Characterizing the Interactions between Trace Metals and Dissolved

Aug 29, 2008 - Chemistry and Biochemistry, Florida International University,. Miami .... 7374 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO...
0 downloads 0 Views 964KB Size
Download PDF
Environ. Sci. Technol. 2008, 42, 7374–7379

Characterizing the Interactions between Trace Metals and Dissolved Organic Matter Using Excitation-Emission Matrix and Parallel Factor Analysis YOUHEI YAMASHITA AND ´* RUDOLF JAFFE Southeast Environmental Research Center, and Department of Chemistry and Biochemistry, Florida International University, Miami, Florida 33199

Received May 16, 2008. Revised manuscript received July 15, 2008. Accepted August 3, 2008.

Natural dissolved organic matter (DOM) is composed of a variety of organic compounds, which can interact with metals in aquatic environments. The interactions between DOM and two metals of environmental concern (Cu(II) and Hg(II)) were studied using fluorescence quenching titrations combined with excitation-emission matrix (EEM) spectra and parallel factoranalysis(PARAFAC).Thisallowedcharacterizingthespecific interactions between eight fluorescent components in DOM and two metals. Triplicate titration experiments showed good reproducibility when assessing the interactions between humiclike components with Cu(II). Our data show clear differences in metal-DOM interaction for samples of different DOM composition and between two different metals. The results demonstrate that the combination of fluorescence quenching titrations with EEM-PARAFAC was reproducible and sensitive to determine the binding properties of humic-like components with trace metals. The enhancement in fluorescence intensity after its initial decrease for the protein-like components with addition of Cu(II) was observed at mangrove-dominated sites, suggesting changes in the molecular environments of proteinlike components due to increased Cu(II) interaction. The application of EEM-PARAFAC in fluorescence quenching studies is a useful tool to evaluate intermolecular DOM and DOM–trace metals interactions.

Introduction Metal toxicity, bioavailability and mobility in the environment are controlled by their speciation (1-3). The composition of dissolved organic matter (DOM) in diverse aquatic environments has been reported to be quite variable (4) and can interact with metal ions to form organometal complexes, thus strongly affecting the metal speciation. Therefore, knowledge on the metal-DOM binding properties is necessary to better understand the biogeochemistry of metals. Even though not all of the organic ligands do fluoresce, it is well-known that fluorescent components in DOM are quenched by metals, mainly due to static quenching and possibly through collisional (dynamic) quenching (5, 6). Thus, metal-binding parameters of DOM have often been studied * Corresponding author phone: 305-348-2456; fax: 305-348-4096; e-mail: [email protected]. 7374

