Environ. Sci. Technol. 2005, 39, 2522-2528
Quantifying the Contribution of Different Sorption Mechanisms for 2,4-Dichlorophenoxyacetic Acid Sorption by Several Variable-Charge Soils SEUNGHUN HYUN AND LINDA S. LEE* Department of Agronomy, Purdue University, West Lafayette, Indiana 47907-2054
Previous research with phenolic, carboxylic, and urea type organic acids demonstrated that hydrophilic sorption was primarily due to anion exchange, which was linearly correlated to chemical acidity (pKa) and the soil anion exchange capacity. However, for dichlorophenoxyacetic acid (2,4-D), sorption by a kaolinitic soil was much higher than expected relative to all other organic acid-soil data. The enhanced sorption was hypothesized to involve calcium bridging of 2,4-D to hydrophilic domains. In this study, the mechanisms contributing to 2,4-D sorption by variable-charged soils were probed and quantified by measuring sorption from CaCl2, KCl, CaSO4, KH2PO4, and Ca(H2PO4)2 solutions. Linear sorption coefficients estimated for 2,4-D sorption from the different electrolytes decreased as follows: CaCl2 > KCl > CaSO4 > Ca(H2PO4)2 = KH2PO4. Differences in 2,4-D sorption from CaCl2 and phosphate solutions were attributed to sorption by hydrophilic domains, which ranged between 46 and 94% across soils. Differences in 2,4-D sorption from CaCl2 and KCl were attributed specifically to Ca-bridging between 2,4-D’s carboxyl group and the silanol edge on kaolinite and quartz and ranged from negligible to 40% depending on the soil mineral type. Differences in sorption from CaCl2 and CaSO4 was attributed to anion exchange, which ranged from 16 to 91%, followed the trends with pKa developed previously for other organic acids, and correlated well to the soil anion to cation exchange capacity ratio (AEC/CEC). The sum of anion exchange and Ca-bridging contributions agreed well with the sorption estimated to be from hydrophilic domains. All other sorption was attributed to hydrophobic processes, which correlated well to a linear freeenergy relationship between pH-dependent organic carbonnormalized sorption coefficients and pH-dependent octanolwater partition coefficients developed for several other organic acids.
Introduction 2,4-Dichlorophenoxyacetic acid (2,4-D), one of the U.S. EPA priority pollutants, has been widely used as a herbicide for nearly the past three decades. Because 2,4-D is an acidic pesticide with a low pKa (2.8), it exists predominantly as an anion within an environmentally relevant pH range (1). It is * Corresponding author phone: (765)494-8612; fax: (765)496-2926; e-mail:
[email protected]. 2522
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005
weakly retained by most soils and sediments similar to other anionic compounds, which has important implications with respect to its efficiency for weed control as well as its potential for environmental pollution. In previous studies, we demonstrated that hydrophilic sorption (anion exchange in most cases) is significant for several model organic acids (e.g., chlorinated phenols, 2,4-D, and prosulfuron) and that the extent of anion exchange correlates well with chemical acidity (pKa) and the ratio of the pH-dependent anion exchange capacity (AEC, cmol(+)/kg) to cation exchange capacity (CEC, cmol(-)/kg) or the total number of charges on the soil surface (AEC + CEC) (2-4). However, for 2,4-D on a kaolinitic soil, sorption was greater than expected relative to the other organic acid-soil combinations. The enhanced sorption was hypothesized to be from hydrophilic sorption interactions involving divalent cations (4). Electrostatic Interactions. 2,4-D sorption by model colloids is primarily through electrostatic interaction (5-7), and to a lesser extent hydrogen bonding (8, 9), with anion exchange being the most important process for controlling 2,4-D retention on oxide surfaces (10-13). 2,4-D anions preferentially replace weakly sorbed anions such as chloride on oxide surfaces (10, 13) whereas phosphate outcompetes 2,4-D for exchange sites even at pH values lower than the point of zero net charge (PZNC) (8, 12, 14). 2,4-D sorption by smectites, which are predominantly negatively charged, is usually negligible (6, 15) unless the smectite surface is at least partially saturated by Al or Fe. Increased sorption for the latter scenario may be due to metal-enhanced electrostatic interaction and H-bonding with metal-coordinated water (9, 14). Cation Bridging. In the presence of divalent cations such as Ca2+, the silanol surface becomes more positive because of surface exchange reactions (Si-OH + Ca2+ / SiO--Ca2+ + H+) at silanol edges (16, 17). Sorption of 2,4-D by Nasaturated kaolinite or from phosphate buffers was negligible (7, 15); however, from CaCl2 solutions containing no phosphate, 2,4-D sorption was considerable (18, 19). Clausen et al. (19) proposed 2,4-D sorption via the formation of a Cabridge between anionic pesticides and the silanol edge (Si-OH) site of layer silicate mineral. In our previous studies with several organic acids, sorption of 2,4-D by a kaolinitic soil was greater than was expected on the basis of a correlation developed between chemical acidity (pKa) and sorption of prosulfuron (pKa of 3.76) and several chlorinated phenols (pKa of 4.69-7.85) (4). The greater 2,4-D sorption was eliminated in the absence of Ca2+. Ca-bridging on mineral surfaces has also been reported as a potential retention mechanism for several other organic acids (20-23). Hydrophobic Interactions. The contribution of soil organic carbon (OC) to 2,4-D sorption by variable-charge soils has not been clearly demonstrated. 