J. Agric. Food Chem. 2001, 49, 2899−2907 2899 Potential Contributions of Smectite Clays and Organic Matter to Pesticide Retention in Soils Guangyao Sheng,† Cliff T. Johnston,‡ Brian J. Teppen,§ and Stephen A. Boyd*,§ Department of Crop, Soil, and Environmental Sciences, University of Arkansas, Fayetteville, Arkansas 72701, Department of Agronomy, Purdue University, West Lafayette, Indiana 47907, and Department of Crop and Soil Sciences, Michigan State University, East Lansing, Michigan 48824 Soil organic matter (SOM) is often considered the dominant sorptive phase for organic contaminants and pesticides in soil-water systems. This is evidenced by the widespread use of organic-matter- normalized sorption coefficients (KOM) to predict soil-water distribution of pesticides, an approach that ignores the potential contribution of soil minerals to sorption. To gain additional perspective on the potential contributions of clays and SOM to pesticide retention in soils, we measured sorption of seven pesticides by a K-saturated reference smectite clay (SWy-2) and SOM (represented by a muck soil). In addition, we measured the adsorption of atrazine by five different K-saturated smectites and Ca-saturated SWy-2. On a unit mass basis, the K-SWy-2 clay was a more effective sorbent than SOM for 4,6-dinitro-o-cresol (DNOC), dichlobenil, and carbaryl of the seven pesticides evaluated, of which, DNOC was sorbed to the greatest extent. Atrazine was sorbed to a similar extent by K-SWy-2 and SOM. Parathion, diuron, and biphenyl were sorbed to a greater extent by SOM than by K-SWy-2. Atrazine was adsorbed by Ca-SWy-2 to a much lesser extent than by K-SWy-2. This appears to be related to the larger hydration sphere of Ca2+ (compared to that of K+) which shrinks the effective size of the adsorption domains between exchangeable cations, and which expands the clay layers beyond the apparently optimal spacing of ∼12.2 Å for sorption of aromatic pesticide structures. Although a simple relation between atrazine adsorption by different K-smectites and charge properties of clay was not observed, the highest charge clay was the least effective sorbent; a higher charge density would result in a loss of adsorption domains. These results indicate that for certain pesticides, expandable soil clays have the potential to be an equal or dominant sorptive phase when compared to SOM for pesticide retention in soil. Keywords: Sorption; pesticide; clay; organic matter; smectite INTRODUCTION tion coefficients (KOM) to predict soil-water distribution of pesticides implicitly assumes that SOM is the domi- Clay minerals and soil organic matter (SOM) are nant sorptive phase (4, 7, 8) and ignores the contribution considered the two most chemically active components of clays (hereafter used to indicate clay minerals rather of soils. Among clay minerals commonly found in soil than clay-sized particles) to pesticide sorption. The environments, expandable 2:1 layer silicate clays are surfaces of clays have often been viewed as being polar especially important because of their high surface areas in nature (9-12). It has been suggested that the and cation exchange capacities (CECs), as well as their preferential adsorption of water by these purportedly surface reactivities (1). Although often present in rela- polar surfaces rendered clays ineffective as sorbents for tively low amounts in soils, SOM disproportionately neutral organic compounds (13-15). influences many important soil processes, including sorption of aqueous-phase organic contaminants and The prevailing view of organic solute sorption by SOM pesticides. Currently, about 4.5 billion pounds of chemi- generally ignores the role of soil minerals (including cals are used as pesticides each year in the U.S., and clays) in the sorption of organic contaminants and agricultural usage accounts for ∼77% of the total (2). pesticides. Although this may be valid for relatively Sorption of soil-applied pesticides is an important nonpolar molecules (e.g., benzene and trichloroethyl- determinant of their environmental fate and behavior, ene), it is possible that other neutral organic molecules, including bioavailability, persistence, and potential for including important categories of pesticides, are ef- leaching. Over the past two decades, research on fectively adsorbed by clays even in the presence of bulk pesticide sorption by soil has focused mainly on the water. In fact, a few recent studies support this conten- singular role of SOM as the dominant sorptive phase tion. Atrazine adsorption by 13 Ca-saturated smectites (e.g., 3-6). Reliance on organic-matter-normalized sorp- ranged from very low to nearly complete removal from water, and was inversely dependent on clay properties such as surface charge density and cation-exchange * Corresponding author. Tel: 517-353-3993. Fax: 517-355- 0270. E-mail: boyds@msu.edu. capacity (16). With the measured pH of clay-water † University of Arkansas. suspensions ranging between 4.75 and 6.45, and a pKa ‡ Purdue University. of 1.7 for atrazine, it seemed apparent that atrazine was § Michigan State University. adsorbed as the neutral species. Atrazine adsorption by 10.1021/jf001485d CCC: $20.00 © 2001 American Chemical Society Published on Web 05/23/2001 2900 J. Agric. Food Chem., Vol. 49, No. 6, 2001 Sheng et al. a montmorillonite (from Clay Spur, Wyoming) saturated a reference smectite clay to that of a muck soil repre- with Ca2+ or Al3+ was attributed to the formation of senting SOM. These data were used to provide some H-bonds between atrazine and polarized waters of additional perspective on