Equilibrium-Based Sampler for Determining Cu2+ Concentrations in

The gellyfish is a thin disk of polyacrylamide gel embedded with iminodiacetate (Id) groups bound to immobile beads. The sampler operates on the princ...
0 downloads 0 Views 115KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 3381-3386

Equilibrium-Based Sampler for Determining Cu2+ Concentrations in Aquatic Ecosystems DAVID B. SENN, SARAH B. GRISCOM, CHRISTOPHER G. LEWIS, JENNIFER P. GALVIN, MARTHA W. CHANG, AND JAMES P. SHINE* Exposure, Epidemiology, and Risk Program, Department of Environmental Health, Harvard School of Public Health, 665 Huntington Avenue, Boston, Massachusetts 02115

The bioavailability of potentially toxic metals in aquatic systems is frequently related to the dissolved free metal ion (M2+) concentration. However, typical methods used to determine M2+ are labor intensive or require sophisticated equipment. We developed an inexpensive, in situ sampling devicesthe “gellyfish”sthat simplifies Cu2+ determinations in seawater. The gellyfish is a thin disk of polyacrylamide gel embedded with iminodiacetate (Id) groups bound to immobile beads. The sampler operates on the principle that the immobilized Id groups equilibrate with the Cu2+ concentration of the surrounding solution. Cu is then backextracted into a known volume of 10% HNO3 and measured by inductively coupled plasma mass spectrometry (ICPMS). In laboratory tests, we varied Cu2+ concentrations between 10-12 and 10-8 M and salinity between 5 and 35 ppt. Id-bound Cu (CuIdmeasured) did not respond to changes in total Cu. However, CuIdmeasured does increase in a predictable manner with increasing Cu2+, and prototype gellyfish precision (average coefficient of variation ) 10%) is sufficient to resolve small differences in Cu2+ ((30%). Modeled Cu uptake, based on thermodynamic equilibrium speciation of Id within gellyfish, is a good predictor of CuIdmeasured (r2 ) 0.96 and n ) 45).

Introduction For many toxic metals (such as Cu, Zn, Cd, and Ni) in marine and freshwater systems, the free metal ion (M2+) is the form of dissolved metal most readily taken up by aquatic organisms, ranging from direct uptake by unicellular planktonic organisms to uptake across gill membranes in fish (1, 2). M2+ is in many cases the metal species most readily transported across cell membranes due to its freedom to interact with transmembrane metal uptake proteins (3). Not surprisingly, therefore, numerous studies have found that M2+ concentrations best predict the toxicity of many dissolved metals to aquatic plants and organisms (4-6). In natural waters, only a small percentage of dissolved metal is typically present as M2+. Reactive groups associated with dissolved organic carbon (DOC)snaturally occurring and anthropogenic (such as EDTA)shave high binding affinities for transition metals and often complex a large percentage of the total dissolved metal present (>99.9% is * Corresponding author phone: (617)384-8806; fax: (617)384-8849; e-mail: [email protected]. 10.1021/es0353614 CCC: $27.50 Published on Web 05/11/2004

