effect of soil composition and dissolved organic matter on pesticide sorption

The Science of the Total Environment 298(2002) 147–161

0048-9697/02/$ - see front matter� 2002 Elsevier Science B.V. All rights reserved.PII: S0048-9697Ž02.00213-9

Effect of soil composition and dissolved organic matter onpesticide sorption

K.M. Spark *, R.S. Swift1,

Department of Soil Science, University of Reading, Whiteknights, Reading, Berkshire, RG6 2AA, UK

Received 13 August 2000; accepted 1 May 2002

Abstract

The effect of the solid and dissolved organic matter fractions, mineral composition and ionic strength of the soilsolution on the sorption behaviour of pesticides were studied. A number of soils, chosen so as to have different claymineral and organic carbon content, were used to study the sorption of the pesticides atrazine(6-chloro-N -ethyl-N -2 4

isopropyl-1,3,5-triazine-2,4-diamine), 2,4-D ((2,4-dichlorophenoxy)acetic acid), isoproturon(3-(4-isopropylphenyl)-1,1-dimethylurea) and paraquat(1,19-dimethyl-4,49-bipyridinium) in the presence of low and high levels of dissolvedorganic carbon and different background electrolytes. The sorption behaviour of atrazine, isoproturon and paraquatwas dominated by the solid state soil components and the presence of dissolved organic matter had little effect. Thesorption of 2,4-D was slightly affected by the soluble organic matter in the soil. However, this effect may be due tocompetition for adsorption sites between the pesticide and the soluble organic matter rather than due to a positiveinteraction between the pesticide and the soluble fraction of soil organic matter. It is concluded that the major factorgoverning the sorption of these pesticides is the solid state organic fraction with the clay mineral content also makinga significant contribution. The dissolved organic carbon fraction of the total organic carbon in the soil and the ionicstrength of the soil solution appear to have little or no effect on the sorptionytransport characteristics of thesepesticides over the range of concentrations studied.� 2002 Elsevier Science B.V. All rights reserved.

Keywords: Pesticides; Sorption behaviour; Organic matter state; Mineral content

1. Introduction

The retention and mobility of a pesticide in soilis determined by the extent and strength of sorptionreactions, which are governed by the chemical and

*Corresponding author. Tel.:q61-7-5460-1336.E-mail address: [email protected](K.M. Spark).Present address: University of Queensland, Gatton Cam-1

pus, Gatton Qld 4343, Australia

physical properties of the soils and pesticidesinvolved. The sorption interactions of pesticides inthe soil environment may involve either the min-eral or organic components, or both. For soils thathave higher organic matter levels()5%), themobility of the pesticides has been related to thetotal organic matter content, with the nature of theorganic matter having little apparent influence onsorption processes(Bailey and White, 1964;

148 K.M. Spark, R.S. Swift / The Science of the Total Environment 298 (2002) 147–161

Hayes, 1970; Riise et al., 1994; Arienzo andBuondonno, 1993; Jenks et al., 1998; Bekbolet etal., 1999). Total organic matter includes both thesoluble and insoluble fraction of organic matter,although the proportion of soluble organic matterin a soil is usually very small.For soils, which have low organic matter con-

tents, the mobility of the pesticide is often relatedto the active components of the inorganic fraction,which is predominantly the clay-sized fraction. Anincrease in the clay content results in decreasingmobility of the pesticide, with the composition ofthe clay and the identity of the major cations inthe soil solution also being important(Walker andCrawford, 1968; Borggaard and Streibig, 1988;Murphy et al., 1992; Barriuso et al., 1992; Wel-house and Bleam, 1992; Baskaran et al., 1996a).There is considerable evidence that pesticides

can interact with the soluble form of soil organicmatter in the absence of the solid components ofthe soil. The extent and nature of this interactiondepends on factors such as molecular weight andpolarity of the pesticide(Burns et al., 1973;Gamble et al., 1986; Chiou et al., 1986; Madhunet al., 1986; Kan and Tomson, 1990; Maquedo etal., 1993; Spark and Swift, 1994; Devitt andWiesner, 1998). Pesticides such as atrazine havebeen shown to be capable of solution-phase com-plexing with dissolved organic matter(Devitt andWiesner, 1998; Gamble et al., 1986); paraquatassociates with dissolved organic matter(Burns etal., 1973; Maquedo et al., 1993; Spark and Swift,1994) and there is evidence for the association ofhydrophobic pesticides with dissolved organic mat-ter (Kan and Tomson, 1990; Kulovaara, 1993;Fitch and Du, 1996).There is some evidence that, in the presence of

soils, an increase in the concentration of dissolvedorganic matter results in an increase in the mobilityof atrazine(Gao et al., 1997) and 2,4-D(Baskaranet al., 1996b), and of hydrophobic organic chemi-cals such as DDT(Kan and Tomson, 1990; Fitchand Du, 1996). As many of these pesticides arefound in groundwaters it is apparent that there issignificant transport of these chemicals throughthe soil profile(Johnson et al., 1995; Masse et al.,1994; Miller et al., 1995; Ro et al., 1997). Thistransport may be the result of processes such as:

the formation of soluble complexes with soil solu-tion components such as dissolved organic matter;or the incomplete interaction of these pesticideswith the solid state organic or inorganic matter inthe soil.Pesticides which interact with organic matter

will react with both the soluble and solid phasefractions, therefore, competitive effects, the revers-ibility of these two types of interaction, and massaction effects will govern the distribution of thepesticides between the solid and soluble phases ofthe organic matter. For example, the sorption ofatrazine on clays is generally reversible whereasthe sorption from organic matter is less so(Harrisand Warren, 1964; Moreaukervevan and Mouvet,1998).As discussed above, it is has been shown that

there is a strong relationship between total organiccarbon in the soil and the mobility of manypesticides. However, the extent to which the solu-ble organic carbon is involved in the transportprocess is not known. If the interaction with thesoluble organic matter relative to the insolubleorganic matter is significant, the transport of thepesticides through the soil may be enhanced by anincrease in soluble organic matter moving throughthe soil profile. The concentration of soluble organ-ic matter would therefore influence the concentra-tion of the pesticides in the drainage waters,surface waters and ground waters associated witha particular soil.The aim of this work was to determine the

relative importance of soil components on theadsorption of a variety of pesticides. The researchwas carried out using a range of soils differing inorganic matter content as well as clay mineralcontent, and varying the concentration of the sol-uble organic matter in the soil solution.

