Measurement of Cadmium (II) and Calcium (II) Complexation by Fulvic

Overall KCa values range from 4.6 to 7.8 × 102 M-1. View: PDF | PDF w/ ... Cadmium accumulation by muskmelon under salt stress in contaminated organi...
4 downloads 0 Views 87KB Size
Download PDF
Environ. Sci. Technol. 2001, 35, 1463-1468

Measurement of Cadmium(II) and Calcium(II) Complexation by Fulvic Acids Using 113Cd NMR WILLIAM H. OTTO,‡ 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, and Department of Chemistry, Wichita State University, Wichita, Kansas 67260

Aquatic and terrestrial fulvic acids are environmentally important in pollution transport because they affect the bioavailability and transport of metal ions. The complexation of the metal ions, Cd(II) and Ca(II), with several fulvic acids is examined in this study using 113Cd NMR. Our results indicate that Cd(II) predominately binds to the oxygen containing functional groups of the fulvic acids. A single 113Cd NMR resonance is observed in NMR spectra of Cd(II)fulvic acid solutions indicating fast exchange between free and complexed cadmium species. An average association equilibrium constant, KCd, is determined from NMR spectra measured for the titration of fulvic acid with Cd(II). The KCd values determined for the four fulvic acids studied range between 1.2 and 3.5 × 103 M-1. Competitive binding between Ca(II) and Cd(II) is used to indirectly determine an average association equilibrium constant, KCa, for Ca(II) with each fulvic acid. Overall KCa values range from 4.6 to 7.8 × 102 M-1.

Introduction Nuclear magnetic resonance (NMR) spectroscopy is a useful technique to probe metal binding. As a noninvasive technique NMR measurements do not perturb metal binding equilibria. 113Cd NMR was used in this paper to study complexation of Cd(II) with aquatic and soil fulvic acids. Because of the large chemical shift range of 113Cd, the measured chemical shift strongly reflects the chemical environment of the Cd(II) binding site, making113Cd NMR a useful technique to probe Cd(II) binding. Cadmium is an environmentally important toxic heavy metal that exists in water systems both naturally and as a pollutant. However, the concentration of Cd(II) is generally lower than other naturally occurring metal ions in aquatic ecosystems. Therefore, competing metal-humic substance interactions should be considered when modeling the distribution and bioavailability of trace metals in aquatic ecosystems. Humic substances are important because they play a major role in the transport mechanisms of various organic pesticides and metal ions (1-4). Humic substances are also known to affect the bioavailability and toxicity of metal ions (4, 5). 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 * Corresponding author phone: (785)864-4269; fax: (785)864-5396; e-mail: [email protected]. † Wichita State University. ‡ University of Kansas. 10.1021/es991372e CCC: $20.00 Published on Web 02/21/2001

 2001 American Chemical Society

components with an average molecular weight of 800 Daltons with relatively high oxygen and low nitrogen and sulfur content. The origination location of a fulvic acid affects its specific chemical characteristics. On average there are 4-5 carboxylate moieties per fulvic acid. Soil fulvic acids are generally larger than aquatic fulvic acids, with average molecular weights near 3000 Daltons, and tend to have higher oxygen contents. Classical competition reactions between ligands and metals have been utilized when direct methods are not available to determine binding constants (6). Competition experiments utilizing fulvic acids have been performed in prior investigations of Mn2+ binding (7) and Cu2+ competition with Ca2+ and K+ (8). This study investigates the competing interactions between Cd(II) and Ca(II) for complexation by fulvic acids. Ca(II) is a naturally occurring prevalent metal ion and is chemically similar to Cd(II) in ionic size and charge, thus competition for similar metal binding sites on natural ligands is expected between these two ions. Indications that Ca(II) effectively competes for some humic acid sites has been observed using anodic stripping voltammetry (9). Direct detection of Ca(II) via NMR is not feasible; however, the displacement of Cd(II) upon addition of Ca(II) can be used to indirectly probe Ca(II) binding. For example, 113Cd NMR has been used in a prior investigation of Ca(II) and Cd(II) competitive binding to serum albumin (10). In this study, the binding of Cd(II) as well as the competitive binding of Ca(II) to Suwannee River reference fulvic acid (SRFA), Laurentian soil fulvic acid (LSFA), Clinton Lake fulvic acid (CLFA), and Wakarusa River fulvic acid (WRFA) are examined.