 2004 American Chemical Society

not uncommon for Cu(II)). In addition, inorganic ligands (e.g., Cl-, SO42-, OH-, and S2-) can play important roles in metal speciation. These complexing anions become increasingly important in seawater, at higher pH, and in anoxic, sulfidic waters. Complexation of heavy metals by both DOC and inorganic ligands typically decreases the bioavailability of dissolved metals and protects organisms from potentially toxic doses of those metals (7, 8). For example, although the San Francisco Bay is often in violation of the Clean Water Act criteria for total copper concentration, the binding of copper by DOC results in concentrations of Cu2+ that are generally below levels thought to be of ecological concern (9, 10). Because of situations such as this, new water quality criteria have been proposed that are based on metal speciation (the biotic ligand model) (11). One challenge with such approaches is that metal speciation can change in space and time due to changes in environmental factors such as pH, dissolved organic carbon levels, and salinity (2, 12). Accurate measures of M2+ are necessary for predicting dissolved metal uptake and toxicity. In general, though, current techniques are complex and cumbersome, and some can be prohibitively expensive or sufficiently labor intensive as to considerably limit the number of samples that can be measured. Measurement techniques include cathodic stripping voltametry (13), anodic stripping voltametry (14-17), bioassay techniques (18), ligand exchange/separation techniques (19), ion-exchange columns (20), and ion-selective electrodes (21). The availability of an inexpensive, in situ, and relatively simple method for determining M2+ would greatly increase the amount of data generated and thereby permit more accurate and comprehensive assessments of risk posed by toxic metals in aquatic systems, especially those that have traditionally been understudied (e.g., in under-resourced regions). As a step toward developing such a technique, this paper explores the utility of an in situ equilibrium sampler for determining Cu2+. Davison and Zhang pioneered efforts to develop in situ trace metal speciation samplers (e.g., ref 22). Their gel-based flux samplers, referred to as diffusive gradients in thin film (DGT) samplers, utilize the flux of metal ions through a porous gel matrix into an underlying layer of cation exchange beads (Chelex) to quantify dissolved concentrations of labile metal. The concentration of metal in the water can be backcalculated from the mass of a metal that accumulates on the beads using Fick’s First Law. The flux (mol cm-2 s-1) is calculated based on the mass of metal collected, sampler cross-sectional area, and sampler exposure time. Because gel thickness and metal diffusion coefficients within the gel are known, one can estimate the metal concentration in the water that was driving the flux. The design has many inherent advantages (small, relatively inexpensive, and fairly simple ability to measure multiple metals simultaneously), and DGT samplers have been used to study metals in coastal and freshwater systems as well as in sediments and soil slurries (i.e, refs 22-24). However, a potential disadvantage of this technique is that metal uptake by DGT samplers does not necessarily correlate well with M2+ measured by other techniques (24). Instead, the concentration of DGT-labile metals tends to be better correlated with total dissolved metal concentration under certain conditions (24, 25). Sizeexclusion gels, in conjunction with standard DGT, have been used with some success to reduce uncertainty about the metal species measured (26). We hypothesized that a simple and inexpensive in situ equilibrium-based metal sampler for determining M2+ was VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3381

feasible. Such a sampler would not continually take up metals but would instead be suspended in the water column until equilibrium is achieved between the M2+ concentration in the surrounding solution and the metal bound by ligands embedded within the sampler. In situ M2+ concentration could then be determined based on the amount of metal that accumulates in the sampler using a known (thermodynamic, semiempirical) relationship between M2+ and ligand-M2+, with the amount of ligand-M2+ complexes being proportional to [M2+] and a thermodynamic stability constant (K). The equilibrium samplersthe gellyfishsconsists of a pliable polyacrylamide gel wafer embedded with a known quantity of iminodiacetate (Id) metal-binding resin. To experimentally test the gellyfish’s ability to determine Cu2+ across concentration and salinity gradients, we conducted laboratory experiments using prototype gellyfish in various dilutions of artificial seawater that contained known levels of Cu2+, controlled by varying concentrations of total Cu and a well-characterized copper-binding dissolved ligand (nitrotriloacetic acid, NTA).

