Article pubs.acs.org/ac
Simultaneous Measurement of Trace Metal and Oxyanion Concentrations in Water using Diffusive Gradients in Thin Films with a Chelex−Metsorb Mixed Binding Layer Jared G. Panther, William W. Bennett, David T. Welsh, and Peter R. Teasdale* Environmental Futures Centre, Griffith School of Environment, Griffith University Gold Coast campus, Queensland 4222, Australia S Supporting Information *
ABSTRACT: A new diffusive gradients in thin films (DGT) technique with a mixed binding layer (Chelex-100 and the titanium dioxide based adsorbent Metsorb) is described for the simultaneous measurement of labile trace metal (Mn, Co, Ni, Cu, Cd, and Pb) and oxyanion (V, As, Mo, Sb, W, and P) concentrations in freshwater and seawater. The mixed binding layer (MBL) DGT technique was evaluated against the Chelex-DGT and Metsorb-DGT techniques, and all elution efficiencies and diffusion coefficients have been remeasured for the above analytes. Diffusion coefficients (D) measured using MBL-DGT generally agreed well with those measured by Chelex-DGT (DMBL/DChelex = 0.97−1.05), MetsorbDGT (DMBL/DMetsorb = 0.97−1.01), and diffusion cell experiments. The measurement of trace metals and oxyanions by MBL-DGT was independent of pH (5.03−8.05) and ionic strength (I = 0.001−0.7 mol L−1). MBL-DGT accurately measured the concentration of trace metals and oxyanions in synthetic freshwater (CMBL/CSol = 0.82−1.18) over the 4 day deployment and also agreed well with Metsorb-DGT (CMBL/CMetsorb = 0.84−0.94) and Chelex-DGT (CMBL/CChelex = 0.88−1.11) measurements. In synthetic seawater, MBL-DGT accurately measured the concentration of metals and oxyanions (CMBL/CSol = 0.85−1.12) over 4 days, with the exception of Monone of the DGT techniques were capable of measuring Mo in seawater. MBL-DGT measured the Mn concentration accurately over the entire 4 day period, whereas ChelexDGT only measured Mn accurately up to 2 days. The MBL-DGT method described in this study offers significant advantages over the ferrihydrite-Chelex-DGT method reported previously. These advantages include the commercial availability of both Metsorb and Chelex-100, the higher accuracy of Metsorb for measuring some oxyanions in freshwater and seawater, and the possibility of measuring Fe, which would not be possible using the Chelex-ferrihydrite binding layer.
T
For DGT applications, Chelex-100 is the binding agent of choice for the measurement of trace metals, whereas Metsorb (a commercially available titanium dioxide-based adsorbent) and ferrihydrite (an iron-oxide mineral phase) have been used extensively for measuring oxyanion species.8,10−23 Typically, if the concentration of both metals and oxyanion species are to be measured by DGT, separate samplers are deployedone sampler employing Chelex-100 as the binding agent and another utilizing Metsorb or ferrihydrite as the binding agent. Mason et al.24 developed a DGT method that utilized a single sampler to measure both cations and oxyanions simultaneously. This sampler contained a mixed binding agent, Chelex-100 and ferrihydrite, and was evaluated for the measurement of P, Mn, Cu, Mo, Zn, and Cd. Recently, Huynh et al.25 extended the use of the Chelex-ferrihydrite binding layer by measuring As, Pb, Cd, Cu, and Zn in both water and soil impacted by mining activities. However, none of these Chelex-ferrihydrite studies
he use of in situ aquatic monitoring techniques to provide reliable analytical data for environmental assessments is well recognized.1 In situ techniques can minimize problems associated with sample collection, preservation, and storage, such as analyte contamination, change in analyte speciation, and/or loss of analyte during storage.1,2 In many cases the data provided by in situ methods can be more representative than grab sampling, which requires a sample to be collected, in some cases filtered or preserved, transported, and stored prior to analysis.1,3 During this process, the integrity of the sample may be compromised. Passive samplers are diffusion-based in situ techniques that have been extensively used for measuring both organic and inorganic pollutants in water. These techniques offer several advantages over conventional grab-sampling methods.4−6 One such method is the diffusive gradients in thin films (DGT) technique7,8a kinetic-regime passive sampler.4 When it is combined with a suitable analytical instrument, the DGT technique is capable of determining the concentrations of both metals and oxyanions in water, soil, and sediment.7−9 © 2013 American Chemical Society
Received: July 21, 2013 Accepted: November 20, 2013 Published: November 20, 2013 427
