
After 5 min of incubation, 47% of the atrazine in the 0.05 µM solution was sorbed onto the soil (Fig. 2) . Atrazine sorption decreased to 38% when the initial solution concentration was increased to 7.0 µM. After 24 h of incubation, 62% of the atrazine in the lowest concentration was sorbed, compared with 53% at the highest atrazine concentration.
|
|
In atrazine batch experiments, the N value of the Freundlich equation was approximately 0.93 for both 5-min and 24-h isotherms (Table 1). The reported values of N for atrazine are 0.59 to 1.04 (Sonon and Schwab, 1995), 0.76 (Ma and Selim, 1994), 0.70 to 0.88 (Pignatello and Huang, 1991), 0.73 to 1.04 (Rao and Davidson, 1979), 0.79 (Clay et al., 1988), 0.85 to 1.00 (Swanson and Dutt, 1973), 0.96 to 1.00 (Huang et al., 1984), and 0.71 to 0.98 (Hamaker and Thompson, 1972). Since the N values for the 5-min and 24-h isotherms are equal, it can be assumed that the rapid and gradual sorption components will behave similarly over a wide range of concentrations.
Weber and Huang (1996) and Xing and Pignatello (1996) addressed the influence of incubation time on the nonlinearity of hydrophobic organic contaminants (HOCs) by proposing a phase distribution relationship in a multiple domain model. Increasing time of incubation in batch experiments consistently increased the nonlinearity of the Freundlich equation, resulting in decreasing N values. Hydrophobic organic contaminants appear to first sorb to lower energy, exposed inorganic mineral sites (Domain I) and partition into amorphous soil organic matter (SOM) in Domain II. Initially, sorption is limited by diffusion (i.e., partitioning) of the solute through amorphous SOM. Partitioning-dominated sorption is more linear than adsorption to specific reactive sites (Weber and Huang, 1996; Xing and Pignatello, 1996). The amorphous soil organic matter is typically a monolayer of fulvic and humic acids overlaying condensed SOM at the mineral surface of soil and sediment (Murphy et al., 1990). This layer is easily hydrated and swells to form micelle-like bulges that may serve as a bipolar bridge to the underlying condensed SOM (Aiken et al., 1985; Engebretson and von Wandruszka, 1994; Weber and Huang, 1996; Wershaw, 1989; Xing and Pignatello, 1996).
After the solute diffuses through the amorphous SOM, Domain III (condensed SOM) sites begin to fill. Sorption of HOCs by condensed SOM appears to be energetically more favorable and more nonlinear than Domains I or II. It is theorized that sorption by condensed SOM proceeds very slowly, is limited by intradomain diffusion, is predominated by adsorption to specific reactive sites, and, thus, is nonlinear in nature (Weber and Huang, 1996; Xing and Pignatello, 1996).
Because the N values of atrazine in our experiments did not decrease with increasing incubation time, the results appear to contradict the results of Weber and Huang (1996) and Xing and Pignatello (1996). However, the apparent contradiction may result from different experimental parameters. Neither Weber and Huang (1996) or Xing and Pignatello (1996) allowed the soil slurry to equilibrate with the aqueous environment before fortification. As a result, the degree of hydration of the amorphous SOM would be much lower in their experiments compared with ours. It is possible that the time required to fully hydrate amorphous SOM after fortification resulted in decreasing N values during the first 100 min of incubation by Weber and Huang (1996). When fulvic and humic acids are fully hydrated, as in our experiments, sorption of HOCs is typically complete in less than 10 min (Backhus and Gschwend, 1990; McCarthy and Jimenez, 1985; Schlautman and Morgan, 1993).
In our experiments, the amorphous SOM was more fully hydrated from the initial equilibration period, possibly allowing solute penetration to proceed at a faster rate compared with the results of Weber and Huang (1996). Our results support the hypothesis of Weber and Huang (1996) and Xing and Pignatello (1996) and suggests that sorption of atrazine during the rapid and gradual phases is primarily the result of partitioning into SOM.
