Evaluation of Permeable Reactive
Barrier Performance
Federal
Remediation
Technologies
Roundtable
< www.frtr. gov >
Prepared by the
Member Agencies of the
Federal Remediation Technologies Roundtable
-------
-------
Evaluation of Permeable Reactive Barrier Performance
Prepared for the
Federal Remediation Technologies Roundtable (FRTR)
by the
Tri-Agency Permeable Reactive Barrier Initiative
Members:
U.S. Department of Defense
U.S. Department of Energy
U.S. Environmental Protection Agency
and
Interstate Technology and Regulatory Council
December 9, 2002
-------
ACKNOWLEDGMENTS
This report on the performance of permeable reactive barriers (PRB) for groundwater
remediation was prepared under the auspices of the Federal Remediation Technologies Roundtable, a
collaborative effort among federal agencies involved in hazardous waste site cleanup. Three
United States (U.S.) government agencies, the Department of Defense (DoD), Department of Energy
(DOE), and United States Environmental Protection Agency (EPA), as well as the Interstate Technologies
and Regulatory Council (ITRC) contributed to the report, which is a concise summary of the conclusions
and recommendations of the three agencies' individual studies at different sites.
Naval Facilities Engineering Service Center (NFESC), Port Hueneme, California and
Battelle, Columbus, Ohio are the coordinating partners for the DoD study. Battelle coordinated the
preparation of this report for the Tri-Agency PRB Initiative, a consortium that was formed to leverage the
resources, expertise, and site experience of the three agencies. The Tri-Agency PRB Initiative operates
under the umbrella of the Federal Remediation Technologies Roundtable. Charles Reeter is the project
officer at NFESC for this study. Arun Gavaskar (Project Manager), Bruce Sass, Neeraj Gupta, Eric
Drescher, Woong-Sang Yoon, Joel Sminchak, James Hicks, and Wendy Condit from Battelle conducted
the investigation of PRB performance at DoD sites.
Oak Ridge National Laboratory (ORNL), Oak Ridge, Tennessee is the coordinating partner
for the DOE study. Liyuan Liang, Gerry Moline, and Olivia West from ORNL conducted the
investigation of PRBs at DOE sites.
The U.S. EPA study was coordinated by the Subsurface Protection and Remediation Division
in the National Risk Management Research Laboratory, Office of Research and Development, U.S.
Environmental Protection Agency, Ada, Oklahoma. Robert Puls and Richard Wilkin conducted the U.S.
EPA investigation.
Matt Turner with New Jersey Department of Environmental Protection (NJDEP) coordinated
the ITRC contribution and played an important role in disseminating the study results through other ITRC
members and through the ITRC website.
Other participants who contributed strongly (to the DoD study, in particular) include Steve
White with U.S. Army Corps of Engineers (USAGE), Timothy McHale with Air Force Research
Laboratory (AFRL), and Robert Edwards with Waste Policy Institute (representative for the Air Force
Center for Environmental Excellence). The DoD study was funded by the Environmental Security
Technology Certification Program (ESTCP) and the Strategic Environmental Research and Development
Program (SERDP). Other individuals who made significant contributions to the EPA study include Frank
Beck, Cindy Paul, Mary McNeil, and Patrick Clark, all with U.S. EPA's Office of Research and
Development, National Risk Management Research Laboratory.
-------
CONTENTS
ACKNOWLEDGMENTS ii
FIGURES iv
TABLES iv
ABBREVIATIONS AND ACRONYMS vi
1.0 INTRODUCTION 1
1.1 Goal of Tri-Agency PRB Projects 1
1.2 Project Coordinators 1
1.3 Project Objectives and Technical Approach 2
2.0 PRB EVALUATION AT DEPARTMENT OF DEFENSE SITES 5
2.1 Methods 5
2.1.1 Longevity Evaluation Strategy 5
2.1.2 Hydraulic Performance Evaluation Strategy 8
2.2 Former NASMoffett Field (Mountain View, CA) 9
2.2.1 Site Description 9
2.2.2 Results and Discussion 9
2.2.2.1 Groundwater Chemistry Evaluation 9
2.2.2.2 Evaluation of Iron Cores and Silt Deposits 11
2.2.2.3 Evaluation with Accelerated Column Tests 11
2.2.2.4 Hydrogeologic Evaluation 12
2.3 Former Lowry AFB (Denver, CO) 13
2.3.1 Site Description 13
2.3.2 Results and Discussion 13
2.3.2.1 Groundwater Chemistry Evaluation at Former Lowry AFB 13
2.3.2.2 Iron Coring at Former Lowry AFB 13
2.3.2.3 Evaluation with Accelerated Column Tests 15
2.3.2.4 Hydrogeologic Evaluation 15
2.4 Other DoD Sites 16
2.4.1 Seneca Army Depot (Romulus, NY) 16
2.4.1.1 Site Description 16
2.4.1.2 Hydrogeologic Evaluation 17
2.4.2 Dover AFB (Dover, DE) 17
2.4.2.1 Site Description 17
2.4.2.2 Hydrogeologic Evaluation 17
3.0 PRBs AT SITES EVALUATED BY U.S. EPA 19
3.1 U.S. Coast Guard Support Center (Elizabeth City, NC) 19
3.1.1 Site Description 19
3.1.2 Methods 19
3.1.2.1 Groundwater Sampling 19
3.1.2.2 Core Sampling and Analysis 23
3.1.3 Results and Discussion 23
3.2 Denver Federal Center (Lakewood, CO) 25
3.2.1 Site Description 25
3.2.2 Methods 26
3.2.2.1 Groundwater Sampling 26
3.2.2.2 Core Collection and Analysis 26
3.2.3 Results and Discussion 27
in
-------
4.0 EVALUATION OF PRBS AT DOE SITES 29
4.1 Y-12 S-3 Ponds/Pathway 2 PRB 29
4.1.1 Site Description 29
4.1.2 Methods 29
4.1.3 Results and Discussion 33
4.2 Former Uranium Milling Site, Monticello, UT 35
4.2.1 Site Description 35
4.2.2 Methods 35
4.2.3 Results and Discussion 35
5.0 MONITORING AND REGULATORY ISSUES WITH PRBS 37
5.1 Compliance and Performance Monitoring 37
5.1.1 Compliance Point 37
5.1.2 Sampling Parameters 37
5.1.3 Sampling Frequency 38
5.1.4 Sampling Methods 38
5.1.5 Monitoring Well Location 38
5.2 Contingency Sampling Plan 39
5.3 Other Regulatory Issues 39
5.3.1 Biostat 39
5.3.2 Contingency Plans 39
6.0 SUMMARY OF CONCLUSIONS 40
7.0 REFERENCES 42
FIGURES
Figure 2-1. PRB at Former NAS Moffett Field Relative to Lithologic Variations in the
Surrounding Aquifer 10
Figure 2-2. Design Hydraulic Flow Regime at Former Lowry AFB PRB 14
Figure 2-3. Hydraulic Conductivity Values (ft/d) from Slug Tests at the Seneca Army Depot
CRB Showing Variations in Hydraulic Conductivity at the Site 16
Figure 2-4. Plan View of PRB at Dover AFB 18
Figure 3-1. Plan View Map Showing Compliance Well, Bundle and Well Cluster Locations
Relative to Granular Iron Barrier and Cr Plume (Elizabeth City) 20
Figure 3-2. Plan View Map of the Denver Federal Center PRB 20
Figure 4-1. Schematic (a) and Plan View (b) of Y-12 Pathway 2 PRB 32
Figure 4-2. Schematic Summary of Coring Results; Cores Collected 4 Years After Y12
Pathway PRB was Installed 34
Figure 4-3. Plan View of Monticello PRB and Detailed Schematic of Permeable Section 34
TABLES
Table 2-1. Design Features of PRBs at DoD Sites 6
Table 2-2. Site Hydrogeology and Hydraulic Parameters of the PRB at DoD Sites 6
Table 2-3. Site Groundwater Geochemistry at DoD Sites 6
Table 3-1. Design Features of PRBs at Sites Evaluated by U.S. EPA 21
IV
-------
Table 3-2. Site Hydrogeology and Hydraulic Parameters of the PRB at DoD Sites at Sites
Evaluated by U.S. EPA 21
Table 3-3. Site Groundwater Geochemistry at Sites Evaluated by U.S. EPA 22
Table 4-1. Design Features of PRBs at DOE Sites 30
Table 4-2. Site Hydrogeology and Hydraulic Parameters of the PRB at DOE Sites 30
Table 4-3. Site Groundwater Geochemistry at DOE Sites 31
-------
ABBREVIATIONS AND ACRONYMS
AFB Air Force Base
AFRL Air Force Research Laboratory
AFCEE Air Force Center for Environmental Excellence
Cl" chloride
CVOC chlorinated volatile organic compounds
DCE dichloroethylene
DFC Denver Federal Center
DNAPL dense, nonaqueous-phase liquid
DO dissolved oxygen
DoD Department of Defense
DOE Department of Energy
EM Environmental Management
ESTCP Environmental Security Technology Certification Program
Fe (II) ferrous iron
FHWA Federal Highway Administration
FY Fiscal Year
GSA
General Services Administration
ICPS inductively coupled plasma spectrometry
ITRC Interstate Technology Regulatory Council
MCL maximum contaminant level
NAS Naval Air Station
NFESC Naval Facilities Engineering Service Center
NO3" nitrate
NRMRL National Risk Management Research Laboratory
O&M operation and maintenance
ORD Office of Research and Development
ORP oxidation-reduction potential
PCE perchloroethylene
PLFA phospholipid fatty acids
PRB permeable reactive barriers
PV present value
RFI RCRA Facility Investigation
RI/FS Remedial Investigation/Feasibility Study
RPM remedial program manager
S2" sulfide
SO42" sulfate
VI
-------
SERDP Strategic Environmental Research and Development Program
TCE trichloroethylene
TDS total dissolved solids
U.S. United States
U.S. EPA United States Environmental Protection Agency
USAGE U.S. Army Corps of Engineers
USCG-SC U.S. Coast Guard Support Center
VC vinyl chloride
VOC volatile organic compound
vn
-------
1.0 INTRODUCTION
Permeable reactive barriers (PRB) are developing into an entire new class of technologies for
groundwater remediation. A permeable barrier is a porous "barrier" that is placed in the path of a
groundwater plume, in various configurations. The barrier, or at least the permeable portion of the
barrier, contains a reactive or adsorptive medium that helps remove the contaminants from the plume, as
the groundwater flows through the barrier. The primary advantage of permeable barriers is their passive
operation and the resulting potential for long-term cost savings.
The technology emerged in the mid-1990s with the use of granular zero-valent iron as a
reactive medium for treatment of groundwater contaminated with chlorinated volatile organic compounds
(CVOCs), such as trichloroethylene (TCE) and perchloroethylene (PCE). More recently, there is interest
in developing other treatment media and methods of construction to address a broader variety of
contaminants and sites.
1.1 Goal of Tri-Agency PRB Projects
In February 2000, representatives of the United States (U.S.) Department of Defense (DoD),
U.S. Department of Energy (DOE), and U.S. Environmental Protection Agency formed the Tri-Agency
PRB Initiative to coordinate the evaluation of this important technology. The combined expertise and
experience of these three government agencies resulted in critical information sharing and strategy
formulation that maximized the efficiencies of the three studies. The Interstate Technology and
Regulatory Council (ITRC), a consortium of 40 state environmental agencies, partnered with DoD to
support this initiative.
Under this initiative, the three agencies conducted field performance evaluations of several
PRBs installed at sites under their purview. The general goal was to evaluate the longevity and hydraulic
performance of several PRBs in various geologic settings. These are the two issues that the agencies
identified as being the most important to address, based on the experience at several PRB sites across the
country. The results of these studies are being provided to the remedial program managers (RPMs) at
government owned sites to aid in decision-making at both existing PRB sites and sites where PRBs may
be applicable. In addition, the results of these studies are being widely disseminated to potential
government and industrial users through distribution of the project reports on government websites, as
well as through more targeted distribution to interested parties.
The three agencies (and ITRC) coordinated their efforts through periodic teleconferences and
meetings. At these conferences, the agencies updated each other on their ongoing field evaluation efforts,
important results, and future monitoring plans. This constant flow of information allowed each agency to
adjust its evaluation strategy with every new piece of information. In this fashion, lessons learned were
quickly incorporated and efforts were realigned in appropriate directions. This report contains a summary
of the conclusions and recommendations from the three studies.
1.2 Project Coordinators
Naval Facilities Engineering Service Center (NFESC), Port Hueneme, California and its
coordinating partner Battelle, Columbus, Ohio conducted the performance evaluation of PRBs at DoD-
owned sites. The Strategic Environmental Research and Development Program (SERDP) and the
Environmental Security Technology Certification Program (ESTCP) sponsored the DoD study. Other
partners in the DoD study included the following organizations:
-------
a U.S. Army Corps of Engineers (USAGE)
a Air Force Research Laboratory (AFRL)
a Air Force Center for Environmental Excellence (AFCEE), through its
coordinating partner, Waste Policy Institute, Texas.
Oak Ridge National Laboratory, Oak Ridge, Tennessee, the coordinating partner for DOE's
Environmental Management (EM) 50 Program, conducted the evaluation at DOE-owned sites. The U.S.
EPA study was conducted by the Subsurface Protection and Remediation Division of the National Risk
Management Research Laboratory, U.S. EPA, Ada, Oklahoma. The U.S. EPA study was facilitated by
cooperation with the U.S. Coast Guard Support Center (USCG-SC), Federal Highway Administration
(FFIWA), and General Services Administration (GSA).
In addition to this summary report, the three agencies prepared detailed reports describing the
methodology and results of their evaluations (Gavaskar et al., 2002; Liang et al., 2001; and Wilkin et al.,
2002).
1.3 Project Objectives and Technical Approach
The two primary objectives of the three agencies' studies were:
• Assessing the longevity of PRBs made from iron, the most common reactive
medium used so far. Longevity refers to the ability of a PRB to maintain its
reactivity and hydraulic performance (residence time and capture zone) in the
years following its field installation.
• Assessing the hydraulic performance of various PRBs in terms of their ability to
capture the targeted portion of the upgradient plume and to provide the influent
groundwater with the required residence time in the reactive medium.
The general technical approach followed by the three agencies consisted of one or more of
the following elements:
• Reviewing existing field data from existing PRBs
• Conducting additional treatability studies, field PRB monitoring, and
computerized modeling at selected PRB sites to fill in any data gaps
• Recommend suitable long-term design/monitoring strategies for existing and new
permeable barriers.
