Formation of Cu(I) in Estuarine and Marine Waters: Application of a

Formation of Cu(I) in Estuarine and Marine Waters: Application of a New ... Cu(I) in water samples collected from the River Scheldt estuary and the No...
0 downloads 0 Views 97KB Size
Download PDF
Environ. Sci. Technol. 2004, 38, 1843-1848

Formation of Cu(I) in Estuarine and Marine Waters: Application of a New Solid-Phase Extraction Method To Measure Cu(I) DIANE BUERGE-WEIRICH† AND BARBARA SULZBERGER* Department of Limnology, Swiss Federal Institute for Environmental Science and Technology (EAWAG), Ueberlandstrasse 133, CH-8600 Duebendorf, Switzerland

Cu(I) is a key species with respect to the bioavailability and hence toxicity of copper. Therefore, it is important to elucidate the factors that control Cu(I) steady-state concentrations in natural waters. In this study, a solid-phaseextraction-based method was developed that allows Cu(I) measurements at ambient concentrations. Cu(I) is selectively enriched as a bathocuproine complex on a hydrophobic polymer column, whereas Cu(II), bound to ethylenediamine, is not retained on the column. After elution with acidic methanol, Cu is analyzed with graphitefurnace atomic absorption spectroscopy. The detection limit of the whole analytical procedure is below 1 × 10-9 M, and the mean recovery of Cu(I) is ∼70%. We then applied this method to determine Cu(I) in water samples collected from the River Scheldt estuary and the North Sea. Upon irradiation of these filtered water samples in the laboratory (with ∼5 kW m-2), Cu(I) steady-state concentrations ([Cu(I)]ss) were established within a few minutes, and [Cu(I)]ss ranged from 5% to 80% of total dissolved copper, depending on the origin of the water samples. Measured [Cu(I)]ss can be interpreted by considering light-induced reduction of Cu(II) and stabilization of Cu(I) by chloride at high salinity, thermal reduction of Cu(II) by sulfidecontaining compounds at low salinity, and fast reoxidation of Cu(I) due to stabilization of Cu(II) by strong organic ligands present at intermediate salinity.

Introduction Divalent copper is generally assumed to be the predominant oxidation state of copper (Cu) in natural waters. Many studies deal with the speciation of Cu(II), and it was shown that more than 90% of Cu(II) is complexed by organic ligands (1-5). Carboxylic and amino functional groups are assumed to be major Cu(II)-binding groups. However, several studies indicate that Cu(I) is an important species with respect to the bioavailability and hence toxicity of Cu (6-8). In natural waters, sulfide-containing compounds (in the following denoted as sulfides) (9) and glutathione (6) are possible ligands of Cu(I). Sulfides seem to be involved in the detoxification and storage of heavy metals in phytoplankton. * Corresponding author phone: +41 +1 823 54 59; fax: +41 +1 823 50 28; e-mail: [email protected]. † Present address: Swiss Federal Research Station for Agroecology and Agriculture (FAL), Reckenholzstrasse 191, CH-8046 Zuerich, Switzerland. 10.1021/es034845x CCC: $27.50 Published on Web 02/05/2004

