SUMMARY OF THE REMEDIATION TECHNOLOGIES DEVELOPMENT FORUM
PERMEABLE REACTIVE BARRIERS ACTION TEAM MEETING

Hilton Melbourne Airport
Melbourne, Florida
February 16-17, 2000

INTRODUCTION AND WELCOME
John Vidumsky, DuPont

John Vidumsky, co-chair of the Remediation Technologies Development Forum's (RTDF's) Permeable Reactive Barriers (PRB) Action Team, opened the meeting by welcoming the meeting's attendees (see Attachment A) and speakers (see Attachment B). He also provided an overview of the meeting's scheduled events, noting that the following were planned: presentations and discussion sessions, a poster session and reception, and a field trip to NASA's Cape Canaveral Air Station.¹ The meeting, he said, was being held to share information about (1) hydraulic issues that impact PRB system performance, and (2) lessons that have been learned during PRB installation activities. Before turning the meeting over to invited speakers, Vidumsky reported on the Action Team's current activities, the status of PRB technologies, and trends that he expects to see.

Vidumsky said that some PRB Action Team members have been intimately involved with developing and executing a PRB training course. This course, which was developed by several organizations, is currently touring the country. It has already been presented in seven locations, and is scheduled for delivery in California, New York, Colorado, Illinois, and Missouri before the end of September 2000. To date, Vidumsky said, more than 1,000 regulators have participated in the training sessions; people from other sectors have also attended, but in smaller numbers.

Vidumsky said that there are currently about 31 full-scale PRBs that are using iron as a reactive material. Of these, about 20 are continuous reactive walls, 8 are funnel-and-gate systems, and 3 are in situ reactive vessels. A wide variety of construction methods have been used to install these operational systems, Vidumsky said; some methods are more popular than others. For example, he continued, while trenching machines and cofferdams have been used to install continuous trenches at several sites, hydrofracturing and jetting have only been used at a few. In the future, he predicted, these more innovative techniques will probably be used more widely. Vidumsky said that performance data have been collected for several full-scale systems. Results indicate that volatile organic compounds (VOCs) have been degrading consistently for several years. The results also indicate that microbial fouling is not a significant problem. In addition, the data show that precipitants are forming on reactive materials and that most carbonate precipitants are present at the upgradient interface of reactive materials. Even though evidence suggests that the accumulation of precipitants causes porosity/permeability changes over time, Vidumsky said, no significant hydraulic plugging has been reported and no sites have required "retrofitting" or iron replacement due to losses in reactivity. However, he said, some sites have required upgrades or extensions; changes in design had to be made as new information about hydraulic conditions became apparent. Although the data that have been collected so far are promising, Vidumsky said, additional years must pass before firm conclusions can be made about the long-term performance of PRBs.

Vidumsky showed a chart that indicated the number of pilot and full-scale PRBs installed between 1995 and 1999. He pointed out a decrease in the number of installations between 1998 and 1999, but said that he fully expects the technology's popularity to rise as more information is disseminated about its performance and cost-effectiveness. In fact, Vidumsky said, he thinks that the technology may find a wide variety of applications over the next few years. He expects to find PRBs being used in deeper, more challenging situations in the future. Also, researchers are currently evaluating the technology's potential in a variety of treatment scenarios. For example, he said, some investigators are evaluating the technology's potential as a source control remedy; others have expressed interest in combining the PRB technology with natural attenuation remedies.

REVIEW OF THE FIELD PERFORMANCE OF PERMEABLE BARRIERS AT MULTIPLE DEPARTMENT OF DEFENSE SITES
Arun Gavaskar, Battelle Memorial Institute

Arun Gavaskar said that the Department of Defense (DOD), the Department of Energy (DOE), and the U.S. Environmental Protection Agency (EPA) are collaborating to evaluate the long-term performance of PRBs. Each of the participating agencies, he said, is working independently to evaluate a specific set of sites, but efforts are being made to communicate regularly so that similar evaluation protocols are used across the agencies. Under this project, he said, DOD, DOE, and EPA will work together to create three joint products: a Web site, a one-page brochure, and a final PRB performance evaluation guidance report. The latter, he said, will be completed at the end of the project. (The project was initiated in 1999 and will be completed in 2001.) Gavaskar said that the study is being funded by the Strategic Environmental Research and Development Program (SERDP) and the Environmental Security Technology Certification Program, but that several other organizations (e.g., the Air Force Research Laboratory, the Air Force Center for Environmental Excellence, the U.S. Army Corps of Engineers, and the Interstate Regulatory Cooperation Working Group) are also providing support. Gavaskar said that his organization, Battelle Memorial Institute, is working as a partner to DOD, and is performing field investigations. He said that the remainder of his presentation would focus on information that has been collected on DOD sites.

Gavaskar said that DOD has identified the following as its primary objective for this project: identify and evaluate the factors that affect the hydraulic and longevity (geochemical) performance of PRBs at multiple DOD sites. DOD hopes that collecting information on these factors will make it easier to determine:

• Whether PRBs are achieving desired hydraulic performance (capture zone and residence time). Gavaskar said that desired hydraulic performance can be difficult to achieve and to verify in the field due to aquifer/plume heterogeneities, site characterization limitations, and seasonal variability.

• How long reactive materials maintain their reactivity in the subsurface. Gavaskar said that it is difficult to predict the long-term geochemical performance of PRBs based on available tools, such as ground-water monitoring, coring, and modeling. If a prediction about longevity cannot be made, he said, it is difficult to predict remedial costs. (The lifetime cost or savings associated with PRBs is heavily influenced by how frequently, if at all, reactive materials must be replaced.)

To collect the data that DOD requires, Gavaskar said, a two-step plan has been implemented. The first step involves collecting existing field performance data and identifying data gaps. The second step involves going to selected DOD sites and conducting field investigations to supplement ongoing monitoring efforts. Gavaskar said that the data-gathering process has reached the second step at two DOD sites: Lowry Air Force Base (AFB) and Dover AFB. He described the performance evaluation efforts that have been initiated at these sites.

Lowry AFB

Gavaskar said that a funnel-and-gate system was designed and installed at Lowry AFB in 1995. Each funnel is about 10 feet long, he said, and the gate, which is filled with reactive iron filings, is about 10 feet wide and 5 feet thick. At the time the PRB was being designed, Gavaskar said, designers were focusing primarily on achieving required contaminant degradation efficacy. Less emphasis was placed on designing the system so that it achieved a certain capture zone and residence time. Efforts are now being made, however, to better understand the local hydrology, to obtain a clearer picture of ground-water flow patterns and velocities, and to estimate the capture zone and residence time.

Gavaskar said that a hydraulic evaluation was initiated at Lowry AFB between September and November 1999. As part of this effort, he said, new monitoring wells have been installed upgradient of the PRB, water level measurements are being collected, slug tests were performed, and seasonal variations are being tracked. In addition, two types of ground-water velocity sensors--HydroTechnics^™ sensors and colloidal borescopes--have been implemented to measure hydraulic parameters. The latter, developed by DOE, tracks the movement of colloids using a mini-camera. As the probe passes through a well, Gavaskar said, the colloids are tracked on a computer screen and their velocities are averaged to produce a velocity vector. Gavaskar identified the benefits and limitations associated with the instrument. The colloidal borescope can be used in multiple wells; in fact, it was used to measure velocities in about eight wells at the Lowry AFB site. However, the borescope only works in wells that have stable flow--a potential limitation.

Gavaskar summarized some of the data that have been collected at Lowry AFB since the hydraulic evaluation was initiated. Preliminary results indicate that (1) ground water on the western funnel wall appears to be moving toward the gate, whereas part of the flow along the eastern funnel wall is drawn towards a nearby stream, (2) gradients are relatively flat in the gate and on the PRB's upgradient side, (3) velocity vectors, as measured by the colloidal borescopes, generally point toward the center of the barrier, and (4) velocity vectors, as measured by the HydroTechnics^™ sensor, provide conflicting information--one location points toward the gate and the other points away (investigators are still trying to interpret these results).

Gavaskar said that a geochemical evaluation has been initiated at the Lowry AFB. As part of this evaluation, he said, ground-water samples have been collected and are being analyzed for inorganics. Also, iron cores--collected from the PRB's reactive material zone--are being analyzed for carbonates, silica, ferrous compounds, and other precipitates. (Gavaskar said that silica is being evaluated because it has been found deposited on iron particles at other sites and in column tests.) The results of the geochemical analyses are not yet available.

Dover AFB

Gavaskar said that a funnel-and-gate system was installed at Dover AFB in December 1997, as part of a study that is funded by SERDP through the Air Force Research Laboratory. He gave a brief description of the site's geology: it is fairly heterogeneous, it has a clay-like layer about 15 feet below the ground surface, and the ground-water aquifer starts about 40 feet below the surface. He also gave a brief description of the PRB's major design components. He said that the system has two caisson gates, each of which contains 100% iron in its reactive cell, and that pretreatment zones are located in front of each cell. One gate has a 10% iron/sand material in its pretreatment zone; the other has a 10% pyrite/sand mixture. Gavaskar said that the pretreatment zones have been established to remove high levels of dissolved oxygen from ground water. In addition, he said, researchers thought that pyrite might carry an added benefit: helping to achieve some degree of pH control (and therefore precipitation control) within the PRB system. Pyrite lowers pH values. Thus, he said, designers hoped to minimize pH increases within the reactive cell by lowering the pH of ground water before it entered this zone.

Evaluations have been initiated to assess the performance of the funnel-and-gate system at the Dover AFB. Gavaskar said that the following conclusions have been made:

• Caissons are useful when installing reactive material in relatively deep aquifers (down to about 50 feet) or between several underground utility lines.

• Dissolved oxygen can be removed from ground water before it reaches the 100% iron reactive zone by passing it through a pretreatment zone consisting of 10% iron/sand or 10% pyrite/sand. It does not appear, however, that using pyrite in the pretreatment zone helps achieve any substantial pH control within the reactive iron cell.

• The hydraulic capture zone is asymmetrical around the gates. Based on available results, researchers suspect that (1) flow directions change on a seasonal basis, perhaps by as much as 30 degrees, and (2) flow is not perpendicular to the PRB at all times. (Gavaskar said that additional information about the site's hydrology was obtained from water level measurements, two HydroTechnics^™ sensors, and from a colloidal borescope. Learning about the site's hydrology has been fairly challenging, he said, noting that measurements were difficult to make because gradients are relatively flat. Also, the colloidal borescope sometimes points the flow in contradictory directions; efforts are currently underway to evaluate the meaning of these data.)

• Ground-water contaminants are below maximum contaminant levels (MCLs) on the downgradient side of the iron.

Future Evaluation Activities

Gavaskar said that DOD will continue to assess the long-term performance of PRBs over the next two years. During this time, additional efforts will be made to evaluate:

• Geochemical properties. Gavaskar said that geochemical evaluations will continue at Lowry AFB. In addition, (1) iron cores may be collected from the former NAS Moffett Field, (2) long-term column studies will be initiated with ground water from up to two sites, and (3) geochemical models will be further evaluated. Despite all their well-known limitations, Gavaskar said, long-term column studies may provide information about the expected life of reactive media. Although inorganic analysis of ground water, iron coring, and geochemical modeling can be used to identify the types of precipitates that will form, translating the results from these methods to a prediction of PRB life expectancy has not been successful so far. (The expected life of the reactive medium has a strong bearing on the economics of a PRB application.) Gavaskar said that results will be coordinated with those collected from iron mechanism studies that are being conducted by John Hopkins University.

• Hydraulic performance. Gavaskar said that hydraulic evaluations will continue at Lowry AFB and Dover AFB. Another evaluation may be initiated at a DOD site (e.g., the Watervliet Arsenal) that has a continuous reactive barrier. After conducting evaluations at several sites, Gavaskar said, DOD hopes to make some recommendations about which monitoring/modeling methods may be most useful. Gavaskar acknowledged that it is difficult to obtain accurate estimates for hydraulic parameters (e.g., hydraulic conductivity and ground-water velocity). In fact, he said, estimates often vary by a factor of five or ten. This lack of accuracy makes it difficult for designers to optimize the efficacy of PRBs. In addition, he said, seasonal variations and uncertainties in flow direction make it difficult to determine how to orient PRBs so that they will be to perpendicular to flow. He said that there are two ways that designers can work around these difficulties to optimize field design: (1) model different scenarios that cover a full range of hydraulic properties and flow directions, and (2) incorporate adequate safety factors in a PRB's designed thickness and width.

LESSONS LEARNED FROM OPERATION OF THE Y-12 REACTIVE BARRIERS
David Watson, Oak Ridge National Laboratory

David Watson said that his presentation would focus on PRBs that have been installed at DOE's Y-12 Plant. He began his presentation with a brief overview of the site. Between 1951 and 1983, he said, wastes containing technetium, VOCs, and metals; nitric acid; and uranium-laden materials were shuttled from the Y-12 Plant to the unlined S-3 pond disposal area. When the pond was removed from service, he said, investigators found that it was highly contaminated and very corrosive, registering at a pH lower than 2.0. Watson said that the pond was neutralized in 1984 and covered with a Resource Conservation and Recovery Act (RCRA) cap in 1988. It is now covered by pavement and serves as a parking lot. Contaminants from the pond migrated through the subsurface and along three major contaminant migration pathways: Pathways 1 and 2, which are shallow, and Pathway 3, which is much deeper and located within shale and bedrock layers.

Watson said that PRBs filled with zero valent iron (ZVI) are being used to address Pathways 1 and 2. He said that this technology was chosen because it requires low maintenance, is considered more cost-effective than pump-and-treat technologies, and is considered an effective treatment for the contaminants that underlie the Y-12 site. Expanding on the latter point, Watson explained why remedial managers believe that PRBs will be able to remediate the site's mixed contaminants (e.g., uranium, technetium, heavy metals, tetrachloroethylene [PCE], and nitrate). As a plume travels into a ZVI zone, he said, it enters an area of low dissolved oxygen, low oxidation-reduction potential, and elevated pH. These conditions favor the destruction of organics (through dehalogenation processes) and the removal of metals and radioactive contaminants (through redox-driven precipitation processes and sorption to iron byproducts).

Watson said that the PRBs were installed in November and December 1997 and that efforts are underway to assess their long-term performance. He said that investigators are evaluating (1) the effectiveness of hydraulic capture, (2) the severity of system clogging, (3) changes in the barrier's or surrounding soils' physical, chemical, and/or biological properties, (4) the long-term disposition of treatment media, (5) the potential for remobilization of contaminants, and (6) the impact of byproducts.