Although field data from PRBs at several sites initially were examined, the study
subsequently focused on those sites that afforded the necessary range of site characteristics and PRB
designs. In addition, sites with a longer history of operation were selected, especially for the longevity
evaluation. The DoD study focused primarily on PRBs installed at the following sites:
• Former Naval Air Station (NAS) Moffett Field, California
• Former Lowry Air Force Base (AFB), Colorado
• Seneca Army Depot, New York
• Dover AFB, Delaware.
-------
The DOE study focused on the PRBs installed at the following sites:
a Y-12 Security Complex, Tennessee
a Uranium Mill Tailings Site, Monticello, Utah
a Rocky Flats Site, Colorado
a Kansas City Plant, Missouri
a Paducah Gaseous Diffusion Plant, Kentucky
a Portsmouth Plant, Ohio.
The objective of the research conducted in U.S. EPA's portion of the Tri-Agency Initiative is
to evaluate the geochemical and microbiological processes within zero-valent iron treatment zones in
permeable reactive barriers that may contribute to decreases in iron reactivity and decreases in reaction
zone permeability that, in turn, may eventually lead to system plugging and failure. Using advanced
surface analytical techniques together with detailed coring and water sampling programs at two
geographically, hydrogeologically, and geochemically distinct iron barrier installation sites, specific
research objectives were to:
1) Characterize the type and nature of surface precipitates forming over time at the
upgradient aquifer/iron interface, within the iron zone, and at the
downgradient/iron interface
2) Develop conceptual models that predict the type and rate of precipitate formation
based on iron characteristics and water chemistry
3) Identify type and extent of microbiological activity upgradient, within and
downgradient in at least one of the chosen sites to evaluate microbiological
response or effects from emplaced iron into an aquifer system
4) Develop practical and cost-effective protocols for long-term performance
assessments at permeable reactive barrier installations.
Two field sites were evaluated in the U.S. EPA portion of the TRI:
a U.S. Coast Guard Support Center (USCG-SC) site near Elizabeth City, North
Carolina, and
a Denver Federal Center (DFC) in Lakewood, Colorado.
These sites provided a range of PRB designs and hydrogeologic characteristics that could be
studied so that appropriate guidance could be provided for future applications. The sites studied also had
a broad range of contaminants, such as TCE, PCE, chromium, and radionuclides.
The longevity evaluation conducted by the three agencies consisted of one or more of the
following elements:
• Groundwater geochemistry monitoring
• Iron core collection and analysis
• Geochemical modeling
• Accelerated column tests
-------
The longevity of iron barriers is potentially limited by formation of precipitates in the iron,
upon long-term contact with groundwater. Common mineral precipitates found in field-installed zero-
valent iron barriers and in columns designed to simulate field barrier processes include iron hydroxides
and oxyhydroxides, iron sulfides, iron and calcium carbonates, and iron hydroxy carbonates and sulfates
(green rusts). Mineral precipitation is not consistent from site to site, however, and some barriers contain
many of these precipitates while others contain very little. Understanding the processes that control the
rate and type of mineral precipitation is important in barrier planning and design as well as monitoring the
performance of installed barriers.
The hydraulic performance evaluation conducted by the three agencies used one or more of
the following tools:
• Water level measurements
• HydroTechnics™ in-situ flow sensors
• Colloidal borescope (down-hole instrument)
• Groundwater flow modeling
-------
2.0 PRB EVALUATION AT DEPARTMENT OF DEFENSE SITES
Although field data from PRBs at several DoD sites initially were examined, the project
subsequently focused on those sites that afforded the necessary range of site characteristics and PRB
designs. The longevity evaluation focused primarily on two sites:
• Former Naval Air Station (NAS) Moffett Field
• Former LowryAFB.
These two sites were selected because the PRBs there were installed a few months apart
around the beginning of 1996 (that is, they had sufficient history of field operation) and because the
groundwater at these sites was relatively high in total dissolved solids (TDS), an important factor in
accelerating the determination of precipitation potential and longevity. The hydraulic performance
evaluation focused primarily on four sites:
• Former NAS Moffett Field (funnel-and-gate)
• Former Lowry AFB (funnel-and-gate)
• Seneca Army Depot (continuous reactive barrier)
• Dover AFB (funnel with two gates).
These sites provided a range of PRB designs and hydrogeologic characteristics that could be
studied so that appropriate guidance could be provided for future applications. In addition to these
primary focus sites, PRBs at other sites, such as Cape Canaveral Air Station (Hangar K) and former NAS
Alameda, initially were examined, but were de-emphasized as resources were focused on field
investigations at sites that appeared to offer the most features of interest for the project. Also, a separate
detailed study at former NAS Alameda (Einarson et al., 2000) provided sufficient information for this
evaluation.
Tables 2-1, 2-2, and 2-3 summarize the important features of the PRBs at the two key sites,
former NAS Moffett Field and former Lowry AFB.
2.1 Methods
The performance assessment objectives were achieved by using a select variety of tools that
allowed the project to fill in the data gaps identified in the existing information from the PRB sites. Both
performance objectives, longevity and hydraulic performance, presented significant challenges for the
project. The strategy consisted of a combination of tools to address each objective and overcome the
limitations of each individual tool.
2.1.1 Longevity Evaluation Strategy. From the beginning of the project, it was clear that
developing predictions about the life of a granular iron barrier would be difficult, given the short history
of the technology in the field, the lack of information on kinetic rates of precipitation and reactivity loss
that could be used in predictive models, and the difficulty of conducting any kind of laboratory
simulations that would mimic the exposure of the iron to many pore volumes (i.e., long periods) of
groundwater. Tools that initially were used in the current project to evaluate longevity include the
following:
n Analysis of inorganic constituents in groundwater influent and effluent
n Analysis of iron cores collected from field PRBs
n Geochemical modeling
-------
Table 2-1. Design Features of PRBs at DoD Sites
PRB Site Name,
City, State
Former NAS
Moffett Field
Former Lowry
AFB
Pilot/Full
Scale
(Installation
Date)
April 1996
December
1995
Type of
Barrier
-------
Tools that have become fairly conventional for evaluating precipitation in field PRBs include
groundwater monitoring (influent and effluent) and iron core collection and analysis. By analyzing the
groundwater influent and effluent (or upgradient and downgradient) to the PRB, the loss of inorganic
constituents (e.g., calcium, magnesium, alkalinity, sulfate, silicate, etc.) sustained by the groundwater can
be measured as it moves through the reactive cell of the PRB. The differences in or loss of groundwater
constituents represents the potential precipitation that has occurred in the PRB. However, there are two
challenges to using these tools:
n First, the losses in inorganic constituents measured in the groundwater often do not match the
amount of precipitate observed on core samples of iron collected from the PRB. This
mismatch can partly be explained by the fact that there is considerable uncertainty in the
spatial extrapolation of the amount of precipitate observed on small core samples of iron to
the rest of the reactive cell, as precipitates may be unevenly deposited in different parts of the
iron.
n Second, even if the amount of precipitate formed could be accurately determined, it is unclear
how these precipitates distribute on the iron surfaces (whether in mono-layers that use up
maximum surface area or in multiple layers that conserve the available reactive sites). Also,
because the mechanism through which the precipitates may be bound to the iron and the
process by which electrons are transferred between the iron and the contaminants is unclear,
it is difficult to correlate loss of surface area with loss of reactivity. In other words, could
iron continue to react with the contaminants through a layer of precipitates on its surface?
Geochemical modeling previously has been used to elucidate the precipitation process
(Battelle, 1998; Gavaskar et al., 2000; Sass et al., 2001). Two types of models are available - equilibrium
models (models that assume an infinitely long contact time between the iron and the groundwater
constituents) and kinetic models (models that can be can be calibrated to contact time, if the various
reaction kinetics or rate constants involved are known). Because the kinetics of iron-groundwater
reactions have not yet been documented, although attempts have been made by some researchers
(Yabusaki et al., 2001) to do that, kinetic models have limited applicability. However, equilibrium
models are useful for identifying the types, if not the quantity, of precipitates; these models were used in
the current project to understand the kinds of precipitation reactions occurring in the iron and provide
some indication of what to look for when analyzing the iron cores.
Given the limitations of the indicative tools described above, there was a need for direct
empirical evidence of any decline in reactivity of the iron due to exposure to groundwater. Therefore,
accelerated column tests were conducted to simulate the field performance of PRBs at former NAS
Moffett Field and former Lowry AFB. The objective of the accelerated column tests was to examine if
and to what extent the reaction rates (or half lives) of the contaminants would deteriorate when the iron
was exposed to many pore volumes (i.e., long periods) of contaminated groundwater flow. Unlike tests
conducted by John Hopkins University (Arnold and Roberts, 2000; Totten et al., 2001), which currently is
studying the effect of individual inorganic and organic constituents in groundwater on the iron, the
accelerated column tests in the current project were conducted with actual groundwater from the two sites
(former NAS Moffett Field and former NAS Lowry AFB) simulated. The same iron that is in these PRBs
(Peerless Metal Products, Inc., iron at for NAS Moffett Field, and Master Builder, Inc., iron at former
Lowry AFB) was used to pack the two columns. A small amount of oxygen scavenger was added to the
groundwater influent to the columns to restore the low dissolved oxygen (DO) levels of the native
groundwater, because the groundwater is relatively anaerobic at both sites. Therefore, the interplay of
factors occurring in the two field PRBs were simulated as closely as possible.
-------
Higher groundwater flowrates were maintained in the columns than were present in the field
PRBs, in order to accelerate the exposure of the iron to the groundwater. Previous studies (O'Hannesin,
1993) have shown that contaminant half-lives are independent of the flowrate; this was confirmed through
half-life measurements conducted at different flowrates during the current project. Accelerating the flow
through the column permits an examination of the changes in reactivity of the iron when exposed to many
pore volumes (or several years) of groundwater flow. Given the short history of field PRBs (6 years
maximum), this simulation provides valuable insights into the future behavior of the iron-groundwater
systems at these sites.
2.1.2 Hydraulic Performance Evaluation Strategy. The permeable reactive barriers technology
relies upon the use of hydraulic characteristics of the site for successful performance over the short- and
long-term. Therefore, a careful consideration of the hydrogeologic issues must be incorporated at all
stages of the project: site screening, characterization, design, construction, and performance assessment.
Most of the reports about sub-optimum performance at some PRB sites may be attributed to hydraulic
factors. The issues of concern include insufficient residence time resulting in contaminant breakthrough,
inability to verify flow through the reactive cell, plume bypass around, under, or over the barrier, seasonal
fluctuations in groundwater flow that result in variation in performance, and effect of nearby site features
such as drains, surface water, operating pump-and-treat systems, etc. Almost all of these issues can be
related to the two primary objectives involved in designing a PRB and monitoring its hydraulic
performance:
n Ensuring that the PRB will capture the desired portion of the plume, and
n Ensuring that the desired residence time in the reactive cell will be met.
Thus the two primary interdependent parameters of concern when designing a PRB are hydraulic
capture zone width and residence time. Capture zone width refers to the width of the zone of groundwater
that will pass through the reactive cell or gate (in the case of funnel-and-gate configurations) rather than pass
around the ends of the barrier or beneath it. Capture zone width can be maximized by maximizing the
discharge (groundwater flow volume) through the reactive cell or gate. Residence time refers to the amount
of time contaminated groundwater is in contact with the reactive medium within the gate. Residence times
can be maximized either by minimizing the discharge through the reactive cell or by increasing the
flowthrough thickness of the reactive cell. Thus, the design of PRBs must balance the need to maximize
capture zone width (and discharge) against the desire to increase the residence time. Contamination occur-
ring outside the capture zone will not pass through the reactive cell. On the other hand, if the residence time
in the reactive cell is too short, contaminant levels may not be reduced sufficiently to meet regulatory require-
ments.
The basic tools and methods that can be used at various stages of a PRB project for
improving the probability of successful implementation have been discussed in details in the design
guidance (Gavaskar et al, 2000). The two classes of design used in the current study are:
n Site Characterization - this includes developing a detailed understanding of the site
geology, hydrogeology, contaminant distribution, and seasonal fluctuations and
incorporating the ranges in these aspects into the PRB design to maximize successful
implementation.
n Groundwater Flow Modeling - this includes incorporating the site parameters into the
computer simulation tools so that the spatial and temporal variations in these parameters
can be evaluated and the appropriate safety factors can be determined for PRB design and
monitoring system configuration.
-------
The hydraulic performance evaluation strategy consisted of two major elements. One, an
effort was made to conduct more detailed characterization of the flow regime around existing field
barriers. Two, groundwater modeling was used to obtain a better understanding of the various factors that
determine flow at these PRB sites. The objective was to get a better understanding of the groundwater
capture zone and residence time at these sites. Therefore, most of the evaluation was conducted on the
upgradient side of the PRBs. Groundwater flow direction and velocity ultimately are the two key
parameters that need to be estimated to make this determination. The evaluation included the following
tools:
n Water-level measurements
n Slug tests
n In-situ flow sensors
n Colloidal borescope
n Groundwater modeling.
Former NAS Moffett Field, Lowry AFB, Seneca Army Depot, and Dover AFB were the sites
subjected to a more detailed evaluation. These sites provided a wide range of site and PRB design
characteristics.
2.2 Former NAS Moffett Field (Mountain View, CA)
Both geochemistry and hydrologic issues were evaluated at this site, which has a pilot-scale
funnel-and-gate system for a regional TCE plume.
2.2.1 Site Description. The funnel-and-gate PRB at the former NAS Moffett Field PRB site has
been monitored and evaluated in significant details as part of a previous ESTCP project (Battelle, 1998).
The surficial aquifer at this site is divided into two aquifer zones-a shallow zone (Al) and a deep zone
(A2). The barrier is installed in the Al zone of the surficial semi-confined aquifer at the site. The Al
aquifer zone is approximately 25 ft deep. Borings at the site suggest that several sand channels exist in
the otherwise silty sand aquifer. The barrier was installed in a funnel-and-gate configuration through a
major sand channel (Figure 2-1) within the lower conductivity silty and clayey layers. In general, the site
reflects channeled groundwater flow in a multi-layered aquifer system. Peerless Metal Powders, Inc.,
Detroit, Michigan, supplied the granular iron used in the PRB.
2.2.2 Results and Discussion. Following are results of the field performance measurements at
NAS Moffett Field and results of the long-term column test with groundwater from NAS Moffett Field.
