**2. Fate of SAs in the terrestrial environment**

#### **2.1. Basic concept of sorption modelling studies**

Sorption is defined as a phenomenon during which chemicals become associated with solid phases. Immensely important, this process affects the fate of chemicals in the environ‐ ment [24-25]. Adsorption/desorption experiments are useful for generating essential information on the mobility of various contaminants and their distribution in the soil, water and air. Assessing the distribution of a chemical between the soil and aqueous phases is not a straightforward process, however. It depends on various factors, such as the chemi‐ cal nature of the substance and the characteristics of the soil (*e.g.* pH, organic matter (OM) content, clay fraction content). Furthermore, climatic factors such as rainfall, temperature, sunlight intensity and wind, which can affect the strength of sorption, also have to be taken into consideration. Thus, the numerous phenomena and mechanisms involved in the adsorption of a chemical by soil cannot be completely defined by simple laboratory models. Nevertheless, such investigations can provide valuable information on the environmental relevance of the adsorption of chemicals [26].

Laboratory sorption experiments can be carried out under static (batch) and dynamic (column) conditions. Static tests are commonly used when the aim of the study is to calculate the distribution coefficient (Eq. 1), in equilibrium time, which is specific to every chemi‐ cal. Column tests, on the other hand, enable time-dependent monitoring of contaminant leaching from soil and waste materials; in addition, the flow-through pattern of such tests resembles actual environmental conditions. Therefore, the release of a contaminant depending on local equilibrium time and advection conditions can be evaluated based on column testing systems [27].

Both column and batch tests can be used to assess the possible leaching/release potential of contaminated materials on the soil – ground water pathway [28-30]. The release potential of water-soluble contaminants can be assessed as an expression of the source term, which gives an indication of their bioavailability. In this case a batch test provides a snapshot of a particular liquid-to-solid ratio. This type of test focuses on constituent solubility and release over a range of conditions by varying a single parameter (e.g. final extract pH, liquid-to-solid (LS) ratio) [27].

The common approach for modelling sorption results involves using only the initial linear part of the isotherm, plotted as cw and cs:

$$
\mathcal{L}\_s = \mathcal{K}\_d \cdot \mathcal{c}\_w \tag{1}
$$

where: *Kd* is the partition coefficient; *cs* is the content of test substance adsorbed on the soil at adsorption equilibrium (mg kg-1); *cw* is the mass concentration of test substance in the aqueous phase at adsorption equilibrium (mg L-1).

The sorption isotherm can be also mathematically described by the Freundlich or Langmuir model.

The Langmuir equation is written as:

need to be addressed, for example: (i) What is the fate of these compounds in the terrestrial environment? (ii) What are the effects of their presence in the soil environment? (iii) Do they

For these reasons, the aim of this chapter is to review and summarize existing knowledge of

Conventionally, the environmental fate of antimicrobials in the soil ecosystem is assessed with respect to their persistence and sorption onto soil. In the case of SAs, as they are very stable, only photodegradation process could have recognizable influence on their elimination from the environment [1,10]. However this process in the soil ecosystem is of lesser importance. Therefore, sorption processes influence the environmental behaviour of SAs to the greatest

Although a few review papers have been published summarizing the available ecotoxicity data of pharmaceuticals, including some SAs [14-15,17-18], they focus on aquatic organisms rather than soil. In this chapter, therefore, we shall discuss the available data on SA toxicity towards different soil organisms on various trophic levels like bacteria, invertebrates and plants. These results will be discussed with respect to the existing requirements for the environmental risk assessment of veterinary pharmaceuticals (VICH, 2000 [19]; VICH, 2004 [20]; EMEA, 2007 [21]; EMEA, 2008a [22]; EMEA, 2008b [23]). In addition, we shall identify some areas where future work is warranted as well as the needs for further investigations.

Sorption is defined as a phenomenon during which chemicals become associated with solid phases. Immensely important, this process affects the fate of chemicals in the environ‐ ment [24-25]. Adsorption/desorption experiments are useful for generating essential information on the mobility of various contaminants and their distribution in the soil, water and air. Assessing the distribution of a chemical between the soil and aqueous phases is not a straightforward process, however. It depends on various factors, such as the chemi‐ cal nature of the substance and the characteristics of the soil (*e.g.* pH, organic matter (OM) content, clay fraction content). Furthermore, climatic factors such as rainfall, temperature, sunlight intensity and wind, which can affect the strength of sorption, also have to be taken into consideration. Thus, the numerous phenomena and mechanisms involved in the adsorption of a chemical by soil cannot be completely defined by simple laboratory models. Nevertheless, such investigations can provide valuable information on the environmental

Laboratory sorption experiments can be carried out under static (batch) and dynamic (column) conditions. Static tests are commonly used when the aim of the study is to calculate the distribution coefficient (Eq. 1), in equilibrium time, which is specific to every chemi‐ cal. Column tests, on the other hand, enable time-dependent monitoring of contaminant

pose a risk to different soil organisms and also to human health?

