&EPA
RTDF
United States Office of Research and Office of Solid Waste and EPA/600/R-98/125
Environmental Development Emergency Response September 1998
Protection Agency Washington DC 20460 Washington DC 20460
Permeable Reactive Barrier
Technologies for
Contaminant Remediation
Permeable Reactive Barrier
-------
EPA/600/R-98/125
September 1998
Robert M. Powell
Powell & Associates Science Services
Las Vegas, Nevada
David W. Blowes
University of Waterloo
Waterloo, Ontario
Robert W. Gillham
University of Waterloo
Waterloo, Ontario
Schultz
Dupont Company
Newark, Delaware
Timothy Slvavec
GE Corporate Research and Development Center
Schenectady, New York
Robert W. Puls
U.S. EPA National Risk Management Research
Laboratory
Ada, Oklahoma
John L. Vogan
EnviroMetal Technologies Inc.
Guelph, Ontario
Patricia D. Powell
Powell & Associates Science Services
Las Vegas, Nevada
Rich Landis
Dupont Company
Wilmington, Delaware
Work Assignment Manager
Robert W. Puls
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
Ada, Oklahoma 74820
Technical Innovation Office
Office of SolidWaste and Emergency Response
U. S. Environmental Protection Agency
Washington DC 20460
National Risk Management Research Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
Notice
The U.S. Environmental Protection Agency through its Office of Research and Development funded and
collaborated in the research described here as indicated through the authors' affiliations. ManTech Environmental
Research Services Corporation is an in-house support contractor to NRMRL/SPRD under 68-C3-0322 (Roger L.
Cosby. Project Officer). Dynamac is an off-site support contractor to NRMRL/SPRD under 68-C4-0031 (David S.
Burden. Project Officer). Powell and Associates prepared the document as a subcontractor to ManTech
Environmental Research Services Corporation. Dupont and General Electric provided additional support by
providing time for Mr. Rich Landis. Dr. Dale Schultz, and Dr. Tim Sivavcc to work on this report. This report has
been subjected to the Agency's peer and administrative review and has been approved for publication as an EPA
document. Mention of trade names or commercial products does not constitute endorsement or recommendation
for use.
The Remedial Technology Development Forum (RTDF) provided support by way of authorship and document
review. All authors are members of the Permeable Reactive Barriers Action Team of the RTDF. Members of the
steering committee of the Permeable Reactive Barriers Action Team of the RTDF, not involved in authorship,
reviewed this document as well as members of the U. S. EPA's Ground Water and Engineering Forums. Specific
individuals who reviewed the document and provided extensive comments which significantly enhanced the
technical quality included Mr. John Vidumsky (Dupont), Mr. Don Marcus, Mr. Mat Turner (New Jersey
Department Environmental Protection), Mr. Steve White (U.S. Army Corps of Engineers), Major Ed Marchand
(U.S. Air Force, AFCEE), and Mr. Peter McMahon (USGS).
All research projects making conclusions or recommendations based on environmentally related measurements
and funded by the U.S. Environmental Protection Agency are required to participate in the Agency Quality
Assurance Program. This project did not involve environmental related measurements and did not require a Quality
Assurance Plan.
-------
FOREWORD
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's land, air, and
water resources. Under a mandate of national environmental laws, the Agency strives to formulate and implement
actions leading to a compatible balance between human activities and the ability of natural systems to support and
nurture life. To meet these mandates, EPA's research program is providing data and technical support for solving
environmental problems today and building a science knowledge base necessary to manage our ecological resources
wisely, understand how pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory (NRMRL) is the Agency's center for investigation of tech-
nological and management approaches for reducing risks from threats to human health and the environment. The
focus of the Laboratory's research program is on methods for the prevention and control of pollution to air, land,
water, and subsurface resources; protection of water quality in public water systems; remediation of contaminated
sites and ground water; and prevention and control of indoor air pollution. The goal of this research effort is to
catalyze development and implementation of innovative, cost-effective environmental technologies; develop scien-
tific and engineering information needed by EPA to support regulatory and policy decisions; and provide technical
support and information transfer to ensure effective implementation of environmental regulations and strategies.
Environmental scientists are generally familiar with the concept of barriers for restricting the movement of con-
taminant plumes in ground water. Such barriers are typically constructed of highly impermeable emplacements of
materials such as grouts, slurries, or sheet pilings to form a subsurface "wall". The goal of such installations is to
eliminate the possibility that a contaminant plume can move toward and endanger sensitive receptors such as
drinking water wells or discharge into surface waters. Permeable reactive barrier walls reverse this concept of
subsurface barriers. Rather than serving to constrain plume migration, permeable reactive barriers (PRB's) are de-
signed as preferential conduits for the contaminated ground water flow. A permeable reactive subsurface barrier is
an emplacement of reactive materials where a contaminant plume must move through it as it flows, typically under
natural gradient, and treated water exits on the other side. The purpose of this document is to provide the most recent
information on PRB technologies in a format that is useful to stakeholders such as implementors, state and federal
regulators, Native American tribes, consultors, contractors, and all other interested parties. It includes information
on treatable contaminants, design, feasibility studies, construction options, site characterization needs and monitor-
ing, as well as summaries of several current installations. It is hoped that this will prove to be a very valuable
technical resource for all parties with interest in the implementation of this innovative, passive, remedial technology.
Clinton W. Hall, Director
Subsurface Protection and Remediation Division
National Risk Management Research Laboratory
-------
Table of Contents
1.0 Executive Summary 1
2.0 Introduction 4
2.1 Defining Permeable Reactive Subsurface Barriers 4
2.2 Research Support and Application ofThe PRO Technology 5
2.3 Regulatory Acceptance of PRB Technology 6
3.0 Treatable Contaminants, Reactants and Reaction Mechanisms 8
3.1 Desirable Characteristics of Reactive Media 8
3.2 Treatable Contaminants 8
3.2.1 Halogenated Organic Compounds and Iron 8
3.2.2 Redox-sensitive Inorganic Anions and Iron 11
3.2.3 Biologically Mediated Reduction and Removal of Anions 13
3.2.4 Adsorption and Precipitation of Inorganic Anions 14
3.2.5 Reduction of Inorganic Cations 14
3.2.6 Biologically Mediated Reduction and Precipitation of Cations 15
3.3 Enhancing Iron Reaction Rates 15
4.0 Remediation Feasibility, Laboratory Trcatability, and PRB Design Studies 16
4.1 Laboratory Treatability Studies 16
4.1.1 Batch Studies 16
4.1.2 Column Studies 18
4.1.2.1 Column Test Methodology 18
4.1.2.2 Interpretation of Column Data 18
4.1.2.3 Interpretation of Changes in Non-Contaminant Inorganic Constituents 20
4.2 Determination of Required Residence Time in PRBs 22
4.2.1 Chlorinated Volatile Organic Compounds (VOCs) 22
4.2.2 Inorganic Constituents 22
4.3 Ancillary Laboratory Studies 23
5.0 Site Characterization for Permeable Reactive Barriers 24
5.1 Hydrogeologic Characterization 24
5.2 Contaminant Characterization 25
5.3 Geochemical Characterization 25
5.4 Microbial Aspects 26
5.5 Implementing the Field Characterization 27
6.0 Permeable Reactive Barrier Design 28
6.1 The Continuous PRB and Funnel and Gate Designs 28
6.2 Emplacement Methods and Comparisons 30
6.2.1 Conventional Excavation 30
6.2.2 Trenching Machines 31
6.2.3 Tremie Tube/Mandrel 31
6.2.4 Deep Soil Mixing 31
6.2.5 High-Pressure Jetting 32
6.2.6 Vertical Hydraulic Fracturing and Reactant Sand-Fracturing 32
6.3 Emplacement Verification 32
7.0 Monitoring Permeable Reactive Barriers 33
7.1 Planning the Monitoring Effort 33
7.2 Compliance Monitoring 33
7.2.1 Objectives 33
7.2.2 Compliance Monitoring System Design 33
7.2.3 Compliance Sampling Methods 34
7.2.4 Compliance Sampling Frequency 35
7.2.5 Contaminant Breakthrough/Bypass and Formation of Undesirable Products 36
7.2.6 General Water Quality Parameters 36
IV
-------
7.3 Performance Monitoring 37
7.3.1 Objectives 37
7.3.2 Performance Monitoring System Design 37
7.3.3 Performance Sampling Methods 38
7.3.4 Performance Sampling Frequency 38
7.3.5 Contaminant Degradation/Transformation 39
7.3.6 Geochemical Indicator Parameters 39
7.3.7 Coring for Precipitate Buildup, Microbial Effects 39
7.3.8 Hydrologic Testing for Permeability Alteration 39
8.0 Field Installations " 40
8.1 Chlorinated Hydrocarbon and Chromium Removal in Field Scale Systems 40
8.1.1 Industrial Site, Sunnyvale, California (January 1995) 40
8.1.2 Industrial Site, Mountainview, California (September 1995) 40
8.1.3 Industrial Site, Belfast Northern Ireland (December 1995) 43
8.1.4 Industrial Site, Kansas (January 1996) 43
8.1.5 USCG Facility. Elizabeth City, North Carolina (June 1996) 43
8.1.6 FHA Facility, Lakewood, Colorado (October 1996) 46
8.2 Interpretation of VOC Monitoring Data from the Field-scale Systems 47
8.3 Inorganic Constituent Removal in Field-scale Systems 47
8.3.1 Nickel Rim Mine Site, Sudbury, Ontario (August 1995) 47
8.3.2 Langton, Ontario On Site Wastewater Treatment (July, 1993) 49
8.4 Biological Effects on Field-scale PRBs * 50
8.4.1 Microbial Effects on Iron PRBs 50
8.4.2 Microbial Effects in PRBs for Inorganic Constituents 50
8.5 Effects of Mineral Precipitation on Field-scale PRB Performance 51
Appendix A. Reference Table A-53
Appendix B. Explanation of Relevant Physical/Chemical Phenomena B-74
Appendix C. Scoping Calculations C-78
Appendix D. Acronyms D-83
Appendix E. Glossary E-86
References 89
-------
List of Tables
Table 1. Full-Scale in situ Remediation Projects 5
Table 2. States That Have Approved the Installation of Iron PRBs 6
Table 3. Contaminants Treatable By Reactive Materials in PRBs 9
Table 4. Contaminants Presently Not Treatable By Fc(0) 9
Table 5. Contaminants With Unknown Treatability 10
Table 6. Typical Treatability Study Results, Elizabeth City Example 17
Table 7. Construction Details of Six PRB Installations 41
Table 8. VOC Degradation Rate Data 49
Table 9. Required Weight-Per-Area (W/A) of Granular Iron. Calculated with n = 0.33,
P0/P = 1000. and F = 3.5). Information and Equations from Appendix C C-79
Table 10. Summary Information on Full-Scale Permeable Reactive Barriers Using Fe(0) C-80
Table 11. Cost Elements of Funnel and Gate Systems: High Estimate
(Gate @ $1000/yd3; Funnel @ $25/ft2) C-82
Table 12. Cost Elements of Funnel and Gate Systems: Low Estimate
(Gate @ $200/yd3; Funnel @ $10/ft2) C-82
List of Figures
Figure 1. Example of plume being treated by a permeable reactive barrier wall 1
Figure 2. Schematic of the column apparatus used in the bench-scale tests 19
Figure 3. Conceptual residence tme calculation for VOC removal 21
Figure 4. Correlation of TCE degradation rates with temperature 21
Figure 5. Potential effects of ground water flow diversion on plume direction 24
Figure 6. Effects of plume center-of-mass movement downgradient, and higher
concentrations impacting the barrier. Incomplete remediation is now occurring 25
Figure 7. Potential effects of precipitation buildup on the iron overtime 26
Figure 8a. Plume capture by a continuous PRB trenched system.
The plume moves unimpeded through the reactive gate 28
Figure 8b. Plume capture by a fiinncl-and-gate system. Sheet piling
funnels direct the plume through the reactive gate 28
Figure 9. Continuous trencher used at Elizabeth City, North Carolina 34
Figure 10. Bundled tubes with short screens for performance monitoring 38
Figure 11. PRB configuration. Sunnyvale. California 40
Figure 12. PRB configuration. Mountainview. California 42
Figure 13. PRB vessel configuration, Belfast, Northern Ireland 44
Figure 14. PRB configuration, Kansas 44
Figure 15. PRB configuration. USCG facility, Elizabeth City, North Carolina 45
Figure 16. PRB configuration. FHA facility, Denver, Colorado 46
Figure 17. Pilot scale PRB configuration, Colorado and New York 48
Figure 18. Example of an electrochemical corrosion cell B-75
VI
-------
VII
-------
1.0
Executive Summary
Perhaps no recent remedial technology has generated as much interest as the use of subsurface permeable reactive
barriers (PRBs). This is due to the perceived PRJB cost/benefit ratio and the potential of PRBs to mitigate the spread
of contaminants that have proven difficult and expensive to manage with other cleanup methods. The concept of a
PRB is relatively simple. Reactive material is placed in the subsurface where a plume of contaminated ground water
must move through it as it flows, typically under its natural gradient (creating a passive treatment system) and
treated water conies out the other side (Figure 1). The PRB is not a barrier to the water, but it is a barrier to the
contaminant. When properly designed and implemented, PRBs are capable of remediating a number of contami-
nants to regulator}' concentration goals. It is currently believed that these systems, once installed, will have
extremely low, if any, maintenance costs for at least five to ten years. There should be no operational costs other
than routine compliance and performance monitoring.
The majority of installed PRBs use iron metal, Fe(0), as the reactive media for converting contaminants to non-
toxic or immobile species. Iron metal has the ability to rcductivcly dchalogcnatc hydrocarbons, such as converting
trichloroethene (TCE) to ethene. It can also reductively precipitate anions and oxyanions, such as converting
soluble Cr(VI) oxides to insoluble Cr(III) hydroxides. Organic materials are being used as reactive media in some
PRBs to biologically remediate certain other contaminants, such as nitrate and sulfate. Both laboratory and field
results have shown that the rate of transformation of these and many other contaminants is sufficiently rapid for
PRBs to be successfully used as full-scale remediation systems. Numerous other reactive materials are being
investigated, as are means to enhance both the iron and biological reactions. Many of the references for these
investigations are listed and described in Appendix A.
Commercial PRBs arc currently built in two basic configurations (although others arc being evaluated), the funncl-
and-gate and the continuous PRB. Both have required some degree of excavation and been limited to fairly shallow
depths of fifty to seventy feet or less. Newer techniques for emplacing reactive media, such as the injection of
slurries, hydrofracturing, driving mandrels, etc., may serve to overcome some of these emplacement limitations.
The funncl-and-gatc design PRB uses impermeable walls (sheet pilings, slum' walls, etc.) as a '"funnel" to direct the
contaminant plume to a "gate(s)" containing the reactive media, whereas the continuous PRB completely transects
the plume flow path with reactive media. Due to the funnels, the fiinnel-and-gate design has a greater impact on
altering the ground-water flow than does the continuous PRB. In both designs it is necessary to keep the reactive
zone permeability equal to or greater than the permeability of the aquifer to avoid diversion of the flowing waters
around the reactive zone.
Waste
61 Flow-
Plume
Figure 1. Example of plume being treated by a permeable reactive barrier wall.
-------
Several important issues must be addressed when considering contaminant remediation through the use of PRB
technology. These include the nature of the contaminant and the availability of reactive media that can transform
the contaminant yet remain reactive, in situ, for relatively long time periods. For contaminants of unknown
treatability or media of unknown reactivity, addressing these issues will require laboratory studies using both batch
and column techniques. The mobility, toxicity and stability of the transformation products resulting from the
contaminant and media interactions must also be assessed. If these transformation products are regulated
compounds, they must not exit the reactive zone of the PRB without themselves being immobilized or transformed
to innocuous compounds.
A thorough understanding of system hydrogeology and plume boundaries is needed prior to implementing a PRB,
due to the need for the plume to passively flow through the reactive zone of the PRB. The hydrogeologic
characterization must also yield information suitable for determining the rate of ground-water flow through the
reactive zone of the PRB. This is necessary to establish the ground-water/contaminant residence time per unit
thickness of reactive media which, when combined with the contaminant transformation rate as it passes through
the media, determines the total thickness of reactive media that is required. During PRB installation the reactive
media must be made accessible to the contaminant by some emplacement method and. as with most remedial
technologies, this becomes increasingly difficult at greater contaminant depth or for contaminants in fractured rock.
Once installed, the PRB should be carefully monitored for both compliance and performance; compliance to
ascertain that regulatory contamination goals are being met. and performance to assess whether the PRB
emplacement is meeting its design criteria and longevity expectations.
As for any remedial technology, it is important to fully understand the factors that can result in either successful
implementation and remediation or failure to achieve the remedial design goals. This document addresses the
factors, such as those mentioned above, that have been found to be relevant for successfully implementing PRBs for
contaminant remediation. Additionally, it provides sufficient background in the science of PRB technology to allow
a basic understanding of the chemical reactions proposed for the contaminant transformations that have been
witnessed both in the laboratory and in field settings. It contains sections on PRB-treatable contaminants and the
treatment reaction mechanisms, feasibility studies for PRB implementation, site characterization for PRBs, PRB
design, PRB emplacement, monitoring for both compliance and performance, and summaries of several field
installations. The appendices supplement this information with a detailed table of information available in the
literature through 1997, summarizing the significant findings of PRB research and field studies (Appendix A), a
further examination of the physical and chemical processes important to PRBs, such as corrosion, adsorption, and
precipitation (Appendix B), and a set of scoping calculations that can be used to estimate the amount of reactive
media required and facilitate choosing among the possible means of emplacing the required amount of media
(Appendix C). Appendix D provides a list of acronyms and Appendix E a glossary of terms that are used within this
document.
Hie goal of this Issue Paper is to provide the most recent information available on PRB technologies and to do so
in a format that is useful to stakeholders such as implementors. state and federal regulators. Native American tribes,
consultants, contractors, and all other interested parties. Other documents are also available which address PRB
topics that are not discussed in detail in this report to avoid duplicative effort, such as regulator}? issues related to
PRB technology and cost information. For example, the Interstate Technology and Regulatory Cooperation (ITRC)
Workgroup (Permeable Barrier Wall Subgroup) has prepared a document titled "Regulator}? Guidance for
Permeable Barrier Walls Designed to Remediate Chlorinated Solvents" (ITRC, 1997) and the Environics Director-
ate, U.S. Air Force, has published "Design Guidance for Application of Permeable Barriers to Remediate Dissolved
Chlorinated Solvents" (Battelle. 1997). Documents on the costs of PRB technology are being prepared by U.S.
EPA's Technology Innovation Office (TIO) and by its Office of Research and Development, National Risk
Management Research Laboratory (ORD-NRMRL). Several web sites also provide information about PRB
technology. Among these are:
RemedialTechnologies Development Forum
• http://www.rtdf.org
Ground-Water Remediation Technologies Analysis Center
• http://www.gwrtac.org
-------
U.S. EPA's Robert S. Kerr Environmental Research Center of the National Risk Management
Research Laboratory
• http://www.epa.gov/ada/eliz.html
EnviroMetal Technologies, Inc.
• http://www.beak.com/eti.html
Powell & Associates Science Services
« http://www.powellassociates.com
Oregon Graduate Institute of Science and Technology
* http://www.cgr.ese.ogi.edu/iron
-------
2.0 Introduction
A great deal of money and effort has been spent on environmental restoration during the past 30 years. Significant
progress has been made improving air quality, dumps and landfills, and surface water quality, although challenges
still exist in these areas. Among the more difficult and expensive environmental problems, and often the primary
factor limiting closure of contaminated sites following surface restoration, is contaminated ground water. The most
common technology used for remediating ground water has been to pump the water and treat it at the surface.
Although still useful for certain remedial scenarios, the limitations of pump-and-treat technologies have recently
been recognized, along with the need for innovative solutions to ground-water contamination (Keely, 1989;
National Research Council, 1994).
One of the most promising of these innovative solutions is the use of permeable reactive barriers (PRBs) filled
with reactive material to intercept and decontaminate plumes in the subsurface. The concept of PRBs is relatively
simple. Reactive material is placed in the subsurface to intercept a plume of contaminated ground water which must
move through it as it flows, typically under its natural gradient, thereby creating a passive treatment system. As the
contaminant moves through the material, reactions occur that transform it to less harmful (nontoxic) or immobile
species.
Many reactive media combinations can be envisioned for use in PRBs and numerous media and mixtures of media
are being investigated for a variety of contaminants (Appendix A). As of this writing, iron metal, variously
designated as Fe°, Fe(0), or zero-valent iron, is the most common reactive media in the majority of field-scale and
commercial implementations. Scrap iron is not expensive and can be obtained in a granular form in the large
quantities needed. It has the ability to reduce oxidized inorganic species and reductively dehalogenate hydrocarbon
compounds. Reactive iron barriers depend upon corrosion to drive these reactions (Appendix B). For example,
chromate plumes are reduced from Cr(VI) to Cr(III) and precipitated, in situ, as chromium (III) hydroxides or
chromium-iron hydroxide solid solutions. An example from among the amenable halocarbon plumes are those
resulting from the dense, nonaqueous phase liquid (DNAPL) halogenated hydrocarbons. These include chlorinated
ethenes such as perchlorocthylcne (PCE), trichlorethyiene (TCE), dichloroethylenc (DCE), and vinyl chloride
(VC). The reductive dehalogenation of these compounds occurs due to electron transfers from the iron to the
halocarbon at the iron surface. This results in the halogen ions (e.g., Cl") being replaced by hydrogen species,
ultimately yielding ethene or ethane that can be mineralized via biodegradation.
The ability to perform dehalogenation in a reactive barrier is significant since the sources of DNAPL contamina-
tion, such as residual saturation, often cannot be easily located and may continue to generate a continuous plume of
dissolved halocarbons (National Research Council, 1994). Although these plumes can often be controlled by pump-
and-treat, the systems require extensive maintenance and continual energy input. PRBs are also designed for plume
control, but with significant differences from pump-and-treat systems. PRBs arc in situ systems, are intended to
operate in a totally passive manner, do not routinely bring the contaminant to the surface, and should operate for
years with minimal, if any, maintenance.
Although simple in concept, there is a great breadth of science and technology involved in the selection of reactive
materials for different contaminants and in the design, installation, and monitoring of these emplacements in the
subsurface. The purpose of this document is to provide an introduction and a guide to the science and technology of
PRBs.
2.1 Defining Permeable Reactive Subsurface Barriers
Environmental scientists are generally familiar with the concept of barriers for restricting the movement of
contaminant plumes in soil and ground water. Such barriers are typically constructed of highly impermeable
emplacements of materials such as grouts, slurries, or sheet pilings to form a subsurface "wall." The goal of such
barriers is to minimize the possibility that a contaminant plume can move toward and endanger sensitive receptors,
such as drinking water wells, or discharge into surface waters. Rather than serving to constrain plume migration,
PRBs are designed as conduits for the contaminated ground-water flow. As contaminated water passes through the
reactive zone of the PRB, the contaminants are either immobilized or chemically transformed to a more desirable
(e.g., less toxic, more readily biodegradable, etc.) state. Therefore, a PRB is a barrier to contaminants, but not to
ground-water flow. A permeable reactive subsurface barrier is defined as:
an emplacement of reactive materials in the subsurface designed to intercept a contaminant plume, provide
a flow path through the reactive media, and transform the contaminant(s) into environmentally acceptable
forms to attain remediation concentration goals downgradient of the harrier (Powell and Powell, 1998;
'Powell and Puls, 1997a).
-------
In addition to PRBs, research is being conducted on a similar class of subsurface remediation technologies which,
in tliis document, will be referred to as reaction zone formation (RZF) technologies. These are somewhat different
from PRBs because they do not necessarily emplace solid phase reactive media in the subsurface. One approach, for
example, is the injection and withdrawal of a sodium dithionite solution into an aquifer matrix (Fruchter et al.,
1996). The dithionite reduces Fe3^ that is naturally available on the mineral matrix surface to Fe2+, thus increasing
the reducing capacity of the aquifer material itself. This has been shown to be an effective means of removing
chromate from ground water by reductive precipitation as the Cr(III) hydroxide, in much the same manner as was
previously discussed for its precipitation by Fe(0). Another RZF technology is the use of in situ electrodes, not to
directly move or remediate the contaminants, but rather to supply hydrogen as an electron donor to the subsurface
microbes. The microbes can then potentially remediate the contaminant, e.g., a chlorinated hydrocarbon, by using
it as the electron acceptor. This would reductively dehalogenate the compound, as was described above for the
electron transfer between Fe(0) and PCE. Although these RZF technologies are not the subject of this document, it
is important to realize that such processes are being investigated and implemented in the field.
2.2 Research Support and Application of the PMB Technology
The U.S. Environmental Protection Agency has supported the development of this innovative in-situ technology
through active collaboration on research involving the National Risk Management Research Laboratory (NRMRL)
and the National Exposure Research Laboratory (NERL) of USEPA's Office of Research and Development (ORD),
through the Remediation Technologies Development Forum (RTDF) Permeable Barriers Action Team, and from
support provided by USEPA's Technology Innovation Office (TIO). In addition, support has been provided from
several regional offices where sites are testing the technology at pilot-scale. Both DOD and DOE have also actively
supported research on PRO technology, financially supporting both laboratory and field trials.
As with any emerging technology, the field installation of this technology is limited relative to the research and
laboratory testing that has been performed and the overall potential of the technology for contaminant remediation.
Most full-scale PRBs are currently being used for the treatment of plumes of chlorinated hydrocarbons and
chromate. using granular Fe as the reactant. Laboratory studies have shown, however, that many other contaminants
can also be treated using PRB technology with the selection of the proper reactant (Appendix A). As of this writing
(March 1998), at least thirteen full-scale reactive barriers have been installed in the field (Table 1).
Table 1. Full-Scale in situ Remediation Projects
Industrial facility
Mine Site,
Industrial facility
Industrial facility
Industrial facility
USCG facility
Government facility
Industrial facility
Industrial facility
Industrial facility
Industrial facility
Superfund site
U. S. DOE facility
Sunnyvale, California
Sudbury, Ontario
Mountainview, California
Belfast, Northern Ireland
Coffeyville, Kansas
Elizabeth City, North
Carolina
Lake wood, Colorado
South Carolina
Colorado
Oregon
Upstate New York
New Jersey
Kansas City, Missouri
gate and slurry wall
continuous wall
gate and HDPE liner
in situ reactive vessel and slurry
wall
gate and slurry wall
continuous reactive wall
sheet pile funnels and 4 gates
continuous iron/sand reactive
wall
gate and slurry funnel
gates (2) and slurry funnel
continuous wall
continuous wall -
hydrofracturing
continuous wall
December 1994
August 1995
December 1995
December 1995
January 1996
June 1996
October 1996
October 1997
November 1 997
November 1 997
December 1997
Apnl 1998
Apnl 1998
-------
2.3 Regulatory Acceptance of PMB Technology
The USEPA recognizes this technology as having the potential to effectively remediate subsurface contamination
at many types of sites with significant cost savings compared to more traditional approaches (e.g., pump-and-treat).
The USEPA is actively involved in the evaluation and monitoring of this new technology to answer questions and
provide guidance to various stakeholder groups.
The Interstate Technology and Regulator}? Cooperation (ITRC) Workgroup (Permeable Barrier Wall Subgroup) is
also actively involved in defining the regulatory implications associated with the installation of permeable reactive
barriers in the subsurface and in providing guidance on regulatory issues where possible.
The first full-scale commercial PRB was approved for use in the State of California by the San Francisco Regional
Water Quality Control Board (RWQCB) in 1994. Since that time, ten other full-scale systems have been installed
in the U.S.. along with one in Sudbury, Canada, and one in Belfast, Northern Ireland. Table 2 lists some of the states
where permeable barriers have been approved for installation, either at pilot-scale or full-scale. The regulator}'
approach to the technology at several sites has been to treat the installations as '"at risk" remedies. That is, the owner
would be required to implement a more conventional remedy if the permeable barrier failed to meet performance
criteria. In some cases, the lack of existing groundwater use. particularly in industrial areas, facilitated implemen-
tation of PRBs. At sites in California and Colorado, there was a consensus among the regulators, site owners and
site consultants that existing pump and treat systems were not achieving the desired level of groundwater
remediation, making PRBs a favorable alternative solution.
From a federal perspective, one of the more significant advances for PRB technology occurred when a "'chemical
treatment wall" was identified in June 1995 as the preferred alternative in the Record of Decision (ROD) at a
Superfund site (the Somersworth Municipal Landfill in Somersworth, New Hampshire). This decision, coupled
with the 1996 directive from the U.S. EPA Office of Solid Waste and Emergency Response (OSWER) to evaluate
alternatives to pump and treat remedies has encouraged consideration of the technology at other Superfund sites.
Active research programs at several U.S. EPA laboratories have also led to greater acceptance of the technology at
the federal level. A pilot-scale PRB for removal of VOCs in groundwater was evaluated in 1995 under the
Superfund Innovative Technologies Evaluation (SITE) program. The Technology Evaluation Report for this project
Table 2. Some States that have Approved the Installation of Iron PRBs.
State
California
Colorado
Delaware
Florida
Kansas
Illinois
Missouri
New Hampshire
New Jersey
New York
North Carolina
Oregon
South Carolina
Totals
Full-Scale Installations
2
2
0
0
1
0
1
0
1
1
1
1
1
11
Pilot-Scale Installations
2
1
1
3
0
1
0
1
0
1
1
0
1
11
-------
remains to be published at the time of this writing (March 1998). The inclusion of PRBs in the SITE program and
Vendor Information System for Innovative Treatment Technologies (VISITT) has also increased the awareness of
the technology. Recent developments in interstate environmental technology verification programs may further
expedite regulatory acceptance.
In summary, regulatory acceptance of PRBs is expected to increase as the number of site installations increases
and more long-term performance data becomes available from existing installations.
-------
3.0 Treatable Contaminants, Mechanisms
3.1 Desirable Characteristics of Reactive Media
Reactive media used in permeable barriers should be compatible with the subsurface environment. That is. the
media should cause no adverse chemical reactions or byproducts when reacting with constituents in the contaminant
plume, and should not act as a possible source of contaminants itself. This requires that the material be well
understood and characterized. To keep PRB costs to a minimum, the material should persist over long periods of
time. i.e.. it should not be readily soluble or depleted in reactivity, and the material should be readily available at a
low to moderate cost. The material selected should minimize constraints on ground-water flow by not having
excessively small particle size and it should not consist of a wide range of particle sizes that might result in blocked
intergranular spaces (i.e., it should preferably be unimodal with respect to grain size). Worker safety, with regard to
handling the material, should also be considered.
3.2 Treatable Contaminants
Table 3 lists contaminants that have been shown to be successfully treated by zero-valent iron or other media.
Tables 4 and 5 list some contaminants that have been shown to be not affected by zero-valent iron or have not yet
been fully evaluated, respectively.
3.2.1 Halogenated Organic Compounds andiron
Considerable research during the past several years has focused on the degradation of chlorinated solvents, such as
TCE and PCE, by reactions at the surfaces of Fe(0). Although met with initial skepticism, the degradation process
is now widely accepted as abiotic reductive dehalogcnation, involving corrosion of the Fc(0) by the chlorinated
hydrocarbon.
Iron corrosion processes in aqueous systems have been studied extensively. Until recently, the fate of corrosion
processes in dilute aqueous concentrations of chlorinated solvents acting as the oxidizing agents have not been
investigated. The net reductive dcchlorination reaction promoted by Fc(0) (Equation 3) may be viewed as the sum
of anodic and cathodic reactions occurring at the iron metal surface (Equations 1 and 2, respectively), resulting in
hydrocarbon products if the dechlorination proceeds to completion.
Fe° —» Fe2^ + 2e~ Anodic Reaction (1)
RC1 + 2e- + FT -» RH + Cl' Cathodic Reaction (2)
Fe° + RC1 + H+ -> Fe2+ + RH + Cl' Net Reaction (3)
Under aerobic conditions, dissolved oxygen is usually the preferred electron acceptor and can compete with the
chlorinated hydrocarbon as the favored oxidant (Equation 4). Indeed, chlorinated hydrocarbons such as PCE and
carbon tetrachloride have oxidizing potentials very similar to that of O2 (Archer and Barter, 1978). When sufficient
oxygen is present, the Fe2+ generated in Equation 4 further oxidizes to Fe3" (Equation 5) and can precipitate as ferric
hydroxide or (oxy)hydroxides (Equation 6) at the elevated pH typical of corroding Fe systems. Corrosion of the iron
can generate large amounts of iron oxides and (oxy)hydroxide precipitates that can exert significant additional
chemical and physical effects within the reactive system (Powell et al., 1994; Powell et al, 1995a). The rapid
consumption of dissolved oxygen at the entrance to an iron system (column or barrier) has been shown to result in
these precipitates that might impact a system's hydraulic performance at its upgradient interface (MacKenzie et al.,
1995; Mackenzie et al., 1997).
2Fe° + O2 + 2H2O -> 2Fe2+ + 4OIT (4)
4Fe2+ + 4H+ + O2 -» 4Fe3+ + 2H2O (5)
Fe3+ + 3OH-^Fe(OH)3(s) (6)
Fe° + 2H2O -» Fe21 + H2 + 2OFf (7)
Fe2^ + 2OH- -> Fe(OH)2(s) (8)
Anaerobic corrosion of iron by water (Equation 7) proceeds slowly. Both reactions 4 and 7 result in an increased
pH in weakly buffered systems, yielding ferric (oxy)hydroxides in aerobic systems (Equation 6) and ferrous
(oxy)hydroxides in anaerobic systems (Equations). The aqueous corrosion of iron is mediated by the layer of
oxides, hydroxides and oxyhydroxides that are present at the iron-water interface. The formation of these
precipitates might further occlude the iron surface and affect its reduction-oxidation properties. However, this
-------
Table 3. Contaminants Treatable Bv Reactive Materials in PRBs.
Organic
Compounds
Methanes
Ethanes
Ethenes
Propanes
Aromatics
Other
tetrachloromethane
tr ichloromethane
dichloromethane
hexachloroethane
1,1,1 -trichloroethane
1 ,1 ,2-trichloroethane
1 ,1 -dichloroethane
tetrachloroethene
trichloroethene
cis-1 ,2-dichloroethene
trans- 1 ,2-dichloroethene
1 ,1 -dichloroethene
vinyl chloride
1 ,2,3-trichloropropane
1 ,2-dichloropropane
benzene
toluene
ethylbenzene
hexachlorobutadiene
1 ,2-dibromoethane
freon 113
N- nitrosodimethylamine
Inorganic
Compounds
Trace Metals
Anion
Contaminants
Chromium
Nickel
Lead
Uranium
Technetium
Iron
Manganese
Selenium
Copper
Cobalt
Cadmium
Zinc
Sulphate
Nitrate
Phosphate
Arsenic
Table 4. Contaminants Presently not Treatable by Fe(0)
Organic Compounds
Inorganic Compounds
dichloromethane
1,2-dichloroethane
chloroethane
chloromethane
chloride
perchlorate
-------
Table 5. Contaminants with Unknown Treatabilitv
Organic Compounds
chlorobenzenes
chlorophenols
certain pesticides
PCBs
Inorganic Compounds
mercury
passive coating appears to be converted to magnetite (Odziemkowski et al. 1998), which is non-passivating, and
seems to allow sufficient contaminant degradation rates that can be sustained over years of operation in the ground.
Zero-valent iron is a mild reductant. and consequently the dehalogenation rates vary for the various chlorinated
solvents of environmental interest. A number of studies have shown that the primary determinant of degradation
rate is the specific surface area, or the surface area of iron per unit volume of pore water (Matheson and Tratnyek,
1994; Sivavec et al., 1995). Degradation rates are typically pseudo-first-order with respect to the halogenated
hydrocarbon, with the rate constant relatively insensitive to the initial hydrocarbon concentration. Studies of
published degradation rate data for individual halogenated hydrocarbons show that transformation rates are
proportional to iron surface area concentration (Johnson et al., 1996) and that observed rate constants can be
normalized to iron surface area to vield a specific rate constant, or k , for the halocarbon.
1 ' sa'
The reaction pathways by which Fe(0) reduces halogenated hydrocarbons have been determined for a few major
classes of chlorinated hydrocarbons. Such information is significant to the optimal design of a permeable reactive
barrier, as incomplete dechlorination of a highly chlorinated ethene, for example, could produce an intermediate
product, such as VC, which is more hazardous and more persistent than the parent compounds. Even very low
concentrations of undesirable by-products in the reactive barrier effluent must be avoided.
Typically, permeable reactive barriers are designed to provide adequate residence time in the treatment zone for
the degradation of the parent compound and all toxic intermediate products that are generated. At sites where the
ground-water contamination includes a mixture of chlorinated hydrocarbons, the design of the PRB is usually
determined by the least reactive constituent. Perhalogenated hydrocarbons tend to be reduced more rapidly than
their less halogenated congeners, and dechlorination is more rapid at saturated carbon centers (for example, carbon
tetrachloride) than at unsaturated carbons (for example, TCE or VC).
Excellent product mass balances have been determined for the transformation of several chlorinated ethenes (Orth
and Gillham, 1996) and methanes. These mass balance studies are best accomplished in a flow-through column
design, as sorption effects can be minimized. TCE. a common ground-water contaminant, is largely converted to
ethene and ethane. Generally less than 5%-10% of the initial TCE appears as chlorinated degradation products,
including the three dichloroethene products (cDCE is dominant) and VC. As shown in Equation 9, zero-valent iron
reduces TCE via two interconnected degradation pathways: (A) sequential hydrogenolysis and, (B) reductive 6-
elimination (Roberts ct al., 1996). The intermediate products, cDCE and VC. are produced in the sequential
hydrogenolysis pathway, and are slower to degrade than is TCE itself. The chloroacetylene intermediate produced
via the 6-elimination pathway, by contrast, is a very short-lived intermediate and is very rapidly reduced to ethene.
The 6-elimination pathway accounts for the rapid conversion of TCE to ethene and ethane, with relatively minor
intermediate product formation.
10
-------
H CI
B
4r S*.
2C1-
Chloroacetylene H—C=C—CI
2e- 4- H+
H
H * H
)==/ Vinyl chloride
2e-
H a
2e- + H+
•ci-
Acetylene H—C=C—H
2e- + 2H+
H
ci-
H
H H
2e- + 2H+
f
H H H
>v^<
H H H
Ethenc
Ethane
(9)
The degradation pathway for chlorinated methanes is a much simpler one than that for chlorinated ethenes, as the
6-elimination route is not available to the methane family. Carbon tetrachloride has been shown to undergo
sequential hydrogenolysis to chloroform and dichloromethane (Equation 10) (Matheson and Tratnyek, 1994). The
formation of chloroform accounted for about 70% of the carbon tetrachloride lost. Methylene chloride appeared,
after carbon tetrachloride decreased to the detection limit, and typically accounted for about 50% of the chloroform
lost. No further reaction of dichloromethane has been detected in an unamended zero-valent iron system.
CC1
CHd3+cr
CH2CI2.
2CF
(10)
3.2.2 Redox-sensitive Inorganic Anions and Iron
Negatively charged anions and oxyanions, the dissociation products of Lewis acids formed by the hydrolysis of
cations, are important ground-water contaminants. Elements which occur as anions or oxyanions under natural
ground-water conditions include arsenic, selenium, chromium, technetium and antimony. In addition, the dissolved
nutrient species nitrate and phosphate occur as anions, as does sulfate. Because of the negative charge, anionic
species are not attracted to negatively charged mineral surfaces, which are usually predominant in aquifers under
neutral pH conditions. This limited tendency for adsorption, and the high solubility of minerals containing
oxyanions, result in the potential persistence of high concentrations of these elements in aquifers. Treatability
studies, pilot-scale field trials and full-scale demonstration projects have been conducted for a series of anionic
inorganic contaminants, including SO42, NO3\ and those containing Cr(Vl), Se(VI), As(III), As(V), Tc(VII).
Treatment of chromate. CrO42~ which contains Cr(VI), has been the most extensively tested and demonstrated of
the anionic inorganic contaminants. Chromium commonly occurs in two oxidation states, Cr(lll) and Cr(Vl), in the
environment. Trivalent Cr(IlI) is relatively non-toxic and a micronutrient. It forms sparingly soluble hydroxide
precipitates under conditions prevalent in most surficial aquifers and is also readily adsorbed by some minerals.