 2004 American Chemical Society

An increase in the concentration of free Cu in culture media resulted in an increased sulfide production by oceanic phytoplankton cultures (10-16). Sulfide concentrations in the western North Sea and the English Channel ranged from 0.7 to 3.6 nM (thiourea equivalents) (17). In some rivers in Connecticut and Maryland, sulfide concentrations up to 600 nM have been detected (18, 19). Some of these sulfides could be identified as thiols (20, 21). Glutathione or similar thiols are likely candidates for the “unknown ligands” (7). Thiols are known to reduce Cu(II) in dark reactions (22). Sulfides were previously considered to be unstable in oxic waters; however, it was reported that sulfides could persist in seawater for months (23). Sulfides are stabilized by complexation with Fe, Zn, and mainly Cu, forming clusters (24). Copper-sulfide clusters were shown to be more stable than other Cu(I) organic complexes. Chloride is another important Cu(I)stabilizing ligand, owing to the much higher stability constants of Cu(I)-chloride complexes, as compared to Cu(II)chloride complexes (see ref 8 and references therein). Moffett and Zika (8) showed that the half-life of Cu(I) in distilled water, to which chloride was added, increased by more than a factor of 10 when the chloride concentration was increased from 0.1 to 0.7 M. In natural waters, the highest concentrations of Cu(I) were found in surface water layers (26, 27), indicating photochemical formation of Cu(I), whereas at depths below the photic zone, Cu(I) could not be detected (27). Bad weather conditions and strong Cu(II) binding by allochthonous humic acids were supposed to be responsible for the lack of Cu(I) in some regions (25, 27). Light-induced formation of Cu(I) may occur via photolysis of Cu(II) complexes, since most ligands with carboxylate and amino functional groups form Cu(II) complexes that are photoreactive (25). In sunlit surface waters, Cu(I) also is produced through reduction of Cu(II) by photochemically formed reactive oxygen species. Superoxide is considered to be the main reductant of Cu(II), but H2O2 also contributes to Cu(I) formation (25). Inorganically and, above all, organically complexed Cu is an important sink of superoxide (28, 29). Several methods have been published to measure Cu(I) (27, 28, 30). Moffett et al. (30) presented a photometric method, in which Cu(I) is detected photometrically as an orange bathocuproine (2,9-dimethyl-4,7-diphenyl-1,10phenanthroline) complex (λmax ) 484 nm,  ) 12700 M-1 cm-1). This method has the advantage that it is easy to use, but the disadvantage that the detection limit is too high to measure Cu(I) at ambient concentrations in natural waters (detection limit (DL) >1 × 10-8 M, mean of 10 blanks plus 3 times the standard deviation). This value, determined in our experiments, is in the same range as that reported in the literature (30). In other methods published by Moffett and Zika (1988) (27) and Voelker et al. (2000) (28), Cu(I) was fixed with neocuproine (2,9-dimethyl-1,10-phenanthroline) and enriched with a two-step liquid-liquid extraction, first with methylene chloride and then with diluted nitric acid. In this acidic fraction, Cu was measured by atomic absorption spectroscopy (AAS). This method has the advantage that the detection limits are low enough to measure Cu(I) in natural waters, but the disadvantage of high consumption of toxic solvents. In this study, a method based on solid-phase extraction (SPE) was developed to measure Cu(I) at ambient concentrations. This method is described in detail in the Experimental Section. We then applied this method to measure Cu(I) in water samples from the River Scheldt estuary and the North Sea. Water samples were irradiated with a VOL. 38, NO. 6, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1843

laboratory light source, and Cu(I) steady-state concentrations ([Cu(I)]ss) were measured with our new method. This study indicates that different factors such as the presence of sulfides, chloride, and Cu(II) organic ligands influence steady-state concentrations of Cu(I) in natural waters.

Experimental Section Chemicals. CuIBr (purity g99%) (stored in the dark in an exsiccator, to protect it from moisture, and under a highpurity nitrogen atmosphere) and salicylaldoxime (puritiy g98%) were from Fluka, Switzerland. Bathocuproine disulfonic acid disodium salt (2,9-dimethyl-4,7-diphenyl-1,10phenanthroline disulfonic acid disodium salt) (∼90%), in the following denoted as bathocuproine, ethylenediamine (g99%), NaCl (g99.5%), boric acid (g99.9999%), H2O2 (>30%, suprapur quality), ammonia (>25%, suprapur quality), HNO3 (65%, suprapur quality), and HCl (30%, suprapur quality) were from Merck, Germany. Methanol (99.9%) was from Labscan Analytical Sciences, Ireland, and Cu(II) stock solution (1000 µg/mL copper nitrate in 2% nitric acid) from Baker Analyzed, Netherlands. Humic acid sodium salt was from EGA-Chemie, Germany. The gas used for deaerated experiments was a high-purity nitrogen gas, Alpahgaz2 from Carbagas, Switzerland, with an oxygen content of 1 × 10-8 M). Solutions with natural water were prepared by weight. The densities of the different water samples were determined at 25 °C by repeated weighing of a well-defined volume of the sample on a high-precision balance (Sartorius R200 D, Switzerland, precision (0.01 mg). A correction factor for the balance was defined as the ratio of the weight of a welldefined volume of Nanopure water and its density at 25 °C. Sampling of Water from the River Scheldt Estuary and the North Sea. Water from the River Scheldt estuary was sampled during a cruise with the Dutch research vessel 1844

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 6, 2004

FIGURE 1. Map showing the sampling sites in the River Scheldt estuary and the North Sea. The numbers correspond to the sample number listed in Table 1. Sample 11 was taken in the North Sea at a site located outside the boundaries of this map.

TABLE 1. Salinity, Concentrations of Total Dissolved Copper and DOC, and pH of the Different Water Samples from the River Scheldt Estuary and the North Sea sample no. 1 2 3 4 5 6 7 8 9 10 11a a

salinity 0.3 2.6 5.8 8.1 11.2 13.8 18.4 22.5 26.6 30.0 34.1

[Cutot] (M) 10-9

3.9 × 3.5 × 10-8 9.2 × 10-9 1.6 × 10-7 1.6 × 10-8 8.9 × 10-8 3.7 × 10-8 2.6 × 10-8 6.2 × 10-9 6.4 × 10-9 1.5 × 10-8

DOC (mg/L)

pH

7.0 7.4 7.0 5.8 5.1 4.8 4.8 3.3 4.6 1.9 1.2

8.21 8.16 8.22 8.73 8.36 8.63 8.47 8.88 8.42 8.32 8.35

Sample taken at 51°6′N and 1°41′E, at 5 m depth.