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

by fluorescence-quenching methods (5-12). This technique traditionally used single excitation and emission wavelengths, which are typical of humic-like fluorophores (7, 8). However, this technique is not ideally suited to determine binding properties to DOM, since quenching of bulk fluorescence properties will not account for differences in the metal interactions with many different types of fluorophores coexisting in natural DOM. To address the issue of specificity, synchronous fluorescence and excitation-emission matrix (EEM) fluorescence have been combined with quenching methods (5, 6, 9-12). These multidimension fluorescence techniques have demonstrated that changes in fluorescence intensity due to metal interaction were different among the different DOM components. Thus, different peaks in synchronous fluorescence spectra or different regions of the EEM spectra (i.e., pairs of excitation and emission wavelengths) were found to be quenched to a different degree, clearly suggesting various degrees of interaction (5, 9, 12). However, since the synchronous fluorescence spectra and EEMs are often composed of various type of overlapping fluorophores, better resolution is required to obtain accurate binding parameters. Recently, Ohno et al. (13) used EEM fluorescence and parallel factor analysis (PARAFAC), which statistically decompose EEMs into different independent groups of fluorescent components (14-17), in combination with fluorescence quenching to assess the interaction of three forest soil-derived humic-like and fulvic-like components with iron(III) and aluminum(III). Thus, combining fluorescence quenching titrations with EEM-PARAFAC is a potentially useful technique in the assessment of interaction constants between dissolved metals and DOM. It should be noted that DOM from surface waters is not only likely to contain a higher variety of fluorescent groups compared to soil-derived DOM (e.g. (15-17),), but also that the number of samples in the data set will affect the number of components obtained through PARAFAC modeling (16). In fact, recent EEM-PARAFAC models for DOM have identified in addition to humic type components microbially derived humic-like and protein-like components, leading to a substantial number of identifiable fluorescent components (7-13 components) in DOM from various aquatic environments (15-17). The application of a combination of a large EEM-PARAFAC modeling database and fluorescence quenching to study the interaction of trace metals with DOM in aquatic environments has not been performed and could provide additional insight into the biogeochemistry of trace metal complexation. The objectives of this study were (1) to test the application of the combined techniques of fluorescence quenching titration with EEM-PARAFAC and to evaluate the analytical reproducibility of the method; (2) to investigate the interaction between different PARAFAC modeled DOM components with trace metals with different complexing characteristics using DOM samples with different composition; and (3) to evaluate possible changes in the molecular environment of the DOM mixture resulting from enhanced metal complexation. For objective 1, triplicate fluorescence quenching titrations were performed to examine the sensitivity and reproducibility of this technique. For objective 2, surface water samples were collected from various aquatic ecosystems with reported differences in DOM composition (18), and copper (semihard metal) and mercury (soft metal) were used as quenching agents for the titrations. Objective 3 was achieved through observed DOM component pattern changes throughout the quenching experiments. 10.1021/es801357h CCC: $40.75

 2008 American Chemical Society

Published on Web 08/29/2008

TABLE 1. Salinity, pH and Fluorescence Intensity (QSU) of Eight Components in DOMa June SRS2 salinity pH component component component component component component component component total

1 2 3 4 5 6 7 8

SRS4

October SRS6

TS2

SRS2

0 6.5 23.8 0 0 7.7 7.6 7.5 7.3 7.0 145 (24) 128 (23) 119 (23) 52 (25) 166 (24) 134 (22) 141 (26) 135 (26) 46 (22) 151 (22) 109 (18) 92 (17) 80 (16) 27 (13) 135 (20) 77 (13) 63 (11) 59 (11) 28 (14) 89 (13) 41 (7) 58 (10) 58 (11) 17 (8) 55 (8) 31 (5) 38 (7) 37 (7) 12 (6) 41 (6) 44 (7) 15 (3) 14 (3) 8 (4) 22 (3) 26 (4) 16 (3) 14 (3) 17 (8) 25 (4) 606 550 515 207 683

a Relative abundance of fluorescence intensity of each of the component in total fluorescence (%) are also shown in parentheses.

Experimental Section Sampling Sites and Sample Treatment. The surface water samples containing the DOM were obtained from two slough systems; that is, Shark River Slough (SRS) and Taylor Slough (TS), in the Florida Costal Everglades (FCE, Supporting Information) during the wet season (June-October, 2007). Two of the four sampling sites were located in freshwater marshes dominated by peat (SRS2) and marl (TS2) soils, and others were collected at mangrove-dominated brackish sites (SRS4 and SRS6). The molecular composition and optical characteristics of DOM for these sampling sites can be found elsewhere (18, 19). The samples were collected using precleaned brown high-density polyethylene bottles, stored on ice, and filtered through 0.22 µm filters (Durapore, Millipore) in the laboratory to be stored under refrigeration until use. Salinity of water was measured in the field using a YSI 30 salinity, temperature, conductivity meter (model 30/10FT). Salinities of the sampling sites are summarized in Table 1. Fluorescence Titration. Fluorescence-quenching titrations were conducted according to Lu and Jaffe´ (10) and Fu et al. (12). Experiments were carried out by adding 0.01 mol L-1 Cu(NO3)2 or Hg(NO3)2 to a series of brown high-density polyethylene bottles that contained 20-60 mL of the 0.22 µm filtrates. All water samples after addition of Cu(II) or Hg(II) were shaken for 24 h at room temperature to ensure complexation equilibrium. EEM fluorescence spectra were obtained using a Horiba Jovin Yvon SPEX Fluoromax-3 fluorimeter (18). Several postacquisition steps were involved in the correction of fluorescence spectra. First, the UV-visible absorption spectra, measured with a Varian Cay-50 bio spectrophotometer in a 1 cm quartz cuvette, were used for inner filter corrections according to McKnight et al. (20). After this procedure, the EEM of Milli-Q water was subtracted from sample EEMs. Second, the excitation and emission correction files supplied by the manufactures were applied for correction of specific instrument component. Fluorescence intensities were also corrected to the area under the water Raman peak (excitation ) 350 nm), analyzed daily (15), and then were converted to quinine sulfate units using a calibration with quinine sulfate monohydrate (21). Finally, the blank EEMs, containing only Cu(NO3)2 or Hg(NO3)2 in Milli-Q water, were corrected using the same procedure as for the samples and then subtracted from corrected sample EEMs. PARAFAC Modeling. The approach of PARAFAC modeling of EEMs has been described in detail elsewhere (14, 22). The PARAFAC modeling was carried out in MATLAB 7.0.4 (Mathworks, Natick, MA) with the N-way toolbox 2.1.0 (23). The FCE-PARAFAC model (see the Supporting Information)