2. Materials

Reagents were analytical grade unless otherwisestated. All solutions were made up in distilleddeionised water and kept in airtight containers.

2.1. Soil characterisation

The UK soils used in this work were from theSonning and Thames series(Jarvis, 1973) from

149K.M. Spark, R.S. Swift / The Science of the Total Environment 298 (2002) 147–161

Table 1Summary of characteristics of soils

Sonning I Sonning II Sonning III Broad Denchworth

Clay mineralsa

Smectite Small Small Moderate Moderate LargeMicas Moderate Moderate Moderate Small SmallKaolinite Moderate Moderate Moderate Small Small

Particle size(% content of total soil))600mm 1.1 2.3 2.3 5.4 0.8212–600mm 30.7 28.3 28.1 6.4 2.763–212mm 29.8 29.5 25.7 15.9 7.42–63mm 12.4 15.3 12.2 19.6 16.3-2 mm 10.8 9.4 12.9 32.7 44.2

Soil solution componentsb

pH 7.6 7.0 7.0 7.3 7.4Cond. mSycm 0.19 0.26 0.27 0.80 0.37DOC(ppm) 23 51 60 117 55Soluble Ca(II) (mmolyg) 2.6 1.8 2.8 10.0 4.4(Molyl) 0.0013 0.00090 0.0014 0.0050 0.0022Absorbance @ 400 nm 0.051 0.158 0.162 0.211 0.083Abs yDOC (=1000)400nm 2.2 3.1 2.7 1.8 1.5

Organic carbon characteristics% OC (total soil) 1.25 2.05 3.30 6.3 3.25DOCyOC (%) 0.18 0.25 0.18 0.19 0.17

X-Ray diffraction analysis of the clay mineral fraction.Note: All samples contain quartz.a

Soluble components in the supernatant of the soil suspension prepared as described for the sorption studies.b

the University of Reading farm, Sonning, Berk-shire and the Denchworth series(Avery, 1990)from Brimstone, Oxfordshire. A summary of thecharacteristics of the soil samples taken from thesesites is shown in Table 1. Three samples withdifferent organic carbon(OC) contents were takenof the Sonning soil, a free-draining, brown, coarse,sandy-loam soil. Sonning I soil(OCs1.25%) hadbeen under continuous cultivation and cropping,Sonning II (OCs2.05%) was soil under naturalgrassland vegetation, and Sonning III soil(OCs3.30%) was taken under woodland vegetation. TheDenchworth soil (OCs3.25%) had a similarorganic matter content to the Sonning III soil, butwas a very fine clay from the Ap horizon. TheBroad soil, from the Thames series, had the highestOC content in the soils studied here(OCs6.3%)and was a dark brownish-grey, alluvial, loamy-claysoil from a flat, flood-prone plain under grasslandvegetation.The soils were dried, sieved to-1 mm and

stored in sealed containers at room temperature.The soils were characterised for particle size using

sedimentation characteristics of the soil(Averyand Bascomb, 1982) and total carbon using a LecoSC-444 Analyser. The results of the particle sizeanalysis and mineralogical analysis of the soilsusing X-Ray Diffraction are given in Table 1.The composition of the supernatant of the soils

when suspended using the same conditions as forthe pesticide adsorption experiments(describedbelow) but in the absence of pesticides, wasanalysed to determine the pH, the backgroundconcentration of Ca (Perkin Elmer 3030 Atomic2q

Absorption Spectrophotometer), the absorbance at400 nm (Perkin Elmer Lambda 2 UV-VisibleSpectrophotometer), the conductivity and the con-centration of dissolved organic carbon(ShimadzuTOC 5000 Total Organic Carbon Analyser) (Table1).

2.2. Leaching solution containing dissolved fulvicacid

The soluble humic substance used in this workwas a fulvic acid (FA) isolated from a water


draining from a sapric histosol(Methwold Fens,Norfolk). The pH of the drain water was adjustedto two before passing it through XAD-8 resin toseparate the FA from the other components ofhumic substances. Additional treatment associatedwith purification of the FA in the eluent isdescribed in Watt et al.(1996). The FA solutionswere made up daily at pH 7 and stored in airtightvolumetric flasks.The adsorption of FA by the soils in the absence

of pesticides was determined using the same meth-od as for the pesticide adsorption studies(described below), with three types of backgroundelectrolytes; 0.005 M CaCl , 0.01 M CaCl and2 2

0.05 M NaCl, none of which caused precipitationof the FA from solution. The concentration oforganic matter in solution was determined asdescribed in the previous paragraph.

