**Fate of Pesticides in Soils: Toward an Integrated Approach of Influential Factors**

Véronique Chaplain1, Laure Mamy1, Laure Vieublé-Gonod2, Christian Mougin1, Pierre Benoit2, Enrique Barriuso2 and Sylvie Nélieu1 *1INRA, UR 251 PESSAC, Versailles, 2INRA-AgroParisTech, UMR 1091 EGC, Grignon, France* 

#### **1. Introduction**

534 Pesticides in the Modern World - Risks and Benefits

White, J. C., Mattina, M. I., Eitzer, B. D., et al. (2002), 'Tracking chlordane compositional and chiral profiles in soil and vegetation'. *Chemosphere,* Vol. 47, No. 6, pp. 639-646. White, J. C., Mattina, M. I., Lee, W.-Y., et al. (2003), 'Role of organic acids in enhancing the

Yang, H., Zheng, M. and Zhu, Y. (2008), 'Tracing the behaviour of hexachlorobenzene in a

Yang, Y., Ratte, D., Smets, B. F., et al. (2001), 'Mobilization of soil organic matter by

Yao, F., Yu, G., Bian, Y., et al. (2007), 'Bioavailability to grains of rice of aged and fresh ddd

Zhu, Y., Zhang, S., Huang, H., et al. (2009), 'Effects of maize root exudates and organic acids

desorption'. *Chemosphere,* Vol. 43, No. 8, pp. 1013-1021.

and dde in soils'. *Chemosphere,* Vol. 68, No. 1, pp. 78-84.

*Pollution,* Vol. 124, No. 1, pp. 71-80.

20, No. 1, pp. 56-61.

21, No. 7, pp. 920-926.

desorption and uptake of weathered p,p'-dde by cucurbita pepo'. *Environmental* 

paddy soil-rice system over a growth season'. *Journal of Environmental Sciences,* Vol.

complexing agents and implications for polycyclic aromatic hydrocarbon

on the desorption of phenanthrene from soils'. *Journal of Environmental Sciences,* Vol.

Despite constraining legislation and increasing efficiency of pesticides (with a decrease in the applied amounts), their use still cause a contamination of environment (air, soil and water). To conciliate agricultural and environmental interests, a better understanding of the fate of pesticides is needed, in particular because it will determine the exposure and consequently the impact of pesticides on the target and non-target organisms. This goal requires new efforts of research at different scales (from molecular to field scale). Following application, most of the pesticides reach the soil either after direct application or after foliage wash-off. As a major interface between other environmental compartments, the soil plays a preponderant buffering role in the fate of pesticides. Apart volatilization, the main processes that control the fate of pesticides in soils are retention on soil particles and degradation (biotic and abiotic). These coupled bio-physico-chemical processes can lead to a transitory or permanent accumulation of pesticides in soils or, on the contrary, to their elimination from the environment. They determine the pesticide concentration in the soil solution, and have a large influence on pesticide transfer toward ground or surface waters and on their ecotoxicological impacts on soil organisms as well. The main difficulties in studying and predicting the retention and degradation of pesticides in soils are the diversity of chemical structures and reactivities of pesticides, the high diversity of soils and their heterogeneous composition and structure. In addition, the pedoclimatic conditions, in particular soil temperature and water content, have a strong influence on retention and degradation because of their effect on soil biological, chemical and physical properties. Therefore, the objective of this chapter is to provide an overview of the factors involved in the retention and degradation of pesticides in soils and to discuss and clarify the needs of new integrated approach. In particular, this work will examine (i) the pertinent scales (among elementary constituents, aggregates and mesoscopic scales) for both retention and degradation studies, (ii) the integrative properties that should be considered, such as hydrophobicity of the organo-clay granulometric fraction or soil structure, and (iii) the primordial role of water.

#### **2. Retention**

#### **2.1 Definitions**

The retention of pesticides in soils is mainly due to the adsorption, which is the passage of a solute from an aqueous phase to the surface of a solid adsorbent (Calvet, 1989). The solid

Fate of Pesticides in Soils: Toward an Integrated Approach of Influential Factors 537

content. The soil solution is sampled with glass microfibre filters laid on the soil surface. The volume of the soil solution and the dissolved pesticide retained in the filter are determined.