The rapid and gradual phases of atrazine sorption accounted for 74 and 26% of the total observed sorption (Fig. 3) . The contribution of each sorption phase to the total amount of atrazine removed from solution remained constant over initial solution concentrations ranging from 0.05 to 7.0 µM. This represents a 140-fold increase in concentration from the lowest solution concentration to the highest solution concentration. These data suggest that atrazine is moderately to strongly sorbed by the rapid-phase sorption sites located on the soil surface.
|
The imazaquin isotherms in the batch experiments also best fit the Freundlich equation. The rapid phase isotherm Kf equaled 0.056, compared with 0.349 for the gradual phase (Table 1). The linear imazaquin Kd equaled 0.030 and 0.256 after 5 min and 24 h of incubation, respectively. The weak sorption observed on the Demopolis silt loam in the batch experiments agrees with results reported by other researchers. Imazaquin is moderately sorbed in the nondissociated form but weakly sorbed in the anionic form (Mangels, 1991). Imazaquin sorption is negatively correlated with soil pH (Basham et al., 1987; Goetz et al., 1986; Loux et al., 1989; Loux and Reese, 1992; Renner et al., 1988). The imazaquin Kd with a sandy loam ranged from 0.049 (at pH 8.0) to 2.76 (at pH 3.0) (Renner et al., 1988). The soil pH in our experiments was 7.6. Thus, imazaquin would occur predominately in the anionic form and be weakly sorbed given the experimental parameters.
For the 0.05 µM imazaquin solution, the rapid phase of sorption accounted for 30% of the total amount of herbicide removed from the soil solution (Fig. 3). The contribution of the rapid phase of sorption consistently decreased as the initial imazaquin solution concentration increased. With the 7.0 µM initial imazaquin concentration, the rapid sorption phase contributed only 12% to the total sorption process (Fig. 3). O'Dell et al. (1992) found that as imazethapyr [2-(4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl)-5-ethyl-3-pyridinecarboxylic acid] initial concentrations increased, a greater proportion of the total sorption was contributed by the gradual phase. Divincenzo and Sparks (1997) also found that as pentachlorophenol (PCP) concentrations increased, the soil's ability to rapidly adsorb PCP from solution was overwhelmed. As a result, the relative amount of total sorption accounted for by the rapid phase decreased, but this was offset by a greater percentage of the total sorption in the gradual phase. The fact that imazaquin and imazethapyr are weakly sorbed and PCP is strongly sorbed is irrelevant to this argument. The important consideration is that when rapid binding sites are saturated, slow sorption is more pronounced and accounts for a higher percentage of the total sorption (Divincenzo and Sparks, 1997; O'Dell et al., 1992).
Organic matter and clay content also greatly influence imazaquin sorption at moderate pH values. The reported Kd for a Sharkey clay (very-fine, smectitic, thermic Chromic Epiaquert) was 2.57, compared with 0.20 on a Goldsboro sandy loam (fine-loamy, siliceous, subactive, thermic Aquic Paleudult) (Loux et al., 1989). The pH for Sharkey and Goldsboro soils was 5.1 and 5.2, respectively. However, the organic carbon content for the Sharkey soil was 16.2 g kg-1, compared with 3.0 g kg-1 in the Goldsboro soil. Additionally, the Sharkey and Goldsboro soils had clay contents of 600 and 110 g kg-1, respectively. Given the other reported imazaquin partitioning coefficients, our observed partitioning coefficients are within the expected values.
In the batch isotherm experiments, the imazaquin N value was 0.712 for the rapid phase, compared with 0.858 after 24 h of incubation (Table 1). These values are more curvilinear than most of the reported N values for imazaquin in the literature. Literature suggests the N value for imazaquin is very close to unity (Seifert, 1999; Renner et al., 1988; Loux et al., 1989). Loux et al. (1989) reported that N for imazaquin ranged from 0.90 to 1.10 for most soils. However, for a few soils, N was outside this range (Loux et al., 1989). The rapid-phase isotherm observed in our experiments had only 5 min to associate with sorption sites, compared with a much longer incubation time for the Kf values reported by other researchers. These results suggest that adsorption to heterogeneous reactive sites dominate the rapid phase of sorption while linear partitioning dominates the gradual phase (Xing and Pignatello, 1996).
The rapid and gradual phases of imazaquin sorption observed in our experiments contradict those observed for hydrophobic organic contaminants, which tend to have increasingly nonlinear isotherms with increased incubation time (Weber and Huang, 1996; Xing and Pignatello, 1996). This disparity is potentially due to different physicochemical properties of the solutes of interest. In our experiments, imazaquin would have existed primarily as an anion and would interact very differently with SOM than the more hydrophobic and neutrally charged solutes investigated by Weber and Huang (1996) and Xing and Pignatello (1996).