Navicula (Royal NIOZ) between April 2 and April 12, 2001. The sampling sites along a pronounced salinity gradient are shown in Figure 1. All sampling devices were metal-clean. Water was sampled with a fish, filtered (0.2 µm) on board, and deep-frozen immediately after filtration on the vessel. The River Scheldt originates in the north of France and flows across Belgium and The Netherlands into the North Sea through the Western Scheldt estuary. The River Scheldt and its estuary are one of the most important tidal river systems in Europe. The River Scheldt estuary is a typical hypernutrified ecosystem, with very low oxygen concentrations (down to 10-20% saturation). Indeed, the River Scheldt catchment presently receives sewage discharges with only partial treatment from cities such as Lille, Antwerp, Brussels, and Gent. The River Scheldt is one of the most contaminated rivers in Europe (31), although legislation in Belgium and The Netherlands aims to reduce industrial and domestic wastewater discharges. As a consequence of more severe legislation, water quality is improving (32, 33). Salinity, pH, total copper, and DOC concentrations of the water samples are listed in Table 1. The water from the North Sea was sampled during a cruise on the Dutch research vessel Pelagia (Royal NIOZ) between May 26 and June 4, 2001. Seawater was collected using a CTD rosette equipped with 10 L Go-Flow bottles, which were coated with a thin film of Teflon. Water was transferred (without filtration) to metal-clean polypropylene bottles and deep-frozen on the vessel. Samples were kept deep-frozen until use. They were unfrozen slowly at 4 °C, wrapped in aluminum foil, and used within 3 days. Measurement of Total Copper Concentration. The total Cu concentration in the natural, filtered water samples

([Cutot]) was measured by standard addition employing ligand-exchange differential pulse cathodic stripping voltammetry (LE-DP CSV) on a 647 VA stand, connected to a 646 VA processor, both from Metrohm, Switzerland. Previously, organic matter was oxidized by addition of H2O2 (15 mM final concentration) and irradiation during 18 h with a mercury lamp in a photooxidation unit (La Jolla Scientific Co., California) (method adapted from refs 1, 34, and 35). After this treatment, DOC concentrations, measured on a Shimadzu TOC 5000A analyzer, were below the detection limit (0.2 mg L-1). The pH value of 20 mL aliquots of the natural water samples was adjusted to 8.35 with borate buffer (final concentration 0.01 M), followed by addition of salicylaldoxime (final concentration 0.01 M). The solution was then deaerated by purging with water-saturated high-purity nitrogen gas for 5 min and subsequently stirred for 15 s. The deposition potential was set to -80 mV; four mercury drops were discarded before a new mercury drop was exuded. The solution was stirred for a preset period of 1 min, then the stirrer was stopped, and a quiescent period of 15 s was allowed. The potential was scanned from -150 to -600 mV. The scanning parameters were the following: differential pulse modulation, 2 pulses/s; scan rate, 12 mV/s; pulse height, 6 mV. Blanks consisting of Nanopure water were treated in the same manner to control copper contamination. Total copper concentrations were corrected for these values, which were generally approximately 1 nM. Total Cu concentrations measured with LE-DP CSV were verified with GF-AAS measurements in the samples at low ionic strength. These measurements were performed with both irradiated and nonirradiated water samples. The two methods gave the same results within 6%. Assessment of Cu(II) Speciation. The chemical speciation of Cu(II) was assessed with ligand-exchange cathodic stripping voltammetry and equilibrium calculations (3, 36). The analytical procedure was similar to the measurement of total Cu, but without oxidation of organic matter. The final concentration of salicylaldoxime was 15 µM, and the preset period was 4 min (for water from the North Sea, 1 min). Stabilization of Cu(I) for the Solid-Phase Extraction Method. One of the main difficulties in measuring Cu(I) is its stabilization to prevent oxidation. Cu(I) can be stabilized with bathocuproine, a strong and selective chelator of Cu(I) (30). The Cu(I)-bathocuproine complex is stable at pH values around 8. Furthermore, a masking ligand for Cu(II) is needed since bathocuproine can reduce Cu(II). In this study, ethylenediamine was used. Appropriate concentrations of bathocuproine and ethylenediamine were chosen considering speciation calculations and literature data (27, 30) and on the basis of experimental results. Speciation calculations were done with the program Visual Minteq (37, 38), using a Gaussian model for DOC complexation properties and equilibrium constants from the literature (39-41). Because equilibrium constants for bathocuproine were not available, calculations were performed with the constants for neocuproine (2,9-dimethyl1,10-phenanthroline), which is likely to have similar complexing properties. A bathocuproine concentration of 1 µM was found to be high enough to quantitatively bind Cu(I) in the water samples used in this study, and at the same time, low enough to avoid saturation of the adsorption sites of the solid phase used for extraction (see later). A 5-fold excess of ethylenediamine (5 µM) over bathocuproine proved to be optimal to ensure complete complexation of Cu(II) by ethylenediamine. A larger excess would lead to complexation and subsequent oxidation of Cu(I) (27, 30). To minimize oxidation of Cu(I) by oxygen before stabilization, the bathocuproine/ethylenediamine/borate solu-

