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

Garden Plaza Hotel
Oak Ridge, Tennessee
November 17-19, 1998

WELCOME AND INTRODUCTIONS
Dr. Robert Puls, U.S. Environmental Protection Agency (EPA), National Risk Management Research Laboratory (NRMRL)

Dr. Robert Puls, co-chair of the Remediation Technologies Development Forum (RTDF) Permeable Reactive Barriers (PRBs) Action Team, began the meeting by welcoming participants (see Attachment C). The meetings are held every 6 to 8 months. This schedule will continue, and any suggestions for future locations and topics are welcome. Dr. Puls thanked Dave Watson from Oak Ridge National Laboratory (ORNL), Carolyn Perroni from Environmental Management Support, Inc., and Kimberly Coerr from Eastern Research Group, Inc. for their roles in organizing and supporting this meeting.

Dr. Puls provided an update of the RTDF's four main activities. He said the RTDF acts primarily as a coordinator of long-term research and projects. The RTDF also maintains and updates a Web site (www.rtdf.org) that describes various sites that have used PRBs. In order to update this Web site and add information about new sites or ongoing actions at existing sites, the RTDF made a form, which requests site information, available at the meeting. This past year, Dr. Puls noted, the RTDF completed a comprehensive document addressing PRBs using reactors other than zero-valent iron (ZVI) and treating contaminants other than chlorinated volatile organic compounds (cVOCs). This document will be placed on the RTDF Web site soon, but is currently available at www.epa.gov/ada/reports. A hard copy should be available in December and will be mailed to RTDF members. In addition to Web site updates, the RTDF is planning a training program, expected to begin in spring 1999. More information about this program was provided later in the meeting.

The U.S. Department of Energy (DOE) and Permeable Reactive Barriers (PRBs)
Mr. Skip Chamberlain, DOE

Mr. Skip Chamberlain started by commending the RTDF for its activities and work. Mr. Chamberlain is the headquarters lead for the Subsurface Contaminant Focus Area (SCFA), which is part of the Office for Science and Technology. His team has initiated projects involving containment, dense non-aqueous phase liquid (DNAPL), metals, and radionuclides. The projects his office funds are mostly for metals and radionuclide projects. In this presentation, Mr. Chamberlain discussed current and future DOE projects.

Current Activities
The Oak Ridge and Rocky Flats projects started at the same time. The intent was to have these serve as test beds for PRB technology and to evaluate different materials. Activities at these sites are ongoing.

The Accelerated Site Technology Deployment (ASTD) initiative provides incentives for sites to deploy the available technologies. Under this program, an iron barrier was installed in Kansas City, Missouri, and there are plans to install an iron barrier in Grand Junction, Colorado, in spring 1999.

An in situ reduction-oxidation program is ongoing in Hanford, Washington, that involves using a medium similar to kitty litter in a barrier: the substance reacts with chromium IV and immobilizes it as chromium III. This project has been very successful and has won a Research & Development 100 award. This project required extensive and exhaustive review with the stakeholders, who were concerned that this site would turn into just another landfill. The stakeholders wanted insurance that the chromium III would not migrate to their property.

DOE has tested hydraulic fracturing in Portsmouth, Ohio. The SCFA is studying the possibility of using deep reactive barrier systems for contamination greater than 100 feet deep. In the past, DOE had discussions with DuPont about ways of jetting iron filings that deep, but no conversations about this project have occurred recently. DOE has also spoken to the company that produces Humasorb, a material that may be able to stabilize contaminants in deep ground water.

Mr. Chamberlain's other major project is obtaining funding to do a study on existing barriers and developing a design document. The intent is to gather information from the many barriers that have been installed, so that less research work needs to be done and the technology can be implemented more easily. This document would be agreed upon by the Interagency Technology Regulatory Cooperation (ITRC) group. This work has been started and some funding has been committed by different agencies.

An ASTD workshop will take place in Oak Ridge, Tennessee, in approximately 3 weeks. The intent of the workshop is to bring the technology users together with DOE and regulators to educate each group. This workshop will focus on the Oak Ridge, Tennessee, and Savannah River, South Carolina, sites. Two more workshops will be held at a later date, one in the Southwest and one in the Northwest.

Over the past couple of years, the Environmental Management Science Program (EMSP) has solicited proposals for basic science work within the subsurface. They will put out another call for proposals on January 1, 1999.

The Office of Energy Research has a program which focuses on bioremediation of metals and radionuclides. There is a project to study microbial actions in PRBs under their direction.

In spring 1998, there was a meeting with Florida State University and an environmental group affiliated with NATO. This meeting focused entirely on reactive barriers. One goal was to bring together the work accomplished in the United States and Europe.

Future Activities
Future activities will focus on material work, such as researching different materials to use for reactions and methods for installing those materials deep into the subsurface (100+ feet deep). The most important aspect of this effort will be evaluating long-term monitoring issues. Mr. Chamberlain said that this is critical to successfully implementing PRB technology. Currently, EPA Forum members are organizing a 1999 conference on long-term monitoring. DOE has invested research and project effort in potential cost savings for long-term monitoring. They are considering new micro-sensors and monitoring methods.

Another ongoing activity is work on the Hanford, Washington vadose zone problem. Two workshops have been held and a third is scheduled. The first workshop addressed characterization, the second addressed modeling, and the third will address containment, remediation, and monitoring. In conjunction with the workshop, DOE is compiling a catalogue of all the work ongoing at federal agencies related to vadose zone contamination. This publication will define the state of the science and identify further investigation needs. The National Academy of Science is compiling a similar document for all subsurface zones, including the ground water zone and the vadose zone.

Markets for PRBs and Other Innovative Remediation Technologies: National Research Council's (NRC) Findings
Ms. Jacqueline MacDonald, NRC

Ms. Jacqueline MacDonald presented results from a NRC (part of the National Academy of Sciences) study questioning why there are few ground-water remediation alternatives to pump and treat systems, since the ability of these systems to achieve cleanup goals is limited. This study indicated that remediation technology companies were weak economic performers, and that this fact was significant in driving alternative technology development.

NRC identified three reasons why there was a weak remediation technologies market. These reasons include:

• A lack of incentives. In most situations, the "customer" for remediation technologies is reluctant to begin remediation. Financially, it is preferable for companies to delay cleanup activities for as long as possible rather than have large capital expenditures on remediation. This means that developers of technologies cannot predict when their product will be purchased and when returns on their investment will occur.
• A fragmented regulatory system. A developer must deal with multiple regulations for testing and implementing a technology, including multiple local, state, and federal requirements. Since each program has different requirements, there is no clear path for having a technology approved for use.
• Limited cost and performance data. This is not as much a deterrent as the other two, but a customer will be hesitant to select a technology if there is minimal data proving that the technology is better and cheaper than other options.

Based on these findings, the NRC developed a series of recommendations addressing market weaknesses and encouraging alternative treatment technologies. To address the lack of incentives, the NRC recommended, the government should create financial incentives to address contamination. This includes requiring full reporting of all environmental liabilities and removing reporting loopholes. A company would be under pressure to remove the environmental liability and improve its financial standing. For companies with excessive liabilities, a fraction of these liabilities could be reported annually, somewhat like a mortgage payment.

The NRC recommended that EPA lead an effort to bring consistency to the regulatory process and define a clear pathway for approving and implementing new technologies. Evaluating use of national standards was also recommended. Another option is to institute a national program modeled after the Massachusetts Licensed Site Professional (LSP) program. A LSP in Massachusetts can make decisions about appropriate soil remediation without prior approval from the state Department of Environmental Protection (DEP).

The NRC recommended several programs for improving distribution and availability of cost and performance data. Different agencies and groups maintain databases containing cost and performance information. The scope and compatibility of these databases could be increased to improve information sharing. In turn the information within the databases should undergo peer review to improve the credibility of performance data. Verifying technology performance could be done by establishing a national review program that asks key questions about the technology, such as "does the technology reduce risk?" and "how does the technology reduce risk?" As a result of the verification process, short (1-page) summaries could be written and posted, similar to the site summaries available at the RTDF Web site. Once a technology has been verified, demonstrations and site-specific testing could be minimized for highly treatable sites. Costs would be easier to compare if template sites are created. These template sites would represent a range of possible site conditions and provide comparisons of different technologies.

Ms. MacDonald also discussed the NRC's current report for DOE sites. This report evaluates remediation options for DNAPL, metals, and radionuclides in soil and water. The NRC is also evaluating the SCFA, which Mr. Chamberlain heads. This report is due in June. Further information can be found on the National Academy Press Web site (www.nap.edu).

Efforts to Establish RTDF European Group and PRB Technology Development in Europe
Dr. Liyuan Liang, University of Wales

Dr. Liyuan Liang's presentation focused on the history and process of European efforts to develop a group similar to RTDF. In February 1998, the NATO Committee on the Challenges of Modern Society (CCMS) met in Vienna to discuss the program on Evaluation of Demonstrated and Emerging Technologies for the Treatment of Contaminated Land and Ground water. A special technical session on treatment wall technology was presented, with topics including reactive materials, construction of reactive walls, and full-scale projects. Dr. Liang said that meeting participants expressed strong interest in laboratory and field studies on dechlorination and metal immobilization. Dr. Liang noted that a forum with a technology development and demonstration focus was much needed even though there were already a number of European networking initiatives for contaminated land issues. Dr. Liang said participants at the Vienna meeting agreed to investigate the potential for European collaboration and how this might link with efforts in the United States. In June 1998, a meeting was held at the Dutch Organization for Applied and Natural Sciences Research (TNO) in Appeldoorn, Netherlands with nine experts attending from Belgium, Denmark, France, Germany, the Netherlands, and the United Kingdom. The attendees agreed to form the European Treatment Zones Team (ETZT), a team that mirrors the North American RTDF PRB Action Team. Dr. Liang said the ETZT promotes the use, research, and development of this technology in Europe and communicates with North American experts.

