Sediment heterogeneity arising from steep gradients in solute concentrations associated with the well-known sequence of redox reactions prevents accurate replication of measurements at high spatial resolution. Moreover, efforts to homogenize sediments are generally unsuccessful and new structure in solute distributions develop within hours as local redox gradients are re-established. Consequently, it has not been possible to apply and test the DIFS models on simple homogenised sediment systems. From the outset, the primary use of DGT in sediments has been to investigate the distribution of solutes at high spatial resolution [2].

W. Davison, H. Zhang and K.W. Warnken 16.5.1 Practicalities for deployments in sediments

DGT sediment probes have been designed to be pushed into the sediment, so the binding gel, diffusive gel and filter are held in a slim probe, typically 5 mm thick (Fig. 16.5). To ensure a tight seal between the frame and the front pre-filter, a second pre-filter membrane is usually placed behind the 0.4 mm binding gel and 0.8 mm diffusive gel. It also serves as a support for the binding gel when it is removed. The probes have a large window area to allow the binding layer to be sliced prior to analysis. Before they are deployed in anoxic sediment, probes must be deoxygenated by immersing them in a container of 0.01 M NaCl solution, which is moderately bubbled with oxygen-free nitrogen or argon gas for 24-48 h. The container should be either closed except for a small vent or housed in a glove bag. For low-level trace metal work, care must be taken at this stage to prevent contamination and probes that serve as blanks should follow the same procedure. Once removed from the oxygen-free solution, probes should be deployed in the sediment as quickly as possible (seconds), using a smooth insertion action to minimize sediment disturbance and avoid any cavitation, which could allow solution to flow along the probe face. Upon retrieval from the sediment, the surface of the probe is rinsed using a wash bottle of high-purity water, ensuring that there are no particles remaining on the window

Backing plate Resin gel layer

Backing plate Resin gel layer

Fig. 16.5. Schematic depiction of the layers of a DGT sediment probe.

area. The sediment-water interface can usually be seen at this stage as a stain on the filter and should be marked.

Using the sampling window as a guide, a Teflon-coated razor blade can be used to cut through the gel and the filter membrane layers, which can then be carefully removed from the device and placed on a clean flat surface. The top filter membrane and the diffusive gel can then be removed and discarded leaving only the binding gel layer and the supporting filter membrane. The binding gel can then be sliced to the desired resolution using a new, rinsed Teflon-coated razor blade. The smallest practical gel slices that have been used have either been 1mm wide by ~ 1-2 cm long [2,16] or 3 x 3 mm squares [11]. Even at this size, it is difficult to slice accurately, and there is a risk that the Chelex 100 resin beads are not distributed uniformly within the resin-gel layer. For trace metal analysis, each gel strip is placed into a micro-centrifuge tube (0.5 or 1.5 mL) and the gel is eluted using an appropriate volume of 1 M HNO3. The gel should be left in the dilute acid for at least 24 h, to ensure complete elution and diffusional mixing.

16.5.2 Analyte distributions from gel slicing

The first DGT measurements made in sediments using this gel-slicing approach showed steep vertical gradients of metals in the surface sediments of a productive lake [2]. The DGT measurements were reported both as localized fluxes to, and as concentrations (CDGT) at, the interface of the device. The pronounced maxima of metals, particularly Cu and Zn, immediately below the sediment-water interface, were attributed to release from rapidly oxidized organic material. While the various maxima were interpreted in terms of local supply processes, comparisons between interfacial concentrations measured by DGT and concentrations in the bulk pore-waters were made using a theory of supply from the solid-phase, which later formed the basis of the DIFS model [13]. Model calculations for the diffusion-only and well-buffered cases were performed in this early paper. From measurements made using different diffusive gel layer thicknesses, it was concluded that the supply of Zn was fully sustained (rapid) while that of Ni was only partially sustained (slow). These results agree well with conclusions drawn from more recent measurements made on soils, where the dependence of the DGT measurement on deployment time was interpreted using the DIFS model [19].

In view of the significant advances made by this first use of DGT in sediments, it is surprising how few subsequent studies have been carried out using this rather simplistic slicing approach. Zhang et al. [61] used the slicing approach to show pronounced localized features in DGT measurements of As, Co, Fe, Mn and Ni in a marine sediment, with good correspondence between Co and Mn, and Co and Ni. DGT measurements of Fe in this same sediment were interpreted in terms of production rates and used to infer a significant recycling of Fe within the sediment through interactions with sulphide [62,63]. The vertical distributions of Fe and Mn measured by DGT have also been compared with the presence of particular species of micro-organisms [64].

DGT devices were also deployed in Black Sea sediments using an autonomous benthic lander [65]. Distinct maxima in Co and Cd at 4 and 6 cm depth in the sediment coincided precisely with maxima in Mn, whereas Fe maxima were offset by several millimetres (Fig. 16.6). Fones et al. [16] also performed DGT measurements on North Atlantic sediments using both a lander and deployments in retrieved cores. Their data showed highly localised features superimposed on the usual redox-associated gradients. Coincident maxima of Cd, Cu, Ni and Zn were not directly linked to the redox-sensitive elements Co, Fe and Mn.

The difficulties of precise interpretation were considered for measurements of Cd, Fe, Mn and Pb in a lake sediment using diffusion layer thicknesses of 0.4 and 1.2mm [66]. Reproducible results for Cu and Fe measured close to the sediment-water interface have been reported [67]. When Roulier and Motte [68] compared DGT measurements of Cu with concentrations in the pore-waters, they deduced that there was a good supply of Cu from the solid-phase in one reservoir, but not in

0 0

Post a comment