tions were prepared with deaerated water (method as described above) in a glovebox under a high-purity nitrogen atmosphere. These solutions were prepared in 240 mL amber glass screw top bottles with a Teflon/silicone septum. We tested that the septa were tight, and once closed, the bottles could be taken out of the glovebox for irradiation or dark experiments (see later). A volume of typically 100 mL of a Cu(I) solution was then added with a syringe through the septum into 100 mL of the bathocuproine/ethylenediamine/ borate solution (final concentrations: bathocuproine, ∼1 µM; ethylenediamine, ∼5 µM; borate, ∼0.01 M; in the following denoted as bathocuproine solution). This large volume of bathocuproine solution was necessary to outcompete oxidation of Cu(I) by oxygen (contained in natural water samples) by diluting its concentration and thus favoring the complexation reaction of Cu(I) with bathocuproine, present at high concentration. Cu(I)-bathocuproine solutions were equilibrated for 3 h in the dark to allow exchange of Cu(I) between natural ligands and bathocuproine. Equilibrium should be reached after this time, as a nearly diffusion-controlled water loss rate can be expected for a large, monovalent cation with a d10 configuration (42). It could be shown that the Cu(I)-bathocuproine complex remained stable for at least one week, if the samples were kept in the dark and at 4 °C. Solid-Phase Extraction. Solid-phase extraction of the Cu(I)-bathocuproine complex was performed with an Abselut Nexus Bond Elut LRC column from Varian. It contains 60 mg of a cross-linked, polymeric sorbent with a hydrophobicity comparable to those of classical C-8 columns and does not need conditioning. A single, orange-colored narrow band formed in the column, while Cu(I)-containing samples were passed, indicating a high retention of the Cu(I)-bathocuproine complex in this column. Other columns, e.g., C-18, C-12, C-8, and phenyl columns, were also checked, but either the retention was not sufficient or Cu(I) could not be eluted with methanol, needed for AAS measurements (see later). For solid-phase extraction, a vacuum manifold (type Baker SPE 12 G, Switzerland) equipped with Teflon luers and ventils was used. The columns were precleaned by passing sequentially 6 mL of HCl (pH 1.5), 3 mL of H2O (Nanopure), and 7 mL of acidic methanol (“pH 1.5”). After this procedure, the columns were clean, as verified by analyzing the different washing eluates for Cu by GF-AAS (detection limit 0.1 µg/L). To remove methanol from the column and to establish the conditions of the water sample on the column with regard to pH, 2 mL of borate buffer was passed through the column prior to the sample. A volume of typically 400 mL of the water sample was then passed through the column at 1-2 mL min-1. The total amount of bathocuproine passed through the column should not exceed 3 mg to avoid saturation of the solid phase, particularly in the presence of hydrophobic compounds in natural water samples that may also react with the sorbent sites. Thereafter, the remaining Cu(II)-ethylenediamine was washed from the column with 10 mL of Nanopure water. Several alternative washing solutions, such as a mixture of ethylenediamine/borate buffer, were checked, but Nanopure water proved to be best suited. We checked that all Cu(II) was washed from the column in this cleaning step by passing Cu(II)/bathocuproine/ethylenediamine/borate buffer solutions containing no Cu(I) through the column. Finally, Cu(I) was eluted with 2 mL of acidic methanol (pH 1.5), which allowed direct measurements with GF-AAS. AAS Measurements. Atomic absorption analyses were performed with a heated graphite-furnace atomic absorption system with a Zeeman background corrector (Perkin-Elmer 5100 ZL). For methanolic Cu solutions, the temperature program of the furnace had to be adapted (Table 2). Cu was measured at a wavelength of 324.8 nm and a slit width of 0.7. VOL. 38, NO. 6, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1845

TABLE 2: Furnace Conditions for GF-AAS To Measure Total Copper Concentrations in Methanolic Solutions step

temp (°C)

ramp time (s)

hold time (s)

dry step 1 dry step 2 ash step read step clean step

60 110 1300 2500 2400

20 20 25 0 1

10 15 10 5 7

FIGURE 3. Cu(I) formed upon irradiation in the laboratory and in dark controls in natural water samples with different salinities from different locations in the River Scheldt estuary and the North Sea. Error bars indicate the standard deviation of typically three replicates.