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
Analytical Chemistry
Article
mixed together prior to analysis. Chelex binding gels were eluted in 1 mL of 1 mol L−1 HNO3 for 24 h, and Metsorb binding gels were eluted in 1 mL of 1 mol L−1 NaOH for 24 h. For Sb measurements, the MBL and Metsorb gels were also eluted with 1 mol L−1 NaOH/1 mol L−1 H2O216 after the initial two-step or one-step elution, respectively, and all of the respective eluents were combined before analysis. All eluents were diluted 10-fold prior to analysis. Measured elution efficiencies are presented in Table 1. For synthetic seawater and freshwater experiments, the MBL and Metsorb binding gels were rinsed in 5 mL of deionized water for 1 h prior to elution.19 Calculation of DGT Concentration. The DGT-measured concentration (CDGT, ng mL−1) was calculated using the DGT equation8 along with the diffusion coefficients (D) measured in the current study corrected to 25 °C (Table 2). The D value for P was not determined in this study, and the reported value of 6.05 × 10−6 cm2 s−1 was used.23 For high ionic strength deployments (0.7 mol L−1 NaNO3 and seawater) a value of 0.9 × D was used for all analytes.7 Sample Analysis. Measurements were performed using an inductively coupled plasma mass spectrometer (ICP-MS, Agilent 7500a). All ICP-MS samples and standards were prepared in a 2% (v/v) HNO3 matrix. Independently certified QC standards and matrix matched QC standards at 10 μg L−1 were analyzed every 20−30 samples to evaluate interferences from ions in the synthetic seawater and freshwater samples, and analytical recoveries were between 95 and 103%. Sc, Y, and In were added to all samples (final concentration of 10 μg L−1) as an internal standard to account for instrument drift. Phosphorus measurements were performed using a Merck phosphate test kit (EPA 365.2, APHA 4500-P E). For P measurements, seawater and freshwater certified reference materials (Queensland Health Scientific Services) were used for quality assurance purposes; typical recoveries were between 92 and 106%. Laboratory Evaluation. Uptake and Elution Efficiencies. Uptake and elution efficiencies were measured by immersing five MBL, Chelex, or Metsorb gel disks (4.91 cm2) in 4.5 mL of 100 μg L−1 analyte solution (prepared in 0.01 mol L−1 NaNO3, pH 5.85). After 24 h the gels were removed, eluted, and analyzed. Time-Series Accumulation. The time-series accumulation of analytes by MBL-DGT, Chelex-DGT, and Metsorb-DGT were evaluated at pH 6.06 (0.01 mol L−1 NaNO3 and 0.004 mol L−1 Mg(NO3)2) for up to 24 h in a solution containing 10−27 μg L−1 of each analyte. Effective diffusion coefficients were calculated using the slope of the linear regression of the mass of analyte accumulated in the binding gel over time.10 This and all subsequent experiments were performed using DGT water samplers. Diffusion Coefficient Measurements using a Diffusion Cell. A diaphragm diffusion cell27 was used to measure diffusion coefficients at pH 4.01 in 0.01 M NaNO3. The source solution contained 10 mg L−1 of each analyte. The source and receiving solutions were stirred continuously using magnetic stirrers. After 20 min, 1 mL samples were taken from both sides every 10 min for 2 h. Diffusion coefficients were calculated as described by Zhang and Davison.27 pH and Ionic Strength. To examine the efficacy of MBLDGT and Chelex-DGT across wide pH and ionic strength ranges, DGT samplers were deployed in mixed analyte solutions (15−35 μg L−1 of each analyte) of varying pH
evaluated the performance of the binding layer in seawater for long-term deployments and, additionally, the limitations of the ferrihydrite binding agent for measuring oxyanions in freshwater and seawater have been well documented.11,16,19,26 The present study focuses on the development of a ChelexMetsorb mixed binding layer (MBL) for the simultaneous measurement of trace metals and oxyanions in freshwater and seawater. The following series of laboratory experiments were conducted to validate the Chelex-Metsorb MBL-DGT technique: determining elution factors, measuring diffusion coefficients using a diffusion cell and time-series DGT deployments, evaluating the effect of variations in ionic strength and pH, and measuring analyte accumulation over time for 4 days in synthetic freshwater and seawater. This is the first time that the Chelex-Metsorb MBL has been described and evaluated against the existing Metsorb-DGT and Chelex-DGT methods, and it is shown to offer significant advantages by allowing simultaneous measurement of a wide range of metal cations and oxyanions in both freshwater and seawater using a single DGT sampler.