660 Environmental Risk Assessment of Soil Contamination

the fate and effects of SA residues in the terrestrial environment.

extent, so it is these that we shall be discussing in detail.

**2. Fate of SAs in the terrestrial environment**

**2.1. Basic concept of sorption modelling studies**

relevance of the adsorption of chemicals [26].

$$\mathbf{c}\_s = \frac{\mathbf{c}\_{\text{max}} \cdot \mathbf{K}\_L \cdot \mathbf{c}\_w}{\mathbf{1} + \mathbf{K}\_L \cdot \mathbf{c}\_w} \tag{2}$$

where: *cs* is the content of test substance adsorbed on the soil at adsorption equilibrium (mg kg-1); *cw* is the mass concentration of test substance in the aqueous phase at adsorption equilibrium (mg L-1); *cmax* is the maximum sorption capacity of the sorbent*; KL* is the Langmuir constant.

The linear form of this equation is:

$$\frac{1}{c\_s} = \frac{1}{K\_L c\_{\text{max}} c\_w} + \frac{1}{c\_{\text{max}}} \tag{3}$$

In the Langmuir model the mass of solute sorbed per unit mass of sorbent (*cs)* increases linearly as the solute concentration increases. The model is based on three main assumptions:

**•** the sorption energy is the same for all sites and is independent of the degree of surface coverage;


Based on these findings, it is justified that the Langmuir isotherms are observed only very rarely in case of sorption of when organic compounds are sorbed in such a complex and heterogeneous sorbent as soils [31].

The Freundlich model takes the following form:

$$
\mathcal{L}\_s = \mathbf{K}\_F \cdot \mathbf{c}\_w^{1/n} \tag{4}
$$

where: *cs* is the content of test substance adsorbed on the soil at adsorption equilibrium (mg kg-1); *cw* is the mass concentration of test substance in the aqueous phase at adsorption equilibrium (mg L-1); *KF* is the Freundlich adsorption coefficient; *n* is the regression constant

Based on the value of *1/n*, which describes the relative magnitude and diversity of energies associated with sorption, the mechanism of this process can be defined more accurately [31-36]. If *1/n* = 1, adsorption is linear, which indicates the occurrence of homogeneous energy sites on the sorbent. This type of adsorption is generally typical at very low solute concentrations with many sites on the sorbent still available for possible interaction with the adsorbate. A value of *1/n* > 1 indicates a concave, upward-curving isotherm, which is sometimes referred to as a solvent-affinity type isotherm (S-type), where the sorption energy increases with increasing surface concentration. However, phenomena unrelated to the sorption of the analyte to the sorbate surface, such as strong adsorption of the solvent, strong intermolecular attraction within the adsorbent layers, penetration of the solute in the adsorbent, and the monofunctional nature of the adsorbate, can also affect the shape of the isotherm [31]. When *1/n* < 1, the isotherm is of the convex, downward-curving Langmuir-type (L-type), where the marginal sorption energy decreases with increasing surface concentration. This may occur when the competition of solvent for sites is minimal or the absorbate is a planar molecule [36].

In order to calculate the Freundlich factors as *1/n* or *KF*, Eq. 4 can be linearized by a logarithmic transformation:

$$
\log c\_s = 1/n \log \ c\_w + \log K\_F \tag{5}
$$

It has been shown that the sorption of pharmaceutical compounds (PCs) to soils is influenced by solution chemistry and the type of mineral and organic sorbents [24,25,37]. Pharmaceuticals can exist as either neutral or charged species (e.g. cations, anions, zwitterions) [24], depending on the pH. Various adsorptive forces may therefore be acting. Whereas neutral molecules partition to solid matrices via relatively weak van der Waals and electron donor-acceptor interactions, charged species can interact via stronger electrostatic mechanisms, such as cation exchange, cation-bridging and complexation. The acid-base equilibrium of sulphonamides resulting from either the loss or the gain of a proton is similar for all sulphonamides apart from sulphaguanidine (see Białk-Bielińska et al. [3]). The basic properties are due to the aniline group present in all SAs and specific to each SA heterocyclic base. SAs can thus be described by three dissociation constants. Nevertheless, since the presence of the protonated form of the heterocyclic functional group is extremely unlikely, only two of the possible ionized functional groups existing in the structure of SA molecules are taken into consideration – during sorption experiments [38-45]. Hence, Ka2 is the dissociation constant for the equilibrium between the positively charged, protonated amino group of a SA and its electrically neutral conjugate base, whereas Ka3 refers to the equilibrium involving the loss of the SA proton to yield its negatively charged conjugate (Figure 1) [7].

