No job name

Critical Review
Sorption of Veterinary Pharmaceuticals in Soils:
A Review
J O H A N N E S T O L L S *Environmental Toxicology and Chemistry, Institute of Risk Assessment Sciences,Utrecht University, P.O. Box 80176, 3508 TD Utrecht, The Netherlands Veterinary pharmaceuticals (VPs) are used in large Hence, a considerable portion of the VPs can reach the soil amounts in modern husbandry. Due to their use pattern, environment as constituent of urine, feces, or manure (2, 5, they possess a potential for reaching the soil environment.
6). Tetracycline, for instance, has been detected in liquid To assess their mobility in soil, the literature on sorption manure-treated agricultural fields at concentrations of about of chemicals used as VPs is reviewed and put into perspective 10 µg/kg (7), and a monitoring program in the United Statesdirected at waters that are suspected to be contaminated of their physicochemical properties. The compilation of with antibiotics used in husbandry detected trimethoprim sorption coefficients to soil solids (Kd,solid) demonstrates that and sulfamethoxazole in 30% of the samples (8). Therefore, these chemicals display a wide range of mobility (0.2 < the question arises of what happens with the VPs once they 6000 L/kg). Partition coefficients for association of tetracycline and quinolone carboxylic acid VPs to dissolved This question is, among others, addressed formally in the organic matter (Kd,DOM) vary between 100 and 50 000 environmental risk assessment, which is part of the registra- tion procedure for VPs in the European Union since 1996 (9).
d,solid for a given compound in different soils can be significant. For most of the compounds, the One important aspect of this question is the mobility of VPs variation is not considerably lower for the organic carbon- in the soil environment. Highly mobile VPs have the potential to leach to the groundwater (10) and be transported with the groundwater, drainage water, and surface runoff to surface of log Koc by log Kow leads to significant underestimation waters. There is thus a possible exposure pathway for aquatic of log Koc and log Kd,DOM values. This suggests that mechanisms organisms. Second, VPs may be encountered in drinking other than hydrophobic partitioning play a significant water if they, besides being highly mobile, are sufficiently role in sorption of VPs. A number of hydrophobicity- stable toward degradation in the treated animal, the manure, independent mechanisms such as cation exchange, cation and the soil and water purification processes. Strongly sorbing bridging at clay surfaces, surface complexation, and VPs can accumulate in the top layer of the soil. In this case, hydrogen bonding appear to be involved. These processes the availability of these compounds to soil-dwelling organ- are not accounted for by organic carbon normalization, isms becomes relevant. Hence, the assessment of sorption suggesting that this data treatment is conceptually and mobility of VPs in soil environments is of importance inappropriate and fails to describe the sorption behavior.
with regard to the risk of the use of VPs to human andenvironmental health.
Moreover, prediction of log Koc based on the hydrophobicity Sorption to Solids. The reversible sorptive exchange of
parameter log Kow is not successful.
chemicals between the water phase and a solid-phasesorbent, either soil or sediment, is represented by the sorptioncoefficient Kd,solid, which is defined as the ratio of the Introduction
concentrations of a compound in the sorbent phase (Cs) andin the water (Caq) at equilibrium (eq 1): Veterinary pharmaceuticals (VPs) are physiologically highlyactive substances used in husbandry for combating parasites, prevention and treatment of bacterially transmitted diseases, and acceleration of meat production. In the EU, antibiotics and anthelmintics (parasiticides) are the most importantgroups of VPs, both with a market volume of more than 200 Please note, that Caq refers to the concentration of the freely million Euros in 1999 (1). Of the total usage of 5000 t of dissolved molecules rather than to the total concentration antibiotics, 3500 t is used for therapeutic purposes (2) while in the soil solution, which might include fractions that are the remaining 1500 t is added to the feed in order to promote sorbed to suspended particles or to dissolved organic matter.
the growth of farm animals (3). In the United States, the Standard methods for determination of Kd,solid are column estimated use of antibiotics in livestock in 1985 amounted displacement studies (11) or batch sorption experiments (12).
to 8300 t (4). VPs are administered to the animals with In the column displacement experiment, Kd,solid is determined medicated feed, via injection, or by external application.
from the breakthrough curve, usually at one single concen- Depending on the chemical and the animal species, they are tration. Batch sorption experiments at multiple concentra- excreted as the parent compound, as conjugates, or as tions allow for construction of sorption isotherms from which oxidation or hydrolysis products of the parent compounds.
the dependence of Kd,solid on Caq can be determined. TheFreundlich isotherm equation (C ) used empirical isotherm representation in which Kf and n are the Freundlich sorption coefficient and the linearity VOL. 35, NO. 17, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3397
parameter, respectively. In the simplest case, n ) 1 and Kf is equivalent to Kd,solid, irrespective of the magnitude of Caq.