the potential importance of hydration (17). Nitro-substituted aromatics, including smectite clays and SOM as sorbents for aqueous-phase explosives and some dinitrophenol pesticides, are also pesticides. significantly adsorbed by clays (18-20). Adsorption of these compounds was shown to be affected by the type EXPERIMENTAL PROCEDURES of clay and its charge, the type of exchangeable cation, and the type and position of substituents on the Pesticides. Seven pesticides were used in the adsorption aromatic ring. High adsorption of nitroaromatics by experiments (Table 1). The pesticides were purchased from clays was attributed to the formation of an electron ChemService, Inc., West Chestnut, PA, with a purity of >99%, and used as received. Six of them were chosen to represent donor-acceptor complex, in which polarized aromatic major classes of pesticides. The seventh, biphenyl, was in- rings parallel to the basal surfaces accept electrons from cluded as a representative nonpolar unsubstituted pesticide. the siloxane oxygens to form the pesticide-clay complex. The pesticides were also selected to encompass a variety of In addition to the proposed electron donor-acceptor different structural and physicochemical properties. They mechanism, smectite clays may also adsorb neutral contained various substituents, including strongly electron- organic molecules by hydrophobic interactions. Recently, withdrawing groups such as -NO2 and -CN, and also have a Laird and Fleming (21) presented data on the sorption wide range of water solubilities, dipole moments, and octanol/ of butylpyridine from water by Ca-smectite. As much water partition coefficients (KOW). The gas-phase dipole mo- as 95% of added butylpyridine was adsorbed, and Ca2+ ments were computed using B3LYP density functional calcu- lations (28-30). These are quantum mechanical calculations release measurements indicated that ion exchange of reasonably good quality using the Becke three-parameter accounted for <20% of the amount sorbed. These results hybrid method (28) for the exchange functional along with the suggested a hydrophobic interaction between the butyl Lee, Yang, and Parr correlation functional (30). These elec- group of butylpyridine and the siloxane surface of tronic structure calculations show that molecular thicknesses, smectite. Earlier work by Jaynes and Boyd (22) indi- as estimated by the 0.02 electron/Å3 isodensity surfaces, were cated that the siloxane surfaces of smectites possessed 3.5 (0.2 Å for 4,6-dinitro-o-cresol (DNOC) and a wide variety a hydrophobic character consistent with the results of of other planar nitroaromatic molecules. To estimate the Laird and Fleming (21). effective surface area occupied by DNOC on clay surfaces, we used our force field (31) for organic-clay interactions to perform The studies cited above, along with substantial ex- molecular mechanics calculations. We used Grand Canonical perimental evidence from earlier studies (e.g., 23, 24), Monte Carlo calculations (32) to model adsorption of DNOC clearly documents the ability of pure clays, particularly into an uncharged clay with just enough interlayer space (12.8 expandable 2:1 clays, to effectively bind organic mol- Å d spacing) for the DNOC. For three simulations at two ecules, including pesticides. These studies suggest the different unit cell sizes (in-layer unit cell vector products ab potential importance of clays in organic contaminant were 4534 and 5922 Å2), we found similar values of 51.9, 54.3, and pesticide retention by soils. Using sorption data and 52.1 Å2 per DNOC. These values should be the upper limit, as the DNOC packing was not perfect. In summary, the surface from soils of variable clay and SOM contents, Karickhoff area of DNOC should be e 52 Å2, and its thickness should be (25) attempted to evaluate conditions under which about 3.5 Å. The interlayer space available in a 12.5-Å mineral-phase sorption of organic contaminants in monolayer hydrate clay mineral can be estimated by compar- whole soils was important. For the N-heterocyclic si- ing the 9.19-Å d spacing of pyrophyllite (nothing in the mazine and biquinoline, mineral contribution to overall interlayer), which yields about 3.3 Å for the available inter- sorption became apparent (i.e., greater than that pre- layer space, agreeing moderately well with the thickness of dicted from Kom values) at clay/organic matter ratios of the organic molecule. >30. Sorption of pyrene, however, was not affected by Sorbents. A reference smectite, SWy-2, was used in most clay content. It was concluded that the contribution of of the measurements of pesticide adsorption by clays. Five mineral-phase sorption was a direct function of the other clays were also studied for atrazine adsorption measure- ments. The clays were chosen to provide a range of surface polarity of the compound. Grundl and Small (26) evalu- charge densities, and differences in the location of charge ated the role of mineral-phase sorption of atrazine and deficit and the structure of octahedral sheet (dioctahedral vs alachlor by a suite of sediments, and concluded that the trioctahedral) (Table 2). All clays were obtained from the Clay pesticides were sorbed by both natural organic carbon Minerals Society Source Clay Repository (Columbia, MO) (OC) and clays. The critical clay/OC ratios at which except beidellite which was from Ward’s (Rochester, NY). The mineral-phase sorption accounted for 50% of the overall <2 µm clay particles were separated by wet sedimentation. sorption were ca. 62 for atrazine and 84 for alachlor. The clays were subject to K+ saturation, and