2.2.2.1 Groundwater Chemistry Evaluation. At former NAS Moffett Field, TCE, PCE, and cis-1,2
DCE in the effluent from the reactive cell iron continues to be below their respective MCLs and below
detection. Most of the treatment occurred in the upgradient half of the iron. A noticeable clean
groundwater front is not clearly identifiable in the downgradient aquifer, although there are some
preliminary signs that it could occur in the future. After five years of PRB operation in the sand channel
enclosed by silty clay sides, it was expected that introduction of CVOC-free groundwater effluent would
lead to a noticeable improvement in downgradient groundwater quality, despite some contrary site
conditions. One or more of the site conditions that could be acting to delay or prevent an improvement in
downgradient groundwater quality are:
n Less groundwater flowing through the more conductive reactive cell or gate than is
predicted or than is flowing around or below the PRB. In some wells screened at
shallower depths, a proportionate relative decline in CVOC and inorganic constituents
(e.g., calcium) is noticeable overtime, which would support this scenario. CVOC levels
-------
Silt/Clayey Silt
K=0.5
n = 0.40
Silt/Clayey Silt
K=0.5
n = 0.40
Silty Sand
K=30
n = 0.35
Silty Sand
K=30
n = 0.35
I
Silty Clay
K= 0.05
n = 0.45
/ X ' Channel Sand and Gravel | ^
I
T / K=150 T
• i
/ n = 0.30 '
|< 10tt H I
; : Pea Gravel,
Iron cell
I Pea Gravel ;
T '
T!\
«4_ \
CD ° ^
1 1 \
1
si 1
oun |
1 1
' 1
\ |
s
?
\ \ v
\
Explanation
Hydraulic conductivity (ft/d)
Total porosity
— Boundary of geologic unit
Pea Gravel K = 2,830 n = 0.33
Iron K = 283 n = 0.33
N
t
1 Silty Clay
\ K=0.05
\ n = 0.45
\
\
\
\
NOT TO SCALE
MODELCELL01.CDR
Figure 2-1. PRB at Former NAS Moffett Field Relative to Lithologic
Variations in the Surrounding Aquifer
10
-------
have declined somewhat over time in the upgradient aquifer too, making the
determination more difficult.
n Recontamination of cleaner groundwater effluent from the PRB with contaminated
groundwater flowing under the PRB (the pilot-scale PRB intentionally was not keyed into
the clay layer for fearing of breaching a thin aquitard) or from the lower aquifer zone.
The downgradient monitoring wells that are screened at a depth near the base of the PRB
continue to be the most contaminated, indicating that there is underflow. However,
vertical gradients that were upward in the vicinity of the PRB before PRB installation
have consistently turned downward after the installation; this would tend to reduce the
mixing of groundwater flowing under and through the PRB.
n Contaminated groundwater flowing around the funnel walls of the pilot-scale PRB that
was designed to capture only a small part of a regional plume. This is less likely because
the sand channel, which probably accounts for most of the groundwater flow in the local
region of the PRB, directs flow mostly through the gate. The funnel walls encounter
minimal additional groundwater flowing through the silty-clay deposits around the
channel.
n Diffusion of CVOCs trapped in the silty clay layers surrounding the sand channel. This
type of contaminant persistence has been observed at other sites, even with pump-and-
treat systems. However, diffusion is a slow process and water quality improvement
immediately downgradient of the PRB would still be expected.
2.2.2.2 Evaluation of Iron Cores and Silt Deposits. At former NAS Moffett Field, geochemical
analysis of iron cores from the PRB showed the following:
n Calcium, silicon, and small amounts of sulfur were the elements identified on the iron
particles.
n Aragonite, calcite (both forms of calcium carbonate), and iron carbonate hydroxide
(similar to siderite) were the mineral species identified on the iron particles these
minerals were concentrated in the iron samples collected from the upgradient edge of the
reactive cell, indicating that the rest of the iron had not encountered much precipitation.
Calcite, iron oxyhydroxide (FeOOH) or goethite, ettringite (calcium-aluminum sulfate), and
katoite (calcium-aluminum silicate) were the mineral species identified in the silt from the silt traps in the
monitoring wells in the PRB at former NAS Moffett Field. The elements iron and magnesium were
identified in the silt, but could not be associated with any particular mineral species. Some mineral
species (such as feldspar, muscovite, mica and clay minerals) that probably originated from the pea gravel
(granite) were also identified. The presence of minerals in the silt traps that are traceable to the
groundwater indicates that not all the precipitates formed deposit on the iron medium. Finer, colloidal
particles can be transported by the flow to other locations within the PRB, some of which become trapped
in the monitoring wells.
2.2.2.3 Evaluation with Accelerated Column Tests. Long-term accelerated column tests were
conducted with groundwater from the field PRBs at former NAS Moffett Field and former Lowry AFB.
The columns were packed with fresh iron obtained from the same sources that were used at these two
sites. The two columns were adjusted to a flow rate whereby pH and ORP reached a plateau (indicating
that the majority of the reactions between the iron and groundwater had occurred in the column), but was
fast enough that many pore volumes of groundwater could be passed through the column (or many years
of PRB operation could be simulated). After some trial-and-error, a flow rate of 12.5 ft/day was
eventually established as optimum for the column test. At this flow rate, all the precipitates generated
11
-------
stayed in the column (at higher flow rates, there was a tendency for finer precipitates to be transported out
with the flow. If a representative normal flow rate of 0.5 ft/day is assumed at both sites, than the flow in
the columns is accelerated 25 times. The 1,300 pore volumes of groundwater passed through each
column and the 1.5 years of column testing simulate 30 years ormore of operation of the field PRBs. A
related test conducted with the same columns showed that the TCE half-life was independent of the flow
rate over a wide range of flow rates.
The column tests show that over the 1,300 pore volumes of flow that the iron was exposed to,
the half-life of TCE increased approximately by a factor of 2 in the Moffett Field column. While some
effects of aging may be intrinsic to the iron, itself, or to the manufacturing process, the loss of reactivity is
probably due to the inorganic content of the water and the subsequent precipitation of dissolved solids on
the iron surfaces. Former NAS Moffett Field has groundwater with a moderate level (between 500 to
1,000 mg/L) of dissolved solids.
2.2.2.4 Hydrogeologic Evaluation. The purpose of hydrogeologic investigations conducted under
the project was to evaluate the major issues related to capture zone and residence time based on these
existing two classes of tools. These two hydraulic issues were investigated by:
n Conducting a field evaluation of PRBs at various DoD sites, and
n Conducting computer simulations to evaluate the effects of hydraulic variations and
characterization uncertainties.
PRBs have been installed at DoD sites with a variety of site characteristics. Overall, the
PRBs have been fairly effective over a wide range of site conditions.
Water level surveys provide information on groundwater gradients and capture zones for
PRBs to demonstrate that groundwater is flowing through the barrier at a rate, which will ensure adequate
destruction of the contamination. Several rounds of water level surveys were performed at the selected
DoD PRB sites during the project. In general, the groundwater surveys demonstrated a positive gradient
in the expected flow direction through the PRBs, that is, when gradients were measured from upgradient
to downgradient aquifer. For example, positive gradients were observed in periodic monitoring of PRBs
at Dover AFB, former NAS Moffett Field, Seneca Army Depot, and former Lowry AFB.
Within the PRBs themselves, hydraulic gradients were extremely flat, which is expected of
highly permeable and porous media. A few transient flow reversals were reported, for example, at the
Moffett Field site, but these occurrences appear to have been temporary and generally within the
measurement error (Battelle, 1998). At former NAS Moffett Field, monitoring conducted during a
previous project showed that some mounding appeared to be occurring at the downgradient end of the
PRB, which may indicate that groundwater discharge from the highly permeable PRB media to the
generally less permeable aquifer meets with some resistance. Among all the PRB sites evaluated under
the current project, the PRB at former NAS Moffett Field provided the most certainty in terms of
verifying a groundwater capture zone and occurrence of flow through the PRB, probably because the sand
channel surrounded by silty-clay deposits constrained flow from diverging to the sides. Close
examination of the water level data reveals flow divides occurring about halfway across the length of
each funnel wall. Based on these water levels an approximate estimate of capture zone is 30 ft. The
capture zone includes the flow directly upgradient of the 10-ft-wide gate and halfway across 20-ft-wide
funnel wall. Water-level surveys are a key monitoring activity for confirming gradients at PRB sites.
Based on atypical hydraulic gradient of 0.007, observed during water level mapping events,
and atypical hydraulic conductivity of 30 ft/day, representative of slug test results in the sand channel, a
12
-------
typical groundwater velocity of 0.7 ft/day and a residence time of 9 days are estimated. This residence
time estimate matches the results of a tracer test (Battelle, 1998) conducted during a previous project.
The wide variability in the hydraulic conductivities measured at different locations in the aquifer and the
likelihood of preferential pathways in the iron medium itself, as seen in the tracer test, create substantial
uncertainty in the groundwater velocity and residence time estimates.
2.3 Former Lowry AFB (Denver, CO)
Lowry AFB has one of the first PRBs installed in the field; it was installed in December 1995
to address a TCE plume.
2.3.1 Site Description. The aquifer at former Lowry AFB is comprised of 11 ft of silty-sand to
sand and gravel in an unconfined aquifer which overlies weathered claystone bedrock 23-30 ft bgs
(Versar, Inc., 1997). Some degree of heterogeneity is present in the form of sand and clay lenses. The
barrier was set up in a funnel-and-gate arrangement with funnel walls at an angle to the reactive cell
(Figure 2-2). The iron for the barrier was supplied by Master Builders Supply, Streetsboro, Ohio.
2.3.2 Results and Discussion. The results of the field measurements and accelerated column tests
for Lowry AFB are described in this section.
2.3.2.1 Groundwater Chemistry Evaluation at Former Lowry AFB. Groundwater samples were
collected from the PRB at former Lowry AFB in the current project in September 1999, approximately 4
years after installation of the PRB. Groundwater samples were collected in all wells inside the reactive
cell and in the upgradient and downgradient pea gravel zones that are adjacent to the reactive cell. In
addition, aquifer wells were sampled immediately upgradient and downgradient of the reactive cell.
Results of groundwater sampling shows that TCE is the major contaminant in the
groundwater; smaller concentrations of cis-DCE and trans-\,2-DCE also were observed in the aquifer.
CVOC concentrations declined slightly in the upgradient pea gravel due to quick horizontal and vertical
mixing in the porous zone. The contaminants were undetectable in most of the reactive cell wells and are
entirely below detection in the downgradient portion of the cell. These results demonstrate that the
reactive cell is degrading the contaminants to below their respective MCLs (<5 (ig/L for PCE and TCE;
<70 (ig/L for DCE). TCE, cis-DCE, and trans-\,2-DCE are present in the downgradient aquifer as a
result of mixing with contaminated groundwater flowing around the pilot-scale PRB. Trends such as
rising pH, declining ORP, and declining DO as water moves into the reactive cell indicate that the barrier
was functioning normally, after four years of operation. Lower conductivity values in the reactive cell
wells compared to aquifer wells suggests some precipitation of solids inside the reactive cell.
Results of inorganic analysis shows a considerable decline in alkalinity, calcium, magnesium,
silica, and sulfate as the groundwater flows through the reactive cell, which suggests mineral precipitation
inside the barrier.
2.3.2.2 Iron Coring at Former Lowry AFB. Approximately 18 months after the former Lowry AFB
barrier had been in operation, iron core samples were collected for analysis (Versar, 1997). The cores
were sent to the University of Waterloo for mineralogical and microbiological analysis and the results
were reported by EnviroMetal Technologies, Inc. (ETI, 2000). The mineralogical analysis showed that
calcite and aragonite were the main carbonate minerals detected; however, siderite was found in one
sample. A greater concentration of carbonates was found in the upgradient portion of the barrier than in
the middle and downgradient portions. Core samples collected nearest the upgradient face contained 4
grams calcium carbonate per 100 grams of sample. Several other compounds were found throughout the
reactive barrier including green rust, magnetite, and amorphous iron hydroxide. Microbiological analysis
13
-------
REACTIVE WALL'IRCW f LNJ3
W51.H
_
• •:;i : . ", ' -,-f • ,' --•
Dm CHracHan Pip
n/U/rt
Figure 2-2. Design Hydraulic Flow Regime at Former Lowry AFB PRB
14
-------
showed slightly higher microbial populations at the influent end than elsewhere within the wall. The
microbial populations in the wall were thought to be of the same order of magnitude as in similar types of
aquifers and soils.
Iron cores were collected at former Lowry AFB in September 1999, approximately 4 years
after PRB installation. Results from the earlier study (ETI, 2000) differ in some aspects from those
observed in the recent project. Most notable is the absence of calcium carbonate from the iron core
samples collected in this project (in September 1999). Normally, XRD is very sensitive to calcite and
aragonite, so these minerals are unlikely to have been overlooked in diffraction patterns. Also, the total
carbon content was only about 2 percent and the analysis did not reveal an excess of carbon in the barrier
samples compared to the control (unused) sample. Moreover, the carbon that was detected in the iron
samples was attributed to the reduced (graphitic) carbon coatings. One explanation is that the recent
samples were not collected sufficiently close to the upgradient interface where most of the carbonate
precipitation is occurring. Another possibility is that the different analytical methods used in the two
studies gave different results. The reason for the difference in carbonate detection in the two studies still
is unclear.
2.3.2.3 Evaluation with Accelerated Column Tests. Long-term accelerated column tests were
conducted with groundwater from the field PRBs at former NAS Moffett Field (see Section 2.2.2.3) and
former Lowry AFB. The column tests show that over the 1,300 pore volumes of flow that the iron was
exposed to, the half-life of TCE increased approximately by a factor of 4 in the Lowry AFB column, as
compared to a factor of 2 in the Moffett Field column and. While some effects of aging may be intrinsic
to the iron, itself, or to the manufacturing process, other differences may be due to the inorganic content
of the water and the subsequent precipitation of dissolved solids. Former NAS Moffett Field has
groundwater with a moderate level of dissolved solids (between 500 to 1,000 mg/L) and former Lowry
AFB has groundwater with relatively high levels of dissolved solids (greater than 1,000 mg/L);
consequently, Lowry AFB showed a greater decline in reactivity over the same period of exposure to
groundwater as the Moffett Field column.
The column test results indicate the following:
• The geochemical constituents of the groundwater do affect the reactivity of the iron upon
long-term exposure to groundwater.
• The rate of decline in iron reactivity over time is dependent on the native level of certain
dissolved solids (e.g., alkalinity, sulfate, calcium, magnesium, and silica) in the
groundwater.
