Examination of Cadmium(II) Complexation by the Suwannee River

Nov 16, 2001 - Aquatic and terrestrial fulvic acids are environmentally important because they affect the bioavailability and transport of metal ions...
3 downloads 10 Views 68KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 4900-4904

Examination of Cadmium(II) Complexation by the Suwannee River Fulvic Acid Using 113Cd NMR Relaxation Measurements WILLIAM H. OTTO,† SARAH D. BURTON,‡ W. ROBERT CARPER,§ AND C Y N T H I A K . L A R I V E * ,† Department of Chemistry, University of Kansas, Lawrence, Kansas, 66045, Pacific Northwest National Laboratories, Richland, Washington, and Department of Chemistry, Wichita State University, Wichita, Kansas 67260

Aquatic and terrestrial fulvic acids are environmentally important because they affect the bioavailability and transport of metal ions. Prior studies demonstrated that Cd(II) binds to the oxygen containing functional groups of fulvic acids. The complexation of Cd(II) is further investigated in this study using 113Cd NMR relaxation measurements for solutions of the Suwannee River fulvic acid (SRFA). Spinlattice (T1) and spin-spin (T2) relaxation times are measured over a range of Cd(II):SRFA ratios. The results clearly indicate two types of Cd(II) binding sites for the SRFA. A series of model ligands was also examined to gain further understanding of the two types of binding motifs present in the fulvic acid. The results for a model compound containing several carboxylate functionalities in near proximity correspond very closely to the results obtained for the strong binding sites of the Cd(II)-SRFA complexes.

Introduction Humic substances are a heterogeneous mixture of decomposition products of natural organic matter. They are the predominate source of dissolved organic carbon in surface waters and are environmentally important because they play a major role in the transport mechanisms of various organic pollutants and metal ions. Humic substances are also known to affect the bioavailability and toxicity of metal ions. Fulvic acids are a subclass of humic substances, operationally defined by their solubility after extraction. In general, aquatic fulvic acids are a heterogeneous mixture of components characterized by an average molecular mass around 800 Da with relatively high oxygen, low nitrogen, and low sulfur content. The specific chemical characteristics of a fulvic acid sample are influenced by the environment from which it originates (1, 2). On average there are 4-5 carboxylate moieties per fulvic acid, which are important determinants of their metal binding properties. Cadmium is an environmentally important toxic heavy metal that exists in water systems both naturally and as a pollutant. Thus, the complexation between Cd(II) and fulvic acids is an important determinant of the environmental fate and transport of Cd(II). * Corresponding author phone: (785)864-4269; fax: (785)864-5396; e-mail: [email protected]. † University of Kansas. ‡ Pacific Northwest National Laboratories. § Wichita State University. 4900

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001

113 Cd NMR is a useful technique to probe Cd(II) binding because the large chemical shift range of 113Cd strongly reflects the chemical environment of the Cd(II) binding site. Prior studies have utilized 113Cd NMR chemical shifts to study the complexation of Cd(II) with aquatic and soil fulvic acids (35). Average binding constants were measured for several fulvic acids, and the predominate binding site was determined to be the fulvic acid carboxylate moieties (3). Early in the development of pulsed Fourier transform NMR techniques, relaxation parameters, specifically spin-lattice (T1) and spin-spin (T2) relaxation times, were recognized for the valuable information they could yield regarding structural characteristics and molecular dynamics in solution (6-10). For example, direct information about associative processes such as the binding of organic ligands to fulvic acids has been determined from T1 and T2 measurements (11, 12). Parameters representative of the molecular dynamics, such as correlation times, can also be directly determined from measured spin-spin and spin-lattice relaxation times (7, 8, 13, 14). Such a study using the relaxation rate ratio method to directly calculate correlation times has been performed for the Cd-CyDTA complex, concluding that this complex is less labile than Cd-EDTA complexes because of the lack of water in the inner coordination sphere (15). Relaxation measurements can potentially provide information about binding that cannot be revealed through NMR chemical shift measurements. Therefore, this study further investigates the binding of Cd(II) with Suwannee River fulvic acid (SRFA) using 113Cd NMR relaxation measurements. Because of the lack of relaxation rate data for 113Cd complexes in the literature, relaxation times for a series of model compounds were measured to provide a more complete understanding of the effect of different binding motifs on the relaxation rates of 113Cd(II). Since evidence has shown that Cd(II) binds to the carboxylate moieties of fulvic acid (3-5), the model ligands were chosen to represent a series of carboxylate binding sites. The first model compound, acetate, represents a single carboxylate ligand. Because the chemical shift of Cd-SRFA complexes suggests the possibility of bidentate carboxylate-hydroxyl binding (3), salicylate was chosen to represent this binding motif. Evidence has been presented that indicates that nitrogen may be a donor in some instances (5), thus a carboxylate site with a nitrogen donor is represented by nitrilotriacetate (NTA). A polydentate carboxylate binding site is modeled using 1,2,3,4-cis-cis-ciscis-cyclopentanetetracarboxylic acid (CPTA).

