Sorption of paraquat on clay components in Taiwan's oxisol

The sorption of herbicides in soils is mainly influenced by clay components. The objectives of this study were to evaluate the contribution of clay components on paraquat sorption. The surface soils (0–20 cm) of a Laopi pedon (Fine, mixed, Hyperthermic Typic Hapludox) were separated clays into whole (< 2.0 μm), coarse (0.2–2.0 μm), and fine (< 0.2 μm) fractions with the treatments of removals of organic matter (OM) and free Fe (Fe_d) oxides. Results indicated that sorption isotherm of paraquat was fitted by the nonlinear Freundlich equation with R² values ranged in 0.79–0.96, respectively. The shape of paraquat adsorption isotherm on the fine fraction was H‐type, but their shapes on the whole and coarse fractions were L‐types. The fine clay fractions gave higher contribution on paraquat sorption than the coarse clay fractions identified by their K_f values. Organic matter associated with fine clay fraction had high CEC contributing to relatively high affinity for paraquat. The DCB treatment created high‐affinity sites for paraquat on the fine clay, but had little effect on paraquat sorption for the coarse clay. Chemisorption is the major mechanism for retention of paraquat on clay components, not ion exchange. However, the silicate clay had the highest affinity for paraquat and free Fe compound had the lowest.

Keywords: Paraquat; Sorption; Freundlich equation; Clay; Organic matter; Free Fe compound

Adsorption from aqueous solution to solid surfaces is one of the key processes determining the concentration and rate of transport of pesticides in soils and sediments.[1] Soil is complicated by the heterogeneous nature of its components. Thus, interaction among different soil components is likely. Although organic matter and oxides are important in adsorption reactions, differences exist in their relative importance.[2] Past studies of cationic pesticides adsorption by individual components such as organic matter, silicate minerals, and sesquioxides have indicated relatively strong bonding and high capacities of the various materials to adsorb pesticides. Several studies have recognized the need to separate total soil loads into several size fractions to better simulate transport of soil and sorbed pesticides in the environment.[[3]] Celis et al.[6] indicated that mineral and organic coatings on montmorillonite altered the nature of the surface exposed for sorption of thiazafluron, thus affecting the sorption behavior of colloidal soil particles. Beck and Jones[7] found that no significant differences between the sorption of atrazine and isoproturon by < 2 mm and < 250 μm fractions in a clayey soil.

Paraquat (1, 1′‐dimethyl‐4, 4′‐bypyiridinium ion) is one of the contact herbicides widely used for the control of various broad‐leaved weeds and grasses in planting crop such as sugar cane, pineapple and tea on the terrace of Taiwan. It is divalent cation and highly soluble in water. The environmental impact of paraquat is highly influenced by their interactions with the clay components of the soil.[8] Damanakis et al.[9] reported that the strong adsorption capacities (SAC) of paraquat on the peats were much higher than the mineral soils. Gamar and Mustafa[10] reported that the variability of the SAC of paraquat was basically due to CEC. Weber and Scott[11] indicated that paraquat was bound within the interlayer spacing of the montmorillonite clay by coulombic and van der Waals forces and to the surfaces of particles of kaolinite clay by coulombic ion exchange forces only. Rytwo et al.[12] found the adsorption behavior of paraquat corresponded to H‐type or high‐affinity class adsorption isotherms and the affinity of paraquat for adsorption to montmorillonite were much stronger than those observed for inorganic divalent cations. However, little is known about the relative contributions of the soil clay components to paraquat sorption. But it is likely to be important in controlling the concentration of paraquat in the soil solution and mobility in the aquifer. The objectives of this study are to 1) quantify the relative importance and contributions of various soil colloidal components to the sorption of paraquat and 2) understand the role of organic matter, free Fe oxides and silicate clays associated with different clay fractions and their interactions on the sorption of paraquat.

The soil used in this study was collected from the Ap horizon (0–20 cm) of the Laopi series (Fine, mixed, Hyperthermic Typic Hapludox) located on the alluvial fan terrace that was derived from Quaternary‐aged materials in southern Taiwan that main crop is sugar cane or pineapple. The selected soil properties were made including the particle size distribution by the pipette method,[13] the pH of a mixture of soil and deionized water (1:1, w/v) with a glass electrode,[14] the organic carbon content by the Walkley‐Black wet oxidation method,[15] cation exchangeable capacity (CEC) with the ammonium acetate method (pH 7.0),[16] and free Fe (Fed) oxides extracted by means of the dithionite‐citrate‐bicarbonate (DCB) method.[17] The selected physical and chemical properties of the soil as shown in Table 1.