also Ca2+ Hassett et al. (27) evaluated the effect of clay content saturation for SWy-2, by dispersing 10-g clay samples in 1 L on sorption of R-naphthol by soils and sediments with of KCl (0.1 M) or CaCl2 (0.1 M) solution. The clay suspensions were shaken for 24 h, and then fresh chloride solutions were different OC and montmorillonite contents. They con- used to displace the original solutions after centrifugation. This cluded that the contribution of mineral-phase sorption process was repeated four times to ensure complete K- or Ca- to overall uptake was apparent at clay/OC ratios of >10. saturation. Distilled water (1 L) was used to wash the clays Although these studies established the importance of to remove excess K+ or Ca2+. The clays were freeze-dried and clays as sorbents for organic contaminants in soils, they stored for later use. The Houghton muck soil was collected did not isolate or quantify the individual sorptive from the Michigan State University Muck Farm and air-dried. contributions of clays and SOM. Its OC content (49.5%) was measured by a carbon analyzer (Rosemount Analytical, Inc., Santa Clara, CA). The adsorption of aqueous phase pesticides by clays Sorption Isotherms. Pesticides were dissolved in 0.1 M differs significantly among the various combinations of KCl solution with concentrations up to 50% of their water pesticides and clays. It is often unclear whether the solubilities. A pH 3 KCl solution was used for DNOC. In the sorption of pesticides by clays is comparable to that by case of Ca-SWy-2 and the muck soil, 0.1 M CaCl2 was used. SOM. In this study, representative pesticides from Up to 4.4 mL of KCl (or CaCl2) solution was pipetted into the several major classes were used to compare sorption by 7-mL borosilicate glass vials containing clay or muck soil (0.05 Pesticides in Smectite Clays and Organic Matter J. Agric. Food Chem., Vol. 49, No. 6, 2001 2901 Table 1. Selected Physicochemical Properties of Pesticides Used for Sorption Study Table 2. Properties of Clays and the Freundlich Coefficients for Atrazine Adsorption by K-clays cation-exchange % Kf capacity tetrahedral tri-/di- (mg/kg)/ sorbent mineralogy (cmolc/kg) charge octahedral (mg/L)n n K-SHCa-1 hectorite 43.9 0 tri- 120 0.821 K-SWy-2 montmorillonite 81.6 0 di- 74.6 0.784 K-BPC beidellite 84.2 52 di- 31.4 0.925 K-SapCa-2 saponite 94.9 >50 tri- 135 0.841 K-SWa-1 nontronite 107 73 di- 95.3 0.899 K-SAz-1 montmorillonite 130 0 di- 8.47 0.864 to 1 g). Pesticide solution was then added into each vial to After sampling for HPLC analysis, the remaining DNOC- make up a total volume of 5 mL. For DNOC (pKa 4.35 to 4.46), K-SWy-2 suspensions were used for X-ray diffraction analysis. sorption was measured at pH 3 to ensure the molecular form The supernatants of 1-2 mL were retained in the vials to of the pesticide. The pH of clay suspension before DNOC resuspend the clay by hand-shaking, and then dropped on sorption was adjusted until it remained stable at ∼3, and was glass and air-dried overnight to obtain the oriented films. X-ray also measured after sorption to further ensure the stable pH. diffraction patterns were recorded using Cu-KR radiation and Clay dissolution under such a pH was not noted. Consistent a Philips APD 3720 automated X-ray diffractometer using an with previous studies (18, 19) we found that DNOC sorption APD 3521 goniometer fit with a θ-compensating slit, a 0.2- by clays reached equilibrium within 10 min. In the case of the mm receiving slit, and a diffracted-beam graphite monochro- muck soil the pH of the aqueous CaCl2 suspension was mator, from 3 to 14 °2θ, in steps of 0.02 °θ, at 1 s/step. adjusted to 3.0 ( 0.1 before DNOC solution was introduced. The pH adjustment was repeated as needed until it remained RESULTS AND DISCUSSION stable after 3 days. The vials were continuously rotated overnight at room temperature and then centrifuged at 1667g Pesticide sorption isotherms based on per unit mass for 20 min to separate the liquid and solid phases. The of clay or muck (SOM) versus the equilibrium aqueous concentrations of pesticides in supernatants were analyzed, concentration of each pesticide are shown in Figure 1- by direct injection of supernatants (between 10 and 190 µL), (a-g). The Houghton muck soil was used to evaluate using a Perkin-Elmer reversed-phase HPLC (Perkin-Elmer, the potential contribution of soil organic matter (SOM) Norwalk, CT) fitted with an UV-visible detector set at the to pesticide sorption. The representation of SOM by the maximum absorption wavelength for each pesticide (Table 1). muck soil is supported by the similarities of organic- A platinum extended polar selectivity (EPS) C18 column was used. The mobile phase was a mixture of methanol and water carbon-normalized sorption coefficients (Koc) between a ranging from 55% to 75% methanol with a flow rate of 1.0 mL/ peat soil (49.3% OC) and a mineral soil (1.26% OC) for min. The amount of pesticide sorption was calculated from the both nonpolar (ethylene dibromide, EDB) and polar difference between the amount added and that remaining in (dichlorophenol, DCP) compounds (33). The reported log the final solution. Koc values for the peat and mineral soils were 1.28 and 2902 J. Agric. Food Chem., Vol. 49, No. 6, 2001 Sheng et al. Figure 1. Sorption isotherms representing pesticide uptake from water by a reference homoionic K-smectite (SWy-2) and muck soil representing soil organic matter (a-g), and atrazine uptake by several different K-saturated smectites and Ca-SWy-2 (h). 