■
EXPERIMENTAL SECTION General Experimental Details. All solutions were prepared using deionized water (Milli-Q Element, Millipore). As and P standard solutions (1000 mg L−1) were obtained from High Purity Standards and Merck, respectively. Solutions of Mo, V, Sb, and W (1000 mg L−1) were prepared fresh using Na2MoO4·2H2O, NH4VO3, KSb(OH)6, and Na2WO4·2H2O (all AR grade), respectively. Solutions of Mn, Co, Ni, Cu, and Cd (1000 mg L−1) were prepared from MnSO4·H2O, Co(NO3)2·6H2O, NiCl2·6H2O, CuSO4·5H2O, and CdSO4· H2O (all AR grade), respectively; metal solutions were prepared in 2% (v/v) HNO3. Spectrosol standards (1000 mg L−1) were used for Cu, Pb, and Zn. All salts used to prepare experimental solutions were AR grade or higher. Plastic containers and DGT components were acid-cleaned in 10% (v/v) HNO3 (AR grade, Merck) for 24 h and rinsed thoroughly with deionized water prior to use. DGT Procedures. Gel Preparation. Diffusive gels and Chelex gels were prepared according to Zhang and Davison8 with minor modifications: 2 g of dry Chelex-100 (BIO-RAD, 100−200 dry mesh) was used. Metsorb (Graver Technologies, Glasgow, DE, USA) binding gels were prepared according to Bennett et al.10 The mixed binding layer contained 2 g of dry Chelex-100 and 1 g of Metsorb and was cast using the same procedures as for other binding gels. DGT Deployment. Water sampler housings were supplied by DGT Research Ltd. The samplers were assembled as described previously using a 0.45 μm cellulose nitrate membrane filter (Millipore).8 All DGT deployments were carried out in clear polypropylene containers. The temperature, pH, and salinity/ conductivity of the deployment solutions were regularly monitored throughout each experiment. Dilute NaOH and HNO3 solutions were used to adjust the pH. Samples of the experimental solution were removed at the beginning of each experiment and whenever DGT samplers were retrieved, filtered (0.45 μm), acidified to 2% (v/v) HNO3, and stored at 98% (n = 6). Uptake efficiencies for Chelex and Metsorb were >99% for Mn, Co, Ni, Cu, Zn, Cd, Pb, V, As, Mo, Sb, and W (n = 6), respectively. MBL elution efficiencies (1 mol L−1 HNO3 followed by 1 mol L−1 NaOH) were between 80 and 95% for all analytes, with the exception of Sb (18.6%; Table S1, Supporting Information), Chelex elution efficiencies were between 91 and 95% for all metals, and Metsorb elution efficiencies were between 87 and 95% for all oxyanions, with the exception of Sb (20.2%; Table S1). For complete elution of Sb from the MBL and Metsorb binding layers, 1 mol L−1 NaOH/1 mol L−1 H2O2 was used after the initial two-step elution and all of the sample eluents were combined prior to analysis. This procedure yielded overall elution efficiencies for Sb of 89% and 91% for MBL and Metsorb, respectively. Method detection limits (MDL) for MBL-DGT, ChelexDGT, and Metsorb-DGT are presented in Table 1. The MBLDGT detection limits for Co, Cd, V, As, Mo, Sb, and W were similar to the corresponding Chelex-DGT or Metsorb-DGT detection limits. For Mn, Ni, Cu, Zn, and Pb, the MBL-DGT detection limits were 1.5−6.0 times higher than the ChelexDGT detection limits, indicating a low-level contamination of the Metsorb binding agent with these metals. Precleaning the Metsorb binding agent with HNO3 and NaOH, prior to mixing with Chelex-100, lowered the detection limits considerably (values in parentheses, Table 1), but they were still slightly higher than the corresponding Chelex-DGT detection limits. The DGT detection limits achieved in this study are well below the relevant water quality guideline values specified by the Australian and New Zealand Environment and Conservation Council (ANZECC), the United States Environmental Protection Agency (USEPA), and the EU Water 429
