Published in Vadose Zone Journal 2:728-738 (2003)
© 2003 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
ORIGINAL RESEARCH PAPERS
Miscible Displacement, Sorption and Desorption of Atrazine in a Brazilian Oxisol
Fábio Prataa,
Arquimedes Lavorentib,
Jan Vanderborght*,b,
Peter Burauelb and
Harry Vereeckenb
a Escola Superior de Agricultura "Luiz de Queiroz"/Universidade de São Paulo, Caixa Postal 09, CEP: 13418-900, Piracicaba, São Paulo, Brazil
b Institut für Chemie und Dynamik der Geosphäre, ICG-IV, Forschungszentrum Jülich GmbH, 52425 Jülich, Germany
* Corresponding author (j.vanderborght{at}fz-juelich.de).
Received 14 January 2003.
 |
ABSTRACT
|
|---|
We investigated atrazine [2-chloro-4-(ethylamino)-6-(isopropyl-amino)-s-triazine] leaching, sorption, and desorption in a Brazilian Oxisol under no-till (NT) and conventional (CON) agricultural management. The specific objectives were (i) to infer characteristics of the sorption process (equilibrium vs. nonequilibrium sorption, reversible vs. nonreversible sorption) from breakthrough experiments, (ii) to compare sorption parameters derived from breakthrough and batch experiments, (iii) to predict the distribution of atrazine on reversible and irreversible sorption sites within the soil columns using transport parameters derived from breakthrough data, and (iv) to evaluate the atrazine metabolites during breakthrough experiments. Breakthrough curves of radiolabeled atrazine and of a nonreactive tracer, Br-, were measured during miscible displacement experiments in unsaturated soil columns. The concentration of metabolites in the leachate was determined at three different times. At the end of the displacement experiment, which lasted 494 h, the distributions of atrazine and its metabolites within the soil columns and their partitioning in desorbable, extractable and nonextractable fractions were determined. Atrazine breakthrough curves were fitted with a three-site chemical nonequilibrium convective dispersive transport model considering irreversible sorption. The three-site chemical nonequilibrium model predicted that around 40% of the applied atrazine was irreversibly sorbed at the end of the leaching experiment. This corresponded with sum of the measured extractable and nonextractable fractions. The Kd values for reversible sorption sites derived from the BTCs were similar to those derived from 24-h batch adsorption experiments. However, irreversible sorption was not observed in a four-step consecutive batch desorption experiment that lasted 120 h in total. Differences in desorption between the column and batch experiments were probably due to differences in time-scale and sorption-desorption conditions. More than 90% of the radioactivity in the leachate and 80% remaining in the soil column, respectively, were characterized as atrazine, indicating a slow decay. Hydroxyatrazine was the most important metabolite of atrazine and showed higher retention in the column than the other metabolites.
Abbreviations: BTC, breakthrough curve • CDE, convective–dispersive equation • CON, conventional management • DEA, deethylatrazine • DIA, deisopropylatrazine • HA, hydroxyatrazine • LSC, liquid scintillation counter • NT, no-till • TLC, thin layer cochromatography
|
INTRODUCTION
|
|---|
IN THE LAST TWO DECADES, the world-wide pesticide use in agriculture and, as a consequence, the positive analytical findings of these pesticides in shallow groundwater have increased. The movement of pesticides in soils depends on the physical and chemical properties of the compound, the type of soil (including organic matter, physical and chemical composition, pH), and the climatic conditions (Hornsby et al., 1995). Soil-applied herbicides that are highly mobile in the soil may fail to control weeds due to dilution, may damage desirable vegetation by moving to off-site areas, and may contaminate drinkable water sources. Conversely, persistent nonselective herbicides that are immobile in soils can cause problems with carryover. Therefore, the mobility of herbicides in soils is both of agronomic and of environmental concern. Scientists have concentrated their efforts on the elucidation of herbicide mobility in different studies on sorption, dissipation, and leaching of pesticides using soil samples in the laboratory or under field conditions.
Atrazine is a herbicide widely used in Brazil to control broadleaf weeds in corn (Zea mays L.) and sorghum [Sorghum bicolor (L.) Moench] under both NT and CON agricultural managements. It is applied in preemergence treatment and inhibits the Hill reaction in photosynthesis (Rodrigues and Almeida, 1995).
Atrazine and its dealkylated metabolites deethylatrazine (DEA) and deisopropylatrazine (DIA) have been detected in surface and groundwater (Thurman et al., 1991, 1994; Cai et al., 1994; Lerch et al., 1998). To date there are no data for Brazilian conditions. The principal causes for finding of atrazine in streams and shallow groundwater are its widespread use in conjunction with its moderate persistence in soil (Mersie and Seybold, 1996).
Several studies of atrazine behavior in soils have shown a wide range in the sorption partition coefficient (Kd), which was found to be proportional to the soil organic C content (e.g., Mersie and Seybold, 1996; Moreau and Mouvet, 1997; Moorman et al., 2001). In the database compiled by Wauchope et al. (1992), an average Koc (Koc = Kd/organic C% x100) of 100 mL g-1 and a range of Koc values from 38 to 174 mL g-1 is listed for atrazine. In addition, the soil organic matter has been related to atrazine metabolism. Atrazine metabolization to hydroxyatrazine (HA), its major metabolite in most surface soils (Lerch et al., 1999), occurs mainly by chemical hydrolysis (Miller et al., 1997), and is catalyzed by sorption to organic matter at low soil pH (Lerch et al., 1998). For instance, Martin-Neto et al. (2001) showed that proton transfer between atrazine and soil humic substances can transform atrazine to HA and also increase its retention. Therefore, soil organic matter can contribute in various ways to the fate of atrazine in soils.
