Environ. Sci. Technol. 2007, 41, 3153-3158
Spectroscopic Investigation of Ciprofloxacin Speciation at the Goethite-Water Interface PARAS TRIVEDI,† AND DHARNI VASUDEVAN* Department of Chemistry, 6600 College Station, Bowdoin College, Brunswick, Maine 04011
We investigated ciprofloxacin (a fluoroquinolone antibiotic) speciation as a function of pH in aqueous solution and in the presence of dissolved ferric ions and goethite using ATR-FTIR and UV-vis spectroscopy. The presence of dissolved and surface bound ferric species induced the deprotonation of the ciprofloxacin carboxylic acid group at pH < pKa1. The resultant ciprofloxacin zwitterions appeared to interact via both carboxylate oxygens to form bidentate chelate and bridging bidentate complexes within colloidal iron oxide-ciprofloxacin precipitates and bidentate chelates on the goethite surface. However, the structure of the aqueous ferric-ciprofloxacin complexes remains unclear. Our evidence for bidentate chelates (involving only the carboxylate oxygens) on the goethite surface was distinct from previous IR studies of fluoroquinolone sorption to metal oxides that have proposed surface complexes involving both the keto and the carboxylate groups. We find that the distinct ciprofloxacin surface complex proposed at the goethite-water interface may be a result of differences in metal oxide mineralogy or assignment of the carboxylate antisymmetric stretch in the metal oxide-fluoroquinolone spectra.
aluminum oxides (4, 7), and amorphous (4) and crystalline iron oxides (8). On the basis of ATR-FTIR analyses between pH 5.5 and 7.2, the ofloxacin zwitterions are assumed to sorb onto alumina via a pH-independent ligand exchange mechanism involving the ketone and carboxylate groups (7). Likewise, Gu and Karthikeyan (4) proposed from their ATRFTIR studies between pH 5 and 9 that ciprofloxacin zwitterions sorbed as monodentate mononuclear complexes onto hydrous aluminum oxide (HAO) through the involvement of one carboxylate oxygen along with a weak hydrogen bonding association of the keto group and as mononuclear bidentate complexes onto hydrous ferric oxide (HFO) via the involvement of one carboxylate oxygen and the keto group. Furthermore, Gu and Karthikeyan (4) pointed to a similarity between the IR spectra of these ciprofloxacin-HFO surface complexes and the IR spectra of a 1:1 ferric-ciprofloxacin dissolved complex at pH 3. However, the potential dissolution of these metastable amorphous oxides constrained the pH range of these studies to pH 5-9. In contrast, crystalline oxides, such as goethite, that are also environmentally ubiquitous are less susceptible to ligand promoted dissolution than HFO (9). Hence, spectroscopic investigation of ciprofloxacin sorption onto goethite, the subject of this study, will foster additional understanding of fluoroquinolone interactions at the mineral-water interface. This research is focused on a systematic elucidation of ciprofloxacin speciation in the presence of aqueous ferric species and goethite, between pH 2 and 9, through a combination of attenuated total reflectance Fourier transform infrared (ATR-FTIR) and ultraviolet-visible (UV-vis) spectroscopy. Furthermore, these studies of the ferric-ciprofloxacin and goethite-ciprofloxacin complexes will build on previous studies and provide a broader understanding of the impact of ciprofloxacin on ferric ion-bearing minerals present in most aquatic and soil environments. The results of these investigations will aid in accurate risk assessment and management of these antibiotics in soil and aquatic environments.
