O'Donnell Et Al. 1995 - Assessment of Barrier Containment Technologies - A Comprehensive Treatment For Environmental Remediation Applications (EPA NTIS PB96-180583)

1
SECTION 1
INTRODUCTION
Although investigations and efforts to remediate contaminated subsurface

zones have been underway in the United States and elsewhere for some two
decades, it has become evident that cost-effective clean-up at many sites can
not yet be achieved. Factors contributing to this situation include the dearth
of cost-effective technologies for in situ remediation and the relative
ineffectiveness of pump-and-treat approaches at sites where dense
nonaqueous phase liquids are present. In view of this, other approaches for
controlling subsurface contaminant zones need to be investigated and
evaluated.
In 1992, under the guidance of Calvin C. Chien, leader of the Containment
and Transport Modeling (C&TM) team, one of five environmental remediation
technology application teams established within the DuPont Corporate
Remediation Group, an internal study was made of the current status and
applicability of barrier containment technologies in remediation. A major
internal summary report resulted from that study. Following that internal
study, a more exhaustive external investigation was undertaken to further
assess the state of the art of containment technologies. This external
investigation was conducted under the auspices of the New York State Center
for Hazardous Waste Management, located at the State University of New
York at Buffalo, with the participation of Calvin C. Chien, Terry D. Vandell,
and Richard Landis, all members of the C&TM team. Andrew Bodocsi, David
E. Daniel, Jeffrey C. Evans, and James K. Mitchell were invited by DuPont to
participate in the investigation. That external investigation resulted in a report
to DuPont in 1993. That report was subsequently published (in its entirety)
in book form by John Wiley & Sons, Inc. [Barrier Containment Technologies
for Environmental Remediation Applications, edited by Rumer and Ryan,
1995].
From the outset of its investigation into containment technologies, the
DuPont C&TM team had planned to organize an international workshop as a
follow-up to the external report prepared under the auspices of the New York
State Center for Hazardous Waste Management [Rumer and Ryan, 1995]. The
primary purpose of the planned international workshop was to continue to
address the gap between what was known and understood about
contamination containment technologies and the level of information needed
to support consistent decision making relative to their application in
remediation. This was to be accomplished by convening a larger group of
experts from academia, industry, and government agencies in the U.S. and
abroad to conduct a careful review and discussion of the applicability and
INTRODUCTION
reliability of these technologies and to identify information, research, and

development needs. Because of mutual interests and similar needs in this
area, the U. S. Department of Energy and the U. S. Environmental Protection
Agency collaborated with DuPont in organizing and sponsoring the
International Containment Technology Workshop, held in Baltimore,
Maryland, August 29 - 31, 1995. This publication is a product of that workshop.
Over 100 workshop participants from eight countries were invited on the
basis of the knowledge, experience, and perspective each could bring to the
discussions concerning the status of containment technologies. Participants
were invited from a wide spectrum of endeavors so as to broaden the
perspective as well as to facilitate the achievement of the workshop objectives.
The workshop was organized into small working sessions, each dealing
with a particular aspect of barrier containment technologies, to provide
opportunity for the invited participants to share information, exchange
opinions, and achieve consensus on what was known, what was not known,
and what was needed concerning the application of containment technologies
in remediation. The working sessions dealt with a range of topics, including:
design and construction of vertical barrier walls (including sheet piles), barrier
floors (indigenous and artificial), caps, geomembrane applications, barrier
materials (soil-based and chemical-based), permeable reactive barriers,
contaminant transport modeling, performance monitoring, and emplacement
verification. Each working session was chaired by a recognized leader(s) in
the field who was responsible for preparing the summary report for his or
her working session.
This publication contains the edited summary reports from each working
session. Although particular individuals have been credited with the
preparation of each section, other significant contributors are also identified.
It is virtually impossible to acknowledge the specific contributions of all
participants attending the workshop, and this is regrettable. However, a full
listing of all participants can be found in an appendix to this publication.
Each section of this publication has been reviewed for technical content
by James K. Mitchell, Virginia Polytechnic Institute and State University. The
initial compilation of the draft report and the final editing were carried out
by Ralph R. Rumer, State University of New York at Buffalo. The workshop
session leaders participated throughout the editing process.
The responsibility for planning, organizing, and conducting the workshop
rested in the hands of an executive committee, co-chaired by: Grover H. (Skip)
Chamberlain, Office of Technology Development, U. S. Department of Energy;
Randall Breeden, Office of Emergency and Remedial Response, U.S.
Environmental Protection Agency; and Calvin C. Chien, C&TM team leader,
Corporate Remediation Group, DuPont Company. To continue to provide a
forum for future discussions on advancements and experience gained in the
application of containment technologies, the three sponsors decided to
organize future biannual international conferences, with the first one
scheduled for February 1997.
1.0 INTRODUCTION
Acknowledgments
The workshop sponsors and all of the contributors to this book express their
gratitude to Phyllis Adams, DuPont Co., for her very able assistance
throughout the planning and execution of the workshop. The artwork was
prepared by Barbara E. Evans, Art and Photographic Services, State University
of New York at Buffalo. The design and preparation of this publication was
carried out by Sue A. ODonnell, Peripheral Vision, Buffalo, NY. The index
was written by Kathryn W. Torgeson, Ithaca, NY. Loreen Kollar, Florida State
University [under a cooperative agreement with the U. S. Department of
Energy], handled the travel arrangements for the workshop attendees, assisted
at the workshop, and arranged for the printing of this publication.
SECTION 2
SOIL- AND CEMENT-BASED VERTICAL BARRIERS
WITH FOCUS ON MATERIALS
prepared by
Jeffrey C. Evans, Bucknell University

with contributions by
Marita Allan, Brookhaven National Laboratory

Steven R. Day, Geo Con Inc.
George M. Filz, Virginia Polytechnic Institute and State University
Hilary I. Inyang, University of Massachusetts at Lowell
Stephan Jefferis, Golder Associates Ltd, United Kingdom
Lawrence E. Kukacka, Brookhaven National Laboratory
Lisa Martinenghi, Institut fur Geotechnik, Switzerland
Richard Mitchell, RM Consultants
Kelvin Potter, ICI Engineering Technology, United Kingdom
2.1 INTRODUCTION
The soil and cement materials used to construct vertical barriers that include;
1) externally mixed barriers, such as soil-bentonite slurry trench cutoff walls,
and 2) in situ mixed walls, using deep soil mixing and grouting techniques,
are discussed in this section. Slurry trench cutoff walls composed of soilbentonite, cement-bentonite, plastic concrete, and recent innovations (both
proposed and implemented) such as attapulgite based barriers, bioenhanced
barriers, and other mineral enhanced barriers are considered. The differences
between European and U.S. practices are described. Barrier construction
techniques are not discussed, except as necessary to understand differences
in barrier material performance. Regardless of construction method, the focus
is on the barrier materials used and their properties, as determined by both
long and short term performance measures.
Soil-bentonite slurry trench cutoff walls are the most widely used type
in the U.S. Their application in remediation is reviewed in Rumer and Ryan
(1995). For this reason, they are not given detailed attention here. The
construction and performance of soil-bentonite slurry trench cutoff walls are
well understood and they are usually considered reliable when used as
subsurface vertical barriers for containment.
SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)
2.2 BACKGROUND
At virtually all remediation sites, vertical barriers are needed as a component
of a constructed containment system. Typically, vertical barriers are used in
conjunction with some form of pump and treat remediation. The vast majority
of vertical barriers are currently constructed as either soil- or cement-based
walls. In the U.S., the vertical barrier of choice is usually a soil-bentonite
slurry trench cutoff wall. In Europe, the choice is usually a cement-bentonite
slurry trench cutoff wall, often incorporating slag or fly ash in the slurry mix.
The experience with soil- and cement-based vertical barriers is extensive and
long-term. There is also growing experience with geosynthetics in/as vertical
barriers.
Monolithic grout and soil cement barriers may be considered for
containment applications below the water table, in the vadose zone, and at
arid sites where bentonite-based materials are at risk of desiccation and
cracking. Many sites controlled by the U.S. Department of Energy (DOE) are
located in such environments. Contaminants of concern at these sites include
chemical, radioactive, and mixed wastes. Cementitious barriers are also
applicable under saturated conditions.
Subsurface barriers are usually required to exhibit low hydraulic
conductivity (<10-7 cm/s), be durable, and have adequate strength. Installed
barrier walls may be exposed to leachates, contaminants, and aggressive
chemical species that naturally occur in ground water (e.g., high soluble
sulfates). Under these conditions, the properties of grout and soil cement
barriers can be enhanced through selection of appropriate admixtures .
2.3 STATE OF PRACTICE

The methods used to construct soil- and cement-based vertical barrier walls
are well known and well documented. Excavation with clamshells and
backhoes, using bentonite-water slurry to maintain trench stability, is well
understood and has been widely employed. These excavation techniques
allow for the placement of soil-bentonite backfill within the slurry filled trench
to form the soil-bentonite slurry trench cutoff wall. Alternatively, the backhoe
and clamshell excavation techniques can use a slurry of bentonite-watercement to maintain trench stability. This slurry can be left to harden in place
to form a cement-bentonite slurry trench cutoff wall. Soil- and cement based
barrier materials have also been used to construct vertical barriers using in
situ excavation/placement techniques, such as deep soil mixing and grouting
(permeation and jet).
The slurry mix proportions critically affect the short and long term
performance of the completed cutoff wall. For soil-bentonite slurry trench
cutoff walls, the optimum mix is composed of a well-graded material
containing sand and gravel along with 20% to 50% fines of low plasticity/
activity. The mixture typically has a bentonite content of approximately 1%.

This bentonite is added to the backfill in the form of bentonite-water slurry
used to achieve workability (slump). Although, supplementing the mix with
large quantities of bentonite can reduce the initial hydraulic conductivity,
mixes with larger proportions of bentonite can be more susceptible to
degradation because of incompatibility between the mix and the surrounding
contaminated ground water. Non-aqueous phase liquid contaminants pose
a particular problem in this regard.
In the U.S., cement-bentonite slurries are typically composed of water,
5% bentonite, and 15% to 25% cement. The resulting hydraulic conductivity
is on the order of 10-5 to 10-6 cm/s. In Europe, slag or fly ash is frequently
added to the mix. These latter mixes have a hydraulic conductivity typically
less than 10-7 cm/s. Cement-water grouts or soil-cement mixtures can achieve
a hydraulic conductivity as low as 10 -10 cm/s by the inclusion of
superplasticizers. Water/cement ratio is one of the major factors that
determine the final properties of cementitious barrier materials.
2.3.1
Material Choices as Influenced by Construction Method
The cementitious grouting materials chosen for use in a vertical barrier must
be compatible with currently available placement techniques. Since numerous
techniques (e.g., permeation grouting, jet grouting, soil mixing, etc.) are
available, the choice depends on the properties of the site soils, the consistency
required for placement by injection or mixing, the required properties of the
barrier, and the installed cost of the final product. Also, the grouting material
chosen must be non-toxic and acceptable to the Environmental Protection
Agency (EPA) and other regulatory agencies. Portland cement-based grouts
are generally acceptable. However, some commercially available chemical
grouts such as polyacrylamide, polyacrylic acid, phenol-formaldehyde, and
urea-formaldehyde may not meet with EPA or other regulatory agency
acceptance (Kukacka, 1995).
The mechanical and physical properties of the uncured and cured grouting
materials are important. In their uncured state, the materials must exhibit
stability and have viscosity compatible with the placement technique to be
employed. Bleeding of water and settling of particles should not occur. For a
solid grout barrier, the flow properties of the grout should enable it to
completely fill the required cavity.
The set time for grouts must be long enough to permit placement and
equipment clean up. Long set times are not as critical, unless continuation of
work depends on solidification of previously placed material. A period of 12
to 24 hr is generally acceptable. However, placement costs may increase as
the cure time is extended.
After curing, the solidified in situ barrier materials must exhibit low
hydraulic conductivity, low shrinkage, adequate strength, and long-term
durability. The strength requirements are typically minimal, since lateral soil
pressures are frequently equal on both sides of a vertical barrier wall.

Therefore, the design thickness of the barriers is normally controlled by
hydraulic conductivity requirements, rather than strength. However, strength
and compressibility considerations can be important when other structures
are present or in seismically active regions.
Low hydraulic conductivities are needed in vertical barrier walls to control
the migration of contaminants through the barriers by advection. EPA 40
Code of Federal Regulations, Part 264, Subpart N (landfills) stipulates that
containment barriers for waste impoundments must have hydraulic
conductivities less than 10-7 cm/s.
Durability of the barrier is essential throughout its service life. Although
service life times have not been well defined, a few hundred to a thousand
years may be needed in the case of radioactive waste containment. Therefore,
the placed barrier material must be stable, and retain the desired mechanical
and physical properties for long periods of time.
The method used to mix the formulation before placement may be an
important consideration. For example, attapulgite clay cannot be mixed in
ponds because it settles too quickly. Both attapulgite and guar gum, used in
biopolymer trenches, require high energy mixing for adequate dispersion.
a. Permeation Grouting Materials. Permeation grouting refers to the method
of construction by which a barrier is constructed by pumping a grout into the
ground to fill the pores and voids of the subsurface soil and rock matrix (Rumer
and Ryan, 1995). The grout is often a cementitious material and, once hardened
in the voids, seepage and contaminant transport are greatly reduced. For
environmental remediation applications, permeation grouting is used
primarily in fractured rock stratigraphies, while slurry walls tend to be used
for sites underlain by excavatable soils (Karol, 1983; Hausmann, 1990;
Xanthakos, et al., 1994).
Use of slag in treatment grouts can contribute benefits such as resistance
to sulfate attack. The heat generated during grout hardening may cause
cracking, and should be considered. If heat dissipation is inhibited by the
surrounding soil, excessive thermal stresses may develop. The addition of
slag and optimization of the soil/grout ratio can reduce the risk of thermal
cracking.
b. Jet Grouting Materials. Jet grouting employs a very high pressure rotating
jet to construct a subsurface barrier. The grout, often a cementitious material,
mixes with the native soil materials to form a soil-cement column of low
hydraulic conductivity. These columns are constructed sequentially in an
overlapped fashion to form the subsurface barrier wall (Xanthakos, 1994;
Rumer and Ryan, 1995). For environmental remediation applications, jet
grouting can be considered for construction of an artificial bottom barrier, or
for a vertical barrier wall when trench excavation is not possible.
c. Deep Soil Mixing Materials. Deep soil mixing is accomplished using a

special auger mixing shaft that is rotated into the ground while simultaneously
injecting either a bentonite and water slurry or a cement, bentonite and water
slurry (Rumer and Ryan, 1995). The advantages of deep soil mixing include
reduction in the volume of soil to be disposed, greater stability during
construction, structural capability of the completed column, and the ability
to carry out soil vapor extraction during the construction process (Day and
Ryan, 1995). Sampling of cementitious barrier materials, such as those
resulting from the deep soil mixing process, can produce unrealistically high
values of hydraulic conductivity when tested. This has been attributed to
sample disturbance and is discussed further in Section 3.
Polypropylene fibers can be added to mixes (0.2 % volume fraction) used
for producing soil cements by soil mixing, bringing about improved flexural
strength and ductility and reduced widths of potential shrinkage cracks.
Resistance to wet-dry cycling can also be improved by the addition of
fibrillated fibers.
d. Summary of Material Choices for Grouted Barriers. Cementitious grouts
selected from mixtures of Portland cement, sand, bentonite, and various
additives are excellent choices for barrier wall formulations when using in
situ placement techniques. These mixtures are low cost compared to chemical
grouting materials, compatible with commercially available placement
technologies, nontoxic, and, upon curing, yield strong, durable and low
hydraulic conductivity products. The cements can be used as bulk materials
or can be mixed with on-site soils at ratios up to 5 parts soil to 1 part cement.
Hydraulic conductivities in the range of 10-10 cm/s may be attainable by
adding superplasticizers.
2.3.2 Properties of Cementitious Grouts
In its simplest form, cementitious grout consists of ordinary Portland (ASTM
Type I) cement and water. The properties of cementitious grout are controlled
by the water/cement ratio, cement type, addition of bentonite or attapulgite
clay, partial replacement of cement with mineral admixtures (supplementary
cementing materials), addition of retarders or accelerators, and use of water
reducing agents. Grouts also may be modified with latex. Sand may be used
as a filler material in some situations. Each of these additives has a distinct
influence on the properties of the unhardened (uncured) and hardened (cured)
grout.
a. Properties of Cementitious Grout in the Unhardened State. The flow
behavior of the unhardened grout is critical, since it determines the ability to
mix, pump and place the grout using conventional equipment. The important
flow properties include; viscosity, gel strength, bleeding, and set time.
Specification of flow property values will depend on the placement technique
10
to be used and site conditions.

Viscosity. In addition to the strong influence of the water/cement ratio, the
viscosity of grout can be altered using chemical admixtures. Water reducing
and superplasticizing agents can be used to minimize the water demand while
still maintaining low viscosity (see ASTM C 494 and ASTM C 1017).
Reductions in the water/cement ratio are beneficial for properties such as
strength, hydraulic conductivity, and durability. Superplasticizers are
generally more effective than regular water reducers.
The effectiveness of superplasticizers decreases with mixing time, and
the rate of superplasticizer addition can exert a significant effect on grout
rheology. An example of this effect is shown in Figure 2-1 (Allan, 1995) which
presents a plot of shear stress versus shear rate for a Type I cement grout with
a water/cement ratio of 0.4. When superplasticizer was added last, the yield
stress decreased at a given shear rate as compared to adding the
superplasticizer to the water before mixing. Cement type will also influence
grout viscosity. For example, at a given water/cement ratio, a microfine (ultra
fine) cement produces a higher viscosity than would be produced using coarser
cement.
100
Shear Stress (Pa)
Type I Cement
80
Superplasticizer
added to water
60
40
Superplasticizer
added last
20
0
200
400
600
800
1000
1200
Shear Rate (s-1)
Figure 2-1 Effect of sequence of superplasticizer addition on shear stressshear rate relationship of Type I cement grout (Allan, 1995)
The most common mineral admixtures used with cement grout include;
fly ash (ASTM C 618), silica fume (micro silica), and mechanically granulated
blast furnace slag (ASTM C 989). These materials, when used as partial
replacements for cement, can be beneficial by reducing heat of hydration and
improving durability. Silica fume has a strong effect on rheological properties,
while fly ash and slag have less effect. The effect of silica fume replacement
(for cement) on the shear stress/shear rate relationship is shown in Figure 22 (Allan, 1995). The apparent viscosity at a given shear rate is significantly
11
increased, even at relatively low replacement levels. Thus, the practical cement
replacement level with silica fume is limited to 5-10%. Higher replacement
levels can be used with fly ash or slag, with less dramatic effects on rheology.
This is illustrated in Figure 2-3 (Allan, 1995) for 60% replacement of Type I
cement with slag. The type of fly ash used (i.e., pozzolanic or nonpozzolanic)
will also influence rheological properties.
160
10% Silica Fume
Shear Stress (Pa)
140
120
100
5% Silica Fume
80
60
0% Silica Fume
40
20
0
200
400
600
800
1000
1200
Shear Rate (s-1)
Figure 2-2 Effect of partial replacement of Type I cement with silica fume on
shear stress-shear rate relationship (Allan, 1995)
Bentonite is a common cementitious grout additive, used primarily to
improve grout stability and reduce bleeding. Mix viscosity increases with
increasing bentonite content. Superplasticizers are not effective with high
bentonite content grouts. The addition of latex to grout also increases viscosity.
Other factors that affect the viscosity of cementitious grouts include; shear
and thermal history, temperature, and mixing equipment used.
Gel Strength. When the application of stress ceases and the grout is quiescent,
the liquid grout begins to transform to a colloidal gel. Gel strength is a measure
of the stress required to re-initiate flow after the grout has been at rest.
The addition of silica fume and bentonite increases gel strength
significantly. In contrast, adding slag and fly ash tends to reduce gel strength.
Adding superplasticizers also reduces gel strength, the effect again being
related to the sequence of addition. By adding superplasticizer after cement,
rather than with the mixing water, the gel strength can decrease significantly.
Incorporation of latex in grout decreases gel strength.
Bleed. Excessive bleeding (the separation of water from solids) and
segregation of grout are undesirable. The most commonly used additive to
control bleeding and segregation is bentonite. Silica fume is also highly
beneficial in reduction of bleeding. Superplasticizers can act to increase bleed,
12
particularly if overdosed.
100
Slag Modified
Shear Stress (Pa)
60%
80
0%
60
40
20
0
200
400
600
800
1000
1200
Shear Rate (s-1)
Figure 2-3 Effect of 60% replacement of Type I cement with blast furnace
slag on shear stress-shear rate relationship (Allan, 1995)
Set Time. Set time is strongly dependent on temperature, with higher
temperatures accelerating the set. Accelerators or retarders can be added to
counteract adverse temperature conditions. Common retarders include borax,
sugars, and calcium lignosulphonate; while calcium chloride and sodium
silicate are common accelerators. Combination types of water reducersretarders or water reducers-accelerators are also available (refer to ASTM C
494).
Admixtures, used to control other properties, may indirectly affect set
time. For example, superplasticizers act as retardants. Fly ash and slag also
tend to increase set time. Since it is usually necessary to combine a
superplasticizer with silica fume, the set time for such grout mixtures may be
retarded. High proportions of bentonite increase set time.
When grout is mixed with soil, the set time may be altered, depending on
the soil chemistry. Many contaminants, particularly organics, increase the
set time of cement-based materials. Other contaminants may accelerate set.
Therefore, if cementitious grout is to be mixed with a contaminated soil, it is
essential that the potential for interference with cement hydration and setting
be investigated.
b. Properties of Cementitious Grout in the Hardened State. The physical
and mechanical properties of hardened grout and grout treated soil (soil and
cement) depends on other factors, in addition to the presence of admixtures.
Some of these include; mix proportions, curing conditions, age of the hardened
grout, temperature, soil type, and presence of contaminants. Mix proportions
(i.e. grout to soil ratio) of soil and cement mixtures produced by jet grouting
or deep soil mixing can be controlled by the system parameters.
13
Hydraulic Conductivity. Water/cement ratio is the most significant parameter

controlling the properties of hardened cementitious grouts. By reducing
water/cement ratio through use of superplasticizers, the hydraulic
conductivity can be reduced. The hydraulic conductivity increases with
decreasing grout content or with increasing water/cement ratio. This effect
on hydraulic conductivity for soil cements produced from superplasticized
grout with simulated in situ curing is illustrated in Figure 2-4 (Allan, 1995).
The results are for laboratory prepared specimens; thus, some deviation from
field-placed materials is expected. However, the trends for the different
variables are considered relevant.
Hydraulic Conductivity (cm/s)
10-6
10-7
10-8
10-9
10-10
S/C
W/C
0
0.40
1
0.48
2
0.81
3
1.15
4
1.38
5
1.60
Figure 2-4 Effect of soil/cement (s/c) and water/cement (w/c) ratios on

hydraulic conductivity of in situ cured soil cements containing
superplasticizers (Allan, 1995)
Partial replacement of cement with silica fume, fly ash, or slag generally
reduces the hydraulic conductivity of grout, assuming the water/cementitious
material ratio remains constant. The use of sulfate resistant cement may also
be beneficial in reducing hydraulic conductivity. Figure 2-5 (Allan, 1995)
depicts the effect of slag content, water/cementitious material ratio, and soil/
cementitious material ratio on the hydraulic conductivity for superplasticized
samples made from soil and cement under simulated in situ curing conditions
in a vadose zone (i.e. dry curing).
Latex reduces the hydraulic conductivity of grouts due to pore blockage
by the latex film. Findings for samples made from soil and cement produced
using latex-modified grouts have been inconsistent. In general, the hydraulic
conductivity of samples made from soil and cement is reduced by latex, but
not always. Latex cannot be expected to compensate for very high water/
cement ratios.
Strength.
Normally, subsurface barriers are not required to possess
14
particularly high strength, and unconfined compressive strengths of 100 to

200 kPa (15 to 30 psi) are usually adequate. If the constructed barrier wall
serves a structural purpose, higher strength may be required. The hydraulic
conductivity and long-term durability are usually more important than
strength considerations. The factors that control hydraulic conductivity and
durability also affect strength. Thus, there is an indirect correlation between
the strength and the hydraulic conductivity and durability of cementitious
barriers.
Hydraulic Conductivity (cm/s)
10-5
10-6
10-7
0% Slag
40% Slag
60% Slag
80% Slag
10-8
10-9
10-10
10-11
10-12
S/C
W/C
1
0.48
2
0.72
5
1.40
Figure 2-5 Effect of soil/cement (s/c) and water/cement (w/c) ratios and
slag replacement level on hydraulic conductivity of in situ cured
soil cements containing superplasticizers (Allan, 1995)
The strength of cementitious grout barriers made from soil and cement
generally exceeds that required in the field. Strength is largely controlled by
water/cement ratio and the proportion and type of soil. Reducing the water/
cement ratio using water reducers or superplasticizers will result in greater
strength. The effects of soil/cement and water/cement ratio on compressive
strength of superplasticized materials made from soil and cement are
illustrated in Figure 2-6 (Allan, 1995).
The effect of supplementary cementing materials on compressive strength
depends on the type of supplementary material, the proportion added, and
the curing conditions. Certain materials can increase the strength of
cementitious grouts under favorable curing conditions, although the initial
rate at which the strength is gained may decrease, particularly with the
addition of fly ash materials. When grout is mixed with soil, the effects of
supplementary cementing materials on compressive strength tend to diminish
as the proportion of soil and/or water increase. An example of the effect of
replacing cement with slag on the compressive strength of soil-cement is
depicted in Figure 2-7 (Allan, 1995). It is seen that slag has minimal influence
on strength at high soil/cementitious and high water/cementitious material
15
ratios, even at replacement levels of 80%. Therefore, high proportions of slag

can be used without detrimental effects on the strength properties of the
cementitious grout. Similarly, high levels of fly ash can be incorporated in
soil-cements. In contrast, the proportion of silica fume that can be added is
controlled by the viscosity requirements needed for the placement technique
to be used. High proportions of bentonite and other clay-type additives reduce
the compressive strength of grouts and soil cements. Latex also causes a
decrease in strength, particularly under wet conditions and at high water/
cement ratio.
Compressive Strength (MPa)
45
40
Wet Cured
In-situ Cured
35
30
25
20
15
10
5
0
S/C
W/C
0
0.40
1
0.48
2
0.81
3
1.15
4
1.38
5
1.60
Figure 2-6 Effect of soil/cement (s/c) and water/cement (w/c) ratios on

compressive strength of soil cements containing superplasticizers
(Allan, 1995)
Although compressive strength is a commonly measured property,
flexural strength may also be important. The flexural strengths of soil cements
produced from superplasticized grouts can range from 2 to 3 MPa (depending
on mix proportions).
Sulfate Resistance. Many ground waters and soils contain sufficiently high
levels of soluble sulfates so as to be detrimental to the integrity of Type I
cement-based barrier materials. This potential problem can be overcome either
through partial replacement of Type I cement with silica fume, fly ash, blast
furnace slag, or by substitution with a sulfate resistant cement (Type II or V).
Such grout-based materials exhibit improved resistance to sulfate attack.
Sulfate resistance is also improved by minimization of water/cement ratio.
Leach Resistance. Soft waters and many leachates, particularly those with
low pH, are aggressive towards cementitious materials. Silica fume is highly
effective in improving leach resistance.
16
Compressive Strength (MPa)
40
0% Slag
20% Slag
40% Slag
60% Slag
80% Slag
35
30
25
20
15
10
5
0
S/C
W/C
1
0.48
2
0.72
5
1.40
Figure 2-7 Effect of soil/cement (s/c) and water/cement(w/c) ratios and

slag replacement level on compressive strength of soil cements
containing superplasticizers (Allan, 1995)
Shrinkage. If a barrier shrinks, the resulting void may result in greater flow
through the barrier than if the barrier material does not shrink. Shrinkage of
cementitious materials can be minimized by using Type K cement or adding
calcium sulphoaluminate, calcium sulfate, or gas forming admixtures. These
admixtures should be used with caution since continued expansion during
the hardened state can result in cracking. Shrinkage is also reduced by
lowering the water/cement ratio. Although superplasticizers increase
shrinkage, they are less detrimental than a high water/cement ratio. The
effect of supplementary cementing materials on shrinkage will depend on
the proportions and type used. Sand and other filler materials can reduce
shrinkage, but require careful selection of the appropriate size fractions and
proportion of filler.
Cracking may occur in hardened or nearly hardened stage when shrinkage
is restrained. Crack resistance is enhanced by low water/cement ratios. The
widths of cracks developed under restrained shrinkage conditions have been
monitored in laboratory tests at BNL. Grouts develop smaller crack widths
than soil cements. The addition of fibers to the parent grout reduces shrinkage
crack widths significantly, as depicted in Figure 2-8 (Allan, 1995). The fibers
used in this example were fibrillated polypropylene.
Fibers also improve ductility, flexural strength, and toughness. Fiber
reinforced materials subjected to flexural load or to shrinkage form a network
of fine cracks, rather than a single large crack. Even after ultimate load has
been reached, fiber reinforced grouts and soil cements remain contiguous
due to crack bridging by fibers. Fiber-enhanced grouts are incompatible with
jet grouting or permeation grouting, but may have applications in bentonitebased materials placed by backfilling techniques.
17
3.50
Crack Width (mm)
3.00
2.50
Plain
2.00
1.50
1.00
0.1% Fibres
0.2% Fibres
0.50
0.00
0
10
15
20
25
Time After Failure (days)
Figure 2-8 Crack widths of plain and fiber reinforced grouts under restrained
shrinkage (Allan, 1995)
c. Admixtures and Cementitious Grout Properties. A variety of admixtures
can be used to improve the properties of cementitious grouts used to construct
subsurface barriers, particularly if conventional materials are unsuited to the
site conditions. The most advantageous of these admixtures are water reducers
or superplasticizers, which enable reduction of water demand. Reduction of
the water/cement ratio results in enhanced strength, lower permeability, and
greater durability.
Admixtures cannot compensate for other inadequacies; e.g., poor mix
design, unsuitable materials, improper placement techniques for the site
conditions, and substandard work practices. Laboratory and field trials are
necessary for evaluating the suitability of any admixture for a given
application.
In many cases, the addition of admixtures results in increased material
costs. Added capital expenses may occur if additional batching equipment is
needed. Field trials are required to become familiar with the attributes of
specific admixtures. Low bentonite cementitious grouts containing admixtures
are probably most applicable to situations in which the performance and
durability of soil-bentonite, cement-bentonite, or soil-cement-bentonite
materials will not be adequate.
2.3.3 Properties of Soil-Bentonite Materials
a. Hydraulic Conductivity. Soil-bentonite slurry trench cutoff walls can be
designed and constructed to have a hydraulic conductivity less that 10-7 cm/s.
Data compiled by Evans (1993) show that the stress state in the soil-bentonite
backfill can have a strong influence on in-service hydraulic conductivity. Also,
contaminated permeants can increase the hydraulic conductivity of barrier
18
soils, but the effect is less significant when the soil is under high confining
pressure (Acar et al., 1985; Daniel, 1987; and Mitchell and Madsen, 1987).
These investigations indicate that the stress state in the soil-bentonite cutoff
wall can affect its performance as a containment barrier.
b. Compressibility. DAppolonia (1980) showed that the compressibility of
soil-bentonite backfill mixtures increases as the fines content of the base soil
increases, and that the compressibility is greater when the fines in the base
soil are plastic. DAppolonia states that Compressibility depends chiefly on
the percentage of granular bulky-shaped particles in the gradation.
Comparatively low compressibility results when there is sufficient granular
material in the mix to allow grain-to-grain contact between the granular
particles.
Consolidation tests reported by Khoury et al. (1992) exhibited greater
compressibility than those of DAppolonia. It was also found that large
amounts of gravel, i.e., at least 30%, would need to be added to the mix to
produce significant reductions in compressibility. Consolidation tests
performed on laboratory prepared soil-bentonite mixes by Evans et al. (1995)
also found that compressibility increased as the fines content increased.
c. Strength. DAppolonia (1980) presented the results of consolidated-drained
and consolidated undrained triaxial compression tests performed on three soilbentonite mixtures. As shown in Figure 2-9, all three mixes were initially
contractive. Mix A, which is the coarsest of the three mixtures tested, showed
a tendency for dilation at axial strains above about 3%. The consolidation
pressures for the tests were not reported. Effective stress friction angles ranged
from 31 to 33 degrees. Undrained shear strengths for mix A ranged from 32%
to 70% of the consolidation pressure; for non-dilative mixes B and C, the
undrained shear strengths were, respectively, 32% and 40% of the consolidation
pressure. Using the normalized plots shown in Figure 2-9, one can obtain
approximate values of the initial tangent Youngs modulus of the three mixes,
expressed in terms of the consolidation pressures listed in Table 2-1.
TABLE 2-1 Approximate Initial Tangent Youngs Modulus Values for SoilBentonite Mixes (from data presented by DAppolonia, 1980)
Mix
Approximate Initial Tangent Youngs Modulus

CU Test
CD Tests
80con
60con
40con
50con
30con
30con
19
Consolidated-undrained triaxial compression tests performed on samples of

soil-bentonite backfill used in the cutoff wall at a contaminated site in Silicon
Valley, California, resulted in an effective stress friction angle of 32 degrees
and an effective stress cohesion intercept of zero (Filz, 1995, personal
communication).
1.5
100
#20
#200
q/c =(1-2)/2c
% Finer by Weight
Sieve Size
#4
80
60
B C
40
20
0
Gravel
100
Sand
10
Silt
0.1
0.01
1.0
0.5
B
C
Consolidated Undrained
(CU) Tests
0
CD Tests
Clay
0.001
0.5
1.0
1.5
2.0
2.5
1.5
10 12 14 16
q/c =(1-2)/2c
q/c =(1-2)/2c
33
33
31
P/c = (1-2)/2c
0.6
0.2
0.70
0.40
0.32
0.4
Su
c
A
B
C
CU Tests
Particle Diameter, mm
0.8
Sample
1.0
A
B
0.5
Consolidated Undrained
(CU) Tests
0
Axial Strain, %
10 12 14 16 18 20 22
Axial Strain, %
Figure 2-9 Stress-strain and strength behavior for various SB backfill

gradations (DAppolonia, 1995)
2.3.4 Slurry Wall Barriers in the United Kingdom

a. UK Historical Perspective. (Jefferis, 1995). The first cut off walls in the
UK were soil-bentonite systems. However, few records exist concerning these
walls or the procedures used to form them. One of the few reported walls
was installed in 1963 to keep a gravel workings dry. Since 1963 soil-bentonite
walls have been used occasionally in the UK - mostly at gravel workings.
Overall, it can be said that applications of soil-bentonite walls have passed
largely unreported and there is minimal UK literature on such walls. They
are certainly not the cut off wall of choice today.
By the late 1960s to early 1970s, cement-bentonite-aggregate (plastic
concrete) systems were in use. Little (1975) reports the use of plastic concrete
to repair the cores of a number of earth dams. These walls achieved a depth
of over 46 m and were two-phase walls excavated in 6 m panels. It is interesting
to note that one of the walls was designed as a high pulverized fly ash (pfa)
concrete to resist a reservoir water of pH of 3.8. No data are available regarding
the performance of this wall over the last two decades.
20
The first reported cement-bentonite cut off wall in the UK was installed
in 1973 to control seepage from an agrochemicals manufacturing plant into a
stream. The mix used was a blend of ordinary Portland cement, sodium
bentonite (of UK origin and, therefore, a natural calcium bentonite converted
to the sodium form with sodium carbonate) and water. The wall successfully
reduced the leakage of chemicals from the contaminated soils around the
plant into the stream. The 1973 wall was based on a mix design from France
where there had been a substantial program of work by Caron and others.
However, it was found that the mix design had to be revised considerably to
take account of UK materials and, in particular, the cement.
Following the 1973 wall, a research program was initiated at Kings
College, University of London under the direction of Dr. Stephan Jefferis.
This work demonstrated that the properties of cement-bentonite materials
could be radically altered if a substantial proportion of the cement in the mix
was replaced with ground granulated blast furnace slag, a by-product of steel
production. In the UK, such slag is slightly cheaper than Portland cement
and is widely used as a cement replacement material in concrete. The effects
of adding slag are comparable to those of Portland cement, but the slag has
less influence on cement-bentonite mixes. Results of this research were
incorporated in the design of a mix for the cut off under the Kielder Dam
which created one of the largest man-made reservoirs in Europe. The wall,
up to 25 m deep, was constructed in panels using a clam shell grab. The wall
was required to have a permeability of less than 10-6 cm/s and a strain at
failure of greater than 5%. A cement-slag-bentonite wall was used for the
dam, the first use of such a mix in the UK. It may also have been the first use
of a mix in which the slag-Portland cement ratio had been specifically adjusted
to optimize properties of the cut off material. Slag cements had been used in
cut offs in Europe prior to this, but at pre-blends prepared for the general
construction industry at a slag-cement ratio of the suppliers choice which
were often subject to quite wide tolerances. Following the Kielder cut off,
there was a slow expansion of barrier work in the UK. The purpose of most of
these walls was to reduce seepage around hydraulic structures, rather than
prevent contaminant migration. Cement-slag-bentonite materials were used
for most of these walls.
In the late 1980s to early 1990s, the emerging issues of contaminated land
and gas/leachate control at old landfill sites led to a rapid increase in the use
of cut off walls. During this period, a further innovation was the incorporation
of jointed geomembranes into the walls. The first actual slurry trench/
geomembrane wall was installed in 1979 in the Jordanian sector of the Dead
Sea, as part of a system of ponds for salt extraction by solar evaporation. The
geomembrane panels were lapped, not joined. Cement-bentonitegeomembrane walls are now widely used in the UK, especially for the control
of gas and leachate migration from old landfill sites.
At the present time, the material used for slurry trench walls is most
commonly a cement-slag-bentonite mix, though it is recognized that pfa may
21
give better chemical resistance. A few cement-pfa-bentonite walls have been

installed. Slurry trenches in the UK are usually 0.6 m (2 ft) wide. Excavation
normally begins using a backhoe (draglines are not used). Below a depth of
20 m, a cable hung or Kelly-mounted grab is used and walls are excavated in
panels. In hard grounds, a cutter-miller machine (hydrofraise) may be used.
b. Materials for Barriers Constructed using the Slurry Trench Method. Soilbentonite, cement-bentonite, cement-pfa-bentonite, plastic concrete, soilattapulgite, cement-slag-bentonite, and cement-attapulgite have been used
to form cut off elements in the UK. The basic materials are bentonite and
attapulgite clays, Portland cement, slag, and pfa (in the concrete industry
these are often described as cement replacement materials).
Bentonite. In the UK and Europe, almost all the indigenous bentonites have
calcium as the predominant exchangeable cation. Calcium bentonite does not
readily disperse in water to form the desired viscous, nonsettling suspension
at low concentration (perhaps 2% to 6% by weight of water). While it has
applications in slurries, it is not a material that can be used to control settlement
of cement grains, aggregate, etc. in cut off materials. For this reason, sodium
bentonite is used. The main role of calcium bentonites, when included in cut
off mixes, is to provide extra clay solids without the resulting mix becoming
unacceptably viscous. Depending on the source, sodium bentonites may be
used up to a maximum concentration of perhaps 5% to 7% by weight of water
without a cement-bentonite mix becoming unworkable. In contrast, calcium
bentonite may be used at concentrations over 15%.
Calcium bentonites are rarely used in cut offs in the UK. When used, the
indigenous calcium bentonites are pretreated (by the suppliers, rather than
on site) with sodium carbonate to precipitate the calcium ions as calcium
carbonate and convert the clay to the sodium form. The amount of sodium
carbonate used is often above the theoretically required levels and this gives
UK and European converted sodium bentonites a pH of 9 to 10.5. The extra
sodium carbonate also may slightly flocculate the clay and, as a result, these
bentonites may show higher gel strengths than natural sodium bentonites,
such as those from Wyoming.
Many bentonites used in UK and Europe have also been treated with
polymers to improve their properties (details of the polymer types and
quantities added are normally proprietary to the suppliers). It follows that in
the UK and Europe, bentonites from different suppliers may exhibit varying
properties. Thus, when developing cut off materials, it is necessary to consider
the source of the materials to be used. For major cut off projects, trial mixes
will be carried out at the mixing plant with the same materials to be used for
the main job. In so far as the materials can be specified, there can be problems
in ensuring that even a single supplier provides material from a single source
throughout a long project.
22
Portland Cement. Portland cement is a product of clinkering and grinding

natural silica and calcium bearing materials. Gypsum (calcium sulfate
dihydrate) may be added to moderate the set time. As cement is produced
from natural materials, some variation must be expected between projects
and even during a single project over time. Portland cements are generally
used because of their ability to provide a particular minimum 28-day strength
in concrete. Few, if any cements (outside the special products prepared for
oil-well cementing), are tailored to applications other than structural concrete.
For example, the fluid properties of cement-water systems will be of marginal
significance for concrete, since the workability of a concrete will be largely
controlled by the aggregate. Thus, the very properties of Portland cements
considered essential to grout and slurries, i.e., their early age fluid properties,
may be largely ignored by the manufacturers. Cut off wall designers may
have to test cements from a number of sources in order to achieve an optimum
mix. While not a major issue, additional research is needed on the early age
behavior of clay-cement materials.
Portland cement, due to its high calcium content (typically >60%) and its
gypsum content (designed to rapidly release calcium sulfate into the solution
on wetting), will be a ready source of calcium ions. Therefore, the pore fluid
of a cement-bentonite slurry will contain sodium ions from the bentonite and
other ions (calcium, sodium, potassium, and other cations) from the cement.
The effect of these ions will depend on the particular bentonite and its
sensitivity to the ions. However, the calcium ions are likely to dominate the
sodium. Thus, some time after mixing, the bentonite in a cement-bentonite
mix will be effectively a calcium bentonite. The chemistry of a bentonite first
mixed as a sodium bentonite that later converts to the calcium form may be
indistinguishable from a calcium bentonite which has never been converted.
However, the physical form may be very different and it is this feature that
skilled cement-bentonite mix designers recognize and exploit.
The bentonite in a cement-bentonite mix is short lived (Jefferis, 1995).
Smectite clays will react with lime that is produced by the cement hydration.
X-ray diffraction analysis of old cement-bentonites shows bentonite to be
absent (or below detection limits, which can be relatively high). The reaction
products of the cement and bentonite are calcium silicate and calcium
aluminate hydrates, comparable to those formed in the normal hydration of
cement. That cement and bentonite will react is obvious, but the nature of
this reaction is not well understood. This aspect is often ignored in the
literature, or even worse, special features of the mix are ascribed to the
chemistry of the sodium or calcium bentonite used to prepare the mix.
Research is needed on this aspect, particularly on the quantity of bentonite
that can react with a given quantity of cement.
Ground Granulated Blast Furnace Slag. Slag is a by-product from the blast
furnace. To make it hydraulically active, slag must be rapidly cooled after
tapping from the furnace. For this reason, not all slag is suitable for use in
23
cement-slag-bentonite mixes. When used as a cement replacement material,

it is ground to a fineness similar to or slightly finer than Portland cement. In
comparison to Portland cement, slag has a higher silica content and lower
calcium content (typically about 40%); thus, it does not release calcium ions
into solution at the same rate as Portland cement. Therefore, when slag and
bentonite slurries are mixed, there is no immediate stiffening (and subsequent
relaxation), which differs from the behavior of Portland cement. Also, the
mix shows little bleed and thus the slag does not appear to damage the
bentonite dispersion. However, pure slag-bentonite systems set very slowly.
The slurries may be still very fluid at 14 to 28 days after mixing. Ultimately,
they do set to give high strengths and low hydraulic conductivity. Therefore,
it has been standard practice to blend Portland cement and slag in UK cut off
mixes. The lime in the Portland cement triggers the slag hydration, although
this can also be achieved by the addition of other alkalis. Slag has a number
of desired effects when used in cement-bentonite mixes, including; reduction
of the bleed (settlement of solids with expulsion of free water), increase in the
strength of the mix, and reduction of the hydraulic conductivity of the mix.
The research findings suggest that the effect of the slag is best understood
by considering the proportion of slag by weight to the weight of slag plus
cement in the mix. Cement-bentonite mixes can then be specified by the total
quantity of cementitious material in the mix and the slag replacement level.
The effect of the slag on strength may be quite modest up to a replacement
level of about 60%; but, above this, the strength substantially increases (see
Figure 2-10). Up to 60% replacement, the hydraulic conductivity of a slag mix
is generally comparable to that of a pure Portland cement mix. However,
between 60% and 80%, the hydraulic conductivity typically will decrease by
more than one order of magnitude. Above 80%, there is little further reduction
in hydraulic conductivity.
Shear Strength (mn.m-2)
0.30
Agitation Time, hr.
6
24
95S 80S 95S 80S
60S
0.20
40S
60S
20S
0.10
20S
40S
00S
00S
9 12
Hours
14
28
56 91
Days
Time
Figure 2-10 Shear strength development in slurries containing slag (Jefferis,

1991)
24
When cement-bentonite mixes are used, a hydraulic conductivity of less

than 10-7 cm/s is typically required. To achieve the same low hydraulic
conductivity with a pure Portland cement system may require a total
cementitious content of over three times that of a slag mix. Therefore, the use
of slag can be essential to the economy of the system. However, the resulting
increase in strength is regarded as less desirable, a fact that is not always
considered.
Pulverized Fuel Ash (pfa). In concrete mix design, pfa and slag are often
regarded as comparable replacement materials. Although each has its own
particular features, the final concrete mix, containing pfa or slag, may produce
comparable values of strength. However, in cement-bentonite mixes, pfa and
slag behave quite differently. If used as a cement replacement, pfa does not
show the strength increase or reduction of hydraulic conductivity produced
by slag. Indeed, if it is used at replacement levels over about 30%, it may lead
to very soft mixes. Therefore, it is normally used as an additional material,
rather than as a replacement material, and total cement plus pfa contents
may be as high as 300 to 400 kg/m3 of slurry, with pfa as perhaps 50% or
more of this. These are much higher solids contents than would be normal for
cement-slag-bentonite mixes, where the total cementitious matter seldom
exceeds 175 kg/m3 in the UK. The high solids content of pfa mixes gives
them good resistance to drying and the high pfa content means that strength
may continue to increase and hydraulic conductivity decrease, at least up to
one year after mixing. The early age hydraulic conductivities of pfa mixes are
often comparable to those of pure Portland cement mixes. Also, pfa
replacement may offer better resistance to chemical attack than slag.
Attapulgite Clay. Attapulgite clay has a structure different from that of
bentonite. The particles are lath shaped, rather than plate-like. When mixed
with water, the attapulgite laths can tangle, much as a bundle of straw,
producing a viscous suspension, although with poorer fluid loss control than
bentonite. This effect of attapulgite can be quite limited, but it may be improved
by prolonged mixing (tens of minutes to hours) as the laths can split
longitudinally to produce more and thinner laths, thus greater tangling.
Attapulgite-soil walls and attapulgite-cement walls have both been used in
the strongly saline environment of the Dead Sea. Although their use in other
situations may be appropriate, arguments to use attapulgite in place of
bentonite in a cement-bentonite mix (because of its better resistance to saline
environments) would be spurious. As already noted, the bentonite in a cementbentonite is reacted and removed from the mix to leave cement hydrates.
These hydrates also will be present in an cement-attapulgite mix.
Aggregates. In soil-bentonite mixes and in cement-bentonite-aggregate plastic
concrete, the aggregate should be graded in order to achieve a dense mix in
which the aggregate grains do not settle. Also, the mix should not consolidate
25
unacceptably under the in situ soil stresses. For a soil-bentonite, the carrier
fluid for the aggregate (the bentonite slurry) will be weak in comparison with
most soil stresses; thus, the mix will consolidate until the aggregate experiences
grain to grain contact, or until the slurry has consolidated to the point where
it has sufficient strength to resist the soil stresses.
In plastic concrete, two types of mix are possible. The mix may be designed
so that the aggregate is in grain to grain contact or it may be designed so that
the cement-bentonite mix is of sufficient strength to support the aggregate.
The latter can have two advantages: 1) it may increase the strain at failure,
and 2) the soil stresses will be carried by the cement-bentonite matrix and not
the aggregate. This will ensure that some stress is placed on the matrix, thus
limiting the damage that may be caused by chemical attack.
In the UK, plastic concrete mixes are rarely used in two-phase walls
(separate excavation and replacement processes) since they are more expensive
than single phase walls. Two phase walls are used when an excavation is so
deep or complex that it cannot be achieved before hardening of the cementbentonite mix. Such walls are excavated under a bentonite slurry. Plastic
concrete backfill is necessary to ensure full displacement of the excavation
slurry. Without the aggregate, a cement-bentonite mix may be insufficiently
dense to ensure this displacement.
c. UK Specifications. UK specifications include: specifications for materials
as supplied, specification of fluid properties, specifications for the hardened
properties of the wall material, and durability specifications.
Materials. There are well-tried standards for cement, slag, and pfa in the UK
and, although not developed solely for cut off walls, they do provide an
acceptable base-line; however, there is a problem for bentonite. Traditionally,
the specification used for bentonite has come from the Oil Companies
Materials Association Specification, DFCP-4, Drilling fluid materials,
Bentonite. However, it has been recognized that bentonites optimized (by
the suppliers) for use in civil engineering projects may be outside this
specification and to adhere to specifications would not only be costly, but
would also impair the material properties required in the civil engineering
project.
Unfortunately, the optimum bentonite may depend on the mix
formulation, preparation equipment, and site practice. These are normally
contractor dependent and outside the control of the owner. This highlights a
fundamental issue concerning specifications related to construction work.
Should the work be specified by method and/or materials to be used or should
the work be specified in terms of the expected performance after completion?
In the UK, it is almost universal practice to specify construction work by
performance, especially in the case of cut off walls systems. Although, this
situation does not apply worldwide, there is interest in moving towards
performance specifications in other countries.
26
Specifications based on method may address only the ingredients, but

may not allow the specifier to control the mixing process. Different cut off
wall systems may be more or less sensitive to the balance between good
ingredients and the skills of a contractor to obtain the optimum mix of the
ingredients. Thus, the type of cut off wall used for a specific project may
depend on the type of specification used. Experience in the UK suggests that
cement-bentonite walls cannot be specified by ingredients. The skills of the
contractor, the mixing equipment, the mixing, and special proprietary
additives are all key components. In contrast, soil-bentonite walls may be
more easily specified by method; i.e., defining soil grading and bentonite
content, etc.
Fluid Properties. The fluid properties of a slurry are of little significance,
provided that the slurry performs its proper role in supporting the excavation.
Nevertheless, measurements of fluid properties are often proposed as a
method of quality control for the batching and mixing. As a result,
measurements have been made for nearly every property for which a test
existed. The range of tests has reduced substantially over the last decade. In
the past, Marsh funnel, Fann viscometer, density, bleeding, pH, and filter loss
all might be measured, for both bentonite slurries and cement-bentonite
slurries. However, it was found that the fluid properties (except density and
pH) were strongly influenced by the age and shear history of the slurry.
Accurate batching of the raw materials is now regarded as more important
than tests on the fluid slurry. Typically, all materials are required to be batched
to 3%. Currently, tests on the fluid slurry are generally limited to the
following:
Marsh funnel flow time (if the slurry will pass through the funnel),
density, measured to 0.002 g/ml, using a fixed volume cylinder and
an electronic balance, and
bleeding using a 1000 ml measuring cylinder.
Assessment of the Hardened Material. Ideally, tests should be carried out in
situ. Some success has been achieved with in situ cone penetration tests and,
occasionally, techniques based on pore pressure probes inserted into the wall
have worked; but the reliability is much less than 100%.
Golder Associates (Eiben and Jefferis, 1994) have developed a procedure
for the assessment of walls using two rows of wells on either side of the wall
with the wells spaced at intervals of the order of 20 m in each row). The
procedure is to inject water into each of the wells, in turn, at a defined rate for
a defined period, and to measure the pressure response in the injection well
and the adjacent wells. In this manner, the full length of the wall can be
inspected, with hydraulic conductivities of 10-7 cm/s or greater being
measured to within half an order of magnitude .
As in situ tests are not yet reliable, specimens for testing the hardened
27
material in situ are cast from samples taken from the trench after excavation,
but before the slurry hardens. Cored samples from a hardened wall are
invariably so damaged as to be useless. Despite this, regular attempts are
made to core walls, but the results are invariably of little value.
In general, specifications for the hardened (set) material have required
the assessment of three properties of the mix (plus any durability
requirements). These properties are; hydraulic conductivity, strength, and
strain at failure.
Permeability. For hydraulic structures, a hydraulic conductivity of 10-6 cm/s
may be acceptable, but for contaminant containment, a hydraulic conductivity
of 10-7 cm/s is typically specified.
Strength. Specifications call for a strength criterion or a permissible range of
strengths. At present, there is no consensus as to what strength level is
appropriate (except as it may influence the strain at failure). Typical values
might be >50 kPa (7.25 psi), <1000 kPa (145 psi), 100 to 300 kPa (14.5 to 43.5
psi), etc. A common feature of strength tests from many cement-bentonite
cut off walls is the wide scatter of the results. The reasons for this are currently
the subject of research.
Strain at Failure. A strain at failure of greater than 5% is almost invariably
specified in the UK. This specification is seldom justified. Unfortunately, this
specification cannot generally be met with cement-bentonite barrier materials
except under confined drained triaxial conditions with a confining pressure
on the order of 50% to 100% of the unconfined strength of the material. Even
under these high confining pressures, the strain at failure is not specified as a
strain without increase of hydraulic conductivity (which logically is what is
important and for which there is no test, unfortunately). The specified strain
is referred to as the strain capacity.
For a cement-bentonite material with a strength of 200 kPa (29 psi), it is
unlikely that the 5% strain condition can be achieved, except under a confining
pressure of order 100 kPa (14.5 psi). This pressure level will only exist at
substantial depth in the ground, if at all. Thus, in the upper region of any
cement-bentonite wall, the strain condition will not be achieved and the
material is likely to fail at a strain of order 0.2 to 2%. In an attempt to maximize
strain capacity, one approach is to specify a maximum strength for cementbentonite materials. Despite the evident inconsistency of the criterion, this
latter requirement persists in many specifications. However, the Draft UK
National Specification for Cut off Walls omits this requirement, noting that if
high strains are required, special mixes must be designed (although none has
been found to date) or a geomembrane must be included in the wall.
Durability. Durability is a major issue for vertical barrier walls designed to
prevent the migration of contaminants. The literature on containment systems
28
is replete with references to compatibility tests, which may fail to identify

underlying issues. Further, because of general practice to maximize the
number of pore volumes passed through a sample in minimum time, many
tests have been undertaken at such high confining pressures as to ensure that
any serious chemical damage, erosion, or cracking will likely be masked by
consolidation (Jefferis, 1992). There is need for much greater guidance on test
procedures. Passing contaminants through barriers is a favorite topic of
research, but such efforts will have limited value unless standard procedures
are developed.
For cement-bentonite mixes, it is just as necessary to consider the effect
of water as it is to consider the effect of contaminant chemicals (Jefferis, 1995).
Lime (calcium hydroxide) can be leached from cement-bentonite materials.
Leaching of the lime (and other soluble alkalis) will reduce the pH of the mix
and tend to destabilize the calcium silicate hydrates, causing further release
of lime to yield materials of lower calcium to silica ratio, until all the calcium
has been removed and only hydrated silica remains. This process has been
studied in the laboratory and found to cause significant consolidation of the
sample as the loss of lime weakens the cemented structure. However, it does
not lead to an increase in hydraulic conductivity, but rather to a marked
reduction (even at low confining pressures). The hydraulic conductivity of a
cement-bentonite mix may decrease by over two orders of magnitude over a
year of permeation. It has been suggested that the hydraulic conductivity of
cementbentonite materials will continue to decrease, almost without limit,
during permeation. This seems unlikely, since it fails to recognize the leaching
effect of water.
In contrast to the leaching power of water for lime (water will dissolve of
order grams/liter of lime), the effects of many contaminants dissolved in
groundwater, such as heavy metals at low concentrations, may be trivial except
over very long times. While the pH of the slurry is high, many metals will be
immobilized by precipitation. However, once leaching of lime is complete,
they may be re-released unless they remain sorbed in the hydrated silica.
There is little data on the effects of organic contaminants on cement-bentonite
materials and more research is needed.
Certain materials, such as sulfates, can cause expansion of cementitious
barrier materials. This expansion can lead to spectacular damage to unconfined
samples, in effect, total disintegration. However, the effects are much more
limited if the sample is confined, even under quite modest pressures. As the
cement-bentonite material is of only moderate strength, restraint can force
the expansive phases (such as ettringite and thaumasite) to expand within
the sample rather than to expand the whole sample. Thus, under confined
conditions, sulfate appears to cause some softening and only a rather limited
increase in hydraulic conductivity.
Under confined conditions, the change of hydraulic conductivity, due to
contaminant permeation (with the exception of strong acids), may be limited
to the order of a 10 to 50 fold increase, a significant, but often not catastrophic,
29
increase. For many wastes, such as landfill leachate, the immediate effect of
the waste may be to substantially reduce the hydraulic conductivity, perhaps
one order of magnitude, due to the precipitation of metals, carbonate, etc., in
the alkaline-lime rich environment of the cement-bentonite. This reduction
in hydraulic conductivity may substantially reduce the advective flux of
contaminants; thus, extending the service life of a wall.
The effect of mechanical strain and cracking is also of concern. As already
noted, the strain at failure of a cut off wall is likely to be of order 0.2% to 2%.
Thus, if ground deformations occur, cracking is possible. In this respect, soilbentonite materials may offer an advantage over cement-bentonite materials,
since they can have a higher strain capacity. However, it should be
remembered that cement-based materials have the potential for self-healing
(autogenous healing) if the cracking is not so severe as to allow rapid flow of
water (seeps of water through concrete retaining walls often block and cease).
The blocking mechanism seems to be partly due to deposition of calcium
carbonate and partly further hydration/migration of the cementitious phases.
This phenomenon has been demonstrated with cement-bentonite materials
in the laboratory. In the field, conditions may promote self-healing, since
carbon dioxide concentrations in soil are often higher than in air; thus,
carbonate deposition will be encouraged. Also, flow through cracks could be
slowed by deposition of soil particles.
Drying. Cement-bentonite materials are sensitive to drying and will not swell
on re-wetting. This can be expected, given the calcium rich environment of
the cement and the fact that some or all the bentonite will have reacted with
the cement products to form calcium silicate and calcium aluminate hydrates.
Therefore, capping of cement-bentonite walls to prevent cracking needs to be
done soon after excavation. Capping also provides some useful effective stress
in the upper regions of the trench that will improve durability.
2.3.5 Field Performance of Soil- and Cement-Based Barriers
a. U.S. Field Performance Experience. Soil- and cement-based vertical barrier
walls have been successfully constructed to control groundwater seepage and
to prevent migration of contaminants from a waste disposal site. For barrier
walls constructed to control groundwater seepage and/or provide structural
control, their integrity has been demonstrated by safely excavating from
within the barrier for many hundreds, if not thousands, of projects.
Unfortunately, however, it now appears that reliable estimation of hydraulic
conductivities based on post-construction testing of samples from wall
constructed of cementitious materials cannot be obtained.
It was the unanimous view and experience of the workshop panel that
post-construction sampling and testing of walls constructed of cementitious
materials leads to a high estimate of the hydraulic conductivity for the barrier
material. It is believed that these high values are due to the mechanical damage
30
resulting from the drilling and sampling process (and the associated stress
relief) which alters the microstructure of the cementitious barrier material.
For very soft materials, such as soil-bentonite, on the other hand, postconstruction sampling and testing is reliable, since the consolidation process
during testing appears adequate to overcome any disturbance to the soilbentonite backfill microstructure.
Soil- and cement-based barrier walls are generally considered compatible
with a wide range of organic and inorganic contaminants, if at relatively low
concentrations. However, should organic contaminants be present as nonaqueous phase liquids, the potential for wall deterioration increases and
should be carefully evaluated. Cement-bentonite and cement-bentonite-slag
are subject to cracking, if allowed to dry excessively. Cement-bentonite walls
containing fly ash exhibit somewhat better drying resistance. Leaching
resistance is related to the hydraulic conductivity, i.e., all else being equal, the
lower the hydraulic conductivity the greater the leaching resistance.
b. UK Field Performance Experience. Cement-bentonite based walls are in
regular use in the UK. Indeed, at the end of 1994 it was estimated that cut off
walls were being installed at of order 4 km/month, with most of these walls
constructed at old landfill sites, while very few were constructed for seepage
control around hydraulic structures. Earliest barrier walls for environmental
remediation applications were installed in the late 1960s and early 1970s. There
have been no reported failures in service, even for walls subjected to high
piezometric head gradients, and thus subjected to powerful water leaching.
Current wall materials are capable of achieving hydraulic conductivities
of order 10-7 cm/s using only 40 kg of bentonite and 150 kg of cement
(including slag) per cubic meter of slurry, thus yielding economic mixes. Work
is underway to develop cement-bentonite mixes with hydraulic conductivities
less than 10-9 cm/s using cement-slag-bentonite mixes.
Mixes are presently available with hydraulic conductivities below 10-9
cm/s and with good resistance to inorganic materials, such as acid and sulfates.
These semi-setting materials are based on graded aggregate systems filled
with a silicate-silane gel. They have a strain capacity of order 3%. However,
the costs are high and, in many circumstances, it may be less expensive to use
a geomembrane, unless the diffusion of organics through the membrane is
significant.
c. Factors Affecting Field Performance of Soil- and Cement-Based Barriers.
The factors affecting the field performance of soil- and cement-based barriers
include: curing time, wet-dry cycling, freeze-thaw cycling, fluctuating water
table, state of stress, compatibility, and potential construction defects. These
are discussed below along with ways to minimize negative impacts. Needs
for additional research are also highlighted.
Curing time. There is overwhelming evidence that the hydraulic conductivity
31
of cementitious vertical barrier materials continues to decrease with time over

periods of months to years (depicted in Figure 2-11, Jefferis, 1995). The practice
(in the U.S.) of measuring the hydraulic conductivity after 28 days of curing
will, therefore, result in hydraulic conductivity measurements that will be
higher than those after longer curing periods.
10-8
k-Value (m/s)
10-9
10-10
10
Mixture A
Mixture B
Tap Water
Test Liquid
-11
10-12
10-13
20
60
100
140
180
220
260
Age of Sample (Days)
Figure 2-11 Hydraulic Conductivity values of mixtures A and B as a function

of age and test liquid (Jefferis, 1995)
Cycles of Wetting and Drying. Little work has been done to investigate the
effect of wetting and drying cycles on soil- and cement-based barrier materials.
It is known that when a barrier constructed of soil-bentonite is permitted to
dry, the hydrated bentonite shrinks, and hydraulic conductivity increases.
In the U.S., little is known about cement-bentonite walls subjected to drying.
In the UK, it is believed that cement-bentonite-fly ash cutoff walls are more
resistant to desiccation cracking than cement-bentonite-slag walls. For all
types of barriers, the height of the barrier extending above the water table
may remain intact, if that height is less than the height of capillary rise
(assuming the rate of drying is less than the rate at which moisture is replaced
by capillary suction).
Cycles of Freezing and Thawing. Cycles of freezing and thawing can cause a
general deterioration of vertical barrier materials. However, this deterioration
can be avoided by providing a cover, sufficiently thick to protect the barrier
wall from freezing temperatures.
Fluctuating Water Table. Although the portion of vertical barriers constructed
above the highest water table elevation may be subject to the harmful effects
of desiccation, little is known about the portion of the barrier that may be
subject to the long-term effects of a fluctuating water table. In virtually all
32
applications, variations in the elevation of the water table can be expected,

primarily due to seasonal and long term variations in climate. The effect of
these water table fluctuations on the hydraulic conductivity of the cutoff wall
is unknown. Although it may be assumed that there is no effect, there are
few data to support or refute this assumption. One sampling study found
that the hydraulic conductivity was higher for soil-bentonite cutoff wall
material permanently above the water table than that for material permanently
below the water table (Evans and Cooley, 1993). Additionally, when the soil
sample from above the water table was re-saturated in the laboratory, the
original low hydraulic conductivity was not restored. This may not be
surprising, given that the effort required to hydrate the bentonite is
considerable.
State of Stress. The hydraulic conductivity is influenced by the consolidation
stresses during testing; the higher the stresses, the lower the measured
hydraulic conductivity. Thus, in order to accurately reproduce these stresses
in the laboratory, it is necessary to know the state of stress in the constructed
vertical barrier.
Compatibility. The permeation of organic contaminants at low concentrations
generally has little negative impact on the hydraulic conductivities of clayey
barriers (Mitchell and Madsen, 1987). However, concentrated organic
contaminants, e.g., nonaqueous phase liquids, may cause increases in
hydraulic conductivity (Day, 1993; Rumer and Ryan, 1995). The permeation
of inorganic contaminants, e.g., strong acids or bases, into clayey materials
may bring about an increase in hydraulic conductivity.
Construction Defects. The potential for construction defects is of course
dependent on the construction technique used, the materials employed, and
the QA/QC procedures followed during construction.
2.3.6 Vertical Barriers Enhanced to Impede Contaminant Transport
Recent investigations on different clays have shown that a judicious choice of
clay minerals can enhance the adsorption of some contaminants, thus
impeding their diffusive transport through the barrier (Amann, et. al., 1994).
At present, containment barriers are not designed to inhibit advective
and diffusive transport, nor to treat contaminated water that might permeate
the barrier. New types of containment barriers have been proposed that not
only impedes contaminant migration, but also treats contaminants permeating
the barrier wall. This innovative concept, referred to as an enhanced barrier,
combines containment with other technologies, including: stabilization,
physical-chemical treatment, and biological treatment.
Two fundamentally different types of enhanced barriers have been
proposed. The first, termed a minerally enhanced barrier, must still have a
2.4 ASSESSMENT OF SOIL- AND CEMENT-BASED BARRIERS
33
low hydraulic conductivity (as are present barriers), but will also retard organic
and inorganic contaminants by adsorption, precipitation, and possibly other
chemical reactions. In the second type, a biologically enhanced barrier, the
containment barrier material must still have a low hydraulic conductivity
and be capable of supporting and promoting the biological degradation of
permeating organic contaminants. This second type may also be able to bind
organic and inorganic contaminants, similar to the first type.
Enhanced barriers offer improvements over present barriers, which are
designed only to prevent or impede the migration of contaminants. Although
enhanced barrier technologies are attractive, much research and development
remains to be done in order to make them acceptable, cost effective,
commercially available, and effective throughout the service life of the
constructed containment system. Readers are referred to Sections 10 and 11
for treatments of the mathematical modeling of enhanced barrier systems
and of the design and performance of reactive permeable barriers, respectively.
[It should be noted that enhanced barriers are meant to continue to impede
advective transport, in contrast to the unimpeded flow through permeable
reactive barriers discussed in Section 11.]
2.3.7 Basis of Vertical Barrier Design
Vertical barriers are designed to have hydraulic conductivities of 10-7 cm/s,
regardless of the site specific conditions. However, at some sites, where strength
was an also important consideration, cement-bentonite mixes having hydraulic
conductivities as high as 10-5 to 10-6 cm/s have been permitted. A rational
basis for design may involve a more detailed consideration of the time
dependent contaminant transport through the wall, including retardation
mechanisms, biotic reactions, and abiotic reactions. Materials having the same
hydraulic conductivity may perform quite differently when evaluated in this
manner. Such assessments might lead to evaluation of the time varying
reliability of an installed barrier wall (Inyang and Tumay, 1995). The
performance of a barrier wall could also be based upon the concept of barrier
efficiency, computed by comparing the estimated flux of the contaminant before
and after barrier installation. Performance efficiency could also be based on a
comparison of the concentration of a contaminant within the containment to
the concentration of the contaminant predicted to occur outside the containment
at the end of the design service life of the installed barrier wall.
2.4 ASSESSMENT OF SOIL- AND CEMENT-BASED BARRIERS

For soil-bentonite backfill, it is necessary to select a base soil with a low
hydraulic conductivity in order to minimize potential for deterioration
in hydraulic conductivity that may develop as a result of the imposed
environmental stresses.
34
The addition of fibers to a barrier wall formulation improves the posthardened mechanical properties without significant changes in hydraulic
conductivity.
It is essentially impossible to reliably measure the hydraulic conductivity
of core samples obtained from cement-bentonite barrier walls, due to
disturbance of the sample. The measured results generally produce high
estimates for the hydraulic conductivity.
The hydraulic conductivity of cementitious materials declines as curing
time increases.
The integrity of vertical barrier systems has been verified by excavating
from within the barrier at many projects constructed for ground water
and/or structural control. Much less is known about the integrity of
vertical barriers installed for environmental remediation applications.
Additional monitoring and testing is needed to verify the integrity for
these latter applications and to establish if there are differences in reliability
for walls constructed of different materials. Suggestions for system
performance verification include the use of boxouts, pump tests, injection
tests, construction QA/QC, geophysical exploration methods, and postconstruction sampling and testing. However, the measured values for
hydraulic conductivities obtained from post-construction samples can be
higher than the in situ values, due to sampling disturbances.
Although evaluation of the quality of the materials used for the barrier
wall is important, quality control of the construction process is equally
important.
2.5 RESEARCH AND DEVELOPMENT NEEDS

There is concern about the variation in choice of materials used in vertical
barriers for containment (Potter, 1995). The UK experience indicates widespread use of cement-bentonite, whereas the U.S. experience indicates soilbentonite as the material of choice. The basis for this difference in practice
needs to be clarified. Such clarification would aid in the selection of one
material over another. Also, there remain concerns about the long-term
durability of materials used to construct vertical barriers. Thus, better ways
are needed to predict the service life of a given barrier that include the effects
of environmental stresses such as freeze/thaw cycles, wet/dry cycles,
desiccation, dynamic loads (earthquakes), and fluctuating water table.
Durability is of particular concern for barriers constructed in the vadose zone.
As noted, the integrity of the vertical barrier has been verified for
conventional cutoff walls by excavating from within the cutoff wall, but little
has been done to verify the integrity of vertical barriers installed for
environmental remediation applications. There is a recognized need for system
integrity verification of these barrier walls. Additional monitoring and testing
is needed to verify the system integrity (see Section 12) and to establish if
2.5 RESEARCH AND DEVELOPMENT NEEDS
35
there are differences in the reliability of walls constructed from different

materials. Research that addresses the issue of system integrity should include:
1) investigations into the cause/impacts of variations between mix batches,
including considerations of mixing sequence; 2) study of the long-term
performance of cutoff walls constructed in a contaminated environment; 3)
study of performance differences between cutoffs constructed using different
techniques (e.g., slurry trench, jet grouting, deep soil mixing, etc.) and different
materials (soil- or cement-based); and 4) study of the long term effects of cycles
of freezing and thawing and cycles of wetting and drying on barrier wall
performance and the effects of desiccation of cutoff wall materials.
Additional study is needed regarding the analysis of contaminant fate
and transport through soil- and cement-based vertical barriers. Enhancements
to the vertical barrier (e.g., higher clay fraction in a European wall and zeolitic
clays in U.S. studies) can improve wall performance by lengthening the time
for contaminant breakthrough and reducing contaminant flux through the
barrier. For walls needing added strength, the addition of fibers can be
considered. Fiber addition does not increase the hydraulic conductivity of
the barrier. Several options exist for enhancing the performance of low
hydraulic conductivity walls, including the concept of utilizing the barrier as
an in situ reactor, capable of biotic and abiotic transformation and degradation
of permeating contaminants. Further research, both laboratory and field, is
needed to optimize the use of enhancements and to verify expected field
performance. Permeable reactive walls have also been proposed (see Section
11). They differ from those discussed in this section, in that they are designed
to permit the flow of contaminated groundwater through a permeable reactive
media.
Currently, there is no accepted way of testing walls in situ. Various
techniques are being investigated but have yet to be proven reliable. Thus,
there is a serious lack of information on the actual in situ behavior of
constructed vertical barrier walls at contaminated sites. Data are available
for laboratory tests on samples taken from constructed walls, but few walls
have been monitored for extended periods of time.
The major issues relating to barrier wall materials include: hydraulic
conductivity, strain capacity, durability, cost, and the observed variations of
hydraulic conductivities. Hydraulic conductivities of order 10-7 cm/s are
currently achievable, but the causes of observed variations of measured
hydraulic conductivities between batch mixes need to be better understood.
Although very low hydraulic conductivities can be achieved with current
barrier wall technologies, the incorporation of geomembranes may offer some
advantages (see Section 5).
UK specifications often call for a 5% strain without mechanical failure.
However, this cannot be achieved except under moderate to high confining
pressures. The development of low cost materials with greater strain capacity
is of considerable interest, but is unlikely to be achieved with cement based
materials, thus requiring new and different cutoff material formulations.
36
Compatibility testing needs to be standardized. At present, compatibility

testing is conducted using a variety of procedures and methods. Additional
data are needed to characterize the stress-strain, strength, and compression
characteristics of a variety of soil-bentonite mixes. Standardized design
procedures are also needed for predicting 1) the stress-state in soil-bentonite
backfilled trenches and 2) post-construction adjacent ground movements.
A better understanding of the microstructure of low hydraulic
conductivity barrier materials is required. Even with better data on
microstructure, long-term durability will remain an issue. Chemical
compatibility tests will continue to be essential for some contaminants. Existing
walls that have been in service for many years should be investigated to
determine the effects of permeation and the fate of the permeated
contaminants.
2.6 REFERENCES
Acar. Y B., Hamidon, A., Field, S. D., and Scott, L. (1985). The Effect of Organic
Fluids on Hydraulic Conductivity of Compacted Kaolinite. Hydraulic
Barriers in Soil and Rock, ASTM, STP 874, 171-187.
Allan, M. L. (1995). Cementitious Barriers and the Effect of Admixtures,
Material prepared for the Containment Technology Workshop, Baltimore,
MD, August 29-31.
Allan, M. L. and Kukacka, L. E. (1984). In situ Containment and Stabilization
of Buried Waste, Annual Report FY 1994. BNL 60977, October.
Allan, M. L. and Kukacka, L. E. (1993). In situ Containment and Stabilization
of Buried Waste, Annual Report FY 1993, BNL 49709, October.
Amann, P., Madsen, F. T., Martinenghi, L. S. (1994). Waste Immobilisation,
Soil Liner and Slurry Wall Research in Switzerland, First National
Conference on Environmental Geotechnics, Edmonton, 10-15 July.
Blight, G. E. (1973). Stresses in Narrow Cores and Core Trenches of Dams.
Eleventh International Congress on Large Dams, Madrid, Vol. 3, pp. 3-80.
DAppolonia, D. J. (1980). Soil-Bentonite slurry Trench Cutoffs. Journal of
the Geotechnical Engineering Division, ASCE,106(4),.
Daniel, D. E. (1987). Earthen Liners for Land Disposal Facilities. Geotechnical
Practice for Waste Disposal 87, ASCE, Geotechnical Special Publication
No.13, 21-39.
2.6 REFERENCES
37
Day, S. R. (1994). The Compatibility of Slurry Cutoff Wall Materials with

Contaminated Groundwater, Hydraulic Conductivity and Waste
Contaminant Transport in Soils, ASTM STP 1142, D. E. Daniel and S. J.
Trautwein, Eds., American Society for Testing and Materials, Philadelphia.
Day, S. R. and Ryan, C. R., (1995). Containment, Stabilization and Treatment
of Contaminated Soils Using In Situ Soil Mixing, ASCE Specialty
Conference on Characterization, Containment, Remediation and Performance
in Environmental Geotechnics, The Geoenvironment 2000, ASCE Geotechnical
Special Publication No. 46, February.
Eiben, Th. and Jefferis, S. A. (1994). Implementation of Interactive Numerical
Modeling and Hydraulic Testing in the Design and Quality Control of
Low Permeability Barrier Systems, Meeting on Measurement of Low
Permeability in Soil and Rocks, University of Surrey, Guilford, England,
Nov.
Evans, J. C., and Fang, H. Y. (1982). Geotechnical Aspects of the Design and
Construction of Waste Containment Systems, Proceedings of the 3rd
National Conference on the Management of Uncontrolled Hazardous Waste Sites,
Washington, D. C., November, pp. 175-182.
Evans, J. C. (1991). Geotechnics of Hazardous Waste Control Systems,
Chapter 20, Foundation Engineering Handbook, 2nd ed., Ed. H. Y. Fang,
Von Nostrand Reinhold Company, New York, NY.
Evans, J. C. (1993). Vertical Cutoff Walls, in Geotechnical Practice for Waste
Disposal, Ed. D. E. Daniel, Chapman and Hall.
Evans, J. C., Adams, T. L., and Dudiak, K. A. (1995). Enhanced Slurry Walls
as Treatment Zones for Inorganic Contaminants, Toxic and Hazardous
Wastes: Proceedings of the Twenty Seventh Mid-Atlantic Industrial Waste
Conference, Lehigh University, Technomic Publishing Co., Lancaster, PA,
July, pp. 712-721.
Evans, J. C., and Cooley, B. H. (1993). Slurry Wall Performance. Progress in
Geotechnical Engineering, Hershey, Pennsylvania.
Evans, J. C., Costa, M. J., Cooley, B. (1995). The State-of-Stress in Soil-Bentonite
Slurry Trench Cutoff Walls. Geoenvironment 2000, ASCE Geotechnical
Specialty Publication No. 46,1173-1191.
Evans, J. C., Sambasivam, Y. and Zarlinski, S. J. (1990). Attenuating Materials
in Composite Liners, Waste Containment Systems: Construction, Regulation,
& Performance, ASCE Geotechnical Special Publication No. 26, November,
38
pp. 246-263.
Filz, G. (1995). personal communication.
Handy, R. L. (1985). The Arch in Soil Arching. Journal of Geotechnical
Engineering, ASCE, 111 (3), 302-318.
Hermanns, R. U. (1993). Waste Deposit Encapsulation using Vertical Barriers,
Proceedings of the Joint CSCE-ASCE National Conference on Environmental
Engineering, Montreal, Quebec, Canada, pp. 805-811.
Inyang, H. I. and Tumay, M. T. (1995). Containment Systems for Contaminants
in the Subsurface, Chapter 5, Encyclopedia of Environmental Control
Technology, Gulf Publishing Company.
Janssen, H. A. (1985). Versuche uber Getreidedruck in Silozellen. As cited
in Handy.
Jefferis, S. A. (1981). Bentonite-Cement Slurries for Hydraulic Cut offs, Proc.
International Conference on Soil Mechanics and Foundation Engineering,
Stockholm, June, Vol. 1, pp. 435-440.
Jefferis, S. A. (1992). Contaminant-Grout Interaction, Proc. ASCE Specialty
Conference on Grouting, Soil Improvement and Geosynthetics, New Orleans, LA,
Feb.
Jefferis, S. A. (1995). UK Practice: Materials and their Properties, Material
prepared for the Containment Technology Workshop, Baltimore, MD,
August 29-31.
Khoury, M. A., Fayad, P. H., and Ladd, R. S. (1992). Design, Construction
and Performance of a Soil-Bentonite Cutoff Wall Constructed in Two
Stages. Slurry Walls: Design, Construction, and Quality Control, ASTM
STP 1129, 289-308.
Kosmatka, S. H. (1990). Cementitious Grouts and Grouting, Portland
Cement Association, EB111.OIT.
Kukacka, L. E. (1995). Materials Suitable for in situ Construction, Materials
prepared for the Containment Technology Workshop, Baltimore, MD,
August 29-31.
Little, A. L. (1975). In situ Diaphragm Walls for Embankment Dams,
Institution of Civil Engineers Conference on Diaphragm walls and
Anchorages, London.
2.6 REFERENCES
39
Martinenghi, L. S. (1995). Slurry Wall Material Research with Focus on

Immobilization, material prepared for the Containment Technology
Workshop, Baltimore Maryland, 29-31 August.
Martson, A., and Anderson, A. O. (1913). The Theory of Loads on Pipes in
Ditches and Tests of Cement and Clay Drain Tile and Sewer Pipe. lowa
Engineering Experiment Station Bulletin, No. 31, Ames, lowa.
Mitchell R. A. (1995). Evaluation and Design of Subsurface Barrier between a
Hazardous Waste Landfill and Hazardous Waste Surface Impoundments,
Material prepared for the Containment Technology Workshop, Baltimore
Maryland, 29-31 August.
Mitchell, J. K, and Madsen, F. T. (1987). Chemical Effects on Clay Hydraulic
Conductivity , Geotechnical Practice for Waste Disposal 87, ASCE,
Geotechnical Special Publication No. 13, pp. 87-116.
Potter, K. J. (1995). Owners Perspective in Selection of Vertical Barrier
Materials, written material prepared for the Containment Technology
Workshop, Baltimore Maryland, 29 August.
Rumer, R. R., and Ryan, M. E., eds. (1995). Barrier Containment Technologies
for Environmental Remediation Applications, John Wiley and Sons, Inc.,
New York.
Terzaghi, K. (1945). Stability and Stiffness of Cellular Cofferdams.
Transactions, ASCE, Vol 110, Paper 2253,1083-1119.
Tsitsishvili, G. V., T. G. Andronikashvili, and B. N. Kirov L. D. Filizova. (1992).
Natural Zeolites. New York: Ellis Horwood Limited, pg. 1.
U. S. Environmental Protection Agency. (1987). Construction Quality Control
and Post-Construction Performance Verification for the Gilson Road
Hazardous Waste Site Cutoff Wall, EPA/625/287/065, Cincinnati, OH.
U. S. Environmental Protections Agency. (1984). Slurry Trench Construction
for Pollution Control, EPA/540/2-84/001, Cincinnati, OH.
Xanthakos, P. P., Abramson, L. W., and Bruce, D. A. (1994). Ground
Improvement and Control, John Wiley & Sons, Inc., New York, NY.
40
2.7 BIBLIOGRAPHY
The following papers, which have not been specifically cited in Section 2, are
valuable sources of detailed information.
Aitcin, P.-C., Ballivy, G. and Parizeau, R. (1984). The Use of Condensed Silica
Fume in Grouts, in Innovative Cement Grouting, SP-83, American
Concrete Institute, Detroit, pp. 1-18.
Al-Manaseer, A. A. and Keil, L.D. (1992). Physical Properties of Cement Grout
Containing Silica Fume and Superplasticizer, ACI Materials Journal, V.
89, No. 2, March-April, pp. 154-160.
Allan, M. L. and Kukacka, L. E. (1994). Grout Treated Soil For Low
Permeability Barriers Around Waste Landfills, American Concrete
Institute Materials Journal, Volume 91, No. 4, 355-361.
Allan, M. L. and Kukacka, L. E. (1994). Permeability and Microstructure of
Plain and Fibre Reinforced Grouts, Cement and Concrete Research,
Volume 24, No. 4, 671-681.
Allan, M. L. and Kukacka, L. E. (1995). Strength and Durability of
Polypropylene Fibre Reinforced Barrier Grouts. Cement and Concrete
Research, V. 25, No. 3, 511-521.
Allan, M. L. and Kukacka, L. E. Performance of Cementitious Barrier
Materials. ASTM STP 1240 on Stabilization/Solidification of Hazardous,
Radioactive and Mixed Wastes, in press.
Allan, M. L. and Kukacka, L. E., Blast Furnace Slag-Modified Grouts for InSitu Stabilization of Chromium Contaminated Soil. Waste Management,
in press.
Allan, M. L. and Kukacka, L. E., Comparison Between Slag- and Silica FumeModified Grouts.American Concrete Institute Materials Journal, in press.
Alther, G. R., Evans, J. C., Witmer, K. A., and Fang, H. Y. (1985). Inorganic
Permeant Effects Upon Bentonite, Hydraulic Barriers in Soil and Rock,
ASTM STP 874, pp. 64-74.
Alther, G. R., Evans, J. C., Zarlinski, S. (1989). A Composite Liner to Retain
Inorganic and Organic Contaminants Proceedings of the 10th National
Conference on the Management of Uncontrolled Hazardous Waste Sites,
Washington, D. C., November, p. 543- 546.
2.7 BIBLIOGRAPHY
41
Alther, G. R., Evans, J. C. and Pancoski, S. E. (1988). Organically Modified

Clays for Stabilization of Organic Hazardous Wastes Proceedings of the
9th National Conference on the Management of Uncontrolled Hazardous Waste
Sites, Washington, D. C., November, 1988, pp. 440-445.
Alther, G. R., Evans, J. C., and Andrews, E. (1988). Organic Fluid Effects
Upon Bentonite Proceedings of the 5th National Conference on Hazardous
Wastes and Hazardous Materials, Las Vegas, Nevada, April, pp. 210-214.
Brookhaven National Laboratory. (1993). In-Situ Containment and
Stabilization of Buried Waste, Annual Report for Mixed Waste Landfill
Integrated Demonstration, FY 1993, BNL 49709, Brookhaven National
Laboratory, November.
Brookhaven National Laboratory. (1994). In-Situ Containment and
Stabilization of Buried Waste, Annual Report for Mixed Waste Landfill
Integrated Demonstration, FY 1994, BNL 60977, Brookhaven National
Laboratory, November.
Domone, P. L. J. and Jefferis, S. A., eds. (1994). Structural Grouts, Blackie
Academic and Professional, Cambridge.
Domone, P. L. J. and Tank, S. B. (1986). Use of Condensed Silica Fume in
Portland Cement Grouts, in Fly Ash, Silica Fume, Slag, and Natural
Pozzolans in Concrete, SP-91, American Concrete Institute, Detroit, pp.
1231-1260.
Evans, J. C. (1994). Hydraulic Conductivity of Vertical Cutoff Walls,
Hydraulic Conductivity and Waste Contaminant Transport in Soils, ASTM
STP 1142, D. E. Daniel and S. J. Trautwein, Eds., American Society for
Testing and Materials, Philadelphia, pp. 79-94.
Evans, J. C. and Alther, G. (1991). Hazardous Waste Stabilization using
Organically Modified Clays, Geotechnical Engineering Congress, 1991,
ASCE Geotechnical Special Publication No. 27, June, pp. 1149 - 1161.
Evans, J. C. and Pancoski, S. E. (1989). Organically Modified Clays
Transportation Research Record 1219, Geotechnical Engineering 1989,
Transportation Research Board, National Research Council, pp. 160-168.
Evans, J. C. and Fang, H. Y. (1988). Triaxial Permeability and Strength Testing
of Contaminated Soils, Advanced Triaxial Testing of Soil and Rock, ASTM
STP 977, ASTM STP 977, American Society for Testing and Materials,
Philadelphia, pp. 387-404.
42
Evans, J. C., Stahl, E. D. and Droof, E. (1987). Plastic Concrete Cutoff Walls,
Geotechnical Practice for Waste Disposal 87, ASCE Geotechnical Special
Publication No. 13, June, pp. 462-472.
Evans, J. C., Prince, M., Bernardo, M. and Faulkner, M. (1994). Biologically
Active Slurry Trench Cutoff Walls, Toxic and Hazardous Wastes: Proceedings
of the Twenty Fourth Mid-Atlantic Industrial Waste Conference, University of
Delaware, August.
Evans, J. C. and Manuel, E. N. (1985). Geotechnical Property Testing of
Hazardous Materials and Contaminated Soils, Proceedings of the 6th
National Conference on the Management of Uncontrolled Hazardous Waste Sites,
Evans, J. C., Lennon, G. P., and Witmer, K. A. (1985). Analysis of Soil-Bentonite
Backfill Placement in Slurry Walls, Proceedings of the 6th National
Evans, J. C., Fang, H. Y., and Kugelman, I. J. (1985). Containment of Hazardous
Materials with Soil-Bentonite Slurry Walls, Proceedings of the 6th National
Evans, J. C., Kugelman, I. J., and Fang, H. Y. (1985). Organic Fluid Effects on
the Strength, Deformation, and Permeability of Soil-Bentonite Slurry
Walls, Proceedings of the Seventeenth Mid-Atlantic Industrial Waste
Conference, Lehigh University, Bethlehem, PA, June, pp. 275-291.
Evans, J. C., Fang, H. Y., and Kugelman, I. J. (1985). Organic Fluid Effects on
the Permeability of Soil-Bentonite Slurry Walls, Proceedings of the National
Conference on Hazardous Wastes and Environmental Emergencies, Cincinnati,
OH, May, pp. 267-271.
Houlsby, A. C. (1990). Construction and Design of Cement Grouting, John
Wiley and Sons, New York.
Jefferis, S. A. (1982). Effects of Mixing on Bentonite Slurries and Grouts,
Proc. Conference on Grouting in Geotechnical Engineering, ASCE, Feb.
Kosmatka, S. H. (1990). Cementitious Grouts and Grouting, Portland Cement
Association, Illinois, U.S.
LaGrega M. L., Buckingham, P. L., and Evans, J.C. (1994). Hazardous Waste
Management, McGraw-Hill Book Company, New York, NY.
2.7 BIBLIOGRAPHY
43
Manuel, E. M., Evans, J. C., and Singh, R. D. (1987). Discussion of Hydraulic

Conductivity of Two Prototype Clay Liners, by S. R. Day and D. E. Daniel,
Journal of the Geotechnical Engineering Division, ASCE, Volume 113, No.
GT7, July, pp. 804-806.
Onofrei, M., Gray, M., Pusch, R., Borgesson, L., Karnland, O., Shenton, B.,
and Walker, B. (1992). Sealing Properties of Cement-Based Grout
Materials, Stripa Project Technical Report, 92-28, SKB, Stockholm.
Shackelford, Charles D. (1993). Contaminant Transport, Geotechnical
Practice for Waste Disposal, David E. Daniel, Ed., New York: Chapman
and Hall, pp. 40-41.
Smoak, W. G. and Mitchell, K. D. (1993). Effect of High-Range Water Reducers
on Cement Grout, Concrete International, V. 15, No. 1, January, pp. 56-61.
Tosca, S. Z., and Evans, J. C. (1992). The Effects of Fillers and Admixtures on
Grout Performance, ASCE Specialty Conference on Grouting, Soil
Improvement, and Geosynthetics, 1992, ASCE Geotechnical Special
Publication, January.
Wakeley, L.D. and Buck, A. D. (1986). Effects of Different Fly Ashes and Silica
Fume on Selected Properties of an Expansive Grout, in Fly Ash, Silica
Fume, Slag, and Natural Pozzolans in Concrete, SP-91, American Concrete
Institute, Detroit, pp. 1261-1278.
Weaver, K. D., Evans, J. E. and Pancoski, S. E. (1990). Grout Testing for a
Hazardous Waste Application Concrete International, Journal of the
American Concrete Institute, Vol. 12, No. 7, July, pp. 45-47.
45
SECTION 3
DESIGN, CONSTRUCTION, AND PERFORMANCE
of
SOIL- AND CEMENT-BASED VERTICAL BARRIERS
prepared by
George M. Filz and James K. Mitchell

Virginia Polytechnic Institute and State University
Loren R. Anderson, Utah State University

J. David Betsill, Sandia National Laboratories
Ernest E. Carter, Carter Technologies
Richard R. Davidson, Woodward-Clyde Consultants
Steven R. Day, Geo Con Inc.
Annette Esnault, SIF Bachy, France
Jeffrey C. Evans, Bucknell University
Stephan Jefferis, Golder Associates Ltd., United Kingdom
Mario Manassero, Ingegneria Geotecnica, Italy
Lisa Martinenghi, Institut fur Geotechnik, Switzerland
Robert L. Stamnes, U.S. Environmental Protection Agency
George J. Tamaro, Mueser Rutledge Consulting Engineers
G.A.M. van Meurs, Delft Geotechnics, The Netherlands
David S. Yang, SMW Seiko, Inc.
3.1 BACKGROUND INFORMATION

Vertical barriers for cutoff of fluid and chemical flow in the ground are
frequently used as part of waste containment and remediation schemes at
contaminated sites. The purposes of constructing a vertical barrier are to
capture, contain, or redirect the flow of clean or contaminated groundwater,
soil vapor, or nonaqueous phase liquids (NAPLs). This section focuses on
soil- and cement-based vertical barrier technology, including such important
aspects as site investigation requirements, analysis, design, construction, and
performance. Backfill material issues are discussed where they are relevant
to the other topics of this section, but materials are covered more completely
in Section 2, Soil- and Cement-Based Vertical Barriers with Focus on Materials.
Excavation/replacement and mixed-in-place technologies, e.g., backfilled
trenches and jet-grouting, are discussed; permeation grouting is not covered.
Much information on soil- and cement-based vertical barriers is available
from several sources, including Boyes (1975), Evans (1991 and 1993), Jefferis
46
DESIGN, CONSTRUCTION, AND PERFORMANCE OF VERTICAL BARRIERS
(1981), LaGrega et al. (1994), Manassero and Pasqualini (1993), Millet and
Perez (1981), Ryan (1987), and Spooner et al. (1984). Chapter 3, Vertical
Barriers, in Barrier Containment Technologies for Environmental Remediation
Applications (Rumer and Ryan, 1995) provides a comprehensive overview of
vertical barrier technology. The intent in the current chapter is to expand
rather than duplicate the material in Rumer and Ryan (1995). Accordingly,
only brief background material is given here; the purpose being to provide
the reader with descriptions of technologies that are discussed in more detail
in later sections.
Soil- and cement-based vertical barrier types include soil-bentonite cutoff
walls; cement-bentonite cutoff walls; plastic concrete cutoff walls; cutoff walls
backfilled with mixtures of cement, bentonite, fly ash, ground granulated blast
furnace slag, and/or natural clay; and walls constructed by deep mixing and
jet grouting.
Mixing
bulldozer
Backfill
mixing area
Soils excavated
from trench
Excavating backhoe
Soil-bentonite
backfill
Level of
bentonite-water
slurry
Trench ke
Excavation
face
Unexcavated
soil
Aquitard
Figure 3-1 Soil-bentonite cutoff wall (Rumer and Ryan, 1995)

As shown in Figure 3-1, soil-bentonite cutoff walls are constructed in long
trenches. The trench can be excavated by backhoe, clamshell, milling
equipment, or some combination of these devices. Trench stability is
maintained by excavating under a bentonite-water slurry, which is maintained
in the trench at a level above the adjacent groundwater level. In areas of high
groundwater, it may be necessary to construct a berm along the cutoff wall
alignment in order to maintain the necessary differential head. The differential
fluid pressure between the slurry and the goundwater in the formation creates
a tendency for the bentonite-water slurry to flow into the formation; however,
a
47
the formation filters the bentonite particles from the slurry to create a bentonite
filter cake at the trench wall, provided that the void spaces in the formation
are not too large. The low hydraulic conductivity of the filter cake causes a
large pressure drop to develop across the filter cake. This pressure difference
is important in maintaining the stability of the trench wall.
Either the excavated soil, if it is suitable, or imported soil is thoroughly
mixed with bentonite and water to form the cutoff wall backfill material. If
the backfill is properly designed and mixed, it will have the consistency of a
high slump concrete and will exhibit a low hydraulic conductivity. As the
backfill is placed in the trench, it displaces the bentonite-water slurry to form
the cutoff wall.
Cement-bentonite cutoff walls are constructed by using a cementbentonite-water slurry to stabilize the excavation. The cement-bentonite-water
slurry is left in place to harden and form the cutoff wall. (In Europe, this is
referred to as a one-phase method, as opposed two-phase methods in
which the excavation support slurry is replaced by a separate backfill material.)
After hardening, the cement-bentonite typically has the consistency of a stiff
clay. Since the excavation spoils are not reused as backfill, they must be
disposed elsewhere. Material handling is simpler with cement-bentonite cutoff
walls than it is for soil-bentonite cutoff walls, and a cleaner operation results.
Cement-bentonite cutoff walls can be constructed in long trenches, similar to
those used for soil-bentonite cutoff walls, or they can be constructed as a
series of overlapping panels.
Bentonite
slurry
a
Bentonite
slurry
Bedrock or
aquiclude
Panel 1: Concrete or plastic concrete tremie backfill

Panel 2: Excavation process under bentonite slurry
Panel 3: Unexcavated
Figure 3-2 Plastic concrete cutoff wall
Plastic concrete is a mixture of cement, bentonite, aggregate, and water.

Plastic concrete cutoff walls are usually constructed in panels excavated with
clamshells or milling equipment, as shown in Figure 3-2. Panel support during
48
excavation is provided by bentonite-water or cement-bentonite slurry. The

plastic concrete is placed by the tremie method to avoid both segregation of
the backfill and entrapment of the bentonite-water slurry. Plastic concrete
mixes with high unit weights are especially useful when backfilling deep
trenches. After setting, plastic concrete walls are significantly stiffer and
stronger than cement-bentonite walls.
A recent variation of plastic concrete is soil-cement-bentonite (SCB). SCB
walls are usually constructed using the long-reach backhoes typical of soilbentonite cutoff walls, but with a soil-based (usually clay) backfill containing
bentonite and a small proportion (usually less than 10 percent) of cement.
The amount of cement is minimized to avoid formation of cold joints during
construction. After setting, SCB is similar to cement-bentonite in strength
and to soil-bentonite in hydraulic conductivity.
Other backfill mixtures are also used, especially in Europe and Japan.
Such backfills have included cement, bentonite, fly ash, slag, and natural clays.
When fly ash or slag, for example, are added to cement-bentonite-water
slurries, the solids content may increase, which can result in decreased
hydraulic conductivity and increased resistance to chemical attack.
The preceding methods may all be termed excavation/replacement
methods. Another category of soil- and cement-based vertical barrier
construction technology consists of the mixed-in-place methods, which include
deep mixing and jet grouting. An advantage of the mixed-in-place methods
is that smaller quantities of excavation spoils are created. If excavation spoils
are classified as a hazardous waste, special arrangements will have to be made
for their safe disposal.
Hydraulic motors
to rotate augers
In-line augers
Mixing paddles
Discontinuous
auger flights
3rd
penetration
2m
(Typical)
1st penetration
Soil
mixed
wall
2nd penetration
Figure 3-3 Deep soil mixed cutoff wall
49
Deep mixing (DM) is accomplished by use of a rig fitted with a bank of

counter-rotating augers. As shown in Figure 3-3, overlapping panel
construction is used to create a continuous cutoff wall. Additions such as
cement and/or bentonite are introduced during the mixing process to produce
the desired low hydraulic conductivity barrier. Hard ground or the presence
of boulders can cause problems for the DM augers. The deep mixing
technology is described in more detail by Ryan (1987) and Yang (1994).
Drilling
Jet Grouting
starts
Soilcrete Column
under construction
Repetition of
the process
Cement
Air
Water
Additives
Suspension
Figure 3-4 Jet-grouted cutoff wall

Jet grouting uses high pressure jets to erode and mix the soil with added
cement and/or bentonite. If the jets are rotated, cylindrical piles are formed.
As shown in Figure 3-4, cylindrical piles can be constructed in an overlapping
fashion to create a continuous wall. Good control of vertical alignment is
important to assure wall continuity. Various arrangements of the jets are
possible. In the single-rod system, only cement-bentonite grout is pumped
out of the jets to accomplish both erosion and mixing. The single-rod system
produces a fairly small amount of spoils at the ground surface. In the doublerod system, some jets release compressed air, which can aid the cutting process
and increase removal of cuttings from the ground. In the triple-rod system,
various jets release compressed air, water, and cement-bentonite grout. The
triple-rod system can be operated to accomplish almost complete removal of
the soil and replacement with grout. In a relatively recent innovation, the jets
are not rotated so that panels from a few to several centimeters thick are formed
instead of columns. The jetted panels are angled with respect to one another
50
so that they will intersect and form a continuous cutoff.

The SoilSawTM is a relatively recent technology that uses a reciprocating
beam fitted with jets that erode the soil and mix added bentonite and/or
cement to produce a cutoff wall. In effect, the SoilSawTM is a jet grouting
method in which the jets reciprocate rather than rotate. Figure 3-5 illustrates
the construction process.
Soil/Grout Mixture
Figure 3-5 Cutoff wall installed using a Soilsaw

Deep mixing and jet grouting have an excellent history of successful
application in conventional civil engineering applications and are now being
used in environmental applications. The vendor of cutoff walls constructed
using the SoilSawTM considers this to be a still developing technology.
There is another category of cutoff wall construction that could be termed
a displacement method, in which grout is forced into the ground and adjacent
soil is displaced. The vibrating beam is one such approach. Displacement
methods are not discussed here. Some coverage of the vibrating beam method
is provided in Rumer and Ryan (1995).
In the US, soil-bentonite backfilled trenches are common. In Europe,
cement-bentonite and slag-cement-bentonite walls are used most often. In
the Far East, especially Japan, deep mixing is extensively used.
The following sections present 1) the state-of-practice for design and
construction; 2) currently available field performance verification techniques;
3) an assessment of the technology; and 4) research, development, and
technology transfer needs.

3.2.1 Site Investigation Requirements
Subsurface investigations at contaminated sites where vertical waste
51
containment barriers are being considered have several purposes, including

to: 1) identify the nature, extent, and distribution of contamination; 2)
characterize the geological strata; 3) characterize the hydrogeological
conditions; 4) provide geotechnical data for design of the remediation scheme;
5) provide data and samples for the contractors use; and 6) provide
information for use in establishing a performance monitoring program. In
seismic areas, it is also necessary to determine the proximity to active faults
and to estimate the ground shaking potential.
Site investigation techniques include: 1) reviews of previous studies and
reports, 2) geological studies, 3) surface and subsurface geophysical
investigations, 4) trenching, 5) drilling and sampling, 6) use of monitoring
wells, 7) penetrometer soundings, and 8) in situ testing. Several new site
investigation technologies are being developed, including probes for detection
and analysis of inorganic chemicals and hydrocarbons and in situ hydraulic
conductivity probes (Stienstra and van Deen, 1994).
Special considerations that apply to investigating contaminated sites
include: 1) avoiding sample contamination from the drilling and sampling
equipment, 2) avoiding aquifer cross-contamination by proper drilling and
abandonment techniques, and 3) health and safety issues.
Typically, a phased approach is used for site investigation, with separate
feasibility phase and design phase investigations almost always performed.
Additional investigations are sometimes undertaken during construction and
post-construction periods.
Site investigation requirements during the feasibility phase include: 1)
describing land use; 2) determining site topography; 3) evaluating construction
equipment accessibility; 4) drilling at least two to four borings to obtain
samples; 5) making soil classifications (the Unified Soil Classification System
is preferred); 6) acquiring SPT N-values and/or CPT logs; 7) determining
depth to groundwater; 8) evaluating the hydrogeologic properties of the
subsurface materials; 9) assessing the likely extent of groundwater level
fluctuations; 10) determining groundwater chemistry; 11) locating, identifying,
and describing the bottom aquitard in situations where the wall will be keyed;
and 12) assessing the suitability of in situ materials for reuse as backfill.
Site investigation activities during the design phase should be selected to
provide more detailed information needed for final design and construction.
In particular, the design phase investigation should include: 1) detailed
topography along the wall alignment; 2) borings along the wall alignment
spaced at 100- to 1000-foot intervals, depending on the subsurface variability
and the length of the wall; 3) Standard Penetration Test blow count
determinations (SPT N-values) with Unified Soil Classification System
classifications of the soil samples at 5-foot intervals; 4) index property tests
on the soil samples, including grain-size distributions, Atterberg limits, water
contents, and unit weights; 5) identification of fractures and/or slickensides
in clay strata for evaluation of their influence on open trench stability; 6)
strength tests for situations in which open trench stability is an issue; 7)
52
consolidation tests for situations in which adjacent ground movements after

backfilling are an issue; 8) slurry loss testing in the boreholes if gravel/cobble/
boulder zones are encountered; 9) representative samples for mix design and
compatibility testing; and 10) structure, depth, thickness, hardness, hydraulic
conductivity, etc., of the bottom aquitard in situations where the wall will be
keyed. These data will permit evaluation of the impacts of the identified site
conditions on construction of alternate containment barrier systems.
The samples and data obtained during site characterization studies should
be saved for use during the construction and post-construction monitoring
phases of a project. It is especially important that all relevant site data and
soil property information be made available to contractors to assist them in
bid preparation.
An integrated approach to site characterization is essential. This requires
using complementary investigation approaches, including geophysical
methods, penetrometer soundings, and conventional drilling and sampling.
It is especially important that there be participation by all the relevant
disciplines from the very beginning of the site characterization effort.
Geologists, geohydrologists, geotechnical engineers, environmental engineers,
and others must all be involved. In many past projects, the initial studies
have focused only on characterizing the hydrogeologic regime and on
identifying the nature and extent of subsurface contamination. While this is
clearly very important, other important data that are needed for design and
construction can and should be obtained during the early stages of the site
characterization studies. The time and effort expended for the development
of an integrated interdisciplinary site characterization plan can result in better
design, greater construction efficiency, and less total cost to the owner.
There can be no cookbook approach to site characterization. Each site
and problem is unique. The expertise of qualified professionals in several
disciplines working together is needed to accomplish an appropriate site
characterization. Since uncertainty in site characterization can never be
eliminated, it is always important that the owner be made aware of the
uncertainties and the possible deviations from expected conditions. The need
to educate owners, regulators, and the public of the value and, in fact, the
necessity of the observational method for site characterization requires
continuing emphasis.
3.2.2 Analysis and Design
Clear understanding of the purpose and function of the cutoff wall must be
developed prior to designing the wall and other associated components of a
waste containment system. For example, a cutoff wall can be constructed to
provide full containment, to provide underseepage control, to serve as a
seepage diversion wall, or to be part of a temporary dewatering system. A
cutoff wall can be either a hanging wall, or it can be keyed into an aquitard.
The function of the wall is different in each of these cases, and these differences
53
have impacts on wall design.

