Environ. Sci. Technol. 2008, 42, 6903–6909
Interpretation of In Situ Speciation Measurements of Inorganic and Organically Complexed Trace Metals in Freshwater by DGT KENT W. WARNKEN, WILLIAM DAVISON,* AND HAO ZHANG Department of Environmental Science, Lancaster Environment Center (LEC), Lancaster University, Lancaster LA1 4YQ, United Kingdom
Received February 5, 2008. Revised manuscript received May 6, 2008. Accepted May 14, 2008.
The dynamic speciation technique, diffusive gradients in thinfilms (DGT), has been used in freshwater to determine simultaneously, from a single set of in situ measurements, (1) the equilibrium distribution of metal ions between simple inorganic complexes and larger organic complexes and (2) information on the rates of dissociation of these complexes. DGT devices with different diffusion layer thicknesses (0.3, 0.54, 1.34, and 2.14 mm) were used to estimate the in situ dissociation kinetics. Information on the species distribution was obtained by using two types of gel, which allow relatively free (polyacrylamide, APA) and more retarded (restricted, RES) diffusion of the metal complexes. The full theoretical basis of the technique is developed and applied to in situ measurements of Mn, Fe, Co, Ni, Cu, Cd, and Pb in a pristine river (Wyre, UK), with high DOC (15 mg L-1), assuming that organic complexes are dominated by fulvic acid. These first DGT measurements that do not rely on assumptions about complex lability or the distribution of species, are compared to total dissolved measurements, previously reported speciation calculations and measurements using alternative speciation techniques. Examination of calculation consistency suggests that the effective mean diffusion coefficients of metal complexes with organic matter under in situ conditions may be larger than those measured in the laboratory using extracted fulvic acid.
Introduction Despite decades of research, measurement of trace metal speciation in fresh water has proved challenging. The importance of the free ion activity in models describing biological uptake, as incorporated in the free ion activity model (FIAM) (1) and biotic ligand model (BLM) (2), has focused attention on equilibrium based techniques. Ion selective electrodes (ISE) (3), competitive ligand exchange (CLE) (4), the Donnan membrane technique (DMT) (5) and permeable liquid membranes (PLM) (6) have all been used to estimate free ion activity. Recognition that the assumption of the FIAM and the BLM, of neglecting the dynamics of metal exchange in solution, may be inappropriate for some situations, has led to the development of more complete uptake models that consider complex dissociation kinetics (7). This has renewed interest in dynamic measuring * Corresponding author phone: +44 (0) 1524 593935; fax: +44 (0) 1524 593985; e-mail:
[email protected]. 10.1021/es800359n CCC: $40.75
Published on Web 08/16/2008
2008 American Chemical Society
techniques, which remove metals from solution in a controlled manner (8). The measured response is a flux, which is sensitive to two key factors: kinetic, as exemplified by the dissociation rate of metals from their complexes, and mobility, as determined by the diffusion coefficients of metal complexes. The long-serving dynamic technique of voltammetry is prone to interferences from natural organic substances, but this problem has been overcome by using a gel integrated microelectrode (GIME) (9). GIME has been used for in situ measurements that can be obtained at frequent (minutes) time intervals (10). DGT (diffusive gradients in thin-films) is a dynamic technique that is readily used in situ (11, 12). It accumulates metals in a binding layer after they have diffused through a hydrogel. The flux is an integrated value, as it is calculated from the metal accumulated during the whole deployment time, which is typically 1 day to several weeks. Dynamic techniques directly measure the flux of metals to an infinite sink. As it is difficult to distinguish the effects of mobility and lability (defined here as the kinetic limitation to supply), their measured signal in natural waters can not generally be interpreted directly in terms of concentrations of individual species or groups of species. Anodic stripping voltammetry was used to try to determine solely inorganic components in marine systems by assuming that contributions from organic complexes were negligible, when deposition potentials and hydrodynamic conditions were optimized for their discrimination (13). By using hydrogels with different compositions and diffusional properties, which were calibrated with respect to the diffusion coefficients of the inorganic and organic complexes (assumed to be dominated by fulvic acid, FA), it was possible to use DGT to determine the concentrations of metals in two groups: simple inorganic species and organic complexes (metal-FA) (14, 15). However, it was necessary to assume that all complexes were labile. DGT has previously been used in situ in freshwater to measure the overall effective rates of dissociation of metals from their complexes (16). The approach used several devices with different diffusion layer thicknesses, exploiting the fact that the extent of complex dissociation is determined by the time taken to traverse this layer. However, interpretation as a dissociation rate constant was only possible if prior knowledge of the distribution of species was assumed. By combining the two approaches, i.e., one using diffusion layers with different diffusional properties and one using diffusion layers with a range of thicknesses, it should be possible to obtain, simultaneously, information on the distribution of species and their dissociation kinetics. This paper reports the development of the required theory and its application to in situ DGT data, to provide the first measurements of the distribution of species corrected for kinetic limitations.