2.3. Pesticides

The pesticides used in the studies were: atrazine(6-chloro-N -ethyl-N -isopropyl-1,3,5-triazine-2,2 4

4-diamine, technical grade(99.0%)) kindly sup-plied by Ciba-Geigy, UK; 2,4-Dw(2,4-dichloro-phenoxy)acetic acidx supplied by Sigma ChemicalCompany, St. Louis, USA; isoproturonw3-(4-iso-propylphenyl)-1,1-dimethylurea, technical grade(99.0%)x kindly supplied by Rhone-Poulenc Agri-culture Ltd., Essex, England; paraquatw1,19-dimethyl-4,49-bipyridinium, technical grade(98%)x supplied by Aldrich Chemical Co. Ltd.,Dorset, England.The radiolabeled chemicals of the above pesti-

cides were atrazine(atrazine-ring-UL- C), kindly14

supplied by Ciba-Geigy, 2,4-D(2,4-dichlorophen-oxyacetic acid-carboxyl- C,)98% purity) sup-14

plied by Sigma Chemical Company, isoproturon(isoproturon-ring-UL- C), kindly supplied by14

Ciba-Geigy and paraquat(paraquat-methyl- C14

dichloride,)98% purity), supplied by SigmaChemical Company.

3. Methods

3.1. Adsorption studies

Stock solutions of 2,4-D and paraquat contained0.010 gyl of the pesticide dissolved in water passed

through a Milli-Q Reagent Grade Water System(Millipore Corporation, Bedford, Massachusetts).Stock solutions of atrazine and isoproturon wereprepared by dissolving 0.010 g of the pesticide in10 ml of AR methanol and diluting the solutionto 1 l using Milli-Q water. The concentration ofpesticides used in this study ranged from 250 ppbto 10 ppm, the background solution was an elec-trolyte (0.005 M Ca -, 0.01 M Ca , or 0.05 M2q 2q

Na ) and soluble organic matter(0 or 500 ppmq

FA).Adsorption experiments were carried out in

triplicate using soil:solution ratio of 1 g of soil to2 ml of solution unless otherwise stated. Theadsorption solution was shaken on an end-over-end shaker for 16 h. Trials carried out usingSonning I and Broad soil indicated that, for bothsoils, adsorption after 16 h was;96% of thatamount adsorbed after 24 h of shaking for atrazineand isoproturon. The corresponding values for 2,4-D on Sonning I and Broad soils were;88% and;94%. Essentially 100% of the paraquat wasadsorbed even at the shorter shaking time of 16 h.At the end of the 16 h allowed for adsorption,

the samples were removed from the shaker andthe soil suspension was centrifuged at 20 000=gfor 20 min. Duplicate aliquots of the supernatantwere removed for analysis. The amount of pesti-cide in the supernatant solution was determinedusing a liquid scintillation counter which had beencalibrated for solutions containing 0–1000 ppmFA in solution to account for quenching effects ofthe FA.

3.2. Desorption studies

For desorption experiments, the initial adsorp-tion was as described above. After the removal ofthe two samples for analysis, further supernatantwas removed until only 25% of the original vol-ume of solution in the vessel remained. Thesolution in the vessel was then made up to theoriginal volume with the same background electro-lyte and either 0 or 500 ppm of added FA. Thesuspensions were equilibrated on a shaker for 16h, centrifuged and the supernatant was again ana-lysed. This process was repeated three more timesin the desorption studies.


4. Results

4.1. Characterisation of the soils

The results of the characterisation of the soilsare shown in Table 1. The three Sonning soilshave a similar particle size distribution with 55–60% of the particles being in the size rangeassociated with coarse sand(63–600 mm) andapproximately 10% of the particles being in thesize range associated with fine clays(-2 mm).The clay composition(Table 1) in these samplesis similar. However, the Sonning soil samplesdiffer in the amount of organic matter present with1.25, 2.05 and 3.30% total organic carbon forSonning I, II and III, respectively.Broad and Denchworth soils are much heavier

textured soils with a particle size distributionconsisting of a coarse sand content of approxi-mately 35 and 25%, respectively, and a fine claycontent of 35 and 45%, respectively. The natureof the clay in these samples is similar(Table 1).The main difference between these two smectiticclay soils is that the Broad soil has approximatelytwice the total organic carbon to that of Dench-worth (6.3 and 3.25%, respectively).The pH of the soil solutions were similar(Table

1), ranging from 7.0 to 7.6. The conductivity ofthese solutions ranged from 0.19–0.8 mSycm, thewDOCx ranged from 23–55 ppm, the solubleCa extracted using water ranged from 0.18–2q

10.0 mmolyg soil and the absorbance at 400 nmranged from 0.046–0.219(Table 1).The values of DOCyOC (%) shown in Table 1

are relatively similar for all soils except SonningII indicating that the organic matter in this soil isrelatively more soluble, possibly due to lowerwCa x (lowest of all soils). The values of2q

abs yDOC w(absyppm)=1000x show a great-400 nm

er range. In general it has been observed that forthe same DOC concentration the relative absorb-ance at 400 nm of humic acids is greater thanfulvic acids (Senesi et al., 1989). The highervalues of abs yDOC of the Sonning II and III400 nm

soils suggests that the organic matter in these twosoils has a higher ratio of humic:fulvic acid char-acter than that of the other soils. And correspond-

ingly, it suggests that the Denchworth soil has thelowest ratio of humic:fulvic acid character.The adsorption of the FA by the soils was found

to be directly proportional to the amount of FAadded to the soil solution(Fig. 1). The slope ofthe straight line representing this linear proportion-ality K for each of the soils is shown in Table 2,fa

together with the initial absorbance of the soilsolutions. The total amount of FA adsorbed wasleast in 0.005 M Ca (lowest K value) and2q

fa

greatest in 0.01 M Ca (highestK value) for2qfa

all soils. (It is important to note that the fulvicacid fraction used in this work remained in solutionin the presence of 0.01 M Ca background2q

electrolyte solution in the absence of soils.) Theamount of FA adsorbed for a particular type ofbackground electrolyte, was also dependent on thenature of the soil, increasing in the order SonningII-Sonning I s Sonning III-Broad-Dench-worth. This order correlates with the order ofrelative abundance of clay fraction(-2 mm) inthe soils(Table 1) with a statisticalR value for2

fit of 0.98, 0.94 and 0.98 for 0.005 M Ca , 0.052q

M Na and 0.01 M Ca , respectively. As Broadq 2q

soil had the highestwCa x in soil solution the2q

adsorbance of FA would appear to be moredependent on the clay content than the solublecalcium content.