A difference in function is also illustrated by the following two
hypothetical examples. In both examples, a cutoff wall is part of a complete
containment system with a cap and a floor that together enclose an area
contaminated with a concentrated, toxic material.
In the first example, the containment system is operated with an inward
gradient maintained by modest groundwater extraction from within the
enclosed area. Such a system can limit contaminant fluxes out of the
contained region to very small amounts owing to the opposing directions
of advective and diffusive flows through the barrier. In such a system
both the pumping rates and the treatment costs would be low. It should
be noted that even though treatment of the extracted groundwater may
be necessary, the groundwater extraction is considered to be a hydraulic
control in this case. It is not part of a pump and treat remediation system,
because the intent is not to restore a contaminated aquifer to a pristine
condition.
In the second example, the inward hydraulic gradient is not maintained.
Instead, the contaminant is allowed to slowly move out of the contained
area under the influence of hydraulic and chemical concentration
gradients. Outside the contained area, the contaminant is at a low
concentration, and natural attenuation processes are utilized to maintain
concentrations at safe levels. This example illustrates use of a containment
system to extend the range of situations over which the natural attenuation
approach is applicable.
The foregoing examples point out that there should not be universally
prescribed approaches for the design and operation of waste containment
facilities. Once the design function is clearly identified and understood, the
following aspects of cutoff wall analysis and design can be undertaken:
establishing the wall geometry, considering backfill stresses and adjacent
ground movements, selecting the backfill type and mix design, evaluating
the suitability of potential construction methods, designing panel joints,
evaluating costs, designing a monitoring system, and considering QA/QC
requirements. A useful design checklist and reference list for soil-bentonite
cutoff walls is available from the US Army Corps of Engineers (1995).
a. Establish Wall Geometry. The aspects of wall geometry that must be
designed include alignment, depth, and thickness. Considerations in
establishing the wall geometry include containment of the contamination,
property ownership, and containment facility function.
The alignment is ordinarily chosen to contain the contaminant within its
present boundaries. In some cases, however, contaminated soil from outside
the enclosed area may be moved to inside the enclosed area. There may also
be benefits in locating cutoff walls a significant distance beyond the present
54
limits of the contamination, provided that property ownership and land use
will allow it. These benefits include increased breakthrough time and
decreased flux at steady state. However, the potential disadvantages include
increased cost of the containment facility, increased volume of contaminated
soil and groundwater, and increased land use requirements.
The required cutoff wall depth depends on the depth of contamination,
the depth to a suitable bottom barrier layer, and the mode of operation of the
facility. If the wall is to be keyed into an aquitard, an adequate embedment
depth must be provided, and the seal at the joint between the wall and the
aquitard must be evaluated.
Considerations in selection of the thickness of the cutoff wall include wall
type, contaminant flux, wall continuity, construction method, and cost. The
design basis for cutoff walls in contaminant containment applications should
preferably be contaminant flux, rather than the single hydraulic conductivity
criterion of k < 1 x 10-7 cm/sec that has been so extensively used in the past.
When flux is considered, thick walls will generally be preferred to thin walls,
because hydraulic and concentration gradients are lower and because there
are more opportunities for chemical adsorption of contaminants by the wall
materials. Thick walls may also be easier to construct without defects than
thin walls. The practical range of wall thicknesses is often defined by the
construction equipment and methods. For example, cutoff walls excavated
using a backhoe or clamshell are usually 2 to 3 feet thick, with occasional
walls as narrow as 1.5 feet and as thick as 5 feet.
b. Consider Adjacent Ground Deformations and Backfill Stresses. Potential
impacts of vertical barrier wall construction on adjacent ground include
instability and subsidence. These are more likely when long open trenches
are excavated and when non-cementitious backfill is used. These risks should
be evaluated by performing stability and stress-deformation analyses.
In addition to providing information about adjacent ground movements,
stress-deformation analyses are also useful for estimating the consolidated
backfill stress-state. Data compiled by Evans and Cooley (1993) clearly show
that the stress state in soil-bentonite backfill has a strong influence on inservice hydraulic conductivity. Similar data for cement-bentonite materials
is provided by Jefferis (1992) and Manassero et al (1995). In addition,
contaminated permeants can increase the hydraulic conductivity of barrier
materials, but their effect is less significant when the material is under high
confining pressure (Acar et al., 1985; Daniel, 1987; and Mitchell and Madsen,
1987). Thus, the stress state in the trench has a critical influence on cutoff
wall performance.
There are at least two mechanisms important in the development of the
final consolidated backfill stress state in soil-bentonite cutoff walls: arching
and lateral squeezing. Arching in soils has been described by Janssen (1895),
Marston and Anderson (1913), Terzaghi (1945), Blight (1973), and Handy (1985)
and has been applied to soil-bentonite backfilled trenches by Evans et al. (1995).
55
In arching theory, the trench walls are rigid and the backfill is compressible.
Consolidation of the backfill results in the deformed shape shown
schematically in Figure 3-6. This deformed shape mobilizes shear stresses at
the trench walls, thereby providing partial support for the backfill and
reducing vertical stresses in the backfill below overburden pressures. The
magnitude of the trench wall shear stress depends on such factors as trench
depth, trench width, position of the water table, and shear strength at the
interface between the backfill and the trench wall. In arching theory, the major
principal stress in the backfill at the trench centerline is vertical and the minor
principal stress is horizontal. Principal stress directions rotate from this
alignment for positions in the trench away from the centerline.
Cutoff wall
Deformed shape
Shear stress mobilized at

trench wall
Figure 3-6 Schematic diagram of arching mechanism

The lateral squeezing mechanism for soil-bentonite backfilled trenches
(Filz, 1996) is illustrated schematically in Figure 3-7. In this mechanism, the
trench walls are not rigid, but can deform as construction proceeds: inward
trench wall movement during excavation, possible rebound during backfill
placement, and further inward movement during consolidation of the backfill.
The backfill is placed as a low-strength slurry and, immediately after
placement, excess pore water pressures are very high. As the excess pore
water pressures dissipate, the backfill consolidates, and the trench walls move
towards the trench centerline. The process can be visualized as if the backfill
were being squeezed laterally between parallel plates. For narrow and deep
trenches, a reasonable analogy would be to consider the trench backfill as if it
were in a 1-D consolidation test laid on its side, with the consolidation load
being applied by the inwardly moving trench walls. In the lateral squeezing
mechanism, the major principal stress at the trench centerline is horizontal
and the minor principal stress is vertical. When this mechanism comes into
play, the magnitudes of the stresses in the consolidated soil-bentonite depend
on the properties and initial stress state of the adjacent ground, in addition to
the factors listed previously for the arching theory.
There is probably a transition and interaction between the two
56
mechanisms, with arching dominant for shallow, wide trenches and lateral
squeezing dominant for deep, narrow trenches. Preliminary calculations have
been performed to estimate the magnitudes of the major (1) and minor (3)
principal effective stresses at the 75-foot depth in the soil-bentonite backfill of
a 100-foot deep cutoff wall constructed in a medium dense sand deposit. The
results are shown as a function of trench width (B) in Figure 3-8. It can be
seen in this figure that the lateral squeezing mechanism dominates for any
realistic trench width, under the particular geometric and material
assumptions of this example.
Building
Cutoff Wall
a) Trench excavation under

bentonite-water slurry
b) Possible rebound during soil- b) Consolidation of the soil-bentonite

bentonite backfill placement
backfill after placement
Figure 3-7 Schematic diagram of lateral squeezing mechanism

The sequence of events shown in Figure 3-7 also illustrates the kind of
adjacent ground movements that can occur during and after cutoff wall
construction, even when there is no risk of instability. The magnitude of
adjacent ground movement depends, in part, on the compressibility of the
trench backfill. As an example, a soil-bentonite cutoff wall was installed
recently as part of a remediation scheme at a contaminated site in Silicon
Valley, California. The cutoff wall passed within about 20 feet of an existing
building, which was being used for a rather sensitive manufacturing operation
at that time. Construction of the cutoff wall caused several inches of settlement,
and the building was rendered unusable. A large lawsuit resulted. In this
case, it may have been more appropriate to use a cementitious backfill and a
panelized construction method.
c. Select Backfill Type and Mix Design. The following backfill types have
been used for cutoff walls: soil-bentonite, cement-bentonite, soil-cementbentonite, plastic concrete, concrete, and other mixtures of cement, bentonite,
fly ash, slag, and natural clay. Soil-bentonite and cement-bentonite walls are
most common in the USA; whereas, cement-bentonite and slag-cementbentonite walls are widely used in Europe, and soil-cement-bentonite walls
installed by deep mixing have been extensively used in Japan. New backfill
materials are being developed to provide some treatment and/or adsorption
57
of chemical contaminants, as well as containment. Selection of material type

is influenced by cost, design function, compatibility with the contaminant,
construction method, and impacts on adjacent ground.
Selection of the backfill type and backfill mix design are discussed in
detail in Section 2, Soil- and Cement-Based Materials for Vertical Barriers,
where the primary focus is on contaminant transport, contaminant
compatibility, and durability issues. There are, however, several additional
design issues relating to material selection that must be considered. These
issues concern stress-strain behavior of the backfill, density, fluidity, and filter
criteria, and they have important influences on construction, post-construction
performance of the wall, and adjacent ground movements during and after
construction.
3000
1500
Composite
Action
1000
1000
Lateral
Squeezing
g
in
ch
Ar
' (psf)
2000
' (psf)
Composite
Action
500
ch
Ar
10
B (ft)
15
Lateral
Squeezing
g
in
20
10
15
20
B (ft)
Figure 3-8 Preliminary calculations of the principal stresses in a soilbentonite cutoff wall at depth 75 feet
As described above, the compressibility of soil-bentonite backfill materials
has an important influence on the stress-state that develops in the consolidated
backfill. The backfill stress-state, in turn, influences the hydraulic conductivity
of the soil-bentonite, the resistance of the soil-bentonite to chemical attack,
and the magnitude of adjacent ground deformations.
For walls subjected to movement, e.g., from nearby excavations, adjacent
surcharge loads, or seismic shaking, ductile backfill materials can maintain
low hydraulic conductivity better than brittle backfill materials, and a soilbentonite backfill material might be favored over a cement-bentonite backfill.
Conversely, if movements of the adjacent ground due to cutoff wall
construction are of concern, a panel construction method with cementitious
backfill might be favored.
The backfill must have a density sufficient to displace the excavation
support slurry, and it should have sufficient fluidity to completely fill the
excavated space without bridging across small openings.
58
The particle size distribution of a non-cementitious backfill should satisfy

filter criteria so that it is internally stable and so that it will not be eroded out
into the adjacent natural ground under the applied hydraulic gradients.
All of these backfill material issues should be considered during cutoff
wall design. In some cases, the design criteria work in opposite directions, as
indicated in the following examples:
A high water content in a soil-bentonite backfill can have the desired effect
of making the backfill more fluid, but can also make the backfill more
compressible, which could lead to larger adjacent ground deformations.
A high natural clay content in a backfill can decrease hydraulic
conductivity and increase the potential for contaminant adsorption, but
can also make the backfill less dense, which could make complete
displacement of the excavation support slurry more difficult.
A high cement content in a backfill will increase strength, but it can also
make the barrier less ductile and more likely to crack.
d. Evaluate Suitability of Potential Construction Methods. The suitability
of different cutoff wall construction methods depends on such factors as the
type and hardness of the formation to be excavated, the required depth, the
existence of natural and man-made obstructions, and the desirability of
continuous versus panelized construction. For example, chisels or milling
equipment may be necessary to excavate through hard materials or to deal
with obstructions. Compared to panelized construction, continuous
construction methods such as backhoe excavated trenches or the SoilSawTM
reduce the number of joints in the completed cutoff wall. Panelized walls
constructed using clamshells, milling equipment, deep mixing, or jet grouting
provide an opportunity to use cementitious backfill materials with short setup times and high strengths. Panelized construction methods open only a
small part of the ground at any one time, thereby reducing the potential for
adjacent ground instability and movement problems.
e. Design Panel Joints. Depending on the panelized construction technique
used, cold joints can be formed either with or without waterstops. For
example, waterstops can be used on panels excavated with a clamshell, but
not on panels constructed using deep mixing.
The consensus among European experts is that cold joints formed without
waterstops during overlapping panel construction do not compromise the
integrity or function of panel cutoff walls.
f. Evaluate Costs. A soil-bentonite cutoff wall constructed in a trench
excavated by a backhoe is one of the least expensive types of cutoff wall.
Costs can increase when cement is used and when other excavation or mixing
techniques are employed.
59
g. Consider QA/QC Requirements. Because the designers have intimate

familiarity with the assumptions, analyses, and details of the wall design,
they should be involved in establishing the QA/QC requirements for
construction.
3.2.3 Construction
a. Mixing Considerations. Almost all soil- and cement-based vertical barriers,
except for cement-bentonite, require at least two mixing operations: one to
mix the slurry and another to mix the backfill. Even in-situ mixing requires a
slurry or grout mixing operation to supply the DM augers. Different mixing
methods produce different mixtures, and different materials may require
different mixers. As an example, a pugmill supplies a high mixing energy,
but mixes only 1 to 2 yd3 at a time and adds air to the backfill; whereas,
bulldozer methods mix hundreds of cubic yards per batch. As another
example, attapulgite clay cannot be mixed in ponds because it will settle; a
very high energy mixer is needed to develop a high gel strength.
Mixtures containing cement require special consideration, because if they
are disturbed after the start of the hardening process, the set will be lost. For
soil-cement-bentonite mixtures, even the order of mixing the ingredients is
important. Day (1995) cites the following example. Two mixes of sandy claybentonite-cement were prepared, each containing 4 percent bentonite and 10
percent cement. In the first, bentonite was mixed with water and added to
the soil, with the cement then added dry. The 28-day unconfined compressive
strength was 25 psi. In the other, the cement and water were added to the soil
as a slurry, and then the bentonite was added dry. The 28-day compressive
strength was 249 psi.
For the majority of slurry walls, mixing is done beside the trench. Remote
mixing may be required when there is a limited mixing area available near
the trench, when borrow is to be used as the backfill, when cement is an
ingredient of the backfill, or when unusually strict quality control is necessary.
Whatever mixing location is chosen, there must be adequate area available
for material storage, equipment, and mixing. Guidelines for soil-bentonite
walls are that the working platform should be as wide as the trench is deep,
plus 10 ft. For deep mixing, the working platform should be as wide as the
pretrench, plus the width of the DM crane. A slurry mixing plant area for
cement-bentonite and deep mixing should be about 75 ft by 100 ft.
b. Construction Costs. Many factors influence construction costs including:
construction method, backfill material, soil conditions, local labor and
materials cost, length and depth of the barrier walls, working room and access,
season and weather, contractor experience, and level of personal protection
required. Some current installed barrier costs (Day, 1995), exclusive of
mobilization, engineering, work platform, and disposal, are summarized in
the table below.
60
WIDTH
(FT)
DEPTH
(FT)
UNIT COST
($/SF)
PRODUCTION RATE
PER 10 HRS
(SF)
Soil Bentonite
2-3
80
2-8
2500 - 15000
Cement Bentonite
2-3
80
5 - 18
1000 - 8000
Biopolymer Drain
2-3
70
7 - 25
1500 - 5000
Deep Mixing
2.5
90
6 - 15
1000 - 8000
DM Structural
2.5
90
15 - 30
1000 - 3000
1.5 - 3
200
30 - 80
300 - 2500
one row
200
40 - 100
200 - 1000
WALL TYPE
Jet Grouting
Grout Curtain
Another set of estimated costs is provided by Yang (1995):
MATERIAL TYPE
INSTALLATION METHOD
PRICE
($/SF)
Soil-Bentonite
Slurry Trench
3-8
Cement-Bentonite
Slurry Trench
6 - 14
Soil-Bentonite
In Situ Mixing
5-9
Soil-Cement
In Situ Mixing
6 - 14
For comparison, the following table provides information about vertical

containment barrier costs in Europe (Esnault, 1995).
COST
($/SF)
PRODUCTION
RATE PER 10 HRS
(SF)
Slurry/Backhoe
6 - 10
3000 - 5000
Slurry + Membrane/Backhoe
10 - 16
2000 - 3000
Slurry/Clamshell
16 - 20
1000 - 1500
Slurry + Membrane/Clamshell
18 - 22
800 - 1200
Plastic Concrete/Clamshell
25 - 30
1000 - 1500
Plastic Concrete/Cutter
50 - 80
TYPE
Vibratory Beam
3-5
3500 - 5000
Jet Grouting
15 - 30
2500 - 3500
Colmix (DM)
12 - 16
250 - 500
61
3.2.4 Specifications
The technical specifications, along with the drawings, are the means by which
the design engineer communicates the location, dimensions, materials, and
workmanship of the proposed construction to the contractor. If the
specifications are inadequate, the desired integrity and performance of the
barrier cannot be assured. The technical specifications may include sections
covering the scope of the work, site conditions, required contractor
qualifications, submittals, materials, equipment, workmanship, required
quality control activities, and measurement and payment.
As more contractors enter the vertical barrier construction market, it
becomes increasingly important to require specialized contractor experience.
Good general contractors do not necessarily make good specialty contractors.
The experience of the contractors job site supervisor is especially important.
It is important to provide bidders, at least by reference, the data regarding
subsurface conditions. As the time set aside for bid preparation is often very
short in comparison with the time devoted to the site investigation and design
phases, it is unreasonable to expect a bidder to conduct his or her own
investigations during the bid period and to be aware of all the conditions that
may influence constructability, quality, and costs.
Serious difficulties can occur when both performance (the end result) and
procedures (the method) are specified for the same item, as unfortunately
often happens. For example, the contractor may be provided a formulation
for the barrier, told how to construct it, and then held responsible for the
performance of the end product; e.g., attainment of a hydraulic conductivity
less than a specified value. Large claims are likely in such cases. Nevertheless,
both performance and procedure specifications can successfully be used for
different items in the same specification. The objective should be to provide
the most flexibility possible for the contractor, while maintaining sufficient
control of the materials and construction that the design requirements and
the owners needs are satisfied.
Specifications for mixing are often vague on equipment and mixing
quality. Both the designer and the contractor must be aware that different
mixing methods can produce different results. In the case of mixing bentonitewater slurry for excavation support, low shear equipment requires use of
storage ponds to complete hydration of the bentonite over a 24-hour period;
whereas, use of a high shear, colloidal mixer may produce the same quality
slurry in 5 minutes, and the 24-hour hydration requirement that exists in some
specifications becomes unnecessary.
Most vertical barrier construction specifications will contain at least the
following sections:
Introduction
Scope of Work
Site Conditions
Definitions
62
Qualifications
Experience of both the firm and the key personnel
Submittals
Work Procedures
QA/QC Procedures
Materials
Requirements for the raw materials should be specified
Equipment
Equipment capabilities should be defined
Construction
Equipment Decontamination
Site Work and Site Maintenance
Location and Dimensions
Methods of Slurry Mixing
Slurry Backfill and Mix Properties
Clean-up
Inspection and
Testing
Should assign roles and responsibilities among

contractor and owner/engineer
Health and
Safety
An on-site health and safety officer is usually required
Measurement
and Payment
For most projects, a unit area basis is used
References
Appropriate ASTM or other procedures should be cited
Examples of guide specifications for slurry walls are available from Evans
(1995), the Institution of Civil Engineers (1995), and the US Army Corps of
Engineers (1994). Additional information on specifications is presented by
Millet et al. (1992).
3.3 FIELD PERFORMANCE OF THE TECHNOLOGY

3.3.1 Construction QA/QC
Construction quality assurance and quality control (QA/QC) activities are
critical for success of a cutoff wall project. The purpose of quality control
(QC) activities, which are the responsibility of the contractor, are to generate
data that the contractor can use to make judgments about the suitability of
materials and workmanship, and to modify operations if they are found to be
inadequate. Quality assurance (QA), which is the responsibility of personnel
independent of the contractor, provides the means by which the owner and
the owner s engineer can verify that the contractor is satisfying the
requirements specified in the contract documents. In order for quality
63
assurance to be effective, the owner and/or the owners engineer must possess
the knowledge to identify, and the courage to reject, inferior work.
QA/QC activities can include the following types of tests and
observations: tests of the excavation stabilization slurry, tests of the backfill
mix, observations of the excavation alignment and depth, observations of the
backfill mixing and placing operations, observations of ground movements
adjacent to the trench, observations of overlap in panelized wall construction,
and evaluation of the cutoff wall connection to the bottom aquitard and the
overlying cap. Some items of particular importance to the QA/QC program
include the following:
Equipment must possess adequate size and capability to perform the
work.
The excavation support slurry must have sufficient density to support
the excavation. It must not be so dense from suspended solids that the
solids settle out or that the backfill is not easily capable of displacing the
slurry.
Constant vigilance is necessary with respect to the depth of excavation
into the aquitard and clean-out of the key. A seemingly minor deviation
can seriously compromise the effectiveness of the barrier. Continuous
surveys of the bottom depth are necessary.
Proper batching and mixing to create a homogeneous backfill are essential.
QA/QC activities must include observation of the mixing process and
testing of the backfill.
Backfill placement must be accomplished so as not to entrap pockets of
high hydraulic conductivity material or excavation support slurry.
Continuous observations of backfill placement procedures and surveys
of backfill slope are necessary.
Post-construction surveys of backfill consolidation should be made. These
can occur within a few days for cementitious backfill, but may need to be
carried out over a period of several months for soil-bentonite backfill.
The quality of the connection between the vertical barrier and the cap
should be verified.
Continuous observation of the cutoff wall construction should be made,
and all irregularities should be noted on as-built drawings.
Additional description of important QA/QC activities is presented by
Tamaro and Poletto (1992).
European and US QA/QC practices are similar in that both the contractor
and another agency perform tests and observations of the construction. The
specific tests performed are also similar.
Experience has shown that it is virtually impossible to obtain undisturbed
samples of low-strength, cementitious backfill material because of the
disruptions and stress changes associated with the sampling process. As a
result, samples of such material obtained from the completed cutoff wall are
64
likely to exhibit hydraulic conductivities more than an order of magnitude

too high because of internal structural changes (Hollenweger and Martinenghi,
1994). Manassero (1994) has proposed use of the piezocone penetration test
(CPTU) to assess the in situ hydraulic conductivity of cement-bentonite cutoff
walls. This method shows promise, but it is still in the developmental stage.
An essential element of the QA operation is that continuity of personnel
should be maintained throughout the life of a project. For example, the cutoff
wall designer should observe construction activities to determine whether
actual conditions deviate from design assumptions sufficiently that
modifications to the original cutoff wall design are necessary.
The best assurance of high quality construction is to select qualified and
experienced designers and contractors to do the work.
3.3.2 Performance Monitoring
Performance monitoring is essential to establish whether the barrier function
of the wall is being met. Although expert consensus is that soil- and cementbased vertical barriers can achieve, and have achieved, a high degree of
contaminant containment at many sites, there presently appear to exist fewer
well-documented cases of satisfactory wall performance in environmental
applications than would be desireable to establish widespread confidence in
the technology. On the other hand, documented poor performance at some
cutoff wall projects has lead to skepticism by some regulators and members
of the public. Such skepticism can only be overcome by performance
monitoring of cutoff walls that do their job as designed. Performance
monitoring can also serve the following purposes: to provide an early warning
system for release of contaminants, to discover when and where to implement
corrective repairs, to obtain performance data for future use, and to reduce
the costs of remedial activities.
There are important interactions among site characterization, design,
construction, QA/QC activities, and performance monitoring. As one
example, site characterization studies provide information that is essential
for locating performance monitoring equipment. Another interaction is that
both QA/QC activities and performance monitoring provide information
useful to assess the integrity and effectiveness of a cutoff wall.
Performance monitoring of cutoff walls can include tracer tests, global or
local pumping tests, continuity verification, contaminant concentration
measurements, and periodic sampling and testing of the wall itself.
Tracers have been used in aquifer studies and other applications, but have
been underutilized for cutoff wall performance monitoring due to lack of
knowledge of existing tracer technology and insufficient research and
development to overcome limitations of the technology. Some of the
limitations relate to tracer toxicity, the ability of tracers to simulate
contaminants, and the ability of tracer and sensor networks to reliably monitor
the overall performance of the large surface area of a typical cutoff wall (Betsill
65
and Gruebel, 1995).

Global pumping tests can be used to test the permittivity of an entire
containment facility. While conceptually simple, there are significant
difficulties involved in the geohydrologic interpretation of data from pumping
tests that influence the vast regions of the subsurface that can be enclosed by
cutoff walls. In addition, global pumping tests may not be very useful for
determining defect locations.
Pump tests performed in small box-out enclosures improve the ease and
reliability of barrier permittivity estimates along specific parts of the wall.
However, the results of such tests apply directly only to the boxed-out region.
The idea is that if the same construction procedures are used for the rest of
the barrier, it should have the same permittivity as the barrier forming the
box-out region. Such box-out tests are performed frequently in Europe, and
it is recommended that they be used more often in the US.
Local pumping tests involve injecting or withdrawing water on one side
of the wall and measuring the response on the other side of the wall (Eiben
and Jefferis, 1994). A series of local pumping tests can provide useful
information to help locate a defect. Like global pumping tests, the
interpretation of local pumping tests is dependent on interpretation of
geohydrologic conditions. Slug tests within barrier walls have produced
results that are not reliable.
In principle, continuity verification can be accomplished by probing
methods. However, a great deal of probing is required to obtain a high
probability of encountering a randomly located defect. In addition, probing
can potentially damage the completed wall.
Geophysical methods, while attractive from the standpoint of speed, low
cost, and their non-intrusive or minimally intrusive character, suffer from
imprecision in their ability to detect and locate defects and the ambiguity
often associated with their interpretation. The conclusions from an evaluation
of geophysical techniques done by Sandia National Laboratories (Borns, 1995)
were that high-density networks of sensors, closely spaced vertical and/or
directionally drilled access holes, and high frequency geophysical methods
are needed to obtain the resolution (e.g., decimeter) needed for containment
monitoring.
Contaminant concentration measurements made outside the contained
area represent an attempt to directly monitor the ultimate objective of a
containment facility, i.e., how much contaminant is actually escaping. While
expected statistical variations imply that fixed values at any one location
should not be specified, the general principal of monitoring contaminant
concentrations outside the enclosed area should be supported. There are,
however, very great challenges associated with installing a sensor network
capable of reliably monitoring an entire cutoff wall. It should also be noted
that there is a Catch 22 in this approach to monitoring, and that is that
contamination has breached the barrier when it is detected, and the purpose
of the barrier system is to inhibit contaminant escape in the first place.
66
The preceding considerations are applicable primarily to monitoring

below the water table. At present there are no specific regulations under
CERCLA or RCRA for vadose zone monitoring. Technical guidance for vadose
zone monitoring currently is being developed for RCRA Subtitle C facilities
(EPA, 1995) and by ASTM Section D-18.21.02.
The measurement of quantities in addition to chemical concentrations
and groundwater conditions can provide valuable information about the
condition of the wall and changes that may be occurring with time. These
include weather conditions, surface water conditions and drainage patterns,
gas accumulations, and temperature. Data on gas generation and temperature
distributions, and their changes with time, provide indications of chemical
and biological transformations, either of which could impact the barrier
effectiveness. For example, high temperatures can adversely impact high
water content clay-water mixes through drying and shrinkage cracking.
Settlement and inclinometer measurements in the vicinity of slurry walls can
provide evidence of ground movements that could impact wall integrity.
Long-term monitoring of a barrier wall by periodic sampling and testing
of the wall itself, is, at first glance, an attractive approach for continued quality
assurance. There are difficulties in this approach, however, in that the samples
can only represent a very small fraction of the total wall volume, thus making
it almost impossible to detect gross wall imperfections such as cracks, fissures,
and holes. In addition, evidence is now strong that weakly cemented materials
are sufficiently disturbed by the sampling process that internal expansion
and microcracking can result in unrealistically high values of hydraulic
conductivity being measured in subsequent laboratory tests, as mentioned
earlier. According to Jefferis (1995), samples cored from a hardened wall are
invariably so damaged as to be useless.
3.3.3 Long-Term Performance
Unfortunately, little field information is available that can be used as a basis
for definitive conclusions about the long-term durability of soil- and cementbased vertical barrier walls in environmental applications. On the other hand,
the results of long-term observations of soils and cement-treated soils in other
field applications, as well as the results of chemical compatibility testing in
the laboratory, provide some confidence that, except in highly aggressive
chemical environments, or when subjected to repeated wetting-drying and
freezing-thawing cycles, properly prepared containment wall materials may
be able to maintain their barrier function over long time periods. Clays are
very stable materials, and their properties are not significantly affected by
dilute solutions of organic contaminants. However, these considerations in
themselves are not sufficient to assure that initial design criteria will remain
satisfied over the full design life of a barrier. Thus, there is continued need
for long-term studies of the field performance of soil- and cement-based
vertical barriers.
3.4 ASSESSMENT OF THE TECHNOLOGY
67
Jefferis (1995) summarized several important aspects of the durability of

cement-bentonite walls. Water permeation causes leaching of lime, which in
turn can result in weakening of the structure and consolidation, accompanied
by a significant decrease in hydraulic conductivity. Sulfates can cause
expansion of cement-bentonite which, under unconfined conditions, can result
in disintegration of the material; however, under confined conditions, the
expansive phases (ettringite and thaumasite) develop within the material
rather than expanding it, with only limited effects on the hydraulic
conductivity. Jefferis (1995) noted also that changes in hydraulic conductivity
as a result of contaminant interaction under confined conditions may be
limited to the order of a 10 to 50 fold increase, which while significant, may
not be catastrophic. In some cases, interactions between waste and cementbentonite may cause a decrease in hydraulic conductivity because of the
precipitation of metals and carbonates in the high pH environment.
Mechanical strain and cracking of cement-bentonite walls caused by wall
movements may be of concern because brittle failure can occur at low strains
(0.2 to 2 percent) under confining pressures less than the unconfined
compressive strength (Manassero et al, 1995). Jefferis (1995) points out,
however, that cement-based materials also have the capacity for self-healing
(autogenous healing) provided the cracking does not allow rapid water flow
through the cracks. Drying of cement-bentonite walls is also to be avoided,
as once dried, they will not swell on rewetting.
Standardized or accepted methods for assessing wall system durability
are not currently available. The need for, and approaches to the development
of, damage accumulation models are discussed by Inyang (1995), and
additional work in this area is needed.

The expert consensus is that soil- and cement-based vertical barriers, if
properly designed and constructed, can serve a very useful waste containment
function. The U.S. Army Corps of Engineers (1995) has used soil- and cementbased vertical barriers extensively as cutoff walls in dams and levees, and
has also been very successful in using these barriers for controlling pollutants,
contaminated groundwater, and landfill leachate migrating from waste sites.
Because of these successes, slurry walls have largely replaced traditional cutoff
barriers such as steel sheet pile walls and grout curtain walls at hazardous
waste sites. A successful application of a vertical barrier at a contaminated
site in Kent, Washington, is described in US EPA (1995b).
Jefferis (1995) states that cement-bentonite cutoff walls are being installed
in the UK at a rate of about 4 km per month, with most being for old landfill
sites. These types of walls were first used there starting in the late 1960s, and
no failures have been reported, even for walls in high water leaching
environments. He notes further that mixes are being developed that will
68
have hydraulic conductivities less than 1 x 10-11 m/s.

3.4.1 Applications
Soil- and cement-based vertical barriers are potentially applicable wherever
it is necessary to contain, capture, and redirect the flow of clean or
contaminated groundwater, vapor, and free phase liquids within the ground.
3.4.2 Limitations
Limitations of cutoff wall technologies are principally related to depth,
obstructions, site access, ability to excavate the formation, reuse of the
excavation spoils, exposure of the excavation spoils leading to contamination
in another medium, uncertainties about the long-term properties and integrity
of barrier walls, concern about potential incompatibility between the wastes
and wall material, and lack of acceptance by regulatory agencies and the
public. Most of these limitations can be addressed by proper design,
construction, and monitoring, as well as documentation of successful case
histories.
3.4.3 Reliability
Reliability issues can be classified in two groups: defects that occur during
construction and changes that can occur with time. Defects occurring during
construction result from inadequate mixing and placement of materials, and
include holidays (gaps in the continuity of the wall), leakage at panel joints,
leakage at the connection to the bottom aquitard, and the possible development
of a gap between the cutoff wall and cap. Long-term effects can include
property changes of the intact barrier material and structural changes, i.e.,
cracking, in the wall that could be caused by excessive ground movements,
drying, and the like. Long-term property changes are primarily materials
issues and are covered in Section 2. Design details can be implemented,
however, to address long-term property changes caused by desiccation and
freeze-thaw cycles, for example. If significant ground movements are
anticipated, the wall material type should be selected to accommodate for
them.
3.5 NEEDS
3.5.1 Needs for Research and Development
The most important research and development needs for improved acceptance
of soil- and cement-based vertical barrier wall technology are in the areas of
performance monitoring and evaluation of in-service walls. Important
3.5 NEEDS
69
monitoring topics include the improved use of tracers and pumping tests,
development of methods for continuity verification, application of new sensor
technologies, and the development of geophysical methods that can provide
the level of resolution needed for detection and isolation of leaks.
A particularly important need in the current regulatory environment is
for reliable methods to determine the in situ hydraulic conductivity of the
cutoff wall material. As noted previously, samples of cementitious material
obtained from barriers after they have set exhibit hydraulic conductivity values
that are too high, apparently due to sample structure disturbance that is not
corrected by reapplication of in situ confining pressures. In situ slug tests
have also been shown to be less than reliable, which may again be due to
disturbance associated with the drilling process. Larger scale field pump test
methods avoid disturbance problems, but they are costly and often difficult
to interpret. It would be very useful to show that hydraulic conductivity
values from laboratory cured specimens of backfill correlate well with in situ
values, or to develop some other cost effective and reliable means for
determining the in situ hydraulic conductivity of vertical barriers.
Important related research needs are development of improved QA/QC
procedures and development of improved understanding of the long-term
properties of barrier materials and damage accumulation under in-service
conditions. Advances in both of these areas can increase confidence in the
integrity and function of these types of cutoff walls.
Other areas of research need include improved site characterization tools
and development of improved analysis procedures to estimate the backfill
stress state, adjacent ground deformations, stress-strain behavior, and related
hydraulic conductivity.
There is also need for a large-scale model user facility in which
construction technologies could be evaluated, panel joints could be studied,
and performance monitoring technologies could be developed and tested
(Betsill, 1995).
Well-documented and well-monitored demonstration projects would be
helpful in the conduct of these types of research and development activities,
as well as in building increased confidence in the effectiveness of soil- and
cement-based barrier walls as components of environmentally protective
waste containment systems.
3.5.2 Needs for Technology Transfer
The major technology transfer need is the communication of well-documented,
successful cutoff wall case histories to wall designers, to owners, to regulators,
and to the public.
There is also need to communicate the factors that contribute to successful
cutoff wall installations. Some of the more important items include the need
for: 1) integrated site characterization studies that utilize experts from all the
relevant disciplines, 2) cooperation and understanding among the owner,
70
regulatory agency, contractor, and engineer, and 3) consideration of all the

important design issues, which can include wall function, wall geometry,
backfill type and mix design, contaminant flux, construction feasibility, trench
stability, adjacent ground movements, costs, QA/QC requirements, and longterm performance and durability.
3.6 SUMMARY AND RECOMMENDATIONS

Soil- and cement-based vertical barriers are potentially applicable wherever
it is necessary to contain, capture, and redirect the flow of clean or
contaminated groundwater, vapor, and free phase liquids within the ground.
The expert consensus is that these barriers, if properly designed and
constructed, can serve very useful and environmentally protective
containment functions at contaminated sites. Limitations of cutoff wall
technologies are principally related to depth, obstructions, site access, ability
to excavate the formation, reuse or disposal of the excavation spoils, exposure
of the excavation spoils leading to contamination in another medium,
uncertainties about the long-term properties and integrity of barrier walls,
concern about potential incompatibility between the wastes and wall material,
and lack of acceptance by regulatory agencies and the public. Most of these
limitations can be addressed by proper site characterization, design,
construction, and the monitoring of successful case histories.
An integrated approach to site characterization is essential. This requires
using complementary investigation approaches and participation by all the
relevant disciplines from the very beginning of the site characterization effort.
A clear understanding of the purpose and function of the cutoff wall must
be developed prior to designing the wall and other associated components of
a waste containment system. Important design issues that must be addressed
include wall geometry, backfill type and mix design, contaminant flux,
construction feasibility, trench stability, adjacent ground movements, costs,
QA/QC requirements, and long-term performance and durability.
The quality of the technical specifications have an important influence
on the quality of the completed construction. Specification writers must
understand not only the technical objectives of the work, but also the
construction processes involved. The specifications should provide, at least
by reference, the data regarding subsurface conditions.
A soil-bentonite cutoff wall constructed in a trench excavated by a backhoe
is one of the least expensive types of cutoff wall. Costs increase when cement
is used and when other excavation or mixing techniques are employed.
Well-designed and defined construction quality assurance and quality
control (QA/QC) activities are critical for success of a cutoff wall project.
Constant vigilance on the part of the QA/QC personnel is necessary in order
to achieve and document good quality construction. This objective is best
accomplished by selecting qualified and experienced designers, contractors,
3.7 REFERENCES
71
and QA/QC personnel to do the work.

Performance monitoring is essential to establish whether the barrier
function of the wall is being met. Performance monitoring of cutoff walls can
include tracer tests, global or local pumping tests, continuity verification,
contaminant concentration measurements, and periodic sampling and testing
of the wall itself. There are advantages, disadvantages, and limitations to
each approach, and further research is recommended to develop improved
methods. A particularly important need in the current regulatory environment
is for improved methods to determine the in situ hydraulic conductivity of
the cutoff wall material. In the case of cementitious backfill, for example,
experience has shown that it is virtually impossible to obtain undisturbed
samples suitable for evaluating in situ hydraulic conductivity because of the
disruptions and stress changes associated with the sampling process.
Unfortunately, little field information is available that can be used as a
basis for definitive conclusions about the long-term durability of soil- and
cement-based vertical barrier walls in environmental applications. However,
indirect evidence for long-term durability from observations of soils and
cement-treated soils in other field applications and from laboratory chemical
compatibility testing suggest that these materials are well-suited for cutoff
wall applications. The lack of field data from environmental applications
highlights the need for long-term performance monitoring studies.
The major technology transfer need is for documentation of successful
cutoff wall case histories. There is also need to communicate the factors that
are essential for high quality cutoff wall design and construction to owners,
designers, regulators, and the public.
3.7 REFERENCES
Acar, Y. B., Hamidon, A., Field, S. D., and Scott, L. (1985). The Effect of Organic
Fluids on Hydraulic Conductivity of Compacted Kaolinite, Hydraulic
Barriers in Soil and Rock, ASTM, STP 874, 171-187.
Betsill, J. D. (1995). Verification and Monitoring of Subsurface Barrier
Emplacement and Performance: Current Research Needs and Directions,
Presented at the International Containment Technology Workshop,
Baltimore, Maryland, August 29-31.
Betsill, J. D., and Guebel, R. D. (1995). VAMOS, The Verification and
Monitoring Options Study: Current Research Options for In-Situ
Monitoring and Verification of Contaminant Remediation and
Containment within the Vados Zone, 6621/MS-0719. Albuquerque, NM:
Sandia National Laboratories, SAND95-1689.
Blight, G. E. (1973). Stresses in Narrow Cores and Core Trenches of Dams,
72
Eleventh International Congress on Large Dams, Madrid, Vol. 3, 63-80.

Borns, D. J. (1995). Geophysical Characterization of Subsurface Barriers,
6116/MS-0750., Albuquerque, NM, Sandia National Laboratories.
Boyes, R. G. H. (1975). Structural and Cutoff Diaphragm Walls, Applied
Science Publishers, Ltd., London.
Carter, E. E. (1995). In-Situ Slurry Wall Formation with the SoilSaw Barrier
System, Stabilization and Solidification of Hazardous, Radioactive, and
Mixed Wastes, ASTM STP 1240, T. Michael Gilliam and Carlton C. Wiles,
Eds., American Society for Testing and Materials, Philadelphia.
Daniel, D. E. (1987). Earthen Liners for Land Disposal Facilities, Geotechnical
Practice for Waste Disposal 87, ASCE, Geotechnical Special Publication No.
13, 21-39.
Day, S. R. (1995). Costs, Materials, and Material Mixing, Contribution to the
International Containment Technology Workshop, Baltimore, Aug. 29-31.
Eiben, T. and Jefferis, S. A. (1994). Implementation of Interactive Numerical
Modeling and Hydraulic Testing in the Design and Quality Control of
Low Permeability Barrier Systems, Meeting on Measurement of Low
Permeability in Soil and Rocks, University of Surrey, Guilford, England.
Esnault, A. (1995). Construction QA/QC, Contribution to the International
Containment Technology Workshop, Baltimore, Aug. 29-31.
Evans, J. C. (1991). Geotechnics of Hazardous Waste Control Systems,
Chapter 20, Foundation Engineering Handbook, 2nd ed., H. Y. Fang, ed.,
Van Nostrand-Reinhold Company, New York.
Evans, J. C. (1993). Vertical Cutoff Walls, Geotechnical Practice for Waste
Disposal, D. E. Daniel, ed., Chapman & Hall, London.
Evans, J. C. (1995). Guide specifications are available from Professor Evans at
the Department of Civil Engineering, Bucknell University, Lewisburg,
Pennsylvania. Phone (717)524-1371.
Evans, J. C., and Cooley, B. H. (1993). Slurry Wall Performance, Progress in
Geotechnical Engineering, Hershey, Pennsylvania.
Evans, J. C., Costa, M. J., Cooley, B. (1995). The State-of-Stress in Soil-Bentonite
Slurry Trench Cutoff Walls, Geoenvironment 2000, ASCE Geotechnical
Specialty Publication No. 46, 1173-1191.
3.7 REFERENCES
73
Filz, G. M. (1996). Consolidation Stresses in Soil-Bentonite Backfilled

Trenches, Proceedings of the Second International Conference on Environmental
Geotechnics, Osaka, (in press).
Handy, R. L. (1985). The Arch in Soil Arching, Journal of Geotechnical
Engineering, ASCE, 111(3), 302-318.
Hollenweger, R., and Martinenghi, L. (1994). Design and Performance
Verification of a Soil-Bentonite Slurry Wall for the Hydraulic Isolation of
Contaminated Sites, ASTM Symposium on Dredging, Remediation, and
Contaminated Sediments, Montreal.
Institution of Civil Engineers. (1995). Specification for Slurry Trench CutOff Walls (As Barriers to Pollution Migration), Institution of Civil
Engineers, London.
Inyang, H. I. (1995). Relating Damage Accumulation to Contaminant
Transport, Contribution to the International Containment Technology
Workshop, Baltimore, Aug. 29-31.
Janssen, H. A. (1985). Versuche uber Getreidedruck in Silozellen, As cited
in Handy (1985).
Jefferis, S. A. (1981). Bentonite-Cement Slurries for Cut-Offs, International
Conference on Soil Mechanics and Foundation Engineering, Stockholm, Vol.
1, 435-440.
Jefferis, S. A. (1992). Contaminant-Grout Interaction, Grouting, Soil
Improvement, and Geosynthetics, ASCE Specialty Conference, New Orleans.
Jefferis, S. A. (1995). UK Practice: Materials and Their Properties,
Contribution to the International Containment Technology Workshop,
Baltimore, Aug. 29-31.
LaGrega, M. D., Buckingham, P. L., and Evans, J. C. (1994). Hazardous Waste
Management, McGraw-Hill Book Co., New York.
Manassero, M. (1994). Hydraulic Conductivity Assessment of Slurry Wall
Using Piezocone Test, Journal of Geotechnical Engineering, ASCE
120(10), 1725-1746.
Manassero, M., Fratalocchi, E., Pasqualini, E., Spanna, C., and Verga, F. (1995).
Containment with Vertical Cutoff Walls, Geoenvironment 2000,
Geotechnical Special Publication No. 46, ASCE, 1142-1172.
74
Manassero, M. and Pasqualini, E. (1993). Surveying and Construction in

Urban, Suburban, and Polluted Areas - Construction, General Reports,
The Environment and Geotechnics: From Decontamination to Protection of the
Sub-Soil, Paris, April, 1993.
Martson, A., and Anderson, A. O. (1913). The Theory of Loads on Pipes in
Ditches and Tests of Cement and Clay Drain Tile and Sewer Pipe, Iowa
Engineering Experiment Station Bulletin, No. 31, Ames, Iowa.
Millet, R. A. and Perez, L. Y. (1981). Current USA Practice: Slurry Wall
Specifications, ASCE Journal of the Geotechnical Engineering Division,
107(GT8), 1041-1056.
Millet, R. A., Perez, L. Y., and Davidson, R. R. (1992). USA Practice Slurry
Wall Specifications 10 Years Later, Slurry Walls: Design, Construction,
and Quality Control, ASTM STP 1229, Paul, D. B., Davidson, R. R., and
Cavalli, N. J., eds., Philadelphia.
Mitchell, J. K. and Madsen, F. T. (1987). Chemical Effects on Clay Hydraulic
Conductivity. Geotechnical Practice for Waste Disposal 87, ASCE,
Geotechnical Special Publication No. 13, 87-116.
Rumer, R. R. and Ryan, M. E. (1995). Barrier Containment Technologies for
Environmental Remediation Applications, John Wiley & Sons, New York.
Ryan, C. R. (1987). Vertical Barriers in Soil for Pollution Containment,
Geotechnical Practice for Waste Disposal 87, ASCE Geotechnical Special
Publication No. 13, 182-204.
Stienstra, P., and van Deen, J. K. (1994). Field Data Collection Techniques Unconventional Sounding and Sampling Methods, Engineering Geology
of Quaternary Sediments, Rengers (ed.), Balkema, Rotterdam, 41-56.
Spooner, P., Wetzel, R. S., Spooner, C. E., Furman, C. A., Tokarski, E. F., and
Hunt, G. E. (1984). Slurry Trench Construction for Pollution Migration
Control, Environmental Protection Agency, EPA 540/2-84-001,
Cincinnati, OH.
Tamaro, G. J., and Poletto, R. J. (1992). Slurry Walls - Construction Quality
Control, Slurry Walls: Design, Construction, and Quality Control, ASTM
STP 1229, Paul, D. B., Davidson, R. R., and Cavalli, N. J., eds., Philadelphia,
26-41.
Terzaghi, K. (1945). Stability and Stiffness of Cellular Cofferdams,
Transactions, ASCE, Vol. 110, Paper 2253, 1083-1119.
3.7 REFERENCES
75
U.S. Army Corps of Engineers (1994). Soil-Bentonite Slurry Trench for HTRW
Projects, CEGS-02444, Department of the Army, Washington, D.C.
U.S. Army Corps of Engineers (1995). Checklist for Design of Vertical Barrier
Walls for Hazardous Waste Sites, ETL 1110-1-158, Department of the
Army, Washington, D.C.
U.S. Environmental Protection Agency (1995). Permit Writers Guidance
Manual for Monitoring Unsaturated Regions of the Vadose Zone at RCRA
Subtitle C Facilities, (Proposed RCRA technical guidance currently under
review within the U.S. EPA).
U.S. Environmental Protection Agency (1995b). Superfund at Work: Western
Processing Company, Inc., Kent, Washington. EPA 520-F-95-003.
Yang, D. S. (1994). The Applications of Soil Mix Walls in the United States,
Geotechnical News, December, 1994, 44-47.
77
SECTION 4
VERTICAL BARRIERS: SHEET PILES
prepared by
Donald R. McMahon, McMahon & Mann Consulting Engineers

Rene Fuentes, U.S. Environmental Protection Agency

Edward W. Gleason, Encor Coatings, Inc.
Brian M. LaRue, Spencer, White & Prentis
Pedro C. Repetto, Woodward-Clyde Consultants
David J. A. Smyth, University of Waterloo, Canada
Andrew Street, RUST Environmental, United Kingdom
4.1 BACKGROUND
Steel sheet pile walls have had limited application for waste containment
due to concerns over leakage through the sheet pile interlocks (Rumer and
Ryan, 1995). The amount of leakage is greater than typically acceptable for
vertical barriers used for environmental containment applications. The
principal advantage of sheet piling is its strength, which is useful for
retaining earth and water. This high strength provides good resistance to
hydrofracturing under high head conditions compared to other barrier
materials; however, environmental containment barriers rarely are used in
situations where hydrofracturing is a concern. Although the use of sheet
piling in environmental applications is likely to be limited until the issue of
interlock leakage can be solved, there may be circumstances in which
acceptable subsurface containment of contaminants can be achieved using
conventional sheet piles.
Interlock leakage was the first issue considered by the workshop panel
convened to address the use of sheet piles for environmental containment
applications. Attempts have been made to seal the interlocks on sheet pile
barriers, and this has introduced the issue of chemical compatibility of the
sealing material with the below ground environment.
The issue of corrosion in a potentially chemically aggressive environment
has also been expressed as a concern with the use of steel sheet piling for
long-term environmental containment applications.
Geologic conditions can limit the use of sheet piling. They cannot be
driven reliably in soil deposits having cobbles and boulders. Driven sheet
piles can carry contaminants from an upper water bearing zone into and
through an aquitard during driving. This may also create a contaminant
78
pathway along the soil/sheet pile interface.

As is the case with most vertical barrier systems, the ability to monitor
the integrity of the wall, its depth of embedment into a confining layer or its
seal with an underlying bedrock formation are of concern.
Finally, steel sheet piles are generally an expensive alternative compared
to other vertical containment barrier types. It is primarily this feature that
limits the use of steel sheet piling for containment applications in the United
Kingdom.

4.2.1 General
Sheet pile walls have been used for civil engineering applications for many
years and are used frequently in construction. They are a proven technology.
However, there are some limitations associated with their use for
environmental applications that owners should consider, such as a lack of
regulatory acceptance, inadequate QA/QC procedures, high cost (this should
include installation, material and disposal costs), and proprietary issues.
Nonetheless, sheet pile walls have many potential uses for environmental
applications. They have been used to:
1) reduce leakage from reservoirs,
2) limit settlement of cohesive soils due to ground water seepage during
excavation,
3) reduce construction dewatering efforts,
4) contain contaminated ground water,
5) control ground water/surface water interactions,
6) isolate subsurface zones for: performing pilot tests, applying remedial
technology, preventing dilution of reagents and controlling ground
water flow direction,
7) construct reactive walls (i.e., install the sheets, excavate, dewater, backfill
with the reactive barrier material and remove the sheets), and
8) construct temporary enclosures during excavation of contaminated
soils.
4.2.2 Interlock Leakage
In civil engineering applications, reducing leakage at sheet pile interlocks
first became a concern in cellular cofferdam construction and earth retaining
walls, before it became an issue in waste containment applications. The
volume of leakage through the interlocks can be observed from within
excavated cofferdams, but can not be observed directly in sheet piles that are
buried on both sides, such as those used for containment barriers.
79
The volume of water passing through a cofferdam wall should not be

taken as a measure of the volume expected through a sheet pile wall used for
a containment application (Tulett, 1995). Retaining sheet pile walls flex from
the water or lateral earth pressure applied to them. This flexure tightens the
interlocks, making them more water resistant than they might be without
these stresses applied, which is the case in cutoff wall containment
applications.
One technique that has been used when water leakage through the
interlocks of an offshore cellular cofferdam is excessive, is to drop flyash or
bottom ash into the water outside the cofferdam at the leaking interlock
location. Water flowing to the interlock carries the ash to the leaking location
and the ash plugs the leak, Williams and Waite (1993).
Another means for sealing the interlocks on earth retaining sheet pile
walls involves drilling holes outside the leaking interlock location and
pumping grout into the hole. This approach has limitations in that the flow
of grout can not be monitored or controlled, making verification of the integrity
of the external grout seal difficult.
A further method for addressing the interlock leakage problem has been
to place hot rolled sheet piles horizontally on the ground before driving and
to fill the female interlock with an asphalt/grease mixture. The high viscosity
of this sealant limits the volume of soil intruding into the interlock when the
pile is driven; however, the sealant deforms when the male interlock is driven
into it, forming a tight seal. Again, the integrity of the interlock seal can not
be inspected in situ.
The inability to check interlock seals in environmental sheet pile walls
has led to the development of sealable interlocks on sheet piles. One of the
first type of sealable interlocks was formed by welding a steel angle section
to the sheet pile near the female interlock (see Section A in Figure 4-1) prior to
being driven. After driving both sheets, the void created by the angle section
can be flushed clean and filled with a sealant.
2x21/2 x 3/8"
Weld full
length of pile
Sealable cavities
2x2x 3/8"
Figure 4-1 Steel angle section welded to sheet pile interlock
80
Researchers at the University of Waterloo recognized the uncertainty

associated with interlock grouting and developed a sealable joint process
(Waterloo BarrierTM). One of the first experiments at the University of
Waterloo involved welding angle sections to the female edges of hot rolled
sheet pile interlocks, driving the sheets, washing the cavity left by the angle
section, and grouting the cavity with a sealant. A foot plate was welded to
the base of the angle section before the pile was driven to prevent intrusion of
coarse grained soil particles inside the void formed by the angle section and
the sheet pile. The male interlock was driven into the female interlock, and
the angle section served to form a cavity around the interlock. The cavity left
by the angle section could be filled with a sealant. This created some confidence
that the grout would fill the interlock, thus improving the effectiveness of
the sheet pile wall as a containment barrier. A second angle section (Section
B, Figure 4-1) could be used to create a second sealable cavity, if necessary.
Sheet Piling
w
3/8"
3/8"
3/8"
a.
PZ27 and PZ22
24.0 in.
(610 mm)
Sealable
cavity
Sealable
cavity
10.67 in.
(271 mm)
b.
WEZ95
Figure 4-2 Hot-rolled and cold-rolled pile sections. (a) Bethlehem Steel
hot-rolled sheet pile, (b) Canadian Metal Rolling Mills coldrolled sheet pile
81
This approach led to the development of the Waterloo BarrierTM. This

system involves driving cold rolled steel sheet piling along the barrier
alignment with standard pile driving equipment. The cold rolled sheets have
modified loose fitting interlocks which differ from hot rolled sheets, as
illustrated in Figure 4-2. This feature produces a small cavity in the interlocks
when the piles are driven. Foot plates are attached to the tip of the female
interlock to limit the intrusion of coarse grained soil particles into the interlock
during driving (see Figure 4-3). The sheet pile is driven to the design depth.
Next, the adjacent sheet (with male interlock) is driven into the female
interlock, creating a continuous cavity for the length of the sheet piles.
Sealed
Cavity
Sealable cavity allows

use of sealant injection
hose up to 3/4" O.D.
Foot plate prevents

bulk of soil from
entering sealable cavity
Figure 4-3 Bottom plug for a cold-rolled sheet pile interlock

The cavity is cleaned of finer soil particles by lowering a tube discharging
high pressure water into the cavity. The turbidity of the wash water is observed
and washing continues until the return water is clear when the pipe has been
lowered to the base of the sheet pile. This indicates that the cavity had been
cleaned. This interlock cavity can be inspected for defects and plumbness.
Downhole cameras have been lowered into the interlock to observe the
condition of the cleaned interlock. This can increase confidence concerning
the integrity of the barrier installation. A sealant is then placed in the cavity
using the tremie method to seal the interlock, see Figure 4-4. Figure 4-4
shows that a second cavity can be created by attaching an angle section to the
female interlock.
The sealant type is selected to be compatible with the surrounding
82
environment. The University of Waterloo has experimented with several

sealant types, including clay based sealants, cement based sealants, Epoxy,
Urethane and others. The criteria used for selecting the sealant include its
permeability, pumpability, ability to withstand anticipated differences in
hydraulic head across the barrier, ability to be removed, stability, chemical
compatibility with contaminants and resistance to variations in temperature.
Sealant
Internal cavity sealable joint sheet pile
Sealant
Sealant
Dual cavity sealable joint sheet piling
Figure 4-4 Water Barrier sealable interlock

Privett, et. al. (1994) describe a process to limit interlock leakage by joining
two sheet piles together in the factory and driving the sections in pairs. This
technique reduces the number of interlocks requiring field treatment (grouting,
sealing, etc.) by a factor of two.
4.2.3 Corrosion
British Steel reports the results of studies on corrosion and protection of steel
piling. Corrosion of steel sheet piling in undisturbed soil is considered
negligible, regardless of the soil type. This is attributed to low oxygen levels.
Piles extracted from aggressive environments (e.g., from a lagoon with a pH
of 2.8) show insignificant corrosion. However, a large amount of corrosion is
caused by atmospheric exposure, which affects above-ground retaining walls,
and by exposure to sea water in marine applications.
Much of the experience with corrosion of steel piling has been gained in
marine applications. Marine applications offer several different exposure
conditions, each having different corrosion rates (see Figure 4-5). Very little
corrosion occurs below the sea bed, where both sides of the sheet pile are
exposed to soil. In the continuous sea immersion zone, i.e., the portion of the
pile that is always below the water surface but above the sea bed, corrosion is
83
normally retarded due to a blanket of protective marine growth and a lack of

oxygen.
The portion of the pile between high and low tides, the so-called tidal
zone, is also protected by biological growth, albeit a different type of growth
from that occurring in the immersion zone. This biological growth acts to
limit pile exposure to oxygen, thus reducing the rate of corrosion.
Highest corrosion rates occur where oxygen is most abundant, i.e., at the
low water zone and the splash zone. British Steel recommends design
corrosion rates for the various conditions shown in Figure 4-5. The
recommended corrosion rates are shown in Table 4-1. British Steel concludes
that sheetpile corrosion in undisturbed soils is very slight and recommends a
maximum corrosion rate of 0.015 mm/side per year.
Side 1
Atmospheric
(Rural, urban
or industrial)
Side 2
Atmospheric
Side 1
Tidal Zone
Atmospheric
Soil
Side 2
Splash Zone
Soil
Low Water Zone

Immersion Zone
Soil
Soil
a. Atmospheric/Soil
b. Marine/Soil
Side 1
Side 2
Splash Zone
Splash Zone
Tidal Zone
Tidal Zone
Low Water Zone
Low Water Zone
Immersion Zone
Immersion Zone
c. Marine/Marine
Figure 4-5 Different corrosion environments for sheet piles (British Steel,
1995)
84
TABLE 4-1 Recommended Corrosion Rates for Steel Surfaces

(from British Steel)
Pile Face
Environment
Fig. 4-5a
Fig. 4-5a
Fig. 4-5c
Fig. 4-5b
Fig. 4-5c
Fig. 4-5b
Fig. 4-5a
Tidal or
Immersion
Soil
Side 1
Atmospheric Atmospheric
Splash or
Splash or
Tidal or
Low Water Low Water Immersion
Side 2
Atmospheric
Soil
Splash or
Low Water
Soil
Tidal or
Immersion
Soil
Soil
Corrosion*
Rates
(mm/year)
0.07
(mean)
0.05
(mean)
0.15
(mean)
0.09
(mean)
0.07
(mean)
0.05
(mean)
0.03
(mean)
*Represents total loss of section (Side 1 + Side 2)
Notes to Table 4-1

1. The corrosion rates quoted are based on investigations carried out by British Steel and others
on steel exposed in temperate climates. For most environments, mean values are quoted since
they are considered to be most relevant to the design and performance of most sheet piling
structures. However, in some circumstances, the designer may wish to take account of higher
values. It is suggested that in these circumstances, a reasonable practical upper corrosion rate
limit would be that corresponding to a 95% probability value. The mean 95% probability values
for various combinations of UK environments are given as follows:
Pile Face
Environment
Fig. 4-5c
Fig. 4-5b
Fig. 4-5c
Fig. 4-5b
Side 1
Splash or
Low Water
Splash or
Low Water
Tidal or Immersion
Tidal or Immersion
Side 2
Splash or
Low Water
Soil
Tidal or Immersion
Soil
Corrosion Rate (mm/year)

Mean 95% Probability Value
Total Loss of Section
0.33
0.18
0.19
0.11
2. For combinations of environments where low water corrosion is involved, in a small number
of locations, higher rates than those quoted have been observed at or just below the low water
level mark.
3. A maximum value is quoted for soil corrosion and this applies to natural undisturbed soil or
well compacted and weathered fill ground where corrosion rates are very low. Recent fill ground
or waste tips will require special consideration.
4. Corrosion losses due to fresh water immersion are generally lower than for seawater, however,
fresh waters are very variable and general advice can be given to quantify the increase in life.
85
British Steel lists three methods for extending the design life of steel placed
below ground: 1) using a heavier section, 2) using high yield steel, and 3)
using organic coatings. Steel piles are normally coated under controlled
environment conditions (such as in a fabricating shop), and the applied coating
should be damage resistant for transportation and handling. A simple coaltar pitch mixture has been used for some time, but the coating is soft, thin (up
to about 50 microns) and easily damaged. Synthetic resins have been added
to the coal-tar to obtain thicker (100 to 250 microns) and harder coatings.
British Steel has produced three other coatings: tar vinyl (PC1) which is an
aromatic pitch modified with vinyl resins; high-build isocyanate cured epoxy
pitch (PC2) which is a coal tar pitch modified with epoxy resins that make it
harder than PC1; and colored, vinyl ether primer/finish (PC3).
Coal Tar Epoxy coating has been used in the United States for many years
and has been demonstrated to be effective for resisting corrosion in difficult
environments, such as in marine applications. Steel sheet piles treated with
this process and installed in marine environments 35 years ago have been
exhumed recently and have shown no visible defects. This favorable
performance attests to both the survivability of the coating during the driving
process and its long-term resistance in the marine environment.
Uniloc Configurations
Single Lock Connection
Box Connection
Double Lock Connection
S-Loc
B-Loc
D-Loc
Sheet Piling Configurations
Figure 4-6 Uniloc interlock configurations

The U.S. Army Corps of Engineers has a specification C-200 for Coal Tar
Epoxy. Piles coated with this material have shown resistance to chemical
86
solutions, hydrocarbons, acids, alkalis, and alcohols. Newer products have

been introduced, but none have the proven performance of Coal Tar Epoxy.
Cathodic protection can be provided to extend the life of underground
steel structures, but this is a complicated process that requires on-going
maintenance, and it may be operationally expensive.
The concern for steel sheet pile corrosion in a chemically aggressive
environment has lead to the development of the Uniloc high density
polyethylene sheet pile by Spencer White & Prentis, DuPont and Polymer
Construction Products, Ltd. The material used is DuPonts Zemid PG which
is highly resistant to chemical degradation. The Uniloc system consists of
trapezoidal-shaped sheet pile sections having a 3/8 inch wall thickness. The
interlocks differ from those found on steel sheet piles (see Figure 4-6) in that
the flow path through them is longer, thus producing a lower leakage rate for
the same hydraulic head difference. The technique for installing Uniloc sheets
is under development. A hydrostatic pile driver may be used. This would
involve a hydraulic piston alone or a piston with a vibratory hammer (in
conjunction with water jet techniques).
In addition to the Uniloc system, there are other products being advertised,
such as the GeoGuardTM vinyl sheet piling, which is promoted as a corrosion
resistant sheet piling.
4.2.4 Sealant Compatibility
For sheet pile barriers with sealants, the sealant used to treat the interlocks
will depend on site-specific circumstances, and it must be compatible with
the anticipated subsurface environmental conditions. Candidate sealants
include clay based (bentonite and attapulgite) grouts, cement-based grouts,
epoxy polymers, urethane polymers, and others. If field circumstances dictate,
different sealants can be used at different elevations within the interlock. The
sealant selected must also be compatible with the hydraulic head applied,
must be pumpable, have a satisfactory permeability once placed, and must
have thermal expansion properties compatible with the interlock. The sealant
material selected may influence the cost of a project; however, because the
volume of the sealable cavity is small relative to the size of the wall, it may be
feasible to select effective sealants for sheet pile barrier wall interlocks that
would be impractical or too expensive for other types of barrier walls, e.g.,
slurry walls.
4.2.5 Construction Monitoring
A detailed subsurface characterization of the site is necessary for monitoring
the key of a sheet pile wall into an aquitard. This includes making test borings
along the barrier alignment at intervals necessary to fully define changes in
subsurface conditions.
Another method to verify that the sheet piles are keyed into the
87
appropriate aquitard calls for monitoring the effort required to advance the
piles with depth. Advancement rates differ in different soil types. Monitoring
the driving rate (blow counts) of impact-type pile hammers provides
quantifiable data; whereas, monitoring vibratory hammer performance is more
qualitative.
On occasion, downhole cameras have been dropped into the interlock
(after cleaning) to observe its condition. In situations where the integrity of
the sheet may be jeopardized during placement, observation tubes can be
welded to the sides of selected sheet piles (between interlocks) prior to
placement, thus providing a means for later inspection of the sheet.
4.2.6 Costs
Costs for sheet pile barriers depend on many factors, e.g., the distance to
mobilize equipment, size of the project, type of sealant used, the use of a
coating, special installation conditions, weight of the section, etc. Approximate
costs range from about $15 U.S. to $40 U.S. per square foot for the sealable
interlock joint sheet pile wall (Smyth, et. al., 1995).

The bulk hydraulic conductivity of sheet pile barriers is an appropriate
property to use in comparing sheet pile barriers to other barrier types, because
this property considers the discontinuities in the sheet pile wall. Field studies
of sheet pile barriers report this property.
Another appropriate property is the barrier permittivity (the ratio of
hydraulic conductivity to the barrier thickness), because this considers not
only the bulk hydraulic conductivity, but the actual potential for flow through
the barrier, i.e., the flux. For example, a soil-bentonite wall, two feet thick,
with a hydraulic conductivity of 10-7 cm/s has a permittivity of 1.6 x 10-9 sec-1.
The bulk hydraulic conductivity of a sheet pile wall might be 10-9 cm/s (two
orders of magnitude less than the soil-bentonite wall), but the permittivity is
1 x 10-9 sec-1, which is comparable to the soil-bentonite wall.
The University of Waterloo has conducted field tests on more than 20
research test cells and on commercial installations. Hydraulic tests have been
undertaken on concentric test cells that extended into an underlying aquitard
to measure the bulk hydraulic conductivity of the sheet pile wall. In these
cells, a constant head of water was maintained in the moat between the two
cell walls and a higher head was applied within the interior cell (see Figure 4-7).
The difference between the two heads was monitored with time.
One test case was discussed by Smyth, et. al. (1995) in which a bentonite
slurry was used as the interlock sealant. Accounting for evaporation and
assuming that no seepage occurred through the aquitard, the bulk hydraulic
conductivity of the cell double-wall system was calculated to be 6x10-9 cm/
88
sec. An organic polymer sealant when similarly tested resulted in a bulk

hydraulic conductivity of less than 1x10-9 cm/sec. If these walls were 0.375
inches thick, their permittivity would be 6.3 x 10-9 sec-1 and 1 x 10-9 sec-1,
respectively.
A
Plan View
Section A-B
Figure 4-7 Plan and cross section of Waterloo barrier test cell for hydraulic
conductivity (Smyth, et al., 1995)
A field test program was executed for DOW Chemical Company in
Freeport Texas, ENSR (1993). In this field test, Arbed AZ18 piles were selected
to withstand the difficult driving conditions. These sheet piles did not have
sealed interlocks, but the interlock configuration differs from the Bethlehem
Steel interlock shown in Figure 4-2. The Arbed interlock is illustrated in Figure
4-8.
To conduct the field test, a pumping well was installed (with four
observation wells) and pumped to measure the hydraulic characteristics of a
silty sand deposit that overlayed a clay deposit. A sheet pile enclosure, 50 feet
x 50 feet, was constructed around the pumping well and additional observation
wells were installed inside and outside the test facility. A second pump test
was undertaken to measure the effects of the sheet pile enclosure. The data
indicated that the sheet pile barrier had an estimated permeability of 7x10-7
cm/sec. Since the piles were driven into a bottom clay deposit, the observed
leakage was assumed to have flowed through the sheet piles and not under
the piles.
A separate study was undertaken to address the concern for the potential
of flow along the sheet pile/aquitard interface, Hayman, et. al. (1993). In this
case, a laboratory study was initiated to explore the potential for flow along
the interface of piles penetrating an unconfined aquifer overlying an aquitard
and a confined aquifer. This study was carried out to simulate conditions in
which piles are driven through interbedded sand and clay strata and to
determine the potential for creating a pathway for contaminant migration.
89
AZ 18
0.375"
380 mm
14.96"
9.5 mm
mm "
9.5 375
0.
X
55.4
~348 mm
~13.70"
1260 mm
49.61"
Figure 4-8 Arbed sheet pile interlock

Test vessels were constructed with a sand layer over a clay layer which
was over another sand layer. The researchers saturated the lower sand layer
with deionized water and then consolidated the clay layer to represent a
selected overburden pressure. The upper sand layer was saturated with a
dense non-aqueous phase liquid (DNAPL) and a pile was driven through the
soil layers. The deionized water in the lower layer was purged and replaced
with another volume of deionized water. The first volume was sampled and
analyzed for the DNAPL. This purging and sampling process was repeated
with time after completion of the pile driving.
It was found that DNAPL was carried down with the pile during driving,
since DNAPL was detected in the initial samplings of the deionized water in
the lower layer. However, the DNAPL concentration decreased in subsequent
samples, indicating that the DNAPL was being purged from the lower sand
layer. It was concluded that DNAPL was introduced into the lower layer
only during driving and that there was no detectable leakage along the pile/
clay interface after the sheet pile had been placed.

4.4.1 General
As with any technology, there are advantages and limitations associated with
the use of steel sheet piles for environmental containment applications.
4.4.2 Advantages
1. No excavation required - Sheet piles are driven into the ground and
(usually) excavation of potentially contaminated soils is not required.
90
2. Minimum waste disposal - Since no excavation is required for construction

of the sheet pile barrier, disposal of excavated contaminated soil is not required.
The cleansing of interlocks prior to sealing produces potentially contaminated
water which must disposed of properly.
3. Reduced health and safety concerns - Since there is no excavation of
potentially contaminated soils, the risk of exposure to contaminated soils is
lessened, with a correspondingly reduced concern for worker health and
safety.
4. Joints can be resealed (depending on the sealant used) - If there is concern
for the integrity of a sealed interlock, some sealing materials can be removed
and the process repeated to provide a better seal.
5. Sheet piles can be removed later if required - Some environmental
remediation projects require temporary containment. Sheet piles can be
removed using conventional techniques to restore the site to the preconstruction condition.
6. Standard construction equipment can be used - Sheet piles have been
driven for many years. This technology uses standard equipment to install
the barrier.
7. Rapid installation - Since sheet piling involves standard procedures and
equipment, installation can be rapid.
8. Topography and depth to water have little impact - These site-specific
features affect the construction of many barrier types. For example, the ground
surface must be flat to construct a slurry wall and the ground surface usually
must be above the ground water level to assure trench stability. This may
involve raising ground surface elevations along the barrier alignment to
provide these conditions. These conditions are not required for construction
of sheet pile barriers.
9. Diffusive transport through the barrier is greatly reduced - Properly sealed
steel sheet piles are barriers to diffusion, which may be an important
contaminant transport mechanism. Furthermore, they do not sorb
contaminants.
10. Irregular enclosure shapes are possible - This offers an advantage when
constructing a containment system near buildings, utilities, and at small
properties.
11. Costs may be reduced - Many other barrier types are installed by specialty
contractors. Construction at sites located some distance from these contractors
4.5 NEEDS
91
locations will incur expensive mobilization costs. Since sheet piling installation
is a common technology, many local contractors can reliably construct this
barrier type without the expensive mobilization costs. Additionally, smaller
walls can be cost-competitive with other barrier wall types. Many wall types
require construction of temporary facilities that increase the unit prices on
small projects (fixed costs divided by a small quantity). An example is the
construction of slurry ponds for a slurry wall. These temporary facilities are
not required for sheet piling and can make them cost competitive, especially
on smaller projects.
4.4.3 Limitations
1. Noise and vibration - Sheet piles are driven with either vibratory or impact
hammers that have associated side effects. Impact hammers create noise and
vibratory hammers impart vibrations to the ground that can densify loose,
saturated sand deposits resulting in settlement of the ground surface. The
vibrations can also cause damage to adjacent structures, depending on their
location, construction type, and proximity to the vibration source.
2. Not suitable for stiff clay or soils containing cobbles and boulders Boulders and cobble deposits resist penetration by sheet piles regardless of
the hammer type used. Predrilling may be necessary for advancing piles
through these soil types. Stiff clay deposits can also introduce difficulty for
driving piles. The potential for the piles fissuring stiff clay deposits should
be evaluated. Vibratory pile hammers are not usually suitable for driving
through clay soil deposits.
3. Keying into rock is not possible - Most rock deposits are too hard for
sheet piles to penetrate, making keying into rock impractical.
4. Depth penetration limited - There are limitations regarding the depth
that sheet piles can be driven. The maximum depth depends on the soil
conditions, hammer type and the sheet pile section. It commonly is about 30
to 45 meters, Smyth, et. al. (1995). Predrilling along the alignment can be
undertaken to extend this, but this will add to the project cost.
5. Some technologies are proprietary - Proprietary information sometimes
introduces difficulties in competitive bidding situations.
4.5 NEEDS
There is a need to collect information on the long-term performance of sheet
pile walls used as barriers in containment applications. The potential for
corrosion of the steel requires further evaluation of the damage to coatings
92
caused during installation and from exposure to different chemicals

encountered at waste sites.
The long-term stability of interlock sealants needs to be studied, both
with regard to exposure to different chemicals and to the effect of fluctuating
hydraulic gradients.
Published case histories and demonstration results with accurate
performance measurements are needed before this technology will gain
acceptance by regulatory agencies. Several projects have been implemented
and others are planned, so a better prognosis may be available in the near
future. Quality control procedures need to be developed to measure the bulk
hydraulic conductivity and to establish the locations of defects in the field.
However, several of the quality control procedures currently available for
sealable joint sheet piling are as rigorous or better than some of the quality
control measures applied to other barrier wall construction techniques.
4.6 SUMMARY/RECOMMENDATIONS
Field demonstrations using both conventional and sealable joint sheet piling
have shown that sheet piling is a viable barrier option in containment
applications. Although not appropriate for all barrier applications, sheet piling
is worthy of consideration in the project planning stage because it may be the
appropriate technology.
Leakage through the interlocks of conventional sheet piling is piling
dependent, as the joint tightness varies among different pile types. Concerns
about leakage decrease if the containment wall is used with other remediation
techniques, such as pump and treat ground water systems.
Some project experience is available, and the results are promising. Longer
term trials at contaminated sites are required. These should involve both
hydraulic monitoring and observations of contaminant migration in the
evaluation of containment performance.
Monitoring of sites on a large scale is necessary to develop data on the
success of this technique. This will help to make identification of gross leakage
possible, but detection of small leaks will continue to be difficult, as is the
case with all barrier wall types.
4.7 REFERENCES
British Steel Co. (1995). The corrosion and protection of steel piling in
temperate climates, Document Reference P115, published in February.
ENSR Consulting and Engineering, (1993). Sheet pile wall leakage test,
report prepared for DOW Chemical Company.
4.7 REFERENCES
93
Hayman, J. W., Adams, R. B., and Adams, R. G. (1993). Foundation piling as

a potential conduit for DNAPL migration, Presented at the 86 Annual
Meeting & Exhibition of the Air & Waste Management Association,
Denver, Col.
Koerner, R. A., (1994). Designing with geosynthetics, Third edition, PrenticeHall, Inc., Englewood Cliffs, N.J.
Privett, K. D., Matthews, S. C. and Hodges, R. A. (1994). Remedial treatment
of contaminated land using in-ground barriers, liners and cover systems,
report prepared for CIRIA by Edmund Nutall Ltd. in association with
Rust Environmental.
Rumer, R. R. And Ryan, M. E. (1995). Barrier Containment Technologies for
Environmental Remediation Applications, John Wiley & Sons, Inc., New
York, N.Y.
Smyth, David J. A., Cherry, John A. And Jowett, Robin J. (1995). Sealable
joint steel sheet piling (Waterloo BarrierTM) for groundwater pollution
control. Presented to International Containment Technology Workshop,
Baltimore, Md.
Tulett, R. C. (1995) personal communication.
Williams, B. P. And Waite, D. (1993). The design and construction of sheetpile cofferdams, Construction Industry Research and Information
Association and Thomas Telford Publications, London.
95
SECTION 5
VERTICAL BARRIERS: GEOMEMBRANES
prepared by
Robert M. Koerner, Drexel University

and John L. Guglielmetti, DuPont Environmental Remediation Services
Robert C. Bachus, GeoSyntec Consultants

Patrick T. Burnette, Consultant
Nico Cortlever, Delft Geotechnics, The Netherlands
James M. Cramer, Nilex Corp.
Robert E. Landreth, U.S. Environmental Protection Agency
Stephen M. Mangion, U.S. Environmental Protection Agency
Mark Phifer, Westinghouse Savannah River Co.
William W. Walling, SLT North America Inc.
5.1 BACKGROUND
Background information on geomembranes for use as vertical barriers in waste
containment facilities is summarized in the following paragraphs.
5.1.1 Overview and General Schematics
The essential reason for using a geomembrane as a vertical barrier is to assure
complete continuity by means of an extremely low permeability material.
Such continuity and relative impermeability might be compromised in soilbentonite (SB), soil-cement (SC), cement-bentonite (CB), soil-cement-bentonite
(SCB) walls and other related vertical barriers due to the occurrence of
unobservable factors during construction, such as;
collapse of trench walls during construction or backfilling,

sand deposits during pauses in backfilling,
sand or debris on the bottom of slurry supported trenches leading to
ineffective embedment and
improperly mixed backfill material containing granular soil or large
obstructions, e.g., rock or construction debris.
Furthermore, the low permeability of SB, SC, CB or SCB materials and other
related vertical walltypes might be further compromised due to factors such
as;
96
changes in the quality of backfill material during field mixing,

discontinuities at joints and locations of work stoppage during the
backfilling process,
drying and desiccation above the water level or during water table
fluctuations,
freeze and thaw cycling in the upper portions of the wall, and
chemical incompatibility with the leachate that the site contains.
Geomembranes have been used as vertical barriers either alone or in

conjunction with other relatively impermeable materials for approximately
10 years, with the earliest attempts being made in 1980. Applications of the
technology have included the following;
the geomembrane by itself,

the geomembrane with SB, SC, CB or SCB backfill forming a composite
barrier, and
a double geomembrane with a sand leak detection layer between.
The technology is also available to provide the following;
a double geomembrane with a geonet leak detection layer between,

and
a double geomembrane with geonet between backfilled with SB, SC,
CB or SCBforming a RCRA Subtitle C-type liner system used as a
vertical wall.
Geomembranes used for vertical walls are almost always made from high
density polyethylene (HDPE); however, other polymers can also be used.
The geomembrane sheets can be continuous, but usually finite length panels
interlocked by a number of possible connections are used. The depth of the
final wall can be essentially unlimited, but the design depth will dictate the
method of installation. A number of different installation procedures are
available.
The top of the geomembrane sheets in the vertical wall can be seamed to
the geomembrane in the containment cover making a mechanical seal as
shown in Figure 5-1(a). Such a system can be monitored using conventional
CQC/CQA procedures. Alternatively, the geomembrane in the cover can be
extended horizontally or dropped down vertically making an overlap seal
with respect to the geomembrane in the vertical wall. This is the more common
arrangement, as shown in Figure 5-1(b).
The toe of the geomembrane sheets can be keyed into an aquitard (if one
is available) or can be taken to an adequate depth where under-seepage
becomes negligible per the site-specific design, as depicted in Figures 5-2(a)
and 5-2(b), respectively. The latter case is frequently referred to as a hanging
wall.
5.1 BACKGROUND
97
Cover Geomembrane
Seam
Seam
Wall
Geomembrane
Waste
(a) Geomembrane in vertical wall seamed to geomembrane in cover
Cover Geomembrane
Vertical
Overlap
Horizontal
Overlap
Waste
Wall
Geomembrane
(b) Horizontal or vertical overlap of geomembrane in cover
Figure 5-1 Termination of the top of geomembrane vertical barriers
Waste
Wall
Geomembrane
Possibly
Grouted
Aquitard
(a) Keyed (and grouted) into aquitard
Waste
Adequate
Depth
(b) Deep (hanging) wall beneath waste
Figure 5-2 Termination of toe (bottom) of geomembrane vertical barriers
98
Vertical barrier walls with geomembranes are available in a number of

different forms and have been used to the point where their technical feasibility
is well established. CQC and CQA are regularly practiced during the
installation. The state-of-the-practice, field performance, assessment of the
technology, needs for further implementation, and references that are available
in the open literature are presented in this section.
5.1.2 Types of Geomembranes Used
Essentially all barrier walls constructed to date using geomembranes have
been made from high density polyethylene (HDPE). In this regard, HDPE is
a logical polymer choice for the following reasons;
HDPE is chemically resistant to a wide variety of chemicals, including

organic solvents,
HDPE is readily extruded into sheets and into detailed connection
fittings for joining the sheets together,
the sheets and the connection fittings can be seamed together in the
factory or in the field using readily available welding equipment.
They are then referred to as panels,
variable sheet thicknesses, panel widths, and panel lengths can be
manufactured,
the cost of HDPE is currently very low, traditionally being the least
expensive polymer of all geomembrane polymer types, and
a sufficient number of commercial systems are available, such that a
generic specification can be generated and bid upon accordingly.
HDPE geomembranes consist of a formulation of three components as

shown in Table 5-1. The final geomembrane density is typically 0.941 to 0.943
g/cc, placing it in the high density category of linear polyethylenes; hence,
its identification as HDPE.
TABLE 5-1 Typical HDPE Formulation
Component
Density (g/cc)
Percentage
(by weight)
resin
0.934 to 0.940
97 - 97.5
virgin resin, i.e.,

no post consumer
recycled resin
carbon black
2.5 to 2.8
2.0 - 2.5
for ultraviolet stability

and durability
antioxidants
1.5 to 2.5
1.0 - 0.5
for processing and

long-term durability
Comments
While stiff with respect to physical handling, HDPE sheets or panels can
5.1 BACKGROUND
99
not be directly driven into native soil using conventional pile driving
equipment. Even vibratory pile hammers would deliver excessive stresses to
the sheeting causing buckling, folding or tearing of the geomembrane.
If direct driving is a required feature, a stiffer or (higher modulus) polymer
is required. Polyvinyl chloride (PVC) has been used in this regard. Several
PVCsheet products, replicating steel sheet piling, have appeared on the market
that are capable of direct driving. Their primary use appears to be in the
construction of small walls and bulkheads, but they could conceivably be
used for vertical containment walls. Other rigid polymer alternatives exist in
this category as well.
Due to their widespread use in containment walls, however, HDPE sheets
and panels will be the focus of the remainder of this section.
5.1.3 Intrinsic Impermeance of Geomembranes
HDPE has an extremely low permeability. Note that the permeability of a
geomembrane is not hydraulic conductivity (as used for analysis of SB, SC,
CB or SCB vertical barriers), but is a diffusion-related transport property. The
standard test for its determination is a water vapor diffusion test performed
in accordance with ASTME96. A typical value for 2.5 mm (100 mil) thick
HDPE geomembranes is approximately 0.006 g/m2-day (~ 0.006 gal/acreday). While the calculations are a bit spurious, this value converts to
approximately 1 x 10 -13 cm/sec of equivalent Darcian type hydraulic
conductivity, Koerner (1994). Haxo (1988) has determined comparable vapor
transmission values for a range of solvents, see Table 5-2. Note that the values
TABLE 5-2 Solvent Vapor Transmission Values of HDPE after Haxo (1988)
Property
0.80 mm (32 mil)
2.6 mm (102 mil)
Methyl alcohol
0.16
Acetone
0.56
Cyclohexane
11.7
Xylene
21.6
6.86
Chloroform
54.8
15.8
Solvent vapor transmission, g/m2-day
are dependent on thickness and the type of solvent vapor. They are related
to the solubilities of the permeating liquids, which in these cases were 100%
neat organic solvents. While these vapor-transmission values are lower than
any other available polymer barrier system, they apply only to the
geomembrane sheet and not for the interlocks. The interlocks used to join
the individual panels are of the following three different types;
100
a hydrophilic gasket,
grouted, with a variety of sealants, or
mechanically welded.
Note that values shown in Table 5-2 can refer to interlocks that are
mechanically welded, but not to gasket or grouted type interlocks. The gasket
or grouted interlock types are generally used in vertical barrier wall
construction. Owing to their importance, interlocks will be treated in more
detail later in this section.
5.1.4 Scope of Section 5
The state-of-the-practice as presented by the eight containment technology
workshop panelists is central to the text that follows. Field performance via
case histories will also be presented. Emphasis is on those case studies that
are presented in the literature rather than internal reports or manufacturers
brochures.
The technology of geomembranes as vertical barriers is assessed in light
of costs for the various installation techniques and in comparison to other
types of vertical barrier systems. Emerging trends and technologies are also
noted. Lastly, the research, development, and implementation needs related
to the use of geomembranes as vertical barriers is assessed. As will be seen,
although the technology already exists, greater exposure is needed in the
technical literature so that the advantages and disadvantages can be assessed
for the wide variety of site-specific conditions encountered in practice.

This section reviews the current practices, materials, designs, and details
relative to the use of geomembranes as vertical barriers and concludes with a
set of case histories illustrating the various nuances of the technology.
5.2.1 Installation Methods
Early applications of geomembranes as vertical barriers utilized a slurry
supported excavation with the inserted geomembrane displacing the slurry
in a progressive manner. One installation technique, called Envirowall,
employed a folded U-shaped geomembrane that covered both the sides and
bottom of the trench. The space within the U-shaped geomembrane was
backfilled with sand, as installation proceeded. This procedure resulted in a
double lined vertical wall with a sand leak detection system between the two
geomembranes, see Koerner (1986). Unfortunately, construction of this system
was difficult. With the slurry at a density of approximately 1.2 g/cc and the
HDPE geomembrane at 0.94 g/cc, it was difficult to submerge the folded
101
geomembrane beneath the slurry and to hold it in place until sufficient sand
mass had been placed to counterbalance the buoyancy. Other early efforts
used a single geomembrane against the side of the trench, with the bottom
weighted down with pipes, rails, etc. Most of these early techniques were
rather unwieldy.
TABLE 5-3 Current Installation Methods for Geomembrane Vertical Walls
Method
No.
Method or
Technique
Geomembrane
Configuration
Trench
Support
Typ. Trench
Width
mm (in.)
Typ. Trench
Depth
m (ft.)
Typ.
Backfill
Type
trenching
machine
continuous
none
300-600
(12-24)
1.5-4.5
(5-15)
sand or
native soil
vibrated
insertion plate
panels
none
100-15
(4-6)
1.5-6.0
(5-20)
native soil
slurry
supported
panels
slurry
600-900
(24-36)
no limit,
except for
trench stability
SB, SC, CB,

SCB, sand
or native soil
segmented
trench box
panels or
continuous
none
900-1200
(36-48)
3.0-9.0
(10-30)
sand or
native soil
vibrating
beam
panels
slurry
150-220
(6-9)
no limit
SB, SC, CB,

SCB slurry
Since these early attempts, many clever installation methods have been
developed. Five newer methods which have been developed and used on a
regular basis are described in Table 5-3. Each will be described separately.
ll
Ro
e
ran
em
om
Ge
ll
Ro th
d
Wi
Self supporting trench

(depth=site specific)
Figure 5-3 Continuous geomembrane in self-supporting trench excavated

by a trenching machine
Figure 5-3 presents the essence of the trenching machine installation
method. A large bucket trencher or disc cutter is used to excavate an
102
unsupported trench in the ground. The geomembrane roll is lowered vertically

into the trench and progressively unrolled. The geomembrane can also be
unrolled at the ground surface, but the sheet must be distorted into an Sconfiguration in order to reach its final position. A major advantage of this
method is that no seams are necessary (until the end of the roll is reached).
Native soil is commonly used for backfill, since a visual inspection can be
made of the completed installation, at least to the level of water in the trench.
Alternatively, a sand backfill with a pipe leak detection system can be used.
Amajor disadvantage is that the unsupported trench depth is limited by sitespecific circumstances.
Vibratory
Hammer
Insertion
Plate
Soil Backfill
Pins through folded

geomembrane panel at
bottom of insertion plate
Geomembrane
Panels
Interlocks
Figure 5-4 Vibrated insertion plate method with geomembrane panel

attached.
The vibrated insertion plate method utilizes geomembrane panels
approximately 3 m (10 ft.) wide on the side of a steel truss insertion plate,
see Figure 5-4. The geomembrane is fixed at its base with pins protruding
from the bottom of the insertion plate. The geomembrane panel is folded
around the bottom of the insertion plate for penetration by the pins. A
vibratory pile hammer forces the entire assembly (insertion plate plus
geomembrane panel) to the desired depth. The insertion plate is then
withdrawn, leaving the geomembrane panel behind. Interlock connections
are required to join one geomembrane panel to the next. This relatively fast
method requires neither trench support nor special backfill. It can also be
used in running sands, which may be problematic with other methods. A
disadvantage is that only certain soft and/or loose soils can be penetrated to
substantial depths. Also, damage to the geomembrane panel during insertion
may be a concern with certain soil strata.
103
ry
Slur y
Entr
ry
Slur oval
Rem
il,
kfill: d so c.
Bacavate C, et
Exc, CB, S
SB
el
rry
Slu
Lev
Geomembrane
Panels
Interlocks
Figure 5-5 Backhoe excavated, slurry supported method with geomembrane

panels and engineered backfill as secondary operation.
The slurry supported installation method begins by using a conventionally
excavated and slurry supported trench, see Figure 5-5. The geomembrane
panels are then inserted into the slurry to the desired depth utilizing a steel
frame. The panels are held in their final position with weights or toed into
the subsoil stratum and interlocked to form a continuous liner. The backfill
is usually SB, SC, CB or SCB materials. There are other options, including
low permeability native soil. If geomembranes are placed on both sides of
the trench, sand is used as the intermediate backfill, creating a double lined
vertical wall with intermediate leak detection. Depth is limited only by the
stability of the trench, and the deepest vertical walls have been constructed
by this method. The maximum depth constructed to date is 30 m (100 ft.), but
considerably greater depths are being considered.
The segmented trench box method uses a modified steel trench box to
support the side walls of the trench immediately following excavation, see
Figure 5-6. It is moved laterally within the trench which is excavated by a
backhoe or other construction equipment. The geomembrane panels are placed
between the two outer segments of the box. Backfill is usually native soil, but
other materials such as sand with a perforated pipe collection system can
also be used. The technique is depth limited, hence quite site specific.
Although untried, the technique could possibly be used with a continuous
geomembrane roll being fixed to the advancing end of the trench box and
unrolled as the trench box is moved along, thus avoiding the need for
interlocks.
104
t
Spli ch
Tren
x
o
B
kfill
Bac
kfill
Bac
Geomembrane
Panels
Interlocks
Figure 5-6 Trench box method with geomembrane panels and subsequent
backfill
The vibrating beam method utilizes a modified form of the slurrysupported trenching method, see Figure 5-7. The slurry is introduced as the
vibrating beam is inserted. Once the trench is slurry supported and the beam
removed, the geomembrane panels are installed as a secondary operation.
The usual backfill for the relatively narrow trench is a SB, SC, CB or SCB
slurry. The technique is commonly practiced but is somewhat dependent on
site-specific soil conditions.
Vibrating
Beam
l
Soi kfill
Bac
ry
Slur oval
Rem
l
eve
yL
rr
Slu
ite
ton
Be rry n
u
l
o
S ecti
Inj
Geomembrane Panels
Interlocks
Figure 5-7 Vibrating beam/slurry trench support method with geomembrane

panels as secondary operation
105
There may be other methods for vertical geomembrane wall installation,

but those discussed above appear to be the most commonly used at present.
5.2.2 Types of Interlocks
Three types of interlocks are used to connect the geomembrane panels (or
rolls) together. These are:
hydrophilic gasket types,
grouted chemical or tube types, and
continuous seaming by welding.
They generally require a special extruded HDPEshape to be seamed to
the edges of the geomembrane sheets. Quite often they are lock-and-key
mated, requiring the alternating type to be seamed on opposite sides of the
panels. Seaming is done in the factory by wedge welding. Either single or
double wedge welds are made, the latter having the capability for air pressure
testing before the panels are shipped to the field, EPA (1993).
a. Hydrophilic Gasket Interlocks. Figure 5-8 illustrates four different interlock
configurations, each of which requires a hydrophilic gasket for liquid tightness
of the installed panels. The gaskets are either circular or rectangular in cross
section. They are threaded into their groove or slot during connection in the
field. The length of gasket that is fed into the interlock from the ground surface
provides reasonable assurance that the interlock connection has been made
down to the bottom of the previously placed panel.
Figure 5-8 Various hydrophilic gasket type geomembrane panel interlocks

Although the gasket material is usually proprietary, most are made from
rubber (chloroprene or neoprene) formulated with a hydrophilic polymer.
106
Unconfined swelling in water is five to eight times the original volume.

Chemical compatibility of the gasket can be evaluated by standard laboratory
techniques, e.g., the modified EPA9090 test procedure.
b. Grouted Interlocks. Alternatively to a gasket and for certain types of
interlocks, a liquid-tight interlock can be made using various slurries or grouts,
see Figure 5-9. Here the pumped slurry or grout flows down one channel or
tube, and, after reaching the bottom of the interlock returns to the surface in
an adjacent channel or tube. The grout can be any flowable material, e.g.,
cement-bentonite or polymers of a wide variety.
Figure 5-9 Various grouted geomembrane panel interlocks

c. Welded Interlocks. Lastly, it might be possible to create a completely
seamed system by in situ welding of the sheets or panels. This possibility is
currently being investigated. To date, there has been only one field installation
of this type. This interlock category is added for completeness and as an
indication of what might become more established.
5.2.3 Connections to Covers and/or Floors
The end connections (top and bottom) of geomembrane vertical barriers are
important in providing a completely contained system. The vertical wall must
be installed first; then the cover (or cap) is placed on the contained waste.
Depending on the details, in particular the layering of geosynthetics in the
cover, there are several connection possibilities (see Section 6 on caps for
details).
Since there is usually a geomembrane in the cover, the possibility of
mechanically attaching the geomembrane in the cover to the geomembrane
in the vertical wall should be considered, as was shown in Figure 5-1. The
two geomembranes must be compatible insofar as their seaming is concerned.
107
Cover geomembranes are often designed for out-of-plane flexibility; thus,

very flexible polyethylene (VFPE), flexible polypropylene (fPP) or polyvinyl
chloride (PVC) are generally used. While difficult to accomplish, HDPE can
be seamed to VFPE and fPP. HDPE cannot be seamed to PVC. For the
polyolefins (VFPE and fPP), extrusion fillet welding is required, a difficult
process that requires welding over the interlocks. However, it is possible, as
depicted in Figure 5-1(a). Alternatively, the cover geomembrane can be
extended horizontally or dropped down vertically over the outside of the
geomembrane vertical wall as shown in Figure 5-1(b).
The termination of a geomembrane vertical wall at the base, and in
particular its connection to the lower aquitard or confining layer, is important,
especially since the termination cannot be visually inspected. If a low
permeability aquitard is available, an adequate geomembrane penetration
depth into such a layer should be specified as depicted in Figure 5-2(a). The
backfill material at the base of the geomembrane is important and a low
hydraulic SB, SC, CB, SCB, or polymer grout, should be considered. In the
absence of an aquitard, a hanging wall can be created, as depicted in Figure
5-2(b). The depth of a hanging wall is determined on the basis of a seepage
analysis to limit flow beneath the bottom of the geomembrane. This depth is
somewhat subjective and is essentially a design/regulatory decision.
The bottom of a geomembrane vertical wall can also be keyed into an
artificially placed floor. In general, the artificially placed floor should be
constructed before the geomembrane vertical wall is installed.
5.2.4 Case Studies
Several significant case histories have been described in professional journals
and at conferences that illustrate the use of geomembrane vertical barriers.
See Table 5-4 for an overview.
Burnette and Schmednecht (1994) describe a geomembrane vertical barrier
that was constructed around a Great Lakes chemical plant to prevent off-site
migration of contaminated groundwater. The vertical barrier was constructed
through layers of sand, gravel, cobbles, and boulders. The vibrating beam
installation method with a clay/cement slurry was used to construct a 120
mm (5 in.) wide trench. HDPEpanels, 1.2 m (4 ft) wide and 2.0 mm (80 mil)
thick, were inserted into the trench with a steel template. Trench depths were
up to 10 m (35 ft). Jet grouting was used to construct vertical barriers at the
locations with underground or above ground interferences that could not be
rerouted. Average daily production for the entire project was approximately
100 sq. m (1100 sq. ft), with an improved average of approximately 160 sq. m
(1700 sq. ft) during the second half of the project.
Bliss and Burnette (1995) describe 20 km (12.5 mi) of existing earthen
dike embankments in Arizona that were modified with a geomembrane
vertical barrier to provide both a seepage barrier and a filter zone. A 600 mm
(24 in.) wide trench was excavated using a biopolymer slurry for stability.
108
Large, 2.0 mm (80 mil) HDPE geomembrane panels, 7.5 m (24 ft) wide by 16.5
m (55 ft) long, were inserted into the trench with a large steel frame. Granular
backfill material was tremied into the trench as a drainage layer. A full-scale
test was conducted to evaluate the constructability and effectiveness of the
design. A 15 m (50 ft) deep by 340 m (1100 ft) long test section was constructed.
Quality assurance verification of all joints was accomplished with the
electronic circuit method described later in this section. Instrumentation was
installed and the test section impounded with water for 30 days. Seepage
results based on several flow rate monitoring stations indicated the
geomembrane provided an excellent vertical barrier.
TABLE 5-4 Case Studies Utilizing Geomembrane as Vertical Barriers
Depth of
Wall
Length of
Wall
Reference
Type of
Contained
Waste
Type of
Installation
(m)
(ft)
(km)
(mi.)
Burnette &
Schmednecht
hazardous
waste
vibrating
beam
10
35
0.3
0.2
Curtain Wall clay/cement

slurry
Interlock
Bliss &
Burnette
earth dam
cutoff
slurry
supported
15
50
20
12.5
Curtain Wall
Interlock
sand
Burnette &
Pierce
petroleum
wastes
slurry
supported
4.5
15
0.4
0.2
Curtain Wall
Interlock
sand
Burnette &
Pierce
hazardous
wastes
slurry
supported
14
45
0.5
0.3
Curtain Wall
Interlock
sand
Scuero, et al
earth dam
cutoff
slurry
supported
30
0.1
0.1
Michalangeli ash & MSW

Hansen &
Crotty
contaminated
drilling
waste
slurry
supported
trenching
machine
various
3
10
various
34
21
Type of
Interlock
Geolock
Backfill
Material
CB slurry
grouted
CB slurry
welding
natural soil
Scuero, et al. (1990) describe a 2.0 mm (80 mil) HDPEvertical barrier that
was constructed in southern Italy at the upstream toe of a 110 m (360 ft) long
earth dam to limit seepage under the dam. A 500 mm (20 in.) wide by 9 m (30
ft.) deep trench was excavated and a geomembrane installed and keyed to a
minimum of 1 m (3 ft) into an underlying confining stratum. Cement-bentonite
slurry was used to seal the geomembrane into the key. Concrete was placed
at the top of the HDPE vertical barrier to provide a water tight connection
between the vertical barrier and the PVC liner on the downstream slope of
the earth dam.
Michalangeli (1995) describes a series of four case histories using
geomembrane panels as vertical barriers around landfill sites in Italy. In all
cases, 2.0 mm (80 mil) thick HDPEgeomembranes were used with groutedtype interlocks. The projects were extremely large, up to 130,000 sq. m (155,000
sq. yd).
Hansen and Crotty (1995) describe the installation of 34 km (21 mi) of
geomembrane vertical barrier in Alaskas North Slope oil fields to contain
5.3 PERFORMANCE OF THE TECHNOLOGY
109
drilling waste fluids. Various geomembranes were used, including; HDPE,

reinforced PVC, and reinforced polyurethane. Single and double lined systems
were installed. Double lined systems comprised about 85% of the installations.
A trencher was used to excavate a 300 mm (12 in.) wide trench. Typical depths
were 3.0 m (10 ft) and the barriers were keyed into permafrost. Geomembranes
were rolled out in the trenches and roll ends were seamed by a thermal fusion
method. Penetrations through the vertical barrier by underground utilities
were sealed with boots and welding. The liners were installed in temperatures
as low as 42C (44F). Anticipated ground temperatures were 32C (26F).
Excavated soil was used for backfill. Production was about 150 m (500 ft) per
10 hour shift in spite of snow, extremely cold weather, and wind.

This section on the performance of geomembrane vertical barriers discusses
both laboratory and field issues. It is focused completely on HDPE, due to its
general use in vertical barriers.
5.3.1 Durability of HDPE
The chemical resistance of HDPE geomembranes is well established, based
on thousands of EPA 9090 compatibility tests. A wide range of leachates have
been evaluated for landfill/waste pile/heap leach materials. For this reason,
blanket approval of HDPE has often been requested of the regulatory
community and, while not granted at the federal level, many states currently
sole source HDPE geomembranes for the containment of all types of solid
waste. It is used for 80 to 90% of the liners beneath nonhazardous and
hazardous landfills, waste piles, and surface impoundments. It is generally
thought that low level radioactive waste poses no degradation concern;
however, high level radioactive waste must be evaluated on a site specific
basis.
The service lifetime of HDPE geomembranes should be considered for
the containment of all waste materials, especially for low-level radioactive,
low-level radioactive mixed, and hazardous wastes. The common mechanism
for long-term degradation of HDPE is chemical oxidation. Several studies
have been conducted to assess the lifetime of HDPE polymer, particularly in
the gas transmission pipe, electrical cable shielding, and geomembrane
application areas. It is generally considered that HDPE passes through the
following three different stages during its service lifetime:
depletion of anti-oxidants,
an induction time preceding the onset of degradation, and
time for degradation of 50% of a relevant engineering property, such
as strength or elongation (the so-called half-life).
110
These three stages are shown in Figure 5-10 as A, B and C, respectively.

Since field degradation of buried HDPE geomembranes has not been reported
since their introduction as lining materials in 1978, direct field data are not
available. Thus, emphasis must be placed on laboratory simulation and
testing. (Note that uncovered geomembranes will degrade more rapidly when
exposed to ultraviolet radiation and/or high temperatures; hence, this
discussion pertains to covered, or buried, geomembranes).
Selected Properties
T1 > T2 > T3 > T4
100
50
T1
T2
T3
T4
Exposure Time
A = Depletion of anti-oxidants
B = Induction time
C = Time to half-life
Figure 5-10 The stages of lifetime of HDPE geomembrane utilizing time

temperature superposition technique
The accepted method for lifetime simulation of polymer property behavior
is based on time-temperature testing at laboratory-controlled temperatures
combined with Arrhenius modeling (fitting an exponential-type curve to the
experimental data) and using the model to estimate the anticipated response
at a site-specific temperature. The procedure involves incubation of samples
at several elevated temperatures, determination of the time-dependent
behavior of the selected polymer property, and extrapolation of the behavior
to the site specific temperature. The estimated lifetime of the sample property
is predicted by this procedure.
Work at the Geosynthetic Research Institute has resulted in experimental
estimates of the time for antioxidant depletion between 50 to 150 years, an
unknown incubation time (perhaps decades), and half-life degradation times
of 200 to 750 years, Koerner, et al. (1990) and Hsuan, et al. (1993). The half-life
degradation estimates utilized activation energies from the literature, see the
references cited in the above articles.
As a result of the above laboratory simulation studies, in-service lifetimes
for buried HDPE could approach 1000 years.
111
5.3.2 Diffusion through HDPE

Diffusion of water or solvent through HDPE geomembranes can only
occur in a vapor state. Upon adsorption of the liquid phase into the
geomembrane as a vapor, it enters the polymer structure and diffuses through
the amorphous phase. After diffusing through the geomembrane, the vapors
condense to liquid on the opposite side of the geomembrane.
HDPEhas the lowest diffusion rate of any commonly used polymer for
geomembranes. Vapor diffusion rates for a number of solvents were given in
Table 5-2. Note that solvent vapors may have higher or lower diffusion rates
compared to water vapor. HDPE sheets used in vertical wall panels are
typically 2.0 mm (80 mil) thick, i.e., thicker than most geomembranes used
for landfill liners and covers.
While there is certainly a finite diffusion rate of solvents through
HDPEgeomembrane walls, Park, et al. (1995), it is so low that the focus of
attention is usually placed on the integrity of the interlocks.
5.3.3 Interlock Continuity and Testing
Geomembrane panels must be joined together by interlocks, with the most
common type being the hydrophilic gasket type shown in Figure 5-8. To test
the continuity of such interlocks, it is possible to place a contact element and
conductive wire on each paired interlock, Bliss and Burnette (1995). The
contact elements are positioned at the bottom of the interlocks so as to touch
one another when the two panels are properly installed and fully mated to
one another. A battery is connected to the wires at the ground surface, creating
an electrical circuit. Measuring the resistance in the open versus closed circuit
verifies the continuity of the interlock at the intended depth.
Field testing for interlock liquid tightness is very difficult; however, a
laboratory test is available. The experimental test setup is shown in Figure 511, Gundle (1994). Test results at pressures up to 800 kPa (100 lb/in2) exhibited
only nominal seepage.
The testing of grouted interlocks as shown in Figure 5-9 can be
accomplished by observing the flow of grout coming out of the adjacent
channel or outer pipe, respectively. Although not known to have been
performed in the field, a volumetric check probably could be developed.
5.3.4 CQC and CQA
In addition to the techniques described above, both construction quality
control (CQC) and construction quality assurance (CQA) should be employed
during installation of geomembrane vertical barriers. These procedures are
well advanced in the landfill liner and cover application areas, EPA(1993). A
certified field inspectors program exists through the National Institute for
Certification of Engineering Technologists, NICET (1992) based on experience.
112
Level I (entry), Level II (2-year experience) and Level III (5-year experience)
examinations, resulting in individual certifications are available for both CQC
and CQApersonnel. A certification program is not yet available for inspection
of geomembranes in vertical barriers, the possible reasons being;
the field is too new,

the field is not large enough, or
a demand has not existed.
It should be possible to expand the existing certification program to include

geomembrane vertical barriers. CQC/CQA for geomembrane vertical walls
is similar in concept to the installation of geomembranes in caps. However, it
may be more difficult and require a greater level of effort than when the
geomembrane is horizontal or on a slope, but certainly proper CQC/CQAis
possible.
Rubber
Membrane
Inflow
Port
Filter
Paper
Hydrotite
Flow Rate (gal/min/ft)
10-4
Male
Section
Filter
Paper
Female
Section
Outflow
Port
10-5
10-6
10-7
10-8
10-9
0 10 20 30 40 50 60 70 80 90 100 110
Pressure Difference (psi)
Figure 5-11 Hydrophilic gasket interlock test developed by GeoSyntec for

Gundle Lining Systems, Inc.
5.3.5 Long-Term Monitoring
There has only been indirect evidence of the effectiveness of geomembrane
113
vertical walls based on long-term monitoring. For example, downstream

monitoring wells have not indicated the presence of contaminant plumes
emanating from landfills contained by geomembrane vertical walls. More
direct evidence could be provided by lysimeters located downstream of the
geomembrane. Construction of a double walled geomembrane with a leak
detection layer between the two geomembranes, as shown in Figure 5-12,
would provide even better direct evidence of the effectiveness of a geomembrane
vertical wall. Possible arrangements include geomembranes being placed on
both sides of the excavated trench and sand filled between, or with a
geomembrane/geonet/geomembrane composite being installed as a self
contained unit. To our knowledge, the latter has not yet been implemented.
100-mil HDPE
geomembrane
Monitoring well
HDPE
geomembrane
Existing grade
Soil permeated
with bentonite
Grout
monitoring
layer
HDPE
geomembrane
Backfill
(Excavated soil,
SB, CB, SC, etc.)
Sand backfill
Clay layer
or bedrock
3'0"
3'0"
Figure 5-12 Two possible doublewall schemes for leak detection monitoring
5.3.6 Warrants
Geomembrane manufacturers and installers do give warrants on materials
and workmanship, but it must be recognized that the duration of the warrant
will affect the project cost. This is not because of trepidation on the part of the
manufacturer and/or installers, but because funds will have to be escrowed
to insure the warrant.
The service lifetime of HDPE will extend long beyond the normal 20year warrant. Furthermore, any installation flaw should evidence itself soon
after completion of the project; thus, a long-term installation warrant is
generally not necessary. For example, a hole or broken interlock in a
geomembrane vertical wall caused during construction will leak immediately,
114
rather than after many years of service.

Thus warrants, while available, are felt to be of limited value, especially
since they are pro-rata and are reflected in the bid cost of the installation.

An assessment of geomembrane vertical barriers, as presented by the
workshop panelists, is presented utilizing a generalized table, with
approximate costs, followed by miscellaneous items, including emerging
technologies.
5.4.1 General Assessment, Including Estimated Costs
An overall assessment of geomembrane vertical barrier technologies is given
in summary form in Table 5-5. It follows the order of the five installation
methods presented earlier in this section. The methods have been placed in
the approximate order of increasing cost per unit area of wall. Table 5-5 also
includes some comments concerning advantages or disadvantages of each
method. It should be noted, however, that the prices and associated comments
reflect site specific information and are very generalized.
TABLE 5-5 General Assessment of Geomembranes as Vertical Walls
Method No.
Method
Name
trenching
machine
Typical Cost
$/ft2
2-5
$/m2
20-50
Some Advantages
Some Disadvantages
no seams
rapid installation
no slurry
depth limited
soil type limited
trench stability
necessary
rapid installation
narrow trench
no material spoil
no slurry
soil type limited

possible panel
stressing
bottom key is a
concern
vibrated
insertion plate
3-7
30-70
slurry
supported
5-15
50-150
segmented
trench box
16-18
can weld seams

visual inspection
160-180
no stress on panels
no slurry
vibrating
beam
18-25
requires slurry
narrow trench
no material spoil slow incremental
180-250
no stress on panels process
usually CB slurry soil type limited
no stress on panels requires slurry

conventional
buoyancy concerns
slow process
method
choice of backfill
depth limited
soil type limited
slow incremental
process
115
5.4.2 Leak Detection Methods

As with most vertical barriers, post-construction leak detection is performed
via downgradient monitoring wells. The number and spacing of these wells
is a critical decision. It is, perhaps, a more critical decision with geomembranes
as vertical walls as compared to other wall types, since leakage would probably
be from a single (or few in number) point source where a flaw is located.
Thus, the plume from such a leak would not be spread over a large area where
it could be detected by widely spaced monitoring wells. The actual situation
is obviously very site specific.
For critical applications, however, a different strategy for leak detection
should be considered when using geomembranes as vertical walls. Two
geomembranes should be employed, with an intermediate sand or geonet
drainage layer acting as a leak detection system, as shown in Figure 5-12.
With sand as the leak detection drainage layer, the geomembranes would be
placed on each side of the excavated trench. With a geonet as the drainage
layer, the composite system (geomembrane/geonet/geomembrane) is placed
against one side of the trench and the remainder backfilled as desired. For
such a double liner system placed against the side of the trench facing the
waste and the opposite side backfilled with a low hydraulic conductivity soil,
cement, or grouted material, one has the vertical equivalent of a hazardous
waste liner system, i.e., Subtitle C type. Such double liner systems can be
the most secure of the various alternatives within the vertical wall category.
5.4.3 Geomembranes vs. Other Vertical Barriers
In comparing geomembrane vertical walls to other vertical walls described
in this report, many considerations must be assessed. For example, the
following aspects must be considered;
hazard potential of the waste,

presence (or not) of an aquitard,
site stratigraphy,
depth of wall,
configuration of wall,
space considerations/site availability,
allowable hydraulic conductivity through barrier,
anticipated ground movements,
allowable diffusion through barrier,
allowable leakage (if any),
disposal of excavated material from trench,
durability (aging) of system,
capability of long-term monitoring,
risk assessment of the selected system,
availability of materials, installers and equipment, and
116
cost of the installed system.
With this many considerations, it is impossible to prioritize or even

compare many of the alternative strategies for design, construction and
monitoring of different types of vertical walls. It is, however, relevant to note
that for critical containment applications, a double wall with leak detection
capability and a low permeability soil (or other material) backup can provide
the most secure containment of the various types of wall systems that are
available.
5.4.4 Emerging Technologies
Emerging geomembrane vertical wall technologies can be grouped according
to materials, installation methods, and interlock connections. As a material,
HDPE dominates the currently available systems. On the basis of cost and
availability it is likely to continue to do so in the future. This is not to say that
other, better systems might not become available. For example, high modulus,
high strength polymers would permit direct driving (like steel sheet piles),
providing an additional installation method to those described in this section.
For those situations where diffusive transport is of concern, a cross-linked
HDPE is a distinct possibility as is the potential for using geosynthetic clay
liners, Heerten, et al. (1996). Even further, a geomembrane cured in situ is
conceivable. Fiber optic sensors within the geomembrane, i.e., the use of smart
geomembranes, are being evaluated for both leak detection/location and
estimation of material stresses.
There are many possible installation methods in addition to the five
described in this chapter. Undoubtedly, others will appear in the future. The
new excavation method of soil sawing is interesting in this regard and could
be considered as a subset of the trenching method depicted in Figure 5-3. The
future may bring about full fusion welding of connections in the trench to
join the individual panels together. The segmented trench box method
described by Figure 5-6 provides such an opportunity. Welding may even
become possible while in the slurry of the supported trench. Such methods,
however, are in the development stage at this time.
5.5 NEEDS
In assessing the needs within the state-of-the-practice of geomembranes
used as vertical barriers, the Panel did not call for laboratory tests nor
additional R & D to be initiated and performed. The general feeling of the
Panel is that the technology exists and is available for application by the
owner/user community. The question was raised as to why geomembrane
vertical barriers are not used more than currently. Clearly, part of the answer
is the added cost compared to vertical barriers without geomembranes. This
5.6 REFERENCES
117
cost can range from a 10 to 30% increase above the cost of a conventional SB,
SC, CB or SCB backfilled vertical wall without a geomembrane. However,
the added environmental safety and security that a continuous barrier, even
a composite barrier, offers may justify such an increase in cost.
What is needed for increased use of geomembrane vertical barriers is
greater exposure of the technology. Such exposure can come about from
published case studies, of which there are relatively few in the published
literature, (see Table 5-4). The development and eventual publication of
performance case studies should generate increased interest, especially if the
related advantages and disadvantages of the system are noted for the
particular situation.
Increased exposure through publications would eventually bring about
performance based designs. With such design methodologies, direct
comparisons could be made between geomembrane vertical barriers and other
vertical barriers without the inclusion of geomembranes. Such a methodology
would likely be a significant factor in decision-making by owners, designers
and regulators alike.
Finally, a few full scale demonstration projects would be significant in
illustrating the potential for application of geomembrane vertical barriers in
environmental remediation. Such demonstrations have not been undertaken
to date. Double geomembrane technology combined with leak detection
capability between the two geomembranes offers an attractive option for
construction of containment barriers.
5.6 REFERENCES
Bliss, M. and Burnette, P. T. (1995). Reach 11 Dikes Modification: A Vertical Barrier
Wall of HDPE Geomembrane, Proc. Geosynthetics 95, IFAI, pp. 147-160.
Burnette, P. and Schmednecht, E. (1994). Vibrating Beam, Curtain Wall and
Jet Grouting Used to Form a Vertical Barrier Wall, Proceedings, Polluted
and Marginal Land 95, London, England.
Burnette, P. and Pierce, D. (1994). Recent Developments for the Containment
of Lateral Migration of Petroleum Hazardous Waste in Ground Water,
Proceedings, Petro-Safe 94.
Gundle Lining Systems, Inc. (1993). Laboratory Test Results of Gundwall
Locking Section with Hydrotite Leakage Potential Testing, GeoSyntec
Consultants, Project No. GL3398, Houston, Texas.
Hansen, P. G. and Crotty, G. R. (1995). Use of Geomembranes as Vertical
Barrier Liners for Containment on the North Slope of Alaska, Proc.
Geosynthetics 95, IFAI.
118
Haxo, H. E. (1988). Lining of Waste Containment and Other Impoundment

Facilities, U.S. EPA 600/2-88/052,1600 pgs.
Heerten, G., Saathoff, F., von Maubeuge, K. P. and Hahn, A. (1996).
Geosynthetic Clay Liners as Gas and Liquid Barriers in Caps and Polluted
Areas, Proc. ASTM Symp. on Geosynthetic Clay Liners, Atlanta, ASTM, in
press.
Hsuan, Y. G., Koerner, R. M., and Lord, A. E., Jr. (1993). A Review of the
Degradation of Geosynthetic Reinforcing Materials and Various Polymer
Stabilization Methods, Geosynthetic Soil Reinforcement Testing
Procedures, ASTM STP 1190, S. C. Jonathan Cheng, Ed., American Society
for Testing and Materials, Philadelphia, PA, pp. 228-244.
Koerner, R. M. (1986). Designing with Geosynthetics, 1st Edition, Prentice
Hall Publ. Co., Englewood Cliffs, NJ, 424 pgs.
Koerner, R. M. (1994). Designing with Geosynthetics, 3rd Edition, Prentice
Hall Publ.Co., Englewood Cliffs, NJ, 783 pgs.
Koerner, R. M., Halse, Y. H. and Lord, A. E., Jr. (1990). Long-Term Durability and
Aging of Geomembranes, Proc. ASCE Symp. on Waste Containment Systems,
R. Bonaparte, Ed., Geotech. Spec. Publ. No. 26, ASCE, pp. 106-134.
Koerner, R. M. and Lord, A. E., Jr. and Halse, Y. H. (1989). Geosynthetic
Concerns in Landfill Liner and Collection Systems, Proc. 15th
EPAResearch Symposium, pp. 369-375.
Michalangeli, A., Italian Experience with HDPE Geomembranes in Landfill
Liners, Proc. Geosynthetics 95, IFAI, pp. 569-584.
NICET (1992). Geosynthetic Materials Installation Inspection, Program
Detail Manual 1030-10-92, National Institute for Certification in
Engineering Technologies, Alexandria, Virginia.
Park, J. K., Sakti, J. P. and Hoopes, J. A.. (1995). Effectiveness of Geomembranes as
Barriers for Organic Compounds, Proc. Geosynthetics 95, IFAI, pp. 879-892.
Scuero, A. M., Nuzzo, G. and Salvemini, A. (1990). A Continuous Barrier
from Bottom to Crown in Hydraulic Earthfield Structures, Proc. 4th IGS
Conference on Geosynthetics, G. den Hoedt, Ed., Balkema, pp. 425-429.
U.S. Environmental Protection Agency (1993). QualityAssurance and Quality
Control for Waste Containment Facilities, Technical Guidance Document,
EPA/600/R-93/182, 305 pgs.
119
SECTION 6
CAPS
prepared by
David E. Daniel and Beth A. Gross, University of Texas

Robert C. Bachus, GeoSyntec Consultants

Craig H. Benson, University of Wisconsin
Jack Boschuk, Jr., J&L Testing Co., Inc.
Subijoy Dutta, U.S. Environmental Protection Agency
Lorne G. Everett, Geraghty & Miller
Harley Freeman, Pacific Northwest National Laboratory
Glendon W. Gee, Pacific Northwest National Laboratory
Ed. Kavazanjian, Jr., GeoSyntec Consultants
Robert M. Koerner, Drexel University
Robert E. Landreth, U.S. Environmental Protection Agency
W. Eric Limbach, Idaho State University
Stefan Melchior, University of Hamburg, Germany
Robert W. Ridky, University of Maryland
Paul R. Schroeder, U.S. Army Corps of Engineers
Kenneth Skahn, U.S. Environmental Protection Agency
John C. Stormont, Sandia National Laboratories
Andrew Street, RUST Environmental, United Kingdom
Francke C. Walberg, U.S. Army Corps of Engineers
Thomas F. Zimmie, Rensselaer Polytechnic Institute
6.1 BACKGROUND
Caps, also called cover systems or surface barriers, are one component
of the engineered systems used to manage wastes at modern waste disposal
units and, not infrequently, at contaminated sites as part of remedial actions.
The purpose of a cap is to perform one or more of the following functions
(Rumer and Ryan, 1995):
minimize percolation of water into the underlying contaminated

materials;
raise the ground surface and provide appropriate slopes to promote
surface-water runoff;
control the release of gas from the contaminated materials; and
separate the contaminated materials from humans, animals, and plants.
120
CAPS
Caps are composed of six basic components, from top to bottom: surface
layer, protection layer, drainage layer, barrier layer, gas collection layer, and
foundation layer. The layers are illustrated in Figure 6-1, and the primary
functions and potential materials for each layer are listed in Table 6-1. All
layers must have adequate durability so that they function over the design
life of the cap and adequate shear strength so that the cap surface slopes are
stable. Some layers may contain several materials. For example, a hydraulic
barrier layer may consist of a geomembrane upper component and a compacted clay lower component (i.e., a composite). Not all layers are needed
for all sites. For example, a drainage layer may not be needed at an arid site.
All caps, however, require a surface layer.
yer
Surface La
yer
Protection La
Drainage Laye
Barrier Layer
Gas Collection
Layer
yer
Foundation La
Figure 6-1 Cap cross section

The current knowledge and practices for cap technology have been recently reviewed and evaluated (Rumer and Ryan, 1995). The review included
an evaluation of the materials that are used to construct caps, a description of
construction quality control procedures used for barrier layers, and an assessment of factors that can impact the overall performance of caps. The
primary factors that influence cap performance were identified as:
layer types included in the cap, layer materials, and layer thicknesses,
annual precipitation,
surface slope angle of the cap,
compressibility of the underlying waste,
gas generation, and
presence of burrowing animals.
6.1 BACKGROUND
121
TABLE 6-1 Cap components

Layer
Primary Functions
Potential Materials
Separate underlying layers from the

ground surface
Resist wind and water erosion
Reduce temperature and moisture
extremes in underlying layers
Topsoil (vegetated)
Geosynthetic erosion control
layer over topsoil (vegetated)
Cobbles
Paving material
Protection
Layer
Store water that has infiltrated the

surface layer until the water can be
removed by evapotranspiration
Separate the waste from humans,
burrowing animals, and plant roots
Protect underlying layers from wettingdrying cycles, which may cause cracking
Protect underlying layers from freezingthawing cycles, which may cause cracking
Soil
Cobbles
Recycled or reused waste
(e.g., fly ash, bottom ash,
and paper mill sludge)
Drainage
Layer
Reduce the head of water on the barrier Sand or gravel

layer, thereby decreasing water
Geonet or geocomposite
percolation through the cap
Reduce pore water pressures in the
overlying cap layers, thereby increasing
slope stability
Reduce the time the overlying layers are
saturated following rainfall events,
thereby decreasing erosion
Surface
Layer
Hydraulic
Barrier
Layer
Gas
Collection
Layer
Foundation
Layer
Impede percolation of water through the

cap
Restrict outward movement of gases
from the waste
Compacted clay
Geomembrane
Geosynthetic clay liner
Asphalt
Sand or gravel capillary barrier
Collect and remove gases to reduce the

potential for uncontrolled gas migration
Collect gas for energy recovery
Sand or gravel
Geonet or geocomposite
Geotextile
Serve as a foundation for the cap,

especially during construction of layers
requiring compaction (e.g., compacted
clay barrier layer)
Sand or gravel
Soil
Select waste
In that review, it was noted that little published information was available
concerning the performance of caps. The performance data are limited because modern multicomponent caps have only been in existence for about a
decade, giving insufficient time for much data to be gathered. Also, it is difficult to quantify the percolation of water or the migration of gas through a cap
on the basis of the available data because the data are not normally collected
for the purpose of evaluating the performance of the cap.
122
CAPS
The purpose of this section is to review the current technological status

of caps, including material performance, design methods, field performance
(i.e., cap failures), and field monitoring. The focus is on field performance
and field monitoring, as this aspect of caps was identified as an area requiring further research in the earlier publication by Rumer and Ryan (1996). To
this end, researchers and practitioners in the U.S. and Europe involved in the
conduct of field evaluations were asked to share both published and unpublished data. These data are summarized in this section.
It should be recognized that much of the current knowledge about caps
has been developed from landfill capping applications rather than from experiences from containment of contaminated lands. It is important, however,
to learn from the landfill capping experience, since this is the principal source
of the best available information.

