Chemosphere 81 (2010) 351–358

Contents lists available at ScienceDirect

Chemosphere

journal homepage: www.elsevier .com/locate /chemosphere

Assessing pesticide leaching and desorption in soils with different agriculturalactivities from Argentina (Pampa and Patagonia)

Mariana Gonzalez a,b,d,*, Karina S.B. Miglioranza a,d, Julia E. Aizpún a, Federico I. Isla c,d, Aránzazu Peña b

a Laboratorio de Ecotoxicología, Facultad de Ciencias Exactas y Naturales, Universidad Nacional de Mar del Plata, 7600 Mar del Plata, Argentinab Departamento de Geoquímica Ambiental, Instituto Andaluz de Ciencias de la Tierra (CSIC-UGR), Granada, Spainc Instituto de Geología de Costas y del Cuaternario, Facultad de Ciencias Exactas y Naturales, Universidad Nacional de Mar del Plata, Mar del Plata, Argentinad Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Argentina

a r t i c l e i n f o

Article history:Received 17 February 2010Received in revised form 13 July 2010Accepted 14 July 2010Available online 11 August 2010

Keywords:Organochlorine pesticidesPyrethroid insecticidesAnionic surfactantsNon-ionic surfactantsOrganic acidsDOC

0045-6535/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.chemosphere.2010.07.021

Abbreviations: cmc, critical micelle concentraticarbon; ENDO, sum of a-endosulfan, b-endosulfanhumic acids; LD, La Dulce; OC, organic carbon; OCPSDS, sodium dodecyl sulfate; Sl, sewage sludge; TW80WTP, wastewater treatment plant; WW, wastewater.

* Corresponding author at: Laboratorio de EcotoxiExactas y Naturales, Universidad Nacional de Mar del P7600 Mar del Plata, Argentina.

E-mail address: mariana.gonzalez@conicet.gov.ar (

a b s t r a c t

Pesticide distribution in the soil profile depends on soil and pesticide properties as well as on the com-position of irrigation water. Water containing surfactants, acids or solvents, may alter pesticide desorp-tion from soil. The distribution of organochlorine pesticides (OCPs) in two Argentinean agricultural areas,Pampa and Patagonia, was evaluated. Furthermore, pesticide desorption from aged and freshly spikedsoils was performed by the batch technique, using solutions of sodium oxalate and citrate, dissolvedorganic carbon (DOC), wastewater and surfactants. Patagonian soil showed the highest OCP levels(46.5–38.1 lg g�1 OC) from 0 to 30 cm depth and the predominance of p,p0-DDE residues reflected anextensive and past use of DDT. Pampean soil with lower levels (0.039–0.07 lg g�1 OC) was mainly pol-luted by the currently used insecticide endosulfan. Sodium citrate and oxalate, at levels usually exudedby plant roots, effectively enhanced desorption of p,p0-DDT, p,p0-DDE and a-cypermethrin, while noeffects were observed for a-endosulfan and endosulfan sulfate. The non-ionic surfactant Tween 80behaved similarly to the acids, whereas the anionic sodium dodecyl sulfate enhanced desorption of allpesticides. Increased desorption of the hydrophobic pesticides also occurred when DOC from humic acidsbut not from sewage sludge or wastewater were used. Soil profile distribution of pesticides was in accor-dance with results from desorption studies. Data suggest pesticide leaching in Pampean and Patagoniansoils, with risk of endosulfan to reach groundwater and that some organic components of wastewatersmay enhance the solubilisation and leaching of recalcitrant compounds such as p,p0-DDT and p,p0-DDE.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Pesticides are necessary in modern agriculture in order to im-prove crop yield, however they are toxic substances that can alsorepresent a risk to the environment, depending on their character-istics and application rates. Particularly most of the organochlorinepesticides (OCPs) are classified as persistent organic pollutants(Stockholm Convention UNEP, 2004) since they persist in the envi-ronment, bioaccumulate and have the potential to undergo longrange transport (Wania and MacKay, 1996). Besides DDT (dichloro-

ll rights reserved.

on; DOC, dissolved organicand endosulfan sulfate; HA,s, organochlorine pesticides;, Tween 80; VR, Villa Regina;

cología, Facultad de Cienciaslata, Funes 3350, nivel +1.80,

M. Gonzalez).

diphenyltrichloroethane), its metabolites and HCH (hexachlorocy-clohexane) isomers, this group includes compounds such asdieldrin and heptachlor (all of them with forbidden or restricteduse), as well as the currently used technical endosulfan (a-/b-iso-mers in a 70:30 ratio). On the other hand, we can mention pyre-throid insecticides which are assumed to be less toxic, but whichcould also constitute a hazard to the environment due to their highhydrophobicity and the similarity of their structures to that of DDT(Demoute, 1990). All these features and their joint use with endo-sulfan have compelled us to consider them as targets of soil pollu-tion studies.

The target of pesticide application is soil or crop surface, butdepending on its own properties and the interactions with soilmineral fractions and organic matter it could also affect the sur-rounding environment (Sabljic, 2001). As the time of contact be-tween pesticides and soil increases, there is a decrease in theirchemical and biological availability (Semple et al., 2003; Katayamaet al., 2010). Pesticide availability, considered as its potential tomove through soils, depends on the mechanisms and kinetics of

352 M. Gonzalez et al. / Chemosphere 81 (2010) 351–358

pesticide sorption and desorption from soil particles (Moormanet al., 2001; Walker et al., 2005). Once in soil, aged and fresh pes-ticide residues could be prone to desorption under certain circum-stances. So, as their availability increases pesticides could leave thesoil by volatilization or degradation, leach to deeper soil layers orrun off. Moreover, they could be taken up by plant roots with thepossibility of being translocated to aerial plant parts (Trapp andMcFarlane, 1995; Mo et al., 2008).

