JEQ Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (30)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Clausen, L.
Right arrow Articles by Madsen, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Clausen, L.
Right arrow Articles by Madsen, L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Clausen, L.
Right arrow Articles by Madsen, L.
Related Collections
Right arrow Ground Water Quality
Right arrow Agricultural Pesticides
Right arrow Water Pollution
Journal of Environmental Quality 30:846-857 (2001)
© 2001 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America

TECHNICAL REPORT
Ground Water Quality

Adsorption of Pesticides onto Quartz, Calcite, Kaolinite, and {alpha}-Alumina

L. Clausena, I. Fabriciusa and L. Madsenb

a Dep. of Geology and Geotechnical Engineering, Technical Univ. of Denmark, Building 204, DK-2800 Lyngby, Denmark
b Dep. of Geochemistry, Geological Survey of Denmark and Greenland, Thoravej 8, DK-2400 Copenhagen, Denmark

Corresponding author (igglc{at}pop.dtu.dk)

Received for publication October 20, 1999.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
The fate of pesticides in aquifers is influenced by the small but not insignificant adsorption of pesticides to mineral surfaces. Batch experiments with five pesticides and four minerals were conducted to quantify the contributions to adsorption from different mineral surfaces and compare adsorption characteristics of selected pesticides. Investigated mineral phases included quartz, calcite, kaolinite, and {alpha}-alumina. Selected pesticides comprised atrazine (6-chloro-N2-ethyl-N4-isopropyl-1,3,5-triazine-2,4-diamine), isoproturon [3-(4-isopropylphenyl)-1,1-dimethylurea)], mecoprop [(RS)-2-(4-chloro-2-methylphenoxy)propionic acid], 2,4-D (2,4-dichlorophenoxyacetic acid), and bentazone [3-isopropyl-1H-2,1,3-benzothiadiazin-4-(3H)-one 2,2-dioxide]. Specific surface area and mineral surface charge proved to be important for the adsorption of these pesticides. Detectable adsorption of the anionic pesticides (mecoprop, 2,4-D, and bentazone) was only measured when positive sites were present on the mineral surface. However, when CaCl2 was added as an electrolyte, a detectable adsorption of mecoprop and 2,4-D was also measured on kaolinite (which exhibits a negative surface charge), probably due to formation of Ca–pesticide–surface complexes. Adsorption of the uncharged pesticides (atrazine and isoproturon) was detected only on kaolinite. The lack of adsorption on {alpha}-alumina indicates that the uncharged pesticides have a greater affinity for the silanol surface sites (=SiOH) than for the aluminol surface sites (=AlOH) in kaolinite. No measurable effect of ionic strength was found for the uncharged pesticides. The results indicate that quartz and calcite play a smaller role than clay minerals.

Abbreviations: HPLC, high performance liquid chromatography • SEM, scanning electron micrograph • TOC, total organic carbon


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
SORPTION–DESORPTION is one of the key processes affecting the fate of agrochemicals in the sediment–water environment, so a thorough understanding of sorption is paramount for the prediction of pesticide movement in soils and aquifers. Pesticides are adsorbed on both organic and inorganic soil constituents; their relative importance depends on the amount, distribution, and properties of these constituents and the chemical properties of the pesticides (Bailey and White, 1970; Adams, 1973). The sorption of uncharged organic compounds by soils has been shown to be highly correlated with soil total organic carbon (TOC) content (Chiou et al., 1979; Briggs, 1981; Karickhoff, 1984). It is therefore generally accepted that the predominant sorbent of uncharged organic compounds is organic carbon, as long as the TOC is larger than >0.1% (Schwarzenbach and Westall, 1981). In sediments with low organic carbon content the influence of the mineral surfaces may, however, also be important (Stevenson, 1976, p. 180–207; McCarty et al., 1981; Karickhoff, 1984; Haderlein and Schwarzenbach, 1993). Ionic pesticides can interact with the surface sites through, for example, electrostatic interactions, ion-exchange reactions, or by surface complexation, and for these pesticides adsorption to mineral surfaces may be significant (Brownawell et al., 1990; Schwarzenbach et al., 1993; Celis et al., 1996, 1999). Whereas much research has been devoted to study the sorption of pesticides in soils with high organic carbon, much less information is available on the sorption of pesticides to aquifer sediments with low organic carbon content. Sorption of pesticides on aquifer minerals and the role of parameters other than organic matter therefore deserve further investigations in order to evaluate the risk of ground water contamination.

The purpose of this study is to evaluate the relative importance of the different minerals in aquifer sediments on pesticide adsorption using pure aquifer minerals. Earlier work has demonstrated that the use of model sorbents is a helpful tool to determine the role of the different aquifer constituents in adsorption, and to understand the adsorption mechanisms (Cox et al., 1995, 1998; Celis et al., 1999).

The pesticides selected for the work represent both uncharged and ionic herbicides. The five selected pesticides are atrazine, isoproturon, mecoprop, 2,4-D, and bentazone. All of the selected pesticides have been or are being used in Europe in amounts greater than 500 Mg/yr (Fielding et al., 1991), and they have all been detected in surface water and ground water in Denmark. The pesticides were also chosen for a related study of the sorption of pesticides to Danish aquifer sediments (Madsen et al., 2000).