Experimental Section Since Cd(ClO4)2 was used to reference 113Cd chemical shifts and the perchlorate complex is considered to be free Cd(II) (16), all metal solutions were made using the metal perchlorate to prevent the introduction of other ligands into the solutions used in this study. A 113Cd stock solution was prepared by dissolving 113CdO (95%+ enriched, Cambridge Isotope Laboratories) in a minimal amount of HClO4 (70%, Aldrich) and then diluting with D2O. The exact Cd(II) concentration was determined by flame atomic absorption spectrophotometry. Stock solutions of 30.0 mg/mL fulvic acid were prepared by dissolving a weighed amount of lyophilized Suwannee River Reference fulvic acid in D2O. The solution pD was adjusted to 6.40 ( 0.04 using standardized 1.60 M NaOD (Isotec) and 0.121 M NaOD. The ionic strength added by the pH adjustment was 0.07 M. pH measurements were corrected for the deuterium isotope effect using the relationship: pD ) pH meter reading + 0.4 (17). 10.1021/es0108032 CCC: $20.00

 2001 American Chemical Society Published on Web 11/16/2001

Solutions used for most NMR measurements were made with a constant 113Cd(II) concentration, 2.0 mM, and the fulvic acid concentration was varied from 0.50 to 10.0 mg/ mL. To maintain a constant ionic strength over the wide range of fulvic acid concentrations examined, NaClO4 was added to yield an ionic strength of 0.07 M for each solution. The Cd-SRFA solutions were then prepared by adding measured aliquots of the fulvic acid, 113Cd(II), and NaClO4 stock solutions. The final solution pD was adjusted to 6.40 ( 0.04 with NaOD, and D2O was added to yield a total volume of 3.00 mL. The appropriate concentration ratios of ligand and Cd(II) to produce a 1:1 ligand-metal complex for the relaxation studies were determined for each model ligand based on literature values of the ligand-Cd(II) formation constant or the experimentally determined bound 113Cd(II) chemical shifts (18). Samples for the relaxation measurements of the model compounds were prepared such that >85% of the Cd(II) was bound as a 1:1 Cd(II):ligand complex. An acetate stock solution was prepared by dissolving 28.7 mg of sodium acetate (Fisher) in 200 µL of D2O. Subsequently, 171 µL of this acetate solution was mixed with 1.80 mL of D2O and 37.5 µL of 113Cd(II) stock solution and diluted to 3.00 mL with D2O to yield a solution containing 5.25 mM 113Cd(II) and 110 mM acetate at pD 6.30. A salicylate stock solution was prepared by dissolving 65.1 mg of salicylic acid sodium salt (99%, Aldrich) in 1.00 mL of D2O to yield a 406 mM solution. The (113Cd-salicylate)+1 solution used for the measurement of relaxation rates was prepared by adding 1.97 mL of D2O and 36.4 µL of 113Cd(II) stock solution to the 406 mM salicylate solution to yield a solution containing 5.00 mM 113Cd(II) and 135 mM salicylate at pD 4.53. A NTA stock solution was similarly prepared by dissolving 28.2 mg of nitrilotriacetic acid disodium salt (99%, Aldrich) in 200 µL of D2O. A total of 170 µL of this 600 mM NTA solution was added to a solution containing 2.50 mL of D2O and 10 µL of the 113Cd(II) stock solution to yield a solution of 1.8 mM 113Cd(II) and 38.0 mM NTA at pD 3.40. The solution used to measure the relaxation times of the Cd(II)-CPTA complex was prepared from a stock solution obtained by dissolving 57.0 mg of CPTA in 1.00 mL of D2O; 100 µL of this 232 mM CPTA stock solution was added to 2.00 mL of D2O and 10 µL of 113Cd(II) stock. The pD was adjusted to 5.78, and the solution diluted with 500 µL of D2O to produce final concentrations of 1.8 mM 113Cd(II) and 8.89 mM CPTA at pD 5.78. NMR Measurements. All 113Cd NMR spectra were externally referenced using an internal sealed capillary containing 1.5 M Cd(ClO4)2 with respect to 0.7 M Cd(ClO4)2, a concentration where the 113Cd chemical shift was determined to be equal to -1.196 ppm. The bound chemical shifts of the 1:1 complexes of 113Cd(II) with the model ligands were determined by titrating aliquots of each ligand stock solution into a 4.0 mM 113Cd(II) solution in D2O. Measurements of the 113Cd chemical shift for solutions of the model compounds were performed at the University of Kansas using a Bruker AM-360 MHz NMR equipped with a 10-mm broad band probe at a spectral frequency of 79.87 MHz. The measured chemical shift was plotted versus the inverse ligand concentration, and the bound Cd(II) chemical shift was determined from the y-intercept of the line obtained from linear regression. These determined values are Cd-acetate, -23.2 ppm (in good agreement with the value reported by Chung and Moon; 20); Cd-salicylate, -17.3 ppm; Cd-NTA, 6.9 ppm; and Cd-CPTA, -21.8 ppm. 113Cd NMR spectra for the determination of relaxation parameters were measured remotely with a Varian/Chemagnetics Infinity 500 MHz NMR at the Environmental Molecular Sciences Laboratory located at Pacific Northwest National Laboratory equipped with a 10-mm broad band