In the remaining portion of his presentation, Watson provided additional details about each of the barriers used at Y-12 and summarized available performance results.

Y-12 Plant; S-3 Disposal Ponds: Pathway 1

Watson said that Pathway 1 extends from the water table (located about 10 feet below the surface) to the shale bedrock (encountered at a depth of about 25 feet below the ground). The ground water at Pathway 1 is acidic (pH of 5.5) and contaminated with high concentrations of nitrate (1,400 milligrams per liter [mg/L]), uranium (2.6 mg/L), and technetium (10,000 picocuries per liter). A funnel-and-gate system has been installed to address this contamination. The system is designed to capture ground water in a gravel-filled, high-density-polyethylene (HDPE)-lined trench, and to treat it within a vault that contains ZVI.

Although the system was installed in 1997, Watson said, it is not yet meeting its full potential. Ground water is not flowing as anticipated, so the PRB is exhibiting suboptimal capture rates. Watson said that this disappointing outcome might have been avoided if more thorough site characterization had been conducted. He said that the system was designed based on inaccurate inputs: the heads in certain parts of the Pathway 1 site are higher than modelers thought they would be. Researchers are trying to determine if site activities have unexpectedly introduced a recharge zone or have impacted vertical gradients.

Watson noted another problem: ferrous iron is being discharged from the PRB at concentrations as high as 100 parts per million. He said that this discharge may be unacceptable to some regulators because it is above secondary drinking water standards and can cause a red discoloration if it reaches surface water.

Y-12 Plant; S-3 Disposal Ponds: Pathway 2

Watson said that Pathway 2 contains the same contaminants as Pathway 1, but at significantly lower concentrations. A permeable reactive trench has been installed to address these contaminants. The trench is about 2 feet wide and about 225 feet long. It consists of a 26-foot-long zone of ZVI flanked by zones of gravel backfill.² The trench, which has been installed into a saprolite and silt area and extends to the shale/bedrock layer, ranges from about 22 to 30 feet deep.

Watson described some of the monitoring activities that have been conducted to assess PRB performance at the Pathway 2 site. He said that water levels have been measured, wells have been sampled for contaminants, and cores have been collected. One meeting attendee asked him to describe the method used to collect cores, and to indicate whether any recovery problems were encountered. Watson said that a simple push probe was used. No recovery problems arose when investigators drove through native soils and sand/gravel layers, he said, but recovery was a bit more difficult within the reactive ZVI zone.

To date, Watson said, the monitoring efforts have revealed the following:

• ZVI effectively removes uranium, technetium-99, and nitrate from ground water. Watson presented tables that showed contaminant concentrations over time across a series of wells. He said that contaminant concentrations within and downgradient of the reactive barrier are much lower than those found in upgradient areas. Also, he said, some evidence indicates that contaminants are degraded very rapidly once they enter the ZVI zone. (Core samples were collected at an angle so that researchers could see how conditions varied between upgradient and downgradient portions of the barrier. Uranium concentrations were high in the upgradient portions of the barrier, but dropped off rapidly with increasing distance into the barrier.)

• Material precipitates out as ground water passes through the reactive zone. Watson said that data indicate that specific conductance is lower in downgradient portions of the barrier than in upgradient portions. Specific conductance is related to total-dissolved-solid quantities, he said; thus, this finding suggests that minerals are precipitating out of the ground water as it passes through the barrier.

• Minerals are precipitating onto the ZVI. Watson said that investigators subjected core samples to scanning electron microscopy, x-ray diffraction, and energy dispersive x-ray analysis to obtain information about the types of precipitants that have been deposited onto the ZVI. Results indicated that iron oxides (goethite, akagneite, amorphous) are the most predominant precipitants and that these are fairly well-distributed throughout the reactive zone. In addition, Watson said, some iron carbonate (siderite), calcium carbonate (aragonite), and iron sulfide has been detected. The latter, he said, is thought to form as a byproduct of sulfate-reducing bacteria activity. The spatial distribution of the iron sulfide supports this conclusion. (Very little iron sulfide is found in the deep portions of the barriers, where the presence of nitrate prevents sulfate-reducing bacteria from dominating. In the more shallow portions of the barrier, where microbial activity is much more pronounced, more iron sulfide is present.)

• Some clogging has occurred, but only in a small area. For the most part, Watson said, core samples from the barrier have proven to be loosely packed and highly permeable. Thus, clogging does not appear to be a problem across the majority of the barrier. However, Watson continued, cementation has been observed in one very small area in the reactive zone's upper regions. This material, which is very hard and difficult to break apart, is surrounded by a thin coat of iron sulfide. Watson said that the permeability of the cemented material has not been calculated yet.

• Clogging does not appear to be impeding ground-water flow. Watson said that local gradients do not appear to be impacted by cementation within the reactive barrier. If any plugging were occurring, water levels would rise in upgradient wells. Such increases have not been observed over the long term. (All detected increases in upgradient wells have been transitory, apparently related to seasonal trends or storms.)

• The estimated lifespan of the PRB ranges from 5 years to more than 30 years. Watson said that investigators calculated the expected lifespan of reactive materials based on rates of iron corrosion. The lifespan was estimated to be more than 30 years for the non-cemented zones of the reactive barrier. The lifespan of the cemented portions, however, was only estimated to be about five years.

• Guar gum can increase biological activity. Watson said that guar gum was used to keep the trench open during PRB installation. This material, he said, has a high carbon content and serves as an ideal food source for microbes. Several lines of evidence suggest that the guar gum did promote microbial activity at the Pathway 2 site. For example, Watson said, phospholipid fatty acid content was higher within the barrier than it was upgradient of the barrier. Also, colloidal borescopes photographed protozoa within the barrier wall. Watson said that he is not sure how increased biological activity will affect the PRB. He did caution, however, that such activity can alter geochemical conditions, lead to biofouling, and promote the formation of iron sulfide.

• Ferrous iron is being discharged. Watson said that ferrous iron was being discharged from the barrier at high concentrations when the PRB system first became operational. Over time, discharge concentrations have decreased significantly; they are now very low. Watson reminded the audience that ferrous iron is also being discharged at the Pathway 1 site. At this site, however, discharge concentrations have not decreased to low levels over time. He said that this may be because the ground water at Pathway 1 is so corrosive.

• More ground water flows through the barrier in the spring than in the fall. In the fall, Watson said, gradients are fairly flat. In the spring, they are a little steeper and more flow is directed toward the PRB.

Watson said that much remains to be learned about the PRB at Pathway 2. In the near future, he said, investigators plan to evaluate whether contaminants are being remobilized as colloids.

THE IMPACT OF HYDRAULIC HETEROGENEITIES ON FLOW THROUGH BARRIERS
David Smyth, University of Waterloo

David Smyth said that his presentation would discuss how hydraulic heterogeneity affects the way that water flows through PRBs. He said that he would talk about what has been learned at three field sites: (1) a site near Sudbury, Ontario, (2) the Elizabeth City site, and (3) the CFB Borden site. Smyth said that he would also talk about how computer models can be used to help predict flow patterns. He said that Shawn Benner, David Blowes, and Uli Mayer contributed to the presentation that he was giving.

A Site Near Sudbury, Ontario³

Smyth described a site near Sudbury, Ontario, that was formerly used for mining. At this site, contaminants seeped from tailings, infiltrated the subsurface, and contaminated a bedrock-bounded shallow aquifer. Ground water flows through the aquifer at a rate of about 10 to 20 meters per year and discharges to Moose Lake. Sulfate and iron concentrations in the plume are high; sulfate has been detected at about 2,500-5,000 mg/L and iron has been detected at 250-1,600 mg/L.

In 1995, Smyth said, a sulfate-reduction barrier was installed in the bedrock layer to treat the contaminated plume. The barrier, which was placed in the middle of the plume, treats water flowing into the wall, but the contaminated ground water that is downgradient of the wall will continue to discharge into Moose Lake for some time. The barrier, which is about 15 meters long, 4.6 meters thick, and 3.6 meters deep, uses an organic carbon mixture as its reactive material. This mixture consists of 40% municipal compost, 40% leaf and plant compost, and 20% wood chips. After the mixture was prepared, the materials were mixed with pea gravel, and placed in the ground. Monitoring suggests that sulfate reduction and metal sulfide precipitation is occurring within the wall, and that the acid-producing potential of the plume water has been significantly reduced.

Smyth said that the compost had high quantities of chloride, which served as a temporary tracer. Immediately following installation, investigators started tracking the chloride. As time progressed, the chloride was slowly flushed out of the barrier and transported downgradient. By studying its distribution patterns, environmental investigators were able to determine how ground water was flowing through the barrier; results indicate that there is differential transport. Chloride concentrations were not only measured in the field: they were also modeled using a transport code, then used to estimate hydraulic conductivity. Analysis of the measured and modeled data indicated the following: (1) downgradient of the barrier, flux is higher at the base of the aquifer than in its upper portions; and (2) within the barrier, flux is highest in the central portions and lower near the barrier's top and bottom. Expanding on the latter point, Smyth said that there are a couple of possible explanations for the hydraulic heterogeneity within the barrier: (1) the organic carbon mixture/pea gravel materials may not have been well mixed, and (2) air may have been trapped during installation or generated after installation. Heterogeneous flow through the barrier, Smyth said, causes treatment rates to differ across different parts of the wall. Smyth said that the treatment rates in the high flow zones will determine how long the barrier will be able to function without needing to have its reactive materials replaced.

The Elizabeth City Site⁴

Smyth said that a ZVI barrier, about 0.6 meters thick, has been installed at the Elizabeth City site to remediate chlorinated solvents and chromium (VI). Researchers at the site modeled chloride concentrations using field hydraulic conductivity data. They learned that there are zones of high hydraulic conductivity within the aquifer and within the barrier. Smyth presented some data showing trichloroethylene (TCE), vinyl chloride, and chromium (VI) concentrations in areas upgradient of, within, and downgradient of the barrier. The results suggest, he said, that heterogeneous flow fields influence barrier treatment. He said that TCE and vinyl chloride distributions near and within the barrier appear to be controlled by zones of high hydraulic conductivity in the aquifer and barrier.

The CFB Borden Site⁵

Smyth said that an experimental site at the CFB Borden site has been contaminated with polyaromatic hydrocarbons. For example, naphthalene has been detected at concentrations up to about 2,000 parts per billion (ppb). A funnel-and-gate treatment system has been installed to remediate the contaminants, about 30 meters downgradient of the contaminant source. This system has a Waterloo™ funnel and a treatment gate that consists of four media-filled treatment cassettes. Each of the cassettes is separated by an open chamber of water, from which samples can be collected. Monitoring has shown that the plume at this site has a discrete top and bottom boundary within the treatment gate, even with the open water-filled chambers between the treatment cassettes.

Modeling Simulations

Smyth said that model simulations can provide a great deal of information about flow patterns in the vicinity of PRBs. He presented results from some simulations--performed by Benner--that used a suspended barrier as a model and were conducted using 2-dimensional models. (In the future, Smyth said, Benner hopes to use 3-dimensional models.) The following table summarizes what the modeling exercises taught Smyth about the impact that design parameters have on flow patterns.

Scenario	Parameter Modified	Impact on Flow Patterns through the Barrier
Homogeneous aquifer	Hydraulic conductivity (K) of the barrier	Less convergence of flow as Kbarrier approaches that of Kaquifer
• Heterogeneous flow system (high K layer) • Flux conditions when Kbarrier approaches that of high K layer	Kbarrier	Convergence of flow does not increase significantly as Kbarrier increases more than an order of magnitude larger than that of the high K layer
• Homogeneous aquifer • Kbarrier = 10 Kaquifer	Barrier thickness	• Thicker barrier exhibits greater convergence of flow • Thin barrier has less impact on flow systems • There is higher flux along the edges of thick barriers
• Heterogeneous aquifer • Kbarrier equivalent to highest Kaquifer	Barrier thickness	Thicker barrier redistributes flow, but thin barrier does not

Smyth also talked about what modeling results have indicated about:

• Mixing zones and their impact on water flow. Simulations were performed to determine how mixing zones impact flow though a PRB. Results showed that mixing may not be adequately achieved by including just one upgradient mixing zone. In fact, the best mixing appears to be achieved when mixing zones are established upgradient and downgradient of the barrier.

• How heterogeneous zones within barriers impact flow. In a thick barrier that exhibits variability in hydraulic properties, high hydraulic conductivity zones may have significant effect on the flow-through and the treatment achieved by the barrier.

Summary of Presentation

Before closing his presentation, Smyth summarized what he has learned from field studies and model simulations:

• Hydraulic heterogeneities in aquifer and barrier systems influence the performance of PRBs.

• Thin barriers are sensitive to the flux characteristics of an aquifer system.

• Thicker barriers are likely to behave better in heterogeneous aquifer systems, but homogeneous barrier characteristics are also important.

• It is important to know the relationship between flow and contaminant flux distributions.

THE DANISH REACTIVE BARRIER DEMONSTRATION PROGRAMME: HYDRAULIC PERFORMANCE OF THREE FULL-SCALE REACTIVE BARRIERS
Peter Kjeldsen, Technical University of Denmark

Peter Kjeldsen said that interest in innovative technologies is growing in Denmark. In fact, he said, in 1997, the Danish Environmental Protection Agency initiated an effort that promotes field testing of such technologies. PRBs have generated much interest among researchers, Kjeldsen said; thus, the Danish PRB Demonstration Programme has been created. Kjeldsen, who serves as the program's scientific coordinator, said that the program's objective is to test, dimension, and evaluate full-scale barriers, and to offer recommendations about future usage of the PRB technology. To date, he said, five full-scale projects have been installed in Denmark. All of the systems use ZVI as a reactive medium, he said, but they differ in design. Three of them are classified as full-scale barrier systems and two use canister systems. Kjeldsen focused his talk on the three full-scale barrier systems.

The Freight Yard Site

Kjeldsen provided background on the Freight Yard site, a train repair yard that is underlain by a shallow aquifer and contaminated with chlorinated aliphatics. TCE, which was used as a degreasing agent, has contaminated the site, but natural attenuation has converted much of this to dichloroethylene (DCE). Concentrations of cis-1,2-DCE and trans-1,2-DCE have been recorded at concentrations as high as 3,000 ppb and 700 ppb, respectively.

In 1998, Kjeldsen said, a PRB was installed to address the site's contaminated plume. The barrier is about 15 meters long, 2.5 meters wide, and 0.9 meters thick. Reactive materials are covered, he said, by a geotextile, a compacted clay layer, a gravel layer, and pavement. Construction activities, which were completed between May and July 1998, were straightforward: sheetpiling was installed, the area was excavated, and the trench was backfilled with ZVI.