Earlier investigations have been carried out on the adsorption of atrazine to clays (Terce and Calvet, 1978; Borggaard and Streibig, 1988; Laird et al., 1992; Moreau-Kervévan and Mouvet, 1998) and on the adsorption of atrazine to iron oxides (Borggaard and Streibig, 1988; Moreau-Kervévan and Mouvet, 1998). The sorption of isoproturon to clays has been investigated by Terce and Calvet (1978) and Worrall et al. (1996), and to chalk aquifers by Johnson et al. (1998). Also, adsorption investigations of the ionic pesticide 2,4-D have been carried out on clays (Frissel and Bolt, 1962; Aly and Faust, 1964; Weber et al., 1965; Hermosin and Cornejo, 1993; Sannino et al., 1997; Celis et al., 1999) and oxide minerals (Whatson et al., 1973; Kavanagh et al., 1980; Celis et al., 1999). However, not much work has been devoted to the adsorption of pesticides on calcite and quartz, which are major constituents of aquifers. Therefore, these minerals were selected for this work. In order to address the role of clays, kaolinite was included as sorbent. Of the common clay minerals, kaolinite was chosen due to the nonexpanding character. Experiments with {alpha}-alumina were also performed in order to evaluate the role of aluminol surface sites (=AlOH), which are present on many minerals in the environment. The selected minerals represent silicate minerals, carbonate minerals, and oxide minerals, and therefore represent different surface functional groups. The selected minerals are thus expected to have different adsorption capacities.

The specific goals of the study were to (i) quantify the contributions to adsorption from the selected constituents of aquifer solids and characterize the observed adsorption isotherms, (ii) study the effect of ionic strength, (iii) evaluate the importance of various surface functional groups on mineral surfaces, and (iv) compare adsorption characteristics of the selected pesticides.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Sorbent Characteristics
Scanning electron micrographs (SEMs) of the four selected mineral powders are shown in Fig. 1. We chose to work with synthetic minerals whenever possible because they are of high purity and well defined. The identity of each mineral was confirmed by X-ray diffraction (CuK{alpha}).



View larger version (145K):
[in this window]
[in a new window]
  		Fig. 1. Scanning electron micrographs (SEMs) of mineral powders. (A) quartz, (B) calcite, (C) {alpha}-alumina, (D) kaolinite. The scalebar in the SEM image of the quartz powder is 20 µm, whereas the scalebar in the other images is 2 µm.

 
The quartz lattice consists of SiO4 tetrahedrons linked together in a three-dimensional framework. The hydroxylated quartz surface is dominated by silanol groups (=SiOH) (Parks, 1984). The SEM image (Fig. 1) shows that the quartz sample is composed of irregular fragments ranging in size from 1.5 to 15 µm.

Calcite is a carbonate mineral with a rhombehedral structure. The surface of calcite in water is assumed to be dominated by the uncharged sites =CaOH and =CO3H, but ionic sites are also present in the form of =Ca(OH2)+ and =CO3(OH2)- (Geffroy et al., 1999). The synthetic calcite powder was washed four times with MilliQ water (Millipore, Molsheim, France) to remove impurities of Ca(OH)2 and freeze-dried to minimize agglomeration (Madsen et al., 1998). The SEM image (Fig. 1) shows that the powder is composed of crystals in aggregates with crystal diameters between 0.2 and 0.7 µm.

The structure of the oxide mineral {alpha}-alumina consists of oxygen in hexagonal packing and Al3+ in octahedral coordination. In the presence of water, metal oxide minerals are generally covered with amphoteric surface hydroxyl groups (=AlOH), and a hydroxylated oxide particle can therefore be thought of as a polymeric oxo-acid or -base (Kummert and Stumm, 1980). We used {alpha}-alumina to simulate some of the surface properties in clays. The studied powder was baked at 550°C to remove organic impurities. The SEM image (Fig. 1) shows aggregates of equidimensional crystals with a diameter of 0.13 µm.

The clay mineral kaolinite consists of layers of octahedrally coordinated Al3+ alternating with layers of tetrahedrally coordinated Si4+ (1:1 stoichiometric ratio). The surface functional groups are aluminol (=AlOH), silanols (=SiOH), and Lewis acid sites (AlOH+2) (Sposito, 1989). The basal surfaces are thus largely neutral or, in case of Al substitution for Si, permanently slightly negative, whereas the edge surfaces are charged (Davis and Kent, 1990). The silanol groups in kaolinite are hardly comparable with those in quartz because of the wide structural difference between tectosilicates and phyllosilicates. To remove contamination on the surface, the kaolinite powder was washed three times with MilliQ water and freeze-dried. The SEM image of the powder (Fig. 1) shows stacked pseudohexagonal platelets (approximately 100 nm thick) with diameters in the range 0.2 to 2 µm. Relevant properties of the studied minerals are summarized in Table 1.


View this table:
[in this window]
[in a new window]
  		Table 1. Mineral characteristics.