Dr. Liang said that ETZT identified several areas requiring further research. These areas included developing reactive materials, sequencing processes in the treatment zone, designing and modeling hydraulic systems, evaluating technical solutions for depths greater than 10 meters, gaining experience in field-scale demonstrations, developing cost-effective and accurate monitoring, and integrating treatment systems. The data needs dictated the mission statement developed by ETZT. Their first mission is to promote the implementation of treatment zone technologies in Europe. ETZT also promotes public, regulatory, and economic acceptance of these technologies. Other ETZT missions include initiating and supporting development, as well as transferring the technologies to other countries. Overall, the group's mission is to develop a forum for information transfer rather than to conduct research itself. Planned ETZT activities include (1) promoting the group to increase recognizability, (2) creating a database of European projects, (3) developing a state-of-the-technology report, and (4) creating an ETZT Web site that features a database of PRB information. Funding for these projects is being pursued.

After describing ETZT, Dr. Liang provided information on different treatment technologies used in several European countries. In the Netherlands, the TNO has several projects treating chlorinated hydrocarbons, oils, and chlorinated pesticides with bio-activated zones or biosparging. Two of these projects are full-scale demonstration projects. In Germany, there are several ongoing projects, ranging from reaction mechanisms to field design and construction projects. These projects have led to several patents for technologies such as solid free trench technology, treatment barrier construction, slurry wall technology, and overlapping boreholes. One German organization is investigating bioremediation and abiotic remediation methods. Denmark has strongly supported investigations on environmental treatment technologies. Studies in this country include ZVI-canister soil venting, forced leachate systems, funnel-and-gate systems, ZVI treatment barriers, and laboratory experiments focusing on iron aging. In the United Kingdom, bioremediation and natural attenuation methods are used because of cost concerns. The French government does not strongly support research, but does focus on implementation. In Belgium, the regional environmental regulations have encouraged in situ remediation technologies, but very little remediation has been completed.

In conclusion, Dr. Liang stated that the ETZT is seeking to collaborate with the RTDF. A meeting will be held in France in May 1999 and the participation of RTDF members is encouraged.

Update of RTDF Training Program
Mr. Scott Warner, Geomatrix Consultants, Inc.

Mr. Scott Warner announced that the RTDF is planning a series of training programs offered throughout the country (divided by EPA region) over the next 12 to 28 months. The intent of these training programs is to educate federal and state regulators about available PRB technologies, which in turn eases technology use. This is a comprehensive program with subjects covering a wide range of topics from "what is a PRB?" to "how is a PRB constructed and used?" The program is designed to be interactive; for example, subgroups may be asked to go through the design, permitting, and construction process for a PRB at a chosen site. Mr. Warner presented the tentative schedule for conducting the training program in 10 cities in the next 18 months. The audience was asked to give the RTDF any thoughts or ideas about the program. Mr. Warner also asked for volunteers to serve as instructors when the program is in their region. In addition, he asked the audience to announce this program to other interested parties.

PANEL DISCUSSION ON LONG-TERM PERFORMANCE OF PRBs
Mr. John Vidumsky, DuPont

Mr. John Vidumsky, co-chair of the RTDF PRBs Action Team, introduced the panel discussion. The topic of the panel discussion is long-term monitoring of the effectiveness of PRBs. This is a timely topic as the RTDF is working to improve acceptance of PRBs and, since they represent relatively new technologies, there is limited data about the long-term effectiveness and cost of this technology.

EPA Projects
Dr. Robert Puls, EPA-NRMRL
Mr. Timothy Sivavec, General Electric Corporate Research and Development

Dr. Puls began the overview of EPA projects by discussing the RTDF's accomplishments. When the group was created, long-term performance and accompanying concerns were identified as the most pressing issues that limited PRB use. Funding to address these problems has been obtained and studies are ongoing. Dr. Puls provided an overview of the research conducted under this funding effort.

The primary issue being studied is the decrease in barrier performance over time. The research objectives are to develop testing requirements that can predict PRB longevity, monitoring programs that can warn of PRB degradation, and long-term operations and maintenance protocols that reduce costs. The attractive feature of the PRB technology is that after the barrier has been installed, operation and maintenance costs should be minimal. In researching PRBs, the research approach has been to investigate data needs for geochemical, hydrological, and microbiological variables that affect PRB performance. EPA, DOE, and the Department of Defense (DOD) have worked together over the past year to address research needs.

The RTDF, with funding from EPA, is conducting research at three sites: Elizabeth City, New Jersey; Denver Federal Center, Colorado; and Somersworth Landfill, New Hampshire. The focus of RTDF's studies has been examining sites with similar types of iron in the barrier, but of varying geology, hydrology, and barrier age. Dr. Puls discussed Elizabeth City as an example. At this site, issues related to iron reactivity have been the research target. Quarterly ground-water sampling is conducted: geochemistry, water quality, and contaminant concentrations are examined. The sample analysis results are reviewed to identify changes over time. Some examples of trends observed in ground-water are a decline in Eh and an increase in pH in the barrier. Values generally recover beyond the barrier. Dr. Puls has not observed much variability over the 2.5-year study. The study also involves collecting core samples from the barrier. In core samples, trends in biomass, total organic carbon, calcium carbonate, and sulfur-containing compounds (e.g., sulfite and sulfate) have been observed, but data are limited so generalizations cannot be made.

Dr. Timothy Sivavec discussed the results of an 18-month pilot study conducted at Somersworth Landfill. This study evaluated the long-term performance of a 100 percent granular iron zone using (1) pumping tests to provide information on the distribution of hydraulic conductivity near the iron zone, (2) VOC monitoring upgradient, within, and downgradient of the PRB, (3) monitoring ground-water parameters (pH, dissolved oxygen, ORP, and SC) and inorganic parameters (e.g., metals, major ions, and nutrients), (4) microbial characterization of soil and iron, and (5) surface characterization of cored iron material.

In December 1996, a funnel-and-gate barrier system was installed at the Somersworth Landfill using an 8-foot diameter caisson to form a conical-shaped reactive iron zone. Pea gravel was added upgradient and downgradient of the iron, providing an iron flow-through thickness of 4 feet. Low permeability funnels were installed on each side of the gate using a 4-foot diameter auger and soil-bentonite mixture. The chemicals of concern at the site include the VOCs trichloroethane (TCE), tetrachloroethene (PCE), dichloroethene (DCE) isomers, and vinyl chloride.

Over the 18-month study, VOC reductions of 50 percent were observed between the upgradient aquifer and entrance to the PRB. This indicates that biodegradation is occurring. VOCs were reduced to non-detect levels at the first monitoring point in the PRB, approximately 14 inches downgradient from the entrance to the barrier. Bicarbonate, calcium, magnesium, iron, manganese, and sulfate reductions were also observed within the iron zone. As ground water exited the iron zone, the concentrations of these constituents increased to near their natural levels due to mixing with unaffected ground water and the natural buffering effects of the aquifer materials.

One concern for long-term effectiveness is loss of porosity from mineral precipitation within the iron zone, which can be quantified from ground-water inorganic profiles. Bulk digestion analysis and X-ray photoelectron spectroscopy (XPS) of iron cores were used to quantify calcium carbonate and ferrous carbonate precipitates. As observed in other 100 percent iron PRBs, mineral precipitation was highest at the upgradient iron/pea gravel interface and decreased to background levels within the first 6 inches of iron. An approximately 3 percent porosity loss was measured, which is less than predicted by laboratory and pilot-scale column studies. Another concern for long-term performance is biofouling. During the pilot study at this site, microbial growth within the barrier was equal to that in the surrounding aquifer. Sulfate-reducing bacteria were also identified within the iron zone. Dr. Sivavec said that he does not believe biofouling will affect short-term performance.

Dr. Sivavec said that future efforts will focus on further developing cost-effective, long-term monitoring techniques and protocols to minimize operation and maintenance costs. In particular, one proposed approach uses multi-parameter ground-water monitoring probes, each dedicated to individual wells. These probes measure dissolved oxygen, pH, ground-water elevation, conductivity, and temperature. These probes may be linked to an online Web site so that real-time data can be obtained quickly and easily.

At the conclusion of Dr. Sivavec's presentation, audience members asked about design parameters, microbial activity, and the impact of silica to the system. Dr. Sivavec noted that the design parameters, particularly ground-water velocity, differed from values measured in the field. Ground-water velocity was approximately 25 percent of that expected, possibly due to soil densification caused by the caisson installation method. There are limited data about the microbial community composition or activity, and long-term trends cannot be established. Audience members stated that they are also working on issues related to biofouling and microbial activity in PRBs and feel that precipitation is a greater threat to long-term performance than biofouling. Dr. Puls and Dr. Sivavec have not studied silica previously, but they noted that silica analysis will be conducted in the future.