TABLE 1. Summary of Results of 113Cd Chemical Shift and Relaxation Time Measurements as a Function of SRFA Concentration sample SRFA (mg/mL) with 2.0 mM SRFA Cd(II)/ Cd(II) (mM) SRFA 0.50 0.60 0.75 1.00 1.25 1.70 2.00 3.00 5.00 10.0

0.62 0.75 0.94 1.25 1.56 2.12 2.50 3.75 6.25 12.5

3.2 2.7 2.1 1.6 1.3 0.95 0.80 0.53 0.32 0.16

δobs

T1 (s)

T2 (ms)

R 2 /R 1

-2.04 -5.68 -7.13 -7.37 -9.72 -11.77 -16.96 -18.60 -20.26 -21.55

1.8 ( 0.1 1.60 ( 0.05 1.5 ( 0.2 1.14 ( 0.06 0.66 ( 0.03 0.56 ( 0.02 0.33 ( 0.04 0.26 ( 0.01 0.22 ( 0.01 0.17 ( 0.01

9.4 ( 0.4 4.6 ( 0.2 3.9 ( 0.2 3.3 ( 0.2 2.3 ( 0.1 2.00 ( 0.06 1.25 ( 0.06 1.18 ( 0.04 0.84 ( 0.04 0.60 ( 0.03

190 350 380 350 290 280 260 220 260 280

multinuclear probe, at a spectral frequency of 110.872 MHz. One-dimensional 113Cd spectra were obtained with the reference capillary inserted into the NMR tube by applying a 60° pulse (approximately 12 µs). A standard spin-echo experiment was performed using a sample containing 3.00 mg/mL SRFA and 3.0 mM Cd(II) to determine the T2 relaxation time of the 113Cd(II) in this sample. A range of values of the spin-echo delay times (τ) from 0.1 to 1.4 ms was used. Cd(II) T1 measurements were obtained without the reference capillary using the inversion recovery pulse sequence. An appropriate relaxation delay was utilized so that the repetition time was at least 5 times the measured T1, and an array of seven values of the inversion recovery delay time was used with the longest value set to about 6 times the expected T1. For all experiments the spectral width was 11.12 kHz and 5120 data points were collected. 113Cdfree induction decays were zero-filled to 16384 points and apodized using an exponential decay corresponding to 20 Hz line broadening, Fourier transformed, and baseline corrected with a matched fifth-order polynomial using Spinsight or Felix 97.0 (Biosym).