• The PRB is likely to be passivated before the entire mass of zero-valent iron is used up,
unless some way of regenerating or replacing the reactive medium is developed and
implemented.
2.3.2.4 Hydrogeologic Evaluation. At Lowry AFB, gradients were relatively strong in the
upgradient aquifer and indicated not only flow progressing in the expected direction toward the reactive
cell, but also the asymmetric nature of the capture zone due to the effect of an adjacent stream on the east
side. The capture zone at Lowry AFB appears to be approximately 20 ft wide, with 10 ft of capture
directly upgradient of the gate and 10 ft along the western funnel wall. Most of the flow upgradient of the
eastern funnel wall appears to be directed towards the flowing stream on the east. Based on the hydraulic
conductivities measured during slug tests and the hydraulic gradient obtained from water level
measurements, atypical groundwater velocity of 0.2 ft/day and atypical residence time of 25 days are
estimated. A moderate variability in the hydraulic conductivity estimates in the sandy aquifer creates
some uncertainty in these estimates.
15
-------
At Lowry AFB, all the slug tests showed an exceptionally narrow conductivity range
indicating a relatively homogeneous aquifer.
2.4 Other DoD Sites
Primarily hydrologic evaluations were conducted at two additional sites, Seneca Army Depot
and Dover AFB, to obtain a broader perspective on hydraulic performance issues and the monitoring tools
involved.
2.4.1 Seneca Army Depot (Romulus, NY). Seneca Army Depot has a continuous reactive barrier
that is one of the relatively longer PRBs that has been installed to capture a fairly wide TCE plume.
2.4.1.1 Site Description. Groundwater flows through fractured shale and overlying glacial till at
Seneca Army Depot (Parsons Engineering Services, Inc., 2000). The aquifer is unconfined. The PRB at
Seneca is a 600-ft-long continuous trench, approximately 1 ft wide and keyed into competent shale
bedrock 5-10 ft bgs (Figure 2-3). The barrier consists of a 50/50 mixture of sand and iron. Overall, the
Seneca Army Depot site reflects a shallow glacial till aquifer with a long, thin PRB designed to treat a
diffuse plume spread over a large area. During the current project, 14 new 2-inch monitoring wells were
installed (two inside the PRB and 12 in the surrounding aquifer, near the northern end of the PRB) to
determine the flow divide and the capture zone.
MW-T10
38.2
Bat-9
1'4 1*5
Bat-1 Bat-12
36.5 1*5
Bat-4 Bat-14
24.3 Bat.1() 12.3
•
1 -9
Bat -2 „ , _,
29.0 Ba.'-7 Bat-1
1.1
Bat-5
(129.1)
Bat-1 1
Bat-8
67-5 Explanation
BAT-1 Well Name
•
Bat-6 36.5 Hydraulic Conductivity (ft/d)
•
^C C
Continuous Reactive Barrier
Figure 2-3. Hydraulic Conductivity Values (ft/d) from Slug Tests at the Seneca Army
Depot CRB Showing Variations in Hydraulic Conductivity at the Site
16
-------
2.4.1.2 Hydrogeologic Evaluation. At Seneca Army Depot and Dover AFB, the flow divide and
therefore the capture zone were difficult to determine. At Dover AFB, the native gradient itself is low.
At Seneca Army Depot the difficulty was that the PRB was relatively thin (1 ft flowthrough thickness)
and generated a very minor disturbance in the natural flow patterns.
At both these sites, uniformly screened monitoring wells and multiple monitoring events led
to at least some events that afforded discernible groundwater flow trends. To conserve limited resources,
the monitoring well network at Seneca Army Depot was limited to one end of the relatively long PRB.
The water level map for this site for April 2001 shows a steep gradient immediately upgradient of the
PRB and flat water levels farther away. It also shows that the flow lines are pointing towards the PRB at
the northern end of the site indicating capture of the plume from that area. However, during July 2001 the
water levels are flat upgradient of the PRB showing the seasonal effects on the flow patterns and
residence times. In both cases there is a downward gradient from upgradient to downgradient wells
indicating the flow is occurring through the PRB.
2.4.2 Dover AFB (Dover, DE). Area 5 at Dover AFB has a funnel-and-gate type PRB that
intercepts a PCE plume.
2.4.2.1 Site Description. The funnel-and-gate PRB at Dover AFB was designed, installed, and
monitored as part of a SERDP-funded project by Battelle (Battelle, 2000). The aquifer at the Dover AFB
site consists of unconfmed silty sand deposits overlying a thick clayey confining layer. The aquifer is
approximately 20-25 ft thick and fairly homogenous, except for several silty-clay lenses in the upper
portion of the aquifer. The hydraulic gradient in the area is fairly low (0.002) and variable, with
noticeable seasonal fluctuations. The PRB consists of a funnel-and-gate system with two gates (Figure 2-
4). Interlocking sheet piles (Waterloo Barrier™) constitute the funnel and caisson excavations filled with
reactive media (iron) constitute the two gates. The Dover AFB site represents a low-flow velocity setting
in a thick, homogenous aquifer. As part of the current project, water level measurements and colloidal
borescope measurements were performed at this site.
2.4.2.2 Hydrogeologic Evaluation. Seasonal fluctuations in the gradient must be accounted for in
the analysis of water level data. For example, at Dover AFB, historical measurements indicated that
groundwater flow direction changed by about 30° on a seasonal basis (Battelle, 2000). This had a
considerable effect in determining an optimum design and orientation of the PRB so that the PRB was
perpendicular to the flow during most times of the year. At least four quarters of water level data should
be obtained to account for seasonal fluctuations in groundwater velocity and direction, before designing a
PRB. In addition, information on long-term extremes in water levels and flow directions obtained from
historical records, where available, should be considered in the designing PRBs.
17
-------
Building 639
Parking Lot
Groundwater Flow
Evreux Street
N
\
NOT TO SCALE
AREAS 01.CDR
Figure 2-4. Plan View of PRB at Dover AFB
18
-------
3.0 PRBs AT SITES EVALUATED BY U.S. EPA
The evaluation conducted by U.S. EPA focused on PRBs at Elizabeth City and Denver.
3.1 U.S. Coast Guard Support Center (Elizabeth City, NC)
A continuous trencher was used for the first time to install a PRB at this site.
3.1.1 Site Description. In June of 1996, a 46 m long, 7.3 m deep, and 0.6 m wide permeable
reactive barrier (continuous wall configuration, Figure 3-1) of zero-valent iron was installed at the U.S.
Coast Guard-Support Center site located in Elizabeth City, North Carolina (USCG-SC). The reactive
wall was designed to remediate hexavalent chromium-contaminated groundwater, in addition, to treating
portions of an overlapping, larger plume of trichloroethylene (TCE). A monitoring network of over 130
subsurface sampling points was installed in November of 1996 to provide detailed information on spatial
and temporal changes in pore water geochemistry (Puls et al., 1999). Information about the design of this
PRB and initial performance data were published in Blowes et al. (1999a,b).
3.1.2 Methods. Groundwater sampling, iron core analysis, and geochemical modeling were the
methods used to evaluate the performance of this PRB.
3.1.2.1 Groundwater Sampling. Groundwater was sampled from monitoring wells using peristaltic
or submersible pumps. In all cases, low-flow (150 to 250 mL/min) purging and sampling methods were
used to minimize chemical and hydrological disturbances in an around the monitoring wells.
Groundwater was pumped through a flow-through cell equipped with calibrated electrodes for pH,
oxidation-reduction potential (ORP), specific conductance, and dissolved oxygen. Stabilization of
electrode readings was tracked as a function of time (every 1 minute). Final values were recorded after 3
successive readings within ±0.10 for pH, ±10 mV for ORP, ±3% for specific conductance, and ±10% for
dissolved oxygen. After stabilization of the electrode read-outs, turbidity was generally less than 5 NTUs.
Filtered samples (0.45 |im) were collected for the analysis of anions and cations. Unfiltered samples were
collected for the analysis of volatile organic compounds and dissolved gases.
Colorimetric methods were used in the field for determining concentrations of dissolved
oxygen, Fe (II), and hydrogen sulfide. Ferrous iron and sulfide concentrations were measured using the
1,10 phenanthroline and methylene blue indicators, respectively. Dissolved oxygen was determined by
using tests kits that utilize the indigo carmine indicator (DO>1 mg/L), but more typically the rhodazine D
(DO<1 mg/L) colorimetric indicator was employed. Alkalinity determinations were conducted in the
field by titrating samples with standardized sulfuric acid to the bromcresol green-methyl red endpoint.
Quality assurance and quality control practices for field measurements included frequent
checks of electrodes against buffer solutions (pH, ORP, specific conductance). Dissolved oxygen
measurements were checked by reading air-saturated water and comparing results with the temperature-
dependent solubility of oxygen in water. In addition, sodium sulfite was added to water to test the
performance of dissolved oxygen electrodes at low DO levels. Alkalinity measurements were checked by
determinations of prepared sodium carbonate solutions and prepared ferrous ammonium sulfate solutions
were used to check ferrous iron measurements. In general, the methods employed in this study were
found to be suitable for the analysis of geochemical parameters at the PRB sites investigated in this study.
It is worthwhile to note that ferrous iron measurements, ORP measurements, and DO measurements can
be challenging at PRB sites and extra effort must be expended in order to collect high-quality data for
these parameters. The high pH conditions frequently encountered at PRB sites favor rapid oxidation of
Fe (II). Consequently, ferrous iron must be analyzed immediately after sample collection and the Fe (II)
19
-------
Pasquotank River
CO
r-v
CD
Q
5
DQ
®MW46
ML35
MW47.
ML34
ML33
MW35D
e
N
ML14
e
MW38
ML32
ML31
WC21
MW48 ^MLll
o
"Q.
PRB
MW Compliance well
® (5 - 10 ft. screen)
ML Multilevel bundle
0 (6 in. screens)
WC Well cluster
0 (6 in. screens)
'CL
?>
\
MW13
e
%
0 10m
Approx. Scale
HANGAR 79
— r-
j Plating
IShon *
^ —
tGroundwater
flow direction
Figure 3-1. Plan View Map Showing Compliance Well, Bundle and Well Cluster Locations
Relative to Granular Iron Barrier and Cr Plume (Elizabeth City) (June 1994 Data)
Figure 3-2. Plan View Map of the Denver Federal Center PRB
20
-------
Table 3-1. Design Features of PRBs at Sites Evaluated by U.S. EPA
PRB Site Name,
City, State
U.S. Coast Guard
Support Center,
Elizabeth City, NC
Denver Federal
Center, Lakewood,
CO
Pilot/Full
Scale
(Installation
Date)
Full (June 1996)
Full (October
1996)
Type of
Barrier (a)
CRB
F&G
(4 gates)
Reactive Medium
100% Fe
Reactive medium is 100%
Fe; Pretreatment zones are
100% pea gravel
PTZ(b) and
Reactive
Medium
Thickness
N/A
Each gate is
10 ft thick
with 2 - 4 ft
PTZ
Gate/
Barrier
Width (ft)
2
Width of
Fe:
Gate 1 6'
Gate 2 4'
Gate 3 2'
Gate 4 2'
Barrier
Depth (ft)
24
Gate 1 28'
Gate 2 31'
Gate 3 24'
Gate 4 24'
Amount
of Iron
(tons)
280
Source of
Iron
Peerless
Peerless
Notes
PRB length is
150ft
All gates 40 ft in
length
(a) F&G = Funnel and gate; CRB = Continuous reactive barrier
(b) Pretreatment zone (PTZ) is any medium used for homogenizing flow or chemically pre-treating the groundwater.
Use N/A for not available or not applicable.
Table 3-2. Site Hydrogeology and Hydraulic Parameters of the PRB at Sites Evaluated by U.S. EPA
PRB Site
U.S. Coast
Guard Support
Center,
Elizabeth City,
NC
Denver Federal
Center,
Lakewood, CO
Aquifer
Conductivity
(ft/day)
1-30
0.1-100
Groundwater
Gradient
(ft/ft)
0.0011-
0.0033
0.02
Groundwater
Velocity
(ft/day)
0.4-0.6
0.1-1
Aquifer depth
(ft bgs)
24
20-30
Water table
depth (ft bgs)
5-6.5
10-18
Primary Contaminants and
concentrations
Cr(VI) (<10 mg/L); TCE
(<20,000 |^g/L); c-DCE
(<200 |^g/L); VC (<70 |^g/L)
TCE (<700 |^g/L); c-DCE
(<360 |^g/L); TCA (<200
|J,g/L); DCE (<230 |j,g/L)
Notes
From Pacific Western
Technologies, LTD.
(2000)
-------
to
to
Table 3-3. Site Groundwater Geochemistry at Sites Evaluated by U.S. EPA
PRB Site
U.S. Coast Guard
Support Center,
Elizabeth City, NC
Denver Federal
Center, Lakewood,
CO
pH
5.5-6.5
7.1-7.9
Eh
(mV)
-100 to
400
-100 to
300
DO
(mg/L)
0.2-3.0
0.1-4.0
TDS
(mg/L)
250-
400
900-
1200
Ca
(mg/L)
5-20
90-110
Mg
(mg/L)
5-10
17-32
Alkalinity
(mg/L)
30-70
290 - 590
Cl
(mg/L)
20-60
48-81
S04
(mg/L)
5-60
-------
results should be checked against total iron measurements made on acidified samples using methods such
as inductively coupled plasma spectroscopy or atomic absorption spectroscopy. The reducing and iron-
rich environments often encountered at PRB sites are challenging for obtaining reliable electrode and
colorimetric determinations of dissolved oxygen. At low DO levels, electrode response can be slow or
unreliable and the presence of any iron oxidation artifacts can interfere with the rhodazine D colorimetric
tests. While ORP measurements appear to be useful for tracking changes through time of the reductive
capacity of zero-valent iron systems, ORP measurements must be carefully made in the field with
frequent evaluations of electrode performance.