Cr(VI) is a known carcinogen, which forms relatively soluble precipitates, resulting in the persistence of relatively
11
-------
high concentrations of dissolved Cr(VI) in affected aquifers (Palmer and Puls, 1994; Puls et al., 1995). Cr (VI) is
usually speciated as chromate, CrO42~, under typical ground-water pH and Eh conditions. This results in a high
degree of mobility for the Cr(VI) because anions are not readily adsorbed to the predominantly negatively-charged
aquifer materials. Treatments to remove Cr(VI) from ground water typically use reduction to Cr(III) and
precipitation of insoluble Cr(III) hydroxide precipitates.
A range of solid-phases containing reduced iron have been observed to promote the reduction and precipitation of
Cr(VI), including elemental iron (Gould, 1982; Bowers et al., 1986; Bostick et al., 1990; Blowes and Ptacek, 1992;
Powell et al., 1994; Powell et al., 1995a), iron-bearing oxyhydroxides (Eary and Rai, 1989) and iron-bearing
aluminosilicate minerals (Eary and Rai. 1989; Kent et al., 1994). Iron bearing reductants have been evaluated as
potential candidate materials to promote Cr(VI) reduction in reactive barriers. Comparison of reaction rates
observed using Fe(0), pyrite (FeS2), and siderite (FeCO3), indicate that more rapid Cr(VI) reduction is attained
using Fe(0) (Blowes and Ptacek, 1992; Blowes et al., 1997). It has also been shown that the rates of Cr(VI)
reduction are dependent on the Fe(0) itself (i.e., how it was refined and manufactured, its level of impurities, etc.)
and whether certain aluminosilicate-containing aquifer materials are present and mixed with the iron (Powell et al.,
1995a; Powell et al., 1995b; Powell and Puls, 1997).
The overall reactions for the reduction of Cr(VI) by Fe(0) and the subsequent precipitation of Cr(III) and Fe(IIT)
oxyhydroxides are:
CrO42- + Fe° + 8H+ -> Fe3^ + Cr3^ + 4H,O (11)
(l-x)Fe3+ + (x)Cr3^ + 2H2O -» Fe(1.x)CrxOOH(Bj + 3FT (12)
Tlie extent and rate of Cr(VI) removal by elemental iron has been evaluated in laboratory batch tests and stirred
batch reactors (Blowes and 'Ptacek, 1992; Powell et al., 1995a), column tests (Blowes and Ptacek. 1992; Powell et
al., 1994; Blowes et al., 1997a), pilot-scale field trials (Puls et al., 1995) and a full-scale field demonstration
(Blowes et al., 1997b). The results of these experiments indicate that the rate of Cr(VI) reduction and precipitation
is sufficient for use in ground-water remediation systems. The removal of Cr(VI) from solution is accompanied by
a sharp decrease in the Eh. from initially oxidized conditions (Eh > 100 mV) to very reduced conditions
(Eh < -300 mV), and a sharp increase in the pH from initially near neutral conditions (6.5 < pH < 8.5) to more basic
conditions (pH > 9.5). Details of these reactions and the mechanisms of CrO_,2~ reduction by Fe(0) have been
described (Powell et al., 1995a; Powell and Puls, 1997a).
Scanning probe microscopy of the reaction precipitates has indicated a one-to-one correspondence of the locations
of Cr(III) and Fe(III) hydroxide phases on the surface of reacted iron filings, indicating coprecipitation and the
likelihood of the formation of a solid solution phase of the general formula (Crx, Fe1_x)(OH)3 (Powell et al., 1994;
Powell et al., 1 995a). More detailed mineralogical studies indicate that the dominant reaction product derived from
the laboratory experiments is a mixed Fe-Cr oxyhydroxide phase with the mineral structure of goethite (FeOOH) as
shown in Equation 12. The distribution of Cr throughout the structure of this phase is variable, suggesting
incorporation through solid-solution substitution (Blowes et al., 1997b). X-ray Photoelectron Spectroscopy (XPS)
indicates that the Cr within the precipitate is exclusively in the Cr(III) oxidation state, and that Fe present in the
precipitate is in the Fe(III) oxidation state (Pratt et al ., 1 997). Auger Electron Spectroscopy indicates that the Fe(III)
and Cr(III) of the precipitate surface occur in the hematite (Fe2O3) structure, that is distinct from the goethite
structure of the bulk phase, suggesting Cr(III) exolution and formation of a chromite-like (Cr2O3) phase at the
precipitate surface (Pratt et al., 1997).
Reduction and precipitation of other anion-forming elements by elemental iron have also been investigated at the
laboratory scale. Laboratory experiments conducted to evaluate the treatability of Tc indicate rapid reduction using
elemental iron (Bostick et al., 1990; Del Cul et al., 1993; Clausen et al., 1995). Reduction of Se(VI) by elemental
iron, and precipitation of elemental selenium (Se°) is favored thermodynamically. McRae et al. (1997) observed
removal of 1000 jjg/L of Se(VI) from solution in laboratory batch experiments using elemental iron, and proposed
the reaction:
HSeO,,- + 3Fe° + 7HT -» 3Fe2^ + Se°(s) + 4H,O (13)
Although removal of Se(VI) was observed, the secondary reaction product was not isolated or identified.
Arsenic commonly occurs as a dissolved species in two oxidation states. As(V) and As(III), and less commonly in
other oxidation states, including As(0), As(-T) and As(-II). The As(V) oxidation state occurs as H3AsO4 and its
dissociation products. The As(III) oxidation state occurs as H3AsO3 and its dissociation products. Although' As(III),
which has greater health effects, is considered to be potentially more mobile in ground-water systems, As( V) is also
12
-------
observed to be mobile in natural systems. In addition, the rate of arsenic transformation between the As(III) and
As(V) oxidation states occurs rapidly in response to changing environmental conditions. Arsenic reduction to the
As(0) oxidation state and precipitation of native arsenic, or incorporation into a secondary arsenic sulfide, has been
proposed as a potential treatment technique (McRae et al., 1997). In batch tests conducted using elemental iron,
McRae et al. (1997) observed rapid removal of As(V) from concentrations of 1000 jjg/L to < 3 jjg/L over a
two-hour period. Similar experiments using As(III) and mixtures of As(III) and As(V) indicated equally rapid
removal rates. The mechanisms resulting in arsenic attenuation, and the potential duration of treatment are being
evaluated by continuing experiments (McRae et al., 1997). The applicability of Fe(0) in PRBs for permanent As
removal and retention requires further investigation.
Reduction of NO,~ by Fe(0) has been observed to proceed rapidly, resulting in production of NO,,", and
subsequently ammonium ion (Cheng et al., 1997; Rahman and Agrawal, 1997; Till et al., 1998). The proposed
pathway for the overall reaction is:
4Fe(0) + NO3- + 1OH+ -> 4Fe2+ + NH + + 3H2O (14)
During laboratory studies, NO," reduction was observed to proceed through NO2~ to NH4+ formation. The rate of
reduction was observed to be first order with respect to the Fe° substrate. In addition the rate was observed to
decline with increasing duration, particularly in the presence of chloride (Rahman and Agrawal, 1997). However,
NH4T formation is not desirable since its MCL is lower than that of NO3". Less NH4" is formed when Fe° is used in
conjunction with microbial consortia (Till et al., 1998). [See section on Biologically Mediated Reduction and
Removal]
3.2.3 Biologically Mediated Reduction and Removal ofAnions
Biological processes affect the cycling of numerous elements, including nitrogen, sulfur, iron and manganese.
Treatment strategics employing these biologically mediated reactions have been proposed for direct treatment of
nitrate (Robertson and Cherry, 1995; Vogan, 1993; Till et al., 1998), sulfate (Blowes et al., 1994; Waybrant et al.,
1995). and for indirect removal of other anions through precipitation as sulfide phases.
A denitrification system for removal of nitrate from ground water affected by discharge from site wastewater
disposal systems has been developed (Robertson and Cherry, 1995). This system intercepts a plume of nitrate-
bearing ground water with a reactive barrier containing solid phase organic carbon. In the presence of organic
carbon, under anaerobic conditions maintained below a water cover in the subsurface, reduction of NO3" to N,, gas
is thermodynamically favored. The proposed reaction is:
5CH,O(B) + 4NO3- -> 2N2 + 5HCO; + 2H2O + ff- (15)
CH2O represents a simple form of organic carbon which is catalyzed by bacteria of the Pseudomonas group. These
bacteria use NO3~ as an electron acceptor in the oxidation of organic carbon. In a laboratory study a scries of
inexpensive organic carbon sources were evaluated to assess their potential use in reactive barriers for treatment of
NO "(Vogan, 1993). Results to date indicate that readily available materials, including sawdust and wood waste, are
suitable materials for a reactive barrier system. Robertson and Cherry (1995) evaluated permeable reactive barriers
for treating NO3~ at several domestic and institutional septic systems. The results of these studies indicate that
sufficient denitrification occurs rapidly and reduces NO3" from concentrations typically observed in the effluent of
onsite wastewater disposal systems (5 - 90 mg/L NO3" - N) to below the World Health Organization drinking water
standard (10 mg/L NO3~ - N). Reactive barriers using a variety of design configurations have been implemented. No
evidence of assimilatory NO3~ reduction and NH4+ formation has been observed.
A combined system of Fe(0) and the denitrifying bacteria Paracoccus denitrifwans to reduce NO3~ has been
investigated (Till et al., 1998). The hydrogen produced by the corrosion of the iron was used as the electron donor
for nitrate-based respiration. Using steel wool as the Fe° they found that the combination increased nitrate removal
rates, relative to steel wool alone, and transformed a greater portion of the added NO3~ to innocuous gases rather
than to NH/ .
Biologically mediated reduction of sulfate to sulfide. accompanied by the formation of metal sulfides occurs
through the reaction sequence
>H,S, . + 2CO,, . + 2H,O (16)
2 (aq) 2(aq) 2 \ /
+ 2ff- (17)
,, x /
where CH2O represents organic carbon and Me24" represents a divalent metal cation in solution. Biologically
mediated sulfate reduction has been proposed to treat metal cations derived from mine sites in wetlands (Mclntire
13
-------
et al., 1990; Kleinmann et ah. 1991; Eger and Wagner, 1995). bioreactors (Dvorak et ah, 1992) and permeable
reactive barriers (Blowes et ah, 1995; Waybrant et ah, 1995; Benner et ah, 1997; Blowes et ah, 1997b). Although
these systems are designed to promote the removal of dissolved metals, metal removal is accompanied by removal
of sulfate. In laboratory studies, (Waybrant et ah, 1995, 1997a) sulfate removal was observed at rates of 0.14 to
4.23 mg L"1 day1 g"1 of media.
3.2.4 Adsorption and Precipitation of Inorganic Anions
Inorganic anions which are not susceptible to reductive or oxidative processes must be removed from solution by
other means. These anions, as well as rcdox-scnsitivc species, may be removed by precipitation, adsorption or
coprecipitation on mineral surfaces.
Article and Fuller (1979) proposed the use of barriers containing crashed agricultural limestone to neutralize the
pH of landfill leachate and precipitate Cr. Laboratory tests indicated successful removal of the cation form, Cr3+,
and less successful treatment of the anionic form of Cr(VI), CrO42~. Thomson et al. (1991) and Longmirc ct al.
(1991) described the potential use of sphagnum peat, crushed limestone, and hydrated lime, to remove U, As, Mo,
and Se in laboratory batch tests. Morrison and Spangler (1992, 1993) proposed the use of a series of industrial
byproducts as reactants to remove U, As and Mo in precipitation or sorption barriers. Phases evaluated included
bases, rcductants, and sorbcnts, including hydrated lime, ferrous sulfate, hydrous ferric oxides, peat, and phosphate-
bearing phases to remove both anionic and cationic species from solution. The system described by Morrison and
Spangler (1992; 1993) is potentially suitable for a wide range of anionic contaminants.
Phosphate (PO43~) derived from anthropogenic sources is a limiting nutrient in many aquatic ecosystems. Release
of excess phosphate results in accelerated biological activity, and ultimately in cutrophication of the aquatic-
ecosystem. In many regions, phosphate released from onsite wastewater disposal systems, such as septic systems or
sewage lagoons, results in excess phosphorus. Baker et al. (1996. 1997) used amixture of iron oxide, calcium oxide
and limestone to promote adsorption and coprecipitation of phosphorus. Phosphate was removed rapidly in batch
experiments. In column experiments extending over 3 years and 1200 pore volumes, influent PO 3~ . at a
concentration of 4 mg/L PO43" - P, was removed to <0.3 mg/L PO,t3~ - P, indicating more man 90% removal of the
phosphate over a prolonged period. Field-scale demonstration at an institutional septic system resulted in removal
of phosphate from influent concentrations of 1 to 2 mg/L PO43~ - P to less than 0.1 mg/L PO43"- P. Test systems,
established at onsite wastewater treatment systems, indicate potential for removal of PO 3" from wastewater under
continuous flow conditions. The effluent concentrations achieved by this technique are sufficiently low to prevent
eutrophication of surface water flow systems in which PO43~ is the limiting nutrient (Baker et ah. 1996).
The mixture used by Baker et al. (1996) was evaluated for the potential treatment of arsenic (McRae, 1997). Batch
experiments indicated a decrease in As(III) from 1000 jig/L to < 3 ug/L in less than 2 hours. Experiments showed
decreases in As(V) concentrations, and in mixtures of As(III) and As(V) from 1000 ug/L to <3 ug/L over a similar
period. A column test conducted using this material indicated decreases in mixed As(III) and As(V) concentrations
from a total of 1000 ug/L to <3 ug/L for more than 120 pore volumes (McRae et ah, 1997). Research on this system
is continuing.
3.2.5 Reduction of Inorganic Cations
Positively charged inorganic cations, including the metals Cd, Co, Cu, Mn, Ni, Pb, Zn, and complex cations such
as UO,,2' are important ground-water contaminants. High concentrations of these metals are associated with
industrial wastes, mine wastes and nuclear waste disposal sites. Ground water within mine waste piles, and leachate
derived from mine wastes commonly contains high concentrations of dissolved metals (Dubrovsky et ah, 1984;
Morin et ah. 1988). Leachate derived from waste disposal areas, containing high concentrations of dissolved
uranium and technetium, has been reported from the Oak Ridge National Laboratory (Olsen et ah. 1986).
Reduction and precipitation of sparingly soluble solids has been proposed for inorganic cations, as well as anions.
Treatment of cation -form ing electroactive metals, including mercury, uranium, copper and technetium by reduction
with elemental iron and coprecipitation within secondary precipitates has been investigated at the laboratory scale.
Laboratory batch experiments indicate rapid removal of U(VI) and Tc(VII) from solutions in contact with elemental
iron (Bostick et ah, 1.990; Liang et ah, 1.996). The reaction between U(VI) and Fe(0) can be expressed as
(18)
j x 7
where UO,s) is an amorphous or crystalline uranium oxide precipitate. Strongly reducing conditions must be
attained for uranium reduction and precipitation to proceed. In addition, if oxidizing conditions recur, the reduced
uranium may become reoxidized and remobilized. Other metal cations potentially treatable by reduction with Fe(0)
include Cu and Hg.
14
-------
3.2,6 Biologically Mediated Reduction and Precipitation of Cations
Biologically mediated reduction reactions can promote the attenuation of inorganic cations. These reactions
include direct reduction of the cation and precipitation of a sparingly soluble solid, and indirect precipitation
resulting from the oxidation or reduction of an inorganic anion. An example of direct reduction is the reduction of
U(VI) through a bacterially mediated reaction. Bacteria have been isolated that are capable of reducing U(VI) to
U(IV), and sustaining metabolism based on this reaction (Lovely et al., 1991; Lovely and Phillips, 1992a; Lovely
and Phillips, 1992b). It was found that U(VI) reduction by organic matter or H, was insignificant without microbes
present. The proposed reaction for catalysis of U(VI) reduction by acetate, using the microbe G. malallireducens is:
CH3COO + 4UO,(CO3),2- + 4H2O -> 4UO2 + 10HCCX; + H+ (19)
It was also proposed that the A. puirefaciens could oxidize hydrogen to reduce U(VI) by the reaction:
H2 + UO,(CO3)22- -» UO, + 2HCCX; (20)
Indirect precipitation of inorganic cations results from the reduction of an anion-forming species, usually sulfate.
Sulfatc reduction generates hydrogen sulfide, which combines with metals to form relatively insoluble metal sulfidc
precipitates, described in Equations 16 and 17. Laboratory studies indicate that many metals are treatable using this
approach, including Ag, Cd, Co, Cu, Fe, Ni, Pb, and Zn (Waybrant et al., 1995; Way brant et al., 1997a). Column
experiments, conducted using a range of organic substrates demonstrated the potential to remove a range of
dissolved metals at ground-water velocities similar to those observed at sites of ground-water contamination. A
field-scale reactive barrier for the treatment of acid mine drainage and removal of dissolved Ni was installed in
1995, at the Nickel Rim mine site near Sudbury, Ontario. It was composed of municipal compost, leaf compost and
wood chips. Monitoring of the reactive barrier indicates continued removal of the acid generating capacity of the
ground water flowing through the PRB and decreases in dissolved Ni concentrations from up to 10 mg/L to
<0.1 mg/L within the PRB. Monitoring is continuing at this site.
3.3 Enhancing Iron Reaction
The range of reactivity of halogenated hydrocarbons and other contaminants with zero-valent iron illustrates the
value of achieving faster degradation rates and more complete degradation of less reactive products. Research into
increasing these rates has investigated metals that are stronger reducing agents than iron, the addition of
aluminosilicate minerals, and metal couples. Metals that are stronger reducing agents than iron have higher
oxidation potentials; i.e., yield their electrons more readily to an oxidized substance, hence they corrode faster.
Aluminosilicate mineral addition and metal couples result in increased corrosion rates for the iron that is present.
Many aluminosilicate minerals seem to enhance corrosion of Fe(0) by buffering reaction pH to lower values
(pH=7-8 rather than > 9). This buffering comes from dissolution of the aluminosilicate minerals (Powell et al.,
1995a; Powell and Puls. 1997b) with the generation of hydrogen as protons. These protons can coordinate with the
cathodic regions of the iron surfaces, accepting electrons and thereby increasing corrosion rates. In laboratory batch
tests using Fe(0) with an aluminosilicate-containing aquifer material from Elizabeth City, North Carolina, the half-
life of Cr(VI) in solution was decreased by an order of magnitude relative to a system with Fe(0) and silica sand.
The Cr(VI) half-life was about two orders of magnitude shorter than tests where only Fe(0) was present (Powell and
Puls. 1997). The protons (H+) or their reduced species, surface coordinated monoatomic hydrogen atoms (H*), may
also serve as replacement moieties for the halogen atoms on the hydrocarbons that are being reductively
dehalogenated at the cathodic surfaces (Powell and Puls, 1997b).
Bimetallic systems (metal couples) prepared by plating a second metal onto a zero-valent iron surface, including
Fe/Cu, Fe/Ni"(Sivavec et al., 1997; Gillham et al., 1997) and Fe/Pd (Muftikian et al., 1995), have been shown to
accelerate solvent degradation rates relative to untreated iron metal. Palladized iron has been shown to be effective
in dechlorinating halogenated aromatic compounds such as polychlorinated biphenyls (PCBs) in addition to
chlorinated aliphatic compounds (Grittini et al., 1995). The rate enhancement observed in bimetallic systems may
be attributed to corrosion-inducing effects promoted by the second, higher reduction potential metal and possibly
some catalytic effects. However, some investigators have found the enhanced reactivity of these systems to
diminish relatively quickly, whereas others have found no apparent loss of reactivity (Gillham et al, 1997). These
differences may be related to ground-water chemistry or the method used for plating the iron, but further
investigation is needed. It is important to note that zero-valent iron systems have not shown similar losses in
reactivity in long-term laboratory, pilot and field investigations.
15
-------
4.0 Remediation Feasibility, Laboratory Treatability, PRB
Most, if not all, PRB installations have been designed and implemented based on the results of laboratory batch
and column studies used to test reactant materials and the kinetics of contaminant removal. These data are used in
combination with site-specific information (such as ground-water velocity, contaminant type and concentration,
and the total mass flux of the contaminant requiring treatment). The following sections present information on these
laboratory studies and how the information is used in the development of a PRB as a remedial solution. However,
there will often be a need to develop some approximate design configurations and cost estimations prior to carrying
out extensive laboratory studies, to determine whether PRB technology is more suitable for a site than other options.
Appendix C provides some scoping calculations that have been proposed to allow such cost estimation. These can
also provide some insight into the PRB emplacement methods and designs that could provide iron in thicknesses
sufficient to meet the regulator}' requirements of the site.
4.1 Laboratory Treatability Studies
The need for laboratory treatability studies for PRB design is primarily dependent on the contaminants present,
their concentration and the geochemical conditions at the site. For contaminants where an extensive database of
removal rates exists (TCE or chromate in granular iron, for example), these rate data can be used for design
purposes and the treatability tests can potentially be omitted. When there arc mixtures of contaminants, the
geochemical conditions are significantly different from sites previously tested, or where reactive mixtures or
sequential zones of reactive materials are proposed, treatability tests can be highly instructive and are strongly
recommended. Laboratory treatability studies can be used to compare the reactivity and longevity of reactive
materials under uniform and controlled conditions, as well as to estimate half-life (t1P) information useful for PRB
design. In addition, performing studies at differing experimental temperatures allows determination of the
temperature dependence of the reaction rates. This temperature dependence might be described by an Arrhenius
expression, allowing prediction of rate constants for various types of reactions (i.e.. abiotic reductive dechlorination
or biologically mediated contaminant removal) over a range of temperature.
Laboratory treatability studies should be conducted using ground water from the subject plume. Although VOC
degradation rates in granular iron are not greatly influenced by the inorganic chemistry of ground water,
information on inorganic geochemical changes has proven to be very useful. Removal of inorganic contaminants
may be strongly influenced by the background geochemistry of the plume. Every effort should be made to maintain
the oxidation-reduction (redox) state of the ground water used in the studies. This requires both proper sampling
and storage in the field and as the water is transferred from the field sample bottles to the influent reservoir. The
possible effects of sample storage on redox state can be evaluated by comparing pre-batch and column laboratory
results to field pH, DO, and Eh measurements.
4.1.1 Batch Studies
Batch treatability studies are most suitable for screening candidate reactive barrier materials. Results obtained
using various types of materials give relative rates that can be useful in selecting the most appropriate material(s)
for subsequent testing and/or field application. Batch tests are usually faster, cheaper, and simpler to set up than
columns, and allow7 rapid comparison of varied parameters on the experimental results (Powell et al., 1995a).
Laboratory batch treatability tests should include blanks, which contain only site water, and reactive samples
containing the candidate reactive materials along with the site water. It is simple, for example, to determine which
of two types of reactive media is most effective for remediating a contaminant using batch tests. Three or more
replicates for each of the two media types can be set up in capped tubes or bottles using a constant aqueous volume,
contaminant concentration, and media mass (or surface area). After shaking for some interval, the concentration of
the contaminant can be analyzed in each of the replicates and a determination made whether a difference exists in
the effectiveness of the two types of media.
Increasing the experimental complexity somewhat by adding multiple sampling intervals allows the determination
of the rate of contaminant removal using the equations of classical kinetics. Batch systems are somewhat limited in
this regard, however, since shaking of the tubes or bottles negates many of the mass transport and diffusive effects
that would limit reactions in unshaken systems. Also, batch tests usually have very low ratios of reactive material
to solution relative to column tests and actual field implementations. Although Johnson et al. (1996) found that
batch and column studies did not give distinguishable distributions of results when rate data for VOCs was
normalized to iron surface area (ks,J), they did observe that experimental conditions determined whether the
degradation rates were reaction or diffusion limited. To develop a table of representative kinetic data for
dehalogenation by iron metal, they averaged reaction rates for a given halocarbon from column studies, batch tests,
16
-------
and regression analyses of observed rates (kobs) versus iron surface area concentration (i.e., m21/1). This resulted in
average ks/v values with relatively large standard deviations for each halogenated organic compound. For example,
TCE had an average kgA value of (3.9±3.6) x 1O4. Therefore, some caution is warranted when results from
laboratory experiments are extrapolated to field-scale systems.
The test procedures can vary significantly depending upon contaminants of interest and characteristics of the
reactive material. For a site at the U.S. Coast Guard Air Support Center, Elizabeth City, North Carolina, a series of
batch experiments were conducted to evaluate the effectiveness of various commercial iron sources for the
simultaneous removal of both hexavalent chromium, Cr(VI), and trichloroethene (TCE) from the site water. The
results are shown in Table 6, together with the column test results using the same site water. From Table 6, it is
apparent that using TCE batch test half-lives would result in an overdesign of the field treatment system. However,
Table 6. Typical Treatability Study Results. Elizabeth City Example.
6.1 Trichloroethene Half-Lives from Batch Tests.
Iron Source (1, 2, or 3)
1 plus silica sand
2 plus silica sand
3 plus silica sand
1 plus aquifer material
tia (hr)
32
73
31
22
6.2 Time to Non-Detectable Chromium Levels in Batch Tests.
Iron Source (1, 2, or 3)
1 plus silica sand
2 plus silica sand
3 plus silica sand
1 plus aquifer material
time to non-detect (hr)
0.6
1.2
0.4
0.72
6.3 Trichloroethene Half-Lives from Column Tests (v = + 2 ft/day).
Reactive Mixture
100% Source 1 Iron
100% Source 3 Iron
50% Source 1 Iron, 50%
Silica sand
50% Source 1 Iron, 25%
sand, 25% aquifer material
50% Source 2 Iron, 25%
sand, 25% aquifer material
tiQ (hr)
0.61
0.80
0.97
0.48
0.99
Note: chromium removal to detection limits was obtained within 0.1 ft of travel distance in all columns
(i. e., within ±1.2 hrs)
17
-------
the data does indicate which of the commercial mixtures tested (in this case iron sources I and 3) were the best
candidates for further testing using columns.
A second use of batch tests involves the evaluation of new PRB emplacement methods. Several existing
construction techniques, used historically to build low permeability (containment) walls, are being modified to
facilitate the construction of permeable iron barriers at greater depths. Several of these involve the use of a finer
grained fraction of iron in a biodegradable slum? mixture to allow injection or placement of this mixture in the
subsurface. Laboratory-scale and limited field-scale testing has shown that the reactivity of the iron is maintained
following contact of the material with guar based slum?; however, this should still be confirmed on a site-specific
basis. If an enzymatic or acid-based breaker is used to promote slurry breakdown, then batch tests can also be used
to confirm that the breaker used will function in the high pH, low Eh conditions established in the Fe(0).
4.1.2 Column Studies
Laboratory column tests are useful in determining contaminant removal rates under conditions that more closely
approximate the operating conditions anticipated in the field, such as flow velocity. These rates are the basis for the
design parameters used to determine the required residence time for the contaminant in the reactive material. Using
the residence time and the flow rate, the thickness of the treatment zone can be determined. The laboratory column
tests may also include sampling of the column profile and/or influent and effluent, in addition to the contaminant,
to assess changes in the major ion composition of the water. These data provide information concerning potential
mineral precipitation in the reactive material caused by changing redox potential (Eh) and pH conditions, which are
also important parameters affecting the removal of inorganic contaminants in PRBs. Mineral precipitation rates
may also impact the operation and maintenance requirements for VOC removal systems.
Though column tests are more costly and time consuming than batch tests, they typically yield more realistic field
performance rates, provide a better opportunity to examine products of the reactions and can provide useful
information concerning long-term performance.
4.1.2.1 Column Test Methodology
Columns are typically 10 to 100 cm long, and 2.5 to 3.8 cm inside diameter, with sampling ports at the influent and
effluent ends and preferably, also along the length (Figure 2). The sampling ports are designed to allow water
samples to be collected along the central axis. Details concerning column test methodology can be found in several
publications (see Appendix A).
Ground water obtained from the site is supplied to the influent end of the column at a constant flow velocity using
a laboratory pump. The flow velocity is selected to approximate the velocity expected in the field-scale treatment
zone. Contaminant concentrations are measured at the inlet, outlet, and sampling ports along the column even? 5 to
10 pore volumes (one pore volume is equal to the total volume of liquid within the column) until a steady-state
concentration profile is achieved (i.e., concentrations at a point remain relatively constant overtime). For example,
this generally occurs after 40 to 100 pore volumes of flow for VOCs in granular iron. Eh and pH profiles are also
measured periodically during the test period. Concentrations of the major cations, anions, and alkalinity are usually
monitored at less frequent intervals to help predict the potential for mineral precipitation within the reactive
material. If necessary, other chemical parameters relevant to a particular site can also be measured.
4.1.2.2 Interpretation of Column Data
For each test column at each velocity, contaminant concentrations are plotted as a function of distance along the
column. The flow rate is used to calculate the residence time at each sampling position (relative to the influent) for
each profile. The contaminant degradation or disappearance rate constants are calculated for each contaminant in
the influent solution ground water, using kinetic models. For VOCs and/or chromate a first-order model is used:
C = C0e-kt (21)
where:
C = contaminant concentration in solution at time t,
C0 = initial contaminant concentration of the influent solution,
k = first order rate constant, and
t = time.
By rearranging and taking the natural log, Equation (21) becomes:
In (C/CJ = -kt (22)
18
-------
Effluent Samples
Pump
Figure 2. Schematic of the column appartus used in the bench-scale tests.
The time at which the initial concentration declines by one-half. (C/C0 = 0.5), is the half life, which, by rearranging
equation (22), is given by:
t1/7 = 0.693/k (23)
The first-order rate constant, k, has the units of time*1 (e.g., sec"1, hr1) and is computed from the slope of the first-
order model obtained by fitting equation (22) to the experimental data. Half lives, along with corresponding
standard error of estimate (s Rvalues can then be calculated.
When the contaminant is a VOC, breakdown products from the VOCs in the influent solution (e.g., dichloroethene
isomers from TCE) may reach maximum concentrations at an interior sampling port. In this case, an approximate
degradation rate for the breakdown product can be calculated using the maximum concentration at this port, rather
than the influent concentration, as C0. Ideally, both parent and intermediate concentration data should be fitted
using a first-order kinetic model to determine degradation rates and conversion factors.
A variety of mechanisms are potentially suitable for the removal of inorganic contaminants. Many of the
mechanisms employed in treatment of inorganic species rely on the precipitation or adsorption of a chemical
constituent. Laboratory batch and column data should be combined with gcochcmical modeling to assess the
stability of potential precipitates, adsorbates, and to assess the potential utility of reactive mixtures for remediation
of inorganics. Altering the concentrations of potential nutrients (e.g., NH4+, PO43~) in biologically mediated systems
permits an assessment of potential nutrient limitations on the rate of contaminant removal.
A comprehensive characterization of the water chemistry within a batch test permits the use of gcochcmical
speciation calculations. Geochemical models commonly used to perform these calculations include MINTEQA2
19
-------
(USEPA, 1993) and PHREEQC (USGS, 1998). These models contain large and comprehensive databases that can
be used to evaluate the tendency for mineral phases to precipitate or dissolve based on the saturation state of
anticipated minerals. MINTEQA2 also includes a variety of surface ionization/complexation models, which can be
used to evaluate the potential for contaminant adsorption onto mineral surfaces.
The results from laboratory experiments can be augmented by the application of mineralogical characterization
techniques and surface analytical procedures. These techniques allow the isolation and identification of reaction
products, identification of the oxidation state of adsorbed or precipitated elements, and characterization of the
mineral structure of secondary precipitates. Mineralogical techniques can also be used to verify the identity of
secondary phases that are inferred from geochemical modeling. Accumulations of secondary minerals and the
occurrence of coatings can be used to infer reaction progress and to assess potential transport limitations arising
from reaction products blocking reactive surfaces.
When used in conjunction with geochemical speciation/mass transfer calculation, kinetics and column flow data
can also be used to assess the rate-determining steps within an overall reaction sequence. The results of long-term
column experiments can be used to assess variations in the reaction rates resulting from consumption of substrate
materials, growth of bacterial species, and the formation of potentially deleterious reaction products (Waybrant et
ah, 1997a). These data can also be used to assess changes in the dominant reaction mechanisms.
The effectiveness of treatment of contaminant species may be limited by changes in the accessibility of the
reactive material due to the formation of coatings by secondary phases. The degree and duration of treatment may
also be limited by depletion of the mass of reactive material, or in the case of adsorbent barriers, by the availability
of reaction sites on the substrate surface. Continuing controlled column experiments for prolonged periods of time
provides an opportunity to drive the treatment system to failure, to evaluate the potential limitations of proposed
field installations (Blowes et ah, 1997a).
4.1.2.3 Interpretation of Changes in Non-Contaminant Inorganic Constituents
The carbonate equilibrium present in the contaminant plume may be significantly affected as the ground water
passes through the reactive barrier. In Fe(0) columns, calcium and alkalinity concentrations normally decrease as
ground water passes through the iron in response to increasing pH caused by the corrosion of iron. Corrosion also
causes the Eh of the ground water to decrease (see Appendix B). As the pH of the solution increases, bicarbonate
(HCO^) ions are converted to carbonate ions (CO,2"). The CO32~ ion formed can then combine with the cations
present in solution (Ca2+, Fe2+) to form carbonate mineral precipitates such as calcite (CaCO,) and siderite (FeCO3).
At some sites, Mg2+ may precipitate in solid solution with CaCO3.
Independent corrosion rate measurements of metallic iron (Reardon. 1995) indicate that several mmol/L Fe2
would be introduced to ground water in these columns due to iron corrosion. In many instances, iron precipitates
including siderite (FeCO3), iron hydroxides and oxyhydroxides, and '"green rusts" will form in the column.
Observed higher carbonate alkalinity loss relative to the amount of calcium loss can be used to indicate that siderite
formed in addition to calcium carbonate. Iron (oxy)hydroxides which form in the column are converted over time
to magnetite (Odziemkowski and Gillham, 1997). Concentrations of iron are often observed to increase in the
upgradient portion of a column due to metal corrosion, followed by decreases in iron concentrations in the
downgradicnt part of the column as iron precipitates arc formed.
Results of the influent and effluent inorganic chemical analyses can be input to a geochemical speciation model
such as MINTEQA2. In MINTEQA2, aqueous concentrations are used to calculate the saturation indices (Sis) of
various mineral species. These Sis can be used to gauge the potential for these minerals to precipitate. A negative
SI indicates undcrsaturation with respect to the particular mineral phase, while a positive SI indicates ovcrsaturation.
Some caution must be used in interpreting these results, especially since such models assume equilibrium
conditions and often do not take into account the potential surface reactions involving the solid phases. Kinetic
controls on the reaction rates could, in some circumstances, result in lower rates of precipitate formation than
predicted by the speciation models.
The reliability of the column data to gauge the potential effect of precipitates on performance of field-scale PRBs
is dependent on the location of precipitation sites in-situ, and on the extent and kinetics of precipitation under field
conditions. From the calcium and alkalinity profiles collected during column tests, and the data from other field
trials, it appears that most of the carbonate precipitation occurs in the upgradient section of the reactive zones.
20
-------
A
B
.g
^P
S
•4-1
o
u
o
o
Residence Time = tc > tA > tB
Residence Time (hr)
A —> B —> C = reaction pathway for compounds A, B and C
t = Time
Co = Initial Concentration
MCL = Maximum Concentration Limits (Performance Criteria)
Figure 3. Conceptual residence time calulation for VOC removal.
3.000 •
2.500 -
c
o
••s
g 2.000 -
o
o:
"5
& 1.500 -
1.000
0.500 -
0.000 .
Correction Factor
10 15
Temperature (C)
measured value
calculated correction factor
20
25
Figure 4. Correlation of TCE degradation rates with temperature.
21
-------
4.2 Determination of Required Residence Time in PRBs
4.2.1 Chlorinated Volatile Organic Compounds (VOCs)
In the presence of granular iron, chlorinated organic contaminants degrade with first order kinetics. Thus required
residence times can be determined from the degradation half-life. Using half-lives and influent concentrations as
input, models are available that will calculate the time required to reach the desired treatment level. The most useful
of these models will accommodate simultaneous degradation of several compounds and will also accommodate the
simultaneous formation and degradation of breakdown products. As an example, the residence time calculation is
shown conceptually in Figure 3 for PCE degradation.
In the model, potential breakdown products are concurrently produced and degraded as described by first-order
kinetic equations. The model is an expression of the chemistry that is observed in the solution phase. For PCE, TCE,
cDCE and VC. the model takes the form
(24)
Etene
where:
f = mole fraction
k = first-order rate constant
In order to determine the VOC concentrations at a given time, first-order equations are used either directly in
commercially available software, or in their integrated form.
The residence times determined from VOC degradation rates measured in laboratory column studies of iron
materials need to be adjusted due to the effects of lower ground-water temperatures. Laboratory column studies are
normally performed at a temperature of about 20° to 25° C. Rate constants determined during laboratory studies can
be adjusted for temperature effects using the Arrhcnius equation. As an example, experimental data from controlled
temperature column tests at the University of Waterloo were used to generate the linear plot of TCE degradation
rate constant vs temperature (Figure 4, Stephanie O'Hannesin, personal communication). The plot indicates that at
15° C, TCE rates could be expected to decline by a factor of 1.4 from those measured at 23° C.
There has been little published to date concerning similar relationships for other VOCs such as PCE, TCA,
cisDCE, and vinyl chloride. Limited testing at the University of Waterloo has shown little dependence of cDCE and
vinyl chloride rates on temperature. Field observations at some sites with ground-water temperatures in the order of
8°-12° C have shown apparent decreases in TCE degradation rates by factors of 2 to 2.5, but often uncertainties in
flow rates, etc., affect the interpretation of results.
4.2,2 Inorganic Constituents
Reaction rates for inorganic species vary widely, and depend on the site specific characteristics of the aquifer, the
ground water and the reactive material. Laboratory tests using site ground water, or pilot-scale field tests may be
required to accurately estimate reaction rates representative of field conditions. Estimates of reaction rates can be
drawn from batch tests, however, the accumulation of secondary reaction products may not be accurately
represented by the short-term tests. Hie results of column experiments may provide more accurate estimates of
reaction rates under dynamic flow conditions, but again are limited by the duration of the test.
Reaction rates drawn from laboratory experiments can be incorporated into reactive solute transport models for
estimating anticipated reaction progress within a reactive barrier. Direct application of laboratory rates may be
possible for reactions that show little temperature dependence, or dependence on site specific parameters. Reaction
rates for Cr(VI) reduction using Fe(0) shows only slight temperature dependence, allowing direct application of
laboratory rates to predict field-scale performance. Biologically-mediated systems are anticipated to be more
susceptible to variations in temperature and in nutrient concentrations. Direct transfer of laboratory-measured rates
to these systems is less certain. Use of field-scale pilot tests is warranted until a greater understanding of the
limiting factors in biologically-mediated systems is obtained.
22
-------
Estimates of field reaction rates and determinations of residence time requirements must also include an
assessment of the potential variability of the constructed barrier. Reasonable factors of safety need to be included
in design parameters to account for construction defects and site characteristics. Ground-water flow velocities
frequently vary over small spatial intervals, resulting in sharp changes in residence times within a barrier system. In
addition, variability in the material reactivity and emplacement density may substantially influence the performance
of a barrier system (Bennett et al., 1997).
To assess residence requirements for the Elizabeth City project, five columns containing 100% granular iron and
various mixtures of iron and aquifer materials were tested using ground water from the site which contained both
TCE and Cr (VI). The column flow rates (about 2 ft/day) were similar to the flow rates predicted by a ground-water
transport model of the installed system. TCE half-lives obtained from the column tests are shown in Table 6.
Similar to other sites, TCE degraded at relatively high rates and chromium was removed quite rapidly in all
columns. The column data also indicated that VOC degradation rates rather than chromium removal rates, had the
greatest influence on the required residence time in the PRB. TCE half-lives (and the half-lives determined for
breakdown products cis-L2-dichloroethene and vinyl chloride) were subsequently used in the field design (Bennett
et al., 1997).
4.3 Ancillary Laboratory Studies
A variety of more specialized laboratory studies may be conducted to provide additional information which can be
used in PRB design. These studies could include:
(i) laboratory hydraulic measurements (falling head tests, etc.) along with field data to provide input data
for system modeling;
(ii) surface area measurements of various candidate reactive materials. (For some contaminants, these can
be used to predict the relative performance of the respective materials);
(iii)breakdown product and reaction pathway analysis. (This could be accomplished either in batch or
column tests and may be useful in examining compounds not previously tested);
(iv) microbial analyses of solid phase or water samples from batch/column testing. (These tests may be of
particular value in evaluating PRB materials that rely on the stimulation of biological activity);
(v) tracer studies can be conducted on core samples of PRB materials to assist in evaluating the effects of
precipitate formation on system performance (Mackenzie et al., 1997; Sivavec et al., 1997).
