Metal adsorption and partitioning

Like in sediments, metals in soils are part of a complex mixture of solid phase compounds of varying particle size and morphology, including discrete mineral phases, co-precipitated and sorbed species associated with soil minerals and organic matter, dissolved species complexed by a variety of inorganic and organic ligands, including a vast variety of macro- and micro-organisms. Although geochemically similar, soils are exposed to the ambient air and the result of natural weathering processes, while sediments are exposed to water and built up by solutes and particulates entering or formed in the overlying water. Based on their geochemical similarity, similar speciation methods for soils and sediments have been developed.

A first approximation of the speciation of metals in soils can be achieved by assessing their partitioning to the soil particulate matter, and to soil solution. Unlike organic contaminants, metal adsorption in soils is much metal- and soil-specific, meaning that the assessment requires the actual experimental determination of solid-liquid partitioning coefficients (kd). Partitioning coefficients simply help to differentiate between metals dissolved in the soil solution and those bound to the soil particulate phase. From this, however, no information is available concerning the relative bioavailability of dissolved metal species. Recent studies emphasize that pH is still the most beneficial and universal site-specific kd adjustment parameter, although the sorption capacity of a sorbent also depends on a variety of other soil parameters. This is why kd values from low-contaminated soils cannot be applied to assess the partitioning of metals at contaminated sites. Measured kd values depend in particular on total dissolved metal concentrations and on dissolved organic matter (DOM), why every factor influencing these main parameters also will affect metal partitioning and speciation. Today, competitive adsorption models rather than solid-solution ratios (as expressed by kd) are increasingly used to predict dissolved metal concentrations.

The realization that metals contribute to plant uptake also from pools other than soil solution has been delayed during the 60ies, mainly due to the lack of simple realistic methods to assess the rate and extent to which solid-bound metals are released and taken up by biota. But we now know that the bioavailable pool of metals includes the free metal ion in solution, and a labile pool, which buffers the free metal concentration in solution, and includes both species on soil solids or soluble organic matter (DOC). A first approximation to the speciation of metals in soils may be simply done (like for sediments) by partitioning between metals bound to the soil particulate matter, and those dissolved in soil solution. It is the dissolved metal pool that is in general reflecting metals that are easily taken up by plant roots or other soil biota, or leached from the soil, to contaminate ground and surface waters. However, unlike organic contaminants, metal adsorption in soil is much metal- and soil-specific, and partitioning requires the actual experimental determination of solid-liquid partitioning coefficients (kd). It should be noted, however, that such site-specific kd's may not ensure a correct assessment of the fate of metals under changing conditions, but they allow an estimate of dissolved metals, and to predict metal mobility (compare also to section 5.2.2 and 5.4.7.2).

Sauve et al. (2000) provided a compilation of soil solid-liquid metal partitioning data to identify default kd-values (L/kg) for various risk assessment models, but also to evaluate the dependence of kd on soil characteristics, like pH, total metal load and organic matter. The authors point out that sorption models using a single-valued kd approach presume that the sorption capacity of a sorbent is independent of any soil characteristics, which we know that it is not. An example is given (ref. therein), where Zn and Pb reacted in opposite ways to increasing total soil concentrations. In this case, higher dissolved Pb concentrations increased adsorbed Pb and the corresponding kd, in contrast to Zn, where higher soil levels decreased kd values reflecting a lowered affinity of solids for Zn at higher concentrations. Also for this reason, kd values from low-contaminated soils cannot be applied to study the partitioning of metals at contaminated sites. In addition, it remains difficult to distinguish between the influence of pH and other soil characteristics, as they are mostly autocorrelated.

The authors (Sauve et al. 2000) further explain that kd values above all depend on total dissolved metal concentrations, i. e. the sum of free metal pools, inorganic ion pairs, like MeOH+, Me(OH)20, Me(OHy, MeHCO3+, MeCOa0, Me(COa)22-, MeNO3+, MeCl+, MeSO40, etc., and of dissolved organic matter (DOM) complexes, which in turn are particularly affected by pH. DOM strongly influences metal solubility and hence the level of total dissolved metals (the authors give an example, where > 98 % of dissolved Cu was bound to DOM in non-acidic soil solution). So, for example, every factor influencing soil organic matter (SOM) will also affect metal solubility. Another example is given, where Ca was found to promote the coagulation of DOM, which in turn reduces its solubility, and so also the solubility of other trace metals bound to DOM. In contrast, Ca and other cations compete with metals for surfaces and so may again enhance their solubility. The partitioning of SOM to particles, in particular its fulvic:humic ratio is strongly a function of the given soil solution pH. Multiple binding site models have been developed (ref. given) to adequately describe the interaction between organic matter and metals over a wide range of concentrations.

