-------
Speaker's Brief Biographical Information
Conference:
USEPA Region IV Technical Conference on Groundwater
Remediation/Stabilization, Atlanta, December 1-3, 1993
Presentation
Title:
Authors:
Speaker:
Case Study:
Groundwater Remediation System
Occidental Chemical Corporation
Mobile, Alabama
Linda M. McConnell, PE, G&E Engineering, Inc.
Richard B. Adams, PE, DEE, CGWP, G&E Engineering, Inc.
Ed Seitz, Occidental Chemical Corporation
Linda M. McConnell
Linda M. McConnell is Vice President of G&E Engineering, Inc. A
registered professional engineer, Ms. McConnell has over 18 years
engineering and environmental experience. Her primary areas of
responsibility include RCRA permitting, investigations, and corrective
actions. She has also prepared several risk assessments and applica-

-------

-------
Case Study
Olin Corporation
Mcintosh, Alabama
Effectiveness of the RCRA Groundwater
Corrective Action Program
October 1993
1 Introduction
This report presents an analysis of the RCRA Corrective Action Program (CAP) at Olin's
Mcintosh, Alabama, facility. The CAP is a groundwater remediation program using pump-
and-treat methods to reduce mercury and organics in groundwater to the groundwater
protection standards required by a RCRA post-closure operating permit. This report will focus
on the mercury contamination in groundwater and will discuss the progress in remediation in
terms of mercury. The organic contaminant remediation is progressing in parallel to the
mercury.
2 Description of the Facility
The Olin Corporation Mcintosh plant is located approximately one mile east-southeast of the
town of Mcintosh, in Washington County, Alabama (Figure 1). The property is bounded on
the east by the Tombigbee River, on the west by U. S. Highway 43, on the north by the Ciba-
Geigy Corporation plant site, and on the south by River Road. The Olin Mcintosh plant is an
active chemical production facility located on approximately 1,500 acres, with production
areas occupying approximately 60 acres. The Mcintosh plant today produces chlorine, caustic
soda, sodium hypochlorite and sodium chloride and blends and stores hydrazine compounds.
Current active facilities at the plant include: a diaphragm cell chlorine and caustic production
process area; a caustic concentration process area; a caustic plant salt process area; a hydrazine
blending process area; shipping and transport facilities; process water storage, transport and
treatment facilities; and support and office areas.
The production area is relatively flat, about 40 to 50 feet above mean sea level (msl). The
most distinctive topographic feature is a steep bluff located approximately 4,000 feet east of
the production area. This bluff defines the edge of the low-lying floodplain area, which is
about 25 feet lower in elevation than the upland areas immediately to the west.
2.1 Area Land Use
Residential land use (4 percent of area within a 3-mile radius of the site) includes individual
dwellings and groups of two to about twenty dwellings. The commercial activity (less than 1
percent of the area) is generally related to basic domestic needs and services located along
Highway 43. Industrial use occupies 4 percent of area within the 3-mile radius. The two
main industries are the Olin and Ciba-Geigy facilities. A compressed air power plant
(Alabama Electric) and a cement company are also located within the 3-mile radius. Public

-------
Olin Corporation - Mcintosh, Alabama
Page 2
and cemeteries. Forested uplands make up the largest land use area (65 percent). Floodplains
and streams, including the Tombigbee River, the basin and Bilbo Creek, make up about 27
percent of the three-mile-radius area.
2.2 Site History
Olin operated a mercury cell chlor-alkali plant (constructed in 1951) on a portion of the site
from 1952 through December 1982. In 1952, Calabama Chemical Company began operation
of a chlorinated organics plant on property immediately south of Olin. In 1954, Olin acquired
Calabama and in 1955 began construction of a pentachloronitrobenzene (PCNB) plant on the
acquired property. The plant was completed and PCNB production was started in 1956. The
Mcintosh plant was expanded in 1973 to produce trichloroacetonitrile (TCAN) and 5-ethoxy-3-
trichloromethyl-l,2,4-thiadiazole (Terrazole). The PCNB, TCAN and Terrazole
manufacturing areas were collectively referred to as the crop protection chemicals (CPC)
plant. In 1978, Olin began operation of a diaphragm cell caustic soda/chlorine plant, which is
still in operation.
The CPC plant and mercury cell chlor-alkali plant were shut down in late 1982. The CPC
plant was decommissioned and dismantled and the site was capped under a plan submitted to
and approved by ADEM. .The chlorine plant was decommissioned and dismantled in several
phases from 1982 until 1986. Figure 2 presents the location of these areas as well as all Solid
Waste Management Units (SWMUs) at the site.
In March 1982, Soil and Materials Engineers, Inc. (S&ME) performed a hydrogeological
investigation of the Mcintosh site to assess the migration and extent of organic contaminants in
the groundwater (S&ME, 1982). The investigation included the installation of 32 monitoring
wells and groundwater sampling of both new and the existing 43 wells. The field investigation
was completed in August 1982, with the final report submitted to ADEM and EPA in
November 1982. The report established the direction of groundwater flow and defined the
hydrogeologicalj)arameters of the area. The study also identified two plumes of chlorinated
organic contaminants (predominantly chloroform, benzene, chlorobenzene and
dichlorobenzene) in the Alluvial Aquifer, one moving east-southeast, the other west-southwest.
The two plumes' movement in different directions was the result of a groundwater divide that
existed before implementation of the RCRA corrective action program.
The hydrogeological data indicated the Alluvial Aquifer is separated from the deeper Miocene
Aquifer by a low-permeability clay aquitard. The report further indicated that this aquitard
inhibits downward migration of the contaminant plumes. To further define the migration of
the plumes identified by S&ME, Olin Corporation installed 14 additional monitoring wells
between February and March of 1983.
During the period from 1982 to 1986, Olin closed the RCRA-regulated units at the Mcintosh
plant. Several of the units are regulated under a RCRA post-closure operating permit. Since
1984, Olin has conducted its RCRA groundwater detection monitoring program. Ten
additional monitoring wells have been installed on the eastern perimeter of the plant to further
define the migration of contaminants to the east/southeast. The location of all wells is shown
on Figure 3.
In July 1987, construction was completed on the groundwater corrective action program (CAP)
required by the post-closure openi :ing permit. The five-well system became operational in

