Measurement and computation of zinc binding to natural dissolved organic matter in European surface waters
Introduction
Complexation of trace metals by natural organic ligands has an important influence on the speciation of trace metals in natural waters, with consequent influences on their transport and bioavailability. The primary forms of bound zinc in natural waters are zinc complexed by organic and inorganic ligands, and zinc sorbed onto particulate matter. Complexation by dissolved organic matter (DOM) in natural waters may significantly influence free Zn2+, which is of special interest to ecotoxicologists as it is considered to be the most bioavailable or toxic species [1]. Therefore, in order to understand zinc toxicity in water bodies, we need to study the zinc complexation characteristics of natural DOM.
Modeling is an important tool in quantifying the speciation of metals in the environment. The equilibrium speciation of metals in the presence of simple inorganic and organic ligands with well-defined chemical nature is well established. Less well established are the interactions of protons and metals with the chemically complex and heterogeneous humic substances, which comprise the dominant fraction of dissolved organic matter in surface waters. Despite the difficulties of modeling humic substances, ion binding models have been developed (e.g. Models V, VI, and NICA) that have successfully modeled proton and metal binding to isolated humic substances in the laboratory [2], [3], [4]. Of these Model V has become an integral part of recently developed biotic ligand models (BLM), which can predict metal bioavailability and toxicity at different water chemistries [5], [6], [7], [8], [9]. Such models can in principle be applied to field data to predict metal speciation in surface waters, given suitable data on the concentrations of reactive components in the system. However, the metal binding behaviour of unfractionated organic matter in surface waters may differ on a weight for weight basis from that of isolated humic substances, due to a number of reasons (for example, differences in chemical composition, or chemical alterations to humic substances on isolation). Therefore, it is not possible to directly predict metal speciation in surface water on the basis of a single measurement of DOM. What can be done is to test the ability of the model to describe metal complexation in natural waters by comparing the measured DOM with the concentration of humic substances required by the model to fit the data. This has previously been done by Bryan et al. [10], who used Model VI to describe the binding of copper spiked into natural waters. They found that for each surface water sample, Cu binding data could be described by a single model concentration of fulvic acid (FA). Taken across all samples the model FA concentration was on average 65% of the measured DOM. Taking this mean ‘binding activity’ of DOM, Bryan et al. concluded that for a total copper concentration of 1 μM, free copper could be predicted to within a factor of 3.6 in 95% of natural waters. There is other encouraging, although very limited, evidence that a single unified parameter set can describe the binding properties of DOM over large spatial and temporal scales. McKnight et al. [11] modeled the binding of Cu by DOM from sites across the U.S. and concluded that a single parameter set for DOM binding of Cu could generally predict the free Cu2+ activity in natural waters to within a factor of five, and Lu and Allen [12] also found similarity in copper binding by DOM from different North American surface water sites. In contrast to Cu, information on the comparison of zinc binding characteristics of DOM from different surface water sources is still limited. A very limited number of data, obtained mainly from ion exchange methods, are available for the calibration of Zn–DOM binding in current models. In addition, there are a few membrane-based methods and voltammetric techniques available to measure free zinc activity [13], [14], [15]. Anodic stripping voltammetry (ASV) has also been used for speciation studies due to its very low detection limit. Xue and Sigg [16] determined free zinc activity in lake water samples using ligand exchange and differential pulse anodic stripping voltammetry (DPASV). However, data from ASV measurements are often difficult to interpret because ASV measures the “labile metal” rather than “free metal”. The “free metal” needs to be calculated from the “labile metal” and other system conditions, and such studies have to our knowledge not been reported for Zn in systems that contain heterogeneous natural organic matter as major ligands for metal binding.
This paper reports data from experiments conducted on Zn–DOM systems of a variety of European surface water sources, characterized by a wide range of physico-chemical and topological characteristics (e.g. lakes versus rivers, small versus large). Reverse osmosis (RO) was used to isolate DOM as it is a highly efficient technique for rapid collection of large quantities of DOM in surface waters [17], [18]. Recently it has also been demonstrated that this isolation method does not affect the protective effect of DOM on copper or zinc toxicity to the cladoceran Daphnia magna or the green alga Pseudokircheneriella subcapiata[19], suggesting that alteration of metal binding by this method is minimal. The zinc complexation properties of the DOM samples collected were determined by titration with Zn, using ASV to measure the labile Zn activity and free Zn activity was computed from the labile Zn activity and other system conditions. The robustness of the current representation of equilibrium Zn–DOM complexation in Models V and VI was tested with an approach previously used for similar titrations of natural waters with Cu [11], [20]. The challenge for the model in this situation is to describe the relationship between total and free Zn with a plausible ‘binding activity’ of DOM.