was forced to the EEMs obtained from fluorescence quenching titration experiments. This model was obtained from a large database (n ) 1108 surface water samples) collected from the FCE. On the basis of the comparison with fluorescencecharacteristicswithpreviouslyreportedPARAFAC studies (15-17) as well as with traditional EEM source assignments (24), the eight components determined by the FCE-PARAFAC model were categorized as four terrestrial humic-like, two microbial humic-like, and two protein-like fluorescent components (Figure 1 and Supporting Information). Complexation Modeling. The complexation model reported by Ryan and Weber (7) and modified by Plaza et al. (25) was used to determine the binding parameters between fluorescent components derived from PARAFAC and trace metals (Cu(II) and Hg(II)). This model is based on the assumption of the formation of 1:1 complexes between ligands and trace metals. Since PARAFAC decomposes the complex mixture of DOM fluorophores into independent fluorescent components, the application of this model to PARAFAC components is more appropriate than to the fluorescence intensity derived from peak maxima of synchronous and EEM spectra. The binding properties were determined following a nonlinear model (25): I ) I0 + (IML - I0)

(

)

1 (1 + KMCL + KMCM 2KMCL

√(1 + KMCL + KMCM)2 - 4KM2CLCM)

(1)

where I and I0 are the fluorescence intensity at the metal concentration CM and at the beginning of titration (in the absence of added metals), respectively. IML is the limiting value below which the fluorescence intensity does not change due to the addition of metal. KM and CL are the conditional stability constant and total ligand concentration, respectively. SigmaPlot (SPSS Inc., Chicago, IL) was used for solve for IML, KM, and CL. In addition, the fraction of the initial fluorescence that corresponds to the binding fluorophores (f) was determined using eq 2. f )

(I0 - IML) × 100 I0

(2)

Using this nonlinear model, extremely small CL values were estimated. This data may result from the low levels of binding sites, the values of the conditional stability constants (26), or both and were in agreement with observations from previous studies (6, 9, 23, 26, 27). Therefore, the CL values were not reported in the present study.

Results and Discussion Fluorescence Properties of DOM at the Four Sampling Sites. The fluorescence intensity of eight components derived from the FCE-PARAFAC model (Figure 1) and the relative abundance of each of the component to total fluorescence are summarized in Table 1. The most abundant components were terrestrial humic-like components 1 and 2, irrespective of differences in sampling sites (Table 1). The relative abundance of two microbial humic-like components 4 and 5, as well as two protein-like components 7 and 8, were comparatively lower by up to 1 order of magnitude. Such fluorescence characteristics indicate the strong terrestrial/ higher plant origin of the fluorescent DOM and were consistent with previous results of optical analysis at FCE locations (18). The relative abundances of protein-like component 8 at TS2 and component 7 at SRS2 in June were higher than those at the other sites. Total fluorescence intensity at SRS2 was higher in October as compared to June. In June, total fluorescence intensity was highest at SRS2 and decreased with increasing salinity at the SRS location. Total VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7375

FIGURE 1. EEM contours of the eight components identified by the FCE-PARAFAC model.