-------
Olin Corporation - Mcintosh, Alabama
Page 3
Each well has a fully penetrating, 10-inch-diameter, Schedule 40 PVC screen and a 10-inch-
diameter, Schedule 40 PVC casing. The wells were drilled to the top of the Miocene clay,
approximately 90 feet deep. The five pumping wells operate 24 hours per day (with minor
outages for maintenance and emergencies). The pump discharges are equipped with sampling
points, instantaneous flow measurements and control, flow totalization, and pump discharge
pressure instrumentation. Each pumping well is equipped with its own treatment system.
The RCRA post-closure permit requires groundwater monitoring for clpsed RCRA units,
including the weak brine pond, the stormwater pond and the brine filter backwash pond. The
post-closure permit also requires corrective action for releases of hazardous waste constituents
from any solid waste management units (SWMUs) at the facility. There are no active RCRA
units at the facility.
3 Regional Geology
Washington County is located in the Southern Pine Hills District of the East Gulf Coastal Plain
Province. The Mcintosh area is underlain by alternating beds of unconsolidated-to-
consolidated sedimentary rocks that are collectively hundreds of feet thick (Turner and
Newton, 1971). These rocks dip southwesterly at 30 to 50 feet per mile. The general dip of
these rocks is locally interrupted by folds, faults and salt domes. The Mcintosh salt dome is
the most distinctive structural feature of the area. The Olin site lies within the Mobile graben,
a complex north-south oriented fault system that extends in a north-south direction from west-
central Clarke County to east-central Mobile County.
The near-surface sediments of the Mcintosh area are Recent alluvium and Quaternary alluvial
terrace deposits that are collectively as much as 100 feet thick in places (Turner and Newton,
1971). As much as 30 to 40 feet of these strata are exposed in various drainage ditches in the
area and along the west bank of the Tombigbee River east of the site. Based on information
from water wells and tests holes at the Olin site, the thickness of these sediments ranges from
about 80 to 100 feet. Quaternary sediments were deposited unconformably over the
underlying Miocene sediments. These alluvial sediments consist of beds of sand, gravel, silt,
and clay, along with various combinations of these materials. Permeable units of sand and
gravel that exist within this section are aquifers. One of these sand units is located beneath the
plant site, directly overlying the Miocene strata, and varies in thickness from 55 to 80 feet.
This sand unit is referred to as the Alluvial Aquifer (Q2). This Quaternary sand unit is
overlain by 10 to 60 feet of silt and clay. The quality of natural formation water in the
Alluvial Aquifer is generally suitable for domestic and industrial uses. However, iron
concentrations are typically in excess of 0.3 mg/1.
The Miocene series is composed of alluvial sediments that were deposited in complex,
nonmarine environments. West of Mcintosh, Miocene strata outcrop in a large part of
Washington County and are exposed in road cuts and along stream valleys. In the Mcintosh
area, Miocene strata are covered by the younger Quaternary alluvial sediments. The Miocene
series ranges in thickness from less than 275 feet over the Mcintosh salt dome to as much as
600 feet away from the dome. This geologic unit is composed of beds of fine-to-coarse-
grained sand, gravelly sand, sandstone, and beds of light gray and varicolored clay.
The sands and gravel in the Miocene series are the most important groundwater source in the
Mcintosh area. The Miocene series beneath the site is divided into four units. The lower

-------
Olin Corporation - Mcintosh, Alabama
Page 4
site. The Tm4 unit is overlain by the lower Miocene confining unit (Tm3), which is believed
to be laterally continuous and sufficiently impermeable to retard upward vertical movement of
groundwater from the lower Miocene Aquifer. The upper Miocene Aquifer (Tm2) is a highly
permeable sand and is referred to as the Miocene Aquifer in this report. The Miocene Aquifer
is overlain by the upper Miocene confining unit (Tml).
The Miocene Aquifer receives considerable recharge in the large outcrop areas located west
and northwest of Mcintosh. In these outcrop areas, groundwater in the aquifer occurs under
unconfined conditions. Northwest of the town of Mcintosh, the Miocene Aquifer probably
receives considerable recharge by downward vertical leakage from the overlying Alluvial
Aquifer where these aquifers are separated by thin and relatively permeable silt and clay beds.
However, as groundwater in the Miocene Aquifer moves downgradient, it becomes confined
by thicker clay, silt and silty clay beds. The two aquifers are separated by an estimated 80 to
100 feet (the upper Miocene confining unit) at the Olin Mcintosh facility.
Geologic formations of Oligocene and Eocene age underlie the Miocene formations in
Washington County and are sources of groundwater in northern Washington County. These
formations are not sources of potable groundwater in the Mcintosh area. Highly mineralized
groundwater under considerable confining pressure occurs in some aquifers within these
formations. It is postulated that near some fault zones, this highly mineralized groundwater
may have risen upward along and through the fault zones into overlying freshwater aquifers
(Barksdale, 1929, Newton and others, 1972).
4 Site Stratigraphy
Near-surface strata consist of Quaternary alluvial terrace and floodplain sediments deposited by
the Tombigbee River. The Quaternary sediments range in thickness from 80 to 100 feet and
consist of beds of sand, gravel, silt and clay, which form the Alluvial Aquifer system. The
Alluvial Aquifer is underlain by Miocene sediments. The Miocene series is composed of
alluvial deposits.of fine-to-coarse-grained gravel, sand and sandstone and beds of gray-to-
varicolored clay. The Miocene series varies in thickness from less than 275 feet above the
Mcintosh dome to as much as 600 feet away from the dome. The Quaternary alluvial
sediments are divided into two units, designated Q1 and Q2. The Tertiary (Miocene) units
addressed in this study are designated Tml and Tm2.
The Olin site is located within the outcrop area of the upper clay unit (Ql). The lithology of
Q1 is variable, but is composed primarily of red-brown, yellow-brown, and gray, silty/sandy
plastic clay; the silt and sand content varies and generally increases with depth. Thin,
probably discontinuous sand and silt lenses occur interbedded with the clay. The thickness of
Ql varies from less than 10 feet to about 60 feet, as illustrated on the cross sections.
The Alluvial Aquifer (Q2) in the production area varies in thickness from an average of about
55 feet to 80 feet, thinning in the west plant area to approximately 37 feet at the location of
monitor well DH3. The Alluvial Aquifer is divided into two zones. The upper zone of the
Alluvial Aquifer is composed primarily of very fine to fine-grained, silty quartzose,
subangular-to-subrounded sand. The lower zone of the aquifer is composed of fine-to-very-
coarse, orange-brown, quartzose, cherty, subangular to subrounded sands containing varying
amounts of fine-to-large gravel. Although composed predominantly of sands, Q2 also contains
some thin beds of clay or silty, gravelly cl '. One of these beds, a gravelly, silty plastic clay

-------
Olin Corporation - Mcintosh, Alabama
Page 5
from approximately 3 feet to 10 feet thick beneath the southern part of the production area.
However, this fine-grained unit apparently pinches out or becomes sandier in other parts of the
plant site.
In water well and stratigraphic test hole logs, the Miocene confining unit (Tml) is described as
consisting of clays, sandy clays, or clayey sands. Although the lithology may be complex, it
is dominantly clay, with various amounts of discontinuous sand, silt, or sometimes fine gravel.
Monitor well DH2 is screened in a sand unit that was initially interpreted to be part of the
Miocene Aquifer (Tm2). S&ME (1982) later concluded that this well is screened in a
discontinuous sand contained within the upper Miocene confining unit. Boring logs from wells
that penetrate the upper Miocene confining unit indicate that this unit is laterally continuous
beneath the site and approximately 80 to 100 feet thick. Figure 4 is a structure contour map
that illustrates the surface configuration of the top of the Miocene confining unit. The upper
clay consists of blue-gray, sometimes mottled, silty, hard plastic clay with minor amounts of
sand. The vertical permeability of the this upper clay is extremely low, with vertical hydraulic
conductivities (K) of less than 1 x 10-5 feet per day (less than 1 x 10-8 cm/sec) (S&ME,
1982).
5 Site Hydrogeology
The uppermost aquifer at the site is the Alluvial Aquifer described above. It is separated from
the lower Miocene Aquifer by the upper Miocene confining unit. discussed above. This
section describes the hydrogeology of both aquifers, although the RCRA Corrective Action
Program (CAP) addressed remediation of only the Alluvial Aquifer.
5.1 Alluvial Aquifer
The Alluvial Aquifer is composed primarily of sands and varies in thickness from about 55 to
80 feet in the plant area, thinning to less than 40 feet at locations in the west plant area. The
Alluvial Aquifer is generally unconfined throughout the area. The specific yield is estimated
to be 0.20, based on grain size analysis.
Figure 5 depicts the typical potentiometric surface of the of the Alluvial Aquifer, based on
water elevations collected during 1986 and 1987 prior to implementation of the corrective
action program. Figure 5 indicates that groundwater entered the site from the north (there is a
localized recharge from a hydraulic mound near monitoring well PL4D). Recharge was from
direct infiltration where the Alluvial Aquifer outcrops to the north of the Olin facility (S&ME,
1982). A groundwater divide was present near the production area separating the southerly
flow into southeast and southwest components. The occurrence of this groundwater divide
appears to be related to the structure of the top of the upper Miocene clay. Structural lows
occur both to the southwest and to the southeast. Groundwater entering the site west of the
structural high preferentially flowed to the southwest and groundwater entering the site east of
the structural high preferentially flowed to the southeast. The southwest component was
influenced by a hydraulic mound located in the western portion of the site and caused by ponds
resulting from beavers' dam building in low areas. Groundwater elevations near the center of
this mound were approximately 20 feet higher than elevations in wells located approximately
1,000 feet to the east. The hydraulic mound diverted the westerly flow to the south.