 
Sorbates and Solutions
The structural formulae and relevant properties of the five selected herbicides are shown in Fig. 2 and Table 2. The experiments were carried out with unlabeled pesticides and 14C-labeled pesticides. The unlabeled pesticides were used as supplied (Riedel-de Haën, Seelze, Germany) with the following purities expressed in weight percent: atrazine, 99.4%; isoproturon, 99%; mecoprop, 99%; 2,4-D, 99%; and bentazone, 99.9%. The labeled 14C pesticides were purchased from Sigma Chemical (St. Louis, MO; atrazine and 2,4-D), Amersham International (Buckinghamshire, UK; isoproturon and mecoprop), and International Isotope (Munich, Germany; bentazone).



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 2. Structural formulae of herbicides used in sorption studies.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Properties of herbicides used in sorption studies. Data from Shiu et al. (1990), Tomlin (1994), and Hornsby et al. (1996).

 
The specific activity and purity of the 14C-labeled pesticides are: atrazine, 695 MBq/mmol, >95%; isoproturon, 914 MBq/mmol, 97%; mecoprop, 1828 MBq/mmol, >98.4%; 2,4-D, 673 MBq/mmol, >98%; and bentazone, 514 MBq/mmol, >98%. In bentazone the 14C label is located in the carbonyl group and in the other selected pesticides the 14C label is located in the ring structure.

Stock solutions of pesticides were prepared in methanol and kept refrigerated. The methanol was removed under N2 flow before dilution in mineral-saturated low TOC water (MilliQ), resulting in methanol concentrations of less than 0.03 vol. %. In experiments with controlled ionic strength the pesticide solutions were prepared in mineral-saturated 0.005 M CaCl2 solution or in mineral-saturated 0.01 M CaCl2 solution. Calcium chloride was chosen as the electrolyte because it is recommended in OECD Guideline106 (Organisation for Economic Co-operation and Development, 1993) and because Ca is a common cation in aquifers. All solutions were sterile-filtered (0.2 µm) before use.

High performance liquid chromatography (HPLC) measurements were used to evaluate the polarity of the compounds. The following HPLC conditions were used: 5-µm Hypersil (Cheshire, UK) ODS column of 250 mm length and 2.0 mm i.d.; C18 column packing; 0.3 mL/min flow rate; 10:90 (v/v) acetonitrile to 0.001 M ammonium acetate eluent system; 200 nm UV detection.


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Screening Experiments
Screening experiments on the adsorption of all the selected pesticides on all minerals were performed using a batch equilibrium technique based on OECD Guideline 106, similar to the method used by Madsen et al. (2000). The experiments were done at 10 ± 1°C in the dark with an initial pesticide concentration of 0.25 mg/L (1.04–1.21 µmol/L). The solid to liquid ratio in the experiments was 0.4 for calcite, {alpha}-alumina, and kaolinite. In the experiments with quartz the solid to liquid ratio was 1.0 due to the low surface area of this mineral. Glass tubes (10 mL) with Teflon caps were used, and preliminary tests excluded adsorption on these materials. All glassware used in the experiments was baked at 550°C to remove organic contamination. Triplicate samples of 2 g freeze-dried mineral powder (for quartz, 5 g) were equilibrated with 4 mL sterile-filtered mineral-saturated MilliQ water for at least 20 h. In experiments with controlled ionic strength, the mineral powder was equilibrated with 4 mL mineral-saturated sterile-filtered 0.005 M CaCl2 solution (I {approx} 0.015) or 0.01 M CaCl2 solution (I {approx} 0.030). In additional experiments with quartz, pH in the mineral saturated water was adjusted to 2.4 with 1 M HCl (I {approx} 0.004). After equilibration of the mineral–water system, 1 mL pesticide solution was added, and each test tube was placed in a vertical rotator. After equilibration for 96 h the suspension was centrifuged for 40 min at 3200 rpm, and 1 mL of supernatant was removed for analysis. The pH of the remaining solution was determined by electrode measurement. The removed supernatant was mixed with 10 mL OptiPhase HiSafe 3 (Wallac, Turku, Finland) scintillation cocktail and the amount of radioactivity determined by counting for 20 min in a 1414 WinSpectral Liquid Scintillation Counter (Wallac). The specific radioactivities in the experiments were ca. 1000 disintegrations per minute (DPM)/mL, whereas the background counting was ca. 40 DPM/mL. All measurements were corrected for background counting. The concentration of pesticide in the supernatant solution was calculated from the difference between the radioactivity in the supernatant solution and the radioactivity in the reference solution. For each sorption experiment, controls were prepared in triplicate without mineral powders but otherwise handled identically.

Isotherm Experiments
Further experiments with increasing pesticide concentration were performed with atrazine, isoproturon, and 2,4-D (representing the anionic pesticides) on the minerals where a detectable sorption had been detected in the screening experiments. Isotherm experiments with 2,4-D were done with initial concentrations in the range 0.05 to 150 mg/L. Experiments with atrazine and isoproturon were performed with initial concentrations in the range 0.05 to 10 mg/L.