*Soil columns.* The soil columns allow the study of retention in dynamic conditions. The column is filled with disturbed (sieved or small aggregates) or undisturbed soil, a solution of pesticide is applied at the top of the column then water flow is imposed through rainfall simulation or pressure head control (Pot et al., 2005; 2011). Sorption coefficients can be estimated by inverse modelling of the elution curves and retardation factor calculation (Lennartz, 1999). Column experiments allow the possibility to determine the vertical distribution of the pesticide residues in the soil core if recovery in the leachates is incomplete (Benoit et al., 2000; Vincent et al., 2007). In a recent study, Vereecken et al. (2011) analysed the relationships between Koc derived from column data with more classical batch experimental data obtained on the same soils. The authors concluded that such relationships depended on pore water velocity and on the saturation status (saturated *vs* unsaturated) and

*Lysimeters.* A lysimeter consists of an undisturbed soil block or cylinder, embedded in an inert container with a bottom permeable to drainage water or leachate. An outstanding feature of lysimeters is the capability to monitor mass fluxes of water and chemicals under field climatic conditions and representative crop practices (Saison et al., 2008). The distribution of the chemical and of its metabolites in the soil along with their transformation rates can also be determined. Compared to laboratory experiments, outdoor lysimeter studies are closer to field environmental conditions, there is no significant disturbance of the soil structure, but the major limitation is the fact that certain variable experimental conditions such as environmental/climatic parameters are not controlled (OECD, 2000b).

*Surface, volume, and branching.* In general, the adsorption of pesticides increases with the volume and with the degree of branching which is correlated to the surface area (Mamy & Barriuso, 2005; Sabljic et al., 1995). Indeed, the molecular volume is related to water solubility (see hydrophilic/hydrophobic balance) (Calvet, 1989), and the degree of

*Electronic structure.* The nature of atoms and of functional groups determines the electronic structure of the pesticides (therefore their permanent dipole moment and polarizability) that governs the type of interactions of pesticide with soils (donor-acceptor electron, hydrogen bonds) (Calvet, 1989). The different substitutions and their spatial arrangement in the molecule have an effect on the adsorption by influencing the reactivity of functional groups (carbonyl oxygen, amide nitrogen) participating in these bond interactions (Liu et al., 2000). *Ionization.* It determines the charge of the pesticide and depends on its electronic structure. Strong bases always occur in cationic form in soils, but the ionization of weak bases and weak acids depends on the pH of the soil and on pKa or pKb values of molecules (Calvet, 1989). In general, the sorption of cations is strong on negatively charged surfaces like clays, oxides, hydroxides and humic substances. On the contrary, anions are not adsorbed on these surfaces, but their sorption is high in soils with positive charges, like tropical soils. For example, glyphosate has four pKa so that its sorption increases when the soil pH decreases

This method could be adapted to undisturbed soil samples.

packing status (disturbed *vs* undisturbed) of the soil column.

**2.3 Factors controlling the retention of pesticides in soils** 

branching encodes the intermolecular accessibility (Kier & Hall, 2000).

because the number of negative charges of this herbicide decreases.

**2.3.1 Physico-chemical properties of pesticides** 

adsorbents are the different soil constituents. According to the properties of pesticides and adsorbents, several adsorption mechanisms are possible: hydrogen bindings, ion exchanges, interactions with metallic cations, polar interactions, charge transfers, London-Van der Waals dispersion forces and hydrophobic effects (Calvet et al., 2005). As the soil constituents contain polar and ionisable groups, the adsorption of pesticides possessing polar and non polar groups can involve several of these mechanisms. The reverse process of adsorption is desorption. In general, the desorption is inversely related to adsorption, being small when adsorption is great, and conversely (Mamy & Barriuso, 2007). For example, the adsorption of atrazine is fully reversible (Celis et al., 1998); on the contrary that of glyphosate is not reversible and hysteresis is observed (Mamy & Barriuso, 2007). The hysteresis can be due to irreversible adsorption, physical entrapment in organo-mineral aggregates or degradation.