Experimental Section Chemicals. As shown in prior studies, careful solution preparation is vital when investigating Cd(II) binding (11, 12). Since Cd(ClO4)2 is used to reference 113Cd chemical shifts and the perchlorate complex is considered to be free Cd(II) (13), all metal solutions are made using the metal perchlorate, to prevent the introduction of other ligands into the solutions used in this study. A 113Cd stock solution of approximately 500 mM was prepared by dissolving 113CdO (Cambridge Isotope Laboratories) with HClO4 (70%, ACS reagent grade, Aldrich) and D2O (99.9%, Cambridge Isotope Laboratories). Exact concentrations of the Cd(II) stock solutions were measured by flame AAS. pH measurements were corrected for the deuterium isotope effect using the relationship pD ) pH meter reading + 0.4 (14). Ca(II) stock solutions were prepared by dissolving CaO (99.95%, High Purity Standards) in HClO4, diluting with D2O, and adjusting the pD to 6.40 ( 0.05. Stock solutions of each fulvic acid for the Cd(II) titrations were prepared by dissolving 10.0 mg of the lyophilized powder in 750 µL of D2O, adjusting the solution pD to 6.4 with 1 and 0.1 M NaOD (Isotec) solutions, and diluting total volume to 1.00 mL with D2O. Stock solutions of each fulvic acid for the Ca(II) titrations were prepared by dissolving 4.0 mg of the lyophilized powder in 500 µL of D2O and adjusting the pD to 6.4 with NaOD. Four different fulvic acids, from the Suwannee River, the Wakarusa River, Clinton Lake, and a Laurentian soil, were examined in this study. The SRFA reference, isolated from the Suwannee River, GA, was purchased from the International Humic Substance Society. The WRFA and CLFA were isolated from Clinton Lake, KS, and the Wakarusa River, KS, respectively (15). Clinton Lake is a major source of drinking water for Lawrence, KS, and the Wakarusa River is the primary source for Clinton Lake. The LSFA is a soil fulvic acid isolated from the Laurentian area in Canada provided by Robert Cook. VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1463

Isolation and characterization of the SRFA (16), WRFA and CLFA (15), and LSFA (17, 18) are described in detail elsewhere. Preparation of Samples. Samples for Cd(II) titrations were prepared by diluting the FA stock with 1000 µL of D2O and adding an appropriate volume of Cd(II) stock. The solution pD was adjusted to 6.40 ( 0.02 with NaOD solutions. D2O was added to achieve a final volume of 2.50 mL. For the Cd(II) titration, successive aliquots of Cd(II) stock solution were added using a calibrated pipet, and the solution pD was adjusted to 6.40 ( 0.02 with 0.1 M NaOD solution. Samples for the Ca(II) competition studies were prepared by adding an appropriate volume of Cd(II) stock solution such that the concentration of Cd(II) was 3.0 mM. Finally, the solution pD was adjusted to 6.40 ( 0.02 with NaOD solutions, and D2O was added for a final volume of 1000 µL. Successive aliquots of Ca(II) stock were added for each measurement within a titration series. At the start of each titration, the ionic strength of the fulvic acid solution ranged from 20 to 28 mM. NMR Chemical Shift Measurements. 113Cd NMR spectra were externally referenced to 0.7 M Cd(ClO4)2, a concentration where the 113Cd chemical shift is equal to 0 ppm (19). Cd(II) titrations were performed using an internal sealed capillary containing 3 M Cd(ClO4)2 and referenced at -1.78 ppm. Spectra were measured with a Bruker AM-360 MHz NMR equipped with a 10 mm broad band multinuclear probe, at a spectral frequency of 79.87 MHz, using a single pulse experiment with no decoupling at 298 K. The spectral width was 7936 Hz. An 11.0 µs (60°) pulse width was used, and either 2048 or 4096 data points were collected, with acquisition times of 0.129 or 0.258 s. No additional relaxation delays were employed. 113Cd free induction decays were zero filled to 16 384 data points, apodized using an appropriate exponential decay, Fourier transformed, and baseline corrected with a matched fifth-order polynomial using Felix 97.0 (Biosym).