4.2. Adsorption of pesticides

The characteristics of the adsorption at low ionicstrength(0.005 M Ca ) of the pesticides, atrazi-2q

ne, 2,4-D and isoproturon for the soils are shownin Fig. 2a,b,c, respectively. Over the range ofconcentrations used the adsorption is close to alinear relationship between the amount of pesticide(x) adsorbed per g of soil(m) as a function of theconcentration of the pesticide in solution at equi-librium (C).

xymsKd Cp

Where Kd is the distribution constant of thep

pesticide(over the equilibrium solution concentra-tion of 0–2500 ppb) between the soil and thesolution. TheKd values of the plots for all fivep


Fig. 1. The amount of FA sorbed from solution as a function of the initial concentration in the presence of the soils; Sonning I(m),Sonning II(h), Sonning III (j), Broad(s) and Denchworth(�).

Table 2Effect of the background electrolyte type and concentration on the absorbance(400 nm) of the soil solution(1:5 suspension; noadded FA) and on theK value(degree of sorbance of soluble FA from solution) for each of the soilsfa

Electrolyte Ionic Sonning Sonning Sonning Broad Dench-strength I II III worth

Absorbance 0.005 M Ca2q 0.015 0.012 0.135 0.119 0.188 0.075of soil 0.01 M Ca2q 0.03 0.016 nd 0.074 0.166 0.050solution 0.05 M Naq 0.05 0.043 nd 0.155 0.226 0.064

K valuefa 0.005 M Ca2q 0.015 0.26 0.19 0.27 0.52 0.600.01 M Ca2q 0.03 0.57 nd 0.59 0.69 0.800.05 M Naq 0.05 0.39 nd 0.48 0.59 0.75

Note: nd means not determined.

soils are shown in Table 3. The order of adsorptionon all soils is 2,4-D-atrazine-isoproturon-par-aquat. The order of adsorption of the pesticidesatrazine and 2,4-D on each of the soils is SonningI-Sonning II-Sonning III f Denchworth-Broad, which is the order of abundance of OC inthe soils. This order is slightly different for isopro-turon on the soils as the position of Broad andDenchworth are reversed i.e. Sonning I-SonningII-Sonning III-Broad-Denchworth. Paraquatwas completely adsorbed using these experimentalconditions and concentrations(as shown in Fig.

2a) and so other graphical results and theKdpvalues for this pesticide are not included here.The corresponding ratio of theKd value to thep

OC of the soil,Kd yOC, is also given in Table 3.p

The range in these values for all soils is0.56"0.16, 0.12"0.03 and 0.74"0.20 for atrazi-ne, 2,4-D and isoproturon, respectively,(represent-ing 25–30% variation in values). The threeSonning soils have similar mineralogy but differentlevels of organic matter(Table 1). Taking only theKd yOC values for the Sonning soils, and hencep

largely eliminating mineralogy as a factor in the


Fig. 2.(a) The amount of atrazine sorbed by the soil(gyg) as a function of the concentration of the pesticide in solution(backgroundelectrolyte 0.005 M Ca ) for the soils; Sonning I(m), Sonning II(h), Sonning III (j), Broad(s) and Denchworth(�). The2q

amount of paraquat sorbed by the soil(mgyg) as a function of the concentration of the pesticide in solution at low ionic strength(0.005 M Ca ) for Sonning I soil(d); (b) The amount of 2,4-D sorbed by the soil(mgyg) as a function of the concentration of2q

the pesticide in solution(background electrolyte 0.005 M Ca) for the soils; Sonning I(m), Sonning II (h), Sonning III (j),2q

Broad (s) and Denchworth(�); (c) The amount of isoproturon sorbed by the soil(mgyg) as a function of the concentration ofthe pesticide in solution(background electrolyte 0.005 M Ca) for the soils; Sonning I(m), Sonning II (h), Sonning III (j),2q

Broad(s) and Denchworth(�).


Fig. 2 (Continued).

Table 3Kd andKd yOC values for the soilypesticide systems, whereKd is the distribution constant of the pesticide between the soil andp p p

the solution and OC is the organic carbon content of the whole soil

Pesticide Sonning Sonning Sonning Broad Dench-I II III worth

Atrazine Kdp 0.46 1.32 2.23 2.49 2.23Kd yOCp 0.37 0.64 0.68 0.40 0.69

2,4 D Kdp 0.13 0.25 0.51 0.66 0.50Kd yOCp 0.10 0.12 0.15 0.10 0.15

Isoproturon Kdp 1.01 1.53 2.32 2.72 3.21Kd yOCp 0.81 0.75 0.70 0.43 0.99

adsorption characteristics, results in a smaller coef-ficient of variation for isoproturon (now0.75"0.06) but not for atrazine and 2,4-D.