#### **2.2 Methods of measurement**

*Batch experiments.* Most of the time, the retention of pesticides is measured with soil suspensions, known as batch experiments, according to the OECD 106 guideline (OECD, 2000a). A volume (generally 10 mL) of an aqueous solution of pesticide is added to a mass of dry sieved soil (generally 2 g) in glass centrifuge tubes. Soil suspensions are shaken mechanically for 24 hours in darkness and then centrifuged. The duration of 24 hours corresponds to the time needed to reach equilibrium between the adsorbed pesticide and the pesticide in solution. The degradation of the pesticide or the adsorption of the pesticide on the surface of the flask used for the experiment has to be determined (Calvet, 1989). The amounts of adsorbed pesticide in the soil are calculated as the difference between initial pesticide concentration in solution and centrifuged supernatant concentration. This experiment is done at several initial pesticide concentrations to determine the adsorption isotherm of the pesticide (adsorbed amounts as a function of the equilibrium concentration of pesticide). In general, the smaller the concentration, the greater the adsorbed amount per unit mass soil. From this isotherm, distribution coefficients between soil and soil solution can be determined according to the Freundlich (Kf) model (1):

$$\mathbf{K}\mathbf{f} = \mathbf{Q}\mathbf{s} \;/\ \; \mathbf{C}\mathbf{e}^{\text{nf}} \tag{1}$$

where Qs (mg kg-1) is amount of adsorbed herbicide in soil at equilibrium concentration, Ce (mg L-1) is pesticide concentration in supernatant solution, and nf is an empirical coefficient. When nf = 1, the isotherm is linear and Kf = Kd (L kg-1). As organic carbon is a major adsorbent for pesticides (see 2.3.2.1), the Koc (L kg-1) (2) coefficient is often calculated as:

$$\text{Koc} = (\text{Kd} \times 100) \text{ / } \text{Corg} \tag{2}$$

where Corg is the percentage of organic carbon content in soil.

For a given pesticide, the Koc is generally less variable than the Kd among different soils (Calvet, 1989). However, the intensive shaking of soil-pesticide solution leads to dispersion of the soil structure, resulting in a higher availability of sorption sites. Therefore batch overestimates the sorption of pesticides (Walker & Jurado-Exposito, 1998).

*Centrifugation.* The soil sample is prepared at realistic soil moisture, treated with the pesticide, incubated, then centrifuged to collect the soil solution which is directly analysed for pesticide concentration (Benoit et al., 2007; Walker & Jurado-Exposito, 1998).

*Filters.* Gaillardon & Dur (1995) developed an original method using remoulded soil samples that are placed in Petri dishes to give a 3-4 mm thick soil layer at controlled water

adsorbents are the different soil constituents. According to the properties of pesticides and adsorbents, several adsorption mechanisms are possible: hydrogen bindings, ion exchanges, interactions with metallic cations, polar interactions, charge transfers, London-Van der Waals dispersion forces and hydrophobic effects (Calvet et al., 2005). As the soil constituents contain polar and ionisable groups, the adsorption of pesticides possessing polar and non polar groups can involve several of these mechanisms. The reverse process of adsorption is desorption. In general, the desorption is inversely related to adsorption, being small when adsorption is great, and conversely (Mamy & Barriuso, 2007). For example, the adsorption of atrazine is fully reversible (Celis et al., 1998); on the contrary that of glyphosate is not reversible and hysteresis is observed (Mamy & Barriuso, 2007). The hysteresis can be due to irreversible adsorption, physical entrapment in organo-mineral aggregates or degradation.

*Batch experiments.* Most of the time, the retention of pesticides is measured with soil suspensions, known as batch experiments, according to the OECD 106 guideline (OECD, 2000a). A volume (generally 10 mL) of an aqueous solution of pesticide is added to a mass of dry sieved soil (generally 2 g) in glass centrifuge tubes. Soil suspensions are shaken mechanically for 24 hours in darkness and then centrifuged. The duration of 24 hours corresponds to the time needed to reach equilibrium between the adsorbed pesticide and the pesticide in solution. The degradation of the pesticide or the adsorption of the pesticide on the surface of the flask used for the experiment has to be determined (Calvet, 1989). The amounts of adsorbed pesticide in the soil are calculated as the difference between initial pesticide concentration in solution and centrifuged supernatant concentration. This experiment is done at several initial pesticide concentrations to determine the adsorption isotherm of the pesticide (adsorbed amounts as a function of the equilibrium concentration of pesticide). In general, the smaller the concentration, the greater the adsorbed amount per unit mass soil. From this isotherm, distribution coefficients between soil and soil solution

where Qs (mg kg-1) is amount of adsorbed herbicide in soil at equilibrium concentration, Ce (mg L-1) is pesticide concentration in supernatant solution, and nf is an empirical coefficient. When nf = 1, the isotherm is linear and Kf = Kd (L kg-1). As organic carbon is a major adsorbent for pesticides (see 2.3.2.1), the Koc (L kg-1) (2) coefficient is often calculated as:

