Environ. Sci. Technol. 2006, 40, 494-500
Pesticide Sorption and Desorption by Lignin Described by an Intraparticle Diffusion Model W E N D Y V A N B E I N U M , * ,† S A B I N E BEULKE,† AND COLIN D. BROWN‡ Cranfield University, Silsoe, Bedford MK45 4DT, U.K., Environment Department, University of York, Heslington, York YO10 2DD, U.K., and Central Science Laboratory, Sand Hutton, York YO41 1LZ, U.K.
Lignin was used as a model compound for soil organic matter to gain insight into the mechanisms that control the kinetics of pesticide sorption and desorption. Hydrolytic lignin was immobilized in a matrix of alginate gel, and sorption-desorption experiments were undertaken with isoproturon. Sorption increased with time and was close to equilibrium after 14 days. Desorption was measured after sorption for different time intervals and for a number of successive desorption steps of different lengths. The results showed strong differences between the sorption and desorption isotherms. The ratio of sorbed to dissolved pesticide approached and even exceeded the equilibrium ratio, depending on the number of desorption steps and the length of each equilibration period. A numerical diffusion model was developed to describe radial diffusion into the lignin particles in combination with Freundlich sorption inside the particles. Key model parameters were adjusted to fit the sorption data, and the same parameters were then used to predict stepwise desorption. Desorption was well described by the model, which suggests that sorption and desorption were driven by the same mechanism and occurred at the same rate. The observed difference between the sorption and desorption isotherms could be fully explained by the nonattainment of equilibrium due to slow diffusion into and out of the lignin particles.
Introduction Simulations of pesticide fate in soil are often based on the assumption that a characteristic equilibrium between pesticide in soil solution and pesticide sorbed to solid particles is achieved instantaneously (1). However, sorption has frequently been observed to increase with increasing time of interaction with the soil (2-4). This phenomenon is likely to be caused by a combination of processes that delay either the transport of pesticide molecules to sorption sites or the reaction between the sorbate and the sorbent (5, 6). Transport-controlled nonequilibrium may be caused by (i) diffusion of the pesticide through larger pores between soil aggregates, (ii) solute diffusion into the smaller pores inside soil aggregates, and (iii) slow diffusion at the very small scale into the matrix of organic and mineral sorbents. Soil organic matter is the main sorbent for many * Corresponding author phone: +44 (0)1904 462000; fax: +44 (0)1904 462111; e-mail:
[email protected]. † Cranfield University. ‡ University of York and Central Science Laboratory. 494
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 2, 2006
organic solutes. It consists of a large porous network of interlinked organic molecules into which pesticides may diffuse. Extensive research to elucidate the mechanisms that lead to time-dependent sorption has been carried out, and a number of approaches have been proposed to describe these mechanisms mathematically (5, 7). Approaches referred to as two-site models assume that sorption is instantaneously at equilibrium on part of the sorbent and rate-limited on the remaining part, for which sorption is described by a first-order reaction (2, 8). Other modeling approaches describe slow sorption explicitly as a diffusion process (5, 7). An additional simplifying assumption, in tools to simulate pesticide fate, is that sorption and desorption processes are fully reversible and occur at the same rate (1). However, sorption hysteresis of organic contaminants, where sorption and desorption isotherms are not coherent, is a common phenomenon (9-13). Apparent hysteresis is often caused by artifacts during the experimentation, such as changes to the sorbent, a successive decline in the concentration of suspended or dissolved sorbents, or degradation (13). An additional possible cause of hysteresis is the nonattainment of sorption equilibrium during the sorption step (10, 12). In this case, sorption continues during the desorption step, which results in an increase in the proportion of the compound sorbed. True hysteresis may be caused by differences between sorption and desorption reactions. The activation energy of desorption may be larger than the activation energy of sorption, which could lead to kinetic hysteresis (14). Furthermore, desorption can be inhibited by the rearrangement of organic matter caused by the penetration of the solute (6). The relative importance of the various processes leading to nonequilibrium and hysteresis phenomena of pesticides in soils is not fully understood. We used lignin as a model compound for soil organic matter to gain insight into the mechanisms that control the sorption-desorption behavior of the herbicide isoproturon (IPU, 3-(4-isopropylphenyl)1,1-dimethylurea). IPU is a commonly investigated phenylurea herbicide. Numerous studies have shown that sorption of IPU in soil and in sediments is directly related to sorption on soil organic matter (e.g. 15, 16). Sorption is also believed to occur on clay minerals (17) but not on iron oxides (18). The sorbent lignin is a natural component of wood. During decomposition, lignin undergoes chemical reactions and associates with other organic molecules to form soil organic matter. Commercially available hydrolytic lignin was immobilized in a matrix of alginate gel to facilitate the separation of lignin and solution. A homogeneous mixture of lignin and alginate gel was cross-linked into solid spherical gel beads. Alginate is a hydrogel with an open structure through which small molecules can diffuse freely. Sorption and desorption experiments were carried out at different concentrations of IPU and for different time intervals. A numerical diffusion model was developed to describe diffusion into the lignin particles. Key model parameters were adjusted to fit the sorption data. The same parameters were then used to predict stepwise desorption of the compound from the lignin.
