Leaching of terbumeton and terbumeton-desethyl from mini-columns packed with soil aggregates in laboratory conditions

Chemosphere 65 (2006) 1600–1609 www.elsevier.com/locate/chemosphere

Leaching of terbumeton and terbumeton-desethyl from mini-columns packed with soil aggregates in laboratory conditions A. Conrad a

a,*

, O. Dedourge a, R. Cherrier b, M. Couderchet a, S. Biagianti

a

Laboratoire d’Eco-Toxicologie, Universite´ de Reims Champagne-Ardenne, Unite´ de recherche ‘‘Vignes et Vins de Champagne’’, UFR Sciences Exactes et Naturelles, URVVC UPRES-EA 2069, Moulin de la Housse, B.P. 1039, 51687 Reims Cedex 02, France b Chambre d’Agriculture de Lorraine, 9 rue de la Vologne, B.P. 1022, 54524 Laxou Cedex, France Received 11 January 2006; received in revised form 16 March 2006; accepted 17 March 2006 Available online 3 May 2006

Abstract Leaching of terbumeton (TER) and terbumeton-desethyl (TED) from mini-columns packed with natural soil aggregates was investigated. Five soil samples from the Champagne area (France) with different physicochemical parameters were used. The soil samples were hand-packed into a 50 mm column in laboratory conditions. An aqueous solution of TER or TED was percolated through the column and collected effluents were analyzed for TER or TED using HPLC-DAD. The leaching experiments showed that TER and TED were moderately mobile. TED was more mobile than TER, possibly because of its higher polarity. The proportion of organic matter affected the mobility of TER and TED through soil columns (r = 0.971) and leaching was lowest for soil having the highest organic matter content (5.9%). TER and TED were not significantly influenced by leaching solution composition (deionized water or CaCl2 solution), but were strongly affected by soil packing. Packing resulted in less rapid release of compounds suggesting that unpacking may have contributed to preferential pathways through the soil columns. Increasing contact time between TER and soils before leaching decreased the mobility of TER and increased its persistence in soils. Indeed, 76% of TER was released when leaching started after a 15 h contact time whereas it was down to 26% after an aging treatment of 360 h. A proportion of TER (from 8% to 32%) and TED (from 8% to 17%) remained in soil. Associated to its high stability in soils this could in part account for a very slow transfer over the years towards the groundwater.  2006 Elsevier Ltd. All rights reserved. Keywords: Herbicide; Leaching; Mobility; Organic matter; Pollution

1. Introduction In recent years, herbicides and their degradation products have been detected in an increasing number of aquifers all over the world. One of the problematic compounds is terbumeton, a methoxytriazine herbicide (2-amino-4tert-butylamino-ethyl-6-methoxy-1,3,5-triazine-2,4-diamine), which was primarily used in vineyards. Terbumeton acts by inhibiting photosynthesis in sensitive weeds. It is relatively persistent in soil with a DT50 up to 300 d and a Koc value

*

Corresponding author. Tel./fax: +33 391 263410. E-mail address: [email protected] (A. Conrad).

0045-6535/$ - see front matter  2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2006.03.046

comprised between 37 and 158 (Tomlin, 2000). Although it has only been banned in France since April 1998, its use in the vineyards of Champagne had almost stopped since the mid-eighties because of a shift in the weed species. Nevertheless, monitoring has revealed groundwater contamination by this herbicide and its major degradation product (terbumeton-desethyl; 2-amino-4-tert-butylamino-6-methoxy-1,3,5-triazine) (Bardet et al., 2002). Presently, terbumeton concentration has stabilized at ca. 0.07 lg l 1 in groundwater of the Reims area (Champagne, France) (Bardet et al., 2002), close to the international authorized limit of 0.1 lg l 1. So far, no study has been realized about this active compound in order to understand the retardation phenomenon from soil surface to groundwater.

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

To understand and to reduce the mobility of herbicides from topsoil to natural surface- or ground-water it is necessary to identify factors involved in their dissipation. One important source of groundwater pollution is herbicide leaching after application in agriculture. Assuming the mobility is mainly one-dimensional, this process can be simulated in soil column experiments. Laboratory leaching assays using column breakthrough experiments provide estimates of sorption and transport parameters under various flow conditions (Fetter, 1993) and are likely to represent pesticide transport under field conditions (Jackson et al., 1984; Malterre et al., 2000). In comparison, the laboratory studies using batch tests give information concerning the partitioning of organic compound between bulk water and soils in apparent equilibrium state conditions, but do not take into account the texture, and the density of soils implied in the organic compound transfer. Sorption of organic compounds to soils plays an important role as a retardation mechanism in solute mobility process. The sorption phenomenon of organic compounds from bulk water to soils depends on (i) the physicochemical characteristics of organic contaminants (Karickoff, 1981; Schwarzenbach et al., 1993; Beck and Jones, 1996; Luthy et al., 1997) such as their molecular weight, their acidity, their planarity, their polarity, their solubility and (ii) the soil physicochemical parameters (Tye et al., 1996; Ahmad et al., 2001; Novak et al., 2001; Spark and Swift, 2002) such as the aggregation, the density, the water proportion, the composition and proportion of organic and mineral components, and the pH and ionic strength of soil solution. Furthermore, several studies have shown that a higher organic matter proportion enhances the sorption of pesticides (Karickoff, 1981). The aim of this study was to investigate (i) terbumeton (TER) and terbumeton-desethyl (TED) leaching and (ii) the relative importance of soil components on the mobility of TER and TED through the 5–20 cm soil horizon under laboratory conditions. Thus, five soil samples differing in physicochemical characteristics such as organic matter and CaCO3 contents and representative of the area near Reims (Champagne, France) were collected. Different parameters such as lag-time before leaching and packing of soils were tested to observe their influence on the leaching of TER in mini-columns (50 · 80 mm) under simulated rain. 2. Material and methods 2.1. Sites For the study, the soil samples were selected for their variety of physicochemical properties (organic matter and mineral contents . . .) representative of various agrosystems from the area of Reims (Champagne, France). These soils were excavated from the top layer (5–20 cm) of five sites (A, B, C, D, and E). Soil samples were obtained by pooling together five soil sub-samples collected on the same site.