For a given pesticide, the Koc is generally less variable than the Kd among different soils (Calvet, 1989). However, the intensive shaking of soil-pesticide solution leads to dispersion of the soil structure, resulting in a higher availability of sorption sites. Therefore batch

*Centrifugation.* The soil sample is prepared at realistic soil moisture, treated with the pesticide, incubated, then centrifuged to collect the soil solution which is directly analysed

*Filters.* Gaillardon & Dur (1995) developed an original method using remoulded soil samples that are placed in Petri dishes to give a 3-4 mm thick soil layer at controlled water

nf Kf Qs / Ce = (1)

Koc (Kd 100) / Cor = × g (2)

can be determined according to the Freundlich (Kf) model (1):

where Corg is the percentage of organic carbon content in soil.

overestimates the sorption of pesticides (Walker & Jurado-Exposito, 1998).

for pesticide concentration (Benoit et al., 2007; Walker & Jurado-Exposito, 1998).

**2.2 Methods of measurement** 

content. The soil solution is sampled with glass microfibre filters laid on the soil surface. The volume of the soil solution and the dissolved pesticide retained in the filter are determined. This method could be adapted to undisturbed soil samples.

*Soil columns.* The soil columns allow the study of retention in dynamic conditions. The column is filled with disturbed (sieved or small aggregates) or undisturbed soil, a solution of pesticide is applied at the top of the column then water flow is imposed through rainfall simulation or pressure head control (Pot et al., 2005; 2011). Sorption coefficients can be estimated by inverse modelling of the elution curves and retardation factor calculation (Lennartz, 1999). Column experiments allow the possibility to determine the vertical distribution of the pesticide residues in the soil core if recovery in the leachates is incomplete (Benoit et al., 2000; Vincent et al., 2007). In a recent study, Vereecken et al. (2011) analysed the relationships between Koc derived from column data with more classical batch experimental data obtained on the same soils. The authors concluded that such relationships depended on pore water velocity and on the saturation status (saturated *vs* unsaturated) and packing status (disturbed *vs* undisturbed) of the soil column.

*Lysimeters.* A lysimeter consists of an undisturbed soil block or cylinder, embedded in an inert container with a bottom permeable to drainage water or leachate. An outstanding feature of lysimeters is the capability to monitor mass fluxes of water and chemicals under field climatic conditions and representative crop practices (Saison et al., 2008). The distribution of the chemical and of its metabolites in the soil along with their transformation rates can also be determined. Compared to laboratory experiments, outdoor lysimeter studies are closer to field environmental conditions, there is no significant disturbance of the soil structure, but the major limitation is the fact that certain variable experimental conditions such as environmental/climatic parameters are not controlled (OECD, 2000b).

#### **2.3 Factors controlling the retention of pesticides in soils 2.3.1 Physico-chemical properties of pesticides**

*Surface, volume, and branching.* In general, the adsorption of pesticides increases with the volume and with the degree of branching which is correlated to the surface area (Mamy & Barriuso, 2005; Sabljic et al., 1995). Indeed, the molecular volume is related to water solubility (see hydrophilic/hydrophobic balance) (Calvet, 1989), and the degree of branching encodes the intermolecular accessibility (Kier & Hall, 2000).

*Electronic structure.* The nature of atoms and of functional groups determines the electronic structure of the pesticides (therefore their permanent dipole moment and polarizability) that governs the type of interactions of pesticide with soils (donor-acceptor electron, hydrogen bonds) (Calvet, 1989). The different substitutions and their spatial arrangement in the molecule have an effect on the adsorption by influencing the reactivity of functional groups (carbonyl oxygen, amide nitrogen) participating in these bond interactions (Liu et al., 2000).

