United States
Environmental Protection
Agency
Hazardous Waste Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-87/063 Nov. 1987
&ERA Project Summary
Investigation of Slurry Cutoff Wall
Design and Construction Methods
for Containing Hazardous Wastes
Richard M. McCandless and Andrew Bodocsi
Specific technical design standards
for soil-bentonite slurry trench cutoff
walls used to isolate hazardous wastes
have not been established. A review of
current design and construction
methods was performed for summariz-
ing current engineering practice, identi-
fying areas of technical debate, and
initiating necessary research to promote
the development of rational standards.
The review of current methods was
followed by laboratory studies using
specialized test equipment to study
model cutoff walls.
An instrumented slurry test column
was developed and used to investigate
the hydraulic characteristics and im-
portance of bentonite slurry seals
formed on the walls of the cutoff trench
during construction. Testing involved
the penetration of a 5% bentonite: water
slurry into two different sands, the
formation of a different type of slurry
seal in each case, and the measurement
of their hydraulic conductivities based
upon the time-rate of flow and the
measurement of internal pore pressure
conditions. The effectiveness of dif-
ferent slurry seals varied greatly de-
pending upon the degree of filtration of
hydrated bentonite particles during
slurry penetration into granular soils. In
all cases, however, the effectiveness of
the seals alone (ignoring the contribution
of the soil-bentonite backfill) was very
low, suggesting that they cannot be
relied upon to offset the effects of latent
defects in the backfill, and that the
current practice of disregarding the
slurry seal in cutoff wall design should
not be changed.
Laboratory testing also involved an
instrumented slurry wall tank capable
of accommodating 508 mm (20 inches)
diameter, 101.6 mm (4 inches) thick
model cutoff walls. The tank was used
to evaluate the effects of overburden
pressure (vertical consolidation) and
hydraulic gradient (horizontal consolida-
tion), and to evaluate the potential for
self-remediation of hydraulic defects
("windows" through the barrier) via in
situ consolidation of the soil-bentonite
backfill. Various models were permeated
with water under varying hydraulic
gradients and vertical surcharge pres-
sures. The average equilibrium hydraulic
conductivity of the models was mea-
sured under each set of conditions.
Results demonstrated that both over-
burden pressure and hydraulic gradient
have significant and comparable effects
on the average conductivity of the wall.
Moreover, water content, unit weight,
and vane shear strength data measured
on samples of the soil-bentonite backfill
after the test clearly indicated that ef-
fective overburden stress decreased
with increasing depth in the model,
most likely due to friction between the
backfill and sand in which the model
was constructed.
Another model wall was intentionally
breached by two slot-like "windows"
representing small pockets of entrapped
bentonite slurry in the backfill immedi-
ately after construction. By incre-
mentally increasing surcharge pressure
it was possible to "heal" the windows
as evidenced by a return to the predeter-
mined baseline hydraulic conductivity
of the wall. This suggests that in situ
-------
*.,
consolidation of the backfill may help
to eliminate some types of as-built
hydraulic defects or micro-cracks within
the backfill resulting from long-term
chemical degradation.
This Protect Summary was developed
by EPA's Hazardous Waste Engineering
Research Laboratory, Cincinnati, OH, to
announce key findings of the research
project that Is fully documented In a
separate report of the same title (see
Project Report ordering Information at
back).
Introduction
Slurry trench cutoff walls were first
used in the United States in the early
1940's. Since that time, their use has
become more widespread and now in-
cludes application as hydraulic barriers
to control the movement of contaminated
groundwater from hazardous waste dis-
posal sites. Specific technical design
standards for slurry trench cutoff walls
(also known as soil-bentonite walls) have
not been established. Each application is
unique and requires site-specific engi-
neering evaluation. Nevertheless, the
current state-of-the-art involves funda-
mental concepts, performance criteria,
and methods common to all applications.