Material and Methods Preparation of Lignin Gel Beads. Lignin clusters were immobilized in alginate beads to allow their easy and rapid separation from the external solution, which had to be replaced repeatedly during the experiments. This method 10.1021/es051940s CCC: $33.50
2006 American Chemical Society Published on Web 12/07/2005
was preferred over centrifugation because it prevents lignin from becoming attached to the walls and lids of the glass vials. It also maintains the structure of the lignin clusters, whereas lignin particles may aggregate during centrifugation, which could potentially modify their diffusion properties. Lignin suspension was prepared by adding 0.5 g of hydrolytic lignin (Aldrich, organic carbon content ∼67%) to 100 mL of water. The suspension was sonicated in an ultrasonic bath at 40 °C for 15 min to enhance the dispersion of the particles. Then, 0.5 g of sodium alginate was mixed in and left on a magnetic stirrer to dissolve. Gel beads were prepared by slowly dripping the gel solution into a large beaker containing 1 M CaCl2 solution on a magnetic stirrer, using a peristaltic pump. The gel solidifies on contact with calcium and forms spherical beads (19). After complete cross-linking, the gel beads were equilibrated with 5 mM CaCl2 by packing the beads into a glass column and pumping through fresh 5 mM CaCl2 solution for 24 h. The result is a batch of equally sized spherical gel beads (3.2 mm diameter) with microscopic lignin particles dispersed inside the gel. The final size of the gel beads depends on the concentration of CaCl2 they are kept in. To approximate the ionic strength of soil solution, all experiments were performed in 5 mM CaCl2. The average concentration of lignin in the gel beads was calculated from the weight of gel solution used to prepare the beads, its lignin concentration, and the number of beads formed, resulting in an average of 0.344 mg per bead or 21.3 mg g-1. The density of beads was measured in five 5 mL volumetric flasks at 20 °C. Each flask was filled with 100 gel beads and with CaCl2 solution. The volume of the gel beads was calculated by subtracting the volume of CaCl2 (density 1.00 g mL-1) from the total volume. The density, calculated from the weight and volume per 100 beads, was 1.01 ( 0.01 g mL-1, very similar to the density of water. The volume and weight of the lignin in the gel are negligible; therefore, the water content of the gel is around 1.0 mL g-1. Kinetic Sorption-Desorption Experiments. Kinetic sorption was measured with 60 gel beads (∼1 g) in 30 mL glass vials with polyethylene screw caps. To each vial, 3 mL of pesticide solution was added, giving a total solution volume of approximately 4 mL. Solutions were prepared with 3.3, 6.2, 12, or 16 mg L-1 IPU in 5 mM CaCl2. A commercial formulation of IPU (Steffes IPU, 44% w/w suspension concentrate) was used, corresponding to IPU use in the field. The concentration of IPU in solution was measured after shaking it gently at 20 °C for 1 and 3 h and 1, 3, 7, and 14 days. Desorption was measured using 1 g of alginate beads from the same batch as those used for the sorption experiments. IPU solution (3 mL) at two different initial concentrations (6.2 and 16 mg IPU L-1) was added to the beads. After sorption for 1 day, 2.4 mL of the external solution was removed and analyzed for IPU. The solution was replaced by 2.4 mL of 5 mM CaCl2 solution without IPU. The samples were then shaken for 1, 3, and 7 days, and the concentration of IPU in the equilibrium solution was measured by HPLC. An