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
Analytical Chemistry
Article
Table 2. Measured Diffusion Coefficients of Cations and Anions (×10−6 cm2 s−1) using a Diffusion Cell (Dcell) and DGT Samplers with Different Adsorbents (DMBL, DChelex, DMetsorb), Corrected to 25 °C Dcella Mn Co Ni Cu Zn Cd Pb V As Mo Sb W
4.88 5.17 5.21 5.75 5.47 5.52 7.75 3.75 5.26 5.18 4.90 6.22
± ± ± ± ± ± ± ± ± ± ± ±
0.06 0.04 0.03 0.04 0.05 0.04 0.17 0.14 0.04 0.10 0.12 0.11
DMBLb 4.68 5.29 5.13 5.61 6.22 5.56 8.03 7.98 6.02 6.33 6.22 6.88
± ± ± ± ± ± ± ± ± ± ± ±
0.18 0.29 0.21 0.15 0.32 0.24 0.17 0.37 0.32 0.40 0.11 0.47
DChelexc 4.44 5.38 5.29 5.47 6.37 5.72 8.17
± ± ± ± ± ± ±
DMetsorbd
0.17 0.20 0.26 0.32 0.45 0.37 0.28
DMBL/DChelex
DMBL/DMetsorb
1.05 0.98 0.97 1.03 0.98 0.97 0.98 8.13 6.01 6.43 6.16 7.06
± ± ± ± ±
0.64 0.18 0.27 0.33 0.58
0.98 1.00 0.98 1.01 0.97
Conditions: pH 4.01; 0.01 mol L−1 NaNO3. bDiffusion coefficient measured using MBL-DGT: pH 6.06; 0.01 mol L−1 NaNO3/0.004 mol L−1 Mg(NO3)2. cDiffusion coefficient measured using Chelex-DGT: pH 6.06; 0.01 mol L−1 NaNO3/0.004 mol L−1 Mg(NO3)2. dDiffusion coefficient measured using Metsorb-DGT: pH 6.06; 0.01 mol L−1 NaNO3/0.004 mol L−1 Mg(NO3)2. a
Figure 1. Freshwater deploymentoxyanions: average (n = 3) mass of analyte accumulated by MBL-DGT (■) and Metsorb-DGT (○). Error bars represent the standard deviation of triplicate measurements. Dotted lines give the predicted uptakes calculated using the DGT equation. Experimental conditions: pH 7.48 ± 0.11; conductivity 232 μS cm−1; temperature 23.7 °C. Average analyte concentrations were as follows: V, 15.3 μg L−1; W, 15.1 μg L−1; Mo, 14.2 μg L−1; As, 13.0 μg L−1; Sb, 17.4 μg L−1; P, 28.3 μg L−1.
unknown factor contributing to differences between the Dcell and DDGT values for the oxyanion species. For V, a larger discrepancy between the Dcell and DDGT values was observed (Dcell/DDGT = 0.47). Similar results have been reported by Luo et al.13 and Panther et al.16 This difference is probably due to a change in speciation. Speciation calculations using Visual MINTEQ version 3.031 showed that at pH 6.06 and a total V concentration of 15 μg L−1 (conditions under which DDGT was measured), H2VO4− is the dominant species (99.7%). At pH 4.01 and a total V concentration of 10 mg L−1 (conditions under which Dcell was measured), 50% of total V is present as H2VO4− but significant amounts of various
values for the MBL and Chelex were within the range of 0.86− 1.23; however, most values were within 10% of 1.00. For As, Mo, Sb, and W, Dcell/DDGT values for MBL and Metsorb ranged from 0.79 to 0.88. Recently, lower Dcell values have also been reported.10,16 It has been suggested that the difference between the Dcell and DDGT values was due to the presence of a small diffusive boundary layer (DBL) associated with the diffusion cell experiments or the effect of lateral diffusion that occurs within the DGT sampler diffusive layer.10,16 However, good agreement between DCcell and DDGT for the metal species would suggest that this is not the case and there may be some other 430
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
Analytical Chemistry
Article
Figure 2. Freshwater deploymenttrace metals: average (n = 3) mass of analyte accumulated by MBL-DGT (■) and Chelex-DGT (○). Error bars represent the standard deviation of triplicate measurements. Dotted lines give the predicted uptakes calculated using the DGT equation. Experimental conditions: pH 7.48 ± 0.11; conductivity 232 μS cm−1; temperature 23.7 °C. Average analyte concentrations were as follows: Mn, 10.9 μg L−1; Co, 11.2 μg L−1; Ni, 12.5 μg L−1; Cu, 14.4 μg L−1; Cd, 14.8 μg L−1; Pb, 6.8 μg L−1.