4.3. Effect of ionic strength and concentration ofFA

The ionic strength(I.S.) of the 0.005 M Ca2q

solutions referred to in the above discussion is0.015. Using background electrolyte solutions ofhigher ionic strength, 0.01 M Ca (I.S. 0.03) or2q

0.05 M Na (I.S. 0.05) had no significant effectq

on the adsorption of atrazine and isoproturon onthe four soils studied(Sonning I, Sonning III,Broad and Denchworth). Repeating the adsorptionprocedure using 500 ppm FA in the adsorption

solution also had no effect on the adsorptioncharacteristics of these two pesticides in any ofthe three electrolyte solutions used in this work.However, these solution conditions did affect theadsorption of 2,4-D as shown in Fig. 3a, where itcan be seen that increasing the ionic strengthcaused the adsorption of 2,4-D to increase on allfour soils. The effect is greater for higher ionicstrengths solutions containing Ca (I.S. 0.03)2q

than Na (I.S. 0.05) and is reduced in the presenceq

of increased concentration of soluble FA. Paraquatwas completely adsorbed under these experimentalconditions and concentrations and therefore theeffects of ionic strength and added fulvic acid onthe adsorption were not detectable.


Fig. 3. The effect of the absence or presence of FA(500 ppm) and the nature and concentration of the electrolyte(0.005 M Ca ,2q

0.01 M Ca , 0.05 M Na) on the sorption of 2,4-D by soils(Sonning I, Sonning III, Broad and Denchworth) at high(1:2) and2q q

low (1:16) soil to solution ratio(Fig. 3a,b, respectively).


Fig. 4. Desorption of 2,4-D(n,m), atrazine(h,j) and isoproturon(s,d) from Sonning I soil in the absence(open symbolsydottedline) and presence(closed symbolsysolid line) of 500 ppm FA using the same electrolyte solution(0.005 M Ca ) and soil to2q

solution ratio(1:2) as that used in the initial adsorption solution containing 1000 ppb of the pesticide.

4.4. Effect of soilysolution ratio

Decreasing the soil to solution ratio from 1:2 to1:16 had no significant effect on the adsorptioncharacteristics of the pesticides atrazine, isoprotu-ron and paraquat by these soils in the presence ofthe electrolytes and FA solution as outlined in theMethods section. However, the adsorption of 2,4-D was significantly greater from suspensions pre-pared with a lower soil:solution ratio for all soilsfor the two higher ionic strength solutionsw0.01M Ca (I.S. 0.03) and 0.05 M Na (I.S. 0.05)x2q q

(Fig. 3b). The increased adsorption at the higherionic strength was less in the presence of FA,especially in the presence of the Na cations.q

4.5. Desorption of pesticides from soils

The desorption of the three pesticides fromSonning I is shown in Fig. 4. The desorption ofparaquat was negligible and 2,4-D was only slight-ly desorbed under the conditions used here. Atra-zine and isoproturon were significantly desorbedwith a small amount of hysteresis. For all fourpesticides the addition of 500 ppm FA to thedesorption solution had no discernible effect onthe desorption characteristics.

5. Discussion

5.1. Pesticide adsorption

The extent of adsorption of the pesticide underany particular soil type and soil solution conditionswill depend on the nature and properties of thesoil and the pesticide. Atrazine and isoproturonboth have functional groups capable of engagingin H-bonding, van der Waals bonding and ligandexchange. However, the presence of the C_Obond in isoproturon considerably enhances thepotential for H-bonding for this pesticide relativeto that for atrazine. In addition the presence of theC_O bond will result in the overall polarity ofthe isoproturon structure being greater than thatfor atrazine. A more polar molecule is more likelyto move closer to charged surfaces increasing thelikelihood of van der Vaals interactions.Paraquat and 2,4-D both exist as charged species

at pH values approximately 7 and so both couldengage in ionic bonding as well as the otherbonding mechanisms referred to above. As soilorganic matter tends to exhibit predominantly neg-atively charged adsorption surfaces(Beckett and


Le, 1989) the delocalized positive charge associ-ated with the paraquat ion would be expected tobe considerably more interactive with soil com-ponents than the localised negatively charged car-bonyl group of the 2,4-D ion. The chargeinteraction of the 2,4-D ionic species with thenegatively charged soil mineral and organic mattersurfaces is more likely to involve interaction withcationic species such as Ca in salt bridge2q

arrangements, although it may also participate inhydrogen bonding interactions through the carbon-yl group.The four pesticides studied vary in the extent of

adsorption of the pesticide by the soil, from vir-tually complete adsorption for paraquat to limitedadsorption for 2,4-D(Fig. 2a,b,c). The order ofextent of adsorption of the pesticides, paraquat)isoproturon)atrazine)2,4-D exemplifies theabove discussion regarding reactivity of isoprotu-ron relative to atrazine and paraquat relative to2,4-D. For the soils and conditions used in thisstudy the ionic bonding associated with the posi-tive delocalized charge of paraquat appears to bethe most favourable mechanism of bonding to thesoil components.

5.2. Total organic carbon content and adsorption

Apart from the clay mineralogy, the main dif-ferences between the soils are the nature and totalamount of organic matter present. While the humicsubstances in any given soil will vary widely incomposition, most samples will contain similarstructural units and functional groups(Stevenson,1972), although the relative proportions of thesemay vary. As the three Sonning soils have similarmineralogy, differences in interaction of the pesti-cides with the Sonning soils would be expected tobe related generally to the nature and quantity ofcarbon present. From the varyingKd yOC valuesp

for the Sonning soils for atrazine and 2,4-D it isapparent that the adsorption of the two pesticidesis not dependent only on the total carbon in thesoil. For both of these pesticides theKd yOCp

value is less for Sonning I than for the other twoSonning soils.From the results presented in this work it would

appear thatKd yOC values are not reliable indi-p

cators of the extent of adsorption of pesticides bysoils, even for soils of similar clay content. Anexplanation for these discrepancies from conclu-sions generally made in the literature and outlinedin the introduction, may be that the total carboncontent of a soil does not account for the natureor location of the organic matter. It is quite possiblethat some of the organic matter may be inaccessi-ble to the pesticide if it is associated within solidstate humic particles or clay aggregates, reducingthe effective level of active OC in the soil(Skjem-stad et al., 1996; Barriuso and Koskinen, 1996).Alternatively, it is possible that some soils mayhave a lower proportion of reactive to unreactiveorganic matter present, again reducing the effectiveOC of the soil. The variation in the reactivity ofthe organic carbon may not be related just to thenature and origin of the organic matter but it mayalso be related to the nature of the mineral towhich it is sorbed and the solution conditionsprevailing when the sorption occurred(Murphy etal., 1992).