Materials and Methods Introduction Fluoroquinolones, such as ciprofloxacin, have been reported to occur in aquatic and soil environments at concentrations that often exceed the suggested minimum inhibitory levels for various indicator species including many bacterial species (1-3). Such increased antibiotic concentrations in the environment can cause chromosomal mutation of the existing bacteria resulting in the generation of fluoroquinolone-resistant strains (1). Hence, significant concern accompanies the successful application of effective human and veterinary antibiotics. Additionally, these antibiotics can potentially induce physicochemical changes in high surface area soil components, such as metal oxides/oxyhydroxides and clay minerals and thus significantly alter their capability to control the mobility and bioavailability of other contaminants and micronutrients (4, 5). Current risk assessment models and waste management programs lack in mechanistic and thermodynamic information on the fate of ciprofloxacin in the presence and absence of these important soil components. Previous batch studies have revealed the pH-dependent sorption of fluoroquinolone onto aluminosilicate clays (6), * Corresponding author phone: (207) 725-3548; fax: (207) 7253017; e-mail:
[email protected]. † Current address: CH2M HILL, Parsippany, NJ. 10.1021/es061921y CCC: $37.00 Published on Web 03/24/2007
2007 American Chemical Society
Ciprofloxacin hydrochloride was supplied by the Bayer Corporation; the chemical properties of ciprofloxacin are discussed elsewhere (10). The preparation and characterization of goethite employed in the sorption studies has been detailed previously (11). All other chemicals [FeCl3‚6H2O, Fe(NO3)3‚9H2O, NaCl, HCl, NaOH, fluorochloroquinoline carboxylic acid (FCQA), 8-hydroxyquinoline, 4-quinolone, nicotinic acid, and methanol] were purchased from Fisher Scientific or Aldrich at 99.9% or greater purity. All solutions were prepared with deionized water, and the ionic strength was adjusted to approximately 0.01 M using NaCl. All ATRFTIR studies were conducted in triplicate; an average of 512 scans was collected for each sample using a Thermo-Nicolet Nexus 470 FT-IR spectrometer equipped with a ZnSe crystal fitted in a horizontal attenuated total reflectance (HATR) cell (Pike Technologies). Aqueous Speciation of Ciprofloxacin. ATR-FTIR scans were collected on individual 0.5 mM filtered ciprofloxacin solutions with the background electrolyte (0.01 M NaCl) set to specific pH values (between pH 2 ( 0.1 and 9 ( 0.1) with the addition of HCl or NaOH. Vibrations of the dissolved ciprofloxacin molecules (difference spectra) at each pH were isolated by subtracting the spectra of the ciprofloxacin-free background electrolyte solution from the spectra of the ciprofloxacin solution. Complementary UV-vis analyses were conducted on 20 µM ciprofloxacin solutions over the same VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3153
pH range, using an Agilent 8453 UV-vis spectrophotometer. To assist in the IR peak assignments of protonated ciprofloxacin, difference spectra were similarly collected for ciprofloxacin and structurally related compounds [nicotinic acid, FCQA, 8-hydroxyquinoline, and 4-quinolone] dissolved in methanol. Ferric-Ciprofloxacin Complexes. ATR-FTIR scans were collected on equimolar solutions of ciprofloxacin and FeCl3 (final concentrations 0.5 mM each) that were equilibrated for 4 h at a constant pH (between pH 3 ( 0.1 and 9 ( 0.1). The difference spectra obtained by subtracting ciprofloxacinfree ferric spectra from the corresponding ferric-ciprofloxacin spectra provided the vibrations of the ferriccomplexed ciprofloxacin. Preliminary studies showed that the ferric-ciprofloxacin complexes remained dissolved at pH e 6 but formed colloidal precipitates at pH g 6. Thus, for pH g 6, the unfiltered ferric-ciprofloxacin solutions represent ciprofloxacin-iron oxide precipitates rather than aqueous complexes. For UV-vis analyses, 20 µM ciprofloxacin-20 µM ferric mixtures were individually prepared between pH 3 ( 0.1 and 6 ( 0.1 and filtered using a 0.2 µm filter prior to analyses. Ciprofloxacin-Goethite Complexes. Approximately 1 mL of a 4 g/L goethite suspension was evenly spread on the ZnSe crystal surface of a flow-through HATR-FTIR cell and oven-dried at 60 °C to form a uniform oxide coating. First, a 0.01 M NaCl solution at constant pH was input through the cell at a rate of 1 mL/min for an hour to ensure hydration of the goethite coating. The flow was stopped to collect IR spectra of this background electrolyte. Subsequently, a solution of 20 µM ciprofloxacin in 0.01 M NaCl, at the same pH, was input through the cell at 1 mL/min. Equilibrium pH values were continuously monitored for the duration of the flow-through experiment. Periodically, the flow was stopped to collect the IR spectra; the process was continued until there was no change in the relative intensities of the difference spectra. Preliminary studies showed that the 20 µM ciprofloxacin influent falls below the detection limit of our FTIR; therefore, the above mentioned difference IR spectra was solely attributed to ciprofloxacin sorbed onto the goethite surface. For systematic accounting of the ciprofloxacin speciation, IR peak intensities of each spectrum were normalized with the relative intensity of the peak observed at 1485 cm-1 (assigned to the aromatic ring stretch) of the same spectrum since the position of this peak did not change with pH.