1601

Except for the site A (Bouy, France), located in a barley field, the other sites were all located in vineyards near Reims (Cernay-les-Reims, and Nogent l’Abbesse). For the two soils collected from Nogent l’Abbesse: one of them was covered with grass (Soil B) and the other one was bare (Soil C). The two soils collected in Cernay-les-Reims were from a single vineyard: one was taken at the top of the plot (Soil D) and the other at the bottom (Soil E), both of them were bare (without grass, mulch, or rind coverage as topsoil protection). 2.2. Soil characteristics After air-drying, soil samples were passed through a 2mm-mesh sieve, homogenized, and stored at 4 C prior to analysis. Particle size distribution, organic matter, total nitrogen, total alkalinity measured as CaCO3, cation exchange capacity (CEC), and potassium (K2O) were determined in agreement with French standard methods (NF X 31-107 without decarbonation, NF X 31-109, NF ISO 11261, NF ISO 10-693, NF X 31-130, and NF X 31-108, respectively). 2.3. Reagents Terbumeton (TER) and terbumeton-desethyl (TED) (99.5% and 96.1% of purity, respectively) were purchased from Dr. Ehrenstorfer (Ausgburg, Germany). Their chemical structures and physicochemical properties are given in Table 1. All chemicals and solvents were analytical or HPLC reagent grade and used without further purification; methanol and acetone were purchased from Fischer scientific (France). Ultrapure water was purified with a Millipore Milli-Q system. 2.4. Pesticide analysis A 20 ll water sample was injected into a modular chromatographic system, which was equipped with a ASI-100 automatic injector (Dionex, Voisins le Bretonneux, France), a model 480 Gynkotek pump (Dionex), a UVD 340U Gynkotek diode-array detector (Dionex). All data were recorded and reprocessed using a software (Chromeleon software, version 6.60, Dionex, France). A reversedphase column (Kromasil 100 RP-18, 250 · 3 mm I.D., 5 lm particle size, Cil Cluzeau, Sainte-Foy-la-Grande, France), which was connected to a prefilter (A-103X, 0.5 lm, Cil Cluzeau), was used. The mobile phase was prepared volumetrically from individually measured acetonitrile and ultrapure water (60:40, v/v). Mobile phase was degazed and filtrated, prior use through a 0.2 lm Nylon filter (7402-004, Millipore, Maidstone, England). Mobile phase flow rate was maintained at 0.8 ml min 1 at room temperature (ca. 22 C). Stock solutions of standard compounds were prepared in mobile phase at 10 mg l 1 and stored at 4 C. The dilu-

1602

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

Table 1 Chemical structures and some physicochemical properties of terbumeton and terbumeton-desethyl Terbumeton (TER)

Terbumeton-desethyl (TED)

CH3 H3 C O

N

N N

CH3

N

CH3 H3C O

CH3

N

H N CH2 CH3 Molecular weight Solubility (water, 20 C) Log Kowa pKab Ebullition point (C) a b

225.29 130 mg l 2.71 4.68 118.6

N

N N

CH3 CH3

NH2 197.24 Not available 1.93 4.39 60

1

Kaune et al. (1998). Boudesocque et al. (2005a).