FIGURE 2. Recovery of Cu(I) after solid-phase extraction, determined for concentrations of 1 × 10-9 to 7.5 × 10-7 M. Mean recovery is 69%. The inset is an expansion of the data at low Cu(I) concentrations. Measurements were checked with standard addition. The use of a modifier was not necessary. The detection limit (mean of 10 blanks plus 3 times the standard deviation) was 0.1 and 0.5 µg L-1 for aqueous and methanolic samples, respectively. Irradiation Experiments. A 100 mL natural water sample was irradiated in a 100 mL gastight Pyrex syringe (GC syringe, Hamilton, model 1100) in the laboratory using a 1000 W xenon lamp (OSRAM). The light intensity was approximately 4.7 kW m-2, as determined by ferrioxalate actinometry (43). A detailed description of the irradiation unit has been published in ref 44. Samples were irradiated for 15 min, after having tested that steady-state concentration of Cu(I) was reached within this period of time. These tests were performed with Cu(II)-humic complexes (concentrations of humic acid and Cu(II): 14.4 mg/L and 8 µM, respectively), where a steady-state concentration of Cu(I) was reached after about 7 min. The syringe was mounted on a 240 mL amber glass screw top bottle with a Teflon/silicone septum, which contained the bathocuproine solution (see above). After irradiation, the water sample was injected into this bathocuproine solution. To prevent reoxidation of Cu(I), the syringe was kept exposed to light until it was empty. To make injection easier and faster, the amber glass bottle was attached to a vacuum pump via a second inlet through the septum, so that the water sample was sucked from the syringe into the bottle within 2 min. We checked that the temperature of the water samples in the syringe remained constant at 25 °C throughout the experiment. Dark control experiments were performed in exactly the same manner, except that the water in the syringe was not irradiated.

Results and Discussion Recovery, Blank Values, and Detection Limit of This New Method. With this SPE method, we were able to measure Cu(I) reproducibly over a large concentration range. Figure 2 shows recoveries of Cu(I) for a concentration range of 1 × 10-9 to 7.5 × 10-7 M. The mass of Cu(I) (µg) recovered in the methanolic fraction is plotted as a function of the mass of Cu(I) (µg) applied on the column. This kind of representation 1846

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 6, 2004

allows direct determination of the mean recovery via the slope, which was 69%. The dashed lines indicate the 95% confidence interval. Cu(I) concentrations measured in natural water samples were corrected for recovery. Varying the ionic strength between 0 and 0.7 M, using Nanopure water to which the appropriate concentration of NaCl was added, had no influence on the recovery of Cu(I). To determine blank values for samples without Cu(I) and to check for possible interferences with Cu(II), measurements were performed in bathocuproine solutions either with no Cu(II) or with Cu(II) concentrations between 4 × 10-9 and 4 × 10-7 M. Blank values were close to the instrumental detection limit of AAS, independent of the Cu(II) concentration added. This indicates that there is no significant reduction of Cu(II) by bathocuproine on the column, and that Cu(II)-ethylenediamine complexes do not adsorb on the column. The detection limit for the whole analytical procedure is around 1 × 10-9 M or less. The upper boundary of the detection window is given by a slight Cu contamination in the purchased bathocuproine disulfonic acid disodium salt powder as verified by GF-AAS measurements of acidified solutions of this powder. Therefore, the volume of natural water samples applied to the columns has to be big enough such that the mass of Cu applied by the water sample exceeds the mass of Cu applied on the column with the bathocuproine solution. Effect of Salinity on Cu(I) Steady-State Concentrations. Cu(I) steady-state concentrations, [Cu(I)]ss, were measured in filtered water samples from the River Scheldt estuary and the North Sea during irradiation in the laboratory or in dark controls. Figure 3 shows the percentage of [Cu(I)]ss from total dissolved (0.2 µm filtered) copper, [Cutot] (in the following denoted as % [Cu(I)]ss/[Cutot]), as a function of the salinity of the water samples. The error bars represent the standard deviation of typically three replicates; the error was generally below 10%. Three regions with different % [Cu(I)]ss/[Cutot] can be distinguished. A first region covers water samples with low salinity ranging from 0.3 to about 10. In this region, high % [Cu(I)]ss/[Cutot] values (up to 67%) were found upon irradiation of the water samples. However, Cu(I) also was formed in the absence of light, which suggests thermal reduction of Cu(II) and subsequent stabilization by sulfides, e.g., thiols (see the Introduction). Sewage treatment plants and dam sediments are known to be sources of sulfides (24). Sediments in the upper part of the Scheldt estuary represent an additional source of sulfides (45). A second region covers water samples with intermediate salinity, ranging from about 10 to 15. Upon irradiation, much lower % [Cu(I)]ss/[Cutot] values (below 20%) were detected