Experimental Procedures Trace metal cleaned plastic ware (polyethylene, polypropylene, or Teflon) was used for all steps of the experiments: bottles were soaked in 10% HCl for >24 h, rinsed with ultrapure water (Barnstead NANOpure Diamond, Dubuque, IA), stored filled with 0.1% HCl, and then rinsed again with ultrapure water before use. To minimize the potential for Cu contamination, most of the sample handling was conducted on a Class 100 clean bench. Artificial seawater (Aquil) was prepared following the procedure outlined in Price et al. (28) (fluoride and nutrients were excluded), using ultrapure water and trace metal grade salts (Sigma, St. Louis, MO). Desired salinities were achieved by dilution with ultrapure water, and salinity was verified using a salinometer (YSI 30/50 FT; Yellow Springs, OH) or refractometer (VWR Hand-Held Salinity Refractometer; West Chester, PA). To hasten equilibration with atmospheric CO2 concentrations and stabilize pH, Aquil batches were bubbled with HEPA-filtered air for several hours. For experiments in which planned total Cu additions were less than 10-8 M, Aquil was passed in series through a 0.4 µm membrane filter and a conditioned Chelex-100 metal chelating column to remove particles and trace metal impurities, as described in Price et al. (28). Production of Gellyfish Sampler. The gellyfish sampler (GFbeads) consists of iminodiacetate cation-exchange resin beads (Toyopearl AF-Chelate 650M, TosoHaas Biosep LLC; Montgomeryville, PA) embedded in a polyacrylamide gel matrix (modified from ref 29). The resin beads (hereafter referred to as beads) are shipped suspended in 20% methanol, with a mean bead size of 65 µm. The concentration of iminodiacetate groups is approximately 18 µmol per ml of resin slurry. Because of rapid gelling of the matrix, 5 mL batches of gel were prepared, each of which made ∼16 gellyfish. Batches consisted of 2.38 mL of ultrapure water, 0.75 mL of DGT gel cross-linker (2%; DGT Research Ltd.; Lancaster, UK), 1.88 mL of acrylamide solution (40%), 40 µL of ammonium persulfate (10%, prepared within 24 h), 15 µL of TEMED (N,N,N′N′-tetramethylethylenediamine, 99%), and 100 µL of Tosohaas resin beads. Prior to addition to the gel solution, the beads were washed 3 times with ultrapure water to remove the liquid carrier. TEMED and ammonium persulfate quantities were adjusted to yield an optimal gel coagulation rate that allows sufficient time for mixing and pipetting yet minimizes bead settling prior to coagulation. The batch was repeatedly mixed, and 300 µL was pipetted into 16 polypropylene custom drilled molds. After the gel fully set (approximately 30 min), the GFbeads were transferred to a Teflon beaker containing ultrapure water and were rinsed 3382