Kjeldsen said that efforts have been initiated to monitor the performance of the PRB. Toward this end, ground-water flow fields have been evaluated, slug tests have been performed, and ground-water samples have been analyzed for inorganics and chlorinated solvents. Results indicate that contaminant concentrations decrease by about 95% as ground water passes through the barrier. (Influent concentrations have been detected at concentrations as high as 1,000 ppb. Effluent concentrations are less than 50 ppb.)

Kjeldsen said that performance data also suggest that one-fifth of the plume is migrating around the barrier and escaping the PRB's capture zone. This conclusion has been made based on an analysis of ground-water level measurements and contaminant concentration distributions. (Kjeldsen said that ground-water level measurements are collected quarterly. These are used to create flow maps. These maps show that ground water is starting to flow around the site's PRB--a finding that was confirmed when investigators found high contaminant concentrations in wells beside the barrier.) Kjeldsen said that researchers have generated a couple of theories to explain why the water is flowing around the barrier. First, some researchers suspect that the PRB is simply too narrow. Second, some researchers suspect that changes in the barrier's permeability could be causing the bypass. Kjeldsen said that data (accumulated via slug test) suggest that the barrier has become less permeable over time. Kjeldsen is trying to determine why such a reduction has occurred. He has thought of two possible reasons:

• The ZVI may be clogging. Several inorganic salts are dissolved in the site's ground water. Evidence suggests that inorganic constituents precipitate out of solution while they are in the barrier. These particles could be clogging the ZVI.

• Hydrogen formation may be influencing water flow. The anaerobic corrosion of iron produces hydrogen gas. By performing laboratory experiments, the University of Waterloo determined the amount of hydrogen that is produced in a day and the amount of void space that is present in a typical PRB. A volume equivalent to 5% of the PRB's void space is produced daily at the Freight Yard site. Kjeldsen wonders whether filling a portion of the barrier's void space with gas could impact the barrier's overall permeability. In most cases, gas buildup would probably not be a problem because it would diffuse out of the barrier. At the Freight Yard site, however, a compacted clay layer lies over the reactive zone; perhaps this layer is trapping the hydrogen gas.

Kjeldsen said that monitoring efforts will continue at the Freight Yard site. Therefore, researchers will continue to perform slug tests and to monitor chlorinated solvents and inorganics. In addition, the iron within the ZVI zone will be sampled and characterized in the near future.

The VAPOKON Site

For many years, Kjeldsen said, the VAPOKON site was used as a solvent recycling factory. Contaminants entered the subsurface through leaky concrete underground storage tanks. The underlying ground water is contaminated with a variety of chlorinated solvents and breakdown products, including PCE, TCE, trichloroethane (TCA), dichloroethane (DCA), DCE, dichloromethane (DCM), and aromatic compounds (benzene, toluene, ethylbenzene, and xylene). Kjeldsen said that very high ground-water velocities (up to 400 to 500 meters per year) have been documented at the site. The site's geology is simple, he continued: a 3.5-meter-thick loamy soil sits above a 10-meter-thick sandy aquifer; all of this lies above a clay layer.

Kjeldsen said that two technologies will be employed to remediate the VAPOKON site plume: a funnel-and-gate system that uses ZVI, and natural attenuation. The latter is required, he said, because studies indicate that some of the plume's contaminants (e.g., DCA and DCM) cannot be degraded by ZVI. Investigators believe that anaerobic biological processes will degrade the ZVI-resistant compounds before ground water discharges to a nearby creek.

Kjeldsen described the funnel-and-gate system's design. It consists of a 120-meter-long funnel (made out of sheetpile) and a 80-cubic-meter reactive zone, and is equipped with 45 screens. Models were used to help create an adequate design. For example, a two-dimensional ground-water flow model was used to evaluate vertical mixing through the system. Kjeldsen said that an upstream drainage system has been incorporated into the design to reduce the amount of ground water that passes through the gate. This intervention was deemed necessary to avoid overwhelming the gates with the site's high ground-water velocities.

Kjeldsen said that construction activities were just recently completed. For the most part, he said, installation was straightforward, although scuba divers had to be called upon to perform underwater welding after some organic sheetpiling seals were destroyed. Monitoring efforts were only just recently initiated, so no performance data are available yet.

The Haardkrom Site

For many years, Kjeldsen said, the Haardkrom site was used as an electroplating facility. Contaminants were released during operations, polluting ground water with TCE and chromate. The latter of these is present at concentrations up to 100 mg/L. Kjeldsen described the site's subsurface: soils are characterized by a low permeability, and a shallow heterogeneous aquifer lies under the site.

Kjeldsen said that a continuous reactive barrier was installed at the Haardkrom site about one year ago. The system consists of a 60-meter-long wall that is between 1 and 3 meters tall and 1 meter thick. When designing the barrier, Kjeldsen explained, investigators made sure to dimension it so that the ZVI zone had enough capacity for all of the plume's chromate. (When dealing with high concentrations of chromate, Kjeldsen said, it is important to remember that the PRB should be dimensioned with a limited chromate reduction capacity since layers of chromate could cover the iron particles.) Also, Kjeldsen said, the barrier's designers tried to enhance water flow in the system to account for the aquifer's heterogeneities. Toward this end, they included bypassing trenches and recirculation pipes in the system's design.

Kjeldsen said that the barrier was not constructed using the same methods that were employed at the Freight Yard site and the VAPOKON site. Rather, an excavation box was used. (This approach was feasible at the Haardkrom site because site soils had a low permeability.) Kjeldsen said that performance monitoring data have been collected for about one year. Results suggest that there is uneven pollutant distribution along the barrier; the design does not appear to be controlling the flow field very effectively. Kjeldsen said that spatial monitoring will be performed in the near future and that the design may be modified to distribute flow more evenly across the PRB.

WALL-AND-CURTAIN FOR SUBSURFACE TREATMENT OF CONTAMINATED GROUND WATER
David R. Lee, Atomic Energy of Canada Ltd.

David Lee said that his presentation would focus on a PRB that has been installed in Chalk River, Ontario, Canada. The system, he said, is being used to mitigate a strontium-90 plume. The leading edge of this plume, which is about 20 feet wide, is located within the deep portions of a 40-foot-thick aquifer. It has migrated about 1,400 feet downgradient of its initial source area and is heading toward a wetland. Lee said that site managers recognized the importance of preventing the plume from entering the wetland. (Strontium-90, an analogue to calcium, is readily taken up and accumulated in vegetation.) Addressing problems early, Lee said, is a useful approach to take. Site managers gained approval for remedial activities with relative ease because they initiated action before regulators could point to a contaminated surface environment.

Lee said that the remedial technology used at the site had to be easy to maintain and more cost-effective than pump-and-treat technologies. In addition, it had to be set up so that precise performance measurements could be collected with ease. When radioactive contaminants are a concern, Lee explained, site managers must be able to prove, with a high degree of confidence, that remedial technologies are performing effectively. Several different technologies were evaluated before choosing a remedial approach for the site. After years of preliminary testing, a wall-and-curtain treatment system was chosen, in which the wall is a sheetpile cutoff wall and the curtain is a zone of reactive material. The design is unique among PRBs, Lee said, because the treated water is monitored, and, while entirely passive, the system can be hydraulically adjusted to fit plume dimensions.

Lee described the components of the Chalk River site's wall-and-curtain design. The 100-foot-long wall extends into till or to bedrock, which are about 40 feet below ground. The curtain is filled with granular clinoptilolite, a zeolite that has a high affinity for strontium-90. Drainage via a series of pipes, and a level-control manhole have also been incorporated into the Chalk River PRB. The wall-and-curtain has been designed to:

• Allow operators to control hydraulic gradient across the curtain. Ten vertical, perforated pipes are located directly between the wall and the curtain. These pipes discharge to the water level-control manhole, which, in turn, drains to a nearby wetland. Gradient and flow can be controlled by changing the elevation settings of the effluent drainage pipe. Thus, managers can adjust the capture zone.

• Extend the lifespan of the reactive materials. A drainage system has been installed upgradient of the wall-and-current treatment system. It captures and shuttles shallow, uncontaminated ground water to the nearby wetland, bypassing the entire treatment system. By reducing the volume of flow through the curtain, the drainage system extends the lifespan of reactive materials.

• Allow easy and precise monitoring. In contrast with other PRBs, operators can obtain measurements of flow, contaminant concentrations, hydraulic conductivity, and a variety of other parameters whenever they wish. These data can be obtained for any interval that investigators want to evaluate and to as many significant figures as they desire.

Prior to installation, Lee said, the wall-and-curtain design was greeted with some skepticism. To prove that the system would work, a prototype was set up and tested in the laboratory. A sand box model proved to be a useful tool for predicting qualitative performance at different head settings. Using it involved dripping dye onto the sand box and recording the dye movement with a camera. The model demonstrated that a significant amount of water would bypass the treatment system if the hydraulic gradient across the curtain was smaller than the natural gradient. The model also showed that more flow would be captured if the head in the reactive zone was lowered. Numerical modeling was used to quantitatively tailor the system to the actual site.

Two years before building the PRB, Lee said, data about the reactive material performance were collected using in situ column experiments. This method involves placing columns of reactive material into strontium-90-contaminated wells, pumping ground water through the columns for a period of time, and then retrieving and analyzing them. The columns were retrieved after being exposed to field conditions for seven months. The results showed that the clinoptilolite would indeed capture strontium-90, and that the zeolite would not become clogged in the process. No change was detected in the hydraulic conductivity of the clinoptilolite.

One meeting attendee asked whether the clinoptilolite-filled curtain will be removed from the subsurface after it has been loaded with strontium-90. Lee said that a final decision has not been made on this, but that there is a strong possibility that it will be left in the subsurface. Over time, he said, the captured strontium-90 will disappear as a result of radioactive decay. (Strontium-90 has a half-life of 29 years.)

Lee said that the wall was installed in late 1998. He summarized the lessons that were learned during construction activities: (1) the team should have packed more of the dewatering wells with reactive materials, (2) it is not wise to assume that installation contractors will provide available information about the aquifer to dewatering subcontractors, (3) roots should be cleared before construction so that they pose no obstacles during sheetpile installation, and (4) the lateral boundaries of the plume should be clearly defined. Lee said that monitoring data are being collected to assess the wall-and-curtain's performance.

In summary, Lee said, the wall-and-curtain system appears to be cost-effective and to exhibit good performance. He described the technology as an hydraulically adjustable system that can be sized, both vertically and horizontally, to fit treatment requirements.

NASA PRB INSTALLATION USING DEEP-SOIL MIXING: LESSONS LEARNED
Debra Reinhart, University of Central Florida

Debra Reinhart said that her presentation would focus on a PRB system that has been installed at NASA's Cape Canaveral Air Station Launch Complex 34 (LC34). This site, which was used for Saturn rocket launches between 1959 and 1968, is heavily contaminated; some reports indicate that 20,000 kilograms of dense non-aqueous phase liquid may be present in the subsurface. Reinhart said that the site's underlying ground water contains TCE and its daughter products (cis-1,2-DCE and vinyl chloride) at concentrations that exceed MCLs. She provided an overview of the site's hydrology, noting that the hydraulic gradient is extremely low (about 10^-4), ground-water velocities are slow, and ground-water flow follows an unusual radial pattern. The PRB at LC34 is not expected to remediate the entire LC34 area. Rather, it was installed to meet the following objectives: (1) show that ZVI-filled barriers are a viable remedial option at NASA sites, (2) demonstrate an innovative construction technique, and (3) evaluate a rejuvenation technique.

Reinhart described the PRB's design. It consists of a series of 11 overlapping columns (each about 4 feet in diameter) that contain a mixture of ZVI (16% by weight), native soil (79% by weight), and gravel (5% by weight). The column is about 40 feet long, is about 4 feet thick, and is keyed into a clay layer that is about 40 feet below ground. A mixing zone has been established immediately upstream of the iron-filled columns. This zone, which runs parallel to the reactive barrier, was created by driving a series of 10 boreholes into the subsurface. These boreholes, which were left empty, increase the permeability in front of the reactive barrier, and encourage perpendicular ground-water flow through it.

Reinhart said that the wall was installed using a combination of deep-soil mixing and vibro-installation technologies. First, 44 steel casings were driven into the ground using a 325-horsepower vibratory hammer. Each casing was hollow, 40 feet long, 10 inches in diameter, and equipped with a removal steel plug. The latter prevented materials from entering the casings as they were drilled into the subsurface. Once all of the casings were in place, they were filled with iron or gravel. (As noted above, the reactive wall consists of 11 overlapping columns, each 4 feet in diameter. Four casings were placed within each 4-foot column area; one casing was filled with gravel, the other three were filled with iron.) Then an acid wash (2% sulfuric acid) was poured into some of the casings, vibrated gently for 30 minutes, and left to stand overnight.⁶ Once the wash was completed, investigators broke off each casing's steel plug and removed the casings from the subsurface. After this, a deep-soil mixer (ICE Model 55AT) was used to mix the columns. Two mixing passes were conducted for each column to create a homogeneous iron/gravel/soil mixture. A drilling fluid had to be introduced to facilitate mixing; without it, the mixer became mired in the first foot of soil. At first, water was used as the lubricating fluid. After the site became flooded, however, this approach was abandoned and compressed air was used. Once site construction activities had been completed, the reactive wall was mounded and covered with a layer of sod.

Reinhart said that the PRB has been operational for several months, and that researchers have recently started testing an ultrasonic device to determine whether it can be used to maintain the kinetic degradation capabilities of the PRB by reducing buildup on iron filings. (Buildup, Reinhart said, results from materials precipitating out of the ground water and from iron corrosion processes. Expanding on the latter, she said that iron corrodes as it transfers its electrons to chlorinated solvents.) The ultrasonic devices were introduced into monitoring wells that extend into the reactive barriers. Researchers hope to find that the devices' ultrasonic waves can reduce buildup.

Reinhart said that much has been learned from this research project:

• Deep-soil mixing is a cost-effective, efficient installation approach with several benefits. Using this technique requires very little soil excavation. Therefore, site managers do not have to worry about disposing of large quantities of excavated materials, and workers have less exposure to hazardous subsurface materials.

• Mixing using two passes may have been insufficient. It might have been beneficial to perform three, rather than two, mixing passes on the reactive wall's columns. More mixing would have increased the likelihood of obtaining a homogeneous mixture of iron, native soil, and gravel. Reinhart is not sure how well-mixed the wall is, even after two sampling efforts. According to split-spoon samples, the iron percentages (by weight) range dramatically across different depths and sampling points. However, less variation was detected in drill cutting samples--values for percent iron by weight ranged between 13.5% to 20.9%, and were close to the desired target (16%).