The purpose of this project is threefold:
• to compile information on current
design and construction methods
• to identify specific research needs to
promote the development of rational
standards
• to perform initial research in selected
areas of need
The first phase of the project involved
review of published literature on slurry
wall technology, interviews with owners,
engineering consultants and construction
contractors, and a general assessment of
methods and research needs. Based upon
these findings, two subsequent research
phases emphasized laboratory model
studies of slurry seals formed on the
walls of a cutoff trench during construc-
tion and small model cutoff walls in-
corporating both slurry seals and a
standard soil-bentonite backfill.
Specific objectives of the laboratory
studies were to determine or evaluate:
• the depth of penetration of slurry or
filtered slurry into typical granular
soils
• the hydraulic conductivity of various
types of seals derived from slurry
penetration and slurry filtration dur-
ing penetration into typical granular
soils
• the stability of the seals (described
above) after initial development
• in situ consolidation and the effect
of surcharge loading and hydraulic
gradient on soil-bentonite hydraulic
conductivity
• the feasibility of "window" closure
within a soil-bentonite wall due to
overburden consolidation pressures.
Current Methods
The initial phase of this study involved
a survey of current design and construc-
tion methods which form the basis of
present slurry cutoff wall technology. The
survey involved review of published litera-
ture on the subject, interviews with
selected vendors and professional practi-
tioners specializing in slurry wall applica-
tions, and visits to three slurry wall
construction sites, the report does not
attempt to quantify the variability in
present methods but simply documents
the range of philosophy and current prac-
tice in the areas of Design, Specification,
Construction and QA/QC. The specific
considerations that are least standardized,
and therefore most variable, in each
subject area are summarized below:
Design • soil-bentonite mix
design
• method of hydraulic
conductivity testing
• bentonite type
• bentonite content in
the backfill
• the use of
contaminated trench
spoils in the backfill
Specification • performance type or
materials and
methods type
Construction • backfill mixing/
handling techniques
• backfill placement
method
• equipment type
• personnel - level of
experience
QA/QC • verification of trench
depth, width and
continuity
• personnel - level of
training/experience
• responsibility -
contractor,
consultant or owner?
• frequency and
manner of backfill
testing
Laboratory Investigations
Procedures
Slurry Seals
An instrumented slurry test column
was developed to study various bentonite
slurry seals formed on the walls of the
cutoff trench during construction. The
system consists of an acrylic column
equipped with probes to measure in situ
pore pressure after the formation of a
slurry seal in different sands. Spring-
suspended inflow (head) and outflow (tail)
permeant reservoirs were employed tc
achieve constant-head test conditions. A
schematic of the system is shown ir
Figure 1. Pore pressures were monitorec
during permeation to produce data on the
depth of the slurry penetration, the
hydraulic conductivity of the overall seal
and changes in these features as a func
tion of time.
A clean fine sand identified herein a;
the "+200 sand" (retained on the no. 2CK
sieve) was used to study the surfac<
filtration (filter cake) type of slurry seal ii
the slurry test column. This sand i
predominantly fine, of roughly uniforn
size (no. 40 to no. 50 sieve size), witl
about 25 percent medium sand by weigh
A clean medium to coarse sand was use
to investigate deep filtration and rheolog
cal blockage seals. The gradation con-
prised roughly 75% medium sand an
25% coarse sand, with all material bein
retained on the no. 40 sieve ("+40 sand"
All tests involved slurry seals derive
from the penetration of a standard
percent bentonite: water slurry (weigh
volume basis). Slurry was driven into tr
test sands under controlled pressure (se,
formation pressure) for a standard peric
of five hours. Seals formed in this manni
were then permeated by water undi
variable hydraulic pressures sometime
different than the seal formation pre;
sure. Testing comprised both saturate
and unsaturated cases to model cone
tions below and above the groundwat
table, respectively.
In all cases, hydraulic conductivity da
were calculated from several paramete
measured during the test. These parar
eters included the pressure different!
between any two pore pressure probe
the physical distance between the probe
and the volume flow-rate through tl
sample (discharge per unit time).