Additional isotherm experiments were performed with atrazine and 2,4-D to characterize the isotherms in a wider concentration range. The initial 2,4-D concentration ranged from 0.25 to 350 mg/L and the initial atrazine concentration ranged from 0.5 to 30 mg/L. To avoid the dissolution involved in adding the pesticide solution to the mineral phase suspended in water, the pesticide solution was added directly to the dried mineral powder.

Data Analysis
Data from screening experiments were analyzed assuming a linear isotherm:

[1]
where Kd is the adsorption distribution coefficient, Cs is the equilibrium concentration of sorbate associated with the sorbent, and Ce is the equilibrium concentration of sorbate in solution. In environmental science, Kd is often expressed in liters per kilogram. However, the sorption properties of the mineral are related to the surface area rather than the weight of the mineral, and sorption distribution coefficients for minerals are therefore more useful if they are normalized to the solid surface area. Accordingly, Kd in the present paper is expressed in L/m2, Cs in mol/m2, and Ce in mol/L. The adsorption was calculated in percentage from the liquid scintillation counting, and the Kd values (L/m2) were calculated from:

[2]
where A is the adsorption (%), V0 is the initial volume of aqueous phase in contact with the mineral (L), and Smineral is the surface area of the mineral phase (m2). The standard deviations of Kd were calculated from the formula of propagation error.

The sorption isotherm data were analyzed according to Freundlich and Langmuir isotherms. A Freundlich equation is empirical, whereas the Langmuir isotherm indicates that only one type of adsorption site is involved.

The Freundlich equation can be written as:

[3]
where KF is the Freundlich constant ([µmol/m2] [µmol/L]-n) and n is a measure of the nonlinearity involved (dimensionless). The Freundlich isotherm parameters were calculated by the least-squares method using log-transformed equilibrium concentrations. In this calculation we have assumed a constant relative error (Kinniburgh, 1986), which is an appropriate assumption according to our measurements in percent.

The Langmuir equation can be written as:

[4]
where Cs,max is the maximum concentration of sorbate associated with the sorbent (mol/m2) and KL is a constant known as the Langmuir parameter (L/mol). The Langmuir parameters have been calculated using nonlinear least squares (NLLS) analysis (assuming constant relative error), as described by Kinniburgh (1986).


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Linear Distribution Coefficients
The calculated linear adsorption distribution coefficients (Table 3, Fig. 3) were normalized to specific surface area to account for the influence of the solid surface area, which is obviously larger for fine particles than for coarse particles. After normalization to the specific surface area, the Kd values varied significantly between the selected minerals (Table 3). None of the selected pesticides adsorb on quartz at pH 6.5. However, by lowering pH to 2.4, adsorption is found for atrazine, mecoprop, 2,4-D, and bentazone. On calcite and {alpha}-alumina adsorption is only significant for the anionic pesticides: mecoprop, 2,4-D, and bentazone. On kaolinite adsorption is only significant for the uncharged pesticides: isoproturon and atrazine. (Due to high pKb values [Table 2], atrazine behaves as a uncharged compound except under extremely acidic conditions.) The equilibrium concentrations of the anionic pesticides in experiments with kaolinite were in fact found to be slightly larger than the respective initial concentrations. The same was observed in other studies of adsorption of 2,4-D to kaolinite and montmorillonite done by Weber et al. (1965) and Celis et al. (1999). As noted by these authors, the dried clays apparently adsorbed water in preference to the anionic pesticides, leading to an increase in the pesticide concentration in the solution.


View this table:
[in this window]
[in a new window]
  		Table 3. Sorption distribution coefficients (Kd values). The Kd values were obtained at an initial pesticide concentration of 250 µg/L (1.04–1.21 µmol/L).

 


View larger version (31K):
[in this window]
[in a new window]
  		Fig. 3. Distribution coefficients (Kd) normalized to surface area, as a function of equilibrium concentration of sorbate in solution. Error bars indicate standard deviation.

 
To compare the measured Kd values in Table 3 quantitatively, the adsorption experiments must be performed in the concentration range corresponding to the linear portion of the adsorption isotherm, so that the Kd values are independent of sorbate concentration. Accordingly, isotherm experiments were performed with atrazine, isoproturon, and 2,4-D (representing the anionic pesticides) on the minerals where a detectable adsorption had been measured in the screening experiments. From the isotherm experiments a linear sorption coefficient was calculated in the solution equilibrium concentration range where the curvature (n in the Freundlich equation) of the isotherm was not different from 1 within 95% confidence limits (Table 4).


View this table:
[in this window]
[in a new window]
  		Table 4. Linear sorption distribution coefficients (Kd values) calculated from screening experiments and linear sorption coefficients calculated from isotherm experiments.