Results and Discussion Aqueous Speciation of Ciprofloxacin. The ATR-FTIR difference spectra of ciprofloxacin and structurally related compounds dissolved in methanol (Figure 1) aided in the peak assignments of the keto and carboxylic acid moieties of ciprofloxacin. The spectra of all compounds containing a COOH group possessed a peak at ∼1720 cm-1 (Figure 1). However, this peak was absent in the 8-hydroxyquinoline and 4-quinolone spectra, as these compounds lack a COOH group. Therefore, the peak at 1718 cm-1 in the ciprofloxacin spectra (Figure 1) was assigned as the carboxylic acid CdO stretch (νCdOcarboxyl) (12). The peak in the 1640-1600 cm-1 region was observed only in compounds containing a keto carbonyl moiety (FCQA, ciprofloxacin, and 4-quinolone) and was therefore assigned to the ketone CdO stretch (νCdOketone) (12), with the split in the νCdOketone peak in the 4-quinolone spectra attributed to Fermi resonance (13). The spectral features in the 1300-1260 cm-1 region of all COOH containing compounds were assigned to coupled carboxylic acid C-O stretch (νCOOH) and O-H deformation (δC-OH) (14). Notably, these spectral features were also present in the 8-hydroxyquinoline spectra, although this compound lacks a COOH group but possesses a phenolic OH group. In keeping with previous studies (14), the peak at 1270 cm-1 in the 8-hy3154
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
FIGURE 1. ATR-FTIR difference spectra of nicotinic acid, fluorochloroquinoline carboxylic acid (FCQA), ciprofloxacin, 8-hydroxyquinoline, and 4-quinolone dissolved in methanol (normalized with the relative intensity of the peak observed at 1485 cm-1). roxyquinoline spectra was attributed to the phenolic C-OH stretch (νC-OHphenol). The ATR-FTIR spectra of aqueous ciprofloxacin (Figure 2a) at low pH (pH 2-4) were similar to the ciprofloxacin spectra in methanol with a few important exceptions: (i) the νCdOcarboxyl stretch appeared as a broad band centered at 1710 cm-1 in aqueous solution instead of a sharp peak at 1718 cm-1 observed in methanol and (ii) the νCdOketone stretch was broader in aqueous solution and shifted to lower frequency (1618-1622 cm-1) in contrast to a defined peak at 1630 in methanol. The differences in methanol and low pH aqueous spectra can be ascribed to the effect of differences in solvation (13). Interestingly, previous studies have also noted a broad νCdOcarboxyl stretch of lower intensity than the coupled νCOOH and δC-OH modes in aqueous solution spectra of compounds with a COOH group involved in intra- and intermolecular H-bonding (14). Our peak assignments and analysis of speciation within 1800-1350 cm-1 for the aqueous ciprofloxacin spectra corroborated with previous studies of fluoroquinolones (4, 7). With an increase in pH from 2 to 5, the νCdOcarboxyl (1710 cm-1) peak significantly weakened in intensity and was undetectable at pH g 6. Concurrently, two bands around 1580 ( 2 and 1380 ( 2 cm-1, associated with the antisymmetric (νCOOas) and symmetric (νCOOs) stretches of the carboxylate group, respectively, strengthened in intensity with an increase in pH. Between pH 6-9, the νCOOs peak was shifted to 1380 cm-1 and exhibited a significant increase in intensity as compared to the νCOOs peak at 1390 cm-1 in the pH 2-5 spectra. Similarly, the νCOOas stretch also increased in intensity and exhibited bifurcations possibly due to solvation effects. In addition, the complementary UV-vis spectra of ciprofloxacin (Figure 2b) exhibited a hypsochromic shift (from 277 to 272 nm) between pH 5 and 6 characteristic of the deprotonation of carboxylic acid groups (15, 16). These observations were consistent with the pKa1 of 6.10 ( 0.5 for the ciprofloxacin carboxylic acid group (17-19) and with assignments for νCdOcarboxyl, νCdOketone, νCOOas, and νCOOs peaks noted in previous IR studies of ciprofloxacin (4) and ofloxacin (7).