In addition to the partition coefficient, which describes the equilibrium distribution of the pesticide between sorbed and dissolved states, the rate of achieving this equilibrium also determines the fate of pesticides. The sorption of pesticides can be divided into a fast and a slow phase (Gamble et al., 2000). The fast sorption phase includes surface processes, whereas the slow phase is related to diffusion into and out of humic substances. This very slow diffusion process also leads to a slow release of pesticides that are sorbed in the interior of the humic matrix back into the soil solution. In some cases, this release is so slow that pesticides that are sorbed in the humic matrix are considered to be irreversibly sorbed. This apparently irreversibly sorbed fraction is often called the "bound residue" fraction. However, the concept of bound residue is still under discussion. Führ et al. (1998) defined bound residues as follows: "Bound residues represent compounds in soil, plant or animal which persist in the matrix in the form of the parent substance or its metabolite(s) after extraction. The extraction method must not substantially change the compounds themselves or the structure of the matrix." Irreversible sorption, or bound residue formation, plays an important role in the dissipation of hydrophobic pesticides in soil (Barriuso and Koskinen, 1996). Nakagawa et al. (1995) showed that around 50% of atrazine applied in an organic Brazilian soil was found in a bound residue form 180 d after the application. Similar results were observed by Barriuso et al. (1997) for bound residue formation of atrazine in soil after 8 mo incubation.
In addition to batch and incubation experiments, miscible displacement experiments in soil columns have been performed to infer distribution constants and sorption rates from breakthrough curves (BTCs) (Ma and Selim, 1996; Meyer-Windel et al., 1999). These experiments can be performed in undisturbed soil samples or in columns packed with disturbed soil. Using the latter, the effect of the soil chemical properties on solute movement is emphasized.
Several mathematical models, based on the convective–dispersive equation (CDE) and a kinetic sorption model, were used to model the observed BTCs (Ma and Selim, 1996). For atrazine transport, a two-site first-order rate constant (e.g., Gamerdinger et al., 1990; Gaber et al., 1995) and 
distributed sorption rate constants (Chen and Wagenet, 1997) were considered to describe the sorption kinetics. Although these models consider a rate-limited sorption and desorption, they do not consider a bound residue fraction. Beigel and Di Pietro (1999) modeled the formation of a bound residue fraction as a first-order decay process. To distinguish the formation of a bound residue fraction from real decay or transformation, the atrazine remaining in the soil column after the leaching experiment must be analyzed. Studies of atrazine displacement in soil columns with a detailed distribution of the remaining concentration, and comparisons of this internal distribution with that predicted by mathematical models are practically nonexistent in literature, particularly for Brazilian soils.
The main objective of this research was to study atrazine leaching, sorption, and desorption in a Brazilian Oxisol, which had been maintained under NT or CON agricultural management for 25 yr. Oxisols are the most important soils in the tropics. In 2000, about 13.5 million hectares were cultivated in Brazil using a NT management system, which represented almost 35% of the area cultivated with soy [Glycine max (L.) Merr.], corn, and cereals. The specific objectives were (i) to infer characteristics of the sorption process (equilibrium vs. nonequilibrium sorption; reversible vs. nonreversible sorption) from atrazine breakthrough data using transport model fits, (ii) to compare sorption parameters derived from breakthrough data with those derived from batch sorption experiments, and (iii) to predict the atrazine distribution within the soil columns, using parameters that were derived from breakthrough data. To distinguish irreversible sorption from decay, atrazine metabolites were analyzed in the leachate and in the soil column at the end of the leaching experiment.
|
MATERIALS AND METHODS
|
|---|
Atrazine sorption and desorption batch experiments, as well as Br- and atrazine miscible displacement experiments in disturbed soil columns were performed in samples of a Brazilian loamy Oxisol in the Paraná region. The soil is classified as a Rhodic Hapludox and had been maintained under NT or CON agricultural management for 25 yr.
Soil
For all analyses, soil samples were collected between the 0- and 20-cm depth, air dried, and sieved using a 2-mm mesh. Soil samples were analyzed for various chemical, physical, and mineralogical properties that are considered to be important for the sorption and mobility of pesticides (Table 1). The chemical analyses were performed according to van Raij and Quaggio (1983). Particle size analyses were performed following the pipette method (Day, 1965). Total Fe (Fe2O3) and Al (Al2O3) were extracted by H2SO4 (18 mol L-1) (Vettori, 1969). The predominant clay minerals were identified by X-rays diffraction (Jackson, 1969).
View this table:
[in this window]
[in a new window]
|
Table 1. Chemical characteristics and physical and mineralogical properties of a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural soil systems.
|
|
Chemical and Standards
Atrazine (Fig. 1) was available as both radiolabeled and nonlabeled compound, with an analytical purity >97%. Carbon-14 atrazine had a specific activity of 1.786 MBq mg-1 and the 14C radiolabeling positions were uniform in the ring. For the preparation of stock solutions, atrazine and its standard metabolites, DEA, and DIA were dissolved in acetone, and HA in 0.1 mol L-1 HCl. Only the atrazine was radiolabeled and applied later to the sorption batches and soil columns. The stock solutions of metabolites were used for the thin layer cochromatography (TLC). Water solubilities are 33 mg L-1 for atrazine, 2700 mg L-1 for DEA, 980 mg L-1 for DIA, and 16 mg L-1 for HA (Seybold and Mersie, 1996). The pKa for atrazine is 1.7, DEA is 1.4, DIA is 1.5, and HA is 5.2 (Seybold and Mersie, 1996).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 1. Molecular structure and some physicochemical properties of atrazine [2-chloro-4-(ethylamino)-6-(isopropylamino)-s-triazine]. * denotes position of the 14C.
|
|
Sorption and Desorption Batch Experiments
Five concentrations of atrazine isotopic mixture (14C atrazine with radioactivity of 0.24 kBq mL-1 and 12C atrazine) were used in the batch experiments: 0.14, 0.25, 0.36, 0.68, and 2.38 mg L-1, each with three replicates. Two grams of dried soil was mixed with 10 mL of atrazine solution and shaken for 24 h in 25-mL glass centrifuge tubes. The tubes were then centrifuged, and the supernatant was decanted. The radioactivity in the equilibrium solution was determined in aliquots of 1.0 mL, using a liquid scintillation counter (LSC) (Packard 2500 TR, Hewlett-Packard Analytical, Wilmington, DE). Desorption was performed with soil from the 0.68 mg L-1 sorption batch. The batch was decanted and weighted to determine the volume of solution (around 0.7 mL) and the total mass of atrazine remaining in the batch. Subsequently, 10 mL of atrazine-free 0.01 mol L-1 CaCl2 solution was added and the centrifuge tube was shaken for 24 h, centrifuged, and decanted. The atrazine concentration was measured in a 1-mL aliquot taken from the decanted solution using a LSC. The sorbed atrazine concentration was calculated from the difference between the total atrazine mass in the batch and in the solution. The decanted batch was used for a consecutive desorption step, of which, in total, four were made.