polymeric V species (43.4%) and VO2+ (6.6%) are present. Cationic species generally have lower diffusion coefficients than anionic species, and the polymeric V species are also likely to have appreciably slower diffusion coefficients; thus, speciation differences and the relative diffusion coefficients of the species may explain the difference between the measured DDGT and Dcell values for V. DDGT values are a combined measurement of the rate of diffusion of analyte through the DGT diffusive layer and the binding ability of the binding agent for the analyte; hence, DDGT values more accurately reflect both the mass transport and binding of the analyte within the DGT sampler and therefore were used for all subsequent calculations. pH and Ionic Strength. The effect of pH (5.03−8.05) and ionic strength (I = 0.001−0.7 mol L−1) on the performance of MBL-DGT for measuring metals and oxyanions in solution was investigated. The concentrations measured by MBL-DGT (CMBL) were also compared to those by Chelex-DGT (CChelex) and Metsorb-DGT (CMetsorb). In general, MBL-DGT accurately measured the concentration of all analytes in solution to within 15% of the dissolved concentration across the pH and ionic strength range investigated. CMBL/Csol ratios are given in Table S2 (Supporting Information). The measurement of metals by Chelex-DGT and oxyanions by Metsorb-DGT were also in good agreement with the MBL-DGT measurements; CChelex/ CMBL and CMetsorb/CMBL values were generally between 0.90 and 1.10 (results not shown). Collectively, these results demonstrate the utility of the MBL-DGT technique for the simultaneous measurement of metals and oxyanions over a wide pH and ionic strength range. Synthetic Freshwater. DGT samplers containing MBL, Metsorb, and Chelex binding layers were deployed in synthetic
freshwater spiked with V, As, Mo, Sb, W, P, Mn, Co, Ni, Cu, Cd, and Pb (6−28 μg L−1) for up to 4 days. The concentration of analyte in the experimental solution decreased over the 4 day deployment. For most analytes this decrease was between 7 and 21% of the original concentration; however, for Pb and P the concentration decreases were 38% and 39%, respectively. This loss can be attributed to adsorption of analytes to container walls at low ionic strength.32,33 In the case of P, microbial uptake over the 4 day deployment may also have contributed to the decrease in solution concentration. Because of the decrease in analyte concentrations, the dashed lines in Figures 1 and 2, which illustrate the predicted mass of analyte accumulated by DGT, are nonlinear. Plots of analyte mass accumulated over time for MBL-DGT showed good agreement between the mass of analyte accumulated by the DGT sampler (solid squares in Figures 1 and 2) and the predicted mass of analyte for all analytes investigated. CMBL/Csol values ranged between 0.82 and 1.18 (results not shown); however, in general, values were typically between 0.9 and 1.10. These results demonstrate that MBLDGT quantitatively measured the analyte concentration in freshwater over the entire 4 day deployment. Metsorb-DGT samplers were deployed in triplicate at 76 and 100 h for comparison with MBL-DGT (Figure 1, open circles) for the measurement of the oxyanion species (V, As, Mo, Sb, W, and P). Good agreement was obtained between CMetsorb and Csol; however, the CMetsorb/Csol ratio was consistently greater than 1.00 (CMetsorb/Csol = 1.02−1.12) and CMBL/CMetsorb was less than 1.00 for all analytes (CMBL/CMetsorb = 0.84−0.94). However, independent t-tests confirmed that there was no statistical difference between the CMetsorb and Csol values and between the CMBL and CMetsorb values. 431
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
Analytical Chemistry
Article
Figure 3. Seawater deploymentoxyanions: average (n = 3) mass of analyte accumulated by MBL-DGT (■) and Metsorb-DGT (○). Error bars represent the standard deviation of triplicate measurements. Dotted lines give the predicted uptakes calculated using the DGT equation. Experimental conditions: pH 8.02 ± 0.14; salinity 32.4; temperature 23.9 °C. Average analyte concentrations were as follows: V, 14.0 μg L−1; W, 13.5 μg L−1; Mo, 14.7 μg L−1; As, 13 μg L−1; Sb, 27.1 μg L−1; P, 31.5 μg L−1.
Figure 4. Seawater deploymenttrace metals: average (n = 3) mass of analyte accumulated by MBL-DGT (■) and Chelex-DGT (○). Error bars represent the standard deviation of triplicate measurements. Dotted lines give the predicted uptakes calculated using the DGT equation. Experimental conditions: pH 8.02 ± 0.04; salinity 32.4 μS cm−1; temperature 23.9 °C. Average analyte concentrations were as follows: Mn, 12.7 μg L−1; Co, 10.4 μg L−1; Ni, 10.8 μg L−1; Cu, 10.6 μg L−1; Cd, 12.8 μg L−1; Pb, 8.5 μg L−1.