additional experiment investigated desorption of IPU in successive steps. Sorption of IPU on the lignin for 1 day was followed by five successive 1 day or 3 day desorption steps. Sorption for 7 days was followed by five successive 3 day desorption steps. At each step, 2.4 mL of the external solution was removed and replaced with 5 mM CaCl2 (2.4 mL). The concentration of IPU in the external solution was measured after each sorption and desorption step. Again, two initial concentrations of IPU were tested (6.2 and 16 mg L-1). There were at least two replicates in all sorptiondesorption experiments. Parallel samples without alginate beads (blanks) were prepared and kept refrigerated until
analysis in order to determine the actual concentration of the solution added to the system. Sorption on the vial caps was found to be negligible in a separate experiment. Supporting Experiments. Sorbed concentrations of IPU were derived from concentrations in the external solution and from the total concentration in the system minus any IPU degraded during the experiment. To determine the rate of IPU degradation, the total amount remaining in the system (sorbed + dissolved) was measured after sorption for 1, 7, and 14 days. The lignin was destroyed by adding 10 mL of a 0.2 M NaOH solution to each vial and leaving the lignin to dissolve for 48 h. An additional experiment was carried out to investigate whether diffusion of IPU into the alginate gel beads is a ratelimiting factor. Alginate beads without lignin (1 g) were shaken gently with 3 mL of the IPU solution (16 mg L-1). Parallel batches were shaken gently for 10, 20, 30, 40, 50, and 60 min, and samples were taken for analysis. The study was carried out with two replicates per equilibration time. Sorption on lignin immobilized in the alginate beads was compared with sorption on lignin in suspension. The suspension (0.5 g of lignin to 100 mL of demineralized water) was sonicated in an ultrasonic bath at 40 °C for 15 min to enhance dispersion of the lignin and then adjusted from pH 3.5 to pH 6 by adding sodium hydroxide. The lignin was centrifuged and washed several times with 5 mM CaCl2 solution. Sorption was measured by adding IPU to the lignin suspension, resulting in parallel samples of 10 mL with 4.9 g L-1 lignin and 16 mg L-1 IPU. The suspension was shaken gently for six different periods with a maximum equilibration time of 7 days (two replicates per time step). The samples were centrifuged at 2000g for 20 min, and the concentration of IPU in the supernatant was measured by HPLC. Analysis of IPU. The sample (20 µL) was injected directly into a DX600 (Dionex, Sunnyvale, CA) HPLC equipped with a PDA100 photodiode array detector and a Discovery C-18 column (150 mm long × 4.6 mm i.d., 5 µm particle size, Supelco, Bellefonte, PA). The mobile phase was a 40:60 mixture of acetonitrile and aqueous solution. The aqueous solution was made up of 0.04% v/v orthophosphoric acid in HPLC-grade water. The flow rate of the mobile phase was 1 mL min-1 with a run time of 7 min per sample, and the detection wavelength was 200 nm. The typical retention time was 6 min. The limit of quantification was 70 µg L-1, and the limit of detection was ca. 10 µg L-1. Intraparticle Diffusion Model. Lignin consists of associated macromolecules that may form clusters in suspension. The kinetic sorption process was modeled as radial diffusion into these clusters in combination with Freundlich sorption and degradation inside the porous particles:
(
)
∂S ∂2C 2 ∂C ∂C - k(θC + FS) + F ) θDe 2 + ∂t ∂t r ∂r ∂r