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004

and resuspended into clean water at least 3 times over the course of 24 h. GFbeads were stored in ultrapure water for up to 2 weeks before use and were rinsed periodically during that time. Fully hydrated gellyfish had dimensions of diameter ∼2 cm and thickness ∼2 mm and were >95% water with a wet volume of approximately 0.65 mL. The Idtotal concentration in each gellyfish was approximately 1.5 × 10-4 eq/L (6 µL beads/gellyfish). Gellyfish blanks containing no beads (GFblank) were prepared following the same recipe as above (minus the beads) and deployed alongside GFbeads in all experiments to quantify Cu within the gel matrix that was not bound by iminodiacetate groups (CuGFblank). Equilibration Time Experiment. Aquil was dispensed into two series of triplicate 1 L bottles. We used MINEQL+ chemical equilibrium modeling software (version 4.07; Environmental Research Software) to calculate the required total Cu(II) and NTA additions. MINEQL+ inputs included concentrations of all major anions and cations, total Cu(II), and NTA. MINEQL+ calculated Cu(II) speciation (i.e., the concentration of Cu2+ as well as the amounts of Cu(II) complexed by various ligands) and determined pH by solving the electroneutrality equation. Bottles were then spiked with predetermined amounts of stock solutions of CuSO4 (0.001 M) and Na3NTA (0.01 M) prepared from reagent-grade salts (Sigma) to yield a desired Cu2+ (10-10 or 10-8 M) at salinity ) 20 ppt (final concentrations: total Cu(II) ) 2.0 × 10-6 and 1.5 × 10-7 M and Na3NTA ) 6.5 × 10-6 and 5.0 × 10-5 M for Cu2+ ) 10-8 and 10-10 M, respectively). Bottles were allowed to equilibrate overnight, after which one GFbeads and one GFblank were added to each bottle. At various times (0.25, 1, 3, 7, and 11 days), gellyfish were removed from bottles, rinsed quickly with ultrapure water to remove residual surface Aquil, and placed in polypropylene centrifuge tubes containing 5 mL of 10% HNO3 for at least 2 days to back-extract copper. A time-series experiment showed that Cu could be >95% back-extracted into the 10% HNO3 within 24 h (data not presented). Aliquots of the samples were further diluted into 5% HNO3 acid and measured for copper by inductively coupled plasma mass spectrometry (Perkin-Elmer Elan 6100 DRC; Boston, MA) using indium as an internal standard to account for instrument drift. We measured 65Cu to avoid polyatomic interferences that occur when measuring 63Cu+ (from 40Ar23Na+). The concentration of total Cu originally within the gellyfish was then calculated ([Cuextract]‚volumeextract/volumegellyfish). The amount of Cu bound to Id groups within the gellyfish was determined by subtracting the amount of Cu in GFblank from the Cu in GFbeads ([CuIdmeasured] ) [CuGFbeads] - [CuGFblank]). Approximate ICP-MS detection limits were 0.2 µg/L total Cu (3 × average of 10 acid blanks). Recovery of NIST 1643-d standards, measured periodically throughout analyses, were generally 100 ( 10%. Salinity and Cu2+ Gradient Experiments. Using the same experimental design as stated previously, we measured Cu uptake by gellyfish across salinity (5, 20, and 35 ppt) and Cu2+ (10-13, 10-12, 10-11, 10-10, 10-9, and 10-8 M) gradients. As before, MINEQL+ was used to determine appropriate total Cu and Na3NTA additions (Supporting Information). Combinations of total Cu and total NTA were selected to yield the desired nominal Cu2+ and also guaranteed that less than 5% of total Cu in each bottle would be taken up by GFbeads, based on data from preliminary experiments and modeling calculations. When possible, total Cu was set such that [CuGFblank] < 0.1[CuGFbeads]. To quantify any Cu contamination in clean Aquil, bottles with no Cu additions but spiked with the highest NTA concentration, lowest NTA concentration, and no NTA were run alongside copper-spiked bottles. Bottles were mixed daily, and gellyfish were removed after 7 or 11 days. Constant Cu2+, Variable Total Cu Experiment. The sensitivity of CuId to changes in total Cu concentration was

FIGURE 1. Gellyfish equilibration time experiment for [Cu2+] ) 10-10 M (A) and 10-8 M (B). Each data point represents the mean of triplicate samples. Error bars are one standard deviation. Data were fit to a curve of the form C(t) ) Cequilibrium(1 - e-kt), where t is time, k is a first-order rate constant (days-1), and Cequilibrium is the equilibrium concentration at t ) ∞. also tested by varying total Cu while maintaining a constant Cu2+ concentration (at salinity ) 20 ppt). Bottles containing Cu2+ ) 10-10 M were prepared using several combinations of NTA and total Cu in which total Cu varied over 2 orders of magnitude.

Results and Discussion Equilibration Time. Results from equilibration experiments show that copper concentrations within GFbeads were elevated relative to total Cu in the surrounding solution and that Cu concentrations within GFbeads increased over time (Figure 1 A,B). In parallel experiments in which [Cu2+] differed by a factor of 100, the calculated times to 90% equilibration (t90) from best fit curves for Cu2+ ) 10-10 M and 10-8 M were 8.9 and 6.6 days, respectively (95% confidence intervals: 6.912.7 days and 4.8-10.7 days, respectively). Equilibration of GFblank with total Cu was reached within 1 day for both Cu2+ ) 10-10 M and Cu2+ ) 10-8 M. GFblank total copper concentrations were comparable to those in the surrounding solution, indicating that CuIdmeasured complexes ([CuIdmeasured] ) [CuGFbeads] - [CuGFblank]) were primarily responsible for elevated copper inside GFbeads. Salinity and Cu2+ Gradients. When Cu2+ was varied between 10-12 and 10-8 M, CuIdmeasured increased with increasing Cu2+ at all salinities (Figure 2A-C). CuIdmeasured asymptotically approached the total Id concentration (Idtotal ∼150 µM) at higher Cu2+, indicating that Id sites were becoming saturated with Cu. Gellyfish measurement uncertainty was small, with an average coefficient of variation of 10% in the triplicate analyses. On the basis of this, it is estimated that gellyfish deployed in triplicate are capable of detecting a 30% difference in mean concentration between