It is possible that the proportion of neutral to anionic imazaquin molecules increased with increased time since the pH of the solution in close proximity to soil–organic matter complexes may be several pH units lower than the bulk solution. If anionic imazaquin molecules that were initially sorbed to, or in close proximity with soil colloids, were gradually protonated during incubation, the type of sorptive forces could have changed from primarily anionic exchange, water bridging, and cation bridging (nonlinear isotherm) to partitioning (linear isotherm). This would explain the observed increased isotherm linearity with increased incubation time.
The mechanisms of gradual-phase sorption in soils were discussed by Brusseau et al. (1989)(1991), Brusseau and Rao (1989), and Pignatello and Xing (1996). They concluded that diffusion into three-dimensional soil components should be the most important processes accounting for slow sorption. This process includes diffusion into micropores of soil aggregates, soil organic matter, and soil colloids. Diffusion of herbicide molecules into three-dimensional structures probably results in more tightly bound residues that are less subject to desorption. Additionally, the gradual sorption phase is not dependent on specific sorption sites on the soil surface that could be easily saturated with weakly sorbed herbicides. Therefore, the gradual component of soil sorption should not be as easily saturated at lower herbicide concentrations as the rapid, surface phase sorption sites. In these experiments, the observed 24-h isotherm is more linear than the rapid-phase isotherm, and the sorptive behavior of imazaquin after 5 min and 24 h of incubation supports the observation of other researchers. Other researchers report that models of diffusion into three-dimensional structures allowed good prediction of sorption kinetics of organic compounds in soil (Ball and Roberts, 1991; Wu and Gschwend, 1986).
Because the isotherms for atrazine and imazaquin sorption are nonlinear and have unequal N values, it is invalid to compare the partitioning of atrazine and imazaquin over a range of observed concentrations (Chen et al., 1999; Xing, 2001). Nonlinear isotherms were confirmed with changes in the concentration-dependent sorption coefficients (i.e., higher Kd values at lower equilibrium solution concentrations). The concentration-dependent Kd for atrazine at an equilibrium solution concentration of 0.05 µM was 1.69 after 5 min of incubation and 3.00 after 24 h of incubation (Table 2) . When the equilibrium solution concentration increased to 4.50 µM, the rapid and gradual Kd values decreased to 1.25 and 2.16, respectively. Given that the gradual to rapid Kd ratio equaled 1.78 at the lowest solution concentration and 1.73 at the highest solution concentration, it is clear that the rapid and gradual sorptive processes behave similarly over a large range of solution concentrations. Thus, it is valid to compare the Kf values of the rapid and gradual phases of atrazine sorption under the experimental parameters (Chen et al., 1999; Xing, 2001).
|
Thin-Disc Flow Experiments
The breakthrough concentration (BTC) curves for the atrazine thin-disc flow experiment are illustrated in Fig. 4
. Figure 4A shows the BTC curve for atrazine flowing through the thin-disc apparatus without soil (no-soil control). This treatment provided an experimental control to test the apparatus for retention of atrazine without the influence of soil. The conservative tracer (3H-water) and 14C-atrazine had equal peak height and shape. Both peaks were asymmetric, with a slight tailing. O'Dell et al. (1992) observed very similar conservative tracer peak shapes from undisturbed soil columns. The post-peak 14C-atrazine BTC curve was slightly elevated compared with the 3H-water curve. This was an experimental artifact resulting from error in dual-label counting. As the total counts approached zero, there was a greater tendency to incorrectly count 3H-water disintegrations as 14C-herbicide. This resulted in elevated 14C counts in the later fractions. Using Full Spectrum Dual Label DPM counting and starting with an initial 3H to 14C ratio equal to 10:1 seemed to minimized this error compared with other dual label counting methods or higher 3H to 14C ratios.
|
Figure 4B shows the BTC curves of 3H-water and 14C-atrazine flowing through the experimental apparatus in the presence of soil. The conservative tracer peak-fraction contained approximately 12% of the applied 3H-water. The conservative tracer BTC curve was somewhat asymmetrical. On average, the amount of 3H-water in the leading edge of the curve increased 4.7% of the originally pulsed solution with each 1-mL aliquot (Fig. 5A) . However, after the peak fraction, the amount of 3H-water in each collected fraction decreased by 1.5% (Fig. 5A). This indicates that diffusional mass transfer of 3H-water into and out of the microporous structure of the soil occurred more slowly than the bulk flow of solution through the soil disc (Bouchard, 1999). If the flow rate of solution through the soil disc was reduced, the 3H-water BTC curve would probably appear more symmetrical.