*Ionization.* It determines the charge of the pesticide and depends on its electronic structure. Strong bases always occur in cationic form in soils, but the ionization of weak bases and weak acids depends on the pH of the soil and on pKa or pKb values of molecules (Calvet, 1989). In general, the sorption of cations is strong on negatively charged surfaces like clays, oxides, hydroxides and humic substances. On the contrary, anions are not adsorbed on these surfaces, but their sorption is high in soils with positive charges, like tropical soils. For example, glyphosate has four pKa so that its sorption increases when the soil pH decreases because the number of negative charges of this herbicide decreases.

Fate of Pesticides in Soils: Toward an Integrated Approach of Influential Factors 539

The soil structure is characterized by the bulk density and the pore geometry which depend on agricultural practices and on the climate. In addition, the bulk density depends on the size of the soil sample (from aggregate to macroscopic scale) due to the spatial variability of the soil structure (Alletto et al., 2010a). Pesticide movement through aggregated soils is mainly controlled by kinetic sorption and diffusion (Beulke et al., 2004). In static conditions, the rate of pesticide adsorption decreases when the density of soil aggregates increases (Chaplain et al., 2008). In dynamic conditions, retention depends on transport parameters such as pore water velocity and residence time (Pot et al., 2005; 2011). Compared to tilled soils, the no tilled or grassland soils are characterized by the presence of biopores (due to earthworm burrows, roots…) and high content of organic matter in the surface layers. The retention of pesticides is therefore generally higher in these soils (Benoit et al., 2000; Larsbo et al., 2009). However, the increase in retention can be counter-balanced by increased preferential transport because no tillage leads to enhanced macropore connectivity (Larsbo

*Water content.* The soil water content defines the specific exchange surface between solid and liquid phases. The adsorption of pesticides increases with water content as it facilitates pesticide diffusion to sorption sites. As water content increases, the organic matter also becomes more hydrophilic with greater sorption potential for hydrophilic pesticides (Roy et al., 2000). For hydrophobic pesticides like trifluralin, the adsorption decreases when the soil water content increases because the hydration of the surfaces of adsorbents decreases the accessibility to adsorption sites (Swann & Behrens, 1972). Low soil moisture content might also favour access to the hydrophobic regions of humus by generating more hydrophobic surfaces, thereby increasing the sorption of hydrophobic substances (Roy et al., 2000). *Temperature.* In general, the adsorption of pesticides decreases when the temperature increases (Ten Hulscher & Cornelissen, 1996). However, fast sorption should be differentiated from slow sorption: the fast sorption increases with decreasing temperature, but the slow sorption is more rapid at higher temperature. This could explain why for some compounds, overall sorption with short equilibration times is nearly independent of temperature. The slow sorption is generally due to diffusion of the pesticide through the organic matter and increasing temperature decreases the density therefore increases the diffusion (Ten Hulscher & Cornelissen, 1996). For some pesticides that exhibit decreasing solubility at higher temperatures, an increase in the sorption with temperatures can be

**2.3.3 Effect of environmental conditions (water content, temperature)** 

observed (Chiou et al., 1979, as cited in Ten Hulscher & Cornelissen, 1996).

decrease in the organic carbon content (Mamy & Barriuso, 2005).

*Spatial variability.* The retention of pesticides varies laterally and vertically. For example, at the scale of one watershed, the variation coefficients of the Koc of several pesticides in the soil surface can reach 30%. It seems mainly due to the variation of organic carbon content (Coquet & Barriuso, 2002). But, the variability of pesticides adsorption is high even at smaller scales (cm to m) (Mermoud et al., 2008, Vieublé-Gonod et al., 2009). The adsorption of pesticides also varies with soil depth: in general, the adsorption decreases because of a

**2.3.4 Spatio-temporal variability of retention** 

**2.3.2.2 Soil structure** 

et al., 2009).

*Hydrophilic/hydrophobic balance.* The hydrophilicity of pesticides is defined by their water solubility and the hydrophobicity by their octanol/water partition coefficient. In general, the adsorption of pesticides decreases when their water solubility increases because of their high affinity for the water phase, and conversely, the adsorption increases with the hydrophobicity of pesticides. However, it also depends on the hydrophilic/hydrophobic balance of the soil adsorbents. The adsorption of polar compounds does not always decreases with increasing water solubility (Calvet, 1989). For example, glyphosate, a polar herbicide, is highly soluble in water (12 g L-1) but is strongly sorbed to soils (Mamy & Barriuso, 2005). Indeed, the sorption of ionisable molecules and molecules with phosphogroups like glyphosate involve high-energy binding adsorption phenomena (ionic and coordination bindings, complex formation with metals in solution or at the solid–liquid interphase) that overbalance the effects of high solubility (Tao & Lu, 1999).