432
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
Analytical Chemistry
Article
accumulation was not adversely affected by competition over the 4 day period. Further research is required to evaluate this observation. For Pb, both MBL-DGT and Chelex-DGT measured similar concentrations across the 4 day deployment; however, both CMBL and CChelex were lower than Csol (CMBL/Csol = 0.80−0.82; CChelex/Csol = 0.70−0.74). This discrepancy is unlikely to be due to the presence of colloidal Pb in solution, as speciation modeling using Visual MINTEQ predicted that colloidal Pb would not form at the experimental pH. It is possible that, in the synthetic seawater solution, a small fraction of the Pb existed in a kinetically inert form that is not labile on the DGT time scale. Although it remains unclear why CMBL and CChelex are appreciably lower than Csol, it is important to note that both MBL-DGT and Chelex-DGT measured similar Pb fractions. Furthermore, this is the first time that long-term seawater DGT deployments have been studied under controlled conditions for Pb and many of the other analytes investigated in this paper. Further research may be required to fully explain this observation.
For the trace metals (Mn, Co, Ni, Cu, Cd, and Pb), ChelexDGT devices were deployed in triplicate for the same sampling times as for the MBL-DGT samplers (Figure 2, open circles). Good agreement was obtained between CChelex and Csol (CChelex/Csol = 0.85−1.19) and, hence, between CMBL and CChelex as well (CMBL/CChelex = 0.88−1.11). Synthetic Seawater. DGT samplers containing MBL, Metsorb, and Chelex binding layers were deployed in synthetic seawater spiked with V, W, Mo, As, P, Sb, Mn, Co, Ni, Cu, Cd, and Pb (8.5−32 μg L−1) for up to 4 days. For all analytes, with the exception of P, the solution concentration of analyte at the end of the deployment was within 10% of the initial concentration. The concentration of P decreased by 20% throughout the experiment, likely due to microbial uptake of P, and as a result the dashed lines in Figures 3 and 4 for the predicted mass of P accumulated are nonlinear. Plots of analayte mass accumulated over time for MBL-DGT, Metsorb-DGT, and Chelex-DGT are shown in Figures 3 and 4. For V, As, Sb, W, and P, CMBL was in good agreement with Csol (CMBL/Csol = 0.85−0.96) and CMetsorb (CMBL/CMetsorb = 0.95− 1.06) over the entire 4 day deployment. The concentration of Mo in seawater could not be measured accurately by either MBL-DGT or Metsorb-DGT (CMBL/Csol = 0.26−0.79; CMetsorb/ Csol = 0.37 and 0.28 at 77 and 99 h, respectively). These results are in agreement with those of Panther and co-workers, who demonstrated that Metsorb-DGT and ferrihydrite-DGT were incapable of measuring Mo to within 20% of the actual solution concentration for deployment times >4 h, presumably due to competition between the Mo and major anions in solution.16 For Co, Ni, Cu and Cd, CMBL was in good agreement with both Csol (CMBL/Csol = 0.89−1.14), and CChelex (CMBL/CChelex = 0.92−1.12) over the entire 4 day deployment. The analytes Mn and Pb require further discussion. For Mn, good agreement was obtained between CMBL and Csol (CMBL/Csol = 0.90−0.99) over the 4 day deployment. However, for Chelex-DGT, the concentration of Mn measured in solution was only accurate up to 2 days (CChelex/Csol = 0.92 and 0.94 at 24 and 52 h, respectively). At 77 h CChelex/Csol was 0.76, and at 100 h the corresponding value was 0.51. In fact, the mass of Mn accumulated by Chelex-DGT at 100 h was less than that accumulated after 77 h, suggesting replacement of Mn on the Chelex adsorbent by other ions in solution. Similar results for Mn have been reported by Tankere-Muller et al.,34 who demonstrated that for solutions containing Ca and Mg at concentrations typical of seawater at pH 5−6, the mass of Mn accumulated by Chelex-DGT was only 15−30% of the predicted mass after a 24 h deployment. Tankere-Muller et al.34 established that this effect was due to Mg2+ and Ca2+ competing with Mn2+ for binding sites on the Chelex adsorbent; similar competition effects on DGT measurements have also been reported by other researchers.15,19,26 The degree of underestimation was not as evident in our work with quantitative