5.3. Dissolved organic carbon content andadsorption

In previous studies(Spark and Swift, 1994) itwas found that there is a strong interaction betweenparaquat and FA in solution, but no significantinteraction between atrazine, 2,4-D and isoproturonand the solution FA. The results from this presentwork support these previous results with respectto atrazine and isoproturon. From the values ofabs yDOC (Table 1) the humic:fulvic acid400nm

character of the pre-existing soluble organic matterin the Sonning soils decreases in the order SonningII)Sonning III)Sonning I. However, there is nocorrelation between this order and the order ofeither theKd or Kd yOC values(Table 3). Inp p

addition, the presence of high concentrations ofFA had no effect on the adsorption properties ofthese pesticides by soils for the conditions used inthese studies. Because of the strong adsorptioncharacteristics of paraquat for all the soils at theconcentrations used in this work, it is not possibleto distinguish the effect of increased dissolvedorganic carbon on the adsorption of this pesticideby soils.


2,4-D was the only pesticide to show a changein sorption behaviour(sorption decreased) withthe soils as a result of an increase in solubleorganic matter, although this pesticide exhibitedno discernible interaction with FA in the absenceof soil (Spark and Swift, 1994). Compared withthe other pesticides studied here, the adsorption of2,4-D from solution by all five soils is relativelysmall. 2,4-D will be negatively charged above pH4 ( K s2.73, Tomlin 1997) and would experienceP a

repulsion by the predominantly negatively chargedsurface or organic matter in the soil matrix. Despitethis, the adsorption of 2,4-D is usually associatedmost strongly with the organic matter content ofthe soil (Reddy and Gambrell, 1987; Hermosinand Cornejo, 1993; Bekbolet et al., 1999).Adsorption of 2,4-D is not likely to involve H-

bonding for these soil conditions as H-bondingwill be limited to acid conditions where carboxylgroups are not ionised(Barriuso et al., 1992). Theadsorption of 2,4-D on humic acid is thought toinvolve physical adsorption possibly involving vander Waals forces and hydrophobic bonding, withthe rate limiting step for the initial process beingthe diffusion of the pesticide molecules to thesurface of the humic acid and for the secondprocess being the diffusion of the pesticide mole-cules into the humic particles(Khan, 1973).The increased adsorption of 2,4-D at high

Ca suggests that the bonding mechanism for2q

part of the uptake could involve a salt bridge withthe Ca cation acting as the bridge between the2q

pesticide and the negatively charged organic matteror negatively charged clay particles. The saltbridge would be favoured at pH)7, and wouldovercome the problem of repulsion between thenegatively charged 2,4-D and the negative humicsubstances or clay minerals. As the adsorption of2,4-D is reduced with increase in FA concentrationit is consistent to propose that the negativelycharged FA added to the soil suspension competeswith the negatively charged pesticide for adsorp-tion sites involving the salt bridging.The effect of increased ionic strength on the

adsorption of 2,4-D at 1:2 soil:solution ratio ismore strongly evident when using a soil:solutionratio of 1:16 for the Broad and Denchworth soils(Fig. 3b). These two soils contain significantly

more clay-sized particles per gram of soil than theSonning soils, and hence would be expected tohave significantly more charged mineral surfacearea in solution. At the Na higher ionic strengthq

the thickness of the double layer around thecharged soil particles would be reduced, enhancingthe likelihood of the 2,4-D overcoming the electro-static repulsion enough to be adsorbed via thephysical adsorption processes mentioned above.

5.4. Soil mineralogy and adsorption characteristics

For soil suspensions of equivalent mass, thequantity and nature of the smaller particle sizefractions will contribute most to the surface areaand hence largely control the overall adsorptioncharacteristics of the soil. This fraction will consistmainly of clays and aluminium and iron oxides.With these considerations, the mineral componentsof Denchworth and Broad soils would be expectedto adsorb the pesticides to the greatest extent basedon surface area availability, with the Sonning soilsadsorbing the least amount of pesticide. The resultsindicate that, similar to the discussion on soilorganic matter content, the sorption of the pesti-cides is not simply a function of the clay contentas the adsorption characteristics of the Sonningsoils vary by as much as a factor of 5(e.g.atrazine) when the difference in clay content is atthe most a factor of 1.5.There are frequent reports in the literature that

the adsorption of pesticides is not solely related toeither the organic matter content or the clay contentalthough it appears to be positively correlated withboth of these soil components. Celis et al.(1999)concluded that for atrazine and simazine adsorp-tion the extent of adsorption on individual soilcomponents (montmorillonite, ferrihydrite andhumic acid) did not relate directly to the adsorptionof the pesticides on inter-associated combinationsof two or more of these individual components. Itwas concluded that the adsorption sites for thepesticide may be reduced in number or change incharacter as a result of the association betweensoil components. Spark et al.(1997) found similarresults for the sorption of heavy metals on individ-ual minerals or humic acid compared with that onminerals coated with humic acid.