FIGURE 2. (A-C) Cu2+ and salinity gradient experiments. Triplicate samples for CuIdmeasured are presented at all Cu2+; differences between triplicates are imperceptible at certain Cu2+. Measured pH of experiments was 7.60, 7.80, and 8.05 at salinities of 5, 20, and 35 ppt, respectively. Actual Cu2+ was not the same as the original calculated value because of differences between measured and modeled pH. Therefore, actual [Cu2+] in the bottles was recalculated using MINEQL with measured pH as an input. The horizontal dashed lines depict [Idtotal] within the gellyfish. The dotted curves represent CuIdmodeled. (D) CuIdmeasured vs CuIdmodeled for all salinities and Cu2+. samples (p-value less than 0.05, power of 0.8) (30). Using this experimental setup and prototype gellyfish, [CuGFbeads] ≈ 2[CuGFblank] when Cu2+ ) 10-12 M; thus, Cu2+ ) 10-12 M represents the approximate overall detection limit at pH ∼8. Triplicate bottles of Cu2+ ) 10-13 M were included in the salinity ) 20 ppt experiment, but differences between [CuGFbeads] and [CuGFblank] were undetectable (Cu2+) 10-13 M bottles were not run at salinity ) 5 and 35 ppt). Modeled versus Measured Data. A spreadsheet-based equilibrium complexation model was constructed to predict copper uptake by the gellyfish. The gellyfish was modeled as if it were a volume of water that freely exchanges dissolved substances with the surrounding solution but contains a high concentration of immobile Cu-binding ligands. The model assumes that, once equilibrium is reached, the concentrations of dissolved constituents (major anions and cations in VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3383

TABLE 1. Equations for Gellyfish Equilibrium Modela

a Key: *, stability constants for the Id groups on Tosohaas beads were not available. K was estimated using published values for iminodiacetic Id acid (IDA) (T ) 25 °C, ionic strength ) 0.1 M; except NaId, T ) 20 °C, ionic strength ) 0.1 M) (27). KId were adjusted at each salinity for ionic strength effects (ionic strength ) 0.1, 0.4, and 0.7 M for 5, 20, and 35 ppt, respectively) using the Davies’ equation (32).

