SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
PERMEABLE REACTIVE BARRIERS
ACTION TEAM MEETING
Portland Community College—Washington County Workforce Training Center
Beaverton, Oregon
April 1516, 1998
WELCOME AND INTRODUCTIONS
Robert Puls, U.S. Environmental Protection Agency (EPA), National Risk Management Research Laboratory (NRMRL)
Dr. Puls, co-chair of the Remediation Technologies Development Forum (RTDF) Permeable Reactive
Barriers (PRBs) Action Team, began the meeting by welcoming the participants. Dr. Puls thanked the
Oregon Graduate Institute (OGI) for hosting the meeting and Environmental Management Systems, Inc.
(EMS), and Eastern Research Group, Inc. (ERG), for organizing the meeting.
EMPLACEMENT TECHNIQUES FOR PRBs
Permeable Barrier Remedy Performed at Caldwell Trucking Site
Mr. John Vidumsky, DuPont
Mr. Vidumsky said the ongoing PRB project at Caldwell Trucking in New Jersey, which is designed to protect receptors at an exposure point (a ground-water seep) as part of a site-wide remediation, significantly advances both PRB technology and its acceptance by regulators and the public. The project's risk-management focus departs from usual PRB application, which is to control contaminant sources. Mr. Vidumsky said that, because the next speaker would discuss the PRB installation method, he would present the project's background and regulatory issues.
The Caldwell site consists of large and small lagoons that cover 11 acres in northern New Jersey, near the Passaic River. The lagoons received industrial waste from the 1950s to 1984. The site, which is adjacent to light industrial and residential areas, was divided into two operable units (OUs): OU 1 is soil and sludge that is contaminated with lead and trichloroethylene (TCE); OU 2 is ground water that is contaminated with TCE. The ground-water plume extends 4,000 feet off the site toward the Passaic River and exceeds 10 parts per million (ppm) TCE in its center. Ground water rises to the surface in a 10-foot-by-10-foot seep before reaching the river. The regulatory history of the site includes two Records of Decision (RODs), two Explanations of Significant Differences (ESDs), and one ROD amendment.
The remedy selected for OU 1 (on-site soils) was in situ stabilization to control lead contamination and limited removal of TCE "hot spots." During the soil stabilization, DuPont found that volatile organic compound (VOC) levels in the soil were so high that soil vapor extraction (SVE) was necessary simply to meet health and safety requirements. The OU 1 remedy eventually incorporated voluntary SVE removal of 10 tons of VOCs from the soil.
The original ROD for OU 2 (ground water) called for a pump-and-treat system to reduce (with minimal treatment) the contaminant plume to potable-use levels within 30 years. The ground-water seep was to be remedied with stone fill. Because ground water in the area is not used for drinking, the identified ground water risks were direct contact with water at the seep and surface-water contamination. Although the original cleanup goal was to attain the maximum contaminant level (MCL), regulators soon recognized it was technically impracticable (pump-and-treat would have taken 100 years to reach MCL), especially considering that other VOC sources were in the area. An ESD changed the remedy to 30 years of pump-and-treat with a limited number of extraction wells (the public was opposed to a large array of extraction wells) to reduce ground water to potable levels with "moderate" treatment. The seep was to be remediated by upgradient extraction wells at a cost of $14 million.
Mr. Vidumsky said it was clear to him that even the modified remedy failed to make sense, especially given that the primary risk area was the small ground-water seep. DuPont reevaluated the plan, remembering the objectives of the remedial action: to eliminate direct contact risk, reduce surface-water concentrations to acceptable risk levels, and reduce contaminant mass. Superfund reforms of 1996 allowed potentially responsible parties (PRPs) to make ROD/remedy changes when justified by new information, changed site conditions, new remediation technologies, or significant possible cost savings. In light of these reforms, DuPont developed an alternate remedial approach that replaced the pump-and-treat system with an in situ PRB of zero-valent iron (ZVI). This PRB was projected to achieve the same mass removal as the proposed pump-and-treat system (500 kg/yr). This remedy also recognized the contributions of natural attenuation (3,000 kg/yr) and prior source reduction through SVE (10,000 kg total). The alternate remedy would also achieve greater and faster risk reduction at a location close to the area of greatest risk (the seep) while saving over $10 million.
EPA agreed to negotiate a ROD amendment if the potentially responsible parties (PRPs) could design and build the PRB and generate 1 year of successful performance data. EPA still wanted an additional remedy to control the pollution source, however. (The regulators and DuPont looked to achieve this source control through ZVI or bioremediation, rather than pump-and-treat.) In this "handshake deal," if the PRB failed, EPA would require the PRPs to install the pump-and-treat system, which had been designed. The remediation goal for water in the seep is 50 ppm, but Mr. Vidumsky said he expected lower levels. The remediation goal is limited not by the treatment capacity of the PRB but by the difficulty of treating all ground water flowing to the seep.
The PRB was installed on site in March 1998. The geology of the site consisted of an uncontaminated upper aquifer (A zone) that is underlain by a clay layer. This clay layer covers a contaminated lower aquifer of glacial stream channel deposits (B zone), which is above bedrock of basalt flows (C zone). There is significant ground-water flow through weathered bedrock in the upper C zone. The ground-water seep occurs at a window where the clay layer is absent. This window allows ground water to the surface under artesian conditions.
Placement of the PRB was greatly facilitated because the lot directly upgradient of the seep was for sale. To intercept ground-water flow in the B and C zones, the PRB extends from the unweathered bedrock to the top of the B zone. The PRB consists of two parallel, 4-inch-thick walls. The first wall is 150 feet wide, and the second is 100 feet wide. The walls were installed by permeation infilling of ZVI gel in the C zone and hydraulic fracturing in the B zone. Installation of the PRB was confirmed through electrical resistivity and hydraulic pulse testing. Performance monitoring is ongoing, but data are not yet available. Mr. Vidumsky thanked the regulators from EPA Region 2 and New Jersey for cooperating and "taking a chance" with a PRB system even though a pump-and-treat system design was complete.
Vertical Hydraulic Fracture Emplacement of PRBs
Dr. Grant Hocking, Golder Sierra LLC
Dr. Hocking began by describing methods of PRB installation. In "funnel-and-gate" PRBs, he said, an impermeable wall intercepts ground water and "funnels" it toward a permeable reactive zone or "gate" that treats the ground water as it flows. Dr. Hocking said that while funnel-and-gate PRBs are viable in many cases, they can be problematic because they alter ground-water flow. The PRBs can be installed through braced excavation, continuous trenching, augering, or vibrating beams. Dr. Hocking said that these methods are well-proven but can disrupt confining layers in the soil. He also said the methods can be quite expensive and/or "messy" (involving significant excavation). Hydraulic fracturing and jet grouting through boreholes are installation methods that do not alter ground-water flow or disrupt confining layers. In hydraulic fracturing, holes are bored to initiate a fracture in permeable sands. (This method cannot be used in rock.) A ZVI-containing gel can then be pumped into the fractures to form a continuous wall of reactive material that intercepts and treats ground water without altering its flow of direction. The gel, which is often made of guar gum (a natural food thickener), soon dissolves, leaving behind a wall of ZVI. Mr. Hocking said the oil industry has used hydraulic fracturing technology for 60 years to improve the yields of oil wells and prevent the wells from drawing sand.
The fracturing fluid used at Caldwell consisted of potable water, guar gum, a borax cross-linker (to link iron to the gel), pH buffer, an enzyme breaker (to break down starch in the guar after injection), and a fine-sand propant. Cross-linking was necessary because, as gel is forced through a fracture its velocity decreases, and as the velocity decreases the iron can fall from the gel without permeating the fracture. "Leak-off" tests ensure that the iron has been carried throughout the fracture. Dr. Hocking said that, because many undesired reactions can occur between gel ingredients (e.g., guar can reduce the reactivity of the iron), extensive compatibility, reactivity, and leak-off testing of the fracture fluid ingredients is essential before this installation method is used.
Dr. Hocking reviewed the geology of the Caldwell site, presented by Mr. Vidumsky. The rugged glacial channel deposits in the B zone made PRB installation very difficult. Golder Sierra installed the barrier in the B zone through hydraulic fracturing and used permeation in-filling of ZVI gel in the C zone. Golder Sierra installed upgradient and downgradient monitoring wells in the B and C zones before PRB installation. The boreholes for the hydraulic fracturing were 6.25 inch. PVC casing was placed 15 feet apart. Iron was loaded from hoppers into a mixing tank and an oil field frac pump. Dr. Hocking said the system required automation because it pumps iron into the ground very quickly (10 tons/hour). The ZVI and gel were cross-linked under pressure in the hose before the fluid reached the fracture. Dr. Hocking's group used special equipment to initiate the fractures in the borings and to ensure that the fractures were at the desired azimuth.
