Soils

16.4.1 Practicalities for deployments in soils

DGT measurements in soils initially used the standard piston assembly supplied by DGT Research Limited (Lancaster, UK) that is regularly used for solution measurements [17,30,31]. More recently, a device specially designed for soils has become available [32]. It is similar in appearance to solution devices, but it has an improved seal between the cap and pre-filter to prevent soil particles from entering the device and biasing measurements. Its exposure window consequently has a smaller area (2.54 cm2). Initially, soils are dried and sieved to remove particles larger than 2 mm. Hydration is carried out in two stages, which allows time for equilibration and avoids anoxic conditions. Typically, the soils are hydrated to 60% of the maximum water holding capacity, well mixed and left to stand for 2 days. They are further hydrated to 80-100% maximum water holding capacity, mixed to a smooth paste or slurry and left for an additional 24 h. Prior to deployment, a small amount of soil paste is gently smeared onto the surface of the exposure window (filter membrane) of the DGT unit, to ensure no pockets of air exist between the DGT sampling face and the soil. It is then gently pressed into the soil surface, using a slight turning action to ensure good contact between the soil and the DGT unit. Upon removal of the DGT device, adhering soil particles are washed off by rinsing with a stream of high-purity water from a wash bottle. Obvious surface water is then removed by blotting with a clean tissue. If the soil paste is very sticky, and it is difficult to rinse the DGT unit with water, it can simply be wiped with clean tissue paper. The subsequent procedures are similar to those used for solution, with the caveat that the derived concentration, CDGT, will be the mean concentration at the surface of the device during the deployment time.

The suitability of these procedures has recently been systematically investigated [32]. Extending the soil hydration times did not affect the

DGT measurement and acceptable results were obtained by sieving a field soil and simply estimating the water needed to bring the soil to a smooth paste. Direct DGT measurements on soils hydrated in situ were acceptable, but less reproducible than measurements on soils returned to the laboratory where they were sieved, homogenised and hydrated prior to deployment [32,33].

16.4.2 Soil dynamics

The first use of DGT in soils measured R values for Cd and Zn in homogenised soils contaminated with sewage sludge and interpreted them directly in terms of the kinetics of the metal's release from the solid-phase to solution [17]. The authors concluded that Cd and Zn each had two pools of metal, characterised by different rate constants for release to solution, but that Cu and Ni each had only one kinetically defined pool. The dependence of the DGT flux on the diffusion layer thickness, used in the above study, has been used more generally to investigate the dynamic availability of metals [34,35]. Another study showed that the amount of metal accumulated by DGT was dependent on the moisture content of the soil [36]. Metal uptake increased with increasing moisture content, reaching a broad maximum corresponding to field capacity, and declined at higher moisture contents. The lower accumulations at low moisture contents were attributed to the more tortuous diffusion path, while the decline at higher values was due to dilution of the concentration of metal in the soil solution. Docekal et al. [34] have confirmed some of these effects of soil moisture on the DGT measurement.

Ernstberger et al. [18] deployed DGT in homogenised slurries from a single soil for various times (4 h to 20 days) and obtained time-dependent values of R for Cd, Cu, Ni and Zn. The 1D DIFS model provided good fits of the plots of R versus time, using Kdl and Tc as adjustable parameters. The work was extended to five soils in a later study, where good fits were also obtained [19] (Fig. 16.3). Cd and Zn showed similar behaviour, with fast supply from the solid-phase of all soils being reflected by relatively high rate constants, while the release of Ni was kinetically limited. Distribution coefficients for labile Cd and Zn agreed well with those measured by isotopic exchange. While they were very dependent on the pH of the soil, the Kdl values for Ni were generally lower and more dependent on soil texture (low in sandy soils) than pH. The good fit to experimental data and agreement between Kdl values determined by different techniques provided validation of the DIFS

Fig. 16.3. Measured values of the ratio, R — CDGT/Csoln, for different deployment times in five soils labelled M, A, G, J and O. The lines show fits to the data using the DIFS model (adapted from Ref. [19]).

Fig. 16.3. Measured values of the ratio, R — CDGT/Csoln, for different deployment times in five soils labelled M, A, G, J and O. The lines show fits to the data using the DIFS model (adapted from Ref. [19]).

model. It was considered that this methodology was unsuitable for Cu, due to pronounced complexation in solution that biased the estimation of R [19]. Nowack et al. [33] used a similar approach to determine Kdl and Tc for Cu and Zn in a contaminated field soil. Best fits of their data were obtained by assuming there were two pools of metal, characteristic of fast and slow release. Ernstberger et al. [18] showed that, as model fits are not very sensitive to the values of Tc and Kdl, the values of these parameters cannot be obtained very accurately. A full sensitivity analysis has defined the precision available for a wide range of parameter values [29].