N H S O O <sup>3</sup>NH + R N H S O O <sup>2</sup>NH R S N O O <sup>2</sup>NH R Ka2 Ka3 + H+

**Figure 1.** Dissociation equilibrium of sulphonamides [7]

**•** sorption occurs only at localized sites, with no interaction between adjoining sorbed

Based on these findings, it is justified that the Langmuir isotherms are observed only very rarely in case of sorption of when organic compounds are sorbed in such a complex and

1/*n*

where: *cs* is the content of test substance adsorbed on the soil at adsorption equilibrium (mg kg-1); *cw* is the mass concentration of test substance in the aqueous phase at adsorption equilibrium (mg L-1); *KF* is the Freundlich adsorption coefficient; *n* is the regression constant

Based on the value of *1/n*, which describes the relative magnitude and diversity of energies associated with sorption, the mechanism of this process can be defined more accurately [31-36]. If *1/n* = 1, adsorption is linear, which indicates the occurrence of homogeneous energy sites on the sorbent. This type of adsorption is generally typical at very low solute concentrations with many sites on the sorbent still available for possible interaction with the adsorbate. A value of *1/n* > 1 indicates a concave, upward-curving isotherm, which is sometimes referred to as a solvent-affinity type isotherm (S-type), where the sorption energy increases with increasing surface concentration. However, phenomena unrelated to the sorption of the analyte to the sorbate surface, such as strong adsorption of the solvent, strong intermolecular attraction within the adsorbent layers, penetration of the solute in the adsorbent, and the monofunctional nature of the adsorbate, can also affect the shape of the isotherm [31]. When *1/n* < 1, the isotherm is of the convex, downward-curving Langmuir-type (L-type), where the marginal sorption energy decreases with increasing surface concentration. This may occur when the competition

In order to calculate the Freundlich factors as *1/n* or *KF*, Eq. 4 can be linearized by a logarithmic

It has been shown that the sorption of pharmaceutical compounds (PCs) to soils is influenced by solution chemistry and the type of mineral and organic sorbents [24,25,37]. Pharmaceuticals can exist as either neutral or charged species (e.g. cations, anions, zwitterions) [24], depending on the pH. Various adsorptive forces may therefore be acting. Whereas neutral molecules partition to solid matrices via relatively weak van der Waals and electron donor-acceptor interactions, charged species can interact via stronger electrostatic mechanisms, such as cation exchange, cation-bridging and complexation. The acid-base equilibrium of sulphonamides

log 1 / log log *s wF c nc K* = + (5)

of solvent for sites is minimal or the absorbate is a planar molecule [36].

*s Fw c Kc* = × (4)

**•** the sorption maximum (cmax) represents a monolayer coverage.

molecules;

transformation:

heterogeneous sorbent as soils [31].

662 Environmental Risk Assessment of Soil Contamination

The Freundlich model takes the following form:

Assessing the sorption of veterinary pharmaceuticals in soils is extremely important for estimating the risk of the large-scale usage of veterinary medicines to human health and environmental matrices, because this affects the fate and transport of chemicals in ground water.

#### **2.2. Sorption potential of SAs to soils**

Although SA sorption is quite a common topic of investigation, authors generally focus on just a few SAs, so that knowledge of the sorption behaviour of some of them (e.g. sulphaguanidine, sulphisoxazole) is still limited. So far, only a few review papers have been published describing the sorption of SAs to soils [16,46-49]. However, they cover a wide range of pharmaceuticals, so SAs are inadequately reviewed. Furthermore, since 2011 (when the last review on SA sorption was published), new data have been published, which are included in the present review. The available information on the sorption of the most commonly investigated com‐ pounds will therefore be discussed in depth.

The most widely investigated SA is sulphamethazine (SMZ). The level to which its undergoing sorption onto soils was investigated already more than thirty years ago by Langhammer [50]. That author calculated adsorption coefficients for four different soils, differing in pH and OM content. Based on the values of the distribution coefficients (from 1.0 to 3.1 L kg -1), this drug can be considered as a very mobile chemical. These results are in accordance with the inves‐ tigations of other researchers, such as Thurman et al. [51] and Tolls et al. [52], who reported low sorption coefficients for SMZ (0.6 L kg -1 and 3.0 L kg-1 respectively) or Thiele-Bruhn et al. [53], who gave a *Kd* value of 2.4 L kg-1 for humus-rich soil. This was also supported by Lertpaitoonpan et al. [54], who examined this SA in terms of the distribution in soils varying in OM content (*Kd* lies between 0.2 and 3.9 L kg-1 depending on the physicochemical parameters of soils). However, Fan et al. [55] reported a higher sorption potential of the polar metabolite of SMZ (*N*4-acetyl-SMZ) during a miscible-displacement experiment (column test). The *Kd* values obtained by these authors range from 7.5 andto 206.2 L kg-1 and are much higher than previous data for the native compound. However, this may well be due to the polar functional group present in *N*4-acetyl-SMZ, which could enhance the association of this compound to the negatively charged soil surfaces via cation bridging or complexes. The high mobility of SMZ was also reported by Kurwadkar et al. [41], who observed a 50-90% release of SMZ from a soil column system. The most recent studies presented by Leal et al. [56] underscore the concern regarding the possible occurrence of this compound in the environment. These authors investigated a number of different Brazilian soils, finding a tendency for SMZ to leach from soil matrices.