θ
(1)
where C is the IPU concentration in solution (mg m-3), S is the sorbed IPU concentration (mg kg-1), θ is the porosity (-), F is the bulk density of the particle (kg m-3), r is the radial distance (m), De is an effective diffusion coefficient (m2 h-1) that accounts for the constricted movement of molecules through the narrow particle pores (tortuosity), and k is the degradation rate constant (h-1). Degradation was calculated assuming first-order kinetics and using the same degradation rate constant for sorbed and dissolved pesticide. Although there is some evidence that the availability of sorbed pesticide for degradation may be limited (20), it was not necessary to account for this effect in our study. Only a very small amount of the initially added pesticide had degraded by the end of the experiment, and the modeling results were independent of the relative contribution of the sorbed and dissolved phase to degradation. Sorption was VOL. 40, NO. 2, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
495
TABLE 1. Model Parameters model parameter
value
particle diametera particle porosityb particle densityc lignin mass per replicated Freundlich sorption isotherme
10 µm 0.5 1.5 kg dm-3 20.7 mg KF ) 226 mg kg-1 (mg L-1)-n, n ) 0.90 1.1 × 10-4 h-1 0.25 1.0 × 10-11 m2 h-1
degradation rate constantf instantaneous fractione effective diffusion coefficiente
a Size of large lignin clusters observed with microscope. b Estimated, see text. c Estimate for lignin (22). d Lignin content of 60 gel beads. e Adjusted values to match sorption data. f First-order constant derived from NaOH extraction results.
described by the Freundlich equation (eq 2) where KF (mg
S ) KFCn
(2)
kg-1(mg L-1)-n) is the Freundlich coefficient and n (-) is the Freundlich exponent. Diffusion was calculated for one single particle and corrected for the theoretical total amount of lignin per vial. The system was defined by the initial and boundary conditions: (i) the initial concentration inside the spherical particles (r < outer radius a) is zero, (ii) the initial concentration outside the particles (r ) a) is determined by the amount of IPU added and the instantaneous sorbed fraction (f), (iii) the system has radial symmetry, so the gradient across the center of the sphere is zero, and (iv) the concentration outside the sphere (r ) a) is uniform and the change in concentration is determined by the diffusion flux into the particles and the particle:solution ratio (kg L-1). The above equations were solved numerically with the ORCHESTRA program (21). Diffusion was calculated between 10 cells, representing 10 concentric layers of a sphere. Each concentric layer has a volume (V) and outer surface area (A) determined by the outer radius of the layer. The diffusion flux between the calculation cells was simulated for small time steps based upon the concentration gradient in the pore solution and the effective diffusion coefficient:
where Jdiff,i f j is the diffusion flux from cell i into the adjacent
FIGURE 2. Sorption of IPU after 1 day sorption (solid circles) and subsequent desorption by replacement of the external solution with pesticide-free solution and equilibration for 1, 3, or 7 days (open symbols). Replacement of the pesticide solution causes a drop in the amount of sorption (desorption indicated by an arrow), but after longer equilibration times, the measurements move closer to the equilibrium sorption line due to continued sorption inside the particle.
FIGURE 1. Sorption of IPU on lignin after different times of equilibration.