FIGURE 2. Representative plots of changes in fluorescence intensity of humic-like components with the addition of Cu(II) and Hg(II): (A) terrestrial humic-like component 2 at four sampling sites with Cu(II), (B) terrestrial humic-like component 2 at four sampling sites with Hg(II), and (C) six humic-like components at site SRS2 (July) with Cu(II) and (D) six humic-like components at site SRS2 (July) with Hg(II). fluorescence intensity at TS2 was considerably lower than those at SRS sites, suggesting quantitative differences between marl- and peat-dominated wetland sites. Interaction of Humic-Like Components with Trace Metals. The triplicate fluorescence quenching experiments show excellent reproducibility of the quenching curves for each of the humic-like fluorescent components (see Supporting Information, Figure S4), even though initial fluorescence intensities were largely different (Table 1). The log K and f values obtained from triplicate titration experiments show small errors, irrespective of humic-like component types (the log K and f values were 4.83 ( 0.03 and 46 ( 0.9, 4.75 ( 0.03 and 51 ( 1.1, 4.67 ( 0.10 and 35 ( 2.6, 5.06 ( 0.04 and 29 ( 1.0, 5.10 ( 0.09 and 31 ( 1.0, and 5.21 ( 0.07 and 32 ( 1.2 for components 1-6, respectively). These results strongly indicate that the combined technique of fluorescence quenching titration with EEM-PARAFAC is reproducible for 7376

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

evaluating fluorophore-specific interactions between trace metals and DOM in aquatic environments. Figure 2 shows representative fluorescence quenching curves of each of the terrestrial and microbial humic-like components with Cu(II) and Hg(II). Since initial levels of fluorescence intensity were largely different among components as well as among samples (Table 1), the fluorescence quenching curves were shown as percent changes from initial levels (F/F0 × 100, where F and F0 are the fluorescence intensity with and without trace metals, respectively). The fluorescence quenching curves of terrestrial humiclike component 2 with Cu(II) and Hg(II) were different among sampling sites as well as between trace metals (Figure 2A and B). Although the degree of quenching with Cu(II) was larger than that with Hg(II), the degree of quenching was the largest at marl-based soil freshwater site (TS2) and the

FIGURE 3. f Values determined by the Ryan and Weber model (A) with Cu(II) and (B) with Hg(II). The decreases in fluorescence intensity with Hg(II) of component 3 at SRS 4 and 6 were not modeled. smallest at the brackish sites (SRS4 and 6), irrespective of trace metal type. Different quenching curves for different DOM samples as well as different trace metals have been observed in the previous studies using fluorescence quenching titration with single excitation and emission wavelengths and synchronous and EEM spectra (6, 10, 12, 22, 25). However, the fluorescence quenching titration using EEM-PARAFAC allows for the assessment of metal interactions with specific fluorophores within one sample, as shown when comparing among humiclike components at site SRS2 (Figure 2C and D). Interestingly, although the largest fluorescence quenching was observed in terrestrial components 1 and 2 with Cu(II), terrestrial humic-like component 6 and microbial humic-like component 5 were most strongly quenched with Hg(II). It should be noted that differences in the quenching degree among humic-like components within one sample were in the same range as those found among different DOM samples (Figure 2). The calculated log K values for the six humic-like components found in the FCE ranged from 4.48 to 6.32 and from 3.92 to 6.76 for Cu(II) and Hg(II) titration experiments respectively (see the Supporting Information, Table S2). These log K values were similar in range to those found for bulk DOM (8, 12, 28) and humic substances (6, 25, 26) using fluorescence quenching titration with Cu(II) and Hg(II). The log K values of components 5 (4.71-5.10) and 6 (5.13-5.45) for Cu(II) were in the same range as those found in the corresponding EEM wavelength range for fluorescence quenching of leaf litter extracts (regions B and C, respectively, in ref 9) with Cu(II). On the basis of emission characteristics, components 1, 2, and 4 (Figure 1) resembled peaks IV, III, and II in synchronous fluorescence spectra reported for the FCE (10). The log K values of Hg(II) for these PARAFAC components (see the Supporting Information, Table S2) and for the synchronous peaks (10) were similar, even though the values obtained for the PARAFAC components were more variable. This is likely as result of the higher resolution of EEM-PARAFAC as compared to synchronous fluorescence spectra. There is no systematic trend of log K values among the sampling sites, as well as the DOM components (see the Supporting Information, Table S2). However, it is interesting to note that the log K values of humic-like components for Hg(II) at SRS6 (4.14 ( 0.15, n ) 5) were significantly lower than those at other sites (4.95 ( 0.52, n ) 17), but those for Cu(II) at SRS6 (5.26 ( 0.63, n ) 5) were similar to those at other sites (4.93 ( 0.33, n ) 16). Such differences could be explained by reduced complexation between Hg(II) and DOM due to the enhanced effect on inorganic complexation between Hg(II) and Cl (10) at this high-salinity site (Table 1). The f values of the six humic-like components found in the FCE ranged from 13 to 61 and from 11 to 49 for Cu(II) and Hg(II), respectively (Figure 3). The f values for Hg(II)

