Enhanced Reductive Dechlorination (ERD) with Emulsified Vegetable Oil (EVO) has been used at hundreds of Department of Defense (DoD) sites to remediate chlorinated solvents, chromate, uranium, perchlorate, and explosives. This process commonly involves injecting EVO, nutrients, pH buffer or base, and microbial cultures to adjust biogeochemical conditions in the immediate vicinity of the contaminant, so that:
- Sufficient levels of fermentable organic substrates are present to support microbial growth and contaminant biodegradation.
- The aquifer pH is appropriate for microbial growth and contaminant biodegradation.
- Critical microorganisms are present in sufficient numbers with the required genetic capability to degrade the pollutants.
The overall objective of this project was to identify the reason(s) why remediation systems using EVO failed to meet cleanup goals at some sites while the technology has been shown to be very successful at other sites. At some locations, there are obvious reasons for not meeting cleanup goals including injection system designs that do not follow published design guidelines. However, at other sites, there is no obvious reason for the remediation system not meeting cleanup goals. This project aimed to identify changes in site characterization, remediation system design, and/or field implementation that would have improved remediation system performance.
The two sites evaluated in this project were both located at the former Naval Training Center (NTC) Orlando: (1) Study Area 17 (SA17); and (2) Operable Unit 2 (OU2). At both sites, the remediation systems were initially successful, resulting in substantial reductions in trichloroethene (TCE) concentrations. However, concentrations of cis-1,2-dichloroethene (cDCE) and vinyl chloride (VC) increased in some wells due to TCE degradation and remain elevated. Results from the project evaluations were used to: (a) identify the reason(s) why the ERD systems failed to meet cleanup goals; and (b) develop new approaches and/or procedures to improve performance.
SA17 ASSESSMENT RESULTS
At SA17, two different depth intervals of a TCE source area were treated with EVO to stimulate ERD; Zone B extending from 15 to 30 ft bgs and Zone C extending from 30 to 50 ft bgs. Bioremediation performance in Zone B at SA17 source area was good with 2.8 to 4.6 Order of Magnitude (OoM) reductions in TCE. While cDCE and VC removal were lower, the sum of organic chlorine (ΣCl) declined by 0.8 to 3 OoM indicating a substantial portion of the parent compound was reduced to non-toxic end-products. TCE removal was also good in Zone C source area at SA17. However, higher levels of cDCE and VC accumulated with ΣCl declining by only 0.5 to 1.5 OoM. EVO distribution in both Zones B and C at SA17 was limited by: (a) injection of too little EVO; and (b) development of stagnation zones during injection. cDCE and VC removal in Zone C was inhibited by the low pH due to injection of too little base to neutralize acidity produced during ERD and the background acidity of the aquifer. While Dhc populations were low at many locations, substantial populations of Dhc capable of growing on VC developed at locations with sufficient substrate and appropriate pH, indicating ERD was not limited by absence of required microorganisms. There was no evidence of significant lower permeability zones near the target treatment zone that would result in substantial back diffusion of contaminants, limiting treatment. In summary, the primary factors limiting bioremediation performance at SA17 were inadequate levels of fermentable substrate and low pH due to injection of too little substrate, too little base to increase pH, and limited distribution of these materials throughout the target treatment zone.
OU2 ASSESSMENT RESULTS
Bioremediation was less effective in reducing chlorinated solvent concentrations downgradient of the EVO Permeable Reactive Barrier (PRB) at OU2. TCE concentrations in individual monitoring wells declined by 0 to 3.2 OoM at OU2 (median reduction of 0.5 OoM) with production of large amounts of cDCE. ΣCl removal at OU2 varied from 0.1 to 0.7 OoM with a median reduction of 0.2 OoM, which is lower than reported for other ERD projects. Effective distribution of EVO at OU2 was limited by: (a) injection of too little EVO; and (b) the presence of high TCE concentrations with and/or immediately adjoining lower permeability zones. Conversion of cDCE to ethene was inhibited by the low pH due to injection of too little base to neutralize acidity produced during ERD and the background acidity of the aquifer. While Dhc populations were low at many locations, substantial populations of Dhc capable of growing on VC developed at locations with sufficient substrate and appropriate pH, indicating ERD was not limited by absence of required microorganisms. While back diffusion of contaminants out of the underlying low permeability unit does occur downgradient of the OU2 PRB, the short travel distance from the PRB to the discharge point would greatly limit the impact of this process. In summary, the primary factors limiting bioremediation performance at OU2 were inadequate levels of fermentable substrate and low pH. The low substrate concentrations and low pH were due to injection of too little substrate, too little base to increase pH, and challenges in distributing these materials within and adjoining lower permeability units.