Mineralogical study could include X-ray diffraction (XRD), to determine the presence and nature of the reaction
products, and Scanning Electron Microscopy (SEM) to determine the morphology of secondary precipitates. When
SEM is coupled with Energy Dispersive X-ray Analysis (EDX), the qualitative distribution of elements within the
reactive materials and the secondary precipitates can also be evaluated.
Other studies could incorporate surface chemical analyses, including Auger electron analyses to determine
elemental composition, X-ray photoelectron spectroscopy to determine elemental concentration and oxidation
state, and secondary ion mass spectroscopy (SIMS) to assess elemental composition and perform depth profiling of
oxidation layers.
23
-------
5.0
Characterization for Permeable Reactive Barriers
A complete site characterization is of critical importance for the design and installation of a reactive barrier. This
characterization should include an evaluation of the surface features, structures and buried sendees to determine
whether the site is amenable to PRB installation and, if so, what types of PRB emplacement technologies are
feasible within these limitations.
The plume location and extent, ground-water flow direction and velocity, and contaminant concentrations must be
accurately known to achieve the required performance. In addition, information on stratigraphic variations in
permeability, fracturing, and aqueous geochemistry is needed for the PRB design. The plume must not pass over,
under, or around the PRB and the reactive zone must reduce the contaminant to concentration goals without rapidly
plugging with precipitates or becoming passivated. The PRB design, location, emplacement methodology, and
estimated life expectancy are based on the site characterization information, therefore insufficient or faulty
information could jeopardize the remediation effort.
In general, four aspects of site characterization should be evaluated before implementing a PRB:
• hydrogeology
* contaminant loading
* geochemistry, and when possible
* microbiology.
5.1 Hydrogeologic Characterization
As with any ground-water remediation technique, adequate hydrogeologic characterization must be done to
understand the ground-water flow patterns and the distribution of the contaminant plume. This is especially
important for the installation of a PRB since the entire plume must be directed through the reactive zone of the
barrier. To attain a "passive" remediation system, the PRB must be placed in a location that allows the plume to
move through the reactive zone under the natural ground-water gradient; i.e., the gradient must do all the work.
Information that must be obtained includes advective velocity parameters such as the piezometric surfaces (if a
confined aquifer) or gradient the hydraulic conductivity, porosity, and the usual hydrologic parameters typical of a
careful and complete subsurface characterization. It is also important to understand seasonal changes in flow
direction and flux due to processes such as recharge, since the PRB should be designed to accommodate these
changes. An awareness of the effects of large pumping systems at the site, such as water supply wells or pump-
and-treat remediation technologies, should also be known. These types of effects on flow velocity and direction can
be intermittent and have unforeseen influences on the flow entering the PRB. Figure 5 conceptually depicts how
changes in ground-water flow can reduce the effectiveness of a reactive barrier system due to incomplete plume
capture.
Beyond general hydrologic factors, the stratigraphy and lithology of the site will often dictate the type of PRB
design chosen. It is desirable to "key" the bottom of the barrier into a low-permeability clay layer to prevent
contaminant underflow, for example. If such a layer is not available at a reasonable depth, then a "hanging" design
Time
New GW Direction
Due to
Recharge
Stream
Stream
Plume & Barrier
at Installation
Plume and Barrier
during Rainfall Event
Figure 5. Potential effects of ground water flow diversion on plume direction.
24
-------
might be necessary. This type of system should be engineered to prevent contaminant underflow. In addition,
stratigraphic and lithologic features might limit the ability to perform certain installation procedures. Buried rock
might interfere with the driving of sheet piling, for example.
Understanding the vertical variation in stratigraphy is also important for choosing the stratigraphic zone(s) that the
PRB will intersect. If the contaminant plume is moving through a highly permeable layer amidst less permeable
layers, the PRB should be placed vertically to encompass this high permeability layer. It is also important that
impermeable materials, such as clays, not be "smeared" by the installation techniques across the permeable aquifer
zone that is expected to provide flow into the PRB. Therefore, a careful evaluation of the stratigraphic variability at
the location of the PRB, and the continuity of the stratigraphy with respect to the upgradient plume, will provide
confidence in the design and installation.
Stratigraphic features such as fractured rock are also important. If the plume-carrying flow zone is contacted by-
zones of fractured rock, allowing some flow diversion into the fractures, then it is possible that some fraction of the
contaminant could be short-circuited around the PRB. This could give the appearance that the remediation within
the PRB is incomplete when, in fact, the PRB simply fails to intersect all the potential flow paths of the
contaminant. Reactive media emplacement into bedrock fractures may be necessary to intercept contaminant
migration pathways and achieve performance goals.
5.2 Contaminant Characterization
Information on contaminant concentrations is necessary for any successful remedial operation, whether pump-
and-treat or in situ technologies are used. The remediation technique must be effective up to the maximum
concentrations and the total mass of contaminant that will be encountered at the treatment point. This is especially
important for PRBs because, once emplaced, it is difficult to change the thickness of the reactive zone. The PRB
must also be designed to eliminate the possibility that portions of the plume could flow around the barrier in any
direction. This requires a complete understanding of the extent of the plume, its width, depth, length, contaminant
concentrations within these spatial dimensions, and how these can be expected to change overtime.
Because PRBs are generally placed downgradient of the plume center-of-mass, it is important that the barrier be
designed to accommodate the higher upgradient concentrations, should they arrive at the barrier unattenuated. This
requires sufficient contaminant characterization to accurately determine this high concentration zone. It is also
desirable to know whether this zone is moving downgradient over time or whether a pseudo-steady-state has been
achieved that would suggest natural attenuation is occurring. If steady-state plume boundaries have been achieved,
the PRB could be designed to transect and attenuate only the lower downgradient concentration(s) for the protection
of nearby receptors or to eliminate contaminant migration beyond site boundaries. Figure 6 illustrates the potential
result if a reactive barrier is not designed to fully handle the contaminant concentrations that reach it over time.
5.3 Geochenikal Characterization
Site/plume geochemical information is needed for PRB design and implementation and to further our understand-
ing of the expected lifetime of these systems. It has been shown, for example, that water passing through a reactive
Time
Incomplete
Remediatio
Stream
Stream
Plume & Barrier
at Installation
Plume and Barrier as Center of
Mass Moves Downgradient
Figure 6. Effects of plume center-of-mass downgradient. and higher concentrations impacting the barrier. Incomplete
remediation is now occurring.
25
-------
barrier of Fe(0) undergoes radical geochemical changes, including an increase in pH up to 9 or 10. elimination of
oxygen with Eh reduction to minus several hundred millivolts, and a reduction in carbonate alkalinity. Often,
sulfate is reduced to sulfide and dissolved iron appears. Other changes might also be observed, dependant upon the
contaminants being reacted and the overall ground-water geochemistry. Some of the geochemical changes can
result in precipitation onto the reactant surfaces, potentially reducing the reactivity and permeability of the reactive
zone overtime. Therefore, waters high in carbonate might result in significant buildup of calcite (CaCO3) orsiderite
(FeCO3). In addition, (oxy)hydroxides can also be expected to precipitate. Of particular potential importance is the
precipitation of ferric hydroxide, Fe(OH)^, at the upgradient reactive zone interface due to the reaction of dissolved
oxygen with the iron. At this time, the significance of these reactions over the lifetime of a field installation is not
fully understood. However, column plugging in lab studies, particularly at the influent end where dissolved O2
contacts the Fe(0), has been observed. Laboratory research is under way to identify methods of "refreshing" the
reactivity of the Fe(0) surfaces should this become necessary.
To illustrate these possible effects, the upper graphic in Figure 7 depicts the flow of TCE-contaminated water
through a portion of a PRO of Fe(0) shortly after installation. TCE enters and ethylene exits, with a velocity of 1 and
a residence time of 1, exactly as the wall was designed. In the middle graphic precipitates have begun to
significantly coat the iron particles, limiting access of the TCE to the surface for dechlorination reactions. The
velocity is still about one, as is the residence time; however, in this system a residence time of one does not provide
sufficient contact with the reactive surfaces due to the coatings. This results in the intermediate products of
incomplete dechlorination appearing downgradient of the PRB. The bottom graphic shows that further precipitation
has blocked the pore spaces between some iron particles. As a result, the entire ground-water flux must pass through
the remaining open channels, making the velocity greater than one, the residence time less than one, and
dechlorination incomplete. This type of plugging could also result in upgradient head increase and flow around the
barrier rather than through it. It should be noted that significant plugging problems have not been observed in field
scale applications using iron to degrade TCE.
5.4 Microbial Aspects
One area of site characterization for PRBs that needs further study is microbial activity. The interactions of native
microbial populations, contaminants, and reactive barrier materials are likely to be quite complex, and have the
potential for either beneficial or detrimental effects on the remediation. Native microbial consortia are often
responsible for natural plume attenuation processes. Other beneficial effects could include enhanced contaminant
degradation within or downgradient from the PRB; for example, the further reduction in Eh due to the presence of
Vparticle = 1
'residence
Ethene
V
particle
residence
Vparticle > 1
t < 1
"residence
Figure 7. Potential effects of precipitation buildup on the iron over time.
26
-------
sulfate-reducing bacteria might increase the rates of contaminant reduction. Adverse effects could include the
potential loss of permeability of the reactive zone due to biofouling. Additional laboratory and field studies are
being conducted to understand these interactions and learn how to enhance the potential positive effects and reduce
the potential negative effects of native microorganisms.
5.5 Implementing the Field Characterization
All known information about the site should be assessed prior to mobilizing the field characterization. Usually
some information has been obtained in the process of discovering that there is a contamination problem. Due to
regulator}' concerns and reporting requirements, monitoring wells may have been installed shortly after this
discovery to evaluate the contaminant(s) and its concentration. This information can be combined with historical
records, the memories of current or former workers, and surface features to obtain at least a general idea of the
location of the source term(s) and its nature. General information on topics such as aquifer locations, yields and
water table depths, water quality, stratigraphy, recharge areas, drainage basins, etc., can often be obtained from
local, state and federal agencies. Although this information might not be sufficient at the (usually) smaller scale of
the contaminated site for effecting a complete remedial design, it can be very helpful in designing the site
characterization effort.
When this information has been compiled, the choice of locations and screen depths for a few, select aquifer
testing/monitoring wells can be made, primarily to obtain general hydrologic information. Historically, plume
dimensions and site characteristics have also been determined during the installation and use of such monitoring
wells. However, these are expensive, time-consuming to install and sample, and require special multi-level
installations to provide adequate delineation in the vertical dimension. A better approach might be to begin with
surface geophysical techniques such as electromagnetic surveys, ground-penetrating radar (GPR) and seismic-
studies. These techniques arc of particular use when a site is completely unknown with regard to subsurface
features. Geophysical techniques can locate buried drums, pipes, and power lines in addition to providing
information on water table and bedrock depth, fractured zones, and strata thickness. Once this information is
compiled, push-tool technologies can be used for further characterization, when possible, otherwise conventional
drilling techniques may be necessary. If conventional drilling is necessary, careful consideration should be given to
the surface locations of the wells and the location of the screens with depth.
Push technologies, such as Geoprobe®, Hydropunch®, and cone penetrometers are rapidly becoming the tools of
choice for evaluating shallow plume locations in fine- to medium-grained soils of low to moderate density. Direct
push samplers can be driven rapidly and relatively inexpensively, allowing more points to be sampled than could be
accomplished with monitoring wells for the same amount of time and money. They can be used to collect soil cores,
water, and soil gas samples. Additionally, they can be used to collect these samples over very thin vertical intervals.
allowing better delineation of the contaminant concentrations in the plume and better stratigraphic characterization
than could normally be acquired with conventional exploratory drilling techniques.
27
-------
6.0 Permeable Reactive Barrier
6.1 The Continuous PRB and Funnel-and-Gate Designs
The residence time required and the anticipated ground-water velocity through the PRB are used to determine the
size of PRB needed to achieve the desired treatment level. Two basic configurations are currently being used for
full-scale field application, continuous PRBs or "'funnel-and-gate"' designs. The effect of system configurations on
flow patterns and system dimensions can be evaluated as part of the modeling effort. Factors which may affect the
choice of configuration are discussed below.
Properly designed and constructed continuous PRBs (Figure 8a) have relatively minimal impact on the natural
ground-water flow conditions at a site. The type of continuous PRB most commonly being installed is simply a
trench that has been excavated and backfilled with granular Fe. Several other emplacement methods are possible
such as hydraulic fracturing and jetting. The continuous design incorporates no funnels and. therefore, allows the
water to pass through the barrier under its natural gradient and at its natural flow velocity. As a result, a continuous
PRB only needs to cover an area comparable to the cross-sectional area of the plume. The ground-water flow
velocity through the PRB will be very similar to the velocity in the aquifer. As long as the hydraulic conductivity of
the aquifer is less than mat of the PRB, underflow of contaminated ground water should not occur.
Ideally the continuous PRB is built to a depth that somewhat over-encompasses the vertical and horizontal
dimensions of the contaminant plume, as a safety factor, and is filled with granular iron or some other reactive
material. The PRB thickness must be sufficient to remediate the contaminant of concern to the established
concentration goals. As with the funnel-and-gate design, it is desirable to place the bottom of the continuous PRB
into an impermeable zone; i.e., "key" it into impermeable strata to mitigate the potential for contaminant underflow.
Hie upgradient and downgradient surface areas of the aquifer material contacting the PRB will be approximately
the same, minimizing disruption in the natural ground-water flow relative to current funnel-and-gate designs
(Figure 8b). _ .. _.„„
v & / Continuous PRB
Contaminant Plume
Figure 8a. Plume capture by a continuous PRB trenched system. The plume moves unimpeded through the reactive zone.
Reactive Gate
Funnel-and-Gate PRB
Funnel
Contaminant Plume
Figure 8b. Plume capture by a funnel-and-gate system. Sheet piling funnels direct the plume through the reactive gale.
28
-------
A difficulty with the trench emplacement involves trenching and filling in the fully saturated zone due to the
problems of immediate water intrusion and potential collapse of the trench walls. This problem requires that the
aquifer materials be removed and the reactive media emplaced nearly simultaneously unless subsurface construc-
tions (sheet pile walls, trench filling with biodegradable slurries such as guar, etc.) are used. This rapid aquifer
material removal/reactive media emplacement is possible using continuous trenching devices such as the one shown
in Figure 9. During June 1996 a 150 foot long, 2 foot wide. 24 foot deep continuous trench PRB of iron was
installed at the U.S. Coast Guard Air Support Center near Elizabeth City. North Carolina, to intercept a mixed
plume of chromate and TCE, following the success of a small-scale field test (Puls et al, 1996). This marked the
first time that a continuous trenching technique had been used to emplace iron in a subsurface reactive barrier. A
drawback of this type of PRB installation is the difficulty tracking the volume of reactive material actually
emplaced during the construction.
In a funnel-and-gate configuration, low permeability funnels direct ground water toward permeable treatment
zones or gates.The basic design of a funnel-and-gate system is shown in Figure 8b. The "funnel" typically consists
of sheet pilings, slurry w7alls, or some other material and is preferably "'keyed" into an impermeable layer (clay,
bedrock) to prevent contaminant underflow. This funnel is emplaced to encompass and direct the flow of
contaminated water to a "gate" or "gates" containing a permeable zone of granular Fe(0) or other reactive material.
Due to directing a large cross-sectional area of water through the much smaller cross-sectional area of the gate.
ground-water velocities within the gate will be higher than those resulting from the natural gradient. The funnel
portion of the design is engineered to completely encompass the path of the contaminant plume and the overall
design must prevent the contaminant plume from flowing around the barrier in any direction. The gate shape may-
be controlled by construction techniques, but have most commonly been rectangular. Recent variations include the
use of backfilled caissons, media filled hollow7 vibrating beams, or emplaced reaction vessels. For the emplaced
reaction vessels, contaminated upgradient waters are directed into the subsurface vessel which contains reactive
media, and the treated water is discharged through a pipe that extends downgradient through the impermeable wall.
The more typical rectangular, or box-shaped gate can be built by driving temporary sheet pilings and/or building
removable subsurface w7alls within w7hich the reactive materials are emplaced. Sheet pilings can be driven to
delineate the sides and ends of the media gate. The interior of this construction can then be dewatered and excavated
to make room for the reactive material (Gillham and Burns, 1992). Water intrusion is reduced by the impermeable
pilings and any seepage can be pumped away. This construction approach also allows the incorporation of other
features, such as additional treatment zones or monitoring zones.
Hie permeability of the reactive material in the gated zone must be equal to or greater than the aquifer
permeability to minimize flow restrictions. At the same time, it is necessary for the flowpath length through the gate
to be long enough for complete contaminant remediation; i.e., sufficient contact, or residence time, must be allowed
between the reactive material and the contaminant. Overall system length depends on the number, location and size
of treatment gates and should be determined through ground-water flow modeling. Particular care is required in
designing and constructing the connection between the impermeable "funnel" section and the permeable "gate"
section in order to avoid bypass of contaminated ground water.
Some problems can accompany the funnel-and-gate design that should be addressed, prior to installation, with a
careful site characterization. Foremost among these is the potential for diversion of ground water around the funnel-
and-gate. Even if the permeability of the gate section is tremendously enhanced relative to the aquifer permeability,
the gate again contacts the same aquifer material on the downgradient side and permeability is immediately reduced
to the initial aquifer value. It is important that the gated zone and the downgradient gate/aquifer interface be able to
pass the flux of contaminated ground water intercepted by the funnels, if not the entire flux of intercepted water
(both contaminated and uncontaminated). It would seem unlikely that the total volume of water in a high flux
system, directed by a high surface area funnel, could be infiltrated through the much smaller surface area of the gate
and aquifer material interface that contacts the downgradient side of the gate. This constriction of flow could result
in a significant head difference across the reactive barrier. In fact, a seven to ten foot head difference reportedly-
developed across the funnel and multiple-gate system that was installed at the Denver Federal Facility, Denver, CO
(dishing et al., 1996) and has continued to persist (McMahon, 1997). The important point when designing a funnel-
and-gate system, however, is not that some water is diverted around the barrier, but that none of the contaminated
water is diverted and it passes through the reactive zone for treatment.
Several combinations or "treatment trains" of different types of reactive materials might be considered at specific
sites containing a combination of contaminants (e.g., metals and VOCs, or VOCs and petroleum hydrocarbons).
The funnel-and-gate PRB configuration may be more appropriate for these systems, providing a "focused"
29
-------
treatment area. This may be of particular value if the materials need to be replaced or replenished periodically over
the life of the system. These '"focused'' treatment areas will also facilitate performance monitoring of the system.
However, research needs to be done on how to integrate differing reaction technologies into a subsurface "treatment
train,"' so that the reactions do not interfere or limit one another.
Many of the reaction mechanisms associated with PRB materials are surface area dependent. The amount of
reactive material needed is, therefore, proportional to the mass flux of contaminant requiring treatment. Therefore,
both the funnel-and-gatc and continuous PRB configurations should theoretically require the same amount of soil
excavation and disposal since the PRB should contain the same volume of reactive material. However, in an actual
field implementation, it is possible that more iron will be required by a continuous PRB using current commercial-
scale installation techniques (trenching), particularly if the plume is very broad and requires a relatively long PRB.
The reason for this is the need to be certain that the PRB has no unrcactivc gaps or flowpaths throughout its length.
There will be some limiting thickness that has to be maintained to assure PRB integrity throughout its volume. The
use of jetting, as an emplacement method rather than trenching, could potentially mitigate this problem for thin
continuous PRBs.
Given the above, the selection of an appropriate configuration should be made on a site-specific basis. The cost of
construction of either configuration will ultimately be dependent on the depth, width and saturated thickness of the
plume, which controls the overall dimension of the system. Costs for reactive materials, 'Tunnel" materials (if
required), and construction will be the major capital cost components of these systems.
When considering emplacement methods and system configuration, several other factors common to subsurface
construction procedures need to be considered, including:
* the need for dewatering during excavation;
* the means and costs of ground water and soil disposal;
• health and safety. The only "hazard" associated with granular iron used to date has been nuisance dust.
However, the entry of construction personnel into gate excavations does create a health and safety issue;
• disruption to site activities
These factors will also influence the cost of the system.
6.2 Emplacement Methods and Comparisons
The installation methods used thus far for PRBs are few in number. Most systems have relied on standard means
of excavating and backfilling to place reactive material in the path of a plume, as was discussed for the continuous
PRB and the funnel-and-gate designs. Likewise, the impermeable walls of funnel-and-gate systems have used sheet
pile driving or slurry wall constructions. Among the other methods being developed, tested, and implemented are
the use of trenching machines, mandrel or tremie tubes, deep soil mixing, high-pressure jetting and vertical
hydraulic fracturing. It is unlikely that any single approach will be found to be feasible and cost-effective for all
types of applications. Rather, the value of having a variety of methods from which to choose will be the ability to
select the best method for each application, with its particular combination of depth requirement, reactive media
thickness requirements, stratigraphy, sensitivity to spoils generation, and other factors. Calculations based on the
equations in AppendixC can be used to estimate whether a given emplacement method can provide sufficient iron
thickness to accomplish site-specific remediation goals, provided the method also achieves the other emplacement
requirements.
6.2.1 Con ventional Excavation
Conventional excavation is a basic construction practice using equipment such as backhoes, excavators, and
cranes. Standard excavators can reach depths of about 35 feet, and modified excavators can reach depths of about
70 feet. Cranes fitted with clamshells can attain greater depths.
Shallow trenches may not require any support to remain open before backfilling, but most PRB applications must
be deep enough that they will need to be supported prior to being backfilled with either permeable or impermeable
materials. There are two means of supporting trenches: structural walls (e.g., trench boxes) and slurries.
The trench box approach has been used to install the gate portions for most PRBs thus far. Often the final stages
of excavation involve time-consuming confined entry procedures, making gate installation the most expensive
component of a PRB project. As a result, some recent projects have tested other means of supporting trenches,
30
-------
namely caissons (used for the Somersworth, NH, and Dover Air Force Base projects), continuous trenching
machines (used at Elizabeth City, NC, Oregon, New York and California) and biodegradable slurries (Vancouver,
Canada).
The use of slurries to support trenches prior to backfilling may provide significant cost savings. Bentonite slurries
are commonly used for this purpose in constructing slum? walls. Bentonite slurries cannot be used to construct
permeable zones, but using a biodegradable polymer solution such as guargum may be feasible and cost-effective.
A biodegradable slum? was used to install a small-scale test PRB (30ft x 6ft x 30ft) in Vancouver, Canada. Tests are
ongoing to evaluate the effects (if any) of the biodegradable slurries on iron and other reactive materials.
6.2.2 Trenching Machines
Trenching machines are an effective means of installing relatively shallow trenches. The type most applicable to
installing PRBs are of the "chain saw" design, which cuts a trench between 12 and 36 inches wide. During
installation, the trench is held open by a trench box attached to the chain saw cutting belt mechanism. The trench
box is pulled along after the cutter chain. A hopper attached to the top of the trench box can be filled with reactive
material and feeds the material into the trench. Typical trenchers are limited to depths of about 20 to 30 feet, but
larger equipment is expected to be more widely available in the future. Based on calculations presented in
Appendix C. for a trench width of 36 inches and a granular iron bulk density of 160 lb/ft2, the maximum weight-per-
area achievable by this method is 480 lb/ft2.
6.2.3 Tremie Tube/Mandrel
A tremie tube approach is a modification of the means of installing wick drains. A hollow rectangular tube with
expendable drive shoe on the bottom is driven to depth with hydrostatic force or a vibratory hammer. The tube can
then be filled with dry granular material or a slurry containing the reactive media. The tube then is extracted,
leaving the drive shoe and added materials in the ground. Then the process is repeated along the desired path, each
emplacement overlapping the previous one by an amount necessary to provide continuity.
This method was used to install iron-containing zones for the Phase ITA Lasagna™ project at the DOE gaseous
diffusion plant in Paducah, Kentucky. In that case, mixtures of iron/coke or iron/clay were installed to a depth of
45 feet. At a bulk density of 160 lb/ft3, the use of 100% granular iron in a 4-inch zone would give a weight-per-area
of slightly more than 50 lb/ft2 (see Appendix C).
6.2.4 Deep Soil Mixing
Deep soil mixing utilizes large augers (e.g., 3 ft to 8 ft diameter) suspended by cranes and driven by large motors.
Once a zone is mixed, the auger and crane are indexed to the next location. Applications carried out to date include
injecting cement or other grouting agents to build containment walls or to stabilize an area. In addition, deep soil
mixing has been used with air or steam injection to volatilize contaminants.
It is thought that relatively large amounts of iron can be emplaced by deep soil mixing, but little data is available.
Assuming that the resulting zone is 50% iron, 3-foot augers could be used to apply about 180 lb/ft2. and 8-foot
augers could apply up to about 480 lb/ft2 (see Appendix C).
An important aspect of deep soil mixing is the relatively high mobilization costs associated with transporting and
setting up the large equipment involved. For small projects, this may be prohibitive.
6.2.5 High-Pressure Jetting
High-pressure jetting is an established practice to inject grouting agents to improve the structural characteristics of
soil for construction purposes. More recently, it has been used to inject grouts to make impermeable walls. Jetting
is an attractive concept for many reasons. An important advantage in certain settings will be the ability to install a
wall around obstructions such as boulders and utility lines. Also, the equipment is small and the mobilization costs
are relatively low.
This approach uses jetting nozzles incorporated into a specialized section of the drill string located above the drill
bit. Once the drill string reaches the desired depth, the pump increases its output (up to 90 gpm and 6,000 psi). As
the slurry is injected into the formation, the drill string is extracted from the borehole at the desired rate. Due to the
jetting process, most of the finer soil fractions are forced to the surface, but a significant portion of the coarser
fractions remain.
If the jetting nozzle is rotated during extraction, a column of injected material forms which is approximately three
to seven feet in diameter. Depending on the pumping and extraction rates, it is anticipated that the columnar zone
31
-------
will contain up to 75% of the injected reactive material. Therefore, a 3-foot diameter permeable reactive zone might
contain up to 360 lb/ft2 granular iron (see Appendix C).
If the jetting nozzle is not rotated during extraction, it creates an injected zone known as a thin diaphragm wall. If
such a wall averages five inches thick and contains 75% iron, then the resulting weight-per-area is 50 lb/ft2 (see
Appendix C).
A pilot test to inject granular iron was conducted recently by DuPont. Guar gum (a natural food thickener) was
used as the viscosifying agent for making slurries of 50-mesh iron particles. Both columnar and thin diaphragm wall
injections were successfully demonstrated. The depth of injection was only fifteen feet in this case, but the principle
of using high pressure jetting was established, and it appears to be ready for use at greater depths.
6.2.6 Vertical Hydraulic Fracturing and Reactant Sand-Fracturing
In vertical hydraulic fracturing (VHP), holes are bored to initiate a fracture in permeable sands. A Fe(0)-
containing fracturing fluid gel is then pumped into the fractures, forming a continuous wall of reactive material. The
gel can be guar gum. which dissolves and leaves a PRB of Fe(0). The fracturing fluid used at the Caldwell trucking
site, New Jersey, consisted of potable water, guar gum, a borax cross-linker (to link iron to the gel), pH buffer, an
enzyme breaker (to break down starch in the guar after injection), and a fine-sand propant (Hocking, 1998). Cross-
linking was necessary to prevent the iron from falling from the gel before permeating the fracture. Because many
undesired reactions can occur between gel ingredients, potentially reducing the reactivity of the iron, proposed
fracture fluid ingredients should be extensively tested prior to using this installation method.
Reactant sand-fracturing (RSF) uses high-pressure fracturing with a sand propant, taking advantage of the
fractures that exist in bedrock, and providing a means of creating reactive fracture zones of Fc(0) within
contaminated bedrock. As with VHP, a reactive fracturing fluid is needed for RSF since granular iron and iron
filings do not have the needed hydraulic properties. An iron foam propant was chosen for a pilot test at a California
facility as a method for placing reactive Fe(0) media into a fractured bedrock aquifer contaminated with chlorinated
solvents and metals (Marcus, 1998). The pilot test consisted of boring holes, with chemical and physical testing to
identify fractures containing contamination, pretreatment hydraulic fracturing (to ascertain that the injection
equipment could fracture the bedrock), injection of the foamed iron propants, and post-treatment confirmation of
emplacement with down-hole geophysical and geochemical testing. Cross-linkers in the propant were not necessary
due to the low density of the iron foam relative to some other Fc(0) forms. The low-density foam also allowed the
use of standard hydraulic fracturing equipment.
6.3 Emplacement Verification
As part of the quality assurance program for the installed system, certain measures should be used to verify
emplacement of the PRB system as designed. This is most critical for continuous barrier configurations, but also
required for funnel-and-gate and other emplacement designs. Whether the PRB is installed with excavation and fill,
continuous trenching, hydrofracturing. or jetting, its emplacement and integrity should be confirmed. This will
entail additional analyses beyond contaminant disappearance in downgradient monitoring wells. Geophysical
methods have been used to a limited extent in this regard. These methods include natural gamma, conductivity,
electrical resistivity, cross borehole tomography and surface radar. Tracer tests and standard hydrologic character-
ization methods can also provide useful information, especially when the data are combined with and compared to
the data from the geophysical studies.
32
-------
7.0 Monitoring Permeable Reactive Barriers
7.1 Planning the Monitoring Effort
A Quality Assurance Project Plan (QAPP) must be developed prior to any ground-water sampling, including
baseline water quality data collection efforts. The QAPP must address the data quality requirements established for
the project. U.S EPA's Pocket Guide for the Preparation of a Quality Assurance Plan (EPA/600/9-89/087) is a
useful resource in this regard. A monitoring plan is an integral part of the QAPP for the project and emphasis should
be placed on data collection methods and monitoring network design equal to that traditionally placed on analytical
methods and sample handling.
All procedures and techniques used for site characterization, ground-water monitoring well installation and
development, sample collection, sample preservation and shipment, analytical procedures, and chain-of-custody
control should be specified in a QAPP. At a minimum, the QAPP should address
• sampling objectives
* pre-sampling activities
• sample collection
* in-situ or field analyses and equipment calibration
• sample preservation and handling
* equipment decontamination
• chain-of-custody control and records management
* analytical procedures and quantification limits for both laboratory and field methods
• field and laboratory quality assurance/quality control
* evaluation of data quality, and
• health and safety.
Other sources of useful information on the topics of QA procedures and ground-water monitoring include the
following: Chapter Eleven "'Ground Water Monitoring" of EPA's manual titled Test Methods for Evaluating Solid
Waste (U.S. EPA 1986c, commonly known as "SW-846"); Ground Water Sampling - A Workshop Summary
(EPA/600/R-94/205, 1995); Subsurface Characterization and Monitoring Techniques - A Desk Reference Guide,
Vol. 1: Solids and Ground Water and Vol. 2: The Vadose Zone, Field Screening and Analytical Methods
(EPA/R-93/003b, 1993).
7.2 Compliance Monitoring
7.2.1 Objectives
Compliance monitoring typically involves the monitoring of the contaminants of interest at a particular hazardous
waste site where dissolved concentrations have been detected that exceed regulatory limits. The focus is on the site
and its compliance points with the monitoring program driven by regulatory requirements. General water quality
monitoring is also often included, such as determinations of major cations and anions and other water quality
indicator parameters such as pH, alkalinity, specific conductance, etc. For permeable reactive barriers, similar
compliance monitoring requirements are necessary; however, the placement and design of monitoring wells or
points and the methods used to sample ground water may be different. In addition, monitoring of degradation
products from reductive dehalogenation reactions or other contaminant transformation products may be required.
7.2.2 Compliance Monitoring System Design
Well placement and design are important to ensure adequate assessment of system performance. In addition to
upgradient and downgradient wells, wells should be located to ensure that contaminated water is not flowing
around, under, or over the barrier wall. The number of wells used will depend on system design and size. Figure 8
shows two examples for a funncl-and-gatc design and a continuous wall design. Use of two inch diameter wells is
usually sufficient and appropriate for compliance monitoring purposes. Smaller diameter wells or piezometers can
be used if the selected sampling devices permit it. Selection of screen length should be compatible with sampling
program objectives and site conditions, particularly with respect to plume location as it exists in three dimensions
within the aquifer (Powell & Puls, 1997c) and the depth of the reactive zone. Compliance wells located near the
permeable reaction zone should be located far enough away to avoid mixing of water from distinct geochemical
33
-------
zones during purging and sampling of the wells. For example, a well located immediately downgradient of the PRB
should be located such that water is not extracted from the reaction zone itself during w7ell sampling. .However, the
well should not be located so far downgradient that it takes a long time to determine the effectiveness of the PRB.
Similarly, wells located upgradient should not be too close to the PRB, such that water is inadvertently pulled from
the reaction zone during purging and sampling operations. Wells located beneath the PRB should also be carefully
located with respect to the target sampling zone. The sampling device and sampling methods employed will dictate
how close wells can be located to the PRB and some thought must be given to the zone of influence during the
sample collection process.
It should also be realized that if the PRB has been located to transect the contaminant plume rather than being
installed entirely downgradient of the plume, compliance wells downgradient of the PRB will probably not show
immediate reductions in contaminants. This is because diffusion and desorption of the contaminant from the aquifer
materials downgradient of the PRB will continue for some time before gradually diminishing. Well number 46 at
the Elizabeth City, North Carolina, PRB site is a good example of this phenomenon. The PRB was installed in June,
1996, with Well 46 (Figure 9) at a distance of about 56 feet downgradient of the PRB. In November of 1996, five
months after PRB installation, the TCE concentration at Well 46 was 256 (ig I/1. The following September (1997),
the TCE concentration was approaching compliance for TCE (10.9 ug I/1 ) and was in compliance for VC.
7,2.3 Compliance Sampling Methods
Low-flow sampling methods are recommended for compliance sampling purposes (Puls and Barcelona. 1996).
Low-flow refers to the velocity with which water is withdrawn from the aquifer and is usually from 0.1 to
0.5 L/min, depending upon site-specific hydrogcologic conditions. The effectiveness of using low-flow purging is
intimately linked with proper screen location, screen length, and well construction and development techniques.
The objective is to pump in a manner that minimizes stress to the hydrogeologic system, to the extent practical,
taking into account established program sampling objectives.
The rccstablishmcnt of natural flowpath equilibrium in both the vertical and horizontal directions following
disturbance of the water in the well and prior to sampling are important for correct interpretation of the data (Powell
and Puls, 1993). Much of the need for purging has been found to be due to passing the sampling device through the
overlying casing water, which causes mixing of these stagnant waters and the dynamic waters within the screened
interval. These disturbances and impacts can be avoided using dedicated sampling equipment set within the
screened interval, which precludes the need to insert the sampling device prior to purging and sampling (Powell and
Puls, 1997c).
Figure 9. Continuous trencher used at Elizabeth City. North Carolina.
34
-------
For high resolution sampling needs, screens less than 1 m in length should be used. Isolation of the screened
interval water from the overlying stagnant casing water may be accomplished using low-flow minimal drawdown
techniques. If the pump intake is located within the screened interval most, if not all. of the water pumped will be
drawn directly from the formation with little mixing of casing water or disturbance to the sampling zone. However,
if the wells are not constructed and developed properly, zones other than those intended may be sampled. At some
sites where geologic heterogeneities are sufficiently different across the screened interval, higher conductivity
zones may be preferentially sampled. This is another reason to use shorter screened intervals, especially where high
vertical spatial resolution is a sampling objective.
It is recommended that water quality indicator parameters be used to determine purging needs prior to sample
collection in each well. Stabilization of parameters such as pH, specific conductance, dissolved oxygen, oxidation-
reduction potential, temperature and turbidity should be used to determine when formation water is accessed during
purging. In general, pH stabilizes first, along with temperature and specific conductance, followed by oxidation-
reduction potential (redox), dissolved oxygen and turbidity (Puls et al., 1992; Puls and Powell, 1992a). Temperature
and pH, while commonly used as purging indicators, are actually quite insensitive in distinguishing between
formation water and stagnant casing water; nevertheless, these arc important parameters for data interpretation
purposes and should also be measured. Instruments are available which utilize in-line flow cells to continuously
measure the above parameters. Most of these same parameters are important performance measures for reactive
barriers (e.g., for iron corrosion reactions occurring within the treatment zone).
It is important to establish specific well parameter stabilization criteria and then consistently follow the same
methods thereafter. Generally, the time or purge volume required for parameter stabilization is independent of well
depth or well volumes for low-flow sampling. Important variables are well diameter, sampling device,
hydrogeochemistry, pump flow rate, and whether the devices are used in a portable or dedicated manner. If the
sampling device is already in place (i.e., dedicated sampling systems), then the time and purge volume needed for
stabilization is much shorter. Other advantages of dedicated equipment include less purge water for waste disposal,
much less decontamination of equipment, less time spent in preparation for sampling as well as time in the field, and
more consistency in the sampling approach which will probably translate into less variability between sampling
results. The use of dedicated equipment is strongly recommended at wells which will undergo routine sampling
over time, as will be the case with monitoring of permeable reactive barriers. The use of low-flow sampling
techniques will also allow the placement of compliance wells closer to the treatment wall system, since the reduced
purge volume will decrease the chance of mixing waters from unintended sources and zones.
Monitoring well sampling should always proceed from the well that is expected to be least contaminated to the
well that is expected to be most contaminated, to minimize the potential for cross-contamination of samples that
may result from inadequate decontamination of sampling equipment. Samples should be collected and container-
ized according to the volatility of the target analytes. The preferred collection order for some of the more common
ground-water analytes is as follows (Barcelona et al., 1985):
1. Volatile organics (VOAs or VOCs);
2. Dissolved gases and total organic carbon (TOC);
3. Semivolatile organics (SMVs or SVOCs);
4. Metals and cyanide;
5. Major water quality cations and anions;
6. Radionuclidcs.
7.2.4 Compliance Sampling Frequency
The frequency with which compliance samples are collected should be based on the hydrogeologic character of
the aquifer, the proximity of sensitive receptors such as water supply wells, and the risk posed by the contaminant(s).
In general, if ground-water velocities are low. then samples can be collected less frequently than when velocities are
high. As a general guideline, quarterly monitoring seems appropriate, at least initially, for most sites, with
variations permitted for exceptionally slow or fast ground-water flow situations. In many cases, when the ground-
water flow rate is very slow, semiannual sampling may be justifiable.
Due to the massive subsurface disturbance created during the construction and installation of a permeable reactive
barrier, initial sampling results are often not representative of system performance. Re-equilibration of the ground-
water flow field is required before sampling results can be properly interpreted. Based on field results thus far. this
seems to take about 1-3 months. Water level measurements should be taken within the first few days, both
35
-------
upgradient and downgradient of the PRB, and continued on a regular basis to ascertain that the system is
functioning as designed with respect to plume capture, and the establishment and maintenance of designed flow
parameters through the system and upgradient of the treatment zone. These data are relatively easy and inexpensive
to obtain and the frequency of collection can therefore be increased compared to samples requiring expensive
analysis.
The compliance sampling program should be rccvaluatcd and potentially revised 1-2 years after the installation
and initial data review. The sampling frequency may be reduced if quarterly sampling shows consistent results and
other performance parameters indicate that the system continues to perform as designed.
7.2.5 Contaminant Breakthrough/Bypass and Formation of Undesirable Products
The primary objective of the compliance sampling program is to determine whether the treatment wall is meeting
design goals for remediating the contaminated ground water. The presence of contaminants which exceed target
cleanup goals in downgradient water samples is the first compliance measure most people, particularly regulators,
will examine. It is equally important to examine other data with regard to meeting compliance goals. For example,
hydraulic head data is arguably of equal importance to contaminant concentration data for compliance purposes.
Changes in water flow patterns may lead to short circuiting of the PRB by the plume. The plume could dip below
or move around the barrier due to changes in hydraulic head which could drastically alter previous flow paths.
Underflow is a concern not only for hanging wall configurations (Figure 1), but also for designs keyed into low-
permeability layers. Also, it should not be assumed that "impermeable" layers are perfectly flat, continuous, or even
impermeable. There is typically never enough site characterization data to provide this assurance, so careful
compliance monitoring is important.
It is also important to thoroughly understand the reactions which cause the transformation and destruction/
immobilization of the contaminants to be able to monitor for undesirable degradation or transformation products as
part of the compliance sampling program. As was shown earlier for trichlorocthylcnc (TCE) and Fc(0), there are
two potential pathways for the reductive dechlorination to ethene and ethane to occur: sequential hydrogenolysis,
and reductive 6- elimination, each leading to ethene and ethane as final products (Roberts et al.. 1996). Possible
intermediate products are cis- 1.2-DCE and vinyl chloride (VC). Monitoring for these intermediate constituents
should be included as part of the compliance sampling program since they are also regulated toxic compounds.