6.2.1 Introduction
The design of a cap for a specific application depends on the requirements of
the application. While the cap requirements for hazardous waste landfills
may differ from those for radioactive waste landfills or municipal waste
landfills, the procedures used to design caps for different applications are
generally the same. The current state of practice for the design of conventional caps for municipal solid waste (MSW) landfills has been described by
Othman et al. (1995). As outlined in their paper, major design aspects to be
considered relate to: (i) flow of water into and through the cap; (ii) impacts of
waste settlement on the cap performance; (iii) static and dynamic stability of
the cap; and (iv) surfacewater management and erosion control. Other considerations in cap design include: (i) gas collection and removal; (ii) the
need for bio-barriers; and (iii) long-term durability of cap materials.
Several issues related to cap materials and cap design, namely, the service life of high density polyethlyene (HDPE) geomembranes, internal shear
strength of geosynthetic clay liners (GCLs), properties of the alternative barrier layer materials asphalt and paper mill sludge, slope stability of caps, and
water balance modeling, are reviewed in this section.
6.2.2 Service Life of HDPE Geomembranes
Polyethlyene was first synthesized in 1933; however, it was not used in liners
until the 1960s. HDPE geomembranes have not been used long enough to
field test their durability. Nevertheless, the service life of HDPE
geomembranes can be estimated using Arrhenius modeling to interpret the
results of laboratory tests conducted at elevated temperatures (see Section 5).
Research on the service life of HDPE geomembrane is currently being con-
123
ducted by R.M. Koerner and Y. Hsuan, both of the Drexel University

Geosynthetic Research Institute (GRI). The approach is described by Koerner,
et al. (1992). They identified eight factors influencing degradation of HDPE
geomembrane: (i) oxidation; (ii) chemical attack; (iii) hydrolytic effects; (iv)
ultraviolet radiation; (v) nuclear radiation; (vi) biological attack; (vii) stress
effects; and (viii) temperature effects. Of these factors, oxidation, hydrolysis,
stress, and temperature are being considered in their study. Koerner and
Hsuan defined service life as the time for depletion of antioxidants, initiation
of degradation (i.e., induction time), and degradation to half-life of relevant
strength properties. All of these factors may affect the effectiveness of a
geomembrane as a barrier component. Strength properties are used rather
than permeation properties because loss of antioxidant has little effect on
permeation characteristics but can cause embrittlement and major changes in
strength and elongation characteristics. The study is scheduled to be conducted over a 10-year time period. Based on the approximately three years
of data currently available, depletion of antioxidants from HDPE
geomembrane may take from 45 to 115 years and the geomembrane service
life may be about 250 to 900 years, depending on the specific product and inplace conditions.
6.2.3 Internal Shear Strength of GCLs
As described by Daniel and Koerner (1993), a cap incorporating a GCL barrier layer can outperform a cap with a compacted clay layer. While a clay
layer will frequently crack due to differential settlement, freezing-thawing
cycles, and wetting-drying cycles, a GCL possesses some self-healing capability when subjected to these stresses. Also, a GCL is easier to install and to
repair if damaged. One of the primary concerns with using GCLs in caps is
the relatively low shear strength at mid-plane, as reported for hydrated specimens of some GCLs when tested in the laboratory (Rumer and Ryan, 1995).
GCLs fabricated with internal reinforcing (e.g., needlepunched GCLs) have a
higher peak strength than unreinforced GCLs, but, have been found to exhibit a significant decrease in strength when subjected in the laboratory to
relatively large shear displacements. A field-scale study is currently underway
to: 1) evaluate the mid-plane shear strength of different types of GCLs, 2)
verify that caps incorporating GCLs will remain stable on 3H:1V slopes, and
3) verify that stable caps have a factor of safety of 1.5 or greater for slope
stability. This study, which is being conducted jointly by the University of
Texas, Drexel University, and GeoSyntec is described below.
Fourteen cap test plots, 20 m (66 ft) long by 3 m (10 ft) wide, were constructed with slopes of 3 horizontal: 1 vertical (3H:1V) and 2H:1V. The cap
cross sections consist of, from top to bottom:
0.9 m (3 ft) thick cover soil layer, covered with a geosynthetic erosion
mat and grassed;
124
CAPS
sand (0.3 m, or 1 ft thick) or geocomposite drainage layer;

textured geomembrane/GCL composite or GCL barrier layer; and
subsoil.
The GCLs used in the study were geotextile-encased and needlepunched

(i.e., Bentomat, Bentofix I, and Bentofix II), geotextile-encased, adhesivebonded, and stitched (i.e., Claymax), and a geomembrane-bentonite composite
(i.e., Gundseal). To try to force shear failure through the GCL, all geosynthetic
layers above the mid-plane of the GCL, including the upper geotextile of
geotextile encased GCLs, were cut at the top of the slope. Also the cover soil
was removed from the toe of the slope to eliminate toe buttressing. The deformations and moisture contents of the GCLs are being monitored at selected depths using tell-tales and fiberglass resistance cells, respectively. Two
cap plots on 2H:1V slopes failed as blocks at 21 and 52 days after the cover
soil was placed. In both cases, the GCLs were internally reinforced, with the
failure surface located between the textured geomembrane and the woven
geotextile of the GCL, rather than within the GCL. Based on the test results
to date, the following observations have been made:
The factor of safety against mid-plane failure of GCLs on 3H:1V slopes

has been at least 1.5, based on the fact that 2H:1V slopes have been
stable. (Theoretically, if a 2H:1V slope has a factor of safety > 1.0,
then the factor of safety for a 3H:1V slope composed of the same
materials is > 1.5.)
For reinforced GCLs, the mid-plane strength shortly after construction hasbeen greater than the interface strength between the woven
geotextile of the GCL and the textured geomembrane.
Bentonite sandwiched between two geomembranes was found be
partially hydrated at one test plot. The mechanism for this wetting is
unclear, though it may be due to water migration along the tell-tale
and fiberglass resistance cell cables. At the two other test plots with
a similar GCL type, the bentonite between the two geomembranes
has remained dry. The GCLs in the remaining 11 test plots became
hydrated within several months after construction as a result of absorption of moisture from the underlying, wet subsoils.
6.2.4 Properties of Asphalt

A composite asphalt barrier, composed of a 5-mm (200-mil) thick fluid applied asphalt membrane (FAM) overlying a 0.15-m (0.5-ft) thick hot-mix asphalt concrete (HMAC) layer, is being considered for the cap at the Hanford
site as an alternative to traditional barrier layer materials (Freeman et al.,
1994). Asphalt may be preferable to traditional barrier materials, such as
geomembranes and compacted clay layers, since asphalt appears to have a
longer service life than those materials. The expectation for long service life is
125
based on the fact that asphalt occurs naturally and is known to have existed
naturally in the subsurface for millions of years. Also, asphalt artifacts, up to
several thousand years old, have been found by archaeologists. Asphalt has
also been demonstrated to have a low permeability and be effective at controlling radon emissions. Asphalt barriers have been installed at several sites.
However, in many cases, asphalt may not be selected over other barrier materials because of its relatively high cost [e.g., $ 96/m2 ($ 80/yd2) for the asphalt barrier proposed for the Hanford site compared to less than $ 7/m2 ($
6/yd2) for installation of a 1.5-mm (60-mil) thick HDPE geomembrane]. Two
concerns with asphalt use in a barrier are: (i) the potential for asphalt to creep;
and (ii) asphalt cracking due to age hardening. Asphalt studies are currently
underway at the Pacific Northwest Laboratory (PNL). Two of these recent
asphalt studies are summarized below.
A 0.6 hectare (1.5 acre) prototype cap incorporating the composite asphalt barrier and an adjacent 18.6 m x 8.6 m (61 ft x 28 ft) barrier test pad were
constructed during 1994 at the Hanford site, located near Richland, Washington. Laboratory tests conducted on five core samples from the HMAC component of the barrier gave an average hydraulic conductivity of 1.3 x 10-9 cm/s.
Field falling head tests conducted at five locations and sealed double ring
infiltrometer (SDRI) tests conducted at two locations gave average hydraulic
conductivity values of 3.7 x 10-8 cm/s and 1.1 x 10-8 cm/s, respectively, for the
HMAC. The hydraulic conductivity of the FAM component of the barrier, as
measured in the laboratory for four field samples, was on the order of 1 x 10-11
cm/s. It should be noted that the potential exists for preferential flow through
the HMAC layer. The highest HMAC hydraulic conductivity of 1.1 x 10-7
cm/s was found for a core sample taken along a vertical asphalt seam. Also,
the SDRI test conducted over a vertical seam gave a somewhat higher hydraulic conductivity than the SDRI test performed over unseamed HMAC.
Lateral flow of water within the asphaltic concrete, between horizontal asphalt seams, was observed during the SDRI tests.
Accelerated aging tests have been developed at PNL, permitting the
rheological and chemical properties of asphalt to be determined as a function
of age. The procedure is being validated by comparison with ancient asphalt
artifacts, ranging in age from 500 to 4000 years, and asphalt from naturally
occurring seeps. Accelerated aging tests still need to be completed on a number
of asphalts to allow the long-term performance of composite asphalt barriers
to be predicted.
6.2.5 Properties of Paper Mill Sludge
Paper mill sludges have been used since 1995 as the barrier layer for some
caps constructed in Massachusetts and Wisconsin (Moo-Young and Zimmie,
1995). However, there is little information in the literature on the engineering properties of this sludge. Moo-Young and Zimmie performed laboratory
tests to determine water content, organic content, specific gravity, permeabil-
126
CAPS
ity, compaction, consolidation, and strength for seven paper mill sludges. They
found that the sludges had a high initial water content ranging from 150 to
270 percent, an initial hydraulic conductivity ranging from about 1 x 10-7 to 5
x 10-6 cm/s, and behaved similarly to a highly organic soil.
Moo-Young and Zimmie (1995) also performed laboratory tests on six
samples of a sludge used as the barrier layer material in a cap. Three samples
were obtained shortly after construction and the other three samples were
taken at nine, 18, and 24 months after construction. The results of the laboratory tests on these undisturbed samples indicated that the water content and
hydraulic conductivity of the sludge decreased somewhat over time as the
sludge consolidated and biodegraded (i.e., mineralized to become more like
a soil). The depth of frost penetration in the sludge barrier layer has been
monitored since 1992. To date, the frost layer has not penetrated into the
sludge layer due to the protection provided by the overlying cap layers and
the high water content of the sludge. Based on the results of laboratory tests
over a range of water contents, if a sludge layer is subjected to freezing and
thawing cycles, the hydraulic conductivity of the sludge may increase by one
to two orders of magnitude (Moo-Young and Zimmie, 1995).
6.2.6 Slope Stability
Static and seismic slope stability analyses are typically carried out as part of
cap design. The general procedures for evaluating the static and seismic stability of landfills caps have been summarized by Othman et al. (1995). One of
the more critical parts of the analyses involves the selection of the appropriate shear strength values to use for the cap materials and material interfaces.
While there are published shear strength values, it is recommended that
project-specific testing be conducted under the expected field conditions (e.g.,
soil moisture content and unit weight, consolidation load and time, interface
wetting conditions, normal stresses, shear direction, shear displacement rate
and magnitude). Other factors which may affect the long-term shear strength
properties of caps, such as freezing-thawing cycles, heating-cooling cycles,
and creep, should also be considered. However, no consistent standard of
practice currently exists for directly addressing the potential effects of these
factors on slope stability. They may be accounted for indirectly through use
of a higher safety factor. Other measures, such as increasing the thickness of
cover soils above the critical layers to provide thermal insulation and isolation from the environment, may be beneficial in some ways but detrimental
in others.
Seismic design of caps has been brought to the attention of the engineering community with the promulgation of the Resource Conservation and
Recovery Act (RCRA) Subtitle D regulations for municipal solid waste landfills
(40 CFR 258) which became effective in 1993 and with the observed field performance of several landfills during recent earthquakes in California. Under
the Subtitle D regulations, the performance of landfills located in seismic
127
impact zones must be evaluated. A seismic impact zone is an area with 10

percent or greater probability that the maximum horizontal accelerations in
bedrock will exceed 10% of the earths gravitational pull within 250 years.
With this definition, seismic impact zones cover most of the western U.S. and
part of the central and eastern U.S. Overall, these zones cover approximately
40 percent of the U.S., including areas where seismic design may not have
been considered in the past. Furthermore, under Subtitle D regulations, a
geomembrane/compacted clay composite barrier layer be incorporated in
caps over certain landfills, e.g., landfills containing composite bottom barriers. Incorporating a geomembrane (as well as other geosynthetics) in the
cap, however, may make the cap more susceptible to instability and deformations induced by seismic loading due to relatively low interface shear
strengths.
Observations of the seismic performance of numerous (over 30) landfills
in California, based primarily on field inspections after recent earthquakes,
indicate that, in general, the landfills performed well but that some limited
damage did occur to soil covers (Anderson and Kavazanjian, 1995). Typical
impacts of earthquakes on landfills included: cracking of interim cover soils,
some downslope movement of the cover soils, and disruption of gas collection systems. No observations are available for landfills with geosynthetic
final caps. However, significant displacements could occur along interfaces
between geosynthetics and other geosynthetics or soils.
The state of practice for evaluating the stability and deformation of
landfills under seismic loading has recently been summarized by Anderson
and Kavazanjian (1995). As described in their paper, the seismic design of
caps involves four steps:
characterization of the ground motions for the design earthquake,

evaluation of the response of the landfill to the design earthquake,
calculation of the stability and deformation of the entire waste mass,
and
evaluation of the ability of the cap to maintain its integrity when subjected to the calculated ground motions.
An important consideration relative to seismic loading is the potential

for the waste mass to amplify free-field ground accelerations. Amplification
is known to occur in soil deposits and has been observed at the OII Landfill
near Los Angeles, the one landfill at which instrument measurements are
available. A threefold amplification of peak ground acceleration from the base
of the landfill to the crest was recorded during the 1992 Landers earthquake
at the landfill. Amplification can be predicted analytically if the design ground
motions and the dynamic properties of the waste are known. However, in
the central and eastern U.S., there is considerable uncertainty about the nature of the design motions. There is also considerable uncertainty about the
dynamic properties of wastes, especially municipal solid waste. Factors in-
128
CAPS
fluencing seismic resistance of the landfill cover are the yield acceleration of
the cover system and the allowable seismic deformations. Yield acceleration
depends on the shear strength of the considered failure mass. For geosynthetic
materials, it is not clear when peak, residual, or deformation-compatible
strengths should be used. Additionally, most interface strength tests involving geosynthetics are run at displacement rates well below the rates occurring during seismic loading. The interface shear strength may be higher or
lower at the actual displacement rate caused by the earthquake, depending
on the material used and on moisture conditions. It is extremely difficult to
design an unconditionally stable cap using geosynthetic materials in areas of
high seismicity. Even in areas of low to moderate seismicity, unconditional
stability may be difficult to obtain for covers containing geosynthetics that
are steeper than 3H:1V. Allowable seismic deformations of the cap are based
on practical considerations rather than rigorous analysis. For noncritical caps,
seismic deformation of the cover may be handled as a maintenance issue.
6.2.7 Water Balance Modeling
One of the principal functions of a cap is to limit percolation of water into
underlying contaminated materials. The design of the cap to limit percolation of water through it requires a model that can predict water percolation
through caps. Such models are called water balance models. Because water balance models are essential to performance-based design, they received
considerable attention and discussion at the workshop.
Several computer models are available to evaluate the hydraulic performance of landfill caps. The most widely used model is the USEPA Hydrologic
Evaluation of Landfill Performance (HELP) model (Schroeder et al., 1994a,b).
The HELP model has the advantage over other models in that it contains
default climatic and material properties data, is relatively easy to use, and is
accepted by the regulatory community. The HELP model, however, contains
a number of simplifying assumptions that make the model inappropriate in
certain cases. For instance, the HELP model, which generally assumes a unit
gradient to model vertical drainage, cannot be used to simulate the performance of partially saturated layers in very dry environments. Another model
that has been used to model caps is UNSAT-H. This model, which was developed by Pacific Northwest Laboratory by Fayer and Jones (1990), does not
contain the default data of the HELP model and requires extensive computer
time to conduct simulations. However, UNSAT-H performs a much more
rigorous analysis of unsaturated flow by solving the relevant partial differential equation, rather than assuming unit gradient as the HELP model does. A
limited number of field studies and analytical assessments have been performed to evaluate the reliability of models as tools to predict trends and
magnitudes of the different landfill water balance components. However,
the findings from these studies are not in general agreement. For example,
some of the studies found that the models overpredicted percolation in hu-
129
mid climates and underpredicted percolation in arid climates, while other studies concluded the opposite. In many cases, the models were unable to predict
short-term trends. However, for a number of cases they appeared to give reasonable predictions of cumulative water balances. The developers of the HELP
model report that the approximate annual errors in water balance components
calculated by the HELP model are:
runoff = 25 percent of actual value or 2 percent of precipitation,

evapotranspiration = 10 percent of actual value or 6 percent of
precipitation,
percolation = 10 percent of actual value or 0.3 percent of precipitation, and
lateral drainage = 7 percent of precipitation.
In a study of field-scale caps [30 m x 30 m (100 ft by 100 ft)] in Atlanta,

Georgia and East Wenatchee, Washington, incorporating compacted clay barrier layers, the measured water balances of the caps were compared to the
simulated water balances using the HELP and UNSAT-H models (Khire,
1995). The caps were constructed with two layers: (i) a 60- or 90-cm (2- or 3ft) thick compacted clay barrier layer; and (ii) a 15-cm (0.5-ft) thick vegetated
surface layer. Collection of climate, runoff, soil water content, and percolation data for each cap began in 1992 and is ongoing. Runoff and percolation
are collected in tanks and measured, while soil moisture is measured using
time domain reflectrometry. The cumulative percolations measured for the
caps in Georgia and Washington were about 24 cm (9.4 in.) and 3.1 cm (1.2
in.), respectively. Khire found that for the Georgia cap, HELP overpredicted
percolation by about 320 percent and and UNSAT-H underpredicted percolation by about 24 percent. The relatively large error in percolation predicted
by the HELP model was attributed primarily to the underestimation of runoff. Both models overpredicted percolation by about 43 percent for the cap in
Washington. This deviation was believed to have been caused by preferential flow through cracks and animal borrows observed in the barrier layer.
Khire also found that while both models captured seasonal variations in runoff, evapotranspiration, soil water storage, and percolation, UNSAT-H simulated the variations more accurately.

Some cap failures and findings from other field studies being conducted in
the U.S. and Europe are described in this section.
6.3.1 Cap Failures
There have been a number of documented cases of cap failures, with most of
130
CAPS
the failures occurring during or shortly after construction and resulting from
excessive erosion, build up of pore water pressures in the cap layers, lack of a
drainage layer, a drainage layer with insufficient capacity, or incorrect estimation of the shear strength between the cap layers (Boschuk, 1991). While
most of these failures did not involve rupture of the barrier layer, they were
costly to repair. In one recent case, severe erosion problems developed because the cap slopes were relatively long (180 m (600 ft)) and the cap drainage layer was designed with an outlet only at the toe. At some locations, the
cap had eroded to the top of the clay barrier layer. The erosion problem was
exacerbated in some cases because the drainage layer outlet at the toe of the
slope had not been constructed. In these cases, the trapped water eventually
caused pore pressures to become excessive, causing sloughing of the overlying soil layers at the toe of the slope. At another landfill, a gabion-lined channel for surface water slid down the cap slope due to liquifaction of the fine
sand beneath it brought about by high pore pressures. It was noted that most
of the failures occurred in states with relatively restrictive, prescriptive cap
designs rather than more flexible performance objectives, suggesting that, in
these states with prescriptive cap designs, greater attention may have been
given to regulatory compliance than to the design itself.
6.3.2 Field Study in Hamburg, Germany
Six test caps [10 m (33 ft) wide x 50 m (160 ft) long] were constructed in 1987
to evaluate the field performance of different cap configurations (Melchior
and Miehlich, 1989; Melchior et al., 1994). The caps were constructed with a
0.75-m (2.5-ft) thick sandy loam topsoil layer, underlain by a 0.25-m (0.82-ft)
thick fine gravel drainage layer. The drainage layer was underlain by one of
four barrier layer types: 1) a 0.60-m (2.0-ft) thick compacted clay layer, 2) a
HDPE geomembrane/clay composite layer with welded geomembrane panels, 3) a geomembrane/clay composite layer with overlapped geomembrane
panels, and 4) a compacted clay layer overlying a 0.60-m (2.0-ft) thick fine
sand wicking layer and a 0.25-m (0.82-ft) thick coarse sand/fine gravel capillary barrier.
Each of the four cap configurations was constructed on 4 percent or 20
percent slopes, and several configurations were constructed for both slopes.
Climate, lateral drainages from the topsoil and drainage layers, runoff, and
percolation data are being collected. Soil moisture data are also being collected from several test caps using neutron probes and tensiometers. The
preliminary findings of this field study are summarized below.
For the caps with the compacted clay barrier layer, little percolation was
observed for the first 20 months after construction. Beginning in August 1989,
percolation began to increase and show a correlation with precipitation events.
The summer of 1992 was very dry, and tensiometers indicated that the clay
layers had undergone more drying than usual. This drying resulting in an
almost tenfold increase in percolation measured during the fall of 1992 over
131
percolation recorded a year earlier. Flow through the capillary barrier under
one of the compacted clay layers was first observed during this time. When
excavations were made into the caps in 1993, the clay layers were found to
have small fissures and contain plant roots. Since 1993, the network of plant
roots has developed further, contributing to preferential flow paths and desiccation cracking. Percolation through the compacted clay layers is still increasing and was about 200 mm (8 in.) in 1994.
The caps constructed with a composite barrier layer performed much better, and no percolation has been observed. However, during the summer and
fall months, the matric potential in the clay layers increased and drainage from
the clay layer was recorded. This drainage, which had typically been less than
1 mm/yr (0.04 in./yr), has been attributed to thermal gradients. During the
summer and early fall, the temperature at the top of the clay layer has been
greater than that at the bottom of the clay layer, and water likely flows in liquid
and vapor phase from the hotter to cooler regions resulting in the measured
drainage. The water loss caused by thermal gradients has not caused shrinkage of the clay, although the potential for future shrinkage exists.
6.3.3 Field Study in Beltsville, Maryland
Six lysimeters [14 m (45 ft) wide x 21 m (70 ft) long x 3.0 m (10 ft) deep] with
20 percent side slopes were constructed between May 1987 and January 1990
to evaluate caps incorporating either a compacted clay barrier layer, a rock
capillary barrier, or bioengineering management, which combines enhanced
runoff and plant transpiration (Schultz et al., 1995). The bioengineering management option used alternating aluminum and fiberglass panels as the surface layer over about 90 percent of the cap with moisture-stressed vegetation
(i.e., Pfitzer junipers) located along gaps in the panels. This latter option
requires periodic maintenance and is intended to be used when significant
subsidence of the underlying waste is expected. The six lysimeters were constructed with the following caps: (1) bioengineering management, with the
initial water level 90 cm (35 in.) above the bottom of the lysimeter; (2) bioengineering management, with the initial water level 190 cm (75 in.) above the
bottom of the lysimeter; (3) reference lysimeter similar to lysimeter 1, except
without the surface panels and vegetated with fescue grass; (4) rip-rap surface layer and gravel drainage layer over a compacted clay layer; (5) vegetated soil surface layer, gravel drainage layer, and compacted clay layer over
a gravel capillary barrier; and (6) vegetated soil surface layer and gravel drainage layer over a compacted clay layer. All of the caps were constructed over
native soil. Rainfall, runoff, deep drainage, and soil moisture content data
were collected for the lysimeters.
The data collected through 1994 reveal that initially ponded water in
lysimeters 1 and 2 was removed by the plants within two years after construction. The soils in these lysimeters have generally become drier over time.
The initial water level in lysimeter 3 rose until it was near the surface of the
132
CAPS
lysimeter and the water had to be pumped out. Deep drainage has been
measured for this lysimeter every year of the study. Except for the deep drainage from lysimeter 5, which occurred during 1994, deep drainage has not
been observed from lysimeters 4 to 6. It has been noted that the moisture
content of the clay layer in lysimeter 4 has been increasing, indicating the
possibility of future seepage through the clay. The moisture contents of the
clay layers in lysimeters 4 and 6 show some seasonal cycling, with the lowest
moisture contents being measured in the summer.
6.3.4 Field Study in East Wenatchee, Washington
Two caps [30 m x 30 m (100 ft x 100 ft)], one with a 60-cm (2-ft) thick compacted
clay barrier layer overlain by a 0.15-m (0.5-ft) thick vegetated soil surface layer
and the other with a 75-cm (2.5-ft) thick sand capillary barrier layer used in lieu
of a clay resistive barrier, were constructed and monitored (Khire, 1995). Climate, runoff, percolation, and soil moisture data have been continuously collected since November 1992. The collected data show that cumulative
percolations through the resistive and capillary barrier layers have been 3.1
and 0.5 cm (1.2 and 0.2 in.), respectively. Most of the water movement through
the capillary barrier occurred in winter 1993, primarily due to snowmelt from
the relatively high snowfall (169 cm (66.5 in.)) occurring that year.
6.3.5 Field Studies in Richland, Washington
Since 1985, PNL and Westinghouse Hanford Co. have been working to develop a cap design for the Hanford site. Field tests have been conducted for
over the past seven years using lysimeters to evaluate the performance of
different cap materials and configurations (Petersen et al., 1995). Currently,
24 lysimeters are being monitored to assess the effects of varying precipitation, surface soil, and vegetative conditions. No drainage has been measured
from lysimeters with vegetated or nonvegetated silt-loam surfaces under
normal precipitation conditions; however, some drainage has occurred from
lysimeters with nonvegetated silt-loam surfaces under extreme precipitation
conditions (i.e., three times normal). Significant quantities of water have
drained from lysimeters with gravel and sand surface layers. The performance of one lysimeter with a 1.5-m (5-ft) thick layer of nonvegetated silty
loam was modeled over a six-year period using the HELP and UNSAT-H
models. The HELP model simulation prediction was 1800 percent greater
than the observed drainage, while the UNSAT-H model simulation prediction was at 52 percent of the observed drainage.
A prototype cap [0.6 hectare (1.5 acre)] was also constructed at the site in
1994 using the following layers, from top to bottom (Gee et al., 1994; Wing
and Gee, 1994; Peterson et al., 1995):
1.0-m (3-ft) thick silt loam/admix gravel surface layer,
133
1.0-m (3-ft) thick silt loam protection layer,

0.15-m (0.5-ft) thick sand filter,
0.30-m (1-ft) thick gravel filter,
1.5-m (5-ft) thick fractured basalt riprap capillary barrier and
biobarrier,
0.3-m (1-ft) thick gravel cushion and drainage layer,
0.15-m (0.5-ft) thick composite asphalt barrier layer,
0.10-m (0.3-ft) thick top course, and
compacted soil foundation.
The composite asphalt barrier layer was discussed previously. The water and wind erosion, biointrusion, revegetation success, and water balance
of the cap continue to be monitored. Water balance components being recorded include: precipitation, runoff, snow depth, soil moisture, and percolation. The water balance is being evaluated under normal and stressed (i.e.,
irrigated) conditions.
6.3.6 Field Study in Idaho Falls, Idaho
A replicate field test program is underway at Idaho National Engineering
Laboratories to compare the hydraulic performance of four caps designs: (1)
1-m (3-ft) thick vegetated soil layer over a geomembrane/ compacted clay
composite barrier layer; (2) 2.5-m (8.2-ft) thick vegetated soil layer with a 0.5m (1.6-ft) thick biobarrier located within it at 0.5 m (1.6 ft) below the ground
surface; (3) the same design as cap 2, except the biobarrier is located 1 m (3 ft)
below the ground surface; and (4) 2.0-m (6.6-ft) thick vegetated soil layer.
The test plots were constructed in 1993. Two vegetation types have been
used, a native mixed plant community and a monoculture of crested
wheatgrass. Both vegetative covers were considered since planted
monocultures may be reinvaded by native plant species in the future, and a
mixed native plant community may be more resilient to environmental fluctuations. The test plots will, at times, be subjected to burrowing animals,
ants, and high levels of irrigation. Climate, soil moisture, and percolation
data are being collected. Soil moisture is being measured using a neutron
probe and time domain reflectometry. No data are available.
6.3.7 Field Studies in Albuquerque, New Mexico
A large-scale field test program is being conducted at Sandia National Laboratories to compare the performance of three caps [each 13 m (43 ft) wide and
100 m (330 ft) long]; one incorporates a compacted clay, another a
geomembrane/clay composite, and a third has a geomembrane/GCL composite barrier layer. The test caps were constructed and instrumented during
1995 (Dwyer, 1995). The hydrology and the erosion of the caps are being
monitored. Another three caps are scheduled to be constructed in 1996, each
134
CAPS
having capillary barriers with one vegetated to enhance evapotranspiration.

No data on performance had been collected at the time of the workshop. Monitoring is expected to continue for the next several years.
In another field study under limited evapotranspiration conditions, the
performances of two gravel capillary barriers were compared (Stormont, 1995).
When a uniformly-graded sand layer was placed between the gravel barrier
and the overlying silty sand, the lateral drainage above the barrier increased
while drainage through the gravel decreased.

Several aspects of cap materials and design, namely the service life of HDPE
geomembranes, the internal shear strength of GCLs, the properties of asphalt
and paper mill waste, the slope stability of caps, and water balance modeling,
have been reviewed. The findings from this review are presented below:
Preliminary results from long-term laboratory testing of HDPE

geomembranes indicate that service life of the geomembranes should be
at least several hundred years, depending on the specific product and
the in-place conditions. If HDPE geomembranes provide this length of
service, they appear to be a good investment relative to other more costly
barrier layer materials such as compacted clay.
Based on ongoing field-scale tests of prototype caps incorporating GCLs,
the factor of safety against mid-plane failure of GCLs on 3H:1V slopes
has been observed to be 1.5 or greater. These results are encouraging
since a cap with a GCL is a more effective barrier against infiltration than
a cap with a compacted clay layer. If GCLs remain stable on cap slopes,
they appear to be preferable to a compacted clay layer.
Asphalt and paper mill waste have been used as the barrier layer material
for caps. However, the service life of these materials is uncertain. While
asphalt may have a longer service life than traditional barrier layer
materials, it is significantly more expensive. Also, the significance of flow
along asphalt seams is unknown. Paper mill waste involves reuse of a
waste product; however, it may be affected by some of the same processes
(e.g., freezing/thawing cycles) that adversely impact compacted clays.
The shear strength properties of cap components should be evaluated by
conducting project-specific tests under the expected field conditions. The
effect of freezing-thawing cycles, heating-cooling cycles, and creep on
the long term shear strength properties should also be considered by using
a higher safety factor or by increasing the thickness of cover soils above
the critical layers to provide thermal insulation and isolation from the
environment.
Observations of the seismic performance of several landfills have indicated
that, while landfills have performed relatively well, cracking and some
135
downslope movement of interim cover soils and disruption of gas collection

systems occurred. The effect of the seismic motions on the integrity of caps
incorporating compacted clay layers or geosynthetics has not been well
documented. However, there is concern that compacted clay layers could
be prone to cracking and that displacements may occur along interfaces
between two geosynthetics and between a geosynthetic material and soil.
There is considerable uncertainty about the dynamic properties of wastes,
especially unconsolidated municipal solid waste, and about the appropriate
shear strength values to use for geosynthetic materials under dynamic
loading.
Allowable deformations of caps under seismic conditions are based on
practical considerations rather than rigorous analysis. For non-critical caps,
seismic deformation of caps may be handled as a maintenance issue.
Water balance models are used to predict the performance of caps in terms
of water percolation through the cap. Available water balance models that
are applied to caps typically contain numerous simplifying assumptions
and have been inadequately verified by field data. Since these models
provide wide-ranging estimates of water infiltration under site-specific
conditions, their main value may be to compare alternative designs utilizing
different cap configurations and materials.
Based on the field monitoring of test caps and cap failures, the following
remarks are made regarding cap performance:
Several examples were given of inadequately protected compacted clay

barrier layers that degraded after a few years as a result of desiccation, root
penetrations, or both. Although covering a clay layer with a geomembrane
provided greatly improved protection, one case suggested that thermally
induced flow could eventually desiccate even a geomembrane-covered clay
barrier layer.
There are few data on the performance of caps incorporating geomembranes
and compacted clay layers and even fewer data for caps that include
capillary barriers or employ surface vegetation to enhance
evapotranspiration. Therefore, we are unable to document how well
existing caps are performing.
The performance life of caps has not been established. Although the service
life of some components of the system can be estimated, the functional life
of the surface layer and barrier layer are not well documented and the long
term performance of constructed cap systems has not been adequately
documented.
A number of cases of cap failures have been documented; however, most
of these failures could have been prevented through proper design and
construction. It is felt that some failures may have resulted from
preoccupation with regulatory compliance rather than engineering design
considerations.
136
CAPS
The primary factors adversely affecting cap performance for each of the
six basic layer components are summarized in Table 6-2.
TABLE 6-2 Factors Affecting Cap Performance
Layer
Factor
Surface Layer
Protection Layer
Erosion
Slope Failure Due to Pore Pressure Buildup
Animal Burrows
Drainage Layer
Clogging
Insufficient Capacity
Insufficient Drainage Layer Outlets
Barrier Layer
Cracking due to Desiccation, Deformations From

Waste Settlement, or Seismic Motions (Clay, Paper
Mill Sludge)
Root Penetration
Resistance to Gas Migration (GCLs)
Stability
Creep (asphalt)
Service Life
Gas Collection Layer
Adequate Cover Over Waste
Foundation Layer
Adequate Strength
Erosion
Evapotranspiration
Native versus Exotic Vegetation
Appropriate Armoring for Side Slopes at Arid Sites
6.5 NEEDS
The following needs related to cap technology are presented based on the
assessment of caps presented in this section.
Data are available that demonstrate that the performance of compacted

clay barrier layers in caps will deteriorate over time. Even so, compacted
clay barrier layers are still being used in caps, primarily because they are
specified in regulations. There is a perception that it may be difficult to
obtain regulatory approval to use alternative barrier materials to a clay
layer. This situation could be improved if guidelines were available for
demonstrating the equivalency of performance among the different
options for cap components.
Few data are available concerning the hydraulic performance of
traditional caps with resistive barriers. There are even fewer performance
137
data for caps containing capillary barriers. More data need to be collected
to assess cap performance. Data are especially needed to bring about
regulatory and community acceptance of alternative cap configurations.
While some of these data can be collected from currently instrumented
sites, other additional sites will probably need to be monitored.
The expected performance life of caps is uncertain. Studies are underway
to assess the service life of some individual cap components. However,
the long term service life of these components in a constructed surface
barrier system has not been adequately studied.
There have been a number of documented cases of cap failures; however,
most of these failures could have been avoided through proper design
and construction. There is a need for more guidance on cap design and
for independent peer review of completed designs prior to construction.
Compliance with regulations is not a sufficient check on a completed
design.
More field observations on the effect of seismic motions on the integrity
of caps incorporating compacted clay layers and/or geosynthetics need
to be made.
The shear strength at interfaces between materials is known to be an
important factor affecting the physical stability of caps on slopes. The
shear strength is also affected by freezing-thawing cycles, heating-cooling
cycles, and creep. Standard procedures for evaluating interfacial strength
need to be developed.
More information is needed on the dynamic properties of wastes, and on
the appropriate dynamic shear strength values to be used for geosynthetic
layer materials.
Other alternative barrier layer materials, such as asphalt and paper mill
waste, appear promising for future use. However, more information needs
to be collected on the long-term performance of these materials.
Available computer models for simulating the hydrologic and hydraulic
performance of caps need to be verified by comparison with field data,
and modified as necessary.
The purpose of this section was to define the current technological status of
waste containment caps, including: material performance, design methods,
field performance, and field monitoring. With respect to material performance and design method, the following issues were considered: service life of
HDPE geomembranes; internal shear strength of GCLs; properties of asphalt
and paper mill sludge; slope stability of caps; and water balance modeling.
Cap failure cases and study results from field monitoring at sites in arid and
temperature climates have been reviewed. Based on these considerations and
review, it is concluded that:
138
CAPS
The service life of caps is uncertain. The performance of a compacted

clay barrier layer in a cap will likely deteriorate over time, while an HDPE
geomembrane barrier layer may perform satisfactorily for several hundred
years. The long-term performance of cap components, and caps as
systems, needs to be studied further.
The slope stability aspects of caps are complex. The shear strength
properties of cap components and component interfaces may be impacted
by moisture conditions, stress, creep, temperature, and seismic motions.
However, there are no standard procedures for evaluating interfacial
strength or accounting for these effects. In addition, the dynamic
properties of buried wastes are generally not well understood. Under
certain conditions, the waste may amplify seismic-caused ground
movements which can be transmitted to the cap. Standard procedures
need to be developed for evaluating the shear strength of cap materials
and additional research needs to be performed on the seismic properties
of various types of buried wastes.
There are few published data on the field performance of constructed
cap systems. More data need to be collected. The collected hydraulic
data can also be used to verify currently used models for simulating the
hydraulic performance of caps.
Over emphasis on regulatory compliance inhibits innovative and creative
cap design, particularly with regard to the selection of alternative materials
for cap components. Greater emphasis needs to placed on how the design
will affect cap performance. Technical guidelines for demonstrating
alternative cap equivalency need to be developed.
There have been a number of documented cases of cap failures; however,
most of these failures could have been avoided by proper design and
construction. This problem may be minimized by independent peer
review of design and more rigorous QA/QC during construction.
Nontraditional materials, such as asphalt, paper mill sludge, capillary
drains, capillary barriers, plants with high tranpiration capability, and
GCLs have undergone limited testing in the field but look promising for
future use in caps.
6.7 REFERENCES
Anderson, D.G. and Kavazanjian, Jr., E. (1995). Performance of Landfills
Under Seismic Loading. Proceedings of the Third International Conference
on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics,
University of Missouri, Rolla, MO.
Boschuk, Jr., J. (1991). Landfill Covers An Engineering Perspective.
Geotechnical Fabrics Report, 9(4), pp. 23-34.
6.3 REFERENCES
139
Daniel, D.E. and Koerner, R.M. (1993). Cover Systems. Geotechnical Practice
for Waste Disposal, D.E. Daniel, ed., Chapman & Hall, pp 455-496.
Dwyer, S.F. (1995). Alternative Landfill Cover Demonstration. Landfill
Closures Environmental Protection and Land Recovery, Geotechnical Special
Publication No. 53, R.J. Dunn and U.P. Singh, Eds., ASCE, pp. 19 - 34.
Fayer, M. and Jones. T. (1990). Unsaturated Soil-Water and Heat Flow Model,
Ver. 2.0. Pacific Northwest Laboratory, Richland, Washington.
Freeman, H.D., Romine, R.A., Zacher, A.H. (1994). Hanford Permanent
Isolation Barrier Program: Asphalt Technology Data and Status Report FY 1994. PNL-10194, Pacific Northwest Laboratory, Richland, Washington,
37 p.
Gee G.W., Freeman, H.D., Walters, Jr., W.H., Ligotkr, M.W., Campbell, M.D.,
Ward, A.L., Link, S.O., Smith, S.K., Gilmore, B.G., and Romine, R.A.(1994).
Hanford Prototype Surface Barrier Status Report: FY 1994. PNL-10275,
Pacific Northwest Laboratory, Richland, Washington.
Khire, M.V. (1995) Field Hydrology and Water Balance Modeling of Earthen
Final Covers for Waste Containment. Environmental Geotechnics Report
No. 95-5, University of Wisconsin - Madison, 166 p.
Koerner, R. M., Lord, A. E., and Hsuan, Y. H. (1992). Arrhenius Modeling to
Predict Geosynthetic Degradation, Journal of Geotextiles and Geomembranes, Vol. 11, pp. 151 - 183.
Limbach, W.E., Ratzlaff, T.D., Anderson, J.E., Reynolds, T.D., and Laundr, J.W.
(1994) Design and Implementation of the Protective Cap/Biobarrier
Experiment at the Idaho National Engineering Laboratory. In-Situ
Remediation: Scientific Basis for Current and Future Technologies, G.W.
Gee and N.R. Wing, eds., Battelle Press, pp. 359-377.
Melchior, S. and Miehlich, G. (1989). Field Studies on the Hydrological
Performance of Multilayered Landfill Caps. Proceedings of the Third
InternationalConference on New Frontiers for Hazardous Waste Management,
EPA/600/9-89/072, U.S. EPA, pp. 100-107.
Melchior, S., Berger, K., Vielhaber, B., and Miehlich, G. (1994). Multilayered
Landfill Covers: Field Data on the Water Balance and Liner Performance.
In-Situ Remediation: Scientific Basis for Current and Future Technologies,
G.W. Gee and N.R. Wing, eds., Battelle Press, pp. 411-425.
Moo-Young, H.K. and Zimmie, T.F. (1995). Design of Landfill Covers Using
140
CAPS
Paper Mill Sludges. Proceedings of Research Transformed into Practice

Implementation of NSF Research, J. Colville and A.M. Amde, eds., ASCE,
NY. pp. 16-28.
Othman, M.A., Bonaparte, R., Gross, B.A., and Schmertmenn, G.R. (1995).
Design of MSW Landfill Final Cover Systems. Landfill Closures
Environmental Protection and Land Recovery, Geotechnical Special Publication
No. 53, R.J. Dunn and U.P. Singh, Eds., ASCE, NY, pp. 218-257.
Petersen, K.L., Link, S.O., and Gee, G.W., (1995). Hanford Site Long-Term
Surface Barrier Development Program: Fiscal Year 1994 Highlights. PNL10605, Pacific Northwest Laboratory, Richland, Washington.
Rumer, R. R. and Ryan, M. E., eds. (1995). Barrier Containment Technologies
for Environmental Remediation Applications. John Wiley & Sons, Inc.
NY, 170 p.
Schroeder, P.R., Lloyd, C.M., and Zappi, P.A. (1994a). The Hydrologic
Evaluation of Landfill Performance (HELP) Model Users Guide for
Version 3. U.S. Environmental Protection Agency, Office of Research and
Development, Washington, D.C., EPA/600/R-94/168a.
Schroeder. P.R., Dozier, T.S., Zappi, P.A., McEnroe, B.M., Sjostrom, J.W., and
Peyton, R.L. (1994b). The Hydrologic Evaluation of Landfill Performance
(HELP) Model Engineering Documentation for Version 3. U.S.
Environmental Protection Agency, Office of Research and Development,
Washington, D.C., EPA/600 R-94/168b, 116 p.
Schultz, R.K., Ridky, R.W. and ODonnell, E. (1995). Control of Water
Infiltration into Near Surface LLW Disposal Units Progress Report of Field
Experiments at a Humid Region Site, Beltsville, Maryland. U.S. Nuclear
Regulatory Commission, NUREG/CR4918 Vol. 8, 20 p.
Stormont, J.C. (1995). The Performance of Two Capillary Barriers During
Constant Infiltration. Landfill Closures Environmental Protection and Land
Recovery, Geotechnical Special Publication No. 53, R.J. Dunn and U.P.
Singh, Eds., ASCE, NY, pp. 77 - 92.
Wing, N.R. and Gee, G.W. (1994). Quest for the Perfect Cap. Civil
Engineering. ASCE, Oct., pp.38-41.
141
SECTION 7
FLOORS AND BOTTOM BARRIERS: INDIGENOUS
prepared by
Gordon P. Boutwell, Soil Testing Engineers, Inc.

and Tomasz Hueckel, Duke University
Mark D. Ankeny, D.B. Stephens Assoc.

Robert M. Cohen, GeoTrans Inc.
Michael T. Dukes, RUST Environmental
Douglas J. Hermann, STS Consultants Ltd.
Beverly L. Herzog, Illinois Geological Survey
Lawrence D. McKay, University of Tennessee
Terry D. Vandell, Conoco, Inc.
7.1 OVERVIEW
Natural subsurface barriers can be utilized as the floor of a waste containment
system. However, the effective utilization of a natural or indigenous barrier
depends on a good understanding of the overall site characteristics, including
the integrity of the indigenous formation and how it can be integrated into an
engineered containment facility. The objective is to eliminate or limit
contaminant flux through the natural and constructed barriers that form the
containment structure.
Generally, engineered containment barriers are designed, constructed, and
then subjected to quality assurance (QA). Natural barriers, on the other hand,
simply exist. Their use as part of a containment system depends on a
determination of their material and physical properties, which may vary from
point-to-point within an otherwise apparently consistent stratigraphic unit.
For European perspectives on this topic, see DalPra and Magnini (1993) and
Dorhofer (1993).
Writing on the art of war in 2500 BC, the Chinese general Sun Tsu stated
that the first requirement of a successful campaign was to know the enemy.
Similarly, the designer of a containment system must know the site. Evaluation
of a specific site depends on careful exploration, guided by a thorough
understanding of transport mechanisms under saturated and partially
saturated conditions. This chapter on indigenous (or naturally occurring)
barriers addresses most of the key issues that must be considered in the
decision to utilize an existing indigenous formation as an integral component
of a containment system. These include:
FLOORS AND BOTTOM BARRIERS-INDIGENOUS
142
characterization of the site,

determination of the natural barrier transport properties, and
evaluation of potential barrier property changes resulting from
exposure to the contaminant and other environmental changes.
7.2 DETERMINING STRATIGRAPHY

7.2.1
Background
Stratigraphic studies should be directed towards finding and evaluating the

conduits, or preferential migration paths, in natural formations that are under
consideration as components of a containment system. Stratigraphy evaluation
is, therefore, ultimately aimed at a determination of the distribution of
hydraulic conductivity in the proposed formation.
A well-executed program to determine site stratigraphy consists of several
elements, including: (1) a review of available literature to understand the
regional and general site stratigraphy; (2) a definition of the objectives of the
stratigraphic characterization (i.e., what sorts of decision making will be made
based on the stratigraphic interpretation); (3) the development of a conceptual
model of the site stratigraphy; and (4) a field program consisting of one or
more exploration methods to generate stratigraphic data for testing and
refining the stratigraphic model. The preferred procedure is for the field
program to be phased, with each phase of testing leading to further refinement
of the evolving stratigraphic model.
7.2.2
State of Practice in Determining Stratigraphy
a. General Guidelines. In most cases, multiple methods should be used in

stratigraphic determinations. This provides greater confidence in the final
stratigraphic model. It is also essential that each phase of the investigation
have well-defined objectives. These objectives should be made clear to all
personnel, including the drillers and instrument technicians involved with
the investigation. Furthermore, all personnel should be qualified and
experienced in their portion of the field work, or at least directly supervised in
the field by experienced persons.
b. Intrusive Methods. Most intrusive methods have been used long enough
so that the geoscience professions have confidence in the procedures and results.
However, intrusive methods produce penetrations/perforations into or through
potential barrier strata which must then be sealed. The primary sealing method
is to grout from bottom-to-top with a cement/bentonite or high-solids bentonite
grout. Field data are available from sites where monitoring wells were doublecased through contaminated zones into uncontaminated zones, indicating that
sealing is effective under even severe conditions.
143
Soil Borings. This well-established technology is still the ultimate geologic

tool (the relevant ASTM standards for many of the procedures include D1452,
D1586, D1587, D2113). Although good drillers and loggers are essential,
they must be fully aware of the investigations objectives. Sampling boreholes
in the normal manner can be slow, allowing time for small amounts of
contaminants to transfer from one stratum to another. The amounts
transferred, although small, may affect groundwater chemical analysis
results. Therefore, multiple-use boreholes (e.g., used also as monitoring
wells) are to be avoided where possible. Continuous sampling is needed,
especially through critical zones containing discontinuities such as lenses
of non-matrix material or fractures. Stress relief is a problem in laboratory
testing of borehole samples. In highly overconsolidated deposits, such as
shales, stress relief effects can invalidate strength or hydraulic conductivity
test results.
Test Pits. The advantage of a test pit is that it provides direct observation of a
larger-scale cross-section of the formation. Structural bracing may be required
for safety, or even special protective apparatus may be needed in some cases.
Cone Penetrometer Testing (CPT). While CPT technology is mature and
accepted (ASTM D3441), new downhole sensing devices are being developed
rapidly. The capabilities of the CPT system are limited by the normal rod
sizes; only a certain amount of sensor cable can be used. An advantage of
CPT (or other push technologies) is that few, if any, cuttings (which can be
expensive to dispose) are produced. Disadvantages include the potential for
carrying contaminants down during the advancement of the cone and the
need for proper grouting. Downhole bottom-to-top grouting without prior
rod removal is recommended for contaminated areas, especially.
CPT has been used to delineate ionic and radioactive contaminant plumes.
The methodology is ideal for downhole ultrasonics because of the intimate
contact of soil and sensor. It is good for DNAPL interface sampling. However,
if exploration is to proceed below a DNAPL or other contaminated zone, it is
necessary to case off the contaminated zone. Thus, deeper exploration requires
casing or nested rod sizes. Further discussion of current trends in CPT
Technology is given in Shinn, et. al., 1995.
Downhole Geophysics. In most cases, geophysical methods can supplement
but not replace detailed intrusive sampling. Although most of the downhole
geophysical methods are well established in geophysics, they have been
sparingly used on remedial site investigations. The techniques and equipment,
originally developed using analog data reduction, need upgrading for digital
computation. In general, gamma logs provide better data than neutron logs.
Shear-wave and ultrasonic logging are good for finding discontinuities such
as fractures. Ultrasonic logging requires a borehole caliper log for accuracy.
These units can be pulled through a slanted or horizontal hole, but at a higher
144
risk of losing the (expensive) equipment because of hole collapse. Crosshole techniques improve the resolution of downhole geophysics. The ideal
situation would involve an integrated sensing network, but this is not currently
done. Other useful downhole techniques include temperature gauges, realtime groundwater pressure transducers, and flowmeters for detecting both
vertical and horizontal flows.
c. Non-Intrusive Geophysics. Currently, surface geophysical techniques are
capable of delineating only gross features. They are useful for evaluating
continuity between intrusive sampling points. Various geophysical methods
that apply to environmental work are described by Benson and Yuhr, 1995.
Seismic Reflection and Refraction. This technique has been successful in the
upper units of the clay-sand geology in the Gulf Coast area. However, it
should be considered qualitative, not quantitative. Data interpretation is
currently quite subjective.
Electromagnetic Induction. This methodology has been successful in
delineating strongly differing strata, such as clays, sands, and limestones. It
has poor resolution for minor subsurface feature changes, but can locate
windows in barrier strata if they are large enough and at a depth that is
relatively shallow compared with their lateral extent. The method has been
used with good success as a tool for interpolating between intrusive sampling
points. When this method is combined with regression analysis on the hard
data points, depths to strata changes can be determined to within about
10% to 20% (Lawrence and Boutwell, 1990). Electromagnetic induction
has been used to locate fractures in limestones. It is fast, and large areas can
be surveyed relatively quickly.
Electrical Resistivity. The statements made above for electromagnetic
induction also apply for electrical resistivity. However, the electrical resistivity
method is generally much slower unless specialized equipment is used.
Ground-Penetrating Radar. This method has been very successful in sands
above the groundwater table. However, heavy clays and groundwater reflect
the signal, preventing deeper penetration. Where applicable, it is excellent
for delineating irregularities such as faults, fracture zones, etc.
d. Interpretive/Presentation/Predictive Methods. In every subsurface
exploration, the first step should be development of a simple conceptual model
of the subsurface conditions based on available data from geologic
publications, previous exploration work, and observation of strata exposed
in existing cuts, etc. A few simple cross-section profiles may suffice to construct
the first version of the model. This model should be continually reviewed
and updated as more information becomes available.
145
At many remediation sites, investigations may extend over many years

and have been performed with varying objectives and/or with different levels
of expertise. This frequently leads to data inconsistencies, especially
concerning the minor features that govern contaminant transport. Eventually,
the amount of data becomes so great that a computerized data management
system is required for efficient interpretation. Unfortunately, there can be a
tendency to be overly dependent on the interpretative capabilities of the
available computer software, with too little independent thinking by trained
persons. Still, computer techniques are useful in the relatively rapid
production of visual presentations, such as isopach maps, cross-sections, and
isometric profiles.
Experience with geostatistical methods, such as kriging, although mixed,
has been generally poor. The levels of confidence achieved can easily be an
artifact of some necessary assumption made in the analyses.
7.2.3
Field Performance of the Technology
The investigative tools currently available to the geoscience professions are

adequate for determining stratigraphy at a point. Except in the case of
fractured media, the interpretive methods also appear adequate. There have
been numerous cases where it has been possible to compare stratigraphy, as
revealed in excavations, with predictions based on the findings from typical
field explorations. The correlations have been found to be satisfactory when
a proper field investigation had been made. Discrepancies (which can be
significant in transport analyses) generally occur in complex and/or fractured
strata situations, especially when only a minimal investigation has been made.
7.2.4
Assessment of the Technology
The selection of exploration methods for determining stratigraphy at a site

will be based on a variety of considerations. Although the circumstances will
vary from site to site, the following three considerations tend to be common
to all:
the anticipated geologic setting, in particular the types of lithology

and their stratigraphic sequence, will generally indicate the preferred
exploration methods;
the decision to implement the indicated exploration methods will be
influenced by the relative costs of implementation, which, in turn,
will be dependent on the scale of the project; and
the objectives of the stratigraphic characterization (e.g., contaminant
delineation, hydrogeologic modeling, consolidation analyses, stability
evaluations, or other) will also affect the selection of exploration
methods to be used.
146
Early development and repeated refinement of the site conceptual

stratigraphic model is essential, particularly for large, multi-phased
exploration programs. Development of a stratigraphic model early in the
project is helpful in the process of defining the objectives of more detailed
stratigraphic characterization. Ongoing testing and refinement of the
stratigraphic model during large field programs helps to ensure that a project
team is reviewing data from a completed phase of a field program before
proceeding to the next phase. A review of this type can:
enhance the performance of field staff by providing them with a better

understanding of the types of field data that are important to a
particular stratigraphic characterization effort; and
confirm success in a stratigraphic characterization effort: if data from
later phases of a field program do not exhibit increasing agreement
with the conceptual stratigraphic model of the site, this may indicate
the need for a different model.
Multi-phase field programs, which can be expensive, ultimately stand a better

chance of being performed efficiently and completely if the data from each
phase is carefully reviewed before embarking on the succeeding phase.
In recent years, the compilation, interpretation, and presentation of
stratigraphic data, as well as decision-making based on the interpretation,
have been greatly facilitated by the advent of software applications that can
manage three dimensional earth science data sets, enabling interpolation (and
extrapolation) of data based on a variety of geostatistical routines. These
applications tend to be used on large, long-term projects, both because these
systems are relatively more economical at a large scale, and because large
scale projects (e.g., several hundred borings over decades) are difficult to
manage in any other way. Despite the advantage these applications offer,
they must be used with care, keeping in mind that the compilation and
interpretation of data are only as good as the quality and comparability of the
underlying data. This applies not only to the accuracy and consistency of
data, e.g., a CPT trace or geologists classification of soil, but also to the
accuracy of the site survey which provides spatial control for such data.
Professional judgment is important in the application of geostatistical
procedures. The modeling approach needs to be compatible with the
character and variability of site stratigraphy. Field data must be collected
with a frequency and spatial distribution that enables robust interpolation.
Finally, the results of interpolation must be checked against localized
stratigraphic irregularities (e.g., faulting, buried stream channels, fill, etc.).
Notwithstanding these caveats, these geostatistical applications offer
enormous promise for cost-effectively managing large stratigraphic data sets
and will aid in the development and ongoing checking and refinement of the
stratigraphic model.
7.3 DRIVING POTENTIALS
7.2.5
147
Needs
Non-intrusive site exploration methods need further development to improve

the level of confidence in quantifiable data. In interpreting data, current
quantification methods by statistics lack field-proven statistical parameters,
e.g., correlation lengths.

7.3.1
Background
While contaminant transport is impeded by the geologic strata, it is driven by

spatial variations (gradients) of potential energy. For a given situation, the
gradient of potential energy is the key factor affecting the direction and rate
of transport. Fluid potential differences drive advective flow. The most
common analysis, that for advective groundwater flow, is based on gradients
of piezometric head (sum of pressure head and elevation). However, the
piezometric head at a point can be modified by the presence of chemicals,
water of varying density, and localized heat sources. Fluids also flow in the
gas phase, driven by variations in gas pressure (pressure gradients). Gas
flow examples include methane migration in/from degrading waste masses
and steam or air stripping of soils. Thermal gradient situations involving
waste containment facilities are not simply limited to high-level nuclear sites;
other examples include chemical reactions (or even fires) in buried waste
masses and in subsurface storage of hot hazardous fluids.
7.3.2
State of Practice of Investigative Methods
a. Point Measurement vs. Field Distribution. The geoscience professions

have relatively adequate instrumentation for determining piezometric head,
chemical concentration, and temperature at a given point. The challenge is in
determining the spatial distribution of these potentials, i.e., the potential field.
b. Hydraulic (Liquid) Potential. For simplicity, the term hydraulic in
this discussion will refer to all fluids, not water alone. The basic technique is
simple: measure the piezometric head at various locations and depths. The
normal direct methodologies involve some form of intrusive sensor placement.
Among these are well and/or piezometer clusters and piezocone CPT
measurements. Measurement methods are covered in ASTM D4750.
Transducers in wells, etc., can determine the potentiometric level to 0.001
feet (0.3 mm) in water and to 0.01 ft.(3 mm) in nonaqueous liquids
(NAPLs). The results from sonic units or tape probes exhibit less accuracy
and consistency. Indirect methods, such as electrical resistivity, seismic
reflection, and ground-penetrating radar, can be helpful; but their current
148
accuracy level is poor. More work is needed if these methods are to become
reliable in a quantitative sense. Water levels in nearby water bodies, such as
streams or lakes, can be useful references when conducting a survey of
piezometric head potential.
While the actual data obtained from direct measurement techniques may
be precise, the interpretation of the results may be less accurate. The potential
(piezometric head) distribution is greatly influenced by the distribution of
hydraulic conductivity in the formation being investigated. This has often
been noted in vertical gradient studies, where low-hydraulic conductivity
zones may cause perching. In fractured materials, local changes in hydraulic
conductivity can significantly alter the head distribution over short distances.
There are many cases where the vertical gradient of piezometric head
reverses over a distance of just a few feet. The potential head distribution
within a fractured mass can be quite complex and easily misinterpreted when
based on a few scattered measurements. There can also be temporal variations
due to precipitation or barometric events, and these must be considered in
potentiometric mapping.
To aid in data interpretation, a recommended technique is to concentrate
point measurements in the pervious zones, where head variations are typically
small. These values serve to define the boundary conditions for the natural
barrier formations fairly accurately, and the gross potential head difference
across (but not necessarily within) the aquitard can then be evaluated.
Overall, the hydraulic head potential distribution can be defined more
accurately than can the distribution of hydraulic conductivity; however, this
may not be true in the case of sparsely fractured materials.
c. Gas Potential. Experience has been limited in measuring gas pressure
potentials, implying that potential fields in gas have not been studied in
detail. There have been some cases of breathing in the unsaturated zone,
where cyclical air pressure changes cause transport of gaseous contaminants
into or from the subsurface. These barometric pressure changes can be on the
order of 2% to 3% of atmospheric pressure.
d. Chemical Potential. The chemical potential field is usually determined
by tests for chemical concentrations at various points. The usual technique
involves analytical chemistry testing on samples taken from wells or boreholes.
These tests are quite precise in determining the chemical concentrations in
the sample. However, they may be misleading in an overall sense unless the
sample is representative of the field conditions. Field methodologies include
scan chemistry (organic vapor analyzers), pH and/or electrical resistivity,
CPT probes, etc. These are far less precise but, because they test larger volumes,
are often more accurate than analytical chemistry. Downhole analytic
chemistry is available today for volatile organics and petroleum products using
CPT (push) technology. Also, downhole optic fiber sensors and portable
analytical chemistry instrumentation are being developed for on-site
149
determination of chemical concentrations (Magee, 1996).

Generally, delineation of chemical potential fields is based on the general
pattern of point chemical concentrations, modified (if possible) to account for
chemical reactions. Local variations from the general pattern will exist, but
complete definition of these variations is normally not achievable without
excessive effort (and cost).
As in the case of hydraulic head potentials, chemical potentials can be
determined to greater levels of accuracy than can be achieved in determining
the transport parameters. While improvements in chemical potential
characterization are desired and possible, limited resources might be better
expended on research and development that improves characterization of
the transport parameters of the aquitard.
e. Thermal Potential.
The thermal potential field is studied through
measurement of the subsurface temperature distribution. The normal
methodology involves intrusive placement of temperature sensors.
Temperature sensors can be permanently emplaced with remote reading.
Temperature profiles can be obtained using available CPT downhole sensors.
Precision levels for temperatures measured utilizing direct methods are quite
good (tenths to hundredths of C).
The field distribution of temperature is subject to considerations similar
to those of hydraulic head distributions (for saturated conditions). However,
significant latent heat can be stored or released during phase changes due to
vaporization or freezing. Overall, however, temperature distribution
determinations have been adequate, as has been modeling of heat flux in the
subsurface using heat conservation equations. Since advective transport tends
to smooth thermal gradients in permeable units, bounding conditions for
aquitard units can be established fairly accurately. Absent phase changes,
the confidence level in the determination of thermal potential distributions is
similar to that obtained for hydraulic head and chemical potential
distributions.
7.3.3
The tools available to the geoscience professions have more than adequate
levels of precision in measuring piezometric head, chemical concentration,
and subsurface temperature at a given point in a geohydrologic system. It is a
more difficult task to define the potential field based on a limited number of
point measurements. In spite of this difficulty, the accuracy of potential field
prediction far exceeds the level of accuracy in the determination of the
transport parameters (e.g., hydraulic conductivity).
7.3.4
Needs
The primary needs in potential field determination involve the improvement
150
of interpretive methods and the development of improved surface scanning

techniques for spatial interpretation of the potentials based on point
measurements.
7.4
7.4.1
LIQUID TRANSPORT PROPERTIES - SATURATED CONDITIONS

Background
The primary transport parameters under saturated conditions are: hydraulic

conductivity, effective porosity, diffusion/dispersion coefficients, and partition
coefficients that account for adsorption. These transport parameters determine
the capabilities of a natural stratum to serve as an indigenous barrier.
Determining representative values of these parameters within acceptable
levels of accuracy is the key to adequate transport analyses.
7.4.2
State of Practice
Current methodologies for determining transport parameters are discussed

below.
a. Hydraulic Conductivity. Laboratory Methods. Laboratory tests to
determine hydraulic conductivity can be conducted in either rigid or flexiblewalled permeameters. Daniel (1994) presented an extensive summary on the
state-of the-art of these tests and concluded that flexible wall tests are best for
fine-grained deposits. The results obtained from these tests are critically
dependent upon the selection and preparation of test specimens. Both constant
or falling-head tests can provide satisfactory results. However, head change
must be small if the test specimen is compressible in order to avoid densifying
the specimen, which reduces the sample hydraulic conductivity. Liquids other
than water can be used in the test if the specimen is adequately backflushed
so that chemical equilibration is reached. Other possible problems with these
tests include wall effects (excessive flow near the wall) and entrapped air
(which reduces the effective porosity). The major problem is sample size; the
sample is frequently too small to accurately represent the effects of macropores
on the flow system.
Field Methods. Selection of a field hydraulic conductivity test method depends
on whether the test can be conducted at the exposed barrier floor or if boreholes
must be used. Trautwein and Boutwell (1994) present an overview of both
types of methods. Comparisons of slug test data on indigenous fine-grained
soils based on permeameter test data from samples of the same soils show
that slug tests generally produce higher values of hydraulic conductivity. This
has been attributed to the larger scale of the slug tests (Herzog and Morse,
1986; Bradbury and Muldoon, 1990). Tests on properly compacted clay liners
7.4 LIQUID TRANSPORT PROPERTIES
- SATURATED CONDITIONS
151
have not shown this difference.

Borehole methods include slug tests, which are summarized by Herzog
(1994), and the two-stage borehole (TSB) test, which is described by Trautwein
and Boutwell (1994). Borehole tests in fine-grained deposits are best conducted
in open boreholes created using a thin-walled sampler to eliminate
complications in the analysis due to the high hydraulic conductivity of sand
pack and from side wall smearing. Herzog reviewed four methods for
analyzing slug test data by Cooper et. al. (1967), Bouwer and Rice (1976),
Hvorslev (1951), and Nguyen and Pinder (1984). Although slug tests primarily
measure horizontal hydraulic conductivity, tests conducted in angled holes
have been used to estimate the vertical component of hydraulic conductivity
(Herzog and Morse, 1986). All four analytical methods produced similar
values of hydraulic conductivity for low hydraulic conductivity deposits, but
the values calculated for an individual test tend to diverge at higher hydraulic
conductivities. The Nguyen and Pinder method provides a composite
hydraulic conductivity value, since it is based on the three-dimensional form
of the flow equation. Unfortunately, the method is not recommended because
of an error in the solution (Butler and Hyder, 1994). There is a need for an
improved 3-D method to evaluate hydraulic conductivity from slug test data.
The TSB test, a variation of the slug test, uses the Hvorslev method of
analysis to provide estimates of both the vertical and horizontal hydraulic
conductivities. Borehole methods can provide information on the spatial
variation of vertical and horizontal hydraulic conductivities, and such testing
can be conducted on slopes or at depth. ASTM is preparing standard methods
for conducting the slug and TSB tests.
b. Porosity. For saturated materials, total porosity can be determined by the
conventional oven-drying method (ASTM D2216). A microwave method
(ASTM D4643) can be used if results are needed more quickly. The microwave
method requires the use of a calibration curve that correlates the values obtained
from the microwave method with those obtained using the conventional ovendrying method for each soil to be tested (Krapac et al., 1991).
Effective porosity represents the volume of void space available for fluid
flow and is more difficult to measure. Peyton et al., (1986) studied the effective
porosity of fine-grained deposits and concluded that it depends on the size of
the molecules being transported compared with the size of the passages
connecting the pores. He also found that water could pass through these
passages in a lacustrine sediment, so that effective porosity and total porosity
were identical for water in that case. In contrast, Horton and Thompson (1986)
found the effective porosity of loess and glacial till to be approximately 10%
to 20% of total porosity. The difference between these two findings has not
been resolved and requires further study.
c. Diffusive Flux. In low hydraulic conductivity environments, such as exist
in waste containment barriers, diffusive flux can be more significant than
152
advective flux (see Section 10). No standard laboratory or field method exists
for the measurement of diffusive flux parameters.
Laboratory Methods. Both steady-state and transient methods are available
for determining diffusion in the laboratory. Shackelford (1991) presents a
comprehensive review of these methods and a summary of published values
for effective diffusion coefficients in low hydraulic conductivity soil
formations. Techniques include column, half-cell, double and single reservoir,
and chemical digestion methods. No standards have been published for these
methods. The degree of saturation is a major factor affecting the measured
value of the effective diffusion coefficient in a porous medium. Reported
values for nonreactive and reactive solutes in saturated soils are up to 10 to
20 times greater than the corresponding values in unsaturated soils. Thus,
degree of soil saturation is critical in the determination of reliable effective
diffusion coefficients. Diffusive transport rates of nonreactive solutes may
appear to be much greater (up to 5000 times greater) than rates obtained for
reactive solutes, unless reversible sorption reactions are taken into account
for reactive solutes.
d. Adsorption. Laboratory Methods. Adsorption coefficients must be
determined for specific soils and solutes. The batch-adsorption or staticequilibration technique is the most common laboratory method for testing
the capacity of geologic materials to adsorb chemicals. The U.S. EPA has
published a Technical Resource Document (Roy et al., 1991) on this method,
which is also addressed in ASTM standard D4646. Procedures have been
developed and tested for both inorganic and organic solutes. Although the
techniques are relatively simple, the results are dependent on several
experimental parameters, including: contact time, solution pH, hydrolysis,
and the presence of other dissolved constituents in the aqueous solution, in
addition to the solute of interest.
Adsorption of organic solutes is affected by the following: presence of
dissolved organic carbon, adsorbate volatility, photodegradation,
biodegradation, and compound stability. German researchers have found
that, in general, retardation coefficients calculated from batch tests are
distinctly higher than those obtained from diffusion and percolation
experiments (Czurda and Wagner 1991). The reasons for these differences
deserve further attention.
Field Methods. No field methods have been developed for measuring
adsorption parameters.
7.4.3
Field Performance of the Technology
Hydraulic conductivity can be adequately measured at a point, but the

accuracy of an overall stratum conductivity estimate is probably only correct
7.5 LIQUID/GAS TRANSPORT
- UNSATURATED CONDITIONS
153
to within one order of magnitude. In fractured media, the accuracy may be

poorer. There are few field data that can be used to evaluate the accuracy of
estimates for the other transport parameters.
7.4.4
As stated above, the actual level of accuracy is within about an order of

magnitude in the determination of hydraulic conductivity and is essentially
unknown for the other transport parameters.
7.4.5
Needs
The predominant need relative to the determination of transport properties

for indigenous barriers is the need for a better understanding, definition, and
measurement of the in situ values of these transport properties. Some progress
has been made in accounting for specimen scale size in the measurement of
hydraulic conductivity. Analogous advances are needed in the evaluation of
diffusion and adsorption coefficients, including their relationship to effective
stress and contaminant chemistry.
7.5 LIQUID/GAS TRANSPORT - UNSATURATED CONDITIONS

7.5.1
Background
Currently, all U.S. regulations and most of the commonly used analytic/finite
element models for contaminant transport are based on saturated flow
conditions. However, the saturated condition does not always exist, especially
at sites in the western U. S., requiring that the more general (and more complex)
situation of contaminant transport under unsaturated conditions be
considered. Clay barriers are a good example. Although the saturated
hydraulic conductivity is generally specified, the common field tests (Sealed,
Double-Ring Infiltrometer and Two-Stage Borehole Test) involve flow into an
unsaturated medium (Trautwein and Boutwell, 1994).
The concepts applied to barriers in the saturated case are no longer valid
in the unsaturated regime. Hydraulic conductivity, for example, becomes a
function of the degree of saturation and the driving potential for transport
may be dominated by suction pressures rather than by hydrostatic pressures.
Also, effective porosity varies with the degree of saturation and with the
concentration and type of ions present in the pore water. Diffusive flux ceases
to be characterized using the aqueous solution value of the diffusion
coefficient, and may be better evaluated in terms of gas flow.
Fredlund and Rajardo (1992) suggest that at saturation levels of 85% or
more, transport tends towards (but not exactly) saturated behavior, i.e., liquid
flow dominates; while at saturation levels less than 15%, soil behavior tends
154
toward dry behavior, i.e., gas flow dominates. Between these two saturation
levels, a full partially saturated analysis is required.
7.5.2
Role of Diffuse Double Layer
In clayey soils, most of the soil moisture is contained within the diffuse double
layer. The moisture content in soils at low levels of saturation exists as thin
films on the particles, rather than as sparsely distributed three-dimensional
pockets of moisture. Film thicknesses are on the same order as the diffuse
double layer thicknesses, so that the soil transport properties are influenced
by the electrical fields. High soil densities are often associated with highly
over-consolidated natural clays. Under these conditions, particle spacings
may be so small that the electric fields influence transport in the virtually
saturated state just as they do in the unsaturated state.
These electrical fields affect transport of dissolved ionic contaminants
more than they do the bulk water flow, acting to trap conservative tracers,
such as chloride and bromide. Thus, pure advective transport does not occur
in soil pore systems dominated by the presence of diffuse double layers.
Adsorption of ionics must also be considered. This sorption depends on the
ions charge density (valence per unit volume of the ion). High charge densities
produce high sorption coefficients and thus, high retardation of ionic
contaminant migration. Anion adsorption is usually low, on the order of 1%
to 5% that of cations.
7.5.3
State of Practice
a. Effect on Conventional Transport Parameters. The level of saturation

influences the magnitude of the conventional transport parameters; i.e.,
hydraulic conductivity, effective porosity, diffusion coefficient, and retardation
coefficient.
Hydraulic Conductivity. While unsaturated flow of liquids may still be
analyzed utilizing Darcys Law, the hydraulic conductivity in unsaturated
flow varies with the degree of saturation. The lower the degree of saturation,
the lower will be the hydraulic conductivity. On the other hand, suction
pressure increases with decreasing levels of saturation. Therefore, the material
with the lowest hydraulic conductivity will also exhibit the highest suction.
The changes in transport resulting from changes in degree of saturation can
be very large, varying by orders of magnitude. Further, the relationships
between hydraulic conductivity and degree of saturation are nonlinear and
exhibit hysteresis (not single-valued). The relationship differs when a soil is
dried from the wet state as opposed to when it is wetted from the dry state
(hysteresis effect). Determining this relationship, even in the laboratory, poses
a significant undertaking for each soil type to be studied.
Hydraulic conductivity is thus a strongly nonlinear function of both the
155
degree of saturation and the piezometric head potential. Sands drain at lower
potentials than do clays. As a consequence, while sands have higher hydraulic
conductivities at full saturation, they generally have lower conductivities at
low levels of saturation. The relationship between hydraulic conductivity
and the degree of saturation is routinely determined by measuring the
hydraulic conductivity under saturated conditions and the moisture
characteristics curve, thus enabling the calculation of the hydraulic
conductivity as a function of level of saturation.
Effective Porosity. Effective porosity indicates the percentage of the total
subsurface volume that is actually available for fluid flow. Practical experience
indicates that the effective porosity is generally less than the total porosity.
The actual portion of the pore system void space available for transport affects
solute breakthrough times in laboratory column experiments, as well as solute
travel times in the field. An exception may be the transport of tritium, where
the effective porosity is frequently the same as total porosity. For anions that
are excluded from the diffuse double layer, effective porosity appears related
to free water content. Total porosity is normally determined by ovendrying a
saturated sample. However, the drying process involves the addition of heat
energy to the water. Each bound layer requires a different heat energy for
vaporization. Thus, pore water evaporates sequentially, corresponding to
these energy levels. Free water is lost at moderately low temperatures,
followed by the diffuse double layer, and then by interlayer water as the
temperature increases to the ASTM D2116 standard of 105 oC. Thus, the
measured porosity becomes a function of drying temperature. Effective
porosity is essentially a measure of the free water (unbound) and probably
could be determined by drying at moderate temperatures. Currently, there is
no standard method for this procedure.
Diffusion Coefficient. Diffusion of solutes in unsaturated materials can not
be described assuming that one diffusion coefficient is applicable, since there
is simultaneous diffusion in the bulk liquid, the liquid film, and the gas phases.
Diffusion of ionics is also a function of charge density. Because of this, the
apparent diffusion coefficient is observed to have a dependence on the degree
of saturation, similar to hydraulic conductivity. Most organic contaminants
and CO2 , both of which partition into both the liquid and gas phases, tend to
show an apparent diffusion coefficient intermediate between that obtained
under full saturation and that obtained in the gas phase alone. Therefore,
under partially saturated conditions, the apparent diffusion coefficient for a
specific contaminant depends on the degree of saturation, as well as Henrys
law constant, which determines how the contaminant partitions between the
liquid and gas phases. As a rule of thumb, the apparent diffusion constant
may be taken as being somewhere between the gas value and the liquid value.
Although this relationship depends on the degree of saturation, an average
of these two limiting values is frequently used.
156
Retardation Coefficient. Ionics can be either retarded (anions) or accelerated

(cations) by the electrical fields of the double diffuse layers. Thus, the
retardation factor can be greater than one (anions) or less than one (cations).
If the double layers do not overlap, ions tend to be accelerated. Overlapping
double layers can result in retardation of the ions dissolved in the solvent
(water), i.e., producing retardation factors greater than one. Anions and
cations can not be easily separated due to the charge imbalances that would
be created. Slight charge separation does cause streaming potentials, but cation
and anions tend to move together. In many experiments, the observed
retardation coefficients for anions are less than one.
Non-Polar Molecules. The transport behavior of organic compounds in
unsaturated materials has not been studied extensively, except in vapor
extraction systems. Therefore, little is known on this topic. Most of the transport
of volatile organics occurs in the vapor (gaseous) phase of an unsaturated soil.
Studies have been conducted on vapor extraction systems and air flow through
porous media. In many extraction systems, approximately 50% of the total
costs can be attributed to the treatment of the extracted contaminants.
Subsurface aeration can enhance subsurface biodegradation of the contaminant,
potentially lowering surface treatment costs.
b. Implications for Design. The purpose of a barrier is to prevent or inhibit
migration of contaminants, not simply to comply with some stated regulatory
value for the saturated hydraulic conductivity. In the relatively dry condition
(low degree of saturation), the hydraulic conductivity of fine materials is greater
than that of coarse materials; the exact opposite to the saturated condition.
Thus, under relatively dry conditions, fine materials may serve as drainage
layers and coarse materials become barriers. This phenomenon has been used
as a basis for constructing barrier systems in arid climates. A fine sand becomes
the leachate collection system and coarse sands, gravels, or even concrete
blocks become the barrier to migration.
c. Monitoring. Monitoring procedures in the (unsaturated) vadose zone
depend upon the specific contaminant being monitored. If aqueous phase
contaminants are of concern, suction lysimeters are often used to sample soil
water. If gas phase contaminants are an issue, soil gas is commonly monitored
with multiport sampling systems. Barometric pressure fluctuations may also
be important in contaminant movement and can be monitored with pressure
transducers.
Moisture content is typically monitored by neutron probe measurements.
However, time domain reflectometry and capacitance measurements are
becoming more popular. Use of these latter devices requires training and
experience in order to obtain accurate moisture content data over time.
Tensiometers are often used to monitor hydraulic head potentials near
saturation. At lower head potentials (1/3-5 bar), head dissipation probes are
157
generally the tool of choice.

The make-up of lysimeter samples, obtained by suction, can depend on
the suction level used in the lysimeter. At lower suctions, a larger fraction of
free water and higher ionic concentrations are obtained. At higher suctions,
water adjacent to the double layer and lower ionic concentrations can be
sampled. Thus, the analytical chemistry results can depend on the suction
level and may not reflect contaminant concentrations in the free, advecting
water. Also, the (necessarily) small pores of the lysimeter will preferentially
transmit small ions and exclude or retard large ions. A truly representative
sample of ionic contaminants can be difficult to obtain.
7.5.4
Field Performance
Vadose zone monitoring can provide early detection of contaminant

movement, thus offering an early opportunity for intervention and the
possibility of preventing groundwater contamination. Since natural vadose
zones frequently exhibit preferential flow paths and heterogeneities, much of
the data obtained from point measurements (e.g., tensiometers) or scans
(neutron probe logs) may not capture important three-dimensional transport.
Nielsen et al., (1990) have concluded The efficacy of accurately predicting
the attenuation and eventual location of solutes or constituents in the vadose
zone remains undeveloped. For these reasons, site monitoring will remain
a requirement for the foreseeable future.
7.5.5
Vadose zone contamination problems must be approached with a sound

understanding of unsaturated soil phenomena. Better use could be made of
our current understanding of transport processes and of the available
exploration technologies. While there are limitations in the ability to
characterize, monitor, and predict transport in the vadose zone, additional
vadose zone work continues to be undertaken because of the possibility of
preventing groundwater contamination and possibly reducing the ultimate
costs associated with remediation. For volatile contaminants, field monitoring
in unsaturated formations will be of increasing importance. In engineered
systems, soil physical properties can be manipulated to enhance the
effectiveness of point or scan measurements of hydraulic properties.
Engineered vadose zones offer some practical advantages for both
containment and monitoring. While this knowledge exists, it is often not
considered in a facility design.
7.5.6
Needs
Tools are being developed to improve the ability to see vadose zone processes.
These tools range from geophysical techniques, such as electrical resistance
158
tomography, to improved capacitance and time domain reflectometry probes.

The understanding of contaminant transport processes in unsaturated soil
materials needs improvement, as does the understanding of and the ability
to predict transport of moisture in the vadose zone. The effects of temperature
gradients, diffuse double layers, and sorption processes are not well
understood. As one example, many vadose zone data sets show large
variations in ionic concentrations; but few models offer an adequate
explanation as to why this occurs. More effort needs to be made in educating
the environmental community regarding vadose zone phenomena, since
incorrect assumptions and incorrect installations of monitoring equipment
can lead to costly consequences. Such educational efforts would eliminate
many common mistakes made in vadose zone hydrology.
7.6 EFFECT OF FRACTURES ON TRANSPORT

7.6.1
Background
Fractures (including joints, faults, bedding-plane separation, etc.) can occur

in any cohesive or lithified geologic material. Fractures can often be seen in
outcrops or core samples, but it is generally not possible to visually determine
whether they are significant pathways for contaminant transport. The
presence of alteration features such as oxidation staining, in-filling, and root
casts, are indications that the fractures were open at some time in the past.
However, open fractures can be closed as a result of changes in the
surrounding stress-field or by in-filling of the fracture with low hydraulic
conductivity material. On the other hand, inactive fractures may become
active if the in-fill material is dissolved and washed away. Such dissolution
can result from both natural or man-induced changes in geochemistry of water.
Finally, pre-existing fractures with no visible alteration can act as significant
pathways for flow and such fractures are very difficult to distinguish from
man-induced fractures caused by drilling or excavation.
7.6.2
State of Practice
a. Cubic Flow Law. Flow through a fracture network is controlled by a

number of factors, including: aperture (or separation distance between the
walls of the fracture), fracture density (i.e., spacing between fractures), length
of fractures, and fracture orientation. Fracture spacing, length, and orientation
control the degree of interconnection of the fracture network and can often be
characterized by visual mapping. Fracture aperture is more difficult to
measure because of the small values (typically on the order of 10 to 100
micrometers) and because stress-relief in excavations or boreholes affects the
aperture value. Aperture values in low-hydraulic conductivity materials are
generally calculated indirectly from measurements of bulk hydraulic
7.6 EFFECT OF FRACTURES ON TRANSPORT
159
conductivity and fracture spacing, length, and orientation using the cubiclaw (Snow, 1969). The cubic-law is based on the relationship between
hydraulic gradient and discharge for a viscous fluid flowing between two
parallel smooth surfaces; hence, aperture values derived from this relationship
are some sort of average for the entire fracture network. In reality, the aperture
can vary greatly within a single fracture, which leads to larger aperture
channels controlling much of the flow. There have been a number of
theoretical investigations of the possible influence of aperture channeling on
contaminant migration, but it is very difficult to get reliable field data. Thus,
the most common method for estimating fracture aperture values is still based
on the cubic-law.
b. Matrix Diffusion.
Migration of a solute in a fractured porous medium
occurs by advection through the fractures, advection in the matrix, and especially
diffusion into the matrix. In most low hydraulic conductivity materials,
advection in the matrix is negligible (especially if the system is fully-saturated)
and diffusion controls the transfer of solutes from the fast moving water through
the fractures into the relatively immobile pore water of the matrix. This process
of matrix diffusion can theoretically retard the overall migration of a solute
by many orders of magnitude relative to contaminant transport in the fracture
alone. The effectiveness of this retardation process is largely related to the
porosity of the matrix, with the effect being much greater in high porosity clays
and shales than in low porosity rock. In recent field and lab experiments in
fractured glacial clays (McKay et. al., 1993 and Hinsby et. al., in-press) nonreactive solutes (bromide and chloride) were retarded by a factor of up to 100
relative to the transport of colloidal virus tracers, which were too large to enter
the matrix pores, and hence were not retarded by matrix diffusion. Modeling
simulations by Sudicky and McLaren (1992) showed that in a scenario which
included matrix diffusion, a nonreactive solute (e.g. chloride) migrating
downwards through a 25 m aperture fracture in a glacial clay would take
about 50 years to reach the underlying aquifer at 10m depth. For the same
scenario, but with no matrix diffusion, the solute would reach the aquifer in
about a half a day. Field and laboratory investigations by Birgersson and
Neretnieks (1990) have shown that solute diffusion in low porosity granitic
rock, although much slower than in the above fractured clay examples, can be
significant. As a result, it is possible that in some scenarios (e.g., radioactive
waste migration in deep fractured granites), matrix diffusion may be a significant
retarding mechanism even in very low porosity fractured rocks.
The downward migration of DNAPLs in fractured media (Kueper and
McWhorter, 1991) usually occurs fairly rapidly after a spill and is unaffected by
diffusion, since the penetration is controlled primarily by fracture aperture and
the properties of the DNAPL (density, interfacial tension, and viscosity).
However, once the DNAPL has become immobilized, diffusion of DNAPL
contaminants into the matrix pores will make it much more difficult to
remediate.
160
c. Assessment Methods. Methods of assessing the significance of fractures in

soil or rock fall into the following categories:
1) geologic methods
mapping of fractures in outcrop or core samples
downhole camera and geophysical surveys
local or regional faulting, folding, and/or other indicators of stress
history
2) hydraulic methods
direct measurements of field or lab hydraulic conductivity
monitoring of variations in hydraulic head profile with depth due
to seasonal or episodic changes in head (or recharge)
borehole flow meter surveys
3) geochemical and isotopic methods
isotopic indicators of groundwater age (3H, 2H, 18O, 14C, 85Kr)
chemical indicators of recent recharge (CFCs, dissolved solvents,
or pesticides; i.e., man-made chemicals)
Each of the above methods has varying degrees of reliability, and sometimes
different methods of investigation can produce contradictory results. In
general, the geochemical/isotopic methods tend to be more reliable, but use
of multiple methods of investigation is recommended.
7.6.3
Field Performance
Geophysical methods, such as ground penetrating radar, are capable of

locating fractures and fault zones. However, methods for determining
transport properties through a fractured mass in the field need to be further
developed to become an engineering tool.
7.6.4
Assessment
The capability exists for integrating the various stratigraphic, hydrological,

and chemical data in the prediction of transport of a specific contaminant.
However, this is a complex task and field verification is difficult. Thus, the
current ability to assess the capacity of indigenous layers (especially if
fractured) to serve as natural barriers is still unsatisfactory. Basic questions
remain concerning fractures that require further intense investigation on such
topics as: the role of confining stress in fracture opening and closure due to
excavation and construction; the extent of irreversibility of such opening and
closing and the related question of fracture healing; fracture opening/
closure in response to pore pressure cycles; the role of in-filling material, which
may be dissolved or precipitated by concentrated chemical contaminants; the
role of matrix diffusion in retardation of different types of dissolved
contaminants; the factors influencing penetration of DNAPLs in fractures;
7.7 EFFECTS OF CHEMICALS
161
and the factors influencing attenuation of colloidal contaminants. Current

research models that attempt to simulate these processes are generally too
complex and user-unfriendly for application in the field.
7.6.5
Needs
From an applied perspective, significant improvements are needed in the

following areas: development of methodologies to detect hydraulicallyconductive fractures, particularly in settings where the fractures are widelyspaced (meters to tens of meters); development of user-friendly computer
models that incorporate fractures and matrix diffusion; and finally,
presentation of rigorously investigated case histories that illustrate the
integration of stratigraphic, hydraulic, geochemical/isotopic and geophysical
information combined with the use of computer models in the assessment of
contaminant migration in different types of fractured, low hydraulic
conductivity deposits. Overall, the biggest need is for field verification of the
processes controlling contaminant transport and continued development of
predictive models.
7.7 EFFECTS OF CHEMICALS

7.7.1
Background
The natural geochemistry of clay minerals, excluding the potential impacts

of synthetic chemicals, is itself a complicated subject. Clay mineralogy
textbooks, such as Grim (1968), address this natural geochemistry; but,
understanding subsurface geochemistry at both a micro and macroscopic level
remains a difficult task. Since the geochemical equilibrium in some clay
minerals is often a delicate balance, most aggressive chemical conditions are
likely to alter this equilibrium which, in turn, may alter both the physical and
chemical performance of any clay mineral. Aggressive chemicals may also
alter the physical behavior of both cohesive and cohesionless soils, which
both represent potential transport pathways in geologic media.
Much of the available information concerning the effects of synthetic
chemicals on geologic media is based on laboratory research (Acar, et. al.,
1985, Anderson, et al., 1985, Hermann, 1985). Much less has been reported
on field observations of the effects of synthetic chemicals on geologic materials
(Hermann, 1978; 1985). Chemical effects include a variety of chemical reactions
and physico-chemical processes which may alter the clay mineralogy, as well
as the physical properties of the material. For example, in the case of a
secondary fracture system that dominates the flow characteristics of a natural
geologic material, the shrink/swell characteristics of clay minerals in or
adjacent to the fracture will affect the hydraulic conductivity of the overall
stratum.
162
The ability to measure chemical effects on geologic material is enhanced

in the laboratory where conditions can be more tightly controlled than in the
field. However, indigenous barrier applications are concerned with in situ
conditions and the in situ effects of synthetic chemicals on the hydraulic
conductivity, clay mineralogy, stratification, and natural geochemical
equilibrium. In this regard, low temperature aqueous geochemistry is an
emerging field that has become particularly important in the prediction of
contaminant transport of highly concentrated chemicals contained by
indigenous earth barriers.
7.7.2
State of Practice
Chemical effects on indigenous barriers are, in principle, similar to their effects

on man-made barriers. The results of laboratory compatibility experiments
have been reviewed in depth by Alther, et. al. (1985), Mitchell and Madsen
(1987), and Shackelford (1994). The principal mechanisms of contaminant
influences on soils are summarized as follows: brackish-brine salts: precipitation,
dehydration, cation exchange, density gradients (Alther, et. al., 1985); acids:
dissolution, geochemistry changes; alkalis: precipitation, geochemistry change
(Peterson and Gee, 1985); organics (non-polar): dehydration, water miscibility
of the contaminant. The most important clay transport property is hydraulic
conductivity. Threshold contaminant concentrations required to produce an
order of magnitude or greater impact on hydraulic conductivity are: brine
concentrations greater than 100,000 ppm (NaCl and other salts), pH less than
or equal to 1 or 2, and non-polar organic concentrations greater than 50-80%.
Hydraulic conductivity changes due to permeation by non-polar organics
depend on the kind of clay and the kind of organic chemical (mainly with
respect to their dielectric constant and miscibility). Typically, the changes are
related to the following mechanisms: dehydration, swelling, flocculation, and
macroscopic cracking (see e.g. Anderson et al., 1985, Daniel et al., 1986). A
70% concentration of an organic liquid in the pore water can increase hydraulic
conductivity by two to four orders of magnitude, if no confining stress is
applied. If there is a compressive stress applied, there can be a two order of
magnitude reduction in hydraulic conductivity (Bowders and Daniel, 1987,
Quigley et.al., 1988, and Fernandez and Quigley, 1991). High effective stress
may play a beneficial role in healing possible damage to a clay liner caused
by a catastrophic inflow of a highly concentrated low dielectric constant
contaminant. The level of stress necessary to prevent or remediate a hydraulic
conductivity increase is an important technical parameter in landfill design.
7.7.3
Field Performance
Most of the data concerning synthetic chemical effects on hydraulic

conductivity and other properties of clays have been obtained in laboratory
studies. The few reported field observations have tended to support the
7.8 THERMAL AND RADIATION EFFECTS ON TRANSPORT
163
laboratory findings in a qualitative sense, but quantitative data is lacking.

7.7.4
Assessment
The effects of chemicals on hydraulic conductivity are progressive and develop

over a long time. Predictions based on short term laboratory and field
experiments may be misleading. In the field, observations based on the transit
time of a tracer may require very long monitoring periods, on the order of
decades. Subtle changes in the clay geochemistry equilibrium may affect
long-term performance.
7.7.5
Needs
Changes in the natural geochemical equilibrium can alter the hydraulic

conductivity of an indigenous barrier, sometimes drastically. Clay-filled
secondary structures (fractures) can cause defects in the barrier when attacked
by contaminants. The available data on the effects of chemical attack are
almost exclusively from laboratory studies. Field data are scarce and poorly
documented. Efforts should be made to publish existing field data, despite
their sensitive nature. The response of field systems to the presence of
chemicals should be studied from both the experimental and modeling points
of view. Special study is recommended on determining which chemicallyinduced changes in hydraulic conductivity are reversible and which are
irreversible.

7.8.1
Background
This section covers two related topics: thermal and radiation effects. Thermal
effects are important in conventional practice, as they can adversely affect
the barrier properties of a stratum. Both thermal and radioactive effects are
critical in design of high-level radioactive waste facilities. Each is discussed
separately below.
7.8.2
Elevated Temperatures
a. State of Practice. Most of the recent experience in Europe with thermal

effects on soils at elevated temperatures (50-150 C) has been gained in the
course of developing technologies for disposal of high level radioactive waste
(HLW) in subsurface clay formations (in Belgium, Italy, Spain, UK, France
and Switzerland). In the commonly accepted concept of a containment system
(consisting of waste materials, the engineered barriers of backfill and buffer,
and a host clay formation) the man-made barriers are assumed to fail, releasing
164
the contaminant within 300 yr. Thus, the ultimate barrier is the indigenous
clay formation component of the containment system. No actual repository
is scheduled for construction before 2020.
The principal design consideration in these technologies is the rising
temperature within the soil mass. Temperature increases the pore pressure
during heating. Thermally induced pore pressure gradients may substantially
increase the natural water flow, enhancing potential contaminant transport if
an early breakdown of the inner engineered barrier occurs. Temperature and
pore pressure gradients determine the spacing of tunnels and/or boreholes,
a principal design parameter. Contaminant transport in clays is assumed to
be mainly by advection with water. Therefore, the hydraulic conductivity of
the clay mass is another critical factor for the design. Current design
methodologies are based on predicting temperature and flow using a linear
heat diffusion law, advection using Darcys law with constant hydraulic
conductivity, dispersion, and mass decay.
b. Field Performance. There are few field data reported on thermal effects,
other than the effects on foundation behavior. A 2.5-year in situ heating test
has shown a pore pressure buildup to 0.75 MPa around the heat source (Picard
et al., 1994). Other in situ experiments and studies of natural analogs have
revealed zones of microfractured clay around heat sources (Leone et al.,1986).
c. Assessment of Current Knowledge. Thermally-induced pore pressure
excess measured in the laboratory under undrained isotropic conditions
ranged between 1.5 to 5 MPa per 100 C (Del Olmo et al., 1994). At deviatoric
stress, the thermally-induced growth of pore water pressure may lead to failure
and/or hydraulic fracturing (Hueckel and Pellegrini, 1992). Formation of
such fractures can produce preferential transport paths for radionuclides
during the radionuclide release phase, long after the thermal phase is finished.
The pore pressure excess results from the difference between the thermal
expansion of pore water and the thermal volumetric strain of the skeleton.
While the pore water expansion is poorly known, the thermal strain during
heating depends critically on isotropic effective stress: it is expansive and
reversible for very low effective stresses, but contractive and partly irreversible
for high effective stress. Water content and hydraulic conductivity are
additional factors affecting this process.
The hydraulic conductivity caused by a temperature of 150 C after one
half year of heating in the presence of potassium in artificial clays was observed
to increase by a factor of 100. Mineralogically, thermally driven dehydration
produces illitization of smectites. This generates larger voids between the
collapsed stacks, depending on the effective stress. Studies on natural analogs
reveal that similar processes may occur in indigenous formations (Pusch and
Guven, 1990). The presence of carbonates affects thermo-mechanical behavior
of clays by increasing their apparent overconsolidation, rigidity and brittleness,
and pore pressure buildup. The variability in carbonate content in natural
165
clays makes site characterization a particularly complex task (Hueckel, 1995).

The assumptions of linear heat diffusion and Darcian hydraulic flow are
limitations of the current evaluation methods. Non-(or de-) saturated and
fractured or faulted clays may yield substantial departures from the linear
predictions. The effect of heat on diffusive transport is not well understood.
Numerical predictions using two-phase, hydro-thermo-plasticity models
have revealed that while the thermal phenomena are limited to a narrow zone
around the heat source, the increased water pressures affect a larger area.
During unloading in the course of cooling, clay may reach failure stress in
tension. However, thermal effects generally terminate before the time the
engineered barrier is expected to fail resulting in radionuclides being released
to the host clay. Therefore, the most relevant issues of the whole thermal
phase are the irreversible changes (dehydration, fracturation, strain), because
of the impact these irreversible changes have on the long term performance
of the repository (Baldi et al., 1987).
d. Other Contexts. In other applications outside of the context of high level
radioactive waste disposal technology, one may envisage having to deal with
chemically active liquids at elevated temperatures. No systematic expertise
exists in this area, and condition specific response of soil is expected. In general,
one would expect that elevated temperatures might enhance a possible
destructive chemical effect on hydraulic conductivity, mechanical strength,
and deformation of indigenous clay barriers. However, it is anticipated that
long term thermal exposure, comparable to that with radioactive waste, is
unlikely to occur.
7.8.3
Radiation Effects
During the waste containment phase, the radiation effects (mainly due to
gamma radiation) will be limited to a radius less than 1m. Limited experiments
with Belgian clay show the formation of radiolitic gases: H2, O2, H2O2, CO2,
and CH4. Migration of these gases will depend on the reaction rates, generated
gas pressure, and the formation hydraulic conductivity. Generally, interest is
focused on gas migration phenomena and possible desaturation effects.
During the release phase, clay moisture may enhance the radiolysis, increasing
the production of gas (Henrion et al., 1988).
7.8.4
Low Temperature Effects
a. State of Practice. The focus in this area has been on freeze/thaw effects
on the hydraulic conductivity of compacted natural clays used as landfill
liners and covers, liners for ponds, etc. However, the concepts also apply to
natural barriers. The technology involved is limited and consists of
demonstrating a sufficiently low hydraulic conductivity, even after a number
of freeze and thaw cycles. The understanding is that pore water during
166
freezing expands over 9% forming ice lenses which generate cracks and cause
suction, shrinkage, and cracking in the unfrozen zone. After a few cycles, the
network of cracks form a highly conductive system of channels yielding up
to three orders of magnitude increase in hydraulic conductivity. Effective
stresses applied to the soil during freeze/thaw cycles (and to a lesser degree
after the process), visibly attenuate the increase of hydraulic conductivity.
High rates of freezing worsen the situation (Othman et al., 1994). Earlier
studies of freezing zone propagation and frost heave support this
understanding.
b. Field Performance. Freeze-thaw effects on hydraulic conductivity have
been found to be significant in large-scale testing (Benson and Othman, 1993).
c. Assessment of Current Knowledge. Current knowledge concerning the
freeze/thaw effects is mainly phenomenological. Little effort has been made
to quantify the thermo-mechanical phenomena involved, to predict the
outcome, and possibly control it. Some earlier developments exist on frost
penetration during monotonic cooling, but outside of the context of concern
about hydraulic conductivity changes. Any possible effects of concentrated
chemicals on freeze/thaw changes on hydraulic conductivity are unknown.
How diffusive transport is affected by freeze/thaw cycling is also an open
question.
d. Needs. A better understanding of the basic mechanisms of thermallyinduced pore water pressures and ice pressure development needs to be
achieved. This includes reversible and irreversible macro- and microstructural thermallyinduced changes in clay, such as fracturing and increase
in hydraulic conductivities. The role of effective stress in controlling the
changes in hydraulic conductivity should be better understood. This effort
should be addressed through academic research involving highly specialized
mineralogical labs. Thermal effects in pre-fractured and in thermally fractured
media should be studied in the lab, and because of scale effects, also studied
with in situ field tests or on large models. Numerical models should be
developed for quantitatively predicting and aiding in the control of thermallyinduced changes in indigenous barriers.
7.8.5
Modeling: Coupled Thermo-Hydro-Mechanical and ChemoHydro-Mechanical Effects
a. State of Practice. The predominant practice in modeling combines linear

heat diffusion and thermal expansion with the advective flow equations for
non-deformable media for radionuclides. For modeling transport of chemical
contaminants, the advective flow equations for non-deformable media are
applied, accounting for diffusive transport and adsorption. An important
step forward would be consideration of hydraulic conductivity in the non-
167
saturated zone for contaminant transport. However, the role of effective stress
on the hydraulic conductivity is rarely taken into account. Similarly, physicochemical phenomena and chemical reactions are frequently not included.
b. Assessment of Current Knowledge. Contemporary knowledge of the
mechanics of two-phase transport in porous media combined with numerical
methods offer the possibility to improve the prediction of contaminant
transport beyond that provided by simple linear models. The most serious
obstacle in the development of more sophisticated models is still the limited
understanding of the mechanisms and processes involved and the lack of
quantitative data needed to evaluate material properties. These shortcomings
constitute a challenge for both experimentalists and theoreticians.
There are number of analogies to be explored between thermal and
chemical effects on indigenous barriers. They are not simply limited to the
nature of heat diffusion and contaminant transport, but also include: the
physico-chemical changes affecting the adsorbed water during a thermal
process and the permeation by organics; the nature of thermally and
chemically induced strain, thermal expansion, chemically induced swelling,
thermal consolidation, and chemical consolidation under high effective stress.
Also, there may be an analogy regarding the nature of the effective stress
dependence and changes in hydraulic conductivity. The thermal expansion
of pore water has an analogy in the permeant density change during
contamination.
However, there are also differences between mechanisms and processes
related to chemical and to thermal effects: temperature is a single agent, while
chemicals in a leachate present an enormous diversity. Chemical reactions,
e.g., dissolution, precipitation, etc., have no analogy in the thermal process.
Despite these differences, the basic structure of the model for thermomechanical and chemo-mechanical clay behavior may be based on the same
principles.
Modeling of coupled thermo-hydro-stress effects. Following the above
strategy, a well-established soil mechanics Cam-clay model has been extended
to include thermal conditions, experimentally corroborated, and coupled to
the flow equations for pore liquids in a manner similar to consolidation theory.
Most importantly, the plastic yield limit has been assumed to decrease with
increasing temperature. The latter assumption allows for such phenomena
as thermal decrease of peak strength, thermal generation of overconsolidation,
and for thermal consolidation strain at high stress (Baldi et al., 1987).
Simulations of high level radwaste disposal in boreholes in clays. Using the
above described model, clays from Belgian, Italian, and Spanish experimental
candidate sites have been characterized. Numerical simulations indicate that
while the higher temperatures affect only the close vicinity of a borehole, the
elevated pore pressures have a much larger influence and may affect
168
neighboring boreholes. The effective stress evolution close to the heat source
indicates a possibility of soil failure during cooling. In the failed zone, a
preferential flow path could arise along the containers. In another simulation,
a significant thermally-induced increase in the hydraulic conductivity was
found within a radius of 5 m around the heat source (Ma and Hueckel, 1992).
Modeling of coupled stress-chemical-hydraulic effects on contaminant flow.
To model the above coupling, one may employ the same strategy as in the
case of thermal effects. Chemical reactions and effects resulting from
sequential miscibility can be taken into account by introducing additional
constitutive equations. To effectively model the response of an indigenous
barrier using such models, specialized experiments are required to provide
material parameters. Calibration of such a model is underway utilizing a
data set generated in the laboratory using a Canadian liner clay subjected to
flow of ethanol (Hueckel, 1995). This work is still in progress.
Simulation of coupled chemical-transport effects in clay layers. Limited
simulations of purely advective flow with concentration-sensitive hydraulic
conductivity have shown that the total discharge is largely unaffected by the
changing hydraulic conductivity for contaminant permeation distances less
than half the thickness of the layer. However, for permeation beyond the half
thickness, contaminant transport is accelerated and the breakthrough occurs
significantly earlier than when chemical permeation is absent (30 yr. in the
discussed case). There is further treatment of modeling presented in Section 10.
c. Needs. There is a need for basic experimentation combined with simulation
modeling in the study of the role of effective stress and coupled chemical and
mechanical effects on contaminant transport in indigenous barriers. Also, the
role of effective stress on the hydraulic conductivity of fractured media needs
to be studied, both in laboratory and in the field.
7.9 EFFECTS OF MAN-MADE DISCONTINUITIES

7.9.1
Background
Mans activities can affect the properties of natural indigenous geologic

formations. One man-made activity involves the perforation of aquitards
with holes (such as borings or wells) or objects (such as foundation piling).
The integrity of indigenous barriers (relative to vertical contaminant transport)
is threatened by: pre-existing wells (known or unknown); wells and borings
drilled for site characterization, monitoring, and/or remediation; deep
structures, such as foundation piles; and stress relief effects from nearby deep
excavations. Investigations have been conducted on methods to locate
abandoned wells and borings and on ways to minimize the potential for
169
promoting contaminant migration as a result of drilling or well construction

activities. The flow through an open man-made hole that penetrates an
aquitard can dwarf the advective flow through the aquitard itself. For
example, the discharge through a 10 cm hole in a 6-meter thick clay stratum
is on the same order as the advective flow through 100,000 square meters of
the intact clay (under similar hydraulic gradients). However, wells and
boreholes are not the only potential pathways for contaminant migration.
There are numerous instances of utility trenches and pipe trenches acting as
conduits for lateral movement of fluids. Even foundation units have been
suspected as providing pathways for contaminants.
7.9.2
State of Practice: Wells and Boreholes
a. Problem Sources. The major problem is from wells abandoned without

proper sealing. The problem is not new; it has been known for almost 90
years. For example, Bowman (1906) describes how upward and downward
cross-contamination of aquifers can result from dissolved chemicals entering
improperly abandoned borings and deteriorated well casings. He
recommended the following: (1) plugs and packers be used to prevent crosscontamination along wells or borings; (2) accurate geologic logs be made for
every well drilled; (3) the yield and quality of groundwater from each zone
be examined; (4) the hydraulic heads in each water-bearing zone be
determined; (5) a competent casing should be installed and maintained over
time, using suitable experiments as needed to check well integrity; (6)
consideration be given to potential casing deterioration when changes in water
quality or hydraulic head are observed to occur; and (7) laws be established
to regulate well construction and to minimize the potential for crosscontamination. Bowmans recommendations remain valid today.
During the past 20 years, the consequences of contaminant migration
through known/unknown pre-existing wells and through wells installed for
site characterization, monitoring, or remediation have been realized at many
sites (U.S. EPA, 1977; U.S. EPA, 1992). Modeling studies have demonstrated
that aquifers can be contaminated by preferential vertical migration at wells
or along other similar pathways (Harrison et al., 1992). The significance of
such contaminant migration depends on a variety of factors, including:
dispersion/dilution of the contaminant and the mobility, persistence, and the
toxicity of the migrating contaminant. The potential magnitude of
contamination associated with preferential migration along vertical conduits
can be evaluated by monitoring and through the use of simulation models.
Overall, the most important well defects include:
well left open,

well annulus or bore improperly grouted, and/or
chemical attack on the well casing or the grout material.
170
The first two defects commonly occur in heavily industrialized areas where
shallow groundwater contamination is more prevalent. The third defect is
more rare, but has been observed in PVC wells that extend through hazardous
waste cells. In one such case, the well was virtually destroyed in a matter of
weeks. Borehole television was used to verify the condition of the well.
b. Locating Abandoned and Missing Wells and Borings. Regulations dealing
with the underground injection of liquid waste require that all abandoned
wells be located within an area of review to minimize the risk of contaminant
short-circuiting. Many remediation sites are in old industrial areas where the
well records, if any, cannot be found. The wellheads are frequently covered,
and relocating the well can be difficult if not impossible. Airphotos,
geophysical surveys, and test pits are the primary location techniques
available. Well location methods described by Aller (1984) and are pertinent
to shallow contamination investigations. At many sites, however, the locations
of all pre-existing wells and borings cannot be determined, even after using
these methods. This is often the case at sites where tens to hundreds (or
more) of borings and wells have been drilled since the 1970s. Well inventory
and integrity management is a significant concern at most sites.
c. Recent Evaluations of Methods and Grouts for Sealing Wells and Boreholes.
Whether plugging and abandoning an old well or installing a new well, the
primary objective is long-term integrity. This is achieved through proper
construction of the well, installation of a full, low-permeability seal, and
regular maintenance of the seal in the presence of a potentially damaging
environment. Recognition of the potential for cross-contamination has spurred
research into methods and materials to adequately seal wells and borings.
The sealing characteristics of selected grouts used for well construction were
recently investigated using a large-scale laboratory model by Edil et al. (1992).
Similarly, Lutenegger and DeGroot (1994) examined methods and materials
used for borehole abandonment. Synopses of these studies are provided
below.
The annular spaces between well casings and adjacent formations must
be properly sealed to eliminate the potential for preferential contaminant
transport along the annulus. Edil et al. (1992) used a laboratory physical
model to assess several popular sealing grout mixtures. The sealants tested
included: (1) Neat Cement (one 94-lb bag of C-150T Type I cement mixed
with 5.5 gallons of water); (2) Bentonite-Cement (5 lbs of Quick-Gel mixed
with 6.5 gallons of water and to this slurry, one 94-lb bag of Type I Portland
cement was added); (3) Volclay (2.1 lbs of Volclay was mixed with each
gallon of water to which 2 lbs of magnesium oxide powder was added to
each 50 lbs of slurry as a setting initiator); and (4) Benseal-Bentonite Slurry
(125 lbs of Benseal was mixed into a slurry containing 30 lbs of Natural Gel
bentonite and 100 gallons of water).
Edils physical model was constructed using a moist sand-filled plexiglass
171
container, 1.5 ft. x 6 ft. x 6 ft. deep (30 x 180 x 180 cm). Completely mixed
batches of drilling mud sealant materials were placed in the annular space
between simulated 4-in (10 cm) ID steel well casing pipes and outer 8-in (20
cm) ID PVC casings. The sealant materials were either injected at the bottom
of the annular space using a tremie tube, or, for the thicker mixtures (bentonitecement and Benseal-bentonite grout slurries), simply poured in the annular
space. After grout placement, the outer PVC casings were slowly lifted from
the sand tank allowing direct contact of the grout slurries with the sand. A
short section of 8-in PVC casing pipe was then pushed one foot below the
surface of the sand at the sand-sealant interface. Infiltration tests were
subsequently conducted (in part using Rhodamine dyed water) for 17 weeks
by allowing water to seep down from the short PVC pipe into the grout
column. At the end of the infiltration tests, the model was taken apart and
the sealants were examined slice by slice to investigate dyed water movement
and the condition of the sealants.
Based on the measured infiltration rates and a finite-element seepage
analysis of the experiments, Edil et al. (1992) found that the hydraulic
conductivities of all of the grouts, except for the Volclay, were on the order
of 10-7 cm/s. The infiltration rate into the Volclay grout seal was two orders
of magnitude greater, apparently due to a separation between the Volclay
and the steel casing, which occurred as the Volclay shrank away from the
well casing. Edil et al. (1992) concluded that: (1) the effectiveness of a well
sealant depends on its structural stability, adherence, and hydraulic
conductivity; (2) the Benseal-Bentonite grout adheres to steel and PVC pipes
and provides an excellent seal because of its low hydraulic conductivity, good
swelling characteristics, and flexibility; (3) neat cement and bentonite-cement
grouts form rigid seals with low hydraulic conductivity and high durability,
however, they allowed some limited infiltration at the seal-casing interface;
(4) the Volclay grout does not adhere sufficiently to the well casing; and, (5)
Quik-Gel bentonite slurries of various viscosities and sand contents, like
the Volclay grout, form poorer seals compared to BensealBentonite, neat
cement, and bentonite-cement slurry grouts.
Lutenegger and DeGroot (1994) reviewed materials and practices
employed to seal/backfill abandoned boreholes, and characterized several
sealant materials on the basis of laboratory tests. Materials examined included
compacted soil cuttings, compacted soil and bentonite mixtures, and bentonite
and cement grouts and mixtures. Based on their evaluation of practices and
materials, Lutenegger and DeGroot (1994) concluded: (1) the seal should have
a hydraulic conductivity that is less than the native soil, on the order of 10-7
cm/s to reduce contaminant movement by advection; (2) the seal material
must have sufficient intrinsic structural integrity to prevent seal loss into the
native host soils; (3) the seal material should be compatible with the native
host soil in order to provide a satisfactory bond at the soil/seal interface; (4)
the seal material should be compatible with instrumentation hardware
materials left in the borehole in order to provide a satisfactory bond at the
172
instrument/seal interface; (5) the seal should be of sufficient length to ensure

that contaminant transport is reduced to an acceptable; and (6) the seal should
be considered a permanent repair.
Several practical considerations should be taken into account when
selecting a borehole seal. The seal must be constructible by drillers and drillers
helpers using readily available equipment and materials. Sealing materials
must be emplaced at appropriate intervals. Generally, it will be important to
place sealant materials in stratigraphic units that act as natural groundwater
protection layers (e.g., aquitards). As a practical matter, it is probably easier
to place the seal over the entire length of the borehole rather than to attempt
to construct a seal over an isolated zone. A longer seal length will serve to
reduce diffusive transport of the contaminant. Normal grout mixes may not
set in highly contaminated situations, especially those involving high
concentrations of dissolved salts. There have been numerous cases where
dissolved salts (ions in solution) prevented hydration of the bentonite and/
or cement. Such conditions require special additives to the mix; proven
additives are available commercially. Finally, material and construction costs
are always important considerations.
In general, the bentonite and cement grouts and mixtures generally used
to decommission boreholes are capable of providing low permeability seals.
Lutenegger and DeGroot (1994) concluded that: (1) native cuttings may be
used to seal boreholes provided that sufficient compaction effort is used or
provided that sufficient bentonite is added (note that this recommendation
may not be suitable at some contamination sites); (2) commercially available
bentonite chips and pellets commonly used to seal boreholes have similar
hydraulic characteristics and can successfully be used provided that they can
be placed within the hole and that sufficient water is available to ensure
hydration; and (3) bentonite slurries, mixed to proportions in the range
recommended by manufacturers (i.e., 20% to 40%), do produce tight seals.
d. ASTM Guides for Decommissioning Wells and Borings and for Monitor
Well Construction. ASTM D5299-92 is the standard guide for decommissioning wells and boreholes. The standard lists criteria for plugging
materials specify that such materials: (1) must not react adversely with
contaminants, groundwater, or geologic media; (2) should have hydraulic
conductivities that are similar to, or less than, the adjacent native formations;
(3) must have sufficient structural strength to withstand anticipated pressures
from native conditions; (4) must maintain sealing characteristics and not
degrade due to chemical interaction, corrosion, dehydration, or other
processes; (5) should not be susceptible to cracking or shrinkage; (6) should
be capable of being placed at the position in the well or borehole where they
are needed and must have properties to reduce their unintended movement
vertically or horizontally; (7) must be capable of forming a tight bond and
seal with well casing and the formation; and (8) must have properties that
eliminate leaching or erosion of the plugging material into the formation.
173
Several procedures are recommended for well and/or borehole plugging.