Results from sorption and desorption experiment were fitted with the Freundlich equation:
 | [1] |
where S is the sorbed herbicide concentration (M M-1), C is the herbicide concentration in solution (M L-3), Kf is the Freundlich constant (M M-1) (L3 M-1)N, and N is the degree of linearization of the isotherm. Values of the linear partition coefficient Kd (L3 M-1) were obtained from the average ratio S/C for all points of the experimental isotherm.
Column Displacements Experiment
Breakthrough curves and the distribution of the resident atrazine concentration in the soil were measured during and after column displacement experiments, respectively. For each treatment (NT and CON agricultural management), soil was packed into three glass columns (40 cm long, 5.1-cm diam.) having a conical end at the bottom. Glass wool was placed at the bottom of the column to hold sterilized sand (treated with HCl and high temperature) that filled the 1.5-cm-long cone of the column and supported the packed soil. A long-beak funnel was used to continuously add the dried and sieved soil, and a rubber hammer was used to pack the soil uniformly. Columns were filled with soil up to a depth of 20 cm. The columns were irrigated using a peristaltic pump. To prevent splashing and to promote a uniform distribution of the water, a glass wool plug was placed on the top of the soil. No ponding occurred so the soil columns were not saturated at the top surface. Since no suction was applied, the columns were saturated at the bottom. The average volumetric water content in the columns was determined by weighing. In further analysis, it was assumed that the moisture content profile was uniform in the soil column. The effect of a nonuniform water content profile on solute transport is discussed by Flury et al. (1999).
The columns were first irrigated for 2 d with distilled water to establish a steady-state flow condition. Then, a KBr solution was applied in a short pulse and leached for 84 h at a steady flow rate. After 84 h, atrazine (radiolabeled and nonlabeled) was applied also in a short pulse. The atrazine solution, of which 1.5 mL was applied to a column, was prepared by mixing 0.1 mL stock solution with 10 mL distilled water. Thereafter, columns were further leached with distilled water using the same irrigation rate. The steady-state flow rate was maintained throughout the experiment by controlling the leachate volumes. Details of experimental conditions are shown in Table 2.
Column leachate was collected manually in a 250-mL Erlenmeyer flask at 2-h intervals until 84 h, and afterwards at 12-h intervals (after addition of atrazine). Until 84 h, at each interval, the total amount of the leachate was collected and bromide was quantified using a Dionex 4000i with CMD-I ion chromatograph (Dionex, Sunnyvale, CA), where the separating precolumn and column were IonPac AG9-HC and IonPac AS9-HC (Dionex, Idstein, Germany), respectively. A 9 mmol L-1 isocratic Na2CO3 solution at 1 mL min-1 at 25°C was used as eluant. Bromide was detected by electrical conductivity. Retention time was 11.6 min and the detection limit was 0.05 mg L-1. After the application of atrazine, two replicates of the leachate (10 mL) were taken at each 12-h interval, and added to 10 mL of an appropriate scintillation solution (Insta-Gel XF) to measure the leachate radioactive content using a LSC (Packard 2500 TR). The chemical nature of the radioactivity in the leachate collected at 192, 248, and 300 h after atrazine addition was determined using TLC after concentration of the leachate in a rotary evaporator at 40°C. The solvent mixture used for separation of the compounds applied to the TLC plate (Al, Silicagel 60 F254) consisted of chloroform, acetone, acetic acid, and water (50:30:15:5). Radiobioimage (Fujix BAS 1000) was used to detect the radioactive compounds on the TLC plates.
After the leaching experiments, pressurized air was used to remove the soil samples from the columns. During this procedure, soils were sectioned into 5-cm layers. For each layer, a desorption was performed using 0.01 mol L-1 CaCl2 solution (solution/moist soil ratio = 2:1) followed by two extractions using acetonitrile–water solution (4:1) (solution/moist soil ratio = 2:1). The extracted soil samples were then used for the determination of soil-bound residues. For both desorption and extraction, the systems were shaken at 200 rpm for 2 h using a horizontal shaker, and subsequently centrifuged at 10 000 g for 15 min. Duplicate aliquots of the supernatant (10 mL) were used to determine the desorbed and acetonitrile extractable radioactivity, respectively. Finally, soil samples were air dried, pulverized in a porcelain grinder, and triplicate 0.5-g subsamples were taken for combustion in a biological oxidizer to determine the amount of bound residues at different soil depths.
Theory and Data Analyses
Equilibrium Convective–Dispersive Model
One-dimensional convective–dispersive solute transport with linear equilibrium sorption and without degradation of the target solute can be described by the CDE, which is written for steady-state flow in a homogeneous soil of uniform water content as
 | [2] |
where C is the solute concentration in the liquid phase (M L-3), t is time (T), z is distance (L), R is the dimensionless retardation factor accounting for linear equilibrium sorption of a solute (R = 1 + 
Kd
-1), is the soil bulk density (M L-3), is the volumetric water content (L3 L-3), Kd is the linear partition coefficient (L3 M-1), D is the dispersion coefficient (L2 T-1), and v is the pore water velocity (L T-1) derived from the flux rate or Darcy flux Jw (L T-1), v = Jw -1.