2,4-D will exist primarily as an anion in most soils (typical pH range of 4-8); therefore, the contribution of soil organic matter (hydrophobic domain) to 2,4-D retention by variable-charge soil is likely to be small (5, 6). For soils collected from less weathered or temperate regions, a positive correlation between 2,4-D sorption coefficients and OC content was reported (24, 25); however, no mineral information was reported. The correlation was improved by accounting for the pH at which the sorption isotherms were measured (26), suggesting OC content and pH are the most important factors controlling 2,4-D sorption on those soils. Hydrophobic 2,4-D sorption can be ascribed to changes in hydrophobicity of both the organic acid and soil organic matter as a function of pH. At pH > pKa + 2, the organic acid will be anion, thus less 10.1021/es048820p CCC: $30.25
2005 American Chemical Society Published on Web 03/02/2005
TABLE 1. Selected Physical and Chemical Properties CECe AECd cmolc/ cmolc/ kg kg PZNCf
pHb
clay (%)
OCc (%)
A1
6.1
41
1.38
0.42
1.71
5.1
A2
5.2
77
1.17
0.53
3.32
3.0
A3
5.6
82
0.70
0.38
3.91
1.7
DRC
4.7
81
1.34
0.72
3.34
3.1
4R
5.3
59
2.04
0.25
8.07
8R
5.7
47
2.33
0.20
10.4
K1
5.2
27
8.97
0.77
16.7
K2
5.0
39
4.45
0.41
21.8
Gb, Go, He, Qz basalt Santo A ˆ ngelo, Rio Grande do Sul, Brazil; 300 m; backslope; well- drained; open cerrado; pasture K >> C > Gb nak Costa Rica; well-drained, subtropical, plant nursery K >> C > Gb nak Costa Rica, gentle undulating slope-5%; well-drained, subtropical, road side pasture K > A, Qz, C basalt Cheji, Korea; 200 m; well-drained; temperate prairie; upland farm K > A, Qz, C basalt Cheji, Korea; 150 m; well-drained; temperate prairie; upland farm S, Qz >> M > K loess West Lafayette, IN; tile-drained; forest-prairie integrate; soybean/corn farm
a Soils used in previous studies (2-4) and properties first reported in Hyun et al. (2). b 1:10 (g/mL H O). c Organic carbon. d Anion exchange 2 capacity at the pH value reported for soil-water slurries. e Cation exchange capacity at the pH value reported for soil-water slurries. f The point of zero net charge (PZNC). g A, allophone; C, chlorite; Gb, gibbsite; Go, goethite; He, hematite; K, kaolinite; M, mica; S, smectite; Qz, quartz. h Greater than signs indicate mineral amount relative to other minerals identified; the absence of a sign between minerals indicates the relative amounts are approximately the same. i Additional information about A1, A1, and A3 soils can be found in ref 53. k Not available.
hydrophobic relative to the neutral acid species (27). Likewise, dissociation of the acidic functional groups on soil organic matter can change the overall hydrophobicity of the soil (28). Therefore, at pH values where both organic acid and acidic functional group are fully deprotonated, hydrophobic sorption should be minimal. The humification status of soil organic matter may also affect 2,4-D hydrophobic sorption by altering the nature of the hydrophobic domain (soil OC) such as the ratio of aromatic rings to alkyl chains and the presence of polar functional groups (29, 30). For example, 2,4-D sorption was better correlated to the quality and not quantity of organic matter with sorption being greater on the more humified soil. (29). Overall 2,4-D sorption by variable-charge soils will depend on both hydrophobic and hydrophilic sorption domains with the contribution of each sorption domain being a function of the quantity and quality of soil components such as minerals and soil OC. From the above 2,4-D sorption studies with wellcharacterized model sorbents and some soils, 2,4-D sorption appears primarily mineral-mediated through anion exchange and cation bridging and possibly supplemented by some hydrophobic interactions with soil OC. However, the quantitative apportionment of the various sorption mechanisms to sorption by natural variable-charge soil and subsequent predictive capability is still lacking. In the current study, we clarify the hydrophilic sorption mechanisms involved in 2,4-D sorption; and quantify the relative contribution of divalent cations, other hydrophilic sorption mechanisms and hydrophobic sorption to 2,4-D sorption by variable-charge soils.