|
Figure 6A illustrates the BTC curves of 3H-water and 14C-imaquin passing through the experimental apparatus in the absence of soil. The BTC curves for 3H-water and 14C-imazaquin peaked at 11% of the applied radioactivity. The BTC curves for 3H-water and 14C-imazaquin in the presence of soil both peaked with slightly more than 14% of the applied radiolabel (Fig. 6B). The 14C-imazaquin BTC curve was similar to the no-soil control and the 3H-water curve with or without soil. This indicates that weakly sorbed imazaquin was not retained when passing through the thin soil disc. However, if the flow rate of the experimental system was reduced, it is more probable imazaquin retention would be observed.
|
|
Figure 7B show the relationship of imazaquin passing through the soil disc relative to the conservative tracer. The slope of the imazaquin passing through the no-soil control and the soil disc equaled 1.0. This indicates that the soil disc did not affect imazaquin retention.
Thin-layer soil chromatography results reported by Goetz et al. (1986) support the observations in our thin-disc flow experiments. In the thin-layer chromatograph experiments, atrazine had a high degree of soil retention, with an average retardation factor (RF) equal to 0.38. Atrazine movement across the soil thin-layer was most retarded by a clay soil (RF = 0.29) and least retarded by a sandy loam soil (RF = 0.49). Imazaquin was highly mobile across the soil layer. The imazaquin RF ranged from 0.83 to 0.90. In our thin-disc flow experiments, a greater amount of atrazine and imazaquin retention may have been observed if the bulk flow of solution through the soil disc was slowed below 1.0 mL min-1. Slowing the bulk flow of solution would allow the diffusional mass transfer of 3H-water and 14C-herbicide into and out of the microporous structure to reach a dynamic state of equilibrium with the bulk flow of solution through the soil disc, and increase the retention of both herbicides.
Figure 8 shows the influence of cumulative flow on atrazine solution and soil concentrations. The peak atrazine solution concentration occurred in the sixth fraction collected, and was 0.540 µM. The peak atrazine soil concentration occurred at the 12th fraction, and was 0.430 µmol kg-1. Because the peak solution concentration was observed 6 mL before the peak soil concentration, it can be concluded that the rate of flow of herbicide solution exceeded the rate of atrazine sorption to the soil surface or diffusional mass flow into soil micropores. These results are supported by the asymmetry of the 3H-water BTC curve in Fig. 4B. If the rate of solution flow through the soil was equal to or less than the rate of sorption to the soil, then the solution and soil concentrations would peak at the same time.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This research also suggests that the thin-disc flow method may provide both rapid partitioning isotherms and sorption kinetics under saturated-flow conditions. The amount of rapid sorption observed in the thin-disc flow method closely agreed with rapid sorption measured by traditional batch techniques. Additionally, the thin-disc flow method can allow for a greater resolution of rapid sorption kinetics that are impractical with batch studies because of the time required for centrifugation. In the future, it would be desirable to use the thin-disc flow method to determine binding kinetics and apply these data to model the movement of a herbicide through a soil profile.
Future experiments should include altering the concentration of pulsed herbicide and flow rate of the experimental system. This would allow for a better understanding of the interaction of residence time and herbicide concentration in a nonequilibrium system. Additionally, this data set could then be mathematically extended to predict the BTC curve of a herbicide passing through a soil column. The model could then be experimentally verified by extending the soil disc to a length of several centimeters.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in JEQ:
This article has been cited by other articles:
|
|
 |
|
  M. C. Smith, D. R. Shaw, J. H. Massey, M. Boyette, and W. Kingery Incubation Time Effects on Imazaquin Desorption as Determined by Nonequilibrium Thin-Soil Disc Flow J. Environ. Qual., March 1, 2004; 33(2): 581 - 593. [Abstract] [Full Text] [PDF]
|
|
| |||||||
and imazaquin
after 5 min and 24 h of incubation. The soil was a Demopolis silt loam, pH 7.6, with 62 g kg-1 organic matter; solution to soil ratio = 2:1 in 5.0 mM CaCl2.