1.23 for EDB and 2.03 and 1.87 for DCP, respectively. Table 3. Distribution Coefficients (L/kg) of Pesticides The isotherms representing pesticide sorption on the Sorbed by K-SWy-2, Muck, or Ca-SWy-2 at the Relative muck soil display some nonlinearity. Isotherm nonlin- Concentration of 0.1a earity has been observed previously for polar and pesticide KK-SWy-2 Kmuck KK-SWy-2/Kmuck nonpolar organic solutes and attributed to a glassy 4,6-dinitro-o-cresol 2.49 × 103 184 13.6 phase in SOM (5, 6, 34, 35), to the presence of a small carbaryl 235 54.2 4.34 quantity of high-surface-area carbonaceous material (33, diuron 103 173 0.593 36), and/or to solute-SOM specific interactions (36, 37). atrazine 54.2 (7.68)b 47.1 1.15 The isotherms for pesticide sorption by K-SWy-2 are dichlobenil 275 179 1.54 either linear, type I, or type III (38), depending on the parathion 125 1.08 × 103 0.116 biphenyl 6.40 791 8.09 × 10-3 specific compound. This reflects different types and/or strengths of pesticide-clay interactions. The calculated a Relative concentration equals equilibrium aqueous-phase distribution coefficients (expressed as the ratio of the concentration/water solubility. b Distribution coefficient for sorp- concentration of sorbed pesticide to the aqueous-phase tion by Ca-SWy-2. pesticide concentration) for both K-SWy-2 and muck is the ratio of the two distribution coefficients for each at a relative concentration (i.e., the ratio of the aqueous- pesticide and allows comparison of the relative effective- phase concentration to water solubility) of 0.1 are given ness of K-SWy-2 clay and SOM as sorbents for pesti- in Table 3 for each pesticide. The term KK-SWy-2/Kmuck cides. It shows that some pesticides, i.e., DNOC, car- Pesticides in Smectite Clays and Organic Matter J. Agric. Food Chem., Vol. 49, No. 6, 2001 2903 baryl, and dichlobenil, were more effectively sorbed by K-SWy-2 clay than by SOM. Others, i.e., diuron, parathion, and biphenyl, were sorbed more effectively by SOM than by clay. Atrazine is sorbed by K-SWy-2 clay and SOM to a similar extent. Direct correlations between pesticide adsorption (Kf values) by K-SWy-2 and water solubility, KOW, or dipole moment were not found, reflecting the complexity of pesticide-clay in- teractions. Biphenyl was highly sorbed by the muck soil. In contrast, K-SWy-2 clay was a relatively ineffective sorbent for biphenyl (Figure 1g). The smectite was approximately 120 times less effective than SOM for biphenyl sorption (Table 3). Biphenyl is a nonpolar, poorly water-soluble aromatic molecule. These charac- teristics manifest its high uptake by SOM, consistent with previous findings (39). These molecular properties, however, were apparently not sufficient for significant sorption by K-SWy-2, despite indications that the siloxane surfaces of smectites are hydrophobic in nature (21, 22). This suggests that, in addition to hydrophobic interactions, other mechanisms may be necessary for substantial retention by smectites. Sorption of DNOC by K-SWy-2 was highest among all pesticides studied (Figure 1a), although it is pH- dependent due to ionization of the phenolic hydroxyl. For example, raising the pH from 3 to 5 caused ∼50% reduction in the sorption of DNOC by K-SWy-2 (data not shown), but the sorption was still substantially higher than that by SOM. We have observed similar or slightly higher sorption of the nonionizable compounds Figure 2. (a) 4,6-dinitro-o-cresol adsorption-dependent varia- 1,3- and 1,4-dinitrobenzene by K-SWy-2 (40). Based on tion in the basal spacing of K-SWy-2, and (b) selected X-ray distribution coefficients at a relative aqueous concentra- diffraction patterns. tion of 0.1 (Table 3), sorption of DNOC was ∼9 times f) greater than the next most highly sorbed pesticide 49 mg/g (dichlobenil); the difference is even greater at lower 198 g/mol × 10-3 mol/mmol × 6.023 × 1023/mol × 52 Å2 × 10-20 m2/Å2 concentrations. Also, K-SWy-2 was ∼14 times more 750 m2 effective than SOM as a sorbent for DNOC (Table 3). × 100% ) 10.3% The basal spacing of K-SWy-2 increased gradually from ∼11.1 to ∼12.2 Å with increasing DNOC adsorp- If the aromatic ring of DNOC is oriented parallel to the siloxane surface, and it simultaneously interacts with tion (Figure 2a). Selected X-ray diffraction (XRD) pat- the opposing clay layers, as much as 21% of the smectite terns (Figure 2b) show that the diffraction peaks were clay surface may be occupied at this loading. broad and generally symmetrical, indicating the random The reasons that K-SWy-2 is a highly effective interstratification of DNOC molecules in the interlayers adsorbent for DNOC are not fully understood. The of K-SWy-2. The X-ray diffraction peaks corresponded molecule is planar and aromatic, and it has two nitro to distributions of clay mineral d spacings from roughly groups that are strongly electron-withdrawing. Accord- 10 to 13 Å. Thus, the clay films typically contained a ing to Haderlein and Schwarzenbach (18), these molec- continuum of domains ranging from dehydrated K- ular characteristics favor the formation of an electron smectite (10 Å) to an interlayer containing a monolayer donor-acceptor complex, in which the aromatic mol- of K+, water, and/or DNOC (12 to 13 Å). As the DNOC ecule acts as an acceptor of electrons donated from sites loading increased, the d spacing corresponding to the of negative charge in the clay. FTIR studies showed that centroid of this diffraction peak increased to 12.2 Å for the aromatic ring of adsorbed DNOC molecule is nearly the DNOC loading corresponding to a solution-phase parallel with respect to the siloxane surfaces of K-SWy-2 DNOC concentration of 20 mg/L. The increase in basal (41). The partial