cell j for the subsequent time step, θ is the porosity, De is the effective diffusion coefficient in the pores, Ci(t) and Cj(t) are the concentrations in solution at time t in cells i and j respectively, and dr is the radial distance between the concentric layers of the sphere. The flux was multiplied with the contact area between the concentric layers (A) and the time step (dt) to obtain the mass transfer between the calculation cells. After calculation of all fluxes, the mass is added to and subtracted from the total mass in the relevant cells, obeying conservation of mass. Degradation is calculated for the same time step from the concentration at time t and subtracted from the total mass in the cells. The model parameters are summarized in Table 1. The size of lignin particles was estimated from observations with a light microscope. Lignin was suspended by sonication consistent with the preparation method for the lignin-gel beads. The size distribution was found to be bimodal with the diameter of most particles between 1 and 2 µm and some larger clusters with an approximate average size of 10 µm. The larger clusters are irregular in shape but are best
Jdiff,i f j ) - θDe
496
9
Cj(t) - Ci(t) for time ) t f t + dt (3) dr
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 2, 2006
FIGURE 3. Stepwise desorption of IPU from lignin after 1 day sorption. Stepwise desorption by replacement of the external solution with pesticide-free solution every day (top) or every 3 days (bottom). Results for initial concentrations 6.2 (left) and 16 mg L-1 (right). described as spheres (as opposed to flat shapes or rodlike shapes) for modeling purposes. Diffusion was assumed to be the rate-limiting step for the larger clusters, whereas diffusion in the small particles is very fast and sorption is, thus, virtually instantaneous. A simple Freundlich equilibrium model was used to describe sorption on the small lignin particles (referred to as “instantaneous fraction”). It should be noted that diffusion-controlled sorption onto the outer rim of the larger lignin clusters is also fast due to short diffusion pathways. The density of lignin particles was estimated to be 1.5 kg dm-3 (22). The reported microporosity of lignin is very small (∼0.001, ref 23). However, we assumed a loose packing of particles inside the larger clusters and set the porosity of the clusters to 0.5. We have no information on the structure of lignin clusters, and the values for density and porosity used in the model are arbitrary. This does, however, not have any significant implications for the interpretation of the modeling results. A first-order degradation constant was fitted to the results of the NaOH extractions. The recovery of the extractions was 97% ( 2% after 24 h. Degradation of IPU in alginate beads with lignin was found to be slow. After an equilibration time of 14 days, 94% of the added IPU was recovered (first-order rate constant 1.1 × 10-4 h-1). Sorption equilibrium was not reached within the time-scale of the experiment. The equilibrium sorption coefficient and Freundlich exponent were, therefore, derived by fitting the model to the data. The fraction of
instantaneous sorption and the effective diffusion coefficient were also optimized to give a good fit to the sorption data in time.
Results and Discussion Sorption. The results of the sorption experiments are shown in Figure 1. The sorption curves for IPU were nonlinear following a Freundlich isotherm. Sorption increased with increasing equilibration times up to 14 days. Only a very small additional increase in sorption was observed between 7 and 14 days, suggesting that sorption after 14 days is close to equilibrium sorption. In the absence of measurements at longer time points, it is not clear when a true equilibrium would be reached. However, the equilibrium sorption value optimized in the model was identical to that measured after 14 days. The sorption behavior of IPU was modeled assuming instantaneous Freundlich sorption on a fraction of the lignin followed by slow diffusion into the lignin particles where the pesticide comes into contact with additional sorption sites. The model parameters are given in Table 1. The fitted model gave an adequate description of the increase in the strength of sorption with time (Figure 1). Although estimates were used for some of the uncertain model parameters, such as the density and the porosity of the lignin particles, it can be concluded that diffusion into lignin is a slow process. The diffusion coefficient of IPU in water was VOL. 40, NO. 2, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
497
FIGURE 5. Simulated IPU concentrations in the lignin particle during different stages of the desorption experiment. The amount of IPU inside the particle decreases during the first day of desorption and then increases during the second and third day of desorption due to continued sorption in the particle center.