#### **2.3.2 Soil properties**

#### **2.3.2.1 Elementary properties (minerals, organic matter, pH)**

*Minerals.* The mineral adsorbents involved in the adsorption of pesticides are clays (as silicate minerals), oxides and hydroxides (Calvet, 1989). Their surfaces are mainly hydrophilic because of hydroxyl groups and exchangeable cations. The adsorption of pesticides on clay minerals is likely to occur on external surfaces of clay particles rather than in interlamellar space and increases with the specific surface of clays (Barriuso et al., 1994). Oxides and hydroxides are frequently associated to clays, they have a high surface activity and their charge depends on the soil pH (Calvet, 1989). For example, the adsorption of glyphosate increases as follows: kaolinite < illite < montmorillonite < nontronite (Mc Connell & Hossner, 1985). The adsorption of glyphosate on iron and aluminium oxides and hydroxides is high at intermediate pH and driven by ionic bindings between the positive surface sites of minerals and the negative acid groups of glyphosate (Morillo et al., 2000). However, sorption is much lower at very acid or very alkaline pH because oxides will bear the same charge as glyphosate.

*Organic matter.* Soil organic matter originates from crop residues, microbial biomass and organic amendments. It has very heterogeneous composition and contains both hydrophilic and hydrophobic groups (Calvet et al., 2005). Even if organic matter only represents few percents of the total dried material in soil, it is a major sorbent of pesticides in soil (Calvet, 1989). This is attributed to its high chemical reactivity towards both mineral surfaces and organic molecules, allowing various types of interaction with pesticides. The sorption capacities of organic matter are not only controlled by their chemical composition, but also by their size, due to a greater number of sorptive sites related to a greater surface area with decreasing particle-size (Benoit et al., 2008). In general, the adsorption of pesticides increases with organic matter, except for ionic molecules.

*Soil pH.* The soil pH plays an important role in particular for the adsorption of ionic pesticides like glyphosate or sulcotrione (Calvet, 1989; Mamy & Barriuso, 2005). Depending on the charge of the pesticide, the adsorption will increase (or decrease) with pH. For example, the retention of glyphosate increases when the soil pH decreases because the number of negative charges of the molecule decreases, allowing the adsorption on negatively charged adsorbents like clay or organic matter.

#### **2.3.2.2 Soil structure**

538 Pesticides in the Modern World - Risks and Benefits

*Hydrophilic/hydrophobic balance.* The hydrophilicity of pesticides is defined by their water solubility and the hydrophobicity by their octanol/water partition coefficient. In general, the adsorption of pesticides decreases when their water solubility increases because of their high affinity for the water phase, and conversely, the adsorption increases with the hydrophobicity of pesticides. However, it also depends on the hydrophilic/hydrophobic balance of the soil adsorbents. The adsorption of polar compounds does not always decreases with increasing water solubility (Calvet, 1989). For example, glyphosate, a polar herbicide, is highly soluble in water (12 g L-1) but is strongly sorbed to soils (Mamy & Barriuso, 2005). Indeed, the sorption of ionisable molecules and molecules with phosphogroups like glyphosate involve high-energy binding adsorption phenomena (ionic and coordination bindings, complex formation with metals in solution or at the solid–liquid

*Minerals.* The mineral adsorbents involved in the adsorption of pesticides are clays (as silicate minerals), oxides and hydroxides (Calvet, 1989). Their surfaces are mainly hydrophilic because of hydroxyl groups and exchangeable cations. The adsorption of pesticides on clay minerals is likely to occur on external surfaces of clay particles rather than in interlamellar space and increases with the specific surface of clays (Barriuso et al., 1994). Oxides and hydroxides are frequently associated to clays, they have a high surface activity and their charge depends on the soil pH (Calvet, 1989). For example, the adsorption of glyphosate increases as follows: kaolinite < illite < montmorillonite < nontronite (Mc Connell & Hossner, 1985). The adsorption of glyphosate on iron and aluminium oxides and hydroxides is high at intermediate pH and driven by ionic bindings between the positive surface sites of minerals and the negative acid groups of glyphosate (Morillo et al., 2000). However, sorption is much lower at very acid or very alkaline pH because oxides will bear