3.1.2.2 Core Sampling and Analysis. To assess the extent of corrosion and mineral build-up on the
iron surfaces, 5 cm i.d. cores were collected at the Elizabeth City PRB. Core barrels were driven using a
pneumatic hammer to the desired sampling location and continuous, up to 110 cm, sections of iron or iron
+ soil were retrieved (Beck et al., 2002). Angle cores (30° relative to vertical) and vertical cores were
collected in order assess the spatial distribution of mineral/biomass buildup in the reactive media. Prior to
pushing the core barrel, an electrical conductivity profile was collected to verify the exact position of the
iron/aquifer interface. The electrical conductivity measurements were instrumental in maximizing the
efficiency of collecting cores that consistently captured the desired aquifer/iron interface. Core materials
from the Elizabeth City site were jet black in color without any obvious signs of oxidation. Cementation
of iron grains was not evident in cores from the Elizabeth City site (after 4 years). Immediately after
collection the cores were frozen. In the laboratory, the frozen cores were partially thawed and then placed
in an anaerobic chamber with a maintained H2-N2 atmosphere. Each core was logged and partitioned into
5 to 10 cm segments. Each segment was homogenized by stirring in the glove box and split into 4 sub-
samples: (1) inorganic carbon analyses, (2) sulfur analyses/x-ray diffraction (XRD), (3) Scanning electron
microscopy (SEM)/x-ray photoelectron spectroscopy (XPS) analyses, and, (4) microbial assays. All sub-
samples were retained in airtight vials to prevent any air oxidation of redox-sensitive constituents prior to
analysis.
Coulometric methods were used to determine the concentrations of inorganic carbon and
sulfur on the iron grains. Sulfur partitioning determinations were made by conducting a series of
chemical extractions to obtain information about the abundances of sulfide, disulfide, and sulfate
precipitates on the iron grains. Compositional information was obtained by using x-ray photoelectron
spectroscopy. Mineralogical analysis was performed using powder x-ray diffraction techniques. X-ray
diffraction patterns were collected on fine-grained materials removed from the iron grains by sonication.
Scanning electron and optical microscopy were utilized to determine the thickness of surface precipitates,
evaluate physical morphology of the iron grains, and the extent of surface coverage. Prior to microscopic
characterization, samples were set in epoxy resin, cured, and ground and polished using standard
techniques. Samples splits were also analyzed for content and distribution of phospholipid fatty acids to
evaluate the abundance and composition of microbial biomass.
It should be noted that while detailed studies of core materials from PRB sites are essential
for: evaluating the geochemical and microbiological processes that impact performance, developing tools
for improving technology selection decisions, and for site characterization investigations, such time-
consuming and expensive studies are not likely to be routinely required as a component of most
performance monitoring programs.
3.1.3 Results and Discussion. At Elizabeth City, concentrations of chromium above 5 |lg/L have
not been observed in any of the reactive barrier compliance wells since November 1996. Over the period
of study, chromium concentrations have remained below 10 |ig/L in wells located up to 20 m
downgradient of the PRB. Similarly concentrations of chlorinated organic compounds (TCE, cis-DCE,
vinyl chloride) at Elizabeth City are below regulatory target levels in downgradient compliance wells.
23
-------
Success of the Elizabeth City zero-valent iron PRB for treating a hexavalent chromium plume
over its first five years correlates with consistent patterns in the commonly measured field parameters
(pH, specific conductance, and Eh). At this site, five years appears to be too short a period of time to
observe a clear correlation between changes in geochemical parameters and declining performance. At
Elizabeth City, subsurface regions of high pH do not necessarily correlate with regions of low Eh. Spatial
and temporal variations in the concentration distribution of terminal electron accepting species (e.g.,
sulfate), specific conductance, and Eh suggest that both anaerobic iron corrosion and microbial activity
play important roles in controlling the oxidation-reduction potential in iron barriers. Low Eh values (<-
100 mV relative to the SHE) and decreases in the specific conductance of groundwater between
upgradient contaminant plumes and sampling points within reactive iron media are consistently indicative
of normally operating PRB systems. Anomalous behavior or trends in these parameters may be useful
indicators of declining iron reactivity.
Mineral precipitates identified in the Fe° barrier at the USCG-SC are broadly consistent with
those predicted to form based on the results of geochemical reaction path models that track the attainment
of chemical equilibrium between selected volumes of groundwater and iron metal (Wilkin and Puls, 2001;
Wilkin et al., 2002). Primary authigenic precipitates identified in the Elizabeth City PRB are calcium
carbonates, iron hydroxy carbonate, carbonate green rust, hydrous ferric hydroxide, ferric oxyhydroxide,
and iron monosulfides (mackinawite and greigite). Microscopy observations indicate that mineral
accumulation mainly occurs on the surfaces of the iron particles collected near the upgradient aquifer/iron
interface where steep gradients in pH and redox potential promote mineral nucleation and growth. After
about 4 years of mineral precipitation and accumulation, a consistent coverage of surface material ranging
in thickness from about 10 to 50 (im is observed on iron grains collected near the upgradient interface at
Elizabeth City (horizontal penetration <8 cm). At greater penetration depths (horizontal penetration >8
cm), surface coatings are discontinuous and <5 (im thick. Accumulation of inorganic carbon precipitates
and sulfur precipitates is greatest near the upgradient aquifer-Fe° interface. Abundance of surface
precipitates decreases with increasing penetration into the iron. Concentrations of inorganic carbon in the
reactive media at Elizabeth City are as high as 2000 mg/kg and authigenic sulfur values approach 1200
mg/kg.
A comparison of groundwater chemistry between upgradient and downgradient wells
indicates that the iron media at Elizabeth City is a long-term sink for C, S, Ca, Si, Mg, and N. Solid
phase characterization studies indicate average rates of inorganic carbon and sulfur accumulation of ~0.1
and -0.05 kg/m2y at Elizabeth City where upgradient waters contain up to 400 mg/L total dissolved solids
(TDS). Carbon accumulation rates, based upon the solid-phase characterization studies, are in good
agreement with estimates made through reactive transport modeling efforts (Mayer et al., 2001).
However, the agreement between measured and modeled sulfur accumulation rates is not as good (D.
Blowes, pers. communication). The reasons for this discrepancy between model predictions and field
measurements are currently being examined.
At the Elizabeth City site, consistent patterns of spatially heterogeneous mineral precipitation
and microbial activity are observed. Mineral precipitates and microbial biomass accumulate the fastest
near the upgradient aquifer-Fe° interface. Porosity loss in the iron media due to precipitation of inorganic
carbon and sulfur minerals was estimated by integrating the concentrations of inorganic carbon and sulfur
as a function of distance in the iron and estimating the volume loss by using the molar volumes of zero-
valent iron, calcium carbonate, iron carbonate, and iron sulfide. The rate of mineral accumulation and the
rate of iron corrosion varies spatially, therefore so does the rate of porosity infilling. The highest
concentrations of mineral precipitates and rates of porosity loss are found adjacent to upgradient
interfaces. At Elizabeth City, a maximum of 5.9% loss of the initial available volume (50%) is estimated
at 2.5 cm into the iron media after 4 years of operation. At distances >8 cm, volume loss decreases
significantly to <0.1% of the initial available volume after 4 years.
24
-------
Microbiological impacts are important to understand in order to better predict how long PRB
systems will remain effective. The presence of a large reservoir of iron coupled with abundant substrate
availability (i.e., hydrogen) supports the metabolic activity of iron-reducing, sulfate-reducing, and/or
methanogenic bacteria. About 35 core samples collected from the USCG-SC PRB were analyzed for
content and distribution of phospholipid fatty acids (PLFA). These organic compounds can be used as
lipid biomarkers to provide a quantitative means to evaluate viable microbial biomass, community
composition, and nutritional status. Biomass concentrations after 4 years at Elizabeth City ranged
between about 5 and 875 pmoles of PLFA per gram of dried iron, or between 1.02 x 105 and 1.78 x 107
cells per gram of iron matrix. The highest concentrations of microbial biomass were again found at the
upgradient aquifer/iron interface. Analysis of PLFA structural groups suggests the dominance of
anaerobic, sulfate-reducing and metal-reducing bacteria. Low concentrations of microbial biomass in
mid-barrier and downgradient samples suggest that the environment at these locations is more challenging
to bacterial growth and survival, which is likely due to substantial decreases in biologically available
electron acceptors such as sulfate and cis-DCE
The principal factors that determine the amount of mineral precipitation and biomass
accumulation in reactive iron media are seepage velocity and groundwater chemistry. After 5 years of
operation, the Elizabeth City barrier has developed a consistent pattern of spatially heterogeneous mineral
precipitation and microbial activity. The development of precipitation and biomass fronts result from the
abrupt geochemical changes that occur at upgradient interface regions coupled with groundwater transport
of dissolved solutes. Complete filling of available pore space has not occurred after 5 years, suggesting
that flow characteristics may not be affected by the accumulation of authigenic components. Even
relatively thin coatings of mineral precipitates that do not affect flow patterns, however, may affect the
reactivity of iron particles with respect to the degradation of chlorinated organic compounds by
diminishing electron flow and the efficiency of reductive degradation processes.
A thorough cost analysis of the Elizabeth City PRB (and 21 other sites) was also undertaken
to complement the long-term performance study by U.S. EPA. Full results for this study will be included
in a forthcoming U.S. EPA report in early 2003. In this analysis, it was found that the largest savings
from use of PRB technology comes in reduced operation and maintenance (O&M) costs. The magnitude
of these savings is dependent on the life of the PRB and changes in the monitoring program over time.
Up front capital costs vary with installation type, size of plume, contaminant concentrations, complexity
of natural site conditions, and other factors. Comparisons were made to PRB and pump-and-treat (p&t)
technologies where comparable levels of data were available for the same site. In some cases capital
costs were greater for PRBs than p&t, while in others, capital costs were less for PRBs. Costs for O&M
were consistently less for PRBs compared to p&t. Indeed, when expressed as fraction of construction
costs, PRB O&M costs were 0.12 times construction costs while p&t O&M costs were 0.41 times
construction costs. Interestingly, the Elizabeth City site had approximately the same construction costs
for both PRB and p&t, but O&M costs were $85,000 and $200,000, respectively. This difference is
actually greater because the largest fraction of the O&M costs for PRBs is monitoring and the $85,000
figure was for the first year of operation. Current annual monitoring costs are $30,000. This is because
the wall is now monitored with much less frequency and analyzed parameters have been optimized. This
is typical for PRBs and indeed recommended by the ITRC (ITRC, 1999).
3.2 Denver Federal Center (Lakewood, CO)
3.2.1 Site Description. In the fall of 1996, the Federal Highway Administration (FHWA) and
General Services Administration (GSA) installed a permeable reactive barrier at the eastern edge of the
Denver Federal Center in Lakewood, Colorado to treat a contaminant plume containing volatile organic
25
-------
compounds, primarily TCE, cis-DCE, TCA, and DCE (Figure 3-2). The DFC PRB has a funnel-and-gate
design configuration. The funnel component of the PRB employs metal sheet pile that was driven into
unweathered bedrock or into resistant, weathered layers of the local bedrock. The DFC PRB has 4
reactive gates, each 12.2 m long, 7.5 to 9.5 m deep, and from 0.6 m (Gate 3 and 4) to 1.8 m (Gate 1) wide.
The design thickness varied because of anticipated differences in contaminant fluxes to the PRB at
different locations along the plume front (McMahon et al., 1999; Parsons Engineering Science, 2000).
3.2.2 Methods
3.2.2.1 Groundwater Sampling. Groundwater was sampled from monitoring wells using peristaltic
pumps. In all cases, low-flow purging and sampling methods were used to minimize chemical and
hydrological disturbances in an around the monitoring wells. Groundwater was pumped through a flow-
through cell equipped with calibrated electrodes for pH, oxidation-reduction potential (ORP), specific
conductance, and dissolved oxygen. Stabilization of electrode readings was tracked as a function of time
(every 1 minute). Filtered samples (0.45 |im) were collected for the analysis of anions and cations.
Unfiltered samples were collected for the analysis of volatile organic compounds and dissolved gases.
Colorimetric methods were used in the field for dissolved oxygen, Fe(II), and hydrogen
sulfide. Ferrous iron and sulfide concentrations were measured using the 1,10 phenanthroline and
methylene blue indicators, respectively. Dissolved oxygen was determined by using tests kits that utilize
the indigo carmine indicator (DO>1 mg/L), but more typically the rhodazine D (DO<1 mg/L)
colorimetric indicator was employed. Alkalinity determinations were conducted in the field by titrating
samples with standardized sulfuric acid to the bromcresol green-methyl red endpoint. Unlike the
Elizabeth City PRB, ground water collected from within and around the DFC PRB frequently contained
concentrations of hydrogen sulfide of up to about 1 mg/L.
Quality assurance and quality control practices for field measurements at the DFC were the
same as those used at the Elizabeth City PRB as described in section 3.1.2.1. These QA procedures for
electrode measurements included frequent checks of the electrodes against buffer solutions (pH, ORP,
specific conductance) and measurements of prepared standard solutions.
3.2.2.2 Core Collection and Analysis. To assess the extent of corrosion and mineral build-up on the
iron surfaces, 5 cm i.d. cores were collected at the DFC PRB. Core barrels were driven using a pneumatic
hammer to the desired sampling location and continuous, up to 110 cm, sections of iron, iron + soil, or
iron + pea-gravel were retrieved. Angle cores (30° relative to vertical) and vertical cores were collected
in order assess the spatial distribution of mineral/biomass buildup in the reactive media. Prior to pushing
the core barrel, an electrical conductivity profile was collected to verify the exact position of the
iron/aquifer interface. Core materials from the Denver Federal Center were jet black in color without any
obvious signs of oxidation. In 2001, after 5 years of operation, some of the iron cores collected from the
DFC showed signs of cementation, nodules of cemented iron grains 1 to 3 cm in diameter were recovered
in some of the retrieved cores. Iron grains from the upgradient interface of DFC Gate 2 were noticeably
enriched in a black-colored, gel-like material. This core consistency was not observed at other DFC gates
or at the Elizabeth City PRB. Immediately after collection the cores were frozen. In the laboratory, the
frozen cores were partially thawed and then placed in an anaerobic chamber with a maintained H2-N2
atmosphere. Each core was logged and partitioned into 5 to 10 cm segments. Each segment was
homogenized by stirring in the glove box and then split into 4 sub-samples: (1) inorganic carbon analyses,
(2) sulfur analyses/x-ray diffraction (XRD), (3) Scanning electron microscopy (SEM)/x-ray photoelectron
spectroscopy (XPS) analyses, and, (4) microbial assays. All sub-samples were retained in airtight vials to
prevent any air oxidation of redox-sensitive constituents prior to analysis.
26
-------
Coulometric methods, wet chemical extractions, x-ray photoelectron spectroscopy, x-ray
diffraction, and high-resolution microscopy were used to examine the chemical properties of iron core
materials retrieved from the DFC PRB. Similar methods were used at the Elizabeth City PRB as
described in section 3.1.2.2. Samples splits were also analyzed for content and distribution of
phospholipid fatty acids to evaluate the abundance and composition of microbial biomass.