In that case, sorption is treated in analogy to Nernst Compounds Investigated. Table 1 displays the structures of
partitioning. The alternative Langmuir adsorption isotherm those VPs for which sorption data exist and demonstrates (eq 2) describes sorbate-sorbent interactions for sorbents that they contain a wide variety of functional groups. In the with a finite number of sorption sites.
column physical-chemical properties, the logarithm of then-octanol-water partition coefficient (log Kow) is specified.
Even though it is not a true partition coefficient for acidic and basic VPs (14), this parameter is considered useful for spanning a hydrophobicity scale to rank the VPs. The maximum value of this frequently applied measure of hydrophobicity is 3.5, indicating that VPs as a group of L and Cs,max are the Langmuir sorption coefficient and the maximum sorption capacity of the sorbent, respectively.
chemicals are not hydrophobic. This is underlined by the water solubilities (S) that exceed 1 g/L for most VPs. In LCaq (i.e., relatively high Caq), Cs addition, the table provides pKa data for tetracyclines, s,max, while for sufficiently small values of Caq quinolone carboxylic acids, efrotomycine, and sulfonamides.
1), Cs is linearly related to Caq and Cs,maxKL equals Many of the pKa values are in the range of pH values encountered in soils, indicating that many VPs are subject For many organic chemicals, the Nernst partition ap- to protonation/deprotonation reactions in the soil solution proach has been successful (13). In addition, it has been and that their speciation depends on the pH of the soil found that the Kd,solid of many neutral hydrophobic organic solution. In addition, the antibiotics of the tetracycline and chemicals depends on the organic carbon content (foc) of the quinolone carboxylic acid families form complexes with sorbent (13, 14). A significant reduction in the variability of multivalent cations. As can be inferred from Table 1 and the data is brought about by normalization of Kd,solid to the data by Ross and Riley (20) for a model quinolone carboxylic organic carbon content yielding the organic carbon-normal- acid, the stability of the Ca and Mg complexes is significantly ized sorption coefficient Koc according to lower than for the Al and Fe complexes.
General Findings. Tables 2 and 3 summarize the available
data for the VPs investigated for solid sorbents and DOM, respectively. The tables give information on the type andorigin of the sorbent, the type of experiment performed, some For such compounds, strong quantitative relationships qualitative information, and the values of the sorption have been established between the 1-octanol-water partition coefficients. In case Freundlich isotherms were used to describe the sorption data, the numbers reported in the ow) (13-15) and Koc. They have been rationalized by sorption to soil and sediment organic matter being column Kd,solid are actually Kf values. In that case, the analogous to dissolution into a bulk organic phase (13, 14).
nonlinearity parameter 1/n is provided as well.
These relationships allow for estimating K Sorption experiments have been reported for a range of easily obtained physicochemical property. Therefore, K VPs employing a variety of soils, sediments, soil constituents the favored measure of sorption in environmental risk such as clay minerals or sand, and dissolved organic matter.
The majority of the Kd,solid data has been obtained from batch Association to Dissolved Organic Matter. While sorption
sorption isotherms. In addition, Kd,solid observations from to soil solids decreases the mobility of chemicals in soils, batch experiments at one single concentration (21) as well association of solutes to dissolved organic matter (DOM) as from column displacement experiments (22, 23) were has the opposite effect by increasing the amount of chemical reported. Methods employed for determination of Kd,DOM for present in the soil water as has been demonstrated for PAHs VPs are batch experiments in which Caq is determined by and PCBs (17, 18). Given that the soil solution can contain equilibrium dialysis (24) or solid-phase microextraction (25).