seawater, H+, Cu2+, dissolved Cu complexed by various inorganic and organic ligands, etc.) within the gel are the same as those in the surrounding solution and calculates the speciation of Id groups within the gel microenvironment. The surrounding solution thus acts as an infinite source of Cu: Cu binds to Id, and CuId is at equilibrium with Cu2+ within the gellyfish. Idtotal is comprised of a number of species (Table 1, eq a), based on known binding affinity of Id for various cations that may be present in quantities that could effect overall speciation. Other trace metals were not considered (discussed below). The concentration of each Idcontaining species is calculated using estimated stability constants (K) (adjusted for ionic strength effects) for Id complexation with Cu2+, H+, Ca2+, Mg2+, Sr2+, and Na+ after (Table 1 eqs b-h). Eq h in Table 1 is obtained by substituting eqs b-g into eq a and solving for the amount of deprotonated, uncomplexed Id (Idfree). Once Idfree is obtained, modeled Cu uptake by Id sites within the gellyfish (i.e., CuIdmodeled) is calculated using Table 1, eq d. Other species are calculated similarly. Concentrations of Cu2+, H+, Ca2+, Mg2+, Mg2+, Sr2+, and Na+ are inputs to the model and are determined by MINEQL+. By varying the salt concentrations (salinity), Cu2+ concentration, pH, and gellyfish-specific parameters (number of Id sites, volume of gellyfish), we calculated CuIdmodeled under a variety of conditions (Figure 3). Modeled Id speciation is influenced by both concentrations of salts and pH. HId and MgId should be the dominant Id species at low Cu2+, and Id speciation is relatively unaffected by increasing Cu2+ when [Cu2+] < 10-10 M. CuIdmodel versus Cu2+ therefore should have a log-log linear relationship at low [Cu2+]. However, at greater concentrations of Cu2+, CuId approaches Idtotal asymptotically as Id sites become saturated with Cu. The equilibrium model serves as a good predictor of CuIdmeasured, including capturing the asymptotic approach to Idtotal (Figure 2A-C). There is close empirical agreement (r2 ) 0.96) and mechanistic agreement (slope ) 0.92 ( 0.03) between CuIdmeasured and CuIdmodeled (Figure 2D). This is encouraging considering that stability constants for iminodiacetic acid were used in the model and might be expected to differ somewhat from those specific to the iminodiacetate groups on Tosohaas beads (Table 1). The salinity ) 20ppt treatment gellyfish were removed after 7 days (as compared to 11 days for the other two salinities), and subequilibrium concentrations may partially explain why CuIdmeasured at Cu2+ ) 10-8 and 10-9 M are slightly lower than CuIdmodeled. At salinity ) 5 and 35 ppt, the model underpredicts 3384

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004

FIGURE 3. Modeled equilibrium speciation of imminodiacetate within the gellyfish as a function of Cu2+ and salinity using calculated equilibrium pH ) 8.19, 8.06, and 7.62 for salinity ) 35, 20, and 5 ppt, respectively. CuIdmeasured at Cu2+ ) 10-12 M, possibly indicating that Cu contamination in the Aquil was comparable to the amount of total Cu added (total Cu ∼10-8). Because Id speciation is controlled largely by HId (Figure 3), CuIdmeasured is not extremely sensitive to salinity. Gellyfish, therefore, have the potential to serve as metal speciation tools for studies conducted in aquatic systems prone to salinity changes in space and time (such as estuaries) and may also be useful in freshwater systems. The ability to accurately model the relationship between Cu uptake and Cu2+ (and other parameters) indicates that the model may eventually serve as a calibration curve to identify the concentration of Cu2+ that was present in the surrounding solution. Constant Cu2+ and Variable Total Cu. Results from the variable total and Cu constant Cu2+ experiment support the

FIGURE 4. Constant Cu2+ (Cu2+ ) 10-10 M) and variable total Cu. Each data point represents the mean of triplicate samples. Error bars are one standard deviation. ANOVA analysis indicates that there is no significant differences in CuIdmeasured among any of the groups (p ) 0.48).