Dr. Hocking monitored barrier installation in real time with electrical resistivity testing. Electrifying the injection fluid caused the barrier to transmit a low-voltage signal, which receiver collars installed in nearby monitoring wells detected. The resistivity testing showed when a barrier at one borehole coalesced with adjacent barriers to create a continuous wall. Golder Sierra determined the thickness of resulting barrier by observing it directly with cameras in the boreholes. By forcing a pulse of water into monitoring wells and detecting the pulse with pressure transducers in adjacent wells (a process called hydraulic pulse interference testing), Dr. Hocking's group could detect any small holes in the wall that needed patching.
Over 6 to 8 weeks, 250 tons of iron were installed at the Caldwell site, at a cost of $700,000. In response to a question from the audience, Dr. Hocking said that "biofouling" of the PRB was not a concern, because the guar gum was very high quality and should leave little residue to encourage microbial activity.
The Use of Oxygen Release Compound and Hydrogen Release Compound for Enhanced Bioremediation
Dr. Stephen Koenigsberg, Regenesis Bioremediation Products
Regenesis has developed Oxygen Release Compound (ORC®), a patented formulation of magnesium hydroxide that releases oxygen slowly when hydrated. Many oxygen-releasing compounds are already used. Regenesis adapted ORC® from an agricultural soil amendment to create a compound that enhances the natural attenuation of organic contaminants. The Regenesis patent pertains to ORC®'s time-release feature. The slow-release property of ORC® results from phosphorous that is intercalated into the crystalline structure of the magnesium hydroxide. ORC® is a fine powder (average particle diameter is 44 microns) that is available in socks for ground-water wells or as an injectable slurry. The ORC® slurry can form barriers through push-point injections (e.g., by Geoprobe™) to "cauterize" a ground-water plume by treating ground water as it flows through the barrier. Dr. Koenigsberg stressed that ORC® is intended only to enhance natural bioattenuation and is generally not intended to clean ground water to MCL levels. ORC® lasts approximately 6 months in typical field applications.
Of the approximately 3,000 ORC® field applications to date, 700 have been significant remediations; 600 have been slurry injections and 100 barriers. The first major ORC®-barrier project was undertaken at a leaking underground storage tank (UST) site in New Mexico. Two barriers, consisting of wells filled with ORC® socks, were installed at the site to intercept ground-water flow. Numerous monitoring wells that were placed upgradient and downgradient of the barriers monitored barrier performance. After 200 days, the barriers required recharge with half the original number of ORC® socks. The system cost $100,000 for 1 year and achieved a 58 percent mass reduction of benzene, toluene, ethyl benzene, and xylenes (BTEX) in the study area.
Researchers from the University of Waterloo tested another ORC® barrier in Ontario, Canada. This system reduced contamination by 50 percent and lasted 7 months without recharge. At a smaller study site in California, ORC® wells saved $100,000 over the alternative technology, air sparging.
Dr. Koenigsberg said that, although ORC® was first used primarily in wells, the preferred approach now is to directly backfill a Geoprobe™ borehole with injectable ORC® slurry. A test site using this approach in Dexter, Michigan, reduced BTEX by 54 percent and benzene by 67 percent and saved $120,000 over alternative remedies. At another test site in Washington, an ORC® barrier with 1,000 pounds of ORC® in 15 boreholes reduced BTEX by 41 percent and benzene by 71 percent over 5 months. This system cost $40,000, while the alternative remedy was estimated at $250,000. Yet another ORC® barrier was tested at a site in Great Bend, Michigan, where the client preferred one-time push-point injection because it was more aesthetically pleasing than an air-sparging system. This site was significant because BTEX contaminants did not rebound, and the ORC® barrier did not require recharge. Dr. Koenigsberg said that ORC® can sometimes be a permanent solution and may not require repeated application.
Dr. Koenigsberg said that of 600 large-scale projects, 20 (3 percent), yielded poor results. Failure can result from poor designs that underestimate the ORC® requirements. Ground-water fluctuation can also cause flux in contaminant levels, and high iron levels at some sites can deplete the oxygen supplied by the ORC® and thereby limit natural attenuation.
Regenesis is now investigating ORC®'s effects on other contaminants, like methyl tertiary butyl ether (MTBE). ORC® appears to break down MTBE, but only after all available BTEX has been broken down. Dr. Koenigsberg said that BTEX may prevent the breakdown of MTBE by competitive inhibition of microbial activity (i.e., the microbes prefer BTEX to MTBE).
Hydrogen release compounds (HRC®) can form hydrogen barriers that cut the flow of chlorinated aliphatic hydrocarbons by promoting microbially mediated reductive dechlorination. The primary approach to forming hydrogen barriers is to slowly release lactic acid into the ground. Lactic acid is metabolized under anaerobic conditions to yield hydrogen. HRC® is a polymer gel that degrades, slowly releasing lactic acid. Because it is a gel, it is easily handled and installed to form barriers. A number of environmental consultants at 10 sites are testing HRC® through various applications, including rods, canisters, impregnated particles for trench backfill, and flowable materials for slurry injection.
New Installation Techniques in Side-by-Side Demonstrations at Cape Canaveral Air Station
Major Ed Marchand, U.S. Air Force
Major Marchand began with a video describing the side-by-side demonstration project at Cape Canaveral Air Station, Florida. Major Marchand cautioned the audience that the video is still under review and asked them to voice any comments.
The demonstration project tested two emplacement techniques for installing ZVI walls: hollow mandrel and jet-assisted grouting (JAG). The two techniques were used to build a 100-foot PRB of overlapping walls to intercept a TCE ground-water plume that flows toward the Banana River. The walls had the same design: a 50-foot main wall followed by a second 10-foot wall 4 feet downgradient and a third 10-foot wall 4 feet downgradient from the second. The Air Force installed in situ flow sensors and monitoring wells to track the PRB's performance.
The first emplacement method used a mandrel (a hollow, vibrating beam) to create a 4-inch void in the ground that was 45 feet deep and 32 inches long. A vibrating hammer drove the mandrel into place. ZVI was poured into the mandrel to fill the void. The mandrel was then withdrawn from the ground, leaving a panel of ZVI. These steps were repeated to form a continuous wall of overlapping wall panels. Each mandrel took 45 minutes to 1 hour to place. The 98 tons of ZVI took 19 days (9 days for construction) to place into 32 panels. Problems during installation included misalignment of the mandrels, vibrations affecting nearby buildings, and noise. Mobilization/demobilization cost $84,000, construction cost $89,000, and the ZVI cost $45,000.
The second emplacement method was JAG of a ZVI slurry mixed with guar gum and a binder. A vibrating hammer drove a 36-inch I-beam into the ground to guide a high-/low-pressure nozzle to 45 feet deep. Water jets in the high-pressure side of the nozzle helped create a void for the beam and iron slurry. The iron slurry was pumped into the void through the low-pressure side of the nozzle. The I-beam guide then was removed and the step repeated to form panels. The panels were overlapped to create a continuous wall. Each panel took 1 to 3 hours to place; the placement of 107 tons of ZVI in 24 panels took 34 days (12 days for construction). Problems during installation included flowing sand that clogged the nozzles, maintenance of consistent iron density in the slurry and constant volume of injected grout, vibrations, noise, and an extreme eye hazard from blowing iron dust. It was also impossible to determine the exact position of each panel after installation. Major Marchand said that one advantage of the pressure injection is that it may produce a thick wall in loosely packed, permeable soil areas (where more reactive material is needed) and a thinner wall in densely packed, less permeable areas. This method produced 4,000 gallons of liquid waste and 24 tons of soil waste. Mobilization/demobilization cost $84,000, construction cost $89,000, and ZVI cost $45,000.
Major Marchand said that while ground-water flow of 0.1 to 0.5 feet/day was expected, it was only 0.1 to 0.01 feet/day. While this increased the time contaminants resided in the barrier, it also increased the time needed to collect downgradient ground-water data for performance monitoring. In situ ground-water flow sensors have shown conflicting flow directions. Major Marchand said he planned to present performance-monitoring data from this project later. Quarterly monitoring would be completed in November 1998 and followed by a report in early 1999.
Major Marchand said that the most difficulty with these placement techniques was aligning adjacent wall panels. Because the panels were only 4 inches wide, there was little room for error during alignment. Major Marchand recommended using two tracks of I-beams in future installations to prevent the guides from shifting laterally during emplacement. Major Marchand's group tried to install several monitoring wells in the wall by placing them directly in the mandrel, but the well materials could not withstand the vibrations during emplacement.
