decreases as cement content increases. Compared with achieving the required compressive strength, adequately reducing
the hydraulic conductivity of the gravel soil was more difficult.
Therefore, reducing the hydraulic conductivity controlled the
design. The addition of 1 percent bentonite decreased the hydraulic conductivity of the gravel stratum by about an order of
magnitude. However, adding more than 1 percent bentonite
brought only negligible further decreases in hydraulic conductivity. Finally, the addition of cement kiln dust resulted in
only a small decrease in hydraulic conductivity. Reduction was
greatest when portland cement content was low.
The treatability study indicated that the two mix designs,
one involving 8 percent portland cement and 1 percent bentonite and the other 11 percent portland cement and no bentonite, would meet the stipulated performance criteria. In
fact, both mix designs were predicted to provide hydraulic
conductivity on the order of 5 × 10–7 cm/s at 28 days, providing a reasonable safety factor.
During final design, closely spaced test borings were made
within the area to be treated by means of ISS to establish the
elevation of the top of the clay barrier layer. These data were
used to define the required depth of ISS treatment at any
location. The ISS area was subdivided by means of a 25 by
25 ft grid, and the required treatment depth was defined by
the deepest clay elevation within each grid box. ISS treatment
generally extended 2 ft into the clay, except along the perimeter of the ISS treatment area, where the toe was increased to
4 ft. The increased depth was intended to afford greater certainty that the ISS treatment would fully extend through the
affected soils along the treatment area boundary, providing
a continuous cutoff. In the interior of the treatment area, a
shorter depth into the clay was considered appropriate because any minor untreated zones above the clay surface would
be isolated and completely confined.
Because the design borings were closely spaced and provided reliable data regarding depth to the clay barrier layer,
no confirmation borings were required during construction.
The specifications also required that the perimeter of the
treatment area be completed first to confine the area contaminated by the former MGP and minimize the risk of NAPL
migrating beyond the treatment area boundary.
The design specified that the ISS grout mix contain 8 percent portland cement and 1 percent bentonite and be placed
by auger in the form of columns. Although bench studies had
indicated that an alternative mix with 11 percent portland
cement and no bentonite also would meet the design performance requirements, this alternative was not considered because of concerns that the resulting high-strength grout mix
would impede efforts to form overlapping columns. However, the contractor was allowed the option of developing alternative grout mix designs. The initial ratio of water to reagent
was specified to be 1: 1 by weight, but higher ratios of water
to solids could be used if necessary. Auger mixing was specified, but bucket mixing was not permitted.
The design required that the entire site to be treated by
ISS first be excavated to a depth of 4 ft, corresponding to local
frost depth. This excavated area would subsequently be backfilled with clean soil to protect the top of the ISS monolith
 Civil Engineering SEPTEMBER 2012
from degradation by freeze-thaw action. The excavation also
served to contain spoils resulting from the grouting process.
Remedial ISS was carried out by WRScompass, of Tampa, Florida, between December 2010 and April 2011. ISS
was performed using a 10 ft diameter mixing auger powered
by a crane-mounted drill platform having 400,000 ft-lb of
torque. Type I portland cement was delivered to the site in
pneumatic trucks and off-loaded into storage bins. Bentonite was delivered in 50 lb bags. Grout was prepared using a
high-shear mixer. The batch tank had a working volume of
approximately 5 cu yd.
The required reagent quantities were calculated for each column, taking into consideration its required depth and overlap
with adjacent columns. The predetermined grout volume was
pumped to injection ports located along the arms of the mixing
auger and blended with the affected soils. To ensure full coverage with no gaps, the 10 ft diameter columns were positioned
in a triangular pattern with nominal spacing of 8. 67 ft. Three
vertical mixing passes typically were used, resulting in a homogeneous mixture of the affected granular soils.
Initial trials indicated that the 8 percent portland cement
plus 1 percent bentonite grout mix was not pumpable when
mixed at the initial 1: 1 ratio of water to reagent. To create a
pumpable mix, the ratio of water to solids was increased to
1.85: 1. However, the additional water needed to make the
grout pumpable increased the volume of grout material injected into the soil, increasing the quantity of spoils. The contractor constructed several test columns to evaluate the feasibility of
reducing the quantity of bentonite and water in the grout mix.
The test columns indicated that the quantity of bentonite could
be reduced to 0.75 percent, and possibly to 0.5 percent, without
significantly increasing hydraulic conductivity. Based on these
tests, the grout mix was modified to 8 percent portland cement
plus 0.75 percent bentonite, mixed at a ratio of water to reagent
of 1.23: 1. The modification reduced the required quantity of
reagents somewhat, but of greater importance, it reduced the
required volume of water and resulting ISS spoils.
Approximately 52,100 cu yd of sand and gravel were solidified in situ using the auger mixing technology. Spoils
amounted to approximately 21,180 tons, or approximately
25 percent, assuming a conversion of 1. 6 tons per cubic yard.
The volume of spoils exceeded initial estimates, resulting in
cost overruns, as excess spoils required off-site disposal.
Treated material was sampled using a sampling tool with
a hydraulically operated lid. The lid is opened after the sampler is lowered to the desired sampling depth, allowing the
treated soil to flow into the sampling device. Representative
samples were typically obtained at two or three depths within
an ISS column, and these samples were subsequently blended
to create a single composite sample for the column. Test cylinders were created from the composite sample after screening out larger particles. The test cylinders were stored and
cured in a water bath in a heated room. At the end of the cure
period, the samples were delivered by courier to the test laboratory. Laboratory testing was performed by Atlantic Testing
Laboratories, Ltd., of Canton, New York.
A total of 113 quality control samples were obtained
by AECOM. Unconfined compression strength testing was