Table 1. The selected properties of the study soil

1 +Organic matter.

• 2 ++Cation exchangeable capacity.
• 3 *Free iron.

The soil was dispersed by 5% hexametphosphate, and clay (< 2.0 μm) was separated by sedimentation. Coarse (0.2–2.0 μm) and fine (< 0.2 μm) clay fractions were separated by centrifugation at the appropriate speed. All three clay fractions, including whole (< 2.0 μm), coarse, and fine clays, were then split into three equal portions. One portion was untreated. The other two portions were treated with 30% H2O2 to remove organic matter. One of the H2O2 treated portion was then treated with DCB extract for removal of free Fe oxides. All 9 samples including three size fractions by three treatments were further freeze‐dried.

Sorption isotherms on the soil clays were obtained using the batch equilibration procedure. Paraquat with 99% purity was obtained at the dichloride salt from Sigma Company. Duplicate 0.2‐g soil clay samples were treated with 20 mL of paraquat solutions ranged up to 1 × 10− 2 M. The suspensions were shaken for 2 hrs at 23 ± 2°C. The pH of the suspensions was adjusted to 5.0 by addition of dilute HCl or NaOH. The supernatant was separated by centrifugation at 18,000 × g for 30 min. After adsorption experiments, desorption was subsequently conducted by an equal volume of 0.01 M CaCl2 solution from the highest initial paraquat concentration. Paraquat in all supernantants was added in a 10 mL of 1% Na2S2O3 in 0.1N NaOH and determined by measuring the absorption at 394 nm with UV/visible spectrophotometer (Hitachi U‐2001, Japan). The amount of adsorbed paraquat was calculated at the difference between the amount added initially and that remaining in solution after equilibration. Experiments were carried out in triplicate. All statistical analyses were performed using SAS.[18]

Paraquat adsorption isotherms for all particle size fractions with different treatments are shown in Figure 1. These isotherms were fitted by the Freundlich equation: Cs = KfCeNf, where Cs is the paraquat adsorbed (mg/kg) of clay, Ce is the concentration (mg/L) of paraquat in equilibrium solution, and Kf and Nf are constants characterizing the adsorption capacity and intensity of paraquat, respectively. Sorption coefficients were calculated by nonlinear least‐square regression to estimate Kf, Nf, and their respective standard errors. These observations coincided with the report from Cheah et al.[19] in describing paraquat adsorption on sandy loam and muck soils by the Freundlich equation, but were contrary to several studies which demonstrated that paraquat adsorption to soils followed the linear form of the Langmuir equation.[10][20] The adsorption isotherms were characterized by a decreasing slope as equilibrium concentration increased.

Graph: Figure 1. Paraquat adsorption isotherms for the different clay fractions of the study soil: (a) untreated; (b) H2O2 treated; (c) H2O2 + DCB treated. Differences between duplicate analyses are smaller than the symbols.

The shapes of paraquat adsorption isotherms on different clay fractions were L‐type.[21] Such adsorption behavior is explained by the high affinity of the clay for the paraquat at low concentration in which then decreases as concentration increases. Therefore, adsorption of additional molecules becomes more difficult due to the presence of already occupied specific sites and/or to the presence of less attractive sites. Different clay components of the study soil exposed highly energy sites to adsorb paraquat apparently associated with their CEC values (Table 2).

Table 2. Cation exchangeable capacities for the whole, coarse, and fine clay fractions

4 +Different letters in a column indicate significant difference between treatments at p = 0.05.

It is noted that the shapes of paraquat adsorption isotherms on fine clay fractions with different treatments in Figure 1 were H‐type that is an extreme version of the L‐type isotherms. This indicates that fine clay had the highest capacity to adsorb paraquat, while the coarse clay exhibited the least capacity. Paraquat sorption on the whole clay was intermediate between that observed for the coarse and fine clay fractions (Figure 1a). These retention capacities were identified by Kf values of the Freundlich equation associated with R2 values ranged from 0.79 to 0.96 in Table 3. In all treatments, the fine clay had the highest Kf values (ranged in 5.98–7.01 L/kg, respectively) while the coarse clay had the lowest Kf values (ranged in 0.59–1.21 L/kg, respectively). The adsorption isotherms for all samples are nonlinear, as indicated by the Nf values in which are much less than 1 to be below unity.