than in the first region. The concentration of Cu(I) in dark controls was below or close to the detection limit. The fact that almost no Cu(I) was detected in the absence of light suggests that thiols or other sulfides play a minor role in Cu(I) stabilization in this part of the estuary. A likely explanation of these low Cu(I) steady-state concentrations is that strong Cu(II) ligands probably favor back-oxidation of Cu(I) (see below). Finally, a third region of water samples with salinity above 18 can be distinguished. In this region, high % [Cu(I)]ss/[Cutot] values (up to 80%) were found upon irradiation, whereas in dark controls Cu(I) was not detectable. We suppose that chloride is responsible for the stabilization of Cu(I) in this region (25). It may seem surprising that such high % [Cu(I)]ss/[Cutot] values were observed. It was shown before that Cu(I) comprises 5-10% of total Cu at various locations in the Atlantic Ocean and the Gulf of Mexico (27), up to 15% in the surface water of the Florida Coast (25), and up to 25% in coastal water at Cape Cod, MA (28). However, these concentrations were all measured in situ in the sunlit water column, whereas the light intensity in our experiments was much higher (see the Experimental Section). The possibility of interferences in Cu(I) measurements by hydrophobic organic Cu(II) complexes being enriched on the column seems unlikely. As ethylenediamine was added in excess, and samples were allowed to equilibrate for 3 h, Cu(II) can be expected to be completely exchanged by ethylenediamine. According to equilibrium calculations (see the Experimental Section) ,1% of Cu(I) and >97% of Cu(II) were present as ethylenediamine complexes with the ligand concentrations used in this study, even at the highest DOC and Cu(II) concentrations measured in water samples from the River Scheldt estuary (7 mg L-1 DOC and 1 × 10-7 M Cu(II)). Furthermore, increasing the concentration of ethylenediamine had no effect on the results. Finally, the fact that no Cu(I) could be detected in dark controls at high salinity corroborates the assumption that no Cu(II) complexes were enriched on the column. On the contrary, the question is rather whether all Cu(I) present in the water samples was detected by our method. Cu(I) could be bound to very strong ligands that are not completely exchanged by bathocuproine and not retained on the solid phase. However, this also seems unlikely, as increasing the bathocuproine concentration and the equilibration time did not affect the results. The fact that no Cu could be eluted from a second column, after the water sample had been passed through the first column, further shows that Cu(I)-bathocuproine complexes were completely retained on the solid phase. Effect of DOC Content and Salinity on Cu(I) SteadyState Concentrations. As shown in Figure 3, Cu(I) steadystate concentrations are not a simple function of the salinity; i.e., stabilization of Cu(I) by chloride is not the only factor controlling [Cu(I)]ss. Figure 4 shows the influence of salinity and DOC content on the percentage of Cu(I) from total dissolved copper, % [Cu(I)]ss/[Cutot], during irradiation. According to this figure, the DOC content in the River Scheldt estuary decreases with increasing salinity, which is mainly due to dilution effects. At high DOC content and low salinity (corresponding to the first region in Figure 3), % [Cu(I)]ss/ [Cutot] decreased with decreasing DOC content, reaching a minimum around 5 mg L-1 DOC and 10% salinity (corresponding to the second region in Figure 3). As discussed above, formation of Cu(I) in the first region mainly occurred through thermal reduction of Cu(II), presumably by sulfides, e.g., thiols, and light had only a minor effect on Cu(I) formation (see Figure 3). Above a salinity of 10, % [Cu(I)]ss/ [Cutot] increased with decreasing DOC content and increasing

FIGURE 4. Contour plot representing the influence of salinity and DOC concentration on the percentage of Cu(I) steady-state concentration from total dissolved copper concentration in irradiated water samples from the River Scheldt estuary and the North Sea.