-------
Olin Corporation - Mcintosh, Alabama
Page 6
River, where discharge occurred. Near the center of the facility, the groundwater divide was
evident by westward flow in the former mercury cell plant, CPC plant, and strong brine pond
areas and eastward flow east of the old plant (CPC) landfill. Although the exact east-west
location of the groundwater divide has varied with local infiltration rates and seasonal
fluctuations, its occurrence is well documented by the hydraulic head measurements from 1980
to 1986. The divide was first described and related to the structure of the Miocene clay during
the S&ME investigation (S&ME, 1982).
Five corrective action wells currently in operation at the site recover groundwater from the
Alluvial Aquifer. Extraction of groundwater from these wells has caused localized depressions
in the potentiometric surface beneath the site. Figure 6 presents a typical, current
potentiometric map for the site.
The average hydraulic conductivity of the Alluvial Aquifer has been estimated based on single
well tests as 15 ft/day. The hydraulic conductivity value from the pump test was estimated as
578 fit/day, which is considered an upper range value since the area of the test is known to be
highly transmissive. Since pump tests generally yield more reliable estimates of hydraulic
conductivity, the average of 15 ft/day and 578 ft/day, which is 296 ft/day, was used as the
estimate for the site hydraulic conductivity. The effective porosity was estimated to be 25
percent.
Based on the above hydraulic conductivity and gradients based on the potentiometric maps, the
groundwater velocity at the site varies from about 0.16 ft/day to about 12 ft/day. These values
provide a general estimate of groundwater velocity in the Alluvial Aqyifer. The groundwater
flow velocities will also vary with the heterogeneity of the aquifer.properties and localized
variations in the lateral gradients.
5.2 Miocene Aquifer
The upper Miocene Aquifer (Tm2) contains two main artesian sands that are separated by a
clayey unit ranging from 10 to 20 feet thick. The sands are considered as one hydrogeologic
unit due to a natural hydraulic connection and connection by gravel-packed wells. Tlie
combined transmissivity of the two sands is considered to be in excess of about 25,000 square
feet per day (ft2/day) (S&ME, 1982).
The regional gradient of the Miocene Aquifer is to the east-southeast (DeJarnette, 1989).
However, Olin continuously pumps two Miocene Aquifer process water wells. The
transmissivity of the Miocene Aquifer zone was estimated to be 187,000 gallons per day per
foot (gpd/ft) or 25,000 ft2/day. The storativity was estimated to be 0.001.
In addition to the effects of Olin's pumping wells, the Miocene potentiometric surface will also
be affected by the regional groundwater gradient and pumping wells from other areas such as
the Ciba Geigy facility to the north.
6 Solid Waste Management Units
Figure 2 presents a site map indicating the location of all Solid Waste Management Units

-------
Olin Corporation - Mcintosh, Alabama
Page 7
SWMUs, both closed and active. Several of the former RCRA units and SWMUs are believed
to have been significant sources of groundwater contamination at the site. These are indicated
in this section.
6.1 SWMUs Closed Or Clean-closed Under 40 CFR 265
Ten SWMUs have been closed (1 unit) or clean-closed (9 units) under 40 CFR 265. Under
current regulations (40 CFR 270.1(c)), surface impoundments, landfills, treatment units, and
waste piles that were clean-closed under 40 CFR 265 are subject to the clean closure
equivalency standards. At the Olin Mcintosh facility, the units subject to equivalency
demonstrations include the three clean-closed surface impoundments (the stormwater pond, the
brine filter backwash pond, and the pollution abatement (pH) pond) and the one clean-closed
waste pile (the mercury waste pile storage pad). The clean closure equivalency demonstration
document was submitted to EPA October 1, 1993.
6.1.1 Stormwater Pond
The stormwater pond is a clay-lined earthen structure approximately 140 x 365 feet, which
contained a maximum volume of 500 cubic yards of settled solids. The pond received
stormwater runoff from the mercury cell chlor-alkali processing area perimeter. Since
stormwater pond solids were contaminated with mercury, the pond was designated a hazardous
waste unit for D009 wastes. The pond was originally constructed with approximately 1.5 feet
of natural clay of low permeability as a liner. The clay liner was compacted to 95 percent
Proctor density.
This unit was clean closed and is not being used. It has been concluded that this unit has never
been a significant source of groundwater contamination.
6.1.2	Brine Filter Backwash Pond
The brine filter backwash pond was an earthen structure approximately 160 x 240 feet, lined
with a geosynthetic membrane. The pond received wastewaters that included washdown, filter
backwash and process water. The maximum inventory of settled solids was estimated to be
approximately 600 cubic yards.
This unit was clean closed and is used as needed as a nonhazardous surface impoundment. It
has been concluded that this unit has never been a significant source of groundwater
contamination.
6.1.3	Pollution Abatement (pH) Pond
The pH pond was a wastewater impoundment approximately 140 x 290 feet. The bottom of
the pond was constructed of a backfilled, low-permeability clay approximately 2.0 feet thick.
This pond was designed and constructed to handle corrosive wastewaters having a pH less than
2.0 or higher than 12.5. Wastewaters received for treatment included washdown from process
areas, cooling tower blowdown, various process streams, and contaminated stormwater runoff.