Figure 2 shows typical pore pressu
distributions during steady flow for t
+40 and +200 sands under rougf
equivalent hydraulic gradients. In ea
case, the data demonstrate a nea
-------
Vacuum
Suspension
System
Outflow
Permeant
Reservoir
Inflow
Permeant
Reservoir
Slurry
Reservoir
Supply
Figure 1. Schematic of the slurry test column system.
Baseline Distributions: +40 Sand & +200 Sand
20
24
28
Figure 2.
12 16
Column Depth (in.)
Typical baseline (no slurry seal) pore pressure distribution for fine (+200) and
medium to coarse (+40) sands used in this study.
constant rate of head loss through the
sample prior to the introduction of slurry.
After development of a slurry seal, the
steady-state pore pressure distributions
for the +40 and +200 sands were as
shown in Figure 3a and 3b, respectively.
Data such as these were used to define
the location, thickness, and hydraulic
gradient across the seals, from which
their hydraulic conductivities were
computed.
Results
Slurry Seals
Numerous tests were performed on
both the +40 and +200 sands at seal
formation and perme.ation pressures
ranging from 9.3 kPa (1.35 psi) to 68.95
kPa (10.0 psi). Of these, only two tests of
the +40 sand and five tests of the +200
sand produced useable data. In most
other tests the slurry seals were breached
by the combined effects of cracking and
erosion (piping) from beneath. The cause
is believed to be related to minor pressure
fluctuations within the system in response
to temperature changes and/or supply
pressure changes from day to night and
vice-versa. These pressure fluctuations
would cause differential expansion/con-
traction between the acrylic column and
the sand. Such disturbance would cause
micro-cracks in the seal followed by pro-
gressive widening of the cracks via
erosion. It was possible, however, to gen-
erate comparative initial permeability data
for the seven tests described above, and
to compute the "breakthrough time" (time
for the first drop of permeant to pass
through the cutoff wall barrier) for the
two types of slurry seals.
Figure 4 is a schematic of two typical
soil-bentonite walls, showing the ex-
pected zone of slurry penetration and
seal formation in the +40 and +200 sands.
Deep slurry penetration accompanied by
Theological blockage occurs in the +40
sand, whereas a surface filtration seal is
shown for the +200 sand. In both sche-
matics, the soil-bentonite backfill is as-
sumed to be the same, having a hydraulic
conductivity of 1.0 x 10'7 cm/sec. The
depth of slurry penetration and the
hydraulic conductivity of the seal in each
case are based upon results obtained
using the slurry test column.
Assuming the same in-service head
differential across each barrier and steady
flow according to Darcy's law, it was
determined that the effectiveness of the
wall in the +40 sand based upon a break-
through criterion would be about three
times as much as that of a similar wall
constructed in a deposit of +200 sand
(93.5 years vs. 31.0 years). Moreover, the
breakthrough times of the two slurry seals
alone (no soil-bentonite backfill) was
determined to be on the order of two
weeks or less.
Procedures
Mode/ Cutoff Walls
The slurry wall tank constructed for
this study accommodates circular cutoff
walls roughly 559 mm (22 inches) in
height, 102 to 152 mm (4 to 6 inches)
thick, and up to 610 mm (24 inches) in
diameter. The tank is of stainless steel
construction and employs a pneumatic
bladder system to vertically confine and
consolidate the model wall during per-
meation in the horizontal direction. A
schematic of the system is shown as
Figure 5.
-------
12
Post Penetration Distribution: +40 Sand
(at
11
/On
7
6
4
3
>W
N
Top of Sample
0
Column Depth (in.)
Post Penetration Distribution: +200 Sand (b)
Q .
#-
4-
\
\
\
\
i
r-^j
Top of
»—«--,
Sample
*—»-^
4
Figure 3.
Column Depth (in)
Typical initial pore pressure distributions after formation of slurry seals in the
(a) +40 and(b) +200 sands
The model walls were constructed
between two concentric PVC (polyvinyl
chloride) slip forms representing the walls
of a circular cutoff trench. The forms
were positioned in the tank and backfilled
with clean fine sand in 102 mm (4 inch)
lifts creating an empty 102 mm (4 inch)
wide annular space between the forms.