Quantities and qualities of pesticides used in Argentina varywith the productive system. Climatic and pedological characteris-tics of the vast Argentinean regions determine the kind and exten-sion of agricultural activities. Extensive soybean and wheatproductions are concentrated on the rolling Pampa Region and ac-count for the 80% of the total arable land of the country. This sys-tem is based on the direct seeding technique (to prevent loss ofsoil) of transgenic soybean together with the application of theherbicide glyphosate, and technical endosulfan and a-cypermeth-rin as insecticides. Rolling Pampa is characterized by severe soildamage due to water erosion during heavy rainfalls that are com-mon in the main pesticide application period (November–March,Jergentz et al., 2005) leading to an increasing risk of ground andsurface water pollution. Occurrence of technical endosulfan in sur-face water at levels over the maximum residue limits establishedfor protection of aquatic biota in the southeast of the Pampa region(Río Quequén Grande watershed) was attributed to losses from soilduring rainy seasons (precipitation P900 mm y�1) (Gonzalez et al.,2009).

On the other hand, Patagonian agriculture, in a semidesertic re-gime with precipitation <200–300 mm y�1, is almost exclusivelybased on the fruit and wine production concentrated mainly onthe Rio Negro watershed. Fruit production from this area suppliesboth local and international markets. The historical and current useof pesticides is reflected on the occurrence of OCPs in surface soils,sediments, macrophytes and fish from the Rio Negro Valley with aclear predominance of residues of DDT and its metabolite DDE fol-lowed by a- and b-endosulfan and the metabolite endosulfan sul-fate and HCHs (Miglioranza et al., 2008). Among the pyrethroids a-cypermethrin is also frequently employed at recommended dosesone order of magnitude lower than technical endosulfan (INTA,2004).

Total pesticide levels in surface soils from both regions indicatethat Rio Negro soils are highly polluted by DDE, with concentra-tions three orders of magnitude higher than those in QuequenGrande. Thus, though soils from both areas (Pampa and Patagonia)are polluted by the same pesticide groups, they differ in total levelsand intrinsic soil characteristics as well as in type of culture andmanagement. Due to this, pesticides will probably behave in a dif-ferent way.

Knowledge of the fate and behaviour of pesticides in agricul-tural soils is required for the assessment of water pollution relatedto human health risk and to the selection of remediation strategies.In the Patagonian region, the use of river water for plot irrigationmay alter pesticide sorption–desorption since it usually containssuspended solids and organic dissolved substances. Among them,root exudates, from riparian vegetation such as willows, could bepresent. Plant roots exude large amounts of organic compoundsthat account for 20–30% of photosynthetic products such as car-boxylic acids (Gao and Zhu, 2003). Oxalic, citric, malic, and succinicacids have been reported to enhance desorption of organic contam-inants in soil (White et al., 2003; Gao et al., 2010). Such processesmay support pesticide movement to deeper horizons and togroundwater levels. Moreover river waters receive domestic andindustrial discharges which could be a source of surfactants, men-tioned as the most abundant organic compounds in wastewatersfrom domestic use (Abu-Zreig et al., 2003). Surfactants may be em-ployed, either as barriers to prevent mobility of contaminants from

a point source of pollution (Rodríguez-Cruz et al., 2007) or to in-crease the solubility of organic contaminants (Cheng et al., 2008).Non-ionic and anionic surfactant micelles have been related to en-hanced solubilisation of hydrophobic organic compounds (Rosen,2004; Cheng and Wong, 2006; Wang and Keller, 2008). Besides,it has been recently demonstrated that various natural materialscan have surfactant properties (Lippold et al., 2008; Adani et al.,2010).

Therefore, it is necessary to know the fate and behaviour of pes-ticides in soils in order to identify more sustainable agriculturalpractices for the watershed. In this sense, the aims of this workwere (i) to study the distribution of OCPs in soil profiles fromtwo typical Argentinean agricultural areas that differ in their phys-icochemical characteristics, agricultural practices and total surfacepesticide content and (ii) further explore desorption of aged andfreshly sorbed compounds, by solutions of different compositionthat could have a bearing on their desorption from soil.

2. Materials and methods

2.1. Field study

2.1.1. Soil samplingIntact soil columns were obtained from Pampa and Patagonia

regions. Sampling sites were selected on the basis of previousknowledge about total surface pesticide levels and soil types. Pat-agonian soil samples (Aridisols) were taken between trees from atypical apple and peach field settled in Villa Regina city (fromnow on named VR) in the upper valley of the Rio Negro watershed(S39�04.901400W67�02.905900). Soils from this area are contaminatedwith p,p0-DDE, p,p0-DDT, a-endosulfan, endosulfan sulfate and a-cypermethrin (Miglioranza et al., 2008). Soils representative ofthe Argentinean Pampa (association of typical Argiudols and Udi-fluvent) were sampled from a soybean field near La Dulce village(named LD) in the Río Quequén Grande watershed (S38�11.702900W59�08.803600). Surface soils had a total pesticide levellower than 2 � 10�6 mg g�1.

Soil columns, separated by at least 100 m, were obtained byintroducing an aluminium corer of 10 cm of diameter to a depth630 cm. Once the corers (n = 3 for each soil) were longitudinallyopened, soil profiles were characterized and subdivided into two(VR) and four (LD) layers according to the observed changes in col-our or structure. A composite sample of each layer from all corerswas used for determination of particle size distribution and organicmatter content in each site. For pesticide analysis each corer wasindividually analyzed considering four levels of 4–6 cm each fromtop to bottom. Subsamples were air dried, kept at room tempera-ture or frozen (�20 �C) until physicochemical or pesticide analyseswere performed. Main soil properties are shown in Table 1.

2.1.2. Pesticide extraction and analysisOCPs were extracted according to Metcalfe and Metcalfe (1997),

with modifications of Miglioranza et al. (2003). Subsamples of 10 gdry soil were homogenized with anhydrous sodium sulfate (purityP99%, Merck) and spiked with 20 ng of PCB #103 as internal stan-dard, to account for possible losses during extraction and analyticalprocess; they were Soxhlet extracted (8 h) with a mixture of hex-ane–dichloromethane (50:50), and then concentrated under vac-uum and nitrogen flow to a final volume of 2 mL. Clean up wasperformed by fractionation on activated (200 �C, 24 h) silica gelcolumn. Extracts were concentrated to 1 mL and kept in sealedvials at �20 �C prior to GC analyses.