negative charge associated with the nitro groups may result in electrostatic interactions spacing suggests that DNOC is intercalated in the between the nitro groups and interlayer K+. This is interlamellar region of the clay, with an increasing indicated by the fact that the FTIR stretching vibration fraction of the clay domains at larger d spacings (12 to bands of nitro groups shift depending on the interlayer 13 Å). The surface area of K-SWy-2 is 750 m2/g and cation (42). Thus, the high effectiveness of K-SWy-2 the estimated cross-sectional area of the DNOC mol- for DNOC appears to result from a combination of more ecule is 52 Å2. The surface area occupied by DNOC than one mechanism. molecules (f) was ca. 10%. This was calculated as Two other pesticides, carbaryl and dichlobenil, are follows: also more favorably sorbed by K-SWy-2 than by SOM 2904 J. Agric. Food Chem., Vol. 49, No. 6, 2001 Sheng et al. (Figure 1b and e). The dichlobenil molecule may form to 15 Å and beyond at 100% humidity in the absence of an electron donor-acceptor complex with the siloxane organic sorbates. All Ca-smectites, on the other hand, surfaces because of the strong electron-withdrawing swell to more than 15 Å. Because there is apparently a character of the cyano group on the aromatic ring, in fine line between ∼12.5-Å or larger swelling for K- analogy with nitroaromatics. The two-ring π-electron smectites, it is quite possible that the presence of system of carbaryl may participate in the formation of pesticides could cause the monolayer structure to be an electron donor-acceptor complex depending on the more favored, even in cases where the smectite might inductive and resonance properties of the N-methylcar- swell further in the absence of pesticide. In this scenario, bamate (-OCONHCH3) moiety. The functional groups each pesticide molecule contacts both clay surfaces of carbaryl and dichlobenil may also interact with the simultaneously and thereby avoids contact with most exchangeable cations in analogy with DNOC, but we water molecules. As the free energy of hydration for have no direct evidence for this currently. many small organic solutes is in the range of +10 to Atrazine adsorption by K-SWy-2 was comparable to +30 kJ/mol (45, i.e., n-hexane, +27.5; cyclohexane, that by SOM (Figure 1d) but considerably less than that +23.7; toluene, +18.4 kJ/mol), removal of these solutes for DNOC. In comparison to DNOC, atrazine is larger from aqueous solution may well provide enough energy and more structurally complex. In addition to the to prevent K-smectite from swelling beyond 12.5 Å. For presence of -Cl, atrazine possesses two large substi- a Ca-smectite, in contrast, the energy penalty for tuted amino groups. These three substituents are compressing the interlayer and thereby dehydrating the positioned meta to each other on the triazine ring and pesticide (and the cation) is presumably too large. may manifest a degree of steric hindrance in the The specific type of smectite clay also affects pesticide adsorption of atrazine by K-SWy-2. Steric effects were adsorption. This is illustrated by atrazine adsorption also found in the adsorption of nitroaromatics with ortho by six different species of K-smectites (Figure 1h). In a substitution or large alkyl substituents (18, 19). Laird study of atrazine adsorption by 13 Ca-saturated smec- et al. (16) reported that smectite clays with lower charge tites, Laird et al. (16) found that the logarithm of the densities and CECs were more effective adsorbents of Freundlich adsorption constant (Kf) was inversely cor- atrazine; these characteristics may increase the size of related to clay CEC. We obtained Kf values (ranging the adsorptive domains between exchangeable cations. from 8.5 to 135 L/kg, Table 2) for atrazine adsorption The adsorption of diuron and parathion is comparably to six K-smectites, but we did not observe an obvious low and less extensive by K-SWy-2 than by SOM relationship between Kf and CEC. However, it was (Figure 1c and f). These two pesticides contain large apparent that the clay with the highest CEC (and substituents on the aromatic ring (two -Cls and N,N- charge density), i.e., K-SAz-1, was the least effective dimethylurea for diuron, nitro and diethylphospho- adsorbent for atrazine. For this clay, it seems likely that rothionate for parathion) (Table 1) which may sterically atrazine adsorption decreased because of loss of adsorp- hinder adsorption. The presence of bulky alkyl substit- tion domains of sufficient size to accommodate atrazine. uents has been observed to substantially diminish the Decreasing adsorption with increasing clay layer charge adsorption of nitroaromatics by clay minerals (19). The has been observed previously for the adsorption of type III isotherm for parathion adsorption by K-SWy-2 nonpolar aromatics from water by tetramethylammo- indicates weak parathion-clay interactions. However, nium- and trimethylphenylammonium-smectites (22, both of these pesticides are sorbed by K-SWy-2 to a 49). greater extent than is biphenyl. The potential contributions of smectite clay (K-SWy- 2) and SOM to pesticide retention in soils (at relative Adsorption of atrazine was affected by the type of pesticide concentration of 0.1, Table 3) can be calculated exchangeable cation on the smectite clay. Homoionic by the following equations, assuming these sorbent Ca-SWy-2 was a much less effective sorbent for atra- phases function independently: zine than homoionic K-SWy-2 (Figure 1h). Atrazine adsorption indicates that K-SWy-2 (Kf ) 54.2) is seven Cclay ) KK-SWy-2 × Ce × clay % times more effective