This space was then filled with a 5%
bentoniteiwater slurry (weight: volume
basis) comprising the same bentonite
used in the soil-bentonite mix. The soil-
bentonite backfilling operation varied
slightly for different models but generally
involved raising both forms about 102
mm (4 inches), allowing the bentonite:
water slurry to penetrate the sand and
form a surface filtration slurry seal, and
then backfilling with soil-bentonite using
a pressurized tremie pipe. This general
procedure was repeated until the surface
of the model wall was level with the
surface of the center core of sand (sand
encircling the model wall).
After construction, the model was
readied for testing by installing a com-
bination membrane/hydraulic cutoff over
its surface and positioning concentric
load-bearing plates over each element of
the model (core sand, soil-bentonite wall,
outer ring of sand). This arrangement
allowed for differential loading and con-
solidation of the soil-bentonite wall rela-
tive to the adjacent sand bodies.
The typical testing procedure used in
evaluating the effects of overburden
pressure and gradient involved saturation
of the sand elements of the model, ap-
plication of a selected surcharge pressure,
consolidation of the soil-bentonite wall
under the applied surcharge (time esti-
mated from conventional consolidation
tests performed on the backfill material),
application of the design hydraulic head
pressure at both the top and bottom of
the saturated center core of sand (Figure
5), and the measurement of hydraulic
head and volumetric inflow at prescribed
time intervals.
Similar procedures were used in the
construction and testing of the third
model wall to evaluate the closure of
artificial slot-like windows via surcharge
pressure. The slots were intended to
model macro-defects such as small
pockets of entrapped slurry remaining
after construction of the wall. Two slots
approximately 7.9 mm (5/16 inch) wide
by 1.6 mm (1/16 inch) high were cut
into the third wall after preconsolidation
under an effective overburden of 41.4
kPa (6.0 psi) as measured at the surface
of the wall. The windows were positioned
180° apart at a depth of about 127 mm (5
inches) below the top of the wall. Both
ends of each slot were covered with a
fabric-covered wire mesh to prevent
washing the core sand into the slot during
permeation. The test procedure involved
incremental increase of overburden (sur-
charge) pressure until the slots were
effectively closed as evidence by a return
to the predetermined baseline hydraulic
conductivity of the model.
Results
Model Cutoff Walls
The testing of model slurry walls in-
volved staged incrementation of over-
burden pressure and hydraulic gradient,
followed by sampling and measurement
of unit weight, vane shear strength and
moisture content as a function of depth
in the model. Three different hydraulic
gradients (i = 21, 42, 83) were applied
under effective overburden pressures of
41.4, 82.7 and 165.5 kPa (6, 12, 24 psi)
as measured at the surface of the wall.
Figure 6 presents a chronological sum-
mary of the final equilibrium conduc-
tivities measured for each set of test
conditions. Initial hydraulic conductivities
are represented by an open triangle and
final equilibrium values by an open circle.
Two incidences of hydrofracture are in-
dicated by solid triangles.
Except for test 2(g), the data suggest a
logical trend of decreasing equilibrium
hydraulic conductivity as a function of
either increasing surcharge pressure or
increasing hydraulic gradient. The data
do not, however, reflect the correcl
magnitude of change in hydraulic con-
-------
+40 Sand Test 7-6
S/B Backfill
- 40 "4-36" -f- 40"-
ks = 557 x 10~* cm/sec
*Sb = 1 0 x W~7 cm/sec
+200 Sand Test 5-7
S/B Backfill
A/7
Surface
Filter Cake.
0.25" [_ 35.5"[ 025"
ks= 1 68 x 10 cm/sec
ksh = 1 Ox W7 cm/sec
Figure 4. Idealized conditions after construction of cutoff walls in the +40 and +200 sands.
Permeant Reservoir
(typ)
V9A
' Pore Pressure
' Probe
1
V9B
Dram
V1
V2
— — — — Air Pressure Line
Permeant Flow Line
Figure 5. Schematic of the slurry wall tank system
ductivity between successive tests. The
reason is that hydrofracture permanently
changed the properties of the wall, thus
artifically offsetting groups of data
measured after hydrofracture from other
groups of data measured before hydro-
fracture.