5.5. Desorption characteristics

The desorption of atrazine and isoproturon fromSonning I showed similarities(Fig. 4) despite thedifferences in adsorption characteristics(Table 3).In accordance with the findings of other workers,the desorption of these pesticides although signif-icant, exhibited a hysteresis, which indicates thatthe interaction is not truly reversible(Clay et al.,1988; Piccolo et al., 1992). Harris and Warren(1964) reported that the desorption of atrazinefrom organic matter was less than that from clays.The use of increased concentrations of FA in thedesorption solution had very little effect on theextent of desorption of these two pesticides inagreement with the findings of Clay et al.(1988)for atrazine.The desorption of 2,4-D was relatively small,

compared to that of atrazine and isoproturon. Asfor atrazine and isoproturon, the higher concentra-tions of FA in the desorption solution had no effecton the characteristics of desorption for 2,4-D. Asionic bonding involving a salt bridge such asCa would be expected to be reversible, it would2q

appear that the 2,4-D was involved in a morespecific interaction with the humic surface follow-ing the initial salt-bridge bonding process. Isaacsonand Frink (1984) found similar results for 2,4-Dwith up to 90% of the pesticide being irreversiblyadsorbed. Brusseau et al.(1989) suggested thatthe sorption non-equilibrium of 2,4-D may be aresult of intra-organic matter diffusion as theyfound that rate constants showed an inverse rela-tionship to organic matter content.There is considerable evidence that paraquat

interacts strongly with soluble organic matter, asdiscussed above. The results in this work showthat there is negligible desorption of paraquat fromthe soils in the absence or presence of increasedcontent of soluble fulvic. This lack of desorptionwould result in the paraquat having limited inter-action with the soluble fraction of organic matterin the soil.

6. Conclusion

The adsorption of pesticides by soils is domi-nated by the interaction with the solid state of the

soil. From the results presented here and those ofothers, the solid state interactions involves boththe organic and inorganic fractions, the relativeimportance of which will depend on the relativeabundance of these two fractions.The soluble organic matter in the soil solution

had no effect on the sorption characteristics ofatrazine, isoproturon and paraquat. This fractiondid affect the sorption characteristics of 2,4-D butthis is thought to be due to competitive effects ofthe organic matter with the sorption of the pesticiderather than due to the interaction of 2,4-D withthe soluble fraction of the organic matter.

Acknowledgments

This work was funded by a research grant fromthe BBSRC, UK. We are grateful to John Cookefor technical assistance.

References

Avery, BW. Soils of the British Isles. CAB International,Wallingford, UK, 1990.

Avery, BW., Bascomb, CL. Soil Survey Laboratory Methods.Harpenden, UK, 1982.

Arienzo M, Buondonno A. Adsorption of paraquat by TerraRossa soil and model soil aggregates. Toxicol Environ Chem1993;39:193–199.

Bailey GW, White JL. Review of adsorption and desorptionof organic pesticides by soil colloids with implicationsconcerning pesticide bio-activity. J Agric Food Chem1964;2:324–332.

Barriuso E, Baer U, Calvet R. Dissolved organic matter andadsorption-desorption of dimefuron, atrazine, and carbetam-ide by soils. J Environ Qual 1992;21:359–367.

Barriuso E, Koskinen WC. Incorporating nonextractable atra-zine residues into soil size fractions as a function of time.Soil Sci Soc Am J 1996;60:150–157.

Baskaran S, Bolan NS, Rahman A, Tillman RW. Pesticidesorption by allophanic and non-allophanic soils of NewZealand. New Zealand J Agric Res 1996a;39:297–310.

Baskaran S, Bolan NS, Rahman A, Tillman RW. Effect ofexogenous carbon on the sorption and movement of atrazineand 2,4-D by soils. Aust J Soil Res 1996b;34:609–622.

Beckett R, Le NP. The role of organic matter and ioniccomposition in determining the surface charge of suspendedparticles in natural waters. Colloid Surf 1989;44:35–49.

Bekbolet M, Yenigun O, Yucel I. Sorption studies of 2,4-D onselected soils. Water Air Soil Pollut 1999;111:75–88.

Borggaard OK, Streibig JC. Atrazine adsorption by some soilsamples in relation to their constituents. Acta Agric Scan1988;38:293–301.


Brusseau ML, Rao PSC, Jessup RE, Davidson JM. Flowinterruption: A method for investigating sorption non-equi-librium. J Contam Hydrol 1989;4:223–240.

Burns IG, Hayes MHB, Stacey M. Studies of the adsorptionof paraquat on soluble humic fractions by gel filtration andultrafiltration techniques. Pest Sci 1973;4:629–641.

Celis R, Hermosin MC, Cox L, Cornejo J. Sorption of 2,4-dichlorophenoxyacetic acid by model soil particles simulat-ing naturally occurring soil colloids. Environ Sci Technol1999;33:1200–1206.

Chiou CT, Malcolm RL, Brinton TI, File DE. Water solubilityenhancement of some organic pollutants and pesticides bydissolved humic and fulvic acids. Environ Sci Technol1986;20:502–508.

Clay SA, Allmaras RR, Koskinen WC, Wyse DL. Desorptionof atrazine and cyanazine from soil. J Environ Qual1988;17:719–723.

Devitt EC, Wiesner MR. Dialysis investigations of atrazine-organic matter interactions and the role of a divalent metal.Environ Sci Technol 1998;32:232–237.

Fitch A, Du J. Solute transport in clay media: Effect of humicacid. Environ Sci Technol 1996;30:12–15.

Gamble DS, Haniff MI, Zienius RH. Solution phase complex-ing of atrazine by fulvic acid: A batch ultrafiltration tech-nique. Anal Chem 1986;58:727–731.

