Development of new, cost-effective technologies for heavy metal remediation is critically needed. Of the most frequently cited metals, arsenic, lead, cadmium, and mercury have been shown to precipitate as stable metal sulfides under highly reducing conditions. The lowered drinking water standard for arsenic (from 50 to 10 ppb) will add significantly to the costs of remediating this particular dissolved metal contaminant. The most common method of remediating soluble metals is pump-and-treat, which can take decades to remove contamination because of subsurface heterogeneities and adsorption of metal species to the aquifer solids. The long timeframe for pump-and-treat and the cost for off-site disposal of metals extracted from the groundwater make this a very expensive remedial option.

The objective of this project was to investigate an in situ strategy for precipitating arsenic and potentially other heavy metals as stable metal sulfides and to demonstrate long-term sequestration of these metals under changing geochemical conditions.

Technical Approach

Unlike organic contaminants, heavy metals cannot be destroyed. Existing groundwater and soil remediation technologies generally focus on in situ immobilization, in situ transformation to a less toxic form, or extraction and ex situ treatment. The appropriate application of any remediation technology for a heavy metal or radionuclide depends on a thorough understanding of the behavior of that metal, particularly its speciation and mobility in different environments. The fate of metals in the subsurface is controlled by many geochemical parameters, including pH, oxidation reduction (redox) potential, and salinity. These factors dictate the predominant forms (i.e., species) of metals in the subsurface and thus control the solubility and mobility of these metals.

In anaerobic environments, the mobility of heavy metals is often directly influenced by bacterial sulfate reduction, a process that is ubiquitous when sulfate and carbon substrates are available. Recent studies have shown that arsenic-sulfide precipitates can be formed under the appropriate conditions by in situ biostimulation using ethanol and/or lactate as electron donors; however, there has been very limited research on the mineralogy and the stability of these precipitates under variable geochemical conditions although it has been suggested that these precipitates are relatively stable with low pH and under oxidizing conditions.

A field research site was established at Site No. ST-65, a former refueling area at the Avon Park Air Force Range in Florida. ST-65 had high concentrations of arsenic in shallow groundwater within a sandy, naturally anaerobic aquifer. Sediment and groundwater were collected from the area with the highest concentration of arsenic in groundwater (1,800 ppb) and were used to construct four flow-through columns for a bench-scale laboratory study conducted at Princeton University. In these experiments, indigenous sulfate-reducing bacteria were stimulated with injections of sodium lactate, ethanol, ferrous iron, and sulfate over a period of several months under strictly anaerobic conditions.

A small-scale field pilot demonstration was later conducted at ST-65 in January 2008. The field pilot was a scaled up version of the column study performed in situ at ST-65 over a 30 ft by 30 ft target zone in the area of highest arsenic concentrations. Four monitoring wells, one injection well, and one recovery well were installed.


Analysis of the sediments from one column indicated the presence of arsenic-bearing sulfides, including arsenopyrite and realgar, at ten times the concentration of the natural sediments. Aerobic water was then passed through one of these columns for 115 days. Analyses of the dissolved arsenic concentrations in the columns over time when compared to the total arsenic in the analyzed sediments suggested that only 2% of the sequestered arsenic was dissolved after 116 pore volumes of aerobic groundwater had passed through the column. Analyses of the arsenic concentration of sediments in 15 mL column experiments leached by the same water over the same time period, which was equivalent to 11,600 pore volumes, yielded an arsenic loss rate that agreed within error with the results of the larger column experiment. A scale-up of the column study model using the measured groundwater flow velocity indicates that 98% of arsenic would be retained in the solid phase after 16 years of fully oxygenated conditions in the groundwater and that the dissolved arsenic concentration would average approximately 1 μM (75 ppb) at the source. The results of the modeling indicates that the key to successful sequestration of arsenic under anaerobic, sulfate-reducing conditions is the co-precipitation of iron sulfide, which consumes oxygen during oxidation, thus reducing the oxygen available to dissolve arsenopyrite, and produces Fe3+ hydroxides, which act to adsorb dissolved arsenic, and the co-precipitated iron and calcium phosphate, which may act to further shield arsenic sulfides from oxidation and may incorporate arsenic into their crystal structure. For many subsurface environments, the groundwater is naturally anaerobic and sulfidic, and for these environments, the sequestration of arsenic would be an effective and permanent solution.

At the field site, measurements of background geochemical parameters indicated that the groundwater was naturally reducing. A slug test and tracer test were performed to assess aquifer characteristics to aid in recirculation system design. Upon evaluation of these parameters, a recirculation system was constructed to induce groundwater flow and provide amendments to stimulate indigenous sulfate-reducing bacteria. Injections of sodium lactate, ferrous sulfate, diammonium phosphate, and ethanol began in April 2008. Subsequent measurements performed over the course of the study period indicated that a gradual shift towards more reducing conditions occurred after initiation of amendment injections. Adjustments to the amendment mixture resulted in further reduction of site geochemical parameters to within the sulfate-reducing range. The final amendment mixture consisted of sodium lactate, sodium sulfate, and diammonium phosphate.

Results from groundwater sampling and analyses indicated that arsenic concentration decreased by up to two orders of magnitude to approximately 0.01 μM (1.4 ppb) between March and September 2008. The concentrations were two orders of magnitude lower than that attained in the laboratory column experiments. The effectiveness of the amendment injection in altering the groundwater geochemistry is also clearly represented by a reduction in total iron, total sulfate, and total sulfide concentrations observed over the study period.

Between March and September 2008, total iron, total sulfate, and total sulfide concentrations decreased by up to 85%, 94%, and 91%, respectively, reflecting a significant decrease in reduction potential and subsequent precipitation of iron with sulfide. In addition, sodium concentrations increased significantly, indicating successful distribution of sodium lactate and sodium sulfate, two of the amendments that were injected at the site. Sequential extraction analyses of sediment pouches, which had been deployed in each monitoring well at the outset of the field study, during the course of the experiment confirmed increasing concentrations of solid-phase arsenic associated with higher iron, sulphur, and phosphorus concentrations. At the end of the experiment, the total arsenic concentration in the sediment was up to three times that of the original sediment. These results are in agreement with the outcome of the column study and demonstrate a proof-of-concept of this anaerobic biostimulation technology under field conditions.


Demonstration of a cost-effective, in situ, anaerobic biostimulation technology for the long-term sequestration of arsenic will aid DoD in complying with the lowered drinking water standards. In addition to arsenic, this technology may be applicable to a number of metals of concern.