With increased deprotonation of the COOH groups (pH >5) and loss of the νCdOcarboxyl stretch, changes were observed in the intensity of coupled COOH modes (νCOOH and δC-OH) at 1272 cm-1 and in the shape of the νCdOketone peak. As expected, the coupled COOH modes at 1272 cm-1 showed a significant decrease in intensity with the loss of a proton from the COOH group. Concurrent sharpening of the νCd Oketone peak was likely a result of the loss of intramolecular hydrogen bonding between the keto and the carboxylic acid groups at pH > pKa1 and the resulting solvation effects on the Fermi resonance (13). Distinct spectral features observed to increase in intensity with an increase in pH in the 13011290 cm-1 range may be attributed to NH2+ twisting and COO- scissoring (20). Ferric-Ciprofloxacin Complexes at pH < 6. The UVvis spectra of ferric-ciprofloxacin solutions at pH < 6 (Figure 3a), indicated the deprotonation of the COOH group at pH < pKa1: the peak at 277 nm ascribed to the carboxylic acid group (Figure 2b) was absent whereas the peak at 272 nm ascribed to the carboxylate group (Figure 2b) was present. Similar evidence from ATR-FTIR spectra for deprotonation of the COOH group at pH < pKa1 was less obvious (Figure 3b), possibly due to the presence of both free and Fe(III)complexed ciprofloxacin. The 1350-1200 and 1370-90 cm-1 regions of the ferric-ciprofloxacin spectra were similar to aqueous ciprofloxacin spectra at pH < pKa1 (Figures 2a and 3b), suggesting the presence of the COOH group, whereas the presence of the COO- group was indicated by the absence of the 1710 cm-1 band ascribed to νCdOcarboxyl. Furthermore, we observed a high level of noise in these spectra due to the low solubility of Fe(III) and the resulting low concentrations of ferric-ciprofloxacin complexes. The presence of multiple species and the higher level of noise excluded unambiguous peak assignments; hence, we offer a cautious interpretation of the IR spectra for aqueous ferric-ciprofloxacin complexes at pH < 6 (Figure 3b). In the presence of Fe(III), the νCdOketone stretch was observed at 1635 cm-1; the ∼10 cm-1 shift to higher wavenumbers in the presence of Fe(III) suggested the strengthening of the CdO bond due to the electron withdrawing coordination between the ortho substituted carboxylate group and Fe(III) (13). Despite this indirect evidence for ferric-ciprofloxacin interactions, we were unable to hypothesize the structure of the aqueous ferric-ciprofloxacin complex at pH < 6 because we were unable to assign νCOOas with confidence. It is likely that the νCOOas mode was located within poorly distinguishable bands in the 1580-1520 cm-1 region or alternately shifted to higher wavenumbers appearing as a shoulder observed in the 1640-1650 cm-1 region. Notably, the shoulder in the 1640-1650 cm-1 region was present as a minor feature in aqueous ciprofloxacin spectra at all pH values but was prominent in noisy ferricciprofloxacin spectra (pH < 6). The likely presence of both Fe(III)-complexed and -uncomplexed ciprofloxacin and the similar intensity of the bands in the 1580-1520 and 16501640 cm-1 regions in the noisy spectra precluded our reliable assignment of the νCOOas stretch. Similar uncertainty in the assignment of νCOOas was also noted in a study of ferricbenzoate complexes (14). Therefore, the structure of the aqueous ferric-ciprofloxacin complexes remains unclear.
FIGURE 2. (a) ATR-FTIR difference spectra of a 0.5 mM aqueous ciprofloxacin solution as a function of pH (normalized with the relative intensity of the peak observed at 1485 cm-1) and (b) UV-vis spectra of a 20 µM ciprofloxacin solution as a function of pH. All studies were conducted at 25 °C with an ionic strength of 0.01 M (NaCl).