Materials and Methods Soils. Four subsurface Oxisols collected from Brazil (A1, A2, A3, and DRC), three subsurface Andisols from South Korea (K1 and K2), two surface Ultisols from Costa Rica (4R and 8R), and one surface Alfisols from Indiana (Toronto) were used. Surface and subsurface soil samples were collected at a depth of 0-20 cm and 80-100 cm, respectively. All soils were acidic but vary in soil OC, AEC, CEC, and mineral type
(Table 1). The PZNC estimated from AEC and CEC as a function of pH ranges from pH 98% and calcium sulfate was purchased from Aldrich Chemical Co. (Milwaukee, WI.) Calcium chloride dihydrate was purchased from Fisher Scientific (Pittsburgh, PA). Calcium bis(dihydrogenphosphate) and ammonium fluoride were purchased from Fluka (Buchs, Switzerland). Acetonitrile and potassium chloride were purchased from Mallinckrodt Laboratory Chemicals (Phillipsburg, NJ). The purity of chemical used was greater than 99% unless otherwise noted. Sorption Experiments. Sorption was measured from pHadjusted aqueous solutions consisting of 5 mM CaCl2, 5 mM Ca(H2PO4)2, 5 mM CaSO4, 500 mM CaCl2, 10 mM KH2PO4, and 10 mM KCl as previously described (4). Diluted HCl/ NaOH adjustment was made to achieve similar equilibrium pH in all electrolyte system near the natural soil pH. 2,4-D solutions (4.5-90 µM) were prepared in each electrolyte matrix by dilution of a methanol stock containing 4500 µM 2,4-D; final methanol content was less than 0.2 volume %. 2,4-D solutions were applied to soils, equilibrated for 24 h, centrifuged, and the supernatants were removed. Soils were extracted for 24 h with 4/1 (v/v) acetonitrile/0.25 M NH4F solution; the solute mass extracted is assumed to be the sorbed concentration. 2,4-D concentrations in supernatants and soil extracts were measured using an automated ShiVOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2523
FIGURE 1. Representative 2,4-D sorption isotherms measured from 5 mM CaCl2, 5 mM Ca(H2PO4)2, 5 mM CaSO4, 10 mM KCl, and 10 mM KH2PO4 for A1, K1, DRC, and Toronto soils. madzu HPLC-UV system as previously reported (4). Sorption data collected from each solution were fit with a linear sorption model: Cs ) Kd Cw, where Cs (µmol/kg) and Cw (µmol/ L) are the equilibrium concentrations in the sorbed and solution phases, respectively, and Kd is the linear sorption coefficient (L/kg). The fraction of sorption from hydrophilic processes (fHphilic), Ca-bridging (fCa- bridging), anion exchange sorption (fAE), and hydrophobic sorption (fHphobic) was estimated by the difference in sorption measured from each pair of electrolyte solutions normalized to the sorption measured from 5 mM CaCl2 in which all mechanisms contributed to sorption (4):
fHphilic ) fAE )
Kd,CaCl2 - Kd,Ca(H2PO4)2 Kd,CaCl2 Kd,CaCl2 - Kd,CaSO4 Kd,CaCl2
fCa-bridging )
Kd,CaCl2 - Kd,KCl Kd,CaCl2
fHphobic ) 1 - fHphilic ) 1 - (fAE + fCa-bridging)
(1)
(2)
(3) (4)
where Kd,CaCl2, Kd,Ca(H2PO4)2, Kd,CaSO4, and Kd,KCl are the linear 2,4-D sorption coefficients from 5 mM CaCl2, 5 mM Ca(H2PO4)2, 5 mM CaSO4, and 10 mM KCl solutions, respectively. In CaCl2 solutions, 2,4-D sorption may occur through anion exchange, Ca-bridging, and hydrophobic interactions. In the presence of phosphate at the concentrations used in this study (2-3 orders of magnitude higher than 2,4-D), 2,4-D sorption from both anion exchange and Ca-bridging will be negligible (12, 23, 32-37); therefore, differences in 2,4-D sorption between CaCl2 and Ca(H2PO4)2 systems are due to anion exchange and Ca-bridging (eq 1). In CaSO4 solutions, 2,4-D sorption occurs primarily through Ca-bridging and hydrophobic interactions; therefore, differences in sorption between CaCl2 and CaSO4 electrolytes are attributed to anion exchange (eq 2). In contrast to phosphate, sulfate sorption is considered to be primarily outer-sphere complexation processes (versus inner-sphere for phosphate) and is revers2524
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005
ible (34, 38-40). Likewise, 2,4-D sorption from KCl solutions will not include Ca-bridging; therefore, the contribution of Ca-bridging was estimated from the differences in Kd values measured from CaCl2 and KCl solutions (eq 3).