• Air can be used as a lubricating agent without adversely impacting wall performance or surrounding air quality.

• It is easier to install casings if soils have been pre-mixed. When casing installation began, the casings penetrated the earth with little resistance and installation times were running at about 35 minutes per casing. As the installation team continued, however, they encountered more resistance and the installation times extended to two hours. To fix this problem, installation efforts were halted temporarily and a deep mixer was used to homogenize subsurface materials. Once this activity was completed, casing installation proceeded rapidly, taking less than 20 minutes per casing.

• Acid washing in the field is effective. In the laboratory, it had been shown that acid washing increases iron reactivity by removing corrosion buildup. At the LC34 site, investigators found that these benefits can also be gained by applying acid washes in the field. To evaluate how acid washing impacts iron reactivity, an acid wash was introduced to 7 of the 11 columns. Then, samples were retrieved from the field and exposed to TCE in the laboratory. First-order rate constants were then calculated for the iron. Results indicated that washing the iron caused a 30% increase in reactivity. (The rate constant of washed iron was 0.155 [hr-mg Fe]^-1; it was 0.120 [hr-mg Fe]^-1 for unwashed iron.)

• Modifications in the casing cap may be necessary. The installation team had problems removing the steel plug at the end of each casing. Although the plugs were supposed to be removable, the team was unable to remove them by tapping them with long steel rods. Therefore, investigators had to design a modified plug, which proved to be much easier to remove in the field.

• The reactive barrier is degrading contaminants. Reinhart presented data for TCE, cis-1,2-DCE, and vinyl chloride in areas upstream of the barrier, within the barrier, and downstream of the barrier. She noted that contaminant concentrations exceed MCLs in the upgradient areas, but that they are reduced to nondetect levels within the barrier. TCE and vinyl chloride have not been detected downstream of the barrier, but cis-1,2-DCE does persist in one downgradient well that is screened in the most contaminated stratum of the aquifer. Although concentrations in this well have declined steadily since May 1999, they are still above MCLs. (The cis-1,2-DCE concentration in the downgradient well is 0.3 mg/L, much lower than the concentration in upgradient wells [40 mg/L].)

• The ultrasound device appears to be rejuvenating iron reactivity. Evidence suggests, that using the ultrasonic device caused iron reactivity to increase by 50% within the LC34 PRB. Given these results, Reinhart said, much enthusiasm has been expressed for this product. The device will be field-tested at another full-scale barrier (the Denver Federal Center) in the near future. Reinhart asked meeting attendees to let her know if they know of other sites that would be willing to test this rejuvenation technology.

INSTALLATION OF A PERMEABLE REACTIVE WALL AT PEASE AIR FORCE BASE USING THE BIOPOLYMER SLURRY TECHNIQUE
Jeff Cange, Bechtel Environmental, Inc.

Jeff Cange said that his presentation would focus on a PRB system that has been installed at Pease AFB. This site, he said, has a large chlorinated solvent plume that extends for more than 1,300 feet. The plume contains TCE, cis-1,2-DCE, and vinyl chloride; total VOCs have been detected above 1,000 ppb. Although a discrete source has never been found, investigators suspect that the contaminants originated from Pease AFB's Site 73. Cange described the geology that underlies the site: a stratified silty sand and till that overlies fractured bedrock. The plume is migrating downward; as a result, it has penetrated the upper portions of the bedrock at locations starting about 325 feet away from the source. Ground-water flow patterns and the site's water table are influenced by the pumping action of a soil vapor extraction system. This system, installed to extract petroleum hydrocarbons, appears to be causing the plume to bend. Much remains to be learned about the plume. For example, Cange said, additional work needs to be performed to identify the plume's downgradient edge.

Cange said that a Record of Decision (ROD) that outlines a preferred remedy for the site has been signed. After closer evaluation, however, site managers agreed that the preferred remedy would not work at the site. They started searching for alternatives, and agreed to participate in a PRB demonstration project. If the PRB proves to be a success, Cange said, it will be added to the list of potential remedial options for the site. The goals of the demonstration are twofold: (1) study the application of ZVI barriers at Pease AFB, and (2) assess potential application at other Air Force sites.

Before designing the PRB, Cange said, investigators conducted a siting study to determine where to install it. This proved to be challenging. The Air Force decided to place the barrier downgradient of the source area and upgradient of the point where ground water enters bedrock. Cange said that he had reservations about this decision because he was not sure whether there would be anything in the subsurface for the PRB to key into. Nevertheless, activities progressed forward.

After identifying the location for the PRB, Cange said, investigators conducted supplemental site investigations in the selected area and used ground-water models to gain a better understanding of local three-dimensional flow fields. (This information had to be obtained, in part, to make sure that introducing a PRB would not disrupt flow fields and drive the plume into the bedrock.) Also, a bench-scale treatability study was conducted to ensure that the site's ground water could be remediated with ZVI.

After a PRB had been designed for the site, Cange said, a performance-based specification and bid package was sent out to installation contractors. Rather than specifying what kind of installation technique had to be used, the package asked potential bidders to propose an approach and to explain how using it would meet the Air Force's criteria. Five companies submitted proposals, four of which suggested a biopolymer slurry trenching technique. This method was selected, and Geo-Con was chosen to perform the installation.

Several activities were performed to prepare for the installation. For example, samples were taken to obtain baseline values for inorganics and field parameters, underground utilities were relocated or abandoned, a security fence was installed, and trees were removed. Accomplishing the latter turned out to be difficult; in fact, the PRB's proposed alignment was modified so that an old sugar maple would not have to be cut down. Also prior to installation, some treatability studies were conducted to evaluate (1) how well the biopolymer slurry would degrade in a high-pH environment, and (2) whether the slurry would affect iron reactivity. The treatability tests were inconclusive.

PRB installation was performed within a one-month period, Cange said. Activities included:

• Mobilizing and grading the site. To prepare for installation, equipment was set up, the soil staging area was prepared, and the site was graded.

• Preparing the biopolymer slurry. Geo-Con prepared the biopolymer slurry by mixing a 50-pound bag of slurry with about 850 gallons of water. The preparation was mixed on site, in a small mobile batch plant that Geo-Con provided.

• Excavating a trench. A large Caterpillar ^® 375L excavator was used to excavate the PRB's trench. The PRB was designed to key into the bedrock layer; Geo-Con felt confident that this could be accomplished. This plan had to be abandoned, however, because the bedrock proved to be less fractured and weathered than originally anticipated. (The excavation equipment broke several times when attempts were made to key into the bedrock layer.) Not only did the PRB fail to penetrate bedrock, it also did not extend to the depths that had been projected during design. This was because the bedrock was encountered several feet shallower than anticipated. In fact, the projected depth to bedrock was erroneous along the entire length of the trench. Because Geo-Con did not have to dig the trench as deep as originally planned, they agreed to extend the trench length.

• Segregating spoils. About 473 cubic yards of soil were removed from the trench and deposited above ground on a liner. An elaborate segregation scheme had been developed, which involved segregating soils based on the depth from which they were extracted and their location relative to the plume. This scheme did not work: the materials were too wet and muddy and they ran together.

• Pumping the biopolymer slurry into the trench. The biopolymer slurry was pumped into the trench, one batch at a time, to hold up the trench's side walls.

• Preparing the backfill. The backfill consisted of a mixture of sand and iron. (About 487 tons of sand and 359 tons of granular iron were mixed together.) The sand was obtained from a local batch plant and the mixture was prepared in a ready-mix truck.

• Pumping the backfill into the trench. The backfill was shuttled from the ready-mix truck to the trench via tremie pipe.

• Performing quality control monitoring. Investigators were worried that preferential sorting was occurring, so they collected samples from the backfilled trench. The results indicated no preferential sorting: the samples showed that the backfill was well mixed and the percent-iron-by-weight values generally matched those targeted in the PRB's design.

• Installing monitoring wells. Eighteen monitoring walls were installed in the trench, but installing them through the thick biopolymer slurry proved to be very difficult. In fact, the wells did not proceed into the reactive trench vertically; as a result, investigators are not sure where the wells' bottoms are. After installation activities were completed, about 20 more wells were installed around the trench.

• Degrading the slurry and flushing the trench. In order to break down the slurry, about 20 gallons of enzyme were injected into the barrier through a 6-inch PVC flushing well. The slurry did not break up immediately; in fact, because so much water backed up, a truck had to be brought in to siphon some off the top. (This water was sent to a nearby waste-water treatment plant.) Geo-Con was asked to stay on site and to keep circulating water through the trench until the enzyme degraded the slurry. This took about two and a half days. Cange said that it is important to make sure that contractors understand that they must stay on site until the trench is adequately flushed.

• Installing a non-woven geosynthetic material and restoring the site. A non-woven geosynthetic material was placed on top of the reactive barrier and then covered with fill.

Cange said that a rigorous monitoring program has been initiated at the site. As part of this program, water quality samples are being collected to determine the effectiveness of the treatment zone, whether any contaminants have bypassed the PRB, geochemical conditions, and downgradient aquifer conditions. In addition, hydraulic data are being collected to determine ground-water flow paths, capture zones, and whether installation activities have impacted local hydraulic patterns. Also, data are being collected to measure targeted VOCs, selected inorganics, intrinsic remediation constituents, and field parameters. Data are just now becoming available, but they have not yet been fully evaluated. Preliminary analysis indicates that (1) the PRB is operating effectively, and (2) construction activities have not impacted ground-water contours. Some results suggest that construction activities liberated a large amount of TCE into upgradient ground-water streams. Cange is hopeful that this TCE will be dechlorinated as it passes through the PRB.

CONSTRUCTION OF A TEST SECTION OF A PRB AT THE SOMERSWORTH SANITARY LANDFILL SUPERFUND SITE
Tom Krug, GeoSyntec

Tom Krug focused his presentation on a PRB designed for the Somersworth Sanitary Landfill Superfund site, at which municipal and industrial wastes were disposed of from the mid-1930s to 1981. He began by describing the site's subsurface. Under the landfill, a stratified layer of sands sits over an irregular and weathered bedrock surface. The sand layer exhibits much heterogeneity, Krug said, and contains silt, gravel, and cobbles. A ground-water plume, contaminated with PCE, TCE, and daughter products, has formed under the landfill and is traveling through the overburden and discharging into a wetland.

Krug said that the landfill was placed on the National Priorities List in 1982. Following much investigation, he said, a preferred remedial approach was designated for the site. Krug said that it would have cost about $26 million to implement this preferred approach, which involved: (1) constructing a conventional slurry wall around the entire site, (2) capping the landfill, and (3) installing a pump-and-treat system. In 1995, efforts were made to identify a more cost-effective way to address contamination at the site. A ROD was signed that allowed for a more innovative remedial scheme that uses four different technologies:

• PRB. A PRB will be used to treat overburden ground water. Remedial managers believe that this will reduce chlorinated solvent concentrations to levels below regulatory standards.

• Natural attenuation. Natural attenuation will be used to treat the portion of the overburden plume that has already moved beyond the edge of the landfill's waste.

• Permeable landfill cover. A permeable cover will be installed over the landfill rather than a prescriptive RCRA-type cover. Remedial managers do not want to entomb wastes and trap contaminants. Rather, they want to flush contaminants into the overburden ground-water plume so that they can be treated by the PRB.

• Ground-water extraction system. Some contaminants are present in the bedrock layer. Remedial managers propose removing these via a ground-water extraction system.

Krug explained how remedial managers identified a design for the PRB system. First, field investigations were conducted to better characterize contaminant distributions and hydraulic conditions. From these activities, remedial managers learned that (1) some parts of the PRB would be exposed to much higher contaminant concentrations than others, and (2) ground-water flow and velocities would differ across the PRB. Also, laboratory tests were conducted during the design phase--for example, to determine how temperature affects iron reactivity. Remedial planners used the data collected to calculate temperature correction factors.

Remedial managers considered using a funnel-and-gate design, but found, through ground-water modeling exercises, that water would flow around, underneath, and over this type of system. Instead, they chose to use a continuous wall system. The PRB is designed to be over 800 feet long, consisting of eight sections which are approximately 100 feet long. The design of each section has been customized to accommodate local conditions and to remediate local contaminant concentration. (For each section of the wall, residence time was estimated by combining information on degradation rates with contaminant concentrations. The amount of iron required to remediate the contaminants was then determined by combining the estimated residence time with information on ground-water velocity and capture. Then the system configuration was determined by factoring in geology, depth, and site conditions.) When designing a long wall, Krug said, it is more cost-effective to design the wall in sections than to use one design across the wall's length. PRB walls designed in sections can have larger amounts of iron in areas that are highly contaminated, and smaller amounts in less contaminated areas. This helps cut down on excessive use of iron.

After a design was identified, Krug said, remedial managers had to decide which approach to use for PRB installation. Identifying an approach was a challenge: some of the more conventional installation methods would not have performed well because the site is characterized by a shallow water table, a heterogeneous subsurface that is interspersed with "bony material," heaving soils, and an irregular, weathered bedrock. Also, some construction methods were discounted automatically because they could not install a PRB to the desired depth (40 feet). The following installation methods were considered and rejected: overlapping caissons, continuous trencher, jetting, and soil freezing. After much consideration, and not without reservations, remedial managers chose to use the "open trench with biopolymer slurry" method. Before field operations were started, laboratory tests were performed to determine whether the biopolymer would have a negative impact on iron reactivity. When the biopolymer and the iron were allowed to come into intimate contact, the tests showed that the biopolymer slurry does indeed reduce iron reactivity. Because the tests were only conducted over a short period, it is impossible to tell whether the negative affects would have diminished over time. Nevertheless, the results were alarming enough to prompt remedial managers to identify ways to minimize interactions between the two media. They decided to do this by (1) pre-wetting the iron, and (2) injecting excess water into the backfill so that voids between iron particles would be filled with water rather than biopolymer. Laboratory tests will be conducted soon to determine whether these procedures help mitigate reductions in iron reactivity. In addition, tests will be performed to determine what components of the biopolymer cause the reductions.

In November 1999, Krug said, Geo-Con installed a 25-foot-long section of the PRB using the biopolymer slurry installation technique. The procedures used were very similar to those described above for the Pease AFB PRB installation, but there were some notable differences. For example, at the Somersworth Sanitary Landfill Superfund site: (1) spoils were deposited on top of the site's landfill rather than being moved to a temporary staging area, and (2) a platform was built because the site's water table was so shallow.