tions were then realized in mobile phase just before injection in HPLC system for calibration. TER and TED were detected and quantified spectrophotometrically at 220 and 210 nm, respectively. The limit of quantification (signal/ noise = 10) were 5 and 2 lg l 1, respectively for TER and TED. The relative standard deviation of the measurement was below 1% and the detection response presented a good linearity (r > 0.9997; p < 0.01). When compared to authorized European concentration (0.1 lg l 1), limits of quantification were relatively high when no sample concentration was performed (5 and 2 lg l 1 for TER and TED, respectively). However, this limit was considered sufficient and no sample concentration was performed since TER or TED were below detection limit only in the first four fractions and it had no significant influence on the final results. TER and TED were recognized from their respective absorption spectrum and their respective retention time in soil water samples. Replacing sample solvent (mobile phase) by a 0.01 M CaCl2 solution had no influence on retention times and absorption spectra of the pesticides. 2.5. Extraction procedure of TER and TED from soil samples Soil samples were air-dried in the dark at room temperature and were ground to powder using a mortar. Ten grams of soil sample were homogenized with 25 ml of methanol for 120 min at 400 rpm using a mechanical shaker (Variomag, H+P Labortechnik AG, Germany). The mixture was separated by decantation for 30 min. The supernatant was replaced by 20 ml of methanol and homogenized at 400 rpm for 60 min. This last step was repeated twice. Then, all supernatants were filtered through a PTFE-membrane of 0.45 lm porosity (Cameo, Osmonics Inc., Great Britain) and were evaporated under nitrogen stream at 30 C. The dried extracts were dissolved in 25 ml of acetonitrile and stored at 4 C for a maximum of 10 h before HPLC analysis. The recovery, determined from soil treated with known quantities of TER or TED,

were 97 ± 5% (n = 3 soil samples) and 100 ± 3% (n = 3 soil samples), respectively for TER and TED. 2.6. Stability of TER in soils Five solutions of TER (0, 0.34, 0.68, 1.35, and 13.51 mg TER kg 1 dry soil) were added to 80% of water holding capacity to soils A, B and C. Each treatment was repeated four times. Samples were incubated in the dark at 25 C during 15 d since it is by far longer than the longest elution time (400 min) and it corresponded to the duration of pesticide contact in aging experiment (360 h). After that, TER was extracted as previously mentioned in 2.5 except that 5 g of soil was used and sample was analyzed for TER by HPLC. 2.7. Leaching procedure Soil columns were made by cutting off the bottom of PolyTetraFluoroEthylene bottles. Each column was hand-packed with soil to create homogeneous soil columns as followed: 50 g of acid washed sand (1 mm diameter) was placed in the bottom of the column to hold the soil particles into place. Then, the column was packed with 250 g of untreated soil to form a packed bed of approximately 50 mm in height and 80 mm in diameter. The column was pre-humidified with a 50 ml CaCl2 (0.01 M) solution and was left to equilibrate for 10 h in order to hydrate the soil samples. Then, 50 ml of CaCl2 solution containing the standard compounds were applied to the top of column at the soil surface so that the deposited amounts were 2278 ± 180 lg (n = 5) and 2254 ± 279 lg (n = 3), respectively for TER and TED. For the leaching experiments, a 0.01 M CaCl2 solution prepared from analytical grade CaCl2 was used. The leaching solution was supplied by a peristaltic pump. All solutions were allowed to warm up to room temperature before the experiments were started. Column leaching experiments were carried out at 23 ± 1 C. Except in

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

‘‘aging experiments’’, percolation began 15 h after the pesticide was deposited and it proceeded regularly up to a total volume of 1500 ml of CaCl2 solution (conductivity of 2.5 mS cm 1), simulating rainfall. The resulting leachate was collected as 50 ml fractions, filtered on a membrane of 0.45 lm of porosity (Cameo, Osmonics Inc., Great Britain), and pesticide concentration was measured by HPLC analysis. The TER and TED losses by adsorption on the membrane was less than 1% (n = 3). The pH values were measured in each fraction. At the end of leaching experiments, the totality of soil columns was dried at room temperature and subjected to the extraction procedure. Analysis of leachate prior to pesticide application showed a number of small peaks due to unknown organic compounds. None of which interfered with the determination of either TER or TED.

Table 2 Physicochemical characteristics of the soil samples (A, B, C, D, and E) Soils

3. Results and discussion

The five soils (A, B, C, D, and E) showed great differences in many of their physical and chemical properties (Table 2). Physically, the soil samples differed essentially in the proportion of silt and sand, which corresponded to size particles comprised between 2 and 600 lm. The silt proportion was greater in the samples A, B, and C (ca. 41%) than in the samples D and E (ca. 27%). On the contrary, the sand proportion was greater in samples D and E (ca. 41%) than in the samples A, B, and C (ca. 27%). The clay proportion (size particles below to 2 lm) was similar in all soil samples (ca. 30%). The greatest variability in soil particle size was observed for the fine silt (RSD of 36%) and fine sand fractions (RSD of 47%), these particles may therefore have an important effect on the mobility of TER. The total nitrogen varied from 0.12% (sample D) to 0.19% (samples A and B) (Table 2). The organic matter proportion varied from 2.8% (sample D) to 5.9% (sample

Bouy

Nogent l’Abbesse

Cernay-lesReims

A

B

C

D

E

28.8 27.1 10.8 14.5 18.8 340 57.6 12.9 5.9 0.23 2.53 7.99

32.0 17.2 11.0 18.8 21.0 287 30.4 14.1 2.8 0.12 2.69 7.85

31.6 15.2 11.3 25.4 16.5 255 24.0 15.6 3.5 0.16 2.70 7.75

Particle size distribution (% content of total soil) <2 lm 29.0 28.9 2–63 lm 36.6 30.6 63–212 lm 9.5 8.4 212–600 lm 6.8 11.1 >600 lm 18.1 21.0 264 342 K2O (mg l 1) 73.6 64.8 CaCO3 (%) CEC (meq 100 g 1) 9.2 10.3 Organic matter (%) 3.6 4.2 Total nitrogen (%) 0.19 0.19 Conductivity (mS cm 1)a 2.53 2.53 pHa 8.05 8.05 a

3.1. Soil characterization

1603

Measured in a 0.01 M CaCl2 solution; CEC: cation exchange capacity.