Results and Discussion 2,4-D mass recovered across all soil-solution combinations at equilibrium in the aqueous electrolyte solution and the solvent extract of the soil relative to the mass applied was 92.7% ( 5.9%. Representative isotherms are shown for A1, K1, DRC, and Toronto soils from the different electrolyte solutions in Figure 1. Linear sorption model fits to the isotherms measured from CaCl2, CaSO4, Ca(H2PO4)2, and KCl systems and the associated equilibrium soil suspension conditions are summarized in Table 2. The Kd values measured for A1, DRC, and K1 soils from 10 mM K2HPO4 system (not summarized in Table 2) are 0.29 ( 0.02 L/kg, 0.99 ( 0.05 L/kg, and 6.72 ( 0.18 L/kg, respectively. For all soil-electrolyte combinations, the goodness of fit (r2 value) for the linear model was greater than 0.96 supporting the use of a linear sorption model and in agreement with previous work by Vasudevan et al. (7). The effect of electrolyte compositions on 2,4-D sorption varied across soils. For a given soil, 2,4-D sorption as a function of electrolyte composition generally decreased as follows: CaCl2 > KCl > CaSO4 > Ca(H2PO4)2 = KH2PO4, which reflects both the effect of the inorganic anion in terms of competition and the cation in terms of bridging interactions. Hydrophilic 2,4-D Sorption. The fraction of hydrophilic sorption (fHphilic) estimated from eq 1 ranged from 0.46 to 0.94 (Table 2). To confirm our assumption that phosphate was present at sufficient concentrations (2 orders of magnitude higher than 2,4-D) to minimize 2,4-D sorption from both anion exchange and Ca-bridging, 2,4-D sorption by 3 representative soils (A1, DRC, and K1) was measured from both KH2PO4 and Ca(H2PO4)2 solutions. The differences in the Kd values from the two electrolytes were not statistically significant at the 95% confidence level confirming our assumption. For 2,4-D sorption by each soil, fHphilic is positively correlated (r2 ) 0.79) to the ratio of AEC (cmol/kg) to CEC (cmol/kg) measured at the isotherm pH values (circles; Figure 2). This trend is similar to what was observed for pentachlorophenol (2) and prosulfuron sorption (3) except that
TABLE 2. Sorption Coefficients from Linear Model Fits to 2,4-D Isotherms Measured from 5 mM CaCl2, 5 mM Ca(H2PO4)2, 5 mM CaSO4, 10 mM KCl, and Solutionsk Kd (L/kg) soil A1i A2 A3 DRCi 4R 8R K1 K2 Toronto
pHa 5.8 4.4 5.0 4.2 4.8 4.8 4.7 4.2 4.3
AECb 0.54 0.74 0.46 0.96 0.32 0.30 1.08 0.58 0.04
CECb 1.46 2.57 3.54 2.76 8.67 7.20 13.5 18.0 11.2
CaCl2 (0.13)c
4.49 7.97 (0.24) 2.32 (0.08) 16.6 (0.4) 2.60 (0.09) 3.49 (0.11) 18.5 (0.7) 4.20 (0.12) 1.28 (0.04)
Ca(H2PO4)2
CaSO4
KCl
fHphilicd
fAEe
fCa-bridgingf
fHphobicg
fAEh
0.26 (0.11) 0.91 (0.04) 0.49 (0.03) 1.06 (0.03) 0.70 (0.03) 1.05 (0.09) 6.91 (0.29) 2.26 (0.15) 0.57 (0.02)
0.41 (0.03) 1.19 (0.06) 1.30 (0.08) 4.48 (0.17) 2.41 (0.11) 2.54 (0.09) 13.9 (0.5) 3.21 (0.11) 1.07 (0.05)
4.59 (0.20) 7.46 (0.14) 1.53 (0.11) 12.6 (0.7) NDj ND 11.1 (0.23) 2.93 (0.12) 0.77 (0.04)
0.94 0.89 0.79 0.94 0.73 0.70 0.63 0.46 0.55
0.91 0.85 0.44 0.73 0.07 0.27 0.25 0.24 0.16
-0.02 0.06 0.34 0.24 ND ND 0.40 0.30 0.40
0.06 0.11 0.21 0.06 0.27 0.30 0.37 0.54 0.45
0.96 0.83 0.45 0.70 ND ND 0.23 0.16 0.15
a Average isotherm pH at equilibrium across electrolytes with a standard deviation 0.96.
FIGURE 2. The fraction of hydrophilic sorption (fHphilic) and anion exchange (fAE) for 2,4-D versus the ratio of anion exchange capacity (AEC, cmol(+)/kg) to cation exchange capacity (CEC, cmol(-)/kg) measured at the isotherm pH. Lines are linear regression fits. Also plotted are values calculated from data reported by Vasudevan and Cooper (5).
hydrophilic sorption of 2,4-D was observed even for soils with AEC/CEC ratios near zero resulting in a nonzero intercept of approximately 0.6 indicating sorption from other than or in addition to simple anion exchange. Cation Bridging. The difference in 2,4-D sorption from CaCl2 and KCl systems used to estimate the contribution from Ca-bridging (fCa-bridging, eq 3) is summarized in Table 2. For the DRC, K1, and K2 soils in which kaolinite is the dominant mineral and the Toronto soil in which quartz dominates, sorption in the absence of Ca2+ (e.g., KCl system) was significantly smaller than that measured from 5 mM CaCl2 (Table 2). For the soils containing primarily gibbsite (A1 and A2) in which Ca-bridging is not considered significant, the absence of Ca2+ had a negligible effect on 2,4-D sorption (Table 2). Clausen et al. (19) reported a decrease in 2,4-D sorption by a pure oxide with increasing CaCl2 concentration (0-10 mM), which further supports anion exchange as the dominant 2,4-D retention mechanism and the relative insignificance of Ca-bridging in the systems. Several researchers propose that Ca-bridging can take place between anionic soil organic matter functional groups (e.g., carbolic and phenolic groups) or the negatively charged sites on constant charge mineral surfaces and anionic pesticides (41-43). Additionally, several studies reported that the formation of stable positively charged monodentate complex between divalent metal ions (e.g., Cu2+) and monovalent organic anions/ligands in aqueous solution may
