Environ. Sci. Technol. 2004, 38, 3587-3594
Redox Speciation of Copper in Rainwater: Temporal Variability and Atmospheric Deposition ROBERT J. KIEBER,* STEPHEN A. SKRABAL, CLIFF SMITH, AND JOAN D. WILLEY Department of Chemistry, University of North Carolina at Wilmington, Wilmington, North Carolina 28403-5932
Copper speciation was determined in 68 rainwater samples collected in Wilmington, NC, from August 25, 2000, to September 24, 2002. Volume-weighted average concentrations of Cutotal, dissolved Cu(II), and dissolved Cu(I) were 5.3, 3.2, and 1.4 nM, respectively, with a significantly higher ratio of Cu(II)/Cu(I) in summer relative to winter events. The concentrations of all Cu species were higher in storms of continental origin relative to marinedominated events, suggesting anthropogenic and/or terrestrial sources are important contributors of Cu in precipitation. Concentrations of strong Cu-complexing ligands were consistently lower than dissolved Cu concentrations, indicating a significant portion, but not all, of the dissolved Cu in rainwater is strongly complexed. A portion of these ligands, in addition to the sulfite and chloride in precipitation, may be Cu(I)-complexing ligands, which may explain the resistance of Cu(I) against oxidation in rainwater. Using our rainwater concentration data along with other published rainwater Cu concentrations and an estimate for total global annual rain, the total global flux of Cu removed from the atmosphere via wet deposition is 150 × 106 kg yr-1. This represents complete removal of the estimated Cu input into the troposphere and indicates essentially all Cu released into the global atmosphere is removed by rain.
Introduction Copper is one of the more abundant trace elements in atmospheric waters, including fogwater (10-6 M), cloudwater (10-7 M), and rain (10-9-10-8 M) (1-4). Copper has been implicated in many important atmospheric redox reactions, all of which depend on the oxidation state of the metal. These reactions directly impact the acid-generating and -oxidizing capacity of the troposphere because they include the conversion of S(IV) to S(VI) and reactions with free radicals such as OH, •HO2 and •O2- (5-12). Copper also plays an important role in the speciation of other trace metals in the troposphere including iron and chromium (10, 13-15). The reactivity of dissolved Cu species toward many of these rainwater components may be greatly influenced by complexation with both organic and inorganic ligands (16-20). Previous studies investigating Cu in rainwater based interpretations of its behavior on “total” Cu concentrations (Cutotal) operationally defined as the Cu present in unfiltered, acidified rain samples. However, quantifying Cutotal concentrations does not fully elucidate the behavior of atmospheric * Corresponding author phone: (910)962-3865; fax: (910)9623013; e-mail:
[email protected]. 10.1021/es030698r CCC: $27.50 Published on Web 05/15/2004
2004 American Chemical Society
Cu because its reactivity is so critically dependent on its speciation. A much more relevant assessment of the role of Cu in the redox chemistry of atmospheric waters can be obtained if levels of individual Cu species are measured. The primary objective of the present study is to determine the temporal and seasonal variability of Cutotal, total dissolved Cu(II)(aq), and total dissolved Cu(I)(aq), including all free and complexed forms, in rainwater. We will also address how anthropogenic activities impact the speciation of Cu in rainwater. In addition, this work addresses fundamental questions regarding how the redox chemistry and speciation of Cu compares to that of Fe and Cr, two other redox-active trace metals in rainwater. Approximately 56 × 106 kg of copper enters the atmosphere annually by anthropogenic activities including iron and steel manufacturing, nonferrous metal production, and fossil fuel combustion (21). An additional 18.5 × 106 kg of copper is added to the environment per year by natural sources including windblown dusts, vegetation exudates, and seasalt sprays (21). The final objective of the research presented here is to evaluate the role of rainwater as a removal mechanism for this atmospherically introduced Cu on a global scale. In addition to measuring a flux of total Cu removed, the importance of wet deposition as a removal mechanism for the various Cu species from the global atmosphere is also discussed.