3.2.3 Results and Discussion. Microscopy observations indicate that mineral accumulation
mainly occurs on the surfaces of the iron particles collected near the upgradient aquifer/iron interface
where steep gradients in pH and redox potential promote mineral precipitation (Wilkin et al., 2002).
After about 4 years of mineral precipitation and accumulation, a consistent coverage of surface material
ranging in thickness from about 10 to 50 (im is observed on iron grains collected near the upgradient
interface at the DFC (horizontal penetration <20 cm). Therefore, coverage of iron particles by mineral
precipitates extends to greater penetration depths at the DFC as compared to the Elizabeth City PRB. The
principal reason for this is related to a higher average total dissolved solids concentrations at the DFC that
results in greater net rates of mineral precipitation. At greater penetration depths (horizontal penetration
>20 cm), surface coatings are again discontinuous and <5 (im thick. Accumulation of inorganic carbon
precipitates and sulfur precipitates is greatest near the upgradient aquifer-Fe° interface. Abundance of
surface precipitates decreases with increasing penetration into the iron. Greater buildup of mineral
precipitates and microbial biomass was identified in one gate of the three investigated at the Denver
Federal Center (Gate 2). Concentrations of inorganic carbon in DFC Gate 2 are as high as 8000 mg/kg
and total sulfur values approach 4500 mg/kg, or about a factor of about 4x the maximum amounts
observed in DFC Gate 1, DFC Gate 3, and the Elizabeth City PRB (based on core analysis results from
2000, about 4 years after installation).
A comparison of groundwater chemistry between upgradient and downgradient wells
indicates that the iron media at the DFC is a long-term sink for C, S, Ca, Si, Mg, N, and Mn. Solid phase
characterization studies indicate average rates of inorganic carbon and sulfur accumulation of ~2 and -0.8
kg/m2y at the DFC where upgradient waters contain up to 1200 mg/L total dissolved solids (TDS). At the
DFC, consistent patterns of spatially heterogeneous mineral precipitation and microbial activity are
observed. Mineral precipitates and microbial biomass accumulate the fastest near the upgradient aquifer-
Fe° interface. Porosity loss in the iron zones due to precipitation of inorganic carbon and sulfur minerals
was estimated by integrating the concentrations of inorganic carbon and sulfur as a function of distance in
the iron and estimating the volume loss by using the molar volumes of zero-valent iron, calcium
carbonate, iron carbonate, and iron sulfide. The rate of mineral accumulation and the rate of iron
corrosion varies spatially, therefore so does the rate of porosity infilling. The highest concentrations of
mineral precipitates and rates of porosity loss are found adjacent to upgradient interfaces. At the DFC, a
maximum of 14.2% loss of the initial available volume (50%) is estimated at 2.5 cm into the iron media
after 4 years of operation in Gate 2. At distances >10 cm, volume loss decreases to <8% of the initial
available volume after 4 years. In Gate 1 of the DFC, the precipitation front is spread out over a greater
distance, which may be the result of higher average flow rates in Gate 1 (0.38 m/d) compared to Gate 2.
A maximum of 6% porosity lost is estimated for Gate 1 near the upgradient/aquifer interface after 4 years,
decreasing to <0.5% porosity lost at horizontal penetrations >10 cm.
Microbiological impacts are important to understand in order to better predict how long these
PRBs will remain effective. About 35 core samples collected from the DFC PRB was analyzed for
content and distribution of phospholipid fatty acids (PLFA). These organic compounds can be used as
lipid biomarkers to provide a quantitative means to evaluate viable microbial biomass, community
composition, and nutritional status. The highest accumulations of microbial biomass were found at the
DFC Gate 2 near the upgradient iron/aquifer interface, where concentrations were as high as 4,100
pmoles of PLFA per gram of dried iron (8.36 X 107 cells/gm). The analysis of PLFA structural groups
suggests the dominance of anaerobic, sulfate-reducing and metal-reducing bacteria. Lower
27
-------
concentrations of microbial biomass in mid-barrier and downgradient samples suggest that the
environment at these locations is more challenging to bacterial growth and survival, which is likely due to
substantial decreases in biologically available electron acceptors such as sulfate.
At the DFC, Gates 1,3, and 4 have been successful in removing VOCs to concentrations at or
below MCLs. The reactive gates similarly affect contaminant concentrations in downgradient compliance
wells located within 2 m of the gates. Breakthrough of contaminants of Gate 2 occurred soon after the
system was constructed in October 1996. At the Denver Federal Center, successful performance in Gate
1 and Gate 3 is reflected in long-term patterns of pH, specific conductance, and Eh. In Gate 2 of the
DFC, detection of 1,1-DCE at downgradient sampling points has been linked to impacts of the funnel-
and-gate system on groundwater flow and/or perhaps residual contamination in downgradient sediments.
Contaminant breakthrough, particularly of 1,1-DCE is likely related, either directly or indirectly, to
anomalous build-up of authigenic precipitates and biomass on the reactive iron surfaces, which would
lead to decreased efficiency of contaminant degradation reactions. In addition to mineral/biomass
accumulation in Gate 2, potential indicators of decreased performance are increased Eh values, decreased
dissolved hydrogen values, and increases in relative specific conductance values between upgradient
monitoring points and monitoring points within the reactive media.
28
-------
4.0 EVALUATION OF PRBS AT DOE SITES
The two sites evaluated under the DOE study are Y-12 Plant, Oak Pudge, TN and the
Uranium Mill Tailings Site, Monticello, UT. The features of the PPvBs at these two sites are described in
Tables 4-1, 4-2, and 4-3 and the results of the evaluation are described in this section.
4.1 Y-12 S-3 Ponds/Pathway 2 PRB
4.1.1 Site Description. The Y-12 Pathway 2 PRB is located at the U.S. Department of Energy's
Y-12 National Security Complex in Oak Pudge, Tennessee. The trench-style barrier was constructed in
November 1997 to intercept contaminated groundwater upgradient of a shallow creek (Figure 4-1,
Watson et al., 1999). The plume at this location (referred to as S-3 Ponds/Pathway 2, Watson et al.,
1999) contains on the order of 1 mg/L of uranium and between 20 and 150 mg/L of nitrate. The
hydrogeologic setting at this site is rather complex, consisting of densely fractured shale-carbonate
bedrock overlain by low-permeability clay-rich residuum (highly weathered shale) and fill materials
emplaced during construction of the Y-12 plant. Transport through both bedrock and residuum is
fracture-controlled, with geologic strike following an east-west direction. Further adding to the
complexity is the presence of a former streambed channel containing permeable alluvial deposits and
likely acting as preferential flowpath for groundwater in the area.
The trench is approximately 225 feet long, 2 feet wide, and 30 feet deep (Figure 4-1). The
base of the trench is seated at the point of refusal of backhoe penetration, or the approximate top of
competent bedrock. The PRB was constructed using a trench-and-fill operation, where the trench was
initially stabilized using guar gum and subsequently broken down by circulating an enzyme through the
trench after filling (Watson et al., 1999). The PRB was designed to intercept and channel shallow
groundwater through a section of the trench filled with Fe(0) (Figure 4-la). The reactive portion of the
PRB is a 26-ft long trench filled with Peerless™ Fe(0) filings from the base of the trench to a depth of
about 10-12 feet below ground surface, corresponding to seasonal high water levels. Sections of pea
gravel were placed upgradient and downgradient along the long axis of the reactive section, providing
high permeability zones to facilitate capture and discharge of groundwater. Because the Fe(0) and gravel
zones were estimated to be more permeable than the surrounding sediment, it was anticipated that
groundwater would flow along the long axis of the PRB (Figure 4-la). The trench was subsequently
extended and a sump was installed at the distal end to further drive flow along the length of the trench,
after an early tracer test demonstrated significant cross-barrier transport.
Highlights of monitoring activities conducted at this site are presented below, while more
detailed results can be found in Watson et al. (1999), Watson et al. (2000), Phillips et al. (2000), and
Moline et al. (2002). Geochemical modeling specific to this site is described in Liang et al. (2002).
4.1.2 Methods. The performance of the Y-12 S-3 Ponds/Pathway 2 PRB was evaluated using an
integrated approach which consists of monitoring contaminant levels, changes in water chemical
parameters, and hydrologic tests (e.g., water level measurements and tracer tests). Material coring was
conducted at the PRB site and the iron cores were used for mineralogical analysis. Monitoring and
groundwater sampling was achieved through more than 50 single and multilevel piezometers and wells
installed in and around the PRB before and during barrier construction (Figure 4-lb). Groundwater
samples were analyzed for major cations, anions, field parameters (e.g., pH, specific conductance,
temperature, dissolved oxygen, Eh, S2+, Fe2+, and alkalinity), and contaminant concentrations. Details
regarding analytical methods can be found in Watson et al. (1999), Phillips et al. (2000), and Moline et al.
(2002).
29
-------
Table 4-1. Design Features of PRBs at DOE Sites
PRB Site Name,
City, State
Y-12 Plant,
Oak Ridge, TN
Uranium Mill
Tailings Site,
Monticello, UT
Pilot/Full
Scale
(Installation
Date)
Full Scale
(November,
1997)
Full Scale,
(June 30
1999)
Type of
Barrier (a)
CRB
F&G
Reactive Medium
PTZ is 100% pea
gravel; 100% Fe in
reactive cell
PTZisl3%Fe/pea
gravel; 100% Fe in
reactive cell
PTZ (b) and Reactive
Medium
Thickness
2 ft. across barrier
8 ft. total; 2 ft. PTZ, 4
ft reactive zone and 2
ft gravel down
gradient
Gate/
Barrier
Width (ft)
26-ft reactive
cell;
97-ft and
240-ft slurry
walls; 100 ft.
reactive gate
Barrier
Depth
(ft)
22-30 ft.
12-24 ft.
Amount
of Iron
(tons)
80
Source of
Iron
Peerless
Peerless
Notes
Reactive cell
placed within
gravel-filled
capture trench,
guar gum used
during
installation
Air sparging
system was
installed in the
down gradient in
the gate section
(a) F&G = Funnel and gate; CRB = Continuous reactive barrier
(b) Pretreatment zone (PTZ) is any medium used for homogenizing flow or chemically pre-treating the groundwater.
Use N/A for not available or not applicable.
OJ
o
Table 4-2. Site Hydrogeology and Hydraulic Parameters of the PRB at DOE Sites
PRB Site
Y-12 plant,
Oak Ridge, TN
Uranium Mill
Tailings Site,
Monticello, UT
Aquifer
Conductivity
(ft/day)
2.9 - 0.0029
10-89
Groundwater
Gradient
(ft/ft)
0.02; increases
during storm
events
Groundwater
Velocity
(ft/day)
6-20
2-24
Aquifer depth
(ft bgs)
Approx. 30 ft of
highly -weathered
fill & overburden
Approx. 8 ft. of
alluvial deposits
Water table
depth (ft bgs)
10-15 ft., with
seasonal and
storm variations
6-9 ft.
Primary Contaminants and
concentrations (jlg/L unless
otherwise specified)
U (100 - 2700), Tc (<600
pCi/L), NO3 (10 - 1400 mg/L)
U (245 -916), Se(3.5-350),
V (0.237- 481), Mn (6.7-
872) and As (1.1 -13.6)
Notes
Groundwater flow through
the reactive cell is both
parallel and transverse to
the barrier.
K determined from slug
tests; velocity from tracer
tests
-------
Table 4-3. Site Groundwater Geochemistry at DOE Sites
PRB Site
Y-12 plant,
Oak Ridge, TN
Uranium Mill
Tailings Site,
Monticello, UT
pH
5.7-7.0
6.0-
6.88
Eh
(mV)
-300 to
304
-158 to
244
DO
(mg/L)
0.2-5.4
0.16-
5.5
Electric
condutivity
(umhos/cm)
626-4308
2450 - 2540
Ca
(mg/L)
105-
547
216-
295
Mg
(mg/L)
17.2-63
53.3-
75.1
Alkalinity
(mg/L)
91-1100
206 - 480
Cl
(mg/L)
15-186
82.7-
173
S04
(mg/L)
2-147
0.014-
1180
NO3
(mg/L)
38-822
118
Silica
(mg/L)
0.9-8.3
NA
Na
(mg/L)
5-57
248-
326
Notes
High nitrate &
TDS
ground water
High TDS
ground water
-------
(a)
*y 1 Uonilcring Lac;
M. KiHip LiKJilnnii * MiiltpnrlW.lt.
• BDVI Crnick 9.mpMng Luc
I., n
Figure 4-1. Schematic (a) and Plan View (b) of Y-12 Pathway 2 PRB. Monitoring Network also
Shown in (a). (From Watson et al., 1999)
32
-------
Three bromide-tracer tests have been conducted at the Y-12 PRB: at 10 months, 2 and 4 years
after the PRB was constructed. Cores from the PRB were collected at approximately the same time
intervals as the tracer tests. These solid samples were stored in Ar-purged polyvinyl chloride tubes and
processed soon after the collection for mineralogical analysis by x-ray diffraction (XRD) and scanning
electron microscopy (SEM).
To obtain more quantitative information, laboratory and field column experiments were
conducted using the Y-12 site groundwater and careful control of flow rates. Groundwater from a well
upgradient of the PRB was pumped through two large-volume (6-in diameter, 36-in length) columns and
a small volume column was set up in the laboratory to study gas production during Fe(0) treatment. The
results of the column study can be found in Kamalpornwijit et al. (2002).
4.1.3 Results and Discussion. Uranium is removed by Fe(0) within the reactive zone of the PRB.
Variations in U concentrations with time do not follow any specific trends but appear to be mainly from
fluctuations in influent concentrations. The deep sampling ports from wells DP20 and DP21 (Figure 4-2),
where high U concentrations were observed, are actually located below the Fe(0) media zone (within the
saprolite), based on collocated cores collected in September 2001. Although the reactive zone of the PRB
is effective in removing U, contaminant levels in groundwater within the downgradient gravel zone are
comparable to upgradient levels. Flow through the PRB is not occurring along the length of the PRB
alone, and the downgradient gravel zone is likely receiving more groundwater from elsewhere (in the
direction across the PRB) rather than through the reactive zone.