-------
Olin Corporation - Mcintosh, Alabama
Page 8
The unit was clean-closed, is now lined with a synthetic membrane, and is used as a
nonhazardous wastewater holding pond. This unit was a significant source of low pH
contamination to the groundwater and probably caused increased mercury and dissolved solids
concentrations in groundwater beneath it.
6.1.4	Weak Brine Pond
The weak brine pond was an earthen structure approximately 340 x 340 feet that was primarily
a process brine unit, but also received several D009 and K071 waste streams from operation of
the mercury cell chlor-alkali plant, including filter backwash that contained mercury.
The weak brine pond was also utilized during the closure of the three on-site surface
impoundment units discussed above (i.e., the stormwater pond, the pH pond and the brine
filter backwash pond). The material removed from each of these impoundments was deposited
in the weak brine pond, stabilized, and solidified with cement dust. The total volume of
consolidated waste in this unit is approximately 33,000 cubic yards.
Closure for this unit consisted of liquids removal, on-site treatment through an existing
activated carbon system and discharge through Outfall 001 in accordance with the NPDES
requirements. After removal and treatment of the liquids, chemical stabilization/solidification
materials were added to the pond bottom sludges, along with dry soil pushed in from the
dikes, to form a stable foundation for the final cover and cap material.
After the pond contents were stabilized, the former pond was capped. A composite support
base of dike soil, native soil and clay was constructed using a 12-inch base of native soil and
dike material followed by a 6-inch cover of compacted clay material having a reported
maximum permeability of 1 x 10-7 cm/sec. The clay was also compacted to a minimum 95
percent Proctor density, with the compaction and permeability being verified in the field by an
independent soils testing laboratory. A 30-mil synthetic liner was then placed over the support
base to provide an impervious membrane over the underlying material. A 12-inch sand
drainage layer was placed over the synthetic liner to function as a drainage system to remove
any water that resulted from rainfall and percolation through the overlying layers. The sand
drainage layer was then covered with a geotextile fabric net and capped with a 6-inch layer of
compacted clay having a maximum permeability of 1 x 10-7 cm/sec. A final cover of 6 inches
of topsoil was installed, with seed (30 pounds/acre) and fertilizer being applied to establish
rapid cover and prevent erosion of the cap system. The top cover was sloped at a five percent
gradient. The cap was encircled by concrete-lined perimeter ditches sloped to drain into the
plant wastewater discharge.
This unit is considered to have been the source of most mercury and dissolved solids
contamination in the groundwater at the site.
6.1.5	Mercury Waste Pile Storage Pad
The mercury waste pile storage pad was a concrete pad, approximately 40 x 60 feet, located to
the south of the former mercury cell plant area. The pad was classified as a RCRA hazardous
waste unit for D009 wastes from 1980 to 1984. Half of the unit was used for drum storage
while the other half was used for storage of material contaminated with mercury from the
mercury cell chlor-alkali process operations (e.g., piping, etc.). .Most of the material was

-------
Olin Corporation - Mcintosh, Alabama
Page 9
process was discontinued. The remaining wastes were removed during clean closure in
November 1984. It has been concluded that this unit has never been a significant source of
groundwater contamination.
6.1.6 TCAN Hydrolyzer
The TCAN (trichloroacetonitrile) hydrolyzer was a glass-lined agitation tank designed to
hydrolyze the reactive residue from the TCAN distillation column. The residue was loaded in
the hydrolyzer after being treated with a 20 percent sodium hydroxide solution for
decomposition of the reactive component (TCAN). The nonreactive waste was then drained
into 55-gallon drums for off-site disposal.
All wastes were removed from the TCAN hydrolyzer tank in February 1982 and the unit was
placed in standby status. The unit was then closed in December 1982 by rinsing with a sodium
hydroxide solution followed by a high-pressure, hot water rinse. All waste residues were
collected and solidified for disposal in an off-site hazardous waste landfill. It has been
concluded that this unit has never been a significant source of groundwater contamination.
6.1.7 Mercury Drum Storage Pad
The mercury drum storage pad was used to store drums that contained D009 wastes including
filters, sump sludges and other process waste containing mercury. The unit consisted of a 40 x
60-foot concrete storage pad used for drum storage of mercury-contaminated solids. The unit
was clean closed by removing all the waste, followed by decontamination. It has been
concluded that this unit has never been a significant source of groundwater contamination.
6.1.8 Chromium Drum Storage Pad
The chromium drum storage unit is a 30 x 50-foot concrete pad that was used for storage of
containerized, chromium-contaminated solids and liquids. Maximum inventory of the unit
reached approximately 6,600 gallons. The unit was closed in January 1986. Materials were
removed and transported to an off-site hazardous waste treatment or disposal facility. It has
been concluded that this unit has never been a significant source of groundwater
contamination.
6.1.9 PCB/Hexachlorobenzene Storage Building
The PCB/hexachlorobenzene building was a 60 x 120-foot steel frame building with ribbed
siding. The building was used mainly for the storage of hexachlorobenzene (K085), a waste
material generated in the manufacture of PCNB. Closure consisted of removal of materials
and transportation to an off-site hazardous waste treatment/disposal facility. It has been

-------
Olin Corporation - Mcintosh, Alabama
Page 10
6.1.10 Hazardous Waste Drum (Flammable) Storage Fad
The flammable waste drum storage facility is a 30 x 40-foot steel frame building with a
concrete floor and partial siding. The building was used primarily for the storage of ignitable
(D001) wastes. This unit was clean-closed in January 1986 by removing all waste from the
site and transporting it to an off-site hazardous waste treatment and disposal facility. It has
been concluded that this unit has never been a significant source of groundwater
contamination.
6.2 SWMUs Not Regulated Under 40 CFR 265
6.2.1	Sanitary Landfills
Between 1977 and 1984, Olin operated two sanitary landfills that received general
nonhazardous waste and plant refuse (Figure 2). Records show that waste disposed of in the
two sanitary landfills included paper, glass, boxes, wood, plastic, grass clippings, pipe,
concrete, and sanitary sludge. The first unit was constructed in 1976 and opera tad from 1977
to 1978. It was approximately 150 x 200 feet and contained about 4,000 cubic yards of
material. The second unit was operated from 1978 to 1984. The second unit was
approximately 600 x 800 feet and contained about 18,000 cubic yards of material. It has been
concluded that this unit has never been a significant source of groundwater contamination.
6.2.2	Old Plant (CPC) Landfill
The site of the old plant (CPC) landfill (Figure 2) was utilized from 1954 until 1972 to
neutralize acidic wastewater from CPC plant operations. Neutralization was conducted by
flowing the wastewater over piles of oyster and clam shells. The flow was then directed by an
overflow ditch to the production wastewater ditch. Plant personnel indicate that the former
landfill also received organic wastes from the CPC plant. From 1972 to 1977 the site was
used for disposal of general plant debris such as paper, cardboard, wood, small metal
containers, scrap plastic and rubber items from the entire plant. The landfill area is
approximately 300 x 400 feet and is estimated to have an 8,000-cubic-yard capacity. In 1977,
prior to RCRA, the landfill was closed with a clay cap, topsoil, and grass, as approved by the
AD EM. The cap was upgraded in 1984 to address erosion problems that had occurred.
During the upgrade, a 2-foot-thick layer of compacted clay was placed; a 3- to 6-inch layer of
topsoil was placed over the clay cap and the area was vegetated.
The old plant (CPC) landfill is considered to have been the source of most of the organic
contamination in groundwater.
6.2.3	Diaphragm Cell Brine Pond and Overflow Basin
The diaphragm cell brine pond is approximately 384 x 205 feet, with a 38,000-cubic-yard
capacity. This pond is used for weak brine solutions prior to recycling the brine to brine
wells. The brine cell pond and overflow basin were built in 1976. The overflow basin is
smaller, approximately 384 x 144 feet, with a capacity of 26,600 cubic yards. It has been