OCPs were identified and quantified using a gas chromatograph,Shimadzu 17-A gas equipped with an autosampler and a 63Ni elec-tron capture detector (GC-ECD) and a capillary column (Supelco

Tabl

e1

Pest

icid

eco

ncen

trat

ion

and

mai

nph

ysic

oche

mic

alpr

oper

ties

ofso

ilpr

ofile

sfr

omPa

tago

nia

(Vill

aRe

gina

)an

dPa

mpa

(La

Dul

ce)

regi

ons.

HCH

s:a

-+

b-

+c-

+d-

isom

ers;

HEP

TA:

hept

achl

or+

hept

achl

orep

oxid

e;al

drin

s:al

drin

+di

eldr

in;

CHLO

R:a�

c-is

omer

s+

tran

s-no

nach

lor;

END

O:a

-,b-

isom

ers

+en

dosu

lfan

sulf

ate;

DD

Ts:

p,p0

-DD

T+

p,p0

-DD

E+

p,p0

-DD

D;

OCP

s:or

gano

chlo

rine

pest

icid

es;

OC:

orga

nic

carb

on.<

LD=

belo

wde

tect

ion

limit

.

Soil

profi

les

Phys

icoc

hem

ical

char

acte

rist

ics

Pest

icid

ele

vels

(ng

(gO

C)�

1,m

ean

±st

anda

rdde

viat

ion

)

Site

Dep

th(c

m)

Wat

er(%

)O

C(%

)Sa

nd

(%)

Silt

(%)

Cla

y(%

)H

CH

sH

EPTA

Ald

rin

sC

HLO

REN

DO

DD

TsR

OC

Ps

Pata

gon

ia(V

illa

Reg

ina)

0–6

17.9

2.8

29.6

a58

.4a

11.9

a14

9.5

±15

5.0

3.6

±4.

5<L

D8.

6±

12.4

1245

.0±

1880

.04.

5�

104

±4.

3�

104

4.6�

104

±4.

5�

104

6–12

15.8

1.6

70.3

±36

.01.

3±

2.2

<LD

8.0

±6.

923

0.0

±93

.13.

3�

104

±0.

7�

104

3.3�

104

±0.

7�

104

12–2

217

.31.

745

.3b

25.8

b28

.9b

111.

5±

59.0

2.0

±2.

2<L

D10

.4±

10.5

197.

0±

103.

82.

9�

104

±1.

4�

104

2.9�

104

±1.

5�

104

22–3

16.

70.

495

.3±

64.0

1.1

±2.

0<L

D12

.8±

4.4

181.

0±

73.4

3.8�

104

±1.

2�

104

3.8�

104

±1.

3�

104

Pam

pa(L

aD

ulc

e)0–

47.

61.

441

.350

.48.

314

.9±

2.4

0.5

±0.

3<L

D4.

3±

0.9

15.6

±6.

14.

2±

1.7

39.5

±7.

54–

1110

.50.

955

.827

.416

.816

.0±

5.4

3.7

±4.

8<L

D6.

6±

2.0

26.5

±1.

97.

8±

1.6

60.7

±6.

811

–20

10.0

0.7

63.9

8.6

27.5

4.5

±1.

4<L

D<L

D8.

6±

1.2

40.6

±3.

114

.5±

5.2

68.1

±10

.720

–30

11.6

1.1

61.2

12.3

26.5

6.8

±4.

5<L

D<L

D6.

2±

1.0

46.9

±18

.47.

7±

2.4

67.7

±15

.0

aV

alu

eob

tain

edfr

oma

com

posi

tesa

mpl

eof

the

0–12

cm.

bV

alu

eob

tain

edfr

oma

com

posi

tesa

mpl

eof

the

12–3

1cm

.

M. Gonzalez et al. / Chemosphere 81 (2010) 351–358 353

Inc.) coated with SPB-5 30 m � 0.25 mm i.d. � 0.25 lm film thick-ness. One microliter was splitless injected at 275 �C. The ECD tem-perature was 290 �C. The oven temperature program was: start at100 �C (1 min), then 5 �C min�1 up to 150 �C (1 min), then1.5 �C min�1 up to 240 �C, and then 10 �C min�1 up to 300 �C(10 min). Ultra-high purity helium was used as carrier gas(1.5 mL min�1) and nitrogen as make-up gas (Miglioranza et al.,2003).

2.1.3. Quality control and assuranceLaboratory and instrumental blanks analyzed throughout the

procedure indicate that there were not contaminations or interfer-ences on samples during laboratory handling. Recoveries, calcu-lated by spiked matrixes, were greater than 90%. Detection limits,according to Keith et al. (1983), ranged between 0.003 and0.005 ng g�1 for HCHs (a-, b-, c- and d-isomers) and between0.008 and 0.033 ng g�1 for the rest of chlorinated compounds[chlordane (a- and c-isomers and trans-nonachlor), DDTs (pp0-DDE, pp0-DDD and pp0-DDT), ENDO (a-, b-isomers and endosulfansulfate)]. The standard solutions used for identification and quan-tification of single compounds were obtained from Ultra Scientific,RI, USA and PCB #103 from Accustandard, INC, CT, USA.

A one way ANOVA followed by t-Student test or Mann–WhitneyU test were used to test significant differences in OCP levels be-tween sites. When parametric requirements were not fulfilled aswell as for comparing OCP levels throughout soil profile nonpara-metric ANOVA Friedman test followed by a t-paired test for depen-dent samples comparison were applied. The significance level wasset at a = 0.05, unless otherwise specified.

2.2. Desorption study

2.2.1. Surface soilsDesorption studies were carried out with samples of the upper

layer (0–15 cm) from both soils. Soils were taken from VR appleand LD soybean fields, air dried, crushed and sieved through a2 mm sieve. VR soil, already contaminated, was used directly whileLD soil was spiked with a mixture of p,p0-DDE, p,p0-DDT, a-endo-sulfan, endosulfan sulfate and a-cypermethrin (all from Dr. Ehren-storfer, Augsburg, Germany, with purity P96%). To spike the soil, asoil aliquot (5.0 g) was finely ground and added with an acetonesolution containing a mixture of the insecticides to give a nominalinitial concentration of 5 lg g�1 each (dry weight, dw) in the finalsoil sample. After solvent evaporation the samples were thor-oughly mixed with the rest of the soil (up to 200 g) to obtain ahomogeneous mixture. The dry contaminated soil was transferredto a bottle and aged for about 2 weeks at room temperature (Zhouand Zhu, 2007). Spiked and aged soil residues were analyzed aswas previously described in Sections 2.1.2 and 2.1.3. The main soilproperties, final pesticides levels and the place where they weresampled are shown in Table 2. Some relevant properties of thecompounds are shown in Table 3.