FIGURE 1. Sample 113Cd spectra of 4.0 mg/mL FA solutions containing approximately 3 mM Cd(II). Samples and measured line widths are (A) LSFA, 893 Hz, (B) WRFA, 385 Hz, (C), CLFA, 233 Hz, (D) SRFA, 338 Hz. All spectra were processed using 80 Hz line broadening. The peak at -1.78 ppm arises from the internal sealed capillary used for referencing.

Results Cadmium Titration. A single 113Cd resonance has been observed in prior 113Cd NMR investigations of Cd(II)-fulvic acid complexes (11, 12, 20), indicating 113Cd exists in relatively fast exchange on the NMR time scale, hence the measured chemical shift is a weighted average dependent on the chemical shift of the free and complexed species. In this study, Cd(II) is titrated into a fulvic acid solution, and the 113Cd chemical shift is used to determine the fraction of free and bound Cd(II). Binding studies of Cd(II) with four different fulvic acid complexes are reported. Subsequent competitive binding studies in which added Ca(II) displaces Cd(II) from the fulvic acid complex are also discussed. In all cases, a single cadmium resonance is observed. Figure 1 shows 113Cd spectra measured for each fulvic acid in solutions containing approximately 3 mM Cd(II) and 4.0 mg/mL fulvic acid. In these spectra the cadmium resonance appears as a broad peak with an observed chemical shift, δobs, between -16 and -20 ppm. This single peak indicates that fast exchange exists for each Cd(II)-fulvic acid solution. Cd(II) titrations were started at 1 mM as this was the lowest 113Cd concentration for which an adequate spectral signal-to-noise ratio could be obtained with our instrument. Addition of Cd(II) aliquots continued until precipitation was observed. Figure 2 shows a stacked plot of spectra measured for the SRFA, illustrating the regular movement of δobs as a function of [Cd2+] toward δfree (0 ppm), indicative of fast exchange. When the population of free Cd(II) increases, the cadmium resonance line width narrows as a result of the longer spin-spin relaxation time of the free Cd(II). However, a change in the relative proportion of Cd(II) species in fast, intermediate, and slow chemical exchange may be an additional contribution. 1464

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

FIGURE 2. Stacked plot of spectra measured for the Cd(II) titration with Suwannee River reference FA. The spectra are processed with 50 Hz of line broadening and normalized with respect to peak height. The Cd(II) concentrations and measured line widths are (A) 1.81 mM, 354 Hz, (B) 2.7 mM, 308 Hz, (C) 3.6 mM, 266 Hz, (D) 4.5 mM, 254 Hz, (E) 5.4 mM, 232 Hz, (F) 6.3 mM, 226 Hz, (G) 7.2 mM, 224 Hz, (H) 8.3 mM, 201 Hz. Metal Ion-Fulvic Acid Binding Constants. The analysis of the Cd(II) titration results is an extension of a prior investigation of complex formation determination using NMR spectroscopy reported by Carper et al. (21). In the case of fast exchange, one may write the equation

δobs ) [metal ionbound]δbd/[metal iontotal] + ([metal ionfree]δfree)/[metal iontotal] (1) The above equation can be rewritten to incorporate the following relationship

[metal ionbound] + [metal ionfree] ) [metal iontotal] [metal ionbound]/[metal ionfree] ) (δobs - δfree)/(δbd - δobs) ) Kf(n[total acceptor] - [metal ionbound]) (2) which represents 1:1 binding at n independent, similar sites. Consequently, a plot of (δobs - δfree)/(δbd - δobs) vs [metal

FIGURE 3. Stacked plot of Ca(II) titration spectra using Clinton Lake FA processed with 20 Hz line broadening. The Ca(II) concentrations and measured line widths are (A) 0 mM, 199 Hz, (B) 0.98 mM, 177 Hz (C) 2.92 mM, 158 Hz, (D) 4.86 mM, 147 Hz (E) 7.8 mM, 152 Hz, (F) 9.7 mM, 137 Hz.

