Groundwater contamination by perchlorate is recognized as a significant environmental issue in the United States and abroad. Current remediation methods for perchlorate-contaminated groundwater generally involve extracting the water and treating it ex situ using either selective ion exchange resins to adsorb the dissolved perchlorate or biological reactor systems to destroy it. In situ remediation of perchlorate has the potential for both cost and safety benefits compared to current ex situ approaches. Extensive laboratory and field studies conducted during the past decade have revealed that perchlorate-reducing bacteria (PRB) are indigenous to most groundwater aquifers and that these bacteria can be stimulated to degrade perchlorate through the addition of a variety of organic electron donors, including various fatty acids, alcohols, sugars, and natural oils. The PRB oxidize the electron donor and subsequently reduce perchlorate to chloride and water, two innocuous products. The main challenge for implementing in situ perchlorate bioremediation is effectively mixing an electron donor into the perchlorate-contaminated groundwater and delivering the mixture to the indigenous PRB, without having to extract water from the subsurface. Other challenges include preventing microbial biofouling of pumping wells and minimizing the mobilization of secondary groundwater contaminants, such as manganese (Mn) and iron (Fe).
An innovative in situ bioremediation technology, known as a horizontal flow treatment well (HFTW) system, was evaluated during this demonstration for delivering electron donor and promoting the biological reduction of perchlorate. The HFTW technology consists of two dual-screened treatment wells, one pumping contaminated groundwater from a deep aquifer region and injecting it into a shallower zone and the other pumping contaminated groundwater from the shallower aquifer region and injecting it into the deeper zone. The two wells work in tandem to establish a groundwater recirculation zone in the subsurface. The electron donor is added and mixed with contaminated groundwater at each well, creating an anaerobic, bioactive zone between and downgradient of the HFTWs during system operation. Contaminated water is never brought to the surface, as treatment occurs in the in situ bioactive zones.
The objectives of this project were to demonstrate the following: (1) in situ biological perchlorate treatment is feasible in the field using electron donor addition; (2) perchlorate can be treated to < 4 mg/L; (3) perchlorate can be treated in a drinking water aquifer without mobilizing significant quantities of Fe and Mn or bringing oxidation-reduction potentials (ORP) to very low levels; (4) the zone of influence and efficiency of the HFTW system are sufficient to make the technology a viable, cost-effective option at many sites; (4) biofouling can be effectively controlled by measures that are easily implemented; and (5) co-contaminants, including nitrate and trichloroethene (TCE), can be treated using the same HFTW technology. As with any pilot-scale technology demonstration, a main objective of this field project was to collect and document information that is relevant to site managers and regulators who are responsible for choosing and implementing technologies.
An HFTW system was installed at Aerojet General Corporation’s 8,500-acre site in Rancho Cordova, California (Aerojet). The pair of HFTWs were installed approximately 34 ft apart and screened within a shallow zone in the aquifer from 46-61 ft below land surface (bls) (upper screen) and within a deeper zone at 80-100 ft bls (lower screen). A group of 19 monitoring wells screened within the shallow and deep zones and placed at various locations upgradient and downgradient of the HFTW pumping wells were used to evaluate overall system performance. The wells were each operated at a net flow rate of 6 gpm, citric acid was used as the electron donor, and chlorine dioxide was periodically added to each of the HFTWs as a biofouling control agent.
The demonstration was conducted in three phases. The objectives of Phase I were to evaluate the overall groundwater mixing and capture by the system and to determine the extent of perchlorate and nitrate reduction possible without mobilizing significant quantities of Fe and Mn. Between the final background monitoring event (Day -15) and the final groundwater event in Phase I on Day 275, perchlorate concentrations in the seven shallow monitoring wells declined by an average of 95% from the starting average of 2230 mg/L to 90 mg/L. One of the downgradient wells reached < 5 mg/L on Day 67, but most of the other wells showed stable perchlorate concentrations ranging from approximately 40-160 mg/L. These concentrations remained reasonably consistent with electron donor dosages up to 2.5 times the calculated stoichiometry.
The consistent decline in perchlorate concentration throughout the entire shallow aquifer zone during Phase I showed that the HFTW system provided good mixing and electron donor delivery within this region. This observation was consistent with conservative tracer tests conducted during background testing. Moreover, a rapid and consistent reduction in perchlorate concentrations observed in a side-gradient monitoring well showed that the region of influence of the HFTW system in the shallow zone met or exceeded initial predictions derived from a site-specific groundwater transport model. The low residual concentrations of perchlorate throughout this region during Phase I operation may reflect a limitation in electron donor in this region (the donor was intentionally limited to prevent mobilization of Fe and Mn) or may be a function of the mixing design and flow field of the HFTW system combined with aquifer heterogeneity. In later testing (Phase III), low concentrations of residual perchlorate were detected in several downgradient wells even in the presence of excess electron donor suggesting that the latter hypothesis is more likely.