than Ca-SWy-2 (Kf ) 7.68) (Table 3). Similar effects were observed by Haderlein and CSOM ) Kmuck × Ce × SOM % Schwarzenbach (18) and Haderlein et al. (19) for the adsorption of nitroaromatics by clays saturated with where Cclay and CSOM are the respective clay-sorbed and different cations. The lower effectiveness of Ca-SWy-2 SOM-sorbed pesticide concentrations normalized to the may be due in part to the higher hydration energy and whole soil at the aqueous phase concentration of Ce. hence larger hydrated radius of Ca2+ than that of K+. Calculations using the distribution coefficients listed in Presumably, the waters of hydration associated with Table 3 for model (hypothetical) soils over a range of Ca2+ obscure a greater portion of the clay surface than SOM (0-10%) and smectite clay (0-40%) contents that those associated with K+, and this may effectively might reasonably be found in mineral soils are graphi- shrink the size of the adsorptive domains between cally shown in Figure 3. These plots can be used to exchangeable cations. Similar effects were observed by compare the potential contributions of smectite clays Kukkadapu and Boyd (43) for the adsorption of aromatic and SOM to the pesticide retention in a soil with known and chlorinated hydrocarbons by tetramethylphospho- smectite clay and SOM contents. For example, from nium- and tetramethylammonium-smectites. Figure 3, we estimate that clay (K-SWy-2) contributes Differences in swelling behavior between K- and Ca- (∼390 mg/kg) ∼35 times more to carbaryl retention than SWy-2 may also contribute to their differential pesticide SOM (∼11 mg/kg) in a soil with 16% clay and 2% SOM. adsorption. Many K-smectites equilibrate with ∼12.5-Å For parathion in the same soil, contributions by smectite layer spacings (monolayer structures) at 100% humidity clay and SOM are about equivalent. For some pesticides, and even in aqueous suspension (44-47), although some even low amounts of smectite clay could potentially low-charge K-smectites can swell more. For example, exert a strong influence on sorption. For example, K-hectorite (48) and K-SWy-2 (47) are likely to swell assuming a soil containing 2% K-SWy-2 and 2% SOM, Pesticides in Smectite Clays and Organic Matter J. Agric. Food Chem., Vol. 49, No. 6, 2001 2905 Figure 3. Potential contributions of smectite clay (represented by K-smectite, SWy-2) and soil organic matter (SOM, represented by a muck soil) to pesticide immobilization in model (hypothetical) soils. the amounts of DNOC sorbed at the relative concentra- cations (e.g., Ca2+) because of differences in cation tion of 0.1 are ca. ∼980 mg/kg by clay and ∼70 mg/kg hydration. Thus, K-rich domains with high affinity for by SOM. Although this illustrates the potential domi- pesticides such as nitroaromatics are likely to exist in nant role of certain clay minerals in the retention of many soils. Indeed, the widely occurring phenomenon specific pesticides in soils, caution must be taken in of K-fixation by soil clays (53-55) proves that K-rich extrapolating these results. Humic substances in soils domains are common. Even for a variety of Ca-saturated may form coatings on clay particles thereby rendering smectites, Laird et al. (16) reported very high removal these surfaces unavailable for pesticide adsorption. of atrazine from water, so the efficacy of smectites as However, this effect is probably limited to the external adsorbents of pesticides is not limited to K-saturated surfaces of expandable clays, as there is no evidence for clays. Improved understanding of pesticide adsorption the intercalation of humics by naturally occurring clays. by clays makes it possible to enhance the retentive This effect is further minimized in subsoils and aquifer properties of subsoils and aquifer materials for certain materials with inherently low organic matter contents. pesticides by controlling the base saturation of the In considering the applicability of these results to real matrix via in-situ electrolyte injection (e.g., KCl) as soil clays, K+ is one of the most abundant exchangeable suggested by the recent findings of Weissmahr et al. cations in soils and subsoils, but it is not normally the (56). dominant exchangeable cation. The phenomenon of In summary, pesticides can be sorbed effectively by cation demixing in clays (50-52) results in entire either clays or SOM. Sorption depends on the specific interlayer regions being occupied by a single exchange- type of clay, saturating cation, and pesticide structure. able cation (e.g., K+) even in the presence of other Whereas pesticide sorption by SOM is generally deter- 2906 J. Agric. Food Chem., Vol. 49, No. 6, 2001 Sheng et al. mined by pesticide water solubility, adsorption by clays (9) Hendricks, S. B.; Jefferson, M. E. Structure of kaolin is dependent on both pesticide and clay properties. Our and talc-pyrophyllite hydrates and their bearing on study builds upon previous work to probe the depen- water sorption of clays. Am. Mineral. 1938, 23, 863- dence of pesticide sorption on the structure of the 875. (10) Bradley, W. F. Molecular associations between mont- pesticide, on the exchangeable cation saturating the morillonite and organic liquids. J. Am. Chem. Soc. 1945, smectite clay, and on the smectite clay layer charge. 67, 975-981. Understanding these dependencies will help constrain (11) Low, P. F. Physical chemistry of clay-water interaction. the interpretation of more mechanistic spectroscopic Adv. Agron. 