FIGURE 4. Stepwise desorption of IPU from lignin after 1-day sorption (initial concentration 6.2 mg L-1) followed by replacement of the external solution with pesticide-free solution every 3 days. (a) Sorbed concentration plotted vs dissolved concentration. The arrows indicate the sequential processes during each desorption step according to the model simulations: (1) dilution of the external concentration by replacement with pesticide-free solution, (2) instantaneous desorption from the small lignin particles, (3) slow sorption/desorption inside the larger particles. (b) Sorbed concentration plotted vs time. The line shows the change in sorption during the model simulations. IPU is released from the particles during each desorption step. Only during the first step is the initial release followed by continued uptake of IPU into the particles. calculated for comparison using a model developed by Othmer and Thakar (24) with slightly modified coefficients (25):
Dw ) 3600 ×
13.26 × 10-9 η1.14Vm0.589
(4)
where Dw is in m2 h-1, η is the viscosity of water (mPa s-1), and Vm is the molar volume of the compound (cm3 mol-1) equivalent of the molecular weight of the compound divided 498
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 2, 2006
by its liquid density. The calculated diffusion coefficient in water of 2.3 × 10-6 m2 h-1 was 5 orders of magnitude faster than the optimized coefficient for effective diffusion into the lignin particles (1.0 × 10-11 m2 h-1). It is likely that the pesticide molecules are relatively large compared with the pores inside the lignin particles. This may slow the diffusion of the molecules into the particles. Diffusion of IPU into the alginate gel without lignin was also investigated. It was found that IPU does not interact with the gel and that diffusion was 98% complete within 50 min. The optimized diffusion coefficient for IPU in the gel (1.4 ×10-6 m2 h-1) was similar to that estimated for diffusion in water (2.3 × 10-6 m2 h-1) with eq 4, although the latter is only a rough approximation based on the molar volume of IPU. Kinetic sorption of IPU to lignin in suspension was found to be consistent with sorption to lignin immobilized in alginate beads. These results suggest that diffusion into the alginate is not a rate-limiting factor. The importance of diffusion into the alginate was investigated by additional modeling that took the diffusion into the gel beads into account (results not shown). The model simulation demonstrated that diffusion into the gel only influences adsorption at the shortest equilibration times of 1 and 3 h. This process is not rate-limiting at longer equilibration times. Desorption. After sorption for 1 day, IPU was desorbed for either 1, 3, or 7 days. The results are shown in Figure 2 (symbols) for two different initial concentrations. For both initial concentrations, the ratio of sorbed to dissolved IPU increased during the desorption step compared with the 1 day sorption curve. The sorption step here was short (1 day), and the system was still far from equilibrium after this time. The replacement of part of the external solution by a solution without pesticide resulted in dilution and increased the ratio of sorbed to dissolved pesticide. Sorption and desorption processes during the subsequent equilibration period moved the system close to the equilibrium state. Successive desorption of IPU was also investigated. The experiments were carried out for two initial concentrations. Figure 3 shows the results for repeated 1 day or 3 day desorption steps following sorption for 1 day. The lines show the sorption and desorption curve calculated with the intraparticle diffusion model. Desorption was simulated using the parameters which were previously adjusted to match the sorption data (Table 1, Figure 1). The desorption line clearly differs from the 1 day sorption line (Figure 3). The ratio between sorbed and dissolved increases with each desorption step. After equilibration for 1 day, the amount of sorption is far beneath equilibrium.