considerable concentrations of dissolved organic carbon (19), Alternatively, Kd,DOM was determined from the change in there is the possibility of DOM-facilitated transport of VP electrophoretic mobility as a result of binding to humic transport in the soil. The association of solute to DOM is usually described by considering DOM as a third phase in Quantitative Extent of Sorption to Solid Sorbents. From
the system soil solids, water, and DOM. The association of Table 2, it appears that the tetracycline and quinolone solutes to DOM can be then be approximated by a partition carboxylic acid antibiotics (oxolinic acid, enrofloxacin, equilibrium in which the concentration associated to DOM ciprofloxacin, ofloxacin) display the highest values of Kd,solid (CDOM) is related to Caq via Kd,DOM, the DOM/water partition (range: 70-5000 L/kg). According to a classification of Kd,DOMCaq). It follows that VPs with pesticide mobility in soil (27), these VPs can be considered a high value of Kd,DOM will partition significantly to DOM, to be immobile. Please note that the extremely low Kd,solid possibly resulting in an increased mobility of the VPs in the values of 0.3 L/kg found for oxytetracycline and oxolinic acid in a marine sediment consisting almost exclusively (99.7%) Scope. In this review, literature data on VP sorption to
of sand (28) were not considered for the ranking of the VP soils as well as information on the physical-chemical classes. Avermectin, tylosin, and efrotomycin display inter- ow, water solubility, pKa values, and d,solid values ranging between 7 and 300 L/kg. In metal complex stability constants are collected from the contrast, the remaining VPs (olaquindox, sulfamethazine, literature through December 2000. The sorption coefficient sulfathiazole, metronidazole, chloramphenicol) appear to data are analyzed in order to investigate the validity of the have little sorption affinity to soil particles, as is evidenced paradigm of organic carbon normalization of sorption by their low values of Kd,solid (0.2-2 L/kg). The latter two coefficients. The literature findings are evaluated with regard groups of VPs can be considered to be low to slightly mobile to sorption mechanisms involved in and the influence of environmental properties on sorption of VPs. The discussion is to stimulate future research into mobility of VP and Organic Carbon Normalization. Organic carbon nor-
implementation of scientifically sound concepts into the risk malization is frequently employed to reduce the variability between Kd,solid data of one compound in different soils. To 3398 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 17, 2001
TABLE 1. Overview of Structures of VPs for Which Sorption Data Are Availablea
VOL. 35, NO. 17, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3399
TABLE 1 (Continued)
3400 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 17, 2001
TABLE 1 (Continued)
a The table includes information on the hydrophobicity (log Kow); aqueous solubility at neutral pH; acid-base reactivity (pKa values); and complexation to metals that are important in soil systems, such as Ca, Mg, Al, and Fe. Key: 1, ref 40; 2, ref 42; 3, ref 44; 4, ref 45; 5, ref 47; 6, KOWWINestimate (48) employing measured value for ciprofloxacin or norfloxacin (49) as starting point; 7, ref 29; 8, estimated on the basis of the effectof an alkyl substituent attached to the piperazine ring (52); 9, values at pH 7.4 (52); 10, ref 52; 11, ref 55; 12, ref 56; 13, ref 41; 14, ref 43; 15, ref23; 16, ref 46; 17, KOWWIN estimate (48); 18, ref 50; 19, ref 51; 20, measured value specified in log KOWWIN (48); 21, ref 53; 22, ref 54; na, notavailable.
evaluate whether this data treatment is applicable to VPs standard deviations as indicated by the error bars (Figure 1).
too, the average values of Kd,solid and Koc determined in The error bars display the large degree of variability in both different soils are plotted against each other along with their parameters. The standard deviation relative to the mean VOL. 35, NO. 17, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3401
TABLE 2. Overview of Literature Data on Sorption of VPs to Soils or Soil Constituentsa
Kd,solid
compound/corollary information
Tetracycline
pure Na-bentonite, Langmuir iso, pH dependency, Cs,max at pH 6.1: 78 µmol/g, KL not specified pure Ca-bentonite, Langmuir iso, Cs,max at pH 6.1: 200 µmol/g, KL not specified bentonite modified with cationic surfactant (C12-trimethylammonium), Langmuir iso, Cs,max at pH bentonite modified with tannic acid, Langmuir iso, Cs,max at pH 6.1: 210 µmol/g, KL not specified pure montmorillonite clay mineral, Langmuir iso, Cs,max at pH 5.0: 540 µmol/g, KL not specified soil organic matter (peat), Nova Scotia; pH 4.55 soil organic matter (peat), Nova Scotia; pH 6.14, iso’s nonlinear Oxytetracycline
marine sediment, sand fraction, 0.6% of particle mass smaller than 63 mm, Freundlich iso freshwater sediment from an eel pond, Taiwanc marine sediment from a shrimp farm, Taiwanc Enrofloxacin
Rhodic ferralsol, 22% clay fraction, kaolinite Glegic cambisol, 21% clay fraction, montmorillonite Haplic podsol, 25% clay fraction, montmorillonite Rendzic leptosol, 8% clay fraction, kaolinite Centric flurisol, 8% clay fraction, montmorillonite pure montmorillonite, sorption between clay layers, leading to expansion of the clay Ciprofloxacin
Centric flurisol, 8% clay fraction, montmorillonite Ofloxacin
Centric flurisol, 8% clay fraction, montmorillonite Centric flurisol, 8% clay fraction, montmorillonite Enro-CO2
Centric flurisol, 8% clay fraction, montmorillonite Oxolinic Acid
marine sediment, 97.3% of particle mass smaller than 63 mm, linear iso (1/n ) 1.02) marine sediment, 41.7% of particle mass smaller than 63 mm, Freundlich iso (1/n ) 1.22) marine sediment, sand, 0.6% of particle mass smaller than 63 mm, Freundlich iso (1/n ) 1.35) Efrotomycin
clay loam, Newton, IA; linear iso (1/n ) 0.96) silt loam, Three Bridges, NJ; Freundlich iso (1/n ) 1.3) loam, Riverside, CA; Freundlich iso (1/n ) 1.1) sandy loam, College Station, TX; linear iso (1/n ) 1.0) Avermectin
clay loam, Newton, IA;CD 92% of radioactivity in upper 20% of soil column, none in the rest, sand, Lakeland, FL;CD 92% of radioactivity in upper third of soil column, none in the rest, silt loam, Three Bridges, NJ;CD 92% of radioactivity in upper third of soil column, none in the sandy loam, College Station TX; >97% of radioactivity in upper 40% of soil column, none in Sulfathiazole
loamy sand, LuFa standard soil 2.2 (pH 5.2) Sulfamethazine
loamy sand, LuFa standard soil 2.1 (pH 5.8) loamy sand, LuFa standard soil 2.2 (pH 5.2) silt loam, Merzenhausen, Germany (pH 4.9) 3402 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 17, 2001
TABLE 2 (Continued)
Kd,solid
compound/corollary information
Metronidazole
Olaquindox
Chloramphenicol
freshwater sediment from an eel pond, Taiwanc marine sediment from a shrimp farm, Taiwanc a In the first column, information is given on sorbent, isotherm (iso) type, eventual pH dependence, and other observations. In Kd,solid and Koc, the sorption and the organic carbon-normalized sorption coefficient are given. If not specified otherwise (by superscript CD, column displaceexperiment), the data were generated in batch sorption experiments. b No information on isotherm linearity. c foc not reported; Koc not calculated.
