Installation Date:
08/01/1997

Contaminants:
Uranium

Reactive Media:
Fe⁰, AFO, PO₄

Construction:
Funnel and Gate

Point of Contact:
David N. Naftz , Ph.D.
U.S. Geological Survey
Tel: 801-975-3389
Fax: 801-975-3424
Email: dlnaftz@
usgs.gov
1745 W. 1700 South
Salt Lake city , UT 84104

A field-scale demonstration of a permeable reactive barrier (PRB) system is underway at an abandoned uranium upgrader site in Fry Canyon, UT. The U.S. Environmental Protection Agency (EPA) is the lead agency on the site. The ultimate goal of the demonstrations is to determine the technological and economic feasibility of using permeable chemical or biological obstacles, placed in the flow path, for removing dissolved metals and radionuclides from contaminated ground water. This project is testing the performance of three permeable reactive barriers at the Fry Canyon site. Anticipated results of the research for each of the PRBs tested will include long-term removal efficiencies for uranium and an evaluation of the commercialization potential for each. Specific objectives of the field demonstration project include: (1) hydrologic and geochemical characterization of the site prior to emplacement of barriers; (2) design, installation, and operation of three PRBs; and (3) evaluation of barrier(s) performance and commercialization potential.

At the Fry Canyon site, the water table is located approximately 8 ft to 9 ft below ground surface, and the underlying aquifer ranges from 1 ft to 6 ft deep. Estimated hydrologic properties and measured hydraulic gradients indicate that ground water in the alluvial aquifer moves at a rate of about 1.5 ft/day nearly parallel to the direction of stream flow. The uranium (U) concentration in the shallow colluvial aquifer ranges from 60 µg/L in water from a background well to 20,700 µg/L in water beneath the tailings. The hydraulic conductivity of the barriers is approximately 1,500 ft/day, while that of the surrounding native material is 1 to 2 orders of magnitude smaller. Native material consists poorly sorted fine- and medium-grained sand.

The funnel and gate system, installed in August 1997, is comprised of three barriers, each constructed of different reactive materials. One is bone char phosphate ( PO₄), another is foamed zero-valent iron (Fe⁰) pellets, and the third is amorphous ferric oxide (AFO). Each barrier is approximately 7 ft wide, 3 ft thick, and 4 ft deep. Approximately 110 ft³ of material was used in each barrier. Each contains 22 monitoring points, a water-quality mini-monitor, four pressure transducers, and a flow-sensor port. According to steady-state modeling results, ground-water velocities in the reactive walls are about 4.5 ft/day.

The EPA and U.S. Geological Survey have estimated that the design cost (engineering design and planning for the funnel and gate construction) for this system totals $30,000. The installation cost, including construction, materials, and the reactive materials, totals $140,000. These estimates do not include bench-scale testing of the candidate barrier materials.

Sampling continues at the site. The last sampling event was in November 2000. Overall, results indicate that PO₄ and ZVI PRBs are still removing > 99% of the incoming uranium, but the AFO PRB has reached chemical break-through, which impacts the long-term performance capability of the barrier. A 50% break-through in the AFO PRB after about 1,000 pore volumes had been processed.

If remediation using the PRBs continues at the site, two additional problems are anticipated that could affect the long-term performance of these barriers: long-term release of PO4 sorbed to iron-rich sediments in the colluvial aquifer after PRB removal, and carbonate and sulfide mineral precipitation in the ZVI barrier.

Results from early in the operation of these barriers showed that the input uranium concentrations differed significantly for each PRB, ranging from less than 1,000 mg/L in the PO₄ PRB to higher than 20,000 mg/L in the AFO ZVI. The input uranium concentrations to each of the PRBs also varied seasonally by approximately 4,000 to 7,000 mg/L. During the first year of operation, the PRBs are removing the majority of incoming uranium; however, the percentage of uranium removal varied with time and barrier material. The ZVI PRB consistently removed greater than 99.9 % of the input uranium concentration in flow-path 1. The percentage of uranium removed in the PO₄ and AFO PRBs was slightly less than the ZVI PRB. Except for two monitoring periods, over 90% of the input uranium concentration was removed in the PO₄ barrier. The AFO PRB removed over 90% of the input uranium concentration through November 1997. From January 1998 through September 1998 the uranium removal percentage was reduced to less than 90%.