were in a similar range but relatively smaller as compared to those reported for the FCE, as determined by synchronous fluorescence (10). There were several interesting points on the spatial and intercomponents differences in the f values. First, the f values of terrestrial humic-like components 1, 2, and 3 with Cu(II) were always higher than those with Hg(II), irrespective of differences in sampling sites. On the other hand, the f values for the terrestrial humic-like component 6 with Cu(II) was always lower than those with Hg(II). Such a difference suggests that the binding sites and, thus, the complexation processes are different among terrestrial humic-like components 1, 2, and 3 and component 6. Second, the f values of two terrestrial humic-like components 1 and 2 for both Cu(II) and Hg(II) decreased with increasing salinity at the SRS sites (Figure 3). The competition between high levels of calcium and magnesium ions in seawater with copper ions for the relatively few available DOM binding sites is suggested as an explanation for this observation (8). In agreement, Fu et al. (12) reported that the fluorescence intensity of Hg(II)-DOM complexation increased after calcium ion addition, but not for magnesium ion addition, suggesting that the decreases in f values of the terrestrial humic-like components 1 and 2 with increasing salinity might, in fact, be controlled by high calcium ion levels in seawater. On the other hand, the f values of terrestrial humic-like component 6 with both of Cu(II) and Hg(II) were similar among sampling sites of different salinity, suggesting a preferential binding of component 6 with Cu(II) and Hg(II), as compared to the other major ions in seawater. The f values of terrestrial humic-like component 3 with Cu(II) were largely different between freshwater and brackish sites, and differences in f values of microbial humic-like components 4 and 5 among sampling sites were different between Cu(II) and Hg(II) (Figure 3). In addition to the potential effects of major ions in seawater, the f value of these components could also be affected by the abovementioned differences in binding sites between Cu(II) and Hg(II) and changes in f values due to biogeochemical alterations such as photo- and biodegradation of fluorescent components. Both decreases and enhancements in binding parameters have been reported after photodegradation of fluorophores (28, 29). In addition, it should be noted that mangrove ecosystems are well-known to be strong sources of fluorescent DOM (30, 31). Thus, fresh inputs of fluorescent components from mangrove ecosystems may affect the variation of the f values found in the present study. Behavior of Protein-Like Fluorescent Components with Addition of Trace Metals. Much of the dissolved organic nitrogen in the FCE has been shown to consist of dissolved proteins (18). It is well-known that fluorescence of protein and aromatic amino acids are quenched due to interaction with trace metals (7, 32). The quenching of protein-like fluorophores in DOM by Hg(II) was also reported using a combination of fluorescence quenching titration with synVOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7377