Selection of a particular method depends, in part, on review and inspection
of site-specific well construction, depth, formations present, and other factors.
Well casings can be removed by either pulling or overdrilling. Depending
upon construction, it may be necessary to leave the casing in place and create
perforations in the screen and casing to allow for plugging material to
penetrate the annular space and the formation. If the grout in the annular
space can be verified to be in good condition, the well can be decommissioned
by cutting the blank casing and filling the screened interval with grout. If a
filter pack is present, it may be necessary to remove the filter pack after
perforating the casing, either by washing or overdrilling. If well construction
conditions are not adequately known, and the site contains hazardous material,
it may not be appropriate to remove the casing and screen, as this may increase
the mobility of the contaminants.
The well may require preconditioning to reduce borehole sloughing. This
usually involves removing mud from the borehole walls (when drilling mud
was used) or stabilizing the borehole prior to plugging (using a drilling fluid).
To achieve an effective seal, the borehole should be free of debris and foreign
matter that may restrict the adhesion of plugging materials to the borehole
wall. One method used to prepare the borehole is to circulate clean water or
a high-quality bentonite slurry through a grout pipe set at the bottom of the
boring until blockages are removed and the formation is stabilized. As soon
as the borehole is prepared, the plugging materials should be slowly injected
via a grout pipe placed at the borehole bottom, forcing fluids upward until
undiluted grout reaches the surface. The grout pipe can be raised at this
point, or when the pumping pressure increases significantly due to the rising
grout column. Grouting of shallow auger holes and other shallow borings
can be accomplished by placing grout through a side discharge grout pipe
that has a funnel attached to the top. As grouting progresses, the pipe is
slowly raised.
ASTM D5092-90 is the standard guide for the design and installation of
groundwater monitoring wells. It provides recommendations on materials
and procedures for installing annular sealants. The guide recommends the
placement of bentonite pellet or slurry seals, with a thickness of 3 to 5 feet,
above the screen interval filter pack, using a tremie pipe if possible. A 6-in. to
1-ft. filter pack can then be placed over the bentonite seal, and grout fill is
then injected under pressure from the bottom up in one continuous operation
until the full strength grout flows onto the ground surface without evidence
of drill cuttings or fluid. The guide also provides guidance on constructing
telescoped, multi-casing wells.
e. Boreholes. All the problems given above for wells, and their solutions,
are also applicable to exploration boreholes. A special problem with
geotechnical and seismic boreholes is that sealing was not required until the
past 10 to 20 years, despite Bowmans (1906) warnings. Therefore, many sites,
174
especially those in industrial areas, have many deep, unsealed old boreholes.
Relocation of these old boreholes is virtually impossible.
The are several detailed summaries of drilling, sampling, and well
construction methods at contamination sites (Aller et al., 1989; Driscoll, 1986;
Hackett, 1987, 1988; Smolley and Kappmeyer, 1991). Drilling at a contaminated
site is most commonly done using hollow-stem augers with splitspoon
sampling (Riggs and Hatheway, 1988). Despite the advantages of drilling
with hollow-stem augers, DNAPL can flow down through the disturbed zone
along the outside of the augers and/or possibly enter the augers through
joints and sink to the boring bottom. Drilling in rock typically poses a
significantly greater risk of promoting vertical DNAPL movement than drilling
in unconsolidated media. This is due to the brittle and heterogeneous,
fractured nature of rocks. Fracture networks in rocks are usually ill-defined.
When drilling in rock, DNAPL can enter and exit the borehole unpredictably
by way of fractures. Drilling can also create or widen fractures in the nearwell environment. There is some risk of vertical DNAPL migration associated
with all drilling methods.
Care should be exercised to avoid causing downward movement of
mobile, perched, DNAPL, or DNAPL-contaminated soil while drilling
through an indigenous barrier layer. Similarly, DNAPL may sink
preferentially along the inside or outside of a well. Specific conditions that
may cause downward DNAPL migration include: (1) an open borehole
during drilling and prior to well construction; (2) an unsealed or
inadequately sealed borehole; (3) a well screen that spans a barrier layer
and connects an overlying zone with a perched DNAPL to a lower
transmissive zone; (4) an inadequately sealed well annulus that allows
DNAPL to migrate through the well-grout interface, the grout, the groutformation interface, or along vertically-connected fractures in the disturbed
zone adjacent to the well; and (5) structural degradation of bentonite or
grout sealant, or well casing, due to chemical deterioration by DNAPL or
the groundwater environment (Cohen and Mercer, 1993).
To minimize the risk of inducing DNAPL migration as a result of drilling,
site investigators should: (1) avoid unnecessary drilling within the DNAPL
zone; (2) minimize the time during which a boring is open; (3) minimize the
length of hole which is open at any time; (4) use telescoped casing drilling
techniques to isolate shallow contamination zones from deeper zones; (5)
utilize a site conceptual model (knowledge of stratigraphy and contaminant
distribution), and carefully examine subsurface materials brought to the
surface as drilling progresses, to avoid drilling through a barrier layer
beneath DNAPL; (6) consider using a dense drilling mud to prevent DNAPL
from sinking down the borehole during drilling (care must be taken to avoid
deleterious mud recirculation); (7) consider drilling horizontal wells; (8)
consider using less invasive site investigation techniques, such as use of the
cone penetrometer; (9) select optimum well materials and grouting methods
based on consideration of site-specific chemical compatibility; and (10) if
175
the long-term integrity of a particular grout sealant is questionable, consider

placing alternating layers of different grout types and sealing the entire
distance between the well screen and the surface to minimize the potential
for vertical migration of DNAPL.
At many sites, the DNAPL or highly-contaminated zones can be
adequately characterized while limiting drilling to shallow depths.
Characterization of deeper units can be accomplished by locating deeper
borings and wells outside the highly-contaminated zone.
7.9.3
Field Performance
There are numerous cases where unsealed wells or boreholes have been
pathways for significant or even massive contaminant transport (USEPA,
1977). However, there is little field documentation on the performance of
properly sealed wells and boreholes. What documentation exists suggests
that the techniques discussed above are adequate for preventing significant
contaminant transport along these man-made conduits.
7.9.4
Assessment
As indicated above, the sparse available data indicates that current technology
is adequate for installing proper seals on boreholes and wells. The problem
lies in finding old wells and boreholes. The current technology for locating
old boreholes and wells is inadequate.
7.9.5
Needs
The primary needs within are for improved techniques for detecting and
locating boreholes and wells and for greater documentation of field studies
that describe sealing behavior.
7.9.6
Foundation Units
Deep foundation units, e.g., drilled shafts and driven piles, can and often do
penetrate into or through aquitard strata (potential indigenous barriers).
Although these deep foundation units have been suspected by some
regulatory authorities as providing pathways for vertical migration, there is
little knowledge of this occurring in the field. A literature search yielded one
oblique reference, where such a pathway was suspected, but the migration
was definitely from a nearby pond (Campbell, et.al., 1984). A model study by
Hayman, et.al., 1993, indicated no migration along steel or concrete piles and
virtually no contaminants carried along with the piles. Untreated wood piles
were found to be subject to wicking the contaminants. Other experiments
(A.D. Little, Inc., 1981) have shown that stainless steel may be preferentially
wetted by organic DNAPLs (thus, possibly facilitating DNAPL movement
176
along steel pilings or casings). It should also be noted that the very action of
driving piles sets up high lateral forces against the sides of the piles which
tend to close potential pathways. Generally, a remolded zone is created around
each pile, destroying any secondary structure in clays.
Drilled shafts, however, are often emplaced by open-hole drilling. This
technique reduces stresses in the soils near the shaft, which can open secondary
structure in the adjacent formation. Although the shaft forms a conduit while
it is open, concrete filling should seal against the soil unless there is excessive
shrinkage of the concrete. If the latter occurred, the installed piles would not
develop the skin friction on which their capacities depend. Shafts installed
using mud-drilling techniques might be more appropriate in known or
suspected soil contamination zones.
For the most part, driven piles (other than untreated wood) can be
considered a safe technique in contaminated zones. The situation for deep
foundation units constructed by augering or similar methods is more
ambiguous. The primary need under this subtopic is for field verification of
the conclusions given above.
7.10
SUMMARY AND CONCLUSIONS
The performance of an indigenous barrier depends primarily on:

barrier stratigraphy, which defines likely flow paths,
barrier material properties, which resist contaminant flux, and
the potential field, which drives the flux.
7.10.1
Stratigraphy
Stratigraphic characterization at a site usually consists of grouping the material

types by their hydraulic conductivities (e.g., sands, clays, etc.). A major thrust
of any investigation is the determination of the presence or absence of
transmissive defects in the bottom barrier strata (aquitards). Small defects,
such as pervious layers, fractures, etc., can dominate the flux characteristics
of the barrier.
Current technology for field characterization includes boreholes, cone
penetrometer testing, etc., and are generally adequate for defining gross
stratification. The interpretive methods, which include computer-assisted
data basing, are also adequate for the gross situation. The major problem is
in defining the small defects; additional work is needed in this area.
Overall, a thorough investigation can usually characterize stratigraphy
to an adequate degree. The effect of stratigraphy on flux can probably be
evaluated to within 20% to 30% in unfractured media, with less accuracy
(say 50% to 100%) in fractured media.
7.11 REFERENCES
7.10.2
177
Material Properties
The material properties which influence contaminant flux are hydraulic

conductivity, diffusion coefficients, and retardation coefficients. Hydraulic
conductivity can be directionally dependent in anisotropic media and also
highly variable and scale dependent in heterogeneous media. Most of the
knowledge regarding diffusion and retardation coefficients is based on
laboratory studies; field data are scarce. The major development needs are to
acquire more field data relating to these properties, including how they are
affected by the subsurface chemical regime.
Definition of the material properties is the least accurate of the three major
factors affecting the performance of indigenous barriers. The level of accuracy
is probably within one order of magnitude for hydraulic conductivity, even
in the absence of chemical or thermal effects. The level of accuracy in the
determination of the diffusion and retardation coefficients is unknown, due
to lack of field verification.
7.10.3
Potential Fields
The driving forces for advective and diffusive transport are the gradients of
hydraulic head, chemical concentration, and temperature. Current technology
is adequate for defining these potentials at a point. However, definition of
potential fields is less accurate. Nevertheless, potential fields can be
determined to a far better level of accuracy than can material properties. The
accuracy is probably similar to that of stratigraphy characterization; i.e., within
20% to 50% of the actual situation.
7.10.4
Man-Made Discontinuities
Man-made defects in natural (indigenous) bottom barriers can provide

contaminant flux pathways which virtually negate the use of a naturally
occurring strata as a physical barrier. The major man-made problem is
improperly sealed wells and/or exploratory borings. Although sealing
methods are adequate, the problem is in detecting perforations. The existing
technology for detection of perforations through indigenous barriers is not
adequate; work is necessary in this area.
7.11
REFERENCES
Acar, Y.B., A. Hamidon, S.D. Field, and L. Scott, (1985). The Effect of Organic
Fluids on Hydraulic Conductivity of Compacted Kaolinite, Hydraulic
Barriers in Soil and Rock, American Society for Testing and Materials,
Philadelphia, pp. 171-187.
178
Aller, L. (1984). Methods for Determining the Location of Abandoned Wells,

EPA-600/2-83-123, NWWA, Dublin, Ohio, p. 130.
Aller, L., T.W. Bennett, G. Hackett, R.J. Petty, J.H. Lehr, H. Sedoris, and D.M.
Nielsen, (1989). Handbook of Suggested Practices for the Design and
Installation of Ground Water Monitor Wells, NWWA, Dublin, Ohio, p.
398.
Alther, G., J.C. Evans, H.Y. Fang, and K. Witmer, (1985). Influence of Inorganic
Permeants upon the Permeability of Bentonite, Hydraulic Barriers in
Soil and Rock, American Society for Testing and Materials, Philadelphia,
pp. 64-73.
Anderson, D.C, W.K. Brown, and J.C. Thomas, (1985). Conductivity of
Compacted Clay Soils to Water and Organic Liquids, Waste Management
Research, vol. 3, pp. 339-349.
Baldi, G., M. Borsetto, T. Hueckel and A. Peano, (1987). Coupling of ThermoPlastic and Hydraulic Effects in a Clay Repository: Near Field Analysis,
in Coupled Processes Associated with Nuclear Waste Repositories, ed.
C.F. Tsang, Academic Press, Orlando, pp. 565-580.
Benson, C.H., and M.A. Othman, (1993). Hydraulic Conductivity of
Compacted Clay Frozen and Thawed In-Situ, Journal of Geotechnical
Engineering, ASCE, vol. 19, no. 2, pp. 276-294.
Benson, R.C., and L. Yuhr, (1995). Geophysical Methods for Environmental
Assessment, Proc. GeoEnvironment 2000, ASCE GSP No. 46, ASCE, 1995,
pp. 57-76.
Birgersson, L., and Neretnieks, (1990). Diffusion in the Matrix of Granite
Rock: Field Test in the Stripa Mine, Water Resources Research., V. 26,
No. 11, p. 2833-2843.
Bouwer, H., and R.C. Rice, (1976). A Slug Test for Determining Hydraulic
Conductivity of Unconfined Aquifers with Completely Penetrating Wells,
Water Resources Research. V. 12, No.3, pp.428-428.
Bowders, J.J., and D.E. Daniel, (1987), Hydraulic Conductivity of Compacted
Clay to Dilute Organic Chemicals, Journal of Geotechnical Engineering,
vol. 113, no. 12, ASCE, pp. 1432-1448.
Bowman, I, (1906). Problems of Water Supply Contamination, USGS Water
Supply Paper #160, U.S. Geological Survey, 1906.
7.11 REFERENCES
179
Bradbury, K. R., and M.A. Muldoon, (1990). Hydraulic Conductivity

Determinations in Unlithified Glacial and Fluvial Materials, Ground
Water and Vadose Zone Monitoring, ASTM STP 1053, D.M. Nielsen and
A.I. Johnson, Eds., American Society for Testing and Materials, p. 138151.
Butler, J. J., and Z. Hyder, (1994). An Assessment of the Nguyen and Pinder
Method for Slug Test Analysis, Ground Water Monitoring and
Remediation, V. 14, No. 4, p. 124-131.
Campbell, P.L., R.C. Bost, and R.W. Jacobsen, (1984). Subsurface Organic
Recovery and Contaminant Migration Simulation, Proc. 4th National
Symposium on Aquifer Restoration and Groundwater Monitoring, NWWA,
Columbus, OH, May, 1984.
Cohen, R.M. and J.W. Mercer, (1993). DNAPL Site Evaluation, C.K. Smoley,
Boca Raton, Florida.
Cooper, H. H., J. D. Bredehoeft, and I. S. Papadopulos, (1967). Response of a
Finite Diameter Well to an Instantaneous Charge of Water, Water
Resources Research, Vol. 3, No. 1, pp. 263-269.
Czurda, K. A., and J. J. Wagner, (1991). Cation Transport and Retardation
Processes in View of the Toxic Waste Deposition Problem in Clay Rocks
and Clay Liner Encapsulation, Engineering Geology, Vol. 30, No.1, p.
103-113, March 1991.
Daniel D.E., D.C. Anderson, and S.S. Boynton (1986). Fixed Wall Versus
Flexible Wall Permeameters, in Hydraulic Barriers in Soils and Rocks,
ASTM STP 874, pp. 107-126.
Daniel, D. E., (1994). State of the Art Laboratory Hydraulic Conductivity
Tests for Saturated Soils, Hydraulic Conductivity and Waste Contaminant
Transport in Soil, ASTM STP 1142, D.E. Daniel and S. J. Trautwein, Eds.,
American Society for Testing and Materials, p. 30-71.
Dal Pra, A., and A. Magnini, (1993). Barriers Geologiques Naturelles dans la
Plaine Alluviale Venetienne, Proc. GeoConfine 93, Montpellier, France,
June, 1993, p. 27-32.
De Olmo,C., V. Fioravante, F. Gera, T. Hueckel, J-C. Mayor, and R. Pellegrini,
(1994). Thermomechanical Properties of Deep Argillaceous Formations,
Engineering Geology, in print.
Dorhofer, G., (1993), The Role of Natural Geologic Barriers for the Siting of
180
Landfills in Germany, Proc. GeoConfine 93, Montpellier, France, June, 1993,

pp. 39-46.
Driscoll, F.G. (1986). Groundwater and Wells, Johnson Division, St. Paul,
Minnesota, p. 1089.
Edil, T.B., M.M.K. Chiang, L.T. Lan, and T.V. Riewe, (1992). Sealing
Characteristics of Selected Grouts for Water Wells, Ground Water, vol.
30, no. 2, pp. 351-361.
Fernandez, F., and R.M. Quigley (1991). Controlling the Destructive Effects
of Clay-Organic Liquid Interactions by Application of Effective Stresses,
Canadian Geotechnical Journal, vol. 28, pp. 388-398.
Fredlund, D.G., and H. Rajardo, (1993). Soil Mechanics for Unsaturated Soils,
John Wiley & Sons, New York.
Grim, R.E., (1968). Clay Mineralogy, 2nd ed., McGraw-Hill Book Company,
Inc., NY.
Hackett, G. (1987). Drilling and Constructing Monitoring Wells with HollowStem Augers, PartI, Ground Water Monitoring Review, vol. 7, no. 4, pp.
51-62.
Hackett, G., (1988). Drilling and Constructing Monitoring Wells with HollowStem Augers, Part2, Ground Water Monitoring Review, vol. 8, no. 1, pp.
60-68.
Harrison, B., Sudicky, E.A., and J.A. Cherry (1992). Numerical Analysis of
Solute Migration Through Fractured Clayey Deposits into Underlying
Aquifers, Water Resources Research, vol. 28, no. 2, pp. 515-526.
Hayman, J.W., R.B. Adams, and R.G. Adams, (1993). Foundation Piling as a
Conduit for DNAPL Migration, presented at 86th Annual Meeting, Air
and Waste Management Association, Denver, CO, June, 1993.
Henrion P.N., A. Fontayne, Y. Jiade and M. van Gompel, (1988). Impact of
Radiolysis on the Migration of Radionuclides in the Boom Clay
Formation,Radioactive Waste Management and the Nuclear Fuel Cycle,
vol. 11,no. 1, pp.41-60.
Hermann, D., (1978). Alum Mud Disposal Site - Two Case Histories,
Presented to ASCE National Conference, Chicago, IL.
Hermann, D., (1985). Effect of Acidic Leachate of Clay Permeability,
7.11 REFERENCES
181
Proceedings of the Eighth Annual Madison Solid Waste Conference, University

of Wisconsin, Madison, WI, 1985.
Herzog, B. L., (1994). Slug Tests for Determining Hydraulic Conductivity of
Natural Geologic Deposits, Hydraulic Conductivity and Waste
Contaminant Transport in Soil, ASTM STP 1142, D.E. Daniel and S. J.
Trautwein, Eds., American Society for Testing and Materials p. 95-110.
Herzog, B. L., and W. J. Morse, (1986). Hydraulic Conductivity at a Hazardous
Waste Disposal Site: Comparison and Laboratory- and Field-Determined
Values, Waste Management and Research, V. 4, No. 2, p 177-187.
Hinsby, K. L., D. McKay, P. Jorgensen, M. Lenczewski, and C. P. Gerba, (1995).
Fracture Aperture Measurements and Migration of Solutes, Colloids and
Immiscible Creosote in a Column of Clay-Rich Till, Ground Water,
accepted pending minor revisions, July, 1995.
Horton, R., and M. L. Thompson, (1986). Determination of Effective Porosity
of Soil Materials in Proceedings of the 12th Annual Research Symposium of
Land Disposal, Remedial Action, Incineration and Treatment of Hazardous Waste.
U.S. EPA, Cincinnati, OH EPA/600/9-86/022, p.489.
Hueckel T., (1992). Water - Mineral Interaction in Hydro-mechanics of Clays
Exposed to Environmental Loads: a Mixture Approach, Canadian
Geotechnical Journal, v.29, pp. 1071-1086.
Hueckel T., (1995). Strain and Contamination History Dependence in ChemoPlasticity of Clays Subjected to Environmental Loads, in Numerical
Models in Geomechanics, ed. by G.N. Pande and S. Pietruszczak, A.A.
Balkema, Rotterdam, pp. 329-336.
Hueckel T. and R. Pellegrini, (1992). Effective Stress and Water Pressure in
Saturated Clays during Heating-Cooling Cycles, Canadian Geotechnical
Journal, v.29, 1095-1102.
Hvorslev, M.J., (1951), Time Lag and Soil Permeability in Ground-Water
Observation, Bulletin 36, Waterways Experiment Station, U.S. Army
Corps of Engineers, Vicksburg, MS.
Johnson, R. L., J. A. Cherry, and J. T. Pankwo, (1989). Diffusive Contaminant
Transport in Natural Clay: A Field Example and Implication for ClayLined Waste Disposal Sites, Environmental Science and Engineering,
Vol. 23, No. 3, p. 340-349.
Krapac, I. G., K. Cartwright, B. R. Hensel, B. L. Herzog, T. H. Larson, S. V.
182
Panno, J. B. Risatti, W. J. Su and K. R. Rehfeldt, (1991). Construction

Monitoring and Performance of Two Soil Liners, Illinois State Geological
Survey Environmental Geology Series 141, p. 118.
Kueper, B. H., and D. B. McWhorter, (1991). The Behavior of Dense, NonAqueous Phase Liquids in Fractured Clay and Rock, Ground Water, V.
29, p.716-728.
Lawrence, T.A., and G.P. Boutwell, (1990). Predicting Stratigraphy at Landfill
Sites Using Electromagnetics, ASTM STP 1070, ASTM, Philadelphia, PA,
1990.
Leoni L., C. Polizzano, and F. Sartori, (1986). Nuclear Waste Repositories in
Clays: The OrciaticoMetamorphic Aureole Analogy, Applied Clay Science,
vol. 1, pp. 385-408.
Lutenegger, A.J., and D.J. DeGroot (1994). Hydraulic Conductivity of Borehole
Sealants, Hydraulic Conductivity and Waste Contaminant Transport in
Soil, D.E. Daniel and S.J. Trautwein, eds., ASTM, Philadelphia, pp. 439461.
Ma, C.M. and T. Hueckel, (1992). Stress and Pore Pressure in Clay Mass
Subjected to Heat of Radioactive Waste: a Numerical Simulation,
Canadian Geotechnical Journal, v.29. 1087-1094.
McKay, L. D., R. W. Gillham, and J. A. Cherry, (1993). Experiments in a
Fractured Clay-Till, 2. Solute and Colloid Transport, Water Resources
Research, V. 29, No. 12, p. 3879-3890.
Mitchell J.K. and F.T. Madsen (1987). Chemical Effects on Clay Hydraulic
Conductivity, in Geotechnical Practice for Waste Disposal 1987, ed., R.D.
Woods, ASCE STP 13, pp. 87-116.
Nguyen, V., and Pinder, G. F., (1984). Direct Calculation of Aquifer Parameters
in Slug Test Analysis, Groundwater Hydraulics, Water Resources
Monograph 9, J. Rosensheim and G. D. Bennet, Eds., American
Geophysical Union, p. 222-239.
Othman M.A., C.H. Benson, E.J. Chamberlain and T.F. Zimmie, (1994),
Laboratory Testing to Evaluate Changes in Hydraulic Conductivity of
Compacted Clays caused by Freeze-Thaw: State-of-the-Art, Hydraulic
Conductivity and Waste Containment Transport in Soil, D.E. Daniel and
S.J. Trautwein eds, ASTM, STP 1142, pp. 227-254.
Peterson, S.R. and G.W. Gee, (1985). Interactions Between Acidic Solutions
7.11 REFERENCES
183
and Clay Liners: Permeability and Neutralization, Hydraulic Barriers in

Soil andRock, American Society for Testing and Materials, Philadelphia,
pp. 229-245.
Peyton, G. R., J. P. Gibb, M. H. LeFaivre, J. D. Ritchey, S. L. Burch, and M. J.
Barcelona, (1996). Effective Porosity of Geologic Materials, Proceedings
of the 12th Annual Research Symposium of Land Disposal, Remedial Action,
Incineration and Treatment of Hazardous Waste. U.S. EPA, Cincinnati, OH.
EPA/600/9-86/022, p. 21-28.
Picard, J. M., B. Bazargan and G. Rousset, (1994). Essai Thermo-hydromecanique dans une Argile Profonde, Science et Technique Nucleaires,
Commission European, EUR 15482 FR.
Pusch R. and N. Guven, (1990). Electron Microscopic Examination of Hydrothermally Treated Bentonite Clay, Eng. Geology, vol. 28, pp. 203-214.
Quigley, R. M., F. Fernandez, and R.K. Rowe (1988). Clayey Barrier
Assessment for Impoundment of Domestic Waste Leachate (Southern
Ontario) Including Clay-Leachate Compatibility by Hydraulic
Conductivity Testing, Canadian Geotechnical Journal, vol. 25, no. 8, pp.
574-581.
Riggs, C.O., and A.W. Hatheway (1988). Groundwater Monitoring Field
Practice An Overview, Ground Water Contamination Field Methods,
ASTM, Philadelphia, pp. 121-136.
Roy, W. R., I. G. Krapac, S. F. J. Chou, and R. A. Griffin, (1991). Technical
Resource Document Batch-Type Procedures for Estimating Soil
Adsorption of Chemicals, U.S. EPA Risk Reduction Engineering
Laboratory, Cincinnati, OH, p. 100.
Shackleford, C. D., (1991). Laboratory Diffusion Testing for Waste Disposal:
A Review, Journal of Contaminant Hydrology, V. 7, No. 3, p. 177-217.
Shackelford, C. D., (1994). Waste Soil Interactions that Alter Hydraulic
Conductivity, in Hydraulic Conductivity and Waste Containment
Transport in Soil, D.E. Daniel and S.J. Trautwein eds, ASTM, STP 1142,
pp. 111-168.
Shinn, J.D., W.L. Bratton, and J.L. Bratton, (1995), Direct Penetration
Technology for Geotechnical and Environmental Site Characterization,
Proc. GeoEnvironment 2000, ASCE GSP No. 46, ASCE, pp. 105-122.
Smolley, M. and J.C. Kappmeyer, (1991). Cone Penetrometer Tests and
184
Hydropunch Sampling: A Screening Technique for Plume Delineation,

Ground Water Monitoring Review, vol. 11, no. 2, pp.101-106.
Snow, D. T., (1969). Anisotropic Permeability of Fractured Media, Water
Resources Research, V. 29, p.716-728.
Sudicky, E. A., and R. G. McLaren, (1992). The Laplace Transform Galerkin
Techniques for Large-scale Simulation of Mass Transport in Discretely
Fractured Porous Formation, Water Resources Research, V. 28 No. 2,
p.499-514.
U.S. Environmental Protection Agency, (1977). Impact of Abandoned Wells
on Ground Water, EPA-600/3-77-095, p. 52.
U.S. Environmental Protection Agency, (1992). Dense Nonaqueous Phase
Liquids A Workshop Summary, EPA/600/R-92/030.
Trautwein, S. J. and G. P. Boutwell, (1994). In-Situ Hydraulic Conductivity
Tests for Compacted Soil Liners and Caps, Hydraulic Conductivity and
Waste Contaminant Transport in Soil, ASTM STP 1142, D. E. Daniel and
S. J. Trautwein, Eds., American Society for Testing and Materials, p.184-223.
185
SECTION 8
ARTIFICIALLY EMPLACED FLOORS AND BOTTOM BARRIERS
prepared by
Mary E. Peterson, Pacific Northwest National Laboratory

and Richard C. Landis, DuPont Co.
George Burke, Hayward Baker

Martin Cherrington, Cherrington Corp.
Brian Dwyer, Sandia National Laboratories
Bruno Gemmi, Fondazioni Speciali, Italy
Alex Iskandar, U.S. Army Corps of Engineers
Guy Loomis, Idaho National Engineering Laboratory
Ray Peters, Scientific Ecology Group
Robert Waters, Tensor
Peter Yen, Bechtel Corp.
8.1 BACKGROUND
Although emplacement of floors and bottom barriers is a potentially viable
technical option for environmental remediation throughout the United States
(U.S.), there is little experience and few techniques are available for the
construction of floors and bottom barriers. Potential functions for these
barriers include the prevention of contaminant migration from subsurface
contaminant sources, e.g., underground storage tanks, landfills, soil columns,
trenches, and at inactive hazardous waste disposal sites. There are challenges
to constructing bottom barriers (depicted in Figure 8-1) that include:
heterogeneities in the subsurface; limited site characterization data; buried
structures, e.g., buried debris, engineered landfills, tanks, and utilities;
complicated site geology and hydrology; limited access; limited available
technologies for construction; and difficulties in verifying the integrity of the
constructed bottom barrier.
Important technical parameters affecting the successful installation of a
bottom barrier include: (1) defining the extent of contamination (for
establishing the barrier emplacement depth and length); (2) determining soil
porosity, hydraulic conductivity, and moisture content; (3) evaluating site
geology to assess drilling and emplacement requirements; (4) selecting the
barrier material, whether it is soil and cementitious grouts, chemical grouts,
or frozen soil moisture; and (5) establishing barrier performance requirements.
In addition, many applications will require that the bottom barrier be
186
interlocked with a vertical barrier. These technical considerations are

addressed in other sections of this report (primarily Sections 2, 3, 5, and 9)
and the information presented in those sections must be considered along
with the techniques presented in this section to produce a competent barrier
system. There is very limited field experience with the emplacement of these
barriers and with the verification of their construction and performance.
Grout
Permeation
grout layer
Figure 8-1 Tank farm with grout enclosure barrier.

Techniques for creating a bottom barrier include: permeation grouting,
jet grouting, directional drilling, hydrofracturing, block displacement, sheet
pile cutoff combined with injection grouting, the Zblin System,
microtunneling with injection grouting, and rectangular microtubes pushed
side by side (Rumer and Ryan, 1995). Recently, other techniques have been
identified and advancements in jet and permeation grouting have occurred.
These recent advancements in the state-of-the-practice for emplacement of
bottom barriers are described in this section.

Advancements of the state-of-the-practice can be attributed to the
development of new grout materials, the recognition that technologies for
establishing bottom barriers warrant development, and the establishment of
linkages between U.S. and international industrial expertise. New grout
materials (see Sections 2 and 9) have been developed with controllable flow
properties and solidification times, making them more compatible with the
expected site-specific variations in soil characteristics. Advancements have
been made in the techniques for jet grouting and permeation grouting, making
them more suitable for bottom barrier construction. However, the key
technical and cost issues affecting the implementation of these techniques are
related to the drilling of the boreholes. A key to successful directional drilling
8.3 GROUTED BOTTOM BARRIERS
187
is the ability to accurately place the boreholes. A guidance system for

monitoring and controlling subsurface boreholes is available and has been
used in the oil fields and for river crossings. The availability of such guidance
systems is an important element in the creation of bottom barriers.

8.3.1
Construction of Bottom Barriers by Permeation Grouting
Recently developed grout materials, specifically chemical or particulate gelling

agents, have significantly advanced the technology of permeation grouting.
Previous attempts at solidification, encapsulation, and barrier construction
have been hampered by the inability of grout materials to flow into fine media,
such as fine-grained sand, silt, and clay. Due to subsurface heterogeneities or
stratification, radial grout travel could vary from location to location. Because
of the apparent lack of control in the placement of conventional permeation
grouts, it has been less favored as a means for constructing a barrier at
hazardous waste sites. Factors affecting the penetration of permeation grouts
include: grout viscosity, gel-time control, injection pressure, host material grain
size, and porosity.
Figure 8-2 Exposed polysiloxane (PSX) grouted plume. The whitish strands of
material in the solidified plume indicate the cross-linked solid PSX.
188
Specific new grout materials that have advanced this technology include
polysiloxane and colloidal silica. These materials have a viscosity less than
that of water, and testing has demonstrated long-term stability in a wide range
of chemical environments. Their low viscosity (< 1 cP) permits these grouts
to travel outward as a wetting front, independent of injection pressure. The
low viscosity, together with controlled gel-times ranging from several minutes
to over 24 hours, enables more efficient placement, by using fewer drill holes
than in normal construction practice. The limited field testing completed to
date has produced a uniform solidified grout bulb (up to 9 ft in diameter) in
heterogeneous alluvial material. The solidified grout bulb is shown in Figure
8-2. This testing demonstrated the ability to permeate fine sands, silts, and
voids in clays. Measurements made in the laboratory have shown that
hydraulic conductivities of 10-8 cm/s were achieved. This is significant
because previously, permeation grouting was only able to achieve hydraulic
conductivities of 10-5 cm/s. The next phase of testing aims to emplace a bottom
barrier in heterogeneous alluvial materials using permeation grouting and
these grout materials. [See European Chemical News, 1995 and Environmental
Engineering World, 1995 for descriptions of this technology.]
8.3.2 Construction of Bottom Barriers Using Directional Drilling and Jet
Grouting
The concept of a thin diaphragm wall system with various inclines, interlocks,
and column elements was developed in Italy and represents an advancement
in the application of high pressure jet grouting. The coupling of this system
with directional drilling has enabled the emplacement of three experimental
floors. This technology is depicted in Figure 8-3. Using this technique, a
vertical, angled, or curved barrier can be created at a depth or length up to or
greater than 1000 m. Both cylindrical and planar barrier shapes can be created
using either single- or multiple-rod systems. It uses standard pumping
equipment, but requires a dedicated directional drill rig and a location/
guidance system. A wire-based location and steering system has been used
in the past. The operation involves the drilling of a horizontal or angled
borehole up to the depth or the length required. The rod-string (generally 60
to 90 mm in diameter) is equipped at the bottom with the drill bit and a nozzle
holder device. Initially, the rod-string is extended through the subsurface to
the ground surface at the other end of the target zone for the emplacement of
a floor or to the target depth, when a complete floor is not needed. Next, the
rod-string is extracted at a constant speed, without rotation, while pumping
the grouting material through the nozzle at the required pressure. Rotational
movement can be added at any time during the extraction, if the application
requires it. The excess fluid and soil is forced to the surface around the drill
rod. The nozzles are 160 apart, thus enabling the creation of two thin
diaphragm walls roughly 6 ft long. The pilot boreholes are repeated on centers
of approximately 10 ft to create intersecting thin diaphragm walls that form a
189
containment floor or wall.

The basic elements of this containment system are the panels which can
be interlocked with other panels to form horizontal and inclined curved thin
diaphragm wall barriers. Because of the greater length achievable with this
technique, a barrier can be emplaced under and around a large contaminated
area without drilling through the contaminated zone. The barrier
emplacement may terminate below the surface at an intermediate layer or
may extend to the ground surface. Single- and double-rod systems are used
extensively and quite effectively in horizontal and sub-horizontal applications.
The radius or depth of penetration is increased with the double-rod system.
Horizontal thin diaphragm
wall floor emplacement
Thin diaphragm walls
Figure 8-3 Creation of a bottom barrier using thin diaphragm walls and
directional drilling technique.
a. Advantages. The advantages of using this technique are as follows:

The technique utilizes the soil at the site to form the barrier. The only
secondary wastes produced are the soil spoils exiting around the rod string
during the treatment. However, this material will have been already
partially treated with the grout material and therefore, is likely to be less
hazardous.
The high pressure system allows the injected volume of grout and the
shape of the barrier to be controlled. This is an advantage over permeation
grouting techniques.
This technique is flexible and enables the creation of barriers over specified
distances in the subsurface.
This system can be used to repair cracks, failures, or defects in the barrier,
as well as to seal leaks in other barriers or structures. The barrier is repaired
190
by drilling a small diameter hole in the area of the failure, followed by

the creation of either a column or wall to seal or close the crack. The
same approach may be used in the future to increase the depth or extend
the barrier.
The directional drilling technique is a dedicated system and is able to
achieve considerable distance, allowing for emplacement of a barrier
outside the contaminated portion of the site.
With modifications, the double-rod system could simultaneously inject
two chemical reagents that are designed to react only after injection into
the subsurface.
The double-rod systems can be modified to carry the wires and
transmission system for the control/guidance system needed to guide
directional drilling.
b. Limitations.
The limitations to this technique are as follows:
The efficiency and continuity remain dependent on site-specific soil

conditions. The technique is feasible in loose soil (silts, sands and gravels)
and is less feasible in rocks and fractured systems. Pure gravels are not
suited for using the directional drilling system.
The presence of subsurface obstacles or structures require special attention,
both in the drilling phase and the injection phase, in order to ensure proper
drilling and sealing with the appropriate barrier shape.
As with any subsurface barrier, the consistency, dimensions, and
continuity of the barrier usually cannot be directly observed. Therefore,
verification is based on results from field tests, the recording and
evaluation of data during operations, and the post operational testing,
including coring and geophysical techniques (sonic, radar, etc.)
Using the control/guidance system, the actual tolerance depends on the
precision of the electromagnetic system used and the number of
intermediate checkpoints.
The rod-string may bind and become stuck in the soil, depending on the
friction generated between the soil and the rod string, the length of the
borehole, the type and stability of drilling fluid, and the soil response to
the high pressure injection. This problem can be overcome by increasing
the pull force of the drill rig, casing the hole to ensure circulation, and
creation of an auxiliary pressure relief hole. These solutions can increase
placement costs and may extend construction time.
8.3.3 Construction of Bottom Barriers by Non-Directional Jet Grouting
Conventional jet grouting has been identified as one technique for creating a
bottom barrier (Burke and Brill, 1993 and Burke and Welsh, 1995).
Conventional jet grouting refers to techniques which use non-directional
drilling to emplace grout columns. Barriers produced during jet grouting
191
can differ greatly by varying the many parameters that are part of the grouting
process, including: the drilling parameters (drill hole annulus size, lift speed
and consistency, and rotation speed and consistency); injection parameters
(number of injection nozzles and their size, velocity of injected fluids, and
volumes of injected fluids); and grout properties (type of grout, consistency
of the mix, and viscosity of the mix). One technique to create bottom barriers
using conventional jet grouting is depicted in Figure 8-4.
Barrier Floor Created by Interconnecting
Jet Grouted Columnar Elements
Vertical columnar
jet grouting
Horizontal columnar
jet grouted floor
Figure 8-4 Creation of a bottom barrier using conventional jet grouting

technique.
In addition, the mono-directional thin diaphragm technique can be
employed to create a slanted wall beneath a contaminated zone (shown in Figure
8-5). This technique was developed in Italy and is a variation of the technique
described in the previous section. Using this technique, an angled, but straight,
structure can be created to a depth or length of 20-30 m. The barrier shape can
be either cylindrical, planar, or some combination. The monodirectional thin
diaphragm wall technique uses the double-rod system and standard drilling
and pumping equipment. The operation involves drilling the vertical or angled
borehole using a single-rod string. Once the required depth or length is reached,
the grout mixture is ejected through the nozzles. The position of the nozzle is
confirmed and extraction of the nozzle then begins at a constant withdrawal
speed with no rotation. Nozzle rotation can be added to the technique at any
point during extraction, depending on the specific requirements of the project.
The monodirectional thin diaphragm wall technique is used for applications
requiring hydraulic control (seepage and filtration) without any substantial
static requirements. Therefore, the important performance parameters include
barrier continuity, competent joints between thin diaphragm wall elements,
and optimization of the thickness and amount of injected material. The double-
192
rod system will maximize the dimensions of the barrier without using water.
This enables a good sealing effect between the thin diaphragm wall elements.
It avoids the clear and defined cutting lines usually associated with the
disruption action of the water in the triple-rod system.
Vertical thin diaphragm walls
Figure 8-5 Creation of a slanted wall using the mono-directional thin

diaphragm wall technique.
The basic elements of the monodirectional system are panels that can be
interlocked with other panels for the formation of an inclined thin-wall barrier.
During the extraction phase, the monodirectional technique can be alternated
with the standard rotational technique so that a continuous structure can be
formed with both panel and column elements using the same borehole. The
formation of columns and panels can be alternated on different holes creating
a mixed structure of full height single panels interlocked with full height single
columns. This technique allows for the enclosure of a contaminated zone
with a V-shaped barrier or connection with a bottom barrier which will
broaden the potential applications for bottom barriers.
a. Advantages. The advantages of using conventional jet grouting include
the following:
It is effective in a wide range of soil types, at great depths (up to 150 ft),
and above and below the water table.
The process can be adjusted to compensate for heterogeneities and
stratified soil conditions.
The size of equipment can vary, depending on the needs of the project.
Unconfined compressive strengths of 10-1000 psi and hydraulic
conductivities of 10-5 to 10-7 cm/s can be achieved.
The geometry of the barrier can take many shapes and lends itself to
repair, when the damaged portion of the barrier has been located.
193
Additional advantages for using the thin diaphragm wall technique were
provided in the previous section.
b. Limitations. The limitations for using conventional jet grouting are
identified below.
It is only cost effective on soils that are easily erodible such as sands, silty
sands, etc.
A mixture of some soil and grout returns to the surface which is a waste
by-product that may require disposal if the barrier is constructed in a
highly contaminated area.
The waste return is essential for the creation of uniform soilcrete. If the
return is compromised for even short periods, it is likely that
hydrofracturing will result instead of erosion, causing inconsistent barrier
quality and geometry.
The sequence and timing of soilcrete formation are very important.
For applications that do not use the control/guidance systems, the
tolerance for drilling the boreholes must comply with specifications
established for the project to ensure the interlocking of the thin diaphragm
wall elements.
Additional limitations for using the thin diaphragm wall technique were
provided in the previous section.
Directional backreaming
horizontal floor emplacement
Figure 8-6 Creation of a bottom barrier using directional drilling and

scarifying technique.
8.3.4 Construction of Bottom Barriers by Directional Drilling and Scarifying

Horizontal directional drilling is a technology used by the oil, gas,
petrochemical, utility, and construction industries for placing underground
pipelines, conduits, and utilities under rivers and other locations where
conventional trenching methods are unacceptable. It has been used to install
small and large diameter pipe over short to long distances in a variety of
194
geologic formations. In construction applications, directional drilling has been

used in tunnelling projects where the walls and ceiling of a tunnel are formed
by drilling a series of parallel and overlapping holes horizontally to overcome
the problems of caving in of soft or unconsolidated formations. The application
of this technology is depicted conceptually in Figure 8-6. To emplace a floor, a
pilot hole would be drilled beneath the target area, exiting at the ground surface
on the other side of the target area. To create a grouted barrier, the drill bit
would be removed and replaced by a reaming device that would blend the
grout with the soil as the reamer is pulled back under the target zone. Borehole
spacing of approximately 5 ft would be required to create a continuous barrier.
Although technically feasible, this technique may not be economically feasible
because of the large number of directionally drilled holes.
8.4 FROZEN SOIL BARRIERS

Ground freezing is used in civil engineering applications to provide loadbearing strength to soils during construction and to prevent groundwater
seepage into excavated areas. Frozen soil barriers have the potential to prevent
the migration of contaminants. The system for freezing soil consists of
subsurface pipes spaced evenly beneath and around the contaminated zone
to be isolated. Inside each pipe is a concentric feed-pipe which supplies a
coolant, such as calcium chloride or ethylene glycol. The chilled coolant flows
through the annulus formed by the two pipes, and then back to the
refrigeration plant through the concentric inside pipe. The cold exterior pipe
freezes the adjacent soil, forming a frozen soil mass around the freeze pipe
over its full length. The thickness of the frozen soil mass increases with time
at a rate depending on the soil, soil moisture, and thermal conditions at a
given site. A series of freeze pipes, properly spaced and aligned, can form a
continuous frozen barrier when the frozen soil mass around one pipe merges
with that from an adjacent pipe.
The frozen soil method offers two approaches for containing a subsurface
contaminated zone. One approach would be to freeze the entire contaminated
zone into a solid mass of frozen soil, thus solidifying the contaminants and
preventing their migration. For small areas where the contaminated soil does
not include vessels that could rupture from frost heave, this approach may be
acceptable. A second approach would be to create a frozen soil containment
barrier around the contaminated zone, as shown in Figure 8-7. Ground
freezing may have limited application in environmental remediation due to
the operation and maintenance costs associated with the need to continually
circulate refrigerant to maintain the barrier. The applicability of ground
freezing in the environmental marketplace may be limited to emergency
responses or interim containment during source removal activities. The
technology has been field tested at an uncontaminated site near Oak Ridge,
Tennessee.
a
8.4 FROZEN SOIL BARRIERS
195
Refrigeration
System
Liquid
Recovery
Cryogenic Barrier
System
Figure 8-7 Creation of a bottom barrier using the ground freezing technique.
8.4.1
Advantages
The advantages of frozen soil barriers for the confinement of hazardous waste
include:
a. Flexibility/Additional Uses
Frozen soil barriers can be installed in all types of soils where greater
than 20% soil moisture is available.
The barrier can be frozen into any configuration or depth, so long as the
freeze pipes can be placed. The containment system can be formed by:
(1) placing evenly spaced vertical pipes around the perimeter of the
contaminated zone to a depth where they intersect an impervious
horizontal strata; (2) inclining the pipes to intercept an impervious barrier,
such as another line of freeze pipes, thus forming a "V" trough-type
enclosure; or (3) by arranging the pipes to form a "U" shape containment
zone, accomplished by directional drilling.
Openings in the barrier can be closed by adding freeze pipes to the areas
in question.
As with other barrier technologies, the contaminated zone isolated by
freeze barrier technology becomes an in situ reactor chamber where other
remediation technologies can be applied without concern for contaminant
migration.
The thermal gradients induced by the slow freezing process may cause
196
migration of soil moisture, thus offering the possibility of concentrating

dissolved contaminants around the frozen mass.
b. Environmentally Friendly
Except for the freeze pipes and associated instrumentation, nothing is
added to the subsurface to create the barrier.
Minor cracks in the barrier are self-healing, if sufficient moisture is
available to freeze within the crack and seal it.
The thickness and extent of the barrier can be controlled by adjusting the
coolant temperature, by the number and location of freeze pipes used,
and by installing heating pipes where necessary to protect utilities located
adjacent to or within the barrier confines from freezing.
The frozen barrier can serve as a structural retaining wall, if excavation
and processing of contaminated soil becomes necessary.
8.4.2
Limitations
The application of ground freezing technology is subject to the following

limitations:
Sufficient groundwater must be available so that, after freezing, the ice
fills the soil voids sufficiently to create an impermeable barrier.
If the barrier is to be installed in the saturated zone, groundwater velocities
(on the order of one meter per day) may prevent adjacent freezing soil
masses to merge and form a continuous barrier.
The mechanical refrigeration systems require continuous inspection and
maintenance.
Frost heave pressures may cause structural failures in adjacent facilities
if the barrier is not properly designed.
When the barrier is thawed during decommissioning, adjacent facilities
may settle causing structural failures if precautions are not taken.
Frozen barriers may not effectively isolate all dissolved contaminants,
especially those that are known to lower the freezing point of water.
The cost of directional drilling may limit the application of this technology
to create barriers beneath large contaminated sites.
There are few commercially available and reliable sensors for detecting
barrier failures.
8.5 GUIDANCE SYSTEMS FOR DIRECTIONAL DRILLING

The most practical technology for monitoring or controlling the trajectory of
a subsurface borehole is the magnetic sensor used in conjunction with a gravity
sensor. The current state-of-the-art permits the application of magnetometers
8.5 GUIDANCE SYSTEMS FOR DIRECTIONAL DRILLING
197
and accelerometers with only small errors in orientation, azimuth, and

inclination. In order to obtain best results, the sensors must be properly
mounted in a common probe, properly calibrated, and used with good
practice.
The advantages of this guidance technology include: 1) a high state of
development; 2) good accuracy; 3) rather robust, reliable instrumentation as
presently applied; and 4) relatively low power requirements. Current systems
are designed for ease in use. The software supplied (i.e., for the all-angle
steering tooled system) has been widely accepted as error-free and userfriendly. Nearly all earth materials are virtually transparent to magnetic fields,
certainly so to gravity fields. The few materials in nature that affect magnetic
guidance, such as magnetite, do so on a relatively short-term basis in terms of
hole length. In addition, a method is available for checking calibration in the
field, without the need for other instruments.
Those who have utilized magnetic instruments for directional control have
developed an operating protocol for using these instruments that avoids
magnetic influences capable of causing unacceptable errors in azimuth
readings. The protocol requires the use of a sufficient length of nonmagnetic
materials around, above, and below the magnetometers such that the
maximum error produced is within tolerable limits. The requirement for
nonmagnetic material around the magnetic sensors currently limits the
application of this technology.
When directionally drilling at river crossings (trajectories under rivers or
other obstacles) or establishing boreholes for strategic placement of drain holes
or instrumentation, magnetic interference often is encountered (or is suspected
to be present) for which the operator has no control. In some of these cases,
precise placement of the boreholes may be of greatest importance. For these
applications, where magnetic anomalies are suspected or increased accuracy
is needed, Tensor has developed a steering tool system augmented by a surface
coil or coils (called TruTracker) that can be used to generate a predictable
magnetic field. The TruTracker technology is currently used by
approximately 25 licensees. TruTracker (shown in Figure 8-8) has been used
to place boreholes across river crossings with great accuracy, particularly as
the exit stake is approached, even when little or no magnetic interference is
expected.
The TruTracker technology is used by first placing a coil at the ground
surface and extending it along the borehole right-of-way. The width of the
coil should be approximately equal to borehole depth. After placing the coil,
the location of the corners are stored in the computer coil input file. At intervals
during drilling (usually when drill pipe is added), two sets of readings are
taken, one set with current flow in the forward direction, the second set with
current flow through the coil in the reverse direction. The electrical current is
accurately measured and the value entered into the computer. The computer
then subtracts one set of readings from the other, thus removing the earths
magnetic field and interference fields. What remains is the field generated
198
by the coil. The algorithms produce a model of the field distribution in terms
of field direction and intensity in the plane at right angles with the desired
trajectory. By comparing actual readings to the model output, a coordinate
position is determined that would bring the difference between the actual
reading and the model output to within the desired margin of error. Thus,
the steering tool probe is reported to be located at X distance right or left of
the centerline and Y distance below the entry-to-exit connecting line elevation.
Surveyed corners
Coil
Drill string
Guidance Packet
Magnetic field
Figure 8-8 Schematic of the TruTracker technology for steering the drilling
of boreholes.
The TruTracker method has been used to place boreholes between
adjacent pipelines, under runways containing reinforcing steel, and below multistory building foundations. Several case studies are briefly discussed below.
Boreholes for three pipelines were stacked with vertical spacings of 10
ft, alongside three others in a limited right-of-way, under a ship channel.
Total width of the right-of-way was 40 ft.
A borehole was placed parallel and between corrugated sheet metal pilings
30 ft apart, under a canal. The total distance was approximately 450 ft.
A borehole was placed approximately 20 ft under 250 ft of tarmac and
180 ft of runway to exit at a prescribed point within 4 ft.
Boreholes were placed at approximately 4 ft intervals around the
circumferences of a pair of tunnels. The boreholes were placed with an
accuracy of +/- 1.5 ft. Their purpose was to reinforce the circumference
of the tunnels by grouting prior to boring the tunnels. The trajectory of a
section of the tunnels, under the downtown area of a large city in Southeast
Asia, was continuously curving, starting side by side at one end, and one
above the other at the other end.
8.6 FIELD TESTING OF BARRIER SYSTEMS
199

Field testing of bottom barrier technologies is necessary to acquire the
experienced-based knowledge needed to apply the technologies under
varying site-specific conditions. Unfortunately, only a few field tests have
been completed for creating bottom barriers.
8.6.1
Construction of Bottom Barriers Using Jet Grouting to Create

Thin Diaphragm Walls
The thin diaphragm wall system with various inclines, interlocking, and
column elements was developed for seepage, infiltration, and erosion
problems on the River Po in Northern Italy. A field test was conducted
near the town of Cremona that demonstrated the ability to create a vertical
monodirectional thin diaphragm wall in sandy soils. A water/cement/
bentonite mix injection material was used. In this field test, a 2.5 m to 4
m radius was achieved which resulted in the treatment of 5 to 7 m2 per
borehole for every 1 m of nozzle movement. The test also showed that
the monodirectional and rotational system could be used in combination.
A continuous barrier was formed by interlocking the panels and columns.
An inclined thin panel was created with just 1 m of cover by orienting the
nozzle in the downward direction.
The first full scale application of a thin diaphragm wall occurred in
Italy to alleviate a deep seepage problem along a flood protection levee
at the confluence of the River Oglio and the River Po. A 3 km-long, 20 mdeep interlocked thin diaphragm wall barrier was formed on the toe of
the levee embankment in a predominantly sandy soil. After four years
and at least one major flood each year, the seepage problem has not reoccurred.
A similar seepage problem in Italy has recently been solved in the
same manner by constructing an 800 m-long, 15 m-deep, 60 cm-diameter
interlocked columnar system in a predominantly gravelly soil. This jetgrouted barrier was installed to remedy an emergency deep filtration
problem on the River Po embankment near the Ticino River.
Directional drilling using thin diaphragm wall technology was tested
in Germany by an Italian-German research group (Fondazioni Speciali Parma/FlowTex-Ettlingen). The test, conducted in the Bitterfield area,
demonstrated the feasibility of constructing a thin diaphragm wall
structure using a mineral wax (Montan wax) grout material. A small tub
structure was successfully created.
In Roblingen, an 80 m directional drilling thin diaphragm wall barrier
was successfully created in conjunction with the formation of a 200 m
column. Both single and double fluid jet techniques were used and the
operational difficulties encountered were successfully overcome.
200
8.6.2
Frozen Soil Field Test
The U.S. Department of Energy Office of Science and Technology performed

a subsurface groundfreezing demonstration at the Scientific Ecology Group
(SEG) facilities in Oak Ridge, Tennessee. The primary goal of the field test
was to demonstrate the frozen soil barrier could be used to prevent migration
of radioactive and hazardous contaminant plumes.
The ground freezing demonstration was initiated in March 1994 using a
configuration shown in Figure 8-9. The site contained a fine clay with chert
fragments, typical of the Oak Ridge area. The moisture content was
approximately 27%. The demonstration configuration consisted of a double
outside ring of freeze pipes and a single inside ring of heat pipes that formed
a V-shaped enclosure. The outside and inside dimensions were 56 ft x 56 ft
and 33 ft x 33 ft, respectively. The freezing pipes were installed at an angle of
45 and extended to a depth of 28 ft below the ground surface. A 750 gal steel
tank was located near the center of the enclosed unfrozen soil. This tank was
used to release water and tracers to test the hydraulic conductivity of the
barrier. Sensors were placed in the barrier and the adjacent area for recording
temperatures, soil conductivity, hydrostatic pressures, and movement of the
tracers. Observation wells were also installed.
Two refrigeration units were used, each with a refrigeration capacity of
40 tons utilizing recycled R-22 coolant. Both units were operated during the
initial 41-day barrier formation. After formation of the frozen barrier
containment system, only one unit was needed part-time to maintain the
barrier thickness.
The demonstration showed that a frozen soil barrier can be created that
is impervious to water and the tracers used. The two outer rings of freeze
pipes formed a 10- to 12-ft thick frozen barrier. The innermost ring of heat
pipes circulated warm brine and was effective in controlling the inner growth
of the frozen barrier. The power requirement for maintenance of the frozen
barrier system was approximately 31% of that needed to initially form the
barrier. To the extent that convective heat transfer coefficients could be
quantified and the thermal properties of the soil could be characterized, the
finite element heat transfer analysis provided an effective analytical tool for
predicting the time it would take to develop the frozen soil barrier to the
design thickness and the shape desired. The results showed that finite element
analysis can be a useful tool for developing freeze wall designs and selecting
refrigeration equipment.
A series of tests was performed during installation of the freeze pipes,
barrier formation, and maintenance operation. These tests are described below.
A barrier diffusion test was conducted to measure the diffusive transport
of water-borne contaminants through the frozen soil barrier. Fluorescein
and Rhodamine tracers, at a concentration of 200 ppm, were used.
Fluorescein was released prior to the freezing to assess the hydraulic flow
201
between the release point and the recovery well. Fluorescein was found
to be distributed both inside and outside the barrier.
Approximately 750 gal of water with Rhodamine were released from the
buried tank to simulate a sudden release of liquids from a tank and assess
the ability of the frozen barrier to contain the liquid. Rhodamine was
found only inside the barrier.
Soil movement was continuously measured after initiation of refrigeration.
Theoretical calculations of soil upward movement during testing
predicted 1.2 ft to 2.24 ft displacement during the first 70 days. The actual
movement measured in the field was 1.65 ft. The maximum heave during
steady state operation was measured to be 2.25 ft, which is within the
range typically seen during civil engineering applications of this
technology.
The strain gauge measurements were used to analyze the mechanical
stresses produced on the tank as a result of soil movement during freezing.
The observed maximum stress was 4000 psi, well below the allowable
stress of 12,000 psi for carbon steel.
The findings from the field test were promising, but longer tests would
be beneficial. The tracers used in the future to quantify diffusive transport
through the frozen barrier need to be more representative of the contaminants
to be encountered in the field.
Refrigeration
equipment/pumps
Refrigeration
equipment/pumps
Contaminant
Watertable
Water-bearing
area or seam
Frozen soil zone
Freeze holes
(vertical)
Impervious clay layer

Impervious clay layer
Freeze holes
(slanted)
Monitoring
points/sensors
Figure 8-9 Configuration of demonstration ground freezing system at Oak

Ridge, TN.
202

8.7.1
Thin Diaphragm Wall Technology
The monodirectional thin diaphragm wall technology is ready for immediate

application at sites with limited drill distance and where no active steering
device is required. The directional drilling thin diaphragm wall technology,
although still in the testing phase to determine the extent of its application, is
ready for use in applications where the maximum distances are in the range
200-300 m.
Single-rod systems have been tested and used with cement- based grout
mixtures. The system has been technically proven and can be used to pump
fluids that are rheologically similar to cement slurries. Mineral wax slurries
that have been previously tested can also be used. Application of more
aggressive chemical mixtures in the slurry need further study, including
evaluation of the appropriate rubber and sealing packs required in the high
pressure pump circuits.
Double-rod systems enable the use of a wider range of possible grouting
materials because of their ability to separately inject two different grouting
components and mix them in the subsurface. These components could be
two liquids or a liquid and a gas. Hard soil also can be treated with appropriate
spacing of the barrier elements. The triple-rod system, with its water cutting
action, may be required for some soil types.
The verification techniques and the characteristics of the barriers produced
using these techniques are described below.
a. Emplacement and Performance Verification. There are two phases to quality
control during emplacement of the barrier. The first phase of construction
quality control is to ensure that the boreholes are in the correct location and
that any deviations are within predetermined tolerances.
For the monodirectional thin diaphragm wall system, there is no active
guidance control system used during operation because of the relatively short
drill length. Precision of placement is based on an accurate on hole mast
orientation, stable on hole drill rig positioning, and rigidity of the drill rod
string. Before injection, the borehole deviation can be verified using an
inclinometer. Depending on the project requirements, deviations normally
should not exceed 2% of the hole depth.
For directional drilling, a guidance control system is used. The accuracy
of positioning the borehole is based on the precision of the guidance/control
device used during the drilling phase, the position and realization of the
reference magnetic field (from the surface with a circuit in the ground, etc.),
and the type of drilling technology. The deviation should not exceed 15 to 20
cm, independent of the length of the borehole.
The second phase of construction quality control involves verification of
the injection of the grout mixture. This is done by monitoring the rheological
203
and physical properties of the grout mixture being pumped as well as other
injection parameters, e.g., pump outlet pressure, nozzle outlet pressure, flow
rate, extraction and rotation speed, and pump rotational speed. After
emplacement, further verification can be accomplished by coring the barrier
or by using non-intrusive geophysical techniques, e.g., cross-hole tomography
and seismic reflection-refraction instrumentation. Also, a series of leachate
detection borehole devices can be placed around and beneath the barrier to
monitor long-term performance.
b. Emplaced Barrier Characteristics. The physical characteristics of the
constructed barrier depend on the specific grout mixture used and site-specific
soil conditions. The constructed barrier is usually characterized by the
compressive strength, hydraulic conductivity, and elasticity modulus.
Compressive strengths can range from 2 to 15 MPa (290 to 2,175 psi) for cementbased mixtures in soil formations ranging from clays to gravel, respectively.
For Montan wax based mixtures, compressive strengths range from 0.2 to 1
MPa (29 to 145 psi). Hydraulic conductivities can range from 10-7 to 10-9 cm/
sec for panels formed using cement/bentonite mixtures and from 10-8 to 10-10
cm/sec for panels formed using Montan wax based mixtures. The elasticity
moduli generally range from 1000 to 5000 MPa (1.45 x 105 to 7.25 x 105 psi).
Barrier shrinkage during hardening is a concern.
During construction of the thin diaphragm walls and columns, a minimum
radius or depth of penetration of the jet stream is required to achieve continuity
and interlocking of the elements. For the formation of thin diaphragm walls,
the minimum depth of penetration is usually 1.5 m to 2 m, guaranteeing at
least 3.0 m2 to 4.0 m2 of thin diaphragm wall formation. For the construction
of columns, the minimum radius is usually 0.4 m to 1.0 m, guaranteeing at
least 0.8 m2 to 2.0 m 2 of column formation. Under standard operating
conditions (two injection nozzles and a 350 hp pump), the minimum barrier
thickness is usually 8 to 10 cm near the drill string and 20 cm to 30 cm at the
end of the panel. By modifying the nozzle, the nozzle position, and increasing
the size of the pump, the minimum thickness can be increased to
approximately 15 cm to 20 cm. Depending upon the specific project
requirements, thin diaphragm walls are typically spaced 1 m to 3 m apart
and columns are typically spaced 0.5 m to 0.8 m apart.
8.7.2 Frozen Soil Barriers
The verification techniques and the characteristics of constructed barriers
using this technology are described below.
a. Emplacement and Barrier Verification. Quality control during construction
of the frozen barrier begins by ensuring that the freeze pipes have been
installed in the proper location followed by the use of temperature probes
and conductivity sensors to confirm the extent of the frozen barrier mass.
204
The location of the freeze pipes in the subsurface is determined by passing a

slope inclinometer down each pipe. At the present time, defects in a frozen
barrier are detected by analyses of continuously recorded temperature,
conductivity, piezometer, and readings taken at critical locations within the
barrier and adjacent areas. In addition, the temperature of the liquid coolant
is monitored.
b. Frozen Soil Barrier Characteristics. The hydraulic conductivity of frozen
soil depends on the type of soil and its temperature, when near the freezing
point of water. Andersland, et al. (1995) found that, at a temperature of -10
C, the hydraulic conductivity of a frozen saturated sand was below 10-5 cm/
sec, the limit of the accuracy of the equipment used. The thickness of a frozen
soil barrier can be controlled to meet the requirements of the site conditions.
Thicker barriers provide greater resistance to chemical erosion and offer added
resistance to contaminant migration. A typical single row of freezing pipes
spaced at about 1.5 m can create a frozen barrier with an effective thickness of
about 1.3 m in about 45 days (Sanger and Sayles, 1979). Multiple rows of
freezing pipes can be used to form a thicker barriers during the same time
period. The demonstration of ground freezing at Oak Ridge (reported by
SEG 1995) developed a barrier thickness of about 12 ft in using two layers of
freezing pipes.
8.7.3
Cost Analysis
The construction of a barrier floor beneath a subsurface zone of contamination

is a relatively unproven technology. Construction approaches can employ
conventional civil engineering methods, sometimes using innovative
combinations of these methods. For example, conventional high pressure jet
grouting technology could be used to create a relatively small slanted V
trough or conical shaped containment floor. Alternatively, conventional high
pressure jet grouting could be innovatively coupled with directional river
crossing technology to construct a much larger barrier containment floor. Since
construction of artificial bottom barriers is a relatively new field, very few
field demonstrations have been conducted, resulting in a paucity of actual
emplacement cost data. Nevertheless, reasonable estimates of costs to construct
bottom barriers can be based on the operating experience from conventional
applications.
For construction of large artificial containment floors, most technologies
will likely rely heavily on directional river crossing technology and guidance
techniques. It is estimated that the construction technologies currently of
interest will require spacings of 6 ft to 10 ft between the directionally drilled
boreholes. Directional drilling of a pilot hole typically costs roughly in the
range of $50/ft to $150/ft for relatively shallow, somewhat cohesive soils.
Once the pilot hole is in place, back reaming or scarifying the pilot hole costs
roughly $50/ft to $75/ft. This does not account for the directional drilling
205
costs to drill beneath the target zone, mobilization and demobilization costs
at the site, disposal costs for wastes produced during construction, and
contingency costs, e.g., when adverse soil conditions are encountered.
a. Jet Grouting.
For conventional civil engineering applications, high
pressure jet grouting typically costs $15 to $20 per square foot for columnar
walls. For thin diaphragm walls using jet grouting, costs are estimated at
$10 to $15 per square foot. For estimating purposes, it is assumed that the
typical high pressure jet grouted wall has a diameter of roughly 6 ft and the
thin diaphragm wall length is 5 ft in both directions. In addition, the cost of
directional drilling is estimated at $8 to $25 per square foot for columnar walls
drilled on 6-ft centers and $5 to $15 per square foot for thin diaphragm walls
drilled on 10-ft centers. The installed cost of a jet grouted columnar floor is
estimated at $23 to $45 per square foot and, for a thin diaphragm wall floor,
$15 to $30 per square foot. Not included in these estimates are the costs of the
grouting materials, waste disposal, and contingencies.
b. Directional Drilling. Directional drilling using river crossing technology
can currently scarify a hole to a 60-in. diameter in one operation after the
pilot hole has been drilled. Conventional directional river crossing technology
typically costs $50 to $150 per linear foot to drill the pilot hole and $50 to $75
per linear foot to backream or scarify the hole to a large diameter. On a square
foot basis, the estimated cost is $17 to $38 per square foot for directionally
drilled holes on 6-ft center lines, with a scarified hole diameter of up to 8 ft.
Not included in these estimates are the costs of waste disposal, grouting
materials, and contingencies.
c. Frozen Soil. Conventional civil applications of frozen soil technology,
when used to stabilize soil walls for deep excavations, typically costs $60 per
square foot to emplace and roughly $2 per square foot for operation and
maintenance. The $60-per-square-foot cost estimate is applicable for smaller
installations where standard non-directional drilling technology is used. For
applications requiring directional drilling, the costs, based on 10-ft spacings,
are estimated at $65 to $75 per square foot. Not included in these estimates
are the costs of waste disposal and contingencies.
d. Permeation Grouting. The cost of directional drilling can dominate the
overall cost for emplacing a large containment floor using permeation
grouting. Similarly, but to a lesser degree, conventional vertical or slant drilling
costs can dominate the overall cost for emplacing a smaller containment floor,
as compared to the cost of the grout. Based on spacings of 10 ft between
boreholes, directional drilling costs are estimated at $7 to $17 per square foot.
Not included in this estimate are the costs for grouting materials, waste
disposal, surface support equipment, and contingencies.
206
8.8 NEEDS
8.8.1
General
The following general needs have been identified relative to the construction
of artificial floors:
advanced technologies are needed to verify the emplacement and the
performance of the installed bottom barriers;
new barrier materials need to be developed and tried; and
improved economic analyses are needed to better estimate the cost of
these technologies.
The lack of good economic analyses is partly due to the lack of practitioners
of the technology. Well planned and integrated field demonstration tests may
provide the most effective means to acquire good cost data.
The major cost and performance factors are associated with directional
drilling. Although directional drilling guidance technologies are currently
available, their accuracy needs to be improved for ensuring cost-effective
emplacement of containment floors. Improved accuracy could reduce the
number of boreholes required, thus reducing emplacement costs. Also, many
potential applications are in high risk areas where drilling accuracy is critically
important.
Currently, there are no cost-effective methods available to verify the
emplacement and continuity of an artificial containment floor. Easily
detectable chemical additives to the grout may be a potential technique for
verifying that a barrier has been properly installed. Also, additional field
experience is needed with the new slow setting or reactive grout materials.
Performance standards need to be developed. The development of such
standards could greatly impact the costs for emplacement, verification, and
monitoring of an artificial containment floor. Will performance standards be
based on estimates of contaminant mass flux passing through the barrier
during its service life, on the absence of defects within the containment floor,
or will they be based on a requirement that the measured bulk hydraulic
conductivity in the field meet a project specification? Finally, there are no
standards for determining when a bottom barrier would be required under
varying site-specific conditions.
8.8.2
Technology Specific Needs
The following technology development needs have been identified and apply
to both the directional and non-directional drilling with jet grouting to create
thin diaphragm walls technologies:
Testing is needed under a variety of site-specific soil conditions to develop
8.8 NEEDS
207
methods and hardware modifications that will prevent rod-string binding

and sticking.
A system needs to be developed for detecting the position of the nozzle
during the extraction phase of grout emplacement.
For directional drilling with jet grouting to create thin diaphragm walls,
the technology development needs include:
Improvements in the measurement of the alignment deviation of a
borehole. A tolerance in the range of 10 cm is desired for all applications
(open-hole, blind-hole, extended depth, etc.).
Ability to directional drill into gravels, cobbles, and noncohesive soils.
Advancement of guidance control systems. Current systems slow down
both the drilling and injection phases of the operation. Wireless systems
appear attractive if they can comply with the requirements for accuracy.
For construction of bottom barriers by conventional jet grouting, technology
needs include:
Technical improvement of standard drilling systems and directional
drilling systems to improve alignment and increase reachable lengths.
Improvement of the operational efficiency and mechanical and electronic
links between a directional drill sector and high pressure injection sector.
Development of the ability to inject separately two chemical reagents
which are designed to react in situ and form the barrier after injection.
Techniques to verify the accuracy of barrier geometry and hydraulic
conductivity, at a macroscopic scale on the order of a meter.
For improved application of soil freezing technologies, the following needs
have been identified:
An effective field method is needed for injecting water into dry soils near
the freezing pipes while freezing of partially saturated soil is in progress.
This must be developed in order to create an impervious barrier.
An acceptable performance standard for the frozen soil barrier needs to
be developed. For example, will the standard vary depending on the
nature of the contaminants being contained?
The rates of diffusion of different contaminants through the unfrozen water
films in frozen silt and clay soil need to be determined.
The chemical deterioration of the frozen soil barrier needs to be assessed
for chemicals that the barrier is likely to encounter. This includes the
effect of brine, nonaqueous phase liquids (NAPLs), and radionuclides on
the unfrozen water films and the ice in the frozen barrier.
A reliable method needs to be developed for detecting and locating defects
within the frozen barrier. Possibilities include use of Frequency-
208
Modulated Continuous Wave (FMCW) radar imaging techniques, acoustic

imaging techniques, and Time Domain Reflectometer (TDR).
Additional field data are needed to evaluate and validate the mathematical
models used for design and prediction of rates of freezing.
Artificially emplaced bottom barriers are viewed skeptically by the public
and the regulatory community because verification of emplacement and
performance is done indirectly. However, the techniques employed have been
routinely used in the construction industry for critical and complex
applications. The acquisition of field experience in constructing bottom
barriers and the establishment of successful performance by well-documented
case histories are key factors to the acceptance and application of these
technologies. As field experience is gained, the techniques for construction,
verification, and monitoring of emplaced bottom barriers will evolve and be
developed for the soil conditions and constraints that exist at hazardous waste
sites. Additional research and development are required to provide costeffective techniques for verifying barrier emplacement and performance. The
current interest in bottom barrier technologies by the federal government and
industry may aid the formation of joint field demonstration projects, thereby
expediting the development and acceptance of this technology. These field
projects should be designed to integrate all the elements (grout materials,
emplacement technique, monitoring technologies, etc.) required to create a
competent bottom barrier system that is acceptable to the regulatory
community and the public.
8.10 REFERENCES
Andersland, O.B., S.H. Davies, and D.C. Wiggert, (1995). Performance and
Formation of Frozen Containment Barriers in Dry Soil, Rust Geotech Inc.
Burke, G.K., and G.T. Brill, (1993). Anchored Cutoff Structure Design and
Construction, Proceedings Third International Conference on Case Histories
in Geotechnical Engineering, ASCE.
Burke, G.K., and J.P. Welsh, (1995). Vertical Cutoffs and Bottom Sealing by
Jet Grouting, Proceedings Geoenvironment 2000, ASCE Specialty
Conference, February.
Injected Gel Stops Toxic Seepages, (1995). European Chemical News, pg. 31,
May 22-28.
8.10 REFERENCES
209
Injectable Silicone Barrier Isolates Underground Leaks, (1995). Environmental

Engineering World, pg. 37, May/June.
Raabe, E.W., and S. Toth, (1989) Producing Sealing Walls and Sealing Slabs
by Applying the Soilcrete Process, Proceedings of the Piling and Deep
Foundations Conference, United Kingdom.
Sanger, F. J. and Sayles, F. H. (1979). Thermal and Rheological Computations
for Artificially Frozen Ground Construction, in Ground Freezing,
(Jessberger, H., Ed.), Elsevier Publishing Co., NY.
Scientific Ecology Group, Inc. (1995). Final Report: Demonstration of Ground
Freezing Technology at SEG Facilities in Oak Ridge, TN.
211
SECTION 9
CHEMICAL-BASED BARRIER MATERIALS
prepared by
Joyce M. Whang, DuPont Co.

John A. Apps, Lawrence Berkeley National Laboratory

Ernest E. Carter, Carter Technologies
Alison Hapka, DuPont Environmental Remediation Services
John Heiser, Brookhaven National Laboratory
Reuben Karol, Rutgers University
Raymond Krizek, Northwestern University
Ingo Sass, FlowTex-GUT, Germany
Charles F. Voss, Golder Federal Services, Inc.
9.1 BACKGROUND
Chemical-based barrier materials may provide performance improvements
over soil- and cement-based materials. Chemical grouts offer the potential to:
lower hydraulic conductivity,
minimize the effects of wet-dry cycling,
improve resistance to degradation by contaminants,
minimize diffusive transport of contaminants, and
enable easier placement due to reduced viscosities and/or little or
no particulate matter.
However, chemical-based grouts are usually more expensive than soil-based
materials, so these potential improvements in performance must be weighed
against the added costs on a case-by-case basis.
Chemical grouts have been emplaced using conventional permeation
grouting techniques for decades. However, in recent years, the interest in jet
grouting for environmental remediation applications has grown. Jet grouting
can provide better control of placement and the specifications on set time
may not be as stringent as is the case with permeation grouting. Permeation
and jet grouting emplacement methods have been described in Rumer and
Ryan (1995). Also, see Sections 2, 3, and 8 in this book.
Selection of the appropriate chemical-based barrier material for a given
site depends on factors such as:
212
CHEMICAL BASED BARRIER MATERIALS
hydraulic conductivity requirements,

compatibility with soil chemistry,
compatibility with the contaminants being contained,
soil moisture content, and
emplacement method to be used.

In this section, the following chemical-based barrier materials are reviewed.
sodium silicate
acrylate gel
colloidal silica
iron hydroxides
montan wax
sulfur polymer cement
epoxy
polysiloxane
furan
polyester styrene
vinylester styrene
acrylic
Sodium silicate and acrylate grouts have been used for decades in field
applications such as construction and water sealing. They can provide good
performance at a modest cost. Colloidal silica, iron hydroxides, and montan
wax have been identified more recently as potentially useful materials for
environmental remediation applications. These materials are derived from
inorganics or naturally occurring materials and may offer improved
performance compared to sodium silicate and acrylate grouts. The remaining
materials provide excellent performance, even in aggressive chemical
environments, but they are expensive. Sulfur polymer cement has low material
costs, but emplacement requires heating of the soil.
Important parameters or properties affecting the performance of chemicalbased barrier materials are listed below. Each material will be described in
terms of these parameters.
Viscosity: As discussed in Rumer and Ryan (1995) the viscosity requirement
for permeation grouting depends on the soil being grouted. With jet grouting,
a viscosity close to that of water is needed for more efficient energy transfer
and to enable the free flow of excess materials to the surface.
Set time: Set time must be controllable. Requirements depend on the
emplacement method. Set time must be long enough for grout to reach its
213
target location and short enough for grout to set before it travels beyond this
location. Depending on the emplacement method, a time lag (to account for
possible operator error) may be needed to avoid premature setting.
Hydraulic conductivity: The primary function of a barrier material is to
prevent or inhibit seepage. Thus, barrier materials are selected for their low
hydraulic conductivities. Hydraulic conductivities of 10-7 to 10-9 cm/s are
commonly attained with soil-based grouts. Because they are more expensive,
chemical-based barrier materials should exhibit hydraulic conductivities at
least equal to, and preferably lower than, those of soil-based grouts. The
hydraulic conductivity of a barrier will depend on the soil conditions, the
grout material used, the extent to which the grout fills the pore space, and on
changes to the soil matrix, e.g. compaction, that occur during grouting.
Wet-dry cycling: Soil-based barrier materials and many of the gel-type
chemical grouts require saturated conditions around the barrier to remain
intact. Saturation is easily maintained below the water table. Moisture loss
can be greatly reduced in vertical barriers by capping them. Barrier materials
used at contaminated sites in arid climates must tolerate low soil moisture
and the fluctuations in soil moisture that result from precipitation events.
Resistance to chemicals: acids, bases, organics: Barrier materials used in
environmental applications may be exposed to a wide range of chemical
contaminants. These chemicals may degrade the barrier properties of soilbased grouts. Some chemical grouts will maintain good barrier properties,
despite exposure to contaminants.
Resistance to irradiation: U.S. Department of Energy has many sites with
radioactive wastes. The ability to withstand exposure to radiation is important
for barrier materials that may be used in these sites.
Reduction of diffusive transport: As discussed in Chapter 10, contaminants
can be transported through a barrier by diffusion, as well as by advection.
The ability of barriers to resist diffusive transport may be critical to long term
containment of hazardous materials. Diffusive transport depends on the
solubility and molecular diffusivity of the contaminant, as well as on the
tortuosity of the barrier pore system and the sorptive properties of the barrier
materials. The sorptive properties of many chemical grouts have not been
measured.
Expected lifetime: The expected lifetime of chemical grouts depends on the
durability of the emplaced grout at the application site. For example, gels
will have longer effective lifetimes in the saturated zone than in the vadose
zone, with shorter effective lifetimes in arid regions. Data for predicting
lifetimes are limited. Sodium silicate and acrylate gels are the only materials
214
that have been used as grouts long enough for lifetimes to be estimated from
field performance. For engineered polymers, there are no natural analogs
and, while lifetime is expected to be long, testing methods have not been
developed that enable accurate predictions of service life in the field. Expected
lifetimes of greater than 25 years have been assigned to materials that are
expected to last indefinitely.
Repairability: It is expected that containment barriers may sometimes need
repair. Some materials may be more easily repaired than others.
Safety, toxicity, regulatory acceptability: These factors must be determined for
resins, catalysts, and the final set grouts.
Cost of materials: The total cost of a barrier system includes the costs of both
materials and emplacement. Only material costs are compared in this report.
The amount of grout needed depends on soil type, emplacement method,
and needed barrier properties. Estimated costs are based on a grout-to-soil
ratio of 1-to-3. More grout may be needed in some situations, especially when
grout flows past its destination during emplacement.
For comparison with other containment methods, e.g., sheet piling and
geomembranes, costs have also been calculated per square meter of barrier
wall. This is done by assuming a 1 m wall thickness for barrier materials
with hydraulic conductivity 10-8 cm/s or greater, and a 0.3 m wall thickness
for materials with hydraulic conductivity less than 10-8 cm/s. For materials
of extremely low hydraulic conductivity, thinner walls may provide adequate
containment, but they may require more grout per volume of soil. Costs for
other wall thicknesses or grout-to-soil ratios can be calculated from the data
provided in Tables 9-1 to 9-12.
Extent of application: Many of the materials described have not been field
tested. Only a few have been commercially applied in the field.
Commercially availability: This question will be answered for each material.
9.2.1
Sodium Silicate
Sodium silicate is the first chemical grout whose use in soil was documented.
In 1886, Jeziorsky was granted a European patent based on the injection of
concentrated sodium silicate into one hole and a coagulant into another nearby
hole (Karol, 1990). Until the early 1950s, sodium silicate was the only chemical
grout in field use. Today, it accounts for 60 to 80% of the total chemical grout
volume used in the United States. It is widely used in construction and
generally emplaced using permeation grouting. It can be used in combination
with cementitious grouts.
Sodium silicate, nSiO2 Na2O, is commercially available as an aqueous
215
colloidal solution. The silica/alkali ratio should be between 3 and 4. Solutions

are typically composed of 20-30% SiO2, 510% Na2O, and 60-70% water. When
the colloid is mixed with a concentrated solution of the appropriate acid,
polyvalent cation, or specific organic reagent, there is nearly instantaneous
gelation. The earliest successful field process is credited to Joosten and uses
CaCl2 as the salt causing gelation. The reaction is given as:
SiO2 Na2O + CaCl2 + H2O > Ca(OH) 2 + SiO2 2NaCl
This process is virtually unused today. Today, ethyl acetate and formamide
are most commonly used to initiate the gelling of the sodium silicate. (Karol,
1990)
30
Setting Time (hours)
25
Compositions:
klebolink S = X liters
klebolink K = Y liters
klebolink H1 = Z liters
Water up to 100 liters
X + Y = 24 liters
20
15
10
5
0
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
Klebolink H1 (liters)
klebolink S=4L klebolink S=4.5L klebolink S=5L klebolink S=5.5L klebolink S=6L klebolink S=8L
Figure 9-1 Set times of glyoxal-modified sodium silicate (Golder Associates,

1993)
Golder Associates (1993) conducted tests with a glyoxal-modified sodium
silicate (GMSS) grout developed by the French chemical company, Societe
Francaise Hoechst. This material was extensively tested in the laboratory at
the Hoechst facilities in Paris and at the Technical University of Clausthal. It
has been used extensively in Europe and the United States as a soil
strengthener in unconsolidated soils. GMSS is composed of the following
four components:
water,
4-8% Klebolink S, an alkaline liquid with SiO2, NaO2, and water,
216
2-3% Klebolink H1, an acidic liquid with glyoxal and additives, and
16-20% Klebolink K, an aqueous suspension of non-agglomerated silica
particles in an alkaline medium.
The set time depends on the additives used (see Figure 9-1) and can be
controlled well for set times of about several minutes, but is poorly controlled
for set times of several hours. Set time increases when the material is flowing.
Golder Associates (1993) measured hydraulic conductivities resulting from
pumping GMSS through soils taken from DOEs Hanford and Sandia sites.
Ungrouted soils had hydraulic conductivities of 10-2 to 10-3 cm/s. Treated
soils had hydraulic conductivities 10-5 cm/s. This is consistent with other
literature on sodium silicate and glyoxal-modified sodium silicate grouts
(Bodocsi et al., 1988).
Sodium silicates are primarily used in the saturated zone. GMSS samples
left to desiccate in petri dishes became dehydrated and fragile in 1 day.
TABLE 9-1 Overview of Sodium Silicate
Viscosity:
GMSS: 2-5 mPas; 5-50 mPas for many other products
Set time:
depends on additives*
Hydraulic conductivity measured:
10-5 cm/s*
Wet-dry cycling:
crumbles when dried*
Resistance to chemicals:
acids: fair*
bases: poor*
organics: fair*
Resistance to irradiation:
unknown
Reduction of diffusional transport: unknown

Expected lifetime:
10-20 years
Repairability:
good
Safety, toxicity, regulatory

acceptability:
non toxic
Cost of materials:
$0.50/l, $130/m3 (25% grout), $130/m2 (1 m

thick wall)
Extent of application:
used commercially for decades for construction
Is material commercially available? yes

Unusual strength or weakness:
hydraulic conductivity not very low
* see text for additional discussion
To test for resistance to chemicals (ASTM C267-82), cylindrical samples

of grouted silica sand were immersed for 28 days in each of the following
chemical solutions (Golder Associates, 1993):
4N hydrochloric acid,
6N sodium hydroxide,
217
20% cupric sulfate,

100% methanol,
100% ethylene glycol,
100% aniline, and
100% xylene,
In most cases, there was a minimal amount of pitting, as indicated by the
presence of sediment in the bottoms of the test beakers. The exception was
sodium hydroxide. This sample completely disintegrated within one hour of
immersion.
9.2.2
Acrylate Gel
Acrylate grouts appeared in the commercial grouting market in the early

1980s, in response to an industrial need for a less toxic substitute for
acrylamide. Acrylate grouts are widely used for seepage control, e.g. sealing
storm and sewer lines (Karol, 1990). AC-400, a product made by Geochemical
Corporation, is a commonly used, relatively nontoxic, proprietary mixture of
the following:
acrylate monomers,
methylenebisacrylamide cross-linking monomer,
ammonium persulfate initiator,
triethanolamine accelerator, and
potassium ferricyanide inhibitor.
As seen in Figure 9-2, the gel time depends on the concentrations of the initiator
(KFe) and the accelerator (or catalyst).
Gel Time (minutes)
100
Percent KFe
0
0.005
0.01
0.02
0.04
10
0.1
Catalyst Concentration (%)
Figure 9-2 Set times of AC-400 acrylate grout (Krizek et al., 1992)
218
Injection of AC-400 into columns of Ottawa 20-30 sand yielded hydraulic

conductivities of less than 10-7 cm/s (Siwulo and Krizek, 1992). Hydraulic
conductivities of various acrylate grouts without soil have been measured to
be 5 x 10-9 cm/s (Clarke, 1982) to 5 x 10-10 cm/s (Bodocsi, 1988).
Acrylate gels are intended for use in the saturated zone. Burial of a grouted
sand sample in saturated sand resulted in sample weight increases of 1-4%.
In an arid environment, acrylate gels will dry and lose their barrier properties
(Krizek, 1992).
TABLE 9-2 Overview of Acrylate
Viscosity:
2 mPas
Set time:
depends on additives*
10-7 to 10-9 cm/s*
Wet-dry cycling:
acids: poor*
bases: good*
organics: fair*
unknown

Expected lifetime:
10-20 years
Repairability:
good

acceptability:
low toxicity
Cost of materials:
$4/kg, $3/l 40% solution, $1/l mixed to 12%,

$230/m3 (25% grout), $230/m2 (1 m thick wall)
used commercially for 10-15 years for water

sealing

Acids and bases were pumped through AC-400-grouted columns of

Ottawa 20-30 sand. For solutions with pH as low as 4 and as high as 10,
hydraulic conductivities remained less than 10-7 cm/s (Siwulo and Krizek,
1992).
The following hydraulic conductivities were measured when other
solutions were pumped through various acrylate grouts without soil (Bodocsi,
1988):
1N HCl: 10-5 cm/s, pH 0;
6N NaOH: 10-10 cm/s;
20% CuSO4: 10-7 cm/s;
25% acetone: 10-8 cm/s;
219
100% acetone: 10-4 cm/s;

100% aniline: 10-5 cm/s;
100% ethylene glycol: 10-8 cm/s;
25% methanol: 10-9 cm/s;
100% methanol: 10-3 to 10-5 cm/s; and
100% xylene: 10-7 cm/s.
9.2.3
Colloidal Silica
Colloidal silica (Moridis et al., 1995b) is produced from saturated solutions of

silicic acid by formation of Si-O-Si (siloxane) bonds. Repeated accretion of
molecules by this mechanism results in the formation of particles, the size of
which can be controlled in the range of 2-100 nm. Uncombined SiOH groups
remain on the particle surface. Gelling of colloidal silica particles occurs when
the particles approach each other, and siloxane bonds are formed between
them (see Figure 9-3).
H
O
O
Si
O
O
Si
O
O
Si
O
Si
Si
Si
Si
Si
O
O
H
Si
Si
O
O
H
Figure 9-3 Formation of siloxane bonds as colloidal silica particles gel

(Moridis et al., 1995)
The colloid is stabilized to prevent gelling during storage and shipment
by producing a charge on the particles. This is typically done by raising the
pH of the colloidal suspension. In this basic environment, a double-layer of
counter-ions, usually Na+ or NH4+, causes a repulsion that prevents particles
from approaching each other closely enough to initiate gelling. For controlled
gelling, the repulsive forces must be carefully reduced so that random motion
220
of colloidal particles results in collisions and siloxane bond formations at the

desired rate. Some ways of accomplishing this are to reduce the pH, to increase
the ionic strength, or to introduce multivalent cations.
Time to Rigid Gel (Stage 8) in hr
100
40 g/L
50 g/L
10
60 g/L
70 g/L
1
6.5
7.0
7.5
8.0
8.5
9.0
9.5
pH
Figure 9-4 Set times of Nyacol 1440 colloidal silica (Moridis et al., 1995)
Figure 9-4 shows sample set times measured at Lawrence Berkeley
National Laboratory. Nyacol 1440, a product of PQ Corporation, was mixed
with various concentrations of sodium chloride in a 5-to-1 colloid-to-brine
ratio. Because injected grout will be buffered by the soil, LBNL staff decided
to use ionic strength as the primary means of controlling set time.
Set time of colloidal silica can be accelerated when soils are present,
because of the ionic content of the soil. Such rapid gelation can be prevented
by (1) preflushing the soil with a 4% NaCl solution to remove multivalent
ions, or (2) by using a newly developed generation of colloidal silica products,
e.g., Nyacol DP 5110, in which some surface Si has been replaced with Al.
In laboratory tests, Hanford sands of hydraulic conductivity 10-2 cm/s
were injected with DP 5110 colloidal silica, which is 30 wt % silica. Hydraulic
conductivity dropped to 10-8 cm/s. For comparison, samples were also
prepared in which Hanford sand was added to the colloidal silica solution.
This represents a best case scenario, because, unlike with the injection tests, it
is certain that pore spaces are filled with grout. These tests also yielded
hydraulic conductivity of 10-8 cm/s.
In other tests (Noll et al., 1992), sandpacks were grouted with DuPonts
Ludox SM colloidal silica diluted to 5 wt %. Hydraulic conductivities before
grouting were from 8 x 10-5 to 9 x 10-4 cm/s. After grouting, hydraulic
221
conductivities ranged from 4 x 10-8 to 5 x 10-7 cm/s.

TABLE 9-3 Overview of Colloidal Silica
Viscosity:
5 mPas
Set time:
depends on pH, ionic strength, multivalent ions*
Hydraulic conductivity measured: 10-8 cm/s*

Wet-dry cycling:
crumbles when dried quickly
acids: expect to be good

bases: expect to be poor
organics: expect to be good
expect to be good
Reduction of diffusional transport: immobilizes leachable metals (Noll et al. 1992)

Expected lifetime:
> 25 years
Repairability:
good*

acceptability:
non-toxic
Cost of materials:
30 wt %: $1/kg, $1/l, $250/m3 (25% grout),

$250/m2 (1 m thick wall); 5 wt %: $40/m3 (25%
grout), $40/m2 (1 m thick wall)
laboratory and field tested

Observations in sandbox experiments at Lawrence Berkeley National

Laboratories indicate that a not-yet-gelled colloidal silica plume merged
seamlessly with a previously gelled plume, suggesting good repairability.
9.2.4
Iron Hydroxides
Methods for precipitating metals in situ have long been investigated for
enhanced oil recovery applications. Environmentally benign methods for in
situ generation of iron hydroxides and oxyhydroxides from low cost iron waste
streams are currently being developed by DuPont (Hapka et al., 1995). High
concentrations of iron are kept in solution at moderate pH (e.g., pH = 3) using
chelating agents. These solutions can be neutralized by (1) mixing with base
during jet grouting or (2) using premixed additives that cause a pH increase
at a controlled rate following emplacement by permeation grouting.
Results indicate that mixtures of iron, citric acid, urea, and urease could
be permeation grouted. The citric acid is used to keep 6 %wt iron in solution
at pH 3. Urease, a naturally-occurring enzyme, breaks down the urea. Urea
breakdown results in generation of ammonia, which causes the pH to increase.
Set time depends primarily on initial pH and amount of urease. Set times
222
of 5 minutes to 5 hours have been measured in the laboratory. It is expected

that longer set times could be obtained by using less urease.
Columns of saturated, graded sand have been grouted in the laboratory
by pumping iron solution through them. Measurements using a flexible wall
permeameter indicate that the hydraulic conductivity of the sand dropped
from greater than 10-2 cm/s before grouting to 10-7 cm/s after grouting.
Diffusion of contaminants through iron-based grouts has not been
measured. However, there is extensive literature concerning sorption of metals
to iron minerals such as goethite (e.g. Rodda et al., 1993). Sorption will retard
contaminant transport through an iron-grouted barrier.
Long-term tests have not been conducted. Nevertheless, because this
grout mimics naturally occurring mineral materials, durability is expected to
be very good in some environments.
TABLE 9-4 Overview of Iron Hydroxides
Viscosity:
1 mPas
Set time:
depends on initial pH, amount of urease*
10-7 cm/s*
Wet-dry cycling:
unknown
acids: expect to be poor

bases: expect to be good
unknown
Reduction of diffusional transport: expect adsorption of metals*

Expected lifetime:
> 25 years*
Repairability:
expect to be good

acceptability:
non-toxic*
Cost of materials:
permeation grouted: $0.70/l, $180/m3 (25%

grout), $180/m2 (1 m thick wall)* jet grouted:
$0.20/l, $50/m3 (25% grout), $50/m2 (1 m thick
wall)*
under development in the laboratory
Is material commercially available? no

can be made from low cost industrial by products
The primary components of the iron solutions tested (i.e. iron, citric acid,
urea, and urease) are all naturally-occurring and non-toxic. However, one of
the attributes of this grout is that it can be made from low cost industrial byproducts. These by-products may contain other materials as well. These
other materials will need to be tested for safety and effectiveness on a caseby-case basis.
223
Because iron-based grouts can be made from inexpensive by-products,

the primary costs involved are transportation and additives. For permeation
grouting, urease is the most expensive component. The costs given in Table
9-4 may be further cut as the grout formulation is refined. For jet grouting,
urea and urease are not needed, so, as listed in Table 9-4, the material is much
less expensive.
9.2.5
Montan Wax
Montan wax is a fossilized plant wax extracted from coal or peat deposits. It
is a hard, high melting point, non-toxic material that has been used
commercially for over 100 years in polishes, carbon paper, as a dispersing
and lubricating agent, and as a mold release agent. 80% of the worlds supply
of montan wax comes from the Romanta plant in Amsdorf, FRG. It has been
used in Germany and tested in the United States by Golder Associates Inc.
(1993).
Montan waxes are mixtures of pure wax (70%), resins (20%), and asphaltlike materials (10%) composed of C-24 to C-32 carbon chain esters of long
chained acids and alcohols. Montan wax grout is a suspension-type grout,
consisting of a stable emulsion of montan wax (20%), water (78%), and an
emulsifier (2%).
To break the emulsion, 2-5 %wt sodium or calcium bentonite clay is added
just prior to injection. The clay binds the emulsifier and causes the emulsion
to break. The resulting wax/bentonite mixture is highly viscous and can
significantly reduce the hydraulic conductivity of a soil matrix.
The viscosity before setting depends on the amount of montan wax and
the amount of bentonite. The set time depends on the type of bentonite used,
the amount of mixing, and the soil type. In laboratory tests, rapid mixing
results in much faster setting. Set time is difficult to control, ranging from
about 20 minutes to several hours.
Hydraulic conductivity was measured by pumping montan wax through
soils from DOEs Hanford and Sandia sites. Hydraulic conductivities of soils
from the Hanford site changed from 3 x 10-2 cm/s without grouting to 2 x 10-4
to 5 x 10-5 cm/s with grouting. Hydraulic conductivity of soils from Sandia
dropped from 1 x 10-3 cm/s to 6 x 10-6 to 6 x 10-8 cm/s after grouting.
Montan wax samples that were left to desiccate in petri dishes lasted
several days before dehydrating and becoming fragile.
To test for resistance to chemicals (ASTM C267-82), cylindrical samples
of grouted silica sand were immersed for 28 days in one of the following
chemical solutions:
4N HCl,
6N NaOH,
20% CuSO4,
100% methanol,
224
100% ethylene glycol,

100% aniline, and
100% xylene.
The xylene and aniline had no observable effect on the grouted sand. The
acid, base, and cupric sulfate resulted in some softening at the surface that
seemed to progress very slowly. The ethylene glycol and methanol caused
the development of a disturbed and loosened zone in the top portion of the
sample.
TABLE 9-5 Overview of Montan Wax
Viscosity: depends on composition:
montan wax
%
bentonite
%
water
%
viscosity
mPas
20
78
5-7
18
78
10-20
16
77
20-50
very sensitive to type of bentonite used and

amount of mixing*
Set time:
Hydraulic conductivity measured: 10-4 to 10-7 cm/s

Wet-dry cycling:

acids: fair*
bases: fair*
organics: fair*
unknown
Reduction of diffusional transport: sorbs organics; attenuation measured for TCE

Expected lifetime:
25 years
Repairability:
expect to be good

acceptability:
non-toxic*
Cost of materials:
$1.50/kg, $1.10-$1.30/l, $300/m3 (25% grout),

$300/m2 (1 m thick wall)
commercially used in the field in Europe

set time is difficult to control
Montan wax is nontoxic. There may be some residual toluene, but in

laboratory tests it did not leach. Recently, evolution of methane gas has been
encountered during long term storage of Montan wax. The methane is thought
to be the result of anaerobic activity, which seems to occur with certain
emulsifiers when stored at high temperatures.
9.2.6
225
Sulfur Polymer Cement
Sulfur polymer cement (SPC) was developed by the U.S. Bureau of Mines to
utilize surplus sulfur. It is formed by reacting elemental sulfur with
dicyclopentadiene and oligomers of cyclopentadiene. Since it is a
thermoplastic material, there is a reversible change of viscosity when heated.
It has been used for construction of chemical vats and for road repairs.
Brookhaven National Laboratories has tested Martin Chemicals Chement
2000 product for environmental remediation applications (Heiser and Milian,
1994).
SPC must be heated above its melting point (119 C) to be mixed with
soil. At 135 C, its viscosity is 28 mPas (5.85 x 105 lb-s/ft2). SPC sets when
cooled below its melting point, so its set time depends on the rate of cooling.
To emplace SPC, soil must be preheated. Local heating could be
accomplished using modified jet grouting equipment. Such equipment might
inject steam from a nozzle just above the nozzle from which grout is injected.
As the stem is lifted from the soil, soil would be heated and grouted in close
succession.
Laboratory tests were conducted by mixing SPC and soil, heating to 124 C,
and pouring into sample holders. Three soil types were tested; blended sand,
blended sand and coarse stone, and Hanford soil.
When immersed in ionized water, SPC-grouted Hanford soils cracked
after 3 to 5 days. This was probably because of swelling clays present in the
soil. When the soil was initially heated to 124 C, residual soil moisture would
have been driven off, and clays would have shrunk to their minimum volume.
On immersion in water, certain clays would swell and induce tensile stresses,
which could cause the cracking observed in the samples. Therefore, SPC can
not be used in clay-containing soils that will get wet.
Grouted samples were tested for resistance to chemicals by immersion
for 30-90 days in:
pH 2 aqueous nitric acid solution,
pH 12.5 aqueous sodium hydroxide solution, and
trichloroethylene (TCE)- saturated water.
Samples were removed every 30 days and tested for compressive strength
(ASTM-C39).
Resistance to acids was excellent. There were no visual or dimensional
changes observed. Composites showed downward trends in the average
strength, but the strengths were within one standard deviation of the baseline.
Resistance to bases was poor. No visual or dimensional changes were
observed, but the composites showed downward trends in the average
strength. SPC is known to be attacked by bases, and so deterioration was
expected.
Resistance to TCE was fair. The surface developed a drip-like pattern
226
and blotchy discoloration. Strength loss was measured after 90 days of

immersion.
TABLE 9-6 Overview of Sulfer Polymer Cement
Viscosity:
28 mPas at 135C*
Set time:
depends on cooling*
Hydraulic conductivity measured: 10-10 cm/s*

Wet-dry cycling:
swells and cracks when immersed in water*

acids: good*
bases: poor*
organics: fair*
good*

Expected lifetime:
>25 years
Repairability:
by remelting

acceptability:
not regulated
Cost of materials:
$0.31/kg, $630/m3 (25% grout), $190/m2 (0.30 m

thick wall) note that emplacement would be
more expensive than usual, because of the need
to heat the ground
laboratory tested
Is material commercially available? no

emplacement requires preheating of soil
To test for resistance to radiation, samples were dosed with a total of 1 x

108 rads at a rate of 1 x 106 to 4 x 106 rad/hr. The samples showed no visual
or dimensional changes. Destructive tests for compressive strength (ASTM
D-695) showed no changes of strength.
9.2.7
Epoxy
CARBRAY 100 is an alcohol-based elastomeric epoxy available from Carter

Technologies. It was designed as a two component epoxy for use with
conventional jet grouting equipment to encapsulate buried radioactive waste.
It may also be used for construction of thin barrier walls. It hardens to a
tough rubber-like material resembling a soft polyurethane. Extending the
product with water produces a softer product with physical properties similar
to polyacrylamide gel polymers. It can also be blended with particulate matter
as an extender.
CARBRAY 100s viscosity of 5 to 20 mPas (1 to 5.22 x 105 lb-s/ft2) is very
low compared to other epoxies. It was designed for jet grouting, but it can
227
also be used for permeation grouting. The grout will fully wet most soils at
mix ratios up to about 1 part grout to 4 parts soil. Jet grouting will usually
produce lower mix ratios (closer to 1 to 1) and a correspondingly more elastic
final product.
TABLE 9-7 Overview of Epoxy
Viscosity:
5-20 mPas*
Set time:
8 hours; can be cleaned with soap and water 6

hours after mixing
Hydraulic conductivity measured: expect to be < 10-10 cm/s

Wet-dry cycling:
shrinks upon drying

expect to be good

Expected lifetime:
>25 years
Repairability:
good

acceptability:
non-toxic
Cost of materials:
$8/l, $1600/m3 (25% grout), $490/m2 (0.30 m

thick wall)
limited laboratory testing

unusually low viscosity for an epoxy
CARBRAY 100 costs $8/l for pure resin. In a Soil Saw application in
damp to wet sandy soils averaging less than 20 blows/foot, resin without
water could form a 0.3 m wide wall using a mix ratio of 25% CARBRAY by
volume. The materials cost for this wall would be about $490/m2. A jet
grouted diaphragm panel of thickness 0.15 m and 1-to-1 mix ratio would also
cost about $490/m2.
9.2.8
Polysiloxane
Polysiloxanes are chemically and biologically inert silicon-based chain

polymers. Lawrence Berkeley National Laboratories (Moridis et al., 1995b)
has conducted laboratory and field tests with Dow Corning products. PSX527 was used in laboratory tests. 2-7154-PSX-10 was used in field tests after
being developed especially for this application. These grouts consist of the
following five components:
228
di-vinyl-terminated polydimethylsiloxane polymer,

di-hydrogen-terminated polydimethylsiloxane polymer,
a cross-linker,
a catalyst, and
an inhibitor which can be employed to control cross-linking after
the catalyst is added.
The lengths of the polymer chains determine the viscosity. The crosslinker and catalyst are present in very small amounts. Cross-linking develops
when reactions occur between the H terminations and the cyclic or terminal
vinyl groups. Several hydrogen-terminated chains can link to a single cyclic
molecule to form a complex polymer network.
TABLE 9-8 Overview of Polysiloxane
Viscosity:
PSX-10: 10 mPas; PSX-527: 35 mPas
Set time:
depends on catalyst*
10-10 cm/s
Wet-dry cycling:
expect little effect

organics: may swell*
expect to get stronger

Expected lifetime:
>25 years
Repairability:
good*

acceptability:
non-toxic; inert
Cost of materials:
$20/kg, $20/l, $5000/m3 (25% grout), $1500/m2

(0.30 m thick wall); when mass produced, may
cost $4/kg, $4/l, $1000/m3 (25% grout), $300/m2
(0.30 m thick wall)

high cost
Set time of PSX is controlled mostly by varying the amount of catalyst,

but there is a lower bound to the catalyst amount, below which a solid phase
is not formed. Should longer set times become necessary during application,
the retardant can be added to effect an additional means of control. Soil seems
to have a slightly accelerating effect on the gelling of PSX, but the degree of
acceleration can be easily compensated for by adjusting the catalyst
concentration.
229
In laboratory tests at Lawrence Berkeley National Laboratories, Hanford

sand columns of hydraulic conductivity 10-2 cm/s were injected with PSX527. The resulting hydraulic conductivities were 10-10 cm/s.
Resistance to chemicals has not been measured. Other polymers similar
to PSX sometimes undergo swelling and changes in elastic properties when
exposed to solvents, but they do not deteriorate.
Merging of setting PSX plumes has been tested in the field and appears
to have been successful. Quantitative measurements are needed to
demonstrate the imperviousness of the plume interface.
9.2.9
Furan
Brookhaven National Laboratory (Heiser and Milian, 1994) has tested a

commercial, furfuryl alcohol based furan polymer, FA-Rok 913, manufactured
by QO chemicals. The furfuryl alcohol resin is biodegradable, nonflammable,
and water-soluble. Furans have been used in the fabrication of polymer
concrete pipes, as an organic cementing and sand consolidating material in
oil wells, as floor coatings, and as chemically resistant containers.
Polymerization of furan occurs through a condensation reaction using a
strong acid catalyst. Brookhaven National Laboratories used FA-Rok catalysts.
Polymerization is inhibited by basic reagents. Basic soil consumes the acid
and prevents polymerization.
Set times and exotherms can not be controlled as well as in other polymer
systems. Free floating resin will polymerize very quickly without the heat
sink that the aggregates provide. Care should be taken to avoid free standing
catalyzed resin in order to prevent localized excessive exotherms.
Laboratory tests were conducted using a grout having a one to two hour
gel time. Samples were prepared by mixing grout with soils and pouring
into sample holders. Two soil types were tested: blended sand, and blended
sand with coarse stone.
Hydraulic conductivity measured using a flexible wall permeameter
(ASTM D-5084) for the furan/sand/stone composite was 5 x 10-11 cm/s.
Hydraulic conductivity for the furan/sand composite was 3 x 10-8 cm/s.
Furan grouts showed excellent performance when subjected to combined
wet-dry cycling and temperature cycling. Samples were cycled 12 times, from
60 C dry to 20 C wet (ASTM D-4843). Samples were weighed between cycles
and compression tested (ASTM D-0695) after the final cycle. Weight changes
of grouted samples averaged 0.23%. Strengths increased by an average of
57%. There may have been further curing induced by the elevated temperature
during the dry cycle.
To test for resistance to chemicals, grouted samples were immersed for
30-90 days in:
230
(ASTM-C39).
Resistances to acid, base, and TCE were excellent. Immersion showed no
visual or dimensional changes. No loss of strength was measured.
To test for resistance to irradiation, samples were dosed with a total of 1 x
108 rads at a rate of 1 x 106 to 4 x 106 rad/hr. No visual or dimensional
changes were observed.
TABLE 9-9 Overview of Furan Polymer
Viscosity:
3-8 mPas
Set time:
not well controlled; builds heat easily*
Hydraulic conductivity measured: 10-8 to 10-10 cm/s*

Wet-dry cycling:
little effect*
acids: good*
bases: good*
organics: good*
good*

Expected lifetime:
>25 years
Repairability:
new polymer adheres to old

acceptability:
non-toxic polymer; non-toxic, non-flammable,

water soluble, biodegradable resin
Cost of materials:
$2.35/kg, $2700/m3 (25% grout), $800/m2 (0.30

m thick wall)
laboratory tested

inflammable, biodegradable resin; does not set in

basic soil
9.2.10
Polyester Styrene
Polyester styrenes (PES) are among the most widely used thermosetting resins.
PES polymers are a mixture of a linear polyester resin and styrene monomer.
Chemical and physical characteristics of the final polymer depend on the ratios
of polyester resin to styrene. A need for strong alkali led Brookhaven National
Laboratory (Heiser and Milian, 1994) to study a modified bisphenol fumarate
resin distributed by Reichhold Chemicals using the tradename Atlac 4010A.
A 6% solution of cobalt naphthenate was used as the promoter. Methylethyl
detone peroxide was added as the initiator.
Laboratory tests were conducted using a grout having a 1 to 2 hour set
time. Samples were prepared by mixing grout with soils and pouring into
231
sample holders. Three soil types were tested: blended sand, blended sand
and coarse stone, and Hanford soil.
PES grouts showed excellent performance when subjected to combined
of grouted Hanford soil averaged -1.2%. Grouted sand/stone had weight
change of 0.004%. There was no change of strength.
TABLE 9-10 Overview of Polyester Styrene
Viscosity:
300 mPas
Set time:
depends on promotor-catalyst combination,

amount and ratio*
10-10 cm/s*
Wet-dry cycling:
little effect*
acids: good*
bases: good*
organics: fair*
gets stronger*

Expected lifetime:
>25 years
Repairability:

acceptability:
non-toxic polymer; toxic, flammable resin
Cost of materials:
$3.10/kg, $3200/m3 (25% grout), $950/m2 (0.30

m thick wall)
laboratory tested

flammable resin

30-90 days in the following:
(ASTM-C39).
Resistance to acid and base was good to excellent. Immersion showed no
Immersion in TCE resulted in moderate degradation characterized by
232
surface blistering, flaking, small depressions, blotchy and drop-like color

patterns, granular texture, and grout dissolution. Strength dropped slightly
after 30 days, but remained constant thereafter. To test for resistance to
irradiation, samples were dosed with a total of 1 x 108 rads at a rate of 1 x 106
to 4 x 106 rad/hr. Samples changed colors from beige to yellow after
irradiation. No other visual or dimensional changes were observed.
Destructive tests for compressive strength (ASTM D-695) showed significant
strength increase after irradiation. This effect is attributed to additional crosslinking of the polymer chains.
9.2.11
Vinylester Styrene
Vinylester styrene (VES) polymers have been used to encapsulate radioactive

wastes and in a wide variety of applications calling for resistance to harsh
chemicals. Brookhaven National Laboratory (Heiser and Milian, 1994) tested
a product of DOW Chemicals with the tradename, Derakane 470-45. It is an
epoxy novolac-based vinyl ester resin dissolved in styrene. Dimethylaniline
promoter was used in conjunction with a 40% benzoyl peroxide catalyst
system. The polymerization occurs through an oxidation-reduction reaction.
This particular Derakane was formulated to exhibit good resistance to
chemicals, retention of properties at high temperatures, and low viscosity.
Laboratory tests were conducted using a grout having a 1 to 2 hour set
time. Samples were prepared by mixing grout with soils and pouring into
sample holders. Three soil types were tested: blended sand, blended sand
and coarse stone, and Hanford soil.
(ASTM D-5084) ranged from 1 x 10-10 to 7 x 10-10 cm/s.
VES grouts showed excellent performance when subjected to combined
of grouted Hanford soil averaged -0.7%. Grouted sand and sand/stone had
weight change of less than 0.2%. There was no change of strength.
(ASTM-C39).
Resistance to acid, base, and TCE was good to excellent. Immersion
showed no visual or dimensional changes. No loss of strength was measured.
233
D-695) showed significant strength increase after irradiation. This effect is
attributed to additional cross-linking of the polymer chains.
TABLE 9-11 Overview of Vinylester Styrene
Viscosity:
100 mPas
Set time:

amount and ratio*
10-9 to 10-10 cm/s*
Wet-dry cycling:
little effect*
acids: good*
bases: good*
organics: good*
gets stronger*

Expected lifetime:
>25 years
Repairability:

acceptability:
non-toxic polymer; toxic, flammable resin
Cost of materials:
$3.50/kg, $3600/m3 (25% grout), $1100/m2 (0.30

m thick wall)
laboratory tested

flammable resin
9.2.12
Acrylic
Methacrylate monomers (acrylic) are a commonly used family of polymers.