C). The sample C had a greater proportion of total nitrogen, and organic matter than the other soil samples. The low proportion of organic matter for soil B may be explained by the fact that grass had only been growing for one year. The total alkalinity measured as CaCO3 content varied from 24.0% to 73.6%. The cation exchange capacity varied from 9.2 to 15.6 meq 100 g 1. Physicochemical properties of soils D and E were slightly different even though they were collected in the same plot. The soil E collected at the bottom of the vineyard had higher nitrogen, organic matter, and CEC than soil D collected at the top of vineyard. The soil pH ranged from 7.75 to 8.05 with an average value of 7.9 ± 0.1 (n = 5 soil samples). 3.2. Leaching of TER and TED The results of the column leaching experiment packed with the five soil samples for TER and TED are summarized in Table 3. After a total percolation of CaCl2 solution, the recoveries of TER in leachate were 90%, 80%,

Table 3 Leachate recoveries of terbumeton (TER) and terbumeton-desethyl (TED) in soil column experiments Soils

IA (lg)

UA (lg)

RA (lg)

CLA (lg)

Rleachate (%)

Qextract (lg g 1)

Rextract (%)

Recovery (%)

A

TER TED

2211 ± 35 2006 ± 7


2211 2006

1986 ± 8 1850 ± 2

90 92

0.279 ± 0.002 0.391 ± 0.014

10.2 7.8

93 97

B

TER TED

2341 ± 24 1934 ± 5


2341 1934

1882 ± 3 1752 ± 2

80 91

0.513 ± 0.002 0.251 ± 0.008

19.6 9.4

86 94

C

TER TED

2556 ± 66 2112 ± 14


2556 2112

1738 ± 6 1759 ± 3

68 83

0.775 ± 0.003 0.618 ± 0.021

32.0 16.7

76 91

D

TER TED

2199 ± 8 –


2199 –

1907 ± 2 –

87 –

0.435 ± 0.01 –

13.3 –

92 –

E

TER TED

2084 ± 13 –

3.1 ± 0.1 –

2081 –

1908 ± 2 –

92 –

0.292 ± 0.001 –

8.3 –

95 –

IA: initial amount; UA: unretained amount by soil column before leaching; RA: retained amount; CLA: cumulative leachate amount; Rleachate: CLA/ RA · 100; Qextract: quantity of TER or TED extracted from soils columns; Rextract: (Qextract · 250)/RA; Recovery: (CLA + Qextract · 250)/RA · 100; LOD: limit of detection; TED: terbumeton-desethyl; TER: terbumeton; –: not realized.

1604

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

Released amount of Terbumeton in eluate (μg / eluate fraction)

68%, 87%, and 92%, respectively for the A, B, C, D, and E soil columns. For TED, the column leaching experiments were realized only with the soils A, B, and C. Recoveries of TED in leachate were 92%, 91%, and 83%, respectively for A, B, and C soil columns (Table 3). Therefore, TER was moderately mobile in all type of soils used under prolonged leaching conditions, with mobility increasing in the

order soil C < soil B < soil D 6 soil A 6 soil E. TED was mobile in all the soils, with mobility increasing in the same order than TER: soil C < soil B 6 soil A. Both TER and TED appeared rapidly in leachate water, respectively 22–54 min (time 0, Fig. 1A), and 4–12 min after the first drop of water. For the five soil columns, typical chromatographic peaks were observed (Fig. 1A). The

400

300

200

100

0 0

100

A

200

300

400

300

400

Elution time (min)

Cumulative Terbumeton in eluate (% of apllied)

100

75

50

25

0 0 B

100

200

Elution time (min)

Fig. 1. Elution of TER through columns of soil A (e), soil B (h), soil C (d), soil D (m), and soil E (n). Elution profile (A) and cumulative leachate amount (B).