FIGURE 4. Changes in fluorescence intensity of protein-like component 7 and 8 with Cu(II) and Hg(II). chronous and EEM spectra (10, 12). However, such peak picking techniques might not evaluate the exact changes in fluorescence intensity of protein-like components because peaks of protein-like fluorophores in synchronous and EEM spectra may overlap with those of humic-like peaks (21, 33). In contrast, fluorescence quenching using EEM-PARAFAC modeled protein-like components will allow such assignments with higher precision. However, the reproducibility of triplicate titrations of the two protein-like components was significantly less when compared to that for the humic-like components, particularly at the higher Cu(II) concentration (see Supporting Information, Figure S4). The reason for such differences for the protein-like components is presently not clear. Therefore, only semiquantitative data for the interaction between protein-like components and trace metals are reported in the present study. Figure 4 shows the changes in fluorescence intensity of protein-like components 7 and 8 with the addition of Cu(II) and Hg(II). The fluorescence intensity of component 7 was quenched with increasing of Hg(II), irrespective of differences in sampling sites, even though the degree of quenching at TS2 was much larger than at the SRS sites. The changes in fluorescence intensity with Cu(II) were, however, different among sampling sites. Although continuous quenching was observed at SRS2 and TS2, the fluorescence intensity gradually decreased until a concentration of Cu(II) of about 30 µM Cu(II) was reached, but then sharply increased with further increasing of Cu(II) at SRS4 and 6. The quenching effect of component 8 with addition of Hg(II) was also observed at SRS2 and TS2 waters, but at TS2, the degree of quenching was not as significant as that for component 7, indicating the differences in binding sites for Hg(II) between protein-like components 7 and 8. The changes in fluorescence intensity of protein-like component 8 with the addition of Cu(II) were similar to those of component 7, although increases in fluorescence intensity at SRS4 and 6 were much larger than those for component 7. Such fluorescence quenching profiles after Cu(II) additions at SRS4 and SRS6 might be the result of a combination of interaction mechanisms, changes in the molecular environmental of the protein-like molecules, or both. The 7378

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 19, 2008

quenching of the fluorescence intensity at the initial stage of the titration is likely due to the metal complexation of protein-like components both for Hg(II) and Cu(II) (Figure 4). On the other hand, the increases in fluorescence intensity at the later stage (for Cu(II)) could have at least two reasons. First, changes in quantum yields of protein fluorescence by three-dimensional structural change of protein molecules (32) may have occurred due to high levels of Cu(II). Second, the fluorescence of protein-like components may be quenched due to interactions with inorganic or other organic components. Thus, the fluorescence intensity of such quenched protein-like components might be enhanced through the replacement of the original quencher with Cu(II). The fluorescence intensity might also be enhance due to the replacement of the protein in the DOM-protein complexes through the interaction with Cu(II) whereby the formation of more stable DOM-Cu(II) complexes result in the release of protein-like components from the DOM matrix. Thus, the increases in fluorescence intensity with Cu(II) at the later stages of the titration is likely due to changes in molecular environments of protein molecules. Further studies are needed to clarify such changes in molecular environments of protein-like components and assess its environmental significance. In summary, differences in specific metal binding properties of individual fluorescent components in DOM, as well as between different metals, are reported in the present study. Such differences demonstrate that fluorescence quenching titration combined with EEM-PARAFAC can be applied as a reliable technique to better understand the interactions between trace metals and individual fluorescent components of DOM in aquatic environments. In addition, the experimental results depicting the changes in fluorescence intensity of protein-like components with Cu(II) addition suggest that the fluorescence quenching with EEM-PARAFAC described here can also provide important information on the molecular environments of DOM.

Acknowledgments This study was funded by the National Science Foundation as part of the FCE-LTER program (DBI-0620409) and the Southeast Environmental Research Center Endowment. Y.Y.

thanks the College of Arts and Science at Florida International University for financial support. The authors thank the Wetlands Ecosystem Laboratory at SERC for logistic support. This is SERC contribution #400.