FIGURE 5. Influence of the complexation capacity of Cu(II) ligands on the percentage of Cu(I) steady-state concentration from total dissolved copper concentration in irradiated water samples from the River Scheldt estuary and the North Sea. salinity, indicating that in this third region stabilization of Cu(I) by chloride is an important factor. Effect of Cu(II) Complexation Capacity on Cu(I) SteadyState Concentrations. As discussed above, the distinct minimum of % [Cu(I)]ss/[Cutot] in the second region may be explained by fast back-oxidation of Cu(I), due to Cu(II)stabilizing ligands present in this region. Another reason for this minimum could be low quantum yields of photolysis of Cu(II) complexes. It has been shown that the structure of organic Cu(II) complexes and hence the relative stability of the carbon-centered radicals, as well as the binding mechanism, e.g., outer- or inner-sphere complexation, have a strong impact on the quantum yield of photolysis of Cu(II) complexes (46, 47). To get a better understanding of the effects of ligands on Cu(I) steady-state concentrations, we elucidated the speciation of Cu(II) in water samples with different salinity. To this aim, the influence of the Cu(II)-complexing capacity of organic ligands on % [Cu(I)]ss/[Cutot] was assessed (Figure 5). The ratio [Cu(II)free]/[Cutot] can be used as a measure for the complexation capacity: low ratios are equivalent with high values of (log KL)[L] and vice versa, where KL represents the complexation constant and [L] the ligand concentration. The data shown in Figure 5 indicate that, in water samples from the River Scheldt estuary with salinities below 15, Cu(II) complexation capacity was higher than in the North Sea with a salinity of 34. Similar results, showing that a higher complexation capacity for Cu(II) was found in coastal waters than in oceanic waters, have already been reported (14). Interestingly, in the presence of a weaker Cu(II) complexation VOL. 38, NO. 6, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

1847

capacity ([Cu(II)free]/[Cutot] > 0.01), significantly higher Cu(I) steady-state concentrations were formed than in the presence of stronger ligands. These experimental results may be rationalized by the fact that strong Cu(II) ligands favor back-oxidation of Cu(I). Thus, strong organic Cu(II) ligands could explain the low Cu(I) steady-state concentrations at intermediate salinities (second region in Figure 3). With increasing salinity, two effects may thus favor Cu(I) stabilization: the increase of the chloride concentration and the decreasing complexation capacity of Cu(II) organic ligands.

Acknowledgments We acknowledge the Swiss Government (Bundesamt fu ¨r Bildung und Wissenschaft) for the financial support of this project, which is part of the EU-project COMET (Composition of Dissolved Organic Matter and its Interactions with Metals and Ultraviolet Radiation in River-Ocean Systems: Impact on the Microbial Food Web). We also thank Marie Boye´ (Royal NIOZ, The Netherlands) and Luis Laglera (ULIV, Great Britain) for taking samples during the Navicula Cruise, and Laurence Meunier (EAWAG, Switzerland) for taking samples during the Pelagia Cruise. Rainer Amon (AWI, Germany) is acknowledged for DOC measurements in water samples from the River Scheldt estuary and the North Sea.

Literature Cited (1) Bruland, K. W.; Rue, E. L.; Donat, J. R.; Skrabal, S. A.; Moffett, J. W. Anal. Chim. Acta 2000, 405, 99-113. (2) van den Berg, C. M. G. Mar. Chem. 1982, 11, 307-322. (3) van den Berg, C. M. G. Mar. Chem. 1984, 15, 1-18. (4) Voelker, B. M.; Kogut, M. B. Mar. Chem. 2001, 74, 303-318. (5) Kogut, M. B.; Voelker, B. M. Environ. Sci. Technol. 2001, 35, 1149-1156. (6) Leal, M. F. C.; van den Berg, C. M. G. Aquat. Geochem. 1998, 4, 49-75. (7) Leal, M. F. C.; Vasconcelos, M. T. S. D.; van den Berg, C. M. G. Limnol. Oceanogr. 1999, 44, 1750-1762. (8) Moffett, J. W.; Zika, R. G. Mar. Chem. 1983, 13, 239-251. (9) Al-Farawati, R.; van den Berg, C. M. G. Mar. Chem. 1999, 63, 331-352. (10) Ahner, B. A.; Kong, S.; Morel, F. M. M. Limnol. Oceanogr. 1995, 40, 649-657. (11) Ahner, B. A.; Morel, F. M. M. Limnol. Oceanogr. 1995, 40, 658665. (12) Ahner, B. A.; Morel, F. M. M.; Moffett, J. W. Limnol. Oceanogr. 1997, 42, 601-608. (13) Moffett, J. W.; Zika, R. G.; Brand, L. E. Deep Sea Res. 1990, 37, 27-36. (14) Moffett, J. W.; Brand, L. E.; Croot, P. L.; Barbeau, K. A. Limnol. Oceanogr. 1997, 42, 789-799. (15) Walsh, R. S.; Cutter, G. A.; Dunstan, W. M.; Radford-Knoery, J.; Elder, J. T. Limnol. Oceanogr. 1994, 39, 941-948. (16) Croot, P. L.; Moffett, J. W.; Brand, L. E. Limnol. Oceanogr. 2000, 45, 619-627. (17) Al-Farawati, R.; van den Berg, C. M. G. Environ. Sci. Technol. 2001, 35, 1902-1911. (18) Rozan, T. F.; Benoit, G. Geochim. Cosmochim. Acta 1999, 63, 3311-3319.