1961, 13, 269-327. studies underway in our laboratories. The existence of (12) Yan, L.; Roth, C. B.; Low, P. F. Effects of monovalent multiple sorption mechanisms appears to favor pesticide exchangeable cations and electrolytes on the infrared sorption by smectite clays. Planar aromatic structures vibrations of smectite layers and interlayer water. J. and electron withdrawing substituents (e.g. -NO2, Colloid Interface Sci. 1996, 184, 663-670. (13) Chiou, C. T.; Shoup, T. D.; Porter, P. E. Mechanistic roles -CN) seem to favor pesticide adsorption, possibly via of soil humus and minerals in the sorption of nonionic the formation of an electron donor-acceptor complex organic compounds from aqueous and organic solutions. between pesticide molecules and siloxane surfaces. Polar Org. Geochem. 1985, 8, 9-14. substituents may interact via a H2O-bridge with hy- (14) Chiou, C. T. Roles of organic matter, minerals, and drated exchangeable cations; substituents with negative moisture in sorption of nonionic compounds and pesti- charge character (e.g. -NO2) may interact directly with cides by soil. In Humic Substances in Soil and Crop exchangeable cations. Hydrophobic interactions between Sciences; MacCarthy, P., Clapp, C. E., Malcolm, R. L., the pesticide and the siloxane surfaces may also con- Bloom, P., Eds.; American Society of Agronomy: Madi- tribute to adsorption. Large substituents associated son, WI, 1990; pp 111-160. (15) Chiou, C. T. Soil sorption of organic pollutants and with pesticide structure may cause steric constraints pesticides. In Encyclopedia of Environmental Analysis that diminish adsorption. A steric effect may also arise and Remediation; Meyers, R. A., Ed.; John Wiley & from the hydration of cations saturating clays, i.e., a Sons: New York, 1998; 4517-4554. large hydration sphere may diminish the size of the (16) Laird, D. A.; Barriuso, E.; Dowdy, R. H.; Koskinen, W. adsorption domains between exchangeable cations. Ex- C. Adsorption of atrazine on smectites. Soil Sci. Soc. Am. changeable cations may also influence sorption due to J. 1992, 56, 62-67. effects on basal spacing. A spacing of ∼12.2 Å, such as (17) Sawhney, B. L.; Singh, S. S. Sorption of atrazine by Al- that associated with K-smectites, appears optimal for and Ca-saturated smectite. Clays Clay Miner. 1997, 45, 333-338. adsorption of nitro-aromatic pesticides. Finally, low (18) Haderlein, S. B.; Schwarzenbach, R. P. Adsorption of surface charge density may increase the size of the substituted nitrobenzenes and nitrophenols to mineral adsorptive domain, hence increasing adsorption. In this surfaces. Environ. Sci. Technol. 1993, 27, 316-326. limited study of seven pesticides and reference smectite (19) Haderlein, S. B.; Weissmahr, K. W.; Schwarzenbach, R. clays, the K-SWy-2 clay was shown to be a more P. Specific adsorption of nitroaromatic explosives and dominant sorptive phase than SOM in over half the pesticides to clay minerals. Environ. Sci. Technol. 1996, cases. 30, 612-622. (20) Weissmahr, K. W.; Haderlein, S. B.; Schwarzenbach, R. P. Complex formation of soil minerals with nitroaro- LITERATURE CITED matic explosives and other π-acceptors. Soil Sci. Soc. Am. J. 1998, 62, 369-378. (1) McBride, M. B. Environmental Chemistry of Soils; (21) Laird, D. A.; Fleming, P. D. Mechanisms for adsorption Oxford University Press: New York, 1994. of organic bases on hydrated smectite surfaces. Environ. (2) Aspelin, A. L. Pesticides Industry Sales and Usage: 1994 Toxicol. Chem. 1999, 18, 1668-1672. and 1995 Market Estimates; EPA 733-K-97-002; Office (22) Jaynes, W. F.; Boyd, S. A. Hydrophobicity of siloxane of Prevention, Pesticides & Toxic Substances, U.S. surfaces in smectites as revealed by aromatic hydrocar- Environmental Protection Agency, U.S. Government bon adsorption from water. Clays Clay Miner. 1991, 39, Printing Office: Washington, DC, 1997. 428-436. (3) Chiou, C. T.; Peters, L. J.; Freed, V. H. A physical (23) Mortland, M. M. Clay-organic complexes and interac- concept of soil-water equilibria for nonionic organic tions. Adv. Agron. 1970, 22, 75-117. compounds. Science 1979, 206, 831-832. (24) Bailey, G. W.; White, J. L. Factors influencing the (4) Chiou, C. T.; Porter, P. E.; Schmedding, D. W. Partition adsorption, desorption, and movement of pesticides in equilibria of nonionic organic compounds between soil soils. In Residue Review - Residues of Pesticides and organic matter and water. Environ. Sci. Technol. 1983, Other Foreign Chemicals in Food and Feeds; Gunther, 17, 227-231. F. A., Gunther, J. D., Eds.; Springer-Veslay: New York, (5) Weber, W. J., Jr.; McGinley, P. M.; Katz, L. E. A 1970; Vol 32, pp 29-92. distributed reactivity model for sorption by soils and (25) Karickhoff, S. W. Organic pollutant adsorption in aque- sediments. 1. Conceptual basis and equilibrium assess- ous systems. J. Hydraul. Engr. 1984, 110, 707-735. ments. Environ. Sci. Technol. 1992, 26, 1955-1962. (26) Grundl, T.; Small, G. Mineral contributions to atrazine (6) Xing, B.; Pignatello, J. J.; Gigliotti, B. Competitive and alachlor sorption in soil mixtures of variable organic sorption between atrazine and other organic compounds carbon clay content. J. Contam. Hydrol. 1993, 14, 117- in soils and model sorbents. Environ. Sci. Technol. 1996, 128. 30, 2432-2440. (27) Hassett, J. J.; Banwart, W. L.; Wood, S. G.; Means, J. (7) Briggs, G. G. Theoretical and experimental relationships C. Sorption of R-naphthol: Implications concerning the between soil adsorption, octanol-water partition coef- limits of hydrophobic sorption. Soil Sci. Soc. Am. J. ficients, water solubilities, bioconcentration factors, and 1981, 45, 38-42. parachor. J. Agric. Food Chem. 1981, 29, 1050-1059. (28) Becke, A. D. Density-Functional Thermochemistry. 