-------
Olin Corporation - Mcintosh, Alabama
Page 11
6.2.4 Ash Pond (Active)
The active ash pond was built in 1981 and is currently used for nonhazardous boiler ash. It is
approximately 480 x 870 feet, with a 232,000-cubic-yard capacity. It has been concluded that
this unit has never been a significant source of groundwater contamination.
6.2.5	Ash Ponds (Inactive)
There are two inactive ash ponds located on the facility. The old ash pond was used as a
settling pond for nonhazardous, coal-fired boiler ash. It was built in 1976 and is
approximately 200 x 300 feet, with a 31,000-cubic-yard capacity. It is now used as a standby
unit. The day ash pond was used for dewatering nonhazardous boiler ash. The pond is
300 feet in diameter, with a 50,000-cubic-yard capacity. The day ash pond was built in 1979.
It has been concluded that this unit has never been a significant source of groundwater
contamination.
6.2.6	Lime Ponds
There are two former lime ponds, the east and west ponds, which were not regulated under 40
CFR 264 or 40 CFR 265 (Figure 2). The ponds were approximately the same size, but the
west pond contained approximately 5,300 tons of lime waste and the east pond approximately
4,200 tons. The ponds contain lime (from the absorption and capture of residual chlorine gas)
and lime sludges. These two ponds operated from 1968 to 1976 and were closed in 1979
(prior to RCRA) with a clay cap, topsoil and grass. It has been concluded that this unit has
never been a significant source of groundwater contamination. However, mercury has been
detected in the groundwater in the vicinity of these lime ponds. The weak brine pond, in
which mercury-containing brine was handled, is the suspected source of mercury to the
groundwater in the area. Based on the pre-corrective action potentiometric surface, the wells
situated around the lime ponds were located hydraulically downgradient of the weak brine
pond.
6.2.7	Hexachlorobenzene Spoil Area
The hexachlorobenzene spoil area site apparently was used in the past (date unknown) to
dispose of soils from earth work in the former PCNB production facility. The area was
discovered on October 11, 1990 while grading adjacent to the day ash pond. On October 26,
1990, the EPA and Olin reached an agreement through an administrative order of consent
(AOC) for removal of the hexachlorobenzene-contaminated soil. The AOC required that any
soils within the site that had hexachlorobenzene concentrations higher than 200 mg/kg be
removed and transported off-site for disposal at an approved hazardous waste facility in
compliance with EPA's off-site policy. A total of 11,407 tons of soil were excavated,
transported, and disposed of from October 27, 1990 to November 6, 1990. The excavated soil
was sent to Chemical Waste Management's hazardous waste-permitted landfill in Carlyss,
Louisiana. It has been concluded that this unit has never been a significant source of

-------
Olin Corporation - Mcintosh, Alabama
Page 12
6.3 Additional SWMUs/AOCs Listed in the RFA
Olin was issued a federal EPA RCRA post-closure permit on July 7, 1986. On September 1,
1986, ADEM issued a state hazardous waste permit to Olin. EPA retained authority for the
1984 Hazardous and Solid Waste Amendments (HSWA) portions of the permit, which
included the HSWA Corrective Action Program. The first phase of the HSWA Corrective
Action Program is a RCRA Facility Assessment (RFA). The RFA at the Mcintosh facility
was conducted by an EPA contractor and consisted of a preliminary review (PR) of files from
EPA Region IV and ADEM and a visual site inspection (VSI) on June 17 and 18, 1991. A
draft RFA report was provided to Olin on October 30, 1991. Olin made comments, and a
final RFA report was provided to Olin on February 4, 1992. The RFA lists 52 SWMUs and
six areas of concern (AOCs). Additional sampling was recommended at the Solid Waste
Management Units (SWMUs) discussed below (other SWMUs had recommendations other
than sampling). The soil sampling was conducted as part of the CERCLA Remedial
Investigation/Feasibility Study conducted at the site.
6.3.1	Old Plant Landfill Drainage Ditch
During operations of the old plant (CPC) landfill as a wastewater neutralization unit (1954 to
1972), there was a drainage ditch that connected the former landfill to the wastewater ditch.
The drainage ditch was unlined except with natural clay. It was filled in concurrently with
closure of the CPC landfill in 1977. Due to the extensive earth work related to construction of
the pH pond and closure of the landfill in the area, there is no surficial evidence of the ditch.
Sampling indicated low levels of contaminants in soil. It has been concluded that this unit has
never been a significant source of groundwater contamination.
6.3.2	Crop Protection Chemicals (CPC) Plant
This unit is the former location of the plant that was constructed in 1952 and initially
manufactured nronochlorobenzene, adding pentachloronitrobenzene (PCNB) in 1956. In 1973,
the plant was expanded to produce trichloroacetonitrile (TCAN) and 5-ethoxy-3-
trichloromethyl-l,2,4-thiadiazole (Terrazole_). The PCNB, TCAN and Terrazole
manufacturing areas were collectively referred to as the Crop Protection Chemicals (CPC)
plant. The CPC plant was shut down in 1982 for market economic reasons. In 1984 the
business was sold and the plant area was decommissioned, dismantled and covered with an
approximate 2-foot recompacted clay cap and topsoil. The capped area was then vegetated.
TTie plan for decommissioning and dismantling die CPC plant area was approved by ADEM in
1983 and the work was completed in accordance with that plan. Soil sampling under the
RI/FS indicated elevated levels of semi-volatile organics in soil (no mercury was detected), but
modeling indicated that the contaminants would not migrate from the soils at concentrations
that would affect groundwater above protection levels.
6.3.3 Mercury Cell Plant
The former mercury cell plant is an area approximately 180 x 250 feet that was the site of the
structures and operations for the former mercury cell chlor-alkali plant. The mercury drum
storage pad, which was clean-closed, and the mercury recovery vystems were SWMUs located

-------
Olin Corporation - Mcintosh, Alabama
Page 13
down in 1982. The area was decommissioned and then capped in 1986. Decommissioning
included removal of all aboveground structures. The concrete pads and foundations were left
in place; the sumps and trenches were backfilled with clay, and the area was covered with
asphalt. Decommissioning of the mercury cell plant, which was a process unit, was not
subject to regulations of AD EM or EPA. Soil samples under the RI/FS indicated a wide
variation in mercury concentrations, however, the TCLP test indicated that the mercury was
not significantly mobile and would not affect groundwater above the groundwater protection
standards.
6.3.4	Well Sand Residue Area
Well sands were generated during the period from 1952 to 1968 from development and
operation of the brine wells for the mercury cell chlor-alkali process. These sands were
residues of the natural insoluble material in the salt dome and were deposited in mounds in the
brine field area. The well sand in these mounds is a cemented, granular material that has the
consistency of sandstone. After 1968, Olin changed the method of removing brine from the
salt dome cavities, leaving the residues in the cavity and thus eliminating their accumulation at
the surface. Sampling under the RI/FS indicated elevated levels of mercury in the cemented
well sand, but non-detectable levels of mercury in the TCLP extract indicated that the mercury
was immobile and not a source of groundwater contamination.
6.3.5	Strong Brine Pond
The strong brine pond was constructed in 1952 and is a former process unit that was
approximately 340 x 340 feet, constructed partially above-grade in natural clay. The strong
brine pond was located adjacent to the weak brine pond described above, but the functions and
regulatory status of these two ponds differed, as summarized below:
¦	The weak brine pond was used to manage process brine and hazardous waste
streams while the strong brine pond was used only to manage process brine.
-Thus, the weak brine pond was a regulated unit under RCRA and was closed
under RCRA Part 265 regulations and had potential for higher mercury
concentrations in any seepage.
¦	The strong brine pond was removed in 1985 at the same time the weak brine
pond was being closed under RCRA. Solids from the strong brine pond were
placed in the weak brine pond before its contents were solidified.
¦	The weak brine pond was closed in place under RCRA. The pond contents
were solidified and capped with a RCRA multi-layer cap. The weak brine pond
is regulated today under the RCRA post-closure operating permit.
Removal of the strong brine pond was conducted by dewatering and scraping out the material
in the pond. As mentioned above, the material was then placed in the weak brine pond. The
strong brine pond area was then graded flat, capped, and vegetated. Even though the strong
brine pond was a process unit, it was closed in conjunction with the weak brine pond (a RCRA
unit). Sampling under the RI/FS indicated that the strong brine pond area is not a source of