An audience member asked if the mandrel insertion, by compressing the surrounding soil, would reduce soil permeability and cause ground water to flow around the wall rather than through it. Major Marchand answered that he used a 4-inch wall rather than 6-inch wall to avoid soil compaction. Major Marchand said he believed that vibrations during mandrel removal would reduce the pressure of the compacted soil and restore its permeability. The in situ flow sensors should help determine if this approach was successful.
Reactant Sand-Fracking Pilot Test Results
Donald Marcus, MacMarcus Resources
Mr. Marcus said that contaminated ground water in fractured bedrock can be very difficult to characterize as well as clean. EMCON Associates performed a pilot test study of reactant sand-fracturing (RSF) as a method for placing ZVI reactive media into a fractured bedrock aquifer contaminated with chlorinated solvents (cVOCs) and metals. The goals of the pilot test were to develop a propant, identify bedrock fractures for treatment, hydraulically inject the media, confirm emplacement of the media, reduce TCE and chromium levels, improve hydraulic conductivity, and increase capture and control of the plume for future pump-and-treat remediation.
The test site was a leach field at a former aerospace facility in Newbury Park, California. The geology of the site consisted of a relatively impermeable alluvium/colluvium layer overlying a permeable zone of weathered volcanic bedrock where ground-water contamination (TCE, tetrachloroethylene [PCE], and chromium) was most prevalent. A pump-and-treat system had been used to contain the plume, but it had removed very little contaminant mass. Mr. Marcus said that part of the problem with a pump-and-treat approach at this type of site is that altering the ground-water gradient can dislodge pockets of dense nonaqueous phase liquids (DNAPL) and cause them to sink farther into the ground. (Because they are denser than water, DNAPLs like pure TCE tend to follow gravity rather than the ground-water gradient.) A PRB would avoid this problem, but existing installation methods were impractical for the site's deep, fractured bedrock aquifer. Mr. Marcus and Dr. Robert Gillam at the University of Waterloo developed RSF as a new method for inserting a PRB into deep, fractured bedrock. RSF uses high-pressure fracturing with a sand propant and takes advantage of the fractures existing in the bedrock.
Because iron filings do not have the correct hydraulic properties for this approach, Mr. Marcus needed to develop a reactive fracturing fluid. Researchers at the University of Waterloo and EPA examined several propants in column tests. Dr. Jim Farrell at the University of Arizona tested an iron foam propant that was chosen for the pilot test. The density of the foam equals that of sand, making it much easier to inject than denser iron filings. The pilot test process consisted of boring holes, chemical and physical testing to identify fractures containing contamination, pretreatment hydraulic fracturing (to be sure the injection equipment could fracture the bedrock), injection of the propants, post-treatment confirmation of emplacement with down-hole geophysical and geochemical testing, and collection and disposal of waste fluids. Later in the process, Mr. Marcus' group observed biofouling in the injected barrier, which was caused by microbial activity. Commercially available acid and surfactant treatments were able to remedy the biofouling.
Mr. Marcus did not use cross-linkers in the propant because the low density of the iron foam made them unnecessary. The low-density foam required less expensive pumping equipment (standard hydraulic fracturing equipment was used) and less money. Bromide tracer tests showed that the propant infiltrated fractures over the desired area. Hydraulic pulse testing and permeability tests showed that the RSF had greatly increased the permeability of the treated area. Mr. Marcus found that high-angle fractures were more difficult than horizontal fractures to infiltrate with the propant. Chemical testing showed that most chlorinated solvents and metals were reduced significantly; chromium and TCE were reduced by as much as 98 percent. Dissolved oxygen in the treatment area decreased after treatment, indicating that reductive dehalogenation of the cVOCs was occurring. Higher levels of cVOCs, which were observed for a short period, were caused by a pocket of DNAPL that was dislodged by the RSF process.
Mr. Marcus concluded that iron foams are a suitable propant for RSF and are sufficiently reactive for long periods. In addition, RSF can be performed at relatively low cost with standard hydraulic fracturing equipment. Mr. Marcus said his ultimate goal for RSF technology was to use overlapping boreholes to create a continuous treatment zone to intercept ground-water flow.
Installation of a Reactive-Iron Permeable Barrier Using Continuous Trenching
James Romer, EMCON Associates
Mr. Romer presented the results of a recent project at a maintenance facility in southern Oregon where a plume of PCE, TCE, and other cVOCs extended one-half mile off the site. Highest concentrations were 500 parts per billion (ppb) PCE. EMCON evaluated both a pump-and-treat system and a funnel-and-gate PRB as part of the site's interim remediation. Mr. Romer said that, although pump-and-treat systems are often discounted as too expensive, recent advances have made them more cost-effective. Mr. Romer cautioned the audience not to dismiss pump-and-treat systems because they can be practical at some sites. At this site, however, the soil's low hydraulic conductivity made pump-and-treat impractical.
To launch the project, EMCON performed geotechnical borings every 50 feet across the plume to characterize site geology. Site geology consisted of an uncontaminated aquifer separated by an aquitard from a higher contaminated aquifer. Because only the upper aquifer needed remediation, a PRB could be keyed into the aquitard below the upper aquifer. EMCON considered a collection trench to intercept and remove contaminated ground water, but it would have been expensive due to the size of the site and the operation and maintenance costs involved. EMCON instead chose a slurry-wall funnel-and-gate PRB with two gates. The funnel-and-gate system was chosen over a continuous wall because it limited the amount of ZVI needed and thus reduced cost. The design called for a wall 650 feet long, 6 inches wide, and up to 34 feet deep, depending on the depth of the aquitard. EMCON chose continuous placement entrenching because it was relatively inexpensive and less disruptive than other methods and did not produce liquid wastes. After trenching, the slurry wall was mixed with water and bentonite in the trench. EMCON used grab samples and Shelby tubes to confirm that the slurry wall had been mixed sufficiently. Shelby tubes taken after installation showed the wall's hydraulic conductivity to be 1.4x10-7 cm/sec, which achieved the design goal. During installation, the contractor had to replace U-joints on the trencher repeatedly. Trenching 650 feet took 2 weeks, but Mr. Romer said it would have been less had the trencher not needed frequent U-joint replacements.
Although the design called for 14-inch thick reaction zones, the installation method allowed only a 9-inch wall of iron. To compensate, EMCON installed a second 9-inch reactive wall behind the first and connected the two with sheet piling. The area of the second gate contained a small cobble zone that the trencher used. Mr. Romer said that this showed that flexibility in the field can be important, because site characterization had failed to detect the cobble area. The contractor held the slurry wall open with sheet piling while clearing the gate with traditional excavation. A drag box was used to keep the gate excavation open. This was less expensive than using sheet piles around the excavation. After the system was completed, the site was paved for use as a parking lot.
Mr. Romer said that continuous trenching with in-place mixing of the slurry wall can be cost-effective (this
slurry wall cost $8 per square foot). Mr. Romer encouraged the audience to develop good relationships
with the team of contractors, consultants, and regulators when performing these projects, because field
conditions may force quick design changes that need regulatory approval. Mr. Romer said it was also
important to plan ways to dispose spoils from the excavation process.
PRBs FOR INORGANIC CONTAMINANTS
In Situ Treatment of Inorganic Contaminants Using Reactive Media Other Than ZVI
Dr. David Blowes, University of Waterloo
PRBs can be used to reductively dehalogenate organic contaminants or remove inorganic contaminants by reducing and precipitating them out of solution. Dr. Blowes said that while much work has been done with ZVI as a reactive media, he has focused his recent research on others. Many metals can be treated by changing their oxidation states or the oxidation states of other compounds that will precipitate with the target metals after oxidation changes. The ultimate goal of the treatment is to precipitate the metal from solution or to adsorb it in solid phase.
Dr. Blowes began by studying mine waste treatment, then moved to other waste sources. At mine waste sites, the treatment approach was to induce a sulfate reduction reaction in the aquifer by adding organic carbon in a reactive wall. Sulfate reduction produces hydrogen sulfide, which in turn reduces dissolved metals to form low-solubility sulfides. This approach has been used to treat ferrous iron as well as copper, nickel, lead, zinc, and cadmium.
Treatment development begins with theoretical transport modeling, followed by laboratory batch and column tests, small-scale field tests, large-scale modeling studies, and, finally, full-scale implementation. Ideal reactive materials are permeable, long-lasting (5 to 10 years), practical (easily obtained), and cost-effective. Previous studies found that mixtures of organic carbon from different sources are more effective than carbon from single sources. Leaf compost works particularly well. One method of testing reactive materials is to identify the secondary precipitates formed by the reaction between the media and the dissolved metals. This is accomplished through mineralogical techniques like optical microscopy, scanning electron microscopy (SEM), energy-dispersion spectroscopy (EDS), X-ray diffraction (XRD) and Debye-Scherer analysis, electron microprobe analysis, and X-ray photoelectron spectroscopy (XPS). These studies have shown that the product of the reaction with organic carbon is a poorly crystalline ferrous sulfide with other metals incorporated; no other discrete secondary precipitates have been observed.