ESTCP Project
Mr. Charles Reeter, Naval Facilities Engineering Service Center

Mr. Charles Reeter briefly presented information about activities at the Navy's Moffet Field, California, site and discussed future activities planned by DOD. A funnel-and-gate system was installed in April 1996 at Moffet Field to address a solvent plume in ground water. The Navy collected performance, monitoring, and cost-benefit data, planning ultimately to write a technology evaluation report for public distribution.

Sampling was completed over six continuous quarters. VOCs were reduced to non-detect levels at the downgradient edge of the PRB. Over the course of the study, gaseous ethane was produced, but dissipated downgradient of the PRB. Microbial growth was observed in the downgradient aquifer, but was not significantly elevated within the iron barrier itself. Over the 2.5-year period, the study proved that the funnel captures ground-water flow, VOCs were reduced below maximum contaminant levels (MCLs) or detection limits, gaseous byproducts were produced, and some precipitates were formed; the PRB proved to be cost-effective. The final demonstration report has been published. A summary cost and performance report will be written in the next month and available at ESTCP and NFESC's Web sites.

Mr. Reeter then addressed the long-term studies being undertaken by DOD. He noted that DOD goals are similar to EPA and DOE goals, since the three agencies have been collaborating on their efforts. The agencies have focused on the following concerns: assessing multiple design and construction techniques, revising an existing design and construction guideline, establishing consistent sampling and analysis procedures, and developing modeling programs. Overall, long-term performance and life expectancies of barriers are unknown.

To address some of these concerns, DOD developed research programs for the next 3 years. These programs will evaluate short- and long-term performance criteria, identify predictive modeling programs, and establish standardized sampling and analytical procedures. Another goal is to develop monitoring programs that will serve as early warning systems for barrier failure. A final technology report on performance and longevity will be prepared. The intent of this research is to gain widespread regulatory acceptance of PRB technology and encourage its use. In summary, Mr. Reeter stressed the need to address issues of long-term performance in order to successfully implement PRB technology.

DOE Projects
Mr. Nic Korte, ORNL

DOE is primarily studying biological issues, metals contamination, and hydraulics as related to DOE sites. Mr. Nic Korte presented data from laboratory experiments that show increased microbial growth in iron barriers. These data indicate that a high carbonate concentration and a steady pH result in greater microbial growth. DOE is hoping that using these results, and results from studies conducted by other groups, a design guideline can be developed.

Mr. Korte presented data from studies of metals contamination. One study indicated that a pure precipitated phase was not formed during reactions. This study speculated that hydroxy polymers were being formed. Researchers have been speculating about the actual form of precipitates, such as reductive, absorption, or inclusion precipitates. This has important impacts on capture of radionuclides, which is of concern at DOE sites, and the potential for radionuclides to migrate out of the barrier in precipitates. The apparent loss of collides or soluble metal complexes from a barrier may be a result of either migration from the barrier or sampling techniques. These issues need to be resolved before PRBs can become an accepted technology at sites with radionuclides.

The other area of concern is barrier hydraulics. One study questioned whether tracer tests provide enough data to answer questions about a site. This leads to the need for guidance focusing on tracer tests and how they should be used. Mr. Korte discussed the need for borehole flow meters that are easy to use and give accurate results. DOE developed a colloidal borescope, which is a camera that can be submerged in a well. The instrument can gather flow direction, flow velocity, and collide density. This instrument can be used to look at relative changes around the barrier. Some limitations are associated with this instrument because the borescope looks at only a small area, it produces data requiring interpretation by an onsite hydrologist, and it reports velocities higher than those found with other methods. Despite these limitations, there are still valuable applications because the tool provides an inexpensive way to examine hydraulics surrounding a PRB.

Mr. Korte noted that DOE has generated a schedule for the next 3 years.. During the first year, field work will continue. In the second year, testing will address issues identified in the first year. In the third year, DOE hopes to consolidate gathered data and develop predictive modeling programs. The overall objectives are similar to those mentioned by EPA and DOD: develop predictive testing requirements, monitoring methods, and a long-term testing protocol to minimize operation and maintenance costs.

One audience member questioned whether colloidal borescopes are affected by well screening. Mr. Korte indicated that this issue has been studied and that well screening does affect results. If well screening varies across a site, this could confound the data and be a limiting factor as to how useful this tool could be.

SERDP Projects
Dr. Lynn Roberts, Johns Hopkins University

Dr. Lynn Roberts discussed a project she is beginning in Spring 1999 that looks at the influence of contaminants on the longevity of iron-based PRBs. Results discussed in her presentation are from an initial study completed as a student's post-doctoral research. There are four objectives to the study:

• Evaluate the effect of ground-water composition on long-term performance
• Develop a prototype electrochemical probe to monitor the barrier performance
• Relate changes in performance to changes in iron surface composition
• Incorporate data to guidelines for predicting long-term performance

Factors that may influence barrier reactivity and performance include passivation of reactive surfaces, formation of corrosion pits, generation of reactive surface species, and decrease in pH. These processes may either increase or decrease reactivity in a barrier. The relative importance of the processes and their impact on performance are not well understood. Preliminary data show that certain constituents can influence the reactivity of iron. Dr. Roberts presented data that indicated that, for 2-nitrotoluene, the decay constants increased with addition of chloride and decreased when silicate was added. This has implications for barriers placed in areas where silicate may be present. The upcoming study will follow up on these findings by examining the influence of constituents on both increasing and decreasing iron reactivity rates. Different minerals in ground water may uniquely affect reactivity with different contaminants. Transport variations can also be influenced by geochemical interactions, and these impacts will also be the subject of investigation.

During post-doctoral work, a student developed an ex situ probe constructed of a single grain of iron. The results found from using this probe indicate that impedance spectra are relatively simple and are sensitive to the known corrosion-promoting effects of chloride. This method may be valuable in characterizing commercial iron used in PRBs. Dr. Robert's research will compare spectral information with rates of reductive dehalogenation in column experiments.

The technical approach revolves around using a series of long-term column experiments to evaluate the impacts of ground-water constituents on iron reactivity. Tracers will be periodically introduced into some columns to measure residence time. In other columns, there will be a continuous input of different organohalides. By using the information gathered, a degradation rate constant can be determined. In further studies, more detailed investigations will evaluate effects of different minerals and organic matter.

In addition, Dr. Roberts and a student are developing in situ probes from the ex situ probes. The in situ probes are composed of several electrodes with individual iron grains. These probes will placed in columns to monitor corrosion potential. Several iron grains are being used so that consistency across iron particles can be evaluated. The researcher's ultimate goal is to relate electrochemical data to changes in reaction rate constants.

A final aspect of the study involves removing iron grains from the columns to complete surface chemical studies. The study will take advantage of the vacuum environment to synthesize individual iron phases. Naturally, multiple phases may be present; this complicates finding the correlation between surface reactivity and iron composition. The results of the surface character studies will be related to changes in reactivity seen in the column experiments. Dr. Roberts is also hoping to use these surface characterization techniques on field samples from PRB field sites.

The primary aim of the study is to improve basic understanding of the impact of aqueous chemistry on the longevity of iron, from an aging and clogging perspective. Basic information about iron reactivity, response by the electrochemical probes, and surface characterization will be gained. In addition, Dr. Roberts anticipates that the electrochemical probes will be developed into useful field monitoring devices. The study results will be used to create guidelines that outline reasonable safety factors for residence times and barrier longevity. Results will be available to the technical community through journal articles and conference presentations.

An audience member asked why nitrogen will be used in the columns, and others questioned the number of columns and the type of iron being used in the study. Dr. Roberts replied that they are trying to imitate a field situation and are using the nitrogen to sparge oxygen. Only six columns are being tested simply because of funding limitations; the type of iron Dr. Roberts is using was selected because it is commercially available and commonly used. A way to examine different irons would be to study coring samples received from different sites. This would also consider different ground-water composition that influences the iron.

Open Discussion
Mr. John Vidumsky, DuPont

Mr. Vidumsky led an open discussion between panel members and the audience members based on presentations from earlier in the day. The first question raised by an audience member was "What is long-term?" Panel members replied that "long-term" has different meanings for different investigators. Chemical changes begin happening as soon as barriers are installed and an important question is how do chemical changes impact a barrier's performance, which many people are hoping extends for at least 30 years. The longest-lasting PRB has been active for only 5 years. Another issue to evaluate in considering long-term performance is the incorporation of safety factors and design considerations.

Cost estimates and performance data were discussed. Assumptions used in drawing cost conclusions are important to consider when viewing cost analysis and performance data. These include the time period over which costs are averaged. Most cost estimates average cost over 30 years. However, the actual maintenance costs after multiple years are unknown, because no barrier has been in place more than 5 years. One audience member who works with industry to conduct remedial actions encouraged the panel and research groups to continue examining cost reduction methods.

Of concern to all was whether a barrier will fail quickly or slowly once it begins to fail. Most panel members believe that a barrier will fail slowly, with the exception of barriers in aerobic aquifers. The first indication that a barrier is ceasing to function may be increased contaminant concentrations at downgradient monitoring locations. Dr. Puls believes that clogging may create preferred flow paths that would decrease residence time within the barrier. This discussion of barrier failure led, in turn, to a discussion of risk factors and how to insure continued protection of the environment. The answer to that question may lie in monitoring locations and, in some instances, source removal.