FIGURE 5. Graph of (δobs-δfree)/(δbd-δobs) versus the calculated concentration of complexed Cd(II) for SRFA (9), CLFA (b), LSFA (2), and WRFA (1). The slope is equal to -KCd and the y-intercept is equal to nKCdFA.

TABLE 1. Summary of Results of Cadmium and Calcium Titrations To Determine δbd, Binding Constants, and the Concentration of Binding Sites, n[FATotal] sample SRFA WRFA CLFA LSFA

FIGURE 4. Graph of the observed chemical shift versus the total Cd(II) concentration for SRFA (9), CLFA (b), LSFA (1), and WRFA ((). The y-intercept gives the value used for δbd for each fulvic acid. ionbound] has a slope equal to -Kf and a y-intercept of nK[total acceptor]. This allows the investigator to determine Kf and n from the same plot. Measuring all chemical shifts relative to a solution of the free metal ion, δfree, and setting δfree equal to zero simplifies the calculations. For a metal titration where the acceptor concentration is held constant, the determination of δbd may be accomplished by plotting δobs vs the metal ion concentration and extrapolating to the point that the metal concentration is zero. This is simply the y-intercept of such a plot (Figure 4). The case of 1:1 noncompetitive binding of 113Cd(II) to fulvic acid is represented by

(δobs - δfree)/(δbd - δobs) ) R ) [Cdbound]/[Cdfree] ) Kf(n[FAtotal] - [Cdbound]) (3) where Kf ) [Cdbound]/[Cdfree](n[FAtotal] - [Cdbound]). Thus, plotting (δobs - δfree)/(δbd - δobs) vs [Cdbound] yields the overall KCd for a particular fulvic acid from the slope and n[FAtotal] from the y-intercept (Figure 5). Although this type of graphical analysis can produce incorrect or misleading results if not used properly, our results fall within a range appropriate for this technique as proposed by Klontz (22, 23). Although the binding site heterogeneity of fulvic acids usually requires complex models such as Model V or the NICA-Donnan