Dr. Blowes' group had begun a pilot study at an industrial site in Vancouver, Canada, where the storage of metal sulfide ores had caused long-term release of metals (copper, lead, nickel, zinc, and cadmium) to ground water. The aquifer was 35 meters deep, sand-to-boulder, coarse-grained alluvium with a hydraulic conductivity of 3 x 10-3 to 3 x 10-2 cm/second. Ground-water velocity was fairly fast (920 to 200 m/year). The contaminant plume extended 15 meters deep into the aquifer and was 400 meters long. The plume rested on a denser salt water wedge below it.
During installation of the PRB, the 2.5-meter wide trench was kept open with a guar gum slurry. The reactive media in this test was 85 percent pea gravel and 15 percent municipal compost, with a small amount of limestone. The gravel was to improve wall permeability; the organic content was sufficient for several years. Hydraulic conductivity of the wall was from 0.9 to 0.15 cm/second, and the residence time within the wells was 3 days. The performance of the barrier was tracked with field monitoring, in situ probes, and the sampling of three rows of multilevel monitoring wells placed perpendicular to the wall. Within 2 months of installation, metal concentrations were detected below target levels in the barrier and moving downgradient from the barrier. Metal concentrations have continued to decrease. The barrier quickly achieved reducing conditions and decreased metal concentrations, and there was little wall compression. Because of these results, a full-scale PRB (400 meters long, 2.5 to 5 meters thick) is now planned for the site.
Dr. Blowes said he is also researching the treatment of inorganic anions, like phosphates, selenium, and arsenic. Because phosphate loading can cause eutrophication in lakes, phosphate treatment is needed at some sites where phosphate-contaminated ground water flows from domestic septic systems into surface water. One removal mechanism is a barrier that enhances phosphate absorption/precipitation. Dr. Blowes has tested a reactive material of iron, limestone, and silica sand. Column tests with 10 ppm phosphate solutions have achieved 95 percent phosphate removal after 1,500 pore volumes. A pilot-scale alternative waste-water treatment system at the Waterloo sewage treatment plant achieved excellent removal of phosphorous, with a residence time of 1 to 2 days.
Dr. Blowes said that arsenic and selenium may be treated through increased anion absorption, coprecipitation, or reduction and formation of insoluble solids. Dr. Blowes has tested metal oxide mixtures and ZVI mixtures as reactive materials for anion treatment. A mixture of iron slag, calcium carbonate, and silica sand rapidly removed 99.9 percent of arsenic and selenium in batch and column tests, with a capacity of 150 pore volumes. Activated alumina achieved similar results, but its capacity is still being evaluated.
In conclusion, Dr. Blowes said that the PRB approach is suitable for treating numerous types of dissolved metals and nutrients.
Moffet Field Hydraulic Capture and Iron Cell Coring Results
Charles Reeter, Naval Facilities Engineering Service Center
Dr. Bruce Sass, Batelle Memorial Institute
Mr. Reeter began by summarizing the results of the first tracer tests conducted at the pilot scale PRB at Moffet Field in California. (Mr. Reeter did not describe the Moffet Field site, because it was described in previous RTDF presentations.) The test showed that the tracer dispersed rapidly through the pea gravel zone upgradient of the reactive cell but took complex, preferential pathways through the cell itself. Linear flow velocities in the cell ranged from 0.05 to 0.45 feet/deep. Mr. Reeter's group recently conducted a second tracer test and tracked 16 monitoring points using continuous bromide monitor probes. The second test confirmed that the funnel-and-gate PRB system captures the tracer and directs it though the reactive zones.
Dr. Sass presented the results of coring analysis at the Moffet Field site. Dr. Sass said this coring was performed to investigate the physical and chemical changes occurring in the reactive cell. The researchers had suspected that the high sulfate content of the ground water was being reduced to sulfides by the ZVI; drops in alkalinity and calcium concentrations across the barrier had indicated that calcium carbonate may have also been precipitating in the reaction cell. Coring was performed to see if these precipitation reactions were occurring.
In addition to vertical cores, Dr. Sass' group collected angled cores to capture the upgradient and downgradient interfaces in one sample. Methods for analyzing the cores included bulk chemical analysis (through acid digestion), SEM, raman spectroscopy, XRD, EDS, and microbial analysis. SEM indicated a fibrous material on the iron that may have been a sulphur compound. EDS showed that the sulphur level tailed off from the upgradient to downgradient interface of the reaction cell, while calcium levels were more constant across the cell. XRD and raman spectroscopy showed mineral species in the unused iron to include hematite, maghemite, magnetite, and graphite, while aragonite, calcium carbonate, and marcasite (a sulphur mineral) dominated in the used portion of the reactive cell. Microbial analysis yielded no measurable colonies under anaerobic conditions, but at least one anaerobic microbe species was detected in the aquifer.
Dr. Sass said that the results to date illustrate the early conditions of the Moffet Field PRB but cautioned that the site has operated for only a short time. While the researchers have observed some interesting mineral changes in the reaction cell, these changes have been minor. These result show that processes including oxidation, precipitation of carbonate and sulfides, and anaerobic microbe growth may be occurring in the reaction cell and should be monitored through coring in 5 years to determine if they are affecting the PRB's performance.
Reactive Barriers for Retention and Removal of Uranium, Technetium, and Nitrate in Ground Water
Dr. Baohua Gu, Oak Ridge National Laboratory
Dr. Gu presented data from two PRB demonstration projects at the U.S. Department of Energy (DOE) Y-12 plant in Oak Ridge, Tennessee. This site is a former disposal pond that received 80 to 100 million gallons of waste from 1951 to 1983. Contaminants included nitrates, uranium, and technetium, as well as PCE and other cVOCs. Ground water at the site is very acidic. There are three major ground-water pathways through the site, each with a different contaminant level. The goal of the PRB project was to use ZVI or other media to remove a mix of contaminants. Two types of barriers were used: a permeable gate and an impermeable funnel and gate.
Laboratory tests showed that ZVI quickly removed uranium, chromate, and technetium. Absorbent materials were less efficient at removing uranium due to their limited absorption potential. Nitrate removal was tested with both ZVI and a ZVI/peat mixture. The ZVI/peat mixture removed nitrate more efficiently, converting it primarily to ammonia. The half-life of nitrate depended on the initial nitrate concentration (lower initial concentration led to a shorter half-life).
The funnel-and-gate system pilot test was conducted at ground-water pathway 1. The barrier consisted of a wall of high-density polyethylene (HDPE) membrane that funneled to a concrete vault containing the treatment cell gate. The treatment cell consisted of a treatment train to treat multiple contaminants. Two treatment systems were designed. The first uses a mixture of iron and lime that is designed to raise the pH of ground water and prevent the discharge of ferrous iron. The second is similar except it includes an electrokinetic system that is intended to raise the pH and prevent iron corrosion, thereby enhancing the uranium-removal efficiency of the reactive media. The treatment cell has not yet been installed in the field.
A permeable trench was installed at ground-water pathway 2. Because the native soil is relatively impermeable, a permeable trench containing ZVI was designed to collect water, treat it, and discharge it downgradient. A guar gum was used to stabilize the trench during construction. Performance data show that dissolved uranium was greatly reduced downgradient of the trench. Higher uranium levels that were detected in one trench area may have been due to an additional ground-water pathway. Tracer tests are planned to investigate. Nitrates were completely degraded in the trench, although initial nitrate levels were lower at pathway 2 than at pathway 1. Sulfate decreased in the trench as it reduced to sulfides. One difficulty is that the system discharges ferrous iron, which eventually appears in localized areas of surface water; this ferrous iron may have been caused by microbial activity encouraged by the guar gum. The ferrous iron discharge has decreased. Dr. Gu said that column studies showed precipitation of ferrous iron, ferrous sulfide, calcium carbonate, and ferrous carbonate, so there is concern that these precipitates will clog the reactive media in the pilot-scale system. Dr. Gu said coarser filings were installed in the trench barrier and are expected to limit the unwanted precipitates. Dr. Gu said that, as with all PRBs, long-term performance is another area of concern. Dr. Gu said he believes the installed system should work as long as the redox potential of the system is low and there is sufficient reactive material. Another consideration for this site is potential remobilization of radionuclides. Unlike organic compounds, which are degraded in the system, radionuclides and other metals are adsorbed by the barrier or precipitated and could remobilize. Possible solutions are to excavate the reactive material after the barrier has ceased functioning or to build another barrier downgradient of the original.