-------
Olin Corporation - Mcintosh, Alabama
Page 14
7 Extent of Groundwater Contamination Before Corrective Action
Figures 7 through 9 present isocontours of mercury concentrations in groundwater both before
corrective action (i.e., groundwater pumping and treating) was initiated in September 1987 and
currently. These figures present typical extents based on the highest concentrations from given
wells regardless of the location of the well screen. This is important in considering
contamination at the Mcintosh plant because of the dense brine present in areas near and
downdip of the weak brine pond, i.e., wells screened near the top of the Miocene clay surface
where the dense brine has come to rest will have higher concentrations than adjacent wells
with shallower screens. This is illustrated on Figure 11 which presents time-series plots for
three well pairs (shallow and deep). The deeper wells have significantly higher concentrations
of mercury because of the presence of the dense brine layer which has a high concentration of
mercury.
The plume of contamination extends east and west of the plant area (the source area) because
of the hydraulic groundwater divide described above and because of the dense brine layer's
tendency to flow by gravity along the surface of the Miocene clay.
8 RCRA Groundwater Corrective Action Program
The corrective action program began operating in September of 1987. The system was
designed using a groundwater model, the USGS McDonald-Harbaugh model MODFLOW.
The design criteria included containment of the plume within three months of well start-up, the
capture of contamination that had appeared at the southern property line (well El), and clean-
up to RCRA groundwater protection standards within 30 years. The system includes five 10-
inch diameter PVC interceptor wells that screen the saturated thickness of the Alluvial
Aquifer. The wells are designated as CA and their locations are shown on Figure 3.
Each well is equipped with its own treatment system. The discharge from corrective action
well CA1 is pumped to a natural draft air stripper to reduce the volatile organic concentrations.
Corrective action well CA1 is not equipped for mercury removal because it is located beyond
the boundary of the extent of mercury contamination as shown on Figure 7. Discharge from
the air stripper flows to the plant effluent ditch. The discharge from CA2 is pumped to a
forced draft air stripper for volatile organic reduction. The stripper then overflows to a single
gravity box filter with 8,000 pounds of activated carbon for mercury removal. Final discharge
from the carbon bed flows to the production effluent ditch. Treatment facilities for CA3 are
similar to CA2, except that the CA3 discharge from the carbon bed is routed through a
separate pH adjustment system before entering the plant effluent. Water from CA4 requires
treatment for volatile organics with a forced draft air stripper. No mercury removal is
required at CA4, and the effluent from the stripper is to NPDES-permitted Outfall 001C.
CA5 is also discharged through a forced draft air stripper to reduce volatile organic
concentrations. The stripper then overflows to a single gravity box filter with 8,000 pounds of
activated carbon for mercury reduction. The discharge from the carbon bed is routed to the
same pH adjustment system as CA3 effluent before entering the production effluent.
The discharge from the treatment systems for wells CA1, CA2, CA3, and CA5 is through
NPDES Outfall 001. The discharge from the treatment system for well CA4 is through
NPDES Outfall 001C. The treated effluent flows through the wastewater ditch and ultimately

-------
Olin Corporation - Mcintosh, Alabama
Page 15
Olin operates the wells and treatment systems as integral parts of its chemical manufacturing
operations. Olin utilizes one operator per shift to operate the treatment systems, the corrective
action wells, and all other NPDES treatment systems throughout the plant. Additionally, Olin
has developed and implemented a preventative maintenance program for the corrective action
wells and treatment systems.
Since implementation of the CAP, the interceptor wells have been operating within the
following pumping rates:
Corrective Action Well Design Pumping Range Approximate Pumping Range
(gpm)	(gpm)
CA1	150	100-150
CA2	100	80-100
CA3	100	100-110
CA4	60	40-60
CA5	180	150-180
The RCRA Corrective Action Program (CAP) has operated efficiently and with few problems
since its start-up. Most operational problems with the wells have been associated with partial
blockage of well screens with iron bacteria. This blockage gradually decreases well yield. If
the yield decreases to 70 to 80% of its design flow, the well is acid- and/or hypochlorite
washed to restore yield. This treatment has been successful in restoring yield in all wells that
have sustained a partial blockage.
The few problems with the treatment systems have been associated with the carbon beds
removing mercury from the well discharge. The beds have a long life (1 to 2 years), and
because of this time in service, carbon fines have carried through the bed, settled in the outlet
chamber, and been subject to resuspension into the treated water flow. This can possibly cause
some periods of increased mercury in the discharge. This has the potential to affect
compliance with Olin's NPDES permit and must be corrected quickly by removing the fines.
The major "problem" with the CAP is the rate of aquifer restoration inherent in all pump-and-
treat remedial actions. If sources have been remediated and the contamination is in the
aquifer, this rate cannot be appreciably increased because of the properties of contaminants and
aquifers, and the time over which the contamination took place. Pump-and-treat systems do an
excellent job of containing plumes of dissolved contaminants. They also remove a substantial
amount of contamination from the aquifer. However, they do not rapidly decrease the
concentrations of contaminants in groundwater. Contaminants released into groundwater many
years ago diffuse into the fine-grained soils of the aquifer. Many contaminants, such as
mercury and chlorinated organics at Mcintosh, adsorb to these fine-grained soils and are
released back into groundwater slowly. Even if groundwater velocities are increased by
pumping more groundwater, the increase will occur predominantly in the courser soils, i.e.,
the paths of least resistance, and the contaminants adsorbed or residing in the pores of the finer
soils will not partition into the groundwater any quicker. Solutions to this "problem" are not
readily available. Many contaminants are not amenable to bioremediation. In situ
bioremediation is gaining some favor for contaminants that are amenable, but understanding of

-------
Olin Corporation - Mcintosh, Alabama
Page 16
9 Results to Date from the RCRA Corrective Action Program (CAP)
Results from the RCRA Corrective Action Program (CAP) can be indicated by a reduction in
either the extent of the plume or in concentrations on monitoring wells over time. Figures 7
through 9 present the extent of the plume, showing the location of isocontours of mercury in
groundwater for concentrations of 2, 10, and 100 ug/1, respectively. The isocontours are
presented for data collected before corrective action was initiated and for data collected during
1993 (current case on the figures). Figures 10 through 14 present time series plots of mercury
concentration for certain monitoring wells and the five corrective action wells (Figure 10).
The monitoring well plots are associated with the dense brine area (Figure 11) or with the
isocontour figures (Figures 12 through 14).
Another key parameter for assessing the effectiveness of a pump-and-treat system is capture
zones created by the pumping wells. Figure 5 presents a potentiometric surface plot before the
pumping wells began operation and Figure 6 presents a plot from 1993 data. The
potentiometric surface depicted in Figure 6 shows that corrective action wells CA1 and CA2
have established a large single cone of depression. CA3, CA4, and CA5 have also established
individual cones of depression.
Water elevations in monitor well PL10S, which is situated between CA4 and the Tombigbee
River, show little change. The influence of the adjacent Tombigbee River probably has
masked any effects from the pumping. Water elevations in background monitor well WP9A
have decreased approximately 2 feet.
Olin's quarterly RCRA monitoring programs consist of collecting samples from 37 monitor
wells and analyzing the samples for mercury (from all 37 wells) and organics (from 21 of
these wells).
Although the system has only been operating for about five years, the following observations
can be made regarding the system's effectiveness:
¦	The system has significantly lowered the water table, and is effective at
controlling contaminant migration from any known sources. Distinct cones
have been developed in the Alluvial Aquifer at the five corrective action wells
(Figure 6).
¦	Distinct cones of influence have been developed at Wells CA1 and CA2 (Figure
6). These wells would capture any westward and southwestward movement of
the mercury and organic plumes.
¦	The CAP is effective at controlling migration of the mercury plume to the south
as shown by the time versus concentration curves. Monitor well El (Figure 12)
had increasing concentrations from about 1984 to 1988, followed by a steep
decline. This decline is attributed to the effect of corrective action well CA5,
situated about 600 feet to the northeast of El, and started up in 1987.
¦	The CAP has had limited hydraulic effects to the east of PL10S, in the
southeastern area of the facility, due to its close proximity to the river.
Mercury concentrations have ranged from 1.0 to 3.8 mg/1 and chloroform from