Degryse et al. [37] used the DIFS model to obtain R values from estimated values of Tc and Kdl and then calculated Csoln for Zn from the measured CDGT. The prediction was generally good, except at very high Zn concentrations, where it failed due to the resin-gel approaching saturation.

Some studies have used values of Kdl determined by comparing solution and extracted concentrations. If this value is combined with a single determination of R, the kinetic parameters Tc and k_1 can be calculated using DIFS. Using this approach, Zhang et al. [38] found that the rate of supply of Zn from solid-phase to solution is much higher in soils freshly spiked with Zn than in contaminated field soils with similar total concentrations of Zn. The rate constant (k_1) for release of As in rhizosphere soil was smaller than in the bulk soil, presumably due to plant roots preferentially removing the most readily available fraction in the rhizosphere [39]. A systematic study of 14 freshly contaminated soils showed that the rate of supply of Zn from the solid-phase was too fast to measure in all except three soils, which had fairly low pH and a silt-sand texture [40]. Rates of supply of Cd could be measured in all except six clay soils.

Several studies have used DGT in soils to measure the flux and interpret it in terms of available metal. Comparisons to conventional leaching procedures were used to aid the interpretation of data on Cd, Cr, Cu, Ni and Pb from three contaminated sites [35]. Rachou et al. [41] focussed on the dependence of available Cd on pH, while Lombi et al. [42] used DGT to demonstrate the in situ fixation of metals in soils treated with a bauxite residue. The mobilisation of metals in organically contaminated soils with high microbial activity has been demonstrated in a series of papers [43-45]. A further series of papers has used DGT to investigate the speciation of Al, Cu, Fe and Zn in solutions extracted from soils [46-50].

Phosphorus has been determined in soils using DGT with ferrihyd-rite in the binding layer, either as a pure phase or mixed with Chelex for the simultaneous determination of metals [51]. DGT-measured P correlated better with P in soil solution than with the standard Colwell and Bray extracts [52], suggesting that the extractions measured more P than that measured by DGT.

16.4.3 Biological mimicry

Like DGT, plants accumulate metals by removing them from soil. The perturbation of the soil is similar if the rate of removal by the plant and DGT is similar. Lehto et al. [53] have modelled the uptake of metals by plants and DGT and shown that fluxes to plants and DGT are generally similar for values of diffusion layer thicknesses typically used in DGT devices. Hyper-accumulator plants will tend to have slightly higher uptake fluxes than DGT.

Processes that are not mimicked by DGT can also affect the supply of metals to plants, including convective transport, the root encountering fresh surfaces as it grows through the soil and the influence of root exudates and microenvironments. For DGT to be effective in predicting plant uptake, the contribution from each of these processes must be small compared with supply by diffusion and associated release from the solid-phase. According to the accepted ranges of mass flow [54], modelling indicates that supply by convection is usually negligible compared with diffusive supply, especially when release from the solidphase is also considered [55]. It is more difficult to model the other terms that are not mimicked by DGT. When good relationships between metal accumulated by DGT and plants are obtained, it is reasonable to suggest that these other terms do not contribute appreciably to metal supply to the plant. Their possible significant contribution to supply may be one of the reasons for such relationships breaking down.

In the first comparison of DGT measurements with metal uptake by plants (Lepidium sativum L.), DGT was deployed and plants were grown in the same soil, but at different moisture contents, representing 50-90% of the maximum water holding capacity [56]. The concentration of Cd, Co, Cu, Ni, Pb and Zn in the plant herbage and the flux of metals to the DGT device both increased systematically with moisture content, while the concentration of the metals in soil solution declined. As only the water content of the soil was varied, the concentration of metals in the soil solution can be expected to be proportional to the free ion activity. The results showed that neither the free ion activity nor the concentration of metal in the soil solution could predict metal uptake by the plants. The fact that DGT could predict plant uptake was attributed to supply by mass transport and associated release from the solid-phase being dominant and a similar dependence for both DGT and the plants on soil moisture content.

In the second comparison, an indicator species for Cu (Lepidium heterophyllum Banth.) was grown in 29 different soils with a range of Cu concentrations [20]. Copper was measured by four methods: DGT, in soil solution, as free ion activity and by EDTA extraction. The concept of effective concentration, CE, was introduced. Diffusional supply of a metal to a plant or DGT device is augmented by release from the solid-phase. Therefore, the effective concentration that the plants experience, CE, is larger than the concentration in solution. The DIFS model was used to convert the DGT measured concentration, CDGT, to the effective concentration, CE. Essentially, the flux for the diffusion-only case is compared to the DGT measured flux. A good correlation

0 0

Post a comment