- Parent compound (TCE) removal was relatively good in the SA17 source area in both Zones B and C, even though the amount of EVO and base injected was much less than that required for optimum treatment. This indicates that ERD with EVO is a fairly robust technology and good parent compound removal can be achieved with a less-than-perfect design.
- For the most effective treatment, amendments need to be distributed throughout the entire target treatment zone. If the treatment amendments are not uniformly distributed, CVOCs can persist in untreated zones, increasing the treatment duration.
- Generating strongly reducing conditions with methane production is a poor indicator of effective substrate distribution. Once produced in an EVO treated zone, methane is relatively unreactive and can be transported away from the residual oil.
- Common rules of thumb used for designing EVO injections can greatly under-estimate the actual oil requirement. Oil retention tests should be run to generate more accurate estimates of the actual amount of EVO required for effective treatment.
- Addition of alkaline materials to increase aquifer pH can significantly increase oil retention, reducing contact efficiency. If possible, the alkaline materials should be added after the EVO is injected to reduce these impacts. When groundwater ionic strength is high due to background geochemistry and/or amendment addition, oil retention tests should be run with solutions representative of the groundwater geochemistry.
- The common practice of simultaneously injecting all wells to reduce injection time can result in stagnation zones, leaving some areas untreated. Injecting every other well in a group, then injecting the remaining wells in a second group can improve amendment distribution.
- While not directly addressed in this project, recirculation systems can be used to more effectively distribute substrate, pH buffers, microorganisms, and the target contaminants, improving treatment.
- When the aquifer pH is less than 6.3, site characterization should include measurement of inorganic carbon, mineral acidity, and aquifer buffering capacity (pHBC). With this information, designers can generate reasonable estimates of the amount of alkaline material require to maintain pH within a suitable range.
- In some cases, the amount of base required to maintain an appropriate pH can equal or exceed the amount of organic substrate required.
- Dechlorinator populations including Dhc and bvcA/vcrA can increase and decrease with time due to temporal variations in amount of organic substrate and/or contaminant concentrations. Low dechlorinators numbers do not necessarily indicate absence of required organisms, but can result from unfavorable geochemical conditions.
- Remedial performance at SA17 and OU2 has improved over time as the treatment system has been modified based on our improved understanding of site conditions and in situ bioremediation processes. Site managers should recognize that it may not be practical to remediate a contaminated site with a single EVO injection. It may be more efficient and effective to employ an iterative process where a lower cost remedial system is installed, followed by monitoring and site characterization to identify treatment issues, and then the system is modified to improve performance.
ESTIMATING BASE REQUIREMENT FOR AQUIFER pH CONTROL
At both sites evaluated in this project, low pH inhibited ERD of TCE to non-toxic end-products with accumulation of cDCE and VC. The low pH was due to: (a) low background pH of the aquifer; (b) acidity produced during ERD; and (3) injection of too little base to raise the pH to appropriate levels. To aid in the design of ERD projects at other sites, an MS Excel based design tool is presented to provide preliminary estimates of the amount of base required to maintain a neutral pH during ERD. The design tool approach and calculations were presented in Appendix D. Required input for the design tool include: (1) treatment zone dimensions and design life; (2) site characteristics including K, porosity, hydraulic gradient, contaminant concentrations and electron acceptors produced or consumed during ERD; (3) background pH, total inorganic carbon, mineral acidity, and pH buffering capacity (pHBC); (5) mass of organic substrate and base; and (6) target pH. The design tool calculates the amount of base required to: a) raise the pH of the aquifer material and influent groundwater, and b) neutralize acidity produced during reductive dechlorination and substrate fermentation.
CONCEPTUAL MODEL OF ERD TREATMENT WITH EVO
Conclusions and Lessons Learned in this project were integrated with prior laboratory and field studies to generate a general conceptual model of ERD with EVO and pH buffer. This conceptual model provides a relatively consise summary of the current understanding of ERD with EVO including: (1) ERD microbiology and organohalide respiration; (2) environmental requirements for efficient dechlorination; (3) EVO properties, transport and retention in the subsurface; (4) EVO consumption during ERD; (5) aquifer pH and buffering; and (6) injection system design.