Much attention has also been given to calculating the sorption potential of sulphadiazine (SDZ). Just recently, Doretto et al. [45] reviewed the available literature data through Freund‐ lich sorption coefficient (*KF*) for SDZ. On this basis they concluded that the potential of SDZ to interact with soil particles is relatively low and depends on the type of soil, thus on the physicochemical properties of the sorbent. In another work, these authors demonstrated the weak interaction of SDZ with binding sites on the soil surface (*KF* values from 0.4 to 2.6 µg1-1/ n(cm3 ) 1-1/ng-1). The column studies of Wehrhan et al. [57] showed that the amount of leached SDZ depends strongly on the duration of the process. The eluted mass fraction was consider‐ ably higher in long-pulse experiments (83 and 61% respectively) than in short-pulse ones, during which only 18% was leached. Furthermore, these authors recorded the highest concentrations at the top of the column, with concentrations steadily decreasing towards the bottom. In the column with the short pulse application, solute concentrations were relatively uniformly distributed. Environmental conditions like rainfall can therefore affect the distri‐ bution of contaminants in soil.

It was also observed that SDZ exhibits, for example, a lower tendency to be retained in solid matrices than SMZ [53,56], with respective *Kd* values for SDZ and SMZ varying from 2.0 to 2.4 L kg-1 as reported by Thiele-Bruhn et al. [53], and from 5.2 to 10.5 L kg -1, as obtained by Leal et al. [56].

Although sulphachloropyridazine (SCP) is not as widely studied a sulphonamide as SMZ or SDZ, this drug has been extensively examined using various tests besides batch or laboratory column tests. The data available in the literature show a sorption potential in soil similar to that of other SAs. For example, Boxall et al. [58] reported low sorption coefficients for SCP in soil and a soil/slurry mixture ranging from 0.9 to 1.8 L kg-1. They also confirmed the high mobility of SCP in field studies, demonstrating the rapid transport of SCP to surface waters for concentrations as high as 590 µg L-1. Other studies are also consistent with this statement [38,44,52,56], giving *Kd* values from 0.7 to 70.1 L kg-1.

Blackwell et al. [59], who also examined the leaching of SCP under field conditions, detected this compound in surface run-off samples even at a concentration of 25.9 µg L-1 following application at a rate of 1.18 kg ha-1. These authors reported the occurrence of SCP in soil water samples at a concentration of 0.8 µg L-1 at 40 cm depth as long as 20 days after treatment. On the basis of their results the authors concluded that SCP poses a moderate risk of contaminating ground or surface water but that its potential to accumulate in soils is low. Further lysimeterbased studies by Blackwell et al. [60] sporadically detected SCP in leachate at levels from 0.7 to 8.5 µg L-1, depending on the irrigation conditions. SCP was applied in slurry (197 mL per lysimeter), which corresponds to a SCP application rate of 5.2 mg (1.18 kg ha-1). The authors concluded that this compound has the potential to reach ground and surface waters. On the basis of a lysimeter study, field investigations and laboratory column tests, Kay et al. [61,62,63] pointed out that soil tillage prior to slurry application can significantly reduce losses of SCP to tile drainage systems, thereby reducing the risk of surface water contamination by SCP residues in the slurry. The observed losses of SCP in a soil column with a soil surface broken as a result of tillage fell from 54.6 % of the applied amount to zero [62].

of SMZ (*N*4-acetyl-SMZ) during a miscible-displacement experiment (column test). The *Kd* values obtained by these authors range from 7.5 andto 206.2 L kg-1 and are much higher than previous data for the native compound. However, this may well be due to the polar functional group present in *N*4-acetyl-SMZ, which could enhance the association of this compound to the negatively charged soil surfaces via cation bridging or complexes. The high mobility of SMZ was also reported by Kurwadkar et al. [41], who observed a 50-90% release of SMZ from a soil column system. The most recent studies presented by Leal et al. [56] underscore the concern regarding the possible occurrence of this compound in the environment. These authors investigated a number of different Brazilian soils, finding a tendency for SMZ to leach from