Like the shallow downgradient wells, the perchlorate concentrations in the deep downgradient monitoring wells at the site also declined significantly during Phase I operation, although the extent and consistency of the reduction was less than for the shallow wells. In the nine deep downgradient wells within the treatment zone, perchlorate concentrations declined by an average of 60% from a starting concentration of 3722 mg/L on Day 0 to 1780 mg/L on Day 275. However, in the five deep wells furthest downgradient, which are beyond the immediate influence of the upgradient water entering the system through the HFTWs, average perchlorate reductions exceeding 93% were achieved by Day 146. Thus, with increased residence time, perchlorate reduction in the deep region of the aquifer was much greater than for the wells close to the HFTWs.
One of the key variables in Phase I was to determine if perchlorate could be degraded without significant mobilization of Fe and Mn. This was accomplished by tightly controlling the addition of citric acid, based on expected concentrations of oxygen, nitrate, and perchlorate. Mobilization of both Fe and Mn was minimal during the course of Phase I operation. With the exception of two shallow wells closest to the upflow-HFTW, soluble Fe concentrations throughout the plot remained well below 500 mg/L. Moreover, Fe that was dissolved and mobilized during the active phase of operation rapidly re-precipitated when the system was shut down. Dissolved Mn concentrations also generally remained low during Phase I. Concentrations reached a maximum of 1470 mg/L in one well but rapidly declined back to < 50 mg/L after electron donor addition ceased at the end of Phase I. During the final sampling event in Phase I in which Mn was measured, concentrations of the metal were below 50 mg/L in 12 of the downgradient monitoring wells.
The key objective of Phase II was to treat perchlorate without promoting significant well biofouling, which was an operational issue in Phase I. The two HFTWs were redeveloped between Phase I and Phase II. The electron donor dosing regimen was switched from a daily addition (as in Phase I) to larger weekly or twice-per-week doses in order to evaluate the impact of dosing schedule on both perchlorate treatment and well fouling. In addition, chlorine dioxide was added to each well on a daily basis (4-8 times per day) as a preventative measure. The wells were operated continuously at 6 gpm during Phase II. The objective of Phase III was to assess an “active-passive” mode of operation. In this case, the HFTWs were used primarily for mixing electron donor with the perchlorate-contaminated groundwater. The pumping wells were then turned off between mixing periods. The key objective was to determine whether this mode of system operation would result in a consistent reduction in perchlorate concentrations and reduced system operation and maintenance (O&M) costs. During Phase III, the HFTW treatment wells were operated in a 15-day cycle consisting of 3 days of active pumping followed by 12 days in passive (non-pumping) mode. Citric acid was added to both HFTWs in three 12-h pulses during the active period, and each HFTW was operated at a net flow rate of 6 gpm. The 15-day cycle was repeated six times during the 3-month test period, and three sampling events were performed.
A total of nine groundwater sampling events were performed during Phase II and Phase III operation. These sampling events included one background event prior to each phase, four events to measure system performance in Phase II, and three events to measure system performance in Phase III. As was observed in Phase I, perchlorate concentrations in all of the downgradient shallow wells declined rapidly during Phase II, but they did not generally go below detection, ranging from approximately 30-110 mg/L despite increasing the electron donor addition rate to approximately four times the stoichiometric requirement in the HFTW-U through most of Phase II. Perchlorate concentrations generally remained low in the shallow wells during the Phase III “active-passive” testing. Concentrations in several wells near the HFTW pumping wells were lower during Phase III than in either Phase I or Phase II testing, likely reflecting an increased residence time of water in the bioactive zone while the HFTWs were not pumping. In addition, with the system shut down during “passive” treatment, upgradient water (containing oxygen and nitrate as well as perchlorate) was not continually circulated throughout the plot. The increased reaction time and absence of new electron acceptor demand (particularly from oxygen and nitrate) probably resulted in the significantly lower perchlorate concentrations in this region during Phase III.