Krug described the monitoring activities that are being performed to determine whether the Somersworth Sanitary Landfill site's PRB is operating effectively and whether the biopolymer has impacted hydraulics or iron reactivity. These activities include collecting core samples, conducting hydraulic tests, and measuring contaminants, pH, oxidation-reduction potential, and conductance. Many of the monitoring activities and analyses are still underway, but some preliminary results are available. These indicate the following:

• Residual biopolymer has not collected on the iron. Iron was extracted from the PRB section after the barrier had been in place for one month. The samples, which appeared to be loosely packed, were subjected to scanning electron microscopy; no significant differences were observed between virgin iron and iron collected from the field.

• The biopolymer does not appear to impact hydraulics. Hydraulic conductivity has been measured (using slug tests and ground-water viscosity measurements) in and around the PRB section. Results indicate that the hydraulic conductivity within the barrier matches design targets and that conductivity around the barrier has not changed significantly following the PRB's installation. No residual high ground-water viscosity has been observed within the barrier, so it appears that all of the biopolymer has disappeared from the barrier.

• Specific conductance, pH, and oxidation-reduction potentials are responding as expected. Probes have been used to measure specific conductance, pH, and oxidation-reduction potentials upgradient, within, and downgradient of the barrier. All of these are behaving as expected.

• The PRB is reducing ground-water contaminants. PCE, TCE, cis-1,2-DCE, and vinyl chloride are being monitored upgradient, within, and downgradient of the barrier. PCE and TCE concentrations, which enter the barrier at concentrations exceeding MCLs, are reduced to nondetect within the barrier and do not reappear downstream. The concentrations of the other two chemicals are also reduced within the barrier, but they have been shown to reappear in downstream wells. Krug said that this was not a surprise, given where the PRB is placed. When the PRB was installed, remedial managers already knew that some downgradient areas had been impacted by contaminants. Remedial managers are optimistic that these contaminants will be addressed through natural attenuation.

CONSTRUCTION OF A FUNNEL-AND-GATE TREATMENT SYSTEM FOR PESTICIDE-CONTAMINATED GROUND WATER
EM>Dean Williamson, CH₂M Hill

Dean Williamson said that his presentation would focus on a funnel-and-gate treatment system that has been installed at the Marzone Superfund site. This site, which operated as a pesticide formulation plant between 1950 and 1983, released several types of chemicals, many of which have impacted soil and ground water. Williamson said that the Chevron Chemical Company, the site's original owner, has been heavily involved with cleanup activities. Several removal actions were conducted in the 1980s, a ROD was signed in 1994, and soil cleanup activities were completed between 1996 and 1998. Now remedial efforts are focusing on ground water.

Williamson said that a shallow aquifer (depth to approximately 20 to 25 feet below land surface) has been impacted by pesticides (e.g., DDT, hexachlorocyclohexanes [BHCs], and chlordane), xylene, and ethylbenzene. The aquifer unit is heterogeneous (composed of various interbedded clay, silt, and sand layers) and is characterized as being relatively "tight" and having low yield. A thick clay layer separates the shallow aquifer from a deeper aquifer, which is used as a drinking water source. The ground-water plume has migrated off site, but there is no indication that it has impacted a nearby creek about 300 to 400 feet downgradient of the site. Flow rates in the shallow aquifer are very slow (10 to 20 feet per year).

The ROD identified pump-and-treat as the preferred remedial approach for addressing ground water, but Chevron, which has had little success using this technology at other sites, expressed interest in testing more passive, innovative in situ technologies. As a result, Williamson said, CH₂M Hill prepared a Feasibility Study addendum so that newer technologies could be assessed before a final remedial action decision is made. Remedial managers considered a revised pump-and-treat system, a continuous wall PRB, and a funnel-and-gate PRB system. They chose to test the latter in a pilot test.

Williamson described the process that was used to choose a design for the funnel-and-gate system. First, he said, bench-scale studies were used to identify an appropriate reactive medium. ZVI and granular activated carbon (GAC) were tested; the latter proved to be more effective for the contaminants at the Marzone site. (ZVI has little impact on xylene, which has been detected at concentrations as high as 55,000 ppb.) Also, ground-water modeling was performed to determine how many treatment gates should be used and where they should be installed. Using MODFLOW, investigators evaluated single- and dual-gate systems. Results indicated that the former would be adequate for the site.

Williamson said that determining the location for the funnel-and-gate treatment system proved to be a bit of a challenge. Remedial managers did not want to place it too far downgradient of source areas, because they did not want decades to pass before the slow-moving ground water would carry contaminants to the PRB. However, as the proposed location moved further upgradient, EPA expressed concern about leaving residual contaminants downgradient of the barrier. A compromise was reached: (1) the PRB would be placed so that ~93% of the plume's contaminant mass would be upgradient of the treatment system, and (2) regular monitoring would be conducted downgradient of the barrier to make sure that the remaining ~7% was being remediated adequately via natural attenuation.

Williamson described the chosen design: (1) an Impermix cutoff wall, (2) a collection trench, (3) concrete vaults, and (4) a distribution channel. The cutoff wall acts as the "funnel," and the collection trench, which is filled with inert gravel, collects ground water and helps establish uniform flow. Pipes shuttle the ground water into the treatment gate, which consists of three subsurface precast concrete vaults filled with GAC. After undergoing treatment, water is piped to the distribution channel. This channel has been placed significantly downgradient of the funnel-and-gate in order to take advantage of better hydraulic forces. Remedial managers realize that the funnel-and-gate system may operate for several decades and that site conditions and remedial focus may change over time. Accordingly, they designed the system to be flexible: (1) the treatment vaults can be operated in parallel or in series, (2) the system can be operated in either an upflow or downflow mode, (3) the system can be backwashed, and (4) the system can accommodate alternative reactive media if necessary. In essence, he said, the system is as flexible as an above-ground GAC system.

Williamson said that the PRB was constructed in August 1998 and Geo-Con led installation efforts. Various methods were used to install the different components of the funnel-and-gate system. These included:

• Biopolymer slurry method. To construct the collection and distribution trenches, geotextiles were placed at the bottom of the trenches, a biopolymer slurry was used to hold the trenches open, and an inert gravel was poured into the trenches. A great deal of guar had to be used to hold the trenches open because the guar degraded within a few days of being placed in the trench. (The barrier was installed during hot weather, which increased the rate at which microbes degraded the guar.)

• Trench box. To install the system's gates, a hole was dug, a trench box was installed, three treatment vaults and the system's piping were lowered into the ground, and OSHA-trained staff worked within the trench box to prepare the gates.

• Vibrated beam installation. The cutoff wall was installed using Slurry Systems, Inc.'s vibrated beam method. This approach was chosen because (1) it is easy to use, (2) it does not generate any excavated soil, and (3) it was suitable given the site's subsurface and above-ground conditions (e.g., nearby railroad tracks). The method proved to be very successful. Within five days, a 400-foot-long wall had been constructed. Williamson mentioned one feature of the installation method he found particularly appealing: the vibrator beam device provides real-time measurements of the subsurface's resistance to penetration. Therefore, he said, remedial managers know when the wall hits sand, gravel, or clay; this lets them know whether the wall has been keyed into clay layers.

The funnel-and-gate system, Williamson said, became operational in August 1998. Two problems were encountered during start-up:

• Biogas collected in pipe systems and caused periodic flow blockages. During the first several months of operation, remedial managers had trouble maintaining flow through the funnel-and-gate system. They could induce flow by flushing the system, but they had to do this every two to three weeks to keep the system operational. After running a mini-camera up the effluent lines, remedial managers found the source of the problem: gas bubbles were accumulating in the pipes that connect the gate and the distribution trench. The remedial managers suspected that guar was playing a large role in the gas production, because biogas is formed when guar degrades. Flow blockage ceased to be a problem when vents were installed to relieve the system of gas. Williamson does not think that much gas is being produced in the system any more, because most of the guar degraded after the system's first year of operation. One meeting attendee asked Williamson whether the biogas created any air emissions problems. Williamson said that some emissions could have escaped from the vaults, but that no efforts had been made to determine whether this actually happened.

• Leaks were detected in the mechanical joints between two precast vault units. Remedial managers suspect that the asphaltic seal degraded when it contacted the ground water's xylene and/or ethylbenzene. To repair the leak, a specialty contractor (trained to work in confined spaces) had to enter the vault and inject a polyurethane foam into the joints.

Williamson said that the funnel-and-gate system has been operating for about one and a half years, and that monitoring results indicate that the PRB is operating effectively. As ground water moves through the PRB, contaminants are reduced to nondetect levels. For example, xylene enters the PRB at concentrations of 16,000 ppb, but leaves at concentrations less than 2 ppb. Likewise, alpha-BHC concentrations decrease from 0.96 ppb to less than 0.03 ppb as ground water moves through the barrier. Given these favorable results, Williamson said, EPA has amended the ROD to make the PRB system the permanent ground-water remedy.

In summary, Williamson said, the following lessons were learned at the Marzone Superfund site: (1) ground-water modeling can be an effective design tool, (2) process flexibility can be incorporated into in situ ground-water treatment systems, (3) site constraints can dictate which construction methods are used, (4) guar produces biogas, and (5) it is wise to collect a lot of data.

A SUMMARY OF CONSTRUCTION EXPERIENCES AND LESSONS LEARNED FROM INSTALLATION OF THREE DIFFERENT REACTIVE BARRIER SYSTEMS
Will Goldberg, MSE Technology Applications, Inc.

Will Goldberg said that his presentation would focus on the lessons that have been learned from installing three different reactive barrier systems. All of these systems, he said, have been installed at DOE sites--two at the Y-12 Plant site and one at the Rocky Flats Environmental Technology Site (RFETS). (Goldberg noted that the PRB systems at the Y-12 Plant site had already been discussed, as part of David Watson's presentation. Goldberg said that his presentation would supplement Watson's discussion rather than offer redundant information.)

The Y-12 Plant's Pathway 1 Site

Goldberg said that a funnel-and-gate system was installed at the Y-12 Plant's Pathway 1 site in late 1997. He described the installation of the system's treatment vault (the gate) and wing walls (the funnel). The vault, which contains ZVI, is at the center of the treatment system, and was installed using a trench box. The trench box was large enough to accommodate the vault, but very little space was available for workers to seal joints. This complicated matters and slowed the workers' pace.

The wing walls, which have HDPE membranes sandwiched in their middles and collection and redistribution trenches incorporated into their design, were a challenge to install. Contractors had to choose an installation method without the benefit of detailed geotechnical information. Prior to installation, very little geotechnical data had been collected in the area of the proposed PRB. Although some push probe tests had been conducted, no standard penetration tests or core sampling had been done. Thus, remedial managers did not have a clear idea of what they would be digging into during construction activities. To get a better idea, they dug a trench to the east of the PRB's footprint. Very little water was encountered during digging and the trench stood open for several hours; therefore, a decision was made to install the walls using an open trench method. Once contractors started digging the walls for the PRB, however, it became clear that this method would not work. Much to their surprise, contractors encountered large quantities of unconsolidated backfill and a 12-inch water main. (The site owner had indicated that a water main was present, but had not located it accurately on site maps.) In the end, a decision was made to use the biopolymer slurry method to install the walls.

Goldberg said that the following lessons were learned during installation activities at the Pathway 1 site: (1) geotechnical data should be obtained from the footprint area prior to construction, (2) independent utility surveys should be conducted, (3) a contingency plan should be generated in case underground site conditions differ from what was expected, (4) trench boxes need to be large enough to provide sufficient working space, and (5) HDPE interlocking panels and frames are easy to install if sufficient operating room is available.

The Y-12 Plant's Pathway 2 Site

In 1997, a permeable reactive trench was installed at the Y-12 Plant's Pathway 2 site. In 1999, several modifications were made in an attempt to make the system more effective and to extend its capture zone. Toward this end, the trench was extended and a siphon well, underground pipeline, treatment box, and redistribution field were added. Goldberg described how the enhancement is designed to work, noting that its overall purpose is to increase ground-water gradient. First, he said, water collects in the siphon well and passes through the underground pipeline to a treatment box about 750 feet downgradient. Then, ground water is treated in the box and shuttled to a 300-foot-long distribution trench filled with coarse gravel. Finally, the ground water is redistributed and released gradually to the subsurface.

Goldberg described the construction techniques that were used to:

• Extend the trench. A continuous trencher was used to extend the existing reactive barrier trench about 120 feet to the west. This machine excavated weathered shale and penetrated the bedrock, simultaneously placing pea-sized gravel into the subsurface. (The pea-sized gravel was not listed in the extension plan's original specifications. Rather, designers had suggested using a fine silica/sand as backfill. It became clear that this fine material could not be used, because it had difficulty flowing through the continuous trencher's hopper, and it did not settle well within the water-filled trench.) The continuous trencher proved to be a feasible and successful installation method to use at the Pathway 2 site; the main benefit of using this type of construction technique is that it does not require artificial ground support (e.g., a biopolymer slurry). The continuous trencher is cost-competitive with other installation techniques when used to install long trenches. For shorter stretches, however, it does not appear to be as cost-effective.

• Install the pipeline. Horizontal boring equipment was used to install a 750-foot-long underground pipeline. The installation was accomplished within two days, and was considered to be cost-effective. By using this technique, the installation team was able to install the pipeline without disturbing surface obstructions or generating too much excavated material. Goldberg stressed the benefits of the latter, saying that reducing the amount of spoils reduces cost and risk--less money is spent on spoils disposal, and workers are less likely to contact hazardous materials.

• Install the treatment box. The treatment box, which is underground, contains two vaults, both of which have been backfilled with ZVI. (These operate in parallel, not in series.) Goldberg said that a geotextile has been placed over the iron to separate it from overlying gravel. Remedial managers decided to use vaults so that the ZVI can be easily retrieved and disposed of in the future.

Goldberg said that the siphon system is operational, but has broken down on a few occasions. Researchers suspect that this has been caused by degassing of extracted ground water. This problem will be addressed soon.

Goldberg said that the following lessons were learned during installation activities at the Pathway 2 site: (1) a contingency plan should be generated to deal with unexpected site conditions, (2) fine sands do not settle quickly in water-filled trenches, (3) horizontal boring techniques work well, (4) continuous trenchers work well, and (5) degassing of extracted ground water can cause problems in siphon systems.

The RFETS Site

At the RFETS site, Goldberg said, contaminants from a drum disposal area leaked and created a TCE ground-water plume. For many years, this ground water was collected from a discharge seep, stored in a holding tank, and sent to a waste-water treatment plant. This approach was expensive, so a PRB was installed during the summer of 1998 in the hope that it would be an effective, passive, cost-effective alternative.