Supporting Information Available Sampling locations (Figure S1); validation of FCE-PARAFAC model (Figures S2 and S3); characteristics of the eight PARAFAC components as compared with those reported in previous studied (Table S1); changes in fluorescence intensities of fluorescent components during triplicate titration experiments (Figure S4); log K and f values of six humic-like components with Cu(II) and Hg(II) for four sampling sites (Table S2). This infromation is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Allen, H. E.; Hansen, D. J. The importance of trace metal speciation to water quality criteria. Water Environ. Res. 1996, 68, 42–54. (2) Scott, D. T.; Runkel, R. L.; McKnight, D. M.; Voelker, B. M.; Kimball, B. A.; Carraway, E. R. Transport and cycling of iron and hydrogen peroxide in a freshwater stream: Influence of organic acids. Water Resour. Res. 2003, 39, 1308. (3) Bianchi, T. S.; Biogeochemistry of Estuaries; Oxford University Press: New York, NY, 2007. (4) Cory, R.; McKnight, D. M.; Chin, Y.-P.; Miller, P.; Jaros, C. L. Chemical characteristics of fulvic acids from Arctic surface waters: Microbial contributions and photochemical transformations. J. Geophys. Res. 2007, 112. (5) Cabaniss, S. E. Synchronous fluorescence spectra of metal-fulvic acid complexes. Environ. Sci. Technol. 1992, 26, 1133–1139. (6) Esteves daSilva, J. C. G.; Machado, A. A. S. C.; Oliveira, C. J. S.; Pinto, M. S. S. D. S. Fluorescence quenching of anthoropogenic fulvic acids by Cu(II), Fe(III) and UO22+. Talanta 1998, 45, 1155– 1165. (7) Ryan, D. K.; Weber, J. H. Fluorescence quenching titration for determination of complexing capacities and stability constants of fulvic acid. Anal. Chem. 1982, 54, 986–990. (8) Ryan, D. K.; Weber, J. H. Copper(II) complexing capacities of natural waters by fluorescence quenching. Environ. Sci. Technol. 1982, 16, 866–872. (9) Luster, J.; Lloyd, T.; Sposito, G. Multi-wavelength molecular fluorescence spectroscopy for quantitative characterization of copper(II) and aluminum(III) complexation by dissolved organic matter. Environ. Sci. Technol. 1996, 30, 1565–1574. (10) Lu, X.; Jaffe´, R. Interaction between Hg(II) and natural dissolved organic matter: A fluorescence spectroscopy based study. Water Res. 2001, 35, 1793–1803. (11) Wu, F. C.; Mills, R. B.; Evans, R. D.; Dillon, P. J. Kinetics of metal-fulvic acid complexation using a stopped-flow technique and three-dimensional excitation and emission fluorescence spectrophotometer. Anal. Chem. 2004, 76, 110–113. (12) Fu, P.; Wu, F.; Liu, C.; Wang, F.; Li, W.; Yue, L.; Guo, Q. Fluorescence characterization of dissolved organic matter in an urban river and its complexation with Hg(II). Appl. Geochem. 2007, 22, 1668–1679. (13) Ohno, T.; Amirbahman, A.; Bro, R. Parallel factor analysis of excitation-emission matrix fluorescence spectra of water soluble soil organic matter as basis for the determination of conditional metal binding parameters. Environ. Sci. Technol. 2008, 42, 186–192. (14) Stedmon, C. A.; Markager, S.; Bro, R. Tracing dissolved organic matter in aquatic environments using a new approach to fluorescence spectroscopy. Mar. Chem. 2003, 82, 239–254.