SITE-SPECIFIC PRESENTATIONS OF DOE PROJECTS

Field Treatability Studies: Permeable Reactive Treatment Wall Project - Monticello, Utah
Dr. Stan Morrison, Roy F. Weston

Dr. Stan Morrison described a DOE project in Monticello, Utah, using a PRB to remediate a plume of uranium, vanadium, selenium, molybdenum, and arsenic released from milling piles. At the last RTDF meeting in Oregon, Dr. Morrison discussed laboratory studies that evaluated various types of reactive materials. Dr. Morrison discussed results of the field treatability study and plans for installation.

Dr. Morrison said that the barrier will be installed between bedrock in an alluvial valley. Design parameters called for 600 tons of iron in a funnel-and-gate system, the gate being 100 feet long and 17 feet deep. ZVI was selected for use in the field treatability study. Columns 4 feet high and 4 inches in diameter were used to test various brands of ZVI. The residence time for each column was approximately 2 hours. Sampling results indicate that all contaminant concentrations, except molybdenum, dropped to below detection limits. All remediation goals were met with residence times under 6 minutes. Other solution parameters were also evaluated, including pH, alkalinity, calcium, iron, and sulfate. Of significance, the pH in this study remained stable; in many other studies using ZVI, pH levels changed as ground water flowed through the barrier.

Using data collected in the treatability study, Dr. Morrison conducted geochemical modeling to explain the observed chemistry. First, he said, equilibrium models were generated considering ZVI only, but these did not compare with actual field data. Then, he continued, equilibrium models were generated considering ZVI with mineral precipitates. Again, the modeled data did not compare with the field data. A third model examined the reaction path through which ZVI achieved equilibrium with the solution. With only a small amount of reaction, the model results compared well with the chemistry of the column experiments. Model correlation was improved using an iron-rich calcite phase. This phase was found by performing electron microprobe analysis of the solids in the column. Calcium concentrations declined through the column indicating that mineral precipitation occurred throughout the column, not just at the inlet. The model also accounted for precipitation of uranium in the columns.

In conclusion, the study revealed that several brands of ZVI react similarly. Contaminants have short residence times. Releases of iron and manganese were observed. At the Monticello site, manganese release is an issue because manganese is a contaminant of concern. However, released concentrations were below remediation goals. The chemistry of the site can be explained by partial equilibrium with ZVI and precipitation of an iron-calcium solution. The full-scale barrier installation is scheduled for winter or spring 1999.

One audience member asked if the researchers knew the structure of the carbonate formed. Dr. Morrison replied that the exact structure is unknown, but it is likely a calcite-like mineral based on stoichiometry observed. Another audience member asked if size of the ZVI particles varied from brand to brand. Dr. Morrison answered that the iron particles were close in size, but there was some variation.

Results of Onsite Comparison of Media for PRBs in Bear Creek Valley
Mr. Chuck Parmele, Science Applications International Corporation

Mr. Chuck Parmele worked on a project at Beaver Creek Valley, Tennessee. The goal was to test various media that might effectively remove contamination, as identified during Phase I studies. The site has a uranium plume in ground water from a former disposal pond. Three migration pathways transport uranium from the site. Testing was conducted along Pathway 1 and Pathway 2, the latter of which has the lower contaminant concentrations of the two pathways. There were four elements addressed in the Phase II study; Mr. Parmele's presentation examined treatment media, one of these four. The other elements are discussed in detail in the DOE study report.

DOWEX, ZVI, peat, and carbonaceous adsorbents showed promise in Phase I for effective removal. The Phase II study evaluated factors that would influence uranium migration, such as corrosion, precipitation, and uranium release. The four substances were tested in core samples--columns 3 inches in diameter and 3 feet long. (The DOWEX was tested in two 1.5-foot columns, because it has a higher reaction capacity and breakthrough capacity was of interest.) Flow from a well was diverted to these columns. Discharge water from the columns was collected and measured. Iron was dissolved in the column discharge, as evidenced by precipitation of iron oxides in the collection bins exposed to air. The columns were sampled two or three times a week for metals (including uranium), nitrate, suspended solids, and dissolved solids.

At Pathway 1, iron was more efficient in removing uranium than peat or DOWEX. Peat is an attractive option because of its low cost, but it does not have the needed removal capacity. The peat did remove some of the other heavy metals, such as nickel and cadmium. The biggest problem in this pathway was that high iron concentrations (100 to 300 milligrams/liter) were detected in the effluent. The study group identified additional technology needs, such as discussing the elevated levels of dissolved iron concentrations with regulators; developing a peat appropriate for use in large quantities; and further testing DOWEX capacity, which did not perform well in Pathway 1.

At Pathway 2, DOWEX proved to have superior capacity and was more economical than the ZVI and peat. ZVI was again more efficient than the peat. The carbonaceous adsorbents did not work and are no longer considered viable materials for a barrier at this site. DOWEX is also more economical than peat, considering its higher reaction potential. This increased capacity means that DOWEX could be used in ex situ treatment systems and disposed after its reaction capacity has been reached. This alleviates the concern of uranium breakthrough over time. The study group also looked at means to reduce dissolved iron concentrations in the effluent. Methods evaluated included treating effluent with peat, using a peat-ZVI barrier mixture, pretreating the influent, and controlling electrochemical factors. None of these methods significantly reduced iron concentrations in effluent. Technology needs include studying DOWEX use in canisters, determining regulatory criteria for release of dissolved iron, and demonstrating technologies for reducing iron in effluent.

An audience member asked about the DOWEX composition. Mr. Parmele stated that DOWEX is an ion exchange resin, which they found does not work as well as ZVI for contaminants with high ionic strength. Another audience member asked if the change in pH was responsible for the dissolved iron in the effluent and if regulating pH had an impact. Toward the end of the study period, Mr. Parmele answered, the researchers attempted to regulate pH through pretreatment. However, this portion of the study did not continue long enough to achieve sound results.

Dechlorination of Trichloroethylene by Electro-Enhanced Iron-Based Treatment Systems
Dr. S.Y. Lee, ORNL

Dr. S.Y. Lee presented information from a study he completed last year with Dr. Yul Roh, discussing problems encountered at the Portsmouth, New Hampshire, site. Dr. Lee believes an ideal reactive medium has a high reaction rate and capacity, is able to remain permeable, does not generate secondary contaminants, and is inexpensive. However, such a medium is not available currently. ZVI is currently used in many barriers because it is reactive, inexpensive, non-hazardous, and applicable to multiple contaminants. Disadvantages of ZVI are the slow dechlorination rates and uncontrollable reactions. To improve barrier performance, Dr. Lee and Dr. Roh applied a current to the iron. The benefits of applying current include an increase in dechlorination with minimal cost input, control of pH and Eh, reduction of ferrous iron in the effluent, and reduction of barrier size or iron volume needed. In addition, the iron can be regenerated. The electro-enhanced iron-based reactive system consists of an iron-based reactive medium, electron sources, a reactor vessel, and monitoring vessels. This technology can also be applied in funnel-and-gate or trench systems.

Dr. Lee and Dr. Roh tested three system configurations. The first system consisted of an iron cathode and anode system for TCE dechlorination with a passive collection system. Data was collected for different flow rates (1.7 to 4.41 milliliters/minute) in systems both with and without the added current. These experiments showed that dechlorination rates were approximately 10 times higher with a current than without. The second system is a semi-conductive system with a palladized iron oxide. Again, decreases in residence times and increases in dechlorination rates were observed. This system had a very short residence time; unfortunately, using the engineered medium may be expensive and would have limitations. This system is recommended only for reactor vessels or funnel-and-gate systems.. A patent is currently pending for this technology. The third system was developed for pH and corrosion control using ZVI. This system proved to be a simple, low-cost means to reduce ferrous iron discharge. The current application also enhanced removal capacity. This system is applicable for reactors, funnel-and-gate, and trench systems.

Dr. Lee concluded by saying that results indicated that current application enhances dechlorination, is required if a palladized iron oxide system is used, and reduces dissolved iron in effluent. These findings are significant in that these systems can treat high-flow-rate systems, reduce the required trench or system size, reduce problems associated with secondary mineral precipitates, and reduce capital and operation costs.

In other laboratory studies, one audience member noted high volumes of gas have been generated. Dr. Lee did not have this problem. The amperage used, he said, was likely too low to produce significant gas volumes; in addition, this study treated actual ground water versus distilled water. Another audience member suggested that the iron reduction is not from pH control, but rather is affected by how the iron is being used as a conductor. Dr. Lee agreed with this statement.

Regeneration and Long-Term Use of Palladized Iron
Mr. Nic Korte, ORNL

Mr. Korte has been conducting investigations with palladized iron and discussed the results of these studies. Palladized iron is produced by a reduction-oxidation reaction. Typically, the mixture is only 0.5 percent palladium, but there are still some outstanding issues surrounding the optimal palladium concentration.

Palladium has the ability to absorb hydrogen so that a hydrogenation reaction occurs when a chlorinated contaminant approaches the palladium. This reaction enhances dechlorination but also accelerates iron corrosion. Usually, reaction rate enhancement of 1 to 2 orders of magnitude is observed. A benefit of palladized iron is that it can degrade dichloromethane, which is not reactive with ZVI. This indicates that a unique mechanism is acting to dechlorinate contaminants. Polychlorinated biphenyls (PCBs) have also been shown to react with the palladized iron, although the reaction time is slow.