hypothesis that CuIdmeasured responds to changes in Cu2+ and not total Cu. At constant Cu2+, changes in total Cu over 2 orders of magnitude (10-7.3 to 10-5.3 M) appear to cause no significant changes in Cu uptake by GFbeads (Figure 4). ANOVA analysis of the data shows no significant differences in CuIdmeasured between any of the groups at different total metal concentrations (p ) 0.48). If Cu uptake by GFbeads was a function of Cutotal, CuIdmeasured would have increased by a factor of >10 for the 2 orders of magnitude increase in Cutotal (based on slopes in Figure 2B at Cu2+ ) 10-10 M). There is a clear need for simple, cost-effective methods for measuring contaminant metal speciation in aquatic ecosystems. The data presented previously demonstrate the gellyfish’s potential utility as an in situ metal speciation tool. Gellyfish are easy to use (effort comparable to that required for measuring total metal concentration), inexpensive (approximately $2 US per pair of GFbeads and GFblank), and once collected can be stored in dilute acid until analysis, thus avoiding concerns about altering water sample chemistry during storage. For these reasons, gellyfish may greatly increase the number of samples that can be analyzed per study or extend investigations of metal contamination to aquatic ecosystems that have been historically understudied due to resource limitations. The laboratory data also highlight several necessary improvements to the prototype gellyfish. Equilibration time needs to be shortened to permit deployments that can identify temporal trends in Cu2+ on time scales of hours to days in natural systems. This should be achievable: we calculate that decreasing the gellyfish thickness by a factor of 4 (to ∼0.5 mm) could reduce equilibration time by a factor of approximately 16 (t90 ∼10-14 h). The gel layer must remain sufficiently thick to hold beads and be rugged enough to allow routine sampler handling. The concentration of Id sites within the gellyfish is also important because it in part determines detection limits (CuId must be large relative to background total Cu that is measured in GFblank). Future designs may have higher concentrations of Id within GFbeads, thereby potentially lowering detection limits to the ranges observed in pristine environments (Cu2+ ) 10-13 to 10-14 M) (31). However, there is an upper bound on the amount of metal-binding ligand that can be included in the sampler before the gel structure is compromised (with the current recipe, gel was stable up to ∼900 µeq/L of Id within the gellyfish). In addition, preliminary experiments (conducted with Zn2+) suggest that greater Id concentrations may lead to proportionally longer equilibration times. Equilibration time may also be influenced by the flux of total Cu into the gellyfish and the subsequent dissociation rate of inorganic-

and organic-complexed Cu. Therefore, equilibration times need to be verified under field-like conditions with seawater containing natural ligand assemblages. Ongoing work with gellyfish includes implementing the design improvements and tests discussed previously and testing the applicability of gellyfish for measuring Cu2+ in more complex solutions containing metal mixtures. Model predictions indicate that the presence of other metals, such as Zn2+, Cd2+, Ni2+, and Pb2+, at environmentally relevant concentrations should not interfere with Cu2+ measurements because of their substantially lower affinities for Id (log KMId ) 7.27, 5.73, 8.13, and 7.41, respectively, as compared to 10.57 for Cu2+). For example, in a system with [Cu2+] ) 10-11 M, the addition of [Zn2+] ) 10-7 M alters [CuIdmodeled] by less than 5%.

Acknowledgments This publication was made possible by Grant 5 P42 ES05947 from the National Institute of Environmental Health Sciences (NIEHS), NIH, and the NIEHS Center for Environmental Health Grant ES00002 from the NIEHS. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIEHS or NIH.