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

3.3. Effect of physicochemical parameters of soils on leaching of TER and TED The mobility of both TER and TED in soils was related to their adsorption on the soils. The leaching potential of TER and TED was less important in soil C. This highest affinity of soil C for TER was correlated to differences in soil properties, such as organic matter content, total nitrogen, and K2O (Fig. 2). Such correlation could not be

K 2O (mg l-1)

400

350

300 r = 0.8470 P<0.05

250

200

Organic matter (%)

6

5

4 r = 0.9711 P<0.01

3

2

0.24 Total nitrogen (%)

elution profiles differed in height and width at half-height of peak. In the case of an unretained compound, the peak is high and narrow as for soils A and D. On the contrary, the peak is less high and its width at half-height is greater when the herbicide is partially retained in the soil column as for soil C. Moreover, TER did not appear immediately in leachate. In order to directly compare the different assays, the cumulative released amount was normalized according to the initial amount of retained TER from the experimental data used in Fig. 1A (Fig. 1B). For the different soil column experiments, the curves were sigmoidal (Fig. 1B). The time necessary to reach 50% of released TER ranged from 90 min (soil E) to 266 min (soil C). In all cases, the profile showed the effect of dispersion, regarding an ideal case with no dispersion and perfectly conservative behavior where concentration goes from 0% to 100% immediately (Mills and Saiers, 1993). Three phases could be observed. The first phase during which no TER was released, corresponded to the time necessary for TER to go through the soil column. Then, TER was released in an initial peak, which represented the release of readily available compounds. Finally, a third phase of steady-state release took place, that was mainly governed by diffusion/sorption phenomena and the physicochemical characteristics of soil and herbicide. For TER, the slope observed on the breakcurve was more pronounced with the soils D (1.053) and A (0.875) than the soils E (0.441), B (0.333), and C (0.292). Similar results were observed for TED for soils A, B, and C. For TED, the slope observed at the inflexion point was more pronounced with the soil A (1.180) than the soils B (1.059) and C (0.794). Thus, the soil C had the highest affinity for TER and TED, whereas soils B and E retained TER in smaller proportion, and soils A and D retained the lowest amount of TER. The slight difference observed between Rleachate and the slope value was due to the volume of leachate collected was different. TED was released in greater proportion than TER (91% and 80%, respectively) (Table 3). Likewise, the slope was more important for TED (1.059) than TER (0.333). TED migrated more rapidly through soil B columns than the TER. Similar results were obtained with soils A and C (data not shown). Physicochemical properties of the organic compounds may be responsible for this different behavior. Indeed, TED has a higher polarity and a slightly lower molecular weight than TER (Table 1) and possibly had less hydrophobic interaction with organic matter.

1605

0.20

0.16

r = 0.8394 P<0.05

0.12 60

65

70 75 80 85 90 95 Cumulative released amount / initial amount (%)

100

Fig. 2. Linear relationship between the cumulative amounts of TER released (% of applied) and K2O (mg l 1), organic matter (%), and total nitrogen (%) for soils A, B, C, D, and E.

observed with the other physicochemical soil parameters such as CaCO3 content, CEC, pH, and particle size, even if particles of 2–63 and 212–600 lm presented an important variability. It is generally admitted that a high proportion of organic matter in soils enhances the retention of hydrophobic organic compounds. Although the dissolved organic carbon fraction appeared to have no effect on atrazine transport (Spark and Swift, 2002), the nature of the organic matter such as humified organic matter content is certainly an important factor controlling the sorption of triazines (Dousset et al., 1994). Aromaticity or polarity in the natural organic matter composition may be of importance (Oliver et al., 2003a,b). The organic matter could be at the origin of the apolar domain formation in soils, which are involved in the sorption of hydrophobic organic compound (Ganaye et al., 1997). Other physicochemical parameters have been described to affect the mobility of organic molecules through soil col-

1606

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

umns, these are the composition of the leaching solution, the pH of bulk liquid, and temperature. However, in the present study the leaching solution composition could be ruled out since no significant difference in leaching amount of TER was observed between the TER leaching through the soil A by deionized water or by CaCl2 solution (data not shown) in accordance with Spark and Swift (2002). However, ionic strength of bulk volume may affect the pesticide adsorption to soil in equilibrium condition (Uren˜aAmate et al., 2005). Moreover, the use of deionized water might increase the release of inorganic and organic matter and thus the swelling of clays or/and the soil aggregates dispersion (Reemtsma et al., 1999), which could enhance the leaching of TER or TED. On the contrary, the dissolved organic matter content was not influenced by the chemistry of the leaching solution. Leaching with CaCl2 led to lower hydrodynamic dispersion, because of shrinking clay minerals (Laegdsmand et al., 2005). Therefore, all experimentations were realized with CaCl2 solution (0.01 M) as leaching solution to limit this phenomena. In the case of hydrophobic compounds containing ionizable functional groups as TER and TED, aqueous phase pH may affect the sorption equilibrium (Schwarzenbach et al., 1993). Thus, such solutes may exist both in neutral and ionic forms depending on the solution pH and pKa value. In our assays, pH values were not adjusted and varied from 6.88 to 8.08 in the column experiments. Thus, TER and TED were predominantly in their neutral form at this pH range (Table 1). Moreover, no influence of solution pH could be observed on TER adsorption onto the soils used in this study in batch equilibrium tests (Boudesocque et al., 2005a).