d Kd calculated from sediment to water ratio and amount of compound recovered.
TABLE 3. Data on Sorption of VPs to DOMa
compound
corollary information
Kd (L/kg)
Koc (L/kg)
a AHA and HAS stand for Aldrich humic acid and for humic acid from a soil, respectively. The experimental techniques employed were equilibrium dialysis (ED), solid-phase microextraction (SPME), and electrophoretic mobility. Kd and Koc are the sorption and organic carbon-normalized sorptioncoefficients, respectively. b Organic carbon content of humic acid is not reported.
FIGURE 1. Plot of the organic carbon-normalized sorption coefficient
(Koc) against the soil weight-normalized sorption coefficient (Kd,solid).
The error bars represent one standard deviation of the sorption
FIGURE 2. Graphical representation of the relationship between
coefficients of one compound measured in different soils.
the individual log Koc data and hydrophobicity expressed as log
Kow. The solid line is a regression line obtained for a wide range
(RSD) is higher than 100% for many compounds and is a of neutral organic chemicals (15).
result of the large variation of Kd,solid and Koc from soil to soil.
The respective average values of RSD of the Kd,solid and the by Karickhoff (15), suggesting a systematic deviation of VP Koc are 75 and 67%, indicating that normalization to foc does sorption behavior from that represented by the regression not result in a considerable reduction of the variability.
The plot of log Koc vs log Kow (Figure 2) demonstrates that Surface Sorption. These considerations prompted us to
revisit the foc Kd,solid relationship reported for the anthelmintic occurs in the narrow hydrophobicity range from 0 to 1.1 log avermectin (23). In that study, the particle size of the sorbents Kow units. This indicates that Koc is not related to hydro- employed decreased concomitantly with increasing organic phobicity for the present VP data set. Moreover, for log Kow carbon content. Hence, two factors leading to increased < 3, all data lie above the classical regression line obtained sorption (foc and sorbent surface area) are varied in favor of VOL. 35, NO. 17, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3403
increased sorption. That means that the sorptive surface areaincreased, as did the sorption coefficient. Therefore, theavermectin data are inconclusive with regard to whetherhydrophobic partitioning or more specific interactions areinvolved in sorption of avermectin.
In contrast, there is ample evidence that sorption of several VPs is a surface-related process. The exchange enthalpy ofthe quinolone carboxylic acid enrofloxacin with different pureclay minerals increases with surface area as measured bymicrocalorimetry (29). Similarly, the enrofloxacin sorptioncoefficients toward different clay minerals increases in theorder kaolinite (nonexpandable, two-layer clay mineral) <illite (nonexpandable, three-layer) < vermiculite = mont-morillonite (both expandable, three layers). This indicatesthat sorption of this compound is a process occurring at the FIGURE 3. Plot of the log Kd,DOM data against hydrophobicity
surface of the clay mineral surfaces. The sorption isotherm expressed as log Kow. The solid line is a regression line obtained
of tetracycline displays Langmuir-type shape, indicating that for a wide range of neutral organic chemicals (32).