 
The Kd values determined in the screening experiments generally agree with the Kd values calculated in the linear equilibrium concentration range (Table 4), showing a good reproducibility of the experimental results. One exception is the Kd values for atrazine on quartz, in which the screening data is higher than the isotherm data, possibly due to the lower pH value in the screening experiment. However, the equilibrium solution concentration range, where the determined Kd values are independent of sorbate concentration, varies significantly between the different adsorption isotherms (Table 4). This is also illustrated in Fig. 3, where the Kd value determined in each isotherm point is presented as a function of the equilibrium concentration in solution. The data for adsorption of atrazine on quartz, 2,4-D on calcite, and isoproturon on kaolinite show constant Kd values in the equilibrium concentration range examined (Fig. 3, Table 4). The Kd values for 2,4-D on quartz can be regarded as a constant up to an equilibrium concentration of approximately 4 µmol/L. Above this concentration the Kd values decrease to a lower constant level. However, all adsorption data on quartz have large errors, due to the relatively low adsorption percentage (A = 2–4%). It was not possible to determine the adsorption more accurately, because the solid to liquid ratio was 1 and as large as possible in these experiments. The Kd values for 2,4-D on {alpha}-alumina are constant up to an equilibrium concentration of 30 µmol/L, but above this concentration Kd increases rapidly. The adsorption of atrazine on kaolinite decreases significantly as the equilibrium concentration increases and, as a consequence, the linear equilibrium concentration range is small (0.2–2.7 µmol/L). In general, the determined Kd values shown in Fig. 3 are independent of the sorbate concentration when the equilibrium concentration is low. This means that the assumption of a linear isotherm is valid as long as the equilibrium concentration is less than 2.7 µmol/L. The Kd values in Table 3 can therefore be compared quantitatively.

Adsorption Isotherms
The determined isotherms are shown on a log scale in Fig. 4. In general, the isotherm data fit the Freundlich equation well, whereas poor correlation to the Langmuir equation was found. As also demonstrated in Table 4 and Fig. 4, the curvature (n) of the Freundlich isotherms for atrazine on quartz, 2,4-D on calcite, and isoproturon on kaolinite were not different from 1 (within 95% confidence interval), and consequently linear isotherms can describe the curves as well as the Freundlich equation. The isotherm for 2,4-D on quartz and the isotherm for atrazine on kaolinite (Fig. 4) show, on the other hand, a curvature (n) significantly less than 1, and these isotherms are therefore best described by the Freundlich equation. The isotherm for 2,4-D on {alpha}-alumina has a curvature significantly greater than 1, implying that Kd increases with increasing sorbate concentration (Fig. 3 and 4). Therefore, this isotherm is of the S-type according to the classifications by Giles et al. (1960). The S-shaped isotherm for adsorption to {alpha}-alumina has also been found for the adsorption of benzoic acid (Madsen and Blokhus, 1994) and for the adsorption of 2,4-D on ferrihydrite (Celis et al., 1999). The S-shaped isotherms illustrate that the more 2,4-D is already adsorbed on the oxide minerals, the easier it is for additional molecules to become fixed, probably through hydrophobic pesticide–pesticide interaction.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 4. Freundlich adsorption isotherms shown on a log scale for the sorption of 2,4-D, atrazine, and isoproturon on minerals. The Freundlich constant (KF) is given in (µmol/m2) (µmol/mL)-n. Error bars indicate standard deviation (may be covered by the symbols).

 
Further isotherm experiments were performed to investigate whether the isotherms shown in Fig. 4 would approach or reach a plateau (Langmuir behavior) if the initial pesticide concentration was increased. From these experiments we hoped to determine the maximum adsorption capacity and thereby the site density on the minerals. The additional isotherm experiments were performed without wetting the mineral powder before addition of pesticide solution. The results are shown on a linear scale in Fig. 5 and 6. The isotherm experiments with atrazine (Fig. 5) were performed with an initial concentration close to the solubility of atrazine (153 µmol/L). However, neither the adsorption isotherm for atrazine on kaolinite nor the adsorption isotherm for atrazine on quartz reach the maximum adsorption capacity (Fig. 5). The isotherm for atrazine on kaolinite fit the Freundlich equation in the lower equilibrium concentration range (Fig. 4). The isotherm for atrazine on quartz (Fig. 5) shows a more significant curvature than the isotherm in the lower concentration range (Fig. 4). In fact, this isotherm shows a tendency toward reaching a plateau, and the Langmuir equation can be fitted to the data points (Fig. 5). The calculated maximum adsorption capacity (Cs,max) of 0.0047 µmol/m2 is, however, nearly twice as high as the highest data point of 0.0026 µmol/m2, and the calculated maximum adsorption is therefore hypothetical.



View larger version (28K):
[in this window]
[in a new window]
  		Fig. 5. Atrazine adsorption isotherms on a linear scale for the sorption on quartz and kaolinite. Full symbols denote adsorption points where the pesticide solutions were added after equilibration of the water–mineral system. Open symbols denote adsorption points where the pesticide solutions were added directly to the dried mineral powder. The Freundlich constant (KF) is given in (µmol/m2) (µmol/mL)-n. Error bars indicate standard deviation (may be covered by the symbols).

 


View larger version (24K):
[in this window]
[in a new window]
  		Fig. 6. 2,4-D adsorption isotherms on a linear scale for the sorption on quartz, calcite, and {alpha}-alumina. Full symbols denote adsorption points where the pesticide solutions were added after equilibration of the water–mineral system. Open symbols denote adsorption points where the pesticide solutions were added directly to the dried mineral powder. The Freundlich constant (KF) is given in (µmol/m2) (µmol/mL)-n. Error bars indicate standard deviation (may be covered by the symbols).