Also inorganic ligands are recognized for their capacity to solubilize metals depending on their concentration and corresponding dissociation constants. However, the influence of main geochemical parameters (i. e. of pH, DOM and competing or complexing ions) on metal adsorption is not well defined by the kd-approach. kd values simply help to differentiate between metals dissolved in the soil solution and those bound to the solid phase. From this, no information is available, for instance, on the relative bioavailability of various dissolved metal species. As part of their study, Sauve et al. (2000) applied a 'competitive adsorption model' to predict the solubility of Cd, Cu, Pb and Zn in soil, assuming that free metal (Me2) and H+ ions compete for sorption on available soil exchange sites. The study confirmed that most of the observed variability depends on solution pH and on total soil metal content, beside DOM. It is suggested that a long-term stabilization of the collected field soils will move metals closer to equilibrium regarding sorption, precipitation or specific retention ("aging effect"). Their work further demonstrate that the linear relationship between kd and soil solution pH can explain more of the observed variability than does total metal burden (i. e. between 29-47 % for Cd, Cu and Pb, and between 56-58 % for Ni and Zn). These studies emphasize that pH is still the most beneficial and universal site-specific kd adjustment parameter. The authors stress, however, that competitive adsorption models are actually predicting dissolved metal concentrations and not simply solid-solution ratios (as expressed by the kd). For this reason it is supposed to use adsorption models rather than kd values to directly predict dissolved metals, when assessing the environmental risk and fate of metals in contaminated soils.

There is an overwhelming evidence today that the lability of metals (i. e. their mobility and availability) vary greatly with the properties of a particular soil for similar total metal concentrations. In order to evaluate the effect of different soil:solution extraction ratios and the role of soil properties on metal desorption processes, Yin et al. (2002) conducted desorption experiments. Their data showed that distribution coefficients (kd = Msoil/Msolution) increase with increasing soil:solution ratios, as well as with soil organic matter (SOM). There was in particular a strong affinity of Cu to operationally defined dissolved organic matter (DOM) compounds, which was reflected by a linear correlation between the kd for Cu and the kd for DOM. Furthermore, the soil-solution distribution of Ni, Zn, Cu2+ correlated closely to SOM but not to DOM. So, normalization of the kd-values for Zn, Cu and Ni by SOM in the same soil improved the linear relationship of non-normalized kd-values and soil-pH. The authors conclude that (organic matter) normalized regression equations are more convenient to predict Ni and Zn solubility and the free Cu2+ activity in the soil as a function of pH.

Aldrich et al. (2002) studied the input of Zn and Cu into agricultural soils because of their increasing use in feed additives (manure), fertilisers and fungicides. Cu and Zn contents in the respective test soils are summarized in Table 5.12. From there it can be seen that the humic soil in 'Baltenswil' exceeds the Swiss soil guideline for total Zn and Cu concentrations by a factor of 2. Dissolved metal concentrations remained below the guideline, but increased in the drainage water during rain events. Dissolved Cu was almost completely complexed by organic ligands (> 99.9%), whereas dissolved Zn in the drainage water was partly electrochemically labile, and partly complexed by strong organic ligands. The authors resume that these metals are hardly bioavailable in the soil drainage water due to the very low free metal ion concentration found.

Table 5.12. Total and soluble Cu and Zn contents (|lg/g dw) in two Swiss soils treated with manure, a Zn-containing fungicide (Propineb with 22.6% of Zn) and Cu salt (50% Cu) (from Aldrich et al., 2002)

Sampling site Cu totala Cu solubleb Zn total Zn soluble Mg/g ^g/g ^g/g ^g/g

Lindau 48 0.05 82 0.2

Baltenswil 78 0.2 284 0.3

Swiss guideline 40 0.7 150 0.5

"Extraction with 2 M HNO3. "Extraction with 0.1 M NaNO3

0 0

Post a comment