Brookhaven National Laboratory (Heiser and Milian, 1994) investigated the
use of a series of acrylics manufactured by the 3M company under the
tradename 3M 4R Concrete Restorer. It is a modified high molecular weight
methacrylate. The system consists of dicyclopentadienyl methacrylate and
isooctyl acrylate. It is polymerized using a cobalt octoate promoter (reducing
agent) and cumene hydroperoxide initiator. The laboratory test results
presented are for the low modulus product 3M 4R 5741. Based on work with
Brookhaven National Laboratory, 3M has recently produced the 5750 series
of resins, which is less than half the cost of the original formulations.
Set time depends on the promoter-catalyst combination, amount, and
ratio. It can vary from minutes to days. Laboratory tests were conducted
with a cobalt naphthenate and cumene hydroperoxide system having a 30-60
234
minute gel time (24C). Field tests were conducted with a TEABPO system
having a 90-100 minute gel time (24C). It is expected that set times as long as
one day are possible.
Laboratory samples were prepared by mixing grout with soils and pouring
into sample holders. Three soil types were tested: blended sand, blended
sand and coarse stone, and Hanford soil.
(ASTM D-5084) ranged from <2 x 10-11 to 4 x 10-9 cm/s.
TABLE 9-12 Overview of Acrylic
Viscosity:
5-10 mPas
Set time:

amount and ratio*
Hydraulic conductivity measured: <10-9 to 10-11 cm/s*

Wet-dry cycling:
little effect*
acids: good*
bases: good*
organics: fair*
gets stronger*

Expected lifetime:
>25 years
Repairability:

acceptability:
non-toxic polymer, toxic resin
Cost of materials:
3M 4R Concrete Restorer 5741: $14/kg,

$15000/m3 (25% grout), $4400/m2 (0.30 m thick
wall); 5750 series costs less than half this; further
reductions are possible
Is material commercially available? 3M 4R Concrete Restorer 5741 is; 5750 is not

very expensive
Acrylic grouts showed excellent performance when subjected to combined

of grouted Hanford soil averaged 1.0%. Grouted sand and sand/stone had
weight changes of less than 0.2%. Acrylic/Hanford composites showed slight
increases of strength; acrylic/sand and acrylic/sand/stone composites
showed no change of strength.
9.3 FIELD PERFORMANCE
235

(ASTM-C39).
Resistance to acid and base was good to excellent. Immersion showed no
After immersion in TCE, there was moderate degradation characterized
by surface blistering, flaking, small depressions, blotchy and drop-like color
patterns, granular texture, and grout dissolution. Strength dropped slightly
after 30 days, but remained constant thereafter.
D-695) showed significant strength increase after irradiation. - P effect is
attributed to additional cross-linking of the polymer chains.

9.3.1 Permeation Grouting Using Colloidal Silica and Polysiloxane
Lawrence Berkeley National Laboratory conducted a field test near Los Banos
in the San Joaquin Valley, California in which unsaturated soils in a quarry
were permeation grouted with colloidal silica (CS) and polysiloxane (PSX)
(Moridis et al. 1995a). The field test was designed to demonstrate various
aspects of grout emplacement via permeation. It was not intended to achieve
the creation of a continuous impermeable barrier.
The colloidal silica used was Nyacol DP5110, manufactured by PQ
Corporation. This colloidal silica has alumina replacing some of the silica on
particle surfaces. Its solid content was 30 wt %, and its pH was 6.5. Before
injection, it was mixed with a 35 %wt solution of CaCl2 (a 4M solution). The
polysiloxane used was a Dow Corning product, 2-7154-PSX-10.
The soil was heterogeneous with discontinuous and lenticular coarser
and finer strata, and occasional lenses of well-sorted cross-bedded sands. The
matrix was predominantly coarse sand (0.1-1 mm) with large gravel and cobble
clasts (mostly 10-100 mm) and a fine film of yellow-brown clay binding much
of the sandy matrix. The moisture content of the ungrouted soil varied from
2.5 wt% to 5 wt%. Soil was more moist at depths below 3 m, due to an
increase in fines content below 3m.
1500 liters of CS were injected in two wells, one meter apart. 450 liters of
PSX were injected in another well. In each well, the liquid grouts were injected
through 3 ports at depths 3.0, 3.7, and 4.3 m, using a tube-a-machette technique.
During injection, there was no significant rise of pressure, which would have
236
indicated premature gelling.

After grout placement, soil was excavated around the grouted plumes to
a depth of up to 21 feet. Both CS and PSX had set satisfactorily. Despite the
extreme soil heterogeneity, both CS and PSX created fairly uniform plumes of
solidified grout.
The CS grout sealed fractures and large pores in the clays. In zones with
gravels and pores on the order of 10 mm, CS did not fully saturate the voids,
but appeared to have sealed access to them.
PSX filled and sealed large pores and fractures, as well as accessible small
pores located in the vicinity of the larger pores and fractures. In extremely
large voids, it coated the individual rocks in the gravel and effectively isolated
these zones. PSX also permeated clays and silts, which is unusual. The
mechanism through which this penetration was achieved is under
investigation.
Cored or carved samples were taken from the boulder-sized grouted
chunks for insertion into a flexible wall permeameter to determine hydraulic
conductivity. Coring could only be done in material containing no pebbles.
Grouted materials were found to have hydraulic conductivities two orders
of magnitude less than the ungrouted sand fraction, which, in itself, was less
permeable than the actual geologic matrix. The final hydraulic conductivities
of cored field samples grouted with PSX ranged from 1 x 10-4 cm/s to 3 x 10-4
cm/s. The carved field samples grouted with CS had final hydraulic
conductivities of 2 x 10-4 to 6 x 10-4 cm/s.
The hydraulic conductivities of the field samples were substantially higher
than had been measured in the laboratory. This could be attributed to
incomplete saturation of the pore space and damage to the samples during
recovery, transport, storage, and trimming. Better sealing may be achievable
by multiple, sequential injections of grout.
9.3.2
Permeation Grouting Using Colloidal Silica
DuPont has conducted a field test in which a saturated, fine- to mediumgrained sand was permeation grouted with colloidal silica (CS) (Noll, 1993).
In one phase of this test, grout was injected to simulate the stabilization of a
contamination hot spot. In another phase, the construction of a horizontal
floor was attempted.
The material used was Ludox SM colloidal silica from DuPont. The
Ludox was mixed with HCl and NaCl and diluted to 5 %wt colloidal silica.
The mixes were designed to have 40 and 72 hour set times for the hot spot
and horizontal floor tests, respectively. Eight batches were mixed during 38
hours of injection for the hot spot stabilization. Thirteen batches were mixed
during the 72 hours of injection for the horizontal floor.
Pump tests indicated the sand had an initial hydraulic conductivity equal
to 1.3 x 10-2 cm/s. Computer modeling was used to design the most efficient
injection and extraction well systems for the two tests.
237
For the hot spot stabilization test, CS was injected in a single vertical well
surrounded by 6 vertical extraction wells, located at a radial distance of 3 m
from the injection well. The pumping rate for the injection well was 12 l/min
and for each extraction well, 2 ml/min. Tests indicated that NaCl tracer,
introduced at the injection well, reached the extraction wells in about 12 hours.
However, when CS was injected, CS reached one extraction well in 17 hours
and had not reached the other five wells after 36 hours. Ground penetrating
radar and soil borings both indicated that CS travelled radially 2 - 3 m from
the injection wells and to a depth of up to 4 m. Slug tests conducted after
emplacement of CS indicated a hydraulic conductivity of 3 x 10-6 cm/s.
For the horizontal floor test, the objective was to create a bottom barrier,
thus completing the construction of a containment system formed by the floor
and four vertical cement-bentonite barrier walls that had been previously
constructed. The four barrier walls formed a 4 m x 4 m square. CS was
injected in a horizontal well parallel to and near the bottom of one barrier
wall. Additional horizontal and vertical wells were used to inject and extract
water, thereby creating paths by which the CS might flow to the bottom of
the opposite barrier wall. Fluids were injected into and extracted from the
horizontal wells at 6 l/min. However, CS did not reach the opposite slurry
wall, as revealed by both ground penetrating radar and soil borings following
the emplacement. Pump tests also suggested discontinuities in the CS floor.
CS did not travel as far as expected in either of these tests. Possible reasons
for this include the following.
When constructing the horizontal floor, increased calcium content in

the mix water caused later batches of CS to have shorter set times.
The vertical component of CS flow was not accounted for during the
computer modeling; thus, some of the injected CS may have flowed
vertically, rather than horizontally as intended.
There may have been unidentified soil heterogeneities.
Attempts at sample collection for laboratory determination of hydraulic

conductivity were unsuccessful. However, pump tests indicated that hydraulic
conductivity was substantially reduced in the hot spot stabilization test.
9.3.3 Permeation Grouting Using Sodium Silicate and Montan Wax
Golder Associates Inc. has conducted two sets of field tests of permeation
grouting using sodium silicate and montan wax. The sodium silicate tested
was a glyoxal-modified sodium silicate grout developed by the French
chemical company, Societe Francaise Hoechst. The Montan wax grout was a
mixture of mineral wax and clay developed by the German companies,
Vereinigte Mitteldeutsche Braunkohlenwerke AG and DBI Gas-und,
Umwelttechnik GmbH.
The first set of tests (Golder Associates, 1994) was designed to evaluate
238
the permeation characteristics of the grouts in a range of soil types and

conditions. Single-borehole injection tests were performed at the MWLID
site at the Sandia National Laboratories. Grouts were injected at multiple
levels within each of four boreholes in soil types ranging from low hydraulic
conductivity silts to high hydraulic conductivity coarse sands and gravel
deposits. Grout was injected using the tube-a-manchette method.
The sodium silicate grout mix was comprised of two components: a
mixture of 3% Klebolink H1, 8% Klebolink S, and water; and 16% Klebolink
K. The grout mix had a viscosity 2 to 5 mPas (4.2 to 10.4 x 105 lb-sec/ft2).
The two components were mixed together using a dual-phase system that
simultaneously drew fluid in specified proportions from two reservoirs and
mixed the components at the wellhead, just prior to injection.
Three montan wax grout formulations were tested. Montan wax content
ranged from 9 to 18%. Bentonite content varied from 1 to 2 %. The montan
wax grout was manually mixed and injected in batches, because the
prehydrated bentonite was viscous and difficult to pump. The grout was
continuously stirred with a recirculating fluid pump to minimize plugging in
the injection line.
Geophysical methods were used to monitor movement of grouts,
including measurements of electromagnetic resistance, moisture content, and
temperature. These methods gave some indication of the arrival of the grouts
in the vicinity of boreholes, but none of the methods gave consistent indications
of grout permeation.
Soil samples were collected using a Shelby tube, but samples taken for
gas permeability tests were disturbed during collection. Soil samples were
manually inspected to assess the presence of grout, by sight and smell.
The grout penetration distances typically ranged from 1 to 2 m, for both
the sodium silicate and montan wax grouts. Grout penetration appeared to
have been either asymmetric, heterogeneous, or both.
These test results suggest that both the sodium silicate and montan wax
grouts can be injected in unconsolidated soils using conventional grout
permeation methods, provided the soils have sufficiently high hydraulic
conductivity. The minimum hydraulic conductivity for permeation grouting,
using either sodium silicate or montan wax grout, appears to be around 5 x
10-4 cm/s. Permeation grouting was not possible within the fine-grained
materials at the site, due to low hydraulic conductivities.
A second set of tests (Voss et al., 1995) was undertaken to demonstrate
the feasibility of constructing subsurface barriers under field conditions. The
primary goal was to produce a continuous horizontal barrier with a hydraulic
conductivity of 10-6 cm/s. Multiple-borehole tests were conducted near
Richland, Washington.
The site had laterally continuous, poorly-graded medium grained sands
at a depth of 5 m below the ground surface. Beneath the medium sands were
finer sand and silt. The hydraulic conductivities of the medium sand and
fine sand were 9 x 10-2 cm/s and 8 x 10-3 cm/s, respectively. The water table
239
was about 38 m below surface.

Grouts were injected between 3 and 5 m deep, using tube-a-manchettes.
Geophysical monitoring was used to monitor resistance and moisture.
Permeation grouting using the montan wax grout was largely
unsuccessful. Grout was observed flowing at the surface along or near the
borehole annulus, despite various attempts to reseal the borehole, inject
additional boreholes, or install sleeves with additional locations to inject
annular sealing material. Hydraulic fracturing of the soil was suspected.
The set time of the montan wax emulsion is highly sensitive to the type of
bentonite clay, the amount of bentonite clay, and the amount of mixing that
occurs. The presence of natural clays in the soil can also affect set time. Under
field conditions, it was very difficult to control either the mixing time or the
amount of mixing. As a result, the break time was likely to have varied and
the lack of permeation may have been due to grout setting soon after it was
injected into the soil. Similar problems have been encountered in the German
montan wax grout program, where permeation grouting has largely been
abandoned in favor of other application methods, such as jet grouting.
The sodium silicate grout was successfully injected, but placement
direction and distribution were uncertain. There were two occurrences where
grout flowed to the surface, but these situations were corrected by resealing
the borehole annulus. Geophysical methods indicated that grout flowed
radially over 1 m from the point of injection, but there was no indication that
the target penetration distance of 3 m was achieved. This may have been due
to significant vertical grout movement, or that the geophysical tools failed to
detect the presence of the grout.
The grouted soils have recently been excavated, but a full report is not
yet available. It is doubtful that a continuous layer was made. Penetration
was sometimes uniform, but sometimes there was much fingering.
9.3.4
Jet Grouting Using Montan Wax
FlowTex, a German company, has been testing their Flowmonta technology

for creating horizontal barriers with montan wax (Sass, 1995). By constructing
a bowl-shaped barrier beneath a waste site, the uncertainties associated with
keying vertical walls into a natural floor are avoided.
Four methods have been proposed for verifying the integrity of a
constructed bowl-shaped barrier:
pump tests,
horizontally drilled wells, utilizing fiber optics and geoelectric seismic
topology,
core sampling, and
utilization of a geoelectric direct control system to measure changes in
electrical resistivities between overlapping borings (still under
development).
240
A combination of horizontal directional drilling and jet grouting is used to

emplace the barrier. Drilling begins at one side of the waste site, passes beneath
the waste, and exits at the other side. At the exit side, a jetting tool is attached
and jet grouting proceeds while the tool is pulled back to the beginning side.
Grout is jetted in only two directions, to create overlapping panels. Figure 95 illustrates how a system of overlapping panels could be used to create a
bowl-shaped barrier.
Vertical
sides
Injection
borehole
Horizontal base
Overlapping
panels
Figure 9-5 FlowTex barrier system (Sass, 1995)

In preparation for constructing a bowl-shaped barrier, grouts of different
viscosities were tested by injecting 200 m long columns. Based on the observed
performance, it was concluded that it would be possible to construct 1000 m
columns. Drilling accuracy is within 0.1 m over 100 m.
The bowl-shaped barrier was to be 9 m deep and 100 m long. The soil
was an overconsolidated heterogeneous glacial till with clay and sand
intercalations. Overconsolidation and shallow depth made fracturing to the
surface a possibility. The soil hydraulic conductivity was as low as 10-5 cm/s,
two orders of magnitude below that normally desired for jet grouting.
A mixture of montan wax and cement bentonite was injected. The cement
bentonite was prepared and mixed with montan wax. The grout did not
appear to follow preferential paths. Its movement was predictable for at least
1.5 m from the points of injection. After penetration over several meters,
horizontal fracturing and uncontrolled flow occurred.
Post-placement coring has shown grout to be located where expected.
Currently, additional tests are being conducted, including: pump tests, tests
for radon penetration, and tracer gas tests using carbon dioxide and helium.
At the completion of these tests, the site will be excavated to examine
undisturbed samples.
9.4 NEEDS
241
This test has revealed the need for well-controlled gel times and better site
characterization, especially when heterogeneous soil conditions are present.
9.3.5
SoilSaw Emplacement of Montan Wax
SoilSaw has been used to emplace montan wax in combination with 4%

pre-hydrated bentonite (E. Carter, 1995). SoilSaw is a modified form of jet
grouting, designed to make continuous, uniform, barrier walls at speeds higher
than other methods. It consists of a heavy beam that is reciprocated in the
ground by a crawler mechanism. Several jet nozzles are located along the
length of the beam. As the beam is passed through the soil, the injected grout
disrupts and mixes with the native soil to form backfill in a single step.
SoilSaw operates at ten times the flow rate of typical jet grouting. See Section
3 of this report for more discussion of the SoilSaw method.
The montan wax-bentonite combination produced a soft, plastic material.
Upon examination, the side cuts looked good. However, the montan wax
did not provide any significant performance improvement over bentonite
alone.
Difficulty was encountered in obtaining undisturbed samples from the
field for laboratory analysis. A method was developed for hand trimming
samples in the laboratory, which seemed to improve things.
9.3.6
Jet Grouting Using Acrylic
At Idaho National Engineering Laboratory, a test was conducted in which

acrylic was jet grouted to stabilize a waste pit (G. Loomis, 1995). A twocomponent resin system was used, one half had catalyst and one half had
promoter. The two components were held in separate tanks and emplaced
through a double-wall drill stem having two nozzles. The two components
did not mix until injection into the soil. By this method, a grout with a short
set time of 90 minutes could be used. Short set times are useful when the soil
is very coarse, since grout would pass through the coarse soil if it did not set
quickly.
In this test, 35 columns were constructed using two different resin blends.
Gel times in the field matched laboratory measurements. Further results are
not yet available.
9.4 NEEDS
9.4.1 Assessment of Field Performance
Extensive laboratory testing has been completed for a wide range of chemical
grouts, but only a few chemical grouts have been tested in field studies. For
the field tests that have been conducted, evaluation of placement methods
242
was emphasized, rather than barrier performance. Additional field tests need
to be conducted with more attention given to barrier performance.
In many of the field tests, investigators found or suspected deterioration
of the collected field samples, making laboratory verification of field
performance difficult. Better sampling methods need to be developed and
disseminated to address this problem.
Based on the findings from additional field tests conducted to evaluate
barrier performance, it may be possible to design laboratory tests that will
better simulate field performance. Field performance can be affected by many
factors, including site soil chemistry, overburden pressure, and the
emplacement method used. In studies using sodium silicate, strengths of
grouted soils measured in the field were about three times those measured in
laboratory-prepared samples. In tests with jet grouting, lower hydraulic
conductivities have been measured in the field than in the laboratory (possibly
due to sample disturbance). Although laboratory testing is much faster and
less expensive than field testing and it allows the variation of more parameters,
care must be given to assure that laboratory test results are true indicators of
field performance.
9.4.2 Assessment of Long Term Performance
Service life expectancy of containment systems for environmental remediation
applications is an important performance criterion. Service life expectancy
will remain uncertain until constructed grouted barriers have been in the field
for extended periods of time. However, laboratory testing can be used as a
basis to estimate life expectancy. Such laboratory tests are not presently being
conducted. A focused effort should be made to adapt testing methods from
other disciplines to testing of barrier materials. In other areas of material
science, tests are routinely conducted to predict long term performance using
the results from short-term tests.
Better test methods are needed for determining the long-term performance
of emplaced chemical grouts exposed to contaminants. For acrylate grouts,
test findings indicate that hydraulic conductivity changes when permeated
by various chemical solutions. For many other chemicals, grouted samples
were immersed in chemical solutions for 30-90 days, with assessment of
resistance to chemicals made visually and by strength testing the samples.
Pumping chemical solutions through grouted samples is a better simulation
of long term exposure, and the effect on hydraulic conductivity is of great
importance in the design of environmental containment systems. However,
such testing can be time consuming for very low hydraulic conductivity
materials. Better test methods are needed for these materials.
As discussed in Section 10, the migration of contaminants through barrier
materials can occur by diffusive transport, as well as by advective transport.
Few measurements have been made on the diffusive transport or sorptive
properties of chemical-based grouts. In the absence of measurements, estimates
9.5 SUMMARY
243
may be possible for some materials based on available published data.

Continued evaluation of diffusive transport and advective transport through
these materials may indicate the point at which further reduction of hydraulic
conductivity has little effect on contaminant transport. For some of the
chemical grouts having extremely low hydraulic conductivity, it may be
possible to develop less costly formulations that yield comparably low
contaminant transport rates.
9.4.3
Properties Affecting Verification
As with all containment systems, verification of emplacement and

performance monitoring are critical. Since chemical grouts can be customengineered, it may be possible to develop materials with specific properties
that make detection in the field easier. For example, if the emplaced barrier
material has an electrical resistance that is dramatically different from the
surrounding soil, detection of defects would become more feasible.
Developers of chemical grouts need to work closely with developers of
performance monitoring systems to determine the grout properties that may
make verification and monitoring easier.
9.4.4 Performance Standards
This review has show that there are many chemical grouts presently under
study, primarily in the laboratory. The development of performance standards
will make it easier to determine which materials deserve continued study,
and the grout properties warranting more aggressive research and
development. Some of the materials presently being studied may perform
better than needed. Noll et al. (1992) showed that good performance can be
attained when colloidal silica is diluted by a factor of six, greatly reducing
material costs. Identification of performance needs may make it possible to
develop more cost-effective materials.
9.5 SUMMARY
Table 9-13 presents in summary form some of the key properties of the
materials that have been reviewed. Selection of chemical grouts will depend
on the site conditions and containment needs. Traditional materials, e.g.,
sodium silicate and acrylate gel, provide modest performance at modest cost.
Other more recently identified materials based on inorganics (e.g., colloidal
silica and iron materials) and naturally-occurring materials (e.g., montan wax)
provide improved performance in specific areas. The remaining materials
either require special emplacement methods (e.g., sulfur polymer cement) or
utilize engineered polymers. These remaining materials may provide excellent
performance for a wide range of conditions, but at a high cost.
244
Data obtained from field tests and demonstrations using these newly
developed chemical grouts are limited. There are some indications that field
performance has been poorer than expected, based on laboratory
measurements. To further evaluate the performance of chemical-based grouts,
additional field tests are needed as well as improved methods for retrieving
field samples for laboratory testing. Based on these additional field tests, the
relationship between field and laboratory performance can be better assessed,
and improved laboratory test methods developed, as necessary.
TABLE 9-13 Summary of Key Properties of Chemical Grouts*
Hydraulic
Resistance
Conductivity
to Acids
(cm/s)
Grout
Resistance
to Bases
Resistance
to Organics
Expected
Lifetime
(Years)
Cost**
($/m3)
Cost***
($/m2)
sodium
silicate
10-5
fair
poor
fair
10-20
130
130
acrylate
gels
10-7-10-9
poor
good
fair
10-20
230
230
colloidal
silica
10-8
good
poor
good
>25
60-330
60-330
iron
hydroxide
10-7
poor
good
good
>25
50-180
50-180
montan
wax
10-4-10-7
fair
fair
fair
25
300
300
sulfur
polymer
cement
10-10
good
poor
fair
>25
630
190
epoxy
10-10
good
good
good
>25
1600
490
-10
good
good
good
>25
1000-5000
300-1500
polysiloxane
10
furan
10-8-10-10
good
good
good
>25
2700
800
polyester
styrene
10-10
good
good
fair
>25
3200
950
vinylester
styrene
10-10
good
good
good
>25
3600
1100
good
good
good
>25
7000-15000
2000-4400
acrylics
*
**
***
-9
10 -10
-11
See 9.2 for discussion of properties

Cost of grout, per unit volume of grouted soil
Cost of grout, per unit area of barrier wall
Better laboratory tests are needed for estimating the long term
performance of chemical grouts. Testing is especially needed to determine:
barrier performance in the absence of unusual stresses,
barrier performance when exposed to chemicals, and
diffusive transport of contaminants.
Performance standards are needed to guide the development of costeffective materials. For materials with extremely low hydraulic conductivity,
lower cost versions of these materials may also provide adequate performance.
Although chemical grouts have been used for decades in construction
and water sealing, development of chemical grouts for environmental
9.6 REFERENCES
245
remediation applications is a relatively new field. Chemical grouts can be

made from a wide range of materials, offering the option of customizing to
meet specific performance requirements. Further laboratory testing, field
testing, and performance standards are needed to develop materials that will
provide long term, cost-effective performance.
9.6 REFERENCES
Bodocsi, A., Bosers, M. T., Sherer, R. (1988). Reactivity of various grouts to
hazardous wastes and leachates, EPA/600/2-88/021, prepared for
Environmental Protection Agency under Contract 68-03-3210.
Carter, E. (1995). personal communication.
Clarke W. J. (1982). Performance characteristics of acrylate polymer grout,
Proc. Conf. Grouting in Geotechnical Engrg., ASCE, New Orleans, 418-432,
Golder Associates Inc. (1993). Laboratory test results for evaluating the
performance of montan wax and sodium silicate barrier materials in
unconsolidated soils, 913-1248.600a Draft Report, prepared for the U.S.
Department of Energy under Sandia National Laboratory Contract No.
97-3497.
Golder Associates Inc. (1994). Preliminary results of permeation tests using
montan wax and sodium silicate barrier materials in unconsolidated soils
at the mixed waste landfill integrated demonstration site, 913-1248.600b
Draft Report, prepared for the U.S. Department of Energy under Sandia
National Laboratory Contract No. 97-3497.
Hapka, A. M., J. S. Thompson, J. M. Whang (1995). Method for precipitating
a solid phase of metal, provisional patent application.
Heiser, J. H., L. W. Milian (1994). Laboratory evaluation of performance and
durability of polymer grouts for subsurface hydraulic/diffusion barriers,
BNL-61292 Informal Report, prepared for U.S. Department of Energy
Contract No. DE-AC02-76CH00016.
Karol R. H. (1990). Chemical Grouting, Marcel Dekker, Inc., New York.
Krizek R. J., Michel, D. F., Helal, M., Borden, R. H. (1992). Engineering
properties of acrylate polymer grout, Grouting, Soil Improvement and
Geosynthetics Proc., ASCE, New Orleans, 712-724.
Loomis, G. (1995). personal communication.
Moridis, G. J., Persoff, P., Apps J. A., Meyer, L., Pruess, K. (1995a). A field
test of permeation grouting in heterogeneous soils using a new generation
of barrier liquids, LBL-37554. Supported by U.S. Department of Energy
Contract No. DE-AC03-76SF00098.
Moridis, G. J., Pruess K., Persoff, P., Apps, J. A. (1995b). Performance and
properties of colloidal silica and polysiloxane grouts, notes presented at
246
the International Containment Technology Workshop, Baltimore, 1995.

supported by U. S. Department of Energy Contract No. DE-AC0376SF00098.
Noll, M. R., Bartlett, C., Dochat, T.M. (1992). In situ permeability reduction
and chemical fixation using colloidal silica, Proc. 6th National Outdoor
Action Conference, National Ground Water Association, Las Vegas, NV,
443-457.
Noll, M.R., Epps, D. E., Barlett, C. L., Chen, P.J. (1993). Pilot field application
of a colloidal silica gel technology for in situ hot spot stabilization and
horizontal grouting, Proc. 7th National Outdoor Action Conference, Las
Vegas, NV, 207-219.
Rodda, D. P., Johnson, B.B., Wells, J. D. (1993). The effect of temperature and
pH on the adsorption of copper (II), lead(II), and zink(II) onto goethite,
J. Colloid Interface Sci. 161:57-62.
Rumer, R. R. and M. E. Ryan, editors. (1995). Barrier Containment Technologies
for Environmental Remediation Applications, John Wiley & Sons, Inc.,
New York.
Sass, I. (1995). Personal communication.
Siwula J. M., R. J. Krizek (1992). Performance of grouted sands exposed to
various water chemistries, Grouting, Soil Improvement and Geosynthetics
Proc., ASCE, New Orleans, 1403-1419.
Voss, C. F., Einberger, C. M., Russel, B. F., Matalucci, R.V. (1995). Evaluation
of two new grout materials for in situ barrier construction. Waste
Management 95, Tuscon, AZ.
247
SECTION 10
CONTAMINANT TRANSPORT MODELING
prepared by
Alan J. Rabideau, State University of New York at Buffalo

Charles B. Andrews, S.S. Papadopolous & Assoc.

Chen Chiang, Shell Oil Co.
Peter Grathwohl, University of Tbingen, Germany
Joel S. Hayworth, Armstrong Laboratory, Tyndall AFB
Patricia Culligan-Hensley, Massachusetts Institute of Technology
Mario Manassero, Ingegneria Geotecnica, Italy
James W. Mercer (session co-leader), GeoTrans Inc.
Henry Mott, South Dakota School of Mines and Technology
Paul R. Schroeder, U.S. Army Corps of Engineers
Charles Shackelford, Colorado State University
Georg Teutsch, University of Tbingen, Germany
G.A.M. van Meurs, Delft Geotechnics, The Netherlands
Ron Wilhelm, U.S. Environmental Protection Agency
Chunmiao Zheng, University of Alabama
10.1
BACKGROUND
10.1.1
Overview
Mathematical modeling is a well-established tool for the analysis and design

of systems for environmental remediation. Due to the lengthy time scales for
transport processes and the anticipated long decision horizons, modeling is
expected to play a significant role in the analysis of containment systems. A
number of aspects of barrier performance are amenable to modeling analyses,
including:
conceptual design of low-permeability barriers;

conceptual design of permeable reactive walls;
performance assessment of barrier systems; and
risk assessment for systems that include barriers.
The focus of this section is on the analysis of barrier performance with

respect to contaminant transport; other issues, such as the structural behavior
of barriers, are considered only as factors that might influence contaminant
migration. The barrier performance analysis can be referenced to a local scale
248
(e.g., transport through a low-permeability wall) or a regional scale where

the entire barrier system is considered. Except where noted, barrier systems
are treated as porous media, under the assumptions generally applied in
groundwater modeling (e.g., Bear, 1979), including the continuum
hypothesis, conservation equations, validity of empirical flux laws (Darcys
Law, Fick s Law), and applicable constitutive relationships.
A considerable body of research has developed related to contaminant
transport in porous media. The vast majority of this work, however, has
been applied to systems in which advective transport is the primary concern.
As pointed out by several researchers, however, under design conditions,
molecular diffusion is expected to be much more significant than advective
transport in barrier systems (e.g., Goodal and Quigley, 1977; Crooks and
Quigley, 1984; Gray and Weber, 1984; Johnson et al., 1989). This distinction
is important, and must be reflected in the modeling strategy applied.
Modeling of subsurface transport phenomena may be conducted at
various levels of sophistication, ranging from simple analytical calculations
to complex numerical simulations that stretch the limits of the fastest
supercomputers. A premise underlying this section is that the current
interest in long-term containment as a strategy for subsurface remediation
will result in a demand for user-friendly models that can be applied by
practitioners, including designers and regulators. The discussion is therefore
focused on the development and application of relatively simplified models
that are intended to balance the competing concerns of accuracy,
conservatism, and computational requirements. This focus should not be
construed as a de-emphasis of the need for continued development of
sophisticated numerical models. On the contrary, complex numerical
models provide an avenue for validating and exploring the limitations of
simpler analytical models and, as discussed subsequently, they may be
needed in practical applications where an assessment of the viability of a
proposed remedial strategy is highly sensitive to factors that are not
adequately represented by simpler models. Complex models are also
valuable tools for exploring the relative importance of various transport
processes.
Modeling applications considered in this section are primarily referenced
to the consideration of engineered low hydraulic conductivity barriers
installed to limit the migration of dissolved contaminants by advective and
diffusive transport under saturated groundwater flow. Models for
permeable reactive walls are conceptually similar and are discussed in this
context. Much literature exists regarding landfill liners (e.g., Rowe et al.,
1995) and a detailed treatment of this topic is therefore avoided. While
modeling is also commonly used in the analysis of caps (e.g., Schroeder et
al., 1994), the focus of such efforts is on the prediction of a water balance
and infiltration rates, rather than contaminant migration, and as such this
subject is not considered here.
10.1 BACKGROUND
10.1.2
249
Mathematical Model
The analysis of contaminant transport in groundwater systems is generally

expressed in terms of aqueous phase concentrations. In the models discussed
herein, the contaminant is assumed to be present only in the dissolved and
sorbed phases. While the behavior of nonaqueous phase liquid (NAPL)
contamination is of importance in the analysis of remedial systems, horizontal
NAPL migration into and/or across a barrier system is unlikely to occur, due
to the large pressure gradients required to overcome the capillary forces typical
of fine-grained materials. For a discussion of conceptual and mathematical
models for multiphase contaminant transport, the reader is referred to Abriola
and Pinder (1985) and Cohen and Mercer (1993).
In the discussion that follows, the contaminant mass flux across a
containment barrier is considered the dependent variable of concern.
Contaminant mass flux is preferred, rather than contaminant concentration,
because: (1) under diffusion-dominated conditions, mass transport is driven
primarily by concentration gradients, rather than the actual concentration; (2)
for some one-dimensional conceptual models of interest, the concentration of
contaminant at the barrier/aquifer interface is assumed to be negligible in order
to produce conservative estimates of mass flux; and (3) correspondence between
simplified one-dimensional barrier models and global or field-scale models
can be more readily established by comparing calculated mass fluxes at the
barrier/aquifer interface.
The flux of a single mobile liquid phase contaminant of uniform density
through a saturated porous media with unidirectional groundwater flow in
the x-direction is given by
J = nC nD
C
x
(10-1)
where J is the flux of contaminant mass per unit area at a specified time t , x is
the direction of flow, n is the porosity of the porous media, C is the aqueous
phase contaminant (or solute) concentration, v is the seepage velocity (average
interstitial) of groundwater in the x direction, and D is the longitudinal
dispersion coefficient.
A common approach is to evaluate the above expression for contaminant
transport after solving the mass conservation equation for the solute, henceforth
referred to as the advective-dispersive reactive equation (ADRE)
b S
b
qs
C
C
+
=
iC
Cs a C s
S (10-2)
Dij
t x i x j x j
n t srp n
n
where i and j are subscripts that indicate the direction associated with the
Cartesian coordinate system, x is a spatial coordinate, v is the seepage velocity,
250
D is the dispersion coefficient, qs is a fluid sources/sink, Cs is the contaminant

concentration for the source/sink, b is the media bulk (dry) density, S is the
solute mass fraction in the sorbed phase, the srp represents a sorption reaction,
a is a first-order decay constant for the aqueous phase, and s is a first-order
decay constant for the solid phase.
The seepage velocity is related to the piezometric head gradient by Darcys
law [Eq. (10-4)]. The mass conservation equation for the bulk fluid [Eq. (103)] is given as
x i
K ij x
j
vi=
+ q s = S s t
K ij h
n x j
(10-3)
(10-4)
where K is the hydraulic conductivity, h is the piezometric head, and Ss is the

media specific storage.
The dispersion tensor is commonly represented as consisting of two
components
Dij = Dh + D
(10-5)
ij
where D h represents (mechanical) hydrodynamic mixing and D* is the

effective molecular diffusion coefficient for the solute of interest.
In Eq. 10-5, Dh is considered to be a function of the seepage velocity and
the pore system geometry of the porous media. The reader is referred to
recent textbooks (e.g., Zheng and Bennett, 1995) for further discussion of
dispersion in field-scale models. For barrier systems, the seepage velocity is
expected to be very small, frequently negligible; therefore, molecular diffusion
is considered to be the dominant transport process. The solute diffusivity in
porous media (at the macroscopic scale) is commonly related to the solute
diffusivity in liquid by a relationship similar to the following
D
D = l
(10-6)
where Dl is the liquid diffusivity and is a tortuosity factor.

As presented here, the tortuosity factor is a lumped parameter that
incorporates several mechanisms that inhibit the diffusion of a solute in porous
media, including constrictions and extended flow paths related to the pore
structure. Shackelford and Daniel (1991) referred to this lumped parameter
as an apparent tortuosity factor (a ). The definition and interpretation of
10.1 BACKGROUND
251
tortuosity and effective diffusion coefficients vary throughout the literature,

and the reader should be careful when comparing results presented by
different researchers. In particular, the definition of tortuosity used in this
report does not reflect the fractional reduction of the cross-sectional area for
diffusion represented by the media porosity or retardation due to sorption,
as these factors are explicitly accounted for in Eq. 10-2. Note also that the
tortuosity appears in the denominator of Eq. 10-6 and thus is understood to
be greater than unity. An alternative definition of tortuosity, as a multiplicative
factor less than unity, also has been used extensively (e.g., Shackelford and
Daniel, 1991).
In Eq. 10-2, it is assumed that only two contaminant phases (sorbed and
aqueous) are present. To close the system, an additional equation must be
provided to represent the distribution of the contaminant between the dissolved
and sorbed phases. The term sorption as used here, encompasses a variety
of processes, including: hydrophobic partitioning into soil organic matter, ion
exchange and other surface reactions, and diffusion into the solid matrix and/
or immobile fluid. Most transport models rely on a phenomenological approach
in which the solid phase is treated as a single compartment and an isotherm
expression is used to define the relationship between the dissolved and sorbed
phases at equilibrium. A commonly employed relationship is the linear isotherm
S = kdC
(10-7)
where kd is the solute distribution coefficient.

A nonlinear sorption equilibrium relationship may be described by the
Freundlich isotherm
S = K FC
(10-8)
where KF and nF are the empirical Freundlich parameters. The Freundlich

isotherm is one of several commonly applied nonlinear isotherms (e.g., see
Kinninburgh, 1986, for discussion of other approaches).
If the sorption process is fast, relative to the transport processes, the local
equilibrium assumption is commonly employed. When a linear isotherm is
appropriate, the governing ADRE reduces to
Rf
=
t x i
C
Dij x
j
bkd
qs

x i C + n C s a + s n
b k d
R f = 1 +
n
where Rf is the retardation factor.
(10-9)
252
Although commonly employed because of its mathematical convenience,

the linear local equilibrium assumption may not be appropriate for all systems.
In particular, much research suggests that nonequilibrium sorption can play
a significant role in contaminant transport and remediation when advection
is the dominant transport process. Models for describing nonequilibrium
sorption are numerous, and a detailed review is beyond the scope of this
report (see discussions by Brusseau et al., 1989; Haggerty and Gorelick, 1995).
In general, nonequilibrium models require identification of distinct solid phase
compartments, the solute mass distribution among the compartments at
equilibrium, and rate expressions for mass exchange across compartments.
Equilibrium expressions may be linear or nonlinear. The form of the rate
expression may vary depending on whether the governing theoretical model
is based on assumptions relevant to mass transfer, diffusion, or chemical
kinetics. In general, models that contain more compartments provide for
more flexible data-fitting, but are difficult to parameterize independently and
may be computationally demanding. For this reason, simpler lumped
parameter models are often used in practice. As an illustration, a commonly
applied two compartment, linear, first-order model is given as
dS
= kdC S
dt
(10-10)
where is the sorption rate coefficient.

The other reaction process explicitly represented in Eq. 10-2 is first-order
decay. This reaction term may be interpreted to represent a number of
processes including radioactive, abiotic, and biotic transformations. The decay
constant is related to the contaminant half-life by
ln 2
t5 0
(10-11)
where t50 is the half-life.

a. One-dimensional Geometry. Consideration of a containment barrier
system in the context of the surrounding aquifer requires a complex model
resolved in sufficient detail to represent the considerably different conditions
that influence the contained area, the barrier itself, and the adjoining aquifer.
For this reason it is often convenient to treat the barrier as a separate onedimensional system, with the containment zone and surrounding aquifer
represented as boundary conditions. If source/sink terms and reactions (other
than decay and sorption) are neglected, the one-dimensional form of Eq 10-2
is given as
10.1 BACKGROUND
C
C
2 C b
=
+D 2
t
x
n
x
S C b S
a
s
n
t srp
253
(10-12)
where x is distance measured from the containment zone side of the barrier
in the direction normal to the barrier-containment zone interface (the
dimensional subscripts are henceforth dropped).
The seepage velocity is related to the hydraulic gradient across the barrier
KE h
nL
(10-13)
where KE is the effective hydraulic conductivity in the direction normal to the

plane of the barrier, L is the barrier thickness, and h is the difference in
piezometric head between the inside and outside of the contained area, with
a positive difference representing a larger head on the outside of the barrier.
According to classical dispersion theory (e.g., Bear, 1972) the dispersion
coefficient in Eq. 10-12 can be represented as Eq. 10-14 where a l, the
longitudinal dispersivity, is a property of the barrier media. Although
independent measurements of a barrier dispersivity have not been reported,
the effect of the hydrodynamic mixing represented by this term is generally
believed to be much less significant than diffusion for the low flow condition
associated with barrier systems (e.g., Rowe et al, 1995).
D = al + D*
(10-14)
Two representations of this conceptual model are shown in Fig. 10-1.

Fig. 10-1a is based on a conceptual model described by Rowe et al. (1995) for
a bottom liner, while Fig. 10-1b is based on a similar geometry applied to a
vertical barrier. While similar in form, distinctions are noted between the
two models:
Each model represents the contaminant zone as a completely mixed region

of uniform concentration immediately adjacent to the barrier. For the
commonly applied landfill model, the contaminant zone is understood
to consist of leachate mounded above the bottom barrier, while for the
vertical barrier, the contaminant zone is generally a dissolved contaminant
plume of specified width, frequently assumed to extend over the full depth
of the barrier.
A number of conceptual models for the exit side of the barrier are possible,
depending upon the boundary condition specified. For bottom barriers,
254
one approach is to consider the aquifer as a finite, completely mixed region

characterized by lateral groundwater flow. For vertical barriers, this
assumption is unrealistic, although various degrees of contaminant
flushing may be represented by adopting the mathematical form of the
finite aquifer condition.
Advective transport is driven by the piezometric head gradient across

the barrier. Although this gradient is frequently assumed to be in the
direction of the uncontaminated aquifer for bottom barriers, the gradient
may be in either direction for both bottom and vertical barriers.
Leachate
flow
Hf
X=0
Horizontal barrier
X=L
Groundwater
flow
Confining layer
(a) bottom barrier

X=0
X=L
Vertical
barrier
h1
h2
Contaminant
plume
Hf
Confining layer
(b) vertical barrier wall

Figure 10-1 Schematic diagram of barrier configurations
10.1 BACKGROUND
255
Application of these conceptual models is discussed in more detail below.

b. Auxiliary Conditions. The form of the solution to the ADRE depends on
the auxiliary boundary conditions specified. For engineered barriers, it is
reasonable to assume that the barrier is free of contaminant at initial time, t =
0; i.e.,
C(x,0) = 0
(10-15)
The form of the specified boundary conditions may greatly influence the
solution, particularly for diffusion-dominated transport. For a onedimensional representation of a barrier system, boundary conditions must
be applied at the contaminated side of the barrier (x=0), denoted here as the
barrier entrance, and at the less contaminated side of the barrier (x= L),
denoted here as the barrier exit. In general, boundary conditions are
classified as: (1) first type or Dirichlet conditions, in which the value of the
concentration is fixed at the boundary; (2) second type or Neumann, in which
the gradient of the concentration at the boundary is specified; and (3) third
type or mixed conditions, in which the concentration gradient at the boundary
is a function of the concentration.
The discussion that follows deals with the specification of boundary
conditions for field applications. For analysis of laboratory column
experiments, different boundary conditions may apply (e.g., Shackelford,
1994a) and caution is urged in generalizing results between the laboratory
and the field. For simplicity, solute decay is neglected. For a more detailed
discussion of field scale boundary conditions the reader is referred to Rowe
et al. (1995) and Rabideau et al. (1996).
Entrance condition. The most common and conservative condition applied
to one-dimensional problems is the first type boundary condition; i.e.,
C(0,t) = C o
(10-16)
where Co is assumed to be a known constant concentration.

Eq. 10-16 is popular because it often leads to closed-form solutions of the
ADRE. The fixed concentration condition is conservative in that it does not
provide for the development of a concentration gradient within the
contaminant zone, nor does it account for the reduction in contaminant
concentration that could result from remedial actions or from contaminant
mass transport into the barrier. Consideration of a time-dependent Dirichlet
boundary condition usually increases the computational burden associated
with the solution. Such an effort may be appropriate, however, if timedependent data are available, as might be the case in the retrospective analysis
of historical performance of a system.
An alternative conceptualization of the entrance boundary is the finite
256
mass condition described by Rowe and Booker (1985) [mathematically, this

condition may be considered a third type condition]. Physically, the condition
can be interpreted to represent a known initial mass of contaminant (difficult
to estimate) within a completely mixed contaminant zone. The uniform
concentration in the contaminant zone is then reduced to account for transport
into the barrier
C(0,t) = C o
1 t
dC
I n C(0,) nD
(0,) d
HF 0
dx
(10-17)
where HF is the leachate height (if applied to a landfill liner) or the width of
the contaminant zone normal to the orientation of a vertical barrier.
The application of the finite mass condition results in a smaller amount
of mass transported across the barrier for a specified initial boundary
concentration, and thus is less conservative, but possibly more realistic, than
the constant concentration boundary condition. The sensitivity of the
predicted flux to the width of the contaminant zone, HF, is discussed by Rowe
et al. (1995) for landfill systems and by Rabideau et al. (1996) for vertical walls.
In general, when HF is on the order of the barrier thickness, L, a significant
reduction in flux is noted compared to the constant concentration condition.
However, when HF exceeds the barrier dimension by an order of magnitude
or greater, there is little difference between the two conditions when diffusive
transport dominates (Rabideau et al., 1996).
Other third-type entrance conditions have been proposed for application
to laboratory test columns, including
C o = C(0,t) D
dC
(0,t)
dx
(10-18)
Sound arguments have been advanced by Parker and van Genuchten (1984b)
for the applicability of Eq. 10-18 to advection-dominated laboratory conditions
(based on the conservation of mass flux across the entrance boundary).
However, this boundary condition produces unrealistic flux predictions when
applied under the very low-flow conditions anticipated in the vicinity of lowpermeability barriers (Rabideau et al., 1996). For permeable reactive walls
expected to operate under advection-dominated conditions, either Eq. 10-18
or Eq. 10-16 may be appropriate, depending upon the dispersive properties
of the treatment wall and, if comparisons are made with field data, the manner
in which the field samples are collected (see discussion by Parker and van
Genuchten, 1984b).
Exit condition. One-dimensional transport is sometimes modeled as a semiinfinite system in which the exit boundary condition is defined at x = +,
rather than at the barrier/aquifer interface, i.e.,
10.1 BACKGROUND
C
( 4 ,t) = 0
x
257
(10-19)
Although this boundary condition implies that the transition at the barriersurrounding aquifer interface does not influence contaminant transport within
the barrier, it can facilitate the derivation of closed-form solutions to the ADRE.
For laboratory studies of advection-dominated transport, Eq. 10-19 is
commonly applied. However, for field conditions where diffusive transport
is dominant within the barrier, it is likely that, in the adjoining aquifer outside
the barrier, groundwater flow will result in more rapid removal of the
contaminant exiting the barrier. This scenario may be represented by the
third type boundary condition [proposed by Rowe and Booker (1985) for
describing conditions at the base of a landfill liner],
t
C(L,t) = I
0
t
b C(L, )
n C(L, )
nD dC
(L, ) d I
d
nb hb
n b h b dx
W
0
(10-20)
where nb is the porosity of the adjoining aquifer, hb is the adjoining aquifer

dimension normal to the barrier, vb is the velocity of groundwater flow parallel
to the barrier in the adjoining aquifer, and W is the width of the barrier.
For diffusion-dominated transport, Eq. 10-20 is more conservative than
the semi-infinite condition, since it results in a higher concentration gradient,
and consequently greater diffusive flux, at the barrier exit. In the limit, Eq.
10-20 reduces to a first-type exit condition
C(L,t) = 0
(10-21)
For diffusion-dominated transport, Eq. 10-21 is the most conservative exit

condition, since it implies instantaneous removal of the contaminant exiting
the barrier. It is arguably the most appropriate condition for one-dimensional
models of low hydraulic conductivity vertical barriers. This boundary
condition is henceforth referred to as the perfect flushing condition.
For laboratory studies in which advective transport is predominant, a
second type boundary condition has been proposed (e.g., Danckwerts, 1953)
C
(L,t) = 0
x
(10-22)
Eq. 10-22 is clearly inappropriate for diffusion-dominated transport in that it

implies zero diffusive flux at the barrier exit. Surprisingly, in practice,
laboratory results from advection-dominated studies are often extrapolated
to the field without modification of boundary conditions (e.g., Acar and
258
Haider, 1990). The extrapolation of such results should be interpreted with

caution.
c. Composite Barriers.
The one-dimensional mathematical framework
described above is based on the assumption of uniform barrier properties
throughout the system, that, although difficult to verify in the field, is
consistent with the design of most engineered systems. A notable exception
to this assumption concerns the case of a composite barrier system, in which
different materials are placed side by side in successive units, for the purpose
of enhanced resistance to contaminant transport or improved reliability of
the system. Conceptually, a composite barrier may be represented
schematically as a layered system with each layer characterized by different
properties. Composite systems could include the use of a geosynthetic liner
on either or both sides of a vertical wall, or alternating horizontal layers of
earthen and/or geosynthetic barriers with differing properties. The onedimensional mathematical framework is applicable to the composite system
as long as the differing barrier properties are properly represented. The
solution of the governing equations for a composite system will, in general,
require greater computational effort.
In computing the seepage velocity through a composite system, the
following two modifications must be made to Eq. 10-13: (1) the effective barrier
hydraulic conductivity, KE, is computed as the weighted harmonic mean of
the layer hydraulic conductivities (weighted by layer thickness), and (2) the
seepage velocities through the individual layers are computed using the
respective layer porosity.
Some composite systems may be represented as an equivalent
homogeneous system, particularly if the contaminant flux or concentration is
desired at a single point in space (e.g., at the barrier exit). If the porosities,
diffusion coefficients, and sorption properties of the individual layers are
similar, then an equivalent one-dimensional barrier may be represented using
the computed effective velocity. Similarly, if advection is negligible and layer
porosities are similar, an effective diffusion coefficient may be computed using
a weighted harmonic mean of the layer values. When both advection and
diffusion are significant for a composite system, specification of an
equivalent barrier may not be feasible, and more sophisticated solution
techniques may be required.
10.1.3
Solutions
Since the governing processes and boundary conditions applicable to a barrier

system will be site specific, some published solutions to the ADRE may be
applicable. Based on the mathematical models presented and site-specific
field conditions, a useful solution to the one-dimensional ADRE for application
to contaminant transport through containment barriers should incorporate
the following features:
10.1 BACKGROUND
259
diffusive transport;
advective transport in either direction;
first-order decay; including, where appropriate, parent-daughter
chains;
linear or nonlinear equilibrium sorption;
nonequilibrium sorption;
a constant concentration, time-varying concentration, or finite mass
entrance condition;
a semi-infinite or flushing exit condition;
spatial variation in parameters (i.e., multiple layers); and
temporal variation in parameters (e.g., variation in hydraulic
conductivity due to damage accumulation, consolidation of the barrier
matrix, and/or accumulation of biomass, etc).
The extent to which the above features need to be represented in the ADRE
dictates the solution technique and computational effort required. In general,
solutions may be divided into three categories: analytical, semi-analytical,
and numerical. The distinction between analytical and semianalytical is
somewhat arbitrary, as some semi-analytical techniques (e.g., truncation of
an infinite series) may be readily performed by hand calculation; while
implementation of some closed-form solutions is sufficiently tedious that
computers are routinely used. A thorough review of published solutions is
beyond the scope of this report; however, a brief summary is provided with
an emphasis on solutions that incorporate as many subsets of the above
features as possible.
a. Analytical and Semi-analytical Solutions. Closed-form solutions or
solutions requiring numerical integration, root finding, or approximation of
an infinite series are included in this category. A large collection of solutions
related to diffusion phenomena are presented by Carslaw and Jaeger (1959)
and Crank (1975), including some that incorporate spatially and temporally
varying diffusion parameters. Solutions including both advection and
diffusion, first-order decay, and a constant concentration entrance boundary
were summarized by Bear (1972) for the semi-infinite exit condition, and by
Owen (1925) for a zero-concentration perfect-flushing exit. The latter, which
is included in Carslaw and Jaeger (1959), is particularly appropriate for low
hydraulic conductivity barriers and is reproduced in an appendix to this
section, along with the appropriate form of Eq. 10-1 (the flux equation) for
this solution.
Most published analytical solutions that include both advective and
diffusive transport assume spatial and temporal uniformity of parameters.
Linear equilibrium sorption may be accommodated through appropriate
modification of parameters in the concentration-based solution utilizing a
retardation factor. However, the calculation of contaminant flux using Eq.
10-1 is based on the nonretarded velocity and diffusion coefficient. An inward
260
hydraulic gradient may be represented by specifying a negative velocity.

Semi-analytical solutions incorporating a constant concentration entrance
condition, a semi-infinite exit condition, and nonequilibrium sorption
according to the linear, first-order model (Eq. 10-10) are described by van
Genuchten and Wieranga (1976). These and other concentration-based
solutions have been implemented in a useful set of computer codes that include
parameter estimation routines (Parker and van Genuchten, 1984).
Calculation of contaminant flux from concentration-based solutions may
be accomplished by differentiation and algebraic manipulation according to
Eq. 10-1; flux-based extensions for some of the solutions listed here are
discussed by Rabideau et al. (1996).
b. Laplace-transformed Solutions.
A powerful semi-analytical solution
technique incorporates the numerical inversion of a Laplace-transformed
solution, as described by Rowe and Booker (1985) and implemented in the
POLLUTE code (Rowe and Booker, 1985). This approach has several
advantages, including: (1) ability to incorporate both the finite mass entrance
and the flushing exit boundary conditions; (2) ability to represent composite
barrier systems; and (3) generally superior computational efficiency compared
to numerical models. A drawback to this solution approach is its complexity;
however, the POLLUTE code is available commercially, and although designed
for application to horizontal barrier systems, it may also be applied to vertical
barrier systems after reinterpretation of parameter definitions. Extension of
the finite layer approach to incorporate nonequilibrium sorption is discussed
by Rabideau and Khandelwal (1996).
c. Numerical Solutions. A detailed discussion of numerical solutions to the
ADRE is beyond the scope of this report. In general, numerical solutions
provide greater flexibility in the specification of boundary conditions and
permit the incorporation of such processes as nonlinear/nonequilibrium
reactions, as well as spatial variability of parameters within the barrier. The
merits of various numerical schemes for advective-dispersive transport are
discussed by Zheng and Bennett (1995). One dimensional solutions to the
ADRE are generally accomplished by traditional Eulerian methods (e.g., finite
difference or finite element). Common numerical difficulties related to sharp
fronts are less likely to be significant in the analysis of low hydraulic
conductivity barriers, due to the reduced role of advective transport.
The inclusion of additional nonlinear and/or nonequilibrium reaction
terms in a numerical solution of the ADRE can be addressed through a variety
of strategies; in particular, split-operator methods (e.g., Kaluarachchi and
Morshed, 1995) have seen considerable application in simulating reactive
transport in porous media. However, the use of split-operator techniques
requires careful selection of the time step. While several recent studies have
addressed the selection of time steps under advection dominated conditions
(e.g., Miller and Rabideau, 1993; Morshed and Kaluarachchi, 1995), there are
261
few, if any, published studies of this method applied to transport through

low hydraulic conductivity barriers. Therefore, additional numerical
experimentation is essential to determine appropriate temporal and spatial
discretization for application of numerical models to reactive barrier systems.
10.2
STATE OF PRACTICE
10.2.1
Contaminant Transport in Low Hydraulic Conductivity Barriers
a. Field-scale Models. Incorporation of a containment system into a fieldscale transport model requires either (1) appropriate spatial mesh refinement
and the ability to represent regions with differing chemical and hydraulic
properties, or (2) treatment of the contained contaminant zone as a timedependent source region within the larger model domain. The initial step in
analyzing the field problem is to simulate the groundwater flow system with
special attention to the influence of the containment system. This is
accomplished utilizing a groundwater flow model, for which data are more
readily available and for which there may be greater confidence in computed
results relative to transport models. Considerable insight can be developed
from flow modeling prior to solute transport modeling.
Several commercial three-dimensional transport codes are currently used
by practitioners, including the Princeton Transport Code (Babu and Pinder,
1984) and MT3D (Zheng, 1992). Incorporation of low hydraulic conductivity
barriers in such models is straightforward from the standpoint of the
groundwater flow system. A detailed treatment of the diffusion-dominated
transport region associated with a containment barrier itself, however, would
normally require use of a very fine discretization in the vicinity of barrier.
The commonly employed multidimensional transport codes also typically
rely upon a simplified representation of reaction processes (generally limited
to first order decay and equilibrium sorption) to reduce the computational
burden.
A considerable amount of recent research has addressed the development
of numerical models for simulating transport at the field scale in conjunction
with more sophisticated descriptions of reaction processes . While models of
this nature are currently not routinely used, there are models available that
incorporate three-dimensional heterogeneity and various reaction processes,
often in conjunction with high-performance computers. For research
applications or for projects where considerable resources are available to
support simulation studies, such a detailed modeling approach is likely to
provide the most satisfying results, assuming that adequate data are available
for parameterization.
An alternative approach for field-scale modeling would be to treat the
contained contaminant zone as a set of source nodes within the larger
discretized domain. While such an implementation is possible with most
262
transport codes, it may not be straightforward to represent the dynamic

interaction between the barrier and the surrounding aquifer. A reasonable
approach to this problem would be to couple a separate model for the barrier
system with the field-scale model. The output from the barrier sub-model
would provide the source term for the field-scale model, while the field-scale
model would provide updated boundary condition information to the barrier
sub-model. The dynamic coupling of the two models could be accomplished
using either an iterative approach or a sequential solution over small time
steps. The barrier submodel could be represented as a single or multiple onedimensional model(s) of sufficient complexity to handle time-varying
boundary conditions and the appropriate chemistry. While this coupled
approach is conceptually straightforward, its implementation has not been
reported.
b. One-dimensional Models. Except for landfill liners, published applications
of contaminant transport models to low hydraulic conductivity barrier design
and analysis has been limited. With appropriate treatment of boundary
conditions, however, the one-dimensional framework may be applied to
vertical barriers and/or emplaced bottom barriers. Closed-form and semianalytical solutions for many of the applicable models are available and may
be implemented in spreadsheets or simple programs (Shackelford, 1990).
Commercial codes are also available for one-dimensional transport models
(see review by van der Heijde and Elnawawy, 1993). The International
Groundwater Modeling Center, Colorado School of Mines, serves as a
clearinghouse for a variety of groundwater models.
The commercial POLLUTE code implements several solutions for the onedimensional conceptual models discussed earlier (Rowe and Booker, 1985).
Although structured for application to liner systems, POLLUTE may also be
applied to the analysis of low hydraulic conductivity walls (with appropriate
assumptions). Useful features of POLLUTE include: the ability to represent
composite barriers; the inclusion of a parameter estimation algorithm that
may be applied to the analysis of experimental data; and the recent addition
of graphical user interface. Published applications of the POLLUTE code are
numerous (see summary by Rowe et al., 1995). Further examples of the
conceptual implementation of one-dimensional models and application to
barrier systems are provided by Shackelford (1989) and Rabideau et al. (1996).
10.2.2
Groundwater Flow Modeling for Low Hydraulic Conductivity

Barriers
Under some circumstances, insight into the potential effect of a barrier

containment system on a proposed remedial action may be obtained by
considering simply the influence of the barrier system on groundwater flow
patterns. Such modeling is implemented using a variety of available numerical
and semi-analytical groundwater flow models. For example, the widely used
263
MODFLOW code (McDonald and Harbaugh, 1988) is adaptable to the

inclusion of low hydraulic conductivity barriers. MODFLOW has recently
been expanded to include a module specifically designed to incorporate
vertical barriers that are much thinner than the model discretization (Hsieh
and Freckleton, 1993). For scenarios where NAPL contamination is present, a
multiphase flow model may be required.
The following case studies illustrate the application of models to evaluate
the impacts of low hydraulic conductivity barriers on remedial activities and
the regional groundwater flow system.
a. Gilson Road, New Hampshire. The Gilson Road site, located in Nashua
New Hampshire, was the site for the first cooperative agreement signed under
the CERCLA program. As part of the remediation system a soil/bentonite
cutoff wall was used to isolate a contaminant plume from the surrounding
aquifer and nearby surface water. The cutoff wall extended approximately
4000 feet in length and up to 110 feet in depth. The Gilson Road project is
significant in that it represents what is believed to be the first slurry wall
designed for hazardous waste containment, and the project was therefore
subjected to extensive quality assessment (Ayres et al., 1983; Schulze et al.,
1984; Barvenik et al., 1985; Barvenik et al., 1986; Barvenik and Ayres, 1987).
Following the installation of the containment system in 1982, a series of
post-construction verification tests was conducted, including a pumping test
designed to determine the transmissivity of the fractured bedrock that formed
the bottom of the contained area. An extraction well was located inside the
contained area adjacent to the slurry wall and operated for 10 days.
Drawdowns were measured at over 70 monitoring wells, and the data were
used to calibrate a three-dimensional model of groundwater flow for the site.
The calibrated model (Trescott, 1975) was then used to analyze the performance
of the containment system, including sensitivity analyses to establish a range
for the bulk average hydraulic conductivity of the slurry wall. Although the
design hydraulic conductivity of the wall was 10-7 cm/s, the results of the
numerical simulations suggested that the effective hydraulic conductivity of
the wall could as high as 10-5 cm/s before significant change would be noted
in the predicted hydraulic head distribution. The value of 10-5 cm/s was
therefore adopted as an upper bound for the hydraulic conductivity of the
installed cutoff wall, and used to obtain conservative estimates of leakage
across the wall.
b. Lipari Landfill, New Jersey. At the Lipari landfill Superfund site, an areal
groundwater flow model (Trescott et al., 1986) was applied to evaluate various
remedial actions (Andersen et al., 1984; Mercer et al., 1987). The numerical
model was used to simulate existing flow conditions at the site and provide
initial conditions for a series of sensitivity analyses, including simulations
designed to evaluate: (1) a slurry wall, (2) location of a drain system, (3)
drain depth, and (4) a clay cap.
264
The Lipari Landfill is a 6-acre former gravel pit and industrial chemical
dump located in Gloucester County near Pitman, New Jersey. On September
1, 1983, this site was placed at the top of the Superfund Priority Cleanup List
(Russakoff, 1983). Model calibration for the Lipari site consisted of matching
observed water-level data and surface water discharge data, and the numerical
model was determined to be adequately calibrated so that it could be used
for conceptual design purposes. After calibration, sensitivity simulations
were performed to evaluate various proposed remedial actions. Numerous
simulations were performed for various configurations of a slurry wall, drain
system, and clay cap. In these simulations, the finite difference block
representing the wall was assumed to have a hydraulic conductivity six orders
of magnitude smaller than that of the aquifer. The drain was approximated
by treating the finite-difference block containing the drain as a constant-head
node. Full and partial caps were simulated by setting the recharge to zero in
the finite-difference blocks representing the cap.
The results from model simulations were used to assess the predicted
performance of various components of the remedial system. For example,
the type of cap (full versus partial) was shown to have little effect on the
predicted discharge to the drain system. The findings from the study were
considered along with economic and engineering factors in the selection of
the final remedial system.
c. Love Canal, New York. For the Love Canal site, a vertical cross section,
variably-saturated flow model (Trescott et al., 1976) was applied to evaluate
various proposed corrective actions (Cohen and Mercer, 1984; Mercer et al.,
1987; Cohen et al., 1987). Approximately 22,000 tons of chemical wastes were
buried in a 3-city block long, 18-m (60-ft) wide excavation (known as Love
Canal) in Niagara Falls, New York between 1942 and 1953. In the mid-1970s,
chemical seepage and odors were observed in the basements of many homes
adjacent to the site. Remedial work conducted in 1978 and 1979 included the
construction of a French drain completely around the site and a clay cover
over the landfill. Following declaration of a state of emergency at the site on
May 21, 1980 by President Carter, the U.S. EPA undertook a major study of
the Love Canal environment (U. S. EPA, 1982). Based on the findings by EPA
and others, the New York State Department of Environmental Conservation
(NYSDEC) directed that further monitoring be undertaken and corrective
actions be considered at the Love Canal hazardous waste site.
Additional proposed corrective actions included encapsulation of the
site with a concrete cut-off wall and a synthetic cover. The impact of these
proposed additional remedial actions on groundwater flow at the site were
evaluated for the following cases: (1) no additional corrective actions, (2)
synthetic cover only, (3) synthetic cover and cut-off wall, and (4) addition of a
second French drain. The potential effectiveness of these measures were
evaluated with regard to: (1) dewatering the shallow flow system, (2) French
drain flux, and (3) reversal of vertical hydraulic gradients.
265
Calibration of the groundwater flow model was achieved in a series of

steps, including: (1) simulating the general steady-state water-table pattern
prior to installation of a French drain and clay cap, (2) matching the drain
flux trend observed between August 1979 and June 1982, and (3) matching
water levels measured in a series of piezometers in November 1982.
Effects of the proposed synthetic cover and concrete cut-off wall were
evaluated by comparing 50-year simulations initiated at 1365 days (after drain
and clay cap installation) in which: (1) no additional corrective actions were
taken, (2) only a synthetic cover was included, and (3) both a cover and cutoff wall were included. Several additional simulations were conducted,
including: (1) a sensitivity analysis of the shallow system hydraulic
conductivity; (2) a simulation in which a second French drain was installed
beneath the synthetic cover just inside the concrete wall; and (3) a simulation
in which a second French drain was installed at the location of and in place of
the concrete wall.
The synthetic cover was treated in the model as an impermeable barrier
that produced a net recharge of zero. It was also assumed that runoff from the
membrane was properly managed, causing no extra net recharge to the area
beyond the cover. The concrete cut-off wall was assigned a hydraulic
conductivity of 3 x 10-7 m/d (3.47 x 10-10 cm/s) for the entire 50-year simulation
period. For the purpose of analyzing the hydrogeologic effects of the proposed
corrective actions, it was assumed that the concrete cut-off wall and the
synthetic cover achieved and maintained their design specifications. Given
these assumptions, numerical simulations of existing and proposed corrective
actions at Love Canal were used to assess various aspects of system
performance, including the flux to the drain system under various conditions,
the time to attain the maximum flux to the drain, and the amount of flow
from various locations within the site that contributed to the drain discharge.
Construction of a synthetic membrane cover has been completed;
however, on July 28, 1983, the New York State Department of Environmental
Conservation announced that the cut-off wall would not be constructed as
earlier planned (NYSDEC, 1983), citing observations from the modeling
studies which suggested that the wall would have little effect on the volume
of leachate collected and would impede collection of contaminants from
outside the wall.
d. S-Area, New York. At the Occidental Chemical S-Area site, a multiphase
flow model was applied to evaluate remedial actions (Faust, 1984; A.D. Little,
Inc., 1983; Cohen et al., 1987), including flow reversal of chlorinated
hydrocarbons present as a dense non-aqueous phase liquid (DNAPL). The SArea Landfill is located on the southeast corner of Occidental Chemical
Corporations Buffalo Avenue Plant in Niagara Falls, New York.
Approximately 63,100 tons of chemical wastes were deposited at the site.
A major concern at the landfill are discontinuities in an underlying confining
bed that allowed DNAPL to enter a bedrock aquifer. Proposed actions
266
included containment of liquids present in the landfill utilizing an integrated

system of barrier walls, plugs, drains, and a cap designed to prevent off-site
migration. The multiphase flow model was used to establish the magnitude
of an upward hydraulic gradient needed to prevent downward migration of
DNAPL.
10.2.3 Permeable Reactive Walls
For permeable reactive walls, modeling is needed to evaluate: (1) the
degree of treatment attainable within the wall, and (2) the degree to which
barriers are needed and their effectiveness in directing contaminated water
through the treatment zone. Conceptual models for permeable reactive
barriers that assume horizontal flow in the vicinity of the treatment zone may
be modeled using the one-dimensional framework discussed previously.
However, funnel and gate systems will concentrate and increase flow velocities
in the treatment zone (Starr and Cherry, 1994). For most systems, the governing
conceptual model is a plug-flow reactor with dispersion. Because advective
transport is expected to dominate, the predicted effluent concentration
distribution will be less sensitive to the form of the specified boundary
conditions and reasonable accuracy should be obtainable by assuming
idealized conditions that lead to closed-form solutions. Accurate
representation of the reaction term is essential, since the objective of the model
will be to determine the dimensions of a treatment wall necessary to achieve
the desired reduction in contaminant levels (via the in situ reaction).
As the development of permeable reactive walls is a relatively new
concept, published modeling studies of actual installations are not available.
A likely application, however, is the consideration of a treatment wall in which
organic contaminants undergo first-order decay in the presence of a reactive
agent. Since the design time frame of such a system is long, a steady-state
solution of the ADRE is suitable for design purposes. Under the assumption
of negligible sorption, the solution for the effluent concentration, C, relative
to the entrance concentration, Co, is given by (e.g., Tchobanoglous and
Schroeder, 1987)
2aPe
C
=
2
C o 1 + a exp 0.5aPe 1 + a 2 exp 0.5aPe
a= 1 + 4 t P e
Pe =
(10-23)
vL
D
in which Pe denotes the Peclet number as defined, L is the length of the

treatment zone, D is the dispersion coefficient for the permeable wall, and t
is the reaction decay rate constant. When the velocity, dispersion coefficient,
267
and decay rate constant for the system are known, Eq. 10-23 may be used to
adjust the length of the treatment zone, L, to achieve the desired reduction in
contaminant concentration. Similar equations may be developed for a variety
of reactive systems, such as sparge gates and sorption barriers. More complex
reaction mechanisms may require the use of a numerical model and/or timedependent analyses.
Regional flow system models may be used to verify that the contaminated
groundwater passes through the treatment zone and to determine the
advective velocity at the wall. Starr and Cherry (1994) demonstrate the use
of a groundwater flow model to analyze the performance of a funnel-andgate permeable reactive wall system. Because regional hydraulic gradients
often exhibit seasonal variation in both magnitude and direction, a simulation
model can provide useful insight into the trade-offs between system design
parameters, e.g., funnel length, and the confidence in the ability of the system
to maintain the desired routing of the contaminant plume.
An important concern, however, relates to site characterization and
assumption of homogeneity in a groundwater flow model when applied to
the analysis of a permeable reactive wall. Because advective transport
dominates these systems, the influence of heterogeneity within the
surrounding aquifer may be significant, particularly when the reactive wall
is not keyed into an underlying confining layer. Teutsch and Schad (1995)
describe an example where incorporation of observed heterogeneity into a
three-dimensional flow model of a permeable wall suggested that bypassing
of the barrier due to vertical flow could potentially occur. Such a condition
could also occur if temporal changes in the pore system of the reactive wall
(e.g., due to biomass growth and/or chemical precipitation) cause a reduction
in hydraulic conductivity. For these reasons, careful consideration should be
given to the use of more sophisticated numerical models, incorporating three
dimensional resolution, density driven flows, and variations in hydraulic
conductivity due to heterogeneities, for the analysis of permeable reactive
walls.
10.2.4 Parameter Estimation
Application of mathematical models to barrier systems requires estimation
of the numerous parameters contained in the governing equations described
above. Parameter estimation for transport models can be accomplished by a
number of strategies including: (1) empirical or theoretical correlation with
other known system properties; (2) use of literature values; (3) contaminant/
media specific laboratory experimentation; and (4) calibration using field data.
For engineered barrier systems, calibration based on field data will generally
not be an option and some combination of the other approaches will be
required.
Values for the barrier parameters (e.g., barrier thickness, porosity, and
density) can be specified as part of the design process and are therefore
268
determined. When advection is included in the model, estimates of the

hydraulic conductivity and piezometric head gradient are also required.
Detailed discussions of hydraulic conductivity estimation are available
elsewhere (e.g., Rumer and Ryan, 1995). The pre-construction hydraulic
gradient can be determined through field measurement and may be subject
to control by the designer through the specification of pumping rate(s).
Groundwater flow modeling may be needed to predict hydraulic gradients
after installation of the containment barrier system. The focus of the remaining
discussion is on the estimation of diffusion coefficients, decay rate constants,
and reaction parameters.
a. Liquid Diffusivity.
As indicated in Eq. 10-5, the maximum expected
value for a contaminant diffusion coefficient is the free liquid solution
diffusivity. Literature values are available for many contaminants of interest
(e.g., Cussler, 1988; Shackelford, 1991; Shackelford and Daniel, 1991). For
other compounds, a variety of correlations are available relating diffusivities
to molecular properties. For organic compounds, a review of applicable
correlations is provided by Lyman et al. (1990). An example of a commonly
applied relationship is that proposed by Wilke-Chang (1955),
Dl =
8
1 /2
(7.4x10 )( W M W ) T
W V 0.6
B
(10-24)
where W is the solution association constant, MW is the molecular weight of

water, W is the dynamic molecular viscosity of water, T is absolute
temperature, and VB is the molar volume of the solute at boiling point.
For inorganic compounds, estimation of the liquid solution diffusivity is
less straightforward, due to a variety of electrochemical factors as discussed
by Shackelford et al. (1989). A reasonable starting point is the Stokes-Einstein
equation (Reid et al., 1977).
Dl =
RT
6 rB
(10-25)
where R is the ideal gas constant and rB is the solute molecular radius.
It is noted, however, that the Stokes-Einstein assumptions of a single
dissolved species and infinite dilution are rarely appropriate for estimating
effective diffusivities of inorganic species in subsurface systems.
b. Tortuosity. As noted in Eq. 10-6, the liquid solution diffusion coefficients
for both organic and inorganic compounds are likely to be reduced by effects
related to the structure of the porous media. Various terms have been
introduced in the literature to describe these effects; however, in this section,
269
porous media influences will be grouped into a single tortuosity parameter,

greater than unity, that is used to scale the liquid diffusivity as expressed in
Eq. 10-6. In doing this, tortuosity is understood to represent the combined
effects of the tortuous flow pathways through the pore system geometry, the
steric hindrance or anion exclusion effects that exclude solute molecules from
the smaller pores, and the increases in fluid viscosity that may occur near
surfaces. The tortuosity parameter discussed here does not include the effect
of the reduced macroscopic cross-sectional area relevant to porous media,
since this effect has been accounted for explicitly in Eq. 10-1 and Eq. 10-2.
The value of the tortuosity factor is difficult to determine directly and is
usually inferred by comparison between a measured porous media diffusion
coefficient for a conservative solute with that obtained in free liquid solution.
A number of correlations have been developed that relate the tortuosity, , to
the media porosity (e.g., Millington, 1959). These correlations can be expressed
in the form
= n
(10-26)
where is an empirically-determined exponent, typically between -2 and -1/3,

although there are few data for low hydraulic conductivity materials.
Application of a correlation similar to Eq. 10-26 generally is based on the
assumption that the reduction of the effective diffusion rate is primarily related
to the geometric tortuosity of the pore system geometry, rather than to various
constrictions that might occur as the size of the solute molecule nears the
average dimension of the pore system voids (see Mott and Weber, 1991a, for
a discussion of these effects). In experiments conducted with a standard soil/
bentonite slurry wall mixture, Mott and Weber (1991a) found that an exponent
of -1/3 provided good agreement with experimental data, and attributed the
majority of the tortuosity effect to the extended diffusion path associated with
the pore structure. Theoretical considerations suggest that the applicability
of Eq. 10-25 would likely produce a range of values for the exponent, ,
considering the wide range of barrier material types.
c. Experimental Determination of Diffusion Coefficients. The most
straightforward means of estimating effective diffusion parameters is through
laboratory experimentation. Shackelford (1991) and Rowe et al. (1995) have
provided reviews of relevant experimental techniques. Laboratory
measurements can provide an estimate of an effective diffusion coefficient
that incorporates the influence of media tortuosity and chemical reactions. If
experiments are performed with a conservative solute, the observed effective
diffusion coefficient may be compared with the theoretical free liquid solution
diffusivity to provide an estimate of the media tortuosity. This information
may then be combined with independent experiments (e.g., sorption
isotherms) to yield a complete set of parameters for reactive contaminants.
The selection of the appropriate experimental technique is dependent
270
upon several factors including: (1) nature of the contaminants; (2) availability
of skilled laboratory personnel; (3) ability to obtain independent estimates of
sorption parameters; and (4) time frame available for measurement. For
strongly sorbing contaminants, it may not be feasible to utilize methods that
call for measurement of spatial or temporal breakthrough curves, because of
the long time frames required (e.g., see Shackelford and Redmond, 1995).
Also, volatile and biodegradable contaminants may require that special
attention be given to experimental procedures and data analysis.
d. Sorption Parameters. Estimation of sorption equilibrium constants is
usually accomplished through independent batch isotherm experiments.
Much literature exists concerning the measurement of sorption isotherms,
and discussion of this topic is therefore omitted here (e.g., U.S. EPA, 1991).
Several correlations have been developed for hydrophobic organic compounds
that relate k d to the soil organic carbon fraction and the contaminant
hydrophobicity, as expressed by the aqueous solubility or octanol-water
partition coefficient (e.g., Fetter, 1994). Most published correlations for
sorption equilibrium parameters are applicable only for dilute solutions of
hydrophobic organics. Reliable methods for a priori estimation of sorption
parameters for other organics and inorganics are not available, and laboratory
experimentation is recommended.
Nonequilibrium sorption may be attributed to mechanisms such as
chemical kinetics, physical diffusion, or both. Most model descriptions of
nonequilibrium sorption are regarded as quasi-empirical, with the model
coefficients determined by calibration using laboratory or field data. Examples
of recent experimental studies can be found in Harmon and Roberts (1994)
and Brusseau et al. (1989). Limited efforts have been made to develop and
apply correlations for determining sorption rate coefficients (e.g., Brusseau
and Rao, 1989a; Mott and Weber, 1992). Such correlations may not be widely
adopted, due in part to the diversity of mechanisms influencing sorption
rate behavior. If barrier performance is expected to be influenced by
nonequilibrium sorption, laboratory kinetic experiments are recommended.
e. First-order Decay.
In the ADRE, contaminant disappearance through
first-order decay is generally understood to include the combined effects of a
number of reaction processes, including biotransformation, hydrolysis, and
radioactive decay. The most straightforward approach to the estimate of
first-order decay constants is through batch experiments performed under
conditions as similar as possible to those likely to be encountered in the field.
A drawback to this approach is that for compounds that decay very slowly,
the duration of experiments will be long and strict controls will be required
to obtain accurate results.
An important class of reactions likely to influence aqueous solutions of
organic compounds is hydrolysis, which is generally represented as a firstorder process. The nature and analysis of hydrolysis reactions, including
10.3 ASSESSMENT
271
correlations for estimating rate coefficients, are discussed by Lyman et al.

(1990). Published hydrolysis rate measurements are available for many
compounds of interest (Mabey and Mill, 1978). A summary of measured
radioactive decay rates for radionuclides is provided by Moody (1982).
Observed rates for some other abiotic reaction processes that influence organic
compounds are presented by Vogel et al. (1987). Few data are available for
transport experiments conducted in low hydraulic conductivity porous media.
Consequently, potential mass transfer effects on rate coefficients are not
accounted for in the available measured data, and the use of published
correlations and rate constants should be viewed as a last resort in situations
where no media-specific experimental data are available. Furthermore, for
radionuclides, inclusion of decay reactions without consideration of daughterproduct formation may produce unrealistic results.
10.3
ASSESSMENT
10.3.1
Field-scale Models
The term postaudit has been used to describe studies in which the validity
of model predictions is tested by comparing model output with field data
collected long after the model simulations were first performed. There have
been few published postaudit studies validating groundwater flow or
contaminant transport models. In a review of five published postaudits,
Anderson and Woessner (1992) noted that, in all cases, the models gave poor
predictions. The primary explanations offered for these poor predictions were
(1) an inadequate representation of the subsurface stratigraphy resulting in
an inadequate representation of the hydraulic conductivity (e.g., too few layers,
incorrect boundary conditions), and (2) an inadequate representation of timevarying operating conditions (e.g., pumping cycles, recharge, etc.).
Since engineered subsurface barrier systems are generally designed on
the basis of parameter conformance to a regulatory standard (e.g., K < 10-7
cm/s), it is probable that the reduced role of advection will result in minimal
errors due to misjudgments in regional groundwater flow boundaries and
stresses. On the other hand, the engineered component of these systems
introduces added uncertainty due to the potential for irregularities introduced
during the construction process, and the possible time-evolution of barrier
properties. The latter are considered in a subsequent section.
The rationale supporting the use of sophisticated numerical models in
the analysis and design of remedial systems is based on their ability to produce
detailed and accurate predictions, assuming that the necessary parameters
and contaminant distributions are known. The merits of this rationale depend
on the objectives of the modeling exercise. As an example, consider the use
of site-specific numerical models to predict cleanup times for pump-and-treat
systems. Accurate predictions of cleanup times require: (1) delineation of
272
the contaminant distribution with respect to the phases present and spatial
variation; (2) knowledge of the spatial structure of hydraulic conductivity;
and (3) knowledge of rate coefficients for reactions (e.g., dissolution,
desorption, etc.). The determination of these factors is often difficult and
may not be feasible for some applications. For this reason, while solute
transport models may be appropriate for some remedial applications (e.g.,
delineation of extraction well capture zones), predictions for cleanup times
and long-term system performance are considered to be highly uncertain.
Modeling of containment barrier system performance is subject to
considerations that are both similar to and distinct from those relevant to
pump-and-treat. For low hydraulic conductivity barriers, if the onedimensional framework is adopted, concerns related to characterization of
contaminant distribution and aquifer heterogeneity become less important.
The integrity of the constructed barrier is a prime concern, however, as is the
role of reactions that influence transport within the barrier. Similarly, for
permeable barrier walls, the initial distribution of contaminant within the
plume is less important than the accurate representation of system hydraulics
and the reactions occurring within the treatment zone, as well as any longterm effects such as clogging and/or reductions in reaction rates. Therefore,
assessment of model output needs to be referenced to the site-specific situation
and the objectives of the modeling exercise.
10.3.2
One-dimensional Models for Low Hydraulic Conductivity Barriers
Ultimately, application of the one-dimensional modeling should be decided

on the basis of whether model output can give an accurate prediction of
contaminant transport under conditions consistent with those likely to found
in the field. Because the relevant field data generally are not available, the
decision to apply a one-dimensional model is likely to be based on (1) the
ability of models to describe the results of well-controlled laboratory studies,
(2) the consistency of model predictions with field observations of transport
in clay landfill liners, (3) insights into the behavior of one-dimensional models
obtained through sensitivity analysis, and (4) a critical assessment of the
underlying assumptions built into the modeling framework.
a. Agreement with Laboratory Studies. Although there is a large body of
literature describing laboratory studies of advection-dominated transport in
porous media, few studies have been conducted using low hydraulic
conductivity materials. Shackelford (1991) and Rowe et al. (1995) provide
reviews of experimental results related to the measurement of diffusion
coefficients in various soils. However, experiments of this nature do not
provide a good basis for evaluating the ability of models to accurately predict
contaminant transport under diffusion-dominated conditions. There have
been few studies designed to directly address this issue. The work of Mott
and Weber (1991a, 1991b, 1992) is a notable exception in that independent
10.3 ASSESSMENT
273
quasi-steady-state and transient experiments were used to test the ability of

diffusion models to predict contaminant distributions. While some column
experiments have been performed using barrier materials (e.g., Acar and
Haider, 1990; Bierck and Chen, 1994), few experiments have addressed the
low-flow conditions likely to be encountered in the field [the recent work of
Shackelford and Redmond (1995) is an exception].
Laboratory soil-column experiments are typically conducted using a
subset of the boundary conditions applicable in the field. A more general set
of experiments has been described by Hensley and Schofield (1991), employing
a centrifuge and a physical model designed to mimic the third type boundary
conditions used in the POLLUTE model (finite mass entrance, flushing exit).
Based on their studies and preliminary results of ongoing research sponsored
by DuPont, Inc. (Khandelwahl, 1995, unpublished data), the following
observations are offered:
measured contaminant concentration profiles or breakthrough curves

have generally been consistent with the hypothesis that diffusive
transport dominates under the low flow conditions associated with
barrier materials;
observed effective diffusion coefficients for organic compounds are
generally consistent with theoretical predictions when barrier
materials are not strongly sorbing;
difficulties in predicting breakthrough curves for inorganics have been
noted and attributed to a variety of chemical processes not well
represented by the local equilibrium assumption;
for some barrier materials, particularly those amended with organic
matter to enhance sorption, the retardation factor approach may be
inadequate due to isotherm nonlinearity and possible nonequilibrium
effects; and
good agreement has been noted between the centrifuge experiments
and predictions generated by the POLLUTE code.
b. Field Studies of Landfills. There have been several investigations of

contaminant distributions within compacted clay landfill liners over time (e.g.,
Johnson et al., 1989; studies summarized by Rowe et al., 1995). While these
studies do not carry the weight of a postaudit, they do provide insight into
the applicability of advective-diffusive transport theory for the analysis of
transport in low hydraulic conductivity materials. Based on the findings from
these studies, the following observations are offered:
as expected, the measured contaminant profiles were consistent with

the hypothesis of diffusion-dominated transport;
measured profiles for nonreactive substances could be reasonably well
predicted by mathematical models using values for diffusion
coefficients obtained in the laboratory;
274
the observed permeation of various reactive organic and inorganic

species varied, depending on the compound, but in general did not
exceed the permeation of chloride;
independent predictions of contaminant distributions based on
hypothesized reaction mechanisms were in general unsuccessful; and
no appreciable increases in hydraulic conductivity with time were
noted for the clay liners associated with municipal solid waste (MSW)
landfills; in some cases, a gradual decrease in hydraulic conductivity
was noted.
Although the ability to describe reactive transport was limited for these
field studies, the systems under consideration present a particularly difficult
challenge due to the complex chemical interactions between MSW leachate
and compacted clay, as well as uncertain temporal variability in leachate
conditions. Systems that involve fewer contaminants and/or engineered
materials may be more amenable to predictive modeling of reactive transport.
In general, however, the results of these studies suggest that for some systems,
modeling of nonreactive solute transport may be sufficiently conservative,
making more detailed simulations of reactive contaminant transport
unnecessary.
c. Sensitivity Analysis.
While the one-dimensional framework may be
suitable for design and analysis of many barrier systems, the required
computations may not be trivial in some cases. In the application of onedimensional models, the following factors must be considered:
the selection of processes to be included (e.g., advection, decay,

sorption, etc.);
the time frame of concern (i.e., steady-state versus transient, design
service lifetime);
the governing boundary conditions; and
parameter estimation.
Each of the above factors may be considered at various levels of complexity.

A decision regarding the complexity of model formulation should be
referenced to the modeling objective and/or performance criteria associated
with the particular barrier application. A common approach to evaluating
the relative importance of various transport processes is to perform sensitivity
analyses using a model that incorporates all of the relevant mechanisms. In
the discussion that follows, unless indicated otherwise, the results of computer
simulations are taken from recent work sponsored by DuPont, Inc. (Rabideau
et al., 1996; Rabideau and Khandelwal, 1966). Modeling results are expressed
in terms of dimensionless groupings referenced to a characteristic diffusion
time T = L2/D which is on the order of 30 to 300 years for a typical slurry
wall (L = 1 meter, K = 10-7 cm/s, D = 10-6 to 10-5 cm2/s). Note that, due to the
10.3 ASSESSMENT
275
negligible role played by mechanical dispersion, for these simulations, the

dispersion coefficient (D) is equivalent to the effective molecular diffusion
coefficient ( D*).
A conservative estimate of the steady-state contaminant flux, J, through a
barrier wall may be obtained by neglecting advection and reaction and
assuming a flushing boundary condition; resulting in the following
expression
J=
nDCo
L
(10-27)
where D is the effective diffusion coefficient as defined in Eq. 10-6, Co is the

contaminant concentration at the barrier entrance, and L is the barrier
thickness.
Eq. 10-27 ignores the early-time transient behavior of the system, and
does not consider advection, sorption and first-order decay. Incorporation of
these processes in a time-dependent solution, while retaining the conservative
boundary conditions, results in the more complicated expression for
instantaneous flux (see appendix to this section). The model equations
described in the appendix to this section provide a more realistic representation
of system dynamics and, although more cumbersome than many common
solutions of the ADRE, may be evaluated using a spreadsheet or simple
computer program. An example of an application of a barrier design
procedure utilizing a transient solution to the ADRE is given by Shackelford
(1990).
Normalized Flux = J(t)*L/(nCoD)
1.0
Exit boundary
Perfect flushing
Semi infinite
0.8
0.6
0.4
0.2
0.2
0.4
0.6
Normalized Time = t*D/(L2 )
Figure 10-2 Pure diffusive flux predictions for different exit boundary
conditions
276
Boundary Conditions.
When advective transport in the direction of the
chemical gradient is negligible, Eq. 10-27 provides a conservative estimate of
contaminant flux because of the form of the boundary conditions and the
assumption of negligible reactions. The effect of assuming a flushing type
boundary, as opposed to a semi-infinite boundary, is shown in Fig. 10-2.
Differences are noted in the shape of the flux profiles as well as the peak flux
values. The most significant difference between the two model applications
is that the flushing boundary results in a fairly rapid rise to the steady-state
condition given by Eq. 10-27, attaining 90% of the steady-state value in less
than half the characteristic diffusion time, while the semi-infinite boundary
condition produces a slow rise to a smaller peak, followed by a slow decrease
in the gradient-driven diffusive flux as the contaminant accumulates in the
vicinity of the barrier exit. While some portions of a containment system
may be reasonably well-represented by the semi-infinite condition, it is likely
that the flushing condition is more realistic, and is clearly more conservative.
2.0
Pe=vL/D
+1
+0.5
0
-0.5
-1
1.6
1.2
0.8
0.4
0.2
0.4
0.6
2
Normalized Time = t*D/(L )
Figure 10-3 Flux predictions with varying amounts of advection

Advection. When a positive (outward) hydraulic gradient exists, the steadystate diffusive flux, as given by Eq. 10-27, will no longer provide a conservative
estimate of contaminant flux. Alternatively, if pumping is implemented on the
inside of a contained area causing an inward hydraulic gradient, Eq. 10-27 may
produce excessively conservative estimates of contaminant flux, since advective
transport will be oppositely directed to diffusive transport. The relative
importance of advection may be represented by the macroscopic Peclet number
Pe = vL/D, which may be considered as the ratio of the advective transport to the
diffusive transport. The results of sensitivity analyses are shown in Fig. 10-3,
with a positive Pe corresponding to an outward hydraulic gradient, and a negative
Pe corresponding to an inward gradient. Note that even relatively small changes
10.3 ASSESSMENT
277
in Pe result in significant changes in the predicted steady state flux, due

primarily to the role of advective transport (i.e., hydrodynamic dispersion is
small relative to molecular diffusion). Based on the definition of seepage
velocity expressed in Eq. 10-13, it is seen that Pe is directly proportional to the
hydraulic gradient across the barrier. For the vertical wall parameters used
in the simulations (K = 10-7 cm/s, D = 10-5 cm2/s, n = 0.4, L = 100 cm), Pe/i =
2.5. Thus, the results shown in Fig. 10-3 suggest that advection might be
reasonably neglected under a typical regional gradient of i = +0.05, while a
large induced inward hydraulic gradient i = -1.0 would produce a significant
reduction in the predicted contaminant flux through the barrier. Similar results
have been shown by Gray and Weber (1984).
1.6
1.2
Dimensionless
Decay Rate
Nonreactive
0.1
1
10
100
0.8
0.4
0.2
0.4
0.6
2
Figure 10-4 Flux predictions for various contaminant decay-type reactions

Reactive Transport: First-Order Decay. Since biotic and abiotic reactions
tend to reduce the rate of contaminant transport through a barrier, their neglect
generally produces conservative estimates for contaminant flux, assuming
suitable advection and diffusion parameters have been specified. However,
when a reactive barrier is specifically being considered as a means of reaching
performance criteria, barrier chemistry is important and needs to be included
in the analysis of contaminant flux.
A number of processes may result in the transformation of contaminants,
including radioactive decay, biodegradation, and abiotic reactions. Inclusion of
a first-order decay term in the ADRE is a convenient means of representing these
processes, although often neglected in practice. Because of the long residence
times for contaminants in barrier walls or floors, the effect of decay-type reactions
may be significant, as illustrated in Fig. 10-4. In general, when the value of the
dimensionless rate coefficient L2/D exceeds 1.0 (half life of approximately 20
years for typical slurry wall parameters: D = 10-5 cm2/s, L = 100 cm ), there can
278
be significant reductions compared to steady-state contaminant flux. Decay

processes should be included in models (if they exist) when design
calculations, based on neglecting decay, do not meet performance criteria or
when greater accuracy in model predictions is desired. It can be seen that as
the dimensionless rate constant approaches 100, reactions within the barrier
wall may be sufficient to negate the outward diffusive transport, virtually
eliminating contaminant flux.
1.6
1.2
Rf
1
2
5
10
0.8
0.4
0.2
0.4
0.6
Normalized Time = t*D/(L2)
Figure 10-5 Flux predictions for various contaminant retardation factors

(Pe = 0.1)
Reactive Transport: Sorption. In the absence of reactive decay processes, the
effect of equilibrium (i.e., rapid) sorption in a barrier represented by a flushing
exit condition is to lengthen the time required to reach steady-state. However,
the magnitude of the peak flux may not be significantly reduced, as illustrated
in Fig. 10-5. If reactive decay processes are present, sorption will still increase
the contaminant residence time in the barrier and result in a further reduction
of contaminant flux.
A difficulty encountered in the modeling of reactive transport occurs when
there is nonlinearity in the sorption isotherm (e.g., Brusseau and Rao, 1989b;
Weber et al., 1991). Incorporation of a nonlinear isotherm in Eq. 10-2 generally
necessitates the use of a numerical solution technique. To facilitate the use of
available analytical solutions, the nonlinear isotherm may be linearized
through one of several approaches, as discussed by Brusseau and Rao (1989b).
If the form of the governing nonlinear isotherm is known, a simple strategy is
to equate the linear and nonlinear expressions at some reference concentration,
typically the entrance boundary concentration, Co. The distribution
coefficient, kd, is then computed in terms of the nonlinear isotherm parameters
and the reference concentration. Linearizing an isotherm in this fashion
generally results in conservative predictions of solute transport (i.e., longer
10.3 ASSESSMENT
279
times for contaminant breakthrough) if the governing isotherm expression is

concave in shape (e.g., Freundlich exponent less than 1.0), as is common for
many contaminants. However, the distribution of contaminants within the
barrier generally may not be accurately portrayed when a nonlinear isotherm
is linearized (Shackelford, 1993).
The specification of the appropriate reaction model is particularly
important when organic matter is added to barrier materials for the purpose
of enhancing sorption. In such cases, a sorption isotherm can be determined
from batch experiments and incorporated in the appropriate ADRE. However,
if the sorption process is slow and local equilibrium was not achieved,
modeling predictions will not be conservative. In Fig. 10-6, the results of
sensitivity analyses are shown for a system described by the two-compartment
first-order sorption model (Eq. 10-8). As the sorption rate constant is increased,
the curves move from a shape consistent with negligible sorption to one
consistent with the local equilibrium assumption. Significant deviations from
the equilibrium model are noted when the dimensionless sorption rate
constant L2/D is below approximately 100, a result consistent with similar
analyses conducted for advection-dominated transport (e.g., Jennings and
Kirkner, 1984). These results reveal that the local equilibrium assumption
may be inappropriate for some sorptive barrier materials.
1.6
Dimensionless
Sorption Rate
0.001
0.01
0.1
1
10
100
1.2
0.8
0.4
0.4
0.8
1.2
1.6
2
2.0
Figure 10-6 Flux predictions for various sorption rate constants (semi-infinite
exit boundary, R = 10, Pe = 0.1)
Few data are available to evaluate the validity of the local equilibrium
assumption for low permeability materials. In the studies of Mott and Weber
(1992), good agreement with measured contaminant profiles was noted using
a retarded diffusion model, combined with independent measurements of
sorption and diffusion parameters for a nonamended soil/bentonite (SB)
mixture. However, when flyash was added to the mixture to enhance its sorption
280
capacity, models based on independent isotherm experiments and the

equilibrium assumption significantly underpredicted the penetration of the
contaminant by diffusion into the barrier material. Possible nonequilibrium
effects were postulated as an explanation for the observed deviations.
Bierck and Chang (1994) added granular organic carbon to promote
enhance sorption in soil/bentonite mixes. However, experiments conducted
with a hydrophobic organic contaminant produced less retardation than
would be consistent with equilibrium sorption based on independent isotherm
experiments (Rabideau and Khandelwal, 1996).
The results described above suggest that nonequilibrium effects may
influence the sorption of hydrophobic organic compounds in amended soilbentonite material. For these systems, application of the local equilibrium
assumption may be inappropriate. Development of a model that includes
rate limited sorption is recommended (e.g., Rabideau and Khandelwal, 1996).
d. Parameter Estimation for Barrier Design. For permeable reactive barriers,
it is expected that detailed laboratory and pilot scale studies will be required
to support design and analysis. For low hydraulic conductivity barriers,
however, the degree of effort needed for parameter estimation will depend
upon the modeling objectives and the acceptable level of conservatism.
Laboratory studies for evaluating diffusion and sorption parameters are, by
nature, time-consuming and difficult. Because the observed range for diffusion
coefficients is small for saturated fine-grained soils (10-6 cm2/s < D* < 10-5
cm2/s) relative to other sources of uncertainty, for design applications it may
be adequate to use literature values, particularly if reaction processes are
neglected in the analysis. A conservative value for D* is 10-5 cm2/s, which is
towards the high end of the range for observed free liquid solution diffusivities.
For more accurate results, the methods described in the previous section
should be applied; it is expected that laboratory experiments with the
contaminant/barrier materials of interest should produce the most reliable
results. However, to account for mass transfer effects during contaminant
migration through barrier materials may require long-term testing under low
flow rates (Shackelford and Redmond, 1995).
Although the neglect of first-order decay in transport modeling results in
conservative predictions of contaminant flux, such an approach may not
provide sufficient accuracy, particularly for radionuclides. As discussed
previously, the determination of first-order decay parameters in general is
not straightforward and is best accomplished by laboratory experimentation
under conditions as similar as possible to the expected field conditions.
Published studies of transport and decay in low hydraulic conductivity media
are not readily available, and it is difficult to generalize about the
appropriateness of using published values in transport simulations.
In the absence of decay, the effect of sorption is to retard migration but
not necessarily reduce the peak flux. Thus, for mildly sorbing barrier materials,
a retardation factor of 1.0 is conservative and not unreasonable. If more
10.3 ASSESSMENT
281
accurate results are required, the methods described in the previous section
are applicable, with direct laboratory experiments recommended if resources
permit.
10.3.3
Deviations from Model Assumptions
The contaminant transport models discussed in this report are based on the
applicability of traditional advective-dispersive-reactive transport theory. For
low-permeability materials, however, some deviations from fundamental
assumptions commonly applied in the analysis of porous media may be
expected. Several concerns are mentioned briefly here: the applicability of
Darcys Law, the assumptions of homogeneity and saturated conditions, and
the role of coupled flow processes. Although Darcy s law does cease to be
valid for nonlinear laminar flows or for turbulent flow in porous media, neither
of these flow conditions can be expected to occur in low hydraulic conductivity
fine grained materials (Bear, 1979). In references cited by de Marsily (1986),
observations of a hydraulic conductivity threshold and a nonlinear
relationship between observed flow rates and low hydraulic gradients for
compacted clays are discussed. However, there is little unambiguous or
incontrovertible evidence to negate the validity of Darcys law at low hydraulic
gradients in fine-grained materials (Neuzil, 1986).
The use of a one-dimensional modeling framework implies that the
properties of the barrier are invariant in the other directions. Although this
may be an appropriate modeling framework, the existence of a uniform
hydraulic conductivity throughout a low-permeability barrier has not been
substantiated by field data. Studies of hydraulic conductivity heterogeneity
for other subsurface remedial systems have suggested that even relatively
small variations in hydraulic conductivity may significantly impact the
predicted system performance (Rabideau and Miller, 1994). While numerical
modeling tools exist for incorporating heterogeneity, such tools require greater
user sophistication, compared to one-dimensional models. The ability to
construct a homogeneous low hydraulic conductivity barrier and the effects
of heterogeneity on transport in barriers are currently not well understood
and deserve further study.
Vertical barrier walls are commonly analyzed assuming fully saturated
conditions. For an unconfined aquifer, however, seasonal water table
fluctuations may produce dewatering in a top portion of the wall, with two
implications: (1) wet/dry cycling may affect barrier properties and
performance, and (2) vapor phase diffusion of contaminant may occur.
Because vapor phase diffusion coefficients are generally much higher than
liquid diffusivities, contaminant transport may be greater in the unsaturated
portion of the barrier wall. While some concerns have been noted with regards
to maintaining saturated conditions in earthen walls (Rumer and Ryan, 1995),
the potential impacts of water-table fluctuations on long-term barrier
performance need to be better understood.
282
At the low flow rates expected for low hydraulic conductivity barrier
systems, coupled flow processes may play a significant role, as demonstrated
in recent efforts to develop remedial technologies suitable for fine-grained
soils (Mitchell, 1992). For example, osmotic and electrokinetic transport may
occur in response to both hydraulic and chemical gradients. The extent to
which these processes occur in low-permeability barrier systems is relatively
unknown, and further research is needed to delineate the conditions under
which these processes should be incorporated into the ADRE.
10.3.4
Long-Term Performance
Generally, groundwater flow models for contaminant transport in porous

media assume that media properties remain constant with time. Although
some models do attempt to simulate changes in flow and transport conditions
due to the presence or removal of contaminant (e.g., NAPL dissolution,
biofouling, etc.), certain fundamental media properties, e.g., total porosity
and intrinsic permeability, are usually assumed to be non-varying. Engineered
barriers, however, are of a fundamentally different nature, and it is reasonable
to assume that barrier properties are subject to change over time. This applies
both to low hydraulic conductivity barriers and permeable reactive walls.
Since the design service lifetime for containment systems may range from
decades to centuries, more consideration needs to be given to the possible
evolution of barrier properties during this period.
Engineered systems have been analyzed using risk-based or reliability
approaches for predicting the time-dependent probability of failure for an
engineered system (Freeze et al., 1990). The Weibul probability model has
been commonly used (Inyang, 1995). Certain questions arise when using the
reliability approach, including: (1) What constitutes failure of a barrier?;
(2) How can failure probabilities and the barrier properties causing failure
be quantified? and assuming that these questions can be answered, (3) What
combination of stochastic and deterministic modeling techniques might be
used to produce a realistic scenario for contaminant transport over the lifetime
of a barrier? For example, using Monte Carlo techniques (Zheng and Bennett,
1995), output from repeated simulations could be combined to produce a
probability distribution for contaminant flux that incorporates the failure
information.
One approach to modeling a damaged low hydraulic conductivity
barrier would be to consider it as a fractured, rather than a porous, medium.
A number of strategies have been proposed for modeling transport in fractured
media (Pinder et al., 1993). Simplified equations for describing a regularly
fractured medium have been described by Rowe et al. (1995), and a
multilayered system that includes fractured zones can be modeled with the
POLLUTE code.
Probabilistic modeling approaches, as described above, have not been
employed in the analysis of containment barriers, and the current knowledge
10.3 ASSESSMENT
283
regarding failure mechanisms and model parameters is insufficient to support

the use of this approach. Nevertheless, it may be equally inappropriate to
rely on deterministic models to predict the performance of barriers over their
design lifetimes, particularly when they do not incorporate the possibility of
barrier failure due to damage accumulation. The development of appropriate
long-term modeling frameworks and identifying the supporting data needed
should be a research priority.
Potential causes of long-term damage to barriers include: freeze-thaw
cycling, wet-dry cycling, thermal effects, chemical action of leachates,
weathering, earthquakes, land subsidence, and, in the case of horizontal
barriers, bearing capacity failure. Thermal cracking, as related to the disposal
of radioactive materials, is a primary concern (see Section 7). In experimental
studies conducted by Britto et al. (1989), burial of high-temperature wastes
in sediments caused fractures that led to increased contaminant transport.
Although high-temperature materials will not generally be placed in the
immediate vicinity of engineered barriers, additional experimental work is
still needed to delineate scenarios under which heat-related damage may
occur.
The chemical compatibility of barrier materials with the permeating
solution is another concern (Shackelford, 1994b). Experimental studies have
identified strong acids and bases, strong solutions of electrolytes, alcohols,
and nonaqueous phase liquid (NAPL) organics as contaminants of concern,
in this regard. The chemical compatibility of a leachate with a barrier material
can be investigated in the laboratory by passing several pore volumes of the
leachate through a permeameter. Manassero and Shackelford (1994) proposed
a compatibility index for identifying potential compatibility problems.
These approaches have tended to identify compatibility problems in terms of
a yes/no evaluation. From a modeling standpoint, however, quantitative
functional relationships are needed relating in situ hydraulic conductivity to
chemical concentrations, if long-term chemical damage is to be incorporated
into ADRE models. Further research is needed in this area, particularly for
inorganic compounds. In the case of NAPLs, it is unlikely that hydraulic
gradients would be large enough to produce significant NAPL penetration
into an engineered barrier.
In summary, while there is an appreciation of the mechanisms that may
produce damage to engineered barriers, there are no mathematical models
presently available that incorporate these mechanisms. Yet, if containment is
to be accepted as a viable option in remediation strategies, new models must
be developed that predict long-term behavior by including mechanisms that
cause time dependent damage accumulation. In the near-term, hydraulic
control (i.e., pumping to create inward gradients) may represent the best
strategy for addressing the uncertainties associated with damage
accumulation to engineered barriers. In conjunction with physical barriers,
such control should be relatively cost-effective, compared to conventional
pump-and-treat systems.
284
10.4 NEEDS
10.4.1
Performance Criteria for Low Hydraulic Conductivity Barriers
Although clean-up criteria in environmental remediation are frequently

concentration-based standards, in the case of engineered containment barriers,
flux-based standards (that provide acceptable risk levels external to the
containment system) would be more appropriate. This statement is based on
the following:
Since transport in low hydraulic conductivity barriers is likely to be

diffusion-dominated, attainment of a low concentration at the barrieraquifer interface does not, by itself, ensure that contaminant flux out
of the containment barrier system is insignificant. Over time, a
significant amount of contaminant mass may be transported across
a barrier due to diffusion, even when low concentrations exist at the
barrier-aquifer interface.
The recommended boundary condition (flushing condition) for onedimensional analysis of a low hydraulic conductivity barrier does
not give meaningful predictions of contaminant concentration at the
barrier-aquifer interface; however, it does yield meaningful and
conservative flux predictions.
Currently, both flux and concentration criteria are under consideration

as elements of proposed regulations related to landfill design in Germany
(Teutsch, personal communication, 1995). While adoption of a flux-based
standard represents a conceptual shift from the traditional approach, it would
facilitate the use of tractable models and produce conservative system designs.
In some situations, it may be advisable to supplement flux-based criteria with
a requirement for some specified level of hydraulic control. In the future,
such control may become unnecessary as more data become available and
models are developed that more confidently predict the long-term
performance of engineered barriers.
10.4.2
Software Development
An attractive approach to modeling barrier systems that would enable the

consideration of the barrier system in the context of a larger field-scale model
is the coupled approach in which a local (e.g., one-dimensional) model of
reactive transport within a barrier system is linked as a source term to a largerscale multidimensional model of flow and transport. Because the conditions
in the surrounding aquifer influence the boundary condition at the barrier, a
dynamic linkage of the two models would be most desirable, enabling a
realistic treatment of the barrier system while maintaining the larger
discretization and simpler chemical model appropriate for the field scale. Such
10.4 NEEDS
285
a linkage, while conceptually straightforward, has not been implemented in

practice, and should be a priority for future model development activities.
10.4.3
Better Understanding of Reactive Transport
In general, accurate modeling of multispecies reactive transport is poorly

developed. For some low-permeability barrier systems, neglecting the role
of chemical reactions may be conservative and appropriate. The development
of reactive low-permeability barriers may be a promising alternative, however,
to the construction of the massive barriers that would be necessary to limit
diffusive flux over time frames of interest. If such an approach is to be
successful, development of appropriate and reliable models for reactive
transport is essential. Currently, the relative importance of various reaction
process is poorly understood, particularly for systems that contain mixtures
of contaminants. Because engineered barriers will typically exhibit chemical
behavior that is different from the surrounding media, a solid understanding
of system behavior is needed to support innovative design.
10.4.4
Field Data for Model Testing
Confidence in model predictions can only be established through

documentation of successful simulations in a post-audit mode. In particular,
considerable uncertainty is associated with modeling of low-permeability
barrier systems, due to:
the need for modeling predications over time horizons much longer
than any experimental or historical observations of similar systems;
difficulty in conducting laboratory experiments that are representative
of field conditions;
uncertainty regarding the correspondence between low-permeability
barrier systems as designed and constructed; and
the lack of information regarding time-evolution of barrier properties.
For these and other reasons, it is crucial that coordinated efforts be initiated
to develop a database for field applications that includes spatial and temporal
measurements of parameters influencing contaminant transport.
10.4.5
Long-Term Performance
From conceptual standpoint, engineered barriers are unlikely to perform as

intended over the design lifetime of the system, and realistic models should
include consideration of the time-evolution of barrier properties. In practice,
however, neither an appropriate modeling framework for barrier systems nor
data for parameterization of failure and damage accumulation models are
available. While there are analogues in other disciplines, the unique nature
286
of subsurface porous media systems complicate the application of models

incorporating failure in a realistic manner. If predictions of long-term barrier
performance are to be credible, a framework must be developed for dealing
with the anticipated long-term evolution of barrier properties. Along with
the acquisition of field data, the development of such a framework should be
a research priority and a subject of ongoing dialogue among modelers and all
affected parties. In the short term, performance criteria may need to include
some level of hydraulic control to compensate for uncertainty regarding longterm performance.
10.5
SUMMARY
10.5.1
Background
The primary focus of the Contaminant Transport Modeling session was on

simulation of groundwater flow and advective-dispersive-reactive
contaminant transport, as applied to low permeability barriers. There was
general agreement regarding the applicability of the fundamental theories
commonly applied to porous media and the availability of a variety of suitable
analytical and numerical models. The primary focus of the discussion was
on identifying appropriate modeling strategies for applied modeling and
decision making, and in this regard a major concern was the establishment of
correspondence between modeling objectives, the degree of detail needed in
a predictive model, and the availability of data for parameter estimation.
10.5.2
State of Practice
The assessment of the state of practice varied according to the modeling

objective under consideration. In general, it is believed that low-permeability
barriers can be readily incorporated into groundwater flow models through
well-established procedures; the popular MODFLOW model, for example,
has recently been expanded to include a barrier module. Numerous examples
are available in which modeling studies had been used to analyze the impact
of low-permeability barriers on groundwater flow.
With regards to contaminant transport, it suggested here that the inclusion
of a low permeability barrier in a regional or field-scale model, while tractable,
is unrealistic for routine applications, due to the grid resolution needed to
handle the discontinuity in media properties between the barrier and adjoining
aquifer. A coupled modeling approach is suggested, in which a local model
of the low-permeability barrier system would be solved separately and
coupled with a field-scale transport code (e.g., MT3D) as a source term. While
this coupling is expected to be straightforward, it has not been implemented.
A considerable amount of discussion addressed the appropriate
application of one-dimensional transport models. One-dimensional
10.5 SUMMARY
287
contaminant transport models for horizontal low-permeability barriers have

been developed and widely used, and with appropriate modification of the
conceptual model may be readily applied to vertical low-permeability barriers.
In general, modeling predictions would be best expressed in terms of flux
from within a contained area, rather than concentrations at a point outside
the barrier.
Performance criteria for low-permeability barrier systems should be
referenced to the flux concept; recent consideration of this approach in
Germany for landfill design was cited.
10.5.3
Assessment
In general, the use of one-dimensional models for low-permeability barrier

systems was viewed as reasonable approach when the barrier performs as
designed, i.e., for diffusion-dominated transport with constant, spatially
uniform barrier properties. A variety of appropriate published solutions are
available for various boundary conditions and reaction scenarios. For a
conservative analysis, a constant-concentration entrance boundary and a zeroconcentration exit boundary are recommended. The Appendix contains an
example of a useful solution that includes these conditions and the processes
of advection, linear equilibrium sorption, and first-order decay.
A number of areas have been identified in which one-dimensional
transport models may be inadequate. In particular, the importance of a realistic
representation of chemical reactions is highlighted. Since most relevant
reactions result in reduced contaminant flux, however, neglecting reaction
terms may result in conservative predictions and may, therefore, be acceptable
for routine applications. For cases in which reactive low-permeability barrier
materials were specified as a means of meeting performance criteria, a more
realistic treatment of reaction terms is essential. For example, the validity of
the local equilibrium assumption has not been established for earthen barrier
materials amended to promote sorption of contaminants. Other areas of
concern include an inadequate understanding of coupled flow processes, and
a lack of experimental studies performed under the low flow conditions and
long time frames relevant to low-permeability barrier systems.
Problems were noted with respect to the use of simplified groundwater
flow models (e.g., two dimensional, homogeneous) in the design of permeable
reactive barriers. For this application, it is important to consider the effects of
hydraulic conductivity heterogeneity, as even minor variations in spatial and
temporal flow conditions could result in contaminants bypassing the treatment
zone. The use of multi-dimensional numerical models and a refined
computational grid was recommended for this application.
Discussion regarding parameter estimation focused on the measurement
of diffusion and sorption coefficients. While laboratory methods for
estimating these parameters are well established, the use of literature values
may be appropriate and conservative for many design applications. As noted
288
previously, more precise estimates may be required for reactive lowpermeability barriers.
An unresolved issue relates to the consideration of barrier damage
accumulation. Engineered barriers are unlikely to maintain constant
properties over the course of decades or centuries, but a consensus has not
been reached regarding how to incorporate damage accumulation and failure
into predictive models. Discussion of this topic among committee members
should continue.
10.5.4
Recommendations
With regards to future research efforts, three needs are highlighted:
10.6
implementation of well-instrumented, long-term field studies, in

which data are collected over an extended time period;
additional laboratory studies of contaminant transport in lowpermeability materials, especially those involving reactive
contaminants and low-flow conditions; and
further development of a conceptual approach for modeling damage
accumulation and failure in barrier systems, eventually leading to an
established procedure for incorporating these considerations into
predictions of long-term performance.
REFERENCES
Abriola, L. M., and Pinder, G. F. (1985). A multiphase approach to modeling

porous media contamination by organic compounds, Water Resources
Research, 21(1), 19-26.
Acar, Y. B., and Haider, L. (1990). Transport of low-concentration
contaminants in saturated earthen barriers, Journal of Geotechnical
Engineering, 116(7), 1031-1052.
A.D. Little, Inc. (1983). S-Area two phase flow model, Report to Wald,
Hardrader & Ross, Reference 84204-31 (May).
Anderson, M. P., and Woessner, W. W. (1992). The role of the postaudit in
model validation, Advances in Water Resources, 15(3), 167-173.
Andersen, P. F., Faust, C. R., and Mercer, J. W. (1984). Analysis of conceptual
designs for remedial measures at Lipari landfill, New Jersey, Ground
Water, 22, 176-190.
Ayres, J. E., Leger, D. C., and Barvenik, M. J. (1983). The first EPA Superfund
10.6 REFERENCES
289
cutoff wall: design and specifications, presented at the 3rd National

Symposium and Exposition of Aquifer Restoration and Ground Water
Monitoring, National Water Well Association, Columbus, Ohio.
Babu, D. K., and Pinder, G. F. (1984). A finite-element alternating direction
algorithm for three-dimensional groundwater transport, Advances in
Water Resources, 7, 116-119.
Barvenik, M. J., Hadge, W. E., and Goldberg, D. T. (1985). Quality control of
hydraulic conductivity and bentonite content during soil/bentonite cutoff
wall construction, Proceedings of the Eleventh Annual Research Symposium,
United States Environmental Protection Agency, EPA/600/9-85/013,
Cincinnati, OH.
Barvenik, M. J., Brown, D. M., and Kern, T. J. (1986). Evaluation of cutoff
wall containment efficiency using aquifer stress analysis - Gilson Road
NPL site, Nashua, New Hampshire, Proceedings of the Twelfth Annual
Research Symposium, United States Environmental Protection Agency,
EPA/600/9-86/022, Cincinnati, OH.
Barvenik, M. J., and Ayres, J. E. (1987). Construction quality control and postconstruction performance verification for the Gilson Road Hazardous
Waste Site cutoff wall, report prepared for the United States
Environmental Protection Agency, EPA/600/2-87/065, Washington, DC.
Bear, J. (1979). Hydraulics of groundwater. McGraw Hill, New York, NY.
Bear, J. (1992). Dynamics of fluids in porous media. Elsevier, New York, NY.
Bierck, B. R., and Chang, W. C. (1994). Contaminant transport through soilbentonite slurry walls: attenuation by activated carbon, Proceedings, Water
Environment Speciality Conference on Innovative Solutions for Contaminated
Site Management, Miami, FL, March 6-9, 1994, 461-472.
Brenner, H. (1961). The diffusion model of longitudinal mixing in beds of finite
length: Numerical values, Chemical Engineering Science, 17, 229-243.
Britto, A. M., Savvidou, C., Maddocks, D. V., Gunn, M. J., and Booker, J. R.
(1989). Numerical and centrifuge modeling of coupled heat flow and
consolidation around hot cylinders buried in clay, Geotechnique, 39(1),
13-25.
Brusseau, M. L., and Rao, P. S. (1989a). The influence of sorbate-organic
matter interactions on sorption nonequilibrium, Chemosphere, 18(9-10),
1691-1706.
290
Brusseau, M. L., and Rao, P. S. (1989b). Sorption nonideality during organic

contaminant transport in porous media, Critical Reviews in Environmental
Control, 19(1), 33-99.
Brusseau, M. L., Rao, P. S., and Jessup, R. E. (1989). Flow interruption, a
method for investigating sorption nonequilibrium, Journal of
Contaminant Hydrology, 4, 223-240.
Carslaw, H. S., and Jaeger, J. C. (1959). Conduction of heat in solids. Oxford,
London.
Cohen, R. M., and Mercer, J. W. (1984). Evaluation of a proposed synthetic
cap and concrete cut-off wall at Love Canal using a cross-sectional model,
Fourth National Symposium and Exposition on Aquifer Restoration and
Ground-Water Monitoring, Columbus, OH (May).
Cohen, R. M., Rabold, R. R., Faust, C.R., Rumbaugh, J. O. III, and Bridge, J. R.,
(1987). Investigation and hydraulic containment of chemical migration:
Four landfills in Niagara Falls, Civil Engineering Practice, Journal of the
Boston Society of Civil Engineers Section/ASCE, 2(1):33-58.
Cohen, R. M., and Mercer, J. W. (1993). DNAPL site evaluation, C. K. Smoley
(CRC Press, Inc.) Boca Raton, FL.
Crank, J. (1975). The mathematics of diffusion, Oxford, London.
Crooks, V. E., and Quigley, R. M. (1984). Saline leachate migration through
clay: A comparative laboratory and field investigation, Canadian
Geotechnical Journal, 21, 349-362.
Cussler, E. L. (1988). Diffusion: mass transfer in fluid systems, Academic
Press, New York, NY.
Danckwerts, P. V. (1953). Continuous flow systems: distribution of residence
times, Chemical Engineering Science, 2, 1-13.
de Marsily, G. (1986). Quantitative hydrogeology, Academic Press, San Diego,
CA.
Faust, C.R. (1984). Affidavit for Civil Action No. 79-988, United States of
America, et al., Plaintiffs v. Hooker Chemical and Plastics Corporation, et
al. (S-Area Landfill) Defendants, in the United States District Court for
the Western District of New York.
Faust, C. R., Guswa, J. H., and Mercer, J. W. (1989). Simulation of three-
10.6 REFERENCES
291
dimensional flow of immiscible fluids within and below the unsaturated

zone, Water Resources Research, 25(12), 2499-2464.
Fetter, C. W. (1993). Contaminant hydrogeology, Macmillan, New York, NY.
Goodal, D. C., and Quigley, R. M. (1977). Pollutant migration from two
sanitary landfill sites near Sarnia, Ontario, Canadian Geotechnical
Journal, 14, 223-236.
Gray, D. G., and Weber, W. J. (1984). Diffusional transport of hazardous
waste leachate across clay barriers, Proceedings of the Seventh Annual
Madison Waste Conference, Madison, WS.
Haggerty, R., and Gorelick, S. M. (1995). Multiple-rate mass transfer for
modeling diffusion and surface reactions in media with pore-scale
heterogeneity, Water Resources Research, 31(10), 2383-2400.
Harmon, T. C., and Roberts, P. V. (1994). Comparison of intraparticle sorption
and desorption rates for a halogenated alkene in a sandy aquifer material,
Environmental Science & Technology, 28(9), 1650-1660.
Hensley, P. J., and Schofield, A. N. (1991). Accelerated physical modeling of
hazardous waste transport, Geotechnique, 41(3), 447-465.
Hsieh, P. A., and Freckleton, C. R. (1993). Documentation of a computer
program to simulate horizontal-flow barriers using the U. S. Geological
Surveys modular three-dimensional finite difference groundwater
model, U. S. Geological Survey, Open-File Report 92-477, Sacramento,
CA.
Inyang, H. I. (1995). A Weibull-based reliability analysis of waste
containment systems, Proceedings of the First International Conference on
Environmental Geotechnics, July 11-15, Edmonton, Ontario, Canada, 273278.
Jennings, A. A., and Kirkner, D. J. (1984). Instantaneous equilibrium
approximation analysis, Journal of Hydraulic Engineering, 110, 17001717.
Johnson, R. L., Cherry, J. A., and Pankow, J. F. (1989). Diffusive contaminant
transport in natural clay: A field example and implications for clay-lined
waste disposal sites, Environmental Science & Technology, 23, 349-349.
Kaluarachchi, J. J., and Morshed, J. (1995). Critical assessment of the operatorsplitting technique in solving the advection-dispersion-reaction equation:
292
1. First-order reaction, Advances in Water Resources, 18(2), 89-100.