Much attention has also been given to calculating the sorption potential of sulphadiazine (SDZ). Just recently, Doretto et al. [45] reviewed the available literature data through Freund‐ lich sorption coefficient (*KF*) for SDZ. On this basis they concluded that the potential of SDZ to interact with soil particles is relatively low and depends on the type of soil, thus on the physicochemical properties of the sorbent. In another work, these authors demonstrated the weak interaction of SDZ with binding sites on the soil surface (*KF* values from 0.4 to 2.6 µg1-1/

1-1/ng-1). The column studies of Wehrhan et al. [57] showed that the amount of leached SDZ depends strongly on the duration of the process. The eluted mass fraction was consider‐ ably higher in long-pulse experiments (83 and 61% respectively) than in short-pulse ones, during which only 18% was leached. Furthermore, these authors recorded the highest concentrations at the top of the column, with concentrations steadily decreasing towards the bottom. In the column with the short pulse application, solute concentrations were relatively uniformly distributed. Environmental conditions like rainfall can therefore affect the distri‐

It was also observed that SDZ exhibits, for example, a lower tendency to be retained in solid matrices than SMZ [53,56], with respective *Kd* values for SDZ and SMZ varying from 2.0 to 2.4 L kg-1 as reported by Thiele-Bruhn et al. [53], and from 5.2 to 10.5 L kg -1, as obtained by Leal

Although sulphachloropyridazine (SCP) is not as widely studied a sulphonamide as SMZ or SDZ, this drug has been extensively examined using various tests besides batch or laboratory column tests. The data available in the literature show a sorption potential in soil similar to that of other SAs. For example, Boxall et al. [58] reported low sorption coefficients for SCP in soil and a soil/slurry mixture ranging from 0.9 to 1.8 L kg-1. They also confirmed the high mobility of SCP in field studies, demonstrating the rapid transport of SCP to surface waters for concentrations as high as 590 µg L-1. Other studies are also consistent with this statement

Blackwell et al. [59], who also examined the leaching of SCP under field conditions, detected this compound in surface run-off samples even at a concentration of 25.9 µg L-1 following application at a rate of 1.18 kg ha-1. These authors reported the occurrence of SCP in soil water samples at a concentration of 0.8 µg L-1 at 40 cm depth as long as 20 days after treatment. On the basis of their results the authors concluded that SCP poses a moderate risk of contaminating ground or surface water but that its potential to accumulate in soils is low. Further lysimeter-

soil matrices.

664 Environmental Risk Assessment of Soil Contamination

n(cm3 )

et al. [56].

bution of contaminants in soil.

[38,44,52,56], giving *Kd* values from 0.7 to 70.1 L kg-1.

Studies of sulphamethoxazole (SMX) have reported a similar sorption potential to that of SMZ [40]. With some exceptions, Leal et al. [56] recorded similar values of the distribution coefficient *Kd* for both SMX and SMZ in an examination of thirteen soils. The results obtained by Yu et al. [64] are in agreement with that. These authors calculated a *Kd* of 18.9 L kg-1 for one of three investigated soils. Their aim was to assess the suitability, inter alia, of SMX as a wastewater indicator. However, owing to the formation of non-extractable residues, such an application of SMX was regarded as limited. On the basis of a few investigations into the sorption of SMX to activated sludge, we can state that the sorption potential of sulphonamides to this sorbent is much greater than to soils [65-69]. Hrsing et al. [65] presented *Kd* values for SMX ranging from 280 to 370 L kg-1, depending on the type of activated sludge. The results are consistent with the investigations of Hyland et al. [66], who studied the sorption of 75 pharmaceuticals onto activated sludge, obtaining a *Kd* value of 269 L kg-1, or with those of Göbel et al. [67], who obtained a similar value of *Kd* for SMX. In contrast, Yang et al. [68] reported a lower sorption of SMX to activated sludge (*Kd* = 28.6 L kg-1). However, these differences may have arisen, for example, from the different methodologies used in the tests. Nevertheless, Yang et al. [68, 69] concluded that sorption of SAs to activated sludge is highly reversible (the amount retained after desorption is 0.9% of the original dose of 100 µg L-1). Therefore, the use of sewage sludge as fertilizer may constitute an additional source of SAs in ground and surface waters. More‐ over, since SAs may be taken up by farmland crops, as demonstrated by Li et al. [70], the use of sewage sludge as fertilizer poses a serious risk to human health as well.