δbd -20.53 -18.33 -17.47 -25.16

n[FAtotal] 10-3

7.6 × 6.4 × 10-3 6.6 × 10-3 3.5 × 10-3

KCd

KCa ×103

2.6 ( 0.2 1.9 ( 0.2 ×103 1.2 ( 0.1 ×103 3.4 ( 0.3 ×103

0.65 ( 0.03 ×103 0.46 ( 0.02 ×103 0.50 ( 0.02 ×103 0.78 ( 0.08 ×103

models (24) to describe the interactions, the rapid exchange observed in these measurements means the Cd(II) itself effectively averages the different binding sites allowing the average binding constant to be determined from the concentration dependence of the 113Cd NMR chemical shift. A limitation to this approach for a system with heterogeneous binding sites is that at low occupancy only the strongest sites will participate in binding; therefore, a deviation from linearity may occur at very low Cd:FA ratios. The bound 113Cd chemical shifts, KCd values for each fulvic acid as well as the n[FAtotal], are listed in Table 1. The value n[FAtotal] is representative of the concentration of the ligand binding sites. Since the ligand, in this case fulvic acid, is a complex heterogeneous mixture, the true concentration cannot be simply delineated. However, if one assumes two binding sites per FA molecule (n ) 2) for the 4.0 mg/mL FA used in these experiments, the average molecular weight calculated for the three aquatic fulvic acids (CLFA, SRFA, and WRFA) would be between 1000 and 1250 from the determined n[FAtotal]. These values fall in the reported molecular weight ranges. Similarly, the calculated molecular weight for the LSFA assuming n ) 2 is 2100, a reasonable molecular weight average for a soil fulvic acid. From the determined n[FAtotal] and the quantity of the fulvic acid used, the LSFA would have 0.93 mmol/g of bidentate carboxylate binding sites assuming bidentate binding. This value is much lower than the reported 5.6 mequiv/g of Cu bidentate binding sites (18). However, stoichiometrically the 0.93 mmol/g makes sense for this set of experiments, as the titration ended at 10 mM Cd(II), thus the Cd:FA ratio is approximately 2. Hence for this set of experiments, all of the binding sites are not being saturated, and only the two strongest binding sites are examined. Thus the average reported binding constants are on average for the two strongest binding sites. Calcium Titration. For the Ca(II) titration, the FA concentration was the same as for the Cd(II) titration, 4.0 mg/mL. To observe displacement of the bound Cd(II), the Cd(II) concentration was chosen so that most of the Cd(II) VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1465

FIGURE 6. Graph of SRFA (9), CLFA (b), and WRFA (() calcium titrations assuming that the concentration of displaced Cd(II) is equal to the concentration of the bound Ca(II). in solution was bound; however, a sufficiently high concentration was needed so a 113Cd spectrum could be measured in a reasonable amount of time (∼24 h). Thus 3 mM Cd(II) was used. Addition of Ca(II) was continued until observable precipitate formed. Figure 3 shows a stacked plot of representative spectra collected using the CLFA as a function of Ca(II) addition. From this series of spectra, it can be observed that over the course of the Ca(II) titration the 113Cd chemical shift and the resonance line width becomes more like that of the free Cd(II), similar to what was observed in the Cd(II) titration. Metal Ion-Fulvic Acid Binding Constants through Competition. In the case of 1:1 competitive binding at n independent and similar sites where Ca(II) competes with Cd(II), it is possible to derive an equation that provides relative formation constants for metal-fulvic acid complexes. Since the 113Cd nuclei are in fast exchange, eqs 1 and 2 are still valid. The only change in eq 3 concerns the definition of KCd

KCd ) [Cdbound]/[Cdfree](n[FAtotal] - [Cdbound] - [Cabound]) (4) Equation 4 now contains a correction for the loss of fulvic acid binding sites to Ca(II). The binding constant for calcium, KCa, can be written in a manner similar to KCd with one modification, designating n′ as the number of binding sites for Ca(II), and assuming that both n and n′ sites can be occupied by either Cd(II) or Ca(II). Changes in binding geometry, energetics, etc. can then be considered.

Kf(Ca) ) [Cabound]/ [Cafree](n′[FAtotal] - (n′/n)[Cdbound] - [Cabound]) (5) If n ) n′, then eq 6 holds

([Cdbound]/[Cdfree])/([Cabound]/[Cafree]) ) KCd/KCa (6) Thus a plot of ([Cdbound]×[Cafree]) vs ([Cdfree]×[Cabound]) would give a slope equivalent to KCd/KCa. If all the Ca(II) that binds displaces a Cd(II), then [Cabound] would be equivalent to [Cddisplaced]. From Figure 6, a plot using the assumption that [Cabound] is equal to [Cddisplaced], it is apparent that although the WRFA and CLFA fit this model reasonably well, the SRFA does not. Close examination of each plot shows a nonlinear portion at the lowest concentrations of Ca(II). Thus, it is apparent from the experimental data that one cannot assume that [Cabound] equals [Cddisplaced]. However, an appropriate n[FAtotal] can be calculated for the Ca(II) titration solution 1466

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

FIGURE 7. Graph of Ca(II) competition experiments measured for SRFA (9, 2.90 mM Cd(II)), CLFA (b, 2.93 mM Cd(II)), LSFA (1), 2.90 mM Cd(II)), and WRFA ((, 2.86 mM Cd(II)) at a concentration of 4.0 mg/mL fulvic acid. Equation 7 is used to determine the concentration of the bound Ca(II). prior to Ca(II) addition and using eq 3, then subsequently determine [Cabound] by eq 7.

