Environ. Sci. Technol. 2008, 42, 8270–8276
Probe Compounds to Quantify Cation Exchange and Complexation Interactions of Ciprofloxacin with Soils ALLISON A. MACKAY* AND DANIEL E. SEREMET Environmental Engineering Program, University of Connecticut, 261 Glenbrook Road, Storrs, Connecticut 06269-2037
Received April 14, 2008. Revised manuscript received August 28, 2008. Accepted August 29, 2008.
Ciprofloxacin (CIP)-soil sorption interactions by surface complexation (via -COOH group) and cation exchange (-NH3+ group) were estimated by use of the structurally related probe compounds flumequine (FQ) (-COOH) and phenylpiperazine (PP) (-NH3+). Comparison of CIP and FQ sorption by surface complexation on goethite indicated a 0.7 times lower driving force for sorption, Kxs, for CIP than for FQ, with a model that assumed functional group interactions were enhanced by the hydrophobicity of the rest of the molecule. Similarly, Kxs was 9.5 times greater for CIP than for PP for sorption by cation exchange on kaolinite and montmorillonite. Use of the pure phase sorbent Kxs scaling factors between PP, FQ, and CIP for eight soils with a range of clay and oxide contents yielded total sorbed CIP concentrations that were within a factor of 2 (at pH 7.2) or less (at pH 5) of observed values. The estimated relative contributions of CIP cation-exchange versus complexation interactions increased for soils with increasing cationexchange capacity. The agreement between independently estimated and observed CIP sorption indicates that the use of probe compounds has the potential to provide quantitative measures of sorption contributions for other complex sorbates with multiple functional groups, including other emerging contaminants and pesticides.
are complicated further by the fact that many compounds have more than one polar functional group and consequently may participate in multiple types of sorption interactions, each with different matrices of an aggregated environmental solid. Thus, there is a need to develop quantitative approaches for evaluating the various sorption interactions that a complex sorbate with more than one polar functional group may have with real environmental solids to improve fate predictions and environmental management decisions for such emerging contaminants. To provide an example of the types of sorption interactions that a complex sorbate with more than one functional group may have with environmental solids, we consider the highuse veterinary fluoroquinolone antibiotic ciprofloxacin (CIP) (Figure 1). Sorption studies with pure phase minerals have demonstrated that the carboxylic acid group participates in sorption interactions with amorphous (9) and crystalline (10, 11) iron oxides. Additional complexation interactions through the carboxylic acid group may be anticipated with aluminum oxides (9) and aluminosilicate clay edges. The piperazinyl amine, when protonated, enables fluoroquinolone compound exchange with clay interlayer cations (12). Sorption of fluoroquinolone compounds to clay cations via bridging by the carboxylic acid group has also been suggested (12). Presumably, this interaction may also occur between CIP and cations sorbed by other matrices, such as organic
Introduction Sorption predictions for the ever-growing list of emerging contaminants present challenges for environmental engineers and scientists who are trying to assess exposure risks for these compounds. Many of the organic contaminants termed to be emergingspharmaceuticals, hormones, and other compounds from personal care products or wastewater originscontain polar ionic groups that facilitate electrostatic sorption interactions with the mineral fractions of environmental solids (1-4). Polar organic compound sorption to mineral phases has been described quantitatively for pure phase sorbents (5, 6), yet these models cannot describe sorption interactions with real environmental matrices because of differences in sorbent properties for environmental solids (7) and because the aggregate structure of environmental solids results in site blocking by material coatings (8). Sorption estimates for emerging contaminants * Corresponding author phone: (860) 486-2450; fax: (860) 4862298; e-mail:
[email protected]. 8270
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 22, 2008
FIGURE 1. Structure and chemistry of ciprofloxacin (CIP) and the probe compounds flumequine (FQ) and phenylpiperazine (PP). Bars show the charge of the dominant species in solution as a function of pH with pKa values (17, 18) of acidic groups noted. (() Zwitterion; (0) neutral compound. 10.1021/es800984x CCC: $40.75
2008 American Chemical Society
Published on Web 10/22/2008
TABLE 1. Sorbent Properties and Modeling Parameters parameter
Adams
Berryland
Burton
Goldsboro
Heiden
Iredell
Kleinpeter
Orangeburg
Kxs (PP) KxsKrxn (FQ) σA (mol/kg) ECECa (mmol/kg) DCB Fe + Alb (mmol/kg) organic carbon (%)
4.5 0.16 0.0085 10 900 2.5
5.7 0.014 0.0052 8 8 1.7
6.3 0.022 0.022 86 850 8.9
13 0.004 0.009 18 120 0.3
7.7 0.0004 0.05 440 120 1.3
48 0.002 0.048 210 450 0.46
26 0.003 0.013 53 100 1.9
14 0.022 0.009 39 550 0.23
CsIE d CsSC e est Cstot obs Cs
7.9 8.5 16 16
7.7 5 13 14
pH 5:c Sorbed CIP (mmol/kg) 19 19 34 21 7.2 4.7 40 26 39 58 20 56
56 2.8 59 60
30 8.9 39 41
25 7.5 32 31
CsIE a CsSC b est Cstot obs Cs
4.6 8.3 13 32
2.7 4.3 7 17
pH 7.2: Sorbed CIP (mmol/kg) 13 6.2 19 19 4.4 2.7 32 11 22 58 21 55
52 2.5 55 59
18 5.5 24 36
12 7.8 20 30
CsIE a CsSC b est Cstot obs Cs
0.3 5.8 6 27
0.2 1.1 1.3 8
pH 8.5: Sorbed CIP (mmol/kg) 2.6 0.4 1 10 0.5 0.2 13 0.9 1.2 53 10 48
18 1.3 19 51
1.3 0.8 2.1 22
f f f
a Effective cation-exchange capacity. b Dithionite-citrate-bicarbonate extractable (crystalline) iron and aluminum. c Values for Heiden and Orangeburg reported at pH 6. d CIP sorbed by cation exchange, calculated with Kxs(CIP) ) (10-1.63 × -0.6)Kxs(PP). e CIP sorbed by complexation, calculated with KxsKrxn(CIP) ) (100.82 × -0.2)KxsKrxn(FQ). f CEC not available.