Knowledge of the sorption potential of other SAs like sulphathiazole (STZ), sulphapyridine (SPY), sulphamerazine (SMR), sulphadimethoxine (SDM), sulphamonomethoxine (SMM), sulphaguanidine (SGD) or sulphisoxazole (SSX) is very limited. In the literature there are only a few reports dealing with the sorption of these pharmaceuticals. *Kd* for STZ adsorption onto soil particles ranges from 1.0 to 62.5 L kg-1 depending on soil properties [56]. In the case of SDM, Sanders et al. [71] pointed out that the linear sorption coefficient for SDM differs somewhat, ranging from 0.4 to 25.8 L kg-1 as a single solute and from 2.5 to 22.1 L kg-1 as a cosolute with ormetoprim, another antimicrobial. Moreover, the sorption of SDM was less linear in combination with ormetoprim. In turn, Maszkowska et al. [72] did not determine the influence of the co-solute on sulphonamide release. SDM exhibited a similar leaching behav‐ iour from the soil when it was tested alone or in a mixture with SGD. Nevertheless, these authors also reported the considerable mobility of three SAs (SDM, SGD, SSX) in three different soils; SDM was released the slowest from the soil column. These results are consistent with those published previously by Białk-Bielińska et al. [43], who showed that SDM had a greater sorption potential than SGD (*Kd* = 0.3 – 107.5 L kg-1 for SDM, 1.0 – 31.0L kg-1 for SGD). SDM was also found to have the strongest tendency of all the SAs investigated to interact with activated sludge [68,69]. In addition, these authors investigated the sorption strength of sulphamonomethoxine (SMM) on activated sludge, finding a lower affinity of SMM than of SDM for activated sludge particles. Jin et al. [73] demonstrated the relatively high mobility of SDM in soil, obtaining a *Kd* of 18.9 L kg-1. These authors also highlighted the influence of different co-contaminants on adsorption. They concluded that anionic surfactants and urea could adversely affect the sorption potential of SDM, whereas cationic surfactants could improve the retention of SDM on soil particles. Figueroa-Diva et al. [40] found that SMR exhibited the lowest level of sorption of the four SAs (SDM, SMX, SMZ and SMR) that they examined. According to Thiele-Bruhn et al. [53], SPY was the most strongly retained SA in the soil matrix, with *Kd* higher than that of SMZ, SDZ and SDM.

Summing up, the available data indicate that determining the environmental fate of SAs in soils is not an easy task, as this depends largely on the physicochemical properties of soils and the chemical structures of the SAs. Nevertheless, one can infer from these results that these pharmaceuticals will tend to leach into ground or surface water rather than persist in soils. These data also show, however, that a certain amount of the SAs entering the soil can be retained there for quite a long time. Furthermore, their sorption to soils can increase or decrease depending on a number of different factors, which are discussed below.

## **2.3. Factors influencing sorption of SAs to soils**

#### *2.3.1. Influence of the organic/mineral composition of soil*

Soils can be regarded as mixtures of mineral and organic fractions. The differences in their texture, structure, consistency, colour, chemical, biological and other characteristics arise from the type of parent material. Soils are therefore diverse matrices in which different sorption mechanisms can occur. The organic matter (OM) content undoubtedly plays a critical role in the sorption capacity of soils [74]. Overall, in accordance with the available literature data, it has been shown that OM positively affects sorption strength of organic compounds. Figueroa-Diva et al. [40] reported values of *Kd* for all examined SAs increasing in the same sequence as the organic carbon content (ƒOC) in the soils they investigated. Białk-Bielińska et al. [43] pointed out that SAs predominantly interact with soil OM by nonbonding van der Waals interactions and hydrogen bonding. Furthermore, such weak bonding forces are susceptible to desorption, resulting in the free release of SAs following their prior surface adsorption, an observation previously made by Thiele-Bruhn et al. [53]. The authors indicated that the influence of soil OM on adsorption depends not only on the organic carbon content, but also on its composition. Sukul et al. [39] demonstrated increased sorption of SDZ in soils in the presence of manure compared to soil without manure, which greatly emphasizes the role of dissolved OM and organomineral soil particles in SDZ sorption. On the basis of original research and literature available data, Hou et al. [75] demonstrated a positive relationship between *Kd* and the organic carbon content (ƒOC) for SMX sorption on soils/sediments with ƒOC > 2%. However, for adsorbents with ƒOC < 2%, a lower ƒOC could result in increased sorption, suggesting competi‐ tion between SMX and organic matter on mineral particles. Hyland et al. [66] confirmed the positive influence of organic matter on SA sorption. The high values of *Kd* for SMX sorption onto activated sludge are fully justified, due in great part to the organic carbon (average ƒOC = 44.1%) in the sorbent. Leal et al. [56] analysed the influence of the OM and clay content on the sorption of several SAs, concluding that hydrophobic partition was important in SA sorption. Nevertheless, they also found that non-hydrophobic interactions with organic and/or mineral surfaces, mainly with Al and Fe oxides and hydroxides (abundant in the investigated soils) were also important in SA retention in soils. Boxall et al. [58] determined the influence of the type of mineral fraction in soil (clay or sand) on *Kd*. Their results showed that clay had a greater sorption capacity for SCP than sand. The same was reported by Ter Laak et al. [38], whose *Kd* value for soil with greater amount of clay was twice as high as that for sandy soil. During field studies, however, Boxall et al. [58] reported faster leaching of SCP to ground water from a clay site than from a sandy site, an observation corroborated by Fan et al. [55]. The *Kd* values for SMZ were positively related to the OM content in case of sorbents without sand. However, the latter authors' *Kd* value was higher for sand (%OM=0) than for soil containing OM. They explained this as having resulted from the transport of SAs on mobile colloids (< 2 µm, dissolved organic matter and clay-sized materials) in accordance with EPA [76], which resulted in faster elution from a soil column with OM content than from sand.