FIGURE 6. Stepwise desorption of IPU from lignin at two different initial concentrations following sorption for 7 days. The desorption steps bring the measured sorption much closer to the modeled equilibrium line, reaching and crossing the equilibrium line after two desorption steps. The model calculations show that this behavior can be explained by slow intraparticle diffusion. The behavior of the model is illustrated in Figure 4. The arrows in Figure 4a show the modeled change in the sorbed and dissolved concentrations after 1 day sorption and successive 3 day desorption steps (initial concentration 6.2 mg L-1). The sorbed concentration is plotted versus time in Figure 4b. Sorption equilibrium was not established at the time of the first replacement of part of the external solution by solution without pesticide (Figure 4). The model predicts slow desorption from the outer region of the particles and continued sorption inside the particles during the subsequent 3 day equilibration period. Dilution and instantaneous desorption immediately after the second sampling shift the modeled ratio of sorbed to dissolved pesticide above the equilibrium ratio. The system then moves back to the equilibrium line. For the subsequent steps, the equilibration period is too short to fully compensate for the decline in concentration caused by dilution. As a result, the model predicts that the ratio of sorbed to dissolved pesticide exceeds the equilibrium ratio (Figure 4a). Slow sorption and desorption occur simultaneously in different regions within the spherical lignin particles and result in a net uptake (sorption > desorption) or release (desorption > sorption) of the pesticide during the equilibration period. This is illustrated in Figure 5. After 1 day sorption, IPU has only diffused into the outer rim of the particle. During desorption, part of the IPU is released from the outer rim of the particles back into the external solution. However, diffusion deeper into the particle continues at the same time, causing additional sorption. Figure 6 shows the results for sorption of IPU for 7 days followed by successive 3 day desorption steps. Again, there was a clear difference between the sorption and desorption lines. The sorbed concentrations were slightly underestimated by the intraparticle diffusion model. General Discussion. Simulations of the desorption behavior of IPU used parameters fitted to the sorption behavior of the compound without any further adjustments. The good fit to the desorption data implies that both sorption and desorption were driven by the same mechanism and occurred at the same rate. While pesticide is desorbed from the fraction of the lignin that is instantaneously at equilibrium and from
the outer rim of the particles, pesticide still continues to diffuse and sorb in the interior of the particle. The observed nonsingularity of the sorption and desorption isotherms could be fully explained by the nonattainment of equilibrium due to slow diffusion into and out of the lignin particles. This is in line with observations by Altfelder et al. (10, 12). There was no indication that the shift toward the sorbed state during desorption was caused by strong or irreversible reactions between the solute and the sorbent in our test system. Irreversible binding is, however, reported as a possible reason for differences between sorption and desorption for some pesticides that are likely sorb to soil organic matter. Moyer et al. (26) found that part of the monuron, linuron, and prometryne added to peat became irreversibly bound when the peat was dried. Work by Celis and Koskinen (11) suggested that a fraction of sorbed imidacloprid was irreversibly bound in soil suspensions. Desorption of IPU was initiated by a replacement of the external solution with pesticide-free solution. This dilution effect occurs in natural soils during intense leaching events. Incoming rainwater without pesticide replaces the resident soil solution, which results in movement of pesticide to depth. Our results suggest that more pesticide may initially be available for leaching than estimated from equilibrium isotherms. Sorption-desorption processes after each leaching event may lead to an apparent increase in the strength of sorption beyond the equilibrium state. However, sorption to natural soils is more complex than sorption to the isolated organic material used in this study. Time-dependent sorption and differences between adsorption and desorption in natural aggregates are likely to be the result of diffusion into pores inside the aggregates followed by intraparticle diffusion. The relative importance of these processes at the macroscopic scale is currently being investigated.
Acknowledgments This work was funded by the British Biotechnology and Biological Sciences Research Council (BBSRC) under research grant 63/D14743.
Supporting Information Available The performance of the diffusion model is compared with the performance of a two-site model to describe sorption and desorption of isoproturon on lignin. This material is available free of charge via the Internet at http://pubs.acs.org. VOL. 40, NO. 2, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
