Arsenic (As) is one of the most common contaminants of concern exceeding risk criteria because soil ingestion is the primary human health risk driver at many DoD sites. Use of contaminant total content instead of bioavailability is often overly conservative and can result in costly and unnecessary soil remedial action. Bioavailability-based in situ remediation of lead (Pb)-contaminated soil by using inexpensive and widely available phosphorus soil amendments is a proven technology. This technology has been used to remediate soil on firing ranges of DoD sites by reducing Pb bioavailability and exposure. Concern over the long-term effect of key biological and chemical processes on bioavailability of sequestered Pb is a barrier for implementation of this successful technology.

The specific objectives of this project were to (1) conduct a comprehensive study to link the binding mechanism of As in soil (i.e., speciation) with in vitro and in vivo methods used to predict current and potential future bioavailability of soil As to humans, and (2) evaluate the effect of key biological and chemical processes on the permanence of binding and bioavailability of Pb in untreated and treated (i.e., remediated) soils.

Twenty-seven As contaminated soils that represent a wide variety of properties and As sources from DoD installations, industrial sites (mining, smelting, and glass production), residential and agricultural sites were studied. Contaminant binding mechanisms were determined via speciation using As and iron (Fe) x-ray absorption spectroscopy (XAS) extended x-ray absorption fine structure spectroscopy (EXAFS) and Mössbauer spectroscopy by U.S. EPA National Risk Management Research Laboratory (NRMRL). Relative bioavailability (RBA) of As was determined by the adult mouse bioassay by U.S. EPA NHEERL/NERL and the juvenile swine bioassay by the University of Missouri. In vitro bioaccessible (IVBA) As was determined by the Ohio State University (OSU), unified BARGE method (UBM), physiologically based extraction test (PBET), California bioaccessibility method (CAB), and U.S. EPA Method 9200 IVBA methods by OSU. In vivo in vitro correlations (IVIVC) were used to evaluate the ability of IVBA methods to predict RBA.

Key biological and chemical processes were evaluated for soils from two contaminated sites which were remediated by phosphorus (P) soil amendment; a firing range at the Oak Ridge National Laboratory (ORNL) and a Pb-smelter impacted site in Joplin, MO. Soils were incubated with fungus that has the potential to increase Pb bioavailability and mobility. Soils were subjected to soil chemical acidification.  Fungal treatments were evaluated pre- and post-incubation by determining (1) IVBA Pb by OSU IVG and U.S. EPA 1340 pH 2.5, (2) RBA Pb from mouse bioassays, (3) Pb speciation by XAS EXAFS, and (4) mobility using synthetic precipitation leaching procedure (SPLP) U.S. EPA Method 1312. The effect of soil chemical acidification was determined by measuring dissolved Pb from liquid solid partitioning extraction across pH 1 – 13. 

Total soil As ranged from 162 to 12,500 mg/kg with a median value of 464 mg/kg. The most abundant As species was As(V) adsorbed to Fe oxides. Other As species included As (-III) metal sulfides, As (III) sulfides, As (III) oxides, As (III) adsorbed to Fe/ Al oxides, lead arsenates, and hydrous ferric arsenates (HFA). RBA As ranged from 6.37 to 81.2%. IVBA As ranged from <1% to 100%. Median and mean IVBA As followed the trend CAB (pH 1.5) > UBM (pH 1.2) ≈ OSU (pH 1.8) ≈ PBET (pH 1.8) ≈ Glycine (pH 1.5). IVIVC analysis showed all of the IVBA methods were predictive of RBA for both the mice and swine bioassays. Linear regression coefficient of determination (r2) values for IVIVC using mouse RBA were OSU IVG (0.89) > UBM (0.84), PBET= Glycine (0.82) > CAB (0.74) and using the swine method were OSU IVG (0.73) > UBM (0.67), PBET (0.63) > Glycine (0.60) > CAB (0.54). Swine RBA As was greater than mouse RBA As. However, variability in RBA As was larger for swine than mouse. CAB was more accurate for low RBA As soils and for soils with high reactive Al and Fe oxides. Despite As(V) adsorbed to mineral surfaces being a major component of most soils (>50%), these soils ranged from ~20-80% in IVBA As and widely ranged in RBA As. Arsenic speciation alone was not predictive of IVBA or RBA As. However, As speciation was very important to provide information on IVBA or RBA As results and/or determine a priori if a bioavailability-based risk assessment is justified.

Fungi isolated from soil produced low molecular weight organic acids (LMWOA) and dissolved PbCO3 in laboratory medium. However, the LMWOA-isolated fungus inoculated into study soils (ORNL, Joplin) had no effect on Pb solubility, mobility, and bioavailability. Soil treatment with P produced Pb minerals, including Pb pyromorphite, that were much more stable to LMWOA than PbCO3. Soluble Pb sharply increased at pH < 4 for the untreated control soil and pH < 3 for the P-treated soil. The P-treatment extended the insolubility of Pb from pH 4 to pH 3. 

Results from this research project provided site managers and risk assessors the following defensible science to implement bioavailability. All of the IVBA methods can be used to predict RBA As. The CAB was accurate for low RBA As soils and for soils with high reactive Al and Fe oxides. Non-CAB methods underpredicted RBA As for several soils with high concentrations of reactive Al and Fe oxides. IVBA methods using gastric extraction provide a more conservative RBA As.  Predictive equations can be used to predict RBA As for all IVBA gastric extraction methods. Arsenic speciation alone is not predictive of IVBA or RBA As. However, As speciation is very important to provide useful information to decide whether to consider adjusting for bioavailability in a risk assessment. Arsenic species with low RBA As are excellent candidates for exposure adjustment using in vitro extraction methods (i.e., IVBA) or in vivo data from animal models (i.e., RBA). 

Results showed sequestered Pb in P-treated soil was stable. Neither biological fungal treatments or chemical acidification (pH > 3) affected Pb mobility, bioaccessibility or bioavailability in P-treated soils. Low bioavailability mineral forms of Pb formed from P treatment in Joplin contaminated soil has remained stable over a long period of time (>10 y). The Pb minerals in the P treated soil are stable to acid inputs (i.e., natural or fertilizer, acid rain, etc.) to very acidic soil pH ≥ 3. Pb contaminated soils with large amounts of sulfidic waste (i.e., mining waste) should be treated with adequate alkaline soil amendments (i.e., agricultural limestone) to neutralize excessive amounts of future acidity and maintain soil pH > 4. 

A major finding of this study was that U.S. EPA method 1340 is not suitable for evaluating the bioaccessibility and bioavailability of P treated soils. A modified Method 1340, extracting solution of pH 2.5 instead of 1.5, or another method (i.e., OSU IVG) should be used to evaluate IVBA and RBA Pb in soil treated with P. These results were consistent with findings reviewed by Henry et al. (2015). Limited studies have reported U.S. EPA method does not accurately measure reduction in IVBA Pb or RBA Pb in P-treated soil and recommend changing the extraction pH from 1.5 to 2.5 in U.S. EPA Method 1340. The ability of U.S.EPA Method 1340 with an extraction pH of 2.5 to predict RBA Pb in soils treated with phosphorus and other soil amendments known to reduce RBA Pb should be researched. Research should be conducted with procedures outlined by the U.S. EPA Technical Review Workgroup and other committees focused on acceptable methodologies for regulatory acceptance including the Interstate Technology and Regulatory Council (ITRC) “Bioavailability in Contaminated Soil” workgroup. (Project Completion - 2017)