In summary, Dr. Gu said that ZVI is an efficient and cost-effective method for removing by reductive precipitation mixed wastes, including uranium, technetium, and chromium. A ZVI/peat mixture effectively degrades nitrate. One barrier system has been installed and initial data has been gathered, and installation of a second system is nearly complete. Potential concerns for these systems are discharge of ferrous iron, clogging caused by precipitation, long-term performance, and remobilization of radionuclides.
Status and Preliminary Results of the Fry Canyon Reactive Chemical Wall Project
Dr. David Naftz, U.S. Geological Survey
The Fry Canyon site in southeastern Utah is a former uranium upgrading facility that was abandoned in the 1960s. Dr. Naftz presented preliminary results of a long-term field demonstration at the site to test PRB as a treatment method for uranium in ground water. Ground water at the site is shallow (10 to 15 feet deep), and uranium concentrations exceed 20 ppm in some areas. The demonstration consists of three funnel-and-gate barriers containing three reactive materials: ZVI, amorphous ferrous oxyhydroxide (AFO), and phosphate in the form of Cercona bone char. Installed in 1997, the reactive materials were placed in excavations held open with a template. The reactive zones are 3 feet wide and 3 feet thick. Performance of each barrier is tracked with transducers, 20 monitoring wells, and water-quality monitors; 27 chemical and physical parameters are measured hourly at each barrier.
Ground-water modeling and preliminary results from the transducers indicate that ground water is flowing through the barriers. Dr. Naftz said that reverse flow through the barrier, which can be caused by flash flooding of a downgradient stream, is a concern at the site. Residence time in the barriers is estimated at 16 to 30 hours and will be confirmed through an upcoming tracer test.
Uranium concentrations have been measured over four monitoring periods to date. The ZVI barrier reduces uranium concentrations to the detection limit in the first foot of the barrier and maintains this low level through the length of the barrier. The AFO barrier reduces uranium to near the detection limit; a small amount of uranium breakthrough was detected in one sampling event. The phosphate barrier reduces uranium to near the detection limit in the first 1.5 feet of the barrier, but there has been an unexplained rise in the uranium level over the second 1.5 feet of reactive material. ZVI is the most efficient reactive material over time, but all three materials have performed efficiently.
Dr. Naftz said that decreases in iron, calcium, and carbonate concentrations across the ZVI barrier indicate that precipitation may be occurring. Based on flow through the barrier, Dr. Naftz estimates that there will be roughly a 50 percent void loss over 20 years due to precipitation of calcite.
Dr. Naftz said that regulators from the state of Utah are concerned about quality of water discharged from the barriers. Potential concerns were high-pH water being discharged from the ZVI barrier and high-phosphate levels being released from the phosphate barrier. Although the barriers are clearly affecting the pH of water in the barriers, thus far no pH changes have been seen in downgradient ground water. Elevated phosphate has been seen in some sampling events in some downgradient monitoring wells, but Dr. Naftz believes that naturally high iron concentrations in the sediment should prevent significant phosphate migration. Increased uranium levels were observed in downgradient wells. Dr. Naftz said this may be caused by desorbtion of uranium from contaminated sediments downgradient of the barriers. If this is the case, these uranium levels would be expected to decrease. Dr. Naftz said that, unlike some of the other sites discussed at the meeting, ferrous iron levels have not been elevated downgradient of the barrier.
In summary, Dr. Naftz said that ground water is flowing through the barriers, each barrier is removing uranium but differently, chemical longevity of the barriers appears promising, and the barriers are not degrading downgradient water quality.
In response to a question from the audience, Dr. Naftz said that in the case of the phosphate barrier, the reactive material should eventually be removed to prevent the remobilization of uranium. Dr. Stan Morrison, who is also involved with the project, said that there is little evidence of remobilization with ZVI media.
Recent Progress on In Situ Redox Manipulation Barriers for Chromate and TCE
Dr. John Fruchter, Batelle Pacific Northwest National Laboratory
Dr. Fruchter updated the status of a project he presented to the RTDF in December 1996. This project involves an in situ redox manipulation (ISRM) barrier installed at the Hanford DOE site in eastern Washington. This site was a weapons material production facility where a number of contaminant plumes, including uranium, technetium, chromate, strontium, PCE, and carbon tetrachloride, flow toward the Columbia River. The site is arid, and ground water is 50 to 350 feet deep. The goal of the project is to remediate deep ground-water plumes with a PRB. Because placing ZVI at such depths is impractical, the selected approach was to use sodium dithionite to reduce the natural iron in the aquifer (called structural iron, Fe III) to Fe II. The sodium dithionite can be pumped into the deep aquifer through standard ground-water wells. Dr. Fruchter said that while the reaction between sodium dithionite and Fe II reduces iron to ZVI, the Fe II that is produced still reacts readily with many contaminants. Sodium dithionite was chosen because it degrades quickly and is plentiful in the area.
Chromate was the initial target contaminant in bench scale tests. In 1995, these tests were followed by a proof-of-concept field test of one injection well, which created a 50-foot thick barrier. This barrier reduced chromate levels from 80 ppb to 0.5 to 2 ppb. Dr. Fruchter's group selected another area with higher chromate levels (1 to 2 ppm) for a larger treatability test. This test will involve injecting dithionite into five wells; the treated areas will overlap to create a continuous barrier that is 150 feet wide. Dithionite has been injected in one of the five wells to date; the other wells will soon follow. Early monitoring data show that the initial barrier section is reducing chromate to below the detection limit in downgradient ground water.
Dr. Fruchter said that stakeholders in the area want assurance that the plume downgradient from the barrier will reoxygenate before reaching the Columbia River. These stakeholders are concerned that deoygenated ground water flowing to the river could harm downstream salmon-spawning areas. The researchers are currently conducting gas tracer tests to respond to these concerns.
Because many more site have TCE contamination than chromate contamination, Dr. Fruchter said he has been eager to study ISRM as a treatment for TCE. Initial bench scale tests have shown that the half-life of TCE in a ISRM barrier is 52 hours. Dr. Fruchter said that there is some evidence that TCE is degraded in the barrier through beta elimination rather than stepwise dechlorination.
Two demonstration projects to tests TCE treatment have been proposed, one at Moffet Field, California, and one at Fort Lewis, Washington. Dr. Fruchter has begun bench scale testing for each. Dr. Fruchter said that because this treatment depends on natural iron in the soil, tests using soil from the actual are site particularly important. Dr. Fruchter estimated that soil must have an iron content of 0.02 percent or more for this method to work.
Dr. Fruchter summarized the advantages of ISRM treatment: It achieves a treatment zone that treats redox-sensitive contaminants in a method similar to a ZVI PRB, pumping is required during the emplacement but not during treatment, and the ISRM barrier can be injected to treat deep ground-water plumes.
In response to a question from the audience, Dr. Fruchter said that dithionite would be needed to recharge the barrier only every 15 to 30 years. The dithionite dissipates very quickly, leaving a stable zone of reactive iron.
Status of a PRB Project for Uranium Containment at Monticello, Utah
Dr. Stan Morrison, Roy F. Weston
Dr. Morrison said that he considers the PRB project at Monticello, Utah, to be a spin-off from the Fry Canyon project discussed by Dr. Naftz. The Monticello project involves the treatment of uranium-contaminated ground water at a uranium mill tailings site. Acid, carbonate, vanadium, and uranium tailing piles are at the site. Ground-water contaminants include manganese, arsenic, uranium, molybdenum, and vanadium. Site geology consists of two bedrock banks on either side of a valley of alluvial sediments that contains a shallow aquifer. The impermeable bedrock sides of the valley funnel contaminated ground water from the tailing piles through the narrow alluvial aquifer, making the site ideal for a PRB to intercept and treat the ground water. In fact, an early conceptual PRB design for the site would intercept an estimated 95 percent of total ground-water flow through the aquifer.
In 1992, a field trial of a ferric chloride injection treatment had been planned but was eventually canceled. The new plan for the site is a full-scale PRB to be installed through traditional trenching and filling. The PRB may be a funnel and gate or a continuous wall. Given the high permeability of the aquifer, some ground-water mounding against the PRB is expected. Because the ground water is 10 to 15 feet below ground surface, however, mounding should not reach the surface.