3.4. Effect of soil packing on leaching of TER The effect of density/porosity of soils on the leaching of TER was also examined by using columns containing the soil A packed with a CaCl2 solution and another column containing the non-packed soil A (Fig. 3). The microstructure of the soils seemed to play a significant role in the leaching of TER through soil column. Indeed, the leached amount of TER was reduced by the packing of soil. TER appeared in the first fraction for the non-packed soil, while it appeared after 69 min in case of the packed soil column. Moreover, a greater amount of TER was retained by the packed soil (43%) as compared with the non-packed soil (21%) at 300 min. Similar results were obtained with the soil B. The unpacking of soil affected TER mobility through the soil column probably because of the formation of artificial fissures or macropores into the soil column, which created a preferential pathway for the hydric flow and thus increased the transport of terbumeton. As previously mentioned by Novak et al. (2001), a preferential flow into the macropore enhanced the leaching of metolachlore through soil. Also, some studies using tracers such as chloride or bromide added to column influent solution showed a preferential flow through the soil in response to packing (Fetter, 1993). 3.5. Effect of aging phenomenon on TER leaching from soil column In this study, aging phenomenon is defined as the contact time between soil and TER from application before

Cumulative Terbumeton in leachate (% of applied)

100

80

60

40

20

0 0

50

100

150

200

250

300

350

400

Elution time (min) Fig. 3. Cumulative amounts of TER released from soil A in packed column (e) and in unpacked column () and effect of aging on the cumulative leaching of TER from soil C (d: 15-h aging; s: 360-h aging).

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

3.6. Environmental impact After the complete percolation of the CaCl2 solution, amounts of TER and TED were still retained within the soil column. The proportions of TER remaining in soil columns were 10.2%, 19.6%, 32.0%, 13.3%, 8.3%, respectively for soils A, B, C, D, and E (Table 3). The proportions of TED remaining in soil columns were 7.8%, 9.4%, 16.7%, respectively for soils A, B, and C (Table 3). The total recovery of

140

Percentages of Terbumeton remaining (% of applied)

the leaching procedure (Beulke et al., 2004). Aging retarded the efflux of herbicide from the soil. Indeed, more TER remained in the soil column and less herbicide was recovered in the leachate of the 360-h-aged soil (26%) than in the control soil corresponding to a soil freshly treated (67%) (Fig. 3). Thus, increasing contact time between TER and soil before leaching modified the mobility of TER. As already mentioned (Pignatello and Xing, 1996; Capri et al., 2001), the aging phenomenon enhanced the retention of some pesticide in soils. It can therefore be proposed that TER was first sorbed onto the adsorption sites that are easily accessible at the surface of the soil aggregates followed by its sorption onto sites within the soil matrix, which are diffusion limited and are reached only later (Lesan and Bhandari, 2001). Diffusion kinetics in soil organic matter can be widely variable depending on polymer structure. Slow absorption may be observed into condensed organic polymeric matter or into microporous minerals such as zeolites with porous surfaces at water saturation, as compared to fast absorption to amorphous natural organic matter, onto water-wet organic surfaces as soot or to exposed water-wet mineral surfaces as quartz (Luthy et al., 1997). This effect of aging is supported by the fact that Abiven et al. (2006) observed that 5 d were necessary in order to reach an apparent equilibrium state between bulk water and soil in batch equilibrium assays. Later, the sorption/desorption behavior of hydrophobic compounds in soils was modified and the initially weak bonds were consolidated, inducing change in the mechanism of sorption/binding, and in the trapping of the pesticide by a steric effect into organic matter. Luthy et al. (1997) suggested that the trapping of organic compound within matrix from which organic compounds cannot readily escape can contribute to the aging phenomenon. Moreover, the dissolved organic matter may also affect the aging phenomenon (Fine et al., 2002) and consequently affect the mobility of pesticide-organic matter associations. To conclude, the prolonged contact time between TER and soil may have led to unexpected persistence of TER in the environment as already suggested by Sharer et al. (2003) for other organic compounds. Thus, the prediction of transport requires the data concerning the characterization of sorption/desorption of herbicide residues in aged soils, particularly in the case of prediction of herbicide transport in soil which contribute to over-estimations if freshly treated soil were used to predict transport (Koskinen et al., 2002).

1607

120 100 80 60 40 20 0 0 0.34 0.68 1.35 Initial amount of TER (mg kg-1 of dried soil)

13.51

Fig. 4. Percentages (% of applied) of terbumeton remaining in soils A (open bars), B (solid black bars), and C (solid grey bars) after the soils have been incubated with different concentrations of terbumeton during 15 d at 20 C (n = 4).

the herbicide (sum of herbicide in the leachate and herbicide remaining in soils) was different from 100%, it ranged between 76% and 95% for TER and 91% and 97% for TED. The possibility that this difference may be due to (bio)degradation was ruled out. Results have shown that biodegradation was not observed during the duration of experiments. Indeed, after 15 days of incubation, the results did not show any degradation of TER. In average, for the four solutions of added TER, the percentages of TER remaining after the soil extraction are 100%, 95%, and 96% for the soils A, B, and C respectively, which was not statistically different from 100% (Fig. 4). Either the biodegradation of TER is very slow or it is not easily accessible to biodegradation. That is in agreement with its high persistence in soil with a DT50 of up to 300 d (Tomlin, 2000). Although the columns used for the experiments were not sterile, the major metabolites (terbumeton-desethyl, terbumeton-2-hydroxy, and terbumeton-deisopropyl) were never observed in the leachates. As previously demonstrated (Boudesocque et al., 2005b) in equilibrium test, a hysteresis was measured, showing that a low proportion of sorbed TER (below to 5%) was released in bulk water of soils. Therefore, it was considered that sorbed TER would be hardly desorbed from soils after aging treatment, and then the leaching potential of TER in soil decreased. These small amounts of TER and TED accumulated in soils associated to their low degradability may be desorbed slowly and contaminate the groundwater during many years. 4. Conclusion The use of soil column is an easy method to understand the movement of herbicide by following the amount of herbicide released in the leachate volume. TER was moderately mobile in soils and TED was more mobile in soils than TER. TER and TED were released mainly in two