Goldberg described the PRB system's design. He said that ground water flows into a gravel-filled collection trench (270 feet long and 20 feet deep) before being intercepted by an impermeable HDPE barrier. Ground water then flows through an underground pipe and into iron-filled treatment boxes, which can be operated in series or individually. The treated water then passes through a flow meter, and is distributed to the subsurface via a french drain.

Goldberg did not go into detail about the installation methods that were used at this site, but he did note some of the problems that were encountered during construction activities. Ground stability proved to be a problem, even though site characterization data had not suggested that this would be an issue. (Intensive ground movement occurred along the site's hillside.) Also, underground utilities (e.g., an alarm line) slowed down trenching activities. At times, obtaining adequate manpower and equipment was a problem.

In summary, Goldberg said, the following lessons were learned during installation activities: (1) it is wise to minimize the length of open trench and the time taken to place backfill, (2) it is wise to minimize equipment operations and stockpiling adjacent to open trenches, (3) maintenance must be considered in a PRB's design, (4) backfill specifications must be rigidly followed, (5) independent utility surveys should be conducted, and (6) adequate manpower and equipment should be provided.

Summary and Conclusions

Before closing his presentation, Goldberg offered some suggestions about PRB construction activities at any site: (1) beware of underground utilities and buried debris, (2) completely understand site requirements and restrictions, (3) carefully assess physical limitations and restrictions as they relate to working room, and (4) identify an effective onsite point of contact.

DUAL REACTIVE BARRIER WALLS FOR THE REMEDIATION OF CHLORINATED HYDROCARBONS, WATERVLIET ARSENAL, NEW YORK: DESIGN AND INSTALLATION
Grant A. Anderson, U.S. Army Corps of Engineers
Russell Marsh, U.S. Army Corps of Engineers

Grant Anderson and Russell Marsh described the dual reactive barrier walls that have been installed to remediate chlorinated hydrocarbons at the Watervliet Arsenal site. Anderson, who opened the presentation, provided a brief site description and summarized the site's geology. He said that the Watervliet Arsenal, the oldest cannon factory in the United States, disposed wastes and operated a burn pit in a portion of the site called the "Siberia Area." Anderson said that the Siberia Area, particularly its northeast quadrant, is heavily contaminated with various VOCs, semivolatile organic compounds, polycyclic aromatic hydrocarbons, and metals. Contaminants have migrated to the subsurface, he said, and a contaminated ground-water plume has formed and migrated to the sewer bedding, which may act as a conduit for off-site transport. Anderson said that the Siberia Area was once a bog, but that it has been filled in. Underneath the fill, he said, there is a tight layer of lacustrine silts, clay, and peat. On the eastern side of the Siberia Area, this layer sits directly over weathered bedrock. On the western side, a layer of sand separates the two. Anderson said that slug tests have been performed to measure the hydraulic conductivity of the different layers. The mean values, in feet per day, are 8.9 for the fill, 0.45 for the lacustrine layer, and 6.3 for the weathered bedrock.

After providing this site background, Anderson turned the presentation over to Marsh, who began by describing the design process for the Watervliet Arsenal's PRB. Bench-scale studies, said Marsh, showed that: (1) a residence time of 60 hours was needed (vinyl chloride served as the driver for residence time), (2) a 0.38-foot-thick wall of 100% iron was needed, and (3) inorganics might precipitate out of solution, but this would probably not impact the PRB system. The design process also involved using ground-water models, which proved to be a powerful design tool. The models were used to evaluate many different designs and to determine the best alignment for the PRB. The results indicated that using a funnel design could create underflow at the site and reduce the efficacy of the reactive wall. The models also showed that underflow or mounding could occur if the PRB were not perfectly perpendicular to ground-water flow. At the end of the modeling stage, investigators decided that dual reactive walls would provide the most efficient configuration.

Marsh described the system that has been installed at the Watervliet site. It consists of two reactive walls: (1) the southern wall, which is about 190 linear feet long and serves as the main barrier, and (2) the northern wall, which is about 80 linear feet long and acts as a polishing wall. Both walls are about 3 feet wide and about 8 to 12 feet deep, are keyed into competent bedrock, are filled with an iron/sand mixture, and are perpendicular to ground-water flow. To install the reactive barriers, trenches were dug using a conventional track-mounted excavator. Two sets of steel shoring plates were then installed, and hydraulic "speed shores" were used to prevent sidewall collapse and to maintain designed trench width (30 inches). After the trenches were backfilled, backfill materials were added behind the shoring plates (to fill voids), and then the plates were removed.

Marsh explained how the backfill material was prepared and delivered. The material, a mixture of sand and iron, was prepared at a nearby concrete plant. First, bulk granular iron was delivered to the plant. At the plant, iron and sand were transferred to hoppers on a conveyor belt, then discharged into a computerized weighing bin so that weight slips, which documented the weights of iron and sand mixed together, could be generated. After this, the backfill mixture was loaded into a truck equipped with a rotary drum, then delivered to the site. A chute transferred the backfill to the trench; for each batch, samples were collected early in the placement process, halfway through it, and at the end. These samples were collected to ensure that the backfill being placed into the trench was well mixed. Unfortunately, remedial managers could not evaluate the mix through simple visual inspection, because the iron was the same color as the sand. Therefore, remedial managers took selected samples, used a magnet to separate the sand and iron, weighed each component, and compared the results with the concrete plant's weight slips.

Marsh said that the PRB system at the Watervliet Arsenal is proving to be an effective treatment system. For the most part, monitoring shows that chlorinated hydrocarbon concentrations fall to nondetect values as ground water flows through the reactive barriers. Also, preliminary cost calculations suggest that the PRB system is very cost-effective; costs associated with designing ($113,000) and constructing ($278,000) the PRB were estimated to be about one-third the cost that would have been associated with designing and building a pump-and-treat-system. As for operation and maintenance (O&M) costs, he continued, the discrepancy between PRB and pump-and-treat systems is very wide: over a 30-year period, he said, it should cost about $3 million less to use the PRB system.

Marsh cited two lessons that were learned during installation. First, it is easy to overestimate the amount of water that will enter the trench during installation. (At the Watervliet Arsenal site, a holding tank many orders of magnitude too large was installed.) Second, it may not always be possible to determine whether backfill is well mixed through simple visual evaluation.

LESSONS LEARNED INSTALLING AN IRON BARRIER AT THE KANSAS CITY PLANT
Paul Dieckmann, Honeywell
John Moylan, URS Greiner Woodward Clyde

Paul Dieckmann and John Moylan described a PRB that has been installed at the Kansas City Plant. Dieckmann said that a VOC plume has formed under this site, and that cis-1,2-DCE (concentrations up to 1,377 ppb) and vinyl chloride (concentrations up to 291 ppb) are the predominant contaminants. He provided a brief overview of the site's subsurface, noting that saturated thickness extends to a depth of 21 feet. Soils in the subsurface consist of a silty clay over a (somewhat heterogeneous) 3-foot basal gravel layer. A bedrock layer, Dieckmann said, is located under the soil formation.

Before designing a PRB system for any site, Dieckmann and Moylan said, it is important to perform site characterization investigations. For example, it is important to collect information on:

• The area's natural ground-water flow patterns. Prior to designing a PRB, Moylan said, it is important to collect information on the area's natural ground-water flow patterns. At the Kansas City Plant, a pump-and-treat system influences ground-water flow patterns; this had to be shut down prior to site investigation studies so that flow distributions could take on their natural patterns.

• The area's topography. Moylan said that understanding a site's topography helps remedial managers understand water flow systems. The Kansas City Plant is characterized by a very interesting and complicated topography. Essentially, the topography at the plant is about 15 feet lower than it was in the past. Also, it has changed dramatically since the 1940s: the Blue River has been cut off, a flood project was initiated, a levee was constructed, and some areas have been excavated.

• The site's subsurface. Dieckmann and Moylan also noted the importance of collecting information on the subsurface's physical properties, layering patterns, and hydraulic conditions. Dieckmann listed some of the input parameters that were used to design the Kansas City Plant PRB:
  
  

  Parameter	Value
  Hydraulic conductivity of clay layer	0.75 feet/day
  Hydraulic conductivity of basal gravel layer	34 feet/day
  Hydraulic gradient	0.01 ft/ft
  Porosity	30%
  Ground-water velocity in clay layer	0.025 ft/day
  Ground-water velocity in basal gravel layer	1.13 ft/day

  Dieckmann said that these parameters were used to design the PRB because they were the best estimates that remedial managers had access to at the time. More recent data suggest that heterogeneity in the soil formation causes fairly significant spatial variations in (1) ground-water velocities, and (2) hydraulic conductivity in the basal gravel layer.

Dieckmann explained how the PRB at the Kansas City Plant was installed, noting that several problems were encountered during construction activities. Originally, remedial managers planned to install the barrier using single-pass trenching equipment. Before trenching activities were initiated, the equipment, which arrived at the site in pieces, had to be put together. Then the installation contractor made some modifications to the trenching train. (Rock teeth were added among the soil paddles, because remedial managers wanted to key into the bedrock layer. They hoped that establishing intimate contact between the reactive material--iron--and the bedrock would eliminate any chance of underflow.) From the start, the installation team had trouble keeping the single-pass trencher operational. For example, the machine ran off its track on its return from refueling. Once trenching activities began, a number of obstacles were encountered: the machine's stinger became mired in mud, the machine's chain broke, and the machine's stinger was sheared off. Given all these setbacks and difficulties, the single-pass trencher was abandoned, and installation contractors decided to use the trench box method to perform the installation. This, too, proved ineffective and had to be abandoned. Dieckmann said that the site's wet, heavy clay was probably too heavy to allow maneuverability for the single-pass trenching equipment and the trench box.

Finally, Dieckmann said, an appropriate installation method was identified for the site: sheetpiles were driven into the ground to brace the trench. A vibratory hammer was used to install the sheet piles and the 130-foot-long PRB was installed in five sections. (The PRB was divided into 26-foot long cells by driving down steel plates at the end of each cell.) Installing the PRB in sections, Dieckmann explained, minimized the stress that was placed on sheetpiles and reduced the amount of dewatering that was required. Before each cell was backfilled, he said, it had to be thoroughly cleaned and dewatered. This proved to be a slow process initially, but went more smoothly once (1) the installation contractor customized his backhoe bucket to make soil removal easier, and (2) workers started using a compressed air line in combination with a long piece of steel angle (attached to a backhoe bucket) to remove soil that was attached to the side of sheet piles.

After installation was completed, Moylan said, monitoring activities were initiated to assess the PRB's performance. The results indicate that introducing the PRB (a highly conductive zone) into the semi-confined aquifer has redistributed flow patterns near the barrier. Ground-water elevations near the barrier's extremities seem to be lower than they were before PRB construction; conversely, the levels are now higher in areas near the barrier's center. Also, during storm periods, flow appears to travel toward the PRB both from upgradient and downgradient sides. In summary, the redistribution of flow has improved plume capture in the northern portions of the barrier, but it has caused a section of the plume to bypass the southern end of the barrier.

Moylan concluded by listing lessons learned at the Kansas City Plant: (1) PRBs can cause redistribution of local ground-water flow patterns, (2) careful monitoring and evaluation should be conducted after installation, and (3) site characterization is extremely important because design and installation performance are influenced by geotechnical properties, horizontal and vertical hydraulic conductivity patterns, hydraulic gradient, water table elevations, contaminant distributions, and temporal variations.

OPEN DISCUSSION/QUESTION AND ANSWER PERIOD

Two discussion sessions were held during the meeting, one before lunch and one right before the meeting adjourned. During these sessions, the meeting's speakers were called to the front of the room to form a panel. Meeting attendees were then allowed to ask them questions and to stimulate discussion in areas that they felt required attention. Discussion revolved mainly around hydraulic, construction, and cost issues, but several other topics were also brought up.

Hydraulic Issues

Tracer Studies

At one point, Donald Marcus said, tracer tests were being hailed as a useful technique for measuring retention times and ground-water velocities. He asked why none of the meeting's speakers discussed tracer tests during their presentations and asked whether anyone had tried using this technique in the field. Kjeldsen responded by saying that he believes that tracer tests, in theory, are an excellent analytical tool; in fact, he said, they may be one of the only techniques available that can be used to obtain direct, rather than estimated, ground-water velocities. Although the tracer test is expensive and labor-intensive, he said, it is probably useful for sites with steep gradients. However, he continued, it is more difficult to obtain meaningful results at sites with flatter gradients. Kjeldsen said that a low-scale tracer test was conducted at the Freight Yard site, but that the tracer could not be found after it had been injected into the subsurface. Other meeting attendees reported similar experiences. (Gavaskar said that a tracer test was performed at the NAS Moffett Field site in 1998, but that researchers could not find the tracer after it had been injected into the ground. At Lowry AFB, Bob Edwards said, the same thing happened--tracer was injected into the ground and never seen again. Vidumsky said that he experienced the same disappointing results, but he did not indicate at which site.) David Watson, however, said that he has experienced some success using tracer tests. He said that he has found that systems must be overwhelmed with tracer in order to obtain meaningful results.

Uncertainties in Ground-Water Flow Measurements

Lee noted that hydraulic conductivity values and hydraulic gradient measurements are used to estimate ground-water flow. Values for both parameters, he said, often carry a high degree of uncertainty; he questioned how accurate flow estimates really are. Meeting attendees thought that Lee had made an excellent point, one that was worthy of additional discussion. Building on Lee's point, Vidumsky said that engineers who design PRBs must determine how much uncertainty can reasonably be eliminated by conducting additional site characterization. Also, he said, designers must decide whether they will include safety factors in their designs to account for unavoidable uncertainties. He said that some researchers are talking about using Monte Carlo analyses to generate more accurate hydraulic information. He asked meeting attendees to comment on ways to reduce uncertainties or to incorporate uncertainties into design plans. Meeting attendees offered the following suggestions:

• Use simple simulations as a starting point. Eykholt said that much can be learned about flow rates using simple simulation models. He said that these models can be used to show investigators how flux varies under different scenarios. As a result, he said, useful information can be collected without conducting overly extensive (and expensive) monitoring efforts.

• Compartmentalize the barrier. Eykholt said that it might be useful to compartmentalize barriers and evaluate the flow that goes through each segment.

• Use tracer tests. Marcus suggested using the following approach to learn about an area's hydraulics: conduct tracer studies, collect information on a timed basis, and subject the data to Monte Carlo analysis.

• Track contaminants to determine flow patterns. Gavaskar said that contaminants can be thought of as tracers and can be used to determine whether (1) ground water is bypassing a PRB or (2) a PRB has changed an area's hydraulics. In order for this approach to work, he said, several monitoring wells must be installed in and around the barrier. Also, prior to PRB installation, remedial managers must have a clear picture of contaminant distributions.