Nonequilibrium Convective–Dispersive Model
In the nonequilibrium convective–dispersive model, the sorption sites are divided into two types: instantaneous and rate-limited sorption sites. When the sorption rate is described by a first-order process (e.g., van Genuchten and Wierenga, 1976), the following two equations are used to simulate chemical nonequilibrium solute transport:
 | [3a] |
 | [3b] |
where 
is a first-order kinetic rate coefficient of mass transfer between sorption sites and soil solution (T-1), f is the fraction of sorption sites that are always in equilibrium, Sk is the sorbed concentration at the rate-limited sites (M M-1), and µl (T-1) is a first-order degradation rate constant.
The sorption process that is represented in Eq. [2] and [3] assumes that the sorption isotherm is single valued and the sorption is a reversible process; desorption occurs when (1-f) KdC < Sk. Hysteresis of a sorption process and irreversible sorption may be attributed to a very slow desorption process as compared with the sorption process (Altfelder et al., 2000). When we describe the sorption or desorption of a compound X to or from an "irreversible" sorption site (Sirr) as a chemical reaction,
 | [4] |
where X is the concentration of compound X, Sirr of the free irreversible sorption sites, SirrX* of sorbed compound X, with X, Sirr, and SirrX* expressed as mass equivalents of compound X per volume of solution, k1 is the sorption rate constant, and k2 is the desorption rate constant. Since we consider a second-order sorption process, the sorption rate constant k1 has the dimension (concentration unit T)-1 whereas the dimension of k2 is T-1. The change of SirrX* with time is then given by
 | [5] |
When the desorption is much slower than the sorption, and the time scale of the desorption process is much larger than the time scale of the experiment, the sorption process can be described as an irreversible process (i.e., k2 = 0). When we further assume the number of available sorption sites are in excess (X << Sirr), and its concentration does not change with time, the formation of "irreversibly" sorbed X (SirrX*) is described as a first-order decay of X. As a consequence, the first-order degradation constant µl in Eq. [3] can describe an irreversible sorption process with µl = k1Sirr. Note that since k1 has dimension (concentration unit T)-1 and Sirr has the dimension of concentration unit, that is, the mass of compound X that can be sorbed to free irreversible sorption sites per volume of solution, the product k1Sirr has dimension T-1. Therefore, Eq. [3] can be used to describe a sorption process to three different types of sorption sites: equilibrium and nonequilibrium reversible sorption sites and irreversible sorption sites. The first-order rate parameter µl lumps the mass loss due to decay and irreversible sorption. The two processes cannot be distinguished based solely on concentration data. Measurements of metabolite concentrations, indirect measurements of irreversibly sorbed compound based on radioactivity measurements in combusted sediment samples, and chemical analysis of the extracted radioactivity are required to discern irreversible sorption from decay.
The CXTFIT 2.0 code (Toride et al., 1995) was employed to fit the BTCs and to estimate transport parameters for atrazine and Br-, as well as to predict the depth distribution and partitioning in nonsorbed, reversibly, and irreversibly sorbed fractions of the atrazine resident in the soil column at the end of the leaching experiment. Bromide BTCs were fitted with the equilibrium CDE (Eq. [2]) with R = 1. The hydraulic parameters D and v, obtained from Br- fits, were kept constant when estimating the parameters of the three-site chemical nonequilibrium model with irreversible sorption from the atrazine breakthrough.
Since the fitted effluent concentrations represent flux-averaged concentrations, the solution of Eq. [2] and [3] for a first-type boundary condition were used:
 | [6a] |
 | [6b] |
where C0 (M L-3) is the input concentration.
The atrazine concentration in the soil solution (C), the sorbed mass to the equilibrium sorption sites (fKdC), and the sorbed mass to the nonequilibrium sorption sites (Sk), were predicted using the solution of Eq. [3] for a third type boundary condition:
 | [7a] |
 | [7b] |
The input concentration, C0, and application time, t0, are given in Table 2.
When dispersion is caused by microscopic variability of advection velocities, transport within the column should remain unaffected by a downstream boundary, except for its effect on the water flow itself. As a consequence, the solution of the CDE in a semi-infinite soil column is more suited to describe concentrations in the column and in the effluent (van Genuchten and Parker, 1984). The bottom boundary of a semi-infinite column is
 | [8] |
The mass of atrazine in the 5-cm-thick soil slices in soil solution and sorbed to equilibrium and nonequilibrium sites (reversible sorption) was predicted from the depth-averaged C, fKdC, and Sk. Assuming that there was no decay, the mass of atrazine bound to irreversible sorption sites in a certain soil slice was calculated from a mass balance, using predicted solute fluxes in and out of the slice and the predicted mass remaining in the slice in the soil solution and sorbed to reversible sorption sites.
|
RESULTS AND DISCUSSION
|
|---|
Sorption and Desorption Batch Experiments
Percentage of atrazine sorbed and desorbed, the linear partition coefficient (Kd), and the Freundlich parameters for the sorption and desorption of atrazine were of similar magnitude in both NT and CON management soil system (Table 3). The sorption and desorption isotherms are shown in Fig. 2. The percentage of atrazine sorbed in both tillage systems was around 43%. The Nsor values were 0.85 and 0.84 for NT and CON, respectively, showing nonlinear isotherms for atrazine sorption at the considered concentration range. The Kd and Koc values are in good agreement with the literature (Clay and Koskinen, 1990; Mersie and Seybold, 1996; Moreau and Mouvet, 1997; Mersie et al., 1999) and suggest a medium sorption when compared with the sorption values for molecules that sorb strongly such as glyphosate (Prata et al., 2000), organchlorines (Roberts and Hutson, 1999), or readily leached molecules such as picloran or aldicarb (Roberts et al., 1998; Roberts and Hutson, 1999).