 
Maximum adsorption capacities could not be determined from the additional adsorption experiments with 2,4-D on quartz, calcite, and {alpha}-alumina (Fig. 6). The isotherm for 2,4-D on quartz (Fig. 6) shows an S-shape tendency in the larger equilibrium concentration range. The data point in the low concentration range from the adsorption on the dried mineral surface is in agreement with the isotherm data (Fig. 4) obtained from the water–mineral equilibrated system. Therefore, the S-shaped isotherm (Fig. 6) is probably not an effect of adding the pesticide solution to the dried mineral surface, but rather an expression of an increasing adsorption with increasing sorbate concentration. The isotherm for 2,4-D on calcite (Fig. 6) fits the Freundlich equation in the entire equilibrium concentration range. In contrast, the isotherm for 2,4-D on {alpha}-alumina changes shape when the pesticide is added directly to the dried mineral surface (Fig. 6). Up to an initial concentration of 230 µmol/L, the isotherm data obtained when adding the pesticide solution to the dried mineral surface follows the same isotherm as obtained from the prehydrolyzed surface. However, above an initial concentration of 230 µmol/L, the adsorption of 2,4-D on {alpha}-alumina increases significantly. This result indicates a change in adsorption mechanism dependent on the initial concentration of the pesticide and the state of hydrolysis of the surface.

Effect of Calcium Chloride as a Background Electrolyte
Experiments where CaCl2 were added as an electrolyte (Fig. 7) illustrate that the adsorption of mecoprop on quartz at pH 6.5 increases with increasing CaCl2 concentration. Increasing mecoprop and 2,4-D adsorption is also found with increasing CaCl2 concentration on kaolinite. In contrast, the adsorption of the anionic pesticides on calcite and {alpha}-alumina decreases with increasing CaCl2 concentration. However, the decrease in adsorption of the anionic pesticides is more pronounced on {alpha}-alumina than on calcite. The results demonstrate that the addition of CaCl2 electrolyte strongly affects the adsorption of the anionic pesticides, whereas no measurable effect of ionic strength was found for the adsorption of the uncharged pesticides.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 7. Effect of CaCl2 as electrolyte on the sorption of selected pesticides by quartz, calcite, {alpha}-alumina, and kaolinite. The solution conditions are: temperature = 10°C, pH (quartz) = 6.7 ± 0.1, pH (calcite) = 8.8 ± 0.1, pH ({alpha}-alumina) = 7.0 ± 0.1, pH (kaolinite with CaCl2) = 5.9 ± 0.1, and pH (kaolinite without CaCl2) = 6.5. Error bars indicate standard deviation.

 
The addition of CaCl2 to the solution may cause several competing effects (Mattigod et al., 1979; Stumm, 1992). Solution effects include (i) an increasing competition for adsorption sites by the electrolyte anions caused by increasing CaCl2 concentration, (ii) an increasing aqueous complexation with the anionic pesticides as a consequence of increasing electrolyte cation concentration, and (iii) a decrease in the activity of the charged ions caused by increasing electrolyte concentration. All three solution effects cause decreasing adsorption of anionic pesticides to mineral surfaces with a net positive charge. In addition, there are electrostatic effects. For example, surface complexation models indicate that the rate at which the potential decays with distance from the surface increases with increasing ionic strength (Dzombak and Morel, 1990). This effect is, however, counteracted by a change in the surface charge, because acid–base titration curves demonstrate that an increase in ionic strength will increase the charge on the mineral surface (Dzombak and Morel, 1990).

The formation of Ca–pesticide complexes is dependent on the type of anion involved in the formation. However, the curves for the adsorption of the anionic pesticides on {alpha}-alumina and calcite are parallel-displaced and the adsorption of the anionic pesticides are therefore affected to the same degree. The decreasing adsorption of the anionic pesticides with increasing CaCl2 concentration is therefore probably not an effect of Ca–pesticide complexes. On calcite, the decrease in adsorption must be due to electrostatic effects because chloride will be inert with respect to the calcite surface. The increasing chloride to anionic pesticide ratio leading to more efficient competition for the surface positive sites can, however, be an important effect on {alpha}-alumina, because chloride is known to adsorb on oxide minerals (Cornell and Schwertmann, 1996). This would also explain the more pronounced decrease on {alpha}-alumina than on calcite when CaCl2 was added to the experiments. Haderlein and Schwarzenbach (1993) have demonstrated that Ca2+ can adsorb onto silanol surface sites, resulting in positive =Ca+ sites to which negatively charged compounds will be attached due to electrostatic interactions. Therefore, the increasing adsorption of anionic pesticides on kaolinite and quartz with increasing CaCl2 concentration can be explained by Ca complexes bridging pesticides and mineral.

Adsorption of Ionic Pesticides
The absence of adsorption of the anionic pesticides on kaolinite in experiments without electrolyte is consistent with results for 2,4-D on clays reported in other studies (Frissel and Bolt, 1962; Weber et al., 1965; Sannino et al., 1997; Celis et al., 1999), where no adsorption of 2,4-D on clays was observed except in experiments with low pH or in experiments with an electrolyte.