Construction/Installation Issues

Types of Contracts

Meeting attendees talked briefly about the different types of agreements that can be established with installation contractors. One participant said that some contractors are willing to sign "turnkey" contracts. He asked whether any contractors are willing to sign fixed-bid contracts. Meeting attendees said that some contractors are willing to do so. (Dieckmann said that the Kansas City Plant PRB was installed under a fixed-price contract. Vidumsky said that DuPont is asking for fixed-priced bids for a PRB that is scheduled to be installed this year.) Goldberg said that contractors are more likely to agree to fixed-bid contracts if extensive site characterization information is available.

Depth Limitations of the Biopolymer Slurry Method

Meeting attendees talked briefly about the depth to which the biopolymer slurry method can be used. Krug said that the technology could only be used to the depth at which a backhoe can reach. One participant noted that some people claim that backhoes can reach as far down as 85 feet, but he questioned whether biopolymer could actually be expected to hold back 85-foot-deep sidewalls. Krug recommended deferring the question to a geotechnical contractor.

Cost Issues

One meeting attendee questioned some of the claims that had been made about the cost efficacy of PRBs. He noted that several of the meeting's speakers had stated that PRB technologies are much more cost-effective than pump-and-treat technologies. He questioned this, saying that one of EPA's reports does not suggest such a finding. Vidumsky said that he thought he knew of the report that was being referred to. Upon closer analysis, Vidumsky said, the report indicates that the capital costs incurred by PRBs and pump-and-treat technologies are comparable, but that the O&M costs are not. In fact, Vidumsky continued, some estimates suggest that the O&M costs associated with PRBs are about six or seven times lower. The magnitude of the difference, he said, will be determined, in part, by the number of times a PRB's reactive materials must be replaced. Vidumsky said that some preliminary cost analyses suggest that PRBs will still be cheaper than pump-and-treat systems even if reactive materials have to be replaced once every five years. Several other meeting attendees expressed their belief that PRBs are more cost-effective than pump-and-treat systems. Karl Hoenke, a representative from Chevron, said that experience has taught him that there are many problems and hidden costs associated with pump-and-treat systems.

One meeting attendee said that he too believes that PRBs will prove to be more cost-effective than pump-and-treat systems, but he questioned whether the gap will be as large as suggested in some cost comparison studies. He reminded Action Team members that pump-and-treat systems have been improved over the last five years by incorporating pressure transducers and telemetry-based systems into their design. He said that these improvements have decreased the amount that must be spent on O&M activities.

Miscellaneous Issues

Several miscellaneous topics were addressed during the open discussion. These included:

• Regulatory acceptance. Richard Steimle said that many regulators have started expressing interest in PRB systems and that regulators are becoming more comfortable granting approval and issuing permits for PRBs. Steimle said that he believes that training programs have played a large role in facilitating regulatory acceptance of this technology. Another participant thanked the Action Team for their role in creating the PRB training course that is currently touring the country. He encouraged the Action Team to keep holding the workshops, saying that he thinks they help to eliminate misconceptions.

• Treatment trains. Steimle said that many regulators have started recognizing the benefit of using PRBs to treat and manage contaminated plumes. He said that some regulators have asked whether this technology could also be used to help eliminate source areas. For example, Steimle said, some people have proposed the following as a potential treatment approach: (1) mobilize contaminants in the source zone using in situ flushing or a thermal treatment, and (2) capture the mobilized contaminants with a PRB as they move downgradient. Steimle asked meeting attendees to comment on the viability of such an approach and on whether they thought that using PRBs as part of a treatment train would become a popular approach in the future. Several meeting attendees responded:
  
  

  —	Jackie Quinn said that some source mobilization technologies (e.g., steam injection technologies) would not be compatible with PRBs.
  —	John Vidumsky advised against using a source mobilization technology that would kill biological activity. In many cases, he said, PRB technologies and biological reduction work well together.
  —	Bob Gillham said that laboratory work suggests that iron can degrade chlorinatedsolvents in the presence of co-solvent flushing liquids. However, he said, the results indicate that the iron's reactivity is diminished. He also said that some researchers are trying to determine whether iron can be used to treat source zones directly. (He said that Texas A&M and DuPont are working together to evaluate this possibility and that the University of Waterloo is also starting to evaluate the feasibility of such an application.)

• Using PRBs to remediate pesticides and polychlorinated biphenyls (PCBs). As noted by one meeting attendee, PRBs are used mostly to address chlorinated solvent plumes. The attendee said that he would like to know more about PRBs' ability to treat other types of contaminants. Watson said that Oak Ridge National Laboratory has evaluated PRBs' ability to degrade PCBs. Another meeting attendee said that some studies have been performed on plumes containing DDT. Williamson reminded the meeting attendee that a funnel-and-gate system has been used to remove BHCs and DDT from ground water. (See Williamson's presentation summary for details.)

• Potential topics for the next PRB Action Team meeting. One meeting attendee suggested reviewing reaction pathways at the next PRB Action Team meeting. She suggested providing an overview of the different biological and chemical reactions that occur in the subsurface.

• PRB implementation: the role of private industry and the government. One meeting attendee noted that many of the barriers discussed during the meeting have been funded with government monies. He asked whether PRB technologies have gained much interest among the private sector. Moylan and Vidumsky said that they have, and that several barriers have been installed at privately owned sites. Still, panel members admitted, most barriers have been installed using government money. Moylan and Marsh said that this was appropriate because the government has many sites to remediate and it actually possesses the money to test innovative technologies. Marsh said that all tax-paying citizens will benefit if the government identifies faster, cheaper, and more effective ways to remediate the environment.

• Future projects and directions. Over the next couple years, Vidumsky said, he expects that groups will start installing PRBs to greater depths. It has already been shown that deep installations are possible, panel members noted; outside the United States, barriers have been installed to depths of 300 feet.

CLOSING REMARKS

Vidumsky closed the day's presentations and discussions by reiterating his enthusiasm for the future of PRB technologies. Also, he thanked speakers for their stimulating talks, meeting attendees for their excellent questions and feedback, and Connely GPM, Battelle Memorial Institute, and Environmental Technologies, Inc., for paying for the meeting's refreshments.

Attachment A: Meeting Attendees List

RTDF Permeable Reactive Barriers
Action Team Meeting

Hilton Melbourne Airport
Melbourne, Florida
February 16-17, 2000

Attendees

M. Talaat Balba Technical Manager CRA Services 2055 Niagara Falls Boulevard Suite 3 Niagara Falls, NY 14304 716-297-2160 Fax: 716-297-2265 Email: tbalba@craservices.com William Baughman Project Manager Cummings/Riter Consultants, Inc. Parkway Building - Suite 201 339 Haymaker Road Monroeville, PA 15146 412-373-5240 Fax: 412-373-5242 Email: wbaughman@cummingsriter.com Shawn Benner University of Waterloo 200 University Avenue, W Waterloo, Ontario N2L 3G1 Canada 519-888-4878 Fax: 519-746-3882 Paul G. Bucens Environmental Engineer Remediation Group, Inc. 62 Whitemore Avenue Cambridge, MA 02140 617-498-2667 Fax: 617-498-2677 E-mail: paul.g.bucens@grace.com Jim Bush Remediation Systems Manager Pacific Northwest National Laboratory P.O. Box 999 Richland, WA 99352 509-376-6555 Fax: 509-372-1704 Email: james.bush@pnl.gov Jean Caufield Program Manager, RCRA Corrective Action Remediation Team General Motors Argo A - 1004 H (MC 482-310-010) Detroit, MI 48202-3220 313-556-0845 Fax: 313-556-0803 Email: jean.e.caufield@gm.com Yoon-Jean Choi Geotechnical Engineer Technical Support U.S. Environmental Protection Agency 1 Congress Street - Suite 1100 (HBS) Boston, MA 02114-2023 617-918-1437 Fax: 617-918-1291 Email: choi.jean@epamail.epa.gov Manoj Chopra Associate Professor Department of Civil and Environmental Engineering University of Central Florida P.O. Box 162450 4000 Central Florida Boulevard Orlando, FL 32816-2450 407-823-5037 Fax: 407-823-3315 Email: chopra@mail.ucf.edu Jean Chytil Geologist Geotechnical Engineering & Sciences Branch Omaha District U.S. Army Corps of Engineers 215 North 17th Street (CENWO-ED-GH) Omaha, NE 68102 402-221-7788 Fax: 402-221-7769 Email: jean.m.chytil@usace.army.mil Lisa Clark Hydrogeologist RMT, Inc. 100 Verdae Boulevard Greenville, SC 29607 864-281-0030 Fax: 864-281-0288 Email: lisa.clark@rmtinc.com Christian Clausen Department of Chemistry University of Central Florida 4000 Central Florida Boulevard Orlando, FL 32816 407-823-2293 Fax: 407-823-2252 Email: clausen@pegasus.cc.ucf.edu Narenra Dave Chief Geologist ET Division Louisiana Department of Environmental Quality P.O. Box 82178 Baton Rouge, LA 70884-2178 225-765-0489 Fax: 225-765-0602 Email: narendra_d@deq.state.la.us Richard deVivero Southeast Region Manager Radian Remediation and Operating Services URS Corporation 315 East Robinson Street - Suite 245 Orlando, FL 32801 407-422-0353 Fax: 407-423-2695 Email: rick_devivero@radian.com Robert Edwards Research Scientist The Texas Center for Applied Technology The Texas A&M University 100 North East Loop 410 Suite 1200 One International Centre San Antonio, TX 78216 210-321-5136 Fax: 210-525-1801 Email: bob_edwards@sa.wpi.org Gerald Eykholt Assistant Professor Department of Civil and Environmental Engineering University of Wisconsin - Madison 3208 Engineering Hall 1415 Engineering Drive Madison, WI 53706-1691 608-263-3137 Fax: 608-262-5199 Email: eykholt@engr.wisc.edu Robert Focht Remediation Engineer EnviroMetal Technologies, Inc. 745 Bridge Street, W - Suite 7 Waterloo, Ontario N2V 2G6 Canada 519-746-2204 Fax: 519-446-2209 Email: rfocht@eti.ca Cherie Geiger Assistant Professor Chemistry Department University of Central Florida P.O. Box 162366 Orlando, FL 32816-2366 407-823-2135 Fax: 407-823-2252 Email: cgeiger@pegasus.cc.ucf.edu Robert Gillham Professor Department of Earth Sciences University of Waterloo 200 University Avenue, W Waterloo, Ontario N2L 3G1 Canada 519-888-4658 Fax: 519-746-1829 Email: rwgillha@sciborg.uwaterloo.ca John Greiner Project Manager Remediation Technology Conoco, Inc. 600 North Dairy Ashford Houston, TX 77252 281-293-5683 Fax: 281-293-3305 Email: john.f.greiner@usa.conoco.com J. Kyle Harris Superfund Consultant Environmental Remediation Exxon Mobil Corporation 800 Bell Street Room 4111D Houston, TX 77002 713-656-9213 Fax: 713-656-9430 Email: kyle.j.harris@exxon.com Mark Hemann Senior Hydrogeologist West Valley Nuclear Services P.O. Box 191 (MS-AOC-09) 10282 Rock Springs Road West Valley, NY 14171-0191 716-942-2213 Fax: 716-942-4473 Email: hemannm@wv.doe.gov Chris Herin GeoSyntec Consultants 621 N.W. 53rd Street, Suite 650 Boca Raton, FL 33487 561-995-0900 Fax: 561-995-0925 E-mail: chrish@geosyntec.com K.A. (Karl) Hoenke Senior Environmental Project Manager Chevron Environmental Management Compnay 6001 Bollinger Canyon Road San Ramon, CA 94583-0712 925-842-9259 Fax: 925-842-0213 E-mail: khoe@chevron.com Kevin Ignaszak Senior Environmental Engineer The Sear-Brown Group 85 Metro Park Rochester, NY 14623-2674 716-475-1440 Fax: 716-424-5951 Vinod Jayaram Project Manager Geo-Con 9942 Currie Davis Drive - Suite B Tampa, FL 33619 813-626-0751 Fax: 813-626-4049 Faruque Khan Office of Research and Development Subsurface Protection & Remediation Division U.S. Environmental Protection Agency Robert S. Kerr Environmental Research Laboratory P.O. Box 1198 Ada, OK 74820 580-436-8704 Fax: 580-436-8703 Email: khan.faruque@epa.gov Stephen Klein President Connelly-GPM, Inc. 3154 South California Avenue Chicago, IL 60608 773-247-7231 Fax: 773-247-7239 Richard Landis Development Engineer DuPont Barley Mill Plaza (27/2264) P.O. Box 80027 Wilmington, DE 19880-0027 302-892-7452 Fax: 302-892-7641 Email: richard.c.landis@usa.dupont.com Terry Liikala Hydrogeologist Environmental Technology Division Battelle Pacific Northwest National Laboratory P.O. Box 999 K6-96 Richland, WA 99352 509-376-4143 Fax: 509-372-1704 E-mail: terry.liikala@pnl.gov