View this table:
[in this window]
[in a new window]
|
Table 3. Sorption and desorption, and parameters of the sorption and desorption isotherms of atrazine in a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems.
|
|
View larger version (9K):
[in this window]
[in a new window]
|
Fig. 2. Sorption (circles) and desorption isotherms (diamonds) of atrazine in a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems. Lines are fits of the Freundlich model to the sorption (dashed line) and desorption (full line) isotherms.
|
|
Almost all of the amount of atrazine sorbed could be desorbed (around 91%) at the end of the desorption experiment independently of the tillage system, as is predicted by the low Freundlich coefficients for desorption (Table 3). However, even with this high desorption rate, the fact that Ndes < Nsor indicates hysteretic desorption; that is, sorption is not completely reversible and requires less energy than desorption. Possible causes include the different types of bonding involved in s-triazine sorption by soil organic matter, such as hydrogen bonds, charge transfers, ionic bonds, cation bridges, hydrophobic interactions, and physical diffusion into the humic substances (Choudhry, 1983; Senesi, 1992; Moreau and Mouvet, 1997; Martin-Neto et al., 2001). The hysteresis of the atrazine ad- and desorption in this Oxisol, however, is small (Fig. 2) compared with the hysteresis observed in some other soil and sediment types (e.g., Fig. 1 in Moorman et al., 2001; Fig. 1 in Seybold and Mersie, 1996). A large atrazine recovery in batch desorption experiments was also observed by Moreau and Mouvet (1997) for soil and aquifer sediments and by Moorman et al. (2001) in soils with relatively high C contents. Whereas a high recovery rate in a desorption experiment is an indication of a reversible sorption process, the opposite is not true. Low recovery rates may as well be the result of a combination of a high solid/liquid ratio, a large sorption, a limited number of sequential desorption steps, and a large fraction of solution remaining in the batch between two desorption steps. Assuming nonhysteretic ad- and desorption and equilibrium conditions, we calculate that only 40% of the initially sorbed mass would be desorbed after four desorption steps if we performed the desorption experiment using a solid/liquid ratio of 1:1 and a 60% replacement of the solution between two desorption steps.
Column Displacement Experiment
In this section, we assume that the detected radioactivity comes from the radiolabeled atrazine, and we neglect atrazine metabolites. An analysis of the chemical nature of the radioactivity, which is presented in the next section, confirms this assumption. The atrazine BTCs were very similar for the two agricultural managements, showing a tailing phenomenon (Fig. 3), which is characteristic for reactive solutes and is generally a consequence of the kinetic component of sorption (Kookana et al., 1993). Since the equilibrium CDE was able to fit the bromide BTCs (Fig. 3), suggesting a physical equilibrium in the soil column, the tail of atrazine BTCs is explained rather by chemical nonequilibrium sorption. The tail of atrazine BTCs was relatively well described by the three-site nonequilibrium model. To fit the mass leached from the soil columns, a first-order decay of atrazine in the liquid phase due to irreversible sorption to a third type of sorption sites had to be incorporated in the model. The fitted model parameters are listed in Table 4. The mass recovered of 14C for both NT and CON managements, as well as mass balance predicted by three-site chemical nonequilibrium model, are presented in Table 5.
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 3. Measured and fitted breakthrough curves of Br- and atrazine in a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems.
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. Fitted parameters of the equilibrium convective–dispersive equation (Eq. [2]) to Br- BTCs and of the three-site nonequilibrium with irreversible sorption model (Eq. [3]) to atrazine breakthrough curves for a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems.
|
|
View this table:
[in this window]
[in a new window]
|
Table 5. Measured and predicted mass balances of 14C atrazine in column experiments using a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems.
|
|
The effect of the NT system on atrazine sorption has already been shown in the literature (Sadeghi et al., 1998; Shelton et al., 1998) and is related with the quantitative and qualitative nature of the soil organic matter. However, the differences between NT and CON systems observed in our study are relatively small. Furthermore, we did not observe significant differences in the sorption of atrazine for NT and CON in the batch experiments (Table 3 and Fig. 2); neither did we observe differences in the fitted sorption parameters for both management systems (Table 4). This suggest the different management systems did not lead to dramatic changes in the chemical properties that are relevant for atrazine leaching (Table 1). Otherwise, since we packed the soil cores with disturbed soil samples, the effect of the soil structure on the atrazine leaching was not considered. However, soil structure is a very important parameter that may override the effect of chemical soil properties on leaching (e.g., Vanderborght et al., 2002). The effect of the management system on soil structure and pesticide leaching requires further investigation.
Radioactivity resident in the columns, expressed as percentage of the total radioactivity applied, was highest at the top and at a depth between 15 and 20 cm for both NT and CON management systems, showing an increment from the 5- to 10-cm to the 15- to 20-cm depths (Fig. 4). The depth distribution of the total radioactivity within the columns was relatively well reproduced by the three-site nonequilibrium model that was calibrated to the atrazine breakthrough.
View larger version (43K):
[in this window]
[in a new window]
|
Fig. 4. Measured and predicted distribution of the radioactivity in the different fractions in the soil columns, packed with a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems, after leaching experiments with 14C atrazine. 100% of radioactivity corresponds with the total applied radioactivity. (Horizontal lines below the bars represent the standard deviation of the total resident radioactivity in the three replicate soil columns.)
|
|
The contribution of the desorbed phase in the total mass balance at the end of the leaching experiment was only 2.86 and 2.46% for NT and CON, respectively. This desorbed phase is the fraction of herbicide displaced with an aqueous solution with ionic strength similar to that of the soil solution and is theoretically equivalent to the exchangeable fraction of the total sorbed herbicide. The radioactivity extractable with acetonitrile represented 17.87 and 14.55% of the total applied in NT and CON, respectively. These high values relative to the percentage desorbed are probably due to the high solubility of atrazine and its metabolites in this solvent.