Results and Discussion Initially a one-dimensional 113Cd spectrum was measured for all samples. As expected on the basis of previous literature reports, the cadmium solutions containing the fulvic acid sample or model ligands yielded only a single 113Cd resonance, indicating that the Cd(II) was in fast exchange on the NMR time scale (3-5, 20). The 113Cd chemical shifts measured with respect to 0.7 M Cd(ClO4)2 for these solutions are reported in Table 1. The 113Cd chemical shift moves from the determined bound chemical shift at low Cd:SRFA ratios toward the chemical shift of free Cd(II) at high Cd:SRFA ratios. However, more detailed information about the nature of the binding sites cannot be readily discerned from the chemical shift measurements. Further evidence that the Cd(II) is in fast exchange was provided by the temperature dependence of the 113Cd resonance line width. 113Cd NMR measurements of a solution containing 3.0 mM Cd and 4.00 mg/mL SRFA were performed as a function of temperature from 291 to 330 K with approximately 15 min for equilibration prior to each measurement. The 113Cd line width did narrow slightly from 222 to 205 Hz (about an 8% change) as the temperature was increased from 291 to 330 K. However, this small change is hardly indicative of a major contribution to relaxation from exchange broadening. Furthermore, this small change in line width was accompanied by a parallel decrease in chemical shift from -21.4 ppm at 291 K to -20.9 ppm at 330 K, VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4901

FIGURE 2. Plot of the peak intensities measured for the spectra shown in Figure 1. The line is the nonlinear least-squares fit of eq 1 to the experimental data. FIGURE 1. Stacked plot of the inversion recovery spectra measured for a solution containing 3.00 mg/mL SRFA and 2.0 mM 113Cd(II). The τ values (in ms) are (A) 0.001, (B) 50, (C) 250, (D) 450, (E) 650, (F) 900, and (G) 1400. The spectra were acquired by coaddition of 8500 transients, with a recycle time of 1.401 s. All spectra are processed with 20 Hz of line broadening. indicating that the change in resonance line width over this temperature regime may result from a temperature-dependent decrease in the binding constant rather than a shift in the rate of chemical exchange. Although the variation of 113Cd chemical shift could not be used over this wide range of concentration ratios to understand the nature of the Cd(II)-SRFA complex, additional information about the binding can be determined by analysis of 113Cd relaxation rates. To determine the T1 relaxation times, a series of 113Cd inversion recovery spectra were measured for each 2.0 mM Cd(II) sample as a function of the relaxation delay (τ) as shown in Figure 1 for the 3.00 mg/mL SRFA solution (21). The 113Cd resonance intensities in each set of spectra were integrated and analyzed by nonlinear leastsquares fitting using

It ) I0(1 - B exp(-τ/T1))

(1)

where It is the peak intensity measured with a delay of τ, I0 is the peak intensity of the fully relaxed magnetization, and B is a variable compensating for imperfect 180° pulses (22, 23). For all measurements, B was determined to be between 1.8 and 2.0 (the value for a perfect 180° pulse). The fit of eq 1 to the data obtained for the 3.00 mg/mL SRFA sample is shown in Figure 2. The quality of the fit of a single-exponential function indicates a single average T1 relaxation time, suggesting fast exchange on the relaxation time scale. The 113Cd T relaxation times measured as a function of the SRFA 1 concentration are reported in Table 1. These relaxation times are much shorter than the value obtained for the free Cd(II) in solution (∼60s). The relaxation rates (R1 ) 1/T1) measured for samples with a constant 113Cd(II) concentration are plotted as a function of the inverse of the SRFA concentration in Figure 3. From this plot there are two observable regimes. The first regime, observed at high fulvic acid concentration, shows a large decrease in R1 until the [Cd] to [SRFA] ratio reaches 1.5. Since the SRFA is a heterogeneous mixture, this regime corresponds to the region where the Cd(II) is bound predominately at the strongest binding sites. A second regime in this graph is observed at lower fulvic acid concentrations, where Cd(II) binding should occur at the weaker fulvic acid 4902