(15) Cory, R. M.; McKnight, D. M. Fluorescence spectroscopy reveals ubiquitous presence of oxidized and reduced quinones in dissolved organic matter. Environ. Sci. Technol. 2005, 39, 8142– 8149. (16) Stedmon, C. A.; Markager, S. Resolving the variability in dissolved organic matter fluorescence in a temperate estuary and its catchment using PARAFAC analysis. Limnol. Oceanogr. 2005, 50, 686–697. (17) Yamashita, Y.; Jaffe´, R.; Maie, N.; Tanoue, E. Assessing the dynamics of dissolved organic matter (DOM) in coastal environments by excitation and emission matrix fluorescence and parallel factor analysis (EEM-PARAFAC). Limnol. Oceangr. 2008, 53, 1900–1908. (18) Maie, N.; Parish, K. J.; Watanabe, A.; Knicker, H.; Benner, R.; Abe, T.; Kaiser, K.; Jaffe´, R. Chemical characterization of dissolved organic nitrogen in an oligotrophic subtropical coastal ecosystem. Geochim. Cosmochim. Acta 2006, 70, 4491–4506. (19) Maie, N.; Yang, C. Y.; Miyoshi, T.; Parish, K.; Jaffe´, R. Chemical characteristics of dissolved organic matter in an oligotrophic subtropical wetland/estuarine ecosystem. Limonol. Oceanogr. 2005, 50, 23–35. (20) McKnight, D. M.; Boyer, E. W.; Westerhoff, P. K.; Doran, P. T.; Kulbe, T.; Andersen, D. T. Spectrofluorometric characterization of dissolved organic matter for indication of precursor organic material and aromaticity. Limnol. Oceanogr. 2001, 46, 38–48. (21) Yamashita, Y.; Tanoue, E. Chemical characterization of proteinlike fluorophores in DOM in relation to aromatic amino acids. Mar. Chem. 2003, 82, 255–271. (22) Ohno, T.; Bro, R. Dissolved organic matter characterization using multiway spectral decomposition of fluorescence landscapes. Soil Sci. Soc. Am. J. 2006, 70, 2028–2037. (23) Andersson, C. A.; Bro, R. The N-way toolbox for PARAFAC. Chemom. Intell. Lab. Syst. 2000, 52, 1–4. (24) Coble, P. G. Characterization of marine and terrestrial DOM in seawater using excitation-emission matrix spectroscopy. Mar. Chem. 1996, 51, 325–346. (25) Plaza, C.; Brunetti, G.; Sensei, N.; Polo, A. Molecular and quantitative analysis of metal ion binding to humic acids sludgeamended soils by fluorescence spectroscopy. Environ. Sci. Technol. 2006, 40, 917–923. (26) Bai, Y. C.; Wu, F. C.; Liu, C. Q.; Li, W.; Guo, J. Y.; Fu, P. Q.; Xing, B. S.; Zheng, J. Ultraviolet absorbance titration for determining stability constants of humic substances with Cu(II) and Hg(II). Anal. Chim. Acta 2008, 616, 115–1521. (27) Bai, Y.; Wu, F.; Wan, G.; Liu, C.; Fu, P.; Li, W. Ultraviolet absorbance titration for the determination of conditional stability constant of Hg(II) and dissolved organic matter determining. Chin. J. Geochem. 2008, 27, 046-052. (28) Wu, F.; Cai, Y.; Evans, D.; Dillon, P. Complexation between Hg(II) and dissolved organic matter in stream waters: an application of fluorescence spectroscopy. Biogeochemistry 2004, 71, 339– 351. (29) Brooks, M. L.; McKnight, D. M.; Clements, W. H. Photochemical control of copper complexation by dissolved organic matter in Rocky Mountain streams, Colorado. Limnol. Oceanogr. 2007, 52, 766–779. (30) Jaffe´, R.; Boyer, J. N.; Lu, X.; Yang, C.; Scully, N. M.; Mock, S. Source characterization of dissolved organic matter in a subtropical mangrove-dominated estuary by fluorescence analysis. Mar. Chem. 2004, 84, 195–210. (31) Maie, N.; Boyer, J. N.; Yang, C. Y.; Jaffe´, R. Spatial, geomorphological, and seasonal variability of CDOM in estuaries of the Florida Coastal Everglades. Hydrobiologia 2006, 569, 135–160. (32) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, Kluwer/ Plenum: New York, 1999. (33) Maie, N.; Scully, N. M.; Pisani, O.; Jaffe´, R. Composition of a protein-like fluorophore of dissolved organic matter in coastal wetland and estuarine ecosystems. Water Res. 2007, 41, 563–570.

ES801357H

VOL. 42, NO. 19, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

7379