3. (8) Hassett, J. J.; Banwart, W. L. The sorption of nonpolar The Role of Exact Exchange. J. Chem. Phys. 1993, 98, organics by soils and sediments. In Reactions and 5648-5652. Movement of Organic Chemicals in Soils; Sawhney, B. (29) Kohn, W.; Becke, A. D.; Parr, R. G. Density functional L., Brown, K., Eds.; American Society of Agronomy: theory of electronic structure. J. Phys. Chem. 1996, 100, Madison, WI, 1989; pp 31-44. 12974-12980. Pesticides in Smectite Clays and Organic Matter J. Agric. Food Chem., Vol. 49, No. 6, 2001 2907 (30) Lee, C. T.; Yang, W. T.; Parr, R. G. Development of the (44) Suquet, H.; de la Calle, C.; Pezerat, H. Swelling and Colle-Salvetti Correlation-Energy Formula Into a Func- structural organization of saponite. Clays Clay Miner. tional of the Electron-Density. Phys. Rev. 1988, B37, 1975, 23, 1-9. 785-789. (45) MacEwan, D. M. C.; Wilson, M. J. In Crystal Structures (31) Teppen, B. J.; Rasmussen, K.; Bertsch, P. M.; Miller, of Clay Minerals and Their X-ray Identification; Brind- D. M.; Scha¨fer, L. Molecular dynamics modeling of clay ley, G. W., Brown, G., Eds., Mineralogical Society: minerals. 1. Gibbsite, kaolinite, pyrophyllite, and beid- London, UK, 1980; pp 197-248. ellite. J. Phys. Chem. 1997, B101, 1579-1587. (46) Suquet, H.; Pezerat, H. Parameters influencing layer (32) Frenkel, D.; Smit, B. Understanding molecular simula- stacking types in saponite and vermiculite: a review. tion: From algorithms to applications; Academic Clays Clay Miner. 1987, 35, 353-362. Press: San Diego, CA, 1996. (47) Sato, T.; Watanabe, T.; Otsuka, R. Effects of layer (33) Chiou, C. T.; Kile, D. E. Deviations from sorption charge, charge location, and energy change on expansion linearity on soils of polar and nonpolar organic com- properties of dioctahedral smectites. Clays Clay Miner. pounds at low relative concentrations. Environ. Sci. 1992, 40, 103-113. Technol. 1998, 32, 338-343. (48) Tardy, Y.; Touret, O. In Proceedings of the International (34) Xing, B.; Pignatello, J. J. Dual-mode sorption of low- Clay Conference; Schultz, L., van Olphen, H., Mumpton, polarity compounds in glassy poly(vinyl chloride) and F., Eds.; The Clay Minerals Society: Bloomington, IN, soil organic matter. Environ. Sci. Technol. 1997, 31, 1985; pp 46-52. 792-799. (49) Lee, J.-F.; Mortland, M. M.; Chiou, C. T.; Kile, D. E.; (35) Weber, W. J., Jr.; Huang, W. A distributed reactivity Boyd, S. A. Adsorption of benzene, toluene and xylene model for sorption by soils and sediments. 4. Intrapar- by two tetramethylammonium-smectites having differ- ticle heterogeneity and phase-distribution relationships ent charge densities. Clays Clay Miner. 1990, 38, 113- under nonequilibrium conditions. Environ. Sci. Technol. 120. 1996, 30, 881-888. (50) Glaeser, R.; Mering, J. Isotherms d’hydration des mont- (36) Chiou, C. T.; Kile, D. E.; Rutherford, D. W.; Sheng, G.; Boyd, S. A. Sorption of selected organic compounds from morillonites bi-ionique (Na, Ca). Clay Minerals Bulletin water to a peat soil and its humic-acid and humin 1954, 2, 188-193. fractions: Potential sources of the sorption nonlinearity. (51) Fink, D. H.; Nakayama, F. S.; McNeal, B. L. Demixing Environ. Sci. Technol. 2000, 34, 1254-1258. of exchangeable cations in free-swelling bentonite clay. (37) Spurlock, F. C.; Biggar, J. W. Thermodynamics of Soil Sci. Soc. Am. Proc. 1971, 35, 552-555. organic chemical partition in soils. 2. Nonlinear partition (52) Levy, R.; Francis, C. Demixing of sodium and calcium of substituted phenylureas from aqueous solution. En- ions in montmorillonite crystallites. Clays Clay Miner. viron. Sci. Technol. 1994, 28, 996-1002. 1975, 23, 475-476. (38) Gregg, S. J.; Sing, K. S. W. Adsorption, Surface Area, (53) Eberl, D. D. Alkali cation selectivity and fixation by clay and Porosity; Academic Press: New York, 1982. minerals. Clays Clay Miner. 1980, 28, 161-172. (39) Chiou, C. T.; McGroddy, S. E.; Kile, D. E. Partition (54) Goulding, K. W. T. Thermodynamics and potassium characteristics of polycyclic aromatic hydrocarbons on exchange in soils and clay minerals. Adv. Agron. 1983, soils and sediments. Environ. Sci. Technol. 1998, 32, 36, 215-264. 264-269. (55) Bouabid, R.; Badraoui, M.; Bloom, P. R. Potassium (40) Boyd, S. A.; Sheng, G.; Teppen, B. J.; Johnston, C. T. fixation and charge characteristics of soil clays. Soil Sci. Mechanisms for the adsorption of substituted nitroben- Soc. Am. J. 1991, 55, 1493-1498. zenes by smectite clays. Environ. Sci. Technol. 2001, (56) Weissmahr, K. W.; Hilderbrand, M.; Schwarzenbach, R. submitted for publication. P.; Haderlein, S. B. Laboratory and field scale evaluation (41) Sheng, G.; Johnston, C. T.; Teppen, B. J.; Boyd, S. A. of geochemical controls on groundwater transport of Adsorption of dinitrophenol pesticides by montmorillo- nitroaromatic ammunition residues. Environ. Sci. Tech- nites. Clays Clay Miner. 2001, in press. nol. 1999, 33, 2593-2600. (42) Johnston, C. T.; deOliveira, M. F.; Teppen, B. J.; Sheng, G.; Boyd, S. A. Spectroscopic study of nitroaromatic- smectite sorption mechanisms. Environ. Sci. Technol. Received for review December 13, 2000. Revised manuscript 2001, submitted for publication. received April 12, 2001. Accepted April 16, 2001. This research (43) Kukkadapu, R. K.; Boyd, S. A. Tetramethylphospho- was supported by USDA-NRICGP Grants No. 98-35107-6348 nium- and tetramethylammonium-smectites as adsor- and 99-35107-7782, and by the Michigan Agricultural Experi- bents of aromatic and chlorinated hydrocarbons: Effect ment Station. of water on adsorption efficiency. Clays Clay Miner. 1995, 43, 318-323. JF001485D