2.2.2. TreatmentsDesorption solutions were prepared in MilliQ water and soni-

cated for 15 min for a correct homogenization and the pH adjustedto 7 in order to avoid possible changes in soil dissolved organicmatter associated to pH changes (Reemtsma et al., 1999). Sodiumcitrate (Merck) and sodium oxalate (Probus) were prepared at0.05 and 0.1 M, concentrations close to reports concerning plantexudates (Gent et al., 2005). Solutions of the anionic surfactant so-dium dodecyl sulfate (SDS, Scharlau) and the non-ionic surfactantTween 80 (TW80, polyethylene glycol sorbitan monooleate, Sigma)were prepared at 2 and 10 times the critical micellar concentration(cmc); cmc values for SDS and TW80 are 2.4 and 0.016 g L�1,respectively (Mukerjee and Mysels, 1971; Chou et al., 2005).

Table 2Pesticide concentration and main physicochemical properties of aged (Patagonia, Villa Regina) and spiked (Pampa, La Dulce) upper layer soils (0–10 cm). OC: organic carbon.

Soil Physicochemical characteristics Pesticide concentration (lg g�1, dry weight; mean ± standard deviation)

Water(%)

OC(%)

Sand(%)

Silt(%)

Clay(%)

CEC meqz/100 g

a-endosulfan

endosulfan sulfate p,p0-DDT p,p0-DDE a-Cypermethrin

Patagonia. Villa Regina(S39�06.401000W67�03.508800)

10.8 2.7 14.1 62.9 23.0 18.6 0.003 ± 0.001 0.024 ± 0.010 0.05 ± 0.01 0.55 ± 0.09 0.003 ± 0.0004

Pampa, La Dulce(S38�15.001600W59�05.704600)

7.6 1.9 60.7 31.8 7.5 7.4 4.9 ± 0.3 5.2 ± 0.3 4.2 ± 0.3 5.8 ± 0.3 5.2 ± 0.2

354 M. Gonzalez et al. / Chemosphere 81 (2010) 351–358

Aqueous solutions containing 30 and 90 mg L�1 dissolved or-ganic carbon (DOC) concentration were prepared by shaking 1 gdewatered sewage sludge or humic acids (Fluka, Madrid) with10 mL Na2HPO4 50 mM (Prolabo) during 24 h (Reemtsma et al.,1999). The supernatant was diluted to give the required contentin DOC. Sewage sludge came from the wastewater treatment plant(WTP) from the city of Granada (Southeast Spain). Wastewater(WW) from the same origin was also used. It corresponded tothe effluents of the secondary sedimentation tank with pH 7.5,conductivity 1112 lS cm�1, 23 mg L�1 suspended solid contentand 30 mg L�1 OC content. All solutions were filtered through a0.45 lm Millipore and adjusted to pH 6–7 by adding few dropsof HCl solution (1 M).

2.2.3. Pesticide desorption and analysisDesorption was measured using batch duplicated experiments.

Soil (4 g) and desorption solutions (40 mL) were shaken in centri-fuge tubes end-over-end for 24 h at 20 �C, then centrifuged at1720 g for 15 min. The supernatant, to which 20 lL of fenarimolat 100 mg L�1 was added as the internal standard, was extractedby solid phase extraction. After conditioning the C18 cartridges,the retained compounds were eluted with 3 � 2 mL ethyl acetate,3 � 2 mL dichloromethane: hexane (1:1) and 3 � 2 mL hexane(de la Colina et al., 1996). Samples were frozen at �80 �C and theorganic phase brought to dryness in a rotary evaporator at 35–40 �C. The dry residue was dissolved in 1 mL hexane, transferredto a vial and analyzed by GC-ECD. Matrix effect was evaluated byanalyzing spiked solution of surfactants, organic acids, DOC andwastewaters with a mixture of all pesticides at 5 and 0.5 mg L�1.Additionally, blanks of the solutions used for desorption indicatedno pesticide contamination. Data are shown as the mean of twoindependent batch analysis injected twice in the GC-ECD.

3. Results and discussion

3.1. Soil profiles

3.1.1. Total OCPs levels and physicochemical characteristicsTotal OCP levels in VR were significantly higher than those from

LD (Table 1, p < 0.5). This difference was mainly due to DDTs andENDO groups; both OCPs together with HCHs were the maingroups in both sites. However, in LD the distribution pattern ofOCPs was ENDO > HCHs > DDTs (p < 0.05) and ENDO > DDTS > restof OCPs in the 0–12 cm and 12–30 cm, respectively. Distributionpattern in VR soil was DDTs >>>> ENDO = HCHs (p < 0.05) andDDTs >>> ENDO > HCHs (p < 0.05) in the 0–6 and 6–30 cm,respectively.

Results from soil profiles reveal that both sites had a differenthistorical use of pesticides, with a high dominance of DDTs in VRas a consequence of the intensive use of this insecticide during longtime. Data also indicate that DDT residues in the soil profile of VRare slowly removed after their application to the soil, because theuse of this insecticide is forbidden in Argentina since 1998 (SAG-PyA, 2005). Although technical endosulfan is currently used in both

watersheds, levels in VR were higher than in LD. Higher soil organ-ic carbon content in VR cannot explain the large differencesencountered. Environmental characteristics such as rainfall or agri-cultural management could lead to a higher ENDO loss in LD soils.If all endosulfan applied would have reached the soil surface andconsidering the recommended dose in soybean fields of0.2 mL (m2)�1 for technical endosulfan (35% active ingredient), to-tal residues in the upper soil layer (5 cm) should account for amaximum of 1.5 lg g�1. Residues found in the upper 0–4 cm ofLD soils are in the order of 0.2 ng g�1, while in VR with similarapplication dose values in the order of 35 ng g�1 were found. Thismeans that pesticides in LD are either not reaching the soil or elsemoving quickly, mainly due to the high precipitation regime. Be-sides, technical endosulfan application in LD is mainly carriedout by plane, with the concomitantly higher pesticide dispersionand reduction of the effective dose reaching soil surface. On thecontrary the use of portable knapsack sprayers for technical endo-sulfan application in VR allows a more effective input of pesticidesin the soil around or below the trees.