Gao JP, Maguhn J, Spitzauer P, Kettrup A. Distribution ofpesticides in the sediment of the small Teufelsweiher pond.Water Res 1997;31:2811–2819.

Harris CI, Warren GF. Adsorption and desorption of herbicidesby soil. Weeds 1964;12:120–126.

Hayes MHB. Adsorption of triazine herbicides on soil organicmatter including a short review on soil organic matterchemistry. Residue Rev 1970;32:131–174.

Hermosin MC, Cornejo J. Binding mechanism of 2,4-dichlo-rophenoxyacetic acid by organo-clays. J Environ Qual1993;22:325–331.

Isaacson PJ, Frink CR. Non-reversible sorption of phenoliccompounds by sediment fractions: the role of sedimentorganic matter. Environ Sci Technol 1984;18:43–48.

Jarvis, MG. Soils of the Wantage and Abingdon District(Memoirs of the Soil Survey of Great Britain-England andWales) Agricultural Research Council, Harpenden, 1973.

Jenks BM, Roeth FW, Martin AR, McCallister DL. Influenceof surface and subsurface soil properties on atrazine sorptionand degradation. Weed Sci 1998;46:132–138.

Johnson AC, Haria AH, Cruxton VL, Batchelor CH, WilliamsRJ, Heppell K, Burt T. Isoproturon and anion transport bypreferential flow through a drained clay soil. Pesticidemovement to water. In: Walker A, Allen R, Bailey SW,Blair AM, Brown CD, Gunther P, Leake CR, Nicholls PH,editors. Proceedings of a symposium held at the Universityof Warwick, Coventry, UK 1995. p. 105–110.

Kan AT, Tomson MB. Ground water transport of hydrophobicorganic compounds in the presence of dissolved organicmatter. Environ Toxicol Chem 1990;9:253–263.

Khan SU. Equilibrium and kinetic studies of the adsorption of2,4-D, and picloram on humic acid. Canad J Soil Sci1973;53:429–434.

Kulovaara M. Distribution of DDT and Benzowax(pyrenebetween water and dissolved matter in natural humic water.Chemosphere 1993;27:2333–2340.

Madhun YA, Young JL, Freed VH. Binding of herbicides bywater-soluble organic materials from soil. J Environ Qual1986;15:64–68.

Maquedo C, Morillo E, Martin F, Undabeytia T. Interaction ofpesticides with the soluble fraction of natural and artificialhumic substances. J Environ Sci Health B 1993;28:655–670.

Masse L, Prasher SO, Khan SU, Arjoon DS, Barrington S.Leaching of metolachlor, atrazine, and atrazine metabolitesinto groundwater. Trans ASAE. 1994;37:801–806.

Miller JJ, Foroud N, Hill BD, Lindwall CW. Herbicides insurface runoff and groundwater under surface irrigation insouthern Alberta. Canad J Soil Sci 1995;75:145–148.

Moreaukervevan C, Mouvet C. Adsorption and desorption ofatrazine, deethylatrazine, and hydroxyatrazine by soil com-ponents. J Environ Qual 1998;27:46–53.

Murphy EM, Zachara JM, Smith SC, Phillips JL. The sorptionof humic acids to mineral surfaces and their role in contam-inant binding. Sci Total Environ 1992;117y118:413–423.

Piccolo A, Scheunert I, Schroll R. Supercritical vs. normalmethanol mass extraction of C labelled atrazine bound to14

humic substances. Fresen Environ Bull 1992;1:627–631.Reddy KS, Gambrell RP. Factors affecting the adsorption of2,4-D and methyl parathion in soils and sediments. AgricEcosystems Environ 1987;18:231–241.

Riise G, Eklo OM, Lode O, Pettersen MN. Mobility of atrazineand tribenuron methyl in the soil water system lysimeterexperiments. Contamination of pesticides from agricultureand industrial areas to soil and water. Norw J Agric SciSupp 1994;13:31–41.

Ro KS, Chung KH, Chung YC, Tsai FJ. Pesticides andherbicides. Water Environ Res 1997;69:664–667.

Senesi N, Miano TM, Provenzano MR, Brunetti G. Spectro-scopic and compositional comparative characterisation ofIHSS reference and standard fulvic acid and humic acids ofvarious origin. Sci Total Environ 1989;81y2:143–156.

Skjemstad JO, Clarke P, Taylor JA, Oades JM, McClure SG.The chemistry and nature of protected carbon in soil. AustJ Soil Res 1996;34:251–271.

Spark KM, Swift RS. Investigation of the interaction betweenpesticides and humic substances using fluorescence spec-troscopy. Sci Total Environ 1994;152:9–17.

Spark KM, Wells JD, Johnson BB. Sorption of heavy metalsby mineralyhumic acid systems. Austral J Soil Res1997;35:113–122.

Stevenson FJ. Organic matter reactions involving herbicidesin soils. J Environ Qual 1972;1:333–343.

Tomlin CDS. The Pesticide Manual, 11th Edition. British CropProtection Council, Surrey, UK, 1997.


Walker, A., Crawford, DV. The role of organic matter in theadsorption of the triazine herbicides by soils. In Isotopesand radiation in soil organic matter studies. InternationalAtomic Energy Agency, Vienna, 1968.

Watt BE, Malcolm RL, Hayes MHB, Clark NWE, ChipmanJK. The chemistry and potential mutagenicity of humic

substances in waters from different watersheds in Britainand Ireland. Water Res 1996;30:1502–1516.

Welhouse GJ, Bleam WF. NMR spectroscopic investigation ofhydrogen bonding in atrazine. Environ Sci Technol1992;26:959–964.

effect of soil composition and dissolved organic matter on pesticide sorption

Documents