Lack of flow through the long-axis of the PRB was confirmed by all tracer tests at the site. A
concentrated Br solution was injected into well TMW11, which is located in the upgradient gravel zone
about 2-ft from the gravel/Fe interface (see Figure 4-la). High Br levels were detected at the "deep" port
of DP22 within 24 hours, but Br was not detected in significant levels in the shallow and intermediate
ports of DP22 which are only ~4 ft from the injection well (Figure 4-2). Br signals disappeared from the
injection well within 24 hours so groundwater was flowing freely through this well but not through the
reactive zone of the PRB. Aside from the deep port of DP22, Br was only detected at high levels within
the reactive zone in the deepest ports of DP21 and DP20, both of which turned out to be located within
saprolite below the PRB. Angled coring of the PRB from the upgradient gravel zone to Fe(0) zone
intersecting well DP22 (Figure 4-2) revealed significant cementation in the Fe(0) within the shallow zone,
consistent with the tracer test which showed blocked flow between the injection well and the shallow and
intermediate ports of DP22. The primary crystalline phases found in the solid samples using XRD
analysis are aragonite, green rust, siderite and quartz. However, the most abundant solid phase based on
thin sections of the cemented Fe(0) filings was amorphous iron oxyhydroxides; work is ongoing to
identify and quantify this amorphous phase. High nitrate levels in the influent groundwater is likely
leading to the significant corrosion of the Fe(0) material and subsequent precipitation of Fe-oxides in the
reactive zone. Precipitation of carbonate phases further contribute to cementation and clogging within the
reactive zone of the PRB. Geochemical modeling is currently being used to integrate the mineralogical
analysis with the groundwater geochemistry, and to determine whether rapid clogging of the PRB could
have been predicted. Results of the geochemical analyses, as well as comparisons of field results with the
column study (Kamalpornwijit et al., 2002) are the subject of forthcoming publications (Liang et al.,
2002; Moline et al., 2002).
33
-------
Side rite (FeCOS)
Green rust
Aragonite (CaCOS)
K DP
TMW11 K 22
DP
21
DP
20
Aragonite
Green rust
TMW9
Aragonite
Quartz (SiO2)
Aragonite (CaCOS)
Green rust
DP/
I/ /
iDP
\3uartz(SiO2)
Siderite (FeCOS)
Aragonite (FeCOS)
t
DP
23
TMW7
Saprolite
Figure 4-2. Schematic Summary of Coring Results; Cores Collected 4 Years After PRB was
Installed. Gray rectangular area corresponds to Fe(0) zone, TMW11, DP22 etc are monitoring
wells. Dotted and diamond-shaded areas correspond to loose and cemented filings
100-ft slurry wall
groundwater
flow •>—
250-ft slurry wall
2 ft of 3/8-in washed
peagravel + 13%ZVI
4 ft of 100%ZVI
2 ft of 100%
washed pea
gravel
103-ft-long by 8-ft-deep
gate; keyed into bedrock
Figure 4-3. Plan View of Monticello PRB and Detailed Schematic of Permeable Section
34
-------
4.2 Former Uranium Milling Site, Monticello, UT
4.2.1 Site Description. A zero-valent iron funnel-and-gate system was installed in July 1999 to
treat contaminated groundwater at a former uranium milling site in Monticello, UT (Figure 4-3; Morrison
et al., 2001, 2002). The contaminant plume, which consists primarily of uranium and other heavy metals,
lies within alluvial deposits underlain by a mudstone/siltstone aquiclude. Results of pumping and slug
tests conducted within the alluvial aquifer prior to barrier installation indicated a hydraulic conductivity of
about 10"2 cm/s (Morrison et al., 2002), although discrete measurements ranged over four orders of
magnitude. The Monticello PRB was installed across an alluvial valley and consists of a 103-ft long
permeable section containing a 4-ft thick layer of Fe(0) (-8+20 mesh Peerless), a 2-ft layer of 77% pea
gravel 713% Fe (0) mix upgradient and a 2-ft m pea gravel pack downgradient of the Fe(0) layer (Figure
4-3). The permeable section is bounded by two (250 ft and 100 ft long) 8-ft thick slurry walls; the south
wall was keyed into the alluvial valley while the north slurry wall stops ~50 ft from the alluvial valley
wall. The bottom of the PRB was keyed at least ~1 ft into the underlying aquiclude -15-20 ft below
ground surface. The top of the Fe(0) is approximately 0.5 ft below ground surface, covered by a
geotextile and topped to grade by native material.
4.2.2 Methods. A network of 52 wells was installed during PRB construction to provide a means
for collecting groundwater samples as well as evaluating the hydraulic characteristics of the PRB.
Routine contaminant and geochemical groundwater monitoring was conducted by the DOE/Grand
Junction Program Office; details regarding procedures can be found in Morrison et al. (2001) and
Morrison et al. (2002). Hydraulic characterization activities conducted by ORNL at the Monticello PRB
site included groundwater flow velocity measurements using a colloidal borescope, a multiple tracer
(bromide, iodide, neon and helium) test. The borescope measurements and tracer test were performed
approximately one year after the PRB was installed. A more detailed description of the borescope and
tracer test procedures can be found in Liang et al. (2001).
4.2.3 Results and Discussion. Data collected in August 1999 through October 2000 show
uranium, selenium, vanadium, arsenic, molybdenum, and nitrate decrease significantly within the Fe(0)
zone of the PRB when compared to upgradient levels (DOE-GJPO, 2002). Uranium concentrations are
predominantly lower in wells immediately downgradient of the PRB, although some exhibited levels
comparable to upgradient values. These high U concentrations were attributed to leaching from
contaminated aquifer sediments (Morrison et al., 2001), and by-pass of contaminated groundwater around
the ends of the slurry walls (Morrison et al., 2002).
The bromide and iodide tracer test results showed significant heterogeneity and preferential
flow path development within the PRB. Velocities on the order of 2 ft/day were initially anticipated
based on the results of groundwater modeling. However, actual velocities based on the bromide and
iodide breakthroughs in the PRB wells ranged from 2.4 to 18 ft/day, with residence times ranging from 22
to 90 hours. Gas tracers were not detected in any of the wells, even in wells where co-injected bromide
and iodide tracers were being detected. It is unclear whether this was due to analytical difficulties (e.g.,
loss of gases during sample collection) or if the gaseous tracers were actually being stripped by a gas
phase at the upgradient section of the PRB.
Water levels show groundwater flow perpendicular to the long-axis of the PRB, but the tracer
test revealed significant lateral transport. The borescope measurement showed flows that were even
directed upgradient for a few wells. The discrepancies among the estimated flow characteristics were
caused by the scale that different methods evaluate with. Potentiometric surfaces simply show average
gradients, and may not be sensitive to heterogeneity within the scale of a typical PRB thickness (i.e., a
few feet). However, they are useful for assessing larger regional flow patterns, and for delineating gross
features such as groundwater mounding, zones of dewatering, potential for vertical transport, and large
35
-------
permeability changes such as at the influent face of the PRB. Tracer tests provide definitive evidence of
transport, and will average over the distance between monitoring wells and, as such, are scalable. The
borescope method measures localized velocities and is subject to perturbations from structures such as the
PRB walls or even from the borehole itself. The scale of the measurement needs to be taken into account
to interpret field data.
36
-------
5.0 MONITORING AND REGULATORY ISSUES WITH PRBS
5.1 Compliance and Performance Monitoring
A permeable reactive barrier (PRB) monitoring program typically consists of both
compliance and performance monitoring. The objective of the compliance monitoring program is to
ascertain compliance with applicable state standards at designated compliance point(s). This monitoring
is typically driven by regulatory requirements. The contaminated site and the compliance point(s) are the
focus of the monitoring.
The objective of the performance monitoring program is to verify that the PRB is operating as
designed. Performance monitoring generally is focused on the PRB system itself (including impermeable
funnel walls, if present), rather than the entire site or the compliance boundaries. The program should be
designed to verify proper installation of the PRB and identify any changes in the system that would affect
treatment effectiveness. Ability to identify changes, such as loss of reactivity, decrease in permeability,
changes in contaminant residence time within the reaction zone, and short circuiting or leakage through
the funnel walls should be taken into account when designing a performance monitoring program.
5.1.1 Compliance Point. The compliance point typically is chosen at a location downgradient of
the treatment system, where the water quality is expected to meet the groundwater quality standards or
criteria. The chosen location for the compliance point must also ensure protection of downgradient
receptors.
Identification of a compliance point at PRB sites can in many cases be complicated by the
location of the PRB. At a majority of the cases, the PRB has been placed within the contaminated plume
rather than at the leading edge of the plume. In cases where the PRB is within the plume, downgradient
points initially are contaminated. Sampling of wells located downgradient of the PRB will reflect the
initial contamination within the aquifer, including any contamination desorbing from the aquifer soil or
diffusing from finer sediments. It may take several months or even years for downgradient wells to show
any improvement in water quality, depending on the specific site's characteristics. Therefore, when a
PRB is installed within a contaminant plume, additional monitoring wells generally are installed within
the PRB itself (near the downgradient edge of the reactive medium) to monitor contaminant removal. If
the thickness of the reactive media is not sufficient to incorporate monitoring wells, the wells can be
located very near the downgradient edge, while taking into account any difference in geochemical
makeup between the reactive medium and aquifer. Given the time lag in achieving an improvement in
downgradient water quality, regulators and site owners should consider establishing a temporary
compliance point within the reactive medium or near the downgradient edge of the PRB, for some period
of time after installation. This allows an evaluation of the treatment system's ability to remove
contaminants from the groundwater to levels that meet the established regulatory standard or criteria. The
location of the compliance point can be re-evaluated and changed to a more suitable downgradient
location (protective of downgradient receptors) once the water quality in the downgradient aquifer begins
to reflect the treatment occurring within the PRB.
5.1.2 Sampling Parameters. Compliance monitoring parameters for PRBs typically include the
contaminants of concern (e.g., TCE), as well as any potentially deleterious reaction products (e.g., cis-1,2
DCE). In addition, general water quality monitoring (field) parameters, such as pH, alkalinity, specific
conductance, dissolved oxygen, redox potential, and temperature, typically are measured with each
sampling round. Water levels also are an important parameter to identify groundwater flow paths and
hydraulic capture of the treatment system.
37
-------
Performance monitoring focuses on parameters useful in the evaluation of the geochemistry
associated with the treatment system. There is some overlap in parameters between the compliance and
performance monitoring programs. The overlap includes monitoring of the contaminants of concern,
byproducts, general water quality parameters, and water level data. In addition, the native groundwater
species, such as alkalinity, calcium, chloride, iron, magnesium, manganese, nitrate, potassium, sodium,
sulfate, silica and TDS, can be monitored as indicators of short- and long-term PRB performance.
Standard U.S. EPA methodologies should be employed for all analysis.
5.1.3 Sampling Frequency. Monitoring frequency should be determined on a site by site basis.
The frequency of monitoring typically depends on the groundwater flow velocity and the location of the
PRB. Monitoring frequency should consider the amount of contaminated ground water that will flow
through the treatment system over a given period of time. At sites where the PRB is placed at the leading
edge of the plume, monitoring may not be necessary until the contaminant plume reaches the PRB.
As a general guide, monitoring for compliance purposes should be completed on a quarterly
basis. This schedule also allows evaluation of seasonal changes. After the first year or two the PRB
system should be evaluated based upon compliance, performance and stability. A reduction in the
monitoring frequency may be appropriate where the system is operating as originally designed,
on a consistent basis.
5.1.4 Sampling Methods. Sampling within and around a PRB requires special techniques in order
to collect representative samples. Groundwater samples should be collected in such a way that the
residence time of the growth in the PRB is not shortened. Passive sample collection methods are
preferred for the collection of samples. Low flow sampling methods or micro purge techniques should be
used for sampling wells, especially those in close proximity to or within the reactive zone. An example of
a passive sampling compounds (VOCs) is the technique for volatile organic compounds of passive
diffusion bag samplers. Vroblesky (2001) provides guidance on the use of diffusion bag samples.
Field parameter measurements should be conducted with a flow through cell and monitoring
instruments for continuous measurement and to minimize any interferences associated with the
introduction of oxygen into the sample. Field instruments can also be employed as in wells (down-hole
probes) for the collection of field parameter data. Down-hole probes can be inserted in wells either at the
time of the sampling episode or left in a well on a continuous basis (Sivavec et al., 2001).
5.1.5 Monitoring Well Location. The location of monitoring wells is a critical element in
determining whether the PRB is meeting compliance and performance criteria. The ground water model
for the site typically is utilized for locating monitoring wells around a PRB. In general, wells are located
upgradient and downgradient and, if possible, within the PRB itself. In addition, wells at each end of the
PRB are necessary to verify hydraulic capture and evaluate potential plume by-pass. Where a PRB
includes impermeable sections (funnel walls), monitoring the impermeability of that barrier typically is
appropriate. The monitoring wells are generally sized to be just large enough to accommodate the
sampling equipment. Smaller wells, such as l-or-2-inch diameter wells, are preferred to limit the amount
of water that is extracted for sample collection, thereby allowing the efficient collection of a
representative sample that minimizing changes to the groundwater residence in the PRB system.
It is important to achieve complete vertical and horizontal delineation of the contaminant
plume and its relation to the PRB. Monitoring wells should be screened at depths where the highest level
of contamination is migrating through the aquifer. It is preferable to screen all the wells in the vicinity of
the barrier at the same depth interval.
38
-------
Monitoring well locations must be evaluated on a site by site basis. It is important, when
considering the number and location of wells, that all aspects of the contaminant plume are characterized
and conceptually understood. Ground water modeling conducted for the site is an essential tool for
determining the placement of monitoring wells. The appropriate locations and number of monitoring
wells will be dictated by the size and geometry of the contaminate plume, the size of the PRB,
groundwater flow rate, the heterogeneity of the geologic formation, and outside influences on ground
water flow in the vicinity of the PRB.
Elder et al. (2001) used computerized modeling to evaluate the locations of monitoring wells
at PRB sites. MODFLOW and adaptive particle tracking were used to identify flow through a
heterogeneous aquifer with both continuous and funnel and gate PRBs. Monitoring well networks were
identified that provided the fewest monitoring points with the highest probability of detecting the median,
75th and 90th percentiles of effluent concentration. For continuous reactive barriers, a horizontal
monitoring spacing of 15 feet and a vertical spacing of 9 feet was recommended by Elder et al. (2001).
For funnel-and-gate system there is recommended a horizontal spacing of 6 feet and a vertical spacing of
12 feet.
5.2 Contingency Sampling Plan
A contingency sampling plan should be developed whenever a PRB is the chosen remedial
alternative. A contingency sampling plan addresses alternative sampling and investigative techniques that
would be used in a situation where the PRB fails to meet compliance or performance criteria. Techniques
or methods which should be considered as part of the contingency sampling plan include, changes in
monitoring frequency, tracer testing, coring and analysis of the reactive media from the PRB, as well as
long term column testing using site groundwater.