[Cabound] ) n[FAtotal] - (R/KCd) - [Cdbound]

(7)

Again if n′ ) n, then a plot of ([Cdbound]×[Cafree]) vs ([Cdfree]× [Cabound]) should be linear with a slope of KCd/KCa. This new method of calculating the [Cabound] is utilized for the plot in Figure 7, with the linear fit showing that n′ ) n is a good approximation for the case of Ca(II) competition with Cd(II) for fulvic acid binding sites. The determined KCa for each fulvic acid investigated is reported in Table 1.

Discussion Binding Sites. In the case of 113Cd NMR, Cd(II) complexed through nitrogen atoms generally have chemical shifts above 100 ppm, while Cd(II) bound through oxygen tend to have chemical shifts below 0 ppm (13). For the Cd(II) bound to all four different fulvic acids in our experiments, the bound chemical shift is near -20 ppm. This indicates the Cd(II) predominately binds to fulvic acid through oxygen components, in particular carboxylate moieties. This result is consistent with the predominance of carboxylate moieties associated with these fulvic acids and a previous study that measured the signal attenuation in 13C CP-MAS spectra upon binding of the paramagnetic ion, Cu(II), that indicated that Cu(II) binds predominately through carboxylate groups as well as carbohydrate moieties (17). Other oxygen containing binding sites are likely part of the strong acid sites identified by Leenheer et al., mainly carboxyl groups associated with cyclic aliphatic R-ester and R-ether structures (25). The bound cadmium chemical shift can reflect the chemical nature of the carboxylate ligands. Cadmium bound to a single carboxylate ligand has a chemical shift between -10 and -23 ppm, dicarboxylate cadmium complexes have chemical shifts of -30 and -40 ppm, and the chemical shift of a bidentate Cd(II)-hydroxycarboxylate is between -20 and -30 ppm (20, 26). From Table 1, an interesting correlation between the strength of the binding constant and the upfield chemical shift of the bound cadmium is observed suggesting that the fulvic acids with a stronger binding constant have a greater population of the stronger hydroxycarboxylate or dicarboxylate binding sites. As illustrated in Figure 1, there are differences in the 113Cd resonance line width in each of the four fulvic acid solutions with a direct correlation between the binding strength and the line width. Although detailed chemical shift information about the nature of 113Cd binding sites is available in the