#### *2.3.2. Influence of pH*

those published previously by Białk-Bielińska et al. [43], who showed that SDM had a greater sorption potential than SGD (*Kd* = 0.3 – 107.5 L kg-1 for SDM, 1.0 – 31.0L kg-1 for SGD). SDM was also found to have the strongest tendency of all the SAs investigated to interact with activated sludge [68,69]. In addition, these authors investigated the sorption strength of sulphamonomethoxine (SMM) on activated sludge, finding a lower affinity of SMM than of SDM for activated sludge particles. Jin et al. [73] demonstrated the relatively high mobility of SDM in soil, obtaining a *Kd* of 18.9 L kg-1. These authors also highlighted the influence of different co-contaminants on adsorption. They concluded that anionic surfactants and urea could adversely affect the sorption potential of SDM, whereas cationic surfactants could improve the retention of SDM on soil particles. Figueroa-Diva et al. [40] found that SMR exhibited the lowest level of sorption of the four SAs (SDM, SMX, SMZ and SMR) that they examined. According to Thiele-Bruhn et al. [53], SPY was the most strongly retained SA in the

Summing up, the available data indicate that determining the environmental fate of SAs in soils is not an easy task, as this depends largely on the physicochemical properties of soils and the chemical structures of the SAs. Nevertheless, one can infer from these results that these pharmaceuticals will tend to leach into ground or surface water rather than persist in soils. These data also show, however, that a certain amount of the SAs entering the soil can be retained there for quite a long time. Furthermore, their sorption to soils can increase or decrease

Soils can be regarded as mixtures of mineral and organic fractions. The differences in their texture, structure, consistency, colour, chemical, biological and other characteristics arise from the type of parent material. Soils are therefore diverse matrices in which different sorption mechanisms can occur. The organic matter (OM) content undoubtedly plays a critical role in the sorption capacity of soils [74]. Overall, in accordance with the available literature data, it has been shown that OM positively affects sorption strength of organic compounds. Figueroa-Diva et al. [40] reported values of *Kd* for all examined SAs increasing in the same sequence as the organic carbon content (ƒOC) in the soils they investigated. Białk-Bielińska et al. [43] pointed out that SAs predominantly interact with soil OM by nonbonding van der Waals interactions and hydrogen bonding. Furthermore, such weak bonding forces are susceptible to desorption, resulting in the free release of SAs following their prior surface adsorption, an observation previously made by Thiele-Bruhn et al. [53]. The authors indicated that the influence of soil OM on adsorption depends not only on the organic carbon content, but also on its composition. Sukul et al. [39] demonstrated increased sorption of SDZ in soils in the presence of manure compared to soil without manure, which greatly emphasizes the role of dissolved OM and organomineral soil particles in SDZ sorption. On the basis of original research and literature available data, Hou et al. [75] demonstrated a positive relationship between *Kd* and the organic carbon content (ƒOC) for SMX sorption on soils/sediments with ƒOC > 2%. However, for

soil matrix, with *Kd* higher than that of SMZ, SDZ and SDM.

666 Environmental Risk Assessment of Soil Contamination

**2.3. Factors influencing sorption of SAs to soils**

*2.3.1. Influence of the organic/mineral composition of soil*

depending on a number of different factors, which are discussed below.

In the context of the acid-base equilibrium of SAs, pH can strongly affect sorption. This has been confirmed in many investigations. The overall trend presented in the literature indicates decreasing sorption of SAs with increasing pH. This is explained by the amphoteric nature of SAs, which consequently can occur in cationic, anionic or neutral form. The strongest possible interactions (ion-exchange mechanism) arise from competition for negatively charged sites on the soil surface between a cationic analyte and other cations present in the solute. Nevertheless, the existence of cationic SAs is limited due to the relatively low *pKa2* value. Cation exchange is therefore not regarded as a favourable mechanism for SA sorption to soil matrices [39]. Although decreasing SA sorption is observed at high pH, Sukul et al. [39] achieved relatively strong adsorption in the case of one soil at a pH where the anionic form of SDZ was dominant, claiming that this was due to possible partition to the positively charged surfaces of pedogenic oxides, very abundant in the clay fraction [39,53]. Kim et al. [42] and Białk-Bielińska et al. [43] also observed a negative correlation between *Kd* and pH. The former authors considered that the changes were better evident for soil with a greater OM content. Pinna et al. [77], in turn, did not observe such a strong dependence on OM content. The addition of cow manure (ƒOC = 30.58) did not significantly affect antibiotic sorption to one of the investigated soils, but did increase the extent of sorption to another soil about three times, even though larger amounts of manure had been added to the first soil than to the second one. These authors concluded that the greater sorption to the second soil prior to the addition of cow manure was most probably due to the low pH of the soil suspension rather than to its high organic carbon content. On the other hand, the high pH of the first soil suspension (7.8) could have been responsible for the reduced sorption, despite the considerable amount of OM in this amended soil.