5.3 Other Regulatory Issues
5.3.1 Biostat. The use of guar gum (a natural food thickener) as a reactive medium or as a support
for trench excavation, is gaining increased popularity for the installation of PRBs. Stabilizing the guar
gum prior to installation typically includes the addition of a biostat to slow microbial breakdown of the
mixture. The addition of the biostat into the aquifer has raised regulatory concerns. The biostat has the
potential to contaminate groundwater, both from the original compound used and any degradation
products. It is important to understand the fate and transport of any biostat before it is utilized in these
applications. In some instances regulators have prohibited the use of biostat (Huber et al., 2001) while in
other instances, additional monitoring requirements have been imposed.
5.3.2 Contingency Plans. In many cases, a contingency plan is required in the event that the PRB
fails to meet the compliance criteria. Contingency plans may range from modification of the PRB system
to use of an alternative technology. For instance, the contingency plan could include such options as
extensions to the PRB system or the ability to install a second PRB downgradient of the initial PRB
system. Alternative technologies could include the ability to operate a pump-and-treat system. The need
for a contingency plan should be evaluated during the design of a PRB system.
39
-------
6.0 SUMMARY OF CONCLUSIONS
The study by the three U.S. government agencies (DoD, DOE, and U.S. EPA) and the ITRC
(a DoD study partner) covered more types of PRBs and site characteristics than would have been possible
for any one agency alone. Through periodic teleconferences and meetings, the agencies were able to
quickly transfer new ideas and lessons learned in a way that maximized the effectiveness of the three
studies.
PRBs have emerged as an entire new class of technologies. Just as with pump-and-treat
systems, different hydraulic capture configurations and different permeable barrier media are making it
possible to address a number of contaminants of concern under a number of different site characteristics.
Because zero-valent iron barriers were the first and most common PRBs installed, the tri-agency study
focused primarily on iron barriers. Especially for the longevity evaluation, it was important to focus on
sites with a history of at least a few years of operation. All the PRBs evaluated in this study were of the
trench type (excavate-and-fill type). The more innovative PRB installations, where the reactive medium
is injected into the ground using special methods, such as jetting or hydraulic fracturing, were not
evaluated. The performance of injected PRBs is more difficult to evaluate in the field and was beyond the
scope and resources of these studies. However, the general conclusions of this study are expected to be
applicable to several different types of PRBs, a technology that typically relies on passive groundwater
capture and treatment.
In the short term, the key performance issue is the ability of the PRB to prevent the target
contamination from progressing beyond the plume cutoff location and thus reducing the risk to
downgradient receptors. In the long term, the key performance issue is one of longevity; in other words,
the question of how long a PRB may be expected to retain its reactive and hydraulic performance.
Although post-installation monitoring was the primary tool used by the three agencies in this study, both
short-term and long-term issues are best addressed in the pre-installation design stage at any prospective
PRB site. Therefore, many of the study's recommendations relate to the measures that can be taken in the
design of a PRB. A PRB is a more or less permanent installation, much more so than a pump-and-treat
system. Once a PRB is installed, modifications can be relatively expensive; therefore, it is more
important to get the PRB designed and installed right.
Pre-installation monitoring (site characterization) is an important tool in achieving a good
design. Post-installation monitoring is required to verify compliance and to identify long-term
performance trends. Lessons learned from this study in terms of the monitoring tools available and their
effectiveness provide important pointers for future sites.
Some general conclusions concerning long-term performance of PRBs gleaned from this
study are the following:
a Adequate site characterization, especially to improve understanding of the hydraulic flow
regime at a prospective PRB location, is imperative to maximize the potential for success
of a PRB meeting cleanup goals.
a Low-flow or passive sampling approaches are required for collecting representative data
from PRBs.
a Over long periods of groundwater exposure, the reactivity of the granular zerovalent iron
declines due to precipitation of native groundwater constituents.
40
-------
The ability of easily measurable water quality indicator parameters, such as pH and ORP,
to provide early warning of reduced PRB performance in the long-term is unclear and
requires further study.
In geologic settings where low flow (fine textured formations) conditions exist, extra care
should be taken to insure good hydraulic connection between the native aquifer material
and the permeable reactive zone during system installation.
Increased microbial activity and biomass in the immediate vicinity of a barrier wall may
contribute to loss of reactivity and/or permeability over time.
Additional studies are needed to monitor PRB longevity and improve lifetime predictions
based on site-specific hydrologic, geochemical, and microbiological conditions.
41
-------
7.0 REFERENCES
Arnold, W. A. and Roberts, A. L. 2000. Inter- and intraspecies competitive effects in reactions of
chlorinated ethylenes with zero-valent iron in column reactors. Environmental Engineering and
Science, 77:291-302.
Battelle. 1998. Performance Evaluation of a Pilot-Scale Permeable Reactive Barrier at Former Naval
Air Station Moffett Field, Mountain View, California. Prepared forNFESC, PortHueneme, CA.
November 20.
Battelle. 2000. Design, Construction, and Monitoring of the Permeable Reactive Barrier in Area 5 at
Dover Air Force Base. Final Report Prepared for Air Force Research Laboratory, Tyndall AFB,
FL. March 31.
Beck, F. P., Clark, P. J., and Puls, R. W. 2002. Direct push methods for locating and collecting cores of
aquifer sediment and zero-valent iron from a Permeable Reactive Barrier. Ground Water
Monitoring and Remediation, 22: 165-168.
Blowes, D. W., Gillham, R. W., Ptacek, C. J., Puls, R. W., Bennett, T. A., O'Hannesin, S. F., Bain, J. G.,
and Hanton-Fong, C. J. 1999a. An in situ Permeable Reactive Barrier for the treatment of
hexavalent chromium and trichloroethylene in ground water, Vol. 1, Design and installation.
U. S. Environmental Protection Agency, EPA/600/R-99/095a, 111 pp.
Blowes, D. W., Puls, R. W., Gillham, R. W., Ptacek, C. J., Bennett, T. A., Bain, J. G., Hanton-Fong, C. J.,
and Paul, C. J. 1999b. An in situ Permeable Reactive Barrier for the treatment of hexavalent
chromium and trichloroethylene in ground water, Vol. 2, Performance monitoring. U. S.
Environmental Protection Agency, EPA/600/R-99/095b, 207 pp.
Department of Energy Grand Junction Program Office (DOE-GJPO). 2002 (date of website access).
http://www.gj o .doe .gov/perm-barr/proj ects/monticello/monticello .htm
Elder, C. R., Benson, C. H., and Eykholt, G. R. 2001. Economic and performance based design of
monitoring systems for PRBs. In Proceedings of the International Containment and Remediation
Technology Conference and Exhibition, Orlando, FL (June 10-13).
Einarson, M. D., Langdon, R. L., and Einarson, J. F. 2000. Hydraulic Performance of a Funnel-and-
Gate Treatment System in a Shallow Tidally-Affected Aquifer. Prepared for NFESC, Port
Hueneme, CA.
Envirometal Technologies Inc. 2002. Probabilistic Design for Granular Iron PRBs. Technical Note
3.04.
Gavaskar, A.R., Gupta, N., Sass, B., Janosy, R., and Hicks, J. 2000. Design Guidance for Application of
Permeable Reactive Barriers for Groundwater Remediation. Prepared for the Air Force Research
Laboratory, Tyndall AFB, FL (March 31).
Gavaskar, A., Sass, B., Gupta, N., Drescher, E., Yoon, W., Sminchak, J., and Hicks, J. 2002. Evaluating
the Longevity and Hydraulic Performance of Permeable Reactive Barriers at Department of
Defense Sites. Prepared for Naval Facilities Engineering Service Center. Port Hueneme, CA.
Prepared by Battelle, Columbus, OH (April 24).
42
-------
Huber, D., Manganaro, D., Hurley, P., Commerford, J., and Richards, P. 2001. The use of the Biostat
Troysan 142 in a Permeable Reactive Barrier and its potential effects on public heath and the
environment. In Proceedings of the 2001 Contaminated Soils, Sediments and Water Conference,
Amherst, MA (October 22-25).
ITRC. 1999. Regulatory Guidance for Permeable Reactive Barriers Designed to Remediate Inorganic
and Radionuclide Contamination. Interstate Technology and Regulatory Council Work Group,
Permeable Reactive Barriers Work Team.
Kamalpornwijit, W., Liang, L., West, O. R., Moline, G. R., and Sullivan, A. B. 2002. Heterogeneity
development and its influence on long-term PRB performance: A column study. (Submitted to
Journal of Contaminant Hydrology).
Liang, L., Korte, N. E., Moline, G. R., and West, O. R. 2001. Long-term Monitoring of Permeable
Reactive Barriers: Progress Report. ORNL/TM-2001/1. Oak Ridge National Laboratory, Oak
Ridge, Tennessee, 51 p.
Liang, L., Sullivan, A. B., West, O. R., and Gu, B. 2002. Predicting the precipitation of mineral phases
in permeable reactive barriers (in preparation).
Mayer, K. U., Blowes, D. W., and Frind, E. O. 2001. Reactive transport modeling of an in situ reactive
barrier for the treatment of hexavalent chromium and trichloroethylene. Water Resources
Research, 37: 3091-3103.
McMahon, P. B., Dennehy, K. F., and Sandstrom, M. W. 1999. Hydraulic and geochemical performance
of a permeable reactive barrier containing zero-valent iron, Denver Federal Center. Ground
Water, 37: 396-404.
Moline, G., Liang, L., Kamolpornwijit, W., and West, O. R. 2002. Interlinked hydrologic and
geochemical processes affecting the performance of zero-valent iron barriers: a field study at the
Y-12 Plant, Oak Ridge, Tennessee (in preparation).
Morrison, S., Metzler, D. R., Carpenter, C. E. 2001. Uranium precipitation in a Permeable Reactive
Barrier by progressive irreversible dissolution of zerovalent iron. Environmental Science and
Technology, 35: 385-390.
Morrison, S., Carpenter, C., Metzler, D. 2002. Design and Performance of a Permeable reactive barrier
for containment of uranium and associated contaminants at Monticello, Utah, USA. In
Groundwater Remediation of Metals, Radionuclides, and Nutrients with Permeable Reactive
Barriers.
O'Hannesin, S. F. 1993. A field demonstration of a permeable reaction wall for the in situ abiotic
degradation of halogenated aliphatic organic compounds. Unpublished M.S. thesis, University of
Waterloo, Ontario, Canada.
Pacific Western Technologies, Ltd. 2000. 7999 Performance Evaluation of the Interim Groundwater
Remediation Measure, Denver Federal Center. Prepared for Federal Highway Administration
and General Services Administration, Denver CO.
43
-------
Parsons Engineering Science, Inc. 2000. Draft Feasibility Memorandum for Groundwater Remediation
Alternatives Using Zero Valent Iron Reactive Wall at the Ash Landfill, Seneca Army Depot
Activity Romulus, New York. Prepared for Army Corps of Engineers, Huntsville, AL.
Phillips, D. H., Gu, B., Watson, D. B., Roh, Y., Liang, L., and Lee, S. Y. 2000. Performance evaluation
of a zerovalent iron reactive barrier: Mineralogical characteristics. Environmental Science and
Technology, 34: 4169-4176.
Puls, R. W., Blowes, D. W., and Gillham, R. W. 1999. Long-term performance monitoring for a
permeable reactive barrier at the U.S. Coast Guard Support Center, Elizabeth City, North
Carolina. Journal of Hazardous Materials, 68: 109-124.
Sass, B., A. Gavaskar, W-S. Yoon, N. Gupta, E. Drescher, and C. Reeter. 2001. Geochemical
investigation of three Permeable Reactive Barriers to assess impact of precipitation on
performance and longevity. In Proceedings of the International Containment & Remediation
Technology Conference and Exhibition, Orlando, FL (June 11-13).
Sivavec, T., Krug, T., Berry-Spark, K., Forct, R. 2001. Performance monitoring of a Permeable Reactive
Barrier at the Somersworth, NH Landfill Superfund Site, submitted for publication in the ACS
volume Innovative Strategies for the Remediation of Chlorinated Solvents and DNAPLs in the
Subsurface.
Totten, L. A., Jans, U., and Roberts, A. L. 2001. Alkyl bromides as mechanistic probes of reductive
dehalogenation: Reactions of vicinal dibromide stereoisomers with zero-valent metals."
Environmental Science and Technology, 35: 2268-2274.
Versar, Inc. 1997. Final Evaluation of Long-Term Performance Reactive Wall Demonstration Project.
Prepared for Air Force Center for Environmental Excellence (AFCEE/ERB) Base Closure
Restoration Division. Prepared for Brooks AFB, TX (December).
Vroblesky, D. A. 2001. Users guide for polyethylene-based passive diffusion bag samplers to obtain
volatile organic compound concentrations in wells. USGS Water Resources Investigation Report
01-4060.
Watson, D., Gu, B., Phillips, D. H., and Lee, S. Y. 1999. Evaluation of permeable reactive barriers for
removal of uranium and other inorganics at the Department of Energy Y-12 Plant, S-3 disposal
ponds. ORNL/TM-1999/143. Oak Ridge National Laboratory, Oak Ridge, TN.
Watson, D., Gu, B., Phillips, D., and Lee, S. 2000. Evaluation of permeable reactive barriers for removal
of uranium and other inorganics at the Department of Energy Y-12 Plant, S-3 disposal ponds (In
preparation).
Wilkin, R. T. and Puls, R. W. 2001. Carbon and sulfur accumulation and iron mineral transformations in
Permeable Reactive Barriers containing zero-valent iron: Impact on long-term performance. In
Proceedings of the International Containment & Remediation Technology Conference and
Exhibition, Orlando, FL (June 11-13).
Wilkin, R.T., Puls, R.W., and Sewell, G.W. 2002. Long-term performance of Permeable Reactive
Barriers for ground water remediation: An evaluation at two sites. U.S. EPA Environmental
Research Brief, U.S. EPA/600/S-02/001.
44
-------
Yabusaki, S., Cantrell, K., Sass, B., and Steefel, C. 2001. Multicomponent reactive transport in an in situ
zero-valent iron cell." Environmental Science and Technology, 35: 1493-1503.
45
-------
-------
Solid Waste and EPA542-R-04-004
Emergency Response February 2004
(5102G) www.epa.gov
www.frtr.gov
National Service Center for
Environmental Publications
P.O. Box 42419
Cincinnati, OH 45242
Official Business
Penalty for Private Use $300
------- |