A detectable adsorption of mecoprop, 2,4-D, and bentazone was only observed when the suspension pH was below or near the point of zero charge (PZC) (Tables 1 and 3), where positive sites are present on the mineral surface. This is especially obvious in the experiments with quartz, where no adsorption was found at pH 6.5. However, by lowering pH to near the PZC, a significant adsorption of all three anionic pesticides was found. Atrazine also adsorbs on quartz at pH 2.4, probably on the negative sites because atrazine is positively charged at this pH.

On calcite and {alpha}-alumina, the adsorption of the anionic pesticides increases in the order bentazone < 2,4-D < mecoprop. At the pH in these experiments, the acidic pesticides are primarily of anionic form. Therefore, it seems probable, as also proposed by other authors (Watson et al., 1973; Sannino et al., 1997; Celis et al., 1999), that the phenoxy acids (mecoprop and 2,4-D) may sorb with their COO- groups oriented toward the positive sites on the surface. The difference in adsorption properties of mecoprop and 2,4-D may be caused by differences in polarity between the compounds. The HPLC measurements confirmed that mecoprop is less polar than 2,4-D. When lowering pH in the experiments with quartz the sequence in adsorption of the phenoxy acids changes, and 2,4-D is now more strongly adsorbed than mecoprop. At the pH in this experiment (2.4), 2,4-D is more dissociated than mecoprop due to a lower pKa value (Table 2), and this result indicates again that it is the COO- group in the molecules that is responsible for the adsorption of these compounds. The adsorption of bentazone on all minerals is less than the adsorption of mecoprop and 2,4-D, but as for these compounds the adsorption of bentazone is dependent of the mineral surface charge. Bentazone is therefore probably also sorbed through electrostatic interaction between the anionic part of the molecule and the positive sites on the surface. However, the electrostatic attraction between bentazone and the surface is not as strong as for the phenoxy acids.

The adsorption of the anionic pesticides is dependent on the surface charge and, as a consequence, it is the charged surface sites that are responsible for the adsorption. On quartz (pH 2.4) the anionic pesticides are therefore attracted by the =SiOH+2 and atrazine is probably attracted by the =SiO- site. On calcite the adsorption of the anionic pesticides is probably related to the =Ca(OH2)+ site on the surface. On {alpha}-alumina the adsorption may be caused by electrostatic interactions between the anionic part of the molecules and the =AlOH+2 site on the surface. Kummert and Stumm (1980) proposed that phenoxy acids can also adsorb through a ligand exchange mechanism, where the organic acids replace the surface hydroxy groups on the Al2O3 surface. However, the strong dependence of ionic strength on the adsorption of the phenoxy acids observed in this study (Fig. 7) illustrates that these compounds are adsorbed through a nonspecific adsorption (outer sphere adsorption) involving electrostatic bonding mechanisms (Stumm, 1992). Specific adsorption through a ligand exchange mechanism is therefore probably not the case in these experiments.

Adsorption of Nonionic Pesticides
Adsorption of the uncharged pesticides was only detected on kaolinite. With regard to calcite, Johnson et al. (1998) found that sorption of isoproturon on chalk was related to small quantities of organic matter or clay located in fractures in the chalk, whereas the chalk matrix itself was relatively inert. This is in agreement with the results obtained in this study, where no sorption of isoproturon on calcite was found. The adsorption of atrazine and isoproturon on kaolinite has been studied by Worrall et al. (1996) and Moreau-Kervévan and Mouvet (1998). Moreau-Kervévan and Mouvet (1998) measured a Freundlich coefficient for atrazine on kaolinite of Cs = 7.1 x 10-5 -n x C0.956e (equilibrium concentration range 0–2 µmol/L, 20°C, 0.01 M CaCl2). Considering the different experimental conditions, this Freundlich coefficient is in reasonable agreement with our Freundlich coefficient of 8.2 x 10-5 -n x C0.903e (calculated from Fig. 4). Worrall et al. (1996) determined a Kd value for adsorption of isoproturon on kaolinite of 2.15 L/kg (20°C, 0.02 M CaCl2). By using the isochore for the sorption of isoproturon determined by Worrall et al. (1996), we get a temperature-corrected Kd,10°C value of 1.15 L/kg. Worrall et al. (1996) did not determine the BET surface area of the kaolinite used, but their results will agree with ours under an assumption of a surface area of 19 m2/g, which is a realistic value (Bailey and White, 1970).