sorption occurs at a limited number of sorption sites on thesurface of the clay minerals (30). Sithole and Guy (30) variedthe surface accessible to tetracycline by performing sorption distinguished from cation bridging as reported for enro- experiments with bentonite exchanged with trimethyldode- floxacin (29) cannot be resolved on the basis of presently cylammonium (C12-TMA) ions and with bentonite coated with tannic acid. In the former case, the clay coagulates, Sorption to DOM. The Kd,DOM values for sorption to
resulting in a significantly reduced the surface area and purified humic materials range between 1500 and 2000 L/kg concomitantly a low value of Cs,max. Moreover, the sorption for tetracycline (24) and between 100 and 53 000 L/kg for coefficients for oxolinic acid and oxytetracycline toward sand two series of quinolone carboxylic acid (25, 26). The large are much lower than those found for sediments that contain differences between the two data sets can be explained by a significant portion of silt and clay (28). Hence, sorption of differences regarding the sources of the DOM, the methods tetracycline and quinolone carboxylic acids appears to be applied, and the pH difference. At the pH value of 9.2 strongly related to the particle size of the solids, which in employed by Schmitt-Kopplin et al. (26), all test compounds turn is related to the specific surface.
carry a net negative charge, and the acidic functional groups Interaction Types. Despite the relative hydrophobicity
of DOM are likely to be deprotonated for the largest part imparting the DOM with a significant negative charge. The the least of several bentonite treatments (30). Apparently, resulting electrostatic repulsion might partly account for the the hydrophobic interactions are not effective in counteract- difference in the Kd,DOM data obtained at lower pH (25).
ing the effect of the reduced surface area and thus do not Figure 3 gives an overview of all reported Kd,DOM data by play a major role in tetracycline sorption. X-ray diffraction plotting them against log Kow. One feature of Figure 3 is that analysis of clay minerals showed that sorption of tetracycline no clear relationship exists between log Kow and log Kd,DOM.
and enrofloxacin widened the clay interlayer spacing, In addition, the Kd,DOM values lie above the solid line, which indicating that the interlayers of expanding clays are also log Kd,DOM relationships established for involved in sorption (29, 31). While the dimethylammonium neutral hydrophobic compounds (32). In analogy to Figure group exchanges with inorganic cations at the cation 2, Figure 3 demonstrates that association of VP to DOM is exchange sites at low pH (31), Sithole and Guy (30) found much stronger than predicted from hydrophobic interactions, possibly due to hydrogen bonding as invoked by Sithole et s,max for sorption of tetracycline to the clay mineral bentonite is 2.5 times higher when the cation exchange sites ¨ tzhøft et al. (25). In analogy with the are occupied with Ca2+ instead of Na+ (at pH 6.1). IR spectra sorption to clay minerals, cation bridging is an alternative of tetracycline sorbed to montmorillonite at various pH values explanation that has been put forward to account for the suggest interaction of tetracycline with Ca2+ at the clay association of quinmerac, a quinoline carboxylic acid her- surfaces to be the prevalent sorption mechanism at inter- mediate pH (6.1). FTIR spectroscopy of the enrofloxacin-montmorillonite system demonstrates that the carboxylic Implications for Future Research
moiety undergoes a specific change during sorption that is Conceptual Model. The wide diversity of functional groups
accompanied by a decrease of the pH in the clay suspension, present in VPs bears resemblance to that found among in agreement with deprotonation of the carboxylic acid. In pesticides and their metabolites, suggesting that in analogy addition, Nowara et al. observed that the Kd of the decar- to pesticides many mechanisms are involved in sorption of boxylated derivative is 60 times lower than that of enrofloxacin VPs (34). Hence, instead of merely considering the contribu- itself (29). All this indicates that an interaction of the tion of hydrophobic partitioning to sorption, a conceptually deprotonated carboxylic acid with the clay surface contributes more complete representation of sorption of (ionizable) significantly to sorption of the quinolone carboxylic acid organic chemicals (eq 4; 14) should be adopted. Equation 4 antibiotics. Hence, the high sorption coefficients of the explicitly accounts for different mechanisms to be involved tetracyclines and quinolone carboxylic acids at typical soil pH values appear to be primarily due to interactions ofanionic VP species at the clay surfaces, either the basal planes or the interlayer spaces exposed in expandable clays such as bentonite and montmorillonite. In addition, the formation of surface complexes with Al at the edges of bentonite is discussed (30). The proposed mechanism for the fluoroqui-noloes is cation bridging in the diffuse double layer at the The different summands in the nominator represent clay surfaces (29). As to how far formation of Ca2+ complexes different contributions to the overall sorption. The specific at clay surfaces, as reported for tetracycline (30, 31), can be processes considered are sorption to organic matter, surface 3404 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 17, 2001