The predicted amount of atrazine that remained in the columns at the end of the leaching experiment but was still available for leaching (i.e., the sum of the mass sorbed to the equilibrium and kinetic sorption sites and in solution) was only a few percentages of the applied atrazine, like the measured desorbed atrazine fraction (Table 5 and Fig. 4). The Kd values derived from the leaching experiment (Table 4) were very close to Kd values derived from the batch experiments (Table 3). The small difference suggests that the sorption isotherm could be approximated by a linear isotherm for the atrazine concentration in the column experiment (up to 0.4 mg L-1). The most notable contrast between the batch and column experiments is related to desorption: 90% of the atrazine was desorbed for both management systems in the batch experiment (Table 3), whereas the sum of the leached and desorbed fraction at the end of the leaching experiment was only 53% for NT and 65% for CON in the column experiment (Table 5). Although the same solution had been used to desorb atrazine in both experiments (0.01 mol L-1 CaCl2), the contact time between the herbicide and soil in the column experiment was much higher than in the batch studies. This might have contributed to a higher physical diffusion into humic substances, apparently increasing the amount of atrazine irreversibly sorbed. However, the fitted rate parameters suggest that irreversible sorption occurs as fast or even faster than sorption to rate-limited reversible sorption sites ( < µ1). This implies that either the physical and chemical conditions in the batch experiments were less favorable for irreversible sorption than in the column experiments, or that the kinetics of irreversible sorption are not completely described as a first-order rate process.
Predicted amounts of irreversibly sorbed atrazine were 2.07 and 1.74 times higher than measured bound residue for NT and CON, respectively (Table 5). However, the predicted amounts of irreversibly sorbed atrazine corresponded fairly well with the sum of the measured extracted and bound residue fractions (Table 5, Fig. 4). According to the definition of bound residues given by Führ et al. (1998), the extraction method that is used to determine the bound residue fraction characterizes a fraction that might be remobilized. However, it does not necessarily represent the fraction that can be desorbed during a leaching experiment or under natural field conditions. However, predictions of irreversibly sorbed atrazine in the soil columns by the three-site chemical nonequilibrium model, which was calibrated to breakthrough data, indicated that both the acetonitrile extracted fraction and the bound residue fraction constitute the irreversibly sorbed fraction during the leaching experiment. Therefore, the leaching experiment does not provide evidence of a difference in leaching potential between the acetonitrile extracted fraction and the bound residue fraction.
Atrazine Metabolites
Figure 5 shows the average percentages of the radioactivity characterized as atrazine and its metabolites in the leachate collected at 192, 248, and 300 h after the atrazine application. The percentage of atrazine metabolites in the leachate was very low (Fig. 5) and almost the total amount of radioactivity in the leachate when the major breakthrough of radioactivity occurred was original compound (94.7% in NT, and 92.9% in CON). The metabolites identified were HA, DEA, and DIA. However, values for the total percentages of atrazine resident inside the column were only 83.52% (NT) and 76.30% (CON) for desorbed radioactivity, and 82.96% (NT) and 81.78% (CON) for extracted radioactivity (Fig. 6), suggesting a higher mobility for atrazine than for its metabolites.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5. Relative amounts of radioactivity characterized as atrazine and its metabolites in the leachate from soil columns packed with a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems. 100% of radioactivity corresponds with the total radioactivity in the leachate. (ATZ, atrazine; HA, hydroxyatrazine; DEA, deethylatrazine; DIA, deisopropylatrazine. Vertical lines represent the standard deviation of the percentages in the three replicates.)
|
|
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 6. Relative amounts of radioactivity characterized as atrazine and its metabolites in the desorbed and extracted fraction at different depths in the soil columns packed with a Brazilian Rhodic Oxisol collected from no-till or conventional agricultural management systems. 100% radioactivity corresponds with the total radioactivity detected in a certain fraction at a certain depth. Horizontal lines represent the standard deviation in the three replicates.
|
|
Distribution and characterization of the desorbable and extractable radioactivity in the column are shown in Fig. 6. With the exception of the desorbable fraction in NT, we observed a higher amount of HA than DEA and DIA, a reduction in HA concentration, and an increase in the concentration of two other metabolites with depth for all treatments. This indicates that HA was the metabolite with the highest sorption in the columns. Several authors also observed higher retention for HA than DEA and DIA (Clay and Koskinen, 1990; Brouwer et al., 1990; Mersie and Seybold, 1996; Lerch et al., 1999). This has been attributed to a lower water solubility and major thermodynamic stability of HA (Erickson and Lee, 1989; Lerch et al., 1999).
The metabolite HA, the major metabolite of atrazine in most surface soils (Lerch et al., 1999), is mainly formed by chemical hydrolysis (Miller et al., 1997), which is catalyzed by sorption processes and low soil pH values (Lerch et al., 1998). In fact, the soils contained a considerable organic C content, which probably contributed in the experiments to the higher metabolization of atrazine in HA. In addition, HA has been shown to sorb to soils by pH-dependent cation exchange and hydrophobic interactions (Lerch et al., 1997), which suggests that part of bound residue fraction could be HA. Martin-Neto et al. (2001) studied the sorption mechanisms for atrazine in humic substances and observed the occurrence of proton transfer. This sorption mechanism can promote the Cl atom displacement of the atrazine molecule, which is metabolized to sorbed HA.
|
CONCLUSIONS
|
|---|
The three-site chemical nonequilibrium model with irreversible sorption successfully fitted the atrazine BTCs. The irreversible sorption of atrazine was modeled as a first-order rate decay process. The existence of an irreversibly sorbed fraction was proven using radioactive labeled atrazine, which was found back on the solid phase. The predicted irreversibly sorbed atrazine that remained in the soil column at the end of the leaching experiment amounted to 40% of the applied atrazine and corresponded well with the sum of the measured extracted and bound residue fractions.
The notion "irreversible" is relative and refers to the atrazine fraction that could not be desorbed during the time scale of the leaching experiment. Whether the fractionation of this "irreversibly" sorbed fraction using an organic solvent into potentially mobile and a bound residue fractions is relevant for the long-term mobility of atrazine could not be confirmed by these experimental data. To predict the impact of long-term application of atrazine on its potential leaching, the process of irreversible sorption must be better understood. Our modeling approach is based on the assumption that irreversible sorption sites are in excess, an assumption that might not hold for long-term application when atrazine accumulates.
Differences in desorption between the leaching experiment, in which about 60% of the applied atrazine was desorbed, and batch experiment, in which 90% was desorbed, suggest that the equilibrium sorption time and the sorption–desorption conditions are very important. The breakthrough data from the column experiment suggested an irreversible sorption rate, which was sufficiently high so that also within the time frame of the batch experiment atrazine could be irreversibly sorbed.