In the case of chromium and Fe(0), the reaction product is an insoluble hydroxide mineral phase. This can only be
confirmed using advanced surface analytical techniques, but can be inferred from non-detection of Cr(VI) and
minute or non-detection of Cr(I'II) in aqueous samples together with ground-water quality data and geochemical
modeling.
Evaluating trends in the data over time is important since this technology is new and many questions remain
concerning long-term in situ performance, especially with regard to maintenance of reactivity and system
permeability.
7.2.6 General Water Quality Parameters
Water quality parameters such as pH. specific conductance, alkalinity, major cations and anions and dissolved
oxygen are routinely collected as part of many site assessments and remedial actions. These same data should be
collected as part of the compliance sampling program for permeable reactive barriers. Also, parameters relevant to
the reactive material used in the PRB should be considered for analysis. For example, when zero-valent iron is the
reactive material, measurements such as redox (Eh) should be made and samples analyzed for dissolved ferrous
iron. All of these parameters arc important as either indicators of the corrosion process for iron-mediated
transformation reactions, or for potentially undesirable secondary effects in downgradient water quality. Most of
the effects due to the emplacement of a reactive iron barrier should disappear within a few meters downgradient of
the reaction zone, due to the buffering of subsurface systems, but changes in downgradient water quality should be
followed as part of the routine compliance monitoring program for this new and innovative remedial technology.
Physically or chemically unstable analytes should be measured in the field, rather than in the laboratory. Examples
of unstable parameters include pH, redox potential, dissolved oxygen, alkalinity, temperature and dissolved iron.
Dissolved oxygen (and other analyses that would be expected to vary significantly with dissolved oxygen or carbon
dioxide concentration), turbidity, and specific conductance should be determined in the field as soon as practicable
after purging. Most conductivity instalments require temperature compensation; therefore, the temperature of the
samples should be measured at the time conductivity is determined unless the monitoring equipment automatically
makes this compensation. Temperature data should also be collected as near the wellhead as possible. This reduces
travel distance through the tubing and minimizes changes due to the atmospheric temperature.
36
-------
Three methods generally are employed for measuring unstable field parameters: use of an in-line flow cell,
collecting discrete samples and analyzing them at land surface, and using analytical equipment with probes that can
be lowered into the well. The preferred method is use of an in-line flow cell. Although some investigators have
experienced freezing of tubing in very cold weather, this method provides results that typically are more precise
than those obtained using down-hole probes or collecting discrete samples and analyzing them at land surface.
Analyzing discrete samples at the land surface involves collecting a sample in a clean bottle or beaker in the same
manner that a sample for laboratory analysis would be collected, and then analyzing the sample using a field test kit
or meter. Use of down-hole probes typically requires that investigators induce ground-water movement past the
probe, either by physically moving the probe (potentially creating turbidity and potentially mixing casing water
with screened interval water), or pumping (potentially producing inconsistent results from well to well). If down-
hole probes (e.g., pH electrode, thermistor) are used to measure unstable parameters, the probes should be
decontaminated in a manner that prevents them from contaminating the water in the well. In no case should field
analyses be performed directly on samples that will be submitted for laboratory analysis.
7.3 Performance Monitoring
7.3.1 Objectives
Performance monitoring is focused on the PRB system itself (including impermeable funnels, if present), rather
than the entire site or the compliance boundaries. Effective performance monitoring begins with adequate site
characterization, to provide a baseline for later comparison, and its objective is to evaluate PRB performance
relative to design. Performance monitoring of PRBs includes the evaluation of physical, chemical and mineralogic
parameters overtime. It should address verification of emplacement and be able to detect loss of reactivity, decrease
in permeability, decrease in contaminant residence time in the reaction zone, and short circuiting or leakage in the
funnel walls. In addition to monitoring the contaminants of concern and general water quality, the following are
also recommended: contaminant degradation products, precipitates, hydrologic parameters and gcochcmical
indicator parameters. Understanding the mechanisms controlling contaminant transformation, destruction or
immobilization within the reaction zone is critical to interpretation of performance monitoring data as these data
provide insight into barrier functionality.
7.3.2 Performance Monitoring System Design
Sample points to evaluate system performance are located within or immediately adjacent to the PRB system. Due
to the relatively small zones within the aquifer where the reactive material is emplaced. these sampling points and
the sampling methods can be somewhat different from the more traditional approaches employed for compliance
monitoring purposes. Also, the volume of aquifer targeted for performance sample collection is usually signifi-
cantly smaller than that targeted for compliance monitoring purposes.
Monitoring wells can be inserted into the reactive barrier zone easily and precisely if a shallow system is being
constructed in a funnel-and-gate design. In some funnel-and-gate installations this has been accomplished by
suspending the wells in a metal framework which is subsequently removed during backfilling of the treatment
material. In a continuous wall design, where a trencher continuously excavates aquifer sediment and backfills with
the reactive material in one step, this is not an option. Monitoring points in such designs must be installed later. This
can be done using small drill rigs or by hydraulic push to place the well points. Precise placement of well points
within the barrier can be problematic, particularly with increasing depth and with thin PRBs. Small vertical
deviations can quickly result in screen placement into unintended zones. For example, in a 20 inch wide PRB where
a well is positioned in the center of the barrier, a 3% deviation from vertical will cause the well point to exit the
reactive media only 16 feet below ground surface. There are, however, rather simple and effective means to verify
proper placement. Certain geochemical indicator parameters can confirm emplacement within zero-valent iron
barriers. The use of a simple conductivity probe (commercially available for use in hydraulic push-type soil probes)
can also verify iron placement because conductivity readings between the iron and the aquifer sediment can differ
by as much as two orders of magnitude.
Smaller diameter wells (e.g., 3/4 inch i.d.) or even bundled tubes with short screens (< 1 ft screens) are often
preferred for performance monitoring purposes (Figure 10). Proper location of the screened intervals in three
dimensions is important for monitoring different vertical zones within the PRB as well as to provide coverage for
different flow paths and allow estimation of degradation or transformation rates as the contaminants pass through
the system. Similarly, these same monitoring points can be used for tracer tests to evaluate changes in permeability.
The placement of monitoring points for flow path analysis requires prior detailed site characterization to determine
flow path direction, flow velocities and system heterogeneities.
37
-------
Distance Below
Ground Surface
(meters)
1 -
t
3-
4 -
5 -
6 -
7 -
8 -I
2" PVC well with
10' screen
-
;
-I
1/2" PVC wells with
4" screens
Figu re 10. Bundled tubes with short screens for performance monitoring.
7.3.3 Performance Sampling Methods
Monitoring of permeable reactive barriers presents many unique challenges. Traditional sampling approaches are
often inappropriate because the withdrawal of large water volumes might compromise PRB sampling objectives.
Data is often sought to confirm not only contaminant reduction or removal but to provide data on contaminant
transformation or degradation rates and changes in the wall reactivity. These rates and changes might need to be
addressed over relatively small spatial intervals. To do this accurately, water from much smaller volumes of the
aquifer or from within the PRB must be withdrawn.
The use of passive or semi-passive sampling approaches can provide the required samples for these objectives
(Powell and Puls 1997c). Several researchers have demonstrated the effectiveness of using discrete-level sampling
approaches (Robin and Gillham, 1987: Powell and Puls, 1993). Puls and Paul (1995) have shown the effectiveness
of a discrete multi-level sampling device which can be used in traditional monitoring wells to provide no-purge
samples representative of formation water. However, this work addressed only inorganic contaminants and
inorganic constituents in ground water. Researchers at'NRMRL-EPA are currently evaluating the use of this device
for chlorinated organics and to understand the effects that highly reducing environments might have on sampling
results.
7.3.4 Performance Sampling Frequency
The frequency for performance sampling will be dictated by site-specific hydrogeochemical conditions, system
design, and performance sampling objectives, which should be specified prior to system installation. These
objectives should be discussed among the responsible parties, the lead regulatory agency and those who will be
tasked with implementing the sampling program and maintaining the remedial system. At current installations the
scope of performance sampling activities ranges from almost no monitoring to extensive and detailed performance
monitoring. Many of the heavily monitored installations are being used as research sites to resolve questions
concerning effective implementation of the technology and long-term performance issues. As these questions are
answered and issues are resolved, it is anticipated that subsequent performance sampling events will be greatly
reduced over what is currently practiced. However, there will always be a need for some level of performance
sampling (or at least the ability to perform such sampling on a contingency basis, should unexpected compliance
results occur) to insure that the system continues to operate as designed, particularly with respect to reactivity and
permeability.
38
-------
7.3.5 Contaminant Degradation/Transformation
As with compliance monitoring, it is important to evaluate or confirm that the desired degradation or transforma-
tion of target contaminant species is occurring as the plume moves through the reaction zone. In addition to those
degradation products which may also be regulated contaminants, it may be desirable to analyze for non-contaminant
species which provide assurance of contaminant transformation.
7.3.6 Geochemical Indicator Parameters
Geochemical indicator parameters which provide some measure of system performance will vary based on the
reactive barrier material and the contaminants present. For a reactive iron barrier these will include: pH, Eh,
alkalinity, dissolved oxygen, total dissolved sulfide. ferrous iron and. potentially, dissolved hydrogen. These
parameters can indicate that iron corrosion is proceeding and provide some indication of the extent of precipitate
formation within the barrier that may eventually decrease wall performance overtime. When included with general
water quality monitoring and used in conjunction with geochemical modeling, these geochemical parameters can
support modeling projections concerning potential precipitate formation.
For example, iron corrosion in subsurface PRB systems causes an increase in pH and can generate free Fc2+ and H2
(Equations 4 and 7). Monitoring for increases in pH, the appearance of ferrous iron and increases in dissolved
hydrogen can confirm that corrosion reactions are occurring. The redox status of the aqueous environment should
decline significantly as the corrosion process proceeds. Levels to less than -400 mv versus the standard hydrogen
electrode (S.H.E) have been observed in some installations.
Similarly, when chromate reduction by Fe(0) is occurring the key indicator parameters are pH increase and ferrous
iron formation, but also include the loss of dissolved Cr from solution accompanied by the formation of a mixed
Fe-Cr (oxy)hydroxide mineral phase solid solution (Equations 11 and 12). As pH increases, a shift occurs in
dissolved carbonate equilibria with decreasing concentrations of carbonic acid and bicarbonate species in solution
and increasing concentrations of carbonate ions. The potential for precipitation of calcite and siderite minerals can
be evaluated by monitoring for changes in alkalinity, ferrous iron and calcium. By knowing which reactions govern
contaminant transformation for a given system of contaminant and reactant, the monitoring program can be tailored
to confirm that the system is performing as designed.
7.3.7 Coring for Precipitate Buildup, Microbial Effects
Laboratory column and field tests of Fe(0) have indicated the formation of iron (oxy)hydroxides, iron sulfide
minerals, calcite and siderite as potential mineral phases which might impact iron reactivity and cause a decline in
permeability within an iron treatment zone. More study is needed to evaluate the occurrence, rate of formation, and
importance of these and other potential precipitates in existing emplaced iron systems. For PRBs comprised of other
reactive materials, even less is known and more research is needed. Techniques that can be used to assess
precipitate formation include scanning electron microscopy with energy dispersive x-ray analysis, auger spectros-
copy, x-ray photoelectron spectroscopy, laser raman spectroscopy and conventional wet chemical extractions. Over
time there will be less need for the collection of such core data, as the rate of precipitate formation and buildup
become more predictable based on aqueous and solid phase site geochemistry and kinetics.
M'icrobial activity may also be important in terms of PRB performance, especially for systems dependent upon
biologically-mediated contaminant remediation. Microbial characterization can utilize epifluorescence microscopy
and scanning electron microscopy for microbial identification and enumeration along with fatty acid methyl ester
(FAME) analyses to characterize the bacterial populations upgradient, downgradient and within the wall. In an iron
PRB, the presence of a large reservoir of iron under conditions of suitable pH and substrate availability may
promote the activity of iron and sulfate reducing bacteria and methanogens. Enhanced activity could influence zero-
valent iron reductive dehalogenation reactions through favorable impacts on redox potential, the iron surfaces, or
through direct microbial transformations of the target compounds. However, this activity enhancement may come at
the expense of biofouling of the permeable treatment zone. Analysis of these effects under field conditions is under
way.
7.3.8 Hydrologic Testing for Permeability Alteration
Hydrologic changes should also be closely monitored. Head measurements, tracer tests and in-situ flowmeters can
be used to monitor changes in system permeability' and alteration of flow paths overtime. Flow heterogeneities of
the natural subsurface system should be evaluated as part of site characterization and serve as the baseline for
comparison to post-barrier installation. Significant changes in upgradient flow patterns or head measurements can
result in hydrologic conditions which deviate from assumed design parameters and perhaps affect contaminant
residence time in the reactive material or cause bypass around the reaction zone. Tracer tests performed in locations
where prior tests have already been performed can be very helpful in this regard.
39
-------
8.0
Field
8.1 Chlorinated Hydrocarbon Chromium Removal in Field Scale Systems
Six of the full-scale PRBs installed to date for ground-water remediation are described below. These six
installations occurred from December 1994 through October 1996. From October 1997 to January 1998. six more
full-scale installations have occurred. The configurations of the first six systems are shown in Figures 11 to 16, and
construction details for these six are provided in Table 7. Summary information on these and other installations can
be found on the internet at http://www.rtdf.org.
8.1.1 Industrial Site, Sunnyvale, California (January 1995)
In January of 1995. after being approved by the California Regional Water Quality Control Board, an iron PRB
was installed at an industrial site in Sunnyvale. California (Szerdy et al., 1996). This in-situ system replaced an
existing pump-and-treat system which was being maintained at a significant cost. The capital cost for the in-situ
system, including the slurry walls used to direct ground water toward the permeable reactive wall, was $770.000.
The reactive wall was 4 feet wide. 36 feet long and 20 feet deep (Figure 11). Since installation, VOC concentrations
have been reported as non-detect from monitoring wells located within the iron wall.
The original pump-and-treat system at the site has been removed and the property has been restored to full
economic use. The monitoring wells provide easy access to the in-situ system for periodic monitoring compliance.
8.1.2 Industrial Site, Mountainview, California (September 1995)
This continuous PRB was installed as a contingency measure for the remediation of residual cis 1,2-dichloroethene
(cDCE) and vinyl chloride (VC) contamination found below the water table at an industrial facility in Mountainview,
California. The water table at the site was 10 feet below ground surface.
A source area containing VOCs was dewatered and excavated to a depth of about 25 feet. As part of the backfilling
procedure, a 44 foot long and 5 foot high zone of iron was placed at the downgradient base of the excavation
(Figure 12). The iron zone was installed in order to treat any residual VOCs which were not removed during the
excavation phase.
A low permeability HDPE liner was used to direct ground-water flow from the upgradient source area through the
iron treatment zone. The HDPE liner was placed on top of the iron zone, and extended up to the high water table
elevation on the upgradient side of the treatment zone. This ensured that any ground water contacting the HDPE
Permeable
Treatment Wall
Soil Cement-Bentonite Slurry Wall
250'
40'
Pea Gravel
Groundwater Flow
T
Cement-Bentonite
Slurry Wall
Building
220'
Historical Range in
Groundwater Flow Direction
Figure 11. PRB configuration, Sunnyvale, California.
40
-------
Table 7. Construction Details for Six Full-Scale Systems
Site
Installation
Date
Contamina-
nt & high
cone.
Design
Reactive
Wall Type
Funnel
Material
Funnel
Length
No.of
Gates
Reactive
Material
Reactive
Zone
Height
Reactive
Zone
Length
(4 x 40 =
160)
Total Mass
of Reactant
Treatment
Wall Depth
Total
System
Length
Special
Features &
Misc.
Cost
Industrial
facility.
Sunnyvale,
Califonnia
Jan 1995
0.5 mgL-1 VC
1.4 nig L-1
cDCE
0.2 mg L-1 TCE
Funnel & Gate
SoI-Bentonite
Slurry
220 ft + 250 ft
1
Fe°
10ft
36 .a
4ft
220 tons
20 ft
506 ft
$770 K
Industrial
facility,
Mountainvie-
w. California
Sept. 1995
2 mgL-1 cDCE
Excavate & fin
Not Applicable
Not Applicable
Not Applicable
Fe°
5ft
44 ft
4.5ft
90 tons
15 to 20 ft bgs
44 ft
HOPE atop Fe
to surface
upgradient
directs H2O
through Fe
$100 K
Industrial
facility,
Belfast,
Northern
Ireland
Dec. 1995
300 mgL-1
TCE
Reaction Vessel
Slurry Walls
100 ft + 100 ft
1 Reaction
vessel
Fe°
16 ft in vessel
NA
16 ft in vessel
15 tons
18to40ftbgs
Approx. 200 ft
WaHs direct
H2O to vessel
inlet gravity
flow to outlet
downgradient
$375 K
Industrial
facility,
Coffeyville,
Kansas
Jan 1996
400 gL-' TCE
Funnel & Gate
SoI-Bentonite
Slum7
490 ft + 490 ft
1
Fe°
lift
20 a
3ft
70 tons
17 to28ftbgs
1000 ft
$400 K
USCG facility,
Elizabeth City,
North
Carolina
June 1995
lOmgL-1
TCE
10 mgL.-1
Cr(Vf)
Continuous
Trench
Not Applicable
Not Applicable
Not Applicable
Fe°
Approx. 23 ft
150 ft
2ft
450 tons
3 to 26 ft bgs
150 ft
Two contain.
treated. Chain
trencher with
immediate Fe
placement
$500 K
Government
facility,
Lake wood,
Colorado
Oct. 1996
700jig L-1 each
TCE & DCE
15 ugL'1
VC
Funnel &
Multiple Gate
Scalable Joint
Sheet Pilings
1040 ft total
4
Fe°
10-15 ft
40 ft each
(4 x 40 = 160)
Gates differed.
low = 2 ft
high=6ft
No Information
10-15 to
20-25 ft bgs
1200 I
Largest of its
kind. Gates
installed using
sheet pile box.
$1000 K
41
-------
Plan
Section
Groundwater
Flow
O",
-^- Monitoring Well
® New Extraction Well
O Old Extraction Well
Area of Residual Soil
Excavation Below Water Table
Clean Native Soil
—— ground surface
HOPE
Liner
(20 mil)
Finer
Grain
Soil
."."."."."."."."."
'.*.'.*.'.*.'.*.'.
Filter
Fabric
Gravel
-— 5 feet bgs
10feetbgs
(low water table)
25 feet bgs
Figure 12. PRB configuration, Mountainvicw, California.
42
-------
would be forced down and through the iron treatment zone. The entire excavation, including the iron treatment
zone, was covered with clean backfill material.
The estimated construction cost of the iron treatment zone was $100.000 including $60.000 for 90 tons of iron.
The 4.5 foot thick iron zone was designed to treat up to 2 ppm of c'DCE and 100 ppb VC.
Monitoring at the site has been difficult because fluctuating water levels and pumpage in the area have resulted in
uncertain flow patterns. It is clear, however, that VOC levels in the iron are significantly below those in the
upgradient source area.
8.1.3 Industrial Site, Belfast, Northern Ireland (December 1995)
A circular in-situ reaction vessel was installed to a depth of about 40 feet at an industrial facility in Belfast to treat
up to 300 ppm of trichloroethene (TCE) and related breakdown products. Two 100 foot long slurry walls direct
water to the inlet of the steel reaction vessel, which is 4 feet in diameter and contains a 16 foot vertical thickness of
iron (Figure 13). Ground water flows by gravity through the iron zone and discharges via a piped outlet on the
downgradient side of the slurry wall. The vessel is equipped with a manhole to access the top of the iron zone,
should periodic scarification of the iron surface prove necessary. The system was designed to provide about 5 days
of residence time.
The total cost of the system, including slurry walls, granular iron, reaction vessel, and engineering was about
$375,000 U.S.
The system was designed to meet ground-water quality criteria of 500 ppb for TCE. These criteria apply to ground
water beneath industrial land slated for redevelopment. Flow rates through the reactor have varied substantially
since its installation, but the TCE levels in the system have decreased to 10s of ppb in the effluent sample ports.
8.1.4 Industrial Site, Kansas (January 1996)
This treatment system was installed at the property boundary of an industrial site in Kansas in order to treat ground
water containing up to 400 ppb of trichloroethene (TCE) and 100 ppb of 1.1,1-trichloroethane (TCA).
The system uses a fiinnel-and-gate configuration to direct ground water through a single, 20 foot long, 3 foot thick
permeable treatment gate (Focht et al., 1996). The funnel section of the system consists of two 490 feet long soil-
bentonite slurry walls on either side of the treatment gate (Figure 14). A low ground-water flow velocity of
0.2 ft/day permitted the use of this relatively high funnel-to-gate ratio. The system is installed to a depth of 30 feet
in a basal alluvial aquifer. The treatment gate contains 70 tons of granular iron. The installation cost for the system,
including slurry walls, treatment gate and granular iron, was approximately $400,000.
No determinations of ground-water velocity through the system have been made to date. Concentrations in the iron
zone are below MCLs.
8.1.5 USCG Facility, City, North Carolina (June
A full-scale demonstration of a permeable reactive barrier (Figure 15) to remediate ground water contaminated
with both chromate and chlorinated organic compounds was initiated at the USCG site by researchers from the U.S.
EPA National Risk Management Research Laboratory (NRMRL) and the University of Waterloo in 1995. A
continuous PRB composed of zero-valent iron was installed in June. 1996 using a trencher that was capable of
installing the granular iron to a depth of 24 feet. The continuous trenching equipment used for the installation was
similar to a large "'Ditch Witch" (Figure 9).
This was the first application of the continuous PRB configuration in a full-scale system to treat a combined
inorganic and organic plume. It was designed to meet cleanup goal concentrations of 0.05 mg/L Cr(VI). 5 ug/L
TCE. 70 ug/L DCE and 2 ug/L VC. The trenched PRB was approximately 2 feet thick. 24 feet deep, and about
150 feet long. The PRB begins about 3 feet below the ground surface and consists of about 450 tons of granular
iron. The total installation cost was $500.000. with the cost of iron representing approximately 35% of the total. It
was installed with the trencher in less than 8 hours.
In addition to the compliance wells, the PRB is monitored using a series of multilevel sampling devices to monitor
the geochemical mechanisms occurring in the barrier, and in the downgradient aquifer. To date there have been
2 years of post-installation performance monitoring performed. For all but one quarterly sampling event, 15 multi-
level samplers (7 to 11 sample ports per sampler) and 9 to 10 compliance (2 in. PVC) wells have been sampled. In
addition to on-site sampling of the full suite of geochemical indicator parameters listed in the site work plan,
samples have been collected for laboratory analysis of the following constituents: TCE, cDCE, vinyl chloride,
ethane, cthcnc. methane, major anions, and metals. In addition, numerous vertical and angle cores have been
43
-------
Figure 13. PRB vessel configuration, Belfast, Northern Ireland.
iron zone
thickness = 3ft
slurry wall -
-pea gravel
suspected
source area
ground water
flow direction
Figure 14* PRB configuration, Kansas.
44
-------
Pasquotank River
Figure 15. PRB configuration, USCG facility. Elizabeth City, North Carolina.
collected to examine changes to the iron surface overtime and to evaluate the formation of secondary precipitates
which may affect PRB performance over time. Coring was done vertically (perpendicular to ground surface) and on
an angle (30°). The former method provided continuous vertical iron cores, while the latter provided a transverse
core through the PRB with the aquifer-iron interfaces intact (front and back of the PRB). These cores continue to be
under study. Inorganic carbon contents, in the form of carbonate minerals, increase dramatically at the upgradient
aquifer sediment-iron interface and decrease within the PRB, reaching background levels within 4 inches
downgradient from the upgradient iron-aquifer sediment interface. Total inorganic carbon content has increased
over time within the PRB.
Results of geochemical sampling on site indicate that iron corrosion is proceeding within the PRB. There are
significant reductions in Eh (to < -400 mv), increase in pH (to > 10), absence of DO, and decrease in alkalinity.
Downgradient of the PRB (5 ft), pH returns to near neutral and Eh is quite variable with depth. Over time there have
been indications that a redox front is slowly migrating downgradient within the first few meters from the PRB.
Water levels indicate little difference (< 0.3 ft) between wells completed and screened at similar depths upgradient
and downgradient of the PRB, indicating that it continues to effectively function as a "permeable" reactive barrier.
Sampling results for chromatc indicate that all chromatc was removed from the ground water within the first
6 inches of the PRB as expected. No chromate is detected downgradient of the PRB either in the multi layer
samplers or in the 2 inches compliance wells located immediately behind the PRB.
The vast majority of the multilayer sampling ports show reduction of the chlorinated compound concentrations to
less than regulatory target levels. Only one port (ML25-1) continues to show levels above target concentrations.
This is the deepest port in the middle of the wall where the solvent concentrations are highest.
45
-------
8.1.6 FHA Facility, Lakewood, Colorado (October 1996)
The first funnel-and multiple-gate system using granular iron was installed to a depth of about 25 feet below
ground level at a site in Lakewood, Colorado. The system comprises about 1,040 feet of funnel section (scalable
joint sheet piling) and four reactive gate sections, each 40 feet wide (Figure 16). The gates were constructed using
a sheet pile "box." Native material was excavated from the box and the reactive material installed, separated from
the aquifer materials by a layer of pea gravel. Capital cost of the system (iron plus construction) was about
$1,000,000. A high degree of lateral geologic heterogeneity and variation in VOC concentrations exists in the
unconsolidated sediments which led to varying iron thicknesses being placed in each gate. Ground water velocities
through the gates were expected to range from 1 ft/day to 10 ft/day, depending upon the hydrogeologic conditions
in the vicinity of the respective gates. Measurements in the PRB gates using a heat-pulse flowmeter have ranged
from < 0.1 ft/day to about 1.5 ft/day (McMahon, 1997).
Design concentrations include up to 290 ppb of trichlorocthanc (TCA). 700 ppb of trichlorocthcnc (TCE). 700 ppb
of cis 1,2-dichloroethene (cDCE), and up to 15 ppb of vinyl chloride (VC). Half-lives of about 1 hour or less were
measured for these compounds in bench-scale design studies. The only VOC exiting the gates above the 5 ppb
reporting level is 1,1-dichloroe thane, which has been measured at up to 8 ppb on the downgradient side of the gates.
There is some evidence of the precipitation of calcitc and sideritc in the gates based on decreases in calcium and
inorganic carbon in the treated ground water. This is estimated to result in a potential porosity loss of 0.5 percent of
the porosity per year of operation.
Reservoir
" GSA-36-^-
GSA-31-f-
GATE 3,
flow
direction
GSA-21 +
GATE 1.
each gate section 40ft in length
— sheet pile funnel
— gate section
+ monitoring well data used in design
200 400r
Figure 16. PRB configuration, FHA facility, Denver, Colorado.
46
-------
Hydraulic head has increased upgradient of the funnel-and-gate system, with up to 10 feet of head difference
measured across the barrier. This increases the possibility for contaminated water to move around the barrier.
Indeed. VOC concentrations are increasing in ground waters moving around the south end of the barrier and there
is some evidence of VOC moving under the barrier in one location.
8.2 Interpretation of VOC Monitoring Data from the Field-scale Systems
Although VOC concentrations from the six full-scale systems indicate compliance with regulator}? criteria, the
monitoring networks installed at most of these facilities do not generate sufficient field data to permit accurate
evaluation of field VOC degradation rates. Pilot-scale systems at Lowry AFB and NAS Moffett Field (installed
under the auspices of AFCEE and the U.S. Navy, respectively), at CFB Borden in Ontario, and at a private facility
in New York state are much better suited to this purpose. Monitoring wells at these sites are located at various
distances in the PRB, as well as on the upgradient and downgradient sides (Figure 17). This allows multiple point
VOC concentrations vs. distance profiles to be obtained, which can be used to calculate field degradation rates.
However, there is significant uncertainty involved in these calculated rates as a consequence of uncertainty in the
ground-water velocity through these systems. Three methods have been attempted to measure ground-water
velocity:
* calculation using water level elevation data and Darcy's equation
* use of a conservative tracer
• use of an in-well heat pulse flow meter
Tracer tests, with bromide as a conservative tracer, may be very useful but the tests are time consuming.
Calculations using water table elevations are limited by the accuracy of measurement (small gradients over short
distances across the PRB) and the uncertainty in hydraulic parameters (porosity and hydraulic conductivity). The
heat-pulse velocity meter has given magnitudes that are in the ranges anticipated but the directional vectors were,
in some cases, suspect. Because of the ease of use. in-well velocity probes such as the K-V meter, or other in-situ
probes (Ballard, 1996), appear to offer the greatest promise for velocity determinations in PRBs.
A second uncertainty in determining field reaction rates is a consequence of the rapid disappearance of the VOCs
at several sites. This leads to the detection limit being used as the reported concentration at the first sample point in
the reactive zone, with subsequent calculation of a degradation rate from a two point curve (i.e., the influent
concentration and the detection limit). The use of detection limits in a two point curve causes an artificially long
half-life to be calculated since, in reality, the detection limit would be reached at some unknown distance upgradient
of the sampling point.
Degradation rate data from three pilot-scale PRBs are shown in Table 8. These data, when compared to the half-
lives predicted from the results of bench-scale studies, compare reasonably well given the above uncertainties.
8.3 Inorganic Constituent Removal in Field-scale Systems
8.3.1 Nickel Mine Site, Sudbury, Ontario (August 1995)
A full-scale continuous PRB was installed in an aquifer downgradient from an inactive mine tailings impound-
ment at the Nickel Rim mine site, Sudbury, Ontario, in August 1995. The PRB was installed by a cut-and-fill
installation technique during which the reactive material was installed within a valley confined by bedrock. The
PRB dimensions are 50 feet long, 14 feet deep and 12 feet wide. The PRB is composed of a reactive mixture
containing municipal compost leaf compost, and wood chips to promote bacterial sulfatc reduction and metal
sulfide precipitation reactions. These organic materials were mixed with pea gravel to attain a permeability greater
than that of the aquifer. Three-foot wide buffer zones containing coarse sand were installed on the ^gradient and
downgradient sides of the reactive material.
After passing through the PRB, water quality shows a significant improvement (Bcnner et al., 1997). Concentra-
tions of sulfate decrease from 2400 - 3800 mg/L to 110 - 1900 mg/L. Concentrations of Fe decrease from 740 - 1000
mg/L to < 1 - 91 mg/L. Alkalinity values increase from 60 - 220 mg/L as CaCO3 to 850 - 2700 mg/L as CaCO3. The
acid producing potential of the water entering the wall is converted to an acid consuming potential. Concentrations
of dissolved Ni up to 10 mg/L upgradient of the PRB arc decreased to < 0.1 mg/L within and downgradient of the
wall. Enumeration of sulfate-reducing bacteria indicates an abundance of these species within the wall, and elevated
numbers in the downgradient aquifer (Benner et al., 1997).
47
-------
• N4
P~] Pea Gravel
\ '/[ Iron Filings
^H Bentonite
Estimated Capture Zone (6.1m)
Monitoring Well
' Sheet Piling
• U1 *U2 • U3
10ft
• P1
15ft
| | Pea Gravel
|\ x| Iron Filings
^| Bentonite
• Monitoring Well
V . S* Sheet Piling
Estimated Capture Zone (7.3m)
1 1 t
Direction of GW flow
Figure 17. Pilot scale PRB configuration. Colorado and New York.
48
-------
Table 8. Field Degradation Rates from Pilot-Scale Systems
(A) In-Situ Installation, New York (May 1995)
voc
TCE
cDCE
VC
Predicted Half-Life3
ti/2, hr
0.4 to 1.1
1.5 to 4.0
2.0 to 6.0
Observed Half-Lifeb
ti/2, hr
<4.0
3.0 to 5.0
5.0 to 10.0
(B) Pilot Installation Moffett Field
VOC
PCE
TCE
cDCE
Predicted Half-Lifea'c
ti/2, hr
0.6 to 1.2
1.2 to 1.8
6.1 to 9.3
Observed Half-Lifeb'd
ti/2, hr
<2.5
<0.84
5.7
(C) Pilot-Scale, Lowry AFB, CO (December 1995)
VOC
TCE
cDCE
Predicted Half-Life3
ti/2, hr
0.9 to 1.3
4.4 to 6.6
Observed Half-Life3
ti/2, hr
2.1 to 4.5
2.6 to 9.3
a - rates at two velocities, temperature adjusted
b - two point curves using detection limit as second point
c - 50% iron
d-100% iron
8.3.2 Langton, Ontario On Site Wastewater Treatment (July 1993)
A funnel-and-gate system designed to remove PO43~ and NO3~ derived from a large-capacity septic system tile field
was installed at a public school in Langton, Ontario, in July 1993 (Charmichael. 1994; Baker et al., 1997). The
funnel consists of two sheet-piling walls extending 12 feet from the central gate area. The gate is 6 feet wide, 5 feet
long and approximately 3 feet deep. It contains two treatment zones, a PO 3" treatment zone 2 feet thick, and a NO3~
treatment zone 4 feet thick. The PO43~ zone contains a reactive mixture composed of 6% Fe/Ca oxide material. 9%
high-Ca limestone, and 85% local aquifer sand. Phosphate is removed by adsorption onto Fe oxides and
precipitation of Ca-PO^ phases. The NO3~ treatment zone contains organic carbon in the form of wood chips. Nitrate
is removed by bacterial denitrification.
Monitoring of the performance of the barrier system for over two years indicates that influent PO43~ concentrations
vary between 1.0 and 1.3 mg/L as P. Effluent concentrations within the treatment gate remained < 0.01 mg/L P for
the first 221 days, then increased to a steady concentration of 0.19 ± 0.04 mg/L P (Baker ct al., 1997). Nitrate
49
-------
concentrations varied from 23 to 82 mg/L as N upgradient of the gate, and remained < 1 mg/L as N within the gate
for a 292 day monitoring period (Charmichael, 1994). The very high organic carbon content of the nitrate treatment
zone resulted in the release of high concentrations of dissolved organic carbon and other constituents from this
portion of the gate. Current research focuses on optimizing the reactive mixtures for both PO43~ and NO," treatment.
8.4 Biological Effects on Field-scale PMBs
8.4. J Microbial Effects on Iron
The predominant concern expressed with respect to biological effects on iron PRBs has been the potential for loss
of permeability due to biofouling. This in turn relates to iron-oxidizing bacterial populations that have been
observed in aquifers at other sites and the related plugging of well screens and other treatment equipment. However,
geochemical conditions and bacterial populations in an in-situ PRB of reactive iron are quite different from those
encountered in normal ground-water pumping and monitoring wells. In a well screen, relatively reduced ground
water containing dissolved iron enters an oxygenated environment in the well bore, creating conditions where iron-
oxidizing bacteria can cause fouling problems. In an iron PRB, ground water becomes even more reducing as it
moves through the iron. The pH within treatment zones of granular iron is generally close to 10, discouraging high
levels of biological activity.
To date, no sliming or other visual evidence of microbial activity has been observed in over 50 laboratory-scale
treatability studies, using ground waters exhibiting a wide range of inorganic and organic chemistry. Though
encouraging, the applicability of these observations across a range of geochemical and microbiological in-situ
conditions is subject to question.
The effects of microbial activity have also been examined at the field scale. Cores of the reactive wall at the
Borden test site, collected two years after installation, showed no significant population of iron-oxidizing microbes,
and only low numbers of sulfate-reducers (Matheson and Tratnyek, 1993; Matheson, 1994). No evidence of
microbial fouling or decrease in performance was observed at the Borden site over a monitoring period of 5 years.
Phospholipid-fatty acid analysis of ground water from an above-ground test reactor at an industrial facility in
California showed no enhanced microbial populations in the reactive iron relative to background ground-water
samples. An above ground reactor has been operating since October 1994 in New Jersey with no indications of
biofouling.
The most detailed microbial ground-water sampling of an in-situ installation has been completed at a pilot-scale
system in New York state. Data on microbial biomass and composition were collected from wells in the iron zone,
in upgradient and downgradient pea gravel zones, and in upgradient and downgradient aquifer monitoring wells six
months following installation. The background microbial community in the aquifer at this site appeared to be
disrupted during construction, and then re-established itself in the new environment created by the gate. There was
no evidence of significant microbial growth in the upgradient gravel and iron zones; the microbial populations in
the upgradient. iron, and background zones appeared to be of similar size and composition. The microbial biomass
in the downgradient gravel and aquifer zones was approximately 10 times greater than the microbial biomass in the
background, upgradient gravel, and iron zones and was of different microbial composition. The difference in
microbial population size and composition can largely be explained by the changes in the geochemistry created by
the treatment wall on the downgradient side of the iron zone, notably the production of hydrogen gas from the iron
which supports the activity of many obligate anaerobic bacteria such as sulphate reducing bacteria. Although the
biomass increased on the downgradient side of the wall, it remained small relative to the microbial biomass
commonly found in surficial soils and shallow aquifers.
Analysis of cores taken from the New York site two years after installation, and cores taken from the Lowry AFB
pilot 18 months after installation, have confirmed the lack of significant microbial activity in iron PRBs. Plate
counts on core samples showed only small numbers of microbial populations. No biofilms were observed on core
samples examined by scanning electron microscopy.
In summary, assessment of microbial activity to date appears to show little effect on the performance of the
reactive zero-valent iron materials at both the laboratory and field scale.
8.4.2 Microbial Effects in PRBs for Inorganic Constituents
Reactive barriers designed to promote biologically mediated reactions exploit the growth and respiration of
anaerobic bacteria. Denitrification barriers are designed to promote biological denitrification by facultative
anaerobic bacteria, such as Pseudomonas sp. These barriers are designed to provide the nutrients required for
bacterial growth and sustained survival, with the exception of dissolved nitrate, which is provided by the plume of
contaminated ground water. Observations of denitrification systems indicate active and sustained denitrification
50
-------
over prolonged periods with no requirement for nutrient addition or mechanical modification (Robertson and
Cherry, 1995). Although no efforts have been made to quantify the bacterial populations of these systems,
continued denitrification indicates that conditions suitable for bacterial growth are sustained.
Barriers designed to promote sulfate reduction and metal precipitation as sparingly soluble precipitates require the
growth and activity of sulfate reducing bacteria. Enumeration of sulfate reducing bacteria in the full-scale reactive
barrier at the Nickel Rim site was conducted annually for the two years following barrier installation. The results of
these studies indicate an abrupt increase in these bacteria within the barrier, increasing to more than five orders of
magnitude greater than in the upgradient aquifer (S.G. Benner, personal communication). Because of changes in
enumeration methods, bacterial numbers derived from the two sampling sessions are not directly comparable.
These two enumerations, however, do show similar trends. In addition to sulfate reducing bacteria, the performance
of the reactive barrier at the Nickel Rim site is affected by the activity of iron reducing bacteria and other
heterotrophic bacteria. Enumerations of these bacteria, and estimates of bacterial activity are underway. In addition
to direct enumeration of bacterial populations and measurements of bacterial activity, the activity of sulfate
reducing bacteria in this system is indicated by measurements of the isotopic ratio of dissolved sulfate, the isolation
of hydrogen sulfide and the identification of iron sulfide precipitates. All of these parameters indicate the
occurrence and activity of sulfate reducing bacteria within the barrier.
8.5 Effects of Mineral Precipitation on Field-scale PRB Performance
Coring activity at pilot-scale PRBs in Canada. New York and Colorado have confirmed the formation of carbonate
precipitates in the upgradient portion of these systems. Aragonite, calcite, iron oxyhydroxides and magnetite have
been identified in core samples. Porosity losses in the range of 10% over the first foot of iron were measured at the
Colorado and New York sites, after 18 months and two years of operation, respectively. Given that the original
porosity of the media was on the order of 0.5, it is not expected that flow patterns have been significantly affected.
The velocity measured at the New York site after two years (immediately prior to coring), was similar to that
measured after six months.
To date, there is no evidence that precipitate formation in PRBs has adversely affected system performance. There
was no discernible change in the performance of the CFB Bordcn trial over the 5-year period of monitoring. Cores
taken after two years showed no measurable precipitate accumulation and only slight oxidation of the upgradient
iron/aquifer interface was observed after four years (O'Hannesin and Gillham, 1998). Ground-water velocity and
VOC removal rates were very similar at six months and 25 months at the New York installation. VOC removal rates
appeared unchanged after 6 months and 18 months of operation of the pilot-scale PRB at Lowry AFB in Colorado.