When it was noted that the reaction rates increased, the question about what happens with soil flushing fluids was raised. Palladized iron proved effective in methanol and ethanol environments. However, it does not work in solutions without water, and reaction rates decrease significantly as the percent of water in a solution decreases. Some limited work was conducted with DDE, toxaphene, and pentachlorophenols. Dechlorination at 50 percent ethanol seems be optimal and additional studies are planned.

Regeneration is the greatest issue involved in palladized iron use. In laboratory studies, breakthrough occurred at approximately 23 days, with a significant decline in removal efficiency observed after initial breakthrough. The system was rinsed with a simple hydrochloric acid solution and reaction capacity returned. There were minimal loses of palladium and iron during these rinses, but not enough to significantly impact performance. Column experiments required rinses every 10 to 12 days, and after numerous rinses a decline in reaction capacity was seen.

After success in the laboratory, palladized iron was applied under field conditions. Because of these field conditions, TCE concentrations varied widely during portions of the experiment. Removal rates were as expected initially, but the palladized iron did not recover after an acid rinse. There was concern that sulfur reduction had poisoned the palladium, but no reduced sulfur species were found in further investigations. To further assess palladium poisoning, water was pretreated. With pretreatment for sulfate, the palladium still was poisoned over time. With an ion-exchange pre-filter, removal capacity was not lost, but acid rinses were required daily.

In summary, a palladized iron mixture is easy to use and may have appropriate applications, such as working with batch systems. Mr. Korte visualizes such a mixture as a material to use in batch processing rather than placement in an underground barrier. More research is needed to find the optimal mixture. A palladium-iron mixture may prove more useful, although it would be more expensive than the palladized iron.

An audience member asked what would happen if the acid rinse concentration were increased. Mr. Korte responded that further optimization studies are needed, but higher acid concentrations may result in greater palladium and iron losses during the rinse. At the levels tested, palladium was not detected or detected at very low concentrations in the effluent; these concentrations were higher in the acid rinse. How to address the dissolved palladium and iron will be an issue of concern when using this technology.

Savannah River Site GeoSiphon Cell Demonstration and Deployment
Mr. Mark Phifer, Westinghouse Savannah River Company

Mr. Mark Phifer is studying alternative treatment systems: GeoSiphon and GeoFlow cells, both of which use natural head differences at a site to drive more flow through a treatment system than a trench or funnel-and-gate system. This technology can be configured as in situ or ex situ, in removable or permanent installations. GeoSiphon uses a siphon and GeoFlow uses an open channel to induce flow. Advantages of this technology over pump and treat systems include: in situ treatment, no power requirements, and lower operation and maintenance costs. Advantages over funnel-and-gate systems and trench PRBs include: ability to use existing wells or structures, accelerated cleanup with increased flow rates, and applicability to multiple site conditions. The GeoSiphon is limited to sites with shallow ground water (less than 25 feet) or artesian conditions because of the amount of lift that can be generated. In addition, measures must be taken to prevent air from entering the siphon line and stopping flow.

At the Savannah River site, a GeoSiphon configuration is in place to treat a TCE plume. The system consists of a 8-foot-diameter well filled with granular iron. A siphon in the iron induces flow through the iron in the well and discharges effluent to an outfall ditch, through which it eventually flows into the Savannah River. An air chamber was placed on the siphon line to collect hydrogen and methane produced in the system. Once installed, the system was pumped to regulate flow rates for the first 6 months. Testing of treated water indicated that concentrations of TCE were below 5 micrograms/liter (µg/l) after traveling 1.5 feet through the iron. At increased flow rates, dechlorination rates increased. Mr. Phifer speculated that this was caused by increased use of pore space at higher flow rates. The optimized siphon system was then set up and a flow rate of 2.7 gallons per minute (gpm) was established, with a head difference of 1.4 feet. A second cell and GeoSiphon line are being installed: these will use a greater head difference and higher flow rate.

The GeoSiphon and GeoFlow technologies are applicable to contaminants that can be addressed with a PRB. Both technologies are applicable only to sites with a head difference. The GeoSiphon technology is only applicable to shallow ground-water sites. Use of the GeoFlow technology is limited only by the depth at the site that can be reached by the installation equipment. Westinghouse Savannah Company has filed an international patent application and is in the process of obtaining a commercial license for the GeoSiphon and GeoFlow technologies.

An audience member asked about permit requirements for discharge. At the Savannah River site, the discharge is monitored monthly for TCE, PCE, DCE isomers, and vinyl chloride. A toxicity test is also conducted monthly. There is a National Pollution Discharge Elimination System (NPDES) permit in place.

Construction of Ground-Water Treatment Systems at the Oak Ridge Y-12 Plant and the Rocky Flats
Mound Plume Site
Mr. William Goldberg, MSE Technology Applications, Inc. (MSE)

Mr. William Goldberg described two PRBs installed by MSE, one at the Y-12 Plant site in Oak Ridge, Tennessee, and the other at the Rocky Flats Mound Plume site in Colorado. These PRBs were installed to demonstrate and evaluate PRB systems. The demonstration also served as a remediation method at both sites. Implementing a PRB as a demonstration results in high costs per square foot, but the cost per square foot is reduced for full-scale projects because of economics of scale.

At Oak Ridge, a funnel-and-gate system with wing walls 150 feet wide on one side of the gate and 75 feet wide on the other side of the gate has been installed. The walls are constructed of a high-density polyethylene (HDPE) membrane and are located in a trench. The gate is a concrete vault containing treatment canisters for evaluating different treatment media. The treatment vault consists of five vertically stacked reactors. An advantage to vertical reactors is the ease of cleaning and replacing used or clogged iron. The primary contaminants of concern are uranium, nitrate, and technetium. One problem at the site was slumping in the trench, so barriers were installed using a guar slurry for support. An enzyme breaker was used to digest the guar. The guar was recycled down the trench as construction progressed. Guar use increases the biological activities seen in the system. A process control system was designed, fabricated, and installed. This system was used to operate a treatment train, composed of various treatment media, designed and operated by ORNL.

At Rocky Flats, a different type of funnel-and-gate system was installed. This system consisted of a 225-foot continuous capture wall with pipe penetration through this barrier that directed flow to underground treatment vaults. The treatment system consists of two reactors containing a total of 60 tons of iron. The bottom 4 feet of the system at the inlet consist of pure iron topped with 1 foot of an iron/sand mix. The water level remains above the top of the iron/sand mix to prevent oxidation, and the reactor is designed so that the reactor does not drain during times of low or no flow. Ground water passes through the system and discharges to a nearby stream. The primary contaminants of concern are TCE and PCE. Slumping has occurred in the trench at this site and side walls have been installed at a slope.

Development and Operation of a Passive-Flow Treatment System for Sr-90-Contaminated Ground Water
Mr. Paul Taylor, ORNL

Mr. Paul Taylor discussed the development and use of a passive flow system using underground canisters to remove strontium-90 (Sr-90). The treatment system was placed at a free-flowing seep at former solid waste storage area 5 (SWSA-5) at ORNL. Surface flow rates varied with the season and contained high concentrations of Sr-90, approximately 25 percent of the Sr-90 contamination at ORNL.

Initially, the full-scale operation was designed as zeolite, a natural treatment medium that is selective for Sr-90, in underground troughs. However, the zeolite clogged almost immediately with iron hydroxide at the test scale operation, so this method was discontinued. Instead, pilot studies focused on a canister treatment system. The pilot-scale demonstration consisted of a modified 5-gallon bucket containing the natural zeolite. The system's design was such that the zeolite was submerged continuously. Clogging was also a problem in the canister pilot study. At the start of the day, flow was strong, but formation of iron hydroxide nearly stopped flow by the end of the day. Mr. Taylor maintained the system daily by decanting any dirt, debris, and iron hydroxide. It was thought that the full-scale system could be engineered to address the clogging problems. After treatment of approximately 3,000 bed volumes, significant Sr-90 breakthrough was seen. At the end of the 60-day pilot study, approximately 23 millicuries of Sr-90 were removed. The full-scale system construction began as the pilot scale study was ending.

A full-scale system is in place at ORNL to treat wastewater containing Sr-90. This is used as a comparison with the pilot study and seep treatment system. The wastewater contaminants are similar to the ground-water contamination, although there is more variation in the wastewater composition. This system is composed of two sand filter columns to remove solids and a zeolite system.. Researches found that changes in influent composition affect treatment capacity, and breakthrough occurs at approximately 3,000 bed volumes.

The full-scale system at Seep C was designed in parallel with the pilot scale system and began operation in November 1994. This system uses a french drain to collect and carry ground water to an underground treatment system composed of eight 55-gallons drums containing zeolite. The drums are arranged in four parallel series of two drums such that ground water runs through two drums for treatment. Filters are located at the inlet to prevent dirt from entering the system and clogging the zeolite. Treated water is discharged to a nearby creek. Flow and water level are continuously monitored. This system has operated for over 4 years and has treated over 3 million liters of water at a removal efficiency of 99.9 percent, approximately 1.6 curies of Sr-90. There is a continuing iron hydroxide plugging problem, which indicates that air is entering the system. In an attempt to solve this problem, a small amount of nitrogen gas has been added in order to sparge oxygen; this improved conditions, but did not eliminate the problem. Drums have also been changed in order to improve flow rates and address system clogging. Other possible solutions have included adding a filter and cleaning the french drain. The filter system did not work because the filter size needed was too small and the filer required continuous replacement. Cleaning the french drain still allowed iron to enter the system. As a result of the clogging, the french drain is continuously full and ground water flows around the french drain and the treatment system itself.