Batch tests are underway to determine the feasibility of both sorbents and reactive media. Materials tested include ferric oxyhydroxides mixed with gravel or coated on glass foam, volcanic rock, or hematite pellets; activated carbon; peat; humate pellets; HUMASORB-CS™; and apatite (fish bones or bone char). The reactive material must increase the pH of the ground water to limit the iron mobility downgradient of the barrier. Most of the iron types tested in the laboratory removed the target contaminants very quickly. Dr. Morrison said that data concerning the types of reactions occurring with these media are limited. He suspected that reductive precipitation is the primary mechanism in the ZVI, although he expects other reactions as well.
More laboratory tests on the media are planned and will be followed by field column tests using site ground water. Dr. Morrison hopes to begin installing the PRB in October 1998.
Temperature-Dependent Reductive Dechlorination of TCE by ZVI and Tin
Dr. Chumming Su, EPA NRMRL
Dr. Su said that several uncertainties surround ZVI treatment of organic contaminants, including the chemical mechanisms, the effect of acid treatments, and the effects of temperature. The objective of Dr. Su's study was to investigate the reactivity of iron and tin metals in reductive chlorination of cVOCs. Dr. Su conducted kinetics testing to determine the effects of HCl washing and temperature on the reactivity of these metals.
Dr. Su studied iron from four vendors: Aldrich, Fisher, Master Builders, and Peerless. HCl washings increased the total surface area on grains of the Fisher, Master Builders, and Peerless irons but decreased the surface area on the Aldrich iron grains. Dr. Su found a pseudo first-order reaction rate and a linear relationship between surface area and TCE degradation with these metals. Dr. Su said that, in addition to increasing the surface area, the HCl washing appeared to reduce the activation energy of the degradation reaction on the metal surface.
Higher temperatures (55C) greatly increased TCE degradation in all the reactive materials, including virgin and HCl-washed metals. Tin generally degraded TCE slower than did iron, but the temperature effect was more significant for tin than for iron. Dr. Su quantified the effect of temperature on TCE degradation with the Arrhenius equation. Half-lives of TCE decreased with increased temperature for all metals tested.
Dr. Su said that the detection of acetylene as a break-down product of TCE indicates that beta elimination may have occurred rather than reductive dechlorination.
Dr. Su said that the results of this research have practical implications. In light of the observed temperature effect, PRBs in cold, northern climates should be relatively thicker than PRBs designed for warmer regions. Also, it might be possible to thermally enhance the degradation of contaminants by heating the reactive material in PRBs, either electrically or through radiofrequency heating.
Coordination of Long-Term Performance Monitoring at PRB Sites
Dr. Robert Puls
Dr. Puls said that last year the PRB Action Team Steering Committee was formed to push the Action Team in a focused direction. The Steering Committee recognized early the importance of researching the long-term performance of ZVI PRBs. In particular, performance can be impaired by lost permeability or reactivity as a result of precipitation of minerals within the reactive zone.
The Steering Committee has tried to coordinate the research of long-term performance to develop (1) testing that can predict the longevity of a PRB, (2) monitoring methods that might give early warning of pending failure or degradation of a PRB, and (3) long-term monitoring protocols that reduce monitoring and operation and maintenance costs. This research requires studying PRBs at different sites with various aqueous chemistries and other site conditions.
In January, Steering Committee members met at Oak Ridge National Laboratory to determine what EPA, the DOE, and the DOD can do to facilitate the research of long-term PRB performance. The Steering Committee's approach was to leverage funding from these organizations to support a concerted effort to study the following basic factors that could influence long-term performance:
The Steering Committee proposed that each lead agency fund long-term performance research at two or three sites to develop a body of comparable data across six to nine sites. To ensure that the data among sites are comparable and data-collection methods are consistent, the Steering Committee will coordinate the research.
Several study sites and cooperating agencies have been selected for this research. EPA will work with the U.S. Geological Service at the Denver Federal Center, Colorado; the U.S. Coast Guard at Elizabeth City, North Carolina; and General Electric at the Somersworth, New Hampshire, landfill. Dr. Puls will lead the EPA research effort. Funding for these EPA sites is in place, and sampling has already begun at the Elizabeth City site. The DOE will study the Oak Ridge National Laboratory Y-12 site in Tennessee; a site in Kansas City, Missouri; and the Rocky Flats site in Colorado. Nic Korte of Oak Ridge National Laboratory will lead the DOE sites. The DOD research will involve the Navy, the Air Force, and the Army Corps of Engineers, as well as sites at Cape Canaveral, Florida; Moffet Field, California; and, perhaps, Dover Air Force Base, Delaware. Charles Reeter will coordinate the DOD research.
Dr. Puls said that other members of the Action Team were welcome to get involved in the research effort.
One participant said that there are many questions about the proper way to decommission PRBs. Dr. Puls agreed and said the decommissioning method would depend on the contaminants and the end-products of the treatment reaction. Inorganic contaminants are clearly more problematic than organic contaminants. Another participant commented that when implementing PRBs, more attention should be focused on potential ecological effects. Dr. Puls said that this is especially true when PRBs are built close to downgradient surface waters.
Ground-Water Cleanup with Scrap Iron: A Case Study in Technology Development for Teaching
Environmental Chemistry
Dr. Paul Tratnyek, OGI
Dr. Tratnyek described an educational CD-ROM he is developing with funding from the Dreyfus Foundation. This CD-ROM will use PRBs to illustrate the principles of chemistry, environmental science, and other scientific subjects. Dr. Tratnyek said the inspiration for this product came to him when he wrote an article about ZVI treatments for Chemistry and Industry magazine. The article was warmly received, which caused Dr. Tratnyek to realize that ZVI PRBs touched on a wide range of science concepts and might be serve as an educational tool. Dr. Tratnyek went on to develop a proposal to repackage the ZVI PRBs information in an educational CD-ROM.
Dr. Tratnyek began developing the CD-ROM with a storyboard describing the product's structure. The CD will have three levels of information to appeal to different audiences: an introductory level, for high school students or industry trustees and CEOs who may want a basic explanation of a PRB; an intermediate level, for college science students and environmental project managers; and an advanced level for teachers, thesis students, and professional engineers. At the introductory level, the program could illustrate basic chemistry principles, such as redox chemistry. For intermediate users, the program could simulate reaction kinetics or product distribution through a reaction zone and, allow users to manipulate parameters. Features for advanced users could include Quicktime/MPEG movies illustrating chemical reactions in three dimensions. The CD-ROM will include a searchable database of references on various topics. This database already exists on the Web and could be packaged into the product or linked "live" to the product so the user can access the most up-to-date database even if the CD-ROM itself becomes out of date.
The "front end," or introduction, of the CD-ROM will describe the technological concept of PRBs and the story of its development. Dr. Tratnyek said he would like help from Action Team members in collecting the content of this section. He would like information and resources from members of the RTDF who have been developing PRBs; especially useful would be stories, graphics, pictures, and videos. In addition, Dr. Tratnyek is looking for software/hardware support, beta testers, and a product title. Contributors would receive exposure and visibility within the product, perhaps in the form of dynamic links to their Web sites.
The Dreyfus Foundation is funding a press run of 3,000 copies, depending on the final price, and intends to have the product ready for Christmas 1998. The first copies will be distributed free to schools and may be followed by an improved version that could be marketed. Because the content of the CD-ROM must be determined within the next 3 months, Dr. Tratnyek asked those who may contribute to do so as quickly as possible.
Electronic Status Report of the Ongoing and Completed PRB Projects Available on the RTDF Web Site
Dawn Carroll, EPA, Technology Innovation Office
Ms. Carroll said that Action Team members have asked the EPA's TIO to make the Action Team's RTDF Web site more comprehensive. In response, the TIO, with EMS and members of the Steering Committee, is adding several features to the Web site. The primary addition to the Web site will be a status report database of ongoing and completed PRB sites. The database will organize PRBs by six contaminant classes: chlorinated solvents, metals and organics, fuel hydrocarbons, nutrients, radionuclides, and other organic contaminants. Information for each site will include installation data, contaminants, reactive media, installation cost, construction, and point of contact. Each site will have a "Lessons Learned" button, which will list specific problems encountered during PRB implementation. There will also be a one-page narrative describing each PRB project. The Web site will also include an area where visitors can register their PRB sites and enter information about them. Ms. Carroll is now developing a list of PRB sites to be included in the database and is looking for relevant hyperlinks. She asked the Action Team members for any suggestions they may have.
The RTDF Web site is http://www.rtdf.org.
Dover Air Force Base PRB Site Update
Lieutenant Dennis O'Sullivan, U.S. Air Force Research Laboratory
Lieutenant O'Sullivan provided an update on the large pilot-scale demonstration project underway at Dover Air Force Base in Delaware. The project is at the site of a cVOC plume that is in a surficial aquifer that is bounded by an aquitard at 40 to 45 feet deep. Ground-water flow at the site is very slow (0.5 to 0.1 feet/day). The upper portion of the site has a high dissolved-oxygen content. The PRB is a funnel-and-gate system with two reaction zone gates. The impermeable walls were made with Waterloo sealable sheet piles, and the reaction zones were installed with 8-foot caissons. Permeable pretreatment zones were installed directly upgradient of each reaction zone. Construction of this PRB was completed in early January 1998.