Kinninburgh, D. G. (1986). General purpose adsorption isotherms,
Environmental Science & Technology, 20(9), 895-903.
Lyman, W. J., Reehl, W. F., and Rosenblatt, D. H. (1990). Handbook of chemical
property estimation methods, American Chemical Society, Washington, DC.
Mabey, W., and Mill, T. (1978). Critical review of hydrolysis of organic
compounds in water under environmental conditions, Journal of Physical
Chemistry Reference Data, 7, 383-415.
Manassero, M., and Shackelford, C.D. (1994). Classification of industrial
waste for reuse and landfilling, Proceedings of the First International
Conference on Environmental Geotechnics, July 11-15, Edmonton, Ontario,
Canada, 103-114.
McDonald, M. G., and Harbaugh, A. W. (1988). A modular three-dimensional
finite-difference groundwater flow model, U. S. Geological Survey, OpenFile Report 92-477, Sacramento, CA.
Mercer, J. W., Faust, C. R., Truschel, A. D., and Cohen, R. M. (1987). Control
of groundwater contamination: Case studies, Proceedings of Detection,
Control, and Renovation of Contaminated Ground Water, American Society
of Civil Engineers, Environmental Engineering Division, 121-133.
Miller, C. T., and Rabideau, A. J. (1993). Development of split-operator,
Petrov-Galerkin methods to simulate transport and diffusion problems,
Water Resources Research, 29(7), 2227-2240.
Millington, R. J. (1959). Gas diffusion in porous media, Science, 130, 100-102.
Mitchell, J. K. (1991). Conduction phenomena: from theory to geotechnical
practice, Geotechnique, 41(3), 299-340.
Moody, J. B. (1982). Radionuclide migration/retardation: research and
development technology status report. ONWI-321, Office of Nuclear
Waste Isolation, Battelle Memorial Institute, Columbus, OH.
Morshed, J., and Kaluarachchi, J. J. (1995). Critical assessment of the operatorsplitting technique in solving the advection-dispersion-reaction equation:
2. Monod kinetics, Advances in Water Resources, 18(2), 100-110.
Mott, H. V., and Weber, W. J. Jr. (1991a). Diffusion of organic contaminants
through soil-bentonite cut-off barriers, Research Journal of the Water
10.6 REFERENCES
293
Pollution Control Federation, 63(2), 166-176.

Mott, H. V., and Weber, W. J. Jr. (1991b). Factors influencing organic
contaminant diffusivities in soil-bentonite cutoff barriers, Environmental
Science & Technology, 25(10), 1708-1715.
Mott, H. V., and Weber, W. J., Jr. (1992). Sorption of low molecular weight
organic contaminants by fly ash: considerations for the enhancement of
cutoff barrier performance, Environmental Science & Technology, 26,
1234-1241.
Neuzil, C. E. (1986). Groundwater flow in low-permeability environments,
Owen, S. P. (1925). The distribution of temperature in a column of liquid
flowing from a cold source into a receiver maintained at a higher
temperature, Proceedings of the Physics Society of London, 23, 238-249.
Parker, J. C., and van Genuchten, M. Th. (1984a), Determining transport
parameters from laboratory and field experiments, Bulletin 84-3, Virginia
Agricultural Station, Blacksburg, VA.
Parker, J. C., and van Genuchten, M. Th. (1984b). Flux-averaged and volumeaveraged concentrations in continuum approaches to solute transport,
Pinder, G. F., Huyakorn, P. S., and Sudicky, E. A. (1993). Simulation of flow
and transport in fractured porous media, in Flow and Contaminant
Transport in Porous Rock, Bear, J., Tsang, C., and de Marsily, G., (Eds.),
Academic Press, San Diego, CA.
Rabideau, A. J., and Miller, C. T. (1994). Two-dimensional modeling of aquifer
remediation influenced by hydraulic conductivity heterogeneity and
nonequilibrium sorption, Water Resources Research, 30(5), 1457-1470.
Rabideau, A. J., Khandelwal, A., and Millar, D. (1996). One-dimensional
modeling of contaminant transport in vertical barriers, submitted to
ASCE Journal of Environmental Engineering.
Rabideau, A. J., and Khandelwal, A. (1996). Analysis of nonequilibrium
sorption in soil/bentonite barriers, submitted to ASCE Journal of
Environmental Engineering.
Reid, R. C., Prausnitz, J. M., and Sherwood, T. K. (1977). The Properties of
Gases and Liquids, 3rd Ed., McGraw-Hill, New York, NY.
294
Rowe, R. K. (1987), Pollutant transport through barriers, Geotechnical

Practice for Waste Disposal, ASCE Special Publication No. 13, 159-181.
Rowe, R. K., and Booker, J. R. (1985). 1-D pollutant migration in soils of
finite depth, Journal of Geotechnical Engineering, 111(4), 479-499.
Rowe, R. K., Quigley, R. M., and Booker, J. R. (1995). Clayey Barrier Systems
for Waste Disposal Facilities. E & FN Spoon, London.
Rumer, R. R., and Ryan, M. R. (Ed.) (1995). Barrier Containment Technologies
for Environmental Remediation Applications, John Wiley & Sons, Inc.,
New York.
Russakoff, D. (1983). EPA adds 133 toxic dumps to Superfund cleanup list,
Washington Post, Washington, D.C. (September 2).
Schroeder, P. R., Dozier, T. S., Zappi, P. A., McEnroe, B. M., Sjostrom, J. W.,
and Peyton, R. L. (1994). The hydrologic evaluation of landfill
performance (HELP) model, EPA/600/R-94/168a, U. S. Environmental
Protection Agency Risk Reduction Laboratory, Cincinnati, OH.
Schulze, D., Barvenik, M. J., and Ayres, J. E. (1984). Design of the soil/
bentonite backfill mix for the first EPA Superfund cutoff wall, Proceedings
of the 4th National Symposium and Exposition of Aquifer Restoration and Ground
Water Monitoring, National Water Well Association, Columbus, Ohio.
Shackelford, C. D. (1988). Diffusion as a transport process in fine-grained
barrier materials, Geotechnical News, 6(2), 124-127.
Shackelford, C. D. (1990). Transit-time design of earthen barriers,
Engineering Geology, 29, 7994.
Shackelford, C. D. (1989). Diffusion of contaminants through waste
containment barriers, Transportation Research Record, 1219, 169-182.
Shackelford, C. D. (1991). Laboratory diffusion testing for waste disposal A review, Journal of Contaminant Hydrology, 7, 177-217.
Shackelford, C. D. (1991). Contaminant transport, Chapter 3, Geotechnical
Practice for Waste Disposal, Daniel, D. E. (Ed.), Chapman-Hall Publ.,
London.
Shackelford, C. D. (1994a). Critical concepts for column testing, Journal of
Geotechnical Engineering, 120(10).
10.6 REFERENCES
295
Shackelford, C. D. (1994b). Waste-soil interactions that alter hydraulic

conductivity, ASCE Specialty Technical Publication 1142, Daniel, D. E.
And Trautwein, S. J. (Ed.), ASTM, Philadelphia, PA, 111-168.
Shackelford, C. D., and Daniel, D. E. (1991). Diffusion in saturated soil: I.
Background, Journal of Geotechnical Engineering, 117(3), 467-484.
Shackelford, C. D. and Redmond, P. (1995). Solute breakthrough curves
through processed kaolin at low flow rates, Journal of Geotechnical
Engineering, 121(1), 17-32.
Shackelford, C. D., Daniel, D. E., and Liljestrand, H. M. (1989). Diffusion of
inorganic chemical species in compacted clay soil, Journal of
Contaminant Hydrology, 4, 441-473.
Starr, R. C., and Cherry, J. C. (1994). In situ remediation of contaminated
ground water: the funnel-and-gate system, Ground Water, 32(3), 465475.
Teutsch, G., and Schad, H. (1995). Performance and reliability assessment
for in situ reactive walls in heterogeneous environments, accepted for
presentation at the MODELCARE 96 conference, 25-27 Sept., 1996, Golden,
CO.
Tchobanoglous, G., and Schroeder, E. D. (1987). Water Quality, AddisonWesley, Reading, MA, p 286.
Trescott, P. C. (1975). Documentation of a finite-difference model simulation
of threedimensional groundwater flow, U.S. Geological Survey Open
File Report 75-438.
Trescott, P. C., Pinder, G. F., and Larson, S. P. (1976). Finite-difference model
for aquifer simulation in two dimensions with results of numerical
experiments, Techniques of Water Resources Investigations of the U.S.
Geological Survey, Book 7.
U.S. Environmental Protection Agency (1982). Environmental monitoring
at Love Canal, Vol. 1, EPA-600/4-82-030a, Washington, D.C.
U.S. Environmental Protection Agency (1984). Summary report: Remedial
response at hazardous waste sites, EPA/540/2-84/002a, Washington,
DC.
U.S. Environmental Protection Agency (1991). Seminar publication: Site
characterization for subsurface remediation, EPA/625/4-91/026, Office
296
of Research and Development, Washington, DC.

van der Heijde, P. K. and Elnawawy, O. A. (1993). Compilation of GroundWater Models, U.S. Environmental Protection Agency, R. S. Kerr
Environmental Research Laboratory, Ada, OK.
van Genuchten, M. Th., and Wierenga, P. J., (1976). Mass transfer studies in
sorbing porous media I. Analytical solutions, Soil Science Society of
America Journal, 41, 272-278.
Vogel, T. M., Criddle, C. S., and McCarty, P. L. (1987). Transformation of
halogenated aliphatic compounds, Environmental Science & Technology,
21, 722-736.
Weber, Jr., W. , McGinley, P. M., and Katz, L. E. (1991). Review paper: Sorption
phenomena in subsurface systems: concepts, models, and effect on
contaminant fate and transport, Water Research, 25(5), 499-528.
Wilke, C. R., and Chang, P. (1955). Correlation of diffusion coefficients in
dilute solutions, AIChE Journal, 1, 264-270.
Zheng, C., and Bennett, G. B. (1995). Applied Contaminant Transport
Modeling, Van Nostrand Reinhold, New York, NY.
Zheng, C. (1992). MT3D Version 1.8 documentation and users guide, S. S.
Papadopulos & Associates, Inc., Bethesda, MD.
10.7 APPENDIX
Transport Equation/Solution for One-Dimensional Transport of a
Miscible Contaminant in a Low Hydraulic Conductivity Barrier
A useful model for describing the transport of a miscible contaminant in a
one-dimensional barrier system is
Rf
C
C
2 C
=
+ D 2 C
t
x
x
where C is the volume-averaged aqueous phase contaminant concentration,

t is time, x is the distance from the entrance side of the barrier, v is the interstitial
fluid velocity, D is the dispersion coefficient (includes hydrodynamic
dispersion and molecular diffusion), is the first-order decay coefficient, and
10.7 APPENDIX
297
Rf is the retardation factor.

As discussed in the text, a conservative set of auxiliary conditions for
predicting contaminant flux under diffusion-dominated conditions is
C(x,0) = 0
C(0,t) = C o
C(L,t) = 0
where L is the barrier thickness and Co is the constant contaminant

concentration on the entrance side of the barrier. A solution to the above
system is provided by Owen (1925), modified here to incorporate the
retardation factor.
=:=
==
==
==
=
=
C = C o ;=
= 2
==
=+ 2
== L
==
=
<
sinh [L x] 2 + 4D
sinh L 2 + 4D
mx
m
4 ( 1) msin
L
exp
3
2
2
2D
m=1
m
+
+
2
D 4D2
L
exp
x
2D
2
2

Dm
exp +
+
2
R f 4DR f R f L
EBE
EE
EE
EE
E
E
CE
E
EE
E
EE
EE
E
D
The concentration-based solution may be modified to represent the

contaminant flux normal to the barrier according to
J = n C nD
C
x
where J is the contaminant flux and n is the porosity.

The development of the flux-based solution is straightforward, as presented
by Rabideau, et al. (1996)
298
=:
=
==
==
==
==
=
==
= 3
== + 2
= L
==
x ==
;
J = nC o exp
L ===
==
=
==
==
==
==
= 2D 2
==
=+ 3
== L
=
=
<
sinh [L x] 2 + 4D / 2D
sinh L 2 + 4D / 2D
m (L x)
( 1) m msin
4
L
3
2
m=1
m22
+
2 +
D 4D
L2
+
Dm 2 2
exp +
+
4DR f
R fL2
2 + 4D cosh [L x] 2 + 4D / 2D
2
2
sinh L + 4D / 2D
m (L x)
( 1) m m 2 sin
L
4
3
2
2 2
m=1
m
+
+
D 4D2
L2
Dm 2 2
2
exp +
+
4DR f
R fL2
EB
E
EE
EE
EE
EE
E
EE
E
EE
E
EE
E
E
CE
EE
EE
E
EE
EE
EE
EE
E
EE
E
EE
E
E
D
The net flux of contaminant crossing the barrier at a specified time may be
determined by evaluating the above expression at (x=L ).
10.8
LIST OF SYMBOLS
al
C
Co
Cs
D
Dh
D*
Dl
h
hb
h
HF
J
K
kd
KF
KE
L
Longitudinal dispersivity (L)

Aqueous phase solute concentration (M/L3)
Known constant concentration (M/L3)
Concentration of source/sink (M/L3)
Dispersion coefficient (L2/T)
Hydrodynamic mixing component of dispersion coefficient (L2/T)
Effective solute diffusion coefficient (L2/T)
Solute liquid diffusivity (L2/T)
Piezometric head (L)
Dimension of adjoining aquifer normal to barrier (L)
Difference in piezometric head between the inside and outside of a
barrier wall (L)
Dimension of contaminated area inside barrier (L)
Solute flux (M/L2-T)
Hydraulic conductivity (L/T)
Sorption distribution coefficient (L3/M)
Freundlich isotherm parameter (L3/M)nf
Effective hydraulic conductivity (L/T)
Barrier thickness (L)
10.8 LIST OF SYMBOLS
MW Molecular weight of water (M/L3)

n Porosity
nf Freundlich isotherm exponent
Pe Peclet number
qs Fluid source/sink (L3/T)
rB Solute molecular radius (L)
R Ideal gas constant (0.082 atm-L/mol-K)
Rf Retardation factor
S Solute sorbed phase mass fraction (M/M)
Ss Specific storage (1/L)
T Absolute temperature (oK)
t50 Half-life (T)
v Seepage velocity (L/T)
vb Velocity of groundwater in aquifer adjoining the barrier (L/T)
VB Solute molar volume at boiling point (L3)
W Width of barrier normal to fluid/concentration gradients (L)
x Spatial coordinate (L)
Sorption rate coefficient (1/T)
Exponent used in correlation between tortuosity and porosity
a First-order decay constant for the aqueous phase (1/T)
s First-order decay constant for the solid phase (1/T)
b Lumped first-order decay constant (1/T)
b Bulk density (M/L3)
W Solution association constant in Wilke-Chang correlation
Tortuosity factor
W Dynamic viscosity of water (FT/L2)
Integration variable
299
301
SECTION 11
PERMEABLE REACTIVE BARRIERS
prepared by
Stephen H. Shoemaker, DuPont Co.,

John F. Greiner, DuPont Environmental Remediation Services,
and Robert W. Gillham, University of Waterloo, Canada
David Blowes, University of Waterloo, Canada

David R. Burris, Armstrong Laboratory, Tyndall AFB
Peter Grathwohl, University of Tbingen, Germany
Stephan Jefferis, Golder Associates Ltd, United Kingdom
Patricia Mackenzie, General Electric Co.
Dianne C. Marozas, Sandia National Laboratories
Steve McCutcheon, U.S. Environmental Protection Agency
Stan Morrison, RUST Geotech
Robert Orth, Monsanto Co.
Steve Paulson, U.S. Bureau of Mines
Mark Phifer, Westinghouse Savannah River Co.
Robert W. Puls, U.S. Environmental Protection Agency
Dale Schultz, DuPont Co.
Georg Teutsch, University of Tbingen, Germany
Bruce Thomson, University of New Mexico
Scott D. Warner, Geomatrix Consultants, Inc.
11.1 BACKGROUND
A permeable reactive barrier (PRB) consists of a permeable curtain containing
appropriate reactive materials, generally constructed to intercept the path of a
contaminant plume. As the contaminated groundwater passes through the
curtain, the contaminants are removed through chemical, physical, or biological
processes (Gillham,1995). PRBs may also be installed downgradient of
contaminant sources to prevent plumes from developing. Although permeable
reactive walls are generally vertically oriented, horizontal applications have
been considered for controlling the downward migration of contaminants. This
latter application may prove feasible in some situations. Discrete areas may be
treated by the use of fencepost-style reactive zones placed into the natural
groundwater flow system (Puls et al., 1995). In addition, PRBs may be used in
low permeability strata when the local hydraulic gradient must be
supplemented with another driving force, such as electroosmosis (Orth, 1995).
302
PRBs are comprised of a reactor system and an optional control system to

guide the groundwater flow through the PRB. The control system may be
advantageous in situations where the contaminant plume is so large that the
dimensions of a continuous PRB would be economically impractical. In these
cases, sheet piling or low permeability slurry walls may be employed to funnel
the plume more efficiently through a small reactor system. Under favorable
hydrogeologic conditions, the natural groundwater flow system may require
no control, and a continuous permeable wall reactor system may be employed.
The reactor system contains a porous reactive medium designed to treat
the dissolved contaminants as the groundwater flows through the PRB. The
reactive media may be permanent or replenishable, depending on the nature
of the site geochemistry, reaction mechanisms, contaminant loading, degree
of contaminant reduction desired, and design lifetime of the PRB. Reactors
may extend the full width of the contaminant plume (as in the case of a
continuous barrier) or be combined (singly or in multiples) with low hydraulic
conductivity barrier walls (funnel elements). Also, PRBs containing different
reactive porous media may be constructed in series to treat individual
contaminant constituents.
PRBs can utilize a wide variety of reactive materials to effect contaminant removal. Removal mechanisms can be biotic or abiotic and include:
sorption, precipitation, dehalogenation, oxidation-reduction, fixation, and
physical transformation. PRBs can be installed by excavating a trench and
emplacing the porous reactive medium, by flushing-type emplacement and
replenishment of the reactant using injection wells, or by other specialized
means discussed in this section.
PRBs can be passive or active, based on the need for post-construction
operations at the site. The passive PRB system utilizes natural groundwater
flow to effect treatment. The active system requires postconstruction
operation of extraction and injection wells to effect treatment. For example,
operation of a series of air sparging wells or continuous in-well aeration is
considered active. For the most part, this section focuses on passive PRB
technology.

11.2.1 Introduction
For an in situ reactor system to be effective in removing groundwater
contaminants, three basic requirements must be met. First, the time required
for the desired reactions to occur must be less than the time required for the
contaminated groundwater to pass through the reactor. If the reaction rates
are too slow, only partial transformation and removal of the contaminant will
occur, leading to incomplete treatment, and possible failure of the system to
meet regulatory compliance goals.
303
The second requirement is that the emplaced reactive treatment medium

perform satisfactorily for an economically viable period. Contaminant plumes
can persist for decades or longer (especially when DNAPL source zones are
present), requiring the treatment system to be effective for a corresponding
period. Under some geochemical settings and contaminant loadings, the
selected reactive materials may require replacement during the design lifetime
of the PRB. Design of a PRB for a given application may call for optimization
in the selection of the reactive media so that it is sufficiently reactive to effect
treatment during the design flow residence times, yet sufficiently stable to be
effective for an economically viable period.
The third requirement is that the reactive medium itself not introduce
contamination that would be unacceptable in the downgradient groundwater
(Blowes, et al., 1995a). In a poorly designed reactor system, reaction byproducts might be formed that are less desirable than the original contaminant.
The goal in most cases is to achieve risk-based groundwater cleanup
requirements; for example, U.S. EPA maximum contaminant levels (MCLs)
in the case of a drinking water aquifer, or levels consistent with surface water
protection where groundwater is not used directly.
Thus, in the design and construction of a permeable reactive barrier, due
consideration needs to be given to the requirements discussed above. The
basic elements which factor into the design and construction of PRBs are:
selection of the proper treatment technology/medium,

system design protocol, including
- site characterization in advance of the design,
- design of the optional control system, when necessary,
- design of the reactor system,
emplacement/construction of the permeable reactive barrier, and
performance monitoring.
11.2.2 Permeable Barrier Treatment Technologies

The selection of a treatment medium is based primarily on its effectiveness to
treat site-specific contaminants. Treatment schemes can be based on organic,
inorganic, and combined (organic and inorganic) reaction processes.
Frequently, several different contaminants can be treated utilizing the same
or related reaction processes. For example, sorption may be effective in
removing hydrocarbons and certain metals from contaminated groundwater.
For this reason, reaction mechanisms are discussed in terms of their
applicability to treat specific types of contaminants. Table 11-1 presents an
overview of various treatment media considered for PRBs, the target
contaminants they address, and the current stage of development (i.e.,
laboratory study stage, field demonstration stage, etc.). Brief descriptions
have been provided for the types of reactions used in PRBs.
304
a. Chemical Precipitation. Precipitation of dissolved inorganic species (e.g.,

heavy metals) may be accomplished by using a slightly soluble material
containing an ion which forms an insoluble salt (such as a phosphate, sulfate,
hydroxide, or carbonate metal salts) with the contaminant. The material itself
should be nontoxic, have a solubility higher than that of the precipitate to be
formed, but sufficiently low so that the material continues to be effective for
the desired period of time.
TABLE 11-1 Treatment technologies applied in permeable reactive barriers
Treatment Medium
Target Contaminants
Technology Status
zero valent iron
halocarbons
commercially applied
zero valent iron
reducible metals (Cr+6, U)
field demonstration
limestone
metals, acid waters
in practice (mining)
precipitation agents
(gypsum, hydroxyapatite)
metals
lab studies
sorptive agents
(Fe hydroxide, GAC*, zeolites,
coal)
metals and organics
field demonstration and/or lab

studies
reducing agents
(organic compost, dithionite,
hydrogen sulfide)
reducible metals
field demonstration
metal couples**
halocarbons
lab studies
biologic electron acceptors

BTEX
(ORC***-oxygen source, nitrate)
field experiments
* GAC = granulated activated carbon

** coupled oxidation of metal and reduction of halocarbon to produce chloride and Fe+2 in solution
*** ORC = oxygen release compound
Other reactants can be added to promote precipitation of insoluble mineral

phases or to promote precipitation through adjustment of pH. Limestone
(calcium carbonate) has been used for many years to neutralize acid and to
precipitate metal contaminants; e.g., to construct reactive flow-through
treatment barriers to neutralize acid mine drainage from coal and metal ore
mining sites. Similarly, hydroxyapatite (calcium phosphate) has been used
as a source of phosphate to precipitate lead from groundwater in the form of
lead phosphate. Morrison and Spangler (1993) investigated the use of
hydrated lime as a precipitation agent in chemical barriers.
b. Oxidation-Reduction Reactions. Oxidation-reduction (redox) reactions
can be used to change the valence state of an inorganic contaminant, thus
reducing solubility and enhancing precipitation (Blowes and Ptacek, 1994).
An example involves the reduction of hexavalent chromium (Cr+6) to the
trivalent form (Cr+3), with subsequent precipitation as an oxide. Redox
reactions can involve chemical, electrochemical, and biochemical removal
processes. Treatment media which have been applied include such reductants
as zero-valent metals, hydrogen sulfide, sodium dithionite, and degradable
305
biomass.
Blowes, et al. (1995a) and Thomson (1995) reported on the use of biomass
to promote redox removal of inorganics. Biological reduction of sulfate to
sulfide by sulfate-reducing bacteria can be used to remove metals from mine
tailings water through precipitation as insoluble metallic sulfides. Also, a
biological approach can be used to remove nitrate from groundwater through
the process of denitrification. Generally, these processes employ the addition
of a substrate to the PRB that promote biodegradation using indigenous
bacteria.
Redox reactions also can be used to convert organic contaminants to
nontoxic by-products. For example, the strong oxidant potassium
permanganate has been investigated as a possible oxidizing agent for
remediation of chlorinated hydrocarbons (Farquhar, 1992).
c. Zero-Valent Metal Dehalogenation. Zero-valent metals (e.g., granular
iron) can be used to promote reductive dechlorination of chlorocarbons, e.g.,
PCE and TCE (Gillham, et al., 1995). The zero-valent iron reaction involves
oxidation (corrosion) of the metal by the chlorocarbons, with the metal serving
as a source of electrons for chlorocarbon reduction. The reduction step removes
chlorine atoms from the chlorocarbon molecule, releasing chloride and ferrous
iron (Fe+2) into solution. The process appears to be surface-mediated, meaning
that direct contact with (and possibly sorption onto) the iron surface is
necessary for the reaction to take place, in that the rate of reaction appears to
be directly proportional to the surface area of granular iron present. The endproducts of PCE and TCE dehalogenation are primarily ethene and ethane.
However, a small proportion of the initial PCE/TCE can appear as partially
dechlorinated daughter products (i.e., cis 1,2-dichloroethene and/or vinyl
chloride) if reaction time is insufficient. Thus, a primary design goal using
reactive iron is to size the PRB to ensure complete dehalogenation. Table 112 (Gillham, et al., 1995) provides a summary of contaminants that have been
tested with granular iron, along with reported degradation half-lives based
on iron surface area normalized to 1.0 square meter per milliliter of solution.
Various enhancements to the standard granular iron degradation process
are being examined. Orth (1995) investigated the use of iron plated with
metals having higher reduction potentials (e.g., copper) and found
significantly greater rates of dechlorination compared to iron alone, apparently
due to the corrosion-inducing effect of the plated metal. Korte, et al. (1995)
had similar results using palladium plated on granular iron (palladized
iron). In addition to increasing the rate of dechlorination, chlorinated
daughter products were not detected. The addition of sulfur-containing
compounds such as pyrite may help to stabilize the iron surfaces, thus
prolonging reactivity (Holser, et al., 1995).
d. Biological Degradation Reactions. Modifying redox conditions can
increase the rates of biodegradation of some common aromatic hydrocarbons,
306
such as benzene, ethylbenzene, toluene, and xylenes (BETX). These

compounds generally are more readily biodegraded under more oxidized
conditions than under highly anaerobic conditions (Wilson and McNabb,
1983). Modifying redox conditions involves the addition of a suitable electron
acceptor (nitrate or oxygen). The rate of biodegradation can be enhanced by
supplying the PRB with rate-limiting nutrients, e.g., nitrogen and phosphorus.
TABLE 11-2 Compounds tested and half-lives normalized to 1 m2 iron surface per
ml solution (from Gillham, 1995)
Pure Iron
t 1/2 (hr)
Organic Compound
Methanes
Ethanes
Ethenes
Other
No Apparent
Degradation
a
b
c
d
Commerical Iron
t 1/2 (hr)
Carbon Tetrachloride
0.20a, 0.003f, 0.023h
0.31-0.85b
Chloroform
1.49a, 0.73f
4.8b
Bromoform
0.041a
Hexachloroethane
0.013a
1,1,2,2-Tetrachloroethane
0.053a
1,1,1,2-Tetrachloroethane
0.049a
1,1,1-Trichloroethane
0.065a,1,4g
1.7-4.1b
Tetrachloroethene
0.28a,5.2g
2.1-10.8b
Trichloroethene
0.67a,7.3-9.7f
1.1-4.6b,2.8e
1,1-Dichloroethene
5.5a,2.8g
15.2e
trans 1,2-Dichloroethene
6.4a
4.9b,7.6e
cis 1,2-Dichloroethene
19.7a
10.8-33.9b
Vinyl Chloride
12.6a
10.8-12.3b,4.7
1,1-2-Trichlorotrifluoroethane
1.02b
(Freon 113)
1,2,3-Trichloropropane
24.0c
1,2-Dichloropropane
4.5c
1,3-Dichloropropane
2.2c
1,2-Dibromo-3-chloropropane
0.72b
1,2-Dibromoethane
1.5-6.5b
n-Nitrosodimethylamine
(NDMA)
1.83b
Nitrobenzene
0.008
Dichloromethane
af,g
, 1,4-Dichlorobenzeneg, 1,1-Dichloroethaneg,
1,2-Dichloroethane , Chloromethaneb
Gillham and OHannesin (1994)

Unpublished Waterloo data
Focht (1994)
Agrawal and Tratnyek (1994)
e
f
g
h
Mackenzie, et al. (1995)

Matheson and Tratnyek (1994)
Schreier and Reinhard (1994)
Lipczynska-Kochany, et al. (1994)
307
Kao and Borden (1992) proposed the use of a two-layer PRB, composed
of a nutrient-laden (nitrogen and phosphorus) concrete briquette layer
followed by a layer of peat. Nitrate is released as the contaminated water
passes around the nutrient briquettes, thus stimulating BETX degradation
through denitrification.
Bianchi-Mosquerat, et al. (1993) describe the use of a solid-phase oxygen
release compound (ORC) to enhance degradation of BETX compounds. Devlin
and Barker (1992) proposed the use of a semi-passive PRB to facilitate nutrient
addition for stimulating bioremediation processes. They envisioned a PRB
with a hydraulic conductivity substantially greater than the surrounding
geologic material, through which a nutrient-amended groundwater stream
would be periodically circulated using injection and withdrawal wells. The
injected amended groundwater stream would then be carried downstream
of the wall by the natural groundwater flow. Periodic injection of the
amendment to the PRB would result in a series of pulses that, while migrating
downstream, would mix (by dispersion) with the contaminated groundwater,
blending to form a continuous treatment zone of nutrient-amended
groundwater.
e. Sorption Reactions. Sorption reactions may be divided into hydrophobic,
hydrophilic, and ion exchange types. Many materials have been investigated
and/or used to sorb dissolved organic and inorganic species from aqueous
solutions. Since sorption reactions typically equilibrate in a short period of
time, sorption media are generally well-suited for application in PRBs.
However, sorption media have finite capacity to sorb; therefore, contaminant
breakthrough occurs when this capacity is exceeded. Thus, in order to avoid
eventual breakthrough of the contaminant, a means to remove and replenish
the treatment medium must be provided in a sorption PRB system.
f. Sorption of Organics. Sorption may be well suited for strongly sorbing
organic compounds that have relatively low water solubilities, hydrophobic
character, and are not easily amenable to biodegradation (such as polynuclear
aromatic hydrocarbons). These compounds are known to partition from the
water phase to the solid phase organic carbon of geologic materials. Hence, a
possible approach for the in situ removal of these contaminants from
groundwater might be to increase the organic carbon content of the aquifer
material in the path of the contaminant plume.
Teutsch and Grathwohl (1995) discuss the application of permeable
sorptive walls for treatment of hydrophobic organic contaminant plumes.
Potential materials applicable to sorption of these organics include: granular
activated carbon, peat, coal, and organic-rich shales. Again, attention needs
to be given to the sorptive capacity of the selected treatment material.
Modification of sorptive materials by surfactant application can increase
sorption capacity and selectivity. The capacity of a porous medium to sorb
hydrophobic organic solutes may be enhanced by injecting a cationic surfactant
308
solution into the subsurface. The ionic end of the surfactant molecule is sorbed
by the mineral surfaces, while the organic solutes are sorbed at the
hydrophobic end of the surfactant molecules.
Immobilized (sorbed) organics may be subsequently biodegraded through
the upstream injection of nutrients (Teutsch and Grathwohl, 1995). This
combined treatment approach might be particularly effective for organics that
are readily biodegradable, such as aromatic hydrocarbons (e.g., BETX).
g. Sorption of Inorganics. Inorganics, e.g., metals, are well suited to
hydrophilic and ion exchange sorption reactions. Materials suitable for sorbing
metals include organic carbon, zeolites, aluminosilicate clays, iron
oxyhydroxides, and other mineral materials.
The most important factors affecting metal adsorption onto organic carbon
are: the carbon type, pH of the contaminated groundwater, and surface loading
rate. The adsorption reaction occurs between the metal ion and their hydroxy
species with the carbon surface to form hydrogen bonded surface complexes.
Sorption of metals on carbon follows the Irving-Williams order of complex
formation as follows: Pb > Cu > Ni > Zn = Mn = Cd = Co; where all the metals
are in the two valence state.
Zeolites have been widely used and studied as ion exchange media. Their
ion exchange capacity depends on the substitution of alumina for silica in
parts of the zeolite structure, resulting in a net negative charge on the mineral.
Sodium, calcium, potassium, and other exchangeable cations balance the
charge by occupying the open channels in the structure. Larger cations, with
low energies of hydration, form the strongest bonds in the zeolite structure.
Factors affecting the sorption of metals by zeolites include the initial form of
the zeolite, (e.g. sodium saturated), the ionic radius and energy of hydration
of the sorbing cation, and the presence of complexing agents in solution.
Unmodified zeolites sorb only cationic metals (i.e., oxyanionic forms such as
arsenate and chromate are not sorbed) and the sorbed metals may be displaced
with concentrated sodium solutions. Clinoptilolite, chabazite, and other
zeolites have been used to sorb inorganic contaminants.
The use of organo-zeolites (formed by reacting natural zeolite with a
cationic quaternary amine surfactant) has been investigated for the sorption
of inorganic oxyanions, such as chromate, selenate, and sulfate (Haggerty
and Bowman, 1993). The precipitated organo salt is stable in the hydrophobic
environment created by the surfactant. Sorption of TCE by a synthetic
hydrophobic zeolite has been investigated by Alvarez-Cohen, et al. (1993)
using a two stage process in which the hydrocarbon is first sorbed onto the
zeolite and subsequently desorbed and biodegraded.
While aluminosilicate clays are suitable ion exchangers and are known
to sorb organics and heavy metals, their small particle size (typically less than
five microns) and associated low hydraulic conductivity excludes them for
use in PRBs.
Iron, aluminum, and manganese oxyhydroxides can adsorb divalent
309
metals such as cadmium, cobalt, copper, lead, and zinc. Adsorption of heavy
metals onto oxyhydroxides depends on pH, (i.e., sorption increases with
increasing pH). The affinities of metal ions for the oxide surface decrease in
the order: Cu > Pb > Zn > Co > Cd. Morrison and Spangler (1996) report on
the laboratory study of amorphous iron oxyhydroxides (AFO) to sorb a wide
variety of metals, including the oxyanions chromate and arsenate.
Oxide surfaces can be modified by polyphosphates to enhance sorption.
The polyphosphates do not form insoluble metal salts, but they will sorb to
the oxide surfaces. Once sorbed, they may enhance the sorption of metals by
forming metal-polyphosphate-surface complexes. Excessive amounts of
polyphosphate (beyond sorption capacity) will cause desorption of metals
from the oxide surface.
Sorption of organics by organo-clay complexes has potential application
in low organic carbon aquifers. Aquifer mineral phases can form organomineral complexes that increase the adsorption capacity of the aquifer
materials. The organo complexes are formed by reaction of the aquifer solids
with various surfactants. Surfactant micelles adsorbed to the aquifer solid
surfaces attract dissolved organics. The strongly bound surfactants can
substantially increase the sorption capacity of the aquifer materials.
TABLE 11-3 Summary of reported permeable reactive barrier materials for inorganic
contaminants* (from Morrison and Spangler, 1995)
Exp.
No.
Contaminant
Material
o
Maximum Capacity
(mol/mol)
-4
Reference
Cr
Fe
5 x 10 mol Cr/mol Fe
Cantrell, et al. (in press)
Cr
Feo
3 x 10-6 mol Cr/mol Fe
Blowes and Ptacek, 1992
Cr
Fe
-2
Powell, et al. 1995
-3
Zachara, et al. 1987
-4
Haggerty and Bowman, 1994
Cr
AFO
Cr
HDTMA
Zeolite
Cr
H 2S
1 x 10-2 mol Cr/mol S
Thornton and Jackson, 1994
Cr
BrY bacteria
1 x 10-6 mol Cr/106 Cell
Gorby, et al. 1994
Feo
1 x 10-4 mol U/mol Fe
Kaplan, et al. 1994
AFO
1 x 10-4 mot U/ mol Fe
Morrison, et al. 1995
10
Peat
2 x 10-3 mol U/mol C
11
Ca(OH)2
1 x 10-4 mol U/mol Ca
12
Mo
Feo
3 x 10-4 mol Mo/mol Fe
Cantrell, et al. (in press)
13
Pb
Phosphate
rock
2 x 10-2 mol Pb/mol P
Ma, et al. 1995
14
Sr
Zeolite
3 x 10 mol Cr/mol Si
-3
9 x 10 mol Sr/mol Si
Cantrell
*Some estimates of chemical compositions (e.g. of peat) and porosity were made.
Table 11-3 (from Morrison and Spangler, 1995) presents a summary of

PRB materials suitable for inorganic contaminants, including several sorption
materials. Removal capacities (based on lab studies) are listed for the various
310
materials. The reported removal capacities range over several orders of

magnitude, and should be used with caution. For example, the values of
chromate retention in experiments 1 and 2 represent minima because the
granular iron capacity had not been depleted in the experiments. In addition,
solution composition (e.g., hardness, pH, concentration of competitive ions)
will affect the degree of sorption. For these reasons, PRB designs should be
based on treatability studies using the sorptive media and groundwater from
the site.
11.2.3 System Design Protocol
a. Site Characterization. PRB systems are designed by focusing on the
factors that most affect their performance. Characterization of the remediation
site is an essential element in the system design protocol. (Since site
characterization is treated elsewhere in this report, discussion here is limited
to PRB applications.) In many respects, the considerations given to the
placement of a low hydraulic conductivity barrier are also applicable to the
placement of a PRB. Since it is desired that all of the contaminated groundwater
flow through the PRB, careful attention needs to be given to development of
the groundwater flow model for the site. The groundwater flow model enables
evaluation of: 1) the effectiveness of various PRB configurations to intercept
the contaminated groundwater flow and 2) the need for hydraulic control
augmentation.
A hydrogeologic conceptual model is developed early in the process to
aid in making large scale decisions concerning the optimal deployment of a
PRB. Simply put, a conceptual site hydrogeologic model consists of defining
the basic dimensions of the contaminant plume and defining the rates and
directions of plume movement within the context of the site stratigraphy.
The conceptual model helps in developing the initial concept of how a PRB
would best be used and also serves as a basis for initial assessment of technical
and cost feasibility. In addition, the model serves as a framework for
developing more detailed information on the site which is needed during the
final design.
The most obvious constraints to emplacement and effective treatment
using a PRB are the geologic setting of the contamination (i.e., aquifer type)
and the depth of contaminated groundwater. The PRB typically will be
constructed through the entire thickness of a contaminated aquifer. Ideally,
the barrier would be keyed into an underlying low hydraulic conductivity
layer, e.g., a clay aquitard or sound bedrock. For buoyant or sufficiently
shallow plumes, a hanging wall system may be considered.
Once the preliminary conceptual model is completed and a PRB concept
(or set of PRB options) developed, a detailed geologic investigation is
conducted in the immediate vicinity of the proposed PRB. Geologic
parameters to be determined include lithology, stratigraphy, grain size
distribution, and structural relationships. A set of geologic cross sections is
311
developed for the PRB site showing details of the stratigraphy at least as deep
as the first aquitard beneath the contaminant plume.
The site groundwater gradient, flow direction, hydraulic conductivity,
and water balance are developed using standard methods. Uncertainties in
these parameters can be particularly significant in the design of a PRB.
Measurements of piezometric head may not be a sufficiently reliable basis for
estimating of flow direction due to small head differences displayed by wells
located in close proximity. Groundwater flow directions can vary due to seasonal
recharge effects and heterogeneity and anisotropy of the aquifer, which may
cause local variations in flow rate and contaminant flux, often by an order of
magnitude or more from one point to another. Aquifer anisotropy is particularly
critical when a hanging gate PRB is being considered. PRB designs must
account for these uncertainties. Alternatively, the designer may build in physical
features or safety factors to compensate for these uncertainties.
Delineation of the plume boundaries and the location of zones with high
concentration are important factors in determining if a PRB will be a costeffective remedial technology. The spatial distribution of the contaminant must
be determined prior to design, as well as contaminant properties (solubility,
vapor pressure, specific gravity, partitioning, etc.), including their chemical
relationship to site geology and geochemistry.
Groundwater flow models are useful in optimizing the PRB design. Twodimensional plan view simulations are appropriate for PRBs that penetrate the
entire thickness of relatively homogeneous aquifers. Systems that extend only
partially through an aquifer (hanging wall systems) are best described by three
dimensional simulations, since low hydraulic conductivity lenses can
significantly affect PRB performance, potentially limiting overall system
efficiency.
Geochemical characterization of site groundwater and host aquifer materials
is critical to proper evaluation of PRB performance. Naturally occurring
groundwater constituents may compete or interfere with sorption and
precipitation reactions. Chemical reactions between treatment reactants,
groundwater contaminants, and the natural groundwater constituents may also
cause formation and reaction zone plugging. Chemical interactions of the barrier
reactants with natural groundwater constituents can cause barrier aging and
failure. For example, amorphous ferric oxyhydroxide undergoes a
transformation to the crystalline goethite form of iron oxide, occurring over a
period of weeks to years. Both iron species will adsorb contaminants from
groundwater but at different reaction rates. In addition, aquifer materials can
buffer groundwater equilibrium reactions in ways that are unexpected and
deleterious to the PRBs performance.
b. PRB Hydraulic Control Systems. Basic types of PRB hydraulic control
systems that have been studied for application to contaminant plume control
include: funnel and gate PRBs, continuous wall PRBs, and injected treatment
zone PRBs.
312
Funnel and Gate PRBs. Funnel and gate PRBs (Starr and Cherry, 1994; Smyth,
et al., 1994; Smyth, Cherry, and Jowett, 1994) utilize an impermeable barrier
(funnel) placed in the path of the contaminated groundwater flow (see Figure
11-1). The funnel shape guides the flow through the permeable reactive zone
(gate). Types of impermeable barriers currently used in practice include slurry
walls, sheet piles, or soil admixtures formed by soil mixing or jet grouting.
The function of the funnel is to widen the capture zone by diverting the
groundwater flow through the relatively small permeable gate. The presence
of the impermeable barriers alters the local piezometric head distribution
which in turn controls the groundwater flow in the vicinity of the funnel and
gate. The funnel must be placed to sufficient depth and width to produce a
zone of capture which encompasses the entire plume and directs it through
the permeable reactive gate. Keying the bottom of the funnel and gate into a
lower aquitard is the typical concept considered.
(a) Single Gates
(b) Multiple Gates

Treatment cells
Treatment cell
Contaminant
plume
Cleaned
Water
Contaminant
plume
Waterloo Sealable
Joint Sheet Piling
Waterloo Sealable
Joint Sheet Piling
(c) Multiple Cells

Treatment cells
Cleaned
Water
Contaminant
plume
Waterloo Sealable
Joint Sheet Piling
Figure 11-1 Conceptual configurations of funnel-and-gate systems in plan

view: 11-4a shows a simple system with a single-cell gate; 11-4b
shows a system with three single-chamber gates in parallel; and
11-4c depicts a system with a sequence of treatment steps
incorporated in the gate. (Smyth, Cherry, and Jowett, 1994)
The funnel and gate system can have one of two basic configurations: (1)
retrievable, or (2) non-retrievable. The retrievable system consists of the
following:
a porous structural member with possible surrounding gravel pack
designed to separate aquifer soils from the treatment medium yet allow
313
transmission of water through the media (e.g., a well screen with gravel
pack), or
porous treatment medium contained in a removable container such as
a filter fabric bag, wick drain, or metal cassette.
The possible advantage of a retrievable gate is that the treatment medium
can be removed and replaced, once its useful life has been reached. However,
care must be given to ensuring that the effective hydraulic conductivity of
the retrievable gate is comparable to that of the surrounding aquifer. Excessive
head losses in the vicinity of the gate will raise local piezometric head levels,
resulting in the potential diversion of contaminated groundwater flow around,
over, or under the gate. In addition, such systems have yet to be tried at fullscale, and retrievable/replaceable materials could prove difficult in practice.
Non-retrievable reactive zone gates place the reactive medium in direct
contact with aquifer soils or with a filter medium (e.g., gravel pack) between
the reactive medium and aquifer soils. One disadvantage of this approach
may be a reduced capability to replace the reactive medium if the design life
is not sufficient for long-term plume control. It remains to be seen if this is a
true limitation, since various methods of replacing or regenerating materials
in place can be envisioned (e.g., soil mixing, vacuum trucks, regenerative
reagent flushing, etc.).
In a funnel and gate system, the resultant zone of capture may be small in
comparison to the required funnel dimensions. The following are potential
concepts to enhance the zone of capture of funnel and gate systems (Phifer,
1995):
multiple funnel and gate systems placed in series,
downgradient pumping to steepen the hydraulic gradient and increase
groundwater flow through the treatment gate,
creation of enhanced permeable zones parallel to the impermeable
barrier on both the up- and downgradient sides of the barrier, and tied
into the treatment gate.
The use of downgradient pumping to increase the hydraulic gradient
through the gate has the disadvantage of bringing groundwater to the surface
which will require handling. Since the water theoretically will be clean,
reinjection upgradient of the funnel and gate may be acceptable. However,
routine operation will be required, so that the system is no longer strictly
passive (a definite disadvantage).
Starr and Cherry (1994) provide a comprehensive modeling study of
various alternative funnel and gate systems and offer guidance on optimizing
the design of funnel and gate systems.
Continuous Wall PRBs. As described by Phifer (1995), a continuous wall PRB
consists of the placement of a permeable reactive zone perpendicular to the
314
path of the contaminant plume across its entire depth and width. Reactive
media is placed in direct contact with aquifer solids or a gravel pack may be
placed between these elements. As discussed in a later section, continuous
wall PRBs could be emplaced by excavation/backfill or soil displacement
methods.
The continuous wall has the conceptual advantage of creating less
disturbance to the natural groundwater flow pattern than the funnel and
gate system (described above). Also, the system may be more easily designed
to achieve the long-term life requirements for addressing groundwater plumes.
The following considerations are important factors in the selection of a
continuous wall PRB versus a funnel and gate system:
up-front cost of the treatment media compared to the on-going costs of
replacing treatment media in a funnel and gate system,
irretrievability of the treatment media; a new wall would need to be
emplaced if plume life exceeded the reactive life of the continuous
wall PRB,
volume of contaminated soils excavated during PRB emplacement
(depending on the emplacement method selected and ability to replace
soils on-site),
potential formation disturbances resulting from PRB emplacement
methods.
Injected Treatment Zone PRBs. An alternate form of PRB involves the injection
of the reactant into the soil or aquifer to form a reactive treatment zone. The
reactant may be delivered into the aquifer as a dissolved phase by injection
through wells. The injected solution then reacts with and coats the aquifer
solids, thereby creating the PRB. For example, Burris and Antworth (1992)
discuss the formation of a sorbent PRB within an aquifer by injection of an
aqueous solution of cationic surfactants through a set of injection wells located
downgradient from a contaminant source. Morrison and Spangler (1993, 1995)
proposed the injection of a ferric chloride solution into calcite-bearing aquifers
to precipitate amorphous ferric oxyhydroxide on the aquifer solids as a sorbent
PRB. Alternatively, the reactive media may be injected as a slurry (suspended
solid phase media) through the process of hydraulic fracturing.
A potential advantage of this approach over the funnel and gate and
continuous wall PRBs is the ability to more cost-effectively emplace materials
at depths greater than are accessible by standard excavation methods. A
potential disadvantage is the relative lack of control during the emplacement
process. These issues are discussed in more detail later in this section under
Emplacement Methods.
Miscellaneous Factors Affecting Hydraulic Performance. Designing the PRB
to Address Non-Uniform Flow. Nonuniform flow can result when regional
groundwater flow directions are not perpendicular to the orientation of the
315
treatment wall (Gallinatti and Warner, 1994). Also, heterogeneities in the

aquifer and/or in the reactive barrier itself can result in nonuniform flow.
The implications of nonuniform flow in the vicinity of the permeable barrier
are serious, since the treatment of the contaminated groundwater flow may
also be nonuniform.
Groundwater flow that is not perpendicular to the orientation of the
permeable barrier may cause groundwater flow to converge within the reactive
zone causing local increases in velocity. Heterogeneities may create
preferential flow paths that also result in locally high groundwater velocity
zones. Two methods for reducing the potential effects of nonuniform flow
are: 1) extending lateral barriers downgradient from the permeable wall to
create a downgradient chute that causes flow convergence downgradient
rather than within the treatment zone, and 2) sandwiching the treatment zone
between zones of very high hydraulic conductivity material; the nonuniform
flow components tend to occur within the higher hydraulic conductivity zones
rather than within the treatment zone (Warner, et al., 1995).
Effect of Reactor Hydraulic Conductivity. The hydraulic conductivity of the
permeable reactive barrier should be as high as possible so as to minimize
local groundwater head buildup upstream of the reactor. However,
maximizing reactor conductivity must be tempered by the kinetics of the
reaction mechanism, whose rate may be dependent on surface area
considerations (generally low for large grain-size porous media). Discharge
through the reactor increases with increasing hydraulic conductivity, but there
is relatively little increase for conductivities greater than 10 times that of the
aquifer. Thus, selection of a properly graded porous reactive media is an
important consideration in balancing hydraulic conductivity and reaction
kinetics. In general, a poorly graded (well-sorted) granular backfill is likely
to be best in practice.
Enhanced Permeable Zones. In the funnel and gate PRB system, enhanced
permeable zones upstream and/or downstream of the gate can be used to
direct the flow of groundwater more efficiently through the reactive zone,
thereby increasing the size of the capture zone. Enhanced permeable zones
can also be used to redistribute groundwater downstream of the gate to
alleviate hydraulic head buildup in the funnel area upstream of the gate (Phifer,
1995). Enhanced permeable zones can also be used to reduce the need for
more elaborate funnel configurations and to reduce the effects of subsurface
heterogeneity.
c. PRB Reactor System Design. The major elements of reactor system design
include: (1) selection of the reactive medium, including the chemical makeup,
particle size distribution, and proportion of admixtures, if needed; (2) setting
the dimensions of the reactor to provide the necessary retention/reaction time;
and (3) sizing the reactor to provide for adequate treatment over the required
design lifetime. Treatability studies (described briefly below) using the
candidate reactive media and site groundwater form the basis for primary
316
design. However, a number of other related activities should be considered

in the reactor design, including: (1) examining the potential effects of flow
dispersion in the reactor, and (2) applying a groundwater flow model to
examine the effects of the proposed PRB on piezometric heads and gradients.
Finally, evaluating and mitigating potential adverse effects of the PRB reactions
on downgradient groundwater chemistry should be an integral part of the
reactor design.
Determining Reactor System Residence Time. The relationship between the
residence time of contaminated groundwater in the PRB and the rate of the
contaminant degradation reactions is of critical importance in designing the
PRB. The necessary residence time depends on the influent contaminant
concentration, the required effluent concentration, and the reaction rate, which
in turn may depend on the seepage velocities through the PRB.
In simplest terms, the mean residence time in the PRB reactor can be
calculated by dividing the pore volume of the PRB by the discharge through
the PRB. For degradation processes that can be described in terms of firstorder reaction kinetics, the residence time necessary to reduce concentrations
a given amount can be calculated using:
N1/2 = [ln (C effluent / C influent)] / ln (1/2)
where Ceffluent is the concentration at the effluent, or downstream side of the
reactor; Cinfluent is the concentration at the influent, or upstream side of the
reactor; and N1/2 is the number of half-lives required.
Reduction in effluent concentrations can be achieved by either increasing
reaction rates or lengthening retention times. In some cases, it may be possible
to increase the reaction rate by increasing the surface area of the reactive
material or by using a higher proportion of reactive material in the reactive
porous mix. Alternatively, longer residence times can be obtained by
decreasing the discharge through the reactor, which causes a corresponding
reduction in the size of the capture zone, or by increasing the size of the
reactor.
The reactor volume can be increased by making the reactor wider (or the
wall thicker). In a funnel and gate design, gate thickness (in the direction of
flow) has a minor effect on discharge through the gate, so retention time
increases almost linearly with gate thickness. The ultimate manifestation of
increasing gate width is the resulting continuous wall PRB design, where the
wall width exceeds the plume dimension. At this point, further increases in
the reactor width causes no further increase in retention time. Long residence
times may also be achieved by completely surrounding a contaminant source
zone with cutoff walls, except for a gap that contains a PRB in the downstream
wall. In general, capture zone size and residence time are inversely related.
It should be noted that plumes are highly variable in concentration (by
orders of magnitude) and often occupy heterogeneous media. For this reason,
317
calculation of average residence time across a plume/PRB system probably

is not of much value for design. Of greater relevance to design is the worst
case combination of plume concentration and groundwater velocity, yielding
the greatest mass flux of contaminant which must be addressed by the PRB.
Treatability Studies for Reactor Design. Typically, reactor system design is
based on results from treatability studies which help in estimating the reaction
rates achievable in the field. Standard treatability test protocol involves batch
reaction tests and lab or field flow-through column tests. At this stage in PRB
development, no standard treatability test protocols have been developed.
However, a number of references contain test methodology applicable to
various treatment approaches (for example, Gillham and OHannesin, 1993;
Matheson and Tratnyek, 1994; Burris and Antworth, 1992; Xu and Schwartz,
1992; Morrison and Spangler, 1993).
Prior to initiating the treatability studies, a concept of the reaction process
should be known. The concept should consider the type of reactions involved,
the sites where various reactions take place (i.e., on reactive particle surfaces;
in solution, etc.), and the likely kinetic rate limitations. One also should
consider whether certain geochemical factors may affect reactivity, such as
dissolved oxygen concentration, pH, Eh, temperature, and/or concentration
of various anions and cations that may compete for reaction sites. Also, the
model should consider the effect of initial contaminant influent concentration
on the rate of reaction.
In general, to ensure representative results, laboratory treatability studies
in support of PRB design should be conducted with actual groundwater taken
from the application site or, if necessary, with a synthesized contaminated
groundwater that has been matched to the site. When taking field samples, it
is often not possible to avoid a degree of sample disturbance in the course of
transport to the lab. For example, dissolved oxygen incursion, Eh and pH
shifts, and temperature variation may cause the sample to be less than
completely representative of the in situ groundwater. The reaction process
concept can be used to predict the potential unrepresentative elements and
help in interpretation of results.
The designer should also consider the potential for competing pathways
or mechanisms of degradation. Of special concern is the potential for alternate
degradation pathways to form intermediates that are more toxic than the
contaminants originally of concern (Burris, 1995).
Batch reaction tests involve reacting a known quantity of contaminated
groundwater with a known quantity of reactant in a closed container. The
container typically is agitated to facilitate contact between groundwater and
reactant. A series of tests may be performed in which the proportion of
groundwater to reactant is varied and reaction times are also varied. Initial
and final contaminant concentrations are compared to evaluate important
measures of reactivity (e.g., degradation half life, sorption isotherms, etc.)
useful for reactor design. Reactivity is expressed in general terms (such as
318
mass of contaminant sorbed per mass of reactant).

Flow-through column tests have been applied both in the laboratory and
in the field in support of PRB design. These tests are conducted by packing a
column with the reactive media and passing the contaminated groundwater
through the column. Influent and effluent concentrations are recorded to
evaluate the effectiveness of treatment. Column tests can be set up to directly
model the flow-through process in a PRB. Flow rate can be chosen to simulate
groundwater velocity and reactor residence time, and the column length can
be scaled to mimic the reactive wall dimension. In some cases, a number of
researchers have devised field columns or canisters that can be operated at
the site, circulating actual contaminated groundwater pumped from wells in
the plume. A number of such ex situ tests are discussed in Section 11.3.
Geochemical interactions between the reactant and the natural
groundwater system can affect the performance of PRBs. Buffering can affect
the solubility of solid and aqueous species. Precipitation of minerals can cause
plugging at the entrance of the reaction zone, porosity reduction within the
reaction zone, thereby inhibiting the performance of the PRB (Mackenzie et
al., 1995). One purpose of performing column tests is to identify such potential
problems in advance.
Detailed micro-scale characterization of reactants, as an adjunct to
standard treatability testing, may contribute to the understanding of the
mechanisms responsible for the degradation of organics and/or fixation of
inorganics (Paulson and Petrie, 1995). Tools available to perform the detailed
characterization of reactant materials include: optical light microscopy,
scanning electron microscopy, electron probe microanalysis, and x-ray
diffraction analysis. Since oxide layers and reaction rims on the surfaces of
the barrier materials can have significant influences on reaction rates, microscale characterization studies also provide information on PRB long-term
performance factors such as precipitate formation, effective barrier life, and
potential plugging problems.
Effects of Flow Dispersion in the Reactor. Emplacing mixtures of reactive
materials that are chemically and physically homogeneous is difficult. As
reactions occur, mixtures that were homogeneous initially can develop
preferential flowpaths over time. For these reasons, predicting the long-term
efficiency of contact between contaminated groundwater and reactive material
is difficult. Diffusional processes or flow in secondary and higher-order
channels may be sufficient to utilize the majority of reaction sites, but mass
transfer rates may become too slow to efficiently reduce contaminant
concentrations. Since the performance of PRBs is scale dependent, accurate
assessment of reaction efficiency is best derived from field data. Unfortunately,
performance of a field pilot in advance of every PRB project is not necessarily
practical or economically justified. As an alternative, safety factors (e.g.,
increasing PRB wall thickness to increase residence time beyond the design
requirement) can be included in the design as a way to account for
319
uncertainties such as dispersion.

Effects of the Reactor Medium on Groundwater Quality. Barrier reactants
may have a deleterious effect on downstream groundwater quality. For
example, the use of hydrated lime to precipitate metals is likely to produce
high calcium concentrations and increase the pH of the groundwater. The
interaction of some reactive materials can also affect the indigenous microbial
consortia. Incomplete reactions may produce reaction degradation products
that are more toxic than the original contaminants. Such effects and their
implications should be considered in the course of developing a PRB design.
Life Span Limitations and Barrier Fate. In sorption barriers, contaminants
will continue to migrate through the sorption zone until released at the
downgradient side. This process, called the chromatographic effect, will limit
the life-span of the barrier. Precipitation barriers have a similar limitation in
that some contaminant-containing minerals will redissolve after the reactive
material or plume is expended. These limitations can be overcome by
emplacing greater amounts of reactive material than required throughout the
compliance period. The barrier material could be removed and disposed or
the metals reprocessed; or the contaminants could be stabilized in place (e.g.,
by aging) so that they would no longer be mobile. If the risk management
goals for the site can not be achieved by managing the contaminants in place,
it may be necessary to excavate them or remove them by in situ leaching
methods (Morrison and Spangler, 1995). For PRBs designed to degrade
contaminants, these issues are less important.
The ultimate fate of sorption/precipitation barriers is a regulatory issue
that may be addressed by using risk-based versus cost considerations. In
some cases, the contaminants in the barrier may be transferred above ground
for recovery or disposal. In other cases, the reactive material may be retrievable
and the entire matrix recovered. The PRB may be stabilized so that the
contaminants are permanently fixed within the reactive material. When
appropriate, the barrier may be left in place strictly on the basis of a riskbased exposure assessment. However, if stabilized contaminants are left in
place, certain states (e.g., Texas) may require some form of deed recordation
under current rules.
11.2.4 Constructability and Emplacement Methods
As described in the previous section, treatment of contaminated groundwater
using permeable reactive barriers involves the capture, treatment, and release
of the treated groundwater. The contaminant plume must be captured and
hydraulically controlled so that it passes through the permeable reactive zone.
Following treatment in the permeable reactive zone, the treated groundwater
is returned to the aquifer. Proper construction and reactive zone emplacement
methods are critical to achieving these objectives. Among the potential pitfalls
320
of the construction process are: (1) aquifer formation disturbance and/or

reactive zone pluggage, resulting in impeded flow through the permeable
reactive zone; (2) emplacement limitations resulting in incomplete plume
capture and potential flow bypass; and (3) production of large volumes of
excavated materials (i.e., soil, debris) which may be contaminated and require
treatment or disposal.
There has been relatively little experience with the actual construction of
permeable reactive barriers to date. The few existing systems have been
shallow (less than 30 feet deep) and have been constructed using standard
design and construction approaches. Emplacement below a 30 foot depth is
unproven to date, although several potential approaches have been proposed.
The cost of emplacement will increase with depth and may limit the extent to
which PRB technology is put into practice.
Various methods for constructing permeable reactive barriers are
discussed in this section. These are categorized based on whether they would
be used for reactive wall emplacement (for example, a continuous wall PRB)
or to form an injected treatment zone PRB. Funnel and gate PRB construction
is not discussed specifically in this section since much of the construction
involves low hydraulic conductivity barrier wall emplacement (discussed
elsewhere in this volume). Approaches for continuous wall PRB emplacement
include:
standard excavation methods,

specialized trenching methods,
jet grouting technology,
mandrel-based technology,
vibrating beam technology, and
soil mixing.
Approaches for emplacement of injected treatment PRBs include:

fluid injection through wells, and
hydraulic fracturing.
The following discussion is intended to reflect the state-of-the-art as
currently applied in the field or being considered for near-term application.
a. Continuous Wall PRB Emplacement. Standard Excavation Methods. After
excavation of a trench using normal construction methods, the trench is
backfilled with the permeable treatment media. A geotextile filter fabric may
be placed adjacent to the trench wall prior to backfilling to control invasion of
soil fines into the treatment media. Various methods have been devised to
maintain an open trench until backfilling is completed. Readers are referred
to Section 3 of this volume for a more complete treatment of construction
methods for installing barrier walls.
321
Modifications to the standard excavation approach have been developed

to address the need for stabilizing the trench. The EnviroWall groundwater
barrier and pass through system was used recently to construct an interceptor
trench system at the U.S. Department of Energys Savannah River Plant in
South Carolina (Phifer, 1995). This method utilizes excavators and movable
guide boxes to maintain a consistent open area in the trench until various
materials such as barrier geomembrane and permeable backfill can be
emplaced. The system allows for up to 120 feet of wall to be constructed at a
time to depths up to 40 feet. The guide boxes are such that geomembrane
barrier or geotextile filter can be placed alongside permeable backfill. Barrier
and pass through zones can be positioned to manipulate groundwater flow
to discrete openings on the downgradient side of the trench (similar to a funnel
and gate approach), while providing for treatment within the reactive media
core of the trench. The Savannah River interceptor trench was constructed to
a depth of 30 feet using this technique. Inert sand backfill takes the place of
the reactive media core in this installation. The interceptor trench is 240 feet
long and consists of two 120 foot long wing sections with a central
groundwater collection assembly toward which groundwater is funneled by
the wings.
The University of Waterloo (Gillham, et al. 1995) constructed a permeable
wall by first driving steel sheet piling to the desired depth (about 20 feet)
around the intended wall location, and excavating the soil from within the
sheet piling. The excavation was backfilled with an iron-sand (22% iron by
weight) mixture. The sheet piling was removed after backfilling was
completed, leaving a wall about 18 feet long and five feet wide.
A similar reactive wall was constructed by excavating within pre-driven
steel sheet piling (Warner, et al., 1995). The final design of the system consisted
of a 40-foot long by 20-foot deep by 8-foot wide permeable wall (four feet of
granular iron sandwiched between two-foot thick sections of pea gravel),
upgradient lateral hydraulic barriers of soil-cement-bentonite slurry walls
(about 300 feet long on one side and 225 feet long on the other), and a
downgradient steel sheet pile lateral barrier extending about 20 feet beyond
one end of the permeable wall. The permeable wall section was constructed
by driving sheet piles around the perimeter of the intended excavation, digging
out the soil with a backhoe, placing steel sheet piling in the excavation to
create three compartments, backfilling the trench compartments with
permeable material, and extracting the piles.
Standard excavator-based construction has several advantages. It utilizes
readily available equipment with proven ability to dig in a broad spectrum of
soils ranging in difficulty up to shallow weathered bedrock. In addition, since
excavation permits the operation to be viewed, treatment media can be placed
more precisely, allowing segregated placement of different treatment media,
treatment and filter media, or treatment media and barrier (sheet piling or
geomembrane), within the same trench. Limitations to this approach are that
the width of the trench is a function of the desired depth of excavation and
322
the width of the boom, stick, and bucket of the excavator and the large costs
of constructing trenches to great depths. The excavated contaminated soil
may require disposal or, alternatively, a special allowance may be needed to
return soil to the trench (under current U.S. EPA rules). Excavation may create
a large soil surface area with higher levels of personnel protection required at
sites involving volatile contaminants, resulting in higher costs and reduced
productivity.
Specialized Trenching Methods. Specialized excavation methods using
trenching machines can be adapted to emplace permeable reactive walls.
Trenching machines come in a wide variety of designs and sizes. Designs
include chain saw and circular saw devices with depth capability varying
from several feet to over 100 feet and trench widths varying from less than
one foot to several feet. There are plans by U.S. EPA and the University of
Waterloo to construct an iron-filled permeable wall using specialized trenching
methods at a pilot site near Elizabeth City, North Carolina in 1996 (Puls, et al.,
1995).
The primary advantage of a trenching machine is that it can productively
trench in a wide variety of materials ranging from soils to soft rock. Trenching
machines adapted for interceptor well construction in particular can be highly
productive since the trench shoring moves along with the cutter head.
Production on the order of one lateral foot of trench per minute to an excavated
depth of 20 feet can be achieved. High production rates may translate directly
into low cost for depths where this method is applicable. Disadvantages are
similar to the standard excavation approach.
Jet Grouting Technology. Jet grouting has been used typically for emplacing
cement grouts to form low permeability barriers and is not believed to have
been used to date for permeable reactive wall emplacement. Application to
reactive wall emplacement would require development, particularly for
identifying suitable carrier agents and nozzle configurations to convey
treatment media to the subsurface. However, the technique can be applied
selectively at discrete depths and in a variety of soils, including cobble and
debris-bearing soils. Also, it appears practical for avoiding underground
utilities by proper positioning of the drill holes. The technique has the potential
for rapidly emplacing large amounts of reactive material in a relatively short
time. Disadvantages include the relative lack of control during emplacement
and the inability to directly monitor construction.
Mandrel-Based Technology. Mandrel-based emplacement using vibratory
hammers is commonly used to install prefabricated vertical drains for soil
consolidation. These drains are constructed of geotextile fabric wrapped
around a plastic spacer to create a thin conduit for groundwater flow. The
drains are installed by inserting the drain into a steel emplacement mandrel
(a rectangular hollow steel casing). A drive shoe/anchor is attached to the
323
drain and placed over the leading edge of the mandrel. A vibratory hammer
then drives the mandrel into the soil to the desired depth. The drive shoe
prevents soil from entering the mandrel during emplacement and securely
anchors the drain. The process is repeated at the next location.
To form a permeable reactive wall, the drain can be filled with reactive
media or eliminated, with the mandrel being used simply as a tremie device.
In the latter case, the treatment zone materials are placed through the mandrel
as it is extracted thus filling the soil void. Mandrel emplacements can be
arranged in panels to form a more or less continuous permeable barrier
perpendicular to groundwater flow.
This approach was used to install treatment zones in the first phase of the
Lasagna project conducted by a consortium of Monsanto, DuPont, and
General Electric at the U.S. Department of Energys Paducah, Kentucky plant.
The Lasagna process uses a combination of in situ electrodes and in situ
treatment zones (i.e., permeable treatment panels) to remediate soils of low
hydraulic conductivity. In the Paducah pilot, two 15-foot long electrode rows
were placed ten feet apart to a depth of 15 feet. Four rows of treatment zones
were placed equally-spaced in between the electrodes. An electric field was
applied to drive TCE-contaminated water into the treatment zones by
electroosmosis. The treatment material in this case was activated carbon
contained within 18-inch wide strip drains. Mandrel size was four inches by
21 inches in outer dimension. After driving the mandrel to the desired depth
(15 feet), the carbon-filled strip drains were inserted into the mandrel cavity.
The mandrel then was pulled out leaving the drive shoe and treatment panel
in the ground. The treatment panels were staggered and overlapped a few
inches so as to approximate a continuous wall. This approach proved
satisfactory within the context of electroosmotic flow.
A Phase II test is being planned at the Paducah site, involving placement
to greater depth (about 45 feet). The mandrel approach will be used, but a
reactive iron-clay mixture will be substituted for the carbon treatment zone
material. The treatment zone mixture will be emplaced directly as a thick
slurry using the mandrel as a tremie device. The emplacement is expected to
be completed in the first half of 1996.
Advantages of this method are that virtually no waste soil is generated
and worker exposure is limited. Since conventional vertical drains have been
emplaced to depths up to 190 feet using this technology, there is good potential
for emplacing treatment media to similar depths.
Since typical mandrel sizes range in outer dimension from 1 inch x 5 inches
up to 4 inches by 21 inches, resulting wall thicknesses are relatively thin. This
may limit the applicability of the technique to PRBs requiring little wall
retention time or to sites having very slow groundwater seepage velocities.
Soil conditions must be conducive to driving the mandrel. Cobbles and debris
may cause refusal. In addition, since the soil is compacted during mandrel
emplacement, there is potential for formation disturbance. While this is not
an issue with respect to electroosmotic flow, it could be a limitation in adapting
324
the technique to a passive flow application.

Vibrating Beam Technology. The vibrating beam technology is used
commercially for emplacing interlocking geomembrane panels for vertical
containment barriers at hazardous waste sites. The technique appears to be
readily adaptable for emplacement of permeable reactive walls. As with the
mandrel emplacement technology, virtually no waste soil is created, and
worker exposure is minimized. The overlapping drive pattern allows
construction of a more or less continuous trench. Wall thicknesses are limited
by the thickness of the I-beam.
As with the mandrel emplacement method, a potential disadvantage is
that soil conditions must be conducive to driving the vibrating beam. Cobbles,
debris, or other impenetrable materials can cause the beam to meet refusal.
Furthermore, emplacement accuracy may be of some concern since the drive
shoe can encounter hard materials which can push the beam off the intended
path, particularly at depths greater than 40 feet.
Soil Mixing. The soil mixing process has been used commercially for several
years as a way to carry out solidification and stabilization of soils and sludges
in situ. This method employs soil augers of various sizes to drill into the soil
and inject and mix reagents.
Soil mixing could be used to emplace reactive material into an aquifer in
such a way as to create a permeable wall. Equipment is commercially available
(e.g., GeoCon, Inc. and others) which can penetrate weak soils up to 40 feet
with eight to 12-foot diameter augers, or up to 150 feet with three-foot diameter
augers. The soil mixing technology has been used to form soil-cement
groundwater cutoff walls by augering in an overlapping, offset pattern (Rumer
and Ryan, 1995).
A pilot-scale version of the soil mixing approach was performed recently
by the U.S. EPA and University of Waterloo at the Elizabeth City, North
Carolina test site. Relatively small diameter hollow stem augers (six-inch
inner diameter) were used with a standard geotechnical drill rig to emplaced
a series of 21 permeable reactive cylinders into a shallow aquifer in a fence
pattern. The fence posts were located on roughly 1.0 to 1.5-foot centers and
arranged in three rows of seven each, oriented transverse to groundwater
flow direction. The treatment media consisted of reactive iron filings, native
aquifer material, and coarse washed silica sand. The estimated diameter of
the cylinders was eight inches and they were installed to depths ranging from
10 feet to 26 feet below ground surface. Performance of this system is described
in section 11.3.5.
b. Injected Treatment Zone PRB Emplacement.
As discussed previously,
injected treatment zone systems differ from the reactive wall approach in the
method of emplacement and the anticipated final form of the treatment zone.
Injected systems involve injection of fluids or fluid/particulate mixtures to
325
distribute a treatment zone through a target strata (i.e., aquifer). The resulting
zone may be oriented vertically or horizontally, but generally will be placed
perpendicular to the primary groundwater flow direction. There has been
less field experience using this approach than with the permeable wall
approach. The primary means considered for emplacing injected treatment
zones are: (1) injection of liquid solutions or gas through wells with subsequent
precipitation of treatment media on the aquifer matrix, and (2) hydraulic
fracturing to delivery treatment media either as a liquid solution or as particles
entrained in a carrier fluid.
Injection Through Wells. In this method, a dissolved or gaseous reactant is
injected through a series of wells and precipitates or sorbs onto the target
aquifer matrix. The treatment zone may be distributed throughout the plume
or as a discrete zone emplaced along the plumes leading edge. The distributed
reactant can perform its treatment function either by reacting immediately
with plume constituents or by reacting over time as the plume migrates
through the treatment zone.
Several concepts have been tested in the laboratory, some of which were
described in the previous section. Rust Geotech of Grand Junction, Colorado
has developed a technique for injecting a ferric chloride solution through
wells to react with alkaline aquifer materials to form a ferric oxyhydroxide
treatment zone distributed on the aquifer matrix (Morrison and Spangler,
1995; Morrison, et al., 1996). Iron hydroxides are excellent adsorbents for
dissolved metals. A field demonstration of this approach was designed and
wells constructed at an abandoned uranium mill tailings site in the western
U.S., but the project was discontinued before injection due to regulatory
concerns.
The use of hydrogen sulfide in aqueous solution to treat mine wastes in
Wyoming has been reported (Marozas, et al., 1995). Hydrogen sulfide is used
to reduce and precipitate dissolved metals as sulfide minerals. A
demonstration by Sandia National Laboratory at Kirtland Air Force Base in
New Mexico involved injection of gaseous hydrogen sulfide through wells
into the vadose zone to precipitate metals present in soil pore water. Other
reducing agents may be used in a similar manner. A field test involving sodium
dithionite injection is being planned to reduce and precipitate dissolved
chromium at the U.S. Department of Energys Hanford, Washington site.
An injection process has been proposed to deliver a cationic surfactant
solution into an aquifer to form a sorptive barrier (Burris and Antworth, 1992).
The injected surfactant molecules would become associated with exchange
sites on the aquifer material. With appropriate arrangement of injection wells,
a continuous sorptive barrier could be created across the plume to remove
sorbable dissolved organics. Since the chromatographic effect will limit the
life-span of a sorptive barrier, it was further proposed that bioremediation
amendments be injected upgradient of the treatment zone to promote
degradation of the sorbed constituents within the zone.
326
Although there have been no definitive field trials of the injected treatment
zone approach to date, observations of naturally occurring permeable barriers
are useful in assessing the potential effectiveness of this technology. The
concept of a chemically-reducing permeable barrier has its origins from
observations of ore-forming processes. Ore deposits have formed at locations
where oxidizing, metal- and uranium-bearing groundwaters encountered a
reducing horizon in the subsurface. These ore-forming systems can be thought
of as natural analogs to injected permeable barriers.
The potential advantages of the injected treatment zone approach are:
costs should be relatively low since trench construction is eliminated; treatment
zones could be emplaced to relatively great depths limited only by the ability
to drill; worker exposure and waste generation can be minimized, since an
open excavation is not involved; and the method lends itself reasonably well
to follow-up emplacement through additional injection to address untreated
flow paths.
A principal concern with this approach relates to the questionable
reliability of injection to create a homogeneous treatment zone. Emplacing
mixtures of reactive materials to form a well distributed treatment zone may
prove difficult. Thus, predicting the efficiency of contact between
contaminated groundwater and reactive material could be difficult.
Furthermore, chemical and microbiological reactions that occur in the
subsurface are not well understood and could affect performance. This
emplacement method has a greater potential to allow untreated flow through
preferential flow paths than the constructed wall approach. Finally, injected
solution chemistry may adversely affect aquifer water quality. As an example,
injection of ferric chloride to precipitate iron oxyhydroxide will release calcium
and chloride ions which may have to be pumped out of the aquifer.
Hydraulic Fracturing. The primary purpose of hydraulic fracturing is to
enhance seepage flow in the vicinity of a well. A secondary application is to
deliver solid compounds into the subsurface (Murdoch, et al., 1991). In this
latter application, the fractures are filled with granular compounds to form
an in situ treatment zone. Fine-grained reactive iron, encapsulated sodium
percarbonate (oxygen source for biodegradation), and various sorption agents
(e.g., activated carbon) have been considered.
Since hydraulically-induced fractures generally have a horizontal to
gently-dipping orientation, flat-lying treatment zones could be created to
address vertically downward migration in the vadose zone, within confining
layers beneath contaminated aquifers, or in fractured bedrock systems.
A detailed description of the hydraulic fracturing process using fluid
injection as applied in remediation can be found in Murdoch, et al. (1991).
Major issues affecting the choice of a carrier agent for treatment zone formation
include the ability to transport the granular material into the fracture during
propagation and avoidance of agents which could impede flow through the
zone after emplacement. Biopolymer gels, such as guar gum and xanthan
327
gum, are biodegradable and have been proposed for remediation applications.
Hydraulic fracturing at shallow depth (ten to 40 feet) have created propped
fractures generally less than one inch thick and 20 to 40 feet in diameter.
No record of field application of fracturing to form treatment zones could
be found to date. However, D. Marcus of EMCON Associates (Burbank,
California), at a recent meeting of the Remediation Technologies Development
Forum, described a plan for delivery of a reactive zinc-based proppant through
fractures in a TCE- contaminated bedrock aquifer. In this case, hydraulic
fracturing is being used to open and prop existing fractures in the bedrock.
Zinc was selected for its dechlorinating ability and its relatively low density
when coated on an inert proppant core. Guar gum is planned as the carrier
agent. Field work is expected to begin at a California test site in January 1996.

Relatively few field demonstrations of permeable reactive barriers have been
undertaken to date. The earliest known application of this concept used
limestone-filled permeable barriers installed by the U.S. Bureau of Mines to
control acid mine drainage (Kleinmann, et al., 1983). However, long-term
performance of these trenches has not been well documented. In response to
the rising interest in the potential application of PRBs in remediation
applications, a handful of well-documented, small scale field experiments
have been initiated. However, one small commercial scale application (a
reactive iron wall for TCE and DCE destruction) for which results were
available was undertaken in early 1995 in Sunnyvale, California. The early
performance of this PRB has been encouraging. However, there has been
insufficient period of record to evaluate long-term effectiveness. A summary
of the documented field experience to date is given below.
11.3.1 Borden, Ontario Reactive Iron Barrier
The University of Waterloo project consists of a permeable reactive iron barrier
placed in the path of a chlorocarbon plume (TCE, PCE, and chloroform) at
the Canadian Forces Base Borden, Ontario site (Gillham, et al., 1995). The
wall, which was constructed in 1991, had been in continuous operation for
about four years at the time the workshop was held. The permeable wall was
constructed about 15 feet downgradient from the contaminant source zone in
an aquifer comprised of fine to medium sand (see Figure 11-2). Average
groundwater seepage velocity is about 0.3 feet per day with flow direction
varying seasonally by up to 30 degrees.
Laboratory column tests using reactive iron indicated half-lives of about
15 hours could be expected for both TCE and PCE. Since TCE degradation to
less than the MCL of 0.005 mg/l would require about 16 half lives (ten days),
the minimum wall thickness was set at three feet. The actual constructed wall
328
was five feet thick and 18 feet long (transverse to flow). The reactive media
backfill consisted of 22% by weight iron grindings (collected from a local
machine shop) and 78% concrete sand. The sand was included to ensure
greater hydraulic conductivity in the wall than the adjacent aquifer. In spite
of a dense sampling network (a total of 348 separate sampling points),
definitive monitoring was hampered to some degree by the narrow plume
dimension and the changing nature of the flow direction.
Bundle
Piezometer
Permeable Wall
Source
Flow Direction
1.25
2.5 m
Figure 11-2 Plan view of the Borden test site showing the contaminant source,
the reactive wall, and the monitoring network (Gillham, et al.,
1993)
Monitoring during the four years since wall construction has shown
consistent removal of about 90% for TCE and 86% for PCE. The downgradient
concentrations are, however, about three orders of magnitude greater than the
MCL. Increasing chloride concentrations (ranging from less than 10 mg/l up
to over 100 mg/l) across the wall is evidence that TCE and PCE are being
dechlorinated (see Figure 11-3). In addition, DCE isomer formation and
disappearance has been tracked across the wall. The principal DCE isomer
formed is cis 1,2-DCE at a peak concentration of 2 mg/l at a distance of 1.5
feet into the wall. The DCE concentration declines to about 0.2 mg/l at the
downgradient monitor point. No vinyl chloride production has been detected
(see Figure 11-4).
329
300
Wall
TCE
PCE
Chloride
Concentration (mg/L)
250
200
150
100
50
0
10
11
12
Distance Along Flow Path (m)
Figure 11-3 Maximum PCE, TCE, and C1 concentrations measured across

the reactive wall 199 days after installation of the wall (Gillham,
et al., 1993)
3000
Wall
Concentration (mg/L)
2500
TCE
PCE
Chloride
2000
1500
1000
500
10
11
12
Distance Along Flow Path (m)
Figure 11-4 Distribution of chlorinated degradation products across the wall,

299 days after installation of the wall (Gillham, et al., 1993)
Core samples collected from the wall after a year of operation for biological
study, indicated no evidence of biological growth and no chemical
precipitation or alteration of the iron (Matheson, 1994). Precipitates were not
detected by x-ray diffraction or scanning electron microscopy. Although losses
330
of 185 mg/l of calcium and 82 mg/l of bicarbonate were noted across the
wall, any precipitates that were being formed were not detectable. There was
no evidence that performance was being affected. Additional sampling of
wall material was planned at the conclusion of the test. The levels of TCE
and PCE destruction were below that expected, based on the initial laboratory
column tests. Subsequent lab tests, designed to better simulate the conditions
in the wall, suggest that a higher proportion of iron to sand might have resulted
in more complete removal.
11.3.2 SGL Printed Circuit Site, Wayne, New Jersey
A test of reactive iron technology for chlorocarbon destruction at Wayne, New
Jersey (EnviroMetal Technologies, Inc., Guelph, Ontario) is being monitored
under the SITE Program of the U.S. EPA (Gillham, et al., 1995). Although the
project has focused on on-site treatability testing of iron using an ex situ reactor,
the experience gained is relevant to permeable wall technology.
Chlorocarbon contaminants (PCE and TCE) are present in eight to 12 feet
of silty clay and the underlying fractured bedrock. Maximum concentrations
measured in the groundwater are 50 mg/l and 3 mg/l for PCE and TCE,
respectively. Much of the groundwater is believed to flow through a thin
permeable zone at the overburden-bedrock interface. It has been proposed
to install a drain at the overburden-bedrock interface to intercept the
groundwater flow and direct it to a subsurface iron treatment zone. To evaluate
the feasibility of treatment, laboratory column tests were conducted and an
above-ground reactor demonstration initiated.
Lab column tests were conducted using water collected from the site and
the commercial granular iron treatment medium being considered for use at
the site. The collected water contained PCE ranging from 4 to 12 mg/l, TCE
at 1 mg/l, and cis 1,2-DCE at 0.15 mg/l. No vinyl chloride was detected in
the collected water. DCE and vinyl chloride were detected at intermediate
points in the iron-filled column, but not in the effluent from the column.
Observed half-lives using 100% granular iron were: PCE, 0.4 to 0.6 hours;
TCE, 0.5 to 0.7 hours; DCE, 1.5 to 3.7 hours; and VC, 1.2 to 0.9 hours.
Design of the above-ground field test was based on a flow rate of two
liters per minute and a maximum PCE concentration of 30 mg/l. Based on
lab results, it was calculated that cis 1,2-DCE and VC concentrations as high
as 3 mg/l and 0.3 mg/l, respectively, could be encountered. The design
residence time of 24 hours for the field reactor was based on a New Jersey
standard of 0.01 mg/l for 1,2-DCE in the effluent. Considering the two liters
per minute flow rate, the packed bed reactor was sized to hold about 280
cubic feet of granular iron. It consists of an eight-foot diameter fiberglass tank
filled with granular iron to a depth of 5.5 feet. Water from the subsurface tile
drain flows into a sump from which it is pumped to the top of the reactor.
Flow through the reactor is vertically downward. Ponded water is maintained
above the iron surface to avoid invasion by atmospheric oxygen.
331
Concentration profiles for PCE, TCE, cis 1,2-DCE, after 30 days and 60
days of operation showed nondetectable levels in the effluent (Figure 11-5).
In fact, disappearance of all three constituents was observed to occur roughly
midway through the reactor. As shown in Figure 11-5, PCE and TCE influent
concentrations were somewhat lower than those in the collected water used
in the lab testing and assumed in sizing the reactor. Nonetheless, results
appear favorable and indicate the reactor is performing as designed.
Precipitate formation was observed at the top of the reactor but analyses of
the precipitates had not been performed at the time of the workshop. Calcium
carbonate, siderite, and possibly iron hydroxide formation were anticipated.
PCE
TCE
cDCE
0.0
55 pv
Depth (feet)
1.0
30 pv
55 pv
30 pv
55 pv
2.0
30 pv
3.0
4.0
5.0
0 1000 3000 5000 7000 0
200
400
150
300
Concentration (ug/L)
Figure 11-5 Concentration profiles of PCE, TCE, and cDCE after 30 and 60
days of operation of the New Jersey above-ground demonstration
test (Gillham, et al., 1995)
11.3.3 Hill Air Force Base, Utah

Laboratory column studies and an above-ground field test using a small
canister were completed at Hill Air Force Base in 1994 as part of the evaluation
of a reactive iron permeable barrier application (Wray, 1995). The source of
contamination under study is a landfill which received municipal waste
between 1955 and 1967, and possibly waste solvents from aircraft maintenance
operations at the base. The contaminant of concern is TCE. Maximum
measured concentration is 18 mg/l with an average of 3 mg/l. The plume
occupies about 69 acres and is located within the upper 25 feet of a shallow
aquifer, the top of which is 30 feet below ground surface.
Laboratory column studies were initiated by the University of Waterloo
332
in January 1994. Three column tests were completed, one with 100% granular
iron, one with 50% iron and 50% silica sand (by weight), and the third with
100% silica sand as a control. Influent to the columns was site groundwater
containing 3 mg/l TCE, 0.1 mg/l PCE, 0.05 mg/l cis 1,2-DCE, and 0.002 mg/l
vinyl chloride. These experiments showed rapid degradation of TCE with a
half-life of about 0.5 hours for the 100% iron column. Both cis 1,2-DCE and
vinyl chloride were observed as intermediate reaction products. They were
completely degraded in the 100% iron column but persisted in the effluent from
the 50% iron column. Retention time in these columns was not provided.
The above-ground canister test was operated from June to October, 1994.
The test used groundwater extracted from the site as the influent water. This
water contained 2 mg/l TCE. The fiberglass canister (4.5 feet long and 1.0
foot in diameter) was filled with 100% Master Builder iron (Strongsville, Ohio),
Blend B, GX-27. About five inches of pea gravel was placed at each end of
the reactor. Flow was upward through the canister and rates varied from
0.01 gallons per minute to 0.1 gallons per minute, corresponding to a retention
time between about two to 22 hours.
Influent pH was stable at about 7.5 throughout the test; however, pH
increased to above 9 during flow through the canister. Influent dissolved
oxygen ranged between 4 to 6 mg/l. The majority of the TCE was degraded
by the first sampling port, located about 8.6 inches into the canister.
Concentrations of cis 1,2-DCE and vinyl chloride were observed to increase
within the canister but were less than 0.001 mg/l in the effluent. Ethene and
ethane were produced and accounted for 60% of the initial TCE mass in the
effluent.
Gradual plugging of the canister occurred, as evidenced by an increase
of pressure drop across the canister from 0.5 psi to 9 psi during the test. X-ray
diffraction tests on iron samples taken from the canister indicated that iron
and calcium carbonate compounds were precipitated and appeared to cause
the plugging. About 5 grams of carbonate per 100 grams of granular iron
were deposited, corresponding to a porosity loss of 14%. Remaining porosity
loss was attributed to trapped particulates carried in the influent.
Implementation of a field scale treatment system is pending the resolution
of a number of issues. First, a determination of whether carbonate precipitation
would occur in an in situ barrier must be made, and if so, a remedy identified.
Second, the best location of the field system is on off-site property for which
access is not currently available. Third, the Superfund Record of Decision
(ROD) for the site, which has been signed and currently does not include this
remedy, would have to be amended.
11.3.4 Sunnyvale, California Semiconductor Site
This full-scale commercial application involved a permeable granular iron
reactive barrier for destruction of TCE, cis 1,2-DCE, and vinyl chloride in
groundwater. A low hydraulic conductivity soil-cement-bentonite and cement-
333
bentonite slurry wall system was designed to route groundwater through the
treatment zone. This system has received regulatory approval as a final
remedy by state and local regulators, replacing an interim groundwater pump
and treat system. Construction at the site was completed in February, 1995
(Warner, et al., 1995; Gillham, et al., 1995).
Site conditions consist of up to 20 feet of heterogeneous silt, sand, and
clay overburden on top of a 65-foot thick clay aquitard. Groundwater in the
overburden contains TCE (0.05 to 0.2 mg/l), cis 1,2-DCE (0.45 to 1.0 mg/l),
vinyl chloride (0.1 to 0.5 mg/l), and CFC-113 (0.02 to 0.06 mg/l). Concerns
over the heterogeneous nature of the overburden and variable seasonal flow
direction were addressed in the barrier design and construction.
An above-ground pilot test (9 month duration) was performed using a
six foot long by two foot diameter flow-through canister. The treatment
medium contained equal amounts of granular iron and sand. Half lives
determined from this test were: TCE, less than 1.7 hours; cis 1,2 DCE, 0.9
hours; vinyl chloride, 2.0 to 4.0 hours; and CFC-113, less than 1.6 hours. The
water at this site is highly mineralized and, although precipitate formation
was evident at the inlet end of the reactor, the rate of degradation remained
relatively constant throughout the test.
The permeable component is about 40 feet wide and 20 feet deep. The eightfoot thick treatment zone consists of four feet of granular reactive iron sandwiched
between two-foot thick upgradient and downgradient pea gravel sections,
designed to distribute flow evenly through the iron. The system included
upgradient slurry walls to guide groundwater flow laterally into the wall. The
eastern wall also was extended with sheet piling a distance of 20 feet downgradient
from the treatment barrier, thereby creating a chute which caused flow
convergence downgradient from the barrier rather than within the treatment
zone. As an added safety factor, the barrier was constructed with 100% granular
iron (i.e., no sand mixture), even though the barrier was sized based on the aboveground test half life results.
The site is regulated under an order issued by the California Regional Water
Quality Control Board - San Francisco Bay Region that requires quarterly
groundwater monitoring. Cleanup standards are developed from California
maximum contaminant levels (MCLs) with the lowest MCL being 0.0005 mg/l
for vinyl chloride.
Results from two quarterly groundwater sampling events indicate that
the concentration of total VOCs in groundwater about ten feet upgradient
from the barrier was approximately 1.0 mg/l during the most recent sampling
event with principal constituents of cis 1,2-DCE at 0.8 mg/l, vinyl chloride at
0.1 mg/l, and TCE at about 0.03 mg/l. Groundwater samples collected from
performance monitoring wells within the barrier showed no VOCs above the
detection limit of 0.0005 mg/l. Field water quality measurements in the barrier
showed high pH and low Eh, indicating that the barrier is affecting
groundwater in the expected manner. An assessment of hydraulic
performance indicated the system is performing within design specifications.
334
The former above-ground treatment system and groundwater extraction wells

have been removed from the property, and the formerly vacant site has been
leased and is now in use.
11.3.5 Elizabeth City, North Carolina Field Study
This small-scale field test at Elizabeth City, completed in 1995, evaluated the
in situ treatment of groundwater contaminated with chromate and
chlorocarbons. The permeable treatment system was composed of a granular
iron/coarse sand/native aquifer mixture installed as fence posts in a
staggered design using hollow stem augers (Puls, et al., 1995; Blowes, et al.,
1995). The project has been a cooperative effort by RSKERL, the University of
Waterloo, and the Oregon Graduate Institute, with assistance by the U.S.
Bureau of Mines.
The site is located at the U.S. Coast Guard Support Center near Elizabeth
City in the coastal plain region of North Carolina. The site geology consists
of coastal plain sediments with variable sequences of sand, silt and clay (Puls,
et al., 1994). Groundwater flow velocity is variable with depth, with highest
flows through a highly conductive layer at roughly 15 to 22 feet below ground
surface. This layer coincides with the highest groundwater concentrations of
chromate and chlorocarbons (TCE, cis 1,2-DCE, and vinyl chloride). While
the test was not specifically designed to remediate the chlorocarbons, their
concentration was monitored during the test. The water table ranges from
five to seven feet below ground surface.
Sources of iron included low-grade steel waste stock turned on a lathe
using diamond bits to produce 200 liters of turnings (source - Ada Iron and
Metal, Ada, Oklahoma), with a size range from 1 to 10 mm, and heated cast
iron in the form of iron chips (source - Master Builders Supply, Streetsboro,
Ohio), with a size range from 0.1 to 2 mm.
Twenty-one eight-inch diameter fence posts were emplaced on 1.0 to
1.5-foot centers in three rows oriented transverse to flow (Figure 11-6). In
addition, 24 monitoring wells were installed within the roughly 60 square
foot treatment zone. Most of these were 5/8-inch PVC wells with 1.0 to 1.5foot long screens set between 14 to 20 feet below ground surface. In addition
to these permanent points, temporary sampling points were used to increase
spatial resolution of the data. Tracer tests using bromide to evaluate
groundwater flow velocity through the fence region were performed prior to
and following emplacement.
Results of the eight month long test are summarized in Table 11-4 (from
Puls, et al., 1995). Untreated groundwater from the test area is characterized
by chromate concentrations in the range of 1.0 to 3.0 mg/l, dissolved iron
levels less than 0.05 mg/l, TCE+DCE about 6.5 mg/l, and Eh greater than 400
millivolts. Treated zones (represented by wells located within or
downgradient of the fence posts) showed chromate decreased to less than
0.01 mg/l, dissolved iron increased to 1 to 20 mg/l, TCE+DCE decreased to
335
1.5 mg/l, and Eh decreased to the range of 0 to 200 millivolts. In addition,

dissolved oxygen was shown to decrease from about 0.6 mg/l to less than 0.1
mg/l. Evidence of sulfate reduction was present, and a slight pH increase
was evident. In all, the data present a picture of chemical reduction and
precipitation of the chromate, reductive dechlorination of TCE, and
accompanying oxidation of iron metal to form dissolved iron. It should be
noted that while TCE was reduced on average by about 75% in the treated
zones and vinyl chloride was reduced to less than 0.002 mg/l, the cis 1,2DCE showed little change. This was attributed to the relatively short residence
time in the fence posts (10 to 16 hours estimated per post).
GW Flow
Shop
GMP1
6
F13
F11
F10
GMP2
F9
FFe
F5
F6
F12
F2
6 GMP3
F7
6
6
F4
F8
6
River
12"
F3
GMP4
Fe cylinders,
diameter = 8"
MW (single)
6 MW clusters
Figure 11-6 Plan view showing site monitoring locations at Elizabeth City,
NC and location of reactive fence posts (Puls et al., 1995)
While considerable variation in hydraulic conductivity exists at the site,
on average the iron-sand mixture was estimated to be about 75 times more
permeable than the aquifer. While this increase was considered desirable
from a standpoint of promoting flow through the fence posts, increasing plume
capture, and mitigating pluggage due to geochemical changes, it also
decreased residence time. This decreased residence time was not a limitation
for chromate reduction, but affected the degree of chlorocarbon degradation.
Tracer test data before and after fence emplacement showed increased
336
groundwater velocity through the fence area. Post-emplacement velocities

were approximately 1.0 and 1.7 feet per day at depths of 15.5 to 16.5 feet and
17.0 to 18.0 feet, respectively. Pre-emplacement velocities were 0.3 and 0.7
feet per day at these two locations. Much of the increase was due to temporal
variation in recharge during the winter and spring months of the test.
TABLE 11-4 Monitoring Results (from Puls, et al, 1995)
Chemical Property
Untreated Zones
Treated Zones
Cr (VI)
1 - 3 mg/L
< 0.01 mg/L
Fe (II)
1 20 mg/L
TCE/DCE
6.5
1.5
Eh
> 400 mv (Pt)
0 - 200 mv (Pt)
pH
5.6 - 5.9
> 6.1
DO
~ 0.6 mg/L
< 0.1 mg/L
Sulfides/SO4
absent/stable
present/slight decrease
Alkalinity
no change
increased
In summary, field test results showed nearly instantaneous reduction and

removal of the chromate forming an insoluble trivalent chromium-iron
hydroxide phase. The reduction of chlorocarbons was variable but followed
the expected order of TCE>>vinyl chloride>cis 1,2-DCE. A Phase II study
using a more traditional permeable reactive wall was in the planning stage
for 1996 as of the workshop date. Improvement in chlorocarbon degradation
was to be a focus of the Phase II study. A trenching machine was being
considered for construction of the reactive wall.
11.3.6
Sudbury, Ontario Field Trial
A small-scale field test of a reactive wall system was initiated in October 1993
at the Nickel Rim tailings impoundment site near Sudbury, Ontario (Blowes,
et al., 1995b). The reactive wall test made use of organic matter to promote
sulfate-reducing conditions thereby treating acid mine drainage and associated
dissolved metals (e.g., iron, calcium, nickel, and zinc).
The reactive wall test cell was excavated into an unconfined sand aquifer
within a shallow alluvial valley about 250 feet downgradient from the Nickel
Rim mine tailings impoundment. A contaminant plume of sulfate and
dissolved metals formed downgradient from the impoundment in the shallow
unconfined sand aquifer. At the location of the test cell, sulfate levels are in
the 3000 to 4000 mg/l range while dissolved iron levels vary from 600 to 1000
mg/l. The Eh is moderately oxidized (300 to 500 mV), pH slightly acidic (4.5
337
to 5.5), and alkalinity is less than 50 mg/l (as CaCO3).