Experimental Section Sample Collection. Rainwater samples were collected on an event basis at the University of North Carolina at Wilmington (UNCW) from August 2000 to September 2002. The UNCW rainwater collection site is a large open area of approximately 1 ha located within a turkey oak, long leaf pine, and wire grass community, typical of inland coastal areas in southeastern North Carolina. The site (34°13.9′ N, 77°52.7′ W) is approximately 8.5 km from the Atlantic Ocean. We have collected rainwater composition data for this site for over 15 yr, providing a large database that was useful for interpreting the metal speciation data and also allows comparison of our data with those from other locations. Due to the close proximity of the collection site to the laboratory, Cu(I) analyses could be initiated within minutes of collection, which reduced the possibility of compositional changes between the time of collection and analysis. Event rain samples were collected using four AerochemMetrics (ACM) model 301 automatic sensing wet/dry precipitation collectors. One of the collectors contained a 2 L muffled Pyrex glass beaker from which samples for dissolved organic carbon (DOC), hydrogen peroxide, pH, and inorganic ions were collected. The trace metal sample collectors, from which all Cu samples were obtained, consisted of a polyethylene funnel connected by FEP-Teflon-lined Tygon tubing to a 2 L FEP-Teflon bottle extensively cleaned using trace metal clean procedures (22-24). Approximately 39% of all rain events that occurred during the 2 yr period of the study were analyzed for all Cu forms (Cutotal, total dissolved Cu(I), total dissolved Cu(II)). Not all rain events were sampled because some lacked sufficient volume and/or too much time expired between collection and analysis. All samples reported were collected and analyzed immediately after cessation of a rain event. Copper concentration averages and standard deviations in rainwater were volume weighted, which minimizes effects of small rain events on averages and is the mathematical equivalent to combining all rain samples into one container prior to analysis. Stability studies on several rain samples analyzed after 4 h of storage showed VOL. 38, NO. 13, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3587
no change in Cu speciation relative to initial concentrations. Meteorological data including rain amounts, rain duration, time of day, surface temperature, wind speed, wind direction, and storm origin were also recorded. Real-time precipitation maps were used to define the end of specific rain events. Analytical Procedures. Dissolved Cu(I). Three 250 mL aliquots were filtered for each rainwater sample (usually within 10 min of collection) through 0.2 µm Millipore Isopore membrane filters and placed into fluorinated high-density polyethylene (FHDPE) bottles. Total dissolved Cu(I) concentrations in these freshly collected rainwater samples were quantified using a previously described solvent extraction procedure (25). This method utilized the Cu(I)-specific chelator 2,9-dimethyl-1,10-phenanthroline (dmp) (2.5 × 10-3 M) in concert with the strong Cu(II) chelator ethylenediamine (1.25 × 10-2 M), which prevents formation of Cu(I) during the extraction procedure. The pH of the extracting solution was adjusted to ∼8 using ultrapure dilute HCl. The addition of this extracting solution to the rain samples did not significantly alter their ambient pH values. Cu(I)-dmp complexes were extracted into methylene chloride and backextracted into 2 mL of ultrapure 5% HNO3, resulting in a preconcentration factor of ∼125. All extractions were performed under subdued light. Concentrations of Cu in the extracted samples were determined using graphite furnace atomic absorption spectrophotometry (GFAAS) with a Perkin-Elmer 5100 PC instrument equipped with a 5100 ZL Zeeman furnace module and an AS 70 autosampler programmed with recommended conditions. Concentrations were calculated using standard curves after standard addition experiments demonstrated the absence of significant matrix effects. The limit of detection for total dissolved Cu(I) in rain samples was estimated to be 0.1 nM (defined as 2 times the average standard deviation of triplicate analyses of seven low-level ( 109) organic ligands. In an earlier study using the tropolone CLE-CSV technique, strong Cu-complexing ligands determined using a similar level of competition (log RCu(trop)2 ≈ 4.1-4.5) were found at concentrations of 12-32 nM in four samples collected and analyzed during the winter at a semiurban location near Norwich, U.K. (17). Using different competition VOL. 38, NO. 13, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3589
FIGURE 1. Total dissolved Cu (TDCu) and relatively strong ligand (L) concentrations (nM) in Wilmington, NC, rainwater.