1608

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609

phases, with an initial leaching of readily available compounds and a second phase of steady-state release that is governed by diffusion. The leaching of TER and TED through soil column depended first on the soil characteristics, particularly the proportion of organic matter, which affected the retention of TER and TED in soils. Leaching of TER could be correlated to total nitrogen and K2O proportions of soils. Secondly, the density/porosity of soil affected the mobility of TER and TED due to the formation of artificial pathway through soil column. We observed also that the increase of contact time between TER and soils decreased the leaching of TER. Third, increasing contact time between TER and soil decreased the leaching of TER. Probably because the herbicide had more time to reach less accessible binding sites within the soil aggregates. Even under a prolonged and continuous leaching procedure, low TER and TED quantities were retained in soils. This low amount retained by soils associated to the high stability of TER in soil and the effect of contact time between TER and soils could be an explanation to the groundwater contamination. Acknowledgements This study was conducted in the framework of the AQUAL project (Ze´rophyto program) of Europol’Agro. Arnaud Conrad was recipient of a post-doctoral fellowship from the ‘‘Fondation du site Paris-Reims’’. We are grateful C. Piquet and M-R. Lejay for their contributions. References Abiven, D., Boudesocque, S., Guillon, E., Couderchet, M., Dumonceau, J., Aplincourt, M., 2006. Sorption of the herbicide terbumeton and its metabolites onto soils: influence of copper (II). Environ. Chem. 3, 53– 60. Ahmad, R., Kookana, R.S., Alston, A., Skjemstad, J.O., 2001. The nature of soil organic matter affects sorption of pesticides 1. Relationships with carbon chemistry as determined by 13C CPMAS NMR spectroscopy. Environ. Sci. Technol. 35, 878–884. Bardet, C., Rouxel-David, E., Auguste, G., Cordonnier, G., Dachy, S., 2002. Synthe`se des e´tudes mene´es sur le bassin versant du champ captant de Couraux (Reims, Marne), rapport BRGM FREDONCA 204, http://europolagro.univ-reims.fr/garde_couraux.pdf (accessed 01:03:06). Beck, A.J., Jones, K.C., 1996. The effects of particle size, organic matter content, crop residues and dissolved organic matter on the sorption kinetics of atrazine and isoproturon by clay soil. Chemosphere 32 (12), 2345–2358. Beulke, S., Brown, C.D., Fryer, C.F., van Beinum, W., 2004. Influence of kinetic dorption and diffusion on pesticide movement through aggregated soils. Chemosphere 57, 481–490. Boudesocque, S., Aplincourt, M., Dumonceau, J., Guillon, E., Abiven, D., Biagianti, S., Couderchet, M., Dedourge, O., 2005a. Re´tention du terbume´ton et de ses principaux me´tabolites par des e´chantillons de sols et un re´sidu lignocellulosique. In: Colloque of PIREN-SEINE, Session 5: Dynamique des phytosanitaires, February 3 and 4, Paris, France, http://www.sisyphe.jussieu.fr/internet/piren/v2/modules/communications (accessed 01:03:06). Boudesocque, S., Guillon, E., Dumonceau, J., Aplincourt, M., Abiven, D., Biagianti-Risbourg, S., Conrad, A., Couderchet, M., Dedourge,