-------
Olin Corporation - Mcintosh, Alabama
Page 17
constituent. However, there are no known sources in the area and it is expected
that concentrations will eventually decrease with continued pumping.
¦	The mercury concentration data from the weak brine pond area show slight
decreasing trends in the monitor wells screened in the upper zone of the Alluvial
Aquifer, which indicates that contamination is being removed. These trends are
shown on the time versus concentration curves for MP9, BR7, and BR8 (Figure
10).
¦	There are no apparent trends in mercury concentrations for wells screened in the
lower zone of the Alluvial Aquifer in the weak brine pond area (Figure 11). It
is concluded that dense, mercury-containing brine that sank to the top of the
Miocene clay constitutes a secondary source of mercury in the area.
Accumulated brine at the base of the aquifer would explain the higher chloride
and mercury concentrations in the lower zone as compared to the upper zone
Alluvial Aquifer wells. The existing corrective action system, which includes
wells screened over the entire saturated thickness of the Alluvial Aquifer, would

-------
.4
	Sea-
\
«rOtlN f>ROPERTY BOUNDARY-
PHOPERTY BOUtOHRY E)
LEC£WP
—*— our
NNN OLt
laffl/FS
MeJNTOSH PLANT SITE
U
'oottocrd-Cftrde <
Consulting Enfptssa.

and EhihiniimM ?tlinUjln
Baton Rougat LoutaJono
QUN CWEKOICAIL CORPORATION-
CHABLESTOW, TEMHE3SEE
SCALE;
uaoz er: SJL
OflTE 12/90
ru «a.
AS SHOWN
aeasD stslrfAt
0K&B./1*
90B440C
RE; U.S.as. 7.3 MINUTE SERIES QUADRA
38YE LOC^TaOW MAft

-------
Olin Corporation - Mcintosh, Alabama
General Plot Plan
Hazardous Waste Drum
(Flammable) Storage Pad
(Cleaned Closed)

-------
fl
3
Olin Corporation - Mcintosh, Alabama

-------
Olin Corporation - Mcintosh, Alabama
Structual Contour Map
Miiocene Clay Surface
feet above MSL
HAZARDOUS WASTI DRUM
(FLAMMAILI)STORAdl PAD
Note Contours are generalized for presentatcn purpotos.

-------
Olin Corporation - Mcintosh, Alabama
Potentiometric Surface Contours
Before Corrective Action
K«fN«RI
Note Contour* are generalized for presentation purposes
. A HAZARDOUS WASTI MUM
14 (FUUMMABLE)STORAQE pad
(dimcletid)

-------
Olin Corporation - Mcintosh, Alabama
Potentiometric Surface Contours
HAZARDOUS WASTE DRUM
(FLAMMAfiLC)STOftAOE PAD
Nate Contour* are generalized foi presentation purpose®

-------
Olin Corporation - Mcintosh, Alabama
Mercury in Groundwater -100 ug/l Contours
Before and After Corrective Action


PROPERTY LINE
Note Contours are generalized for presentation purposes
-	Before
-	Current

-------
Olin Corporation - Mcintosh, Alabama
Mercury in Groundwater -10 ug/l Contours
Before and After Corrective Action
Note Contours are generalized for presentation purposes
- Before
m -Current

-------
Olin Corporation - Mcintosh, Alabama
Mercury in Groundwater - 2 ug/l Contours
Before and After Corrective Action
Note Contours are generalized (or presentaton purposes
-	Before
-	Current

-------
Figure 10
Time Series Plots
Corrective Action Pumping Wells
Mercury Concentrations, ug/l
Corrective Action Well 1
0.6
0.5
0.4
0.3
0.2
0.1
Jan-87 Jan-89 Jan-91 Jan-93
Corrective Action Well 2
50
40
30
20
10
0
Jar»-87 Jan-89 Jan-91 Jan-93

Corrective Action Well 3
12

10

8

"6
n t
4 „
v / \ ^ A. Jt
2

n



Jarv87 Jan-89 Jan-91 Jan-93

-------
Figure 11
Time Series Plots
Wells Illustrating Hg Associated with Dense Brine
Mercury Concentrations, ug/i
Shallow Wells	Deep Wells
Monitoring Well BR-7D
300
Jan-82 Jan-85 Jan-88 Jan-91 Jan-94
Monitoring Well BR-8
250
200
150
100
50
0
Jan-82 Jan-85 Jan-88 Jan-91 Jan-94
Monitoring Well BR-8D
300
250
200
150
100
50
0

-------
Figure 12
Time Series Plots
Wells Illustrating Change in 100 ug/l Contour
Mercury Concentrations, ug/l

Monitoring Well E-1
250
a
200
K
150
JVt
100
T \
50

A
• • ¦ V.	
U

Jarv87 Jan-89 Jan-91 Jan-93

Monitoring Well E-6
250

200
4 A
150
r M t
100
t aL
50

n

U

-------
Figure 13
Time Series Plots
Wells Illustrating Change in 10 ug/l Contour
Mercury Concentrations, ug/l
Monitoring Well PH-7
Jan-87 Jan-89 Jan-91 Jan-93
Monitoring Well PH-2D

Jan-87 Jan-89 Jan-91 Jan-93

Monitoring Well E-1
250
A
200
A
150
/ft
100
f \
50
jJ \ji
n
ti—mm
\J

Jan-87 Jan-89 Jan-91 Jan-93

Monitoring Well WP-3
250

200
I
150
A
100

50
A/W n

y si k
U
V '

-------
Biography
of
Bradley A. Jackson, Remedial Project Manager
U.S. Environmental Protection Agency
Waste Management Division
Education:
Bachelor of Science, Biology; University of West Florida, Pensacola, Florida; April 1983.
Work Experience:
U.S. Environmental Protection Agency, Region IV; Waste Management Division, South
Superfund Remedial Branch; November 1989 - Present. Duties include project management,
technical review, public relations, contract management, CERCLA enforcement activities.
NUS Corporation, Atlanta, GA; EPA Field Investigation Contractor;'January 1984 - November
1989. Duties included project management, CERCLA site inspections and reporting, site
evaluation for National Priorities List, staff supervision.
EPA Library Region 4
025
09