In our modeling approach, we did not consider decay of atrazine. Since 90% of the radioactivity in the leachate and 80% in the desorbed and extracted fractions was characterized as atrazine, the decay of atrazine in the leaching experiment was low. The higher percentage of metabolites in the desorbed and extracted fractions than in the leachate suggests a smaller mobility for the metabolites than for atrazine. Hydroxyatrazine was the most important metabolite of atrazine. Unlike the distribution of the other atrazine metabolites DEA and DIA, the HA concentration decreased with depth. This points at a higher retention of HA than of the other metabolites. Possibly, a significant part of the atrazine bound residue could be HA.
In contrast to other studies, we did not observe a different sorption and leaching behavior of atrazine in soil under CON and NT management. Since we used repacked soil columns, in which the effect of soil structure on leaching is eliminated, this was due to similar chemical properties of CON and NT soil. To account for the effect of management on soil structure, and as a consequence on pesticide leaching, transport experiments in undisturbed soil cores need to be performed.
|
ACKNOWLEDGMENTS
|
|---|
We thank FAPESP for a scholarship to the first author and Dr. Kilian Smith for his English review of the manuscript.
|
REFERENCES
|
|---|
- Altfelder, S., T. Streck, and J. Richter. 2000. Nonsingular sorption of organic compounds in soil: The role of slow kinetics. J. Environ. Qual. 29:917–925.[Abstract/Free Full Text]
- Barriuso, E., S. Houot, and C. Serra-Wittling. 1997. Influence of compost addition to soil on the behavior of herbicides. Pestic. Sci. 49:65–75.
- Barriuso, E., and W.C. Koskinen. 1996. Incorporating nonextractable atrazine residues into soil size fractions as a function of time. Soil Sci. Soc. Am. J. 60:150–157.[Abstract/Free Full Text]
- Beigel, C., and L. Di Pietro. 1999. Transport of triticonazole in homogeneous soil columns: Influence of non-equilibrium sorption. Soil Sci. Soc. Am. J. 63:1077–1086.[Abstract/Free Full Text]
- Brouwer, W.W., J.J.T.I. Boesten, and W.G. Siegers. 1990. Adsorption of transformation products of atrazine by soil. Weed Res. 30:123–128.
- Cai, Z., V.M.S. Ramanujam, M.L. Gross, S.J. Monson, D.A. Cassada, and R.F. Spalding. 1994. Liquid-solid extraction and fast atom bombardment high-resolution mass spectrometry for the determination of hydroxyatrazine in water at low-ppt levels. Anal. Chem. 66:4202–4209.
- Chen, W., and R.J. Wagenet. 1997. Description of atrazine transport in soil with heterogeneous nonequilibrium sorption. Soil Sci. Soc. Am. J. 61:360–371.[Abstract/Free Full Text]
- Clay, S.A., and W.C. Koskinen. 1990. Adsorption and desorption of atrazine, hydroxyatrazine, and s-glutathione atrazine on two soils. Weed Sci. 38:262–266.
- Choudhry, G.G. 1983. Humic substances: Sorptive interactions with environmental chemicals. Toxicol. Environ. Chem. 6:127–171.
- Day, P.R. 1965. Particle fractionation and particle size analysis. p. 545–567. In C.A. Black (ed.) Methods of soil analysis. Agron. Monogr. 9. ASA, Madison, WI.
- Erickson, L.E., and K.H. Lee. 1989. Degradation of atrazine and related s-triazines. Crit. Rev. Environ. Control 19:1–14.
- Flury, M., M.V. Yates, and W.A. Jury. 1999. Numerical analysis of the effect of the lower boundary condition on solute transport in lysimeters. Soil Sci. Soc. Am. J. 63:1493–1499.[Abstract/Free Full Text]
- Führ, F., H. Ophoff, P. Burauel, U. Wanner, and K. Haider. 1998. Modification of the definition of bound residues. p. 175–176. In F. Führ and H. Ophoff. (ed.) Pesticide bound residues in soil. Wiley-VCH, Weinheim, Germany.
- Gaber, H.M., W.P. Inskeep, S.D. Comfort, and J.M. Wraith. 1995. Nonequilibrium transport of atrazine through large intact soil cores. Soil Sci. Soc. Am. J. 59:60–67.[Abstract/Free Full Text]
- Gamble, D.S., A.G. Bruccoleri, E. Lindsay, C.H. Langford, and G.A. Leys. 2000. Chlorotalonil in a quartz sand soil: Speciation and kinetics. Environ. Sci. Technol. 34:120–124.
- Gamerdinger, A.P., R.J. Wagenet, and M.Th. van Genuchten. 1990. Application of two-site/two-region models for studying simultaneous nonequilibrium transport and degradation of pesticides. Soil Sci. Soc. Am. J. 54:957–963.[Abstract/Free Full Text]
- Jackson, M.L. 1969. Soil chemical analysis. Advanced course. M.L. Jackson, Madison, WI.
- Hornsby, A.G., R.D. Wauchoupe, and A.E. Herner. 1995. Pesticide properties in the environment. Springer-Verlag, New York.
- Kookana, R.S., R.D. Schuller and L.A.G. Aylmore. 1993. Simulation of simazine transport through soil columns using time-dependent sorption data measured under flow conditions. J. Contam. Hydrol. 14:93–115.
- Lerch, R.N., P.E. Blanchard, and E.M. Thurman. 1998. Contribution of hydroxylated atrazine degradation products to the total atrazine load in Midwestern streams. Environ. Sci. Technol. 32:40–48.
- Lerch, R.N., E.M. Thurman, and P.E. Blanchard. 1999. Hydroxyatrazine in soils and sediments. Environ. Toxicol. Chem. 18:2161–2168.
- Lerch, R.N., E.M. Thurman, and E.L. Kruger. 1997. Mixed-mode sorption of hydroxylated atrazine degradation products to soil: A mechanism for bound residue. Environ. Sci. Technol. 31:1539–1546.
- Ma, L., and H.M. Selim. 1996. Solute transport in soils under conditions of variable flow velocities. Water Resour. Res. 32:3277–3283.