The final dimensions of the excavated cell were about 4.8 ft long by 3.6 ft
wide by 4.0 ft deep. The reactive mixture consisted of composted leaf mulch,
pine mulch, and pine bark (as carbon sources), creek sediment (as a bacterial
source), agricultural limestone (neutralizing agent), and coarse sand and gravel
(to increase permeability in the cell). The cell was constructed by first driving
steel sheet piling, excavating soil from within the piling and backfilling the
void with the reactive mixture, then removing the piling.
Bundle-style piezometers were used to sample groundwater upgradient,
within, and downgradient of the cell. Results within the test cell and
immediately downgradient over a two year period following construction
showed increased pH (to >6), decreased Eh (to <250 mV), and corresponding
declines in sulfate (to <250 mg/l) and dissolved iron (10 to 50 mg/l). Nickel
levels decreased from roughly 2 mg/l in upgradient groundwater to less than
0.005 mg/l in the test cell. In addition, hydrogen sulfide was measured at
levels up to 5 mg/l within the cell.
Based on these positive results, construction of a full-scale test was
completed about a week prior to the workshop in August 1995. The fullscale system was placed across a narrow margin (50 ft across) in the alluvial
valley downgradient from the impoundment. Construction was completed
with techniques similar to those used in the test cell construction, with the
PRB keyed into underlying bedrock in both the valley bottom and the side
walls. The reactive mix included organic matter obtained from several
municipal compost sources mixed with pea gravel to obtain a hydraulic
conductivity about an order of magnitude greater than the aquifer. Results
from the full scale system were not available at the time of the workshop.
11.3.7 Future Planned Field Tests
The following briefly summarizes the status of some proposed tests.
a. Durango, Colorado Uranium Mill Tailing Site. The U.S. Department of
Energys Uranium Mill Tailing Remedial Action (UMTRA) program is
conducting a pilot scale study through Sandia National Laboratory (SNL) at
an UMTRA-owned site in Durango, Colorado. The focus of the study is to
find effective, cost-efficient methods for remediating sites with uraniumcontaminated groundwater (Marozas, et al., 1995; Marozas, 1995).
The use of reactive iron to reduce mobile radioactive metal species, such
as uranium, was based on the studies on chromate conducted by RSKERL
and the University of Waterloo. In the UMTRA pilot study, uranium and
other inorganic contaminants will be treated with metallic iron in a subsurface
reactive infiltration bed. Three different reactive iron reagents will be screened
for efficiency in removing uranium, selenium, and molybdenum from mill
tailing effluent. Since these are common constituents in uranium mill tailings,
the results will be used to evaluate whether reactive barriers are an acceptable
338
remediation option for groundwater cleanup at UMTRA sites.

The test site is located in the Bodo Canyon Disposal cell near Durango.
Roughly 2.5 million cubic yards of tailings were relocated to the canyon from
another location near Durango in 1990. The tailings are underlain by about
50 feet of silty clay, silt, and sand alluvium which overlie sandstone bedrock.
Shallow groundwater occurs in the alluvium.
The Bodo Canyon disposal cell is designed to limit infiltration and, with
time, the alluvium below the cell is expected to become dewatered. At that
point, the unsaturated zone will attenuate seepage from the bottom of the
cell. At present, fluids draining from the cell are collected by a gravity drain
at the toe of the cell and are treated by active chemical flocculation in a lined
retention pond. No contamination has been detected in downgradient
groundwater. Tailing effluent contains an average of 6 mg/l of dissolved
uranium, compared with a Colorado standard of 2 mg/l.
Like chromium, uranium generally is found in two oxidation states, the
oxidized and relatively mobile U(VI) state, and the reduced and relatively
immobile U(IV) state. Reactive wall chemistries have focused on chemical
reduction and precipitation to stabilize uranium.
An engineered subsurface treatment system will be installed inside the
existing lined retention pond. The pond is 60 feet by 80 feet and six feet deep.
The treatment system will be placed in the path of seepage from the toe drain
within a segmented area of the pond. The installed system will simulate flow
and treatment of mill tailings groundwater, but in a controlled setting.
The new treatment system will distribute tailings effluent into one of four
treatment zones that include two subsurface drain fields and two flow-through
enclosed containers. In both systems, effluent will percolate by gravity through
the treatment media into an underdrain. The treatment systems will be
installed below the subsurface to prevent oxygen from entering the reaction
zone and to prevent freezing during the winter.
Steel wool, steel wool treated with copper, and a metallic iron foam were
selected for this study, based on laboratory testing. These reagents were
selected for their environmentally benign nature, expected longevity without
maintenance, availability, and cost.
b. Shiprock, New Mexico UMTRA Site. A permeable barrier design is being
prepared for an UMTRA site located near Shiprock, New Mexico. The design
uses naturally occurring microorganisms to achieve reduction and subsequent
precipitation in the barrier (Thomson, 1995).
The Shiprock UMTRA site is located on an elevated river terrace
approximately 50 feet above the flood plain of the San Juan River. Underlying
strata consist of coarse alluvial material underlain by shale bedrock. The
flood plain alluvium is highly permeable and consists of 15 to 30 feet of
unconsolidated silts, sands, gravels, and cobbles overlying the shale. A steep
escarpment separates the elevated terrace and the flood plain. Principal
groundwater contaminants include sulfate (mean concentration 13,000 mg/l),
339
nitrate (3300 mg/l), uranium (2.8 mg/l), selenium (0.12 mg/l), and radium226 (0.1 picocuries/l).
Design of the permeable barrier is proceeding with laboratory studies
and groundwater modeling. A recent study at the University of New Mexico
developed a numerical model of groundwater flow in the flood plain. The
USGS MODFLOW code and PATH3D were used to simulate head distribution
and develop streamlines and groundwater travel times. The model was
calibrated using water level data from a system of 30 monitor wells constructed
in the flood plain. The calibrated model was used to consider a variety of
gradient manipulation techniques using permeable or impermeable barrier
locations to facilitate either flushing of the aquifer or remediation. The
optimum design has not yet been determined but will include a trench-based
permeable barrier keyed into the shale.
The permeable barrier will use an anaerobic bacterial consortium
contained in a cellulosic substrate to produce sulfate-reducing conditions in
the barrier. Substrates being considered include various agricultural waste
products, e.g., straw, wheat chaff, corn husks, or sawdust. Substrate will need
to remain in place for a period of at least several years to make the technology
viable. Column experiments are planned to determine rate of substrate
consumption.
The geochemical form of the reduced metals is important in determining
their ultimate fate in the barrier. Well-ordered crystallized mineral phases
are much less soluble than amorphous precipitates and are believed to oxidize
more slowly in aerobic environments. A major objective of current studies is
to characterize the solid phases formed during sulfate reduction. The two
basic options available for management of the spent reactor wall are: (1) leave
the barrier and the associated contaminants in place, or (2) remove the barrier
and dispose of the contaminants in an appropriate facility. If the first option
is selected, the ultimate fate of the metals will depend on the long-term
geochemistry in the barrier. It is reasonable to expect that metals will remain
in place as long as reducing conditions are maintained. Field sampling in
uncontaminated regions of the Shiprock site indicate reducing conditions
presently exit. The alternative is to remove the spent reactive wall once
remediation is completed and the substrate has been consumed. This option
has been proposed for the Shiprock site to remove all possibility of future
release.
c. Gasworks Sites - Southern Germany.
Two former gasworks sites in
southern Germany are under investigation for application of sorptive barriers.
Both of the sites are quite large (several centers of contamination of roughly
80 feet by 80 feet at each site spread over a large area), and broad groundwater
plumes have formed with widths of 150 to 500 feet. Contaminants of concern
include three to six ring polynuclear aromatic compounds (Teutsch and
Grathwohl, 1995). The proposed reactive wall design utilizes the funnel and
gate approach. Because of the expense in building a replenishable in situ
340
reactor, the researchers wish to minimize the size of the reactors.

The two sites under consideration are located in valley settings above
unconsolidated, heterogeneous sand and gravel aquifers. Concern about the
effect of subsurface heterogeneity on the overall performance of the systems
is being addressed by conducting small scale tracer tests to establish local
flow conditions and velocities. Data from the tracer tests will be used in system
design to set reactor sizes and configurations and to determine the feasibility
of using a hanging gate design.
The reactor system under study consists of a number of high organic
carbon materials including activated carbon, bituminous coal, and bituminous
shale. The reactor construction employs a steel frame with individual cells
(length of 7.9 feet and width in direction of flow of 1.7 to 2.6 feet). Two to
three rows of the cells will be arranged perpendicular to flow direction. The
system is intended to have flexibility in filling, monitoring, and replenishment
of the reactor material. The steel frame is sized to allow individual cells to be
emptied entirely. Future studies will focus on extension of the reactor system
to combine a bioreactor with sorption.
11.4 EVALUATION OF THE TECHNOLOGY

The permeable reactive barrier technology is still largely in the developmental/
demonstration stage. However, as indicated by the previous sections, a great
amount of work is being performed and the technology has been applied or
proposed for a wide range of common groundwater contaminants including
metals, chlorocarbons, and petroleum hydrocarbons. The technology has
undergone an exponential increase in research and development activity over
the past four years.
The major conceptual advantage of the permeable reactive barrier relates
to the potential for remediating contaminated groundwater more costeffectively than with pump and treat systems. This appears to be particularly
true for sites which involve persistent sources of groundwater contamination
for which remediation methods have not been demonstrated or are costprohibitive (e.g., DNAPL sites). At these sites, the PRB may actually outlast
the plume life and eliminate the need for institutional controls over
groundwater use at the site. On the other hand, PRBs are not yet considered
a proven technology and a number of potential disadvantages and technical
hurdles have been identified that will require further effort before the
technology is applied routinely.
11.4.1 Cost Evaluation
The incentive to develop PRBs as a viable technology relates directly to the
increasingly recognized limitations of pump and treat. A principal expected
advantage of the permeable barrier technology is the greatly reduced operation
341
and maintenance (O&M) costs. For example, consider the generic case for
the reactive iron PRB applied to chlorocarbon plume containment presented
by Schultz (1995). Conventional methods for treating chlorinated solvents in
groundwater include liquid phase activated carbon adsorption and air
stripping, often combined with vapor phase carbon treatment. A recent review
within DuPont produced the following rules of thumb for the cost of such
systems:
installation costs between $10,000 and $30,000 per gallon per minute
(gpm) of treatment capacity, and
O&M costs ranging between $5 to $20 per 1000 gallons of treated water.
Costs for O&M of permeable reactive barrier systems will not be available
until they have been operating for many years, but it is generally believed
they will be lower than those for pump and treat. On the other hand, barrier
emplacement can be estimated today based on costs to construct barrier
containment systems, such as sheet piles and slurry walls. Assuming a
relatively shallow plume depth of 20 feet, a corresponding emplacement cost
of $5 to $10 per square foot of wall face (from slurry wall construction
experience), and delivered cost for granular iron of $400 per ton (as quoted
by Peerless Iron of Detroit, Michigan), the construction cost of a reactive iron
permeable wall was estimated to range from about $19,000 to $23,000 per
installed gpm (somewhere in the middle of the range for pump and treat
systems). If this estimate proves correct and if the expectation of much lower
O&M costs prove true, this technology would be more cost-effective than
pump and treat for chlorocarbon containment.
This cost comparison is supported by the Sunnyvale, California
commercial PRB application. For this site, a detailed feasibility and cost study
were completed prior to selecting the reactive iron PRB for TCE plume
containment. It was estimated that the cost of removing an existing pump
and treat system from the site and installing the permeable barrier would
result in a savings of nearly $5 million on a present value basis over a 30 year
period. The cost advantage was attributed to reduced O&M, elimination of
treatment system, discharge monitoring and reporting, and revenue gains
from the ability to lease the property for commercial use once pump and treat
surface facilities were removed. It should be noted that significant safety
factors were incorporated into the Sunnyvale design. For example, the reactive
iron zone thickness was doubled from the required two feet predicted by
modeling, and the proportion of iron in the wall was increased from 50% by
weight to 100%. Thus, four times as much iron was used in the wall than the
least cost design optimum. Still, the PRB alternative had substantial cost
advantage over the existing pump and treat system.
The cost to implement the PRB technology is directly related to the cost
of the reactive materials; thus, it is highly dependent on the volume of
groundwater requiring treatment. A decision must be made whether the PRB
342
should be used as a final remedy for all affected groundwater at a site, or if it

may be more cost effective as a source control with another treatment method
(such as intrinsic biodegradation) used to remediate lower concentration fringe
areas. An assessment of the cost effectiveness of the final remedy must also
consider the costs associated with achieving a specified factor of safety for
the final design. Other remedial alternatives may be more cost-effective than
pursuing additional site characterization or employing ultraconservative
design assumptions. Finally, the practicality of PRBs should be assessed
considering site-specific factors and then compared to the feasibility of
implementing other traditional and/or innovative technologies.
11.4.2 Other Conceptual Advantages
It should be possible to design and construct a robust PRB system with
little or no operating equipment. A major problem with pump and treat
systems is the tendency for equipment breakdown and system failure; e.g.,
recovery wells plug up and require redevelopment; treatment systems,
especially air strippers, require frequent maintenance; and pumps fail.
Hydraulic control over the plume is lost for some period whenever a pump
and treat system goes off line for repair or maintenance. Depending on the
aquifer characteristics, the length of time to re-establish the zone of capture
can be a few hours to a few months. As a permanent physical structure, the
PRB (if designed properly) should have the advantage of full-time operation.
The inherently slow nature of groundwater movement works to the
advantage of the technology by allowing use of slow destruction technologies
that are not practical in an above-ground treatment setting. The reactive
granular iron technology is an excellent example. The rate of destruction is
such that reactor residence time of several hours to a few days may be needed
to achieve destruction to the parts per billion levels currently required for
chlorocarbons. It is generally not practical to size an above-ground system to
achieve such long residence times, whereas a below-ground PRB of the
required size is quite feasible.
Other factors to consider are: (1) the inherent energy efficiency of the
process, (2) the fact that little or no surface facilities are required, (3) the process
is conservative of groundwater resources, and (4) the need for surface water
discharge or groundwater recharge is eliminated.
PRBs offer other conceptual advantages when compared with low
hydraulic conductivity physical barriers. For example, PRBs are not designed
to exclude groundwater flow but instead work within the natural flow regime.
This is particularly true of the continuous wall PRB design. It is less true of
the funnel and gate concept. In any event, flaws and imperfections in the
construction are not nearly as critical as in the construction of an impermeable
barrier, where much higher pressure gradients may exist across the barrier.
Also, in contrast to an impermeable barrier, there may not be a need to control
upgradient surface water infiltration (e.g., by capping) to reduce flux.
343
11.4.3 Potential Disadvantages

The potential disadvantages associated with the PRB technology are primarily
related to its early stage of development and the relative lack of field
experience. First and foremost, the viability of this technology relates directly
to its ability to provide long-term, relatively trouble-free service. Since there
is no extended-term performance record beyond four years (e.g., the Borden
reactive iron wall demo), the remediation community may be reluctant to
put trust in the technology until a longer-term track record is established.
With respect to long-term performance, are the following factors are of concern:
possible plugging and reductions in hydraulic conductivity over time
as a result of particle invasion, chemical incompatibility between
reactant and host formation, or microbial activity, and
possible gradual loss of reactivity as the reactant is either depleted or
coated by reaction by-products (i.e., rind formation).
Theoretically, it should be possible to address each of these issues through
laboratory evaluation, modeling, and proper design. However, the tools to
complete such evaluations have not been completely developed and verified
at this stage. There is a lack of tested and proven design procedures and
protocols developed to this point. In addition, a protocol for effective site
characterization in support of PRB design is still evolving. As discussed in
Section 11.2, a different focus is needed in the site characterization from the
standard investigations done at hazardous waste sites. Greater attention must
be given to aquifer characteristics, local and seasonal groundwater flow
patterns, and groundwater geochemistry.
Long-term analogs exist in nature for many of the processes being
proposed for PRB application. For example, the use of redox barriers to
precipitate trace metals has a direct corollary in the creation of uranium
deposits in reducing zones in relative shallow sand aquifers of the western
U.S. Study of such corollaries may lead to a greater level of assurance that
PRBs can operate and remain geochemically stable for very long periods.
Another disadvantage relates to the potential depth limitations for
installation of PRBs. With current technology, the cost of PRB emplacement
is strongly related to depth of emplacement. The current rule of thumb for
emplacement depth limits using excavation techniques is about 100 feet. At
greater depths, alternative emplacement methods such as jet grouting, mandrel
emplacement, or injection well emplacement must be considered. None of
these methods have been demonstrated for this purpose and may require
additional development.
An obvious disadvantage of PRBs, compared with certain other
technologies, is that PRBs are passive; the contaminant must come to the PRB.
This is not necessarily a disadvantage where the cleanup objectives are based
on risk management, but may be a disadvantage when site restoration is the
344
goal. However, both pump and treat done for hydraulic control purposes
and impermeable physical barrier technologies possess this same limitation.
Nevertheless, the PRB approach would appear to have limited flexibility once
installed, compared to a pump and treat system where individual wells may
be added, removed, or replaced relatively easily.
Finally, emplacement of PRBs may face greater logistical constraints than
siting a pump and treat system. Underground utilities, buildings, pavement,
and overhead structures all must be identified, avoided (if possible), or moved
during PRB construction. In addition, excavation may result in the generation
of contaminated soil, requiring management and added costs.
11.5 DEVELOPMENT NEEDS

It is apparent that the technology is still in the development stage and that
there are hurdles to be overcome before the remediation community will
accept and apply the technology. While a number of technical needs have
been identified, the one basic need is to increase the comfort level of the
remediation community and validate the technology through successful field
demonstrations. The success of demonstration projects in turn hinges on
the development of effective and readily available design tools and protocols.
Other identified needs include:
development of effective/proven modeling and design tools,
increased experience with field applications, and wider dissemination
of know-how,
cheaper, more effective emplacement methods, especially for deeper
(i.e., greater than 30 feet) PRBs,
enhanced mechanistic research on treatment processes leading to
development of optimized media,
engineering evaluation of alternative PRB configurations, with
guidance for optimizing hydraulic controls.
Each of these needs in discussed in more detail below.
11.5.1 Modeling and Design Tools
A basic need exists to develop the know-how for modeling and design of
PRB systems. Of greatest priority is the development of recognized and
accepted protocols for the following:
geochemical and hydrologic site characterization supportive of PRB
design;
performance of treatability studies, including guidance concerning
345
what levels of treatability study are sufficient in a given setting;

design of the hydraulic control and reactor systems;
safety factor rules of thumb, and/or methods to mitigate pluggage
and ensure long-term performance; and
development of proven methods for working over PRBs which have
lost hydraulic efficiency and/or reactivity.
Development of a well-recognized basis for designing PRBs for long-term
performance will accomplish several objectives. First and most obvious, the
successful performance of future PRBs will be better assured. Second, the
technology will become better accepted by the remediation community as
greater comfort is acquired with the design and operation of these systems.
Third, a basis will be established for regulatory review and approval.
11.5.2 Accelerated Field Experience and Communications
Coupled with the need for recognized modeling and design tools is the need
for increased field experience, and communication of this experience to the
remediation community. The widespread successful application of any
technology generally follows only after a sufficient level of know-how is
developed. The know-how in turn can only come from the extensive level of
experience derived and shared from many applications. Both successful and
unsuccessful applications can contribute to the experience base.
As previously discussed, there has been little field experience with PRBs,
with one well-documented commercial scale project and a handful of smaller
scale field demonstrations having been undertaken. Thus far, the track record
looks promising, but it is too early to generalize from these results. Confidence
in the use of permeable reactive barriers will increase as the number of
successful field demonstrations increases. Several field demonstrations are
currently in the planning stage and there appears to be a high level of interest
in PRB technology.
Several PRB technologies are ready for field testing. In a few cases,
principally the reactive iron technologies for chlorocarbons and reducible
metals, the field effort has started and valuable lessons are being gained.
However, other types of permeable barriers useful for non-reducible organics
and metals have lagged in development. In particular, the sorption barrier
technology could benefit from performance of a few well-planned field
experiments with a variety of contaminants, treatment media, and geologic
settings.
Timely communication of results and findings associated with laboratory
and field projects is important and forms the basis for effective technology
transfer. This includes not only communication and collaboration among the
technology developers, but also with the remediation community at large.
Formation of a central clearinghouse for demonstration plans and reports,
application costs, and sharing of basic know-how would facilitate a more
346
rapid pace of development.

11.5.3 Development of Better Emplacement Methods
A potential limiting factor to the use of permeable reactive barrier technology
appears to be the inability to cheaply emplace the barriers, particularly at
depths greater than 30 feet. PRB emplacement to a depth of 30 feet is possible
with current technology. Emplacement at greater depths (from 30 to 100 feet)
is more difficult and increases costs. Emplacement below depths of 100 feet
is essentially unproven at this stage with current technology. The existence
of surface obstructions (i.e., buildings, roads, utilities, etc.) presents other
challenges to the emplacement of a PRB.
Emplacement methods that require excavation may generate substantial
volumes of contaminated soil. Under current RCRA rules, these soils may
need to be handled as waste, thereby increasing project costs. Worker exposure
may also be an issue. Excavators and trenching equipment tend to produce a
fixed trench width which may or may not fit the particular PRB design
requirement (for example, retention time considerations may indicate a PRB
could be built with a width of one foot whereas trenching machines are often
two feet wide). This difference could result in over-designed systems with
associated added costs, particularly where expensive treatment media are
involved.
The jet grouting method and mandrel emplacement methods have
potential for producing less waste and emplacing deeper PRB systems than
the excavation-based approach. However, they are relatively unproven and
may have certain drawbacks of their own. For example, adequate control
over emplacement may be an issue with the jet grouting approach, whereas
mandrel emplacement technology has been questioned with respect to
formation disturbance and plugging around the reactor. Nonetheless,
improvement of these methods could help to reduce the cost of deeper
emplacements. The development of improved emplacement methods,
possibly involving new equipment and techniques to achieve greater depths,
could benefit the development of PRB technology. In addition, the
development of no waste emplacement, or alternatively, regulatory policies
which provide greater flexibility in handling excavated soils, should also
contribute to the development of PRB technology.
11.5.4 Research on Treatment Processes
Considerable laboratory-based research has been conducted to date on a
variety of treatment processes. For example, sorption of organics appears to
be fairly well understood considering its widespread use in above-ground
flow-through systems.
Additional mechanistic research on certain key PRB reactive chemistries
is needed. This is particularly true with respect to the use of reactive iron for
347
chlorocarbon destruction. At this stage, much remains to be learned about

the reactive iron dechlorination process. For example, Burris (1995) points
out that two basic reaction pathways for PCE/TCE dechlorination are evident
(one involving sequential dechlorination to ethene and the other involving
acetylene as the end-product). Since the acetylene route appears to produce
less intermediates (e.g., DCE and vinyl chloride), it is regarded as the more
desirable pathway for dechlorination. Research is currently underway at the
Armstrong Lab at Tyndall Air Force Base to identify and influence the factors
involved in this process.
Extensive work is ongoing concerning aspects of the reactive iron
technology. This work has been quite fruitful in understanding the
mechanisms, suggesting improvements, and translating the knowledge gained
into better field designs. For example, the role of dissolved oxygen in causing
premature reactor plugging has been identified as a result of mechanistic work
in the laboratory. Work at U.S. EPAs Athens, Georgia laboratory has yielded
additional suggestions for preventing pluggage and for maintaining reaction
rates by mixing iron sulfide with the granular iron to stabilize pH (Holser, et
al., 1995). Other improvements can be expected based on the knowledge
gained from laboratory research.
11.5.5 Evaluation of Configuration Alternatives
The continuous wall approach involves the construction of the reactor portion
of the PRB across the entire front of the plume, whereas the funnel-and-gate
option reduces the reactor size in favor of low hydraulic conductivity barrier
funnels to guide the groundwater flow. In the continuous wall approach, the
reactor is sized to match the plume life, or at the least to provide for a very
substantial life of treatment (e.g., 30 years). Thus, a substantial up-front cost
is realized associated with the purchase of the reactor media for the given
lifetime. Although the funnel-and-gate design reduces reactor volume, it may
require periodic replenishment of the reactant and periodic O&M.
All things being equal, the choice of a PRB configuration alternative would
be based on economic considerations. The need then, is for a set of economic
evaluations to be performed which would identify the important cost variables
and aid decision-making. A cost model for the reactive iron PRB technology
is nearing completion under a multi-year project administered by the Air Force
through the Strategic Environmental Research and Development Program
(SERDP) (Smith, 1995). This model will function as a module in the Air Forces
ENVEST remediation costing program. The new ENVEST module could serve
as a tool for carrying out these economic evaluations.
11.5.6 Recommendations for Addressing Needs
First and foremost, it is recommended that a suite of definitive field
demonstrations be conducted in a variety of hydrogeologic settings and with
348
a range of groundwater contaminants.

Second, there needs to be a collaborative development of recognized
procedures and protocols for: (a) site investigation, (b) treatability studies,
and (c) hydraulic control and reactor design. These tools should be developed
as consensus documents by leaders in the R&D community and shared widely
with the remediation and regulatory community. U.S. EPAs Remediation
Technologies Development Forum, which features cooperative technology
development between public and private organizations, is viewed as a model
framework for establishing this collaborative effort and in sharing the findings.
Third, a concerted effort is recommended to educate the remediation
community, including regulators, to help pave the way for future applications.
A clearinghouse for information exchange should be developed to effect timely
technology transfer. The clearinghouse would serve as a mechanism for
sharing status of research programs, demonstration project plans and reports,
and notice of meetings. A newsletter format (either as a stand-alone letter or
added as a section to an existing newsletter) would be preferred. Also,
publishing the newsletter on the Internet may be a good mechanism for broad
dissemination of information related to PRB developments.
11.5.7
Intellectual Property and PRB Development
One issue not addressed at the workshop concerns the rights to intellectual
property and the evolving relationships between technology developers,
technology users, and intellectual property owners. While cooperative
development was seen by the panel as the optimum path for moving this
technology into full-scale application, it may not be realistic to expect that
sharing of patents and know-how necessary for true cooperative development
will take place. On the other hand, technology development and field testing,
particularly by the public sector, could result in increased business for
technology owners if performance is verified independently. Thus, there are
incentives which could outweigh the impediments.
This issue certainly is not unique to PRB technology. However, potential
users should be cognizant that several PRB processes have been patented
(e.g., Gillham, 1993; Blowes and Ptacek, 1994) which could limit the potential
for cooperative development of certain aspects of the technology.
11.6 REFERENCES
Alvarez-Cohen, L., P.L. McCarty, and P.V. Roberts. (1993). Sorption of
trichloroethylene onto a zeolite accompanied by methanogenic
biotransformation, Environmental Science and Technology, Vol. 27, 21412148.
Bianchi-Mosquera, G.C., Allen-King, R. M. and D. M. Mackay. (1994).
11.6 REFERENCES
349
Enhanced degradation of dissolved benzene and toluene using solid

oxygen-releasing compounds, Ground Water Monitoring and Remediation.
Blowes, D.W., and C.J. Ptacek. (1994). System for treating contaminated
groundwater, U.S. Patent No. 5,362,394.
Blowes, D.W., Ptacek, C.J., Cherry, J.A., Gillham, R.W., and W.D. Robertson.
(1995). Passive remediation using in situ treatment curtains,
Geoenvironment 2000, Characterization, Containment, Remediation, and
Performance in Environmental Geotechnics, Acar, Y. B. and Daniel, D. E. (Eds.),
Geotechnical Special Publication No. 46, vol. 2, American Society of Civil
Engineers, New York, pp. 1588-1607.
Blowes, D. W., Ptacek, C. J., Bain, J. B., Waybrant, K. R., and W. D. Robertson.
(1995b). Treatment of mine drainage water using in situ permeable
reactive walls, Sudbury 95 Conference on Mining and the Environment, May
28 - June 1, Sudbury, Ontario, Vol. III, pp. 979-987.
Burris, D.R., and C.P. Antworth. (1992). In situ modification of an aquifer
material by a cationic surfactant to enhance retardation of organic
contaminants, Journal of Contaminant Hydrology, Vol. 10, 325-337.
Burris D.R. (1995). Written communication to the International Containment
Technology Workshop, Permeable Reactive Barrier Session, Baltimore,
MD.
Devlin, J.F., and J.F. Barker. (1993). A semi-passive injection system for in
situ bioremediation, Abstract, Second International Symposium, In Situ and
On-Site Bioremediation, April 5-8, 1993, San Diego, CA. Sponsored by
Battelle Memorial Institute.
Farquhar, G. L., Schnarr, M. J., and T. Gonullu. (1992). Enhanced chemical
attenuation: Destruction of the oily phase, in situ chemical oxidation,
Subsurface Restoration Conference, Dallas, Texas, pp. 63-65.
Gallinatti, J. D. and S. D. Warner. (1994). Hydraulic design considerations
for permeable in situ groundwater treatment walls, abstract in
Groundwater, Vol. 32, p. 851.
Gillham, R. W. (1993). Cleaning halogenated contaminants from
groundwater, U. S. Patent No. 5,266,213.
Gillham, R. W., OHannesin, S. F., and W. S. Orth. (1993). Metal enhanced
abiotic degradation of halogenated aliphatics: laboratory tests and field
trials, 1993 HazMat Central Conference, Chicago, IL, March 9-11.
350
Gillham, R.W., and S.F. OHannesin. (1994). Enhanced degradation of

halogenated aliphatics by zero-valent iron, Ground Water, Vol. 32, 958967.
Gillham, R.W. (1995). In situ permeable reaction walls for treatment of
contaminated groundwater, Written communication to the International
Containment Technology Workshop, Permeable Reactive Barriers Session,
Baltimore, MD.
Gillham, R.W., S.F. OHannesin, and J.L. Vogan. (1995). Abiotic degradation
of chlorinated organic compounds by zero-valent metals, Written
communication to the International Containment technology Workshop,
Permeable Reactive Barriers Session, Baltimore, MD.
Grathwohl, P. (1990). Environmental Science and Technology, Vol. 24, 16871693.
Haggerty, G.M., and R.S. Bowman. (1993). Sorption of chromate and other
inorganic anions by organo-zeolites, Extended abstract, Division of
Environmental Chemistry, American Chemical Society, 292-295.
Holser, R.A., McCutcheon, S.C., and N.L. Wolfe. (1995). Mass transfer effects
on the dechlorination of trichloroethylene by iron/pyrite mixtures in
continuous flow column systems, Written communication to the
International Containment Technology Workshop, Permeable Reactive
Barriers Session, Baltimore, MD.
Kao, C-M, and R.C. Borden. (1992). Denitrification of BTEX using nutrient
briquets as the nitrate source: laboratory microcosm and column studies,
Proceedings, Subsurface Restoration Conference, June 21-24, 1992, Dallas, TX.
Karickhoff, S.W., Brown, D.S., and T.A. Scott. (1979). Water Research, Vol. 13,
241-248.
Kleinmann, R.L.P., T.O. Tiernan, J.G. Solch, and R.L. Harris. (1983). A lowcost, low-maintenance treatment system for acid mine drainage using
sphagnum moss and limestone, National Symposium on Surface Mining,
Hydrology, Sedimentology and Reclamation, University of Kentucky,
Lexington, Kentucky.
Korte, N., R. Muftekian, C. Gritteni, Q. Fernando, J.L. Clausen, and L. Liang.
(1995). ORNL/MMES research into remedial applications of zero-valent
metals; bimetallic enhancements, Division of Environmental Chemistry,
American Chemical Society, 209th ACS National Meeting, April 2-7, 1995,
Anaheim, CA, Vol. 35, 752-754.
11.6 REFERENCES
351
Mackay, D.M., and J.A. Cherry. (1989). Groundwater contamination: pumpand-treat remediation, Environmental Science and Technology, Vol. 23,
630-636.
Mackenzie, P., T. Sivavec, and J. Salvo. (1995). Written communication to the
Barrier Session, Baltimore, MD.
Marozas, D.C., B.P. Dwyer, and K. Cantrell. (1995). Reactive barrier treatment
of groundwater at uranium mill tailing sites, Written communication to
the International Containment technology Workshop, Permeable Reactive
Marozas, D. C. (1995). personal communication.
Matheson, L.J. (1994). Abiotic and biotic reductive dehalogenation of
halogenated methanes, Ph.D. Dissertation, Environmental Science and
Engineering, Oregon Graduate Institute of Science and Technology, 125 p.
Matheson, L.J., and P.G. Tratnyek. (1994). Reductive dehalogenation of
chlorinated methanes by iron metal, Environmental Science and
Technology, Vol. 28, 2045-2053.
Morrison, S.J., and R.R. Spangler. (1993). Chemical barriers for controlling
groundwater contamination, Environmental Progress, Vol. 12, 175-181.
Morrison, S.J., and R. Spangler. (1995). Permeable barriers for inorganic
contaminants; the need for field experiments, Written communication
to the International Containment Technology Workshop, Permeable
Reactive Barrier Session, Baltimore, MD.
Morrison, S. J., Spangler, R.R., and S. A. Morris. (1996). Subsurface injection
of dissolved ferric chloride to form a chemical barrier: laboratory
investigations, Ground Water, Vol. 34, pp. 75-83.
Murdoch, L.C., G. Losonsky, P. Cluxton, B. Patterson, I. Klich, and B. Braswell.
(1991). The feasibility of hydraulic fracturing of soil to improve remedial
actions, Final Report U.S. EPA 600/2-91-012. NTIS Report PB91-181818.
298 p.
National Research Council. (1994). Alternatives for Ground Water Cleanup,
Committee on Ground Water Cleanup Alternatives, National Academy
Press, Washington, D.C. 315 p.
Orth, R. (1995). Permeable reactive barriers, Written communication to the
352

Pankow, J.F., R.L. Johnson, and J.A. Cherry. (1992). Air sparging in gate wells
within cutoff walls and in trenches for the control of chlorinated organic
compounds (VOCs), Ground Water, Vol. 31, 654-663.
Paulson, S.E., and L.M. Petrie. (1995). Written communication to the
Phifer, M.A. (1995). Permeable reactive walls - hydraulic and emplacement
overview, Written communication to the International Containment
MD.
Puls, R.W., D.A. Clark, C.J. Paul, and J. Vardy. (1994). Transport and
transformation of hexavalent chromium through soils and into
groundwater, Journal of Contaminant Hydrology, Vol. 3, 203-224.
Puls, R.W., Paul C.J., and R.M. Powell. (1995). In situ remediation of ground
water contaminated with chromate and chlorinated solvents using zerovalent iron: a field study, Written communication to the International
Containment Technology Workshop, Permeable Reactive Barriers Session,
Baltimore, MD.
Robertson, W.D., and J.A. Cherry (1995). In situ denitrification of septic
system NO3 using reactive porous barriers: Field trials, Ground Water,
(in submittal).
Rumer, R. R. and M. E. Ryan. (1995). Barrier Containment Technologies for
Environmental Remediation Applications, John Wiley & Sons, New York,
NY.
Schultz, D.S. (1995). Iron in situ treatment zones for dechlorinating
groundwater, Written communication to the International Containment
MD.
Smith, M. (1995). Tyndall Air Force Base, personal communication.
Starr, R.C., and J.A. Cherry. (1993). In situ remediation of contaminated ground
water: the funnel and gate system, Ground Water, vol. 32, 465-476.
Sweeny, K.H., and J.R. Fischer. (1973). Decomposition of halogenated organic
11.6 REFERENCES
353
compounds using metallic couples, United States Patent No. 3,737,384.

Teutsch, G., and P. Grathwohl. (1995). Permeable sorptive walls for treatment
of hydrophobic organic contaminant plumes in groundwater, Written
communication to the International Containment Technology Workshop,
Thomson, B.M. (1995). Permeable barriers for groundwater remediation at a
uranium mill tailings remedial action project (UMTRA) site, Written
communication to the International Containment technology Workshop,
Vogan, J. L. (1993). The use of emplaced denitrifying layers to promote nitrate
removal from septic effluent, M. Sc. Thesis, Dept. of Earth Sciences, Univ.
of Waterloo, Waterloo, Ontario.
Vogel, T. M., Criddle, C. S., and P. L. McCarty. (1987). Transformation of
halogenated aliphatic compounds, Environmental Science and
Technology, Vol. 21, 722-736.
Warner, S.D., C.L. Yamane, J.D. Gallinatti, F.S. Szerdy, and D.A. Hankins. (1995).
Assessing the feasibility of permeable reactive barriers for treating VOCaffected groundwater in situ: experience from the first full-scale
commercial application in California, Written communication to the
International Containment technology Workshop, Permeable Reactive
Wilson, J.T., and J.F. McNabb. (1983). Biological transformation of organic
pollutants in groundwater, EOS, Vol. 64, 505-507.
Wray, D. (1995). Hill Air Force Base, UT, personal communication.
Xu, Y., and F.W. Schwartz. (1992). Immobilization of lead in groundwater
with a reactive barrier system, Proceedings, Subsurface Restoration
Conference, June 21-24, 1992, Dallas, TX.
355
SECTION 12
PERFORMANCE MONITORING AND EVALUATION
prepared by
Hilary I. Inyang, University of Massachusetts at Lowell

J. David Betsill, Sandia National Laboratories

Randall Breeden, U.S. Environmental Protection Agency
Grover H. (Skip) Chamberlain, U.S. Department of Energy
Subijoy Dutta, U.S. Environmental Protection Agency
Lorne Everett, Geraghty & Miller
Rene Fuentes, U.S. Environmental Protection Agency
Jan Hendrickson, MSE, Inc.
John Koutsandreas, Florida State University
David Lesmes, Boston College
Guy Loomis, Idaho National Engineering Laboratory
Stephen M. Mangion, U.S. Environmental Protection Agency
Dale Morgan, Massachusetts Institute of Technology
Cathy Pfeifer, Idaho National Engineering Laboratory
Robert W. Puls, U.S. Environmental Protection Agency
Robert L. Stamnes, U.S. Environmental Protection Agency
Terry D. Vandel, Conoco, Inc.
Cecelia Williams, Sandia National Laboratories
12.1
12.1.1
BACKGROUND
Containment System Performance Monitoring Objectives
Waste containment system performance data are needed to conduct

assessments that can be grouped into three primary categories: exposure risk
assessment; source term estimation of contaminant concentrations for
contaminant transport models; and facility permitting and future maintenance
needs assessment. Long term performance of containment systems is a
concern because of the long design service lives of most facilities. Typically,
design service lives range from as little as 10 years for slurry walls to more
than 1,000 years for radioactive waste storage structures. The longer the service
life of a containment system, the greater the probability of system failure (i.e.
greater opportunities exist for excessive damage accumulation and
contaminant transport). Considering that most components of waste
containment systems exist underground, constraint is placed on direct visual
356
inspection as a monitoring method. Thus, several traditional and evolving

techniques of indirect observation need to be employed to obtain performance
data for the following categories of assessments.
a. Exposure Risk Assessment. Exposure risk assessment is usually conducted
within the framework of environmental and human health risk assessment.
Many agencies and facility owners have adopted risk-based approaches to
selecting and formulating contaminated site remediation plans. The use of
innovative, cost-effective monitoring methods can provide accurate estimates
of parameters that are included in exposure assessment equations, especially,
when the objective is to quantify the reduction of risk that may attend the
implementation of a containment system.
Many computational methods and recommendations (Marnicio at al.,
1991; Schanz and Salhotra, 1990; Imam et al., 1990; Tonn and Wagner, 1990;
Kostecki et al., 1989; Taylor et al., 1987; Preuss et al., 1986; U.S. EPA, 1990a;
U.S. EPA 1990b; and U.S. EPA, 1990c) have been developed for use in risk
assessment. Equation 12-1 is an example of a general relationship for exposure
assessment. The use of innovative monitoring techniques for contaminants
around facilities can improve numerical estimates of parameters C, EF, and
ED of Equation 12-1.
IN =
C IR EF ED
BW AT
(12-1)
IN = intake = the amount of a specific chemical in a contaminated

medium taken by a receptor organism (mg/kg of body weight/
day).
C = concentration = average chemical concentration contacted over
the exposure period (mg/l, or mg/mg).
IR = intake rate (or contact rate = amount of contaminated medium
contacted per unit time or event) (mg/day or l/day).
EF = exposure frequency (upper bound value), (days/ year).
ED = exposure duration (upper bound value), (years).
BW = body weight = average body weight over the exposure period
(kg).
AT = averaging time = time period over which exposure is averaged =
exposure duration for non-carcinogens, and 70 years for
carcinogens, (years).
b. Source Term Estimation for Transport Models. Uncertainties characterize
numerical estimates of source term concentrations of contaminants that may
travel through barrier systems. Monitoring systems may be installed to
measure source strengths for verification of estimates generated by numerical
12.1 BACKGROUND
357
models. Monitoring data can be incorporated into contaminant transport

models to estimate future concentrations at locations away from the source.
Figure 12-1 shows a surface impoundment of basal width d, which has released
contaminants into the vadose zone. The source term concentration, Ca , as
well as the fluid migration rate q, can be monitored. Leachate plume
concentration, Co, can be monitored and/or numerically estimated at locations
around the containment system, using several relevant parameters among
which are plume velocity Vh, and thickness b. In the case illustrated, the
contaminant release rate is partly proportional to the rate at which the
containment system deteriorates. Within the context of source term estimation,
the structural state of the barriers as well as contaminant concentration
parameters, Ca and Co, can be monitored.
Surface
impoundment
Ground surface
Long term flow
channels Dt
Liner
d
Unsaturated
geomedia
q, Ca
Water table
Saturated
geomedia
Leachate plume
Vh, Co
Bedrock
Figure 12-1 Illustration of source term parameters for a simplified assessment

of contaminant migration from a leaking surface impoundment
(Inyang, 1994b)
c. Facility Permitting and Maintenance. Post-construction performance

monitoring by waste management facility owners is necessary in order to
refine maintenance plans. Also, information on performance is useful to
regulatory personnel who are involved in permitting of initial facility
construction as well as expansions. Some containment system rating and
performance prediction methods have been developed for these purposes
(Inyang, 1994a; and Koerner and Daniel, 1992). Figure 12-2 illustrates the
conceptual long term deterioration pattern and maintenance scheme proposed
by Inyang (1994a) for waste containment systems. It should be noted that
this is a conceptual scheme and does not simulate the exact deterioration
pattern of a specific containment system. Following curve 1, the effectiveness
of a containment system is believed to decrease under service conditions,
358
System Effectiveness, E (fraction)
from an initial value of Eto to a value Etr, at which regulations require the
implementation of repair activities. If maintenance is implemented at time
tm, the effectiveness of the containment system may increase instantaneously
to a value, Etm. A damaging natural event such as an earthquake or flood
may cause a sharp decrease in effectiveness as illustrated in Figure 12-2 by a
drop from E1 to Etg at time tg. In order to verify and improve the accuracy of
this type of modeling approach, monitoring data are required. Due to the
large volume of data required, the use of non-invasive geophysical and
electrochemical sensing systems is desirable.
Curve 1
Eto
Curve 3
Etm
E1
E2
Etg
Etr
Curve 2
to
tg
t2 tm
tr
Time, t (years)
Figure 12-2 A conceptual long term deterioration pattern and maintenance

scheme for waste containment systems (Inyang, 1994b)
12.1.2
Containment System Performance Factors
Containment systems are constructed facilities. All constructed facilities

contain flaws and are susceptible to deterioration processes. Flaws may not
always occur at sufficient scale and frequency to cause functional failure of a
containment system. An appreciation of probable deterioration processes
and damage-producing transient events is essential for optimal design of
containment system monitoring schemes. For example, excessive settlement
of a landfill cover can result from a high rate of waste decay within a landfill.
Settlement estimates can be used to optimize the distribution of monitoring
points on the landfill cover.
a.
Deterioration Mechanisms.
Containment system deterioration
mechanisms fall into two categories: physico-chemical processes; and transient
events. Physico-chemical processes often occur continuously but slowly. They
may wane or intensify with time, depending on the chemical environment,
temperature, and pressure conditions. It is difficult to monitor physico-
12.1 BACKGROUND
359
chemical processes directly for most practical assessments. Typically, the

effects of physico-chemical processes on the containment system are measured,
rather than the processes themselves. Examples of such processes are barrier
flocculation, dissolution, desiccation and slurry settlement. Some transient
events occur dramatically, e.g., ground movements due to earthquakes. The
primary determinant of the probability that a well-constructed containment
system will be affected by a transient event during the postconstruction period
is its location. Zoning systems have been developed for some transient events
such as earthquakes and freeze-thaw action (Inyang, 1991; Wright et al., 1993).
b. Deterioration-Time Pattern. Engineered waste containment systems
have existed for a short time in comparison with other types of structures.
Numerical estimates of the design lives and potential deterioration pattern of
containment systems have been made without the benefit of field data on
past performance of similar systems over reasonably long time intervals. Most
of the data that are available on the damage of components are not timedistributed. Thus, it is difficult to assess the rate at which the damages
accumulated and the relative contributions of various deteriorations processes
to the observed damages. Table 12-1 provides a summary of field experiences
with some waste containment systems compiled by Bass et al. (1985). Laine
and Miklas (1989) also documented leaks in impoundments as shown in Table
12-2.
The paucity of time-distributed monitoring data for a wide variety of
waste containment system configurations has necessitated the development
of general deterioration-time models based on observed deterioration pattern
of most constructed facilities. Inyang and Tomassoni (1992) have discussed
this problem in detail. Furthermore, Inyang (1994a) proposed the use of
Equation 12-2 for time-scaling of the effectiveness or reliability of waste
containment systems.
Rt = exp -{ [t - to]/n}b
Rt =
t =
to =
b =
n =
(12-2)
reliability of the containment system at the future time of

reference, t
future time of reference [T]
time considered to be the start of deterioration for analysis
purposes (location parameter) [T]
shape parameter [dimensionless]
scale parameter [T]
The shape parameter is directly proportional to the design conservatism

of the waste containment system that needs to be monitored. For barrier
containment systems, the shape parameter ranges from 2 to 5, approximately.
This is the range for which the deterioration versus time curve exhibits the
usual negative exponential decay geometry. The scale parameter is a
360
TABLE 12-1
Single
(S) or
Site
Double
ID
(D)
Liner
Summary of some experiences with containment system

performance in the field (Bass et al., 1985)
Primary
Liner
Material
Total Exposed
Problems
Primary
MoniSurface (E) or
Layers in Liner System* Air
with
Liner
toring
Area
Buried
(bottom to top)
Vents
Liner
(mil)
System
(ac)
(B)
V1-1
OR-CPE
(R)
36
10
No
V1-2
CSPE (R)
36
22
Yes
Comp Clay/S/FML
Yes
Yes
LimeRk/S/FML/S/
Lime/Rk
No
V1-3
V1-4
Gr/GeoTex/S&G/GeoYes
Tex/FML/Soil cement
PVC (U)
30
No
PVC (U)
30
10
No
Comp Soil/FML/Soil
No
V1-5
PVC (U),
CSPE (?)
20, 36
Yes
Comp Clay/S/FML/
S/FML
Yes
V1-6
PVC (U)
30
No
Old Fill/Clean
Fill/FML/Clay
Yes
V2-1
CSPE (R)
30
120
No
Comp Clay and

Limestone/FML
Yes
Yes
V2-2
CSPE (R)
30
No
Comp Soil/S&G/FML
No
Yes
V2-3
CSPE (R)
30
2.3
No
Comp Sub-base/FML
No
Yes
V2-4
CSPE (R)
30
4.3
Yes
Comp Fill/FML/S/G
No
Yes
Prepared Limestone/
FML/Stone
No
Yes
V3-1
PO (R)
30
42
V3-2
PVC (U)
20
75
Yes
Comp Clay/FML/S
No
Yes
V3-3
PVC (U)
20
No
Comp Soil/FML/S
No
V3-4
Soil
Sealant
4 in
25
Ye s
Comp Sand/Liner/S
No
Yes
V3-5
Asphaltconcrete
5 in
Ye s
Comp Soil/Asphalt
(2 lifts)
No
V4-1
HDPE (U)
100
18
Yes
Comp Sand/FML
No
V4-2
HDPE (U)
100
18.5
Ye s
Comp Sand/S/FML
Yes
V4-3
HDPE (U)
80
88
No
Comp Subgrade/FML
Yes
V4-4
HDPE (U)
80
E
(sides)
Yes
Clay/S/Comp Soil/
FML/Comp Soil
Yes
V4-5
HDPE (U)
80
3.2
Ye s
Comp Clay/FML/
Comp Clay
No
V4-6
HDPE (U)
100
0.3
Yes
Comp Soil/FML
No
80
66
E
(sides)
No
Subgrade/FML/S
(bottom only)
No
V4-7
HDPE (U)
V5-1 3D, 1S
CPE (U),
CPE (U)
20, 30
1.5
E (CIM
only)
Yes
Subgrade/CPE/Soil/
Concrete/CIM
No
(?)
Yes
V5-2
CPE (U)/
PVC (U)
20, 10
13
Ye s
Nat Soil/FML/Nat Soil
No
Yes
V5-3
CPU (U)
30
0.7
Ye s
Nat Soil/FML/Nat
Soil/Soil Cement
No
V5-4
PVC (U)
20
1.4
Yes
Comp Soil/Clay/
S/FML/Nat Soil
No
(?)
Yes
V5-5
Triple
30, 20
0.75
Ye s
Comp Fill/CPE/G/
PVC/CPE/?
No
(?)
2xCPE(R),
PVC (U)
*Comp = compacted; FML = flexible membrane liner; G = gravel; GeoTex = geotextile;

Gr = ground; Nat = natural; Rk = rock; S = sand
12.1 BACKGROUND
TABLE 12-2
Survey No.
361
Leak detection and survey data for impoundments where the

bottom floor areas were surveyed (Laine and Miklas, 1989)
Size in ft2
Total Leaks
Leaks Located in
Bottom
Seam
Sheet
Leaks per
10,000 ft2
20.9
958
958
31.3
958
31.3
40.0
1,000
1,798
0.0
2,625
22.9
3,000
21
21
21
70.0
3,000
13.3
3,200
0.0
10
4,951
0.0
34.3
11
4,951
17
17
17
12
4,951
4.0
13
5,175
3.9
14
7,007
5.7
15
12,600
5.6
27.3
16
18,346
50
50
35
15
17
26,016
2.7
18
26,016
1.5
19
27,297
2.9
20
32,292
25
25
25
7.7
21
43,560
0.5
22
45,345
0.9
23
50,000
1.2
24
50,400
193
193
188
38.3
25
54,500
29
29
18
11
5.3
26
55,025
12
12
12
2.2
27
58,900
1.4
28
62,500
21
21
19
3.4
29
64,583
29
29
21
4.5
30
65,340
56
56
55
8.6
31
65,369
0.9
32
65,369
1.1
0.8
33
65,369
34
65,500
1.1
35
65,500
0.8
36
74,088
20
20
19
2.7
37
82,500
18
18
15
2.2
38
87,120
0.9
39
87,120
17
17
17
2.0
40
99,050
18
18
14
1.8
41
135,036
17
17
16
1.3
42
150,781
64
64
46
18
4.2
43
152,460
0.1
44
152,460
0.5
45
157,584
12
12
10
0.8
46
164,085
18
18
16
1.1
47
362,690
51
51
37
14
1.4
TOTALS
2,769,336
811
811
709
102
2.9
362
normalization factor for time. It is the time duration that corresponds to a

containment system failure probability of 0.632. The necessity for
implementing automatic monitoring systems increases as (t-to)/n increases
and the shape parameter, b decreases. This temporal reliability concept has
not yet been applied to indexing of long term performance of waste
containment systems. A requirement for the practical application of Equation
12-2 is the formulation of a scheme for relating the shape parameter to initial
design configuration, barrier material properties, and the probable intensities
of possible damaging events and processes.
12.1.3
Monitoring Approaches and System Failure Determination
In terms of containment system effectiveness, two types of failure categories

can be identified:
structural failure: one or more components of the containment system

fails structurally (excessive deformation and/or introduction of flow
channels).
functional failure: the containment system can no longer perform the
function for which it was designed (prevention of excessive
contamination of the protected environment).
Structural failure can occur without functional failure although structural

failure may eventually lead to functional failure. Furthermore, for multicomponent systems, the structural or functional failure of a component does
not necessarily imply that the entire system has failed. It depends on the
criticality of the failed component. With respect to the use of monitoring data
to define containment system functional failure, the following two approaches
can be identified:
arrival of contaminants in excessive concentrations at a critical point

within or just outside the containment system, and
excessive contamination of a protected natural resource by
contaminants that have traveled away from the containment system
(examples of natural resources are surface waterbodies and
groundwater aquifers).
Three principal approaches can be adopted to monitor structural and

functional failures of waste containment systems. In most cases (as described
in section 12.5.3), these approaches are complementary. Various types of
monitoring methods that fit within these three principal approaches are
summarized in Table 12-3 and Table 12-4.
a.
Barrier Integrity Monitoring.
This approach involves monitoring
deformation and flaw development within barriers. Such flaws may
363
12.1 BACKGROUND
compromise the structural integrity of the containment system and serve as

flow channels for moisture and/or contaminants. Figure 12-3 illustrates the
increase in infiltration through cracked portions of a concrete cover, relative
to intact portions of the cover.
lc
lc
lm
Cemented Cover
Figure 12-3 Concrete cover with illustration of infiltration through cracks (Ic)
relative to infiltration through matrix (IM) (Inyang and Myers,
1993)
b. Barrier Permeation Monitoring.
This approach involves sensing of
moisture and contaminant concentrations and flow rates within components
of a containment system. As examples, the infiltration rate of moisture through
a cover system and/or the concentration of specific contaminants in the
leachate within a drainage layer can be monitored.
TABLE 12-3 General application of monitoring approaches

Monitoring Method
Barrier Monitoring
Approach
ElectroWell
Geophysical
chemical
Network Methods
Methods
Mechanical
and Electrochemical
Methods
Electrical
Methods
1 Barrier Integrity
Monitoring
2 Barrier Permeation
Monitoring
3 External Monitoring
Key: C = conventional; G = growing application; R = rare; U = unfeasible

Note: These methods comprise several specific techniques, some of which may not
necessarily fit into these three categories.
364
TABLE 12-4
Examples of traditional and innovative monitoring techniques

for specific waste containment problems
WASTE CONTAINMENT PROBLEM
MONITORING TECHNIQUES
1. Cover slope deformation and failure
Inclinometers, electrical resistance shear

strips, and vibrating wire extensiometers
2. Lateral deformation of vertical walls
Electrical inclinometers, extensiometers,

and tiltmeters
Excessive settlement and cracking of soil

3.
and concrete covers
Sealed Borros anchor, inclinometers,

settlement platforms, electrical crack
guages, convergence guages, microwave
sensors, dye-tracing
4. Excessive infiltration of cover systems
Tensiometers, gypsum blocks,

psychrometers, neutron guages and
piezometers in covers, thermo-electric
flow velocity guages in drainage layers,
and fiber-optic moisture sensor in covers
5.
Voids and high permeability zones in

emplaced and natural barriers
Crosshole seismic surveys, ground

penetrating radar surveys, seismic
tomography, surface-based seismic
refraction surveys, soil gas surveys,
temperature logging, cone penetration
testing, and dye tracing
Leachate plume migration from waste

6. containment systems; and plume capture
by permeable reactive walls
Electro-chemical sensing cables and fiberoptic chemical sensors around

containment structure, DC electrical and
electromagnetic resistance surveys above
ground, ground penetrating radar
surveys, dye tracing, and groundwater
monitoring
Contaminant flow through bottom

7. barriers and low hydraulic conductivity
walls
Electro-chemical sensing cables and fiberoptic chemical sensors within, upgradient,

and downgradient from barriers, and
pumping tests
8.
Leaks in geomembranes and sheet pile

joints
Electrical leak detection systems and fiberoptic monitoring systems for joints
9.
Flow of leachate through reactive

permeable walls
Thermal flow sensors upstream, within,

and downstream from wall, electrochemical sensors, pH and Eh sensors
c. External Monitoring. External monitoring involves the measurement of

contaminant concentrations outside a containment system. It is geared towards
the identification of functional failure. Monitoring locations proximal to and/
or distant from the containment system can be selected. Hydrogeological factors
and containment system configuration are important factors in external
monitoring point selection. Inyang and Tumay (1995) have analyzed various
containment system configurations. External monitoring may involve one or
12.2 MONITORING TECHNIQUES AND TECHNOLOGIES
365
more of the following techniques, depending on the purpose of the monitoring

effort:
12.2
successive measurements at a point to determine the arrival time of

moisture and/or contaminants following the construction of a
containment system,
successive measurements at various points to estimate the transport
rates of moisture and contaminants through the media around the
containment system, and
one-time measurement of contaminant concentrations at various
locations around the containment system to delineate the boundaries
of a plume.
CURRENT AND EVOLVING MONITORING TECHNIQUES AND

TECHNOLOGIES
Four primary categories of techniques and technologies are currently used to

monitor waste containment system performance within the three approaches
outlined in the preceding section. As summarized in Table 12-3, despite the
potential cost-effectiveness of some of the techniques, they have not been
used widely to monitor performance. Current and evolving monitoring
systems and the principles of their operation are briefly discussed below. For
details, the reader is referred to Inyang (1994b), Betsill and Gruebel (1995)
and U.S. EPA (1993a,b).
12.2.1
Groundwater and Leachate Monitoring Well Network
Traditionally, monitoring of containment systems has been primarily

performed through groundwater and leachate collection and testing.
Leachates are usually extracted from sumps within the containment area, or
from water wells sited around the containment system as illustrated in Figure
12-4. Minimum requirements generally specify the placement of one
upgradient and three downgradient wells around a containment system.
Water is extracted periodically from these wells and monitored for specific
contaminants. Leachate collection sumps are part of the Leachate Detection
System (LDS) of landfill systems. A survey of 28 landfills (U.S. EPA, 1992)
indicated that at most of the facilities, sumps are inspected daily. A recent
study (GAO, 1995) revealed that 1209 hazardous waste land disposal facilities
in the United States are required to implement groundwater monitoring
programs. Seventy-seven percent of the facilities use one well to track potential
releases; whereas, the rest require from 2 to 17 wells. An increase in the number
of wells can improve monitoring effectiveness. However, it is emphasized
that properly designed and installed monitoring wells are still an indispensable
366
aa
tool for monitoring containment performance. Several technical guidance

documents (e.g., U.S. EPA, 1991c and U.S. EPA, 1989) have been developed
for use in the implementation of effective groundwater and leachate
monitoring systems.
Upgradient Wall
Downgradient Wells
a
Landfill
Unsaturated Zone
Saturated Zone
Water Table
River
Saturated Zone
Shale
Sandstone
Figure 12-4
12.2.2
Traditional groundwater monitoring system using wells (GAO,

1995)
Strain Gauging and Mechanical Detection of Flaws
Gauges capable of measuring deformation of subsurface and exposed

components of waste containment systems are in use today as some facilities.
These gauges have previously been used for many decades in geotechnical
structures (e.g., embankments, diaphragm walls, building foundations, and
dams) to monitor deformations exemplified by settlement, lateral movement,
inclination, and slope failures. Strain gauges can provide information on the
onset and location of potential structural failures within a containment system.
The deformation of a strain gauge that is compliant with a containment system
in which it is installed can be read from a scale or converted to electrical
impulses for motion indication. Strain causing actions such as shear,
extensions, compression, and deflection can be detected with these gauges.
Dunnicliff (1982), King and Kusenberger (1973), and Dietrich and Salley
(1975) have provided details on various strain sensing principles. Inyang
(1994b) has described the principle of operation of an electrical resistance
shear strip, e.g., the one illustrated in Figure 12-5. The strip can sense
deformation occurring transverse to its orientation. Typically, an electric circuit
runs through two parallel strips that are embedded in the monitored structure.
Since distortions in the atomic lattice of a metal rod during straining influence
aa
367
its electrical resistivity and straining reduces the cross-sectional area through
which current flows, it follows that the change in electrical resistance of a
strain gauge is directly proportional to its change in length. Equation 12-3
represents this proportionality and is used in practice to scale linear
deformation to the electrical resistance of an embedded gauge.
S=
( R ) L
=
Lo
Ro
S
Ro
Lo
R
L
=
=
=
=
=
(12-3)
strain sensitivity factor

initial resistance of the gauge wire
initial length of the wire
change in resistance due to straining of the wire
change in length of the wire
The theoretical value of S ranges from -12 to 6 for common metals (King and
Kusenberger, 1973). Other devices for measurement of deformation of
containment system components include:
interface slip measurement using electrical resistance shear strips,

settlement gauges for covers, liners, waste piles and confined disposal
facilities,
inclinometers for slurry walls, and
extensometers for covers, liners and confined disposal facilities.
Processor and
data storage unit
Parallel electric
circuit
Sliding slope failure
Landfill
cover
Deformed
slurry wall
Waste
Landfill liner
Soft clay
Sand
Lateral soil
movement
Figure 12-5 A schematic illustration of the use of an electrical shear strip to

sense the flow failure of a landfill cover (Inyang, 1994b)
368
Unfortunately, despite several deformations and failures of containment

systems reported by Mitchell et al., 1990; Byrne et al., 1992; Suter et al., 1993;
Kahle and Rowlands, 1982; Yen and Scanlon, 1975; and Dodt et al., 1987, very
few containment systems have been instrumented with strain and structural
failure sensors. The focus has been on the detection of barrier permeation.
12.2.3
Geophysical Systems
The geomaterials and fabricated materials that form the components of waste
containment systems, as well as the natural and subsurface around them,
exhibit physical and chemical characteristics that may change with chemical
contamination and/or internal void development. These changes can be
detected through the use of geophysical methods.
Geophysical techniques are based on the response of the geomedia and
pore fluids to various segments of the electromagnetic (EM) spectrum, or
seismic and/or acoustic energy, or other potential fields, e.g., a magnetic field
or gravitational field. The radiation type employed by a given electromagnetic
technique within the spectrum is described in terms of its range of frequency
or wavelength. Electrical currents in the ground can be induced remotely
using electromagnetic methods, eliminating the need to physically disturb
the specific subsurface location being evaluated.
Electromagnetic measurements can be used to investigate materials either
in the frequency domain or time domain. Frequency domain evaluations
track the materials response to electromagnetic fields at one or a number of
frequencies. This is often termed electromagnetic induction (EMI). In time
domain electromagnetic measurements (TDEM) or transient electromagnetic
(TEM) soundings, the change in response is measured as a function of time
following the switch-off of the transmitters. Application of TDEM methods
in contaminant migration studies is relatively recent.
Seismic methods are based on the speed of travel of compressiongenerated disturbances (or waves) in the medium evaluated (Terzaghi, 1943).
These waves may be reflected or refracted at boundaries between media of
different physical and/or chemical properties. Usually sensors are positioned
at known distances from a seismic source to detect reflected or refracted waves.
The speed of travel of the waves depends on the layering and properties of
the media through which they travel.
Gravitational methods are based on the principle that the differences
between measured gravitational acceleration or potential of media or objects
reflect their mass differences. Thus, a buried drum accelerates differently
from its surrounding soil.
Geophysical methods have been used for many years to verify the
continuity of grout curtains and bulbs emplaced in the subsurface. Similar
techniques are currently being used to monitor emplaced barriers. However,
it should be noted that obtained data may be of limited use for quantitative
assessment of contaminant migration. The resolution of currently available
369
geophysical technologies is too large to reveal the details of flaw sizes and
directions in constructed barriers. With respect to containment system
performance monitoring, the application of geophysical techniques is growing
in two areas.
Assessment of the continuity of natural and emplaced barriers

- Cross-hole seismic imaging
- Surface seismic refraction
- Ground penetrating radar (GPR)
- Microwaves, ultrasonics, and radio waves
Tracking of contaminant plume extent around containment systems

after release from an existing containment system
- Electromagnetic (EM) resistivity
- Ground penetrating radar (GPR)
These geophysical methods are briefly described below.

a. Crosshole Seismic Imaging. This new technique is illustrated in Figure 12-6.
Shear and compression waves, generated from an array of transmitters in a
borehole, travel through the medium of interest and are detected by an array of
receivers in another borehole. The travel times of seismic waves between the
transmitters and receivers are inverted into a two-dimensional velocity structure
map using the principles of tomography. This method is semi-invasive because
of the use of boreholes. However, by placing the boreholes outside the
boundaries of the contaminated zone or the grouted media, the risk of drilling
through the barrier or contaminants can be minimized.
Control
equipment
Borehole
Inert gas cylinders

(Air, nitrogen, etc.used to operate
transmitters)
Sensors connected
to computer
Shock Waves
Transmitters
Receivers
Subsurface
object (fluid,
rocks, waste
drums, etc.)
Figure 12-6 Schematic of a crosshole seismic tomography imaging system

(U.S. DOE, 1994b)
370
Compression waves generated by seismic sources are in general, faster in

dense, tightly cemented, and wet materials than in loose and dry materials.
Thus, seismic wave velocity is directly proportional to the density of the
transport media. Grouting of a loose medium generally increases its density.
If the density increases significantly, wave travel velocity through the grouted
or improved medium will increase. Hence, before and after (grouting) transit
time measurements can be used to evaluate the effectiveness of barriers that
are large enough to exhibit significant changes in seismic wave velocities.
b. Surface Seismic Refraction. This method is effective mainly for use in
delineating relatively horizontal boundaries among strata that exhibit different
physical characteristics and moisture saturation levels. It is very effective
within the upper 30m of the subsurface, which applies to the depth range of
most waste containment and site improvement projects. However, the
difficulties associated with interpreting refraction data from irregularly shaped
structures within the ground hampers the application of this method for
delineating media boundaries in the subsurface. Nevertheless, measurements
taken before and after ground improvement can be used qualitatively in
indexing the effectiveness of projects.
The operating principle is that primary (compressional) waves are
refracted at the interface between media of significantly different physical
characteristics; e.g., density, modulus of elasticity, porosity, and moisture
content. For shallow hazardous waste site investigations, seismic waves can
be generated by a hammer hit or drop weight on the ground. Geophones
placed at distances away from the seismic source detect incoming refracted
waves and are processed using electrical circuits. The data acquired from
geophones are typically processed and plotted as points on time-distance
graphs. Straight lines are used to connect points based on the analyzers
interpretation of the data. Changes in the slope of lines indicate differences
in wave velocities and hence material characteristics of different layers of the
subsurface. A critical assumption in the use of seismic refraction is that within
each stratum, density increases with depth. Also, interpretation is easier in
cases where denser strata underlie less dense strata.
c. Ground Penetrating Radar (GPR). GPR uses electromagnetic waves to
penetrate the ground and delineate differences in pore water content and
quality, soil texture, and soil density. Wave frequencies ranging from 80 to
1000 megahertz (broad band) are normally used. Upon contact with materials
of different properties, a fraction of the wave energy is reflected back to a
receiving antenna located on the ground surface while the remaining fraction
propagates deeper into the ground. Reflected electromagnetic pulses can be
converted to plots of amplitude versus wave travel time.
Soils with high conductivity of electromagnetic waves (e.g., clays and
other fine-grained saturated soils) rapidly dissipate radar energy, thereby
limiting depth penetration. While penetration depths of up to 10 meters are
371
common, 5-10% (by weight) of montmorillonitic clay can reduce the

penetration depth to about 1m (Walther et al., 1986). Conductivity is also
directly proportional to the concentration of dissolved solids in the pore water.
In field applications, the surface receiving antenna can be towed at speeds up
to 8 km/h. In Port Washington, New York, GPR was used at the Roslyn/
Beacon Hill Landfill site to optimize the locations of monitoring wells (Kardos
and Ennis, 1993).
d. Electromagnetic (EM) Resistivity. This method is also based on changes
of the electrical conductivity that usually occurs between subsurface zones of
different physical and chemical properties. Since electromagnetic phenomena
are generated in the media investigated, it is not necessary to use electrodes
embedded in the ground to transmit and receive current. For this reason, EM
resistivity is an effective means for delineating contaminated zones beneath
covered areas (e.g., structural foundations, streets, and parking lots).
The principle of inducement of electromagnetic field in geomedia for
resistivity measurements is illustrated in Figure 12-7. An alternating current
is applied to the terminals of a transmitter coil, fixed on or close to the ground
surface, to induce the flow of a current, thereby creating a magnetic field
which, in turn, causes electric currents to pass through the earth. Within the
earth, a secondary magnetic field is induced. The secondary and primary
magnetic fields are detected by a receiving coil placed near the transmitting
coil. For a fixed spacing between coils and operating frequency, the magnitude
Current
Ground surface
I
Voltmeter
Slurry wall
Buried
waste
a
Direction of
groundwater movement
Contaminant
plume
Figure 12-7 Sketch of Wenner resistivity array for delineating contaminant

plumes behind a subsurface barrier (Inyang, 1994b)
of the secondary magnetic field (or ratio of the secondary to primary magnetic
fields) is directly proportional to the conductivity of the ground. Resistivity
can be computed from conductivity data. In the subsurface, the solid matrix
acts as an insulator while soil moisture acts largely as the conductor of the
372
electric current. Furthermore, the conductivity of the pore fluid is proportional

to the concentration of dissolved ions. Thus, these measurements can be used
to delineate contaminated zones around waste containment barriers. If buried
metallic objects are present near the surveyed site, the readings may be
negatively affected. The effective depth of penetration can be as high as 60m.
This method is often called the Frequency Domain Electromagnetic (FDEM)
method.
There are two other variations of the electromagnetic induction method.
In one variation, the rate at which the magnetic field dissipates is measured
after the transmitter is turned off. An eddy current flows through the ground
at successively greater depths. The data obtained are interpreted to obtain
resistivity variation with depth. This is called Time Domain Electromagnetic
(TDEM) survey. It has been applied by Hoekstra et al. (1992) to trace the
migration of brines from an oil field brine pit in southwestern Texas.
Penetration depths of up to 500m are estimated by Hoekstra et al. (1992). The
other variation of (EM) measurement involves the use of very low frequencies
(15 to 25 kHz). In this method, ground contact is required for the potential
electrodes. It is very suitable for investigating relatively shallow contaminant
plumes (20-50m) around waste disposal sites or containment systems.
Techniques based on microwaves, ultrasonics, and radio waves can also
be used to detect voids in various types of waste containment barriers,
especially in cemented components. Laine et al., (1980) and Koerner et al.,
(1982) have described methods that are based on radiowaves and microwaves,
respectively. As will be discussed in section 12.3, the full potential of
geophysical techniques has not been realized in containment system
monitoring. Geophysical techniques, which are being adapted from site
characterization technology, need further refinement to achieve the spatial
resolutions necessary for containment system monitoring programs.
12.2.4
Electrochemical Systems
Electrochemical sensing systems operate on the principle of changes in the

physico-chemical characteristics of the sensor due to contact with a solid or
fluid. These changes may be directly converted into electrical or optical
impulses which are conveyed to processing equipment for interpretation.
Interpretation of impulses involves the location of points of contact between
the field of concern and the sensor network. Processing units with digital
recorders have been developed to track several sensors concurrently. The
essential difference between these detection systems and geophysical
techniques is that the detection sensor must be in contact with the pore fluid
whereas, using geophysical methods, the presence of the fluid is remotely
sensed, albeit less precisely.
Sensing systems can be divided into three categories (illustrated in Figure
12-8) on the basis of the spatial coverage of the sensors within a waste
containment structure or a potentially contaminated zone. Point systems
373
typically comprise a stem which terminates in a sensor. The sensor may also
be located at a point on the embedded instrument. Linear systems are typically
cables. Areal systems comprise point and/or linear systems linked together
to form a network. For example, a cable can be looped or netted to monitor a
wide area beneath a landfill. This is especially effective, considering that it is
often difficult to determine a priori, where leakage will occur from a
containment system.
Point Sensor
Linear Sensor
Areal Sensor
Processor
and
readout
Connecting
wire
Waste
Waste
Waste
Sensor
Sensor
Slurry wall
Sensor
(entire loop)
Figure 12-8 Illustration of external monitoring of waste containment

structures using point, linear, and areal sensors (Inyang, 1994b)
The Detex Systems and Raychem Corporation sensing cables are depicted
in Figures 12-9 and 12-10, respectively. The operation of both cables is based
on electrical conductivity measurements, but the mechanism by which voltage
drop occurs along the encased wire differs for each sensor. Voltage drop along
the Detex Systems conducting cable is caused by the degradation of the coating
upon contact with the target substance (hydrocarbons). In contrast, the
polymer jacket of the Rachem cable swells upon contacting hydrocarbons,
causing contact with an internal conducting wire, resulting in a voltage drop
in the circuit. Other physico-chemical mechanisms specific to targeted classes
of contaminants or water can be employed. Sensing cables and heads can
also be made permeable, exclusively, to water.
Fiber-optic sensors are based on the transmission of probe signals within
the visible light regime by optical fibers through very long distances to
embedded sensors. Wavelength-dependent optical attenuation of the probe
beam or production of fluorescent emissions by the contaminant can be
measured and scaled as an index of the presence of the contaminant at the
sensor location. Often, an organic dye is added to the sensor to promote
fluorescence upon reaction with low concentrations of target compounds.
Most of the relevant reactions are reversible. Thus, it is possible for changing
contaminant concentration levels to be recorded by a single sensor. The details
374
of a reversible fiber-optic sensor designed to measure carbon tetrachloride

(see Figure 12-11) is being developed by Lawrence Livermore National
Laboratory (U.S. DOE, 1994b). Currently, fiber-optic sensors are designed
for specific classes of contaminants. Figure 12-12 depicts a processing circuit
developed for detecting hydrocarbons (DeFilippi and Cody, 1994).
Metal conductors
Coating degradable
by hydrocarbons
Figure 12-9
Cross section of a Detex Systems electrical conductivity sensor

based on coating degradation by hydrocarbons
Sensor wire
Signal wire
Continuity wire
Conductive
polymer layer
Overbraid
Figure 12-10 Cross section of Raychem Corporation electrical conductivity

sensor based on polymer swelling on contact with hydrocarbons
Traditionally, electrochemical sensors have not been installed within
subsurface barrier systems. However, within the past five years, advances
have been made in sensor development for underground storage tanks.
Assessments are being made for installation of sensors in barrier systems.
Betsill and Gruebel (1995) cite literature surveys that identified about 25
chemical sensor companies in the United States.
375
Fiber optic sensor inside a

protective stainless steel housing
Solar cell
Light into fiber
Electronics
Optical fiber
with cladding
Carbon
Tetrachloride
Bare core of
optical fiber
Flourescent light
out of fiber
Stainless steel
housing
Flourescent material
on optical-fiber core
Perforated stainless
steel fiber tip guard
Figure 12-11 Schematic of a reversible fiber-optic carbon tetrachloride sensor

(U.S. DOE, 1994b)
12.2.5
Electrical Systems
The term electrical systems are herein used generically to cover techniques
that involve the passage of electric current through subsurface media or
physical interactions between an embedded device and surrounding media
that may lead to the transmission of electrical impulse to direct-read systems.
These systems include direct-current (DC) resistivity monitors, piezometers,
infiltrometers, and flow velocity sensors.
Beam
Collimating
Filter splitter Lens
lens
Light
source
Filter
Sensor
Lens
Flourescent coating
Micro
processor
A/D
Photodiode
Optical fiber
Signal readout
Alarm vs Set point
Fiber-optical chemical
sensor (end-coated)
Figure 12-12 Block diagram of a fiber-optic sensor fluorimeter (DeFilippi

and Cody, 1994)
a. Direct-Current (DC) Resistivity. This method can be used to detect flaws
in geomembranes and the boundaries of a contaminated zone around a
containment system. Typically, the frequencies used are so low (under 10
376
Hz) that no significant electromagnetic current is induced in the ground. The

currents are generated by buried electrodes. As in the case of EM resistivity,
the properties of the transport medium affect its electrical conductivity or
resistivity.
Current source
electrode
Remote current
return electrode
I
I
Earth
Moving
measurement
electrodes
Earth
Liquid
Leak path
Membrane liner
Current
flow lines
Figure 12-13 Sketch of the electrical leak location method for geomembrane
liners (Darilek and Parra, 1988)
For containment systems with barriers that are mostly horizontal and
near the ground surface (landfill covers and impoundments), measurements
of electrical self potential can be used to locate leaks. As illustrated in Figure
12-13 (Darilek and Parra, 1988), two electrodes are used at a membrane liner
of a surface impoundment. One of the electrodes is fixed at a location remote
from the impoundment while the other electrode is operated in a scanning
mode over the area to be investigated. In the absence of leaks, the voltage
200
180
160
220
140
240
120
260
100
Current injection
280
80
300
60
320
Leak
340
40
0
20
Figure 12-14 Equipotential lines on a liner indicating a leak on the 340 radial (Peters et al., 1982)
377
imposed on the liner produces a uniform distribution of potential across the

surface of the liner. At a leak, conductive fluids permeate the underlying soil
to close the electric circuit by producing a path for current flow through the
soil to the remote electrode. Figure 12-14 (Peters et al., 1982) illustrates the
recording of an anomaly that can result in the equipotential lines around a
leak. Due to the high electrical resistivities of geomembranes (data by Peters
et al., 1982, suggest values of 108-1013 ohms), they have been the primary
components investigated using this method.
This technique has been successfully used to locate leaks in a one-acre
geomembrane-lined impoundment (Schultz et al., 1984). This impoundment
facility had design dimensions of 65.8m by 65.8m. Leaks in the 100-ml high
density polyethylene liner were identified by the distortion of equipotential
voltage lines plotted from DC-resistivity data. Darilek and Parra (1988)
conducted similar tests on geomembranes placed between soil covers (see
Figure 12-15). Abrupt decreases in recorded voltage during radial traverses
from an electrical current injection point indicate the locations of leaks. Thus,
the electrical resistivity method can be used for leak location in geomembranes,
regardless of whether it is covered by a soil or a liquid. However, the mass of
soil over a geomembrane may reduce signal strength and, therefore, require
a reduction of the separation distance between the scanner and the
geomembrane.
Voltage/Current (volts/amp)
Leak diameter=3mm
15.2 cm Soil
20
25.4 cm Soil
30.5 cm Soil
6.1 cm Soil
Distance (meters)
Figure 12-15 Anomalies in voltage/current ratio recorded during electrical

resistivity scans of various thicknesses of cover above a
geomembrane (Darilek and Parra, 1988)
378
b. Electrical Piezometers and Infiltrometers. The moisture contents of barriers

can affect infiltration rates and barrier structural integrity. Thus, the moisture
content of earthen barrier components can directly influence barrier
performance. Electrical piezometers and infiltrometers are frequently used
to measure pore water pressure and infiltration rates of moisture in earthen
materials, respectively. Sensors can be buried in barrier system components
and data from these instruments can be converted to electrical impulses for
automatic and continuous recording. The most common piezometers operate
through the displacement of a fluid inside a tube due to the build-up of a
pore water pressure outside the tube. In the arrangement illustrated in Figure
12-16, a U-shaped wire is immersed in a mercury tube and connected to a
Wheatstone bridge circuit. The electrical resistance of the wire is proportional
to the displacement of the mercury level. An increase in local pore pressure
depletes the mercury in the left column and unbalances the Wheatstone bridge
of the circuit. As the pore pressure increases in the left tube, the resistance of
the left wire increases while that of the right wire decreases. Thus, the voltage
output is directly proportional to the measured difference in resistance, R.
Equation 12-4 relates the parameters involved.
Vo = V[R/R] = bV[P]
Vo
R
P
b
=
=
=
=
(12-4)
voltage output
resistance of the wire
pressure input
scaling factor
Some methods of barrier infiltration analysis, exemplified by the water

budget methods, involve the use of mathematical models for computation of
the rate of migration of moisture through barriers located in the unsaturated
zone. This is commonly done for horizontal barriers such as landfill covers
and concrete covers. Tensiometers embedded at various depths in a barrier
can be used to track the advance of the wetting front through the barrier
during and after hydrological events. Soil suction, which is measurable
through the use of tensiometers, reduces to zero under saturated conditions.
By monitoring the time distribution of these suction decreases with respect to
the depth locations of the tensiometer sensors within the barrier, infiltration
rates can be determined.
Several additional physico-chemical principles can be applied to measure
moisture content in partially saturated barrier materials so long as water
movement by gravitational forces is not significant. Appropriate instrumentation
can be installed inside and outside barriers in the subsurface (U.S. EPA, 1993b).
The principles of a few of these techniques are briefly outlined below.
12.3 ASSESSMENT OF VARIOUS MONITORING TECHNOLOGIES
379
+V
R
Pressure
Electric
piezometer
Vout
Mercury
R
Wire
Cover
Signal processor
and readout
instrument
Figure 12-16 Schematic of an electrical piezometer embedded in a landfill

cover (not drawn to scale) (Inyang, 1994b)
Gamma ray attenuation: the degree of attenuation of gamma rays in

soil depends on the bulk density and moisture content of the soil.
Thus, variations in moisture content can be scaled using gamma ray
attenuation data.
Electrical resistance of embedded gypsum blocks: the electrical
properties of an embedded gypsum block changes in proportion to
the volume of moisture it absorbs from the surrounding soil.
Thermocouple psychrometry: the relative humidity of voids within
soils is directly related to soil water potential. The latter is the sum of
the osmotic and matrix potential. Thus, curves can be developed to
relate relative humidity, water potential, and osmotic potential for
unsaturated conditions.
Heat dissipation: the rate of heat dissipation into the surrounding
soil from an imbedded heat source is affected by the moisture content
of the soil. Using calibration curves developed for the soil/heat source
combination in the laboratory, in situ moisture content measurements
can be made.
12.3
ASSESSMENT OF VARIOUS CATEGORIES OF MONITORING

TECHNOLOGIES
12.3.1
System Cost-Effectiveness Factors
The cost-effectiveness of the monitoring systems discussed in the preceding

sections depends on the following factors:
380
12.3.2
spatial coverage around the containment source,

resolution and tributary region of the technique and sensor,
release-to-detection time interval,
feasibility of automatic operation, and
invasive versus non-invasive operation.
Cost-Effectiveness of Groundwater Well Systems
Currently, groundwater monitoring wells and leachate collection sumps are

the most common systems for monitoring containment system performance.
In the United States, waste containment units placed into inactive status prior
to November 19, 1980 are generally not required to comply with groundwater
monitoring requirements. The common monitoring well configuration (one
well upgradient and three wells downgradient around the containment
system) is inadequate for cost-effective monitoring of containment system
performance since the release-to-detection time interval may be too long,
allowing contaminants to short-circuit the monitoring system because of
inadequate spatial coverage provided by only a few. Leachate collection and
testing is labor-intensive because it involves a sequence of tasks, including:
collection, storage, and analytical testing of fluids. Furthermore, the
concentrations of contaminants in samples collected from observation wells
represent composite values and do not provide information on concentrations
at specific exit points from the containment system. In a recent study of 890
leaking waste containment facilities (GAO, 1995), 107 facilities had
groundwater monitoring systems incapable of detecting the releases. An
additional 153 facilities had monitoring systems that did not provide adequate
information for assessing the extent of contaminant releases.
12.3.3
Cost-Effectiveness of Geophysical Techniques
Geophysical methods are cost-effective when used to assess the continuity

and dimensions of emplaced and natural barriers. Large areas can be
monitored and some methods are non-invasive. However, the resolution of
some geophysical techniques such as GPR, seismic refraction, and resistivity
profiling may be insufficient for detecting flaws which may serve as moisture
and contaminant flow channels through barriers.
12.3.4
Cost-Effectiveness of Electrochemical Sensing Techniques
Electrochemical sensing systems can provide higher spatial coverage at lower

cost. Unlike the case of groundwater monitoring wells, the unit cost of
increasing monitoring points is relatively modest. Data can be acquired from
sensors automatically, and the release to detection time interval can be
minimized by placing a dense network of sensors around the containment
system. Electrochemical monitoring systems do have one drawback in that
381
sensors are usually fabricated to detect a narrow range of contaminants,

implying that the contaminants (or class of contaminants) have to be identified
a priori. Furthermore, sensor repair and reinstallation techniques have not
yet been developed fully.
12.3.5
Relative Cost-Effectiveness of Various Techniques
The cost-effectiveness of the various monitoring systems depends on the

specific performance aspects to be monitored, the desired accuracy, and the
frequency of measurement. Each technique (for example, geophysical
systems) involves several methods that may have widely different levels of
cost-effectiveness when used to monitor a specific aspect of containment
system performance. Thus, the patterns of cost-effectiveness illustrated in
Figure 12-17 (Inyang, 1994b) are intended to represent general trends only.
Geophysical techniques may be the most cost-effective for external
contaminant concentration monitoring over large areas, and less cost effective
for small monitoring areas, due to the large capital cost investment in
equipment. Electrochemical sensor systems and groundwater well monitoring
systems exhibit a decrease in cost-effectiveness as the monitoring zone
increases in size. This is attributable to the high number of wells or sensors
that need to be installed and operated to maintain an acceptable level of
probability that contaminant leaks will be detected.
Cost-Effectiveness of
Monitoring Technique
Closely spaced
electrochemical
sensing systems
Geophysical
techniques
Array of groundwater
monitoring wells
Size of the Required Monitoring Zone

Around a Waste Containment System
Figure 12-17 Conceptual relationship between the size of monitored area

and the cost-effectiveness of various monitoring methods
(Inyang, 1994b)
382
To illustrate the relative cost-effectiveness of the techniques shown in

Figure 12-17, a hypothetical landfill monitoring case is used as an example.
The cost estimates used are drawn from Inyang (1994b). The hypothetical
landfill contains oily wastes, requiring a 20-year post-construction monitoring
period. For this analysis, the monitoring system design objectives involve:
detection of any volatile organic compounds or metals released into

the groundwater, and
location of the contaminated groundwater plume if a release occurs.
The selected arrangement of monitoring points for each of the techniques

complies with regulatory guidelines (U.S. EPA, 1991d). The monitoring
networks selected are approximately of the same effectiveness, in order to
facilitate comparison of cost-effectiveness. However, the size of the area to be
monitored has been varied in order to assess the extent to which cost is affected
by changes in areal coverage.
Cost estimates for off-site equipment are not included in this example
since, in most monitoring programs, analytical testing is usually done off-site
by a certified laboratory. Such off-site equipment usage is included in the
service charges. Thus, only on-site equipment costs are included in capital
cost estimates. Inflation has been factored into the operation and maintenance
costs; thus, annual average costs given have been computed for the specified
20-year post-construction monitoring period.
The example site conditions include an aquifer ten feet below the bottom
landfill liner; thus, the primary concern is leachate contamination of this
aquifer. A semicircular area of 400 ft radius, bordering the landfill on the
down-gradient side, is to be monitored. The number of monitoring wells or
sensor locations required can be estimated using the following relationships:
Am = R2/2
nm = Am/at = Am/S2
Am
R
nm
at
S
(12-5)
(12-6)
= semicircular area to be monitored [L2]

= radius of Am [L]
= number of monitoring wells or sensor locations
= area to be serviced by each well or sensor [L2]
= linear spacing of monitoring wells or sensors [L]
For this example, S = 80 ft, approximating the average of values reported

in the literature (McNeil, 1982 and Jensen et al., 1992), and R = 400 ft, giving
Am @ 251, 327 ft2, and nm @ 39. The required number of wells or sensors for
different values of Am are presented in Table 12-5.
383
TABLE 12-5 The relationship between the size of the monitoring area and the
required number of monitoring points for the hypothetical case
REQUIRED NUMBER OF
MONITORING POINTS
(approximated to the next
whole number)
MONITORING AREA
(ft2)
LINEAR SPACING OF
MONITORING POINTS
(ft)
20,000
80
50,000
80
16
150,000
80
24
200,000
80
32
250,000
80
40
The following assumptions have been made in computing and comparing

the 20-year costs of installing and operating the three monitoring systems:
Groundwater Monitoring Wells
- installation cost: $5,000 per well (Weber et al., 1991)
- chemical analyses: Samples collected semiannually and tested in
accordance with protocols described in SW-846 (U.S. EPA, 1991d). Cost
estimates are as follows (Brown, 1993): EPA Method 8260 for VOCs at
$600 for two tests per well per year, amounting to $12,000 per well for 20
years; and EPA Method 6010 for metals at $300 for two tests per
monitoring well per year, amounting to $6,000 per well for 20 years.
- operational and management costs: Labor and overhead costs associated
with sample collection, system maintenance, and record keeping
estimated at $5,000 per year, amounting to $100,000 for 20 years (not
dependent on number of monitoring points).
Electrochemical Sensing
- data processor: $5,000 (quotation by Raychem Corp.)
- below ground sensor cables: 100 ft length @ $12/ft = $1,200 for each
sensor (Raychem Corp.)
- above ground connecting cables: $300 per sensor
- installation costs: $400 per sensor
- operational and management costs: estimated at $6,000 per year, or
$120,000 for 20 years
Geophysical System
- electromagnetic resistivity: measurements at pre-selected locations semiannually
- outfitted vehicle: $20,000 (for acquiring measurements)
- data processor/display: $15,000
- electromagnetic wave generator and receivers: $10,000
- data management/interpretation software: $1,500
384
- operational and maintenance costs: $9,000/yr, or $180,000 for 20 years

(vehicle and system maintenance, labor costs, transportation costs, etc.)
The cost-estimates for the three monitoring methods are summarized in
Table 12-6. In this hypothetical example, geophysical sensing has the highest
cost for the smallest monitoring area (20,000 ft2). However, for a monitoring
area of 250,000 ft2, the cost of geophysical resistivity sensing if only about
23% of that required for groundwater monitoring using wells. Electrochemical
sensing is the most cost-effective system for the example given here. It is
conceivable that the geophysical sensing system could become the most costeffective for areas greater than 250,000 ft2.
TABLE 12-6
Cost comparisons of the containment detection techniques

for three monitored area sizes using the assumptions
discussed in the text
20-YEAR COST FOR AREA OF SIZE, Am, TO BE MONITORED
MONITORING
TECHNIQUE AND COST
ITEM
Am = 20,000 ft2
UNIT
Total
COST ($) Number
Required
Item
Cost ($)
Am = 150,000 ft2
Number
Required
Total
Item
Cost ($)
Am = 250,000 ft2
Number
Required
Total
Item
Cost ($)
GROUNDWATER
MONITORING WELLS
Well installation
Chemical analyses
Operation and management
[(a) per well for 20 years]
5,000
18,000 (a)
100,000
3
3
-
20-YEAR TOTALS
15,000
54,000
100,000
24
24
-
169,000
120,000
432,000
100,000
40
40
-
652,000
160,000
720,000
100,000
980,000
ELECTRO-CHEMICAL
SENSING
Central electronic unit
Sensing cables
Connecting cables
Sensor installation
5,000
1,200
300
400
120,000
1
3
3
3
-
20-YEAR TOTALS
5,000
3,600
900
1,200
120,000
1
24
24
24
-
130,700
5,000
28,800
7,200
9,600
120,000
1
40
40
40
-
170,600
5,000
48,000
12,000
16,000
120,000
201,000
GEOPHYSICAL SENSING
Field vehicle
Wave generator and
receivers
20,000
Data interpretation software 10,000 (b)
Data processor and
1,500
displayer
15,000
180,000
[(b) per set]
20-YEAR TOTALS
1
1
1
1
-
20,000
10,000
1,500
15,000
180,000
226,500
1
1
1
1
-
20,000
10,000
1,500
15,000
180,000
226,500
1
1
1
1
-
20,000
10,000
1,500
15,000
180,000
226,500
12.4 CONTAINMENT SYSTEM MONITORING NEEDS
12.4
CONTAINMENT SYSTEM MONITORING NEEDS
12.4.1
Sensor Performance Needs
385
The performance of sensors plays a key role in the quality of monitoring data
obtained for use in the analysis of waste containment system effectiveness.
The most critical sensor performance needs are summarized as follows:
12.4.2
sensor must perform in aggressive physico-chemical environments,

sensors need high bandwidths so that a wide array of contaminants
can be detected,
sensors must be immune to electromagnetic interference but maintain
high sensitivity to contaminants (in the range of ppb to ppm),
sensors must perform dynamically (record contaminant concentrations
regardless of sequence: high before low, or low before high), and
in situ sensors that can track contaminant and/or moisture flow rates.
Barrier-type Specific Needs
A summary of techniques for monitoring each type of containment component

is presented in Table 12-3. The most important monitoring needs for specific
barrier types are summarized as follows:
a. Indigenous and Artificial Floors, and Chemically Based Grouts
monitoring systems that can delineate the perimeter and continuity
of the barrier,
sensors that can detect leaks with spatial resolution on the order of a
few centimeters,
high resolution imaging systems, and
tracers that can effectively delineate possible pathways for moisture
and contaminant transport through the barrier.
b.
Caps (Soil-based and Concrete)

monitoring systems that can detect and quantify settlement and sideslope deformation,
monitoring systems that can detect the creation of cracks and fissures,
monitoring systems that quantify moisture content and flow rates at
critical points within the cover,
monitoring systems that can detect excessive deterioration in concrete
quality, and
monitoring systems that can accurately detect relative movements
between geomembranes and cover soils.
c.
Sheet Piles and Geomembranes

interstitial monitoring system for double-wall configurations,
386
monitoring systems that can measure the tightness of interlocks,

seams, and keys into geomedia,
sensors that can detect flaws through which contaminants can flow,
in situ sensors that can measure the rates of barrier deterioration
processes, and
geotechnical monitoring methods for barrier deformation.
d. Permeable Reactive Barriers

monitoring system for in situ determination of contaminant
concentration and flow velocity upgradient and downgradient,
sensors for tracking the concentrations of reactants and products at
points within reactive barriers, and
monitoring systems for verification of plume capture.
e.
Soil and Cement-based Vertical Barriers

systems for monitoring hydraulic heads on either side of the vertical
barrier,
devices for in situ measurement of hydraulic conductivity,
methods for non-intrusive determination of wall embedment,
automatic moisture sensing systems for wall portions above the water
table, and
monitoring systems for measuring contaminant concentrations upgradient and downgradient from vertical walls.
12.4.3
Installation and Retrieval Needs
Emplacement techniques need to be made more cost-effective, since sensors

embedded in containment barriers during construction may need to be
replaced during the design lifetime of the system. There is a need for minimally
intrusive retrieval and reinstallation methods for sensors. The use of remote
data acquisition techniques could eliminate the need for direct connection of
sensors to above ground data processing systems. A method for pinpoint
location of isolated sensors in subsurface barriers is needed, since ground
movements may displace sensors from their original positions during the
design lifetime of the containment system.
12.4.4
Monitoring Network Needs
Betsill and Gruebel (1995) have presented a detailed discussion of monitoring

system network needs. The most critical needs are as follows:
automated data and sample collection analysis system that can record
and evaluate changes in contaminant concentrations and barrier
characteristics (a sensor network is depicted in Figure 12-18),
12.4 CONTAINMENT SYSTEM MONITORING NEEDS
Electro-chemical and
moisture sensors
Optional protective
soil, concrete, etc.
Berms
Waste pile
Filter medium
Leachate collection system
387
Pipe
Compacted low permeability clay liner

2
3
Contaminated
soil
Native soil
Figure 12-18 Schematic showing installation points of electrochemical and

moisture sensors for detection of contaminants beneath a waste
pile (Inyang, 1994b)
passive monitoring systems and alarms (smart sensors) that can alert
personnel to failures and malfunctions,
data management and archiving systems that can eliminate the need
for excessive duplicate data sets and simplify future data retrieval
and analysis, and
spatial referencing systems for embedded sensors that can indicate
barrier conditions at specific locations in the subsurface (these systems
can be integrated with tomography as illustrated in Figure 12-19).
Source
current
Signal
source
Voltage
ERT tomograph
Electrodes
Electrodes
Boreholes
Figure 12-19 A circuit for an electrical resistance tomography system (U.S.

DOE, 1994a)
388
12.4.5
User Interface Needs

operational computer software for acquisition, interpretation, and
visualization of spatial and temporal monitoring data,
facilities for testing of sensor performance in real-time dirty
environments that simulate field conditions,
development of query systems for sensor networks and systems for
integrating and reconciling monitoring data obtained using different
techniques,
data analysis systems that can identify out-of-compliance situations,
a central data bank for storage of data that can be analyzed to show
relationships among containment system failure, site characteristics,
waste types and system design, and
user-friendly systems for manipulating collected data to verify and
refine both damage accumulation and contaminant transport models.
12.5
SUMMARY AND RECOMMENDATIONS
12.5.1
Barrier Component-Specific Monitoring Methods
There are several monitoring techniques which can be used to track the postconstruction performance of waste containment systems. In most cases, it is
necessary to combine a number of monitoring techniques in order to acquire
the variety of data required to perform comprehensive analyses. Everett et al.
(1984a and 1984b) discuss the applications of various vadose zone monitoring
techniques. The techniques discussed here focus on flow monitoring.
Table 12-4 provides a general summary of the applications of various
monitoring techniques to different types of flaws and deterioration modes in
barriers, moisture and contaminant flow rates through barriers, and the
distribution of contaminant concentrations around barriers. Table 12-3 offers
information on the general utility of the categories of monitoring techniques
to the three main monitoring objectives: integrity monitoring, permeation
monitoring, and external concentration monitoring. Table 12-4 can be used
in conjunction with Table 12-3, with consideration of the following additional
points.
In all cases, groundwater wells with installed piezometers constitute

the most common monitoring method currently used.
A double-walled internally monitored system is a potentially effective
monitoring method for vertical walls, but it is not being implemented.
For indigenous floors, vertical borings/wells can be driven into the
top of the floor, and horizontal wells can be driven below the floor.
The horizontal well can also be used for contaminant recovery, if
necessary.
12.5.2
389
For all barrier types, especially, newly developed systems exemplified

by permeable reactive walls, monitoring systems should be more
conservative for field demonstrations than for commercial
applications.
Liquid and gas tracers are particularly useful for verifying the integrity
of cryogenic barriers.
Piezocone soundings and other penetrometer-borne sensors have
been effectively used to measure in situ chemical concentrations and
hydraulic conductivities in slurry walls. It may be necessary to use
grouts to prevent cross-bed contamination.
Moisture content sensors have been used effectively in cap system
monitoring. However, the following limitations and necessary
improvements apply:
- gypsum blocks saturate easily and are suitable for short-term use;
- neutron moisture sensors need calibration standards;
- pressure-driven vacuum lysimeters need minimization of
gradient to reduce changes in pore-fluid chemistry, and bubbling
pressure needs to be sufficient to operate over the full range of
capillary pressures;
- frequency and time-domain reflectance methods are susceptible
to soil moisture salinity.
Adoption of Some Available Innovative Technologies
The merger of material science, structural mechanics, sensor and actuator

technologies, and advanced data processing techniques has accelerated the
development of smart structures or intelligent structures. This
development has provided opportunities for incorporating effective neural
and responsive systems into waste containment monitoring systems. The
intelligent structure concept and technology are being refined for automobiles,
robots, high performance aircraft, bridges, and highways. Within the next
few years, it should be possible to use these systems to track environmental
and physical changes within and around containment systems, interpret such
changes, and devise appropriate responses to these changes. The evolving
generation of fiber optic sensors promises to be low-cost, lightweight, immune
to electromagnetic interference, and rugged enough to be incorporated directly
into monitored barrier components. They will have high bandwidths and
will perform under temperature extremes. For example, silica-based fibers
have an upper operating temperature range of 700C.
Sensor research conducted by private industry is another potential
resource for monitor development. For example, at the Microswitch Division
of Honeywell, researchers have developed the concept of smart sensors for
heavy manufacturing applications. These sensors use on-board intelligence
to link themselves to programmable logic controllers (PLCs) and transmit
information regarding their working status (e.g., calibration data) to evaluate
390
the systems ability to function within designated parameters (Murray, 1994).

Other companies, such as Motorola, Philips Components, and Texas
Instruments are working on smart sensor technology (Ormund, 1993).
Although smart sensors are currently utilized primarily for heavy
manufacturing, the technology may provide insight for developing smart
monitoring sensors for environmental applications.
Some of the additional specific sensors that have been developed or are
being developed include:
sensors for use with the cone penetrometer for site characterization
activities, including soil moisture and chemical concentration
measurements;
down-hole x-ray fluorescence metals analysis;
reversible, fiber optic sensors for detection of specific VOC
contaminants (e.g., methylene chloride) in the parts per billion (ppb)
range;
real-time detection of radionuclides and other contaminants during
drilling;
robust organic vapor monitors for harsh field environments and
conditions;
on-site analysis of metals using adsorptive stripping analysis;
portable apochromatic/fiber optic detector for VOCs in air and
watera potential sensor with alarm for initial detection in field
applications; and
borehole liners with real-time VOC monitoring systems.
There is the possibility of developing barrier materials that can be more

easily monitored after placement. The physical characteristics of such barriers
should contrast sharply with those of the host media. Also, ultra-conductive
grouts would render geophysical techniques more effective as monitoring
techniques. A number of innovative technologies, developed as part of the
U.S. Department of Energy (U.S. DOE) Environmental Management Programs,
are described below.
a. Heavy-Weight Cone Penetrometer.
This truck-mounted ground
penetration device consists of a rod with a 1.75 inch diameter conical tip. The
cone penetrometer can be equipped with instruments to monitor for
contaminants and geophysical data as the penetrometer is pushed into the
ground or to leave monitoring devices in place as the penetrometer is
withdrawn. Specialty penetrometer probes are capable of measuring electrical
resistivity, hydraulic conductivity, gamma radiation, temperature, soil density,
fluorescence (laser induced), pH, etc.
b. Seamist. This is a liner that is deployed inside an open borehole or a well
and instrumented with sensors and fluid samplers to assist with in situ
391
characterization, monitoring, and remediation. Applications include soil

vapor and groundwater sample collection.
c. Halsnif. This is a portable detection device that interacts helium with
radio frequency signals in an excitation chamber to measure chlorinated
compound concentrations in real time. The device can be used with the cone
penetrometer.
d. Cross Borehole Electromagnetic Imaging. This is a source and plume
detection technique that uses continuous wave and pulsed radar systems to
provide a means of imaging the subsurface with plumes of nonaqueous phase
liquids, drums, tanks and cylinders. It provides a mapping of electrical
permittivity between boreholes.
e. Rapid Geophysical Surveyor (RGS). The RGS characterizes a buried
waste site by collecting high quality, dense sets of magnetic data. It is a
manually-maneuvered vehicle that carries a cesium total field magnetometer,
a data logger, and data storage hardware and software.
f. Surface Acoustic Wave Sensors (SAWS). This sensor is being investigated
by the DOE/EM for contaminant detection. However, the Penn State
University has also developed a version of SAWS for implantation in the
roadbed of highways or bridges to detect ice formation, flooding, dangerously
heavy traffic, and cracks forming in support structures. Each sensor contains
a miniature microwave antenna and a piezoelectric transducer that converts
microwaves into sound waves. As the sound waves travel through the
structures, any strain or physical abnormalities on the structures (bridges or
roads) alters the speed and strength of the sound waves which can be detected
by other SAWS. The sensors are small and can be emplaced throughout the
structure. Signals are fed back to a control center for analysis.
g. Colloidal Borescope. The borescope is used as an in situ instrument to
directly observe the motion of colloidal size particles, thereby determining
groundwater flow direction and rate. The instrument consists of a charge
coupled device (CCD) camera, an optical magnification lens, an illumination
source, a downhole compass to assess direction of natural flow, and a
watertight stainless steel housing. It is approximately 60cm long, with a
diameter of 4.4cm. The electronic image is transmitted to the surface by a
fiber optic cable and the image is viewed on a high resolution monitor. The
instrument can be used in a well as small as 5cm in diameter and
measurements can be obtained in about 30 min. The colloidal borescope is
capable of determining the vertical and spatial distribution of local
groundwater velocity, both magnitude and direction. It is capable of these
measurements in low and high hydraulic conductivity materials.
392
Chemical manufacturing companies are currently installing automatic

materials inventory data acquisition systems. This technology is adaptable
to contaminant leak detection. The Eco Instrumentation Corporation (EIC)
Remote Access Inventory Monitoring System is illustrated in Figure 12-20.
Liquid level is monitored using fiber-optic or other sensors. The sensors
transmit data to a processing unit at the site. Using ratio frequency signals,
the interpreted information is broadcast to computer systems at offices or
control centers off-site. The operator can query the system after the occurrence
of a natural hazard such as an earthquake or flood to determine the extent of
facility damage and loss of stored fluids.
Facility Management Center
Cover
Fiber-optic
monitoring layer
Radioactive
waste
Pre-monitor barrier
External barrier
Radio frequency
signal
Data
processing
unit
Figure 12-20 Conceptual design of EIC remote access inventory monitoring

system
On-site Analyses
In-situ senors
Innovative
Drilling
Membrane
Liners
Real-time
Field Analysis
Borehole
and In-situ
Measurements
Directional
Boring and
Sonic Drilling
Passive or
Active Sampling
Subsurface Monitoring
and Sampling
Borehole
and Sampling
Optimization
Screening
Minimally and
Nonintrusive
Characterization
Historical
Records
Subsurface
Imaging
Data Evaluation
and Risk Analysis
Figure 12-21 The Landfill Assessment and Monitoring System (LAMS) (U.S.
DOE, 1995)
393
The U.S. Department of Energy has developed the Landfill Assessment

and Monitoring System (LAMS). This system (illustrated in Figure 12-21) is
based on the application of a systems approach to monitoring, with resulting
maximizations of data gathering, and minimization of costs, worker exposure,
and sampling time. The Expedited Site Characterization (ESC) program
focuses on the use of a variety of non-intrusive and minimally intrusive
technologies (e.g., surface geophysics, cone penetrometers, hydropunch
sampling, vegetation sampling) to optimize sampling locations and minimize
monitoring well installation (U.S. DOE, 1995).
12.5.3
Use of Multi-Phased Monitoring Approach
A systematic approach to the selection, implementation, and operation of a

containment system monitoring strategy should be adopted. Considering
that there are several alternative technologies, monitoring objectives, and
barrier configurations, a single technology may not be cost-effective for all
applications. Inyang (1994b) has proposed the approach illustrated in Figure
12-22 as the sequence of steps that can be used to select and implement a
containment system monitoring strategy. Each step involves analyses which
provide input into subsequent steps and monitoring activities.
Stage
Hazard identification
Potential failure mode

assessment
Result
Risks
Relevant phenomena
and mechanisms
Identification of vulnerable
components
Potential failure points and areas,

relevant factors of safety
Formulation of the hierarchy

of monitoring technologies
Technical feasibility, durability and

maintenance requirements
Selection of monitoring
system configuration
Monitoring system
implementation
Figure 12-22
Sensor location points; combination of

techniques; data acquisition and processing
plans; and cost-effectiveness data
Sensor calibration, installation and
system testing (where possible)
Flow chart of the evaluation and selection process for

innovative monitoring technologies for subsurface waste
containment barriers (Inyang, 1994b)
394
12.5.4
Specification of System-Specific Failure Rate
Natural and constructed facilities exposed to environmental and humaninduced stresses deteriorate as time progresses. This is especially true of waste
containment systems because they may be exposed to aggressive environments
for time periods that range from decades to centuries. Inyang (1994a), Kargbo
et al. (1993), Flemming and Inyang (1995), Kim and Kim (1991) and Peterson
(1990) have discussed containment system deterioration processes. However,
structural deterioration of some components of a multicomponent system may
not always lead to total functional failure of the system. A uniform approach
needs to be developed for specifying the failure condition of containment
systems. Monitoring data can be combined with numerical models to forecast
future performance levels, risks, maintenance requirements, and clean-up
logistics for constructed containment systems.
12.6
REFERENCES
Bass, J.M., Lyman, W.J. and Tratnyek, J.P. (1985). Assessment of synthetic
membrane successes and failures at waste storage and disposal sites.
Project Summary, EPA/600-S2-85/100. U.S. Environmental Protection
Agency, Cincinnati, Ohio
Betsill, J.D. and Gruebel, R.D. (1995). Current research options for in situ
monitoring and verification of contaminant remediation and containment
within the vadose zone. VAMOS-Verification and Monitoring Options
Study, TTP #AL2-2-11-07, Environmental Restoration Technologies Div.,
Sandia National Laboratories, Albuquerque, New Mexico
Brown, J. R. (1993). An evaluation of leak detection indicator parameters at
hazardous waste disposal facilities. Project Report submitted in partial
fulfillment of the requirements for the M. S. degree, Virginia Polytechnic
Institute and State University, Blacksburg, VA.
Byrne, R.J., Kendall, J. and Brown, S. (1992). Cause and mechanism of failure
of Kettleman Hills B19, phase 1A. Proceedings of ASCE Specialty Conference
on Stability and Performance of Slopes and embankments-11, Berkeley,
California, pp. 1188-1215.
Darilek, G.T. and Parra, J.O. (1988). The electrical leak location method for
geomembrane liners. Proceedings of the 14th Annual Research Symposium
on Land Disposal, Remedial Action, Incineration and Treatment of
Hazardous Waste, Cincinnati, Ohio, EPA/600/9-88/021, pp. 167-176.
De Filippi, R. and Cody, T.C. (1994). Fiber optics crest next wave of leak
12.6 REFERENCES
395
detection technology, Soils, January-February, pp. 14-17.