Ferric-Ciprofloxacin Complexes at pH g 6. At pH g 6, a significant fraction of the ferric-ciprofloxacin complexes precipitated out of the solution as colloidal material, and the IR spectra of the filtered solutions indicated the absence of dissolved ciprofloxacin. Thus, the spectra of the unfiltered colloidal suspension at pH > 6 (Figure 3b) represent ciprofloxacin-iron oxide precipitates and provide important insight into ciprofloxacin speciation at the iron oxide-water interface. VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3155
FIGURE 4. Types of interactions between a carboxylate group and a metal ion.
FIGURE 3. (a) UV-vis spectra of aqueous ferric-ciprofloxacin complexes between pH 2 and 6 and (b) ATR-FTIR difference spectra (normalized with the relative intensity of the peak observed at 1485 cm-1) of ferric-ciprofloxacin complexes between pH 2 and 9; pH 3-5 spectra represent aqueous Fe(III)-ciprofloxacin complexes, and pH 6-9 spectra represent colloidal Fe(III)-ciprofloxacin complexes. All studies were conducted at 25 °C with an ionic strength of 0.01 M (NaCl). 3156
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
As expected at pH > 6, the carboxylate anion was the dominant species as evidenced by the absence of the νCd Ocarboxyl stretch and the diminished intensity of the coupled νCOOH and δC-OH modes at 1272 cm-1. The νCdOketone stretch was shifted to 1635 cm-1 as compared with the corresponding aqueous ciprofloxacin spectra (where the peak occurred at 1618 cm-1) suggesting, as previously noted, a stronger CdO bond resulting from the electron withdrawing nature of the iron-carboxylate bond. Additionally, a shoulder at 1620 cm-1 was noted in the νCdOketone stretch, and a broader νCOOs stretch of lower intensity occurred at the same position (1380 cm-1) as in the aqueous ciprofloxacin spectra. Here, assignment of the νCOOas stretch was key to identifying the type of metal-ligand complex formed. For reasons outlined next, we have assigned νCOOas to the peaks at 1560 and 1575 cm-1 as opposed to the shoulder at 1620 cm-1 (Figure 3b). Typically, the carboxylate group can exist as an uncoordinated anion (ionic) or form monodentate (one carboxylate oxygen bonded to a metal ion), bidentate chelate (both oxygens bonded to a single metal ion), or bridging bidentate (each oxygen bonded to different metal ions) complexes (Figure 4). Nakamoto (21) introduced the use of the difference (∆ν) between the carboxylic antisymmetric and symmetric stretch (∆ν ) νCOOas - νCOOs) to identify the binding mode of acetate groups in pure metal acetates. Despite criticisms of the ∆ν value as a diagnostic tool, this approach provided a meaningful guide to determine the binding modes of acetate-metal complexes in aqueous solution (22). Nakamota (21) and Tackett (22) suggested that a ∆ν of the metal-carboxylate complex (∆νMe-COO) greater than the ∆ν of the free carboxylate complex (∆νfreeCOO) indicates a monodentate complex, a ∆νMe-COO < ∆νfreeCOO bidentate chelate, and a ∆νMe-COO = ∆νfree COO bridging bidentate complex. In addition, the value of ∆ν is known to be influenced by the angle of the O-C-O bonds, polarity of the Me-O bond, cation size, and differences in the two C-O bond lengths (22, 23). A decrease in the O-C-O bond angle is known to decrease the ∆ν value (and the extent of asymmetry), whereas an increase in the difference in the two C-O bond lengths increases the asymmetry and the ∆ν value (23). The position of νCOOs was not significantly different in the aqueous ciprofloxacin and ferric-ciprofloxacin spectra. With νCOOs at 1380 cm-1, ascribing the νCOOas to the shoulder occurring at 1620 cm-1 would result in ∆νMe-COO > ∆νfreeCOO, indicative of a monodentate complex. However, monodentate complexes examined experimentally and via ab initio calculations exhibited a 50-100 cm-1 