result in additional sorption to cation exchange sites (44, 45). However, unlike a Cu2+-anion complex, the calciumcarboxylate monodentate complex is not strong in aqueous solution and thus is unlikely to contribute significantly to 2,4-D sorption (45). For the soils in the current study, the estimated fCa-bridging (Table 2) was poorly correlated with soil CEC (r2 ) 0.25) or CEC/AEC (r2 ) 0.06) (not shown) indicating that cation exchange does not play a positive role on 2,4-D sorption. The latter may be because 2,4-D sorption via Cabridging would take place on silanol edges (Si-OH) of kaolinite or silicate edges, which remain mostly neutral in the pH range investigated; greater than 86% of the silanol edges in kaolinite are neutral between pH values of 3 to 6 (34). Anion Exchange. Values for fAE estimated from the difference in 2,4-D sorption between 5 mM CaCl2 and 5 mM CaSO4 (eq 2, Table 2) is positively correlated to the ratio of AEC/CEC (r2 ) 0.84; slope ) 2.54, filled triangles in Figure 2) but with a zero intercept (unlike for fHphilic, filled circles in Figure 2). The negative charge quantified by CEC would serve to repel organic anions, thereby attenuating sorption to anion exchange sites. Combining 2,4-D data from the current study with literature data for other organic acids (2, 3), a trend for a given soil of increasing contribution of anion exchange (fAE) with increasing acidity of the organic becomes apparent; the slope between fAE and AEC/CEC increases with decreasing pKa: 2,4-D (pKa ) 2.8) > prosulfuron (pKa ) 3.76) > pentachlorophenol (pKa ) 4.69). Vasudevan and Cooper (5) also noted that anion exchange of 2,4-D was a substantial contribution to the sorption process from 2,4-D sorption data collected in phosphate-free distilled water and potassium phosphate systems. They did not attempt to calculate the fAE values; however, applying eq 2 to their 2,4-D sorption data from phosphate-free and 9.69 mM phosphate systems resulted in fAE values for two of their three soils that are in good agreement with what we observe for our soils (open triangles, Figure 2). If we assume that hydrophilic sorption is primarily the sum of sorption by anion exchange and Ca-bridging, then the difference between fHphilic and fCa-bridging estimated using eq 1 and 3, respectively, should approximate the estimate of fAE calculated using eq 3. These two estimates of fAE are generally in very good agreement (Table 2), which supports our assumption. Good agreement also supports that the relative contribution of each sorption mechanism can be reasonably estimated by exploiting sorption measured from different electrolyte solutions. Therefore, although other sorption mechanisms have been proposed such as hydrogen bonding and van der Waals interaction, their relative VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2525
FIGURE 3. Log Koc values calculated from 2,4-D sorption from 0.005 M Ca(H2PO4)2 (A) plotted as a function of pH with the solid line obtained by fitting the data to 2,4-D speciation weighted sum of KocHA and KocA- (42) and (B) plotted as a function of log KowpH values calculated from the speciation weighted 2,4-D speciation weighted sum of KowHA and KowA- values (26) with the solid and dotted lines representing the linear regression fit and one standard deviation, respectively. contribution to the overall sorption does not appear significant for 2,4-D sorption by variable-charge soils. Hydrophobic Sorption. The hydrophobic sorption of 2,4-D (Table 2) was responsible for 6 to 54% of the total sorption. Hydrophobic-type sorption contributed the least (e21%) to the highly weathered soils (Oxisols; A1, A2, A3, and DRC) compared to the relatively less weathered soils. Sorption through hydrophobic forces should primarily vary with the amount of soil OC and the pH-dependent speciation of 2,4-D and organic matter functional groups. To assess this hypothesis for 2,4-D, the log Koc,Ca(H2PO4)2 values calculated from the Kd,Ca(H2PO4)2 values measured from the current study along with those we estimated previously for four soils with a pH range between 3 and 6.5 (A1, DRC, K1, and Toronto; ref 46) were plotted as a function of pH (Figure 3A). Recall that we hypothesized that sorption from Ca(H2PO4)2 represents sorption through primarily hydrophobic forces. Data were fitted with a weighted sum of KocHA and KocA- (27); KocpH ) KocHA fHA + KocA- fA-, where fHA and fA- are the neutral (HA) and anionic (A-) fractions of 2,4-D, which resulted in values of log KocHA and log KocA- of 2.99 and 1.70, respectively. In a similar manner, pH-dependent KowpH values were estimated using log KocHA and log KocA- values of 2.81 and -0.75, respectively (31). These log KowpH values were correlated with the log Koc,Ca(H2PO4)2 values (n ) 40) giving log KocpH ) 0.29 log KowpH + 1.74 (r2 ) 0.61; Figure 3B). The consistency based on the linear free-energy relationship between Kow and Koc suggests that 2,4-D sorption through hydrophobic-type forces can be reasonably predicted from pH, OC, and KowpH. Increasing CaCl2 Concentrations. The presence of anion exchange and Ca-bridging on 2,4-D sorption by variablecharge soils was further assessed by measuring sorption at 2526