#### *2.3.3. Influence of ionic strength*

Another environmentally important factor that can affect SA sorption is ionic strength. But this has not been examined extensively. Ter Laak et al. [38] carried out sorption studies of SCP, among other compounds. Generally speaking, they did not observe any significant influence of ionic strength, except in the case of one soil (clay loam), in which sorption doubled when the CaCl2 concentration was raised from 0.006 to 0.2 M. The authors concluded that this increase in sorption was probably due to the neutral form of SCP increasing from 3.3 to 8.3% because of the decreasing pH. Protons are displaced from the cation-exchange sites by the addition of Ca2+ cations, which are ultimately responsible for the decrease in pH. Elevated cation concentrations near negatively charged soil surfaces, resulting in a decrease in the electrostatic repulsion of negatively charged sorbate molecules and soil particles, is another explanation considered by those authors. Srinivasan et al. [44] reported the different behaviour of SMX under conditions of increasing ionic strength. They explained the increasing *Kd* for SMX in the case of one soil as being due to cation bridging. Since positively charged calcium ions are present in the solution, bonding of anionic SMX to calcium is possible. In addition, the occurrence of a salting out effect, reducing the solubility of SMX in the salt solution so that it precipitates in the soil, was taken into consideration as a possible reason for the increase in sorption. The positive influence of ionic strength on sorption can also be attributed to the replacement of protons from the soil surface as the ionic strength increases, causing a slight reduction in pH, and shifting acidic SMX towards neutral forms that are more strongly sorbed than the anionic forms. Two other soils examined by Srinivasan et al. [44] exhibited an opposite and irregular trend in sorptive affinity of SMX, with elevated ionic strengths resulting in decreased sorption coefficients of SMX in the case of both soils. A slight decrease in sorption with increasing ionic strength of solute was also observed by Białk-Bielińska et al. [43] in the case of SDM and SGD and three soils. Srinivasan et al. [44] concluded that the ionic composition plays an important role in the sorption of ionizable organic compounds. Nevertheless, they, too, highlighted the necessity for further research in view of the conflicting results published in the literature.

#### **2.4. Available data on the presence of SAs in soils**

Although many methods have been developed in the past decade for the analysis of SAs in aqueous matrices, only a few are described in the literature for the determination of these contaminants in soil matrices. This is because the chemical analysis of pharmaceuticals from soil matrices is complicated by the need for extraction. Hence, our knowledge about the quantity of SAs in solid matrices is still limited. Nevertheless, the available literature data indicate their occurrence in agricultural soils after conventional fertilization. In a two-year monitoring study Höper et al. [78] determined SMZ at a concentration of 11 µg kg-1. Pawelzick et al. [79] reported a maximum concentration of 4.5 µg kg-1 for SMZ; these results are in agreement with Hu et al. [80], who demonstrated the occurrence of SMX (0.03 – 0.9 µg kg-1) and SCP (0.18 – 2.5 µg kg-1). Karcı et al. [81] found three SAs in agricultural soils in Turkey at concentrations even two orders of magnitude higher than those reported in previous studies: STZ (0.05 – 0.4 mg kg-1), SCP (0.05 – 0.1 mg kg-1) and SMX (0.05 – 0.1 mg kg-1). There are some discrepancies in the available literature data, which may be due to differences in the physico‐ chemical properties of the solid samples examined. They may also stem from the intensity of fertilization and the initial quantities of SAs applied in animal husbandry. Nevertheless, even low concentrations of SAs reported in soil samples may contaminate other environmental compartments as a result of release via desorption.

In general, concentration limits of antibiotics in the environment are not regulated, even though growing public concern has been taken into account in the prescription of environ‐ mental risk assessments of veterinary pharmaceuticals in the USA and Europe [19-23]. For these reasons, it is still necessary to develop analytical methods for the quantitation of the most important SAs in soil samples: this will help to estimate the exposure potential as well as the concentration of these substances in the terrestrial environment. Nevertheless, for a full risk assessment of these compounds not only is an exposure assessment necessary but also a hazard characterization, which addresses the question whether a substance has the potential to cause harmful effects. This will be discussed below.