Table 3. Adsorption coefficients of the Freundlich equation for paraquat and their correlation (p < 0.05) in the study soil

5 +Different letters in a column indicate significant difference between treatments at p = 0.05. Numbers in the parentheses are standard deviations.

The sorption isotherms for the H2O2 treated samples indicate that the fine clay still exhibited stronger paraquat retention than did the whole and coarse fractions (Figure 1b). The H2O2 treatment in the whole and coarse clays showed less effect on Kf values, but the significantly (p < 0.05) lower Kf value was found in the H2O2 treated fine clay compared to the untreated fine clay (Table 3). If the samples were treated by H2O2, the CEC value of the fine clay was significantly lower than the untreated one, but not for the whole and coarse clays (Table 2). The study soil had low organic carbon content (14.4 g/kg) (Table 1), however, the organic matter gave more contribution in the fine clay fraction to adsorb paraquat with the corresponding CEC data (Table 2). Some studies had indicated that both CEC and organic matter content had positive correlations with inorganic and organic cations sorption.[22], [23] Therefore, this indicates that organic matter associated with fine clay fraction had high CEC contributing to relatively high affinity for paraquat.

Compared with the H2O2 treatment, the removal of free Fe compound by DCB treatment from the H2O2 treated clay enhanced paraquat adsorption for the whole and fine clay, but not for the coarse clay (Figure 1b and 1c). The whole and fine clays in the treatment of H2O2 + DCB had significantly higher Kf value than those in the treatment of H2O2 (Table 3), but the increase of the whole clay was estimated to result from the contribution of the fine fraction. This indicates that the DCB treatment created or exposed highly energy sites, apparently associated with the silicate clays. The surfaces of silicate clay was coated by free Fe compound, blocking access of paraquat to potential sorption sites. On the other hand, higher charge cations such as Al3 +and Fe3 +also electrostatically interconnect the macromolecules of soil organic matter, restricting diffusion of paraqut to exchange sites. The presence of activated C–H bond from the methyl groups in paraquat structure is evaluated to form hydrogen bonds (as C–H...O bonding) with oxygens of the siloxane surface of silicate clays. The coarse clay fraction did not show the significant differences among treatments, because kaolinite and quartz in the study Oxisol were the dominant clay minerals with weaker bindings to paraquat that coincided with Weber and Scott.[11] The strong binding of paraquat to the study soil was indicated by its little amount desorbed from different clay components ranged in 5.89–13.2% (Table 4), respectively. All samples had the lowest amounts of desorption in the treatment of H2O2 + DCB, respectively. Therefore, the observed constraint in desorption of paraquat suggests that ion exchange is not a major mechanism for retention of paraquat in the 0.01 M CaCl2 at pH 5. This is usually indicative of chemisorption reported by McBride.[24] In the meanwhile, this suggestion is the most clear in the treatment of H2O2 + DCB. Summarily, the silicate clay had the highest affinity for paraquat and free Fe compound had the lowest.

Table 4. Desorption of paraquat from the study soil

6 +Numbers in the parentheses are standard deviations.

The sorption isotherms of the clay components were fitted by the nonlinear Freundlich model coincided with past studies about the description of pesticide sorption on soil and sediment. The shape of paraquat adsorption isotherm on the < 0.2 μm fraction were H‐type, but their shapes on the < 2.0 μm and 0.2–2.0 μm fractions were L‐types. After organic matter removal, the whole and coarse clays retained more paraquat than did the fine clay. This indicates that organic matter associated with fine clay fraction had high CEC contributing to relatively high affinity for paraquat. The DCB treatment created high‐affinity sites for paraquat on the fine clay, but had little effect on paraquat sorption for the coarse clay. chemisorption is the major mechanism for retention of paraquat on clay components. However, the silicate clay had the highest affinity for paraquat and free Fe compound had the lowest.

The authors thank the National Scientific Council, Republic of China (Grant No. NSC89‐2316‐B‐020‐003) for providing the financial support in this study.