The adsorption of nonionic atrazine was greater than that of isoproturon on kaolinite. In several investigations of adsorption of nonionic organic compounds to mineral surfaces, the large aqueous solubility of a nonionic compound decreases adsorption (Schwarzenbach et al., 1993; Mader et al., 1997). The adsorption of atrazine and isoproturon follows that tendency, because the solubility of isoproturon is twice the solubility of atrazine (Table 2). The commonly proposed adsorption mechanisms for uncharged pesticides to clay minerals in the neutral pH range are Van der Waals forces and hydrophobic bondings (Bailey and White, 1970; Koskinen and Harper, 1990). However, the hydrophobicity of atrazine and isoproturon was not sufficient to lead to adsorption on the other three studied minerals. No detectable adsorption of atrazine and isoproturon on {alpha}-alumina was found, which indicates that in clay minerals these compounds have a greater affinity for the silanol surface sites than for the aluminol surface sites in clay minerals. This has also been demonstrated to be the case for the adsorption of several nitroaromatic compounds (NACs), where Haderlein and Schwarzenbach (1993) found a specific adsorption of NACs to the silanol surface sites of kaolinite. These authors proposed that an electron donor–acceptor complex (i.e., a {pi} complex) between electron donor functions at the siloxane surface and the aromatic ring system of the NAC is responsible for the observed specific adsorption. This may also be a plausible mechanism for the adsorption of the uncharged pesticides investigated in this study. A larger adsorption of both isoproturon and atrazine on smectite than on kaolinite has been explained by the larger surface area and cation exchange capacity of smectite (Worrall et al., 1996; Moreau-Kervévan and Mouvet, 1998). However, the larger adsorption on smectite than on kaolinite may also be related to the 2:1 structure of these clays, where more silanol surface sites are available for adsorption.

Environmental Implications
The results demonstrate that adsorption of pesticides to mineral surfaces is significant. The measured Kd values for the synthetic calcite are equivalent to a range of Kd for anionic pesticides in natural chalk of 0 to 0.04 mL/g (assuming a surface area of 2 m2/g for natural chalk and a CaCl2 concentration of 0.01 M). The surface-area-normalized Kd values for kaolinite (Table 3) are equivalent to a Kd value for uncharged pesticides in the range 0.4 to 0.8 mL/g for a sediment of pure kaolinite. An adsorption of atrazine, mecoprop, 2,4-D, and bentazone on quartz was found by lowering the pH to 2.4. Here, the measured Kd values are equivalent to 0.03 to 0.08 mL/g for a typical quartz sand with a specific surface area near 1 m2/g. Adsorption studies with atrazine, isoproturon, mecoprop, 2,4-D, and bentazone on natural aquifer sands yield results in the range 0 to 1 mL/g (Madsen et al., 2000). This indicates that quartz and calcite only play a minor role as compared with clay and possible oxide minerals on low-TOC sands.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
(i) Pesticides adsorb only weakly to pure mineral surfaces of calcite, quartz, kaolinite, and {alpha}-alumina from aqueous solutions.

(ii) For a given concentration, the magnitude of the adsorption depends on the type of pesticide, the type of mineral, and the specific surface of the mineral.

(iii) Anionic pesticides (mecoprop, 2,4-D, and bentazone) adsorb to mineral surfaces at pH values where the surface contains positive sites; addition of CaCl2 as an electrolyte causes the adsorption to diminish. This indicates that the anionic pesticides are weakly attached through electrostatic forces. When the mineral surface is dominated by negative sites (kaolinite at pH 5.9, quartz at pH 6.7), addition of CaCl2 induces a rise in adsorption of mecoprop and 2,4-D, probably due to formation of Ca complexes bridging pesticide and mineral.

(iv) Uncharged pesticides (atrazine above pH 4 and isoproturon) adsorb to kaolinite only and not to calcite, quartz, or {alpha}-alumina. The hydrophobicity of these pesticides is apparently not sufficient to lead to adsorption on these minerals. The lack of adsorption on {alpha}-alumina may indicate that the adsorption involves the Si–oxide side of the kaolonite crystal. The addition of CaCl2 has no influence on the adsorption of the uncharged pesticides.

(v) Linear adsorption isotherms are well defined for 2,4-D, isoproturon, and atrazine on all the selected minerals at equilibrium concentrations below 2.7 µmol/L. Kd values in the range 0.012–0.092 mL/m2 were found.

(vi) Above 2.7 µmol/L, most isotherms follow the Freundlich equation or even a linear equation. Two exceptions were 2,4-D on quartz and 2,4-D on {alpha}-alumina, which exhibit S-shaped isotherms.


   ACKNOWLEDGMENTS
 
The present work was supported financially by the Technical University of Denmark (DTU) through a Ph.D. grant for Liselotte Clausen. The technical assistance from Bente Frydenlund is gratefully acknowledged. Vibeke Knudsen is thanked for editing the figures, and Flemming Rasmussen is thanked for the construction of the vertical rotator. The authors also wish to thank Henrik Spliid (Department of Mathematical Modelling, DTU) for advice concerning the statistic data analysis. 14C-labeled pesticides were kindly provided by GEUS (Geological Survey of Denmark and Greenland). Bo Lindhardt (GEUS) and Niels P. Arildskov (DTU) are thanked for helpful suggestions and for critically reading the manuscript.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 J. Environ. Qual.Home page
L. Clausen and I. Fabricius
Atrazine, Isoproturon, Mecoprop, 2,4-D, and Bentazone Adsorption onto Iron Oxides
J. Environ. Qual., May 1, 2001; 30(3): 858 - 869.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (30)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Clausen, L.
Right arrow Articles by Madsen, L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Clausen, L.
Right arrow Articles by Madsen, L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Clausen, L.