*Organic matter.* Soil organic matter originates from crop residues, microbial biomass and organic amendments. It has very heterogeneous composition and contains both hydrophilic and hydrophobic groups (Calvet et al., 2005). Even if organic matter only represents few percents of the total dried material in soil, it is a major sorbent of pesticides in soil (Calvet, 1989). This is attributed to its high chemical reactivity towards both mineral surfaces and organic molecules, allowing various types of interaction with pesticides. The sorption capacities of organic matter are not only controlled by their chemical composition, but also by their size, due to a greater number of sorptive sites related to a greater surface area with decreasing particle-size (Benoit et al., 2008). In general, the adsorption of pesticides increases

*Soil pH.* The soil pH plays an important role in particular for the adsorption of ionic pesticides like glyphosate or sulcotrione (Calvet, 1989; Mamy & Barriuso, 2005). Depending on the charge of the pesticide, the adsorption will increase (or decrease) with pH. For example, the retention of glyphosate increases when the soil pH decreases because the number of negative charges of the molecule decreases, allowing the adsorption on

interphase) that overbalance the effects of high solubility (Tao & Lu, 1999).

**2.3.2.1 Elementary properties (minerals, organic matter, pH)** 

**2.3.2 Soil properties** 

the same charge as glyphosate.

with organic matter, except for ionic molecules.

negatively charged adsorbents like clay or organic matter.

The soil structure is characterized by the bulk density and the pore geometry which depend on agricultural practices and on the climate. In addition, the bulk density depends on the size of the soil sample (from aggregate to macroscopic scale) due to the spatial variability of the soil structure (Alletto et al., 2010a). Pesticide movement through aggregated soils is mainly controlled by kinetic sorption and diffusion (Beulke et al., 2004). In static conditions, the rate of pesticide adsorption decreases when the density of soil aggregates increases (Chaplain et al., 2008). In dynamic conditions, retention depends on transport parameters such as pore water velocity and residence time (Pot et al., 2005; 2011). Compared to tilled soils, the no tilled or grassland soils are characterized by the presence of biopores (due to earthworm burrows, roots…) and high content of organic matter in the surface layers. The retention of pesticides is therefore generally higher in these soils (Benoit et al., 2000; Larsbo et al., 2009). However, the increase in retention can be counter-balanced by increased preferential transport because no tillage leads to enhanced macropore connectivity (Larsbo et al., 2009).

#### **2.3.3 Effect of environmental conditions (water content, temperature)**

*Water content.* The soil water content defines the specific exchange surface between solid and liquid phases. The adsorption of pesticides increases with water content as it facilitates pesticide diffusion to sorption sites. As water content increases, the organic matter also becomes more hydrophilic with greater sorption potential for hydrophilic pesticides (Roy et al., 2000). For hydrophobic pesticides like trifluralin, the adsorption decreases when the soil water content increases because the hydration of the surfaces of adsorbents decreases the accessibility to adsorption sites (Swann & Behrens, 1972). Low soil moisture content might also favour access to the hydrophobic regions of humus by generating more hydrophobic surfaces, thereby increasing the sorption of hydrophobic substances (Roy et al., 2000).

*Temperature.* In general, the adsorption of pesticides decreases when the temperature increases (Ten Hulscher & Cornelissen, 1996). However, fast sorption should be differentiated from slow sorption: the fast sorption increases with decreasing temperature, but the slow sorption is more rapid at higher temperature. This could explain why for some compounds, overall sorption with short equilibration times is nearly independent of temperature. The slow sorption is generally due to diffusion of the pesticide through the organic matter and increasing temperature decreases the density therefore increases the diffusion (Ten Hulscher & Cornelissen, 1996). For some pesticides that exhibit decreasing solubility at higher temperatures, an increase in the sorption with temperatures can be observed (Chiou et al., 1979, as cited in Ten Hulscher & Cornelissen, 1996).