measurements of Mn at 24 and 52 h. However, the experiments performed in the current study were carried out at pH 8.05 and it is likely that the binding of Mn to the Chelex resin is more efficient at higher pH. The reason for quantitative uptake of Mn by MBL-DGT, but not Chelex-DGT, may be due to better selectivity of the Metsorb adsorbent for Mn. The binding sites on the Metsorb adsorbent (titanium hydroxide) may be capable of binding Mn strongly but only binding Mg and Ca weakly; hence, there would be less competition from these ions. Alternatively, the use of two resins may have increased the overall capacity of the binding layer so that Mn
■
CONCLUSIONS This work has described a new mixed binding layer DGT technique for the measurement of labile trace metal and oxyanions in freshwater and seawater. Mixing the Chelex-100 and Metsorb binding agents allows the use of a single DGT sampler, rather than separate samplers, for trace metals and oxyanions. The efficacy of the MBL-DGT technique was comparable to that of the Chelex-DGT and Metsorb-DGT techniques separately and even exceeded the performance of Chelex-DGT for the measurement of Mn in seawater. This is the first time that a MBL-DGT method containing Metsorb and Chelex-100 has been described and is the first comprehensive evaluation of a mixed binding layer for such a large suite of analytes (Mn, Co, Ni, Cu, Zn, Cd, Pb, V, As, Mo, Sb, W, and P). By optimization of the elution procedure, the MBL-DGT technique could be further developed for the measurement of Fe,8 Al,15 reduced Cr,35 U,22 and the rare-earth elements (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Tb, and Y).30,36 This technique could potentially be used to measure at least 33 elements simultaneously. The Chelex-Metsorb MBL described in the current study has considerable benefits over the previously described Chelexferrihydrite binding layer.24,25 The commercial availability of both adsorbents (Chelex-100 and Metsorb) is a significant advantage. Ferrihydrite requires careful preparation under controlled conditions to ensure that the correct mineral phase has been synthesized.23 Ferrihydrite has also been shown to underestimate the concentrations of P, As, Sb, and W in seawater and of Mo and Sb in freshwater, in comparison to the Metsorb binding agent.11,16,19 Furthermore, it would not be possible to measure Fe with the Chelex-ferrihydrite binding layer due to the high Fe background that would be associated with the ferrihydrite; this would not be an issue with the Chelex-Metsorb MBL. The Chelex-Metsorb MBL-DGT technique is a promising tool for the environmental assessment of a variety of surface waters and could be easily adapted to the measurement of metals and oxyanions in soils and sediment.
■
ASSOCIATED CONTENT
S Supporting Information *
Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org. 433
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
Analytical Chemistry
■
Article
(27) Zhang, H.; Davison, W. Anal. Chim. Acta 1999, 398, 329−340. (28) Warnken, K. W.; Zhang, H.; Davison, W. Anal. Chim. Acta 2004, 508, 41−51. (29) McCurdy, E.; Potter, D. Spectrosc. Eur. 2001, 13, 14−21. (30) Garmo, Ø.; Røyset, O.; Steinnes, E.; Flaten, T. Anal. Chem. 2003, 75, 3573−3580. (31) Gustafsson, J. P. A. Windows Version of Visual MINTEQ 3.0; http://www2.lwr.kth.se/English/OurSoftware/vminteq/. (32) Diaz-Cruz, J. M.; Esteban, M.; Van den Hoop, M. A.; Van Leeuwen, H. P. Anal. Chem. 1992, 64, 1769−1776. (33) Giusti, L.; Hamilton Taylor, J.; Davison, W.; Hewitt, C. Sci. Total Environ. 1994, 152, 227−238. (34) Tankéré-Muller, S.; Davison, W.; Zhang, H. Anal. Chim. Acta 2012, 716, 138−144. (35) Ernstberger, H.; Zhang, H.; Davison, W. Anal. Bioanal. Chem. 2002, 373, 873−879. (36) Garmo, Ø. A.; Lehto, N. J.; Zhang, H.; Davison, W.; Røyset, O.; Steinnes, E. Environ. Sci. Technol. 2006, 40, 4754−4760.
AUTHOR INFORMATION
Corresponding Author
*P.R.T.: e-mail, p.teasdale@griffith.edu.au; tel, +61 7555 28358; fax, +61 7555 28067. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank Graver Technologies (www.gravertech.com) for providing the Metsorb product used in this study. We also thank the two anonymous reviewers for their helpful comments and suggestions.