adsorption to mineral constituents, ion exchange, and might also compete with soil solids for VP molecules. Given, reactions (such as complexation and H-bonding). They are the considerable DOM concentrations in soil solution [up to represented by the respective subscripts OM, min, ie, and 50 mg/L in agricultural soils (19) and several hundreds of rxn. A stands for the specific surface area available for each milligrams per liter in liquid manure (36)], association to interaction, and C stands for the concentration of the DOM might lead to an increased mobility of VPs in soil and chemical interacting via the respective mode. The surface to low concentrations of the freely dissolved VP species. In concentration of an interaction site on a given sorbent is addition, sorbent properties are expected to differ depending indicated by σ. The denominator of eq 4 considers the on the properties of the soil solution. The surface charge and presence of neutral and ionized species of an organic the cation exchange capacity of soil constituents are de- pendent on the pH of the soil solution. The cation composi- First, eq 4 is instructive in explaining why Koc is successful tion of the soil solution determines the thickness and for compounds devoid of functional groups, such as neutral composition of the diffusive double layer at the solid-water chlorinated solvents, chlorinated aromatic compounds, interface and has been shown to influence the sorption of polycyclic aromatic hydrocarbons, and many pesticides. In tetracycline to bentonite (see discussion above and ref that case, sorption is driven by hydrophobic repulsion from 30) and montmorillonite (31). Taking all this together, it is the solution and COMfOM is the major contributor to the obvious that aqueous chemistry might influence sorption of sorption coefficient Kd. As a result, eq 4 can be approximated VPs to soil and that VP sorption experiments should (COM/Caq)fOM, demonstrating that eq 3 can be seen be performed under adequate control of the solution as a simplification of eq 4 for hydrophobic neutral com- Other Factors and Processes. The sorption coefficients
Second, eq 4 is valuable for rationalizing Figures 2 and observed by Nowara et al. (29) were higher than expected based on the contents of pure clay minerals presumably due Kd,DOM values (log Kow range: -1.2 to 1) are higher than to the presence of phyllosilicates, which also provide large predicted based on the regression models used for hydro- surface areas for sorption al (29). Considering that surface phobic interactions. This suggests that hydrophobicity-driven complexation is rather similar to solution complexation (37), interactions as represented by log Kow do not explain the the high stability constants of tetracyclines and quinolone high sorption of VPs to solids and DOM observed for carboxylic acids with Al3+ and Fe3+ imply that these VP might compounds with relatively low values of log Kow (<2.5).
undergo complex formation at aluminum and iron hydroxide According to eq 4, other processes must then contribute surfaces. While complex formation is not expected to be of significantly to sorption in the low hydrophobicity range.
importance for crystalline aluminum and iron hydroxides One of them is cation exchange, which has been shown to due to their low specific surface area, amorphous hydroxides account for tetracycline sorption to clay minerals at low pH such as coatings on soil solids and the edges of clay minerals (30). Cation bridging of quinolone carboxylic acids (29) can (30) might provide a sufficiently large number of sites be viewed as adsorption to a mineral surface. Hydrogen available for ligand exchange in order to contribute signifi- bonding has been invoked to explain the association to DOM cantly to sorption of these VPs and efrotomycin (38).
(quinolone carboxylic acids and tetracycline) (24, 25) and Formation of bound residues has been demonstrated to occur for a wide variety of pesticides (39). Given the broad overlap For VPs other than tetracyclines and quinolone carboxylic in functional groups between pesticides and VPs, it appears acids, sorption interactions have not been investigated.
likely that VPs become covalently attached to soil matter, Inspection of the structures of these VPs (Table 1) leads to even though this has not been reported for VPs.
identification of four structural entities present in thesecompounds. First, all VPs under study can undergo hydrogenbonding. However, it is generally believed that water Implications for VP Risk Assessment
molecules outcompete organic sorbate molecules at mineralsurfaces in contact with water (14) such that H-bonding is The current risk assessment schemes (5, 6) operate with one not considered to be of major importance. Second, sulfona- single value of Kd,solid. However, the majority of reported mides, metronidazol, olaquindox, efrotomycin, and chloram- isotherms (Table 2) are nonlinear such that Kd,solid decreases phenicol possess highly conjugated moieties that might form with increasing concentration of VPs in the aqueous solution.
charge-transfer complexes with soil constituents (35). Third, Given that the concentrations encountered in the soil the neutral form of weak acids such as the sulfonamides and environment are generally lower than those employed in the efrotomycin can sorb via hydrophobic interactions. The VPs laboratory investigations, it is likely that Kd,solid values in field with intermediate sorption coefficients (efrotomycin, aver- soils are equal or higher than those reported in the literature.
mectin, and tylosin) are large molecules with the highest Moreover, the factors that might influence VP sorption are not considered. Rather, the organic carbon content is the ow (>3). In their case, deviation from predictions only soil property considered when assessing the mobility oc are small and not systematic. This can be explained by the hydrophobic interactions being so pre- of VP, even though the present analysis has shown that it dominant that they overwhelm the contributions of the other explains little if any of the variability in the VP sorption data.
Therefore, the appropriateness of the above schemes should Influence of Soil Solution Chemistry on VP Sorption.