Materials and Methods Field Site. DGT samplers were deployed during April 19-22, 2004 in the River Wyre, at Garstang, Lancashire (NW England). The mean total dissolved trace metal concentrations (CTF) in this pristine water, calculated from seven samples collected during DGT deployment, are given in Table 1. The primarily moorlands catchment is partially covered in peat, so dissolved organic carbon (DOC) is quite high at 15 mg C L-1. As fulvic acid (FA) is the major organic component in solution in these waters (3), it was assumed that its complexes dominated speciation. Rainfall was monitored at Hazelrigg meteorological field station (DCNN7236), only a few kilometers from the sampling site. VOL. 42, NO. 18, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
6903
TABLE 1. Accumulated Mean Masses of Metals Measured in DGT devices Deployed In Situ and Fitted with APA (MAPA) or RES (MRES) Gelsa ∆g MAPA (nmol) MAPA (nmol) MAPA (nmol) MAPA (nmol) MAPA (nmol) MRES (nmol) SDRES (nmol) gkin(W) (cm) gkin(fit) (cm) CTF (nM) CM (nM) CML (nM) Cnd (nM) lab (nM) CM lab (nM) C ML
0.30 0.54 0.94 1.34 2.14 0.94 0.94
Mn 14.1 10.6 6.92b 5.01 3.62 5.09 0.758 0.013 0.010 242 223 47.4 -28.4 262 -ve
Fe
Co
Ni
27.3 22.7 19.1b 16.0 12.7 17.1 2.71 0.116 0.118 5200 817 1530 2860 947 -ve
0.0357 0.0272 0.0186b 0.0138 0.010 0.0128 0.00212 0.015 0.020 1.63 0.50 0.76 0.37 0.63 -ve
0.367 0.291 0.208b 0.159 0.120 0.110 0.0126 0.031 0.042 22.4 2.53 29.4 -9.6 4.4 12.3
Cu
Cd
Pb
0.119 0.100 0.0730b 0.0583 0.0419 0.0428 0.00721 0.053 0.051 20.4 1.13 8.93 10.4 1.75 2.39
0.00454 0.00332 0.00213b 0.00162 0.00104 0.00142 0.000212
0.00422 0.00275 0.00186b 0.00147 0.000900 0.00125 0.000140
0.197 0.054 0.045 0.098 0.066 0.003
1.83 0.037 0.025 1.77 0.045 -ve
a The standard deviations for RES gels is SDRES. The kinetic terms calculated from the full model fit, gkin(fit) and using WHAM6 to predict the speciation, gkin(W), are shown. The mean total filterable metal, CTF, is given. Inorganic metal species, CM, organically complexed metal, CML, and nondynamic metal, Cnd, are derived from the full model fits. The values of CML lab and Clab respectively. b Calculated values. and CM calculated assuming all complexes are fully labile are shown as CML M
Total Dissolved Trace Metals. Sampling for total dissolved trace metals was generally carried out twice a day: morning and late afternoon. Briefly, duplicate samples were collected using all-plastic syringes, after first rinsing several times with ambient river water, and immediately filtered through a 0.4 µm Nuclepore filter membrane into a Nalgene 15 mL lowdensity polyethylene (LDPE) bottle containing 300 µL of HNO3 (2% v/v). The bottles were immediately capped and placed into double acid cleaned plastic bags, which were then stored in a refrigerator until analysis. The entire sampling procedure takes