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005
higher CaCl2 concentrations (500 mM). Increasing CaCl2 concentrations may suppress anion exchange while enhancing both Ca-bridging and hydrophobic interactions depending on soil properties (13, 19, 47). Four soils (A1, DRC, K1, and Toronto) were selected that varied in mineral type and the major sorption mechanism identified using eqs 1-4 (Table 2). For gibbsite-dominated Al soil, an increase in salt concentration decreased sorption from 4.49 ( 0.13 L/kg to 1.42 ( 0.06 L/kg, supporting that anion exchange is the main 2,4-D sorption mechanism as estimated from eq 2 (fAE ) 0.94, Table 2). For the other three soils, which are kaolinitic, increasing CaCl2 concentrations may also enhance Cabridging and hydrophobic sorption (19). For the DRC soil (fAE ) 0.70, Table 2), 2,4-D sorption decreased from 16.6 ( 0.44 L/kg to 4.05 ( 0.36 L/kg at the higher CaCl2 concentration, which supports the dominance of anion exchange even though Ca-bridging was a contributing mechanism (fCa-bridging ) 0.24, Table 2). For K1 and Toronto soils, an increase in CaCl2 concentration increased sorption about 30% (Kd values are 25.48 ( 1.46 and 1.68 ( 0.10 for K1 and Toronto soils, respectively) indicating the dominance of Ca-bridging and hydrophobic sorption. Nearly 80% of 2,4-D sorption by K1 and Toronto soils was attributed to Ca-bridging and hydrophobic interactions in Table 2, which is supported by the effect of increasing CaCl2 concentrations. Trends in sorption observed with increasing CaCl2 concentrations across several soils support the approach used in this study to delineate the contribution of specific sorption mechanisms to sorption by soils. Functional Group and pKa Effects on Anion Exchange. The contribution of hydrophilic processes to sorption of several organic acids by a gibbsite-rich soil (A1) and a kaolinitic soil (DRC) from CaCl2 solutions was previously observed to be linearly correlated to chemical acidity (pKa) with the exception of 2,4-D sorption to the kaolinitic soil (4). For the latter, sorption was higher than expected, but upon eliminating the Ca-associated sorption contribution, the remaining hydrophilic sorption contribution followed the trend observed with the other organic acids. In the current study, the same phenomenon was consistently observed for additional soils. For the gibbsite-rich A2 soil, fHphilic for 2,4-D is equivalent to fAE whereas for all the kaolinitic soils, fHphilic of 2,4-D is greater than fAE (Table 2). Upon eliminating the Ca-associated sorption contribution, the trend between the remaining hydrophilic sorption contribution, that is, fAE, and pKa appears to follow a linear trend with pentachlorophenol (2) and prosulfuron (3) data (see supporting material). Agricultural Implications. 2,4-D sorption was reduced by more than an order of magnitude in the presence of phosphate or sulfate. Although type and amount of fertilizers and pesticides are site- and crop-dependent, phosphate fertilizer (P2O5, H2PO4-; about 20 ∼ 50 kg ha-1) or gypsum (CaSO4‚2H2O; about 157 ∼ 616 kg ha-1) are often used as plant nutrients or soil amendments (48, 49). Moreover, because deficiencies in phosphate and sulfate are often large in variable-charge mineral soils such as Oxisols and Andisols, farmers frequently apply 10-20 times more of these soil amendments then may be required (48, 49). Recommended field application rates of pesticide 2,4-D ranged from 0.28 to 1.7 kg/ha (1), which may be 2-3 orders of magnitude less than the inorganic anions. The applied inorganic anion will outcompete anionic 2,4-D for hydrophilic sorption sites because of a higher affinity for sites (e.g., phosphate) or a high molar ratio, for example, [sulfate] >>> [2,4-D]. Also specific and preferential sorption of phosphate by variablecharge minerals can modify the soil surface to be more negative (50-52). Therefore, continuous input of phosphate for variable-charge soils through the fertilization may subsequently decrease sorption of postapplied 2,4-D and desorb previously applied 2,4-D. Such processes need
to be incorporated into best management practices to maintain crop production without enhancing the likelihood of off-site leaching to surface and groundwaters.
Supporting Information Available Figure exemplifying the fractions of hydrophilic sorption (fHphilic) and specifically anion exchange (fAE) for several organic acids chemical pKa. This material is available free of charge via the Internet at http://pubs.acs.org.