1848

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 6, 2004

(19) Rozan, T. F.; Benoit, G.; Luther, G. W., III. Environ. Sci. Technol. 1999, 33, 3021-3026. (20) Al-Farawati, R.; van den Berg, C. M. G. Mar. Chem. 1997, 57, 277-286. (21) Tang, D.; Warnken, K. W.; Santschi, P. H. Mar. Chem. 2002, 78, 29-45. (22) Krezel, A.; Bal, W. Acta Biochim. Pol. 1999, 46, 567-580. (23) Luther, G. W., III; Tsamakis, E. Mar. Chem. 1989, 27, 165-177. (24) Rozan, R. F.; Lassman, M. E.; Ridge, D. P.; Luther, G. W, III. Nature 2000, 406, 879-882. (25) Moffett, J. W.; Zika, R. G. In Photochemistry of Environmental Aquatic Systems; Zika, R. G., Copper, W. J., Eds.; Advances in Chemistry Series 327; American Chemical Society: Washington, DC, 1987; pp 116-130. (26) Moffett, J. W.; Zika, R. G. Environ. Sci. Technol. 1987, 21, 804810. (27) Moffett, J. W.; Zika, R. G. Geochim. Cosmochim. Acta 1988, 52, 1849-1857. (28) Voelker, B. M.; Sedlak, D. L.; Zafiriou, O. C. Environ. Sci. Technol. 2000, 24, 1036-1042. (29) Zafiriou, O. C.; Voelker, B. M.; Sedlak, D. L. J. Phys. Chem. A 1998, 102, 5693-5700. (30) Moffett, J. W.; Zika, R. G.; Petasne, R. G. Anal. Chim. Acta 1985, 175, 171-179. (31) Somville, M.; de Pauw, N. Water Res. 1982, 16, 1349-1356. (32) Van Eck, G. T. M.; De Bruijckere, F. L. G.; De Meyer, E.; Maeckelberghe, H. Water 1998, 102, 293-303. (33) van Damme, S.; Meire, P.; Maeckelberghe, H.; Verdievel, M.; Borugoing, L.; Taverniers, E.; Ysebaert, T.; Wattel, G. Water 1995, 85, 244-256. (34) Achterberg, E. P.; Braungardt, C. B.; Sandford, R. C.; Worsfold, P. J. Anal. Chim. Acta 2001, 440, 27-36. (35) Campos, M. L. A. M.; van den Berg, C. M. G. Anal. Chim. Acta 1994, 284, 481-496. (36) Xue, H.; Sigg, L. Limnol. Oceanogr. 1993, 38, 1200-1213. (37) Allison, J. D.; Brown, D. S.; Novo-Gradac, K. J. Minteq A2/Model A2, A Geochemical Assessment Model for Environmental Systems: Version 3.0 User’s Manual, Athens, GA 30605, 1991. (38) Allison, J. D.; Brown, D. S.; Novo-Gradac, K. J. Minteq A2/Model A2, A Geochemical Assessment Model for Environmental Systems: User Manual Supplement for Version 4.0, Athens, GA 30605, 1999. (39) Smith, R. M.; Martell, A. E. Critical Stability Constants. Vol. 4: Inorganic Complexes; Plenum Press: New York, 1976. (40) Smith, R. M.; Martell, A. E. Critical Stability Constants. Vol. 3: Other Organic Ligands; Plenum Press: New York, 1977. (41) Smith, R. M.; Martell, A. E. Critical Stability Constants. Vol. 5: First Supplement; Plenum Press: New York, 1982. (42) Meierstein, D. Inorg. Chem. 1975, 14, 1716-1721. (43) Hatchard, C. G.; Parker, A. C. Proc. R. Soc. London, Ser. A 1956, 518-536. (44) Siffert, C.; Sulzberger, B. Langmuir 1991, 7, 1627-1634. (45) Paucot, H.; Wollast, R. Mar. Chem. 1997, 58, 229-244. (46) Sun, L.; Wu, C.-H.; Faust, B. C. J. Phys. Chem. A 1998, 102, 86648672. (47) Wu, C.-H.; Sun, L.; Faust, B. C. J. Phys. Chem. A 2000, 104, 49894996.

Received for review July 30, 2003. Revised manuscript received December 15, 2003. Accepted December 26, 2003. ES034845X