The consistent decline in perchlorate throughout the entire shallow aquifer during Phase II confirmed that, even with much more periodic dosing of electron donor (i.e., from daily dosing during Phase I to one or two times per week during Phase II), the HFTW system operated well as a treatment technology in the shallow zone. Moreover, the data from Phase III suggest that perchlorate treatment can be achieved by using the HFTW system intermittently as a vehicle to mix electron donor with the contaminated groundwater. Even in the side-gradient well, perchlorate concentrations remained < 100 mg/L throughout Phase III, even though the system was not pumped continuously. This suggests that the capture zone of the system during active pumping was maintained during the “active-passive” phase. The ability to operate this system several days per month rather than continuously could appreciably reduce the O&M costs associated with biofouling and well redevelopment, which is the most significant issue with this design.
The perchlorate concentrations in the shallow zone on Day 801 (the final sampling event in Phase III) represent a 96 + 4% reduction in dissolved perchlorate from the starting concentration in each well prior to Phase I (Day -7) and an average 94 + 3% reduction from perchlorate concentrations prior to Phase II (Day 472). Thus, perchlorate treatment in the shallow zone was very effective. However, with the exception of one well, perchlorate concentrations of < 4 mg/L were not generally achieved in the shallow zone during Phase II and Phase III. Rather, perchlorate stabilized between approximately 30-100 mg/L in most wells. Interestingly, a low residual concentration of contaminant was also observed during previous testing of a HFTW system for cometabolic treatment of TCE. The low residual contaminant was attributed primarily to competitive interactions between toluene (the cosubstrate) and TCE during biodegradation by toluene-oxidizing strains. However, the occurrence of low residual contaminant concentrations in both demonstrations suggests that this may be characteristic of the HFTW system.
The perchlorate concentrations in the deep downgradient monitoring wells showed a less consistent pattern of decrease during Phase II and Phase III than did the shallow wells during the same interval. However, the overall percentage reduction in the deep zone on Day 801 was 80 + 39% from the starting perchlorate concentration in each well prior to Phase I (Day -7) and an average 52 + 29% reduction from perchlorate concentrations at the end of Phase I (Day 275). If one only considers the six deep wells furthest downgradient from the HFTWs, the total perchlorate reduction during the 801-day demonstration was 88 + 9%. Thus, although non-detect concentrations of perchlorate were only achieved in a few wells, reasonable perchlorate treatment occurred in the deep zone, particularly considering results from the far downgradient wells.
The treatment of TCE by the HFTW system was also evaluated during Phase II and Phase III. The electron donor concentration was increased significantly and a commercial culture containing Dehalococcoides spp. (Shaw culture SDC-9) was injected into the HFTWs during Phase II to enhance reductive dechlorination. TCE concentrations in many of the shallow wells declined significantly during Phase II and Phase III. There was a 76 + 23% reduction in total TCE in all of the shallow wells from the beginning of Phase II (Day 472) to the end of Phase III (Day 801). If only the downgradient wells are considered, then the percent loss was 87 + 14%, with average final concentrations being 323 mg/L. cis-1,2-DCE (the initial reductive degradation product of TCE) was detected at high concentrations (>1,000 mg/L) in three of the shallow wells, while vinyl chloride (VC) was only detected during the last sampling event (Day 801) in one well. The relatively rapid and significant decline in TCE during the months after bioaugmentation with Dehalococcoides spp. in many of the shallow wells suggests that the procedure enhanced the dechlorination kinetics. The TCE concentrations in a number of the deep downgradient monitoring wells also declined significantly from the beginning of Phase II to the end of Phase III. Most notably, the TCE concentration in the far downgradient wells declined by as much as 98% from the start of the demonstration. However, as with perchlorate, the average decline in TCE concentrations in all of the deep monitoring wells was appreciably less than in the shallow wells, averaging 71 + 23% in the four wells furthest downgradient from the beginning to the end of Phase III.
The operational data from Phase III suggest that an “active-passive” approach may be the best overall operational strategy for an HFTW system in terms of both contaminant treatment and reduced O&M costs. Pressure increases also occurred in the HFTWs during Phase III, but with the short-term operation and large doses of citric acid (which assists in biofouling control through both acidification of local groundwater and chelation of precipitated metals), these increases did not affect operation during “active” phases. In addition, large additions of chlorine dioxide or other biofouling agents can be applied to wells during the passive phases to assist with long-term biofouling control. Thus, given that the treatment of perchlorate, as well as TCE, during this phase was equivalent to or better than that observed during the continuous-pumping phases, while biofouling was more readily controlled, “active-passive” operation appears to be the most desirable operational approach for this type of in situ design.