Literature Cited (1) Weed Science Society of America. Herbicide Handbook, 8th ed.; Weed Sci. Soc. Am.: Champaign, IL, 2002. (2) Hyun, S.; Lee, L. S.; Rao, P. S. C. Significance of anion exchange in pentachlorophenol sorption by variable-charge soils. J. Environ. Qual. 2003, 32, 966-976. (3) Hyun, S.; Lee, L. S. Factors controlling sorption of prosulfuron by variable-charge soils and model sorbents. J. Environ. Qual. 2004, 33, 1354-1361 (4) Hyun, S.; Lee, L. S. Hydrophilic and hydrophobic sorption of organic acids by variable-charge soils: effect of chemical acidity and acidic functional group. Environ. Sci. Technol. 2004, 38, 5413-5419 (5) Vasudevan, D.; Cooper, E. M. 2,4-D sorption in iron oxide-rich soils: Role of soil phosphate and exchangeable Al. Environ. Sci. Technol. 2004, 38, 163-170. (6) Celis, R.; Hermosin, M. C.; Cox, L.; Cornejo, J. Sorption of 2,4dichlorophenoxyacetic acid by model particles simulating naturally occurring soil colloids. Environ. Sci. Technol. 1999, 33, 1200-1206. (7) Vasudevan, D.; Cooper, E. M.; Van Exem, O. L. Sorptiondesorption of ionogenic compounds at the mineral-water interface: study of metal oxide-rich soils and pure-phase minerals. Environ. Sci. Technol. 2002, 36, 501-511. (8) Haque, R.; Sexton, R. Kinetic and equilibrium study of the adsorption of 2,4-dchlorophenoxyacetic acid on some surface. J. Colloid Interface Sci. 1968, 27, 818-827. (9) Cox, L.; Celis, R.; Hermosin, M. C.; Cornejo, J. Natural soil colloids to retard simazine and 2,4-D leaching in soil. J. Agric. Food Chem. 2000, 48, 93-99. (10) Kavanagh, B. V.; Posner, A. M.; Quirk, J. P. The adsorption of phenoxyacetic acid herbicides on goethite. J. Colloid Interface Sci. 1977, 61, 545-553. (11) Kavanagh, B. V.; Posner, A. M.; Quirk, J. P. Effect of adsorption of phenoxyacetic acid herbicides on the surface charge of goethite. J. Soil Sci. 1980, 31, 31-39. (12) Madrid, L.; Diaz-Barrientos, E. Effect of phosphate on the adsorption of 2,4-D on Lepidocrocite. Aust. J. Soil Res. 1991, 29, 15-23. (13) Watson, J. R.; Posner, A. M.; Quirk, J. P. Adsorption of the herbicide 2,4-D on goethite. J. Soil Sci. 1973, 24, 503-511. (14) Sannino, F.; Violante, A.; Gianfreda, L. Adsorption-desorption of 2,4-D by hydroxy aluminium montmorillonite complexes. Pestic. Sci. 1997, 51, 429-435. (15) Weber, J. B.; Perry, P. W.; Upchurch, R. P. The influence of temperature and time on the adsorption of paraquat, diquat, 2,4-D and prometone by clays, charcoal, and an anion-exchange resin. Soil Sci. Soc. Am. Proc. 1965, 29, 678-688. (16) Schlautman, M. A.; Morgan, J. J. Sorption of Perylene on a nonporous inorganic silica surface: Effects of aqueous chemistry on sorption rates. Environ. Sci. Technol. 1994, 28, 2184-2190. (17) Zachara, J. M.; Ainsworth, C. C.; Cowan, E.; Schmidt, R. L. Sorption of aminonaphthalene and quinoline on amorphous silica. Environ. Sci. Technol. 1990, 24, 118-126. (18) Cason, R.; Reed, L. W. Chemistry of two clay systems and three phenoxy herbicides. Proc. Okla. Acad. Sci. 1977, 57, 116-118. (19) Clausen, L.; Fabricius, I.; Madsen, L. Adsorption of pesticides onto quartz, calcite, kaolinite and R-alumina. J. Environ. Qual. 2001, 30, 846-857. (20) Biggar, J. W.; Cheung, M. W. Adsorption of picloram (4-amino3,5,6-trichloropicolinic acid) on Panoche, Ephrate, and Palouse soils: A thermodynamic approach to the adsorption mechanism. Soil Sci. Soc. Am. Proc. 1973, 37, 863-868. (21) Peterson, J. B. Calcium linkage, a mechanism in soil granulation. Soil Sci. Soc. Am. Proc. 1947, 12, 29-34. (22) Schevchenko, S. M.; Bailey, G. W. Non-bonded organo-mineral interactions and sorption of organic compounds on soil surfaces: a model approach. J. Mol. Struct. (THEOCHEM) 1998, 422, 259-270.