3.1.2. ENDO and DDTs distribution throughout the soil profile3.1.2.1. ENDO. Alpha-endosulfan is homogeneously distributedthroughout the profile in LD soil, except the upper layer, while b-and the sulfate metabolite increase and decrease their levels withdepth, respectively (Fig. 1a), without significant differences. Iso-mers proportion in technical endosulfan is 70:30 (a:b) and endo-sulfan sulfate production is known to be carried out mainly bybiological metabolism of the a-isomer. ENDO distribution in LDsoil profiles is in agreement with the current application of thetechnical mixture followed by the vertical movement of both iso-mers, the transformation of a- to sulfate in the upper layers dueto higher biological activity and accumulation of the most persis-tent b-isomer in deeper layers. In addition to the fast metabolismto sulfate, the absence of a-isomer in the upper layer (0–4 cm)could be related to sampling in July when technical endosulfanwas not being applied.

In VR soil the same behaviour was found both for a-isomer andsulfate metabolite, with a homogeneous distribution and upperlayer predominance, respectively (Fig. 1a). The distribution patternendosulfan sulfate > a- = b- in the 0–6 cm and a- > b- > endosulfansulfate in the 6–31 cm indicates that a- to sulfate metabolismleads to a deviation from the technical mixture. The occurrenceof the a-isomer in the upper layer is in relation to the applicationof technical endosulfan during the sampling time (summer).

3.1.2.2. DDTs. Fig. 1b shows p,p0-DDT, p,p0-DDE and p,p0-DDD levelsin soil profiles from LD and VR. As was discussed in Section 3.1.1,soil properties, different history in pesticide use and environmen-tal conditions are responsible for these OCPs distribution. DDTsdistribution in soil from LD was p,p0-DDT > p,p0-DDE and p,p0-DDT = p,p0-DDE for the 0–4 cm and subsequent soil depths, respec-tively. This result could be explained by lower inputs of fresh DDTand the migration of p,p0-DDE in the soil profile together withlosses by volatilization or plant uptake. Fresh p,p0-DDT could arise

Table 3Selected physicochemical properties of a-endosulfan, endosulfan sulfate, p,p0-DDT, p,p0-DDE and a-cypermethrin.

Properties Pesticides

a-endosulfan endosulfan sulfate p,p0-DDT p,p0-DDE a-cypermethrin

Molecular weight 406.9a 422.9a 354.5b 318.03b 416.3c

Water solubility (mg L�1) 0.32a 0.22a 0.025b 0.12b 0.004c

Log Kowd 3.83 3.66a 6.91 6.96 6.6f

Log Kocd 4.13 3.96e 5.31 4.82 5.5f

a Agency for Toxic Substances and Disease Registry www.atsdr.cdc.gov/toxprofiles/tp41-c3.pdf.b Agency for Toxic Substances and Disease Registry www.atsdr.cdc.gov/toxprofiles/tp155-c4.pdf.c Agency for Toxic Substances and Disease Registry www.atsdr.cdc.gov/toxprofiles/tp35-c4.pdf.d Sabljic et al. (1995).e Calculated as log Koc = 1.02 + 0.52 � log Kow, according to Sabljic et al. (1995).f Laskowski (2002).

Fig. 1. Distribution of (a) endosulfans and (b) DDTs in soil profiles from Patagonia (Villa Regina) and Pampa (La Dulce). Levels are expressed as ng (g organic carbon)�1. Labels:different letters indicate significant differences among isomers and metabolite within each soil depth (p < 0.05), �b-endosulfan values had no standard deviation because onevalue was below the detection limit and no statistical test was applied; + is an indication of statistically significant differences of the isomers or metabolite among soil depths(p < 0.05).

M. Gonzalez et al. / Chemosphere 81 (2010) 351–358 355

from the use of the acaricide dicofol since DDT appears as an impu-rity of its manufacturing (Qiu et al., 2005). In VR soil, the patternp,p0-DDE >>> p,p0-DDT > p,p0-DDD found in all depths indicatesan intensive and continuous use of p,p0-DDT in the past. Althoughdicofol is applied in this field the amounts of DDE entering the soilby this source should be negligible in comparison with found DDElevels. Microbial degradation and limited leaching would have ledto less DDTs concentration in deeper layers and some accumula-tion of p,p0-DDT and p,p0-DDE in the 12–20 and 20–30 cm, respec-tively. However, no statistically significant differences were foundamong depths.

3.2. Desorption assays

3.2.1. Organic acid esters solutionsDesorption of all pesticides was enhanced by both organic acid

esters and it was particularly noticeable for the more hydrophobiccompounds (p,p0-DDT, p,p0-DDE and a-cypermethrin) (Fig. 2a andb). Moreover, oxalate showed a better desorption ability than cit-rate. Pesticide desorption was not modified by increasing acid con-

centration and for sodium citrate a trend to desorption decreasewas observed at 0.1 M with respect to 0.05 M (Fig. 2a and b). Sim-ilar results have been already reported by other authors (Whiteet al., 2003; Luo et al., 2006). Endosulfans, with lower Kow, are al-ready more easily desorbed by water (Fig. 2a and b) with no ob-served effect of citrate or oxalate solutions.

Pesticide desorption was ranged as p,p0-DDE > p,p0-DDT > a-cypermethrin in both soils and for both acids independently ofthe initial soil levels (Table 2, Fig. 2a and b). Comparing treatmentswith their corresponding control, similar trends in desorption per-centage were found, with values between 200% and 500% (Fig. 2aand b). However, although the trend was maintained, the desorp-tion extent was always lower for the aged soil (VR) than for thefreshly spiked soil (LD). Soil profile from VR (Table 1, Fig. 1b) isin agreement with this desorption behaviour, since p,p0-DDE wasthe main compound in upper and all subsurface layers, indicatingthat the vertical movement was probably enhanced by time andsoil solution action.