This study has had mixed success. The zeolite is effective in removing Sr-90, but ground water is flowing around the system and maintenance costs are higher because of clogging. An audience member asked if Sr-90 was seen in the precipitated iron hydroxide. Mr. Taylor reported that a simple precipitation of iron does remove approximately 50 percent of the Sr-90 in a solution. Another audience member asked if biological activity was studied and if this is contributing to system clogging. According to Mr. Taylor, very little biological activity was seen--as a contributor to system clogging, biological activity was of minor importance compared to the iron hydroxide problem. The audience also asked if other treatment media were studied. Only zeolite was used in the pilot study, since earlier studies examined different media and zeolite was chosen for its selectivity and capacity. At the full-scale system, approximately 20 to 25 drums of zeolite have been used. The drums are removed from the system because of iron hydroxide plugging, not because capacity has been reached.

Long-Term Performance Assessment of Zero-Valent Reactive Barriers in High-Sulfate Ground Water
Dr. Liyuan Liang, University of Wales

Dr. Liang presented the results of an ongoing study on ZVI barriers that were installed in high-sulfate ground water. Dr. Liang noted that she initiated the study while she worked at ORNL, but that Dr. West and others have since taken over and collected new data. Dr. Liang said the system is composed of a porous horizontal well within the aquifer that collects and transports the TCE-contaminated ground water to a treatment facility, where several media can be installed and tested. In March 1996, three iron media (master builder, peerless iron, and palladized iron) were tested. Gas buildup in the closed system required venting to prevent flow interruptions. This test indicated that these three media were equivalent in degrading TCE. The palladized iron was more effective in degrading TCE, as seen in the laboratory, but was poisoned over time. Master builder iron clogged after approximately 6 months because of clumping. Peerless iron did not clog and has performed better. Researchers believe this is a function of the different iron particle sizes.

In July 1997, three different iron media (peerless iron, cercona foamed pellets, and slow-diffusion potassium promanganite) were tested to determine if using a coarser material improves the life of the media. The study focused on examining the hydraulic life of the test media. Dr. Liang only discussed the iron media in the presentation. The foamed pellets are larger than the peerless iron, and potentially less prone to clogging. Pressure gauges were used to identify where clogging in the column occurred and microbial activity and surface analysis tests were conducted to identify the precipitates. The treatment train consisted of three parallel canisters with a flow rate of 0.3 gpm induced by gravity. Influent TCE concentrations were 100 µg/l and effluent concentrations were less than 5 µg/l. Daily and monthly sampling was conducted.

Pressure gauges indicated that the hydraulic conductivity changed in the first canister but remained stable within the next two canisters. The pressure drop occurred throughout the first canister for the peerless iron filings, indicating a systematic reduction of permeability throughout the media. The pressure drop occurred at the inlet to the first canister for the foamed pellets. That suggests that the clogging of the foamed material occurred at the inlet. Tracer tests indicated that effective porosity reduced over time, but that this occurred more slowly in the foamed pellets. Coring of the used peerless iron filings showed that filing size reduced over time. Particle size reduction is not as extensive in the foamed pellets. Finer materials were also collected and analyzed to identify the minerals precipitating in the columns.

Hydraulic data and physical characterization indicated that, in the iron filings, microbial activity accelerated iron corrosion and generated fines. Sulfate reduction and sulfide mineral precipitation also contributed to volumetric clogging in the canisters. Less bioactivity occurred in the foamed pellet canisters, but precipitation at the inlet caused clogging. Overall, the foamed pellets performed better in high-sulfate ground water conditions.

Innovative Methods for Hydraulic Manipulation of Ground Water Through Reactive Treatment
Zones at the Oak Ridge Y-12 Site
Dr. Baohua Gu, ORNL
Mr. David Watson, ORNL

Mr. David Watson described the hydraulic manipulation of ground water through barriers at the S-3 disposal pond at the Y-12 site in Oak Ridge, Tennessee. Other contaminant plumes, sources, and remediation methods at the Oak Ridge Y-12 site had already been described by presenters at this meeting. Past disposal of radioactive wastes in former disposal ponds has contaminated ground water with nitrate, uranium, technetium, and VOCs. Contaminated ground water migrates along three pathways; Dr. Gu and Mr. Watson conducted remediation studies along Pathway 1 and Pathway 2. Within Pathway 1, ground water flows into a trench, is directed through a treatment module, and is then discharged toward a creek. There is a significant upward gradient to ground-water flow in the competent shale bedrock, which causes a ground-water high in the trench. Seasonal changes in ground water affect capture area and ground-water flow directions.

More data is available for Pathway 2, which Mr. Watson therefore discussed in more depth. Pathway 2 consists of a 225-foot gravel trench without a membrane and with a 26-foot iron zone. Ground water flows through the gravel to the iron zone. The barrier was installed to prevent contamination from entering a nearby creek. Tracer tests were conducted with bromide and ferrous iron to examine ground-water flow and direction. Monitoring at several points in the iron found increasing pH within the iron and recovery to neutral conditions at the downgradient edge of the iron barrier. Specific conductance exhibited the same type of trend, with reductions in total dissolved solids within the iron. Calcium distribution showed the same trend, with concentration reductions within the iron. What caused these trends is unclear because flow data within the iron is lacking.

The study revealed that ground-water capture may be affected by vertical gradients, clogging from mineral precipitation or guar breakdown, and seasonal variations in ground-water elevation. Ground-water capture can be enhanced by variations in ground-water head, which can be manipulated by pumping or siphoning. Trench modifications are planned for later this year. These include extending the trench and installing a siphon line that discharges to a treatment reactor. The intent is to create a more aggressive capture zone with lower operation and maintenance costs. In the future, ongoing testing will occur at Pathway 1 and studies will continue at Pathway 2 to examine flow paths and assess the impact of microbial activity and the new trench configurations.

Treatment of Radionuclides and Inorganic Contaminants in Ground Water by Permeable Reactive Barriers
Dr. Baohua Gu, ORNL
Mr. David Watson, ORNL

Dr. Baohua Gu continued the discussion of the PRB along Pathway 2 at Y-12 site, specifically the chemical aspects and field operation of the technology used. The overall goal of the project was to demonstrate a technology that can treat a mixture of contaminants, including radionuclides, heavy metals, cVOCs, and nitrate. Initial laboratory studies indicated that different types of iron filings are similarly effective in removing uranium.

Ground water was monitored from wells upgradient, within, and downgradient of the iron barrier. Monitoring over time provided the following results:

• An initial increase in total iron concentrations (750 µg/l) was observed in some wells within the iron barrier. These levels significantly declined over time. Ferrous iron concentrations showed the same peak but at a lower level (160 µg/l). Dr. Gu believes this initial spike is a result of iron dissolution and corrosion related to the guar gum used in barrier installation.
• Nitrate concentrations were degraded within the iron barrier and in some downgradient wells, but remained relatively stable in the other wells sampled. After trench installation, an initial low nitrate concentration was also observed in some wells upgradient of the trench. Dr. Gu believes that these observations might be related to the use of guar gum and enhanced microbial activity.
• pH increased within the barrier overtime. A slight increase was observed in the immediately downgradient well. pH levels were stable in upgradient and further downgradient wells.
• Uranium concentrations remained stable in each well. Lower or non-detectable concentrations of uranium were observed within the iron barrier and some downgradient wells, indicating that the iron filings are acting to remove the uranium. Technetium follows this same pattern.
• Sulfate reduction was observed in the iron barrier. Dr. Gu believes the sulfate is being reduced to sulfide, as supported by sulfide analysis results of field and laboratory samples, or incorporated into mineral forms. A similar trend was observed with calcium and bicarbonate/carbonate.

In summary, ZVI is effective for removal of the contaminant mixture. Calcium and sulfate reduction that creates precipitates could be a long-term concern because of potential barrier clogging. Ferrous iron releases, biological fouling, and contaminant remobilization are long-term performance concerns that should be addressed. The audience asked if natural strontium was also removed. Natural strontium removal was not measured, but Dr. Gu believes that strontium would behave similarly to calcium because of their chemical similarity. An audience member asked what was required by federal and state regulatory agencies for removing and disposing uranium collected in the PRB. Once uranium has been remediated in ground water, the iron barrier will be excavated and disposed. Easy removal and disposal is a benefit of canister use.

Status of the Durango Permeable Reactive Treatment Walls
Dr. Stan Morrison, Roy F. Weston

Dr. Morrison provided a site update of a PRB installed at the Durango site in Colorado. Uranium mill tailings in a disposal cell are the source of uranium contamination in a seep at the toe of the cell. Water from the seep is collected in a gravel trench and discharged to a retention pond. After treatment, water in the retention pond is discharged to a small creek. In this project, flow was diverted from the retention pond to several subsurface reaction cells, each with unique configurations. Cell C is the primary operating system and contains ZVI foam bricks. Cell B began operation recently and contains steel wool and copper wool as the reaction media. Cell A and Cell D contain steel wool and have not been put in operation. Only one cell operates at a time because of flow restrictions.