(23) Varadachari, C.; Modal, A. H.; Ghosh, K. The influence of crystal edges on clay-humus complexation. Soil Sci. 1995, 159, 185190. (24) Moreale, A.; Van Bladel, R. Behavior of 2,4-D in Belgian soils. J. Environ. Qual. 1980, 9, 627-633. (25) von Oepen, B.; Ko¨rdel, V.; Klein, W. Sorption of nonpolar and polar compounds to soils: Processes, measurements and experience with the applicability of the modified OECDGuideline 106. Chemosphere 1991, 22, 285-304. (26) Farenhorst, A.; Muc, D.; Monreal, C.; Florinski, I. Sorption of herbicides in relation to soil variability and landscape position. J. Environ. Sci. Health 2001, B36, 379-387. (27) Lee, L. S.; Rao, P. S. C.; Nkedi-Kizza, P.; Delfino, J. J. Influence of solvent and sorbent characteristics on distribution of pentachlorophenol in octanol-water and soil-water systems. Environ. Sci. Technol. 1990, 24, 654-661. (28) Spadotto, C. A.; Hornsby, A. G. Soil sorption of acidic pesticides: Modeling pH effects. J. Environ. Qual. 2003, 32, 949956. (29) Dorado, J.; Tinoco, P.; Almendros, G. Soil parameters related with sorption of 2,4-D and atrazine. Commun. Soil Sci. Plant Anal. 2003, 34, 1119-1133. (30) Picton, P.; Farenhorst, A. Factors influencing 2,4-D sorption and mineralization in soil. J. Environ. Sci. Health 2004, B39, 367-379. (31) U.S. Department of Agriculture. The ARS Pesticide Properties Database; USDA, 1989. http://www.arsusda.gov/acsl/services/ ppdb/textfiles/2,4-D-ACID (accessed Aug 2004). (32) Appelt, H.; Coleman, N. T.; Pratt, P. F. Interactions between organic compounds, minerals and ions in volcanic-ash-derived soils: II. Effects of organic compounds on the adsorption of phosphate. Soil Sci. Soc. Am. Proc. 1975, 39, 628-630. (33) Hingston, F. J.; Posner, A. M.; Quirk, J. P. Anion adsorption by goethite and gibbsite. I. The role of the proton in determining adsorption envelops. J. Soil Sci. 1972, 23, 177-192. (34) He, L. M.; Zelazny, L. W.; Baligar, V. C.; Ritchey, K. D.; Martens, D. C. Ionic strength effects on sulfate and phosphate adsorption on γ-alumina and kaolinite: Triple-layer model. Soil Sci. Soc. Am. J. 1997, 91, 784-793. (35) Muljadi, D.; Posner, A. M.; Quirk., J. P. The mechanism of phosphate adsorption by kaolinite, gibbsite and pseudobohemite I, II and III. J. Soil Sci. 1966, 17, 212-247. (36) Van Olphen, H. Introduction to clay colloid chemistry; Interscience: New York, 1963. (37) Zhang, H. X.; Tao, Z. Y. Sorption of uranyl ions on silica: Effects of contact time, pH, ionic strength, concentration and phosphate. J. Radioanal. Nucl. Chem. 2002, 254, 103-107 (38) Curtin, D.; Syers, J. K. Mechanism of sulfate adsorption by two tropical soils. J. Soil Sci. 1990, 41, 295-304. (39) Guadalix, M. E.; Pardo, M. T. Sulfate sorption by variable charge soils. J. Soil Sci. 1991, 42, 607-614. (40) Marsh, K. B.; Tillman, R. W.; Syers, J. K. Charge relationship of sulfate sorption by soils. Soil Sci. Soc. Am. J. 1987, 51, 318-323 (41) Deschauer, H.; Ko¨gel-Knabner, I. Sorption behavior of a new acidic herbicide in soils. Chemosphere 1990, 21, 1397-1410. (42) Matallo, M.; Romero, E.; Sa´nchez-Rasero, F.; Pen ˜ a, A.; Dios, G. Adsorption of mecoprop and dichlorprop on calcareous and organic matter amended soils; comparative adsorption of racemic and pure enantiomeric forms. J. Environ. Sci. Health 1998, B33, 51-66. (43) Spark, K. M.; Swift, R. S. Effect of soil composition and dissolved organic matter on pesticide sorption. Sci. Total Environ. 2002, 298, 147-161. (44) Maes, A.; Cremers, A. Stability of metal uncharged ligand complexes in ion exchangers. Part 4.-Hydration effects and stability changes of copper-ethylenediamine complexes in montorillonite. J. Chem. Soc., Faraday 1 1979, 75, 513-524. (45) Ali, M. A.; Dzombak, D. Effects of simple organic acids on sorption of Cu2+ and Ca2+ on goethite. Geochim. Cosmochim. Acta 1996, 60, 291-304. (46) Hyun, S. Soil, solution, and chemical properties attenuating organic acid sorption by variable-charge soils. Ph.D. Thesis, Purdue University, West Lafayette, IN, 2003. (47) Clausen, L.; Fabricius, I. Atrazine, isoproturon, mecoprop, 2,4D, and bentazone adsorption onto iron oxides. J. Environ. Qual. 2001, 30, 858-869. (48) Fox, R. L. Soils with variable charge: Agronomic and fertility aspects. In Soils with variable charge; Theng, B. K. G., Ed.; N. Z. Soc. Soil Sci.: Lower Hutt, New Zealand, 1980; pp 195-223. (49) Sanchez, P. A.; Uehara, G. Management consideration for acid soils with high phosphorous fixation capacity. In The role of VOL. 39, NO. 8, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
2527
phosphorous in agriculture; Khasawneh, F. E.; Sample, E. C.; Kamprath, E. J., Eds.; ASA-CSSA-SSSA: Madison, WI, 1980; pp 471-514. (50) Mekaru, T.; Uehara, G. Anion adsorption in ferruginous tropical soils. Soil Sci. Soc. Am. Proc. 1972, 36, 296-300. (51) Schalscha, E. B.; Pratt, P. F.; Soto, D. Effect of phosphate adsorption on the cation-exchange capacity of volcanic ash soils. Soil Sci. Soc. Am. Proc. 1974, 38, 539-540. (52) Parfitt, R. L.; Atkinson, R. J. Phosphate adsorption on goethite (R-FeOOOH). Nature 1976, 264, 740-742
2528
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 8, 2005
(53) Marques, J. J. Trace element distributions in Brazilian cerrado soils at the landscape and micrometer scales. Ph.D. Thesis, Purdue University, West Lafayette, IN, 2000.
Received for review July 28, 2004. Revised manuscript received January 21, 2005. Accepted January 21, 2005. ES048820P