Removal of precipitates could represent a significant operations and maintenance (O&M) cost for PRB technol-
ogy. Though the evidence indicates that precipitates do not have a significantly adverse effect on reaction rates, they
nevertheless form and one must assume that they will eventually cause an adverse reduction in permeability. Based
on numerous laboratory evaluations of porosity loss, and fewer evaluations at field sites, it is estimated that in
highly mineralized and/or oxygenated ground water some degree of maintenance could be required as frequently as
ever\' five years in order to manage potential problems caused by precipitate formation. In less mineralized waters,
the frequency could be as low as every 10 to 15 years. The certainty in these estimates will increase as the period
of experience grows.
To date, there has been no need for rejuvenation, thus methods have not been developed or tested. Methods mat
include techniques such as hydraulic or mechanical mixing are being considered. Hydraulic mixing with, for
example, jetting equipment would provide the potential for adding chemical descalants. Though the need for
maintenance to control the effects of precipitate formation remains uncertain, as docs the most cost-effective
methods for such maintenance, periodic O&M requirements should be included in long-term cost models for the
technology.
51
-------
52
-------
A
to
A-53
-------
Appendix A. Summary of published results relevant to subsurface remediation using permeable reactive barriers
Metal
(zero valent
unless specified)
Fe, Ultrasound
(US)
Fe
Fe
Fe
Fe
Fe
Fe
Contaminant
TCE
Nifrobenzene
Nifrobenzene
and Carbonate
Nifrobenzene
cDCE, tDCE
CO,
Dithionite
Type test
Batch,
Column
Batch
Batch
Batch
Batch
Batch
Batch,
Column
Description/Conditions
US is co-applied in batch/column to remove deposits from
metal surface. Column 80:20 sand/iron with 50 meshFe,
in N2. Various concentrations of TCE used in studies.
Steady flow conditions maintained throughout the
columns with minimal channeling. HPLC pump used to
sample from columns to provide accuracy at low flow
rates and ensure minimal evaporation.
Batch experiments to investigate nitro reduction by
granular iron in mo del systems.
Adsorbed H2CO3 & HCCV drives metal dissolution by:
Fe» + 2H2C03(,d,) — » Fe^+2HC03-(Kl,)+H2(g)
Fe° + 2HC03-M<,-» Fe'+ + 2CO3^d,)+H2B
Corrosion rates decrease from carbonate precipitates.
18- 20 mesh Fluka Fe turnings, sonicated in 10%HC1,
washed with buffer to remove acidity or Cl". Anaerobic
batch in 60 mL serum bottles with 2 g dry, sieved Fe.
10 g Fisher pretreated 40-mesh filings; SA 1.0 nrVg. 0.20
g powdered pyrite (buffer), DI water. cDCE & OCE at
two Q in ZHEs, anaerobic, at 12 rpm, 22-25°C.
Aging & concentration effects of acetylene & CQ4 in
Ar-purged vials w/ 100 mL DI H2O, 5 g HC1 pretreated
Fisher Fe & 0. 10 g powder pyrite (buffer), 20° C, 6 rpm.
Fresh systems: Fe'H2 O not mixed prior to compound
exposure.
Batch and column with Hanford sediments to predict (1)
longevity of dithionite, (2) efficiency as reductantof Fe
(3) longevity and reactivity of the reduced Fe
Results
Batch: US removed inactive deposits and impurities thus
extending activity of the metal surface. US also found to cleave
Cl-H bond in TCE giving an added benefit . US probe in column
did not extend deep enough but modifications in design show
potential for application. Results w/o US in column: TCE t,,2_
260 min at 3mL/min; no less than 20 PV in the life of the
column. t1/2 = 360 min at 2 mL/min; samples taken after > 150
PV.
Nitrobenzene (disappears in few h) — > nitrosobenzene — >
aniline Reduction by surface; dissolved Fe2* & FT produced
during corrosion.
Anaerobic bicarbonate buffer, Fluka Fe, and nitrobenzene as
oxidant in batch exp. Kob, declined with increased carbonate and
extended exposure of metal to carbonate buffer. FeCQ,
aggregates formed on metal surfaces when using bicarbonate
medium but were not observed on surfaces exposed to DI water.
Rates for nitrobenzene (0.035) — > nitrosobenzene (0.034) — >
aniline 0.008 /min. Rates controlled by mass transfer to metal
surface. Precipitation of siderite on metal surface inhibited nitro
reduction.
Reductive dechlorination & sorption; Cl" 80-85% products:
ethene, ethane; more VC found in cDCE. cDCE not 1st order
reaction. Sorption is described by Freundlich isotherms (tDCE
sorption > more soluble cDCE); quasi-equil. 1.1 h.
CC14 : initial rates in fresh systems were 2 to 4 X > in aged
systems and faster at lower concentration. Acetylene initial rate
(0.1 to2.0/zmol) were> order of magnitude lower in systems
aged as little as 1 d. For both compounds, pseudo- first-order rate
constant independent of concentration in sufficiently aged
systems (3 to 7 d). Fast reacting sites eliminated within fewh due
to precipitation, sorption, corrosion.
Other than initial reaction with ferric iron, primary factor
promoting loss of dithionite in system was disproportionation via
heterogeneous catalysis at mineral surfaces.
Reference
Afiouni, G.F., et al., 212th
National A CS Mee ting,
Orlando, FL, 36:22-29 (1996)
Agrawal, A. & P.O. Tratnyek,
207th ACS Nat 'I Meeting, San
Diego, C4,pp. 492 (1994)
Agrawal, A., et al., 209th
National A CS Mee ting,
Anaheim, CA, April 2-7,
35:720(1995)
Agrawal, A., P.O. Tratnyek.
ES&T, 30(l):153(199o)
Allen-King, R.M., etal.,
Environ. Toxicol. Chem.,
16:424-429(1997)
Allen-King, R.M., etal., 213th
National ACS Meeting, Sen
Francisco, CA, 37:147-149
(1997)
Amonette, }., et al., In Situ
Remed.: Sci. Basis for Current
& Future Technol. Sym.
Battelle Press, pp. 851 (1994)
-------
Appendix A. continued...
Metal
(zero valent
unless specified)
Fe(n) in aquifer
material
Ni-Fe Wall
Al, Fe, Zn
Al, In, Cd, Bi, Sn, Ag,
Ge, Sb, Cu, Hg, Pd, Th,
Pb, Ti, Mn, Co
Al
Zn
Limestone
Zn
Zn
Contaminant
Cr(IV)
Organic s
1 1 Chlorinated
Solvents
CC14
1,1,1-TCA
TCE, PCE rates;
products
Cr
CC14
Hexachloro-
benzene (HCB)
Type test
Batch
In situ- Otis
AFB, MA
Batch
Batch
Batch
Batch
Column
Batch
Batch
Descrip tion/Con diti ons
Batch experiments using sand collected from a shallow
sand and gravel aquifer to determine reduction of Cr(VI).
Q Cr(VI) varied 2 to200,uM; Adjusted pH 4.5, 5.5, 6.5.
Note: Fine fraction Fe(II) minerals (<64 /jm dia.)
dominated Cr(VI) reduction due to greater reactivity and
SA. However, to provide consistent results fines and sands
had to be separated and parallel experiments run on each
Induce fracture to fill w/ Fe filings w/ slurry mixture
moving down and outward, creating series of overlapping
vertical planes thus becoming a "continuous" wall.
Al coupon size 2 x 15 x 0.2cm, Zn& Fe coupon 2.5 x 15
x 0. 1 cm. 65 mL solvent and 5 mL DI water in 125- mL
flask containing solvent and metal coupon.
1 mL CC14 heated at 200° C, sealed tubes, for 14 d using
twice the metal shot, powder, granules, or chips needed to
complete reaction.
5 mL inhibitor -containing solvent refluxed with 0.5 g of
16-32 mesh pure Al pellets in open reaction tubes. Tubes
in oil bath at 7° C. Upper portion extends through a water-
cooled Al blockthat acts as a condenser.
Deoxygenated water (buffered), Zn(0) PCE or TCE.
Sampled for product formation. Early heterogeneous
process, but initial rate does not increase linearly with
increasingCj (expected for pseudo-lst-order system), but
levels off as Q increased.
Limestone or sand2 cm over 10-cm depth of soil in PVC
column. Leachate passed through columns at 1 PV/24 h
until breakthrough. Unamendedleachate dilutedto 25%
and pH 4.0 or 2. 5 to keep Cr in solution. -3,000 ppm TOC
upon dilution and adjusting pH to 2.5 .
Zinc powder ± B12 under N2. Initial cone. 2.2 mM CC14.
Zinc powder ± B12 under N2. Initial concentration of 50
AiM HCB.
Results
As pH decreased (6. 4 to 4. 5), Cr( VI) reduction increased 3 0 to 5 0
nmol/m2 for the sand fraction (645 - 1000 //m) and 130 to 200
nmol/m2 in fines. Amount reduced in both sand and fines
increased from 3 5 to 80 and from 130 to 1000, respectively, for a
10- fold increase. Consistent rate descriptions achieve by
assuming that intraparticle diffusion limited the observed rate of
reduction.
Plan 2 parallel 50' walls, 2' apart, perpendicular to flow path of
600' wide plume with 5 to 150 ppb TCE & PCE. Walls will
begin 80' bgstotop of plume, extending to as deep as 150'.
Reactivity accelerated when water added. Problems with Al and
Zn, but not Fe corrosion in dry systems. 1, 1,2-trichloroethane
only structure with appreciable oxidative breakdown.
Hexachloroethane: end products perchloroethylene, hexachloro-
butadiene. Most reactive Al( 100%), Ti( 100%), Cd(74%),
antimony (58%), In (58%), Ge (47%), Sb(33%).
Inhibitors compete with solvent for A1Q3 produced at micro
corrosion sites by complexing the chemisorbed A1C13 product
Resultant complex insoluble in solvent, acting as a plug covering
original reaction site.
Reductive elimination (RE) important inZn(0). ~ 1 5% PCE — >
dichloroacetylene (0.25 — > acetylene bypass VC). TCE — >
acetylene (20% of original TCE); trace VC. Chloroacetylene
intermed. — ^ acetylene preferred over VC. Manipulating RE
over initial hydrogenolysis would be beneficial goal.
Limestone delayed breakthrough of Cr. Retained Cr(in) more
thanCr(VT). Retention» Be, Cd, Fe, Ni, Zn. TOC & Fe(II)
determine Cr(VT)/Cr(in) (Clay, Fe oxides better at retaining Cr).
pH affects solubility of Cr and limestone. When Cr(VI) & Cr(IH)
in leachate, migration delayed several-fold by limestone barrier.
Zn + B12 dechlorinated CC14 to CH, (50% recov). Rates slowed
when B12 absent [CC14 — > CHC13, DCM, CM, CH4 (80%
recov)].
Product pentachlorobenzene higher rates w/o B12 at 9.6 h"1
compared to 0.3 h"1 w/ B12. B12 may compete w/ HCB for e .
Reference
Anderson, L.D., et al., ES&T,
28(1):178(1994)
Appleton, E.L., ES&T,
30(12):536A(1996)
Archer, W . & E. S impscn,
Industrial Eng. Chern. Product
R&D, 16(2):158(1977)
Archer, W.L. & M.K. Harter,
Corrosion: Narl Assoc. of
Corr. Eng. (NACE), 3<5):159
(1978)
Archer, W.L., Industrial Erg.
Chem. Product R&D, 21:670
(1982)
ArnoH, W.A. & A.L. Roberts,
213thNationalACS Meeting.
SanFrancico, CA, 37:76-77
(1997)
Artiole, J. &W.H. Fuller,
Journal of Environmental
Quality, 8:503-510(1979)
Assaf-Anid,N.&L. Nies,
209thNat'l ACS Meeting,
April 2-7, Anaheim, CA,
35:09-811 (1995)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Metal oxide from
steel manufac.
Limestone
Fe
Zeolite, 3 media
types
Mixture organics,
sulphate-reducing
bacteria
Fe
Fe
Fe
Pyrite or Fe
Mixed organics,
bacteria
Contaminant
Phosphorus
TCE; DCE, VC,
dichloro-
methane
Sr, Cs, TCE
Acid Mine
Drainage
Cr
Solubilized PCE
Hydroxypropyl-
p-cyclodextrin
(HP-P-CD)
Cr
Cr(VI), U, etc
Nitrate - Tile
Drainage
Type test
Column,
Cylinder,
Reactive
Wall
Batch,
Column
Containers,
TN&OH
Reactive
Wall,
Ontario
Reac. Wall
Eliz. City,
NC
Batch,
Column
Batch,
Column
Trench
In-Line
Bioreactor
Description/Conditions
Permeable mix 50% sand, 45% crushed limestone, 5%
metal oxide in acrylic column w/ 3.3 mg/L PO4-P over 3
y (1250 PV). Biofilter effluent in 0.5 x 0.5 m cylinder 2
L/d over 133 d (101 PV). Funnel-&-gate in septic plume.
7 m long funnel, 1 . 8 long; 2 wide; 10 deep (m) gate for
779 d.
Site ground water (GW) from DOE Pinellas Plant. VWR
coarse iron filings used for high reactivity and low cost.
55-gal drums of Na-chabazite zeolite (remove Sr, Cs. at
seep, Oak Ridge National Lab, TN) and using 3 media
types (reduce TCE, Portsmouth Gas.Diff Plant, OH).
Wall installed at the Nickel Rim mine site near Sudbury,
Ontario on 8/95. 15 m long perpendicular to GWflow, 3.6
m deep, 4m thick. Used municipal & leaf compost, wood
chips, and pea gravel (for permeability). Sand buffer, clay
cap. Mon. wells parallel to GW flow.
Models indicate reactive barrier most efficient/cost
effective. 46 long x 0.6 wide x 7.3 deep (m) barrier
installed < 6 h using continuous trencher 6/22/96.
Batch: 100 mesh Fisher Fe powder & - 8 to 5 0 mesh Peer-
less mix 40 mL sealed vials, 10 g Fe, PCE, 40 mg/L
CaCO3; 0, 45, 70 g/L HP-P-CD. Column: 0.47 mx 5 cm,
15 cm 70/100 mesh sand below (30-cm head) 1 0 cm 30/40
mesh sand. 25 mL PCE to form pool. RT 25.3 h.
Batch: Compare 100 g siderite, pyrite, F<0) (~ 0.5
-1mm), chips (~ l-5mm). Use 500 g Cr inCaCQ, DI water
agitated at room temp. Column: 15x6.5 cm, pump tracer
base upward to determine void v & dispersivity of
column.
Excavate trench, fill with active material such as pyrite or
elemental iron to transform & precipitate contaminant.
Two 200-L fixed-bed bioreactors, with coarse sand and
organic carbon (tree bark, wood chips and leaf compost),
to treat 3-6 mg L NQj-N from farm-field drainage tile.
Results
>90% efficiency in column & cylinder. PO4-P 0 to 0.3 mg/L in
column effluent. Phosphate accumulated on oxide surfaces &
precipitated as microcrystallinehydroxyapatite. Cylinder: 3.93 to
0. 14 mg/L-P, 2.50 to 0.05 mg/L-orf/» P. F&G: 4 m above 2 to 3
mg/L; 0.4 m above gate 1.5 to 0.4 mg/L. Average phosphate in
wall~0.19mgP/L.
Batch: fast rates for TCE, DCE , VC in site GW. Dichloro-
methane rates very slow. Column: tU2 TCE = 36-103 min; DCE =
150-200 min. However, rapid plugging of iron by Pinellas GW.
>99.9% Sr, Cs removal ORNL (25% red. total radioactive
discharge). TCEreduced at PORTS X-120 site. Drums predict
flow and allow easy media replacement and monitoring.
After 9 mo sulfate reduction and metal sulfide precipitation.
Sulfate: 2400 to 4600 mg/L to 200 to 3600 mg/L, Fe:250 to 1300
mg/L to 1 and 40 mg/L; pH 5.8 to 7.0. Alkalinity rose 60-220 to
850-2700 mg/L as CaCO3. Fe mono- sulfide precipitate. Cost-
$30,000.00 (half materials/half installation) potential life 15 y.
1 1/9 sampling indicated 2. 5 mg/L Cr(VI) declined to < MCL
within barrier. Further sampling under way determining ground-
water chemistry and organic concentrations.
HP-P-CD enhances solubility w/o decreasing in terfacial tension
of PCE and water. Smaller % PCE degraded at higher HP-P-CD
cone. PCE decreased in both non-recycling and recycling of post-
treatment effluent. Greater degradation at higher iron SA. Plan to
use higher Fe SA and longer RTs to increase degradation.
Batch: Fe(0) > pyrite. Column: Fe chips (remove 50% Cr < 2 h) >
pyrite (w/ calcite 50% 0. 5 h; no calcite 50% 1 h) > coarse Fe
(50% 28 h). Chips: Coating(rf 5.2). No Cr4.5PV. Filings: NoCr
> 1 5 PV. Remain active, litfle coating.
This patent relates to the treatment of GW for the purpose of
removing water-borne contaminants.
Reduced NO3-N < 0.02 mg/L at 10-60 L/d over a 1-y periodby
anaerobic denitrification promoted organic carbon. Design is
simple, economical and maintenance free.
Reference
Baker, M.J., etal., Internat'1
Contain. Technol. Conf. &
Exhib., St. Petersburg, EL, Feb
9-12, pp. 697(1997)
Baghel, S., et al., G. E. Corp.
R&D Center fir USDOE,
Sandia(1995)
Barton, W., et al., Internat'1
Contain. Technol. Conf. &
Exhib., St. Petersburg, EL, Feb
9-12, pp. 827(1997)
Benner, S.G., etal.,
1 ACS Meeting, San Fran.,
CA, April 13- 17, 140
(1997)
2 Internat'1 Contain. Technol.
Conf. & Exhib, St.
Petersburg, FL, 835(1997)
3 GWMR, Fall, 99-107
(1997)
Bennett T.A, et al., 213th
National ACS Meeting, San
Francisco, CA, April 13-17,
pp. 243-245(1997)
Bizzigotti, G.O, et al., ES&T.
31:472-478(1997)
Blowes, D.W. & C.J. Ptacek,
Subsurface Restor. Conf, 3rd
Interna'l Conf. on Ground
Water Quality Res., June
21-24, Dallas, TX, 214-216
(1992)
Blowes, D.W. & C.J. Ptacek,
Patent 5,362,394 (1994)
Blowes, D.W., et al., Journal
of Contaminant Hydrology
15:207-221(1994)
-------
Appe nd ix A. continued...
Metal
(zero valent
unless specified)
Contaminant
Type test
Description/Conditions
Results
Reference
Fe
Cr(VI)
Batch,
Column
Batch: 100 g (50% fine Fe filings, 49% sand, l%calcite),
Cr(VI). 2ndmix 50:50 Fe, quartz. Column: 50%F<0),
50% sand, top 5 cm 1% calcite. 20 mg/L Cr(VI) in
simulated GW.
Cr(VI) 25 to <0.05 mg/L 3 h batch. Column: No breakthrough Cr
after 140 PV. Dissolved & total Cr <0.05 mg/L. Fe(III)
oxyhydroxides form but not sufficient to inhibit Cr(Vi) reduction
at experimental velocity.
Blowes, D.W, etal., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:780(1995)
Fe
Acid Mine
Drainage
Test
Reactive
Wall
Test cell 1.5 long x 1 wide x 1 deep (m) installed 10/93 in
sand aquifer ~ 75 m downgradienttailings impoundment.
Qrganics (leaf, pine mulch, bark), creek sediment (sulfate-
red. bacteria), limestone, coarse sand and gravel.
1 m on flow path, SO, 3500 to 7 mg/L, Fe 1000 to < 5 mg/L, pH
& alkalinity increased from sulfate-reduction. Sulfate-reducing
reactive walls are potentially effective and economical solution
to many acid mine drainage problems.
Blowes, D.W., et al., Mining
& Environ. Conf, CANMET.
Sudbury, Ontario, May 28 to
June 1,3:979 (1995)
Fe
Cr(VI), TCE
Fe
Reactive
Wall
Wall of 100% Fe filings 46 m long, 0.6 m wide, 7.3 m
deep installed in < 6h usinga continuous trenching
technique at Elizabeth City, NC on 6/96. Bench lab
studies, and flow and transport models used in wall
design.
Site GW, reactive materials: Cr(Vr) 11 to <0.01 mg/L and TCE
1700 /zg/L to< 1 /zg/L. Decrease of Cr(VI) from influent 6 mg/L
to < 0.01 mg/L and TCE from 5600 /zg/L to 5.3 /zg/L within
wall. TCE approaches or attains the MCL within the barrier.
Blowes, D.W., et al., Internal!
Contain. Tech. Conf. & Exhih
St. Petersburg, FL, Feb 9-12,
pp. 851 (1997)
Fe
Cr(VI)
Batch,
Column
Batch: 500 g Cr(VI) to 100 g solidmixes (siderite, pyrite,
coarse & fine Fe(0)) in open flasks, agitated, room temp.
Settle 5 min; 10 mL sampled. Column: 6-6.5 cmdia
acrylic, 5-15 cm long, 1-20 cm long with layers reactive
mix. Void volume & dispersivity determined. Cr(VI)
solution introduced.
Batch: Rate of Cr(VI): fine Fe(0) > pyrite & coarse Fe(0).
Column: partial reduction of Cr(VI) by pyrite & coarse Fe(0)'
quantitative reduction of Cr(VI) by fine Fe(0) at rapid velocities.
F<0) reduces Cr( VI) to Cr(HI) with oxidation of Fe(0) to Fe(H)
&Fe(III), and precipitation of sparingly soluble (oxy)hydroxide.
Cr(ni) forms a solid solution or adsorbs on goethite.
Blowes, D.W., etal., ES&T.
31:3348-3357(1997)
Fe
CO, or CT,
CKCLorDCM
Batch
Open circuit potential time measurements using Fe with
CT & DCM in borate buffer and simulated GW of KBr &
CaCO3. Polarization of Fe electrode in borate solution to
which 0.2 mL of CT added. CT acts as an oxidizer of Fe
electrode, while DCM does not.
Injection of CT shows faster and larger potential shifts in Fe than
in freshly cleaned Fe electrode. Potential decay in all GW
studies. Magnitude depended on pH and solution. Borate and
KBr decay mainly from chemical dissolution of films. In CaCO3,
autoreduction/chemical dissolution maybe responsible.
Bonin, P.M.L., etal., 213th
National ACS Meeting, San
Frandsco, CA, 37:86-88
(1997)
Sn, Zn & Mg
CO,
Batch
Vaporization procedure (SMAD or cyro method) to
compare metal powders.
FLO oxidation overwhelmed Mg-CCl4 reaction. Sn, Zn degrade
CCL, but differ in carbon product (Zn —> CH,; Sn —> CO2.
Intermediate C13CMC1 may be protonated by FtO to give CHC13
or eliminate CC12 which subsequently reacts with FtO to form
Ctt and HC1.
Boronina, T., et al., ES&T,
29:1511(1995)
Steel wool
Tc9!
Column
Simulated process & GW from DOE uranium enrichment
plants. Packed column of steel wool, Dowex TM 1-X8.
Use of iron economical but maybe more difficult to accurately
predict its sorptive capacity or functional "lifetime"
Bostick, W., et al., OakRidge
K-25 Site Rep. Martin
Marietta DOE K/TCD-I 141
(1995)
Fe
Cd, Mg, Ni,
TCOT, uaz+
Batch
Shake solid w/ soln 16-24 h in sealed container. Exp 2:
0.01, 0.03, or 1.0-g iron to lOmL w/ 8 mg/L U, shake 18
h, sample day 1 & 30. Exp 3: 1.43x0.15-cm iron coupons
in 500-mL bottle w/ 927-mg/L U. N2 purge.
Iron surfaces passivate at elevated pH (little activity at>9.5).
Sorption to iron corrosion products predominant removal process
for uranyl (Cd, Mn and Ni also). Sorbed products need to be
controlled. Columns needed to determine long-term capacity.
Bostick, W., et al., OakRidge
& Martin Marietta Energy
System^ Inc. for
U.S. DOE. K/TSO-35P (1996)
-------
Appendix A. continued...
Metal
(zero valent
unless specified)
Iron oxide
aggregate
Fe
Iron filings/pyrite
Fe
Fe
Fe Colloid Barrier
Fe Colloids
Barrier
Pd/C
Fe
Fe
Contaminant
U
U
TCE, PCE
PCE
UO22+, MoO42',
TcCV, CrO,2'
Various
Various
4-chlorophenol
Nitrate
PCBs (Aroclor
1221,1254)
Type test
Batch
Batch
Batch
Batch
Batch
Batch
Column
Batch
Batch
Batch
Description/Conditions
Continued from previous reference.
Batch: ~1.4x0.16cmFe coupons, simulated GW, soluble
U asuranyl nitrate. Gas glove box Pure O2 added to yield
equivalent to solution purged with lab air.
4-15 mL vials, anaerobic: 2controls/2 Fe+pyrite(5 g pre-
treated Fe filings, 0.1 g ground pyrite [buffer]).
15 mL serum vials, 5 g iron, 0. 1 g pyrite (ground). Pyrite
added to stabilize the pH to 6.5 - 7.0. Anaerobic
Kinetic studies in 50 mLredox- sensitive-metal solutions
added to polystyrene centrifuge tubes containing l-g40
mesh metallic iron with SA of 2.43 m2/g.
0. 2% Fe colloids, surfactant Polymers (vinyl (VP), vio
(GX), cellulose (CMC )) tested to increased colloidal
Fe(0) mobility in porous media. Turbidimeter meas. Fe
colloids.
1 m columns, 4.4 cm dia. 20-30 mesh sand avgr| = 0.32.
Fe(0) colloid dia = 2 /an, bulk density = 2.25 g-cm"3,
particle density = 7.6 g-cm"3. V = 0. 154, 0.307, 0.614
cm/s. 0.01MCaCl2at 0.2cm/min (~2PV) to simulate
GW.
Palladized graphite and carbon cloth electrode in 3-neck
round-bottom flask (50 mL). Cu wire threaded through 5
mm glass tubing. Central portal used for pH & cathode
potential. Anodic compartment vented for Oi escape.
4 g untreated 325 mesh iron to 50 mL of 12.5 mM nitrate
buffered at pH 5. No effort made to exclude O2.
Vigorously stirred.
Iron powder (0.5 g), 1.5 /zmol PCBs in 10 mL flame-
sealed glass ampule.
Results
Fe(0) w/ sand or pelletized Fe oxide most effective. Reduces
cementation from rust; enhances dilution of hydroxyl ion
reaction product; enhances sorption of cationic contaminant to
pelletized Fe.
Under oxic conditions, U(VI) rapidlyand strongly sorbedto
hydrous ferric oxide particulate ("rust"), whereas U slowly and
incompletely reduced to U(IV) under anoxic conditions.
Reaction orders 2.7 TCE & 1.3 PCE total system cone.
Nonlinear sorption fit generalized Langmuir isotherm. 1 st order
rates.
Rapid initial rate followed by a slower rate. Sorption of PCE to
Fe(0) follows Langmuir-type isotherm.
Particulate Fe(0) effectively removed each of the contaminants
from solution by reductive precipitation. Removal rates
decreased by CrQ.2- > Ted," > TO2* » MoO42'.
VP is superior to GX and CMC because W suspension produced
the lowest backpressure, resulting in the highest hydraulic
conductivities.
Colloidal-size Fe(0) injected into porous media forming chemical
reactive barriers. Relatively even distributions of Fe(0) in sand
column at low cone.; high injection rates. As V increased,
distribution of Fe(0) colloids became increasingly even.
Rapid dechlorination of 4-chlorophenol on palladized carbon
cloth or palladized graphite electrodes. Hypothesize Ft gas
interlaced in Pd lattice powerful reducing agent dechlorinating
compounds adsorbed on palladized electrode surfaces.
Nitrate < 0.2mM in 74 min. Ammonia 103%. Pseudo- 1st order
rate constant 0. 05 30/min. Buffer is key in nitrate reduction.
PCBs undergo dechlor. & other reactions at> 300°Cin presence
of Fe powder. Virtually complete loss of chlorinated congeners.
Reference
Bostick, W., et al ., OakRidge
& Martin Marietta Energy
Systems, Inc. for
U.S. DOE. K/TSO-35P (1996)
Bostick, W.D., et al., Internat'l
Contain. Technol. Conf. &
Exhib, St. Petersburg, FL, Feb
9-12, pp. 767(1997)
Burris, D.R., etal., ES&T,
29:2850(1995)
Campbell, T.J. & D.R. Burris.
209thNat'l ACS Meeting,
Anaheim, CA, 35:775 (1995)
Cantrell, KJ.,etal., J. ofHaz.
Mat., 42:201 (1995)
Cantrell, K.J., etal., J. of
Environ. Erg.-ASCE, 23:786
(1997)
Cantrell, K.J. & D.I. Kaplan, J.
ofEnvircn.
Engineering -ASCE
123:499-505(1997)
Cheng, I. F, etal., ES&T,
31:1074-1078(1997)
Cheng, I.F., et al., 213thACS,
San Fran, CA, 37:165 (1997)
Chuang, F.-W. & R.A. Larson,
ES&T, 29:2460(1995)
-------
Appe nd ix A. continued...
Metal
(zero valent
unless specified)
Fe
Fe
Fe(II)
Fe
Fe
Fe
Fe
Fe
Fe
Contaminant
DCE, TCE, PCE
TCE
TcCV
Background
hydrocarbon
formation
VC
U
Tracers, D2O
Alachlor,
Metalochlor
CC14
Type test
Batch
Column
Batch
Batch
Batch
Riot
barrier in
CO
Column
Batch
Kinetics
Elemental
Fe cathode
Description/Conditions
Solutions incubated with lab grade iron under static
conditions. Concentrations monitored as a function of
time. Pyrite with Feto counterbalance pH increases.
Long-term (-30 d, >300 PV) & fast velocities (5-30
cm/h) w/ 4-600 mL Fe mix columns. Rates varied by
medium, SA, mixture, time. Influent 5 0 /zM TCE at 3.15
mL/min.
Fe(II), in slightly acid to base solution, in 500-cm3 bottle
w/ hydrophobic inner surface, ambient temperature, and
under anaerobic conditions. 0. 1 FeCl2 (pH 4) added,
sampled by syringe through septum.
15. 0-mL sealed vial. 5.00 g Fe with lowCOi water,
anaerobic, 8 rpm in dark at 20° C. pH increased 5-7 when
Fe added. 13CO2to determine if COt is beingreduced by
iron to form the hydrocarbons. Note: Non-Cl-
hydrocarbonsform during ethene reduction and when
there is no ethene. Fischer-Tropsch type synthesis of
hydrocarbons, proposedfor hydrocarbon production.
40 mesh HC1 pretreated Fisher Fe filing s, SA 1 . 1 8 m2/g
(0.2 to 10 g)usingl5mLborosilicate in ZHE filled by
VC solution, anaerobic, 8 rpm, 20° C. Also at 4, 20, 32,
45 °C.
Pilot barrier using Fe to remove U from tailings effluent.
atUMTRA site scheduled 5/96 to 1999. Fe foam SA 0. 1
& 5 m2/g; Fe(0) SA 5.6E3 m2/g. During3 years U, Se,
Mo, other elements monitored as well as costs and
benefits.
Glass column with wet Ottawa sand or VWR coarse iron
filings. 2.65 g/cnf for sand and 6.5-7.6 g/cm3 for iron.
100 mL ZHEs, ~ 40 g coarse Fe filings (40 mesh, SA 13.5
nrVg), 10 mg/L or 100 mg/L alachlor and/or metalochlor,
room temp, 3 rpm. Sampled over even intervals for 5 d.
Experimental reactors using a two- part glass vessel with a
Nafion-1 1 7 proton permeable membrane.
Results
Oxidation reactions of Fe increased pH. Halocarbons convert
slower at high pH. High grade pyrite more reactive. SA important
consideration.
Fe oxidation increases pH, pyrite decreases pH. Pyrite is placed
at column head because precipitation slower at lower pH w/
wider dispersion of precipitation zone throughout column and
lessens plugging at head of column.
3 e- reduction process, although thermodynamically feasible,
slow if at all. Fe(II) [sorbed to wall or precipitate as Fe(OH)2(s)
or FeCO3(s)] reduced TcQT. Rates proportional to sorbed or
precipitated Fe(II). Discuss redox/paths TcQ," to TcO2'nF4O.
13CQ was not incorporated into hydrocarbons produced. Acid
dissolution of gray cast irons containing both carbide and
graphite carbon yielded hydrocarbons and a substantial amount
of graphite residual. The dissolution of metallic irons containing
only carbide carbon yielded total carbon conversion to
hydrocarbons. Carbide carbon in the iron most likely carbon
source for the production of the background hydrocarbons.
5.0 gFe/15.0 mL: VC— > ethylene (Partial absorb, to Fe). Rates
increase as Fe & temperature increases. Activ. E 40 KJ/mole
indicate surface reaction. Ft also produced, but, Fe2+ not directly
involved in reduction.
Field coincides with experimental findings. InFe foam batch U
removed to 700 PV oxygenated water.
2 tracers (D2O, KBF4) for Fe column. D2O more conservative and
KBF4 easier on-line detection. Both inert with respect to Fe.
Rapid dechlorination byFe(O) shown by Cl" and GC/MS
analyses. Apparent Ist-order kinetics, butindication of rate
limited and instant sorption. 2-site batch kinetic model fitted to
results.
Reduction of CQ4 — ^ chloroform in hrs at .0005 to .005 / min.
Suggest H2 serves as intermediate for CC14 hydrogenolysis.
Reference
Cipollone, M.G., et al ., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:812(1995)
Cipollone, M.G., et al ., 213th
ACS Nat'l Meeting, San Fran.,
CA, 37:151-152(1997)
Cui, D. &T.E. Erflsen,
ESiT_30:2263-2269 (1996)
Deng, B. & AT. Stone,
ES&T_30:463-472 (1995)
Deng, B., et al., 2 13th National
ACS Meeting San Frand sco,
CA. 37:81-83(1997)
Dwyer, B.P. & D.C. Marozas,
Internat'l Contain. Tech. Conf.
& Exhib. St. Petersburg, FL,
Feb9-12, pp. 844-850(1997)
EyHiolt, G., et al., 209th Nat'l
ACS Meeting Anaheim, CA,
April 2-7, 35:818(1995)
EyHiolt,G.R.&D.T.
Davenport, 213thNat'l ACS
Meeting San Francisco, CA,
37:79-81 (1997)
Festa,K.D, etal., 209thNat'l
ACS Meeting Anaheim, CA,
April2-7,35:711-715(1995)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Fe
Fe
PRB
PRB
Fe Foam
Fe,Mg,
Ultrasound (US)
Fe
Contaminant
1,2,3-trichloro-
propane
Review
TCE
TCE
TCE, cDCE, VC
CO,, TCE,
CHC13, Cr(VI),
Tc, U
Velocity
measurements
As, Se, Mo, U,
sulfate, nitrate
TCE
Halogenated
compounds
Type test
Column
Reactive
Walls
Funnel-
and-Gate
Funnel-
and-Gate
Funnel-
and-Gate
In situ
Tracer
Batch,
Column
Batch
Batch
Description/Conditions
50 & 93 mg/L 1,2,3-TCP was passed through 6 flow-
through reactive columns containing Fe(0)- silica sand &
simulated GW.
Gillham founded EnviroMetal Technologies in 1992 to
commercialize the reactive wall technology.
In Sunnyvale CA site, 1 00% pure granular iron filing wall
is 4' thick, 40' wide and 20' deep.
1000 foot funnel- and-gate system installed at industrial
facility in Kansas in January 1996 to treat 1 00 to 400 ppb
TCE. Reactive zone 30' to 17' bgsand 3' thick.
NY facility (1995) treats up to 300 TCE, 500 cDCE, 80
VC (ppb). 12' x 3.5' reactive section flanked by 15' sheet
piling on either side.
Used abiotic reagent sodium dithionite atthe Hanford site.
Biotic reagent/nutrient either citrate or glucose.
Bromide considered most appropriate tracer. Pilot system
(3 long x 3 wide x 5. 5 deep (m)) installed 1 1/95 gov.
facility in CO. Models indicate ~ 60 cm/d (2'/d) in
reactive zone.
Batch ( 1 wk) & column (70 d) using Fe foam and steel
wool to compare removal of As, Se, Mo, U, sulfate, and
nitrate. Precipitation and removal mechanisms also
studied.
TCE (20 ppmv) in 3 neck-1 L round bottom reactors, with
US probe inserted in center neck and tip just above 1 .0 g
Fe or Mg, or 50:50 mix, w/ & w/o US. Controls only US,
no metal or US.
10 g 100 mesh iron filings in 40 mL ZHE hypo vials, at 2
rpm. C0for CC14, HCE, PCE were 1630, 3620, 2250 /zg/L,
respectively.
Results
End- product propene. Fe(0) enhanced dechlorination. Rate
increased in proportion to iron SA to solution volume ratio.
Ratios of 1. 16, 3.7, 8 mVmL gave t1/2 17.6, 6.6, 3 h, respectively.
To FJate 5 full-scale in situ treatment zones installed: 2 in
commercial sites in CA, 1 in KS, 1 in N Ireland and 1 at
Elizabeth City, NC
TCElevels of 30-68 ppb entering wall reduced to <0. 5 ppb;
cDCE of 393-1916 ppb to <0.5 ppb
Under optimum conditions, the soil-bentonite slurry wall could
be built in 1 or 2 weeks and gate section in one week. Slurry
wall, gate, 7900 ton of granular iron = -$400,000
Installed in 10 d. VOC reduced to MCLs within 1.5 feet of travel
through reactive media. Cost $250,000 including $30,000 for 45
tons of iron.
Compared abiotic & biotic methods for controlling redox
potential to reduce solids in unconfined aquifer.
Small tracer not detected. Large targets could disrupt flow, time
consuming. Water table calculations are not accurate. Heat-pulse
velocity meter give suspect directional vectors. In situ velocity
probes are most promising/easy to use.
Batch Fe foam removed 100% Se, 86% nitrate, 100%U, 83%
As. Steel wool ~ 80% U, 20% As, 70% nitrate. Neither remove
sulfate or Mo. Column conductivity decreased slightly (0.08 steel
wool; 0.09 cm/sfoam). Reduction/ precip. Se; adsorption As, U.
Higher pH from dechlorination reactions increases deactivation
by precipitating metal compounds on active surface. US can strip
away corrosion keeping metals active longer . Another benefit is
sonication produces H* ion to stabilize pH.
tU2CC\v PCE, HCE = 20, 1100, 13 min., respectively.
Chloroform, produced by CCL, was the only breakdown product
accumulating insignificant quantities.
Reference
Focht, R.M & R.W. Gillham.
209thNational ACS Meeting
Anaheim CA, April 2-7,
35:741 (1995)
Focht, R.M, etal,
Remediation, Summer :81
(1996)
Fruchter, J.S., Pacific NWLab
&U. S. DOE. PNL-SA-21731
(1993)
Focht, R.M, etal., Internat'l
Contain. Tech. Corf. & Exhih
St. Petersburg, FL, Feb 9-12,
pp. 975 (1997)
GaUegos,T.J.,etal., HSRO
WERC Joint Conf on the
Environment, May 20, Paper
75(1997)
Gdger, C., et al., 21 1th ACS
Meeting New Orleans, LA,
March 24-28, 36: 17- 18 (1996)
Gillham, RW. &S.F.
O'Hannesin, IAH Conf.,
Hamilton, Ontario, May 10-13
(1992)
>
o
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe,
Zn, stainless steel,
Cu, brass, Al
Fe
Fe, Ni/Fe
Fe
Fe
Fe, Pd/Fe
Fe, surfactant
Cu
Fe
Contaminant
Chlorinated
hydrocarbons
Chlorinated
hydrocarbons
PCE, cDCE,
TCE
TCE
Cr(VI)
PCB
TCE, PCB
Dioxin, Furans
TCE
Type test
Batch,
Column,
Reac. Wall
Borden,
Ontario
Batch,
Column
Canister,
NJ
Batch
Batch
Batch
Batch,
Column
Batch
Batch
Description/Conditions
Batch: 10 g stainless steel, Cu, brass, Al, Fe, &Zn in 40
mLhypovialsw/ 1, 1, 1-TCA Next, Fe w/ halo-aliphatics.
Batch/columns used wall material. Constructed Borden
wallby driving sheet piling to form 1.6m x 5.5 mcell.
Reactive material 22% Fe grinding s and 78% concrete
sand with 348 sampling points installed within wall.
10 g 100 mesh iron powder, silica sand, 40mg/LCaCO3
added to 40 mL hypovials
Fe canister, NJ site testing up to 15 PCE, 1 cDCE, 0.5
mg/L TCE since 1 1 /94. Initial column used Ni plated Fe,
site water. Reactor used commercially plated Ni-Fe 7/96.