The increased desorption in the presence of citrate and oxalateof the highly hydrophobic pesticides, p,p0-DDT, p,p0-DDE and a-

Fig. 2. Pesticide desorption (mg L�1 in solution) in surface soils from Patagonia (Villa Regina) and Pampa (La Dulce) by sodium oxalate (Oxa) and citrate (Ci) (a and b), bysynthetic surfactants sodium dodecyl sulfate (SDS) and Tween 80 (TW80) (c and d) and by wastewater and dissolved organic carbon from sewage sludge (Sl) or synthetichumic acids (HA) (e and f).

356 M. Gonzalez et al. / Chemosphere 81 (2010) 351–358

cypermethrin, was expected, since the effect of low molecularweight organic acids on the desorption of hydrophobic compoundsis mediated by soil organic carbon and DOC. When these carboxylicacids are added to soil, chelation of metal ions occurs followed bydisruption of organo-mineral linkages and the mobilization of or-ganic compounds into the aqueous phase by increasing DOC con-tents (Yang et al., 2001; Gao et al., 2010). As a consequence, thehydrophobic pesticides complexed with soil organic carbon willbe desorbed into the aqueous phase. Results indicate differencesin the retention mechanisms of endosulfans and DDTs. Both a-endosulfan and the sulfate metabolite, with higher water solubility(Table 3), were easily desorbed from the studied soils under thepresence of water or water-soluble organic acids in spite of the dif-ferent physicochemical characteristics of both soils. This result is of

particular interest since it indicates the leaching potential of thesecompounds to groundwater. This fact could be effectively feasibleconsidering the flooding events that frequently occur in bothwatersheds. Considering field data, endosulfan sulfate levels indeeper soil layers could be better explained as a consequence ofleaching since the a-isomer is homogeneously distributed. Accord-ingly, the metabolism of this isomer in deeper soil horizons, withlower microbial population, may be of less importance.

3.2.2. Surfactant solutionsDesorption enhancement occurred in both soils independently

of pesticide hydrophobicity. However, the magnitude of pesticidedesorption was dependent on surfactant and pesticide concentra-tion (Fig. 2c and d). In LD soil, with high levels of freshly spiked

M. Gonzalez et al. / Chemosphere 81 (2010) 351–358 357

pesticides, desorption with SDS at 2 cmc was 25 to 35-fold greaterthan that obtained with the control, and only a slight desorptionenhancement (up to 45-fold) was observed at 10 cmc (Fig. 2d).Conversely, increases in SDS concentration from 2 to 10 cmc ledto an evident increase of the tightly bound pesticide residues fromVR soil (Fig. 2c), which was contaminated at much lower levels.

TW 80 enhanced desorption of most pesticides at 2 cmc in bothsoils, more in VR than in LD. At 10 cmc pesticide desorption wasenhanced in LD soil, especially for DDTs and a-cypermethrin(Fig. 2d), whereas in VR soil the increase was moderate (30–65%for DDTs or no variation for a-cypermethrin) (Fig. 2c).

Soils with large content of clay minerals reduce considerablythe washing efficiency of non-ionic surfactants, which can besorbed onto the soil matrix resulting in the partition of organiccompounds into immobile sorbed surfactants (Sánchez-Camazanoet al., 2003; Wang and Keller, 2008). Thus, in VR, TW80 may inter-act with clay particles (23%) leading to adsolubilization or reten-tion processes that result in lower pesticide desorption at 10 cmcrelative to 2 cmc (Table 2, Fig. 2c). Conversely in LD, no or littleinteraction between TW80 and soil occurred. Moreover, pesticidedesorption was concomitantly enhanced with increased surfactantconcentration being this result in accordance with the low claycontent (7%) of LD soil (Table 2, Fig. 2d). Results are in agreementwith those reported by Zhou and Zhu (2007) who showed that an-other non-ionic surfactant, TX100, was more effective in desorbingPAHs from soils which had low clay content.

3.2.3. DOC and wastewaterTreated wastewater is normally used for land irrigation due to

the shortage of water for agricultural uses in arid and semiarid re-gions. However, its relatively high DOC content could affect thebehaviour of hydrophobic organic contaminants.

Desorption with DOC from sewage sludge (Sl) in the aged soil(VR) was enhanced for a-endosulfan, p0p0-DDE and a-cypermethrinat 30 mg L�1 (Fig. 2e). A 3-fold increase in DOC concentration led tosimilar desorption results. In LD soil, with freshly spiked pesticidesand at higher concentrations, no effect of DOC from sewage sludgewas observed (Fig. 2f).

DOC from humic acids (HA) enhanced desorption of pesticidesin both soils with respect to control. In aged soil (VR) increasingDOC concentration from 30 to 90 mg L�1 resulted only in theenhancement of p,p0-DDE desorption. However, in the freshlyspiked soil (LD) both polar (about 1.5-fold) and hydrophobic pesti-cides (from 4 to 10-fold) increased their desorption with increasingDOC concentration (Fig. 2e and f).

According to Cox et al. (2007) the presence of DOC may increaseor reduce retention of hydrophobic organic contaminants and theextent of these effects depends on the specific interactions be-tween soil, solute and liquid phases. DOC from pig manure andpig manure compost at concentrations similar to those used in thisstudy (Cheng and Wong, 2006) increased desorption from soil ofvarious polycyclic aromatic hydrocarbons, specifically that of themore hydrophobic compounds. In our case, enhanced desorptionof DDTs and a-cypermethrin occurred mainly when DOC fromHA but not from Sl was used. Sl from the WTP has a 40% OC contentfrom which 1.6% corresponds to HA, as previously reported (Sán-chez et al., 2003). Different nature of DOC could explain this differ-ent desorption behaviour and requires further investigation.