shift (relative to the uncomplexed carboxylate group) to lower wavenumbers for the νCOOs stretch and a 30-80 cm-1 shift to higher wavenumbers for the νCOOas stretch (22, 23). In addition, the two C-O bonds are of distinctly different lengths in a monodentate complex; the free C-O is short and has double bond character, whereas the metal bound C-O bond is long and has single bond character (23). Here, νCOOs was not shifted in relation to the aqueous ciprofloxacin, and double bond character of the free C-O bond was not evident. Hence, we excluded the possibility of a monodentate complex and the shoulder at 1620 cm-1 as representing νCOOas and instead assigned νCOOas to the peaks at 1560 and 1575 cm-1. These assignments corresponded to ∆νiron-ciprofloxacin < ∆νciprofloxacin
TABLE 1. Comparison of the Carboxylate IR Frequencies for Aqueous and Ferric-Complexed Ciprofloxacin Aqueous ciprofloxacin pH 3
νCOOas 1578
νCOOs 1380
∆νciprofloxacin
Fe(III)-ciprofloxacin complexes νCOOas
νCOOs
∆νiron-ciprofloxacin
νCOOs
∆νgoethite-ciprofloxacin
192a
198a
1575 1558
1383
175
1383
192 175
1534
1383
151
1383
192 175
1535
1382
153
1534
1384
150
1533
1383
150
4
1577
1382
195
1575 1558
5
1578
1381
197
1575 1558
1382
193 176
1382
193 176
6
1577
1381
196
1575 1558
7
1578
1381
197
1575 1558
1382
193 176
1382
193 176
8
1580
1381
199
1575 1558
9
1580
1381
199
1575 1558
a
goethite-ciprofloxacin complexes νCOOas
Notation ∆νiron-ciprofloxacin = ∆νciprofloxacin is used when ∆νciprofloxacin - ∆νiron-ciprofloxacin is e5; a difference of 5 units is not statistically significant.
and ∆νiron-ciprofloxacin ∼ ∆ν ciprofloxacin (Table 1) and suggested that the ciprofloxacin binds to Fe(III) via both carboxylate oxygens resulting in bidentate chelate and bridging bidentate complexes. Previous studies have suggested that bifurcations in νCOOas possibly indicate multiple complexes (24, 25) resulting from the interaction of ciprofloxacin with more than one type of ferric species (26). Goethite-Ciprofloxacin Complexes. We found evidence that the goethite surface also induced deprotonation of the carboxylic group at the pH < pKa1 of ciprofloxacin due to the absence of the νCdOcarboxyl (1710 cm-1) peak and the lower intensity of the coupled νCOOH and δC-OH modes in the IR spectra of goethite-ciprofloxacin complexes, collected between pH 4 and 7 (Figure 5). This observation corroborates with the understanding that the diffuse layer of a positively charged metal oxide surface typically possesses a higher pH than the bulk solution due to the presence of surplus hydroxyl ions required for electroneutrality (27). This higher diffuse layer pH facilitated the deprotonation of the COOH group at the oxide-water interface at pHbulk < pKa1. For reasons outlined next, we find that ciprofloxacin complexes to the goethite surface via bidentate chelates involving both carboxylate oxygens. As noted in colloidal ferric-ciprofloxacin spectra (Figure 3b), νCOOs in the goethiteciprofloxacin spectra, although broader, was not significantly shifted in position as compared with the corresponding aqueous ciprofloxacin spectra (Figure 5). In addition, νCOOas was shifted to lower wavenumbers (1535 cm-1) at all pH values as compared with the aqueous ciprofloxacin spectra (Figure 5 and Table 1). Furthermore, values of ∆νgoethite-ciprofloxacin were less than those of ∆νciprofloxacin (Table 1) at all pH values and thus indicated the presence of bidentate chelates that involved the interaction of both carboxylic oxygens with a surface site. As previously noted in ferric-ciprofloxacin spectra (Figure 3b), νCdOketone was also shifted to higher wavenumbers (by 10 cm-1) in the goethiteciprofloxacin spectra reflecting the electron withdrawing nature of the ortho carboxylate-iron bond. However, we do not have unquestionable evidence for the interaction of the keto group with the goethite surface. Our evidence of bidentate chelates on the goethite surface was distinct from mononuclear bidentate ciprofloxacin complexes onto hydrous ferric oxides (HFO) involving both the keto group and one carboxylate oxygen as proposed by Gu and Karthikeyan (4). While the different surface complexes proposed may suggest the influence of oxide mineralogy or sorption densities, it is equally likely that the distinct surface