#### **2.3.4 Spatio-temporal variability of retention**

*Spatial variability.* The retention of pesticides varies laterally and vertically. For example, at the scale of one watershed, the variation coefficients of the Koc of several pesticides in the soil surface can reach 30%. It seems mainly due to the variation of organic carbon content (Coquet & Barriuso, 2002). But, the variability of pesticides adsorption is high even at smaller scales (cm to m) (Mermoud et al., 2008, Vieublé-Gonod et al., 2009). The adsorption of pesticides also varies with soil depth: in general, the adsorption decreases because of a decrease in the organic carbon content (Mamy & Barriuso, 2005).

Fate of Pesticides in Soils: Toward an Integrated Approach of Influential Factors 541

Pesticide metabolism in the environment is also governed by cometabolism in which organisms grow at the expense of a cosubstrate to transform the pesticide without deriving any nutrient or energy for growth from the process. Cometabolism is a partial and fortuitous metabolism and enzymes involved in the initial reaction lack substrate specificity. Generally, it results only in minor modifications of the structure of the pesticide, but these modifications could greatly influence pollutant bioavailability and mobility in soil. Different organisms (mainly fungi) can transform a molecule by sequential attacks, or can use cometabolic products of one organism as a growth substrate. In addition, intermediate products with their own bio-physico-chemical properties can accumulate, thus causing some adverse effects on the environment. The metabolites are generally less toxic than the

Synthesis includes conjugation and oligomerization. Pesticides are transformed into compounds with chemical structures more complex than those of the parent compounds. During conjugation, a pesticide (or one of its transformation products) is linked to hydrophilic endogenous substrates, resulting in the formation of methylated, acetylated, or alkylated compounds, glycosides, or amino acid conjugates. These compounds can be excreted from the living cells, or stored. During oligomerization, a pesticide combines with itself, or with other xenobiotic residues (proteins, soil organic residues). Consequently, they give high-molecular weight compounds, which are stable and often incorporated into cellular components (cell walls…) or soil constituents (soil organic matter). This biochemical process not only affects the activity and the biodegradability of a compound in limiting its bioavailability, but also raises concern about the environmental impact of the bound

*Pure liquid cultures.* Pesticide metabolism can be studied with pure liquid cultures supplemented with fungal or bacterial inocula. These cultures are potent tools to precise the transformation pathways of pesticides and the relevant metabolites. For example, white-rot basidiomycetes have been extensively considered because of their high potential for xenobiotics transformation. These filamentous fungi degrade pesticides using two types of enzymatic systems: intracellular (cytochromes P450) and exocellular (lignin-degrading system mainly consisting in peroxidases and laccases). Each of these systems could also be

*Soil and inoculum.* Studies can be performed in soils favouring microbial development and activity. Some experiments involved bacteria as an inoculum to remove pesticide residues from contaminated soils in laboratory conditions (Duquenne et al., 1996) or *in situ* (Qureshi et al., 2009). In some cases, carriers are developed to ensure fungal growth in the spiked soil. Lignocellulosic materials, that provide also nutrients and easily available carbon for the organisms, have been often retained. Nevertheless, fungal growth or activity are rarely

*Soil incubation (disturbed or undisturbed soils).* The degradation of pesticides is often studied using soil incubation. Soil samples are treated with the pesticide and incubated in the dark under controlled laboratory conditions (at constant temperature and soil moisture). After appropriate time intervals, soil samples are extracted and analysed for the parent substance and for metabolites. Volatile products are also collected for analysis. Using 13C or 14C-

induced or inhibited by pesticides, thus able to modulate their metabolism.

assessed using biological descriptors in biotransformation experiments.

pesticides, but they can be more toxic in some cases (Tixier et al., 2002).

residues (Barriuso et al., 2008; Bollag & Liu, 1990).

**3.1.2 Methods of measurement** 

*Temporal variability.* The retention of pesticides is affected by their residence time in soil because of diffusion into soil micropores, physical entrapment or degradation (Koskinen et al., 2001). In the long term, the interactions responsible for retention evolve to the formation of pesticides non-extractable residues (Barriuso et al., 2008). Non-extractable (bound) residues are pesticides in soils which persist in the form of the parent substance or its metabolite(s) after extraction (Fürh et al., 1998, as cited in Barriuso et al., 2008). A large increase in retention with time is generally observed for weakly adsorbed herbicides, but for strongly adsorbed herbicides, adsorption decreases or remains stable (Mamy & Barriuso, 2007).