Section snippets
Sampling of natural DOM
DOM was concentrated in the field using a portable reverse osmosis (RO) system as described by Serkiz and Perdue [17] and Sun et al. [18]. Sampling was performed at five sites reflecting a large range of physico-chemical characteristics occurring in Europe. Table 1 gives a description of these sampling sites, along with some physico-chemical characteristics.
Bihain is characterized by low pH, Ca and Mg and medium DOC. It is a small creek (width 2 m, depth 1 m) originating from and running through
Results and discussion
In considering the results it is important to appreciate that the application of Models V and VI is essentially a curve-fitting exercise, the objective of which is to determine whether the model parameters (as derived from literature data) can describe the titrations of natural DOM in a chemically plausible manner, i.e. with optimal DOMFA values that are reasonably close to unity.
The zinc titration data of the five DOM samples are presented in Fig. 1. The free zinc activity was calculated by
Acknowledgments
This research was supported by the International Lead Zinc Research Organization (ILZRO) and by the Ghent University Research Fund (BOF No. 01110501). Karel De Schamphelaere was supported in part by a Ph.D. grant from the Flemish Institute for the Promotion of Scientific and Technological Research in Industry (IWT-Vlaanderen), by a postdoctoral fellowship from the Scientific Research Fund – Flanders (FWO-Vlaanderen), and by the ICA Chris Lee Award for Metals Research and the Society of
References (34)
Comp. Geosci.
(1994)- et al.
Comparative Biochem. Physiol. C
(2002) - et al.
Comparative Biochem. Physiol.
(2002) - et al.
Sci. Total Environ.
(1983) - et al.
Water Res.
(2002) Chemosphere
(1999)- et al.
Anal. Chim. Acta
(1994) - et al.
Water Res.
(1990) - et al.
Water Res.
(1995) - et al.
Chemosphere
(2005)
Environ. Int.
J. Electroanal. Chem.
J. Electroanal. Chem.
J. Electroanal. Chem.
Geochim. Cosmochim. Acta
Environ. Pollut.
Geochim. Cosmochim. Acta
Cited by (45)
Predicting trace metal solubility and fractionation in Urban soils from isotopic exchangeability
2017, Environmental PollutionSignificance of particle size and charge capacity in TiO<inf>2</inf> nanoparticle-lipid interactions
2016, Journal of Colloid and Interface ScienceInfluence of dissolved organic matter on photogenerated reactive oxygen species and metal-oxide nanoparticle toxicity
2016, Water ResearchCitation Excerpt :It is noteworthy that the addition of HA into ZnO and Fe2O3 suspensions did not increase 1O2 generation. This is probably because the released metal ions complex with carboxyl and phenolic groups of HA (Fujii et al., 2014; Neubauer et al., 2013; Cheng and Allen, 2006; Cheng et al., 2005), efficiently deactivate the triplet states of HA (Carlos et al., 2012), and subsequently results in the decreased sensitization capacity of HA for 1O2 generation by ZnO and Fe2O3. In the dark, no measurable amount of 1O2 was detected for any of these NPs with or without DOM (data not shown).
Zinc solubility and fractionation in cultivated calcareous soils irrigated with wastewater
2015, Science of the Total EnvironmentCitation Excerpt :The free Zn ion activity (Zn2 +) was estimated using the geochemical speciation model WHAM(VII) (Windermere Humic Aqueous Model, version 7) (Lofts and Tipping, 2011) using the following input parameters: (i) solution concentration of [Zn], [Ca], [CO3] and [NO3] (mol L− 1 of suspension) as dissolved species, (ii) the suspension pH, and (iii) a fulvic acid colloidal phase (from dissolved organic carbon, DOC); the temperature was set to 25 °C and the partial pressure of CO2 was set to 0.03. It was assumed that fulvic acid contained 50% C and also that 50% of the DOC was fulvic in origin (Buekers et al., 2008; Cheng et al., 2005). The descriptive statistics for selected elemental concentrations, pH and %CaCO3 in the soil datasets are summarised in Table 1.