The original objective of the demonstration project was to compare new materials for a PRB reaction zone. Lieutenant O'Sullivan said that after many laboratory and field column tests, the only promising new reactive media was a pyrite mixture. The cost of filling a reaction zone with a pyrite mixture was prohibitive, so ZVI was placed in both reactive zones. The researchers varied the pretreatment zones in front of the reaction zones; one is a 10 percent iron/sand mixture, while the other is a 10 percent pyrite/sand mixture. Arrays of monitoring wells are positioned upgradient, within the reactive zones, and downgradient of the barrier.
Two comprehensive monitoring events are planned for July and December 1998. Weekly water-level measurements have been collected since installation. The researchers use ground-water velocity probes to determine ground-water flow vectors. Preliminary data indicate that ground water is flowing through the gates. The PRB was placed downgradient of the leading edge of the cVOC plume, and contaminated water is expected to take several months to reach the wall. Lieutenant O'Sullivan said that with cVOC levels of 1,000 ppb, the site is not as contaminated as he would have liked for demonstration purposes. The residence time in the reaction zones is a very conservative 23 days.
A tracer test is scheduled, as well as a study to analyze colloid transport across the reaction zones. Lieutenant O'Sullivan also plans to sample cores from the reaction zones and analyze any biofouling of the wall. Future studies may also include microbial adhesion modeling to determine whether microbes transport contaminants across the barrier. Although performance data has not yet been collected, engineers at Dover are already planning to extend the PRB to capture the entire cVOC plume and are effectively using the pilot-scale project as a treatability study for the base. Lieutenant O'Sullivan concluded by saying that the RTDF concept has worked for this project by helping to build an incorporated effort with many contributions from members of the Action Team.
Mixed Waste Remediation Using HUMASORB-CS™—An Adsorbent to Remove Organic and Inorganic Contaminants
Dr. H.G. Sanjay, ARCTECH, Inc.
Dr. Sanjay introduced HUMASORB-CS™ as a patent-pending adsorbent that removes organic and inorganic contaminants. The DOE funded the development of HUMASORB-CS™. Objectives in the development of this product were to use the novel properties of humic acid to capture metal ions and adsorb organics, to characterize humic acid's ion exchange capacity, to impart mechanical strength to HUMASORB-CS™ to improve its handling characteristics, and to develop in situ processes for applying the material in environmental remediation and water treatment.
Properties of humic acid include:
Dr. Sanjay said that ARCTECH, Inc., is one of the largest producers of liquid humic acid products for soil amendment. Humic acid comes from natural dark-brown or black organic macromolecules with a molecular weight between 0.5 to 150 Kda. Humic substances are divided in three fractions based on aqueous solubility: fulvic acids (soluble at all pHs), humic acids, (soluble at high pH), and humins, (insoluble at all pHs). Humic acid is formed by plant and animal decay under moist conditions and is also produced in a process called humification. Humic acid can be extracted with varying yields from sources like coals, soils, peats, plants, water, and sediment. Regardless of the source, humic acid's elemental composition consists only of carbon, hydrogen, oxygen, nitrogen, sulfur, and phosphorous. Recent collaborative research by Temple University, Northeastern University, and Birmingham University (U.K.) has used three-dimensional modeling to investigate the properties of humic acid.
Developed using proprietary methods, properties of HUMASORB-CS™ include:
Laboratory tests have shown that HUMASORB-CS™ can remove contaminants including metals, oxyanions, radionuclides, and chlorinated and nonchlorinated organics. One column test using a 20 percent HUMASORB-CS™/sand mixture removed 20 ppm lead for 2,800 pore volumes. The column then removed 500 more pore volumes of a 200-ppm input solution before breakthrough finally occurred. Tests show that HUMASORB-CS™ has multiple binding sites that enable it to bind more than one type of metal at once.
Bench testing is underway for a HUMASORB-CS™ demonstration project at the Berkeley Pit Superfund site in Butte, Montana. HUMASORB-CS™ will be used to remove iron, copper, zinc, and cadmium from a waste stream. Researchers will use HUMASORB-CS™ in a two-stage system. The first stage will use a liquid humic acid product to remove nutrient metals while allowing toxic metals like cadmium to pass to the second stage. The spent media from the first stage may be used as an agricultural nutrient. The second stage will use HUMASORB-CS™ to remove the toxic metals.
Laboratory tests show that HUMASORB-CS™ removes TCE and PCE faster than ZVI (TCE half-life in ZVI was 16 hours, compared to only 2 hours in HUMASORB-CS™). Adding ZVI to HUMASORB-CS™ did not improve its performance.
Dr. Sanjay said that, although HUMASORB-CS™ has a higher cation exchange capacity than ZVI, the cost of HUMASORB-CS™ ($2,000 per ton) is still three to five time less than ZVI. Dr. Sanjay expects the cost of HUMASORB-CS™ to decrease as the production process is optimized. ARCTECH has begun pilot-scale production of HUMASORB-CS™. HUMASORB-CS™ is 80 to 85 percent regenerable for at least two or three cycles, is denser than water, and can be produced in different grain sizes. Dr. Sanjay said HUMASORB-CS™ avoids common problems that can occur with ZVI, like iron release and contaminant remobilization.
In response to questions from the audience, Dr. Sanjay said that HUMASORB-CS™ does not appear to adsorb calcium while binding other metals, so dissolved calcium does not reduce the capacity of HUMASORB-CS™ to treat target metals. HUMASORB-CS™ can selectively remove different groups of metals, depending on the pH of the input solution.
Development of a Permeable Barrier Process for Pentachlorophenol-Contaminated Ground Water
Dr. Sandra Woods, Oregon State University
Chlorinated phenol (e.g., pentachlorophenol [PCP]) can be degraded by both aerobic and anaerobic processes. In anaerobic conditions, reductive dechlorination replaces chlorine with hydrogen and the phenol become an electron acceptor. In aerobic conditions, the chlorinated phenol donates electrons and is typically hydroxylated. Dr. Woods said that the rationale for her group's research was to combine these aerobic and anaerobic processes in a PRB to treat PCP in ground water. In anaerobic processes, chlorinated phenols, chlorinated benzoates, and other compounds typically are degraded more rapidly when more highly chlorinated. However, nonchlorinated and less chlorinated phenols (e.g., dichlorinated phenol [DCP]) are degraded much more rapidly than PCP when in aerobic conditions. Dr. Woods said that combining aerobic and anaerobic processes in series could take advantage of both the more rapid rates of reductive dechlorination of highly chlorinated phenols and the rapid rates of hydroxylation and ring cleavage for less-chlorinated phenols.
Dr. Woods' group is field testing a system that uses an anaerobic zone and two aerobic zones in series. The anaerobic zone is preceded by a mixing zone where an electron donor (vanilla) and nitrogen gas are mixed. This anaerobic zone is expected to degrade PCP to less chlorinated phenols. These phenols migrate into the first aerobic zone, where more vanilla and oxygen (the electron acceptor) are mixed to promote further degradation. The phenols then move into a second aerobic zone, where more vanilla and oxygen are added to cause ring cleavage and to break down the phenols to chloride and carbon dioxide. Mineralization studies of the combined anaerobic/aerobic process in the laboratory showed a 60 to 70 percent conversion rate of PCP to carbon dioxide.
Although phenol would be good electron donor for this system, regulators and the public would be wary of a system that injects phenol into the ground water. Dr. Wood's group instead used imitation vanilla flavoring as an electron donor. Vanilla flavoring contains the necessary phenolic compounds but is acceptable to regulators and the public because it is a food product.
The field test is underway at the LD McFarland wood preserving facility in Eugene, Oregon. PCP concentrations in ground water are 0 to 3 ppm. The aquifer is shallow and unconfined, ground-water velocity is 1 to 2 feet/day, ground-water temperature is 15C, and pH is neutral. Recovery wells at the site influence local ground-water flow. The test system is in a 34-inch diameter well; the well was placed close to the recovery wells to ensure the proper flow direction. The system is contained in a 3-foot basket and is lowered to a screened section of the well casing. The basket includes sampling points in the reaction zones as well as a continuous sample loop to measure pH and eH. The reaction zones were inoculated with a microbial consortium from activated sludge and anaerobic digester sludge.