TABLE 2. Characteristic Properties of the Fulvic Acids Studied LSFA %C %O %H %N

SRFA

Elemental Analyses 45.1 53.8 49.7 40.9 4.1 4.3 1.1 0.7

WRFA

CLFA

53.38 39.15 5.50 1.64

53.71 38.36 5.85 1.69

19 1

16 2

13C

%carboxylic %phenolic carboxylic mequiv/g phenolic mequiv/g Rcoo- at pH 6.4

NMR Data 33.8 19 8.8 6

Titration Data 8.6 6.4 3.0 1.2 0.74 0.89

6.1 1.7 0.91

6.2 1.8 0.91

literature, we were unable to locate information regarding relaxation parameters for the binding sites mentioned above. However, it is likely that Cd(II) bound to a bidentate site would be more rotationally hindered than Cd(II) bound at a monodentate site. Thus the T2 of the bidentate bound 113Cd would be expected to be shorter than 113Cd bound at a monodentate site. Hence the apparent differences in line width may correlate with the presence of a larger population of the stronger hydroxycarboxylate or dicarboxylate binding sites, in agreement with the conclusions drawn from the trends in the bound chemical shifts. Cadmium Binding Constants. The measured binding constants support the hypothesis that Cd(II) binds predominately through carboxylate functionalities because the KCd and KCa are both close in magnitude to other simple carboxylate ligands such as acetate, oxalate, and citrate. For acetate the KCd is 79 and KCa is 16, with oxalate as a ligand KCd is 7.9 × 103 and KCa is 1.0 × 103, while for citrate the KCd is 1.6 × 104 and KCa is 1.6 × 103 (27). It should be noted that for each of these ligands KCd is larger than the KCa. The measured KCd values are also comparable to values obtained in other investigations using humic substances. A study of a soil fulvic acid from the Okchun Metamorphic Belt yielded an apparent KCd of 7.9 × 103 measured using ultrafiltration (28). This compares favorably with our measured binding constant of 3.4 × 103 for the Laurentian soil fulvic acid. A report of a KCd of 4.7 × 103 for an aquatic fulvic acid at pH 6.0 using an ion selective electrode (29), agrees favorably with our KCd using the aquatic fulvic acids. Using Schubert’s ion-exchange method, Brown determined log K values of 4.42 for Cd(II) and 3.02 for Ca(II) (30). These values will only be relevant when examining the binding of a single metal species at very low metal-to-fulvic ratios, as under the conditions of the Schubert’s ion exchange method the total concentration of metal ions must be negligible compared to the ligand concentration. Since FA is a heterogeneous mixture, Schubert’s ion exchange method effectively samples the strongest FA binding sites, whereby our experiments produce a binding constant averaged over all binding sites accounting for the factor of 10 difference between the values of Brown and those determined in this work. Further differences between the two studies could also arise from differences in ionic strengths used during the two measurements. Although the chemical composition of a fulvic acid varies depending upon the source of the fulvic acid, all four of our apparent binding constants are within a factor of 2. This suggests that there is a homology in Cd(II) complexation by fulvic acids. From Table 2, many bulk characteristics of these fulvic acids have been previously reported (15-18). From the characteristic properties reported such as nitrogen and carboxylate content, there is no obvious relationship between

the trends in binding constants and any of the characteristic properties. Thus differences in the binding constants must arise from differences in macromolecular structure. One interesting trend is that the apparent binding constant decreases regularly relative to the environmental age of a material. The SRFA and LSFA are both sites where decomposition is occurring, thus these fulvic acids should be relatively “new”. The CLFA and WRFA arise predominately from leachate through the soil, thus are more aged by the time they reach the isolation sites, or the stronger binding fractions of the fulvic acid are retained bound to the soil. The trend with age is likely due to oxidation of stronger binding sites as the FA ages, resulting in weaker binding overall. Competitive Binding. The apparent KCa values were all an order of magnitude lower than the measured KCd. Environmentally this is important because in the Ca(II) competition studies the calcium does not effectively displace bound Cd(II) until many of the FA binding sites are populated. Furthermore even a 10-fold excess of Ca(II) would displace less than half the bound Cd(II), indicating that some binding sites have a strong preference for Cd(II). Finally, utilizing the fast exchange of Cd(II) on the NMR time scale, a novel approach to obtaining the apparent binding constants of Ca(II) with FA using NMR competition studies has been demonstrated. Although due to the limitations of NMR, this approach is not applicable to paramagnetic metal ions, this technique can be utilized for any other diamagnetic metal ions or mixtures of metal ions and therefore is useful for the assessment of relative metal binding affinities.

Acknowledgments This work was supported by a NSF-EPA Waters and Watersheds grant, CHE-9524514 and EPA-EPSCoR grant R82758901-0. The LSFA was provided by Dr. Cooper Langford and Dr. Robert Cook.