S A 3 . 1 mVg (before plating 1 . 1 mVg) . 2nd column
commercial Ni-Fe.
Batch 150/zm, 370 /zm mesh Fe & Fe powder in 40 mL
ZHEs, TCE in DI water, shaken 150 rpmthen analyzed
for pH, dissolved Fe and Cl" removal.
Varied Cr(VI), FT, and SA of iron as well as ionic strength
and mixing rate.
HC1 treated Fe particles (<10 /zm). KjPdCl6 w/ Fe powder
(0.05% w/w Pd). 20 ppm PCB ( 1 mL Aroclor 1260 or
1254), methanol/water/acetone (1:3:1) w/0.05% w/w
Pd/Fe in vial, amb. temp., capped, shaken. 1 /zL samples.
TCE batch ZHEs: 20 g of 40-mesh Fe & Pd (0.05%>Fe,
100 mL TCE (2 mg/L), surfactant (2%, 4%), cosolvent
(2%), 30 rpm. PCBs in 5-mL vials, 2 g of 100-200-mesh
Fe-Pd(0. 1%). Columns wet-packed w/ Fe or Fe-Pd,
sampled after >10 PV at various levels, times, rates.
Heat mixtures with Cu to enhance catalyzing reactions to
degrade dioxins and other compounds.
1.5-2.0 g of elemental iron per 100 mL aqueous sample
containing 0.02 mmoles TCE (25ppm); maintained pH.
Results
Batch: 1" order TCA rates, Steel, brass, Cu low rates; Al better,
Zn, Fe rapid (t1/2 100 min.). Fe(0) batch: Highly Cl-organics most
rapid. t1/2 0.22 h HCA to 432 h cDCE. Fe mass to solution
volume ratio important. Rates decline at pH ~ 9. Batch/column:
t1/215 h TCE, PCE. In situ wall: Avg. max. cone, downstream of
wall —10% of influent cone. Performance constant over
14-months.
All 14 chlorohydrocarbons except dichloromethane degraded.
Rates in column independent of velocity and consistent with
batch tests. When normalized to 1 m2 Fe surface/mL solution: tm
0.013-20 h. 5 to 1 5X > natural abiotic degradation.
t1/2 in initial column 10 X (30 to 3 min, TCE) lower than Fe alone
(1 to 1 . 5 versus 24 h RT). Enhanced reactor tia 4X > initial
column test, but 4X lower than Fe reactor. Longer tU2 may result
from inadequacies in commercial plating process.
Amount TCE degraded directly proportional to dissolved Fe in
solution. 2-fold increase in pseudo first order rate constant when
metal particle size decreased from370/zmbyfactorof2. 5. For
iron SA/V of solution <1000 m"1 TCE degradation rate constant
increased linearly with SA/V ratio.
Rate constant 5.45 x 10"5 cm"2 min"1 over wide range of
conditions. 1.33 mol diss. iron for each mole Cr(VI) reduced.
Rapid (few minutes) dechlorination on Pd/Fe surface. Rate
dependant on amount Pd/Fe, % w/w Pd(0) on Fe, & % v/v water.
Pd/Fe surface can be used repeatedly if acid- washed after 3 to 4
uses.
Batch Pd-Fe: tU2 TCE -27. 4 min.; PCB -100 to -500 min (as
surfactant increased). Columns: TCE and PCBs degrade at
enhanced rate t!/2 —1.5 and 6 min due to increased solid to
solution ratio. Fe-Pd filings applicable for ex-situ treatment of
TCE and PCBs in surfactant solutions generated during
surfactant flushing.
Cu catalyzed degradation of PCDD, PCDF, 7 other chlorinated
aromatics at low temp, similar to first observation using fly ash.
With and without citric acid, pH 5.8: rate 5.37 h"1 & 0.85 h"1; tU2
7.74 &48.9min., resp. Citric acid chekting ligand for Fe2+.
Reference
Gfflham, R.W., et al ., HazMat
Central Ccnference. Chicago,
Illinois, March 9-11, pp. 440-
453(1993)
Gfflham, R.W. &S.F.
O'Hannesin, Ground Water,
32(6):958(1994)
Gillham, R.W., etal.,
Internal' 1 Contain. Tech. Conf.
& Exhib. St. Petersburg, FL,
Feb9-12, pp. 85(1997)
Gotpagar, I, etal., Environ-
mental Progress, 16:137
(1997)
Gould, J. P., Water Res.,
16:871 (1982)
Grittini, Q, etal., EST,
29:2898-2900(1995)
Gu, B., etal., Internat'l
Contain. Tech. Conf. & Exhib
St. Petersburg, FL, Feb 9-12,
pp. 760-765(1997)
Hagenmaier, H., et al., ES&T,
21:1065(1987)
Haitko, D.A.& S.S. Baghel,
209thNational ACS Meeting
Anaheim, CA, 35:807 (1995)
-------
Appendix A. continued...
Metal
(zero valent
unless specified)
Fe
Fe
Fe, Sulfur
Fe, Sulfur
Iron, Fe-reducing
microbes
Fe
Pd/C
Fe/Pyrite mixture
Contaminant
G02
Hydrocarbon
(HC) formation
CC14
TCE, PCE
10
Nitrobenzenes
CC14
2-chloro-2-
propen- l-ol, CT,
Chlorobenzene
TCE
Type test
Batch,
Column
Batch
Batch
Batch
Column
Batch
Batch
Column
Description/Conditions
Etetermined that Cl to C5 hydrocarbons are formed by the
reduction of aqueous CQ by Fe(0) and product, have ASF
distribution.
Reduction of aqueous CO2 by Fe(0). Reaction mechanism
proposed for electroreduction of aqueous CQ with M
electrodes, Fe. Anderson- Schultz-Flory (ASF) distrib.
Etetermine if adding sulfur enhances degradation of CC14
by iron.
2 grades Fe, TCE and PCE. Lab grade Fe filings, 420 fjm,
sulfur >180 ppm. Extra pure Fe, at a particle size of 6 to
9//m and S content of 22. 1 ppb.
Columns to assess abiotic/biotic processes in reactive Fe
& Fe-reducing bacteria Nitrobenzene to analines.
CC14 1.5-5.5 fjM, Fe(0) powder: 1 to 10 g per 265 mL
distilled water in anoxic and oxic batch reactors
2-chloro-2-propen- l-ol, CT, Chlorobenzene, both bare
and palladized graphite electrodes in aqueous solution
over 24 h. Chlorobenzene rate limiting step in
dechlorination of compounds such as PCBs.
TCE pumped through stainless steel column w/ mixtures
of granular iron &pyrite. Products not considered.
Results
Pretreat iron with H2 increased hydrocarbon cone. 140 h un-
treated Fe 3.8±1.2//g/L vs. 7.9±2.4//g/Lusing hydrogenation
Indicate absorbed His a reactant in reduction of aqueousCO2.
10 HCs < C5 products ASF distrib. w/ hydrophobics sorbed to
Fe. Fe supply e & catalyst promotes formation/ growth of HC
chains. H2O also reactant HC desorption may be rate-limiting
step.
Sulphur (sulfate, organosulphonicacid, sulfides & pyrite) accel-
erated Fe induced degradation of CC14 under aerobic conditions.
Lab grade Fe resulted in ethyne, ethene, ethane in 24 h. Extra
pure unreactive after 1 mo. despite high SA, but fast production
of ethyne, ethene, ethane after adding sodium hydrogen sulfide.
NAC reduction primarily by a reaction w/ surface-bound Fe
species, which serve as mediators for the transfer of e-
originating from microbial oxidation of organic material by Fe-
reducing bacteria-Regeneration of reactive sites, (not e- transfer
to the NAC) rate-limiting process. Presence of reducible organic
pollutants may significantly enhance the activity of Fe-reducing
bacteria, in that reduction of such compounds continuously
regenerates easily available Fe(ni) species.
Products: CHC13, CH2d2. Anoxic Rate: 0.290 li1, 1 g Fe(0);
1.723 t1, 10 gFe(0); increases with SA(2.4 mg/g) &time.
Slower oxic rates: 0.085 h'1, 1 g Fe(0), DQ 7.4 mg/L. pH rapid
increase after Q, depleted.
Pd reduces CT by factor of 5; Both C and Pd/C cathodes effect
dechlorination, however Pd/C electrode has a greater selectivity
for dechlorination more products dechlorinated and gives much
higher yields of fully dechlorinated products. Favor acidic media
at higher ionic strength, (\ does not affect. 2-chloro-2-
propen-1-ol dechlorinates rapidly with Pd/C but by different
mechanism. Chlorobenzene — ^ benzene w/ Pd/C.
Interaction of GW& geochemical environ, using MTNTEQA2.
Pyrite in granulated Fe mixtures provided of pH control.
Reference
Hady, L.I. &R.W. Offlham.
209thNational ACS Meeting
Anaheim, CA, April 2-7,
35:724 (1995)
Haidy, L.I. &R.W. Offlham,
ES&T, 30:57-65 (1996)
Hams, S., etal. 2C9th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:825 (1995)
Hassan, S.M., et al., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:735 (1995)
Heijman, CO., et al., ES&T,
29:775 (1995)
HeUand, B., etal., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:732(1995)
Hdvenston, M.C, et al., 213th
National ACS Meeting, San
Francisco, CA 37:294-297
(1997)
Holser, R.A., et al., 209th
Nat' 1 ACS Meeting, Anaheim,
CA, April 2-7, 35:778 (1995)
-------
Appendix A. continued...
Metal
(zero valent
unless specified)
Mn2+, Fe2*, steel
wool(Fe)
Fe
Fe
Fe
Fe Colloids
Fe Colloids
Mineral oxides in
presence of Fe(II)
Palladized iron
Pyrite
Contaminant
Cr(VI)
TCE, PCE, TCA
CO,
CC14
Chemical
Barrier
Chemical
Barrier
10 mono sub-
stituted
Nitrobenzene s
Chlorinated
contaminants
CO,
Type test
Batch
Column,
reactor,
Belfast,
Ireland
Column
Column,
Batch
Batch,
Column
Column
Batch
Batch
Batch
Description/Conditions
Low hexavalent, high hexavalent soils had 1 05, 460
mg/kg Cr(VI); 1.8, 104 g total Cr; 8.5, 10.4pH,
respectively.
EnviroMetal provided a treatability study w/ Fe filing, site
water. RT ~ 12 h to reach regulatory limits. Designed 12
m tall x 1 .2 m dia. Fe reactor with 5 m flow path entry and
exitzones to collect and disperse flow. Reactor in cut-off
wall to funnel flow.
15-cm-dia Plexiglass pipe 90 cm long. Holes every 2. 5 cm
first 40 cm, rest 5 cm. Up & downgradient sand zone w/
intermed. Fe zone to simulate permeable iron barrier.
100-mesh untreated Fisher Fe powder (SA 0.057 m2/g)
<325 mesh inN2 purged unbuffered water and 20-32 mesh
Fluka Fe turnings SA 0.019 m2 g"1 in carbonate buffer.
Mixed in dark, 3 6 rpm, 23± 1° C. Table of new and
previous experimental conditions provided.
Hanford Site: Column/batch studies looking at injection
ofmicrometer-sizedFe(0) colloids into subsurface to form
chemical barrier.
PVC column study evaluating Fe(0) colloid injection rate
and concentration on colloid retention by a sand bed.
CaQ2 tracer studies to compare transport rates of colloids.
Suspensions of magnetite, goethite, lepidocrocite,
aluminum oxide, amorphous silica, titanium dioxide in
presence and absence of Fe2+ addition.
Bimetallic process evaluated as a means of increasing
rates of reaction.
1 /jM CCL, reacted with 1.2-1.4m2/L pretreated pyrite at
pH 6. 5, 25 ° C except experiments conducted with sulfide
at pH 7.75 ; aerobic & anaerobic. Pyrite 75-300 /zm.
Results
Mn2+ reduced 50 to 100% Cr(VI) in both soils (nopH adjust.).
Fe2*, steel wool reduced soluble &insol. Cr(VI) reduction
dependent on pH, reducing agent and soil.
/J/2TCE~ 1.2h. Small amount cDCE, Cl" increased, VC formed
from dechlorination (Up to 700 /zg/L) in treatability study.
Solvents degraded rapidly. Site Installation: After 7 mo. TCE 5
/zg/L but slightly higher at exit due to backflow from sampling.
DCE formed but ND (not detected) at 3 mo. VC at 4mo 0.4 /zg/L
and ND 7 mo. At 6 mo TCE 2^g/L and DCE is ND.
CC14 (up to 1.6 mM) fully dehalogenated by first sample port in
Fe zone. CC14 — » CHC13 — » CFi,Cl2. CC14 tia occurred within
0.25 h at a rate of 2.5 cm/h. CHC13 tU2 is slower but increased 2-
fold with a 5-fold increase in flow velocity.
New/previous k,,,, from batch and column varied widely.
Normalization to Fe surface cone, yields specific rate constant
ksa (vary by only 1 order of mag.). Dechlorination more rapid in
saturated carbon centers and high degrees of halogenation favor
rapid reduction. Representative kSA values provided for solvents.
Surfactants in low ionic strength solutions increased length of
time dense colloids (7.8 g cm"3 ) remained in suspension by
250%. Removal effic. sand column partially controlled by
injection rate.
Colloids controlled by rate & influent cone. As colloids accum.,
efficiency decreased due to gravitational settling. Colloids were
evenly distributed & high flow required to mobilize.
Fast nitroaromatic reduction in all Fe hydroxides at 6.5 pH.
Mineral oxides (no Fe2+) show slow reduction, but increased w/
Fe (hydr)oxide coatings. Rates pH-dependent, decreasing w/
increasing compound to solids concentration ratio.
TCE dehalogenation increased 2X using Pd/Fe instead of Fe and
extends process to less reactive dichloromethane.
>90% CC14 transform in 12-36 d (all conditions). Aerobic >70%
CC14 — » CO2 . Anaerobic 50% CC14 — » CHC13. FeOOH coat on
pyrite (aerobic). Pyrite depleted of ferrous Fe in all reactions.
Reference
James, B.R., J. of Environ.
Quality, 23:227(1994)
Jefferis, S.A., et al., Internal' 1
Contain. Tech. Conf. & Exhib.
St. Petersburg, FL, Feb 9-12,
pp. 817-826 (1997)
Johnson, T.L. & P.O.
Tratnyek. 33rd Hanford Symp.
on Health & the Environ — In
Situ Remed., pp. 931 (1994)
Johnson, T.L., et al., ES&T,
30:2634(1996)
Kaplan, D., etal., In Situ
Remed., BatteMe Press, pp. 821
(1994)
Kaplan, D., etal., J. of
Environ. Qual., 25(5):1036
(1996)
Klausen, J., et al., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:716(1995)
Korte,N.,etal.,209thNat'l
ACS Meeting Anaheim, CA,
April 2-7, 35:752(1995)
Kriegman-King, M.R. & M.
Reinhard, ES&T, 28:692
(1994)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Sulfide, Biotite,
Vermiculite
Corrin, reductant
Pd/Zn
Fe
Fe, Pd/Fe
Fe and sulfur
Fe
Fe
Fe
Contaminant
CCL,
CCL,
TCE
TCE
TCE
CC14
Precipitation
TCE, DCE
Precipitation
Type test
Batch
Batch
Batch
Batch
Batch
Batch
Batch
Column
Batch
Descrip tion/Con diti ons
Biotite, vermiculite, muscovite wet-ground to 200-50
mesh (73-300 //m). Ampules w/ 1 3. 5 mL buffer, spiked
with solution saturated with CCL, and flame sealed.
CC14 in vials w/ corrin (B12, cobinamide dicyanide, or
aquocobalamin), reductant (Ti(III), dithiothrietiol, or
S2'cystenine), pH 8.2. Products in headspace and mixture
determined by GC/MS, HPLC, NMR or TLC.
TCE dechlorination by Zn(0) in aqueous solutions at room
temp.
ZHEs w/ 3 Fe filings (Fisher, Columbus Chemical,
MBS). TCE 0. 5 to 20 ppm. Fisher Fe pretreated with
HC1.
25g of 40-mesh Fe filings added to ZHEs, containing 1 25
mL solution (nominally 2 mg/L of TCE) at 3 0 rpm. Pd/Fe
prepared according to Muftikian et al. (1995).
Sodium sulfate, sodium sulfide, ferrous sulfide, pyrite,
organic acid, electrolytic Fe(0) powder, Fe(0) degradation
of CCL, under aerobic conditions.
Tracers, aqueous inorganic profiles, scanning electron
microscopy (SEM), x-ray photoelectron spectroscopy
(XPS), and wavelength dispersive spectroscopy (WDS)to
determine precipitates and porosity loss in Fe systems.
148 Ibs VWR coarse Fe filings in columns. Initially used
buffered DI water (40 mg/L CaCO,). pH 7-8.5. Later used
GW, 400 mg/L CaCO3 and pH 7-8. Flow velocities much
higher than typical to accelerate effects of aging.
Tracers, aqueous inorganic profiles, SEM, XPS, and WDS
to determine precipitates and porosity loss in Fe systems.
Results
~ 80-85% CCL, — > CO2 via intermed. CS2. Chloroform 5-15%,
5%unidentifiednonvolatile and CO. At 25° ImM HS tll2 2600,
160, 50 d for homogeneous, vermiculite, biotite system,
respectively.
Proposed pathway: trichlromethyl radical forms adduct with
reductant. In S2"/cysteine produces CSj orthiazolidinesby way of
thiophosgene. Or radical further reduced to form CHC13 and
CH2C12 or CO and formate by dichlorocarbene intermediation.
Bimetallics Ag, Ni, Pd to enhance Zn. Dechlorination few h to
several d. Best rates w/ cryo-Zn (ultrafme Zn) & Pd. Ethy lene,
ethane, monochlorinated hydrocarbons products.
Rates varied by factor of 2 for 3 Fe's. TCE sorbed then reduced
by MBS. Fisher Fe 7 to 5 h. TCE/Fe ratios changed rate as well.
When 5 ppm TCE reacts with Fe(0), ~ 140 ppb VC persists 73 d.
VC (—10 ppb) remaining with Pd/Fe about an order of
magnitude > w/ Fe(0). Volatile byproducts may be under-
represented in other published data regarding reduction with
Fe(0). Reduction of TCE w/ Pd/Fe (0.05 % Pd) > order of
magnitude faster than with Fe(0). With a 5:1 soluti on-to-solid
ratio TCE tie with Fe(0) =7.41 h, tU2 with Pd/Fe = 0.59 h.
Products: chloroform, potential for carbon disulphide (toxic).
Sulphur significantly increased rates under aerobic conditions.
Pyrite canregenerate ferrous ions, produce sulfate & control pH
Precipitation changes color from black to gray. Loss of alkalinity
and calcium, no signif. magnesium loss. Most loss (5-15%) early
and levels off. SEM shows crystals form on the surface of Fe.
Tracers indicate fairly uniform loss of porosity throughout.
TCE fj/236min at 40, 20, 12 mL/min. tDCE tm "~ 100 min., 1,1-
FJCE tll2 200min, although 1st order fit not as good as for TCE.
cDCE particularly poor. Siderite formed at the top of column 1,
throughout column 2 and at the bottom of column 3.
GW forms precipitates (F^OH^, FeCO,, CaCO,) on Fe surfaces,
which may affect reactivity. However, this effect, to date, small
Also, H2 produced from anaerobic corrosion of Fe afactor
controlling the measured porosity losses in iron systems.
Reference
Kriegman-King, M.R. & M.
Reinhard, ES&T, 26:2198
(1992)
Lewis, T.A., et al., ES&T,
30:292(1996)
Li,W. &K.J. Habunde,
HSRC/WERC Joint Conf. on
Environ., 5/20 Paper 35 (1997)
Liang, L., et al., 209th Nat'l
ACS Meeting, Anaheim, CA,
April 2-7, 35:728(1995)
Liang, L., etal., GWMR
Winter, 122(1997)
Lipczynska-Kochany, E., et al,
Chemosphere, 29(7):1477
(1994)
MacKenzie, P.D., et al.,
Emerg. Technol. in Haz. Waste
Manag. Vn, Sept. 17-20,
Atlanta, GA, pp. 59-62 (1995)
Mackenzie, P.O., et al., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:795(1995)
Mackenzie, P.O., etal.,
National ACS Meeting,
37:154-157 (1997)
-------
Appendix A. continued.
Metal
(zero valent
unless specified)
Ferric
Oxyhydr oxide
Fe
Fe
Fe
Fe
Chemical Barriers
Amorphous ferric
oxyhydroxide
Palladized iron
(Pd/Fe)
Palladized iron
(Pd/Fe)
Fe-Pd
Contaminant
Se
U
TCE, CO,
TCE
CC14, cDCE,
tDCE
Mo,U
Mo,U
TCE, DCE, cis
& trans- 1,2-di-
chloroethylene,
PCE
TCE
chlorophenols
(CPs)
Type test
Batch, Site
Samples
Column
Batch
Batch
Batch
Column
Batch
Batch
Batch
Batch
Description/Conditions
Three sediment samples from Kesterson, Merced County,
CA with elevated levels of Se subjected to in situ Fe(II)
amendment
Rocky Flats seep w/ Cl- organics, metals, radionuclides.
Seep water in glass columns w/ steel wool (Fe(0)) at 5 (1st
d); 10 (2nd d); 30 mL/min (next4 d). Rates increased to
determine effects on U removal. DO 5 to 6 mg/L, pH 8,
and 13 to 21° C.
Evaluated core samples from actual field sites.
Microbiology and geochemistry characterized.
Fe(0) participate by direct reduction, ferrous iron, and
hydrogen produced during corrosion.
Temperature, steric, pH dependance of degradation and
reactions of pollutants in response to untreated iron
powder (finer than 100 mesh), under aerobic conditions.
10 cmdia acrylic pipe w/~l,250 mL sand mixed with
test material. U and Mo measured. Used hydrated lime.
Lab experiments to evaluate material for use in chemical
barrier under a repository containing uranium mill
tailings.
Sealed 1 2 mL glass vial. Pd/Fe [3.6g-10/zm(Aldrich), or
3.6 g Fe filings (Baker & Adamson), or 10 g- 40 mesh Fe
(Fisher)] with 10 mL of chlorinated compounds (20 ppm
in H2O), shaken. Sampled with a syringe for GC analysis.
Ixl-cmpure Fe foil, 0.254 mm thick welded to stainless
steel stub etched w/ 8 keV argon ions. Potassium
hexachloropalladate added and allowed to react.
Batch Fe(0)-Pd(0) in unbuffered, DI water, room temp,
dark, usuallywith [Fe( 0)-Pd(0)]= 69.4 g/L and initial
aqueous chlorophenols [CP] ~ 0.08 mM, 40 rpm.
Results
Both Se(IV) & S 1 ppm. Chloromethanes, CC14, CHC13- CH2C1, also
dechlorinated to methane. CC14 in a few minutes, CHC13 in < 1 h,
CH2C12 in 4-5 h.
Pd(IV) to Pd(H), protons on hydroxylated Fe oxide form Pd(H)-
O-Fe bonds, collapsing to Pd/Fe. TCE forms hydroxylated Fe
oxide film that deactivates Pd/Fe. Dilute acid removes film.
Initial rapid loss CP (tU2 < 0.2 h) due to sorption to Fe(0)-Pd(0)
surface. Rate constant, k,,b!, proportional increase with Pd(0) used
and SA of Fe(0)-Pd(0). Higher cosolvent corresponded to
decrease in k^.
Reference
Manning, B.A & R.G. Burau,
ES&T. 29(10):2639(r995)
Marozas, D.C., et al., Internat'l
Contain. Tech. Corf. & Exhih
St. Petersbmg, EL, Feb 9-12,
pp. 1029-1035(1997)
Matheson, L.J. &P.G.
Tratnyek. 1993. 205th ACS
Meeting 33:3
Matheson, L.J. &P.G
Tratnyek, ES&T 28:2045-2053
(1994)
Milbum, R., etal., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:822(1995)
Morrison, S.J. & RR.
Spangler, Env. Progr.,
12(3):175(1993)
Morrison, S.J. & RR.
Spangler, ES&T, 26(10):1922
(1992)
Muflrkian, R., et al., Water
Research, 29:2434-2439
(1995)
Muflrkian, R, et al., ES&T,
30:353-356(1996)
Neurath, S. K., et al ., 213th
ASC National Meeting, San
Francisco, CA, 37:159-161
(1997)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Contaminant
CC14, TCM,
TCE, PCE
Halo- organic
compounds
TCE, PCE
TCE
TCE
C12
Cr(VI)
Cr(VI)
Cr(VI)
Cr(VI)
Type test
Batch
Wall,
Borden
Reac.
Wall,
Borden
Column
Column
Batch
Batch,
Column
Batch,
Column
Batch
Batch
Description/Conditions
Fe catalyst and aquifer material collected from Canadian
Forces Base, Borden, Ontario. Q CC14-4050, TCM-4650,
TCE4080, & PCE-3970/zg/L at 12° C.
Reactive wall 22% Fe, 78% concrete sand; 5.5 m
downgradient. Cell driven 9.7m to bottom silty clay lens.
Field demonstration in 199 1 at Borden, Ontario. TCE 270
and PCE 43 mg/L. PRB 1.5m wide of 22% granular Fe
and 78% sand placed in path of plume moving 19 cm/d.
Plexiglass columns packed with mixture 15% electrolytic
iron and 85% silica sand. TCE pumped at 0. 1 mL/min. C
TCE= 1.3,4.7, 10.2, 61 mg/L.
Columnsused simulated GW containing 1.3 to 61 mg/L
TCE. Column packed with 15% iron powder mixed with
sieved 35-mesh silica sand. SA of iron 0.287m2/g, iron SA
to solution volume ratio 0.21m2/mL.
Chlorine solutions stirred in 250 mLreactor at 20°C.
Granular Fe 0.2-.5 dia added. 10 mL taken at 5, 1 0, 15,
20, 25 min and analyzed for chlorine & chloride contents.
Experiments also carried out at pH 4, 5, 6, 7, 8, and 9.
Geochemical effects and mineralogy on reduction CrO42"
by Fe(0) using stirred batch reactor (SBRs) and shaken
batch bottles under N2 to evaluate kinetics/mechanisms.
SBRs: under N2 w/ & w/o Elizabeth City aquifer material.
C0 of CrO42" = 136 to 156 mg/L. 1st column portion of
aquifer material. 2nd column mixture of Fe filings (7.5 g)
& aquifer material (67.6 g). 6 mg/L chromate as IQCrQ,
introduced at 0.05 mL/min.
Some types of iron and aquifer material more reactive
than others. Ada Iron & Metal (AI&M) & Master
Builders Supply (MBS); Eliz. City, NC& Otis AFB, MA
aquifer material used along with commercial Si sand.
Aquifer materials from Elizabeth City, NC and Otis AFB,
MA also kaolinite and montmorillonite using simulated
GW. Scrap iron filings from AI&M and cast iron metal
chips from MBS.
Results
tU2 2.2, 850, 1 520 and 4000 minutes for CC14, TCM, TCE, and
PCE, respectively. No Eh change in controls, reactive vials
showed highly reducing conditions, but no significant pH
change.
TCE reduced 95%, PCE 9 1% No TCM downstream of wall. CT
increase consistent with quantity degraded. Traces DCE; no VC.
90% TCE and 8 8% PCE removed from solution and Cf indicated
dechlorination. Major product was cDCE with peak of 2200 /zg/L
followed by tDCE and 1,1-DCE, VCND.
TCE Q = 4.7mg/L. Products: ethene 40%, ethane 18%, Cl to C4
10%. 3-DCE isomers & VC. c 1,2-DCE primary product of
degradation, though sum of all chlorinated was only 3 to 3.5%.
Pseudo-lst-order rates. Products :ethene > ethane >» other
Cl — C4 hydrocarbons. 3.0-3.5% TCE appeared as chlorinated
products. Most TCE probably sorbs to iron surface until
complete dechlorination achieved.
Optimum conditions investigated for pH, particle size, and
contact period. Species of OC1" and HOC1 removed by 100%
between pH 4 and 7 within 25 min.
Chromate reduction rapid and complete in zero valent systems
and natural aquifer material. t1/2 = 1 1 h. When no aquifer material
much slower and incomplete during 146 h.
Rapid changes in Eh from positive to highly negative upon
introduction of Fe metal. Chromate reduction slow in system
with no aquifer material but rapid in system containing the
natural solid phase. Eh and pH changes less dramatic in effluent
from column but influent chromate effectively removed.
Data support conclusion that CrO42- can be reduced to Cr(IH) in
presence of elemental iron. AI&M and Eliz City aquifer material
most reactive. Suitable e acceptors need to form appropriate
couple. Mechanisms proposed.
Coupled corrosion processes responsible for CrO42- reduction &
precipitation, w/ AI&MFe;much greater rates than MBS Fe.
Aluminosilicate dissolution proposed to increase reaction rates.
Reference
O'Hannesin, S.F. &R.W.
Gfflham, 45th Canadian
Geotechnical Soc. Conf, Oct
25-28(1992)
O'Hannesin, S.F.,etal.,
Emerg. Technol. in Haz. Waste
Mangmt VII, ACS, Sep 17-20,
Atlanta, GA, pp. 55-58 (1995)
Orth, W.S. & R.W. Gfflham,
209thNational ACS Meeting
Anaheim, CA, April 2-7,
35:815(1995)
Orth, W.S. & R.W. Gfflham,
ES&T, 30:66-71 (1996)
Ozdemir, M. & M Tufekci,
Water Research, 31:343-345
(1997)
Powell, R.M, M.S. Thesis,
University of Oklahona,
Norman, OK (1994)
Powell, R.M, et al., Water
Environ. Fed. Conf., March,
Miami, pp. 485(1994)
Powell, R.M., etal., ES&T,
29(8):1913(1995)
Powell, R.M., etal., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:732(1995)
>
ON
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Fe
Contaminant
Cr(VI), TCE
Cr(VI)
Atrazine
Cr(VI)
Cr(VI), DCE,
TCE,VC
Cr(VI)
Nitrate, Nitrite
PCP
Type test
Batch
Column
Batch
Char.
Study, Eliz
CityNC
Field
Study, Eliz
CityNC
Field
Study, Eliz
CityNC
Batch
Batch
Description/Conditions
Shaken batch b ottle experiments were used to evaluate
both the dissolution of 7 aluminosilicate minerals in the
presence of Fe(0) and whether the dissolution affected the
rates of chromate reduction by Fe(0).
Chromate reduction in Fe and quartz grains to determine
fate of reduced Cr in wall material. Fe filings reacted with
20 mg/L Cr(VI), as K2CrO7 for more than 1 50 PV. Quartz
grains flushed w/ CaCO, solution following Cr(VI)
breakthrough in the column.
Thiocyanate reacts w/ Fe(in) giving red color (463 nm).
4.5 x 10"5 M Fe2+ (w/ light) accelerates atrazine
decomposition.
48 Cores at various depths. Assessed chemical speciation
and distribution of Cr on contaminated soils and its
leaching potential. Batch adsorption/reduction procedures
used.
25% eachbyvol. EC aquifer material, sand, Fe-lathe
turnings (0. 1-2 mm) & MBS-Fe-chips (1- 10 mm >
sulphur, carbon) in field test. 21 augered boreholes, 3 to 8
m bgs, were filled with the mixture. 21 monitoring wells
installed.
At Elizabeth City, 2 Fe sources (AI&M, MBS) mixed with
native aquifer material and 1 0-mesh washed sand. 20-cm
dia. cylinders installed in three rows 3 to 8 m bgs, 21
total.
Batch in unbuffered, anaerobic, DI water, 22° C, dark, HC1
treated & untreated F<0)(>40 mesh) 69.4 g/L. Inject
nitrate or nitrite[C0 ~ 0. 16 mM), 40 rpm. 1C for NO3" and
NO2" quantification, estimated NH," using a colormetric
Hachkit.
Electrolytic Fe 5 g/20 mL used. Influenced by preparation
of metal surface by treating with HC1, PCP concentration,
pH, temperature, and presence of inorganic compounds.
Results
Support previous hypotheses that aluminosilicate dissolution
promoted Fe corrosion reactions, hence chromate reduction, due
to generation of protons. Proposed mechanisms for chromate
reduction and TCE dechlorination depicted in reaction diagrams.
Reacted Fe developed coatings of goethite withCr(III)
concentrated on outermost edges. In regions of increased Cr,
goethite acquired characteristics similar to Fe2O3 and Cr2O3.
Complete reduction of Cr(VI) to Cr(III). Cr(in) incorporated into
sparingly soluble solid species.
Atrazine fully decomposed atpH 1.5 under sunlight within 2h.
Most important factors under light are surface area and nature of
iron used. No degradation in dark.
Adsorption and reduction capacity of soils were overwhelmed
permitting passage of Cr(Vi). Capacity differences related to clay
content & pH; less to amorphous iron oxide coating.
Disappearance of contaminants with appearance of ferrous Fe,
decrease in oxidation-reduction potential and DO with slight pH
increase. Sulfide also detected downgradient and within 30 cm of
iron cylinders. Less reducing conditions downgradient.
Cr to < 0.01 mg/L Significantreductions in TCE. Siderite not
detected, but Fe sulfides were. Full demonstration scheduled
June 1996 of 50-longx 8-deepx 0.6-wide(m) trench of F<0).
Ist-order nitrate rate constants, ki, increased with HC1 pretreated
Fe(0). First 12 h following treatment, k[ gradually declines in the
presence of C1-. High chloride in otherwise identical systems
cause much smaller decline in k[ . Rate constant for nitrite
reduction, k2, small due to similar acid pretreatment of Fe. k[ &
k2 for nhrate & nitrite reduction by untreated Fe(0) directly
dependent on concentration of Fe(0), ranging 69.4 to 20.8 g L"1.
Prefreatediron improved rates (6 h, 60-70% PCP [2.7 x 10"6 M]
degraded). Keep pH near neutral. Some anions (e.g., Cl) retard
degradation. Results indicated poor remedial choice for PCP.
Reference
Powell, R. M. & R. W. Puls,
ES&T, 31:2244(1997)
Pratt, AR, etal, ES&T
31:2492-2498(1997)
Pulgarin, C., etal., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:767(1995)
Pills, R.W., C.J.Paul, D.A
Clark, J. Vardy.. J. of Soil
Contain. 3(2):203 1994
Ms, R.W., etal, 20 9th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:788(1995)
Puls, R., et al., 4th Great Lakes
Geotech. & Geoenviron. Conf ,
Univ. of Illinois, pp. 23 (1996)
Rahman, A. & A. Agrawal,
213thNational ACS Meeting
37:157-159 (1997)
Ravary, C. & E. Lipczynska-
Kochany, 209th National ACS
Meeting Anaheim, CA, April
2-7, 35:738 (1995)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe
Fe
Fe,US
Fe
Fe
Contaminant
Fe corrosion
TCE, PCE
TCE
tDCE, cDCE
DDT, ODD
Type test
Batch
Site demo,
Moffett,
CA
Batch,
column
Batch
Batch
Description/Conditions
Fe(0) anaerobic corrosion produce Fe2*, OH", H2(g).
Corrosion rates are measured by monitoring the H2
pressure increase in sealed cells containing iron granules
and water. The principal interference is hydrogen entry
and entrapment by the iron. The entry rate is describedby
Sievert's law(R=kPFtOj), and the rate constant, k, is
evaluated by reducing the cell pressure once during a
testMBS Fe, Blend A, 50/50 mix of 10-18 & 18-32 mesh
granules (C, 3.2%; Mn, 0.65%; S, 0.09%; SA, 1.5 rrfg1).
Pilot field demo, Mofiett Field, Mountain View, CA 1/96.
50' long x 10' wide x 22' thick funnel- &-gate installed
across TCE and PCE plume 4/96.
5 to 20 mg/L TCE, 20-kHz US, 0. 16 cm2 in 0. 5-L bag
with 100-mesh Fe, anaerobic, 160 shakes/min.
Column: 20% Fe, 80% sand. 4 Fe's: 50-mesh particles,
Peerless acid-washed chips, unwashed Peerless, MBS
washed chips. 43% 20-mesh; 40% 40-mesh, rest Fe dust.
15 mg/L TCE at 4.7 mL/min. US 15.9 mm-dia, auger drill
bit, 1 5 cm long in 50-mesh Fe, 50% power.
25-mL sealedbottles w/ 1.7 g Fe(0) & 34 mg ground
pyritein 1.7mL Ar-spargedDI water. SAFe(0)0.7m2/g
with either 6.7/j.L tDCE, 4.4 /^L cDCE, 10 //L 1,1-DCE,
30//L VC in methanol,4rpm. Volatiles partition in to high
headspace slowing further reaction and enhance accum.
Sealed 40-mL vials. Fe powder (0.3-3 g) with 20 mL
deoxygenated buffer, pH7, DDT or DDD dissolved in
acetone, and, with and without the presence of nonionic
surfactant Triton X-l 14 at 250 mg/L. Uncapped reactors
flushed with N2, beads to improve mixing. Closed reactors
shaken at 130 rpmat20± 0.5 °C.
Results
For the 10 - 32 mesh iron, k was 0.01 5 initially but decreased to
0.009 mmol kg1 d'1 kPa° 5 in 1 50 d. Corrosion rate in saline GW
0.7-0.05 mmol of Fe kg'd"1 at 25 °C — identical under saturated
or fully-drained conditions. Rates decreased by 50% in 150 d due
to alteration product buildup. First 40 - 200 h progressively
increasing rates of pressure increase. Time before steady-state
rates develop depends on the solution composition. Discarded
this data in calculating corrosion rates. Tests on pure NaCl
solution sat identical equivalent concentrations (0.02 equiv/L)
show the following anion effect on corrosion rate:
HCO3>SO42>Cr. For Nad solutions, corrosion rates decrease
from 0.02to 3.0 mmol kg1 d'1 kPa° 5.
Baseline sampling 6/96 & 9/96 positive. TCE> 1,000 //g/L
upgradient reduced to NDw/in first 2' of cell (gate). Demo
continue until 3/98; Report to be prepared for DOD.
US removes corrosion from Fe surface and prolongs reactive life
Sonication for 0.5 h increased rates about 12%. But rates nearly
tripled to 1 84% after 1 h treatment. Prior to US, lower half
column (highest TCE cone.) ti/2 1.5 times upper section. After
US tU2 dropped. Lower t1/2 decrease 70%; upper t1/2 22%.
2 categories of reductive dehalogenation: Hydrogenolysis
(replace halogen by FT) & reductive elimination (2 halide ions
released), both nettransfer of 2 e. Haloethylenes can undergo
reductive, 0-elimination to alkynes under environmental
conditions. Evidence of this is involved in reaction of
chloroethylenes with Fe(0).
Rates of dechlorination of DDT and DDE were independent of
the amount of iron, w/ or w/o surfactant. Rates w/ surfactant
much higher than w/o. Initial Ist-order rate of DDT
dechlorination was 1.7 ± 0.4& 3.0 ±0.8 d"1 or, normalized by
the specific iron SA, 0.016 ±0.004 and 0.029± 0.008 L riV'/lV11,
w/ and w/o surfactant, respectively. Mechanistic model
constructed that qualitatively fit the observed kinetic data,
indicating that the rate of dechlorination of the solid- phase
(crystalline) reactants was limited by the rate of dissolution into
the aqueous phase.
Reference
Rear don, E.J., ES&T,
29:2936-2945(1995)
Naval Facilities Eng. Service
Center, Enviro. Restor. Div.,
www.Updated April 24, (1 997)
Reinhart, D.R., et al., Internat'l
Contain. Tech. Corf. & Exhih
St. Petersburg, FL, Feb 9-12,
pp. 806-813(1997)
Roberts, A, et al., ES&T,
30:2654(1996)
Sayles, G.D., et al., ES&T.
31:3448-3454(1997)
-------
Appendix A. continued.
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Fe, Mg
Fe, Mg
Fe
Pd,H2
Contaminant
CO,
CO,
CO,
TCE
TCE
chlorinated
ethylenes
Chlorinated
ethylenes
(PCE, VC)
Type test
Batch
Reduction
kinetics
Reduction
kinetics
Batch
Batch
Batch
Batch
Description/Conditions
System similar to that used by Matheson and Tratnyek
(1994) except with the range of Fe(0) concentration
extended from 5-31 mVL to 0.2 - 80 m2/L.
Linear sweep vollammograms(LSV) of Fe(0) at3000
rpm. Potentials set to avoid FT evolution at more (-). than
-700 mV/SHE and Q evol. at more (+) than 800 mV/SHE.