- Martin-Neto, L., D.G. Traghetta, C.M.P. Vaz, S. Crestana, and G. Sposito. 2001. On the interaction mechanisms of atrazine and hydroxyatrazine with humic substances. J. Environ. Qual. 30:520–525.[Abstract/Free Full Text]
- Mersie, W., and C. Seybold. 1996. Adsorption and desorption of atrazine, deethylatrazine, deisopropilatrazine, and hydroxyatrazine on levy wetland soil. J. Agric. Food Chem. 44:1925–1929.
- Mersie, W., C. Seybold, and T. Tsegaye. 1999. Movement, adsorption and mineralization of atrazine in two soils with and without switchgrass (Panicum virgatum) roots. Eur. J. Soil Sci. 50:343–349.
- Meyer-Windel, S., B. Lennartz, and P. Widmoser. 1999. Bromide and herbicide transport under steady-state and transient flow. Eur. J. Soil Sci. 50:23–33.
- Miller, J.L., A.G. Wollum, and J.B. Weber. 1997. Degradation of carbon-14-atrazine and carbon-14-metolachlor in soil from four depths. J. Environ. Qual. 26:633–638.[Abstract/Free Full Text]
- Moorman, T.B., K. Jayachandran, and A. Reungsang. 2001. Adsorption and desorption of atrazine in soils and subsurface sediments. Soil Sci. 166:921–929.
- Moreau, C., and C. Mouvet. 1997. Sorption and desorption of atrazine, deethylatrazine, and hydroxyatrazine by soil and aquifer solids. J. Environ. Qual. 26:416–424.[Abstract/Free Full Text]
- Nakagawa, L.E., L.C. Luchini, M.R. Musumeci, and M.M. Andrea. 1995. Comportamento da atrazina em solos brasileiros em condições de laboratório. (Behavior of atrazine in a Brazilian soils under laboratory conditions.) (In Portuguese.) Pesquisa Agropecuária Brasileira 30:471–476.
- Prata, F., A. Lavorenti, J.B. Regitano, and V.L. Tornisielo. 2000. Influência da matéria orgânica na sorção e dessorção do glifosato em solos com diferentes atributos mineralógicos. (Influence of organic matter on sorption and desorption of glyphosate in soils of different mineralogical composition.) (In Portuguese.) Rev. Bras. Cienc. Solo 24:947–951.
- Roberts, T.R., and D.H. Hutson. 1999. Metabolic pathways of agrochemicals. Part 2. Insecticides and fungicides. The Royal Society of Chemistry, London.
- Roberts, T.R., D.H. Hutson, P.W. Lee, P.H. Nicholls, and J.R. Plimmer. 1998. Metabolic pathways of agrochemicals. Part 1. Herbicides and plant growth regulators. The Royal Society of Chemistry, London.
- Rodrigues, B.N., and F.S. Almeida. 1995. Guia de herbicidas. (Guide of herbicides.) (In Portuguese.) IAPAR, Londrina, Brasil.
- Sadeghi, A.M., A.R. Isennse, and D.R. Shelton. 1998. Effect of tillage age on herbicide dissipation: A side-by-side comparison using microplots. Soil Sci. 163:883–890.
- Senesi, N. 1992. Binding mechanisms of pesticides to humic substances. Sci. Total Environ. 123(124):63–76.[Medline]
- Seybold, C.A., and W. Mersie. 1996. Adsorption and desorption of atrazine, deethylatrazine, deisopropylatrazine, hydroxyatrazine, and metachlor in two soils from Virginia. J. Environ. Qual. 25:1179–1185.[Abstract/Free Full Text]
- Shelton, D.R., A.M. Sadegih, and A.R. Isensee. 1998. Effect of tillage on atrazine bioavailability. Soil Sci. 163:891–896.
- Thurman, E.M., D.A. Goolsby, M.T. Meyer, and D.W. Kolpin. 1991. Herbicides in surface waters of the midwestern United States: The effect of spring flush. Environ. Sci. Technol. 25:1794–1796.
- Thurman, E.M., M.T. Meyer, M.S. Mills, L.R. Zimmernan, and C.A. Perry. 1994. Formation and transport of deethylatrazine and deisopropylatrazine in surface water. Environ. Sci. Technol. 28:2267–2277.
- Toride, N., F.J. Leij, and M.Th. van Genuchten. 1995. The CXTFIT code for estimating transport parameters from laboratory or field tracer experiments: Version 2.0. Research Rep. 137. USDA-ARS, U.S. Salinity Lab., Riverside, CA.
- Vanderborght, J., P. Gähwiller, and H. Flühler. 2002. Identification of transport processes in soil cores using fluorescent tracers. Soil Sci. Soc. Am. J. 66:774–787.[Abstract/Free Full Text]
- van Genuchten, M.Th., and J.C. Parker. 1984. Boundary conditions for displacement experiments through short laboratory soil columns. Soil Sci. Soc. Am. J. 40:473–480.
- van Genuchten, M.Th., and P.J. Wierenga. 1976. Mass transfer studies in sorbing porous media. I. Analytical solutions. Soil Sci. Soc. Am. J. 40:473–481.[Abstract/Free Full Text]
- van Raij, B., and J.A. Quaggio. 1983. Métodos de análise de solos para fins de fertilidade (Methods of soil analysis for soil fertility.) (In Portuguese.) Tech. Bull. 81. Instituto Agronômico, Campinas, SP, Brazil.
- Vettori, L. 1969. Métodos de análise do solo. (Methods of soil analysis.) (In Portuguese.) Tech. Bull. 7. Ministério da Agricultura, Div. Pedologia e Fertilidade do Solo, Rio de Janeiro.
- Wauchope, R.D., T.M. Buttler, A.G. Hornsby, P.W.M. Augustijn-Beckers, and J.P. Burt. 1992. The SCS/ARS/CES Pesticide Database for environmental decision making. Rev. Environ. Contam. Toxicol. 123:1–155.[ISI][Medline]
TOP
ABSTRACT
INTRODUCTION