After the completion of test 2(g) re-
ported in Figure 6, the tank was opened
to permit inspection of conditions and
allow for sampling and testing of the
backfill. Testing involved measurements
of unit weight, vane shear strength and
water content. Data for these parameters
appear as a function of depth in Figure 1.
After sampling and inspecting of the
model a new wall was constructed for
the window closing test. After establishing
a baseline or reference value of hydraulic
conductivity, the two slot windows were
formed at the locations and depths pre-
viously described. Overburden pressure
was than gradually increased causing
the apparent hydraulic conductivity of the
model to decrease until the windows had
been effectively closed as evidenced by a
return to the measured baseline
conductivity.
Conclusions for Slurry Seals
• For seals formed on fine sands by
the surface filtration mechanism: 1)
the density of a seal is proportional
to the density of the sand in which
the seal forms and proportional to
the prevailing hydraulic head under
which the seal forms, 2) the hydraulic
conductivity of a seal is inversely
proportional to the prevailing hy-
draulic head under which the seal
forms and inversely proportional to
the density of the sand in which the
seal forms, and 3) the thickness of
the seal is a function of formation
time only.
• Based upon the unknown frequency
of chemically induced or construc-
tion-related "windows" in a typical
soil-bentonite cutoff wall, it appears
that the current practice of design
on the basis of the permeability of
the soil-bentonite backfill alone
should not be changed.
Conclusions for Model
Cutoff Walls
• The average hydraulic conductivity
of model cutoff walls was observed
to decrease both as a function of
increased overburden pressure
(vertical consolidation), and in-
creased hydraulic pressure (horizontal
consolidation due to hydraulic
gradient), as well as their combined
effect.
• Hydrofracture, or rupture of the
cutoff wall may be induced in the
subsurface at locations where the
hydraulic driving pressure exceeds
the effective vertical overburden
pressure. Although the applied sur-
charge pressure at the top of the
wall in these cases was higher than
the hydraulic pressure, it was not
effective over the full depth of the
wall resulting in general hydrofrac-
ture (presumably near the base of
the wall).
• Density, water content and vane
shear strength data measured on
samples from a cutoff wall after
-------
I
testing all confirm the dissipation of
vertical overburden pressure with
increasing depth in the model.
• The success of the window closing
test suggests that the effective over-
burden pressure in the wall may
serve to close residual slurry win-
dows and may even close a multitude
of micro shrinkage cracks that may
develop in the backfill over the life of
the barrier due to the effects of
chemical leachates.
The full report was submitted in ful-
fillment of contract number 68-03-3210,
07 by the University of Cincinnati under
sponsorship of the U.S. Environmental
Protection Agency.
10
30 60 90 120
Time [Days]
150
180
V Initial K
Q Equilibrium K
1 Breakthrough
• Ruptured Surcharge Bladder
Q Projected Equilibrium K
Figure 6. Chronology and results of hydraulic conductivity tests.
-------
Depth
(in.)
—
1
/
/
/
4 •
IK
?n
Dry Unit Weight
fib/ft3)
114 116 118
/
-K.
^'
'*•'•'
4-
X
X
'**•
*•
f-
/
/
L7*
X
*'
•
Wafer Content (%l
14 16 18 20
~\\
V
\
\
y
*•
•
Figure 7. Results of tests on soil-bentonite backfill after completion of hydraulic conductivity
tests.
Richard M. McCandless and Andrew Bodocsi are with the University of
Cincinnati. Cincinnati, OH 45221.
Naomi P. Barkley is the EPA Project Officer (see below).
The complete report, entitled "Investigation of Slurry Cutoff Wall Design and
Construction Methods for Containing Hazardous Wastes," (Order No. PB 87-
229 688/AS; Cost: $24.95, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati. OH 45268
-------
•C W .U 1.
62SOIC9 t
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PA
EPA
PERMIT No G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-87/063
------- |