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001

FIGURE 3. Plot of R1 vs the ratio of the concentration of Cd(II) (in mM) to the SRFA concentration (in mg/mL). The concentration of Cd(II) was held constant at 2.0 mM in these experiments, and the SRFA concentration was varied. binding sites and the amount of free Cd(II) is increased. The intersection at 1.5 represents a 1:1 Cd(II)-SRFA ratio if a molecular weight of 800 is assumed for the SRFA, a value very close to the reported number average molecular weight. Thus the first regime corresponds to a 1:1 Cd:SRFA complex, while the second regime corresponds to complexes with more than one Cd(II) bound to the SRFA molecules, on average. 113Cd transverse relaxation rates (R ) 1/T ) were measured 2 2 to provide additional information about the Cd(II) binding site. Transverse relaxation times can be measured with the standard spin-echo experiment, which produced a T2 value of 1.4 ( 0.1 ms for a sample containing 3.0 mM 113Cd and 3.00 mg/mL SRFA. The T2 relaxation time can also be determined from the resonance line width, w1/2, after subtraction of any line broadening added during the Fourier transform:

w1/2 ) (πT2*)-1

(2)

However, the relaxation time determined from the line width, denoted T2*, is comprised of both the true T2 and the experimental contributions arising from magnetic field inhomogeneity. For the 3.00 mg/mL SRFA sample described above, a 113Cd T2* value of 1.3 ( 0.1 ms was determined using the resonance line width. The good agreement between this T2* relaxation time and the T2 determined from the spinecho spectrum indicate that experimental contributions do not significantly affect the measured line widths. Thus, all T2 values were determined from 113Cd line width in the onedimensional spectrum of each sample using eq 2. The line

TABLE 2. Results of 113Cd Chemical Shift and Relaxation Time Measurement and for Cd(II) Solutions Containing the Model Ligands

FIGURE 4. Plot of R2 vs the ratio of the concentration of Cd(II) (in mM) to the SRFA concentration (in mg/mL). The concentration of Cd(II) was held constant at 2.0 mM in these experiments, and the SRFA concentration was varied. width of the 113Cd resonance produced by an internal capillary containing 3 M Cd(ClO4)2 was used to further verify that magnetic field inhomogeneity did not fluctuate between samples. The R2 relaxation rates are plotted as a function of the 1/[SRFA] in Figure 4. As was observed in Figure 3, two regimes are noted in Figure 4 with an intersection at a Cd(II)/SRFA of approximately 1.5. Again the strongest binding sites corresponding to the lowest Cd(II)/SRFA yield the larger R2 values. The solutions with higher Cd:SRFA ratios, corresponding to the lower R2 values, should have a greater population Cd(II) bound at the weaker metal binding sites as well as a greater percentage of free Cd(II). Examination of Figures 3 and 4 show that the solutions which would favor complexation of the Cd(II) by the strongest binding sites correspond to the fastest relaxation rates. This would likely arise from efficient relaxation mechanisms provided by a binding site where the Cd(II) is sterically hindered, for example, a multidentate binding site in which the SRFA effectively surrounds the Cd(II). Using the relaxation rate ratio method, R2/R1 ratios act as indicators of relative rotational motion and can be very useful for elucidating relative motion within a given molecule or complex (14). In particular, the R2/R1 ratio for a given nucleus, for example, 113Cd, increases with increasing rotational correlation time (13, 14). For cadmium-ligand complexes, the correlation time can be reflective of the relative mobility. Correlation times measured for the carboxylate carbons of the ligand CyDTA showed dramatic decreases upon cadmium complexation (15). The R2/R1 ratios are listed in Table 1 for the SRFA solutions measured at constant Cd(II) concentration. Assuming an average SRFA molecular mass of 800 Da, the SRFA concentration increases from 0.62 to 12.5 mM and the corresponding ratio of Cd(II)/SRFA decreases from 3.2 to 0.16 (see Table 1). For larger Cd:SRFA ratios (with the exception of the first value), one observes essentially constant R2/R1 ratios of 350380. In the low Cd(II)/SRFA region, essentially constant R2/ R1 ratios of 220-280 are observed. The average high R2/R1 value (360) corresponds to a 113Cd rotational correlation time of 35 ns, and the lower R2/R1 average value (260) corresponds to a 113Cd rotational correlation time of 30 ns (13). A possible interpretation of the results for the larger Cd:SRFA ratios is that the longer correlation time results from an increase in the population of free Cd(II) or from hopping of the Cd(II) between weaker SRFA binding sites. Model Ligands. 113Cd NMR inversion recovery spectra were also measured for a series of model ligands to probe the dependence of the type of Cd(II) binding site on the 113Cd relaxation times. The 113Cd relaxation times and chemical shifts for 1:1 complexes with these model ligands

sample

δobs

T1 (s)