FIGURE 5. ATR-FTIR difference spectra of ferric-goethite complexes as a function of pH. All studies were conducted at 25 °C with an ionic strength of 0.01 M (NaCl). All the spectra are normalized with the relative intensity of the peak observed at 1485 cm-1. structures hypothesized were due to differences in the assignment of νCOOas in the respective iron oxide-ciprofloxacin spectra. Whereas Gu and Karthikeyan (4) ascribed the νCOOas shift to higher wavenumbers (as compared to the aqueous ciprofloxacin spectra), our analyses found evidence for the νCOOas shift to lower wavenumbers. The lack of shift in the νCOOs stretch (as compared to the aqueous ciprofloxacin spectra) in our spectra lends important evidence required to exclude the monodentate complex via a single carboxylate oxygen proposed by Gu and Kartheykeyan (4) for ciprofloxacin-HFO and ciprofloxacin-HAO complexes. VOL. 41, NO. 9, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3157
Goyne and co-workers (7) also found evidence for the binding of ofloxacin to alumina via ketone and carboxylate functional groups. Our assignments of νCOOas in the goethiteciprofloxacin spectra corroborate with those of Goyne et al. (7) in the alumina-ofloxacin spectra. However, a dramatic decrease (as compared with the aqueous solution spectra) in νCdOketone was observed in the alumina-ofloxacin spectra, whereas the intensity of νCdOketone was not significantly altered in our study. Therefore, the differences in the hypothesized structures for the alumina-ofloxacin and goethite-ciprofloxacin complexes may be due to differences in metal oxide mineralogy.
(12)
(13)
(14)
(15)
Acknowledgments
(16)
This research was supported by the National Science Foundation under Grants BES-0225696 and BES-9984489 and the U.S. Department of Agriculture under Grant 2002-3510712258. The authors thank three anonymous reviewers and Drs. Dana Mayo and Richard Broene for their insightful comments and suggestions.
(17)
Literature Cited (1) Halling-Sørensen, B.; Nors Nielsen, S.; Lansky, P. F.; Ingerslev, F.; Lu ¨ tzhøft, H. C.; Jørgensen, S. E. Occurrence, fate, and effects of pharmaceuticals in the environmentsA review. Chemosphere 1998, 36 (2), 357-393. (2) Golet, E. M.; Xifra, I.; Siegrist, H.; Alder, A. C.; Giger, W. Environmental exposure assessment of fluoroquinolone antibacterial agents from sewage to soil. Environ. Sci. Technol. 2003, 15, 3243-3249. (3) Kolpin, D. W.; Furlong, E. T.; Meyer, M. T.; Thurman, E. M.; Zaygg, S. D.; Barber, L. B.; Buxton, H. T. Pharmaceuticals, hormones, and other organic wastewater contaminants in U.S. streams, 1999-2000: A national reconnaissance. Environ. Sci. Technol. 2002, 36, 1202-1211. (4) Gu, C.; Karthikeyan, K. G. Sorption of antimicrobial ciprofloxacin to aluminum and iron hydrous oxides. Environ. Sci. Technol. 2005, 39 (23), 9166-9173. (5) Zhang, H.; Huang, C. H. Oxidative transformation of fluoroquinolone antibacterial agents and structurally related amines by manganese oxide. Environ. Sci. Technol. 2005, 39, 44744483. (6) Nowara, A.; Burhene, J.; Spiteller, M. Binding of fluoroquinolone carboxylic acid derivatives to clay minerals. J. Agric. Food Chem. 1997, 45, 1459-1463. (7) Goyne, K. W.; Chorover, J.; Kubicki, J. D.; Zimmerman, A. R.; Brantley, S. L. Sorption of the antibiotic oxfloxacin to mesoporous and non-porous alumina and silica. J. Colloid Interface Sci. 2005, 283, 160-170. (8) Seremet, D. E.; MacKay, A. A. Probing sorption interactions of pharmaceuticals with multiple functional groups, extended abstract; 228th ACS National Meeting, Division of Environmental Chemistry: Philadelphia, 2005. (9) Stumm, W.; Morgan J. J. Aquatic Chemistry, 3rd ed.; WileyInterscience: New York, 1996. (10) Turel, I. The interactions of metal ions with quinolone antibacterial agents. Coord. Chem. Rev. 2002, 232, 27-47. (11) Vasudevan, D.; Cooper, E. M. Sorption in iron oxide rich mineral soils: Competition between phosphate and organic ions,
