Optimization of an emulsified vegetable oil (EVO) delivery system led to uniform distribution and ongoing biological treatment of a PCE source area last summer at the former Naval Training Center (NTC) in Orlando, FL. Remedial efforts targeted contaminated ground water at a former laundry and dry cleaning facility at NTC's Operable Unit 4 (OU4), where earlier injections of potassium permanganate (KMnO4) failed to reduce contaminant concentrations in the source area. Performance monitoring last fall indicated that the EVO system was operating successfully and aquifer conditions were favorable for enhanced reductive dechlorination (ERD), with total organic carbon (TOC) concentrations approaching 200 mg/L, in contrast to a pre-treatment average of approximately 10 mg/L.
Contamination of OU4 ground water was attributed to past spills and leaks in drain lines, floor drains, and other components of the dry cleaning facility's wastewater collection and conveyance system. Subsurface investigations detected PCE concentrations reaching 22,600 µg/L, which suggested the presence of dense non-aqueous phase liquid (DNAPL). From March through October 2003, a KMnO4 recirculation delivery system was implemented. Subsequent field studies indicated that the elevated natural oxidant demand of the aquifer material had caused formation of manganese dioxide solids, which in turn plugged the KMnO4 injection wells and limited oxidant distribution.
In 2006, an optimization study was conducted to evaluate viable remedial alternatives for addressing the remaining DNAPL source area. Enhanced in situ biodegradation using EVO was selected as a cost-effective strategy to treat source-zone contamination, reduce mass flux, and control plume migration over time.
Recirculation initially was selected as the delivery method most likely to uniformly distribute EVO in the 60- by 80-foot source area. The target treatment zone was defined by a PCE concentration of 2,000 µg/L, approximately 1% of the compound's aqueous solubility. Target depths extended through a shallow source area above a 5-foot cemented sand unit located 20-25 feet bgs. The delivery system comprised a central recovery well surrounded by six equally-spaced injection wells. Based on short-duration aquifer performance tests, an average extraction rate of 2.5 gpm was expected in the shallow zone (5-20 feet bgs) (Figure 2).
EVO application began in September 2007 with ground-water recovery from the single extraction well. Based on the design flow rate, the anticipated duration of recirculation (one pore volume) was 14 days. Performance monitoring indicated that the maximum sustained extraction rate of 1 gpm was well below the design flow rate of 2.5 gpm, and that two monitoring wells (MW1 and MW2) within the recirculation footprint had not been impacted by EVO recirculation. Inadequate distribution of the EVO was attributed to poor extraction from the central recovery well. As a result, the recirculation effort was abandoned after approximately 10% of the design volume was injected (approximately 8,200 gallons of a 1% EVO solution).
Following the initial EVO recirculation event, analysis of ground water in MW1 indicated PCE concentrations had increased from 1,800 to 3,500 µg/L, and daughter products TCE and cis-DCE also increased an order of magnitude. In MW2, PCE and TCE concentrations remained unchanged, but cis-DCE decreased an order of magnitude. Vinyl chloride was not detected in samples from either monitoring well, and no changes in TOC were observed. A significant decrease in redox potential, from -70 to -252 mV, was observed in MW1 after two months of recirculation. In addition, geochemical indicators suggested that sulfate and iron reduction were occurring, but methane concentrations were unchanged.
Evaluation of the initial attempt to recirculate EVO in the shallow zone led to several `lessons learned.` A longer aquifer test would have evaluated sustainable injection and extraction rates better, thereby optimizing design of the recirculation system array. Distances between the injection and extraction wells may have been too large to achieve adequate control of the injected EVO. In addition, injection of EVO at a higher concentration may have resulted in better substrate distribution in the subsurface.
Subsequent optimization activities involved: (1) bench-scale tests to determine the adsorptive oil capacity of the soil and optimal EVO dose; (2) an evaluation of EVO delivery options; and (3) a pilot-scale study to determine optimal injection spacing. Bench-scale tests were performed on two shallow (fine sand) and two deep (fine to medium sand) samples collected from outside the target treatment zone. Results indicated a median adsorptive capacity of 0.00384 (pounds of EVO per pound of soil), which was greater than three times the adsorptive capacity assumed during EVO recirculation.
Direct-push technology (DPT) points rather than permanent injection wells were selected for EVO delivery due to the high cost and extensive field resources associated with installing numerous permanent wells. In addition, deployment of temporary DPT points could focus EVO injection in a discrete vertical interval.
A March 2008 pilot study evaluated 5- and 7.5-foot radii of influence for injection point spacing. The 7.5-foot radius was selected for full-scale delivery due to an increase in TOC and positive results from a bromide tracer test in the monitoring wells. Full-scale injection of EVO, using 22 DPT points spaced 15 feet apart, was conducted in July 2008. A 4-foot screen was used to inject 66,888 gallons of a 2.1-6.3% EVO solution over an interval of 8-20 feet bgs.
Ground-water samples from MW1 and MW2 in November 2008 indicated that TOC concentrations had risen from 10 mg/L to 160-180 mg/L. In MW1, PCE concentrations decreased from 1,800 µg/L to 756 µg/L, while TCE increased from 64 µg/L to 914 µg/L, cis-DCE increased from 87 µg/L to 11,700 µg/L, and vinyl chloride increased from non-detect levels to 92.9 µg/L. Similarly, concentrations of PCE in MW2 decreased from 18,000 µg/L to 184 µg/L, while TCE decreased from 1,200 µg/L to 38 µg/L, cis-DCE increased from 1,500 µg/L to 9,210 µg/L, and vinyl chloride increased from non-detect concentrations to 2,800 µg/L. Ethene was detected in MW2 at concentrations as high as 19 µg/L.
Methanogenesis within the treatment zone was evidenced by an increase in methane concentrations of up to 7 mg/L. In addition, Dehalococcoides ethenogenes was detected within the treatment zone in populations reaching 7.61E+03 cells/mL. Overall results based on formation of TCE daughter products and the shift in aquifer geochemical conditions (from moderately reducing to highly reducing methanogenic conditions) indicate that EVO optimization has improved enhanced reductive dechlorination of OU4 ground water. Further optimization of the ERD system will be conducted based on future monitoring results.