Many VPs undergo pH- and metal ion-dependent speciation Most importantly, however, the use of Koc and thus the in aqueous solution, as indicated in Table 1. This is reflected normalization to the organic carbon content is conceptually in the pH dependence of sorption of tetracycline and inappropriate for VPs. As discussed above, quite a number quinolone carboxylic acids to clay minerals, soil organic of different interaction mechanisms appear to be involved matter, and Aldrich humic acid (24, 25, 30, 31). Increasing in VP sorption. Organic carbon normalization primarily Na+ concentration resulted in decreasing sorption of tet- reduces the variation in the sorption coefficients due to racycline and release of the Ca2+ ion to the soil solution. This hydrophobic interactions (13, 14). Therefore, it cannot explain has been explained by competition between tetracycline variation that is due to non-hydrophobic interactions, and sorption to clay minerals and complexation with Ca2+ ions as a result, it is not successful in reducing the variability in in solution (30). The high values of Kd,DOM for tetracyclines the sorption coefficients. A second consequence is the poor (24) and quinolone carboxylic acids (25) indicate that DOM prediction of log Koc and log Kd,DOM from log Kow.
VOL. 35, NO. 17, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 3405
Acknowledgments
(27) van Loon, G. W.; Duffy, S. J. Environmental Chemistry; Oxford The work was partly funded by the European Union (Grant (28) Pouliquen, H.; Le Bris, H. Chemosphere 1996, 33, 801-815.
ERAVMIS EVK-CT-1999-00003). The author thanks the (29) Nowara, A.; Burhenne, J.; Spiteller, M. J. Agric. Food Chem. 1997,
anonymous reviewers for providing input that greatly (30) Sithole, B. B.; Guy, R. D. Water, Air, Soil Pollut. 1987, 32, 303-
Literature Cited
(31) Porubcan, L. S.; Serna, C. J.; White, J. L.; Hem, S. L. J. Pharm. Sci. 1978, 67, 1081-1087.
(1) Za¨nker, S. FEDESA, personal communication, 2000.
(2) Kay, P.; Boxsall, A. B. Environmental risk assessment of veterinary (32) Burkhard, L. P. Environ. Sci. Technol. 2000, 34, 4663-4668.
medicines in slurry; SSLRC Contract JF 611OZ; Cranfield (33) Deschauer, H.; Ko¨gel-Knabner, I. Sci. Total Environ. 1992, 117/
(3) Alder, A.; McArdell, C. S.; Giger, W.; Golet, M.; Molnar, E.; Nipales, (34) Koskinen, W. C.; Harper, S. S. In Pesticides in the Soil Environ- N. S. Determination of antibiotics in Swiss wastewater and in ment: Processes, impacts and modeling; Cheng, H. H., Ed.; surface water. Presented at Antibiotics in the Environment, (35) Haderlein, S. B.; Schwarzenbach, R. P. Environ. Sci. Technol. (4) Vicari, A.; Landy, R.; Genthner, F.; Morales, R. Antimicrobials 1993, 27, 316-326.
used in animal feedlots: Targeting research on microbial (36) Hsu, J. J.; Lo, S.-L. Water Sci. Technol. 1999, 40, 121-127.
resistance. Presented at the 20th SETAC Meeting, Philadelphia, (37) Stumm, W. Chemistry of the Solid Water Interface; Wiley: New (5) Montforts, M. H. M. M.; Kalf, D. F.; van Vlaardingen, P. L. A.; (38) Tate, R. L., III; Halley, B. A.; Taub, R.; Green-Erwin, M. L.; Lee- Linders, J. B. H. J. Sci. Total Environ. 1999, 225, 119-133.
Chiu, S.-H. J. Agric. Food Chem. 1989, 37, 1165-1169.
(6) Spaepen, K. I.; van Leemput, L. J. J.; Wislocki, P.; Verschueren, (39) Gevao, B.; Semple, K. T.; Jones, K. C. Environ. Pollut. 2000, 108,
C. Environ. Toxicol. Chem. 1997, 16, 1977-1982.
(7) Hamscher, G.; Abu-quare, S.; Sczesny, S.; Hoper, H.; Nau, H.
(40) Wollenberger, L.; Halling-Sorensen, B.; Kusk, K. O. Chemosphere Determination of tetracyclines in soil and water samples from 2000, 40, 723-730.
agricultural areas in lower saxony. Presented at EuroResidueIV, Veldhoven, NL, 2000.
¨tzhoft, H.-C.; Vaes, W.; Freidig, A. P.; Halling-Sorensen, (8) Kolpin, D. W.; Meyer, M. T.; Barber, L. B.; Zaugg, S. D.; Furlong, B.; Hermens, J. L. M. Chemosphere 2000, 40, 711-714.
E. T.; Buxton, H. T. A National Reconnaissance for Antibiotics (42) Mitscher, L. A. The Chemistry of the Tetracycline Antibiotics; and Hormones in Streams of the United States. Presented at SETAC 21st Annual Meeting in North America, Nashville, TN, (43) Duran Meras, I.; Munoz de la Pena, A.; Salinas Lopez, F.; Rodriguez Caceres, M. I. Analyst 2000, 125, 1471-1476.