T2 (ms)

R 2 /R 1

100 mM acetate 5.25 mM Cd(II) 135 mM salicylate 5.00 mM Cd(II) 38.0 mM NTA 1.8 mM Cd(II) 8.89 mM CPTA 1.8 mM Cd(II)

-21.7

30.4 ( 0.5

24 ( 1

1300

-14.7

19.5 ( 1.5

11.2 ( 0.5

1740

6.4

11.5 ( 0.3

1.9 ( 0.1

6000

-20.2

1.8 ( 0.2

10.1 ( 0.5

180

are reported in Table 2. Although the chemical shifts are all indicative of binding at carboxylate sites, the more downfield shift of the salicylate complex, -14.7 ppm, shows the effect of carboxylate-hydroxyl binding, and the effect of nitrogen is observed in the chemical shift of the Cd-NTA complex, 6.4 ppm. The monodentate complex, Cd-acetate, has the longest T1 (30.4 s), which is nearest to that of the free Cd(II). It is possible that, for the Cd-NTA complex, a contribution from CSA or 15N-Cd scalar coupling could affect the value the measured relaxation times (15). The Cd-CPTA complex has the shortest T1 of the model compounds likely due to restricted mobility resulting from complexation by the carboxylate donors of the CPTA complex. The T1 values determined for the other three model ligands, including NTA, are all at least an order of magnitude higher than the CdCPTA complex. However, there must also be additional mechanisms of relaxation for the Cd-SRFA complexes, as the T1 and T2 values of the fulvic acid samples are significantly shorter than the CPTA model. A possible explanation is that if the fulvic acid envelops the Cd(II), it could produce an environment surrounding the metal ion that would place methyl groups nearby providing additional 1H-113Cd relaxation. Chemical exchange could also provide an additional relaxation mechanism for Cd-fulvic acid solutions, as a variety of binding sites are available. However, under the conditions of these experiments, chemical exchange appears to be a minor contributor to 113Cd relaxation. From the R2/R1 ratios and other data in Tables 1 and 2, it is apparent that the Cd-CPTA results are closest to those obtained for the Cd-SRFA samples, in particular for the samples at low fulvic acid content. The weight of evidence based on the relaxation times and R2/R1 ratios all suggest that the Cd-CPTA complex is a reasonable model for the type of binding observed for the stronger SRFA metal binding sites. From our studies of 113Cd relaxation times, it is apparent that two types of Cd(II) binding sites exist in the SRFA. The first is a stronger binding site that provides an efficient relaxation mechanism for Cd(II) and is best represented by the multidentate Cd-CPTA model. Such a complex could easily be considered to be an inner-sphere complex in which these stronger binding sites would have a much higher affinity for a species such as the soft Cd(II) rather than a hard cation such as Ca(II). An inner-sphere complex is defined as linkage between the ligand and the reactive surface with no water of hydration between the adsorbed ion and the surface functional group (24). This type of binding mechanism would explain some of the differences in average binding affinity determined by Otto et al. in previous studies of Cd(II)-Ca(II) competition (3). In this study, we observed that there are a small number of fulvic acid binding sites that preferentially bind Cd(II). These binding sites could correspond to the stronger binding sites observed here based on the Cd:SRFA ratios examined. The second type of binding site, corresponding to an outer-sphere complex, could result from binding to the exterior of the fulvic acid where Cd(II) VOL. 35, NO. 24, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4903

is more superficially bound to carboxylate functionalities. An outer-sphere complex is defined as a site where at least one solvent molecule is interposed between the functional group and the bound ion (24). This mechanism could allow a much higher mobility and thus would not provide as efficient relaxation as an inner-sphere complex. This type of outer-sphere complex has been reported previously for Mn2+ bound to a well-characterized fulvic acid (25). 113 Cd relaxation measurements have been shown to be extremely useful in revealing aspects of the Cd(II)-fulvic acid binding that had not been previously observed. Two distinct types of binding sites are uncovered by examining the relaxation rates as a function of fulvic acid concentration. These two binding sites show that SRFA on average has one strong binding site per molecule and that there exists the potential for the Cd(II) to bind to a second weaker binding site. Of the model compounds examined, the results for Cd(II)-CPTA most closely fit the experimental results obtained for the SRFA solutions, thus a polydentate carboxylate binding site may be representative of the strongest Cd(II) binding site.