499
Literature Cited (1) Jarvis, N. J.; Bergstro¨m, L. F.; Brown, C. D. Pesticide leaching models and their use for management purposes. In Environmental Behaviour of Agrochemicals; Roberts, T. R., Kearney, P. C., Eds.; John Wiley & Sons: New York, 1995; pp 185-220. (2) McCall, P. J.; Agin, G. L. Desorption kinetics of picloram as affected by residence time in the soil. Environ. Toxicol. Chem. 1985, 4, 37-44. (3) Cox, L.; Walker, A. Studies of time-dependent sorption of l inuron and isoproturon in soils. Chemosphere 1999, 38, 27072718. (4) Koskinen, W. C.; Cox, L.; Yen, P. Y. Changes in sorption/ bioavailability of imidacloprid metabolites in soil with incubation time. Biol. Fertil. Soils 2001, 33, 546-550. (5) Brusseau, M. L.; Jessup, R. E.; Rao, P. S. C. Nonequilibrium sorption of organic chemicals: Elucidation of rate-limiting processes. Environ. Sci. Technol. 1991, 25, 134-142. (6) Pignatello J. J. The measurement and interpretation of sorption and desorption rates for organic compounds in soil media. Adv. Agron. 2000, 69, 1-73. (7) Ma, L.; Selim, H. M. Physical nonequilibrium modeling approaches to solute transport in soils. Adv. Agron. 1997, 58, 95150. (8) van Genuchten, M. Th.; Wagenet, R. J. Two site/two region models for pesticide transport and degradation. Soil Sci. Soc. Am. J. 1989, 53, 1303-1310. (9) Weber, W. B.; Huang, W.; Yu, H. Hysteresis in the sorption and desorption of hydrophobic organic contaminants by soils and sediments - 2. Effects of soil organic matter heterogeneity. J. Contam. Hydrol. 1998, 31, 149-165. (10) Altfelder, S.; Streck, T.; Richter, J. Effect of air-drying on sorption kinetics of the herbicide chlortoluron in soil. J. Environ. Qual. 1999, 28, 1154-1161. (11) Celis, R.; Koskinen, W. C. Characterization of pesticide desorption from soil by the isotopic exchange technique. Soil Sci. Soc. Am. J. 1999, 63, 1659-1666. (12) Altfelder, S.; Streck, T.; Richter, J. Nonsingular sorption of organic compounds in soil: The role of slow kinetics. J. Environ. Qual. 2000, 29, 917-925. (13) Delle Site, A. Factors affecting sorption of organic compounds in natural sorbent/water systems and sorption coefficients for selected pollutants. A review. J. Phys. Chem. Ref. Data 2001, 30, 187-439. (14) Pignatello, J. J.; Xing, B. Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 1996, 30, 1-11.
500
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 2, 2006
(15) Cooke, C. M.; Shaw, G.; Collins, C. D. Determination of solidliquid partition coefficients (K-d) for the herbicides isoproturon and trifluralin in five UK agricultural soils. Environ. Pollut. 2004, 132, 541-552. (16) Spark, K. M.; Swift, R. S. Effect of soil composition and dissolved organic matter on pesticide sorption. Sci. Total Environ. 2002, 298, 147-161. (17) Elkhattabi, K.; Bouhaouss, A.; Perrin-Ganier, C.; Schiavon, M. Fate of isoproturon in two Moroccan soils. Agronomie 2004, 24, 177-183. (18) Clausen, L.; Fabricius, I. Atrazine, isoproturon, mecoprop, 2,4D, and bentazone adsorption onto iron oxides. J. Environ. Qual. 2001, 30, 858-869. (19) van Beinum, W.; Meeussen, J. C. L.; Edwards, A. C.; van Riemsdijk, W. H. Transport of ions in physically heterogeneous systems; convection and diffusion in a column filled with alginate gel beads, predicted by a two-region model. Water Res. 2000, 34, 2043-2050. (20) Ogram, A. V.; Jessup, P. P.; Ou, L. T.; Rao, P. S. C. Effects of sorption on biological degradation rates of 2,4-dichlorophenoxy acetic acid in soils. Appl. Environ. Microbiol. 1985, 49, 582587. (21) Meeussen, J. C. L. ORCHESTRA: An object-oriented framework for implementing chemical equilibrium models. Environ. Sci. Technol. 2003, 37, 1175-1182. (22) Skaven-Haug, S. V. Volumetric Relations in Soil Materials. Appendix 2. In Nature and Management of Tropical Peat Soils; Andriesse, J. P., Ed.; FAO Soils Bulletin 59: Rome, Italy, 1988. (23) Rutherford, D. W.; Wershaw, R. L.; Cox, L. G. Changes in Composition and Porosity Occurring During the Thermal Degradation of Wood and Wood Components; USGS Scientific Investigations Report 2004-5292; U.S. Government Printing Office: Washington, D.C., 2004. (24) Othmer, D. F.; Thakar, M. S. Correlating diffusion coefficients in liquids. Ind. Eng. Chem. 1953, 45, 589-593. (25) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M. Environmental organic chemistry, 2nd ed.; Wiley-Interscience: Hoboken, NJ, 2003. (26) Moyer, J. R.; McKercher, R. B.; Hance, R. J. Desorption of some herbicides from montmorillonite and peat. Can. J. Soil Sci. 1972, 52, 439-447.
Received for review September 30, 2005. Accepted October 24, 2005. ES051940S