(9) EMEA. Note for guidance: Environmental risk assessment for (44) Ghandour, M. A.; Azab, H. A.; Hassan, A.; Ali, A. M. Monatsh. veterinary medical products other than GMO-containing and Chem. 1992, 123, 51-58.
immunological products; EMEA/CVMP/055/96; European Agency (45) Tongaree, S.; Goldberg, A. M.; Flanagan, D. R.; Poust, R. I. Pharm. for Evaluation of Medicinal Products: 1996.
Dev. Technol. 2000, 5, 189-199.
(10) Hirsch, R.; Ternes, T. A.; Haberer, K.; Kratz, K.-L. Sci. Total (46) Kaplan, L.; Pink, D. W.; Fink, H. C. Anal. Chem. 1984, 56, 360-
Environ. 1999, 225, 109-118.
(11) OECD. OECD guideline for the testing of chemicalssleaching in (47) Hassan, S. S. M.; Amer, M. M.; Ahmed, S. A. Mikrochim. Acta soil columns; Organization for Economic Cooperation and 1985, 3, 165-175.
(48) Meylan, W. SRC-LOGKOW for Windows, v1.53a ed.; SRC: (12) OECD. Adsorption-Desorption Using a Batch Equilibrium Method; Technical Guideline 106; Organization for Economic (49) Takacs-Novak, K.; Jozan, M.; Hermecz, I.; Szazz, G. Int. J. Pharm. 1992, 79, 89-96.
(13) Chiou, C. T. In Reactions and Movement of organic chemicals in soils; Sawhney, B. L., Brown, K., Eds.; Soil Science Society of (50) Budavari, S. The Merck Index, 11th ed.; Merck & Co: Rahway, America: Madison, WI, 1989; pp 1-30.
(14) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M.
(51) Lin, C. E.; Lin, W.-C.; Chen, Y. C.; Wang, S.-W. J. Chromatogr. Environmental Organic Chemistry; Wiley: New York, 1993.
A 1997, 792, 37-47.
(15) Karickhoff, S. W. Chemosphere 1981, 10, 833-836.
(52) Drakopoulos, A. I.; Iannou, P. C. Anal. Chim. Acta 1997, 354,
(16) van Leeuwen, C. J.; Hermens, J. L. M. Risk Assessment of Chemicals: An Introduction. Kluwer: Dordrecht, 1995; p 374.
(53) Papastephanou, C.; Frantz, M. Anal. Profiles Drug Subst. 1978,
(17) McCarthy, J. F.; Zachara, J. M. Environ. Sci. Technol. 1989, 23,
(54) Szulczewski, D.; Eng, F. Anal. Profiles Drug Subst. 1975, 4, 47-
(18) Magee, B. R.; Lion, R. W.; Lemley, A. T. Environ. Sci. Technol. 1991, 25, 323-331.
(55) Timmers, K.; Sternglanz, R. Bioinorg. Chem. 1978, 9, 145-155.
(19) Maxin, C. R.; Kgel-Knabner, I. Eur. J. Soil Sci. 1995, 46, 193-204.
(56) Ross, D. L.; Riley, C. M. Int. J. Pharm. 1993, 93, 121-129.
(20) Ross, D. L.; Riley, C. M. Int. J. Pharm. 1993, 87, 203-213.
(57) Thurman, E. M.; Lindsey, M. E. Transport of antibiotics in soil (21) Lai, H.-T.; Liu, S.-H.; Chien, Y.-H. J. Environ. Sci. Health. A 1995,
and their potential for groundwater contamination. Presented at 3rd SETAC World Congress, Brighton, UK, May 22-25, 2000.
(22) Rabolle, M.; Spliid, N. H. Chemosphere 2000, 40, 715-722.
(58) Yeager, R. L.; Halley, B. A. J. Agric. Food Chem. 1990, 38, 883-
(23) Gruber, V. F.; Halley, B. A.; Hwang, S. C.; Ku, C. C. J. Agric. Food Chem. 1990, 38, 886-890.
(59) Langhammer, J.-P. Ph.D. Thesis, Rheinische Friedrich-Wilhelms- (24) Sithole, B. B.; Guy, R. D. Water, Air, Soil Pollut. 1987, 32, 315-
¨ tzhøft, H. C.; Vaes, W. H. J.; Freidig, A. P.; Halling- Sørensen, B.; Hermens, J. L. M. Environ. Sci. Technol. 2000, 34,
Received for review December 18, 2000. Revised manuscript received June 5, 2001. Accepted June 18, 2001. (26) Schmitt-Kopplin, P.; Burhenne, J.; Freitag, D.; Spiteller, M.; Kettrup, A. J. Chromatogr. A 1999, 837, 253-265.
3406 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 17, 2001

Source: ftp://ftp.mondoinfo.com/disk3/neurotic/troll-vetpharm-in-soils.pdf