O., 2005b. Re´tention du terbume´ton et de trois me´tabolites par diffe´rents e´chantillons de sols. Sources et transferts de phytosanitaires, rapport d’activite´ 2004, http://www.sisyphe.jussieu.fr/internet/piren/ rapports/2004/2_hydrologie_agriculture/Rapport_d_activites_Zerophyto_01-2005_PIREdoc.pdf (accessed 01:03:06). Capri, E., Camisa, M.G., Flores-Ce´spedes, F., Glass, C.R., GonzalesPrada, E., Trevisan, M., 2001. Imidacloprid and pyrimethanil soil sorption. Agronomie 21, 57–64. Dousset, S., Mouvet, C., Schiavon, M., 1994. Sorption of terbuthylazine and atrazine in relation to the physico-chemical properties of three soils. Chemosphere 28, 467–476. Fetter, C.W., 1993. Transformation, retardation, and attenuation of solutes. In: Fetter, C.W. (Ed.), Contaminant Hydrogeology. Macmillan Publishing Company, New York, pp. 115–162. Fine, P., Hass, A., Prost, R., Atzmon, N., 2002. Organic carbon leaching from effluent irrigated lysimeters as affected by residence time. Soil Sci. Soc. Am. J. 66, 1531–1539. Ganaye, V.A., Keiding, K., Vogel, T.M., Viriot, M.-L., Block, J.-C., 1997. Evaluation of soil organic matter polarity by pyrene fluorescence spectrum variations. Environ. Sci. Technol. 31 (10), 2701–2706. Jackson, D.R., Garrett, B.C., Bishop, T.A., 1984. Comparison of batch and column methods for assessing leachability of hazardous waste. Environ. Sci. Technol. 18 (9), 668–673. Karickoff, S.W., 1981. Semi-empirical estimation of sorption of hydrophobic pollutants on natural sediments and soils. Chemosphere 10, 833–846. Kaune, A., Bru¨ggemann, R., Kettrup, A., 1998. High-performance liquid chromatographic measurement of the 1-octanol-water partition of striazine herbicides and some of their degradation products. J. Chromatogr. A. 805, 119–126. Koskinen, W.C., Rice, P.J., Anhalt, J.A., Sakaliene, O., Moorman, T.B., Arthur, E.L., 2002. Sorption-desorption of ‘‘aged’’ sulfonylaminocarbonyltriazolinone herbicides in soil. J. Agric. Food Chem. 50 (19), 5368–5372. Laegdsmand, M., De Jonge, L.W., Moldrup, P., 2005. Leaching of colloids and dissolved organic matter from columns packed with natural soil aggregates. Soil Sci. 170, 13–27. Lesan, H.M., Bhandari, A., 2001. Evaluation of atrazine binding to surface soils. In: Proceedings of the 2000 Conference on Hazardous Waste Research, Denver, CO. May 23–25, pp. 76–89. Luthy, R.G., Aiken, G.R., Brusseau, M.L., Cunningham, S.D., Gschwend, P.M., Pignatello, J.J., Reinhard, M., Traina, S.J., Weber, W.J., Westall, J.C., 1997. Sequestration of hydrophobic organic contaminants by geosorbents. Environ. Sci. Technol. 31 (12), 3341–3347. Malterre, F., Perrin-Garnier, C., Schiavon, M., 2000. Small size lysimeters: a relevant tool for pesticide transfer assessment. In: Cornejo, J., Jamet, P. (Eds.), Pesticide/Soil Interactions: Some Current Research Methods. INRA Editions, Paris, pp. 267–272. Mills, A.L., Saiers, J.E., 1993. Particles-associated transport of pollutants in subsurface environments. In: Rao, S.S. (Ed.), Particulate Matter and Aquatic Contaminants. Lewis Publisher, Canada, pp. 105–126. Novak, S.M., Portal, J.-M., Schiavon, M., 2001. Effects of soil type upon metolachlor losses in subsurface drainage. Chemosphere 42, 235– 244. Oliver, D.P., Kookana, R.S., Salama, R.B., 2003a. Land use effects on sorption of pesticides and their metabolites in sandy soils. I. Fenamiphos and two metabolites, fenamiphos sulfoxide and fenamiphos sulfone, and fenarimol and azinphos methyl. Aust. J. Soil Res. 41, 847– 860. Oliver, D.P., Kookana, R.S., Salama, R.B., Corell, R., 2003b. Land use effects on sorption of pesticides and their metabolites in sandy soils. II. Atrazine and two metabolites, deethylatrazine and deisopropylatrazine, and prometryne. Aust. J. Soil Res. 41, 861–874. Pignatello, J.J., Xing, B., 1996. Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30, 1–11. Reemtsma, T., Bredow, A., Gehring, M., 1999. The nature and kinetics of organic matter release from soil salt solutions. Eur. J. Soil Sci. 50, 53– 64.

A. Conrad et al. / Chemosphere 65 (2006) 1600–1609 Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M., 1993. Sorption: solid-aqueous solution exchange. In: Schwarzenbach, R.P., Gschwend, P.M., Imboden, D.M. (Eds.), Environmental Organic Chemistry. John Wiley, New York, pp. 255–341. Sharer, M., Park, J.H., Voice, T.C., Boyd, S.A., 2003. Aging effects on the sorption–desorption characteristics of anthropogenic organic compounds in soil. J. Environ. Qual. 32, 1385–1392. Spark, K.M., Swift, R.S., 2002. Effect of soil composition and dissolved organic matter on pesticide sorption. Sci. Total. Environ. 298, 147–161.

1609

Tomlin, C.D.S., 2000. The Pesticide Manual, 20th ed., British Crop Protection Council, Farnham, United Kingdom. Tye, R., Jepsen, R., Lick, W., 1996. Effects of colloids, flocculation, particle size, and organic matter on the adsorption of hexachlorobenzene to sediments. Environ. Toxicol. Chem. 15, 643–651. Uren˜a-Amate, M.D., Socı´as-Viciana, M., Gonza´lez-Pradas, E., Saifi, M., 2005. Effects of ionic strength and temperature on adsorption of atrazine by a heat treated kerolite. Chemosphere 59, 69–74.