matter (13). Which of these sorption interactions observed for single-sorbent systems will be of importance for environmental solids where oxide, clay, and organic matter components are aggregated together has been explored only with statistical tools. Principal component analysis indicated that CIP soil sorption coefficients were highly correlated with soil cation-exchange capacity (cation-exchange and/or cationbridging interactions), although a contribution from complexation for high iron and aluminum oxide content soils was suggested at pH g 7 (14). While these pure phase sorbent and soil studies provide general guidelines for qualitative assessments of CIP sorption to environmental solids, there exists presently no means by which to apportion quantitatively the fraction of overall CIP sorption attributable to specific interaction mechanisms for a given solid. We propose that one approach for differentiating between sorption interaction mechanisms for complex organic sorbate molecules with more than one functional group, such as ciprofloxacin, is through observations of sorption for simpler probe compounds with only one functional group. In the case of CIP, structurally similar probe compounds (Figure 1) of similar size are readily available to evaluate this concept. We anticipated that sorption measurements of the positively charged organic probe compound, phenylpiperazine (PP) (Figure 1), on an environmental solid could be used to discern cation-exchange interactions of CIP with that solid. Similarly, the organic ligand flumequine (FQ) (Figure 1) could indicate CIP complexation interactions with solid-phase or sorbed cations. Such an approach of using probe compounds with one functional group is feasible because observations of charged organic compound sorption to mineral surfaces indicate that analogue compounds with the same functional group exhibit systematic changes in the free energy of sorption with increasing molecular size (15). Quantitative models have been set forth to relate these free energy changes to compound structure (16), thereby providing a framework with which to “scale” measures of probe compound sorption to a solid accordingly to yield an estimate of the fraction of overall sorption of the complex sorbate attributable to the mechanism specific to that probe. By obtaining measurements of probe compound sorption (and scaling appropri-
ately for the complex sorbate structure), the inherent shortcoming of unknown site availabilities in environmental solids is overcome because the probe compound sorption measure implicitly captures the abundance of sites relevant to the type of interaction that the probe functional group undergoes. The purpose of this research was to determine whether structurally related probe compounds (PP, FQ) can provide quantitative sorption estimates for a complex sorbate with multiple functional groups (CIP). CIP and probe compound sorption was first examined as a function of pH (sorption edges) and of sorbate concentration (isotherms) by use of pure phase sorbents to constrain CIP sorption to one interaction mechanism. Consequently, results from these studies were used to evaluate whether existing quantitative models (16) could be used to “scale” probe compound sorption to describe CIP sorption via that mechanism. Finally, sorption of CIP and the two probe compounds was measured on eight environmental solids, soils chosen to have a range of oxide, aluminosilicate clay, and organic matter contents. The quantitative model was applied, using measures of probe compound sorption on soils, to estimate the fraction of overall CIP sorption attributable to complexation and cationexchange interactions with soil components.
Methods Chemicals. Chemical suppliers are provided as Supporting Information. Sorbents. Synthetic goethite (Bayferrox 930, Bayer) had 5.5 sites/nm2 and a zero-point charge of 7.85 (19). Kaolinite (China clay; Sargent-Welch, Buffalo Grove, IL) and montmorillonite (K-10, acid-washed; Aldrich) clays were saturated with sodium by washing twice with 1 M sodium chloride, followed by at least two washes with high-purity water. Soil samples (Table 1) with a range in aluminum and iron oxide content, cation-exchange capacity, and organic matter were collected from agricultural regions of the eastern United States and characterized previously (20). Sorption Edge Experiments. Compound sorption was measured as a function of pH at 25 ( 1 °C with one sorbate per tube to prevent competition for sorption sites. Tube VOL. 42, NO. 22, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8271
preparation methods were the same for all sorbate-sorbent pairs, except where noted in parentheses. Duplicate polypropylene tubes were prepared by prewetting solids in PIPES [1,4-piperazinebis(ethanesulfonic acid)] or carbonate buffer (clays) for 24 h prior to sorbate addition. The final solidto-water ratios (kilograms per liter) were 0.02 (CIP-geothite), 0.005 (kaolinite, FQ-goethite), 0.0005 (montmorillonite), or 0.0025 (soils). Solution pH was adjusted to a value between 4 and 9 by adding small aliquots of concentrated hydrochloric acid or sodium hydroxide. Sorbates were added from stock solutions to give initial concentrations of 0.13 mmol/L for CIP and PP and 0.074 mmol/L for FQ in all experiments. Tubes containing soil received azide biocide (1.5 mM) to prevent compound biodegradation. Sorbent-free control tubes were prepared at each pH to account for sorptive losses to tube walls or thermal degradation. Tubes were wrapped in foil to prevent photodegradation and tumbled end-overend in the dark for 72 h (goethite, soil) or 24 h (clays), as determined with prior kinetic studies (3, 14). At the end of the equilibration time, supernatant was separated from tubes by filtration through 0.2 µm poly(vinylidene fluoride) (PVDF) syringe filters or centrifugation at 12500g for 60 min (clays, soils). Final pH was measured in the supernatant. An aliquot of supernatant was analyzed by HPLC (see Supporting Information for analytical methods) to quantify the equilibrium aqueous sorbate concentration. Sorbed concentrations were calculated by difference, because control tube losses were below detection (