In Cell C, the ZVI foam was formed as a brick with flow-through paths. Data from influent and effluent samples showed effective removal of uranium, molybdenum, and nitrate. A lag time between treatment and decrease in effluent concentration was observed for molybdenum and nitrate. Dr. Morrison suspects that microbial activity caused the delay in reduction. Only one sampling event has been conducted at Cell B. Some contaminant reduction was seen, but removal rates were not as high as those seen in Cell C. Dr. Morrison believes a delay, similar to that seen in Cell C, is occurring. Additional sampling will be conducted and will hopefully provide more information.

In April 1998, Cell C was excavated because the cell had clogged. When the tank was opened, Dr. Morrison observed that the top surface layer was oxidized and the lower surface was not. A gaseous environment had been created; hydrogen and methane had collected in the cell and continued to be released during sample collection. High concentrations of vanadium, uranium, calcium, and total inorganic carbon were found in some samples. Different iron mineral phases were detected during sampling.

Future plans include continued Cell B ground-water sampling, additional Cell C solid phase sampling, and kinetic experiments. The study team will close the main valve to collect water over the winter and allow for studies in the spring. Testing at Cell A and Cell D is also planned. The audience and Dr. Morrison discussed the fact that the foam bricks in Cell C could be a source of the silicate seen as a mineral precipitate; additional research is needed to find a more definitive answer. An audience member noted that dissolved oxygen measurements do not correlate with results of other studies. Dr. Morrison believes that this variation may result from the introduction of oxygen during measurement. A reduction in nitrate concentrations was observed. Dr. Morrison assumed that nitrogen was converted to nitrogen gas; another audience member noted that in studies he had conducted, ammonia was produced. Cell C exhibited good selenium removal in a high-sulfate environment, which is difficult to achieve in most systems. This treatment method may have applications elsewhere. Dr. Morrison stated that the researchers implemented this technology in an ex situ situation. The cost of using the foam bricks is likely comparable to that for other barrier media. An audience member asked why oxidation occurred only on the top surface of the bricks. Dr. Morrison explained that ground water was introduced above the bricks and would have lost its oxygen at the upper surface. Other audience members suggested that sulfate reduction was not seen because of high nitrate concentrations or residence times.

SITE UPDATES

From Demonstration-Scale to Full-Scale--Upstate New York Project Update
Dr. Diane Clark, Sterns & Wheler

Dr. Diane Clark provided an update of the first pilot study for a funnel-and-gate system and the process of moving to a full-scale installation, which was conducted in December 1997.. A former manufacturing operation at an Upstate New York site has caused shallow ground-water contamination with cVOCs. The aquifer is non-potable and separated from a drinking water supply by a clay layer. Flow rates in the aquifer are 42 to 85 feet/year.

The pilot test system consisted of a funnel-and-gate system with an iron reaction zone 10 feet wide, 16 feet deep, and 3 feet thick. The treatment zone retention time was 55 hours. Monitoring wells were installed upgradient of, within, and downgradient of the reaction zone. The pilot test continued for 6 months and revealed a cVOC concentration reduction: during the test period, all parameters met New York state standards at wells within or downgradient of the reaction zone. Angle core samples were collected 2 years after the pilot study started because of concerns about clogging. An approximately 10 percent reduction in porosity was found, and flow velocity had not significantly changed. No evidence of microbial activity was observed.

Based on results of the pilot study, full-scale construction began. At this time, a continuous barrier construction proved to be most economical, based on updated subsurface investigations showing increased plume width, varying clay depths, and residuals management. The final design called for a 1-foot-thick, 21-foot-deep continuous PRB with a second PRB at the area with the highest cVOC concentrations. Preliminary data indicate that cVOC concentrations are declining and pH is increasing, as seen in other studies. An audience member asked if there were problems encountered with till or boulders and use of the continuous trenchers. Dr. Clark stated that the subsurface consists of homogeneous sand and gravel, so no problems were encountered. Another audience member asked if changes in vertical and lateral flow direction affect the PRB's efficiency. Dr. Clark said the site does have flow changes, specifically the site floods during the spring snow thaw, but design did not need to consider this flooding.

Funnel-and-Gate--Coffeyville, Kansas
Mr. John Vogan, EnvironMetal Technologies, Inc.

Mr. John Vogan discussed a full-scale funnel-and-gate system installed in Coffeyville, Kansas. This is still one of the biggest systems installed to date, with a 500-foot slurry wall on either side of a 20-foot treatment zone that is 3 feet thick and 30 feet deep. The iron extends from 19 to 30 feet below ground surface, and is capped with low permeability material. The total cost of the system was $400,000 and the system was designed to treat approximately 300 µg/l TCE and 10+ µg/l of DCE and PCE. Quarterly ground-water monitoring is conducted. The barrier was constructed by first pouring the slurry wall and then excavating the gate area. Dewatering and the use of a metal frame presented several health and safety concerns.

The aquifer is composed of coarse sand and gravel, extending from bedrock at 30 feet to a silty clay at approximately 20 feet. Modeling predicts a 1.4-foot/day flow velocity through the gate. Organic data were collected in wells upgradient of, within, and downgradient of the iron area. Influent levels of organic compounds have been lower than the design parameters, and reduction to below regulatory criteria is occurring. Overall, data are similar to those seen at other sites. For inorganic data, calcium and magnesium concentrations have declined. Minimal concentrations of dissolved iron are exiting the treatment zone; Mr. Vogan speculates that most of the dissolved iron is being precipitated in the treatment zone. There are 2 years of monitoring data that span the installation period and beyond. These data show consistent performance over the monitoring period, however, there are increasing concentrations of cVOCs upgradient of the barrier. Mr. Vogan believes that re-introducing the water collected during construction dewatering to the aquifer via the treatment gate section of the barrier served to dilute the contamination plume.

An audience member asked why a funnel-and-gate system, instead of a continuous barrier, was installed. Mr. Vogan replied that the more cost-effective method of continuous trenching was not available and the funnel-and-gate system was the accepted technology at the time. Under current plans to extend the treatment system, they are still considering a funnel-and-gate system because the site stratigraphy does not lend itself to continuous trenching and there is a strong bedrock layer to contain underflow. The overall move away from funnel-and-gate systems is caused by the potential for under- and over- ground-water flow. However, funnel-and-gate configurations may be more appropriate for use in treatment sequences in which contaminant mixtures are present. Mr. Vogan stated that dewatering problems seen during construction may have been caused by the slurry wall damming ground water flow. Increased hydraulic heads and flow backup have been observed in other funnel-and-gate systems.. This backup may be a cause of over- and under-flow observed in these systems.

Elizabeth City, North Carolina, Permeable Reactive Barrier Site Update
Dr. Robert Puls, EPA, NRMRL

Dr. Puls's presentation focused on lessons learned through facing problems at the Elizabeth City, North Carolina site. Elizabeth City is a Coast Guard base in northeastern North Carolina. A plating shop served as the source of chromate in ground water. The distance from the plating shop to the nearby river is only 200 feet. There is also a plume of TCE migrating toward the river from site-wide sources. The barrier is, by design, 150 feet long, 24 feet deep, and 2 feet wide.

Originally, a natural attenuation project was planned for this site. Investigations indicated that natural attenuation of chromate was occurring in the upper and lower parts of the aquifer, but a preferential pathway existed in the middle. This same pattern was observed for TCE, but there was a dual peak and maximum concentrations were seen at 7.5 meters deep. A full-scale PRB was installed 2.5 years ago to capture these contaminants and ongoing studies have investigated physical, chemical, and mineralogic parameters in an effort to assess system performance.

Conductivity data proved to be very helpful in monitoring and analyzing the barrier. Dr. Puls found that the barrier is somewhat thinner than designed, which supports the use of safety factors. In addition, geochemical parameters have been tracked. Over time, samples in the barrier have shown a steady decrease in the chromate and sulfate influent. Hydraulic studies suggest this decrease was caused by low rainfall levels, since concentrations rebounded after summer rainfall. Data indicate that chromium is treated within the first few centimeters of the barrier. For TCE, concentrations remain at the bottom of the aquifer zone and concentrations have increased over time. Researchers have observed some TCE breakthrough at areas of highest concentrations. This was reviewed with additional ground-water sampling, and TCE was found at concentrations below the barrier. The cause is not known, but Dr. Puls believes some enhanced microbial activity may be occurring. Eh data show trends in levels varying by vertical and horizontal placement in relation to the barrier with levels remaining consistent over time. The same type of trend was observed with pH.

Performance of the Iron Reactive Permeable Barrier at Caldwell Superfund Site
Dr. Grant Hocking, Golder Sierra LLC

Dr. Grant Hocking discussed field performance of an iron PRB installed through vertical hydraulic fracturing at the Caldwell Trucking site in New Jersey. These barriers are installed by pumping a guar, containing iron filings, down a series of boreholes to create a continuous or overlapping barrier. The guar and iron mixture has specific gravity of about 2.2. After injection, the guar breaks down and dissipates, leaving the iron unaffected and in place. Guar selection is important to method success. This barrier is generally 1 to 8 inches thick depending on site stratigraphy.