Recycled wastewater, with a DOC concentration of 30 mg L�1

only induced a relatively low increase of pesticide desorption inVR soil (Fig. 2e), similar to or even lower than that correspondingto 30 mg L�1 of DOC from Sl. Moreover in LD soil no effect ofWW was found, as described with Sl. These results were expectedsince DOC in WW and Sl are from the same origin. Water sourceswith similar characteristics to that from the city of Granada could

be used with low risk of desorption enhancement of these pesti-cides from soil.

4. Conclusions

Organochlorine and pyrethroid pesticides are considered highlyhydrophobic pollutants. The behaviour of these hydrophobic pesti-cides will depend not only on their intrinsic properties but also onenvironmental conditions and agricultural practices. The field sur-vey with soil corers has shown that irrigation and rainfall regimecan greatly affect the movement of these pesticides in the soil pro-file. Moreover, the relatively less hydrophobic pesticides, a-endo-sulfan and endosulfan sulfate, have a greater potential to reachwater sources.

On the other hand, desorption of these pesticides with aqueoussolutions mimicking the composition of low quality water poten-tially used in irrigation has shown a general enhancement of themore hydrophobic pesticides, p,p0-DDT, p,p0-DDE, and a-cyper-methrin. Most of the tested solutions have increased to a greateror lesser extent desorption of all pesticides, but the increase wasmore marked for the more hydrophobic pesticides, compounds ex-pected to be tightly retained on soil and practically immobile. Fur-ther research will be required to determine whether synergistic orantagonistic effects on pesticide desorption exist between acids,surfactants and DOC, since they could be released together intothe environment.

Acknowledgements

This work was supported by the Spanish Ministerio de MedioAmbiente y Medio Rural y Marino through project reference 037/SGTB/2007/6.1 and projects PNUD Proj. BC-42, and PIP 5668 Cons-ejo Nacional de Investigaciones Científicas y Técnicas (CONICET)from Argentina. M. Gonzalez stay at CSIC, Granada was supportedby CONICET and PICTR 327 ANPCyT. Wastewater and sewagesludge was kindly provided by the WTP of Granada (EMASAGRA).

References

Abu-Zreig, M., Rudra, R.P., Dickinson, W.T., 2003. Effect of application of surfactantson hydraulic properties of soils. Biosyst. Eng. 84, 363–372.

Adani, F., Tambone, F., Davoli, E., Scaglia, B., 2010. Surfactant properties andtetrachloroethene (PCE) solubilisation ability of humic acid-like substancesextracted from maize plant and from organic wastes: a comparative study.Chemosphere 78, 1017–1022.

Agency for Toxic Substances and Disease Registry. <www.atsdr.cdc.gov/toxprofiles/tp41-c3.pdf>, <www.atsdr.cdc.gov/toxprofiles/tp155-c4.pdf>; <www.atsdr.cdc.gov/toxprofiles/tp35-c4.pdf> (accessed July 2010).

Cheng, K.Y., Wong, J.W.C., 2006. Combined effect of nonionic surfactant Tween 80and DOM on the behaviors of PAH in soil–water system. Chemosphere 62,1907–1916.

Cheng, K.Y., Lai, K.M., Wong, J.W.C., 2008. Effects of pig manure compost and non-ionic surfactant Tween 80 on phenanthrene and pyrene removal from soilvegetated with Agropyron elongatum. Chemosphere 73, 791–797.

Chou, D.K., Krishnamurthy, R., Randolph, T.W., Carpenter, J.F., Manning, M.C., 2005.Effects of Tween 20 and Tween 80 on the stability of Albutropin duringagitation. J. Pharm. Sci. 94, 1368–1381.

Cox, L., Velarde, P., Cabrera, A., Hermosín, M.C., Cornejo, J., 2007. Dissolved organiccarbon interactions with sorption and leaching of diuron in organic-amendedsoils. Eur. J. Soil Sci. 58, 714–721.

de la Colina, C., Peña, A., Mingorance, M.D., Sánchez Rasero, F., 1996. Influence of thesolid-phase extraction process on calibration and performance parameters forthe determination of pesticide residues in water by gas chromatography. J.Chromatogr., A 733, 275–281.

Demoute, J.P., 1990. A brief review of the environmental fate and metabolism ofpyrethroids. Pestic. Sci. 27, 375–385.

Gao, Y., Zhu, L., 2003. Phytoremediation and its models for organic contaminatedsoils. J. Environ. Sci. 15, 302–310.

Gao, Y., Ren, L., Ling, W., Kang, F., Zhu, X., Sun, B., 2010. Effects of low-molecular-weight organic acids on sorption–desorption of phenanthrene in soils. Soil Sci.Soc. Am. J. 74, 51–59.

Gent, M.P.N., Parrish, Z.D., White, J.C., 2005. Exudation of citric acid and nutrientuptake among subspecies of cucurbita. J. Am. Soc. Hort. Sci. 130, 782–788.

358 M. Gonzalez et al. / Chemosphere 81 (2010) 351–358

Gonzalez, M., Miglioranza, K.S.B., Silva Barni, M.F., Peña, A., Ondarza, P.M., Aizpún, J.,Moreno, V., 2009. Contaminación por Compuestos Organoclorados (COCs) en unagroecosistema típico de la pampa Argentina (Cuenca del Río Quequén Grande)y su relación con la calidad del agua superficial y subterránea. In: Proceeding ofthe IX Congress of SETAC LA. II Congreso SETAC Perú: Ecotoxicología yresponsabilidad social, Lima, Perú, 5–9 October, 2009.

INTA, 2004. Guía de pulverizaciones para los cultivos de manzano, peral, frutales decarozo y vid. Centro Regional Patagonia Norte. Estación ExperimentalAgropecuaria Alto Valle, 132p.

Jergentz, S., Mugni, H., Bonneto, C., Schulz, R., 2005. Assessment of insecticidecontamination runoff and stream water of small agricultural streams in themain soybean area of Argentina. Chemosphere 61, 817–826.

Katayama, A., Bhula, R., Burns, G.R., Carazo, E., Felsot, A., Hamilton, D., Harris, C.,Kim, Y.-H., Kleter, G., Koerdel, W., Linders, J., Peijnenburg, J.G.M.W., Sabljic, A.,Stephenson, R.G., Racke, D.K., Rubin, B., Tanaka, K., Unsworth, J., Wauchope, R.D.,2010. Bioavailability of xenobiotics in the soil environment. Rev. Environ.Contam. Toxicol. 203, 1–86.