■
REFERENCES
(1) Buffle, J. Horvai, G., In Situ Monitoring of Aquatic Systems; Wiley: Hoboken, NJ, 2000. (2) Batley, G. E.; Apte, S. C.; Stauber, J. L. Aust. J. Chem. 2004, 57, 903−919. (3) Allan, I. J.; Vrana, B.; Greenwood, R.; Mills, G. A.; Roig, B.; Gonzalez, C. Talanta 2006, 69, 302−322. (4) Kot-Wasik, A.; Zabiegała, B.; Urbanowicz, M.; Dominiak, E.; Wasik, A.; Namieśnik, J. Anal. Chim. Acta 2007, 602, 141−163. (5) Namieśnik, J.; Zabiegała, B.; Kot-Wasik, A.; Partyka, M.; Wasik, A. Anal. Bioanal. Chem. 2005, 381, 279−301. (6) Vrana, B.; Allan, I. J.; Greenwood, R.; Mills, G. A.; Dominiak, E.; Svensson, K.; Knutsson, J.; Morrison, G. TrAC, Trends Anal. Chem. 2005, 24, 845−868. (7) Davison, W.; Fones, G.; Harper, M.; Teasdale, P. Zhang, H. In In Situ Monitoring of Aquatic Systems: Chemical Analysis and Speciation; ed. Buffle, J., Horvai, G., Eds.; Wiley: Chichester, U.K., 2000. (8) Zhang, H.; Davison, W. Anal. Chem. 1995, 67, 3391−3400. (9) Davison, W.; Zhang, H. Environ. Chem. 2012, 9, 1−13. (10) Bennett, W. W.; Teasdale, P. R.; Panther, J. G.; Welsh, D. T.; Jolley, D. F. Anal. Chem. 2010, 82, 7401−7407. (11) Bennett, W. W.; Teasdale, P. R.; Panther, J. G.; Welsh, D. T.; Jolley, D. F. Anal. Chem. 2011, 83, 8293−8299. (12) Hutchins, C. M.; Panther, J. G.; Teasdale, P. R.; Wang, F.; Stewart, R. R.; Bennett, W. W.; Zhao, H. Talanta 2012, 97, 550−556. (13) Luo, J.; Zhang, H.; Santner, J.; Davison, W. Anal. Chem. 2010, 82, 8903−8909. (14) Osterlund, H.; Chlot, S.; Faarinen, M.; Widerlund, A.; Rodushkin, I.; Ingri, J.; Baxter, D. C. Anal. Chim. Acta 2010, 682, 59−65. (15) Panther, J. G.; Bennett, W. W.; Teasdale, P. R.; Welsh, D. T.; Zhao, H. Environ. Sci. Technol. 2012, 46, 2267−2275. (16) Panther, J. G.; Stewart, R. R.; Teasdale, P. R.; Bennett, W. W.; Welsh, D. T.; Zhao, H. Talanta 2013, 105, 80−86. (17) Panther, J. G.; Stillwell, K. P.; Powell, K. J.; Downard, A. J. Anal. Chem. 2008, 80, 9806−9811. (18) Panther, J. G.; Stillwell, K. P.; Powell, K. J.; Downard, A. J. Anal. Chim. Acta 2008, 622, 133−142. (19) Panther, J. G.; Teasdale, P. R.; Bennett, W. W.; Welsh, D. T.; Zhao, H. Environ. Sci. Technol. 2010, 44, 9419−9424. (20) Stockdale, A.; Davison, W.; Zhang, H. Environ. Chem. 2008, 5, 143−149. (21) Stockdale, A.; Davison, W.; Zhang, H. J. Environ. Monit. 2010, 12, 981−984. (22) Turner, G. S.; Mills, G. A.; Teasdale, P. R.; Burnett, J. L.; Amos, S.; Fones, G. R. Anal. Chim. Acta 2012, 739, 37−46. (23) Zhang, H.; Davison, W.; Gadi, R.; Kobayashi, T. Anal. Chim. Acta 1998, 370, 29−38. (24) Mason, S.; Hamon, R.; Nolan, A.; Zhang, H.; Davison, W. Anal. Chem. 2005, 77, 6339−6346. (25) Huynh, T.; Zhang, H.; Noller, B. Anal. Chem. 2012, 84, 9988− 9995. (26) Panther, J. G.; Teasdale, P. R.; Bennett, W. W.; Welsh, D. T.; Zhao, H. Anal. Chim. Acta 2011, 698, 20−26. 434
dx.doi.org/10.1021/ac402247j | Anal. Chem. 2014, 86, 427−434