Supporting Information Available Table of concentrations of total copper and NTA used in the salinity and Cu2+ gradient experiments. This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Morel, F. M. M.; Morel, N. M. L.; Anderson, D. M.; McKnight, D. M.; Rueter, J. G. In Advances in Marine Environmental Research; Jacoff, F. S., Ed.; EPA 600/9-79-035; U.S. Government Printing Office: Washington, DC, 1979; pp 38-61. (2) Paquin, P. R.; Gorsuch, J. W.; Apte, S.; Batley, G. E.; Bowles, K. C.; Campbell, P. G. C.; Delos, C. G.; Di Toro, D. M.; Dwyer, R. L.; Galvez, F.; et al. Comp. Biochem. Physiol. 2002, 133C, 3-35. (3) Simkiss, K.; Taylor, M. G. In Metal Speciation and Bioavailability In Aquatic Systems; Tessier, A., Turner, D. R., Eds.; John Wiley: Chichester, 1995; pp 1-44. (4) Sunda, W. G.; Guillard, R. J. Mar. Res. 1976, 34, 511-529. (5) Sunda, W. G.; Engel, D. W.; Thoutte, R. M. Environ. Sci. Technol. 1978, 12, 409-413. (6) Campbell, P. G. C. In Metal Speciation and Bioavailability In Aquatic Systems; Tessier, A., Turner, D. R., Eds.; John Wiley: Chichester, 1995; pp 45-102. (7) Hare, L.; Tessier, A. Nature 1996, 380, 430-432. (8) Kim, S. D.; Ma, H.; Allen, H. E.; Cha, D. K. Environ. Toxicol. Chem. 1999, 18, 2433-2437. (9) Sedlak, D. L.; Phinney, J. T.; Bedsworth, W. W. Environ. Sci. Technol. 1997, 31, 3010-3016. (10) Donat, J.; Lao, K.; Bruland, K. Anal. Chim. Acta 1994, 284, 547572. (11) United States Environmental Protection Agency. An SAB report: review of an integrated approach to metals assessment in surface waters and sediments. Ecological Processes and Effects Committee of the Science Advisory Board (1400A) April 6-7, 1999; EPA-SAB-EPEC-00-005; U.S. EPA: Washington, DC, 2000. (12) Renner, R. Environ. Sci. Technol. 1997, 31, 466A-468A. (13) Van Den Berg, C.; Buckley, P.; Huang, Z.; Nimmo, M. Estuarine Coastal Shelf Sci. 1986, 22, 479-486. (14) Huizenga, D.; Kester, D. Mar. Chem. 1983, 13, 281-291. (15) Kramer, C. Mar. Chem. 1986, 18, 335-349. (16) Coale, K.; Bruland, K. Limnol. Oceanogr. 1988, 33, 1084-1101. (17) Kogut, M. B.; Voelker, B. M. Environ. Sci. Technol. 2001, 35, 1149-1156. (18) Herring, J.; Sunda, W.; Ferguson, R.; Morel, F. Mar. Chem. 1987, 20, 299-312. (19) Moffett, J. W.; Zika, R. G. Mar. Chem. 1987, 21, 301-13. (20) Sweileh, J. A.; Weileh, J. A.; Lucyk, D.; Kratochvil, B.; Cantwell, F. F Anal. Chem. 1987, 59, 586-592. (21) Zirino, A.; VanderWeele, D. A.; Belli, S. L.; DeMarco, R.; Mackey, D. J. Mar. Chem. 1998, 61, 173-184. (22) Davison, W.; Zhang, H. Nature 1994, 367, 546-548. VOL. 38, NO. 12, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3385

(23) Morford, J.; Kalnejais, L.; Martin, W.; Franc¸ ois, R.; Karle, I.-M. J. Exp. Mar. Biol. Ecol. 2003, 285, 85-103. (24) Twiss, M. R.; Moffett, J. W. Environ. Sci. Technol. 2002, 36, 10611068. (25) Dunn, J. K.; Teasdale, P. R.; Warnken, J.; Schleich R. R. Environ. Sci. Technol. 2003, 37, 2794-2800. (26) Zhang, H.; Davison, W. Pure Appl. Chem. 2001, 73, 9-15. (27) Martell, A. E.; Smith, R. N. Critical Stability Constants; Plenum Press: New York, 1974. (28) Price, N.; Harrison, G.; Hering, J.; Hudson, R.; Nirel, P.; Palenik, B.; Morel, F. Biol. Oceanogr. 1989, 6, 503-521. (29) Zhang, H.; Davison, W. Anal. Chim. Acta 1999, 398, 329-340.

3386

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 12, 2004

(30) Sokal, R. F.; Rohlf, F. J. Biometry; W. H. Freeman and Company: New York, 1981. (31) Coale, K. H.; Bruland, K. W. Limnol. Oceanogr. 1988, 33, 10841101. (32) Morel, F. M. M.; Hering, J. G. Principles and Applications of Aquatic Chemistry; John Wiley & Sons: New York, 1993.

Received for review December 5, 2003. Revised manuscript received March 14, 2004. Accepted March 23, 2004. ES0353614