Keith, L.H., Crumett, W., Wentler, G., 1983. Principles of environmental analysis.Anal. Chem. 55, 2210–2218.

Laskowski, D., 2002. Physical and chemical properties of pyrethroids. Rev. Environ.Contam. Toxicol. 174, 49–170.

Lippold, H., Gottschalch, U., Kupsch, H., 2008. Joint influence of surfactants andhumic matter on PAH solubility. Are mixed micelles formed? Chemosphere 70,1979–1986.

Luo, L., Zhang, S., Shan, Y.-Q., Zhu, Y.-G., 2006. Oxalate and root exudates enhancethe desorption of p,p0-DDT from soils. Chemosphere 63, 1273–1279.

Metcalfe, T.L., Metcalfe, C.D., 1997. The trophodynamics of PCBs including mono andnon-ortho congeners in the food web of north-Central Lake Ontario. Sci. TotalEnviron. 201, 245–272.

Miglioranza, K.S.B., Aizpún, J.E., Moreno, V.J., 2003. Dynamics of organochlorinepesticides in soils from a SE region of Argentina. Environ. Toxicol. Chem. 22,712–717.

Miglioranza, K.S.B., Ondarza, P.M., Gonzalez, M., Shimabukuro, V.M., Isla, F.I.,Aizpún, J.E., Moreno, V.J., 2008. Persistent organic pollutants in soils, sedimentsand fish from the Negro River basin, Patagonia, Argentina. In: 18th Annual AEHSMeeting and West Coast Conference on Soils, Sediments and Water, San Diego,USA, Book of Abstracts.

Mo, C.H., Cai, Q.Y., Li, H.Q., Zeng, Q.Y., Tang, S.R., Zhao, Y.C., 2008. Potential ofdifferent species for use in removal of DDT from the contaminated soils.Chemosphere 73, 120–125.

Moorman, T.B., Cowan, J.K., Arthur, E.L., Coats, J.R., 2001. Organic amendments toenhance herbicide biodegradation in contaminated soils. Biol. Fertil. Soils 33,541–545.

Mukerjee, P., Mysels, K.J., 1971. Critical Micelle Concentration of AqueousSurfactant Systems, NSRDS-NBS 36. US Government Printing Office,Washington, DC.

Qiu, X., Zhu, T., Yao, B., Hu, J., Hu, S., 2005. Contribution of dicofol to the current DDTpollution in China. Environ. Sci. Technol. 39, 4385–4390.

Reemtsma, T., Bredow, A., Gehring, M., 1999. The nature and kinetics of organicmatter release from soil by salt solutions. Eur. J. Soil Sci. 50, 53–64.

Rodríguez-Cruz, S., Sánchez-Martín, M.J., Andrades, M.S., Sánchez-Camazano, M.,2007. Modification of clay barriers with a cationic surfactant to improve theretention of pesticides in soils. J. Hazard. Mater. 139, 363–372.

Rosen, M.J., 2004. Surfactants and Interfacial Phenomena, third ed. Wiley, NewJersey, USA. 444 p.

Sabljic, A., 2001. QSAR models for estimating properties of persistent organicpollutants required in evaluation of their environmental fate and risk.Chemosphere 43, 363–375.

Sabljic, A., Güsten, H., Verhaar, H., Hermens, J., 1995. QSAR modeling of soilsorption. Improvements and systematics of log Koc vs. log Kow correlations.Chemosphere 31, 4489–4514.

SAGPyA, 2005. Secretaria de agricultura Ganaderia, Pesca y Alimentos. RepublicaArgentina. Guía de Productos Fitosanitarios (para la República Argentina).<www.sagpya.mecon.gov.ar>.

Sánchez, L., Romero, E., Sánchez Rasero, F., Dios, G., Peña, A., 2003. Enhanced soilsorption of methidathion using sewage sludge and surfactants. Pest Manage.Sci. 59, 857–864.

Sánchez-Camazano, M., Rodríguez-Cruz, M.S., Sánchez-Martin, M.J., 2003.Evaluation of component characteristics of soil–surfactant–herbicide systemthat affect enhanced desorption of linuron and atrazine preadsorbed by soils.Environ. Sci. Technol. 37, 2758–2766.

Semple, K.T., Morris, A.W.J., Paton, G.I., 2003. Bioavailability of hydrophobic organiccontaminants in soils: fundamental concepts and techniques for analysis. Eur. J.Soil Sci. 54, 809–818.

Trapp, S., McFarlane, C., 1995. Plant Contamination: Modeling and Simulation ofOrganic Chemical Processes. Lewis Publishers, USA. 254 p.

UNEP-Chemicals, 2004. Stockholm Convention on Persistent Organic Pollutants.United Nation Environment Programme. <http://www.pops.int/>.

Walker, A., Rodriguez Cruz, M.S., Mitchel, M.J., 2005. Influence of ageing of residueson the availability of herbicides for leaching. Environ. Pollut. 133, 43–51.

Wang, P., Keller, A.A., 2008. Particle-size dependent sorption and desorption ofpesticides within a water–soil–nonionic surfactant system. Environ. Sci.Technol. 42, 3381–3387.

Wania, F., MacKay, D., 1996. Tracking the distribution of persistent organicpollutants. Environ. Sci. Technol. 30, 390A–396A.

White, J.C., Mattina, M.J.I., Lee, W.-Y., Eitzer, B.D., Iannucci-Berger, W., 2003. Role oforganic acids in enhancing the desorption and uptake of weathered p,p0-DDE byCucurbita pepo. Environ. Pollut. 124, 71–80.

Yang, Y., Ratté, D., Smets, B.F., Pignatello, J.J., Grasso, D., 2001. Mobilization of soilorganic matter by complexing agents and implications for polycyclic aromatichydrocarbon desorption. Chemosphere 43, 1013–1021.

Zhou, W., Zhu, L., 2007. Efficiency of surfactant-enhanced desorption forcontaminated soils depending on the component characteristics of soil-surfactant-PAHs system. Environ. Pollut. 147, 66–73.