Dietrich, R.J. and Salley, J.R. (1975). Static and dynamic strain measurements
for embankments. ASTM STP 584 on Performance Monitoring for
Geotechnical Construction. American Society for Testing and Materials,
Philadelphia, Pennsylvania, pp. 182-194.
Dodt, M.E., Sweatman, M.B. and Bergstrom, W.R. (1987). Field
measurements of landfill surface settlements. Proceedings of the ASCE
Specialty Conference on Geotechnical Practice for Waste Disposal, Ann Arbor,
Michigan, pp. 406-417.
Dunnicliff, J. (1982). Geotechnical instrumentation for monitoring field
performance. NCHRP Report 89. National Cooperative Highway
Research Program, Transportation Research Board, National Research
Council, Washington, D.C.
Everett, L.G., Wilson, L.G., and Hoylman, E.W. (1984a). Constraints and
categories of vadose zone monitoring devices. EPA/600/J-84/285.
Environmental Monitoring Systems laboratory, Las Vegas, Nevada: .
Everett, L.G., Wilson, L.G., and Hoylman, E.W. (1984b). Vadose zone
monitoring for hazardous waste sites. Noyes Data Corporation, Park
Ridge, New Jersey.
Fleming, L.N. and Inyang, H.I. (1995). The permeability of clay-modified
fly-ash under thermal gradient. Journal of Materials Engineering, ASCE,
Vol. 7, No. 3, pp.178-182.
GAO. (1995). Compliance with groundwater monitoring requirements at
land disposal facilities. Briefing report to the ranking minority member,
Committee on Governmental Affairs, U.S. Senate. GAO/RCED-95-75BR,
U.S. General Accounting Office, Washington, D.C.
Hoekstra, P., Lahti, R., Hild, J., Bates, C.R. and Phillips, D. (1992). Case
histories of shallow time domain electromagnetics in environmental site
assessment. Ground Water Management and Research, Fall, pp. 110-117.
Iman, R.L., Helton, J.C. and Johnson, J.D. (1990). A methodology for grouping
source terms for consequence calculations in probabilistic risk
assessments. Risk Analysis, Vol. 10, No. 4, pp. 507-520.
Inyang, H.I. and Tumay, M.T. (1995). Containment systems for contaminants
in the subsurface. A chapter in the Encyclopedia of Environmental
Control Technology, Gulf Publishing Company, pp. 175-215.
396
Inyang, H.I. (1994a). A Weibull-based reliability analysis of waste

containment systems. Proceedings of the 1st International Congress on
Environmental Geotechnics, Alberta, Canada, pp. 273-278.
Inyang, H. I. (1994b). Cost-effective post construction integrity verification
monitoring and testing techniques for subsurface barrier containment
facilities. Technical Analysis Document, DuPont Chemicals, Wilmington,
Delaware, 78 pages.
Inyang, H.I. and Tomassoni, G. (1992). Indexing of the long term effectiveness
of waste containment systems for a regulatory impact analysis. Draft
Technical Guidance Document. Corrective Action Branch, Office of Solid
Waste, U.S. Environmental Protection Agency, Washington, D.C.
Inyang, H.I. (1991). Hazardous waste facilities in seismic zones. AAAS/
U.S. EPA Environmental Science and Engineering Fellowship report.
American Association for the Advancement of Science, Washington, D.C.
Jansen, J., Haddad, B., Fassbender, W., and Jurcek, P. (1992). Frequency
domain electromagnetic induction sounding surveys for landfill site
characterization studies. Groundwater Monitoring and Research, Fall,
pp. 103-109.
Kahle, R.L. and Rowlands, L. (1982). Overview of low-level radioactive
waste disposal facilities subsidence. Proceedings of the symposium on lowlevel waste disposal, NUREG/CP-0028, Washington, D.C., pp. 223-251.
Kardos, J.J. and Ennis, G.B. (1993). Subsurface interface as an investigative
tool. Public Works, September, pp. 82-83.
Kargbo, D.M., Fanning, D.S., Inyang, H.I. and Duell, R.W. (1993). The
environmental significance of acid sulphate clays as waste covers.
Environmental Geology, vol. 22, pp. 273-278.
Kim, C. and Kim, J. (1991). Source term model for radioactive waste
repository. Proceedings of High Level Radioactive Waste Management, Las
Vegas, Nevada, pp. 1049-1053.
King, R.R. and Kusenburger, F.N. (1973). Strain sensing. In: Nondestructive
Testing. NASA SP-5113. Prepared by Southwest Research Institute for
National Aeronautics and Space Administration, Washington, D.C., pp.
225-247.
Koerner, R.M. and Daniel, D.E. (1992). Better cover-ups. Civil Engineering,
pp. 55-57.
12.6 REFERENCES
397
Koerner, R.M., Lord, A.E. Jr., Bowders, J.J. and Dougherty, W.W. (1982). CW
microwave location of voids beneath paved areas. Journal of geotechnical
Engineering, ASCE, Vol. GT1, pp. 133-145.
Kostecki, P.T., Calabrese, E.J. and Horton, H.M. (1989). Review of present
risk assessment models for petroleum contaminated soils. In: Petroleum
Contaminated Soils. Editors: P.T. Kostechi and E.J. Calabrese. Lewis
Publishers, Chelsea, Michigan, pp. 263-300.
Laine, D.L. And Miklas, M.P. (1989). Detection and location of leaks in
geomembrane liners using an electrical method: case-histories.
Proceedings of the Tenth National (Superfund) Conference, Washington, D.C.,
pp. 35-40.
Laine, E.F., Dines, K.A., Okada, J.T. and Lytle, R.J. (1980). Probing concrete
with radio waves, Journal of Geotechnical Engineering, ASCE, Vol. GT7,
pp. 759-767.
Marnicio, R.J., Hakkinen, P.J., Lutkenhoff, S.D. Hertzberg, R.C. and Moskowitz,
P.D.(1991). Risk analysis software and databases: review of Riskware 90
Conference and Exhibition. Risk Analysis, Vol. 11, No. 3, pp. 545-560.
McNeil, J. D. (1982). Electromagnetic resistivity mapping of contaminant
plumes. Proceedings National Conference on Management of Uncontrolled
Hazardous Waste Sites, Washington, D.C., pp. 16.
Mitchell, J.K., Seed, R.B. and Seed, H.B. (1990). Kettleman Hills waste landfill
slope failure 1: liner system properties. Journal of Geotechnical
Engineering, ASCE Vol. 116, No. 4, pp. 647-663.
Murray, C.J. (1994). Dawn of the smart sensor: sensor technology combines
with bus networking to form the backbone of distributed machine
control. Design News, Vol. 49, no. 9, p. 73.
Ormond, T. (1993). Smart sensors tackle tough environments.
Environmental Design News. Vol. 38, No. 21, p. 35.
Peters, W.R., Shultz, D.W. and Duff, B.M. (1982). Electrical resistivity
techniques for locating liner leaks. Proceedings of the National Conference
on Management of Uncontrolled Hazardous Waste Sites, Washington, D.C.,
pp. 31-35.
Peterson, C.H. (1990). Predicting long term model for radioactive waste
repository. Proceedings of High Level Waste Management, Las Vegas,
Nevada, pp. 1287-1292.
398
Preuss, P.W., Ehrlich, A.M. and Garrahan, K.G. U.S. EPA guidelines for risk
assessment. Proceedings of the 7th National Conference on Management of
Uncontrolled Hazardous Waste Sites, Washington, D.C., pp. 167-172.
Schanz, R.W. and Salhotra, A.M. (1990). Estimating cleanup levels at
hazardous waste sites. Proceedings of the 11th National (Superfund)
Conference, Washington, D.C., pp. 157-160.
Schultz, D.W. Duff, B.M. and Peters, W.R. (1984). Performance of an electrical
resistivity technique for detecting and locating geomembrane failures.
Proceedings of the International Conference on Geomembranes, Denver,
Colorado, pp. 445-449.
Suter, G.W., Luxmore, R.J., and Smith, E.D. (1993). Compacted soil barriers
at abandoned landfill sites are likely to fail in the long term. Journal of
Environmental Quality, Vol. 22, No. 2, pp. 217-226.
Taylor, A.C., Burmaster, D.E., Murphy, B.L. and Boutwell, S.H. (1987). An
exposure assessment modeling system for hazardous waste sites.
Proceedings of the 8th National (Superfund) Conference, Washington, D.C.
pp. 153-157.
Tonn, B. and Wagner, C. (1990). Risk assessment with lower probabilities:
application to toxic wastes. Journal of Environmental Systems, Vol. 20,
No. 2, pp. 169-178.
U.S. Department of Energy. (1994a). Electrical resistance tomography.
Technology summary DOE/EM-0135P on COCs in non-arid soils
integrated demonstration, Office of Environmental Management, Office
of Technology Development, U.S. Department of Energy, Washington,
D.C., pp. 13-14.
U.S. Department of Energy. (1994b). Fiber-optic sensors. Technology
Catalogue DPE/EM-013P, Office of Environmental Management, Office
of Technology Development, U.S. Department of Energy, Washington D.C.
pp. 43-47.
U.S. Environmental Protection Agency. (1989). Statistical analysis of
groundwater monitoring data at RCRA facilities. Interim Final Guidance.
Office of Solid Waste, Washington, D.C.
U.S. Environmental Protection Agency. (1990). Exposure factors handbook.
EPA/600/8-89/043. Office of Health and Environmental Assessment,
U.S. Environmental Protection Agency, Washington D.C.
U.S. Environmental Protection Agency. (1991a). Superfund. Briefing materials
12.6 REFERENCES
399
presented to the administrator, by the superfund program, U.S.

Environmental Protection Agency, Washington, D.C.
U.S. Environmental Protection Agency. (1991b). Design and construction of
RCRA/CERCLA final covers. EPA/625/4-91/025. Office of Research
and Development, U.S. Environmental Protection Agency, Washington,
D.C.
U.S. Environmental Protection Agency. (1991c). Ground Water, Vol. II:
Methodology. Handbook EPA/625/6-90/016b. Office of Research and
Development, Washington, D.C.
U.S. Environmental Protection Agency. (1991d). Chapter 11 of SW-86
groundwater monitoring, Final Draft. Office of Solid Waste, Washington,
D.C.
U.S. Environmental Protection Agency. (1992). Computation of current
practices at land disposal facilities. EPA 530-R-92-003. Office of Solid
Waste, U.S. Environmental Protection Agency, Washington, D.C.
U.S. Environmental Protection Agency (1993a). Subsurface characterization
and monitoring techniques: a desk reference guide; volume II; the vadose
zone, field screening analytical methods; appendices C and D. EPA/625/
R93/003b. Office of Research and development, U.S. Environmental
Protection Agency, Washington, D.C.
U.S. Environmental Protection Agency (1993b). Use of airborne, surface,
and bore hole geophysical techniques at contaminated sites. A reference
guide. EPA/625/R/007. Office of Research and development, U.S.
Environmental Protection Agency, Washington, D.C.
Walther, E.G., Pitchford, A.M. and Olhoeft, G.R. (1986). A strategy for
detecting subsurface organic contaminants. Proceedings of the NWWA/
API Conference on petroleum hydrocarbons and organic chemicals in
groundwater: prevention, detection, and restorations, Houston, Texas, pp. 357381.
Weber, D., Easley, D. and Englund, E. (1991). Probability of plume interception
using conditional simulation of hydraulic head and inverse modeling.
Mathematical Geology, Vol. 23, No. 2, pp. 219239.
Wright, F.G., Inyang, H.I. and Myers, V.B. (1993). Risk reduction through
regulatory control of waste disposal facility siting. Journal of
Environmental systems, vol. 22, No. 1, pp. 27-35.
400
Yen, B.C. and Scanlon, B. (1975). Sanitary landfill settlement rates. Journal
of the Geotechnical Engineering Division, American Society of Civil
Engineers, Vol. 101, No. GT5, pp. 475-487.
APPENDIX
APPENDIX
WORKSHOP PARTICIPANTS & CONTRIBUTORS
Phyllis Adams
DuPont Co.
Wilmington, DE
Gordon P. Boutwell
Soil Testing Engineers, Inc.
Baton Rouge, LA
Marita Allan
Brookhaven National Laboratory
Upton, NY
Randall Breeden
U.S. Environmental Protection
Agency
Washington, DC
Loren R. Anderson
Utah State University
Logan, UT
Charles B. Andrews
S.S. Papadopolous & Assoc.
Bethesda, MD
Mark D. Ankeny
D.B. Stephens Assoc.
Albuquerque, NM
John A. Apps
Lawrence Berkeley National
Laboratory
Berkeley, CA
Robert C. Bachus
GeoSyntec Consultants
Atlanta, GA
Craig H. Benson
University of Wisconsin
Madison, WI
J. David Betsill
Sandia National Laboratories
Albuquerque, NM
David Blowes
University of Waterloo
Ontario, Canada
Andrew Bodosci
University of Cincinnati
Cincinnati, OH
Jack Boschuk, Jr.
J&L Testing Co., Inc.
Canonsburg, PA
Patrick T. Burnette
Consultant
Humble, TX
David S. Burden
Agency
Ada, OK
George Burke
Hayward Baker
Odenton, MD
David R. Burris
Armstrong Laboratory
Tyndall AFB, FL
Hugh Campbell
DuPont Co.
Wilmington, DE
Ernest E. Carter
Carter Technologies
Sugarland, TX
Grover H. (Skip) Chamberlain
U.S. Department of Energy
Germantown, MD
Martin Cherrington
Cherrington Corp.
Sacramento, CA
Chen Chiang
Shell Oil Co.
Houston, TX
401
402
APPENDIX
Calvin Chien
DuPont Co.
Wilmington, DE
Annette Esnault
SIF Bachy
Rueil Malmaison Cedex, France
James H. Clarke
Eckenfelder Inc.
Nashville, TN
Jeffrey C. Evans
Bucknell University
Lewisburg, PA
Robert M.Cohen
GeoTrans, Inc.
Sterling, VA
Lorne G. Everett
Geraghty & Miller
Santa Barbara, CA
Steve Conley
DuPont Co.
Wilmington, DE
George M. Filz
Virginia Polytechnic Institute and
State University
Blacksburg, VA
Richard A. Conway
Union Carbide
South Charleston, WV
Nico Cortlever
Delft Geotechnics
The Netherlands
Patricia Culligan-Hensley
Massachusetts Institute of
Technology
Cambridge, MA
David E. Daniel
University of Texas
Austin, TX
Richard R. Davidson
Woodward-Clyde Consultants
Denver, CO
Steven R. Day
Geo Con Inc.
Monroeville, PA
Michael T. Dukes
RUST Environmental
San Jose, CA
Subijoy Dutta
Agency
Washington, DC
Brian Dwyer
Albuquerque, NM
Harley Freeman
Pacific Northwest National
Laboratory
Richland, WA
Rene Fuentes
Agency
Seattle, WA
Glendon W. Gee
Laboratory
Richland, WA
Bruno Gemmi
Fondazioni Speciali
Parma, Italy
Robert W. Gillham
Ontario, Canada
Edward W . Gleason
Encor Coatings Inc.
Bath, PA
Peter Grathwohl
University of Tbingen
Tbingen, Germany
John F. Greiner
DuPont Environmental
Remediation Services
Wilmington, DE
APPENDIX
Beth A. Gross
University of Texas
Austin, TX
Stephan Jefferis
Golder Associates Ltd.
Berkshire, England
John L. Guglielmetti
Wilmingon, DE
Reuben Karol
Rutgers University
Princeton, NJ
Alison Hapka
Wilmington, DE
Andrea Hart
MSE, Inc.
Butte, MT
Joel S. Hayworth
Tyndall AFB, FL
John Heiser
Upton, NY
Jan Hendricksen
MSE, Inc.
Butte, MT
Douglas J. Hermann
STS Consultants Ltd.
Deerfield, IL
Beverly L. Herzog
Illinois Geological Survey
Champaign, IL
Tomasz Hueckel
Duke University
Durham, NC
Hilary I. Inyang
University of Massachusetts
Lowell, MA
Alex Iskandar
U.S. Army Cold Regions Research
Laboratory
Hanover, NH
David Jaros
U.S. Army Corps of Engineers
Omaha, NE
403
Ed. Kavazanjian, Jr.

GeoSyntec Consultants
Huntington Beach, CA
Robert M. Koerner
Drexel University
Philadelphia, PA
Kevin Kostelnick
Lockheed
Idaho Falls, ID
Michael Kotona
Tyndall AFB, FL
John Koutsandreas
Florida State University
Tallahassee, FL
James M. Cramer
Nilex Corp.
Englewood, CO
Raymond Krizek
Northwestern University
Evanston, IL
Lawrence E. Kukacka
Upton, NY
Richard C. Landis
DuPont Co.
Wilmington, DE
Robert E. Landreth
Agency
Cincinnati, OH
Brian M. LaRue
Spencer, White & Prentis
Swansea, MA
404
APPENDIX
David Lesmes
Boston College
Boston, MA
W. Eric Limbach
Idaho State University
Pocatello, ID
Guy Loomis
Idaho National Engineering
Laboratory
Idaho Falls, ID
Patricia Mackenzie
General Electric Co.
Niskayuna, NY
Mario Manassero
Ingegneria Geotecnica
Torino, Italy
Stephen M. Mangion
Agency
Chicago, IL
Dianne C. Marozas
Albuquerque, NM
Lisa Martinenghi
Institut fur Geotechnik
Zurich, Switzerland
Lynn McCloskey
MSE, Inc.
Butte, MT
Mary McCune
Germantown, MD
Steve McCutcheon
Agency
Athens, GA
Tomiann McDaniel
Washington, DC
Lawrence D. McKay
University of Tennessee
Knoxville, TN
Donald R. McMahon
McMahon & Mann Consulting
Engineers
Buffalo, NY
Stefan Melchior
University of Hamburg
Hamburg, Germany
James W . Mercer
GeoTrans, Inc.
Sterling, VA
Amy Mills
Agency
Washington, DC
James K. Mitchell
Virginia Polytechnic Institute and
State University
Blacksburg, VA
Richard Mitchell
RM Consultants
Novato, CA
Dale Morgan
Massachusetts Institute of
Technology
Cambridge, MA
Stan Morrison
RUST Geotech
Grand Junction, CO
Henry Mott
South Dakota School of Mines and
Technology
Rapid City, SD
Edward ODonnell
U.S. Nuclear Regulatory
Commission
Washington, DC
Robert Orth
Monsanto Company
St. Louis, MO
Phil Palmer
DuPont Co.
Wilmington, DE
APPENDIX
Steve Paulson
U.S. Bureau of Mines
Minneapolis, MN
Ray Peters
Scientific Ecology Group
Oak Ridge, TN
405
Ralph Rumer
State University of New York
at Buffalo
Buffalo, NY
Ingo Sass
FlowTex-GUT
Ettlingen, Germany
Mary E. Peterson
Laboratory
Richland, WA
Kent Saugier
Brown and Root
Houston, TX
Cathy Pfeifer
Idaho National Engineering
Laboratories
Idaho Falls, ID
Paul R. Schroeder
U.S. Army Waterways Experiment
Station
Vicksburg, MS
Mark Phifer
Westinghouse Savannah River Co.
Aiken, SC
Dale Schultz
DuPont Co.
Newark, DE
Kelvin Potter
ICI Engineering Technology
United Kingdom
Charles Shackelford
Colorado State University
Fort Collins, CO
Robert W. Puls
Agency
Ada, OK
Stephen H. Shoemaker
DuPont Co.
Houston, TX
Alan J. Rabideau
State University of New York
at Buffalo
Buffalo, NY
Pedro C. Repetto
Woodward-Clyde Consultants
Blue Ball, PA
Kenneth Skahn
Agency
Washington, DC
David J.A. Smyth
Ontario, Canada
Robert W. Ridky
University of Maryland
College Park, MD
Robert L. Stamnes
Agency
Seattle, WA
Randall Ross
Agency
Ada, OK
Richard Steiml
Agency
Washington, DC
Michael Rubich
MSE, Inc.
Butte, MT
John C. Stormont
Albuquerque, NM
406
APPENDIX
Andrew Street
RUST Environmental
Bristol, UK
Jennifer Su
DuPont Environmental Research
Services
Wilmington, DE
Ron Wilhelm
Agency
Washington, DC
Cecelia Williams
Albuquerque, NM
George J. Tamaro
Mueser Rutledge Consulting
Engineers
New York, NY
Lee Wolfe
Agency
Athens, GA
Georg Teutsch
University of Tbingen
Tbingen, Germany
David S. Yang
SMW Seiko, Inc.
Hayward, CA
Bruce Thomson
University of New Mexico
Albuquerque, NM
Peter Yen
Bechtel Corp.
San Francisco, CA
Terry D. Vandell
Conoco, Inc.
Ponca City, FL
Chunmiao Zheng
University of Alabama
Tuscaloosa, AL
G.A.M. van Meurs

Delft Geotechnics
The Netherlands
Paul Zielinski
Germantown, MD
Charles F. Voss
Golder Federal Services, Inc.
Redmond, WA
Thomas F. Zimmie
Rensselaer Polytechnic Institute
Troy, NY
Francke C. Walberg
Kansas City, MO
William W . Walling
SLT North America Inc.
Conroe, TX
Scott D. Warner
Geomatrix Consultants, Inc.
San Francisco, CA
Robert Waters
Tensor
Austin, TX
Joyce M. Whang
DuPont Co.
Wilmington, DE
INDEX
407
INDEX
above-ground retaining walls, corrosion of, 82

AC-400 [acrylate gel grout], 217
accelerators, for cementitious grouts, 12
accelerometers, use for directional drilling,
196-197
acetone
acrylate grout resistance to, 218, 219
transmission by high-density
polyethylene, 99
acids
effect on barrier materials, 283
mix resistance to, 38
sheet piles resistant to, 86
acoustic imaging techniques, for barrier
fault detection, 133
acrylate gel grout, 212, 217-219, 243
chemical resistance of, 218-219
costs of, 218, 244
expected lifetime of, 213-214, 244
hydraulic conductivities of, 218-219, 244
properties of, 212, 218, 244
acrylic grouts
costs of, 234
expected lifetime of, 244
hydraulic conductivity of, 234, 244
jet grouting of, 241
properties of, 234, 244
use in barriers, 212, 233-235
activated carbon
chemical precipitation by, 303
use in organic compound removal, 330,
340
use in sorptive barriers, 340
adhesive-bonded GCLs, internal shear
strength of, 124
admixtures, cementitious grout
improvement by, 17
ADRE. See advective-dispersive reactive
equation (ADRE)
adsorption parameters, determination of, 152
adsorptive stripping analysis, of metals, 390
advection, contaminant transport by, 213,
248, 254, 266, 268, 272, 273, 276-277
advective-dispersive reactive equation (ADRE)
derivation of, 240-250, 251, 255
first-order decay term in, 270, 277
Laplace-transformed solutions, 260
in one-dimensional transport, 255, 258-271
for permeable reactive barriers, 266

solutions of, 259-261
sorption isotherm in, 279
use in modeling chemical damage, 283
advective-dispersive-reactive transport
theory, 281
application to low-permeability barriers,
286
aggregates, use in soil-bentonite and cementbentonite mixes, 24-25
agricultural waste, use in permeable barriers,
339
agrochemicals, cement-bentonite cutoff wall
for, 20
Alaska, geomembrane vertical barrier use in,
108-109
Albuquerque (New Mexico), cap studies at,
133
alcohols
alkalis, sheet piles resistant to, 86
alumina, in Nyacol DP5110 colloidal silica,
235
aluminosilicate clays, inorganic compound
sorption by, 308
aluminum
inorganic compound sorption by, 308-309
in Nyacol DP 5110, 220
ammonium persulfate initiator, in AC-400
[acrylate gel grout], 217
anaerobic bacterial consortium, use in
permeable barrier, 339
aniline
acrylate grout resistance to, 219
montan wax grout resistance to, 224
sodium silicate grout resistance to, 217
antioxidants, in high-density polyethylene
formulation, 98, 123
aperture, of fractures, 158
aquifers
avoidance of cross-contamination in, 51
contamination of, 169
seasonal water table fluctuations in, 281
aquitard
contamination by sheet piling, 77
geomembranes toed into, 96, 97, 107
perforation of, 168
408
INDEX
sheet-pile keying into, 86

tests on, 52
transport parameters of, 149
walls keyed into, 52, 54, 63
Arbed interlock, for sheet piles, 88-89
arching, 56
in soil-bentonite backfilled
trenches, 54-55
areal sensors, for containment monitoring,
373
Armstrong Laboratory (Tyndall Air Force
Base), permeable
reactive barrier research at, 347
aromatic hydrocarbons,
biodegradation of, 306-307
Arrhenius modeling, 110
of HDPE geomembrane service life, 122
artificially emplaced floors and bottom
barriers
defect detection in, 207
directional drilling for, 193-194
emplacement and performance
verification of, 202-203
field testing of, 199-201
frozen soil barriers, 194-196
grouted, 187-194
guidance systems for, 196-198
limited use of, 185
needs for, 206-208
problems of, 185
technological specific needs for, 206-207
artificially emplaced floors and bottom
barriers, 185-209
applications of, 188
barrier verification for, 202-203, 208
characteristics of, 203
cost analysis for, 204-206
technological assessment of, 202-205
thin diaphragm wall use in, 188-189, 191,
192, 199
as-built drawings, use in cutoff wall
construction, 63
ash and MSW, geomembrane vertical barrier
for, 108
asphalt
in cap layers, 121, 122, 133, 134, 137, 138
properties of, 124-125
asphalt/grease mixture, use to prevent
interlock leaks, 79
ASTM C39, 230, 231, 232, 235
ASTM C267-82, 216, 223
ASTM C494, 12
ASTM C618, 10
ASTM C989, 10
ASTM C1017, 10
ASTM D18.21.02, 66
ASTM D0695, 226, 229, 232, 235
ASTM D1452, 143
ASTM D1586, 143
ASTM D1587, 143
ASTM D2113, 143
ASTM D2116, 155
ASTM D2216, 151
ASTM D4643, 151
ASTM D4646, 152
ASTM D4750, 147
ASTM D4843, 229, 232, 234
ASTM D5084, 229, 232, 234
ASTM D5092-90, 173
ASTM D5299-92, 172
ASTM guides, for decommissioning wells
and borings, 172-173
ASTM W96, 99
Athens (Georgia), permeable reactive barrier
research at, 347
attapulgite
mixing considerations for, 59
as sheet pile sealant, 86
attapulgite-based barriers
mixing of, 8
in slurry trench cutoff walls, 5
attapulgite-cement walls, use in saline
environments, 24
attapulgite-soil walls, use in saline
environments, 24
Atterberg limits, of soil samples, 51
auger holes, grouting of, 173
auger mixing shaft, for deep soil mixing, 9
augers
requirements for use of, 59
use in deep soil mixing method, 48, 49
autogenous healing, of cutoff walls, 29, 67
backfill
cement content effects on, 58
clay content effects on, 58
consolidation studies on, 63
density of, 57
fluidity of, 57
particle-size distribution of, 58
placement observation on, 63
QA/QC tests on, 63
quality changes during mixing, 96
selection of type of, 56-58
stresses of, 54, 69
stress-strain behavior of, 57
use of in situ materials for, 51
INDEX
backfilled cutoff walls, 46, 48
backfilled trenches, 45
backhoe-excavated trenches
joint reduction by, 58
low cost of, 70
barrier containment technologies, 1
workshop on (August, 1995), 2
barrier diffusion test, on frozen soil field
barriers, 200-201
barrier flocculation, in containment systems,
359
barrier floors, 2
barrier infiltration analysis, containment
monitoring by, 378
barrier integrity monitoring, 364-365
barrier layer, of caps, 120
barrier materials, 2
chemical-based. See chemical grouts
chemical compatibility of, 283
barrier permeation monitoring, 365-366
barrier permittivity
of sheet pile walls, 87
of soil-bentonite walls, 87
barriers
long-term damage to, 283
modeling for failure of, 282, 288
monitoring needs for, 385-386, 388
parameter estimation for design of, 280281
performance monitoring and evaluation
of, 355-399
bases, effect on barrier materials, 283
Belgian clay, radiolithic gas formation in, 165
Belgium, radioactive waste disposal in, 163,
167
Beltsville (Maryland), cap field test at, 131132
Benseal-Bentonite Slurry, use as well sealant,
170, 171
Bentofix I and II, internal shear strength of,
124
Bentomat, internal shear strength of, 124
bentonite
as cementitious grout additive, 9, 11, 16
as sheet pile sealant, 86
sodium and calcium compounds
compared, 21
UK specification for, 25
use in UK slurry trench method, 21
bentonite-cement grout, as well sealant, 170,
171, 172
bentonite filter cake, formation at trench
wall, 47
bentonite slurry
409
as borehole sealant, 173

excavation under, 46-47
fluid properties of, 26
as interlock sealant, 87-88
benzene, biodegradation of, 306, 308
Bethlehem Steel hot-rolled sheet pile, 80, 88
BETX, biodegradation of, 306, 307, 308
biodegradable contaminants, 277
diffusion coefficients of, 270
sorption of, 308
biodegradation reactions, in permeable
reactive barriers, 305-307
bioenhanced barriers, 33
biofouling, 282
biologic electron acceptors, chemical
precipitation by, 303
biomass, contaminant treatment by, 304-305
biopolymer drain walls, costs of, 60
biopolymer gels, use in hydraulic fracturing,
326-327
biopolymer trenches, 8
biotic contaminant removal systems, 302
biotic reactions, of contaminants, 277
biotransformation, of contaminants, 270
bleeding
of cementitious grout, 9, 11-12
of slurries, 26
Bodo Canyon Disposal cell (Colorado),
permeable reactive barrier tests at, 338
borax, as cementitious grout retarder, 12
Borden (Ontario), reactive iron barrier tests
at, 327-330
borehole caliper log, use in ultrasonic
logging, 143
borehole methods, for hydraulic activity, 151
boreholes
ASTM guides for decommissioning, 170171
as contaminant pathways, 169
directional drilling of, 188, 197
for floors and bottom barrier
emplacement, 186-187
problems from, 173-175
sealing of, 171-172
sloughing reduction for, 173
TruTracker method for, 198
borescope, for containment monitoring, 391392
borings, for site samples, 51
bottom barriers. See floors and bottom
barriers
bottom plug, for cold-rolled sheet pile
interlock, 81
410
INDEX
boulders
as problem for augers, 49
as problem for sheet pile use, 77, 91
box-out enclosures
pump tests on, 65
for system performance verification, 34
brine, effect on frozen soil barriers, 207
British Steel, recommendations for
combatting steel-pile corrosion, 83, 84
bromide
as conservative tracer, 154
matrix diffusion of, 159
bromoform, dehalogenation of, 306
Brookhaven National Laboratories
acrylic tests at, 233
furan tests at, 229
polyester styrene tests at, 230
sulfur polymer cement tests at, 225
vinylester styrene test at, 232
BTEX, chemical precipitation of, 304
bulldozer, mixing considerations for use of, 59
burrowing animals, cap protection from, 120,
121
cable detection systems, for containment

monitoring, 373
cadmium, adsorption of, 309
calcium, in Portland cement, 22
calcium aluminum hydrates, formation in
cement bentonite, 22, 29
calcium bentonite, sodium bentonite
compared to, 21
calcium carbonate. See limestone
calcium chloride, as cementitious grout
accelerator, 12
calcium hydroxide. See lime
calcium phosphate. See hydroxyapatite
calcium silicate, formation in cementbentonite, 22, 28, 29
calcium sulfate, as grout additive, 16
calcium sulfate dihydrate. See gypsum
calcium sulphoaluminate, as grout additive,
16
California
landfill regulations in, 126-127
maximum contaminant levels (MCLs) of,
333
seismic performance of landfills in, 127
California Regional Water Quality Control
Board, groundwater monitoring by, 333
Cam-clay model, for studies of thermal
effects, 167
camera inspection, of sheet pile interlocks, 81
Canadian liner clay, ethanol flow in, 168

Canadian Metal Rolling Mills cold-rolled
sheet pile, 80
canister tests, on permeable reactive barriers,
332, 333
capacitance measurements, of moisture
content, 156
capillary drains and barriers, possible use in
caps, 138
caps, 2, 29, 119-140, 342
adverse factors affecting, 120, 135-136
aging tests on, 125
asphalt layer use in, 124-125, 136, 137, 138
barrier layer of, 120, 121
capillary barriers for, 136-137
component layers of, 120, 130
contaminant transport modeling of, 248
cross section of, 120
current technology for, 120
design of, 122
drainage layer of, 120, 136
durability of, 135, 136, 137
erosion of, 130
European use of, 130-131
failures of, 129-130, 135
field studies on, 129-134
foundation layer of, 120, 121, 136
gas-collection layer of, 120, 121, 136
geomembrane use in, 112, 121, 122-123,
130
HELP model for, 128-129
needs for, 136-137
paper mill sludge use in, 125-126, 137
performance monitoring of, 385, 389
performance of, 121, 135
protection layer of, 120, 121, 136
purpose of, 119
QA/QC construction of, 138
seismic performance of, 126-127, 134-135,
137
slope stability of, 122, 126, 137, 138
surface layer of, 120, 121, 136
water-balance modeling of, 122
carbon-14, use in fracture studies, 160
carbonates
effects on clay thermo-mechnical
behavior, 164-165
precipitation in cement-bentonite walls, 67
carbon black, in high-density polyethylene
formulation, 98
carbon dioxide
formation in radioactive waste sites, 165
INDEX
carbon tetrachloride
dehalogenation of, 306
detection in waste sites, 374, 375
CARBRAY 100, as epoxy grout, 226-227
carcinogens, risk assessment of containment
systems for, 356
cathodic protection, for underground steel
structures, 86
cement
interlock sealants based on, 82
sulfate-resistant, 13, 15
Type K. See Type K cement use in deep
soil mixing, 9
cement-based vertical barriers, 5-43
applications of, 68
assessment of, 33-34
core sampling of, 34
curing time effects on, 30-31
design and construction of, 45-76
factors affecting, 30-32
field performance of, 29-33
freeze-thaw cycling effects on, 31
long-term performance of, 66-67
mixing operations for, 59
monitoring needs for, 386
research and development needs for, 68-69
types and materials for, 5
wet-dry cycling effects on, 31
cement-bentonite cutoff walls, 46, 56
construction of, 47
cracking of, 30
desiccation resistance of, 31
geomembrane advantage over, 95-96
in situ hydraulic conductivity tests on, 64
strength tests for, 27
cement-bentonite-geomembrane walls, use
in United Kingdom, 20
cement-bentonite materials, 17
aggregate use in, 24-25
as backfill for geomembranes, 103, 104, 107
drying of, 29
durability specifications for, 28
slag addition to, 23
strain at failure specifications for, 27
use in United Kingdom, 19, 23
cement-bentonite-slag cutoff walls, cracking
of, 30
cement-bentonite slurries, fluid properties
of, 26
cement-bentonite slurry trench cutoff walls, 5
cement-bentonite slurry wall, use with
411
permeable reactive barriers, 332-333

cement-bentonite walls
costs of, 60
drying of, 67
durability of, 67
movement of, 67
cementitious backfill, sampling difficulties
for, 63-64
cementitious grouting techniques, 7
for in situ mixed walls, 5
cementitious grouts
admixture improvement of, 17
age of, 12
bleeding of, 9, 11-12
for bottom barriers, 185
cement-type effects on, 10
chemical grouts, 7
contaminants of, 12
curing conditions for, 12
durability of, 14
fiber-enhanced, 16
gel strength of, 9, 11
hydraulic activity of, 7-8, 13, 14
mix proportions of, 12
properties of, 9-17
set time for, 7
as sheet pile sealants, 86
shrinkage of, 16
soil effects on, 12, 14
strength of, 13-15, 16
sulfate resistance of, 15
temperature effects on, 12
uncured and cured properties of, 7
for vertical barriers, 7
viscosity of, 9, 10-11, 12
cement-pfa-bentonite cutoff walls, in United
Kingdom, 21
cement-sandy clay-bentonite mix, mixing
considerations for, 59
cement-slag-bentonite mixes, hydraulic
conductivity of, 23
cement-slag-bentonite trenches, 20
use in Europe, 20, 50
cement-slag cutoff walls, in United
Kingdom, 20
CERCLA, lack of vadose zone
monitoring regulations in, 66
cesium total field magnetometer, in rapid
geophysical surveyor (RGS), 391
CFCs, as indicator of recent recharge, 160
chabazite, inorganic compound sorption by,
308
chain saw devices, for permeable reactive
barrier emplacement, 322
412
INDEX
characteristic diffusion time, 274

charge coupled device (CCD) camera, in
colloidal borescope, 391
chelating agents, use in iron hydroxide grout
production, 221
chemical attack, cutoff wall additive effects
on, 48
chemical barrier materials. See chemical
grouts
chemical compatibility, of barrier materials,
283
chemical compatibility tests
for contaminants, 36
on vertical barrier cutoff wall materials, 71
chemical effects, thermal effects coupled
with, 166-168
chemical grouts, 211-246. See also individual
chemical grouts
advantages of, 211
for bottom barriers, 185
chemical resistance of, 213, 216-217, 218219, 223-224, 225, 229, 231, 232
costs of, 212, 244
expected lifetime of, 213-214
factors affecting choice of, 211-212
hydraulic conductivities of, 213, 244
irradiation resistance of, 213
jet grouting of, 211, 225, 226, 239-241
parameters affecting performance of, 212214
performance standards of, 243
permeation grouting of, 211, 214, 221, 227,
235-239
properties affecting verification of, 243
reduced diffusive transport by, 213
repairability of, 214
research needs for, 241-243
safety, toxicity, and regulatory
acceptability of, 214
set time of, 212-213
viscosity of, 212
wet-dry cycling of, 213
chemical plant, geomembrane vertical
barrier use at, 107
chemical potential, measurement of, 148-149
chemical precipitation, use in permeable
reactive barriers, 304
chemical reactions, in waste facilities, 147
chemical resistance
of acrylate gel grout, 218-219
of acrylic grout, 218-219, 234-235

of chemical grouts, 213
of furan, 230
of montan wax, 223-224
of polyester styrene, 231
of sodium silicate grout, 217
of sulfur polymer cement, 225
of vinylester styrene, 232
chemical sensor companies, 374
chemical vats, sulfur polymer cement use for,
225
chemo-hydro-mechanical models, for
transport studies, 166-168
chloride
as conservative tracer, 154
chlorinated hydrocarbons
daughter products of, 305
redox reaction treatment of, 305
transport of, 265
chlorocarbons
dechlorination of, 305
Halsnif detection of, 391
permeable reactive iron barriers for, 327336, 345
chloroform
permeable reactive iron barrier for, 327
polyethylene, 99
chloroprene, use in geomembrane gaskets,
105
chromate
permeable treatment system for, 334, 335336, 337
sorption of, 308
chromium
adsorption of, 309
chemical precipitation of, 304, 325
circular saw devices, for permeable reactive
citric acid, in iron hydroxide grout
mixture, 221, 222
clay
in cap layers, 120, 121, 123, 130-131
in cutoff wall backfill, 56
hydraulic conductivity of, 155
interlock sealants based on, 82, 86
liquid/gas transport in, 154
matrix diffusion in, 159
radar energy dissipation by, 370
radioactive waste disposal in, 163-164
simulation of radwaste disposal in, 167168
INDEX
stability of, 66
clay-cement materials, 22
clayey barriers
inorganic contamination of, 32
performance of, 35
clay liners
chemical-transport effects in, 168
model predictions of transport in, 272
Claymax, internal shear strength of, 124
clay minerals, geochemistry of, 161
clay-sand geology, 144
clinoptilolite, inorganic compound sorption
by, 308
coal, sorption of organics by, 307-308, 340
coal mine drainage, precipitation treatment
of, 304
Coal Tar Epoxy, use for steel sheet pile
coating, 85-86
coal-tar pitch mixture, for steel sheet piles, 85
cobalt, adsorption of, 309
cobalt naphthenate
as acrylic promoter, 233-234
as polyester styrene promoter, 230
cofferdam wall, water passage through, 79
cold joints, in panel construction, 58
cold-rolled sheet pile, 80, 81
colloidal borescope, for containment
monitoring, 391-392
colloidal contaminants, attenuation of, 161
colloidal silica
as barrier material, 212, 219-221, 243
costs of, 221, 244
expected lifetime of, 221, 244
permeation grouting of, 187, 188, 235-236
properties of, 221
set times of, 220
colloidal virus tracers, transport of, 159
colmix method, European costs of, 60
compacted clay layer, in caps, 130, 134, 135, 136
compacted clay liners
contaminant transport in, 273
hydraulic conductivity of, 150-151, 165
compatibility index, for identifying potential
compatibility problems, 283
compatibility testing, need for
standardization of, 36
complex formation, Irving-Williams order of,
308
composite barriers, one-dimensional
modeling of, 258
compost, use in permeable reactive barriers,
337
413
compressibility, of soil-bentonite materials, 18

compressive strength
of cementitious grouts, 13-15, 16
of jet grouting method, 192
computer modeling
in colloidal silica field tests, 236
of hydraulic performance of landfill caps,
128, 137
computers, use in ADRE solutions, 259, 260,
275
computer software, use for stratigraphic data
interpretation, 145
concrete covers
cracking of, 365
concrete cutoff walls, 56
cone penetration test (CPT)
of hardened material, 26
heavy-weight penetrometer for, 390
logs of, 51
sensor use with, 390
use for containment monitoring, 364, 391,
393
use in stratigraphic studies, 143, 146, 147
connections to covers and floors, installation
methods for, 106-107
conservation equations, in transport
modeling, 248
conservative tracers, entrapment in soils, 154
consolidation tests, for areas of possible
ground movement, 52
construction
costs of, 59-60
large-scale model user facility for tests on,
69
method effects on, 58
mixing considerations in, 59
of panels, 57, 58
QA/QC aspects of, 62-64
of soil- and cement-based vertical
barriers, 45-76
tests used during, 63
Containment and Transport Modeling
(C&TM), 1
containment systems. See also individual
systems
barrier flocculation in, 359
barrier integrity monitoring of, 364-365
barrier permeation monitoring of, 365-366
caps for, 119, 122
desiccation in, 359
deterioration mechanisms in, 358-359
deterioration-time patterns, 359
414
INDEX
dissolution in, 359

earthquake effects on, 359
failure determination in, 362-365
hydraulic conductivity level for, 27
leak detection in, 359-361
monitoring approaches for, 362-365
of, 355-399
physicochemical deterioration of, 358-359
slurry settlement in, 359
time-distributed monitoring data for, 359
wall durability for, 27-29
wall performance for, 35
contaminants
colloidal. See colloidal contaminants
concentration tests for, in vertical barrier
walls, 65
containment systems for. See containment
systems
removal in permeable reactive barriers,
302
subsurface biodegradation of, 156
volatile, field monitoring of, 157
contaminant flow modeling
diffusion coefficients in, 269-270
field data for testing of, 285
first-order decay in, 270-271
liquid diffusivity in, 268
for long-term performance, 285-26
multi-dimensional, 287
one-dimensional models, 272-281, 286-287
parameter estimation in, 267-271
for reactive transport, 277, 281, 285
software needs in, 284-285
sorption parameters in, 270
state-of-practice of, 286-287
symbols used in, 298-299
tortuosity factor in, 268-269
contaminant mass flux, mathematical
modeling of, 249
contaminant transport
chemical effects on, 161-163
driving potentials of, 147-149
electrokinetic transport, 282
fracture effect on, 158-161
laboratory column experiments on, 255,
256, 257
modeling of. See contaminant transport
modeling
multiphase type, 249
osmotic transport, 282

in porous media, 248
potential-energy gradient of, 147
radiation effects on, 163-168
reactive transport, 277, 281, 285
in subsurface media, 248
thermal effects on, 163-168
through pre-existing wells, 169
in unsaturated soils, 158
contaminant transport modeling, 2, 247-299
applications of, 261-271
of caps, 248
deviations from assumptions of, 281-282
entrance conditions in, 255-256, 287
estimation of contaminant concentrations
for, 355
exit conditions in, 256-258, 287
field-scale models, 261-262
in landfill liners, 248
of long-term performance, 281-283
for low hydraulic conductivity barriers,
261-262
in low hydraulic conductivity barriers,
248
mathematical model for, 249-258
of one-dimensional transport, 252-258,
262
of permeable reactive walls, 248, 266-267
probabilistic modeling, 281-282
for reactive transport, 277-281, 284
user-friendly models for, 248
contaminant zone, 253, 255, 256
contaminated subsurface zones,
remediation of, 1, 119
contamination
permeant density change during, 167
studies for bottom barriers, 185
contamination sites
drilling, sampling, and well construction
at, 174
investigation requirements for, 50-52
continuous wall permeable reactive barriers,
311, 313-314
continuum hypothesis, in transport
modeling, 248
convergence gauges, use for containment
monitoring, 364
copper
adsorption of, 309
use in contaminant dehalogenation, 305
copper sulfate
corrosion, of steel piling, 82-86
INDEX
costs
of artificially emplaced floors and bottom
barriers, 204-206
of chemical grouts, 212, 244
of construction, 59-60
of containment monitoring methods, 379384
evaluation of, 58
of frozen soil barriers, 205
of geomembrane vertical barriers, 100,
114-116, 117
of performance monitoring and
evaluation, 379-384
coupled flow processes, effects on barriers,
282
cover systems. See caps
CPT. See cone penetration test (CPT)
CPTU. See piezocone penetration test (CPTU)
cracking
in cement-bentonite walls, 67
of cementitious grouts, 16, 17
of cutoff walls, 29
creek sediment, use in permeable reactive
barriers, 337
creep effects, on caps, 126, 136
cross-borehole electromagnetic imaging, of
sources and plumes, 391
cross-contamination, of aquifers, 169
cross-hole seismic imaging
use for bottom emplacement verification,
203
use for containment monitoring, 364, 369370
cryogenic barriers, liquid and gas tracers for,
389
cubic flow law, for fractures, 158
cumene hydroperoxide, acrylic tests using,
233-234
cupric sulfate, sodium silicate grout
resistance to, 217
curing time, effects on hydraulic conductivity,
30-31
cutter-miller machine, for cutoff wall
installation, 21
cyclohexane, transmission by high-density
polyethylene, 99
cyclopentadiene oligomers, in sulfur
polymer cement production, 225
dams, vertical barrier use in, 67

Darcys law, 154, 164, 165, 248, 250, 281
DCE. See dichloroethenes (DCE)
Dead Sea
415
attapulgite in walls near, 24

slurry trench/geomembrane wall use at, 20
decay constant, derivation of, 252
decay processes, in contaminant transport
modeling, 278
deep mixing method
costs of, 60
working platform for, 59
deep soil mixing method, 12, 46, 48-49
barrier wall construction by, 48-49, 50
effects on wall performance, 35
materials for, 8
for in situ mixed walls, 5
dehalogenation
contaminant removal by, 302
by zero-valent metals, 305
dense non-aqueous phase liquid (DNAPL)
cost evaluation for waste sites of, 340-342
flow through disturbed zones, 174-176, 265
in fractured media, 159
permeable reactive barriers for, 303, 340
transport by sheet piles, 89
transport modeling for, 265-266
density, of slurries, 26
desiccation
in containment systems, 359
effects on vertical barriers, 34, 35, 68, 96
Detex Systems, cable detection system of,
373, 374
1,2,-dibromo-3-chloropropane,
dibromoethane, dehalogenation of, 306
dichloroethenes (DCE)
permeable reactive barriers for, 327, 328,
330-336
dicyclopentadiene, in sulfur polymer cement
production, 225
diffuse double layer, soil moisture in, 154
diffusion
contaminant transport by, 213
laboratory studies of, 280
diffusion coefficient(s)
of contaminants, 268, 275
determination of, 155, 269-270, 287
of indigenous barrier materials, 177
variability of, 251
diffusive flux
determination of, 151-152, 153
for exit boundary conditions, 275
dimethylaniline, as vinylester promoter, 232
direct-current (DC) resistivity system, for
containment monitoring, 375-377
direct driving, of polyvinyl chloride sheets, 99
416
INDEX
directional drilling method

advantages of, 189-190
for bottom barriers, 186, 188-190, 193-194,
199, 202
costs of, 204
guidance systems for, 196-198, 202
limitations of, 190
technological needs for, 207
Dirichlet boundary conditions, in barriers, 255
dispersion coefficient, 253, 296
dispersion tensor, derivation of, 250
dispersion theory, 253
displacement method, of cutoff wall
construction, 50
dissolution, in containment systems, 359
DM structural method, costs of, 60
domain reflectometry method, for moisture
content, 156
double-rod system
for jet grouting, 49, 192, 202
use in directional drilling, 188, 191-192
Dow Chemical Company
Arbed AZ18 sheet piles of, 88
Derakane 470-45 of, 232
Dow Corning, polysiloxane grouts from, 227,
235
downhole geophysical studies, 143
downhole organic chemistry, of potential
sites, 148-149
drainage layer, of caps, 120
drainage patterns, effect on cutoff wall
performance, 66
Drexel University Geosynthetic Research
Institute, 122-123
drilled shafts, in contaminated areas, 176
drilling, at contamination sites, 174
driving potentials, of contaminant transport,
147-149
DuPont
iron hydroxide grout development at, 221
Ludox SM colloidal silica of, 220, 236
permeable reactive barrier studies at, 323
Uniloc sheet piles from, 86
DuPont Corporate Remediation Group, 1, 2
durability, of vertical barrier walls, 27-29
Durango (Colorado), future reactant barrier
tests at, 337-338, 339
dye-tracing, use for containment monitoring,
200-201, 364
dynamic loads, effects on vertical barriers, 34
earth dams, plastic concrete repair of, 19

earthen dikes, geomembrane vertical barrier
use at, 107, 108

earthen walls, water-table fluctuation effects
on, 281
earthquakes
effects on caps, 126-127, 134-135, 137
effects on containment systems, 359
effects on vertical barriers, 34
long-term effects of, 283
East Wenatchee (Washington), cap field test
at, 132
Eco Instrumentation, Remote Access
Inventory Monitoring System of, 392
Eh
effect on permeable reactive barriers, 317,
333, 335
use for containment monitoring, 364
electrical crack gauges, use for containment
monitoring, 364
electrical fields, effect on ionic transport, 154
electrical resistance shear strips, use for
containment monitoring, 364, 367
electrical resistance tomography
containment monitoring by, 387
of vadose zone, 157-158
electrical resistivity method, use in
stratigraphy, 144, 147, 148
electrochemical sensing systems
for containment monitoring, 372-375, 387
cost-effectiveness of, 380-381, 383
installation points for, 387
electrokinetic transport, of contaminants, 282
electrolytes, effect on barrier materials, 283
electromagnetic (EM) resistance method, use
for containment monitoring, 364, 368,
369, 371-372
electromagnetic imaging, of sources and
plumes, 391
electromagnetic induction (EMI) method
use in containment monitoring, 368
use in stratigraphy, 144
electromagnetic system, use in directional
drilling, 190
electron probe microanalysis, on permeable
barrier reactants, 318
electroosmosis, as hydraulic gradient
supplement, 301
Elizabeth City (North Carolina), permeable
reactive barrier studies at, 322, 324,
334-336
EMCON Associates, hydraulic fracturing
method of, 327
EMI. See electromagnetic induction (EMI)
methods empirical flux laws, in
transport modeling, 248
INDEX
emplacement verification, 2
enhanced barriers, 32-33, 35
entrance conditions, in contaminant
transport modeling, 255-256, 287
ENVEST remediation costing program, 347
environmental containment applications, of
steel sheet piles, 77-93
Environmental Management Programs (U.S.
Department of Energy), innovative
monitoring technologies of, 390-393
Environmental Protection Agency (EPA), 2,
7, 324, 330, 347, 348
maximum contaminant levels (MCLs) of,
303
environmental remediation systems
chemical grouts for, 212, 213, 225
mathematical modeling of, 247-299
by vertical barriers, 34
environmental risk assessment, of
containment systems, 356
EnviroMetal Technologies, Inc., permeable
reactive barrier tests by, 330
EnviroWall
as geomembrane installation method, 100
as groundwater barrier, 321
EPA. See Environmental Protection Agency
EPA 40 Code of Federal Regulations, Part
264, Subpart N, stipulation of
hydraulic activity by, 8
EPA Technical Resources Document, for
adsorption parameter determination, 152
EPA 9090 test procedure, 106, 109
epoxy polymers
as barrier materials, 212, 226-227
costs of, 227, 244
hydraulic activity of, 227, 244
as interlock sealant, 82
use for steel sheet pile coating, 85
ethanes, zero-valent metal dehalogenation
of, 306
ethanol, transport studies on, 168
ethenes, zero-valent metal dehalogenation
of, 306
ethyl acetate, as sodium silicate gel initiator,
215
ethylbenzene, biodegradation of, 306, 308
ethylene glycol
ettringite, development in cement-bentonite
417
walls, 67
Eulerian methods, 260
Europe
backfill mixtures used in, 48, 56
box-out test use in, 65
cap field study in, 130-131
cement-bentonite cutoff wall technology
in, 47
cutoff wall construction QA/QC in, 63
floors and bottom barriers in, 141
thermal studies on soils in, 163
vertical barrier practices and costs in, 5, 60
excavation/replacement methods, for soiland cement-based vertical barriers, 46-48
excavations, effect on vertical barrier walls, 57
exit conditions, in contaminant transport
modeling, 253-254, 256-258, 287
Expedited Site Characterization (ESC)
program, 393
exposure risk assessment, of containment
systems, 355, 356
extensiometers, use for containment
monitoring, 364, 367
facility permitting and maintenance, of

containment systems, 355, 357-358
failure
of caps, 129-130, 135
of frozen soil barriers, 196
modeling for barriers, 282, 288
of permeable reactive barriers, 311
falling head tests, of asphalt caps, 125
FAM. See fluid applied asphalt membrane
(FAM)
Fann viscometry, 26
Far East, deep mixing use in, 50
FA-Rok 913 (furan polymer), 229
fault zones, field location of, 160
fence-post permeable reactive barriers, 301,
334-336
ferric chloride
injection into calcite-bearing aquifers, 314
injection method for, 325, 326
ferric oxyhydride
precipitation of, 314, 325
transformation to goethite, 311
ferrous oxide, chemical precipitation by, 303
fescue grass, as cap cover, 131
fiber-optic sensors
use for containment monitoring, 364, 373374, 375, 389
for VOC contaminants, 390
fibers
418
INDEX
addition to barrier wall mixes, 34, 35

use for shrinkage prevention, 16
Ficks law, 248
field-scale modeling, for contaminant
transport, 261-262
fine-grained deposits, hydraulic conductivity
of, 150, 151
fine-grained reactive iron, use in hydraulic
fracturing, 326
first-order decay
contaminant disappearance by, 270-271
modeling of, 277-281
fixation, contaminant removal by, 302
flaws, mechanical detection of, 366-368
flexible polypropylene (fPP), use in cover
geomembranes, 107
flexural strength, of cementitious grouts, 13-15
flood protection levees, thin diaphragm wall
system use in, 199
floods, Remote Access Inventory Monitoring
System use after, 392
floors and bottom barriers, 141-184
artificially emplaced. See artificially
emplaced floors and bottom barriers
block-displacement method for, 186
challenges of, 188
directional drilling for, 186, 188-190
driving potentials for, 141-147, 149, 177
field performance of, 157
frozen soil barriers, 194-196
grouted, 187-194
grout selection for, 185
hydrofracturing method for, 186
indigenous. See indigenous barriers
jet grouting method for, 186, 190-193
liquid/gas transport monitoring of, 153-158
liquid transport properties of, 150-153
man-made discontinuities in, 168-176
microtube method for, 186
microtunneling method for, 186
performance requirements for, 185
permeation grouting method for, 186, 187188
schematic diagram of, 254
sheet pile cutoff combined with injection
grouting for, 186
site monitoring of, 157
stratigraphy studies on, 141-147
technical parameters affecting, 185
techniques for, 186
technological assessment of, 157
vertical barrier use with, 185-186
Zblin System for, 186
flow behavior, of cementitious grout, 9

Flowmonta technology, for montan wax
barriers, 239
Flow Tex company, Flowmonta technology
of, 239
flow-through column tests, for permeable
flow velocity sensors, use for containment
monitoring, 375
fluid applied asphalt membrane (FAM), as
cap material, 124, 125
fluid injection methods, for permeable
reactive barrier construction, 320
fluid properties, of slurries, 26
fluids, hot and hazardous, chemical reactions
in, 147
fluorescent dye
use as hydraulic tracer, 200-201
flushing-type emplacement, of permeable
flux-based standards, for performance
criteria, 284, 287
flux predictions, of contaminant reactions,
276, 277
fly ash
in cement-bentonite walls, 30, 48
effect on grout strength, 14, 15
effect on hydraulic conductivity, 13
effect on soil-bentonite sorption, 279-280
effect on sulfate resistance, 15
effect on wet-dry cycling, 31
pulverized. See pulverized fly ash (pfa)
use to plug cofferdam leaks, 79
FMCW. See frequency-modulated
continuous wave (FMCW) radar
imaging method
foot plates, attachment to sheet pile
interlocks, 80, 81
formamide, as sodium silicate gel initiator,
215
foundation layer, of caps, 120, 136
foundation piling, aquitard perforation by,
168
fractures
assessment methods for, 160
detection in limestones, 144
DNAPL penetration into, 160
effect on contaminant transport, 158-161
field location of, 160
healing of, 160
hydraulic head profile studies of, 160
INDEX
identification in clay strata, 51
methods for studying, 160
France, radioactive waste disposal in, 163
free-phase liquids, vertical barrier walls for
control of, 68, 70
Freeport (Texas), field test on Arbed sheet
piles at, 88-89
freeze-thaw cycling effects
on compacted clay, 134
on containment systems, 359
on geosynthetic clay liners, 123
long term, 283
on natural barriers, 165-166
on sludge barrier layer, 126
on vertical barriers, 31, 34, 35, 66, 68, 96
French drains, use at landfill sites, 264
Freon 113
permeable reactive barrier for, 333
frequency domain electromagnetic (FDEM)
method, use in performance
monitoring, 372
frequency-modulated continuous wave
(FMCW) radar imaging method, for
barrier fault detection, 207-208
Freundlich isotherm, 251, 279
frost, penetration in sludge barrier layer, 126
frost heave, effects on frozen soil barriers, 196
frozen soil barriers
characteristics of, 204
costs of, 205
directional drilling for, 196-198
environmental pluses of, 196
failures of, 196
field tests on, 194-195, 200-201
for floors and bottoms, 194-196
frost heave effects on, 196
limitations of, 196
funnel and gate permeable reactive barriers,
266, 267, 312-313, 315, 339
furan
as barrier material, 212, 229-230
costs of, 230, 244
fusion welding, of geomembranes, 116
GAC. See granular activated carbon (GAC)

gamma radiation
effects on transport, 165
penetrometer probes for, 390
gamma ray attenuation, containment
419
monitoring by, 379

gas accumulations, effect on cutoff wall
performance, 66
gas-collection layer, of caps, 120, 136
gas flow, examples of, 147
gas-forming admixtures, as grout additives, 16
gas migration
control at landfill sites, 20
through caps, 121
gas potential, measurement of, 148
gas release, from contaminated materials,
control by capping, 119
GCLs. See geosynthetic clay liners (GCLs)
gelling agents, for permeation grouting, 187
gel strength, of cementitious grout, 9, 11
General Electric, permeable reactive barrier
studies of, 323
Geochemical Corporation, AC-400 [acrylate
gel grout] of, 217
geochemical methods, for fracture study, 160
geocomposite, use in cap drainage layer, 121
GeoCon, Inc., as soil-mixing equipment
supplier, 324
GeoGuard vinyl sheet piling, 86
geological strata, of containment sites, 51
geologic methods, for fracture studies, 160
geomembranes
as cap material, 112, 120, 121, 122-123, 130,
133
leak detection in, 375, 376, 377
monitoring needs for, 385-386
vibrating beam emplacement of, 324
geomembrane-bentonite composite, internal
shear strength of, 124
geomembrane vertical barriers, 95-118
advantages of, 95-96, 114
case studies on, 107-109
CQC/CQA construction of, 111-112
comparison with other vertical barriers,
115-116
costs of, 100, 114-116, 117
current practices for, 100-109
disadvantages of, 114
double-walled, 113, 115, 117, 124
emerging technologies for, 116
with geonet between backfill, 96
with geonet leak detection layer between,
96
handling of, 98-99
from high-density polyethylene. See
high-density polyethylene (HDPE)
hydrophilic gasket interlock test for, 112
impermeance of, 99-100
installation methods for, 100-105, 116
420
INDEX
interlocks of, 100, 105-106, 111

leak detection in, 115, 117, 375, 376, 377
long-term monitoring of, 112-114
needs of, 116-117
performance of, 109-114
from polyvinyl chloride, 98-100
QC/QA procedures for installation of, 96, 98
segmented trench box installation of, 101,
103, 104, 114
in situ curing of, 116
slurry-supported installation method,
101, 103, 108, 114
smart geomembranes as, 116
top seal of, 96
trenching-machine method for, 101-102,
108, 114
types of, 96, 98-100
use in cutoff walls, 2, 20, 27, 30, 35
vibrated-insertion plate method for, 101,
102, 107, 114
vibrating-beam installation of, 101, 104, 114
warrants for, 113-114
geonet
use in cap drainage layer, 121
use in geomembrane leak detection, 115
geophysical systems
for containment monitoring, 34, 368-372,
393
for containment site studies, 51
cost-effectiveness of, 380, 383-384
for performance monitoring and
evaluation, 368-372
geophysical tests, on vertical barrier walls, 65
Georgia, cap studies in, 128
GeoSyntec
geosynthetic clay liner studies at, 123-124
hydrophilic gasket interlock test of, 112
geosynthetic clay liners (GCLs), 116, 127, 137
durability of, 134
internal shear strength of, 123-124, 134,
137
types of, 124
Geosynthetic Research Institute (Drexel
University)
geomembrane research at, 122-124
polymer-degradation studies at, 110
geotextile-encased geosynthetic clay liners,
internal shear strength of, 124
Germany
adsorption determination methods in, 152
cap field study in, 130-131
flux concept use in, 287
landfill design in, 284, 287
montan wax use in, 223

permeable reactive barrier studies in, 339340
thin diaphragm wall system use in, 199
Gilson Road (New Hampshire), contaminant
transport modeling at, 263
glacial clays, matrix diffusion in, 159
glacial till, porosity of, 151
glyoxal-modified sodium silicate (GMSS), as
chemical grout, 215
goethite, ferric oxyhydride from, 311
Golder Associates, Inc.
tests on hardened material tests by, 26
tests on sodium silicate and montan wax
by, 237-238
GPR. See ground-penetrating radar (GPR)
grain-size distributions, of soil samples, 51
granular activated carbon (GAC)
sorption of organics by, 307-308
gravel, as soil-bentonite additive, 18
gravitational systems, for containment
monitoring, 368-372
gravity sensor, use for directional drilling, 196
Great Britain. See United Kingdom
Great Lakes chemical plant, geomembrane
vertical barrier use at, 107
ground deformation, from vertical barrier
walls, 54, 57, 69
ground movement. See also earthquakes
effects on vertical barriers, 68
post-construction effects of, 36, 57
in vicinity of vertical barriers, 51
ground-penetrating radar (GPR)
use for containment monitoring, 364, 369,
370-371
use in fracture location, 160
use in stratigraphy, 144
groundwater
advective flow of, 147
biological treatment of, 305
contaminated, 35, 67, 68, 70, 170
depth determination at site, 51
flow models for, 248, 262-266, 267, 268,
282, 286, 287, 311, 339
monitoring of, 364, 365-366, 380, 390-391
seamist liner for monitoring of, 390-391
seepage control of, 35
vertical barrier use for, 45, 67, 68
groundwater well systems, 388
cost-effectiveness of, 380, 383, 384
grout, use to plug interlock leaks, 79, 80
grout curtain walls
costs of, 60
INDEX
slurry walls as superior to, 67
grouted interlocks, for geomembrane panels,
106
grout mixing operation, requirements for, 59
grouts. See also individual types of grouts
cementitious. See cementitious grouts
chemical. See chemical grouts
for well and borehole sealing, 170
guar gum
in biopolymer trenches, 8
use in hydraulic fracturing, 326-327
guidance systems, for directional drilling, 197
Gulf Coast, clay-sand geology of, 144
Gundle Lining Systems, Inc., 112
Gundseal, internal shear strength of, 124
gypsum
as cement additive, 22
containment monitoring use of, 364, 379, 389
halocarbons, chemical precipitation of, 304

Halsnif, chlorocarbon detection by, 391
Hamburg (Germany), cap field study in, 130131
Hanford (Washington)
acrylic tests at, 234
asphalt cap studies at, 124-125
cap-design studies at, 132-133
colloidal silica tests at, 220
metal precipitation studies at, 325
montan wax grout tests at, 223
sodium silicate grout tests at, 216
sulfur polymer cement tests at, 225
vinylester styrene grout tests at, 232
hanging gate permeable reactive barriers, 311
hanging wall, cutoff wall as, 52
hazardous contaminant landfills
artificial bottom barriers for, 185, 187
caps for, 122
frozen soil barriers for, 200
HDPE geomembrane barrier use for, 109
monitoring of, 365-366
hazardous materials
diffusive transport of, 213
excavation spoils as, 48
geomembrane vertical barrier for, 108
hazardous waste containment, contaminant
transport modeling of, 263, 264-265
HDPE. See high-density polyethylene (HDPE)
head dissipation probes, 156-157
health considerations, for vertical-barrier
construction, 62
421
heat, effects on transport, 163-168

heat dissipation, containment monitoring by,
379
heating-cooling cycle effects, on caps, 126
heavy-weight cone penetrometer,
containment monitoring by, 390
HELP model, use for landfill cap hydraulic
performance, 128-129, 132
Henrys law coefficient, 155
hexachloroethane, dehalogenation of, 305
high-density polyethylene (HDPE)
cross-linked, 116
degradative stages of, 109-110
diffusion through, 111
durability of, 109-110, 134, 137
formulation for, 98
geomembranes from, 96, 98, 100, 105, 107,
108, 109, 122-123, 137
interlock continuity and testing on, 111
vapor transmission by, 99
warrants for, 113-114
high-temperature wastes, fractures caused
by, 283
Hill Air Force Base (Utah), permeable
reactive barrier field tests at, 331-332
HMAC. See hot-mix asphalt concrete (HMAC)
Hoechst. See Socit Franaise Hoechst
horizontal barrier systems, POLLUTE code
application to, 260
hot-mix asphalt concrete (HMAC), as cap
material, 124, 125
hot-rolled pile sections, 80, 81
human health risk assessment, of
containment systems, 356
hydraulic conductivity
borehole methods for, 151
of cement-bentonite mixes, 24, 28, 30, 54
of cementitious grouts, 7-8, 9, 13, 14
of cement-slag-bentonite mixes, 23-24, 30
of chemical grouts, 213, 216, 218, 220, 221,
222, 223, 224, 226, 227, 228, 230, 231,
233, 234, 238
contaminant transport modeling for, 261262
curing time effects on, 30-31, 34
effect on permeable reactor barriers, 315
effect on piezometric head, 148
effect on system performance, 248
EPA specifications for, 8
field methods for, 150-151, 152-153
heat effects on, 164
increased, from wall contaminant
interaction, 67
jet grouting and, 192
422
INDEX
laboratory methods for, 150

leach resistance relation to, 30
monitoring of, 66
permeating grouting and, 188
saturation effects on, 154-155
of sheet pile walls, 87
in situ, 51, 64, 69
of soil-bentonite backfill, 54
of soil-bentonite slurry trench cutoff
walls, 17-18, 28-29
studies for bottom barriers, 185
thermal effects on, 168
of vertical barriers, 7-8, 33
Waterloo barrier test cell for, 88
hydraulic fracturing method, for permeable
reactive barrier construction, 320, 326327
hydraulic gradient, pre-construction
determination of, 268
hydraulic head potentials, measurement of,
156-157
hydraulic head profile, of fractures, 160
hydraulic methods, for fracture study, 160
hydraulic potential, in geohydrologic
systems, 147, 149
hydraulic structures, hydraulic conductivity
level for, 27
hydrocarbons
in containment sites, 51
detection in waste sites, 373, 374, 375
hydrochloric acid
hydrofraise, for cutoff wall installation, 21
hydrogen, formation in radioactive waste
sites, 165
hydrogen isotopes, use in fracture studies,
160
hydrogen peroxide, formation in radioactive
waste sites, 165
hydrogen sulfide, 337
contaminant treatment by, 304
mine-waste treatment by, 325
hydrogeology, of containment sites, 51
Hydrologic Evaluation of Landfill
Performance. See HELP model
hydrolysis reactions, of organic compounds,
270
hydrophilic gasket interlocks, for
geomembrane panels, 105-106
hydrophilic gasket interlock test, for
geomembranes, 112
hydrophobic organic compounds, sorption

parameters for, 268, 280
hydrophobic partitioning, as type of
sorption, 251
hydropunch sampling, 393
hydrostatic pile driver, use for installation of
Uniloc sheets, 86
hydroxyapatite, chemical precipitation by, 303
Idaho Falls (Idaho), cap field test at, 133

Idaho National Engineering Laboratory
cap studies at, 133
tests on jet grouting of acrylic at, 241
inclinometers, use for containment
indigenous barriers
chemical effects on, 161-163
driving potentials of, 147-149
for floors and bottoms, 141-184
fracture effects on, 158-161
liquid/gas transport in, 153-158
liquid transport properties of, 150-153
man-made discontinuities in, 168-176, 177
material properties affecting contaminant
flux, 177
performance monitoring of, 145, 176, 385
stratigraphy of, 142-145, 176
technological assessment of, 145
thermal and radiation effects on, 163-168
infiltrometers, use for containment
injected treatment zone permeable reactive
barriers, 311, 314, 325-326
injection grouting, for floors and bottom
barriers, 186
injection-well emplacement, of permeable
inorganic chemicals, in containment sites, 51
inorganic compounds
diffusivities of, 268
sorption of, 308-309
inorganic contaminants, biomass treatment
of, 305
inorganic solutes, adsorption of, 152
insertion plate, for geomembrane
installation, 102
in situ mixed walls, cost-effective, 1
in situ mixing method, costs of, 60
in situ remediation, cost-effective, 1
in situ testing, of containment sites, 51
intelligent structures, 389
interlocks
of geomembrane vertical barriers, 100,
INDEX
105-106, 111
leakage through, 77, 79, 92
of polyethylene sheet pile, 86
sealing of, 79, 81-82, 86
International Groundwater Modeling Center,
262
ion exchange, as type of sorption, 251
iron
inorganic compound sorption by, 308-309
permeable reactant treatment of, 326-327
use in permeable reactive barriers, 327335, 345, 346-347
iron foam, use in permeable reactive barrier
tests, 338
iron hydroxide, formation by permeable
iron hydroxides
costs of, 222, 223, 244
iron oxide, chemical precipitation by, 303
iron oxyhydrides
inorganic compound sorption by, 308
precipitation of, 326
iron permeable reactive barriers, 327-328
iron-sand backfill, use with permeable
reactor barriers, 321, 335
iron turnings, use in permeable reactive
barriers, 332
Irving-Williams order of complex formation,
308
isocyanate-cured epoxy pitch, use for steel
sheet pile coating, 85
isotopic methods, for fracture study, 160
Italy
directional drilling for bottom barriers in,
188
geomembrane vertical barrier use in, 107
radioactive waste disposal in, 163, 167
thin diaphragm technique use in, 100, 191
Japan
backfill mixtures used in, 48, 56
deep mixing use in, 50
jet grouting, 7, 12, 16, 45, 46
advantages of, 7, 12, 16, 45, 46, 192-193
barrier wall construction by, 49-50
of bottom barriers, 190-193, 199
of chemical grouts, 211, 225, 226, 239-241
costs of, 60, 205
depiction of, 49
423
European costs of, 60

field tests of, 199, 239-241
limitations of, 193
materials for, 8
for permeable reactive barrier
construction, 320, 322, 346
technological needs for, 207
using montan wax, 239-241
junipers, as cap cover, 131
Kelly-mounted grab, for cutoff wall

installation, 21
Kent (Washington), vertical barrier use in, 65
Kielder Dam, cutoff wall under, 20
Klebolinks, as glyoxal-modified sodium
silicate grouts, 215-216, 238
kriging, use in stratigraphy, 145
krypton-85, use in fracture studies, 160
Landers earthquake (1992), effect on landfill

site, 127
Landfill Assessment and Monitoring System
(LAMS), 393, 394
landfill liners, contaminant transport
modeling of, 248, 256, 262, 272, 273
landfills. See also types of landfills
artificial barriers for, 185
capping of, 122
contaminant transport modeling of, 263-264
cutoff walls for, 20, 67
field-study modeling of, 273-274
geomembrane vertical barriers at, 113
for hazardous contaminants. See
hazardous contaminant landfills
Leachate Detection System of, 365
leachate effect on hydraulic conductivity,
29
of, 355-399
seismic performance of, 127, 134-135
water balance models for, 128-129
land subsidence, long-term effects of, 283
Laplace-transformed solutions, of advectivedispersive reactive equation, 260
Lasagna process, for permeable reactive
lateral squeezing, in soil-bentonite backfilled
trenches, 55, 56
latex
effects on hydraulic conductivity, 13
as grout additive, 9, 11, 14-15
Lawrence Berkeley National Laboratory
424
INDEX
colloidal silica tests at, 220, 221, 235

fiber-optic sensor of, 374
polysiloxane tests at, 227
leachate action, long-term effects of, 283
Leachate Detection System (LDS), of
landfills, 365
leachate migration, control at landfill sites, 20
leachate monitoring, 365-366
leach resistance
of cementitious grout, 15
hydraulic conductivity relation to, 30
lead, adsorption of, 309
leak detection
in containment systems, 361, 392
in geomembrane vertical barriers, 115,
117, 375, 376, 377
levees, vertical barrier cutoff use in, 67
lime, leaching from cement-bentonite mixes,
28
limestone
formation by permeable reactive barriers,
331, 332
fracture detection in, 144
as neutralizing agent for reactive barriers,
337
linear heat diffusion, in clays, 165
linear local equilibrium assumption, in
mathematical modeling of transport, 252
linear sensors, for containment monitoring,
373
liners, seamist for, 391
Lipari Landfill (New Jersey), contaminant
transport modeling of, 263-264
liquid diffusivity, calculation of, 268
liquid/gas transport
role of diffuse double layer in, 154-157
in unsaturated conditions, 153-158
loess, porosity of, 151
long-term damage to barriers, causes of, 283
long-term performance
of cement-based vertical barriers, 66-67
of chemical grouts, 242-243
of soil-based vertical barriers, 66
of vertical barriers, 35, 66-67, 71
Los Angeles, earthquake effects on landfill
near, 127
Love Canal (New York), containment
transport modeling of, 264-265
low hydraulic conductivity barriers
assessment of, 272
boundary conditions in, 276
contaminant transport modeling for, 261-
262
first-order decay in, 277-278
groundwater flow modeling for, 262-266
laboratory studies on, 272-273
one-dimensional models for, 272-281, 286287
reactive transport in, 277-281
sensitivity analysis of, 274-280
sorption in, 278-280
transport in, 284
low-permeability barriers, 301
modeling analysis of, 247, 256
transport equation for, 296-297
low-temperature aqueous geochemistry, 162
low-water zone, corrosion of steel piling in,
83
Ludox SM colloidal silica
field tests on, 236-237
lysimeters
use for cap performance tests, 131-132
use for geomembrane vertical barrier leak
detection, 113
use for performance monitoring, 389
use for soil-water determination, 156, 157
magnetic sensor, use for directional drilling,

196
magnetometers, use for directional drilling,
196-197
mandrel-based methods, for permeable
reactive barrier construction, 320, 322324, 346
manganese oxyhydroxides, inorganic
compound sorption by, 308-309
marine environments, steel-piling corrosion
in, 82-84
Marsh funnel flow time, of slurries, 26
Martin Chemicals, Chement 2000 of, 225
mass conservation equation, for bulk fluid,
250
Master Builder iron, use in permeable
matrix diffusion, in fractured media, 159
maximum contaminant levels (MCLs)
California requirements for, 333
EPA requirements for, 303
3M Company, 3M 4R Concrete Restorer of, 233
mechanical strain, in cement-bentonite walls,
67
metal contaminants, precipitation of, 302
metal couples, chemical precipitation by, 303
metal precipitation
INDEX
in cement-bentonite walls, 67
by hydrogen sulfide, 325
metals, on-site adsorptive stripping analysis
of, 390
methacrylic monomers. See acrylic
methane
migration in/from waste masses, 147
from stored montan wax, 224
zero-valent metal dehalogenation of, 306
methanol
vapor transmission by high-density
polyethylene, 99
methylenebisacrylamide cross-linking
monomer, in AC-400 [acrylate gel
grout], 217
methylene chloride, fiber-optic sensor for, 390
methylethyl detone peroxide, as polyester
styrene initiator, 230
microcracking, effect on hydraulic
conductivity, 66
microstructure, of low hydraulic
conductivity barrier materials, 36
Microswitch Division (Honeywell), smart
sensors of, 389
microwave method, for porosity, 151
microwave sensors, use for containment
monitoring, 364, 369, 372
mineral-enhanced barriers, 32-33
minerals, as cementitious grout additives, 10
mineral wax. See montan wax
mine wastes, hydrogen sulfide in treatment
of, 325
mix-batch variation, effects on vertical
barriers, 35
mixed-place methods, for vertical barrier
construction, 48-50
mixing operations
in construction, 59
QA/QC tests on, 63
specifications for, 61
modeling
of contaminant transport, 2, 247-299
for environmental remediation systems,
247-299
for permeable reactive barriers, 344-345
MODFLOW code, use in contaminant
transport modeling, 263, 286, 339
moisture content sensors, for cap system
monitoring, 389
molecular diffusion
425
in barrier systems, 248

as dominant transport process, 250
molybdenum
adsorption of, 309
reactant barrier removal of, 337
Monsanto, permeable reactive barrier studies
of, 323
montan wax
costs of, 224, 244
grout composition of, 223
jet grouting of, 239-241
SoilSaw emplacement of, 241
use in permeation grouting, 237-239
use in thin diaphragm wall technology,
199, 203
Motorola, smart sensor technology of, 390
MT3D transport model, 261, 286
municipal solid waste (MSW) landfills
cap requirements for, 122, 126
seismic resistance of, 127-128
NAPLs. See nonaqueous phase liquids

(NAPLs)
National Institute for Certification of
Engineering Technologists (NICET),
field inspection program of, 111-112
natural attenuation process, applicability of, 53
Natural Gel bentonite, use in well sealant, 170
NDMA. See n-nitrosodimethylamine
(NDMA)
Neat Cement, as well sealant, 170
needlepunched GCLs, internal shear
strength of, 124
neoprene, use in geomembrane gaskets, 105
Neumann boundary conditions, in barriers,
255
neutron gauges, use for containment
neutron probe test, for soil moisture, 133,
156, 157
New York State Center for Hazardous Waste
Management, 1
New York State Department of
Environmental Conservation, 264, 265
nickel, permeable reactant treatment of, 326327
Nickel Rim tailings site (Ontario), permeable
reactive barrier field tests at, 336-337
nitrate
426
INDEX

as groundwater contaminant, 338-339
nitric acid
acrylic grout resistance to, 235
furan resistance to, 229
polyester styrene resistance to, 231
sulfur polymer cement resistance to, 225
vinylester styrene resistance to, 232
nitrobenzene, dehalogenation of, 306
n-nitrosodimethylamine (NDMA),
nonaqueous phase liquids (NAPLs)
contamination behavior of, 249
dissolution of, 282
effect on frozen soil barriers, 207
potentiometric level of, 147
vertical barrier use for, 45
nonequilibrium models, use in mathematical
modeling of transport, 252
nuclear radiation, effect on HDPE
geomembranes, 123
nuclear wastes, thermal gradients of, 147
Nyacol 1440 and Nyacol DP 5110 colloidal
silicas
field tests on, 235-236
properties of, 220
Oak Ridge (Tennessee), frozen soil barrier

tests at, 194, 200, 204
octanol-water partition coefficient, 270
OII Landfill (Los Angeles), earthquake
effects on, 127
Oil Companies Materials Association
Specification, DFCP-4, Drilling fluid
materials, Bentonite, 25
oil-drilling waste, geomembrane vertical
barrier for, 108-109
oil fields
brine monitoring of, 372
geomembrane vertical barrier use in, 107108, 108-109
oil recovery operations, metal precipitation
for, 221
oil-well cementing, 22
oily-waste landfills, cost of monitoring of,
382
one-dimensional transport models, 252-258,
262
open trench, stability studies on, 51
optical light microscopy, of permeable
barrier reactants, 318
optic fiber sensors, use in site studies, 148

ORC. See oxygen release compound (ORC)
ore deposits, as analogs to injected
permeable barriers, 326
Oregon Graduate Institute, permeable
reactive barrier tests by, 334
ore mining contaminants, precipitation
treatment of, 304
organic carbon
effects on soil-bentonite sorption, 279-280
organic compounds
adsorption of, 152
chemical precipitation of, 304
diffusion coefficients of, 155, 273
hydrolysis of, 270-271
hydrophobic, sorption parameters for, 268
liquid diffusivity of, 268
redox reaction treatment of, 305
sorption of, 307-308
transport behavior of, 156, 274
organo-clay complexes, organic contaminant
sorption by, 309
organo-zeolites, inorganic compound
sorption by, 308
osmotic transport, of contaminants, 282
oxidation-reduction, contaminant removal
by, 302
oxygen, formation in radioactive waste sites,
165
oxygen-18, use in fracture studies, 160
oxygen release compound (ORC), chemical
precipitation by, 303
oxygen-release compound (ORC), use in
BETX degradation, 307
oxyhydroxides
inorganic compound adsorption by, 309
in situ generation of, 221
Pacific Northwest Laboratory

asphalt cap studies at, 125
UNSAT-H model developed at, 128
Paducah (Kentucky), permeable reactive
barrier studies at, 325
palladized iron, use in contaminant
dehalogenation, 305
panel construction method
advantages of, 57
for cement-bentonite cutoff walls, 47
joints in, 58
leakage at joints of, 68
paper mill sludge
in cap barrier layer, 125-126, 134
INDEX
as cap material, 122
parameter estimation, for barrier design, 280281
PATH3D, use in groundwater flow studies,
339
PCE
dechlorination of, 305
permeable reactive iron barriers for, 327,
330, 331, 332, 347
peat
sorption of organics by, 307-308
use in permeable reactive barriers, 307
Peclet number, 266
penetrometer soundings, of containment
sites, 51
Penn State University, surface acoustic wave
sensors of, 391
perfect flushing condition, of boundaries,
257
performance monitoring and evaluation,
355-399
approaches to, 362-365
barrier integrity monitoring, 362-365
cost-effectiveness factors in, 379-384
costs of, 379-384
electrical systems for, 372-375
electrochemical systems for, 372-375, 380,
383, 384
of exposure risk, 355, 356
external monitoring, 364-365
for facility permitting and maintenance,
355, 357-358
flaw detection by, 366-368
for future maintenance needs, 355
geophysical systems for, 368-372, 380
groundwater and leachate monitoring,
365-366, 380, 383, 384
innovative techniques for, 389-393
installation and retrieval needs for, 386
modeling analysis of, 2, 247-248
multi-phased approach to, 393
needs for, 385-388
network system needs for, 386-387
objectives of, 355-358
source term estimation of contaminant
concentrations, 355, 356-357
strain gauging of, 366-368
summary of, 360, 363, 368-393
techniques for, 365-379
user interface needs for, 388
permeable reactive barriers (PRB), 2, 35, 301353
accelerated field experience for, 345-346
427
advantages of, 342

batch reaction tests on, 317-318
biodegradation reactions in, 305-307
configuration alternatives for, 347
constructibility and emplacement
methods for, 319-327, 346
continuous wall type, 311, 313-314, 320322
cost evaluation for, 340-342, 347
definition of, 301
development needs for, 344-348
disadvantages of, 343-344
enhanced permeable zones for, 315
excavator-based construction of, 321-322
factors affecting hydraulic performance
of, 314-315
failure of, 311
fence-post type, 301, 334-336
flow dispersion effects on, 318-319
flow-through column tests for, 318
flushing-type emplacement of, 302
funnel and gate type, 266, 267, 311, 312313, 315, 339
future tests planned for, 337-340
geochemical interactions in, 318
geologic parameters for, 310
hanging gate type, 311
hydraulic control systems of, 311-315
hydraulic fracturing method for, 320, 326327
hydrogeologic conceptual model for, 310
injected treatment zone type, 311, 314
injection-well emplacement of, 302, 325326
installation of, 301, 302
jet grouting method for, 320, 322, 346
life-span limitations and fate of, 319
mandrel-based emplacement of, 320, 322324, 346
micro-scale reactant tests on, 318
modeling and design tools for, 344-345
operation and maintenance costs of, 341
oxidation-reduction reactions in, 304-305,
306
parts of, 302
patents on, 348
performance monitoring of, 389
reactant effects on groundwater, 319
reaction zone plugging in, 311
reactive materials for, 302
reactive vs. passive, 302
428
INDEX
reactor system design for, 315-319

reactor system residence time in, 316-317
recommendations for, 347-348
requirements for, 302-303
safety factors for, 318
site characterization for, 310-311
soil-mixing emplacement of, 320, 324
sorption reactions for, 307-309
state-of-practice for, 302-327
system design protocol for, 310-311
technological evaluation of, 340-344
transport modeling of, 247, 248, 256, 266267
treatability studies for, 317-318
treatment-process research on, 346-347
treatment technologies for, 303-310
trench emplacement of, 302, 320, 322
two-layered, 307
vibrating-beam emplacement of, 320, 324
zero-valent metal dehalogenation in, 305
permeameter, use in leachate studies, 283
permeation, in old barrier walls, 36
permeation grouting, 7, 16
chemical grouts for, 211, 214, 221, 227,
235-239
costs of, 205
description of, 8
for floors and bottom barriers, 186, 187-188
materials for, 8
using colloidal silica, 235-237
using montan wax, 237-241
using sodium silicate, 237-239
personnel, importance of continuity of, 64
PES. See polyester styrene (PES)
pesticides, as indicators of recent recharge,
160
petroleum products, downhole organic
chemistry of, 148
petroleum wastes, geomembrane vertical
barrier use for, 107
pfa. See pulverized fly ash (pfa)
pH
effect on permeable reactive barriers, 317
phenol-formaldehyde, as chemical grout, 7
Philips Components, smart sensor
technology of, 390
pH measurements, on potential sites, 148
physical transformation, contaminant
removal by, 302
piezocone penetration test (CPT)
hydraulic potential measurement by, 147,
148
for in situ hydraulic conductivity, 64, 389
piezometers, use for containment

monitoring, 364, 375, 376, 379
piezometric head potential
determination of, 147, 149
effects on hydraulic conductivity, 155
pipe installation, horizontal directional
drilling for, 193-194
pipe trenches, as contaminant pathways, 169
plants, with high transpiration, use in caps,
138
plastic concrete
aggregate use in, 25
use for earth dam repair, 19
plastic concrete cutoff walls, 46, 56
construction of, 47-48
depiction of, 47
in Europe, costs, 60
plug-flow reactors, in permeable reactive
walls, 266
plumes
electromagnetic imaging of, 391
tracking of, 369, 371
point sensors, for containment monitoring,
373
pollutants, vertical barrier use for control of, 67
POLLUTE code
fractured-zone modeling by, 282
solution of advective-dispersive reactive
equation by, 260
transport predictions generated by, 273
useful features of, 262
polyacrylamide, as chemical grout, 7
polyacrylic acid, as chemical grout, 7
polyester styrene (PES)
polyethylene sheet pile, interlocks of, 86
Polymer Construction Products, Ltd., Uniloc
sheet piles from, 86
polymer grout, as backfill for
geomembranes, 103, 104, 107
polymers, as bentonite additives, 21
polynuclear aromatic hydrocarbons, sorptive
barrier removal of, 339-340
polyphosphates, oxide surface modification
by, 309
polypropylene fibers, as soil cement
additive, 9
polysiloxane
components of, 227-228
INDEX
field tests of, 235
permeation grouting of, 187, 188, 235-236
polyurethane, geomembranes from, 109
polyvinyl chloride (PVC)
geomembranes from, 99, 109
use in cover geomembranes, 107
pore pressure buildup, from heat exposure,
164, 166
pore pressure probe test, of hardened
material, 26
pore system geometry, of porous media, 250
porosity, determination of, 151, 153, 155
porous media, contaminant transport in, 248,
260
Portland cement
in cementitious grouts, 9, 24
grouts based on, 7
slag mixture with, 25
use in UK slurry trench method, 22
use in well sealant, 170
Port Washington (New York), landfill site
monitoring at, 371
post-construction sampling and testing, use
in system performance verification, 34
potassium, in Portland cement, 22
potassium ferricyanide inhibitor, in AC-400
potential-energy gradient, in contaminant
transport, 147
potential fields, of indigenous barriers, 177
PQ Corporation, Nyacol 1440 and Nyacol
DP5110 colloidal silicas of, 220, 235
PRB. See permeable reactive barriers (PRB)
precipitation, contaminant removal by, 302
Princeton Transport Code, 261
probabilistic modeling, of contaminant
transport, 281-282
programmable logic controllers (PLCs), for
sensors, 389-390
protection layer, of caps, 120, 136
PSX-527 (polysiloxane grout), 227
psychrometers, use for containment
monitoring, 364
pugmill, mixing considerations for, 59
pulverized fly ash (pfa), use in cement mixes,
19, 20, 24
pump-and-treat methods, 53, 283
disadvantages of, 1, 342, 344
pumping tests
on box-out enclosures, 65
for system performance verification, 34
429
on vertical barrier cutoff walls, 64, 65, 69

pyrite, use in contaminant dehalogenation, 305
QA/QC construction procedures, 32, 53, 70-71

field performance aspects of, 62-64
need for improvement in, 69
requirements in, 59
use in system performance verification, 34
QC Chemicals, furan polymer from, 229
QC/QA construction procedures, for
geomembrane vertical barriers, 111
Quick-Gel, use in well sealant, 170, 171
radiation
effects on contaminant transport, 163-168
vinylester styrene resistance to, 232-233
radioactive decay, of contaminants, 270, 271,
277
radioactive metals, reactant barrier removal
of, 337
radioactive waste facilities
caps for, 122
chemical grout use in, 213
epoxy grout use at, 226
thermo-hydro-stress models for, 167
transport in, 163
radioactive water storage, service lives of, 355
radiolithic gases, formation in Belgian clay, 165
radionuclides
decay reactions of, 271, 280
effect on frozen soil barriers, 207
real-time detection of, 390
radio waves, use in containment monitoring,
369, 372
radium-226, as groundwater contaminant, 339
rapid geophysical surveyor (RGS), for
containment monitoring, 391
Raychem Corporation, cable detection
system of, 373, 374, 383
RCRA. See Resource Conservation and
Recovery Act (RCRA)
reactive transport, modeling of, 277-281, 284
redox reactions. See oxidation reduction
reactions
regional flow system models, 267
Reichhold Chemicals, Atlac 4010A from, 230
remediation, vertical barrier use for, 45
remediation scheme, design of, 51
remediation sites, excess data collection at, 145
Remote Access Inventory Monitoring System
430
INDEX
(Eco Instrumentation), 392

remote mixing, of slurry walls, 59
resin, in high-density polyethylene
formulation, 98
Resource Conservation and Recovery Act
(RCRA)
lack of vadose zone monitoring
regulations in, 66
regulations for municipal waste landfills,
126-127, 346
retardation coefficient
for barrier materials, 280
determination of, 156
retarders, for cementitious grouts, 12
RGS. See rapid geophysical surveyor (RGS)
Rhodamine, use as hydraulic tracer, 200-201
Richland (Washington)
asphalt cap studies at, 125, 133
cap design studies at, 125-126
permeation grouting tests at, 238
risk assessment, modeling analysis of, 247
river crossings, directional drilling at, 197,
204, 205
roadbed, surface acoustic wave sensors for,
391
road repairs, sulfur polymer cement use for,
225
rock stratigraphies, permeation grouting use
for, 8
rod-string device, use in directional drilling,
188, 190
root penetration, of caps, 136
Roslyn/Beacon Hill Landfill, monitoring
wells at, 371
RSKERL, permeable reactive barrier tests by,
334, 337
Rust GeoTech, injection-well method of, 325
safety considerations, for vertical-barrier

construction, 62
saline environment
attapulgite wall usage in, 24
corrosion of steel piling by, 82-84
sampling, at contamination sites, 174
sand
as grout additive, 9
cap field tests at, 133-134
future reactant barrier tests at, 337
geophysical tests developed by, 65
montan wax grout tests at, 223, 238
sodium silicate grout tests at, 216, 238
San Joaquin Valley (California), colloidal

silica tests in, 235
S-Area (New York), containment transport
Savannah River Plant (USDOE), permeable
reactive barrier system at, 321
SAWS. See surface acoustic wave sensors
(SAWS)
scan chemistry, of potential sites, 148
scanning electron microscopy, of permeable
barrier reactants, 318, 329
scarifying
for bottom barrier construction, 193-194
costs of, 204
Scientific Ecology Group facilities, frozen soil
field test at, 200
SDRI. See sealed double ring infiltrometer
(SDRI) tests
sealants
for potential barrier strata penetrations, 142
for steel sheet pile interlocks, 81-82, 86, 92
for wells, 169-175
Sealed Borros anchor, use for containment
monitoring, 364
sealed double ring infiltrometer (SDRI) tests
of asphalt caps, 125
for liquid/gas transport, 153
seamist, containment monitoring by, 390-391
sea water, corrosion of steel piling by, 82-84
seepage velocity
derivation of, 250, 253
through composite barriers, 258
segmented trench box, for geomembrane
installation, 101, 103, 104
seismic activity. See earthquakes
seismic impact zone, definition of, 127
seismic reflection, use in hydraulic potential
measurements, 147
seismic reflection-refraction instrumentation,
use for bottom emplacement
verification, 203
seismic refraction surveys, use for
containment monitoring, 364, 368, 369,
370
seismic tomography, use for containment
monitoring, 364
selenate, sorption of, 308
selenium
as groundwater contaminant, 339
reactant barrier removal of, 337
semi-infinite system, one-dimensional
transport as, 256
sensitivity analysis, of low hydraulic
conductivity barriers, 274-280
INDEX
sensor research, for monitor development, 389
SERDP. See Strategic Environmental
Research and Development Program
(SERDP)
set time
of cementitious grout, 9, 12
of chemical grouts, 212-213
settlement platforms, use for containment
monitoring, 364
shales
matrix diffusion in, 159
sorption of organics by, 307-308, 340
shear strength
of cap components, 134
slag effects on, 25
shear-wave logging, use in stratigraphy, 143
sheet pile walls, 2, 71-93
barrier permittivity of, 87
cathodic protection for, 86
coatings for, 85-86
construction monitoring of, 86-87
corrosion of, 77, 82-86
costs of, 87, 341
factory-joined piles for, 82
geological limitations of, 77
interlocks of, 77, 79, 85
leakage through, 77
limited application of, 77
in marine environments, 82-85
monitoring needs for, 385-386
polyethylene, 86
recommendations for extending life of, 85
slurry walls as superior to, 67
steel, 77-86
strength of, 77
Shiprock (New Mexico), future permeable
reactive barrier tests at, 338-339
shrinkage, of cementitious grouts, 16
siderite, formation by permeable reactive
barriers, 331
silica-based fibers, for fiber-optic sensors, 389
silica fume
effects on grout strength, 14-15
effects on sulfate resistance, 15
silicate-silane gel, in aggregate systems, 30
Silicon Valley (California)
contamination site in, 19
soil-bentonite cutoff wall in, 56
single-rod system, of jet grouting, 49, 202
site conceptual stratigraphic model, 146
431
site geology, studies for bottom barriers, 185

site investigation requirements
in design phase, 51
in feasibility phase, 51
site investigation techniques, 50-52
SITE program, permeable reactive barrier
test monitoring by, 330
slag
as cementitious grout additive, 9, 10, 11,
12, 20
in cement-slag-bentonite mixes, 22-24,
30, 48
effect on compressive strength, 16
effect on grout strength, 14-15
effect on shear strength, 23
effect on sulfate resistance, 15
slickensides, in clay strata, 51
slope stability, of caps, 122, 126, 137, 138
sludge, from paper mills. See paper mill
sludge
slug tests
within barrier walls, 65
for hydraulic conductivity, 180
slurry, fluid properties of, 26
slurry cutoff walls
costs of, 341
in European costs of, 60
guide specifications for, 62
performance monitoring of, 389
service lives of, 355
slurry QA/QC importance in, 63
slurry mixing operation, requirements for, 59
slurry-supported installation method, for
geomembrane vertical barriers, 101,
103, 108, 114
slurry trench cutoff walls, composition of, 5
slurry trench/geomembrane wall, use at
Dead Sea, 20
slurry trench method
costs of, 60
materials used in UK for, 21
slurry walls
conditions for use of, 8
mixing of, 59
use at hazardous waste sites, 67
smart structures, 389, 390
smectite clay, reaction with lime, 22
smectites, thermal illitization of, 164
Socit Franaise Hoechst, as glyoxalmodifed sodium silicate developer,
215, 237
432
INDEX
sodium, in Portland cement, 22

sodium bentonite, calcium bentonite
compared to, 21
sodium chloride, acrylate grout resistance to,
218
sodium dithionite
contaminant treatment by, 304
use for chromium precipitation, 325
sodium hydroxide
sodium percarbonate, use in hydraulic
fracturing, 326
sodium silicate
as cementitious grout accelerator, 12
chemical resistance of, 216-217
costs of, 216, 244
expected lifetime of, 213-214, 244
gelation of, 215
use in permeation grouting, 237-238
software, for modeling, need for, 284-285
soil, organic carbon fraction of, 270
soil-based vertical barriers, 5-43
applications of, 68
curing time effects on, 30-32
design and construction of, 45-76
freeze-thaw cycling effects on, 31
long-term performance of, 66
types and materials for, 5
wet-dry cycling effects on, 31, 213
soil-based vertical walls, mixing operations
in, 59
soil-bentonite backfill
compressibility of, 57
movement effects on, 57
resistance to chemical attack, 57
stresses of, effect on hydraulic
conductivity, 54
water content effects of, 58
soil-bentonite backfilled trenches
lateral squeezing in, 55
soil arching in, 54-55, 56

use in U.S., 50
soil-bentonite barriers
assessment of, 33
wet-dry cycling effects on, 31
soil-bentonite cutoff walls, 46, 56
barrier permittivity of, 87
construction of, 46-47
contaminant transport modeling of, 263
cost of, 58, 70
depiction of, 46
geomembrane composite with, 96
stresses in, 57
soil-bentonite mixtures
aggregate use in, 24-25
compressibility of, 18
fly ash effects on sorption of, 279-280
strength of, 18-19
use in United States, 34
soil-bentonite slurry trench cutoff walls, 5
hydraulic conductivity of, 17-18
soil-bentonite slurry wall, tortuosity effect in,
269
soil-bentonite walls
costs of, 60
working platform for, 59
soil borings, stratigraphy determination by,
143
soil-cement-bentonite cutoff walls (SCB), 56
construction of, 48
soil-cement-bentonite materials, 17
mixing considerations for, 59
soil-cement-bentonite slurry wall, use with
permeable reactive barriers, 332-333
soil-cement cutoff walls
soil cements
cracking of, 16, 17
strength of, 14, 15
soil-cements, as backfill for geomembranes,
103, 104, 107
soil chemistry, effect on cementitious grouts,
12
soil classification, for stratigraphic studies,
146
soil-column experiments, modeling in, 273
soil columns, artificial barriers for, 185
INDEX
soil-freezing technology, needs for, 207
soil gas, monitoring of, 156, 364
soil grouts, for bottom barriers, 185
soil mixing methods, for permeable reactive
barrier construction, 320, 324
soil-mixing methods, 7
soil moisture, in diffuse double layer, 154
soil porosity, studies for bottom barriers, 185
soils
arching in, 54-55, 56
classifications of, 51
index property tests on, 51
lateral squeezing in, 54
stream or air stripping of, 147
SoilSaw
epoxy grout emplacement by, 227
joint reduction by use of, 58
montan wax emplacement by, 241
use in jet-grouting method, 50
soil sawing, use for geomembrane
installation, 116
soil vapor
seamist liner for monitoring of, 390-391
vertical barrier use for, 45
solvents, as indicators of recent recharge, 160
sonic units, use in potentiometry, 147
sorption
contaminant removal by, 302, 307
laboratory studies of, 280
parameters for, in contaminant flow
modeling, 270
types of, 251
source term estimation, of contaminant
concentrations, 355, 356-357
Southeast Asia, TruTracker use for tunnel
boreholes, 198
Spain, radioactive waste disposal in, 163, 167
Spencer White & Prentis, Uniloc sheet piles
from, 86
splash zone, corrosion of steel piling in, 83-84
spreadsheets, use in transport modeling, 262,
275
stainless steel, wetting by organic DNAPLs,
175-176
Standard Penetration Test blow count
determinations, at contaminated site, 51
steel pilings, wetting by organic DNAPLs,
175-176
steel sheet pile walls, 77-86
case-history need for, 91-92
corrosion of, 82-86
costs of, 90-91
433
diffusive transport through, 90

health and safety concerns of, 90
interlock leakage in, 77-82, 92
of irregular shapes, 90
limitations of, 91
removal of, 90
sealants for, 86
use with permeable reactor barriers, 321
steel surfaces, corrosion of, 84
steel wool, use in permeable reactive barrier
tests, 338
stitched GCLs, internal shear strength of, 124
Stokes-Einstein equation, 268
storm and sewer lines, acrylate gel sealants
for, 217
STP N-values. See Standard Penetration Test
blow count determinations
strain at failure, specifications for, 27
strain gauges
containment monitoring by, 366-368
frozen-soil field test using, 200
Strategic Environmental Research and
Development Program (SERDP),
permeable reactive barrier research by,
347
stratigraphy
conceptual model for, 144-145
data interpretation in, 146
determination for floors and bottom
barriers, 141-147
downhole geophysical studies in, 143-144
electrical resistivity method, 144
electromagnetic induction methods, 144
field performance of, 145
of indigenous barriers, 176
intrusive methods for, 142-144
model for, 146
non-intrusive methods for, 144, 147
seismic reflection and refraction methods
for, 144
soil borings to determine, 143
test pits to determine, 143
strength, specifications for, 27
stress-chemical hydraulic effects, on
contaminant flow, 168
strontium, adsorption of, 309
Sudbury (Ontario), permeable reactive
barrier field tests at, 336-337
sugars, as cementitious grout retarders, 12
sulfate
434
INDEX

cementitious barrier material expansion
from, 28, 67
cement resistant to, 13, 15, 30
as groundwater contaminant, 338-339
slag resistance to, 8
sorption of, 308
sulfur polymer cement
costs of, 226, 244
hydraulic activity of, 226, 244
sumps, for leachate collection in landfills, 365
Sunnyvale (California), permeable reactive
barrier field tests at, 327, 332-334, 341
Superfund Priority Cleanup List, 264
Superfund Record of Decision (ROD), for
Hill Air Force Base site, 332
superplasticizers
for cementitious grouts, 9, 10, 11, 12
effect on flexural strength, 15
effect on hydraulic conductivity, 13, 14
surface acoustic wave sensors (SAWS), for
surface barriers. See caps
surface layer, of caps, 120, 136
surface scanning methods, need for
improvement of, 150
surface water conditions, effect on cutoff
wall performance, 66
surfactants, use to form sorptive barriers, 325
Switzerland, radioactive waste disposal in, 163
symbols, used in contaminant flow
modeling, 298-299
tank farm, with grout enclosure barrier, 186

tape probes, use in potentiometry, 147
tar vinyl, for steel sheet piles, 85
TCE. See trichloroethylene
TDEM. See time domain electromagnetic
measurements (TDEM)
TEABPO system, field tests on acrylic using,
234
TEM. See transient electromagnetic (TEM)
soundings
temperature
effect on cutoff wall performance, 66
effect on HDPE geomembranes, 123
temperature logging, use for containment
monitoring, 364
tensiometers
use for containment monitoring, 364, 378
use for hydraulic head potential

test pits, stratigraphy determination by, 143
tetrachloroethanes, dehalogenation of, 305
tetrachloroethene, dehalogenation of, 305
Texas, regulations on stabilized
contaminants in, 319
Texas Instruments, smart sensor technology
of, 390
thaumasite, development in cementbentonite walls, 67
thermal cracking
prevention of, 8
thermal effects
on barriers, long-term, 282
chemical effects coupled with, 167
on contaminant transport, 163-168
thermal flow sensors, use for containment
monitoring, 364
thermal gradients, of waste facilities, 147
thermal potential, measurement of, 149
thermocouple psychrometry, containment
monitoring by, 379
thermo-electric flow velocity gauges, use for
thermo-hydro-stress effects, on transport, 167
thin diaphragm wall method
for bottom barriers, 188-189, 191, 192, 199
costs of, 205
tidal zone, corrosion of steel piling in, 83-84
tiltmeters, use for containment monitoring,
364
time-domain electromagnetic measurements
(TDEM), use in containment
monitoring, 368, 372, 389
time domain reflectometry (TDR)
for barrier fault detection, 133
for soil moisture, 133
for vadose zone, 158
toluene
biodegradation of, 306, 308
in montan wax, 221
tortuosity factor, 250-251
calculation of, 268-269
tracers, for frozen soil field tests, 200
tracer tests, on vertical barrier cutoff walls,
64, 65, 69
transient electromagnetic (TEM) soundings,
transport codes, use in modeling, 260
transport properties, of indigenous barriers,
153
tremie method
INDEX
plastic concrete placement by, 48
use for interlock sealing, 81
trenches
artificial barriers for, 185
in soil-bentonite and cement-bentonite
cutoff wall construction, 46, 47
trenching methods
for geomembrane installation, 101-102,
108, 114
construction, 302, 320, 322
trichloroethane, dehalogenation of, 305
trichloroethylene (TCE)
biodegradation of, 308
dehalogenation of, 304, 305
electroosmosis treatment of, 323
hydraulic fracturing treatment of, 327
permeable reactive barriers for, 327-336, 347
sorption of, 308
1,2,3-trichloropropane, dehalogenation of, 306
1,1,2-trichlorotrifluoroethane. See Freon 113
triethanolamine accelerator, in AC-400
triple-rod system, of jet grouting, 49, 202
TruTracker, use for directional drilling, 197-198
TSB test. See two-stage borehole (TSB)
test
tunnel boreholes, TruTracker use for, 198
two-phase transport, in porous media, 167
two-phase walls, 25
two-stage borehole (TSB) test, for hydraulic
conductivity, 151, 153
Type K cement, shrinkage reduction by, 16
ultrasonics
use in containment monitoring, 369, 372
use in stratigraphic studies, 143
ultraviolet radiation, effect on HDPE
geomembranes, 123
underground storage tanks, artificial bottom
barriers for, 185
Unified Soil Classification System, use for
site samples, 51
Uniloc interlock configurations, for sheet
pile, 85, 86
United Kingdom
cement-bentonite use in, 20, 30, 31, 34, 56,
67-68
435
field performance of vertical barriers in, 30

radioactive waste disposal in, 163
slurry wall barriers in, 17-29
United Kingdom National Specification for
Cutoff Walls (Draft), 27
U.S. Bureau of Mines, permeable reactive
barrier tests by, 225, 334
U.S. Department of Energy, 2, 213, 321, 323,
337, 390, 391
U.S. Environmental Protection Agency. See
Environmental Protection Agency (EPA)
United States
field performance of vertical barriers in,
29-30
vertical barrier practices in, 5
University of New Mexico, groundwater
model study at, 339
University of Texas, geosynthetic layer clay
studies at, 123
University of Waterloo (Canada)
permeable reactive barrier tests by, 321,
322, 324, 331-332, 334, 337
sheet-pile barrier studies at, 80, 87
UNSAT-H model, for caps, 128-129, 132
uranium
adsorption of, 309
chemical precipitation of, 304
as groundwater contaminant, 339
in ore-forming systems, 326
reactant barrier removal of, 337, 338, 343
Uranium Mill Tailing Remedial Action
(UMTRA), future reactant barrier tests
by, 337, 338-339
uranium mill tailings
future reactant barrier tests for, 337-338
injection well treatment of, 325
urea, in iron hydroxide grout mixture, 221, 222
urea-formaldehyde, as chemical grout, 7
urease, in iron hydroxide grout mixture, 221222
Urethane, as interlock sealant, 82
urethane polymers, as sheet pile sealants, 86
U-shaped geomembrane, installation of, 100
utility trenches, as contaminant pathways, 169
vadose zone
barrier construction in, 13, 34
chemical grout lifetime in, 213
monitoring of, 156-158, 357, 388
monitoring procedures for, 156-157
vertical barrier monitoring in, 66
vapor, vertical barrier walls for control of, 68,
70
436
INDEX
vapor phase diffusion, of contaminants, 281

vegetation, over caps, 121, 131, 133
vegetation sampling, for site monitoring, 393
vertical barrier cutoff walls
case-history documentation for, 71
contaminant concentration tests on, 65
defects in, 68
diagram of, 254
geophysical tests on, 65
leak detection in, 69
limitations of, 68, 70
long-term performance of, 66-67, 71
performance monitoring of, 64-66, 388
probing tests on, 65
pumping tests on, 64, 65
QA/QC construction procedures for, 32,
34, 53, 59, 63, 70-71
reliability of, 68
in situ hydraulic conductivity of, 69, 71
specifications for, 70
technology transfer for, 69-70
tracer tests on, 65
vertical barriers, 2
cement-based. See cement-based vertical
barriers
compatibility testing of, 36
construction defects in, 32
construction method effects on, 58
design basis of, 33
durability of, 8, 71
enhancement of, 32-33
freeze-thaw cycle effects on, 35
geomembranes. See geomembranes
hydraulic conductivities of, 7-8, 33, 35
of low hydraulic conductivity, modeling
of, 257
mix-batch variation effects on, 35
modeling of performance of, 2
monitoring program for performance of, 51
POLLUTE code application to, 260
post-construction sampling of, 29-30, 34
sheet pile walls. See sheet pile walls
site investigation requirements for, 50-52
as in situ reactor, 35
soil-based. See soil-based vertical barriers
stress effects on, 32
technique effects on, 35
use with bottom barriers, 186
water-table effects on, 31-32
wet-dry cycle effects on, 35
vertical barrier walls
analysis and design of, 52-59
depth of, 54
effect on adjacent ground, 54
field performance aspects of, 52-57
function related to design of, 53
modeling of, 281
movement effects on, 57
specifications for, 61-62
thickness of, 54
wall geometry of, 53-54
very flexible polyethylene (VFPE), use in
cover geomembranes, 107
VES. See vinylester styrene (VES)
vibrated-insertion plate method, for
geomembrane installation, 101, 102,
107, 114
vibrating-beam method
of cutoff wall construction, 50
for geomembrane installation, 101, 104,
114, 324
construction, 320, 324
vibrating wire extensiometers, use for
vibratory-beam method, European costs of, 60
vinyl chloride
maximum contaminant level for, 333
permeable reactive barriers for, 330, 332334, 336, 347
vinylester styrene (VES)
costs of, 233, 244
vinyl ether primer/finish, use for steel-sheet
pile coating, 85
viscosity
of cementitious grout, 9, 10-11, 12
volatile organic compounds (VOC)
borehole liners with monitors for, 390
downhole organic chemistry of, 148
fiber-optic sensors for, 390
in groundwater, 333
Volclay, use as well sealant, 170, 171
wall geometry, of vertical barrier walls, 53-54

Washington, cap studies in, 128
waste-containment systems
diffusive flux in, 151-152
failure rate of, 363-365, 394
geomembrane use as vertical barriers in.
INDEX
See geomembranes
HDPE geomembrane barrier use for, 109
of, 355-399
problems in, 364
vertical barrier walls for, 29, 45, 67, 69
water, potentiometric level of, 147
water-balance modeling, of caps, 122, 128129, 133, 136
Water Barrier sealable interlock, 82
Waterloo Barrier, 80, 81
Waterloo barrier test cell, for hydraulic
conductivity, 87, 88
water percolation
from contaminated materials, control by
capping, 119
through caps, 121, 128-129, 130-131
water-reducing agents, for cementitious
grouts, 10
water sealing, chemical grouts for, 212
water-table fluctuations, effects on barrier
performance, 31-32, 34, 281
water-vapor diffusion test, on high-density
polyethylene geomembranes, 99
Wayne (New Jersey), permeable reactive
barrier field tests at, 330-331
weather, effect on cutoff wall performance, 66
weathering, long-term effects of, 283
Weibul probability model, 282
welded interlocks, for geomembrane panels,
105
wells
abandoned, locating of, 170
ASTM guides for decommissioning, 170171
construction at contamination sites, 174
location methods for, 170, 175
sealing-grout model for, 170-171
sealing of, 169, 170-175
as threat to indigenous barriers, 168-173
unsealed, contamination by, 169-175
use in groundwater and leachate
monitoring, 365-366
Wenner resistivity array, for detection of
contaminant plumes, 371
wet-dry cycling
effect on GCL, 123
effects on barrier materials, 31, 34, 35, 66,
281, 283
resistance to, 9
wheatgrass, as cap cover, 133
Wheatstone bridge circuit, in piezometers,
437
378
wood piles, contaminant wicking by, 175
Wyoming, bentonites from, 21
xanthan gum, use in hydraulic fracturing,

326-327
x-ray diffraction analysis, on permeable
barrier reactants, 318, 329
x-ray fluorescence metals analysis, sensors
for, 390
xylene(s)
biodegradation of, 306, 308
polyethylene, 99
Youngs modulus values, for soil-bentonite

mixes, 18
Zemid PG, use for Uniloc sheet piles, 86

zeolites
effect on wall performance, 35
zero valent iron, chemical precipitation by,
303
zero-valent metal dehalogenation, of
contaminants, 306
zero-valent metals, contaminant treatment
by, 304, 305, 306
zinc
adsorption of, 309
permeable reactant treatment of, 326-327
zinc-based proppant, use in hydraulic
fracturing method, 327
zoning systems, for containment systems, 359
Zblin System, for floors and bottom
barriers, 186

O'Donnell Et Al. 1995 - Assessment of Barrier Containment Technologies - A Comprehensive Treatment For Environmental Remediation Applications (EPA NTIS PB96-180583)

Uploaded by

Copyright:

Available Formats

O'Donnell Et Al. 1995 - Assessment of Barrier Containment Technologies - A Comprehensive Treatment For Environmental Remediation Applications (EPA NTIS PB96-180583)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

O'Donnell Et Al. 1995 - Assessment of Barrier Containment Technologies - A Comprehensive Treatment For Environmental Remediation Applications (EPA NTIS PB96-180583)

Uploaded by

Copyright:

Available Formats

1

Although investigations and efforts to remediate contaminated subsurface

reliability of these technologies and to identify information, research, and

Jeffrey C. Evans, Bucknell University

Marita Allan, Brookhaven National Laboratory

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

2.3 STATE OF PRACTICE

2.3 STATE OF PRACTICE

activity. The mixture typically has a bentonite content of approximately 1%.

Material Choices as Influenced by Construction Method

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

pressures are frequently equal on both sides of a vertical barrier wall.

2.3 STATE OF PRACTICE

c. Deep Soil Mixing Materials. Deep soil mixing is accomplished using a

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

to be used and site conditions.

Shear Stress (Pa)

Shear Rate (s-1)

2.3 STATE OF PRACTICE

Shear Stress (Pa)

Shear Rate (s-1)

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

Shear Stress (Pa)

Shear Rate (s-1)

2.3 STATE OF PRACTICE

Hydraulic Conductivity. Water/cement ratio is the most significant parameter

Hydraulic Conductivity (cm/s)

Figure 2-4 Effect of soil/cement (s/c) and water/cement (w/c) ratios on

Normally, subsurface barriers are not required to possess

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

particularly high strength, and unconfined compressive strengths of 100 to

Hydraulic Conductivity (cm/s)

2.3 STATE OF PRACTICE

ratios, even at replacement levels of 80%. Therefore, high proportions of slag

Figure 2-6 Effect of soil/cement (s/c) and water/cement (w/c) ratios on

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

Compressive Strength (MPa)

Figure 2-7 Effect of soil/cement (s/c) and water/cement(w/c) ratios and

2.3 STATE OF PRACTICE

Crack Width (mm)

Time After Failure (days)

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

Approximate Initial Tangent Youngs Modulus

2.3 STATE OF PRACTICE

Consolidated-undrained triaxial compression tests performed on samples of

Figure 2-9 Stress-strain and strength behavior for various SB backfill

2.3.4 Slurry Wall Barriers in the United Kingdom

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

2.3 STATE OF PRACTICE

give better chemical resistance. A few cement-pfa-bentonite walls have been

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

Portland Cement. Portland cement is a product of clinkering and grinding

2.3 STATE OF PRACTICE

cement-slag-bentonite mixes. When used as a cement replacement material,

Shear Strength (mn.m-2)

95S 80S 95S 80S

Figure 2-10 Shear strength development in slurries containing slag (Jefferis,

SOIL- AND CEMENT-BASED VERTICAL BARRIERS (MATERIALS)

When cement-bentonite mixes are used, a hydraulic conductivity of less