In-situ geochemistry of arsenite remediation by nanozerovalent iron in an anoxic aquifer: field versus lab

Arsenic (As) contamination is widespread in soils throughout the world, but remediation options can be limited, especially in anoxic conditions where trivalent arsenite (As(III)) dominates As speciation in the pore solution. As(III) adsorbs weakly to most sorbents and thus spreads readily in anoxic aquifers. Several laboratory assays indicate that in suspensions of zerovalent iron (ZVI), dissolved As(III) can be immobilised in several ways including reduction to insoluble zerovalent As (As(0)) or physical encapsulation in neoformed iron oxides. We previously exposed soil suspensions from an anoxic aquifer, in which As(III) predominates in solution, to ZVI suspensions under anoxic conditions, and quantified As immobilisation kinetics as well as the molecular speciation of Fe and As using X-ray absorption spectroscopy (XAS). As(III) was immobilised by oxidation to pentavalent arsenate (As(V)) which sorbed strongly to these oxides.

The current study investigated the immobilisation mechanisms of As(III) following in-situ injection of polyacrylic acid-coated nano-sized ZVI in the same anoxic aquifer. As concentrations measured at different distances from ZVI injection showed a heterogeneous response where a retarded As reduction, i.e. only after a year of monitoring, was observed in only a selected few groundwater wells. Eh and pH were subsequently measured, 2 years after ZVI injection, as a function of depth and distance away from ZVI injection using dynamic groundwater sampling, and As speciation was determined in extracted porewaters. Moreover, intact cores were sampled using sonic drilling, and the molecular speciation of As and Fe were characterised as a function of depth using XAS.

Lower Eh and higher pH values were found in the field compared to the earlier lab studies on soil suspensions from the same aquifer. As(III) was also not oxidised to As(V). Instead, both XANES and EXAFS analysis suggested that zerovalent As was formed, albeit not homogeneously throughout the treated aquifer. Unexpectedly, ZVI still occurred in zerovalent form to a large part despite the high surface area of nano-sized ZVI. This study thus illustrates the difficulty of reproducing field conditions during laboratory experiments and the sensitivity of immobilisation to geochemical conditions in the field. Moreover, immobilisation as zerovalent As did not sufficiently reduce As(III) concentrations, which remained the predominant As species in most porewaters, most likely because of the observed heterogeneity of the ZVI distribution in the aquifer.