3158
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 9, 2007
(18)
(19)
(20)
(21)
(22) (23)
(24)
(25)
(26)
(27)
extended abstract; 223rd ACS National Meeting, Division of Environmental Chemistry: Orlando, FL, 2002. Coates, J. Interpretation of Infrared Spectra. In A Practical Approach in Encyclopedia of Analytical Chemistry; Meyers, R. A., Ed.; John Wiley and Sons: New York, 2000; pp 10815-10837. Bellamy, L. J. The Infrared Spectra of Complex Molecules. Vol. 2. Advances in Infrared Group Frequencies, 2nd ed.; Chapman and Hall: London, 1980. Tejedor-Tejedor, M. I.; Yost, E. C.; Anderson, M. A. Characterization of benzoic and phenolic complexes at the goethiteaqueous solution interface using cylindrical internal reflection Fourier transform infrared spectroscopy. Part 1. Methodology. Langmuir 1990, 6, 979-987. Crews, P.; Rodrı´guez, J.; Jaspars, M. Organic Structure Analysis; Oxford University Press: New York, 1998. Parikh, V. M. Absorption Spectroscopy of Organic Molecules; Addison-Wesley Publishing: Reading, MA, 1974. Ross, D. L.; Riley, C. M. Physicochemical properties of the fluoroquinolone antimicrobials. II. Acid ionization constants and their relationship to structure. Int. J. Pharm. 1992, 83, 267272. Lee, D.-S.; Han, H.-J.; Kim, K.; Park, W.-B.; Cho, J.-K.; Kim, J. H. Dissociation and complexation of fluoroquinolone analogues. J. Pharm. Biomed. Anal. 1994, 12 (2), 157-164. Lin, C.-E.; Deng, Y.; Liao, W.-S.; Sun, S.-W.; Lin, W.-Y.; Chen, C. C. Electrophoretic behavior and pKa determination of quinolones with a piperazinyl substituent by capillary zone electrophoresis. J. Chromatogr. A 2004, 1051, 283-290. Herlinger, A. W.; Long, T. M. Laser Raman and infrared spectra of amino acids and their metal complexes. III. Proline and biprolinato complexes. J. Am. Chem. Soc. 1970, 92 (22), 64816486. Nakamoto, K. Infrared and Raman Spectra of Inorganic and Coordination Compounds. Part B: Applications in Coordination, Organometallic, and Biorganic Chemistry, 5th ed.; WileyInterscience: New York, 1997. Tackett, J. E. FT-IR characterization of metal acetates in aqueous solution. Appl. Spectrosc. 1989, 43 (3), 483-489. Nara, M.; Torii, H.; Tasumi, M. Correlation between the vibrational frequencies of the carboxylate group and the types of its coordination to a metal ion: An ab initio molecular orbital study. J. Phys. Chem. 1996, 100, 19812-19817. Nordin, J.; Persson, P.; Laiti, E.; Sjo¨berg, S. Adsorption of o-phthalate at the water-boehmite (γ-AlOOH) interface: Evidence for two coordination modes. Langmuir 1997, 13, 40854093. Persson, P.; Nordin, J.; Rosenquivist, J.; Lo¨vgren, L.; Sjo¨berg, S.; O ¨ hman, L. O. Comparison of the adsorption of o-phthalate on boehmite (R-AlOOH), aged γ-Al2O3, and goethite (R-FeOOH). J. Colloid Interface Sci. 1998, 206 (1), 252-266. Allison, J. D.; Brown, D. S.; Novo-Gradac, K. J. MINTEQA2/ PRODEFA2: A Geochemical Assessment Model for Environmental Systems; Center for Exposure Assessment Modeling, U.S. EPA: Washington DC, 1991. O’Reilly, J. P.; Butts, C. P.; l’Anson, I. A.; Shaw, A. M. Interfacial pH of an isolated silica-water surface. J. Am. Chem. Soc. 2005, 127, 1632-1633.
Received for review August 10, 2006. Revised manuscript received February 5, 2007. Accepted February 6, 2007. ES061921Y