Donald Marcus MacMarcus Resources 1771 Ide Court Thousand Oaks, CA 91362 805-379-1471 Fax: 805-967-1529 Email: macmarcus@aol.com Leah Matheson Environmental Engineer MSE Technology Applications, Inc. 200 Technology Way Butte, MT 59701 530-792-1092 Fax: 530-792-1092 Email: leahm@mse-ta.com Tim McHale Dover National Test Site Program Manager Air Force Research Laboratory Arnold Drive Extension Building 909 P.O. Box 02063 Dover AFB, DE 19902 302-677-4147 Fax: 302-677-4100 Email: timothy.mchale@dover.af.mil Wendy Morrison Senior Staff Geologist GeoSyntec Consultants 621 N.W. 53rd Street, Suite 650 Boca Raton, FL 33487 561-995-0900 Fax: 561-995-0925 E-mail: WendyM@geosyntec.com Robyn D. Neely Attorney Akerman, Senterfitt 255 S. Orange Ave, Suite 1700 Orlando, FL 32802 407-843-7861 Fax: 407-843-6610 Email: RNeely@Akerman.com Eric Obrecht Environmental Engineer Division of Environmental Remediation New York State Department of Environmental Conservation 50 Wolf Road Albany, NY 12233-7010 518-457-5667 Fax: 518-457-9639 Email: erobrech@gw.dec.state.ny.us Frederick Pohland Professor and Weidlein Chair of Environmental Engineering Department of Civil & Environmental Engineering University of Pittsburgh 1141 Benedum Hall Pittsburgh, PA 15261 412-624-1880 Fax: 412-624-0135 Email: pohland@engrng.pitt.edu Jacqueline Quinn Remediation Project Manager Environmental Program Office NASA Room 3035 - (JJ-D) Kennedy Space Center, FL 32899 321-867-8410 Fax: 321-867-8040 Charles Reeter Hydrogeologist U.S. Naval Facilities Engineering Service Center 1100 23rd Avenue (411) Port Hueneme, CA 93043 805-982-4991 Fax: 805-982-4304 Email: reetercv@nfesc.navy.mil Walter Richards Engineer Science Applications International Corporation 175 Freedom Boulevard Kevil, KY 42053 270-462-4556 Fax: 270-462-4120 Email: richardsw@saiceecg.com A. Lynn Roberts Associate Professor Whiting School of Engineering Department of Geography and Environmental Science Johns Hopkins University 313 Ames Hall 3400 North Charles Street Baltimore, MD 21218-2686 410-516-4387 Fax: 410-516-8996 Email: lroberts@jhu.edu James Romer Senior Remediation Engineer Remedial Design & Construction Science Applications International Corporation 800 Oak Ridge Turnpike P.O. Box 2502 Oak Ridge, TN 37831 423-481-4676 Fax: 423-481-4757 Email: james.r.romer@saic.com H.G. Sanjay Research Engineer ARCTECH, Inc. 14100 Park Meadow Drive Chantilly, VA 20151 703-222-0280 Fax: 703-222-0299 Email: envrtech@arctech.com Bruce Sass Senior Research Scientist Environmental Restoration Department Battelle Memorial Institute 505 King Avenue Columbus, OH 43201-2693 614-424-6315 Fax: 614-424-3667 Email: sassb@battelle.org Steven Schroeder Senior Project Engineer RMT, Inc. 100 Verdae Boulevard P.O. Box 16778 Greenville, SC 29606-6778 864-281-0031 Fax: 864-287-0288 Email: steve.schroeder@rmtinc.com Thea Schuhmacher Hyrogeologist GeoScience LFR Levine*Fricke 1920 Main Street - Suite 750 Irvine, CA 92614 949-442-7296 Fax: 949-955-1390 Email: thea.schuhmacher@lfr.com Stephen Shoemaker Senior Consultant DuPont Company 6324 Fairview Drive Charlotte, NC 28210 704-362-6638 Fax: 704-362-6636 Email: stephen.h.shoemaker@usa.dupont.com Timothy Sivavec Chemist General Electric Corporate Research and Development One River road Building K1-5A45 Niskayana, NY 12309 518-387-7677 Fax: 518-387-5592 Email: sivavec@crd.ge.com Aamod Sonawane University of Central Florida 12021 Solon Drive Apartment 116 Orlando, FL 32826 407-208-1744 Email: aamodash@yahoo.com Dominique Sorel Project Hydrogeologist Geomatrix Consultants, Inc. 2101 Webster Street 12th Floor Oakland, CA 94612 510-663-4161 Fax: 510-663-4141 Email: dsorel@geomatrix.com Richard Steimle Hydrogeologist Technology Innovation Office U.S. Environmental Protection Agency 401 M Street, SW (5102G) Washington, DC 20460 703-603-7195 Fax: 703-603-9135 Email: steimle.richard@l.epa.gov Jeff Troiano Senior Technical Specialist Environmental Quality Office Ford Motor Company 1 Parklane Boulevard Suite 1400E Deaborn, MI 48124 313-322-3890 Fax: 313-248-5030 Peter Vikesland Postdoctoral Researcher Department of Geography and Environmental Engineering Johns Hopkins University 313 Ames Hall 3400 North Charles Street Baltimore, MD 21218 410-516-5039 Fax: 410-516-8996 Email: pv15@jhunix.hcf.jhu.edu Stephen White U.S. Army Corps of Engineers 12565 West Center Road Omaha, NE 68144 402-697-2660 Fax: 402-697-2673 Email: stephen.j.white@usace.army.mil John Wilkens Research Associate DuPont Central Research and Development Experimental Station 304/A313 P.O. Box 80304 Wilmington, DE 19880-0304 302-695-3143 Fax: 302-695-4414 Email: john.a.wilkens@usa.dupont.com RTDF logistical and technical support provided by: Christine Hartnett Conference Manager Eastern Research Group, Inc. 5608 Parkcrest Drive - Suite 100 Austin, TX 78731-4947 512-407-1829 Fax: 512-419-0089 E-mail: chartnet@erg.com Carolyn Perroni Senior Project Manager Environmental Management Support, Inc. 8601 Georgia Avenue - Suite 500 Silver Spring, MD 20910 301-589-5318 Fax: 301-589-8487 E-mail: carolyn.perroni@emsus.com Laurie Stamatatos Conference Coordinator Eastern Research Group, Inc. 110 Hartwell Avenue Lexington, MA 02421-3136 781-674-7320 Fax: 781-674-2906 E-mail: lstamata@erg.com Elaine (Chipper) Whalan Eastern Research Group, Inc. 2200 Wilson Boulevard Suite 400 Arlington, VA 22201 703-841-0500 Fax: 703-841-1440 E-mail: ewhalan@erg.com

Attachment B: Speakers List

RTDF Permeable Reactive Barriers
Action Team Meeting

Hilton Melbourne Airport
Melbourne, Florida
February 16-17, 2000

Speakers

Grant Anderson Hydrologist U.S. Army Corps of Engineers P.O. Box 1715 Baltimore, MD 21203 410-962-6645 Fax: 410-962-7731 E-mail: grant.a.anderson@nab02.usace.army.mil Jeffrey Cange Remediation Manager Pease Air Force Base Remediation Project Bechtel Environmental, Inc. 151 Lafayette Drive P.O. Box 3500 Oak Ridge, TN 37831-0350 423-220-2255 Fax: 423-220-2108 E-mail: jbcange@bechtel.com Paul Dieckmann Staff Facilities Engineer Federal Manufacturing & Technologies Honeywell P.O. Box 419159 Department 173 - 1B31 Kansas City, MO 64141-6159 816-997-2335 Fax: 816-997-7361 E-mail: pdieckmann@kcp.com Arun Gavaskar Battelle Memorial Institute 505 King Avenue Columbus, OH 43201 614-424-3403 Fax: 614-424-3667 E-mail: gavaskar@battelle.org William Goldberg Chief Consulting Engineer MSE Technology Applications, Inc. 200 Technology Way P.O. Box 4078 Butte, MT 59702 406-494-7330 Fax: 406-494-7230 E-mail: goldberg@mse-ta.com Peter Kjeldsen Visiting Researcher Department of Civil Engineering North Carolina State University Campus Box 7908 Raleigh, NC 27695-7908 919-515-7699 Fax: 919-515-7908 E-mail: PK@imt.dtu.dk Thomas Krug Senior Environmental Engineer GeoSyntec 160 Research Lane - Suite 206 Guelph, Ontario N1G 2M5 Canada 519-822-2230 Fax: 519-822-3151 E-mail: tkrug@geosyntec.com David Lee Research Hydrologist Environmental Technologies Branch Atomic Energy of Canada, Lt. Chalk River Laboratories Chalk River, Ontario K0J 1J0 Canada 613-584-8111 Fax: 613-584-1221 E-mail: leed@aecl.ca

Russell Marsh Environmental Engineer U.S. Army Corps of Engineers P.O. Box 1715 Baltimore, MD 21203 410-962-2227 Fax: 410-962-7731 E-mail: russell.e.marsh@nab02.usace.army.mil John Moylan Senior Consulting Geologist URS Greiner Woodward Clyde 10975 El Monte Overland Park, KS 66211 913-344-1032 Fax: 913-344-1011 E-mail: john_moylan@urscrop.com Debra Reinhart Associate Dean College of Engineering and Computer Science University of Central Florida P.O. Box 162993 Orlando, FL 32816 407-823-2156 Fax: 407-823-5483 E-mail: reinhart@pegasus.cc.ucf.edu David Smyth Research Hydrogeologist Department of Earth Sciences University of Waterloo 200 University Avenue, W Waterloo, Ontario N2L 3G1 Canada 519-888-4567 Fax: 519-746-3882 E-mail: dsmyth@sciborg.uwaterloo.ca John Vidumsky Superfund Program Manager DuPont Engineering Barley Mill Plaza - Building 27 Lancaster Pike & Route 141 Wilmington, DE 19880-0027 302-892-1378 Fax: 302-892-7637 E-mail: john.e.vidumsky@usa.dupont.com David Watson Hydrogeologist Environmental Sciences Division Oak Ridge National Laboratory Building 1505 (MS 6400) P.O. Box 6038 Oak Ridge, TN 37831-6400 423-241-4749 Fax: 423-574-7420 E-mail: watsondb@ornl.gov Dean Williamson Senior Project Manager CH2M Hill 3011 Southwest Williston Road Gainesville, FL 32608-3928 352-335-5877 Fax: 352-335-2959 E-mail: dwilliam@ch2m.com

Attachment C: Poster Presenters

RTDF Permeable Reactive Barriers
Action Team Meeting

Hilton Melbourne Airport
Melbourne, Florida
February 16-17, 2000

Poster Presenters

Jean Chytil Geologist Geotechnical Engineering & Sciences Branch Omaha District U.S. Army Corps of Engineers 215 North 17th Street (CENWO-ED-GH) Omaha, NE 68102 402-221-7788 Fax: 402-221-7769 Email: jean.m.chytil@usace.army.mil Gerald Eykholt Assistant Professor Department of Civil and Environmental Engineering University of Wisconsin - Madison 3208 Engineering Hall 1415 Engineering Drive Madison, WI 53706-1691 608-263-3137 Fax: 608-262-5199 Email: eykholt@engr.wisc.edu Faruque Khan Office of Research and Development Subsurface Protection & Remediation Division U.S. Environmental Protection Agency Robert S. Kerr Environmental Research Laboratory P.O. Box 1198 Ada, OK 74820 580-436-8704 Fax: 580-436-8703 Email: khan.faruque@epa.gov Donald Marcus MacMarcus Resources 1771 Ide Court Thousand Oaks, CA 91362 805-379-1471 Fax: 805-967-1529 Email: macmarcus@aol.com

Leah Matheson Environmental Engineer MSE Technology Applications, Inc. 200 Technology Way Butte, MT 59701 530-792-1092 Fax: 530-792-1092 Email: leahm@mse-ta.com H.G. Sanjay Research Engineer ARCTECH, Inc. 14100 Park Meadow Drive Chantilly, VA 20151 703-222-0280 Fax: 703-222-0299 Email: envrtech@arctech.com David Watson Hydrogeologist Environmental Sciences Division Oak Ridge National Laboratory Building 1505 (MS 6400) P.O. Box 6038 Oak Ridge, TN 37831-6400 423-241-4749 Fax: 423-574-7420 E-mail: watsondb@ornl.gov Dr. Martin Wegner Mull und Partner Ingenieurgesellschaft mbH, Osteriede 5, 30827 Garbsen Germany 0049-5131-4694-59 Fax: 0049-5131-4694-90 E-mail: wegner_mull@compuserve.com

Attachment D: List of Poster Topics

RTDF Permeable Reactive Barriers
Action Team Meeting

Poster Topics

Installation and Operation of a Hanging PRB, Shaw AFB, SC
Jean Chytil and Ted Streckfuss, Dan Gravelding, Randy Rogers

Long-term Column Tests and Modeled Effects of Aquifer Heterogeneity on PRBs
Gerald Eykholt, C.R. Elder, C.H. Benson, and J.R. Fort

Field Evaluation of In-Situ Redox Manipulation for Remediating Chromium Contaminated Ground Water and Sediment
Faruque A. Khan, R.W. Puls, C.J. Paul, and M.S. McNeil

Results of the Reactant Sand-fracking Pilot Test and Implications for the In Situ Remediation of Chlorinated Vocs and Metals in Deep and Fractured Bedrock Aquifers
Donald L. Marcus

Speciation of Uranium in Samples From ZVI Permeable Reactive Barriers at the
Y-12 Plant Site at Oak Ridge TN, and Durango, CO
Leah Matheson

Longterm Observation of a Full-scale Permeable Reactive Barrier in Germany
Wilfried Möller and Martin Wegner

Mineral Precipitation in the Oak Ridge, Y-12 Iron Reactive Barrier
Debra Phillips, David Watson, and Baohua Gu

Humasorb™--A New Multipurpose Remediation Media to Remove Organic and Inorganic Contaminants
H.G. Sanjay, Amjad Fataftah, and Daman Walia

1. This report summarizes the presentations and discussions only. Attachments C and D list the poster presenters and poster topics, respectively.

2. Recently, some modifications were made to enhance the capture zone of the PRB system. These modifications were discussed in William Goldberg's presentation (see below). Although construction has been completed, the system has not restarted operations since the modifications were finished. Thus, all of the performance monitoring data presented in Watson's discussion apply to the original trench design.

3. Smyth provided a reference for the site near Sudbury, Ontario:

Benner, S.G., D.W. Blowes, and C.J. Ptacek. 1997. A full-scale porous reactive wall for prevention of acid mine drainage. Ground Water Monitoring and Remediation. V. XVII (4-fall). pp. 99-107.

4. Smyth provided two references for the Elizabeth City site:

Blowes, D.W., R.W. Puls, T.A. Bennett, R.W. Gillham, C.J. Hanton-Fong, and C.J. Ptacek. 1997a. In situ porous reactive wall for treatment of Cr (VI) and trichloroethylene in ground water. In Proceedings of the 1997 International Containment Technology Conference and Exhibition, February 9-12. St. Petersburg, Florida. pp. 851-857.

Blowes, D.W., R.W. Puls, R.W. Gillham, C.J. Ptacek, T.A. Bennett, J.G. Bain, C.J. Hanton-Fong, and C.J. Paul. 1999. An In Situ Permeable Reactive Barrier for Treatment of Hexavalent Chromium and Trichloroethylene in Ground Water: Volume 2 Performance Monitoring. EPA/600/R-99/095b (In Press).

5. Smyth provided a reference for the CFB Borden site:

Lauzon, F. 1998. In Situ Biodegradation of a Naphthalene Plume in a Funnel-and-Gate System. M. Eng. Thesis, Royal Military College of Canada.

6. Lee asked Reinhart how she could be sure that the acid wash had reached all parts of the iron column. If ground water had leaked into the system and penetrated the column, he said, the acid might not have been able to reach the column's deeper zones. Reinhart acknowledged that this was true. Lee suggested doing the following in the future: drive a hollow tube down to the column's bottom, inject it with acid, and raise and lower the tube.