Lessons Learned

Several important lessons have been learned as a result of the Fry Canyon project. They include:

1. The uneven surface of the underlying confining unit made it difficult to insure that each PRB gate structure or no-flow barrier was in direct contact with the underlying confining unit. If either structure was placed on small lenses of the residual colluvial aquifer, this may have provided a pathway for contaminated ground water to bypass the reactive material. A possible solution to prevent this would be the use of a more powerful track hoe that would be able to excavate into the underlying confining unit. This equipment would allow for a smooth surface and a gradient could be established that would drain the ground water away during excavation, allowing the observation of the underlying confining unit. In addition, the use of pumps with a capacity exceeding the ground-water inflow will allow for visual inspections of the seal between the PRB and the confining layer.
   
   
2. The use of pre-mixed bentonite slurry for construction of the no-flow barriers was problematic. It was difficult to control the movement of slurry from the wing wall to the gate structure of each PRB. It is critical to know exactly where the bentonite slurry is. If the slurry flows into a gate structure it could impact the flow and treatment of contaminated ground water in the finished PRB. In a worst case scenario, the gate structure of the PRB could be sealed off, preventing the treatment of contaminated ground water. A possible solution would be the use of non-hydrated bentonite chips for the construction of PRB wing walls and associated no-flow barriers. After placement, the chips would hydrate with the natural ground water, ensuring correct placement during PRB construction. In addition, the use of bentonite chips would not require the added expense of cement mixers for slurry transport.
   
   
3. The placement of monitoring wells within the wing walls and other no-flow areas between PRB gate structures is important to insure proper operation. Including these wells in the routine monitoring network can provide critical water-level and water-quality data useful in assessing PRB operation. For example, water levels measured in wing wall monitoring wells would be expected to respond more slowly to naturally occurring water-level increases observed within the PRB gate structures.
   
   
4. A large bedrock nose was encountered during PRB installation that resulted in a re-orientation of the PRBs. This re-orientation resulted in the entry of ground water at an oblique angle into the PRB gate structures, rather than the perpendicular angle that was anticipated. In order to prevent this problem in future PRB installations, a more detailed view of the bedrock topography is needed during site characterization activities. Additional data on bedrock topography could be obtained by increased drilling density during pre-installation characterization activities or possibly by subsurface geophysical methods such as seismic or ground penetrating radar techniques.
   
   
5. During the pre-installation characterization, it is important to determine the amount of readily desorbable U contained in the contaminated aquifer sediments. In a remediation scenario where the source term is either removed or stabilized in place, the total mass of readily desorbed U will eventually pass through the PRB. Quantification of this mass is needed to properly design the contaminant removal capacity of the PRB prior to emplacement.
   
   
6. Numerous hydrologic and water-quality characteristics should be considered prior to the selection of an appropriate point of compliance (POC) well when installing a PRB for ground-water remediation. When a PRB is placed within the contaminant plume, the POC well should probably be placed within the PRB. If the POC well is placed downgradient of the PRB it is likely that contaminant free water exiting the PRB could become re-contaminated with readily desorbable contaminants from the aquifer sediments. For some PRB materials, such as ZVI, the placement of a POC well within the PRB could be problematic. For example, the high iron concentrations and high pH values of water within the Fry Canyon ZVI PRB may not meet water-quality compliance standards. However, if the POC wells were placed downgradient of the ZVI PRB, the high iron concentration and high pH values would be significantly reduced. In many situations the location of POC wells may have to be parameter specific depending on the barrier material and onsite hydrologic and geochemical conditions.
   
   
7. Quantification of ground-water flow during pre-installation characterization and after PRB emplacement is critical. This information is needed for PRB design and to monitor changes in PRB hydraulic conductivity after emplacement. Ground-water models provide adequate information to address barrier design issues during the pre-installation characterization phase. After PRB emplacement, tracer injection methods appear to be best suited for monitoring PRB performance over time. Results through 1999 at the Fry Canyon site indicated that in situ flow sensors are of limited use in monitoring changes in either ground-water flow or direction within PRBs. These data are still being evaluated.
   

SITE-SPECIFIC REFERENCES

MORE INFO