Acknowledgments Most of this research was performed in the Environmental Molecular Sciences Laboratory (a national scientific user facility sponsored by the DOE Office of Biological and Environmental Research) located at Pacific Northwest National Laboratory, operated by Battelle for the DOE. This work was supported by EPA EPSCoR Grant R827589-01-0.

Literature Cited (1) Malcolm, R. L. Anal. Chim. Acta 1990, 232, 19-30. (2) McKnight, D. M.; Aiken, G. R. Ecol. Stud. 1998, 133, 9-39. (3) Otto, W. H.; Carper, W. R.; Larive, C. K Environ. Sci. Technol. 2001, 35, 1463-1468. (4) Larive, C. K.; Rogers, A.; Morton, M.; Carper, W. R. Environ. Sci. Technol. 1996, 30, 2828-2831.

4904

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 24, 2001

(5) Li, J.; Perdue, E. M.; Gelbaum, L. T. Environ. Sci. Technol. 1998, 32, 483-487. (6) Farrar, T. C.; Becker, E. D. Pulse and Fourier Transform NMR; Academic Press: New York, 1971. (7) Levy, G. C.; Cargioli, J. D.; Anet, F. A. L. J. Am. Chem. Soc. 1973, 95, 1527-1535. (8) Levy, G. C.; Udlund, U. J. Am. Chem. Soc. 1975, 97, 5031-5032. (9) Farrar, T. C.; Druck, S. J.; Shoup, R. R.; Becker, E. D. J. Am. Chem. Soc. 1972, 94, 699-703. (10) Vold, R. L.; Waugh, J. S.; Klein, M. P.; Phelps, D. E. J. Chem. Phys. 1968, 48, 3831-3832. (11) Nanny, M. A.; Bortiatynski, J. M.; Hatcher, P. G. Environ. Sci. Technol. 1997, 31, 530-534. (12) Dixon, A. M.; Mai, M. A.; Larive, C. K. Environ. Sci. Technol. 1999, 33, 958-964. (13) Carper, W. R.; Keller, C. E. J. Phys. Chem. A 1997, 101, 32463250. (14) Carper, W. R. Concepts Magn. Reson. 1999, 11, 51-60. (15) Dixon, A. M.; Larive, C. K.; Nantsis, E. A.; Carper, W. R. J. Phys. Chem. A 1998, 102, 10573-10578. (16) Summers, M. F. Coord. Chem. Rev. 1988, 86, 43-134. (17) Bates, R. G. Determination of pH: Theory and Practice; Wiley: New York, 1964; pp 219-220. (18) Ramette, R. W. Chemical Equilibrium and Analysis; AddisonWesley: Reading, MA, 1981; pp 727-728. (19) Kostelnick, R. J.; Bothner-By, A. A. J. Magn. Reson. 1974, 14, 141-151. (20) Chung, K. H.; Moon, C. H. J. Chem. Soc., Dalton Trans. 1996, 1, 75-78. (21) Bain, A. D. J. Magn. Reson. 1990, 89, 153-160. (22) Kowalewski, J.; Levy, G. C.; Johnson, L. F.; Palmer, L. J. Magn. Reson. 1977, 26, 533-536. (23) Sass, M.; Ziessow, D. J. Magn. Reson. 1977, 25, 263-276. (24) Sposito, G. The Surface Chemistry of Soils; Oxford University Press: Oxford, England, 1984. (25) Gamble, D. S.; Langford, C. H.; Tong, J. P. K. Can. J. Chem. 1976, 54, 1239-1245.

Received for review March 29, 2001. Revised manuscript received September 19, 2001. Accepted September 21, 2001. ES0108032