TABLE 4. Rainwater Characteristics of the Rain Events Used in Figures 1 and 2 amt event (mm)
pH
33 127 152 163 173 176 304
4.6 4.4 4.6 5.6 4.8 4.6 5.4
35 54 12 48 18 30 33
[DOC] [H2O2] [Cl-] [NO3-] [SO42-] (µM) (µM) (µM) (µM) (µM) 100 112
181 15
11 34 52 39
6 134 24
35 9 68 47 27 15 7
16 15 6 1 12 21 11
24 14 13 4 7 13 10
storm type continental continental mixed continental mixed mixed mixed
levels (log RCu(trop)2 ≈ 2.7-4.5), this work also demonstrated the presence of a spectrum of Cu-complexing ligands in rainwater (17), similar to what has been observed in seawater studies (46). The higher concentrations of strong ligands found in the U.K. study relative to the present work may reflect the contribution of organic constituents from urban sources in the Norwich area. In another study using CLECSV with the very strong competing ligand 8-hydroxyquinoline, 26-32% of the dissolved Cu was strongly complexed in urban and coastal rain samples from northwest England (19). Our results are more similar to this latter study and imply that roughly half the dissolved Cu is bound to strong organic ligands in Wilmington rainwater, with the remainder existing as free hydrated ion or complexed with weaker organic and inorganic ligands. Cu(I) Stability in Rainwater. The significance of Cu to the redox chemistry of the troposphere depends, to some degree, on the lifetime of the various Cu species present in the precipitation. Equilibrium thermodynamic calculations suggest that dissolved Cu in oxygenated systems such as rainwater would be present primarily as Cu(II) species (16). However, data presented in Table 2 and earlier studies with fogwater and cloudwater (12, 16) indicate that appreciable amounts of Cu(I) exist in atmospheric waters. In fogwater and cloudwater, Cu(I) can be produced by reduction of Cu(II) by several classes of organic compounds, by sulfite (at pH values above 6), and by various photoreduction mechanisms. Once formed, Cu(I) may then be stabilized from oxidation by complexation by sulfite, chloride, and organics (12, 16). Concentrations and reactions involving Cu(I) are thus highly dependent upon the composition of the atmospheric water in which it occurs, including the relative levels of Cu(I)-stabilizing ligands and Cu(I)-oxidizing species such as hydrogen peroxide and oxygen. The presence of stabilizing ligands that could protect Cu(I) from oxidation was indirectly assessed in SRW and authentic rain samples performed in the dark to exclude photochemical processes. Concentrations up to 4.8 nM Cu(I) in SRW all rapidly oxidized with only 0.04-0.18 nM Cu(I) recovered after 10 min (Figure 2). In authentic rain initially containing 1.6 nM Cu(I), concentrations of Cu(I) 3590
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 13, 2004
FIGURE 2. Concentrations of Cu(I) (nM) in synthetic and authentic rainwater as a function of time (h). Error bars represent 1 standard deviation based on three replicate analyses. decreased to 1 nM after 4 h and remained relatively stable over 24 h, indicating that Cu(I) is much more stable in authentic rainwater compared to SRW. These results are consistent with earlier studies in authentic and synthetic seawater which found that the oxidation of Cu(I) was too fast to be measured conventionally in synthetic solutions containing no or low concentrations of strong Cu(I) ligands, whereas in solutions containing Cu(I) ligands the oxidation was slowed significantly (47, 48). In various other storage experiments using authentic rain samples, measurements made over periods of 0-24 h showed that, although Cu(I) generally decreased with time, it remained at significant levels, in contrast to the nearly complete oxidation observed in SRW. Thus, we conclude that stabilizing ligands, present in differing concentrations in each rain sample, protect Cu(I) from rapid oxidation in authentic rainwater. These results are similar to earlier kinetic studies with Fe, which demonstrated that Fe(II) is also effectively protected against oxidation in rainwater by organic complexation (24, 44, 49). Possible Cu(I) ligands that exist in atmospheric waters include sulfite, chloride, and various organic compounds. Sulfite is believed to be more effective at stabilizing Cu(I) against oxidation relative to chloride; however, its concentrations in Wilmington rain are typically much smaller (
