In Metal-Aid, layered double hydroxides (LDH) were modified and new reactants designed to enhance their removal capacity towards heavy metals and chlorinated solvents from waters. For example, we modified the LDH interlayer with various anions (incl. organic ions of differing lengths with sulphate, sulfonate or carboxyl functional groups; permanganate) which substantially enhanced sorption affinity and oxidative capacity towards chlorinated solvents. We further showed that green rust (GR), a redox active Fe(II), Fe(III) LDH, cannot degrade chlorinated ethenes, but that it stoichiometrically reduces Cr(VI) to Cr(III) within a few minutes, and this goes even faster if the GR hydroxide sheets are modified with aluminium and/or magnesium ions. Moreover, we showcased that GR is a strong sorbent for As and Ni, but affects As redox state only under specific GR formation conditions. In addition to LDH and GR reactants, we developed a new sulfidised Fe reactant (S-nZVI) that quickly degrades various chlorinated solvents. Moreover, by varying S-nZVI synthesis we demonstrated that S-nZVI nanoparticle structure and stability changes, affecting its selectivity towards the different chlorinated solvents. Next, our developed reactants were tested under relevant groundwater conditions, including reactant transport behaviour, to assess their injectability for in-situ field application. We discovered GR particles exhibit poor mobility in packed fine sand, independent of the applied injection flow rate and GR loading. This demonstrated that although GR reactants are highly reactive towards As and Cr, they are less suitable for in-situ injection. Instead, they could be used in remediation applications such as permeable reactive barriers or filter systems. In contrast, our S-nZVI reactant showed excellent transport properties when amended with a polymer, but mobility was clearly dependent on the polymer concentration, S-nZVI loading, injection flow rate and sand grain size. To determine the interdependencies between injection conditions and S-nZVI mobility, transport data was modelled and the derived parameters used to predict mobility at the next larger scale. In a third step, we assessed the long-term fate of our LDH/GR and S-nZVI reactants under contaminated groundwater conditions. These investigations confirmed that S-nZVI are relatively stable compounds, with little decrease in reactivity observed over 4 months exposure but this depended on S-nZVI structure, i.e. synthesis. Similarly, we showed that As-loaded GRs are stable under groundwater conditions for over a year, and LDHs can be stable for several months, depending on the type of interlayer anions. Furthermore, in GR-Cr interaction studies, we demonstrated that the reaction products (i.e. immobilised Cr) are more stable if Al and/or Mg modified GR rather than pure GRs are used. Lastly, for comparison to the performance of our synthetic reactants, we tested the extent of heavy metal release from iron oxides in Icelandic peat soils. These showed that geochemical conditions have to change dramatically (such as pH) to see the release of heavy metals into the water phase. Overall, our longevity studies were overwhelming positive reaffirming the suitability of our developed reactants for heavy metal and chlorinated solvent remediation. As a last step, we worked at three contaminant sites made available by Metal-Aid partners, and obtained valuable insights into groundwater and soil geochemistry, contaminant dynamics, natural attenuation processes and importantly also into the stability, toxicity and longevity of our refined reactants. A full-scale field injection was not feasible within the frame of the Metal Aid project, however, we performed a tank scale injection test ( ~1 m3). This tank test was carried out with almost all ESRs present and assisting, where 40 L of S-nZVI suspension were injected through a central well, the breakthrough at various distances recorded and the final S-nZVI distribution in the sediment determined. Overall, we observed similar S-nZVI distribution and mobility in this tank experiment as in the sand columns, indicating that column experiments are a useful instrument for injection planning. Furthermore, the results demonstrated that a well injection is a potential S-nZVI delivery method. In addition, we developed a new direct-push tool for S-nZVI particle tracking based on imaging with reactive and inert fluorophores, and successfully showcased it in the tank experiments.