The field test has been underway for 6 months and has been through three phases. In the first phase, background data were collected without the addition of electron donors or acceptors. In the second phase, system operation was begun with addition of vanilla and oxygen. During this phase, the researchers did not see the expected reductive dechlorination, and it appeared that the reaction zones were mixing rather occurring in series. Dr. Woods said this mixing may have been caused by the vigorous addition of oxygen to the system every 15 minutes. In the third phase, the researchers limited the oxygen addition to shorter bursts to promote anaerobic conditions and prevent mixing of the reaction zones. PCP levels vary seasonally at the site, so to measure system performance the researchers must study the metabolites as well as the influent/effluent PCP concentrations. Dr. Woods' group has seen increased reductive dechlorination of PCP to tetrachlorophenol and other less-chlorinated phenols since starting the third phase.
Dr. Woods said that, due to the limited oxygen addition, the system is primarily an aerobic reactor at this time. In the next test phase, the group will study the system under aerobic conditions by adding more oxygen. In the subsequent phase, Dr. Woods hopes to return the system to combined anaerobic/aerobic conditions by optimizing the conditions in each reaction zone. Dr. Wood's group will study the system for simultaneous removal of PCP and polycylcic aromatic hydrocarbons (PAHs) that are also present in ground water at the site. The ultimate goal is to develop this system as PRB for waste mixtures.
A Pilot-Scale Test of a Surfactant-Modified Zeolite Permeable Barrier
Dr. Robert Bowman, New Mexico Tech
Rick Johnson, OGI
Dr. Bowman and Mr. Johnson presented OGI's Large Experimental Aquifer Program (LEAP), which is underway nearby in Beaverton, Oregon. Dr. Bowman began by describing the concept of a surfactant-modified zeolite PRB. Zeolites are microporous crystalline solids that have large internal and external surface areas and high internal and external cation exchange capacity and can be ground to any particle size or permeability. The zeolite used in this pilot-scale test is a natural material that is mined in central New Mexico. Raw zeolite has no affinity for anions or organics, but these properties can be added through treatment with a surfactant. The surfactant used in this project is HDTMA, an amphiphilic surfactant also used in mouthwash. In aqueous solution, HDTMA associates in micelles due to a hydrophobic effect. When added to zeolites, HDTMA sorbs to the negatively charged zeolite surface and forms a bilayer of HDTMA molecules similar to a lipid bilayer. HDTMA reverses the charge of the zeolite surface, creating an affinity for anions, especially oxyanions. The positively charged outer surface of the HDTMA bilayer also has a strong affinity for anions and oxyanions. The internal surface of the zeolite retains it cation affinity, however, and the surfactant does not reduce this affinity for most cation contaminants. The hydrophobic, hydrocarbon interior of the HDTMA bilayer creates a partitioning layer for sorption of organic molecules like PCE. Dr. Bowman said that the sorption properties of PCE and other cVOCs on surfactant-treated zeolites is a function of the contaminant's molecular weight; high molecular weights are more easily sorbed. The surfactant-modified zeolite is an excellent reactive media for environmental remediation because it can simultaneously remove cations, anions, and nonpolar organics from solution.
In preparing for the pilot-scale test, the researchers scaled up production of the surfactant-modified zeolite at the mine site in New Mexico. The process produced 3 tons of reactive material per hour, at a cost of $400/ton. HDTMA was the largest component of this cost. Dr. Bowman said that limited studies of cheaper surfactants have not shown as much promise as HDTMA.
Objectives of the pilot test are to scale up to a PRB-sized installation, test the ability of surfactant-modified zeolite to retard a mixed chromate/PCE plume, and test methods to regenerate the adsorptive capacity of the zeolite as it decreases. Mr. Johnson said the concept of a "box" or tank was developed to allow three-dimensional modeling of an actual PRB. The pilot-test PRB is constructed in a steel/concrete tank that its 8.5 m by 8.5 m and 3 m deep and is lined with geomembrane and HDPE. The tank is filled with sand to simulate an aquifer and has a water injection/extraction system to ensure uniform water flow through the aquifer. The zeolite PRB installed within this aquifer is contained in a row of three removable, hanging steel frames that allow side-by-side testing of materials. The PRB extends 6 meters across the tank, is 1 meter thick, and 2 meters deep. These measurements allow water in the tank to flow beneath and around the barrier for three-dimensional study. The tank contains an extensive sampling array, with 81 five-level monitoring wells located in and around the barrier, for a total of 405 sampling points. The tank was constructed to be reused in additional tests of different materials and contaminants. All contaminant water is recirculated through a series of tanks, and clean water is treated before disposal in the sewer.
Dr. Bowman said that to ensure water flow through the barrier, the zeolite PRB was designed to have a hydraulic conductivity ten time higher than that of the surrounding aquifer. Field tests show, however, that the hydraulic conductivity of the PRB is in fact only equal to that of the surrounding sand.
The contaminant plume of chromate and PCE was injected into the tank in December 1997. After 5 weeks, monitoring data showed that due to the low conductivity of the barrier, the plume was deflecting around and underneath the barrier as well as passing through it. This made it impossible to measure the contaminants in the water that flowed through the barrier. The researchers stopped the contaminant injection and took several steps to improve the hydraulic conductivity of the barrier. Tracer tests indicate that these efforts have been successful in achieving 80 to 90 percent of the flow called for in the original design. The researchers will perform a final tracer test before resuming the contaminant injection.
Field Site Visit to OGI's LEAP Tanks
Approximately 40 meeting attendees visited the LEAP tanks after the close of the meeting. They received a full tour of the site led by Dr. Bowman and Mr. Johnson.
Attachment A
Final Speaker List
RTDF Permeable Reactive
Barriers Action Team Meeting
Portland Community College
Washington County Workforce Training Center
CAPITAL Center
Beaverton, Oregon
April 1516, 1998
Final Speaker List
| David Blowes Professor Department of Earth Sciences University of Waterloo 200 University Avenue, W Waterloo, Ontario N2l 3G1 Canada 519-888-4567, Ext: 5643 Fax: 519-746-3882 Robert Bowman Dawn Carroll John Fruchter Baohua Gu Grant Hocking Rick Johnson Stephen Koenigsberg Ed Marchand Donald Marcus Stan Morrison |
David Naftz Geochemist U.S. Geological Survey 1745 West 1700, S Salt Lake City, UT 84104 801-975-3389 Fax: 801-975-3424 E-mail: dlnaftz@usgs.gov Robert Puls Charles Reeter James Romer H.G. Sanjay Bruce Sass Chunming Su Paul Tratnyek John Vidumsky Sandra Woods |
Attachment B
Final Attendee List
RTDF Permeable Reactive Barriers Action Team Meeting
Portland Community College
Washington County Workforce Training Center
CAPITAL Center
Beaverton, Oregon
April 1516, 1998
Final Attendee List
| J. Frank Abshier Environmental Engineer Olin Corporation Lake City Army Ammunition Plant Highways 7 and 78 - P.O. Box 250 Independence, MO 64051-0250 816-796-5195 Fax: 816-796-5197 David Anderson James Anderson Bruce Barley-Heine Heidi Blischke Randall Boese Jeff Breckenridge Jed Costanza Harry Craig Dave Emilia John Foxwell Bob Gillham Will Goldberg Dibakar (Dib) Goswami Roger Gresh Baohua Gu Kevin Hass Thomas Holdsworth Stephen Hutchins Tim Johnson Mark Kadnuck John Kenneke Larry Kimmel K. Scott King Richard Landis Ed LaRock Jim Leuhart Lindsey Lien Gus Lo Jeff Lockwood Jay MacPherson Leah Matheson Theodore Meiggs Joel Miller Andrew Miller Sheila Monroe Rich Muza Pradeep Naik |
Michael Nimmons Principal Engineer ICF Kaiser Engineers, Inc. 1191 Second Avenue - Suite 1200 Seattle, WA 98101-2933 206-521-5941 Fax: 206-521-5911 E-mail: nimmons@nwlink.com Deirdre O'Dwyer Stephanie O'Hannesin Dennis O'Sullivan Gregory Oberley Eric Obrecht Mart Oostrom Carl Palmer Kevin Parrett Edward Peterson Robert Powell Michelle Scherer Stephen Schmelling Toby Scott Ken Shump Stephen Smith Daniel Sogorka Richard Steimle Peter Strauss Ted Streckfuss Ameer Tavakoli Patty Toccalino Diane Van Schoten John Vogan Daman Walia Ted Wall Scott Warner David Watson Stephen White Bryan Wigginton John Wilkens Mark Williams Randy Wolf RTDF logistical and technical support provided by: Colin Devonshire Susan Brager Murphy Carolyn Perroni Laurie Stamatatos |