This technology was implemented at the Caldwell Trucking site to treat TCE migrating to a seep that discharges to a small stream. The intent of the barrier is to intercept ground water and remove TCE prior to seep discharge. The subsurface is composed of bedrock overlain by an esker (a mixture of sands, gravel, and cobble). The esker is overlain by a clay layer topped by a sand layer.. Design parameters required that two barriers, each 3 to 3.5 inches thick on average, extend from the bedrock to the clay layer. Bedrock was permeated with iron as well. TCE concentrations are as high as 7,000 µg/l and lower concentration of other cVOCs are present. The greatest impedance to using this technology is confirming barrier installation. Salt in the guar is used to confirm installation using electrical pulses, similar to sonar or radar. Testing is conducted in real time during installation. Barrier thickness cannot be monitored, other than by using the amount of iron pumped and the barrier area to calculate an average thickness. Pulse testing can be used to identify any holes in the barrier.

One lesson learned at the site involved the guar's breakdown. The guar was mixed at low temperatures and high pH relative to optimal enzyme breakdown parameters. Dr. Hocking speculated that these factors combined to significantly decrease the breakdown rate. As a solution, a pH buffer and additional enzyme was injected. Guar breakdown then occurred and TCE reductions were observed.

The audience posed several questions, which Dr. Hocking answered as follows. In previous testing at other sites, Dr. Hocking has seen fairly uniform barrier thickness except at the edges of the injection plume. The estimated cost for this site was approximately $85 per square foot for the total installation. Dr. Hocking did notice some ground-water damming when the guar was slow in breaking. At the interface with the barrier, there has not been a concentration of precipitates noted, and TCE concentrations are continuing to decline.

Permeable Reactive Barrier Installation--Fairfield, New Jersey
Mr. Stephen Tappert, VECTRE™ Corporation

Mr. Steven Tappert was involved in PRB installation at a site in Fairfield, New Jersey. Previously, the site was used to manufacture electromechanical devices. The operation was closed, which initiated investigations that identified DNAPL and chlorinated solvent contamination from an etching room, dry well, and septic system. The site is unique in that it is leased to a school for disturbed children and is located in a commercial area. These conditions posed several restrictions on site construction and remediation.

Contamination was limited to the shallow sand aquifer that extended 16 feet below ground surface. The sand aquifer is underlain by a dense clay layer that contains the contamination. Potential migration pathways and resulting exposure points include: (1) a creek 250 feet downgradient of the site, (2) an underlying drinking water aquifer, and (3) indoor air contamination through volatilization. Additional investigations eliminated the indoor air concerns, mapped the clay layer, delineated a DNAPL layer, and evaluated the contamination distribution. Mr. Tappert found that contaminant migration followed the clay layer contours more closely than the ground-water flow direction. Based on the information gathered, a PRB using iron was determined to be the best remediation method.

Under an accelerated remediation schedule, the PRB was permitted for construction within 4 months and installation began during the school's summer recess. Before the barrier was excavated, DNAPL removal was completed to remove the contamination source. This removal was also required by state regulations. Approximately 370 tons of contaminated soil were excavated. The PRB was constructed as a continuous iron barrier located ahead of the highest plume concentrations to prevent offsite migration. The final barrier was 127 feet wide, 25 feet deep, and 5 feet thick. The bottom portion of the barrier used a 4:1 iron/sand mixture and the upper portion of the barrier used a 3:2 iron/sand mixture.. Monitoring wells are located upgradient, within, and downgradient of the barrier. These wells exist solely to prove that the technology is working. Construction was completed in August 1998, and monitoring data indicate that the system is performing as planned. Final construction, operation, and maintenance costs for the PRB were $1.056 million, and Mr. Tappert estimated that a pump and treat system would have cost $1.67 million. In this situation, an advantage of the PRB is that no evidence of ongoing remediation can be seen from the surface, except the monitoring well covers.

Continuous Sand-Iron Reactive Wall--South Carolina
Mr. Steven Schroeder, RMT, Inc.
Ms. M. Elizabeth Rowen, RMT, Inc.

Mr. Steven Schroeder began the presentation by describing conditions at an industrial site in South Carolina. The site consists of a full-scale application of a continuous iron barrier. The PRB was designed to treat a TCE and breakdown product plume within two aquifers underlying the site. There are multiple sources of this contamination, and actions have been taken to control or address these sources. The final installation, which is part of the first phase of remediation, is 325 feet long, 27 to 29 feet deep, and 1 foot thick, and was installed across the migrating plume. A 50/50 mix of iron and sand was installed using a continuous trenching machine. A total of approximately 400 tons of iron was used in the barrier construction.

Ms. Elizabeth Rowen discussed the hydrology and monitoring results at this site. Regional flow is to the northeast, but shallow ditches on the site affect localized ground-water flow. The site geology is consistent with that seen in a coastal plain. Fill of silty sands extends to 7 to 14 feet below ground surface, and clay layers underlie the fill. Discontinuous sand lenses are located within the clay layers. The clay itself is variably saturated due to the sand inclusions. Another silty sand layer is below the clays. Aquifers were identified within the fill, the clay layers, and the silty sand layer. Most contamination was seen in the shallow aquifer, but low contaminant concentrations were detected in the middle aquifer. One design concern was how deep to extend the barrier. As a conservative measure, the barrier was extended through the middle aquifer to capture contamination in that layer and also to establish a solid anchor for the barrier. This approach will also address future remediation needs if increased contamination is detected in the middle aquifer. Monitoring has shown that ground-water flow and elevation within the shallow aquifer have changed somewhat due to the barrier installation. This change is caused by increased flow rates of ground water in the middle aquifer from the sand lenses through the barrier.

Mr. Schroeder continued the presentation by discussing design parameters. Because a sand mixture was used in the barrier, several laboratory experiments were conducted to evaluate barrier performance. Permeability and hydraulic conductivity varied with different sand type and percentage mixtures. Based on these laboratory studies, information was gained on barrier permeability versus the surrounding area. A concern at the site was treating DCE and creating vinyl chloride. Therefore, the barrier residence time was calculated to treat both DCE and resulting vinyl chloride concentrations. These measurements indicated that a 3-inch barrier was required: the final choice of a 1-foot barrier thickness involved several safety factors. Ongoing research will provide a better understanding of barrier performance, and this information may be used to reduce this safety margin. This would lead to a reduction of construction and material costs. Some lessons were learned about the actual construction process, which resulted in excess use of iron and had to be tailored to address slumping in the upper 4 feet of the trench. Trench construction was completed within 2 weeks.

OPEN DISCUSSION AND CLOSING REMARKS
Dr. Robert Puls, EPA, NRMRL

Dr. Puls thanked the speakers, John Vogan and EnviroMetal Technologies, DOE members, Carolyn Perroni of Environmental Management Support, Inc., and Kimberly Coerr of Eastern Research Group, Inc. Dr. Puls requested that attendees complete evaluations forms and provide suggestions for future meetings.

FIELD SITE VISIT TO OAK RIDGE Y-12 SITE
Approximately 50 meeting attendees visited the Oak Ridge Y-12 site after the close of the meeting. They received a full tour of the site, organized by Michelle Silbernagel of EnviroIssues.

POSTER PRESENTATIONS

A poster session was held on the evening of November 17, 1998. Posters presented included:

Pathways and Kinetics of Chlorinated Ethylene Reaction with Fe(0)
William Arnold and A. Lynn Roberts

SIFRA--A Research Project for In Situ Ground-Water Remediation Technologies
Birgit Daus

The ITRC Permeable Barrier Walls Workgroup Poster
Brain Ellis

Transformation of N-Nitrosodimethylamine by Iron and Nickel-Plated Iron
Lai Gui and Bob Gillham

Post-Treatment Evaluation of a Source Control Action By Containment Grouting
Dale Huff

Longevity of Zero-Valent Iron Used in Ground-Water Remediation--Influence of Solution Composition on Reduction of Nitroaromatic Compounds and Electrochemical Properties of Iron Particles
Jörg Klausen

Preferential Designs for Permeable Reactive Barriers: Modeling and Field Installation for a TCE Remediation Project at Maxwell Air Force Base
Theodore Meiggs

Dechlorination of Trichloroethene by Electro-Enhanced Iron-Based Treatment System
Yul Roh and S.Y. Lee

HUMASORB-CS™--an Adsorbent to Remove Organic and Inorganic Contaminants
H.G. Sanjay

The Effect of Competing Chromate and Nitrate Reduction on the Degradation of TCE with Granular Iron
Oliver Schlicker

Chemical-Enhanced Replenishment of Elemental Iron for In Situ Reactive Barriers
Bor-Jier Shiau

Performance Monitoring of the First Commercial Subsurface Permeable Reactive Treatment Zone Composed of Zero-Valent Iron
Dominique Sorel, Scott Warner, and Carol Yamane

Retention of Arsenic by Elemental Iron and Iron Oxides
Chunming Su and Robert Puls

Attachment A

Final Speaker List

RTDF Permeable Reactive Barriers Action Team Meeting

Garden Plaza Hotel
Oak Ridge, Tennessee
November 17-19, 1998

Speaker List

*Facilitator

Attachment B

Final Poster Presenters

RTDF Permeable Reactive Barriers Action Team Meeting

Garden Plaza Hotel
Oak Ridge, Tennessee
November 17-19, 1998

Poster Presenter List

Attachment C

Final Attendee List

RTDF Permeable Reactive Barriers Action Team Meeting

Garden Plaza Hotel
Oak Ridge, Tennessee
November 17-19, 1998

Final Attendee List