LSVs w/ & w/o CCl, . Increased (-) current in CCL,
attributed toreduction of CCL, to CHC13.
F<0) from 99.5% pure F<0) rod 3.0 mm dia, SA 0.071
cm2. Mass transport controlled by polished Fe(0) rotating
disk electrode. Kinetics of CC14 dechlorination in pH 8.4
buffer at potential where oxide film would not form.
4. 1 g/L Fe powder in oxygen-free, pH 7 water, 50° C.
0.5 g either 0.5% Pd alumina or 1% Pd on granular carbon
(GAC) and 1% Pd PAC at 400 rpm. Total aqueous phase
60 mL, N2 purged, capped, 5 mL H2 added for 15 minto
pre-reduce Pd. PCE =1.1 /zmoles in 3. 5 /zL of methanol.
Glovebox 90% N2 / 10% H2, bottles filled with iron and
HEPES-buffered water (pH 7). Placed in 50° C waterbath,
with or without shaking.
125 mL glass bottle, 60 mL N2 purged water, 0.05-0. 5 g
catalyst [5 g either 0. 5% Pd-alumina or 1% Pd on GAC
and 1% Pd PAC or PCI -silica], w/ 1- 1. 1 /zmoles PCE, 400
rpm. 5 mL H2 added at stages; some initial H2 = 0. 1 atm.
Sampled headspace for ethene, ethane, ethylenes.
Results
Greater range of Fe(0) SA shows hyperbolic relationship.
Concentration up to 80 m2/L with sharp increase in rate
indicating heterogeneous catalysis, electrical double layer or
abrasion effects during mixing.
More reducing potential on Fe(0) increases rates of H2O & CCL,
reduction. However, H2O becomes increasingly larger portion. At
potentials more negative than -700 mV/NHE, water reduction
larger portion than CC14 dechlorination, suggesting more
reducing potential would not enhance CCL, dechlorination rate.
Cathodic current independent of electrode rotation rate, Ist-order
rate constant (kc, = 2.3 x 10"5 cms"1) < estimated rate constant for
mass transfer to surface. Rate reduction of CC14 by oxide-free
Fe(0) dominated by reaction at metal-water interface.
TCE tU2 = 20 d using Fe. 1, 1,1-trichloroethane, 1, 1-
dichloroethylene, tetrachloroethylene transformed using buffered
water & landfill leachate.
5 chloroethylenes including PCE and VC reduced in 10 min by
0.5 g of 0.5% Pd on alumina and 0. 1 atm H2. Ethane 5 5-85%,
ethene < 5%. Pd on GAC yielded 55% ethane from PCE. PCE tlfl
= 9 min for 0.055 //mole Pd (583 //,g of 1% Pd on PAC). 1 0%
nitrite decreased rate by 50%. O2 greatly decreased all rates.
The more highly halogenated compounds mostreactive (tDCE >
TCE > PCE = cDCE = 1, 1-DCE) with the exception of VC which
is less reactive than PCE. Product: ethene, ethane. PCE: 15-30%;
VC: 50% reduction. Ethene/ethane ratio larger for VC. TCE
intermediate of PCE. cDCE only intermediate of TCE.
Pd-PCI-silica PCE rate 0.034 min1 (t1/2 =20 min), 65% ethane;
2%ethene. Some PCE dehalogenated before double bond
reduced. Al pellets, GAC or PAC, PCE 99.99% reduced in 10
min.; ethane 65-80%; ethene >5%. Sorption significant on
carbon but did not hinder transformation using Pd-PAC.
Reference
Scherer, M.M. & P.O.
Tratnyek, 209th National ACS
Meeting Anaheim, CA, April
2-7, 35:805 (1995)
Scherer, M.M., et al., 214th
National ACS Meeting, Las
Vegas, NV, 37:247-248 (1997)
Scherer, M.M., et al., ES&T,
31:2385(1997)
Schrder, C.G. & M. Reinhard,
Chemosphere, 29(8):1743
(1994)
Schrder, C.G. & M. Reinhard,
Chemosphere, 31(6):3475
(1995)
Schrder, C.G. & M. Reinhard,
209thNational ACS Meeting
Anahdm, CA, April 2-7,
35:833(1995)
Schrder, C.G. & M. Reinhard,
209thNational ACS Meeting
Anahdm, CA, April 2-7,
35:749(1995)
-------
Appe nd ix A. continued...
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Fe&
Ft/Pd/Al2O3
Fe&
Ft/Pd/Al2O3
Fe
Fe
Fe and FeS
Contaminant
Precipitates
TCA TCE
Al, Cd, Co, Cr,
Fe, K, Mg, Mn,
Ni, Pb, Zn
1 ,2-dibromo-3-
chl or o- propane
(DBCP)
1 ,2-dibromo-3-
chl or o- propane
(DBCP)
Atrazine
TCE, DCE, VC
TCE
Type test
Column
Batch,
Column
Electro-
chemical
Cell
Batch
Batch
Batch, Soil
Batch,
Column
Batch,
Column
Description/Conditions
ID precipitates that form on iron surfaces w/ differing
water chemistry. Two column tests performed using 100
mesh, 99% pure electrolytic iron. A 1 20 mg/L CaCO3
solution passed through one column and a 40 mg/L KBr
solution through other. Rate was at 0.23 mL/min. RT ~
13.3 h. N2 gas passed through 2nd column. Sampled after
158 and 166 PV.
Iron powder for removal of TCA and TCE from waste
water.
Electrochemical cell of massive sulfide-graphite rock
from mine site as cathode, scrap iron as the sacrificial
anode and acidic leachate from mine site as electrolyte.
Iron powder and Ffc/Pd/ALCv Palladium used as a catalyst
with Ft gas as the reductant. Looked at both sterile
(abiotic) buffered and unbuffered conditions.
Compared F<0) and H2/Pd-alumina for DBCP — >
propane. 4 g of 100-200 mesh Fe powder in 125 mL glass
bottle, HOmLdeox. solution, 10/zg/LDBCP, anaerobic,
400 rpm. MilliQ6" (DI) water (pH 7.0) or GW(pH 8.2-
8.7), some amended w/ anions and/or buffer (pH 7.0).
F<0) ( 10% w/w), 0.02 mg/L atrazine in batch. 20 mg
14C-atrazine in Fe(0) (20% w/w) in batch. F<0) (2% w/w)
in soil to determine mineralization and availability of the
pesticide atrazine.
Anaerobic & mildly aerobic conditions; >25 commercial
iron metals in several forms; 0.1- 1325 m2/L iron metal;
several groundwaters used.
Effect of FeS varying the FeS mass in batch. Columns
with Fe filings, SA concentration avg. 6000 m2/L.
Results
CaCO3 treated iron formed whitish gray coating on first
centimeter of column but KBr treated iron did not display any
visible precipitates. CaCft and FeCQ phases were only present
on the surface of the iron removed from the influent end of the
column treated with a CaCO3 solution. Fe surfaces analyzed from
both influent and effluent end of the KBr treated iron and the
effluent end of the CaCQj treated iron indicated presence of
magnetite (FesOi) precipitates.
50% TCA removal from 4 h to 1 h as temperature rose 20 to
50° C Degradation rates highly sensitive to Fe SA, significant
decline at pH values in excess of 8.0.
Cell raised pH of -41 L leachate from 3.0to 5.6 with decrease in
redox potential from >650 to <300 mV. Iron sulfate precipitate
formed with a concomitant lowering of Al, Cd, Co, Cu and Ni .
Fe(0) dehalogenated DBCP under sterile abiotic conditions
buffered & unbuffered; also, Pd w/ Ft gas as reductant in GW.
pH had little effect, however, a solution with pH = 9 inhibited the
reaction.
F<0) in H2O DBCP t1/2 =2.5 min; tU2 = 41-77 min in GW. Q,
NO3- slow reaction. 60 mg/L nitrate removed in 14 min. DBCP
trans, in min. w/ 75 mL GW, 22.5 mg 1% Pd-alumina. Rate in
GW 30% slower compared to Milli-Cf1. Slight inhibition in
Milli-Q111 by SO.,", NO3', Cl- or O2. SO32" » inhibitory effect.
The batch test removed 93% atrazine in 48 h. 5% of that was
adsorbed "readily available"; 33% "restricted"; 2% residues.
88% 14C removed in 48 h of that 6% was available, 72% pool;
the rest was bound. Fe in soil quadrupled mineralization in
120-d. 2% F<0) & 1 00 mg NO3" kg"1 increased by a factor of 1 0;
unextractable residue was greater than two times greater than the
control (noFe(O)).
No significant products from TCE batch or column. Sfrong Fe(0)
it-bonds may prevent DCE & VC products from desorbing.
Direct reduction of adsorbed chloroethene at metal/ water
interface. Reduction by iron oxide and oxyhydroxide not seen.
TCE tU2 —40 min. Rates constant over several hundred PV even
though surface of iron filings coated with FeCQj precipitate.
Reference
Schuhmacher, T., et al., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:801 (1997)
Senzaki, T. & Y. Kumagai,
KogyoYosui, 357:2(1988);
Kogyo Yosui, 369:19 (1989);
Senzaki, T., Kogyo Yosui,
391:21 (1991)
Shelp, G.S., et al., Applied
Geochemistry, 10:705(1996)
Siantar, D.P., etal., 209th
National ACS Meeting,
Anaheim, CA, April 2-7,
35:745(1995)
Siantar, D.P., et al., Wat. Res.
30:2315(1996)
Sinah, J., et al., HSRC/WERC
Joint Conference on the
Environment, May 20, Paper
36(1997)
Sivavec, T.M. & D.P. Horney,
209thNational ACS Meeting
Anaheim, CA, April 2-7,
35:695(1995)
Sivavec, T.M. et al. Emerg.
Technol. in Haz Waste Mngt
VII, Atlanta, pp. 42-45 (1995)
>
o
-------
Appe ndix A. continued...
Metal
(zero valent
unless specified)
Fe, Ni/Fe, Pd/Fe
Fe
Fe and sulfur
Funnel-and-gate
Fe
In Situ Fe Wall
H2Q;/Fe Powder
Organics,
inoculated w/
bacteria
Steel
Zn
Contaminant
TCE
Cd
Mo,U
Configurations
Halogenated
organic s
VOCs
AzoDyes
U
Cr(VI)
2,3-dibromo-
pentane
Type test
Column
Packed Bed
Batch
Column
Model
Batch
Field Site,
CA
Batch
Batch,
Column
Batch
Batch
Description/Conditions
Bimetallics accelerate degradation relative to untreated
Fe. Column of Ni-treated granular Fe with TCEr
contaminated site GW (TCE 2. 1-3.3 mg/L).
Packed bed of Fe sorbent supported on fine mesh stainless
steel screen & Teflon® flakes at 0. 12-0.70 cm/s.
Batch test of test material with synthetic U mill tailing
pore fluid. The column consisted of a solids chamber and
a water sampling chamber. 5 bottom chambers filled w/
FeSO4 and sand, 5 top chambers w/ lime and sand.
A variety of configurations simulated u sing FLOWNET
ver. 2.0, a 2-D steady-state flow model based on dual
formation of flow.
First environmental application for removal of chlorinated
organic compounds from aqueous solution.
VOC degradation rates in GW through 7' Fe canister. 42.9
h RT determined 2.2' width wall. Steel plates divided
-220 tons granular Fe in center from outer pea gravel. 4
monitoring wells downgradient; 2 piezometers
upgradient.
Open batch system of 1000 mL. FtSQ, or NaOH used to
adjust pH. Optimal pH 2-3, well mixed, ratio H2O2 to iron
0.001 M:l g/L
250 mL sealed bottles w/ organics, nutrients, pH 7.0,
Shi prock bacteria. Columns w/ str aw, alfalfa, sawdust,
sand (25% f,c). GW 1 5 mL/ h ( 1 d RT), 1 PV bacteria
Steel nuts put in barrel immersed in 2-L so In containing 90
ppm Cr(VI); at 16 rpm.
Determine e transfer during reductive dehalogenation .
Evidence from stereospecificity of reductive elimination
of vicinal dihalide stereoisomers synthesized in lab.
Results
Ist-order TCE rates and products in > 250 PV in Fe/Ni. In 76 PV
rates accelerated above untreated Fe. Catalytic dehydrohalo-
genationto hydrogenation caused enhancement. But, decreased
until rate similar to untreated iron. Gray precipitate after 100 PV
GW (250 mg/L carbonate). Fe catalysts prone to deactivation.
Similar losses not shown in granular Fe(0) systems.
Cd(II) 5 mg/L at pH 7, flow 1.6 mL/min. ~ 8,000 bed volume of
synthetic waste tr eated before breakthrough of Cd(II).
Redox front coincided w/ precipitation of ferrous iron by contact
w/ Ca(OFI)2. Mo & U successfully removed for 6 & 9 d, respec.
U reduced to UO; & precipitated as CaUO4 from elevated pH.
Mo reduced to Mo3Os or MoS2, or precipitated as FeMoO4.
2-D model shows width of capture zone proportional todischarge
through gate. Most efficient configuration is sides 180° apart,
oriented perpendicular to the regional hydraulic gradient.
Catalyzed metallic iron powder was shown to degrade a wide
range of halogenated organic contaminants.
First commercial in situ iron wall treating VOCs at former
semiconductor facility. Design & construction from 1 1/94 to
1/95. Operation and regulatory issues summarized. Monitoring
shows water quality objectives being met.
H2O2/iron powder is better than the Fentoris reagent due to
continuous dissolution of ironpowder and dye adsorption to
powder even though Fentoris reaction major decolorizing agent.
Sulfate, nitrate, U(VI) monitored 90 d. Precipitated U(IV)
crystalline UO2(s). Batch and column results support use of
cellulosic substrates as candidate barrier materials.
After 7.2 h, final Cr concentration < 0.5/zg/mL.
me,so2,3-dibromopentene — ^ >95% frara-2-pentene; D,L-2,3-
dibromopentane >95% ds-2-pentene. Reduction at metal surface
where 2 e transferred w/ no free radical intermediate.
Reference
Sivavec, T.M., et al. hternat'1
Contain. Tech. Conf. & Exhih
St. Petersbuig, EL, Feb 9-12
(1997)
Svavec, T.M., et al., 213th
National ACS Meeting,
37:83-85 (1997)
Smith, E. H., Emerg. Technol.
in Haz. Waste Mngmt. VII,
Atlanta, GA, pp. 1205(1995)
Spangler, R.R. & S.J. Morrison,
Pasco, WA, U.S. DOE Report
(1991)
Starr, R.C.& J.A Cherry,
aoundWater, 32(2):465
(1994)
Sweeny, K.& J. Fischer, Patent
3,649,821 (1972); Patent
3,737,384 (1973); Patent
4,382,865(1983)
Szerdy, F.S.,etal, ASCE
National Convention, Nov 1 2-
14, Washington, D.C. pp. 245-
256(1996)
Tang, W.Z.&RZ Chen,
Chemosphere, 32(5):947-958
(1996)
Thombre, M.S., et al., Internat'l
Contain. Tech. Conf. & Exhib
St. Petersburg, FL, Feb 9-12,
pp. 744(1997)
Thornton, R.F., Patent
5,380,441 (1995)
Totten, L.A& A.L. Roberts,
209thNational ACS Meeting
Anaheim, CA, April 2-7,
35:706(1995)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Fe
Fe,Zn
Fe
Organic mix,
anaerobic bacteria
Methanogenic
bacteria and Fe
Fe Kinetics
Contaminant
CO,
Chlorinated
organic s
PCE, TCE,
cDCE, VC
chlorinated
solvents
PCE, TCE,
cDCE, VC
CO,
PCE
Mine Waste
CHC13
dibromopentane
Type test
Column
Reactive
transport
model, ksA
Column
Column
Pilot- scale
Test
Batch
Surface
chemistry
Batch
Batch
Batch
Description/Conditions
1st column: CO, in air -sat. DI water. All O2 consumed by
Fe(0) giving anoxic downgradient region. 2nd similar but
Fe zone longer and diluted by mixing with sand.
Average estimates of rate constants. Assume reductive
dechlorination. Deviations with longer exposure due to
precipitates. Ist-order predictions vulnerable to changes in
mechanism or rates for less reactive constituents.
NJ site GW. Maj or VOCs- PCE < 5 0 mg/L, TCE < 3
mg/L. TDS 425-450 mg/L. 100% iron in column.
Column using site water and Single Layer Analytic
Element Model to evaluate treatment zones, flow
velocities, and residence times.
NJ site. Above-ground reactor influent PCE of 30 mg/L.
RT 1.1 d. Flowrate 0.5 gpm for 3 mo.
Kinetics dependent on pH, SA of metal, CC14 cone., buffer
and solvent composition (volume fraction 2-propanol).
XPS identified surface elements, valence state. Alfa Aesar
Fe (10 x 1mm) cleaned (HA, hydrofluric acid, H2O).
Vapor purified PCE, adsorbed to an Fe (100) single
crystal (10 mm in diameter and 1 mm in thickness)
purchased from Alfa Aesar by exposing 1 x 10"' Torr 100
s or 5 x 10"' Torr 200 s atroom temp.
Sealed, glass flasks simulated mine drainage, organic
mixtures, measured permeability. Consortium of bacteria
from creek facilitated reducing conditions & degradation.
Limestone ensured optimum pH; anaerobic conditions.
Anaerobic, 200 rpm, 20° C, 25 mL, methanogenic culture,
iron powder, iron filings, steel wool. CHC13 tested in iron-
cell, iron-supematantand resting cell.
Synthesized stereoisomer to demonstrate e- transfer.
Results
DO reacts with Fe(0) slowing dechlorination. But oxygen from
air accelerated reactions, possibly due to pH effect from
carbonate, changes in pathway, catalytic role of Fe2*, or O2
creating active corrosion sites.
Pseudo 1 st-order rates normalized to Fe-SA(kSA). Solvents com-
pared over range of conditions. kM varies by concentration, Fe
type, etc. Representative kSA,s and reactive transport model
calculate minimum barrier width for flow velocities and
halocarbon.
tU2 PCE, TOE, cDCE, VC = 0.5, 0.5, 1.5, 1.2 h, respectively. 2nd
test similar for PCE, TCEbutcDCE = 3.7& VC 0.9 h.
Corrosion increased pH & promoted precipitation of CaCO3,
FeCQ,, FeOH2.
GW flow model & degradation rates to design and estimate cost
for full-scale fiinnel-and-gate system at shallow sand aquifer (30-
40 ft) at Army Ammunition Plant, MN.
Assumptions: Time for PCE degradation sufficient for any TCE
to degrade; 10% cDCE, 1% VC from PCE & TCE degradation.
Reduced CC14 to chloroform in few h. Rate was 1 st-order with
respect to CCL, at concentrations <7.5 mM.
PCE adsorbed to metal surface, activatedbychemisorption. Cl-
from e transfer from Fe to adsorbed species. Adsorbed water can
dissociate and provide FT for C surface species from PCE
dissociation. Hydrocarbon can be produced from this reaction.
Reactivity& permeability (> 10"3 cm/s) suitable. Higher sulphate
reduction rates and longer effectiveness from organic mixture.
Geochemical reactive and transport models will be used to assess
effectiveness in treatment of mine drainage using reactive walls.
k = 0. 1 1 (Fe-cell), 0.003 (Fe-supernatant) & 0.007 hf ' (resting
cell). Biodehalogenation to abiotic reactions 37:1. Biocorrosion
of Fe & biodehalogenation of CHC13 via cometabolism using H*
from H2O.
Experiments indicate reduction takes place at metal surface.
Reference
Tratnyek, P.O., etal.,
EmergingTechnol. in Haz.
Waste Manag. VE. Atlanta,
pp. 589(1995)
Tratnyek, P. G. et al., GWMR,
Fan, pp. 108(1997)
Vogan, J.L., etal., (1994)
Vogan, J.L. et al., 87th Ann.
Mtg, Air & Waste Manag.
Cindnatti, OH, (1994)
Vogan, J., et al. 209th National
ACS Meeting Anaheim, CA,
April 2-7, 35:800(1995)
Warren, K.,etal.,J. of Haz.
Mat., 41:217 (1995)
Wang, C.-B.& W.-X Zhang,
2 13th ACS Meeting, San
Frandsro, CA, 37:163-164
(1997)
Waybrant, K. R.,etal,
Sudbury '95, Mining and the
Environ. CANMET, Ottawa,
Ontario, 3:945-953(1995)
Weathers, L.J., etal., 209th
Nat'l ACS Meeting, Apnl 2-7,
Anaheim, CA, 35:829 (1995)
Weber, E.J., 209th Nat'l ACS
Anaheim, CA, 35:702 (1995)
-------
Appendix A. continued..
Metal
(zero valent
unless specified)
Fe
Fe
Fe
Nano-Fe or Pd/Fe
Particles
Pyrite fines
Contaminant
4-amino-azo-
benzene (4-
AAB)
1,2-DCE, TCE,
Freon, VC
Nitrate
PCB, TCE
Cr(VI)
Type test
Batch
Treatment
Wall
Batch
Batch
Batch
Description/Conditions
4-AAB has reactive amino group for attaching molecule
to nonreactive surface. Reducing azo linkage suggests
aqueous reductant, if not, then surface- mediated process.
Granular Fe(0) & gravel at former semiconductor facility,
San Francisco. GWflow l'/d. 2 d RTrequired for VC.
150 mL flasks in air-dry & sat. (parafilm seal) 3 d.
Samples exlracted in 50 mL DI. Purged with N2.
"Artificial" soil contain, with nitrate (10 g clay + 10 g sand
+ 1 0 g Fe powder + 1 0 mL of 50 mg-NCV-N/L) .
20 mg/L TCE, 1.0 gnano-Fe or Pd/Fe in 50-mL vial
sealed, 30 rpm. PCBs, 50 /zL 200 /zg/mL Aroclor 1 254 in
methanol w/ 1 mL ethanol/water ( 1/9), 0. 1 g wet Fe or
Pd/Fe in 2-mL vial, 30 rpm, 17 h. Commercial Fe tested.
Pyrite fines collected near mine area. Pyrite crushed to -45
/zm. Used mixed batch reactors.
Results
Circumvented surface mediated contactto surface of Fe(0) by
adding appropriate water soluble e mediators. Species that can
function as e mediators were found present in the soil.
Wall 40' long, 20 to 7' bgs, Fe 4' wide allow 2 d RT. Slurry
walls east and west side for hydraulic control. 4 mon. wells in
wall.
94.4% Nitrate removal using 0.01M HEPES & 6% (w/v) Fe(0).
Treating 60 mg NCV-N/L w/ 6%Fe(0) at pH 1.0 transformed all
in 24 h; nitrate transformation is inversely related to pH. Nitrate
reduced to ammonium. Optimal removal at Q NQj'-N = 50-80
mg/L. In soil up to 97% transformation in air-dry samples, 99%
in wet samples; 2% in controls.
Synthesized sub-colloidal metals. Nanoscale Fe more reactive
than commercial Fe powder, due to high S A, active Fe surface,
less surface coverage by iron oxide layer. Nanoscale Pd/Fe more
active than pure Fe but nano-Fe inactivated by Fe oxide
formation.
Pyrite found to act as efficient Cr(VI) reducing agent. The Cr(III)
hydroxide precipitated onto pyrite particles.
Reference
Weber, E.J., ES&T, 30:716
(1996)
Yamane, C., et al., 209thNat'l
ACS Meeting Anaheim, CA,
April 2-7, 35:792(1995)
Zawaideh, L.L., etal.,
HSRC/WERC Joint Conf. en
the Environment, May 20
(1997)
Zhang, W.-X. & C.-B. Wang,
2 13th ACS Meeting, San
Frandsro, CA, 37:78-79
(1997)
Zoiboulis, A., et al., Wat. Res.
29(7):1755(1995)
-J
OJ
-------
Appendix B. Explanation of Relevant Physical/Chemical Phenomena
Corrosion
Corrosion processes are the required chemical underpinning of contaminant remediation by metallic iron. Until
recently, the science of corrosion was concerned almost exclusively with studying the process in order to minimize.
and if possible eliminate, corrosion reactions. Corrosion results in rusting of the iron and steel in cars, pipes,
bridges, buildings, and other structures. It is perhaps the most expensive aspect of infrastructure deterioration in
modem society. Although corrosion is familiar to everyone as rusting, few are aware of the extremely complex
reactions occurring within the corroding metals, at their surfaces and, in the case of electrochemical corrosion, in
the surrounding electrolyte solutions. To understand how these reactions can be used for remediating contaminants
requires at least a fundamental awareness of the manner in which corrosion reactions proceed.
Zero-valence-state metals, such as metallic iron (Fe(0), can serve as electron donors for the reduction of oxidized
species (Scullcy, 1975). These metals arc unstable in the natural environment and have to be created using high-
temperature metal refining processes (Evans, 1960; Sculley, 1975; Snoeyink and Jenkins, 1980). Zero-valence-state
metals tend to revert to a form that is more thermodynamically stable; for example, iron metal oxidizes to Fe2O3 in
the earth's oxygen-rich atmosphere. At low temperatures the rate of simple atmospheric oxidation of iron and steel
is negligible, however, due to the formation of oxide films that inhibit further surface exposure.
When a metal is immersed in an aqueous salt solution, as would be the case for a reactive barrier of iron chips or
filings in an aquifer, an electrochemical corrosion mechanism will occur. Electrons are given up by the metal in one
area (the anodic region) forming soluble cations of the metal, and taken up by oxidized species that become
reduced, at another part of the metal surface (the cathodic region). The instability' of the iron itself can provide the
necessary energy for oxidation-reduction reactions without external energy input, provided suitable coupled
electron-accepting reactions can occur with reducible species at the cathode.
Typically dissolved oxygen is the preferred oxidant, or electron acceptor, during aerobic corrosion processes.
These systems can, however, become anoxic or anaerobic if oxygen is depleted by the reactions. When present,
inorganic contaminants such as chromate (CrO42~) or highly halogenated organic compounds such as PCE and TCE
can serve as the oxidants, accept electrons, and become reduced. As long as electron acceptors are present,
corrosion processes and electron transfer within the metal can continue.
An electrochemical corrosion cell (ECC) can form in a number of ways, including:
1) the simple contact of two different metals. One will become the anode, the other the cathode. The
position of the metals in the galvanic series determines the direction of electron flow; i.e., which
becomes the anode and which the cathode.
2) when anodic and cathodic regions develop on the same metal surface. This can result from compositional
variations within the metal (i.e., other metal contaminants or inclusions with differing galvanic
potentials), surface defects, differences in grain structure orientation, stress/strain differences, and
chemical variations in the surrounding electrolyte solution (Evans. 1960; Sculley. 1975; Snoeyink and
Jenkins, 1980; Adamson, 1990).
These electron transfer processes within a metal or between contacting metals are said to occur within the external
circuit of the electrochemical cell. An internal circuit is also required to complete the cell. This requirement is
fulfilled by the contacting electrolyte solution. This electrolyte can consist of water containing salts (such as ground
water) and reducible, (i.e.. electron-accepting), solute(s). In some cases the water itself can accept electrons via
reductive dissociation. This is easily observed when the zero-valent metal is very low in the galvanic series, such as
Mg. When Mg metal is added to even deionized water (i.e., no solutes present), bubbles of H, gas rapidly appear on
the Mg surfaces. Mg is so low in the galvanic series that the water itself serves as the electron acceptor and is
rapidly dissociated. As you proceed to metals higher in the galvanic series, the dissociation of water becomes less
and less pronounced.
The ECC is basically a low-power version of the same type of circuit that is established in batteries. Figure 18
shows a simple ECC in a beaker of electrolyte solution with the external circuit resulting from the contact of two
dissimilar metals, tin and iron. The electrons travel through the external circuit from the anode (Fe°) to the cathode
(Sn°), where they can reduce oxidized species.
B- 74
-------
External
Circuit
(contact oT —-
dissimilar
metals)
^f—
^
— — — " — \
:~_/r~ -^.^
Fe
T3
II
^
""—f
e" i
— :
O :!
-t-» :
OJ :
o i
^ —
Corrosion Pit
/
-MH+ + 2e-->H
Internal Circuit
(electrolyte)
Figure 18. Example of an electrochemical corrosion cell.
Some of the potential corrosion and contaminant reduction reactions in zero-valent iron systems are:
Fc° —> Fe2+ + 2c~ Anode
2H+ + 2e- -> H,. , Cathode
2 (gas)
and, when oxygen is present (aerobic system)
4Fe°--> 4Fe3++ 12e~
3O, + 12c-
Net reaction
Anode
Cathode
Net reaction
(25)
(26)
(27)
(28)
(29)
(30)
4Fe° + 3O2 + 12FT -> 4Fe31 + 6H2O
where the increase in pH due to proton consumption (Equation 30) results in the precipitation of the Fc3+ as
Fe(OH)3. Should chromate be present as an oxidized species
Fe° -> Fe3^ + 3e~ Anode (3 1)
CrO/ + 4H2O + 3c- 4-> Cr(OH)3 + SOff Cathode (32)
Fe°+ CrO42- + 4H2O -> Fe(OH)3 + Cr(OH)3 + 2 OH- Net reaction (33)
Gillham and O'Hannesin (1994) have proposed that the reductive dechlorination of chlorinated organic compounds
by iron metal corrosion may proceed as
2Fe° -> 2Fe21 + 4e" (34)
3H2O <-> 3ff- + 3OH- (35)
(36)
X-H
2Fe° + 3H,O + X-C1 <~> 2Fc2+ + 3 OH- + FL, , + X-H + Cl~
2 2 (gas)
(37)
(38)
B- 75
-------
Which of these reactions, or others, is dominant will depend on the conditions and contaminants present; for
example, the presence or absence and the concentrations of other reactive species (including the partial pressures of
gases such as O2 and CO2) and mineral surfaces, pH, etc. It should be noted that these reactions tend to increase the
pH of the corrosion system. It is also important to note that although chromate is reduced and chlorinated
hydrocarbons are reductively dechlorinated in the presence of Fe°, current research indicates that the mechanisms
of the reactions are different.
Sorption. Adsorption. Dissolution. Precipitation Processes in Zero-Valent Iron Systems
As a contaminant plume moves through the subsurface, chemical processes occur that can affect both contaminant
concentration and the overall hydrogeochemistry of the system. Three of the most important physical processes are
sorption, dissolution, and precipitation. These processes also occur in iron PRJBs. where the radical geochemical
changes can exert profound effects on the ground water and contaminants.
Sorption and adsorption are loosely described as processes wherein chemicals partition from a solution phase into
or onto, respectively, the surfaces of solid phase materials. Both sorption and adsorption at particle surfaces tend to
retard contaminant movement in ground water. Retardation of the contaminant at the iron surface is a positive
result, allowing increased time for reactions.
In the subsurface context, sorption usually implies movement, or dissolution, of a chemical into a surface coating
on an aquifer material mineral grain. For example, hydrophobic organic compounds such as PCE tend to be sorbed
into organic carbon coatings on mineral grains. In fact, evidence seems to indicate that graphitic inclusions/coatings
on some granular iron surfaces might enhance reductive dechlorination relative to iron without such organic
materials being present at its surfaces. It is possible that sorption of the chlorinated compounds to the graphitic
constituents increases the contact time between the contaminant and the iron surface. This would allow increased
reaction time and better proximity for the requisite electron transfers to occur.
Adsorption implies attachment of a chemical to reactive sites on mineral surfaces. These sites usually result from
an excess or either positive or negative charge on the surfaces. These surface charges can be constant (fixed) due to
ion substitutions in the mineral matrix (isomorphous substitution), variable with pH, or a mixture of both. In
addition, the adsorption can result from cither inner-sphere or outer-sphere complcxation. Inner-sphere complcx-
ation is due to actual covalent and ionic chemical bond formation. In outer-sphere complexation adsorption results
from ion-pair bonding due to electrostatic forces and hydration water separates the solvated ion from the surface.
Many metal oxides and some clay minerals have net surface charges that vary with pH due to the proportion of
protonatcd versus dcprotonatcd surface sites. Among these variably-charged materials arc the iron oxyhydroxidcs
(rusts) that result when zero-valent iron corrodes. These materials are very significant to adsorption of both
inorganic and organic charged solution species (ionic species). The charges of both the surface and the solution ion
control whether adsorption will occur or whether the surface and the ion will repulse one another. The pHz , or pH
of zero-point of charge, is the pH at which negatively and positively charged surface sites exist in approximately
equal numbers on the mineral. Above pHzic the surface will have a net negative charge, enhancing cation
adsorption; below pH x the surface will have a net positive charge, enhancing anion adsorption. Research has
shown that the reduction of the negatively-charged Cr(VI) chromate ion (CrO42") to Cr (III) by zero-valent iron
occurs more rapidly when the system pH is below the pH. = 8.5 of the iron oxyhydroxidc rusts; i.e., the rusts have
a net positive charge (Powell et al., 1995). This increase in reduction is probably partially due to the contaminant
being maintained at the iron surface by adsorption.
Although flowing subsurface systems are not in true chemical equilibrium, they can establish a hydrogeochemical
pseudo-steady state condition. This condition can be altered when a contaminant front/plume passes through (Puls
and Powell, 1992). Contaminant fronts typically have pH. Eh. ionic strength, chemical species and complexants,
and other features that differ from the intruded pseudo-steady state system. These differences can, among other
things
* dissolve cementation between the mineral grains,
• change the adsorptive nature of the solid phase (potentially causing charge reversal and/or allowing desorption
of previously immobilized species and/or adsorption of previously mobile species),
* alter mineral-bound and aqueous elemental oxidation states, or
* precipitate new phases onto the solid surfaces.
B-76
-------
The implementation of a reactive iron barrier wall in the subsurface can be viewed as having effects on the aquifer
in the vicinity of the w7all that are analogous to a plume moving through the zone. Intra-wall and downgradient
effects will include radically lowered Eh and loss of dissolved oxygen, higher pH, and increases in Fe2+ that will
precipitate downgradient possibly as iron oxyhydroxide colloids. Loss of cementation and precipitate formation
can generate mobile colloidal particles, that can themselves transport adsorbed contaminants, or contain toxic
materials that were naturally occurring in the geologic matrix. Additional research is needed on these disruptive
effects to determine whether certain types of intercepted contaminants can be transported colloidally or, alterna-
tively, whether naturally-occurring but immobilized metals in or on the aquifer minerals might be liberated and
mobilized by the dramatic geochemical changes.
The dissolution and precipitation processes associated with the iron metal are, however, essential for contaminant
remediation. The reduction of Cr(Vl), for example, occurs primarily due to the anodic dissolution of Fe2+ during the
corrosion process (Powell et al., 1995). The subsequent immobilization of the resulting Cr(FH) is due to
precipitation of the chromium as hydroxides or mixed iron/chromium oxyhydroxide solid solutions. Although the
chlorinated hydrocarbons are not precipitated during remediation, dissolution of the iron still occurs as Fe(0) yields
electrons to the hydrocarbon and the resultant Fe2+ is exposed to the solution. Understanding the dissolution/
precipitation geochemistry of these zero-valent iron systems and the contacting aquifer materials is an area of
ongoing research.
B-77
-------
Appendix C. Calculations
Cost-effective use of permeable reactive barriers for ground-water treatment requires proper estimation of the
amount of reactive material required and choosing the best means of emplacing it in the ground. The weight of
reactive material per unit cross-section of the plume may be estimated from laboratory reaction kinetics data and
basic knowledge of the plume and the remediation goals. The value of this parameter has implications regarding the
choice of permeable barrier design and emplacement method. The use of tremie tubes, trenching machines, high-
pressure jetting, and deep soil mixing may be appropriate for different situations, depending on the amount of
reactive material required, the dimensions of the plume, and other factors. The specific application considered here
is granular iron to treat ground water contaminated with chlorinated solvents, but the principles are applicable to
other types of media and contaminants.
Reaction rate parameters from laboratory studies of iron-mediated degradation of a variety of chlorinated solvents
have appeared in the literature in the past several years (Johnson et al., 1996; Shoemaker and al., 1996). The work
of Johnson et al. (1996) has been especially helpful in establishing the high degree of consistency between kinetics
data obtained from batch and column studies. By expressing rate data in a way that accounts for the iron surface
area concentration, it was demonstrated that results reported in the literature varied by less than had previously been
thought. This makes it possible to obtain a fairly reliable estimate of the iron requirement for a potential application
even before site specific laboratory feasibility tests arc conducted. The bulk of the data reviewed by Johnson ct al.
suggest a surface-area-specific rate parameter (kgv) of about 0.2 cm3 h~' nr2 for TCE and of about 0.04 cm3 h~' nr2 for
cDCE. For the examples considered here, 1.0 m2/g will be used for the specific surface area, a value typical of the
granular irons which currently appear to be the most practical for permeable barrier applications. Further, the rate
of reaction will be decreased by 50% to adjust for subsurface temperatures being lower than room temperature
(Sivavec and Homey, 1995). Therefore, the effective rate parameter to be used is 0.1 cm3 g~' h~' for TCE and
0.02 cm3 g-1 h-1 for DCE.
Two example cases are considered below. The first and simplest involves degradation of a chlorinated compound
(e.g., TCE) where the levels of intermediate products (e.g., DCE) are low enough that they do not influence the iron
requirement. The second case involves significant generation of an intermediate product that degrades more slowly
than the parent and thereby determines how much iron is required.
Case 1; Parent Products Only
The rate of reaction may be expressed as
dP
-£ = -lP.P (39)
where P is the concentration of dissolved chlorocarbon, t is the contact time between the dissolved chlorocarbon
and iron particles, kt is the first-order rate parameter, and pm is the mass of zero-valent iron particles per solution
volume. This equation may be integrated to give
where PQ is the initial concentration of dissolved chlorocarbon. In a batch laboratory experiment, kt may be derived
from the slope of a semi-log plot of PQ/P vs. time.
For the case of steady-state flow in a packed bed reactor, an expression analogous to Equation 40 may be derived
by expressing the residence time (t) as the product of the bed void fraction (e) and the reactor volume (V), divided
by the liquid flowrate through the bed, yielding
The term p eV is the mass of zero-valent iron, W, that the fluid encounters as it flows through the bed. With this
rln ' ? o
substitution, and by representing the flowrate as the product of the cross-sectional plume area (A), the soil porosity
(n), and the average flow velocity (u), the amount of iron required per unit cross-section of plume to effect a desired
decrease in chlorocarbon concentration may be expressed as
C-78
-------
A k,
P
(42)
Tliis is a useful expression because it allows estimates to be made without assuming a particular design (such as
funnel-and-gate) or calculating parameters such as residence time, but rather expresses a key aspect of the design
(W/A) in the most fundamental terms. However, it does not reflect uncertainties and fluctuations in parameter
values that must be considered in any design. These can be accounted for in terms of a factor of safety (F) which
increases the amount of reactive material employed:
A k,
P
(43)
A Monte-Carlo simulation has been developed to estimate appropriate factors of safety for permeable reactive
barrier systems (Eykholt, 1997). With influent concentrations varying 10%, the reaction rate parameter varying
30%, and the ground-water velocity varying 100%, achieving a 1000-fold decrease in contaminant concentration
with 95% confidence was found to require a safety factor of 3.5.
As shown in Table 9. calculations based on a safety factor of 3.5 and a range of practical values for reaction rate
parameters and ground-water velocities suggest that W/A should be expected to vary from as little as about 20 lb/ft2
to perhaps 1,000 lb/ft2.
Case 2: Significant Intermediate Generation
If significant amounts of intermediate products arc generated during the degradation of a parent chlorinated
compound, the slower rate of degradation of the intermediate product may be the factor mat determines how much
iron is required (Focht et al.. 1996). The kinetics of intermediate product generation and degradation may be
expressed as
dP
-^7
dt
(44)
—=aklpmP-k2pmD
(45)
Table 9. Required Wcight-pcr- Area (W/A) of Granular Iron. Calculated with n=0.33, P0/P=1000, andF=3.5. Information
arid Equations from Appendix C.
ki
(cmVh-1)
0.1
0.02
u
(ft/day)
0.1
0.25
0.5
1
0.1
0.25
0.5
1
W/A
(lb/ft2)
21
52
100
210
100
260
520
1000
C-79
-------
where D is the concentration of the intermediate product, is the fraction of parent compound which appears as the
intermediate product, and k2 is the rate parameter for intermediate product degradation. These equations can be
solved to vield
= \Dn
a
kl-K2
(46)
where D0 is the initial concentration of the intennediate product.