Literature Cited (1) Carter, C. W.; Suffet, I. H. Environ. Sci. Technol. 1982, 16, 735740. (2) Magee, B. R.; Lion, L. W.; Lemley, A. T. Environ. Sci. Technol. 1991, 25, 323-331. (3) Chakrabarti, C. L.; Lu, Y.; Gregoire, D. C.; Back, M. H.; Schroeder, W. H. Environ. Sci. Technol. 1994, 28, 1951-1967. (4) Winner, R. W. Aquat. Toxicol. 1984, 5, 264-274. (5) Campbell, J. H.; Evans, R. D. Sci. Total Environ. 1987, 62, 219227. (6) Holloway, J. H.; Reilley, C. N. Anal. Chem. 1960, 32, 249-256. (7) Gamble, D. S.; Langford, C. H.; Tong, J. P. K. Can. J. Chem. 1976, 54, 1239-1245. (8) McKnight, D. M.; Wershaw, R. L. In Humic Substances in the Suwannee River, Georgia: Interactions, Properties and Proposed Structures; U.S. Geological Survey Open File Report 87-557; Averett, R. C., Lennheer, J. A., McKnight, D. M., Thorn, K. A., Eds.; U.S. Geological Survey: Denver, 1989; pp 61-79. (9) O’Shea, T. A.; Mancy, K. H. Water Res. 1978, 12, 703-711. (10) Sadler, P. J.; Viles, J. H. Inorg. Chem. 1996, 35, 4490-4496. (11) Larive, C. K.; Rogers, A.; Morton, M.; Carper, W. R. Environ. Sci. Technol. 1996, 30, 2828-2831. (12) Li, J.; Perdue, M.; Gelbaum, L. T. Environ. Sci. Technol. 1998, 32, 483-487. (13) Summers, M. F. Coord. Chem. Rev. 1988, 86, 43-134. (14) Bates, R. G. Determination of pH: Theory and Practice; Wiley: New York, 1964; pp 219-220. (15) Pommes, M. L.; Larive, C. K.; Thurman, E. M.; Reed Green, W.; Orem, W. H.; Rostad, C. E.; Coplen, T. B.; Cutak, B. J.; Dixon, A. M. Environ. Sci. Technol. 2000, 34, 4278-4286. (16) Malcolm, R. L.; McKnight, D. M.; Averett, R. C. In Humic Substances in the Suwannee River, Georgia: Interactions, Properties and Proposed Structures; U.S. Geological Survey Water Supply Paper 2373; Averett, R. C., Lennheer, J. A., McKnight, D. M., Thorn, K. A., Eds.; U.S. Geological Survey: Denver, 1994; pp 13-19. (17) Cook, R. L. Ph.D. Dissertation, University of Calgary, 1997. VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1467

(18) Wang, Z.; Gamble, D. S.; Langford, C. H. Environ. Sci. Technol. 1992, 26, 560-565. (19) Kostelnik, R. J.; Bothner-By, A. A. J. Magn. Reson. 1974, 14, 141151. (20) Chung, K. H.; Rhee, S. W.; Shin, H. S.; Moon, C. H. Can. J. Chem. 1996, 74, 1360-1365. (21) Carper, W. R.; Buess, C. M.; Hipp, G. R. J. Phys. Chem. 1970, 74, 4229-4234. (22) Klontz, I. M. Science 1982, 217, 1247-1249. (23) Munson, P. J.; Rodbard, D. Science 1983, 220, 979-981. (24) Christensen, J. B.; Tipping, E.; Kinniburgh, D. G.; Gron, C.; Christensen, T. H. Environ. Sci. Technol. 1998, 32, 3346-3355. (25) Leenheer, J. A.; Wershaw, R. L.; Reddy, M. M.. Environ. Sci. Technol. 1995, 29, 399-405.

1468

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001

(26) Chung, K. H.; Moon, C. H. J. Chem. Soc., Dalton Trans. 1996, 75-78. (27) Ramette, R. W. Chemical Equilibrium and Analysis; AddisonWesley: MA, 1981; pp 727-728. (28) Lee, M. H.; Choi, S. Y.; Moon, H. Bull. Korean Chem. Soc. 1993, 14, 453-457. (29) Saar, R. A.; Weber, J. H. Can. J. Chem. 1979, 57, 1263-1268. (30) Brown, G. K. M.S. Thesis, Colorado School of Mines, 1996.

Received for review December 13, 1999. Revised manuscript received January 5, 2001. Accepted January 5, 2001. ES991372E