For most situations involving chlorinated compounds, k2 is substantially less than kr For the time domain where
net degradation of the intermediate product is occurring, the first term on the right side of Equation 46 dominates,
therefore the concentration is approximated as
(47)
Making the same substitutions used in deriving Equation 42 yields
W _ra/,
— ~F—In
DO
D
,
D
(48)
As TCE degrades. DCE and vinyl chloride often appear in solution at concentrations corresponding to a few
percent of the TCE originally present (Focht et al., 1996). Depending on the actual amount of conversion (a) and
the values of the other parameters, Equation 48 may indicate the need for more reactive material than would be
suggested by Equation 42. For example, with no intermediate products initially present and with k2/kl = 0.2,
-------
Each of these W/A values can be understood in site-specific terms. Those sites with low W/A are examples of sites
with low ground-water velocity and without significant concerns regarding slow-reacting intermediate products.
The Sunnyvale, California, site has a high W/A value for several reasons. First, the reaction rate parameter (k) is
low because the principal contaminants, DCE and VC, degrade much slower than TCE. Second, the ground-water
velocity is relatively high, at approximately 0.8 ft/day. The safety factor applied in this case was 4 (Warner et al..
1995).'
Cost Estimation
Characterizing a plume in terms of the weight-per-area of reactive material required lends itself to evaluating costs
on a per-area basis. Installation costs for impermeable barriers are often expressed in this manner. For example,
shallow impermeable barriers are often quoted to cost between $10/ft2 and $25/ft2 excluding mobilization. In this
section, it is shown how this practice may be extended to permeable reactive barriers.
Many of the iron PRBs installed thus far are of the funnel-and-gate design, in which one or more discrete
permeable trenches ("'gates'') is installed along with impermeable containment walls ('"funnels") to direct ground
water through the gate. The funncl-and-gate cost analysis presented here will focus on three elements: the iron
itself, the creation of the gate, and the funnel. All incremental cost components will be expressed on a per-area
basis. It is important to understand that these costs are on the basis of the total cross-sectional area of the plume
being treated, not the area of the gate.
The cost of reactive material on a per-area basis is simply the product of its cost per weight and the wcight-pcr-arca
(W/A) required:
^Reactive materials^
cost per plume
cross - sectional
area
(W
= (Cost per weight)x\ —
v F U; (49)
A typical price for the granular iron currently being used for PRBs is $375/ton. This translates into incremental
costs of $9/ft2 for a W/A of 50 lb/ft2, $47/ft2 for a W/A of 250 lb/ft2, and $188/ft2 for a W/A of 1,000 lb/ft2.
The cost of installing a gate is most readily quoted on a per-volume basis. To obtain the gate cost on the basis of
plume cross-sectional area, this value is multiplied by the required gate volume per plume area. The gate volume
required per plume area is simply W/A divided by the bulk density of granular iron. So,
Gate installation
f Installation costA W/A
cost per plume = I I
I per volume )
cross - sectional area
x
where pb is the bulk density of granular iron, typically about 160 lb/ft3. In most funnel-and-gate systems installed
thus far, gate installation has proved expensive. By using a trench box method and confined entry procedures for the
final stages of installation, costs of $1.000/yd3 have not been unusual. This corresponds to a cost of $12/ft2 to install
50 lb/ft2 of iron gate, $58/ft2 to install 250"lb/ft2 and $23 I/ft2 to install 1,000 lb/ft2.
Tables 11 and 12 present cost estimates for several scenarios. In each table, costs are estimated for three values of
W/A: 50, 250, and 1,000 lb/ft2. For Table 11, relatively high installation costs are used: $25/ft2 for funnels and
$l,000/yd3 for gates. For Table 12, significantly lower costs were assumed: $10/ft2 for funnels and $200/yd3 for
gates.
Comparing Tables 11 and 12 demonstrates that there is considerable incentive for employing less costly means of
installing both the impermeable and permeable components of funnel-and-gate systems. Further cost savings may
be realized in some cases by designing continuous PRBs, thereby eliminating the gate installation cost altogether,
assuming that the costs of installing continuous permeable barriers arc similar to the costs of installing impermeable
barriers (funnels).
C-81
-------
Table 11. Cost Elements of Funnel and Gate Systems: High Estimate (Gate @ $1000/yd3; Funnel @ $15/ft2)
Component
Iron
Install gate
Funnel
Sum
W/A =
50 lb/ft2
9
12
25
46
Incremental Costs
($/ft2)
W/A =
250 lb/ft2
47
58
25
130
W/A =
1000 lb/ft2
188
231
25
444
Table 12. Cost Elements of Funnel and Gate Systems: Low Estimate (Gate @ $200/yd3; Funnel @ $10/ft2)
Component
Iron
Install gate
Funnel
Sum
W/A =
50 lb/ft2
9
2
10
21
Incremental Costs
($/ft2)
W/A =
250 lb/ft2
47
12
10
69
W/A =
1000 lb/ft2
188
46
10
244
C- 82
-------
Appendix D. Acronyms
AFB
AI&M
ASF
bgs
C.
Co
CF
CM
CT
DCE
tDCE
cDCE
DCM
DI
DO
DOD
DOE
DNAPLs
DL
e-
ECC
EDX
fc
F&G
FAME
GAC
GC/MS
GPR
GW
h
HC
HCB
HPLC
ISTZ
ITRC
k
MBS
MCL
MW
NACE
NAS
ND
NERL
NMR
NRMRL
O&M
Air Force Base
Ada Iron & Metal
Anderson-Schulz-Flory distribution
below ground surface
initial contaminant concentration of the influent solution
contaminant concentration in solution at time t = 0
chloroform (CHQ3)
chloromethane (CH3C1)
carbon tetrachloride (CC14)
dichloroethene (C2H2C12)
fraMs-dichioroethene (C2H2C12)
c/s-dichloroethene (C2H2C12)
dichloromethane (CH2C12)
deionized
dissolved oxygen
Department of Defense
Department of Energy
dense nonaqueous phase liquids
detection limit
electron
electrochemical corrosion cell
Energy Dispersive X-ray Analysis
fraction of organic carbon
funnel-and-gate
fatty acid methyl ester analyses
granular activated carbon
gas chromatography/mass spectrometry
ground-penetrating radar
ground water
hour
hydrocarbon
hexachlorobenzene
high-performance liquid chromatography
in situ treatment zone
Interstate Technology and Regulatory Cooperation Workgroup
rate constant
observed rate constant
surface-area-specific rate parameter
Master Builders Supply iron
maximum concentration limit
monitoring well
National Association of Chemical Engineers
National Academy of Sciences
nondetectable
National Exposure Research Laboratory
Nuclear Magnetic Resonance Spectroscopy
National Risk Management Research Laboratory
operations and maintenance
D- 83
-------
ORD
ORNL
OSHA
OSWER
PAC
PCE
PEI-silica
PRB
psi
PV
QAPP
rf
ROD
rpm
RSF
RT
RTDF
RWQCB
rxn
RZF
SA
SBR
S.H.E
SEM
SI
SIMS
SITE
SMAD cryoparticles
SMVs or SVOCs
soln
SPM
l\r
TCA
TCE
TCM
TCP
TDS
TER
no
TLC
TOC
UMTRA
US
USCG
vol or v
VC
VISITT
VOCs
VOAs
W/A
XRD
Office of Research and Development
Oak Ridge National Laboratory
Occupational Safety and Health Administration
Office of Solid Waste and Emergency Response (U.S. EPA)
powdered activated carbon
tetrachloroethene (perchloroethene)
polyethylenimine-coated silica beads
pemieable reactive barrier
per square inch
pore volume
Quality Assurance Project Plan
retardation factor
Record of Decision
revolutions per minute
reactant sand-fracturing
residence or retention time
Remediation Technologies Development Forum
Regional Water Quality Control Board (State of California)
reaction
reaction zone formation technologies
surface area
stirred batch reactor
standard hydrogen electrode
scanning electron microscopy
saturation index or indices
secondary ion mass spectroscopy
Superfund Innovative Technologies Evaluation
metal vapor-solvent codeposition method (in preparation of active metal surfaces)
semivolatile organics
solution
scanning probe microscopy
half life
trichloroethane
trichloroethene
tri chl oroni ethane
trichloropropane
total dissolved solids
Technology Evaluation Report
Technology Innovation Office (U.S. EPA)
thin layer chromatography
total organic carbon
Uranium Mill Tailings Remedial Action Program
ultrasound
United States Coast Guard
volume
vinyl chloride
Vendor Information System for Innovative Treatment Technologies
volatile organic compounds
volatile organic aromatics
weight per area
X-Rav Diffraction
D- 84
-------
XPS X-Ray Photoelectron Spectroscopy
y year
ZHE zero head-space extractors
zpc point of zero charge
ZVI zero-valent iron
D- 85
-------
anode
Arrhenius equation
bentonite
biotite
calcite
cell potential
cation
cathode
Eh
electrochemical cell
Fenton 's Reagent
goethite
half-life
hematite
iron metal
Appendix E. Glossary
The electrode in an electrochemical cell toward which anions are drawn and where
oxidation occurs.
An equation that expresses the logarithmic relationship between the rate constant of a
reaction and the reciprocal of the temperature (expressed in K).
The term bentonite is used as a commercial name for clays that are predominantly sodium
montmorillonite. Wyoming bentonite is the most common drilling fluid additive used in
the water well industry.
An important mineral of many intrusive igneous rocks, pegmatites, lamprophyres, some
lavas and metamorphic rocks.
CaCO3. A sedimentary mineral formed by chemical precipitation.
Specific Gravity ~ 2.95 g/cm3.
The voltage of an electrochemical cell.
A positively charged ion, attracted toward the cathode in an electrolytic cell.
The electrode in an electrochemical cell toward which cations are drawn and where
reduction occurs.
or Nernst equation, expresses the relationship between the standard redox potential of a
given redox couple, its observed potential, and the concentration ratio of its electron-donor
and electron-acceptor species.
A system containing an oxidation-reduction reaction in which oxidation and reduction
reactions are physically separated and the transferred electrons pass through an electrical
circuit.
A way to generate OH* by Fenton reaction:
Fe2^ +HA -^Fe3+ +OH'+OH- (1)
aq 2 2 aq v '
Irradiation with light X<580nm effects photoreduction of Fe3^ to Fe2^ together with the
production of OH* radicals
Fe3^. + HO + h
(2)
An iron mineral of the general formula FeOOH (alpha-FeOOH, see lepidocrocite).
Specific Gravity 4.28 g/cm3.
The time required for 50% of a material or compound to undergo transformation or decay.
An iron oxide mineral, Fe2O3, corresponding to an iron content of approximately 70%.
Specific Gravity ~ 5.26 g/cm3.
Variously designated as Fe°, Fe(0), or zero-valent iron. The most common reactive media
in the majority of field scale and commercial PRB implementations. Sources of Fe used in
experiments and installations referenced in this document are: Ada Iron and Metal,
Aldrich, Alfa Aesar, Peerless, Fluka, Fisher, VWR, MBS.
E- 86
-------
K
Tliis is the rate at which the contaminants are observed to degrade.
Normalized kobs to Fe surface concentration yields a specific rate constant, kSA for a
number of solvents.
Lewis acid
Lewis acid-base
lepidocrocite
MINTEQA2
monlmorillonile
Muscovite
Nernst equation
A species that accepts a pair of electrons to form a covalcnt bond.
theory The idea that an acid is a species that accepts a pair of electrons to form a covalcnt bond
and a base is a species that donates a pair of electrons to form a covalent bond.
gamma-FeOOH (see goethite)
A geochemical equilibrium speciation model for dilute aqueous systems. It is an update of
MINTEQ, that was developed by combining the fundamental mathematical structure of
MINEQL with the thermodynamic data base of WATEQ3. For more information see
http://www.cee.odu.edu/cee/model/minteq.html
An aluminosilicate clay mineral in the smectite group
[KAl0[AlSi3O10](OH,F),]. One of the more common minerals in rocks, especially plutonic
igneous rocks rich in silica or aluminum and low or medium to high grade mctamorphic
rocks. Specific Gravity=2.8 g/cm3.
An equation that expresses the exact electromotive force of a cell in terms of the activities
of products and reactants of the cell.
oxidation
(1) a reaction in which there is an increase in valence resulting from a loss of electrons.
Contrast with reduction. (2) A corrosion reaction in which the corroded metal forms an
oxide; usually applied to reaction with a gas containing elemental oxygen, such as air.
passivation
(1) A reduction of the anodic reaction rate of an electrode involved in corrosion. (2) The
process in metal corrosion by which metals become passive. (3) The changing of a
chemically active surface of a metal to a much less reactive state.
PHREEQC
A program for aqueous geochemical calculations. For more information go to
http://watcr.usgs.gov/softwarc/phreeqc.html
potential-pH diagram A plot of the redox potential of a system versus the pH of the system, compiled using
thermodynamic data and the Nernst equation. The diagram shows regions within which the
metal or mineral itself or some of its compounds are stable.
pyrite
An iron sulfide mineral with the general formula of (FeSJ. Specific Gravity=5.0 g/cm3.
Common in plutonic. volcanic, sedimentary and metamorphic rocks.
redox potential (Eh) The subscript H is used to emphasize that the potential only has meaning in reference to the
standard hydrogen electrode reaction.
saturation indices
siderite
Gauges the potential for minerals to precipitate. A negative SI indicates undersaturation
with respect to the particular mineral phase, while a positive SI indicates oversaturation.
An iron oxide mineral with the general formula FeCO3 Specific Gravity~3.95 g/cm3.
E-87
-------
surfactant A surface-active agent; usually an organic compound whose molecules contain a hydrophilic
group at one end and a lipophilic group at the other.
vermiculite A mica-like silicate mineral of the general formula (Mg.Fe2+,Al)3(Si.Al)4O10(OH),
zeolites Complex inorganic framework mineral structures made up from SiO4 and A1O4 tctrahcdra
joined together to form a series of interconnected channels and pores. Small molecules can
diffuse through the zeolite and undergo chemical reactions catalyzed by active sites in the
channel walls. The high acidity of their protonated form, coupled with the heterogeneous
and controllable nature of the reactions which proceed within them has made them ideal
choices for a wide range of catalytic processes.
E-
-------
References
Adamson, A. W. (1990). Physical Chemistry of Surfaces. New York. John Wiley and Sons. Inc.
Archer. W. and Simpson, E. (1977). "Chemical profile of polychloroethanes and polychloroalkcncs." Industrial
Engineering Chemistry' Product Research and Development 16(2): 158-162.
Archer. W. L. (1982). "Aluminum-l,l,l-trichloroenthane. Reactions and inhibition." Industrial Engineering
Chemistry' Product Research and Development 21: 670-672.
Archer. W. L. and Harter. M. K. (1978). "Reactivity of carbon tetrachloride with a series of metals." Corrosion-
National Association of Corrosion Engineers (NACE) 34(5): 159-162.
Article, J. and Fuller, W. H. (1979). "Effect of crushed limestone barriers on chromium attenuation in soils."
Journal of Environmental Quality 8: 503-510.
Baker. M. J., Blowes. D. W. and Ptacek, C. J. (1996). "Development of a reactive mixture to remove phosphorous
from on-site wastewater disposal systems." Waterloo Centre for Groundwater Research Annual Septic System
Conference. Disposal Trenches, Pre-treatment and Re-use of Wastewater, Waterloo, Ontario, May 13.
Baker. M. I, Blowes. D. W. and 'Ptacek, C. J. (1997). "Phosphorous adsorption and precipitation in a permeable
reactive wall: applications for wastewater disposal systems." International Containment Technology Conference
and Exhibition. February 9-12. St. Petersburg, Florida.
Ballard. S. (1996). "The in situ permeable flow sensor: A ground-water flow velocity meter." Ground Water 34(2):
231.
Barcelona, M. J., J. P. Gibb, J. A. Helfrich and E. E. Garske (1985). Practical Guide for Ground-Water Sampling.
U.S. Environmental Protection Agency, Office of Research and Development, Robert S. Kerr Environmental
Research Laboratory, U.S. EPA.
Battelle (1997). Design Guidance for Application of Permeable Barriers to Remediate Dissolved Chlorinated
Solvents. Prepared for the Armstrong Laboratory/Environics Directorate, U.S. Air Force.
Benner, S. G., Blowes, D. W. and Ptacek. C. J. (1997). "Porous reactive w7all for prevention of acid mine drainage:
results of full-scale field demonstration." International Containment Technology Conference and Exhibition.
February 9-12, St. Petersburg, Florida.
Bennett, T. A., Blowes, D. W., Puls, R. W., Gillham, R. W., Hanton-Fong, C. J., Ptacek, C. J., O'Hannesin, S. F. and
Vogan, J. L. (1997). "Design and installation of an In situ porous reactive wall for treatment of Cr(VI) and
trichloroethylene in groundwater." 2.13th National ACS Meeting, San Francisco, California. April 13-17.
Blowes. D. W. and Ptacek, C. J. (1992). "Geochemical remediation of groundwater by permeable reactive walls:
removal of chromate by reaction with iron-bearing solids." Subsurface Restoration Conference. 3rd International
Conference on Ground Water Quality' Research. June 21-24, Dallas, Texas.
Blowes, D. W.. Ptacek. C. J., Benner, S. G., Waybrant, K. R. and Bain, J. G. (1997). In press. "Porous reactive walls
for the prevention of mine drainage: a review7." Mineral Processing.
Blowes, D. W., Ptacek, C. J., Cherry, J. A., Gillham, R. W. and Robertson, W. D. (1.995). "Passive remediation of
groundw7ater using in situ treatment curtains." Geoenvironment 2000, Characterization. Containment.
Remediation, and Performance in Environmental Geotechnics. New York. American Society of Civil Engineers.
1588-1607.
Blowes. D. W., Ptacek. C. J. and Jambor, J. L. (1997a). "In-Situ remediation of chromate contaminated
groundwater using permeable reactive walls." Environ. Sci. Technol. 31: 3348-3357.
Blowes, D. W., Puls, R. W., Bennett, T. A., Gillham, R. W., Hanton-Fong, C. J. and Ptacek, C. J. (1997b). "In-situ
porous reactive wall for treatment of Cr(VI) and trichloroethylene in groundwater." International Containment
Technology Conference and Exhibition. February 9-12, St. Petersburg, Florida.
Blowes, D. W., Robertson, W. D., Ptacek. C. J. and Merkley, C. (1994). "Removal of agricultural nitrate from tile-
drainage water using in-line bioreactors." Journal of Contaminant Hydrology 15: 207-221.
Bostick, W. D., Shoemaker, J. L., Osborne, P. E. and Evans-Brown, B. (1990). "Treatment and disposal options for
a heavy metals waste containing soluble technetium-99." ACS Symposium Series 422. Emerging Technologies
in Hazardous Waste Management. 345-367.
Bowers, A. R., Ortiz, C. A. and Cardozo, R. J. (1986). "Iron process for treatment of Cr(VI) wastwaters." Metal
Finishing 84: 37.
89
-------
Charmichael, P. A. (1994). "Using wood chips as a source of organic carbon in denitrification: a column experiment
and field study implementing the funnel-and-gate design." University of Waterloo, Waterloo, Ontario.
Cheng, I. F., Muftikian, R., Fernando, Q. and Korte, N. (1997). "Reduction of nitrate to ammonia by zero-valent
iron." 213th ACS National Meeting, April 13-17, San Francisco, California.
Clausen, J. L.. Richards. W. L., Korte, N. E. and Liang, L. (1995). "ORNL/MMES research into remedial
applications of zero-valence metals. 3: Removal of TCE. cis-l,2-DCE, vinyl chloride, and technctium." 209th
ACS National Meeting. American Chemical Society, April 2-7, Anaheim, California.
dishing, R., D. Gravelding and P. B. McMahon. 1996. "Overview of the Denver Federal Center." Remediation
Technologies Development Forum, Permeable Barriers Action Team meeting. Denver. Colorado.
Del Cul, G. D., Bostick, W. D., Trotter, D. R. and Osborne, P. E. (1993). "Tcchnctium-99 removal from process
solutions and contaminated groundwater." Separation Science and Technology 28(1-3): 551-564.
Dubrovsky, N. M., Morin, K. A., Cherry, J. A. and Smyth, D. J. A. (1984). "Uranium tailings acidification and
subsurface contaminant migration in a sand aquifer." Water Poll. Res. J. Canada 19: 55-89.
Dvorak, D. H., Hcdin, R. S., Edcnborn, H. M. and McINtirc, P. E. (1992). "Treatment of metal-contaminated water
using bacterial sulfate reduction: Results from pilot-scale reactors." Biotechnology and Bioengineering 40:
609-616.
Eary, L. E. and Rai, D. (1989). "Kinetics of chromate reduction by ferrous ions derived from hematite and biotite
at 25°C." American Journal of Science 289: 180-213.
Eger, P. and Wagner (1995). "Sulfate-reduction for the treatment of mine drainage: long term solution or short term
fix?" Sudbury '95. Mining and the Environment. CANMET, Ottawa. Ontario.
Evans, U. R. (1960). The Corrosion and Oxidation of Metals: Scientific Principles and Practical Applications.
London, Edward Arnold (Publishers) LTD.
Eykholt, G. R. (1997). "Uncertainty-based scaling of iron reactive barriers." In situ remediation of the geoenvironment.
American Society of Civil Engineers, New York, 41-55.
Focht, R., Vogan, J. and O'Hannesin, S. (1996). "Field application of reactive iron walls for in-situ degradation of
volatile organic compounds in groundwater." Remediation Summer: 81-94.
Fruchter, J. S., Amonette, J. E., Cole, C. R.. Gorby. Y. A., Humphrey. M. D., Isok, J. D., Spane, F. A., Szecsody, J.
E., Teel, S. S., Vermeul, V. R., Williams, M. D., and Yabusaki, S. E. (1998). In Situ Redox Manipulation Field
Iniection Test Report-Hanford 100-H Area. PNNL-11372, UC-602. Battelle. U. S. Department of Energy,
Pacific Northwest National Laboratory.
Gallinatti, J. D. and Warner, S. D. (1994). "Hydraulic design considerations for permeable in-situ groundwater
treatment walls." Ground Water 32(5): 851.
Gillham, R. W. (1995). "Resurgence in research concerning organic transformations enhanced by zero-valent
metals and potential application in remediation of contaminated groundwater." 209th ACS National Meeting.
American Chemical Society, April 2-7, Anaheim, California.
Gillham, R. W. and Burris, D. R. (1992). "In situ treatment walls—Chemical dehalogenation, denitrification. and
bioaugmentation." Proceedings from the Subsurface Restoration Conference. Dallas. Texas. June 21-24, Rice
University, Department of Environmental Science and Engineering, Houston, Texas.
Gillham, R. W. and O'Hannesin, S. F. (1994). "Enhanced degradation of halogenated aliphatics by zero-valent
iron." Ground Water 32(6): 958-967.
Gould, J. P. (1982). "The kinetics of hexavalent chromium reduction by metallic iron." Water Res. 16: 871-877.
Grittini. C.. Maicomson, M.. Fernando, Q. and Korte, N. (1995). "Rapid dechlorination of polychlorinated
biphenyls on the surface of a Pd/Fe bimetallic system." Environ. Sci. Technol. 29(11): 2898-2900.
Hagenmaier, H.. Brunner, H.. Haag, R. and Kraft. M. (1987). "Copper-catalyzed dechlorination/hydrogenation of
polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans. and other chlorinated aromatic compounds."
Environ. Sci. Technol. 21: 1085-1088.'
Hocking, G. (1998). "Vertical Hydraulic Fracture Emplacement of PRBs." Summary of the Remediation Technologies
Development Foaim Permeable Reactive Barriers Action Team Meeting, April 15-16. Beaverton, Oregon.
ITRC (1997). Regulatory Guidance for Permeable Barrier Walls Designed to Remediate Chlorinated Solvents.
Interstate Technology and Regulator}? Cooperation (ITRC) Workgroup (Permeable Barrier Wall Subgroup).
90
-------
Johnson, T. L, Scherer, M. M. and Tratnyek, P. G. (1996). "Kinetics of halogenated organic compound degradation
by iron metal." Environ. Sci. Technol. 30(8): 2634-2640.
Keely, J. F. (1989). "Performance evaluations of pump-and-treat remediations". U.S. Environmental Protection
Agency. Superfund Ground Water Issue. EPA/540/4-89/005. Robert S. Kerr Environmental Research Laboratory.
Kent, D. B., Davis, J. A.. Anderson. L. C. D., Rea, B. A. and White, T. D. (1994). "Transport of chromium and
selenium in the suboxic zone of a shallow aquifer: Influence of redox and adsorption reactions." Water
Resources Research 30(4): 1099-1114.
Klcinmann, R. L. P., Hedin, R. S. and Edcnforn, H. M. (1991). "Biological treatment of mine-water an overview."
Second International Conference on the Abatement of Acidic Drainage. Sept. 16-18, Montreal.
Liang, L., Gu, B. and Yin, X. (1996). "Removal of tcchnctium-99 from contaminated groundwater with sorbcnts
and reductive materials." Separations Technology 6: 111-112.
Longmire, P. A., Brookins, D. G., Thomson, B. H. and Eller, P. G. (1991). "Application of sphagnum peat, calcium
carbonate, and hydrated lime for immobilizing uranium tailing leachate." Mat. Res. Soc. Symp. Proc. 212: 623-
631.
Lovely, D. R. and Phillips. E. J. P. (1992a). "Bioremediation of Uranium Contamination with Enzymatic Uranium
Reduction." Environ. Sci. Technol. 26(11): 2228-2234.
Lovely, D. R. and Phillips. E. J. P. (1992b). "Reduction of uranium by Desulfovibrio desuljuricans." Applied
Environmental Microbiology 58(3): 850-856.
Lovely, D. R., Phillips, E. J. P., Gorby, Y. A. and Landa. E. R. (1991). "Microbial reduction of uranium." Nature
350: 413-416.
MacKenzie, P. D., Baghel, S. S., Eykholt, G. R, Homey, D. P., Salvo, J. J. and Sivavec, T. M. (1995). "Pilot scale
demonstration of chlorinated ethene reduction by iron metal: Factors affecting iron lifetime." Emerging
Technologies in Hazardous Waste Management VII. Special Symposium of the American Chemical Society.
American Chemical Society, September 17-20, Atlanta, Georgia, pp. 59-62.
Mackenzie, P. D., Sivavec, T. M. and Horney, D. P. (1997). "Mineral precipitation and porosity losses in iron
treatment zones." ACS National Meeting, American Chemical Society, Division of Environmental Chemistry,
Preprints of Extended Abstracts. 37(1): 154-157.
Marcus, D. L. (1998). "Rcactant Sand-Fracking Pilot Test Results" Summary of the Remediation Technologies
Development Forum Permeable Reactive Barriers Action Team Meeting. April 15-16. Beaverton, Oregon.
Mathcson, L. and Tratnyek, P. (1994). "Reductive dehalogcnation of chlorinated methanes by iron metal." Environ.
Sci. Technol. 28: 2045-2053.
Mathcson, L. J. (1994). Abiotic and biotic reductive dehalogcnation of halogenated methanes. Oregon Graduate
Institute of Science and Technology.
Mathcson, L. J. and Tratnyek, P. G. (1993). "Processes affecting reductive dechlorination of chlorinated solvents by
zero-valent iron." 205th ACS National Meeting. American Chemical Society, March 28-April 2, Denver,
Colorado.
Mclntire, P. E., Edenborn, H. M'. and Hammack, R. W. (1990). "Incorporation of bacterial sulfate reduction into
constructed wetlands for the treatment of acid and metal mine drainage." National Symposium on Mining,
University of Kentucky, May 14-1.8, Lexington, Kentucky.
McMahon, P. B. (1997). "Field Evaluation of a Permeable Reactive Barrier Containing Zero-Valence Iron at the
Denver Federal Center." Remediation Technologies Development Forum, Permeable Barriers Action Team
meeting. Virginia Beach. Virginia.
McRae, C. W., Blowes, D. W. and Ptacek, C. (1997). "Laboratory-scale investigation of remediation of As and Se
using iron oxides." Sixth Symposium and Exhibition on Groundwater and Soil Remediation. March 18-21,
Montreal, Quebec, Canada.
Morin, K. A.. Cherry, J. A., Dave, N. K., Lim. T. P. and Vivyurka, A. J. (1988). "Migration of acidic grouiuhvater
seepage from uranium-tailings impoundments. 1. Field study and conceptual hydrogeochemical model."
Journal of Contaminant Hydrology 2: 271-303.
Morrison, S. J. and Spangler, R. R. (1992). "Extraction of uranium and molybdenum from aqueous solutions: A
survey of industrial materials for use in chemical barriers for uranium mill tailings remediation." Environ. Sci.
Technol. 26(10): 1922-1931.
91
-------
Morrison. S. J. and Spangler. R. R. (1993). "Chemical barriers for controlling groundwater contamination."
Environmental Progress 12(3): 175-181.
Muftikian, R., Fernando. Q. and Korte. N. (1995). "A method forthe rapid dechlorination of low molecular weight
chlorinated hydrocarbons in water." Water Research 29(10): 2434-2439.
National Research Council. (1994). Alternatives for Ground Water Cleanup. Washington, D. C., National Academy
Press.
Odziemkowski, M. S. and Gillham, R. W. (1997). "Surface redox reactions on commercial grade granular iron
(steel) and their influence on the reductive dechlorination of solvent. Micro Raman Spcctroscopic Studies."
213th ACS National Meeting. April 13-17, San Francisco, California.
Olsen. C. R., Lowry, P. D., Lee, S. Y., Larsen, I. L. and Cutshall, N. H. (1986). "Gcochcmical and environmental
processes affecting radionuclide migration from a formerly used seepage trench." Geochimica et Cosmochimica
Acta 50: 593-607.
Orth, W. S. and Gillham, R. W. (1996). "Dechlorination of trichloroethene in aqueous solution using Fe°." Environ.
Sci. Technol. 3 0: 66-71.
Palmer, C. D. and Puls, R. W. (1994). "'Natural attenuation of hexavalent chromium in ground water and soils." U.S.
EPA Ground Water Issue. EPA/540/S-94/505.
Powell, R. M. and Powell, P. D. 1998. "Iron Metal for Subsurface Remediation". The Encyclopedia of Environmental
Analysis and Remediation. Robert A. Myers, ed. John Wiley & Sons, Inc., New York. 8:4729-4761.
Powell, R. M. and R. W. Puls. July 1997a. "Permeable Reactive Subsurface Barriers for the Interception and
Remediation of Chlorinated Hydrocarbon and Chromium (VI) Plumes in Ground Water". U.S. EPA Remedial
Technology Fact Sheet. EPA/600/F-97/008.
Powell, R. M. (1994). Geochemical Effects on Chromate Reduction and Remediation Utilizing the Thermodynamic
Instability of Zero-Valence-State Iron. The University of Oklahoma, Norman, Oklahoma.
Powell, R. M. and Puls, R. W. (1997b). "Proton generation by dissolution of intrinsic or augmented aluminosilicate
minerals for in situ contaminant remediation by zero-valence-state iron." Environ. Sci. Technol. 31(8): 2244-
2251.
Powell, R. M. and R. W. Puls. (1997c). "Hitting the BulPs-Eye in Groundwater Sampling". Pollution Engineering.
June, pp. 50-54.
Powell, R. M. and Puls, R. W. (1993). "Passive sampling of groundwater monitoring wells without purging:
multilevel well chemistry and tracer disappearance." Journal of Contaminant Hydrology 12: 51-77.
Powell, R. M., Puls, R. W., Hightower, S. K. and Clark, D. A. (1995b). "Corrosive and geochemical mechanisms
influencing in situ chromatc reduction by metallic iron." 209th ACS National Meeting. American Chemical
Society, April 2-7, Anaheim, California.
Powell, R. M., Puls, R. W., Hightower, S. K. and Sabatini, D. A. (1995a). "Coupled iron corrosion and chromate
reduction: Mechanisms for subsurface remediation." Environ. Sci. Technol. 29(8): 1913-1922.
Powell. R. M., Puls, R. W. and Paul, C. J. (1994). "Chromate reduction and remediation utilizing the thermodynamic
instability of zero-valence state iron". Proceedings Water Environment Federation. Innovative Solutions for
Contaminated Site Management. Water Environment Federation. March 6-9. Miami, Florida.
Pratt, A. R., Blowes, D. W. and Ptacek, C. J. (1997). "Products of chromate reduction on proposed remediation
material." Environ. Sci. Technol. 31:2492-2498.
Puls, R. and Paul, C. (1995). "Low-Flow Purging and Sampling of Ground Water Monitoring Wells with Dedicated
Systems." GWMR Winter 1995: 116-123.
Puls, R. W., Clark, D. A., Bledsoe, B., Powell, R. M'. and Paul, C. J. (1992). "Metals in ground water: sampling
artifacts and reproducibiiity." Hazardous Waste & Hazardous Materials 9(2): 149-162.
Puls, R. W.. Paul. C. J. and Powell, R. M. (1996). "In Situ Immobilization and Detoxification of Chromate-
Contaminated Ground Water Using Zero-Valent Iron: Field Experiments at the USCG Support Center,
Elizabeth City, North Carolina." The Fourth Great Lakes Geotechnical and Geoenvironmental Conference. In-
Situ Remediation of Contaminated Sites. University of Illinois at Chicago. Chicago, Illinois.
Puls. R. W. and Powell. R. M. (1992a). "Acquisition of representative ground water quality samples for metals."
Ground Water Monitoring Review Fall: 167-176.
92
-------
Puls. R. W. and Powell, R. M. (1992b). Surface-charge repulsive effects on the mobility of inorganic colloids in
subsurface systems. Transport and Remediation of Subsurface Contaminants; Colloidal. Interfacial, and
Surfactant Phenomena, pp. 40-55. American Chemical Society. Washington. B.C.
Puls, R. W. and M. J. Barcelona. April 1996. ''Low-Flow (Minimal Drawdown) Ground-Water Sampling
Procedures". U.S. EPA Ground Water Issue. EPA/540/S-95/504.
Puls. R. W.. Powell, R. M. and Paul, C. J. (1995). "In situ remediation of ground water contaminated with chromatc
and chlorinated solvents using zero-valent iron: A field study." 209th ACS National Meeting. April 2-7,
Anaheim, California.
Rahman. A. and Agrawal, A. (1997). "'Reduction of nitrate and nitrite by iron metal: Implications for ground water
remediation." 213th ACS National Meeting. April 13-17. San Francisco. California.
Reardon, E. J. (1995). "Anaerobic corrosion of granular iron: Measurement and interpretation of hydrogen
evolution rates." Environ. Sci. Technol. 29: 2936-2945.
Reynolds. G. W., Hoff. J. T. and Gillham, R. W. (1990). "Sampling bias caused by materials used to monitor
halocarbons in groundwater." Environ. Sci. Technol. 24: 135-142.
Roberts, A. L., Totten, L. A., Arnold, W. A.. Bums, D. R. and Campbell, T. J. (1996). "Reductive elimination of
chlorinated ethylenes by zero-valent metals." Environ. Sci. Technol. 30(8): 2654-2659.
Robertson, W. D., Blowes, D. W., Ptacek. C. J. and Cherry, J. A. (1995). "Waterloo dinitrification barrier: longer
term performance of pilot scale field trials." Waterloo Centre for Groundwater Research Annual Septic System
Conference "Alternative Systems: Nutrient Removal and Pathenogenic Microbes." Waterloo, Ontario.
Robertson, W. D. and Cherry, J. A. (1995). "In situ denitrification of septic-system nitrate using reactive porous
media barriers: Field trials." Ground Water 33(1): 99-111.
Robin, M. J. L. and Gillham, R. W. (1987). "Field evaluation of well purging procedures." Ground Water
Monitoring Review Fall: 85-93.
Sculley, J. C. (1975). The Fundamentals of Corrosion. Pergamon Press. New York.
Seaberg, J. K. and J. L. Vogan. April 1994. "'Use of Analytic Element Groundwater Flow Modeling to Evaluate
Funnel-and-Gate Technology for In-Situ Groundwater Remediation". International Conference on Analytic
Element Modeling of Groundwater Flow. Indianapolis, Indiana.
Shikaze, S. G., Austins, C. D., Smyth, D. J. A., Cherry, J. A., Barker, J. F. and Sudicky, E, A. (1995). "Hydraulics
of a funnel-and-gate system: Three-dimensional numerical analysis." I.A.H. Conference. Edmonton. AB,
Shoemaker, S. H., Greiner. J. F., and Gillham, R. W. (1996). "Permeable Reactive Barriers." Chapter 11 in
Assessment of Barrier Containment Technologies: A Comprehensive Treatment for Environmental Remediation
Applications. Rumer, R. R. and Mitchell. J. K. (eds.). U.S. DOE, U. S. EPA, and DuPont Company, pp. 301-
353.
Sivavcc, T. M. and Horncy, D. P. (1995). "Reductive dcchlorination of chlorinated ethenes by iron metal." 209th
ACS National Meeting, April 2-7, Anaheim, CA.
Sivavec, T. M., Homey. D. p. and Baghel, S. S. (1995). "Reductive dechlorination of chlorinated ethenes by iron
metal and iron sulfide minerals." Emerging Technologies in Hazardous Waste Management VTI, Special
Symposium of the American Chemical Society. September 17-20. Atlanta, GA.
Sivavec. T. M., Mackenzie, P. D. and Homey, D. P. (1997). "Effect of site groundwater on reactivity of bimetallic
media: Deactivation of nickel-plated granular iron." ACS National Meeting, American Chemical Society,
Division of Environmental Chemistry, Preprints of Extended Abstracts. 37(1): 83-85.
Snoeyink, V. L. and Jenkins. D. (1980). Wrater Chemistry. New York, John Wiley and Sons.
Starr. R. C. and Cherry, J. A. (1994). "In situ remediation of contaminated groundwater: The funnel-and-gate
system." Ground Water 32(3): 46.
Sweeny. K. H. (1980). Treatment of reducible hydrocarbon containing aqueous stream. U. S. Patent Office.
Szerdy. F. S., Gallinatti. J. D., Warner, S. D., Yamane. C. L.. Hankins, D. A. and Vogan. J. L. (1996). In situ
groundwater treatment by granular zero-valent iron: Design, construction, and operation of an in situ treatment
wall. Geomatrix Consultants. San Francisco. CA.
Till, B. A., Weathers. L. J.. and Alvarez, P. J. J. (1998). "Fe(0)-Supported Autotrophic Denitrification." Environ.
Sci. Technol. 32: 634-639.
93
-------
Thomson, B. M., Shelton, S. P. and Smith, E. (1991). "Permeable barriers: a new alternative for treatment of
contaminated groundwater." 45th Purdue Industrial Waste Conference Proceedings. Lewis Publishers. Chelsea,
Ml.
USEPA. (1986). "Ground Water Monitoring". Chapter 11 in Test Methods for Evaluating Solid Waste.
USEPA. (1993). MINTEQA2/PRODEFA2, Metal Speciation Equilibrium Model for Surface and Ground Water.
Version 3.11. Center for Exposure Assessment Modeling, U.S. Environmental Protection Agency, Office of
Research and Development, Athens, GA.
USEPA. (1993). Subsurface Characterization and Monitoring Techniques - A Desk Reference Guide. Vol. 1: Solids
and Ground Water and Vol. 2: The Vadose Zone. Field Screening and Analytical Methods. EPA/625/R-93/
003a and b.
USEPA. (1995). Ground Water Sampling - A Workshop Summary. EPA/600/R-94/205.
USGS. (1998). PHREEQC-A program for aqueous gcochcmical calculations. Version 1.6. U. S. Geological Survey,
Reston, VA.
Vogan, J. L. 1993. "The Use of Emplaccd Denitrifying Layers to Promote Nitrate Removal from Septic Effluent".
University of Waterloo. Waterloo, Ontario, Canada.
Warner, S., Yamane, C., Gallinatti, J. and Szcrdy, F. (1995). "Assessing the feasibility of permeable reactive
barriers for treating VOC-affected groundwater in situ: Experience from the first full-scale commercial
application in California." International Containment Technology Workshop. August 29-31. Baltimore, MD,
Waybrant, K. R., Blowes, D. W. and Ptacek, C. J. (1995). "Selection of reactive mixtures for the prevention of acid
mine drainage using in situ porous reactive walls." Sudbury '95. Mining and the Environment. CANMET.
Ottawa, ON.
Waybrant, K. R., Blowes, D. W. and Ptacek, C. J. (1997a). "Prevention of acid mine drainage using in situ porous
reactive walls: Selection of reactive mixtures." Environ. Sci. Technol. : in submittal.
Waybrant, K. R., Blowes, D. W., Ptacek, C. J. and Benner, S. G. (1997b). "A geochemical and mineralogical
investigation of porous reactive walls for the prevention of acid mine drainage." Fourth International
Conference on Acid Rock Drainage "Application of Technology". May 31-June 6. British Columbia, pp.14.
94
------- |