-------
SCALE
80 FEET
Delta Gas wen
Railroad Track
	 Fence
		 Fence Gale
Adiacent B smessei
Y///A Concrete Siructure
• Well Location
uw-n Monitoring Well IGenhhcation
Source: GCO Well Installation Plan, April 1990
FIGURE 2
MONITORING WELL LOCATIONS
GOLD COAST OIL SITE

-------
TOTAL VOC CONCENTRATION
MONITORING WELL NO. 13
GOLD COAST OIL SITE
CONCENTRATION (UG/L)

-------
An Abstract
Groundwater Remediation
of the
Gold Coast Oil Superfund Site
Miami, Florida
by
Bradley A. Jackson, Remedial Project Manager
U.S. Environmental Protection Agency
Waste Management Division
The cleanup of the Gold Coast Oil Site is truly one of tlu, success stories of the Superfund
program. The Site is the former location of a used-solvent reclamation facility located in Miami,
Florida. The Site overlies the Biscayne aquifer which is the sole source of drinking water for
the Miami area. The aquifer is composed primarily of sand, limestone, and sandstone. The
water table of the aquifer is only a few feet below ground surface, leaving the aquifer relatively
unprotected, and susceptible to contamination.
Following the closure of the facility, the following sources of contamination were identified at
the Site: 1) a storage area of approximately 2,500 corroded and leaking drums that contained
sludges from the solvent reclamation process, contaminated soil, and paint sludges; 2) 26
horizontal storage tanks containing hazardous substances; and 3) extensively contaminated surface
soils and contaminated groundwater. Numerous hazardous organic compounds and heavy metals
were detected in surface soils at the site with levels ranging up to the hundreds and thousands
of parts per million, respectively. The levels were well above acceptable health-based limits.
Groundwater contamination from hazardous organic compounds was also detected in a
60-foot-deep on-site well. Subsequent groundwater investigations confirmed the presence of
extensive groundwater contamination in the upper 50 to 60 feet of the aquifer. The contaminant
levels detected were also well above acceptable health-based standards.
Cleanup activities in 1982, 1989, and 1990, focused on removal of surface debris, bulk liquid
wastes, contaminated sludges and soils, and leaking drums. During these activities, five tank
truck loads of contaminated bulk liquids; 1,660 cubic yards of contaminated sludges and soils;
and 2,500 leaking drums were removed from this one and one-half acre site for proper disposal.
A groundwater treatment system began operation in June 1990 and has treated approximately
77,500,000 gallons of contaminated groundwater. Through these cleanup activities, soil
contamination has been reduced to within acceptable health-based levels and groundwater
contamination has been reduced to a level which is very close to achieving the cleanup standards

-------

\
INNER TIER i NT L UENT
EOU/NEU L
TANK j
C. >
SOD 1UM
HYDROX i DE
DAY TANK
J GRAVITY
i SE f1 AR AT GR j

sludge (
"ULtJ!".;	r
T ANK '
IEP
oEOHEtiC ! NG BATCH RE AC ! OR
( 5BR I
DECANT
:OLLECTION
TANK \
inner tier eefluent
TO tf 1 POLiSHINC POND
LARDOI-.
ADSORPTION
A I
:ripper
LEGEND
EOU= EOUAL IZA1 I ON
NEU = NEUTRAL IZAT I ON
- - - = SLUDGE
	 = GROUNDWATER
FI CURE 3
INNER TIER

-------
/
SOURCE MATERIALS
250		500
APPRO* IMA1E SCALE IN FEET
LEGENO
OT-1 -if~ OUTER TIER EXTRACTION WELL
I T-1
+
INNER TIER EXTRACTION WELL
FIGURE 1

-------
CERCLA Groundwater Remediation Project
Capture zones for the system are estimated from drawdowns in 8
monitor wells surrounding the inner tier. Typical drawdowns
range from one to five feet. Drawdowns measured in October, 1993
averaged 3.32 feet. These drawdowns indicate that the inner tier
system is capturing the contaminated groundwater at the source
area
The inner tier treatment; system was designed to treat the highly
contaminated groundwater from the inner tier area. The overall
effectiveness of the system is measured by volatile organic
compound and TOC concentration reductions.
Based on overall average influent and effluenu values, the inner
tier system has averaged a 94% reduction of targeted volatile
organic compounds, 99% of TIC volatile organics, and 93% of TOC
concentrations.	Targeted volatile organic concentrations
averaged 3 92 ug/'l in the influent and 22 ug/1 in the effluent
The average TIC volatile concentrations are 1619 ug/1 in the
influent and 16 ug/1 in the effluent. TOC concentrations
averaged 1,974 mg/1 in the influent 128 mg/1 in the effluent.
Operational Problems
Both the inner and outer tier systems have proven to be effective
and reliable m operation. The ::: ly ~a;cr problems encountered
were with the inner ti.er pumps and _ -iss 'jpsets in the SBR.
The original pumps in the inner ' . wells experienced severe
clogging problems due to iron pr- :: c :*. .at;on It was believed
that the pumping action was cans:.".} ' . rcr. to precipitate out.
The pumps were replaced in 1990 w:* • . lider pumps. Although the
bladder pumps have performed we). : : <* q-er.t replacement of the
bladders is necessary. Two-inch •* *. : . v submersible pumps are
currently being tested in two of	. r.r.er tier wells. To date
they appear to be providing reliable 1 r.i effective performance.
The sequencing batch reactor performed very well for the first
eight months of operation Starting in the ninth month, it began
to experience biomass upsets with increasing frequency. Within
several weeks, the upsets were occurring on a weekly basis. It
was originally thought that high levels of ethylene glycol in the

-------
CERCLA Groundwater Remediation Project
The 9 inner tier wells are located directly downgradient of the
source materials (Figure 1). They are installed in saprolite to
a total depth of 42.5 feet and are spaced on approximately 80
foot centers.
The inner tier wells are constructed of 2-inch stainless steel
casing with 10-foot stainless steel screens. Seven of the inner
tier wells use pneumatic bladder pumps and controls. Two-inch
electric submersible pumps are currently being tested in the two
other wells. The inner tier influent is transmitted below grade
in polypropylene piping to the treatment system.
The inner tier treatment units (Figure 3) include flow
equalization, iron removal, biological treatment in an activated
sludge Sequencing Batch Reactor (SBR), air stripping, and carbon
adsorption.
The inner tier influent is pumped to a 4500 gallon equalization
tank where sodium hydroxide is added in the tank to neutralize
pH. The groundwater is then pumped to a gravity separator where
iron precipitates out as a flocculate. Polymer is added at the
separator to improve separation. Sludge from the separator is
dewatered in a filter press and then disposed of ac an off-site
landfill.
Groundwater flows from the plate separator to the SBR. The SBR
uses an activated sludge biological process to remove organics
from the water. The operation is a batch process, where aeration
and settling, both occur in one vessel.
From the SBR, the biologically treated groundwater is pumped to a
30 gpm air stripping tower, and finally to an activated carbon
adsorption unit. The treated inner tier groundwater is also
discharged to the first polishing pond of the plant's wastewater
treatment plant.
SYSTEM OPERATION
Outer Tier
The primary goal of the outer tier extraction system is to
control further migration of the contaminated groundwater toward
the site perimeter. The system was designed to use the minimum

-------
CERCLA Groundwater Remediation Project
and semi-volatile organics, and metals in the groundwater
samples. Concentration ranges for selected chemical constituents
identified during the Remedial Investigation are provided in
Table 1.
The chemicals detected Ln the groundwater were similar to those
identified in the burn pit materials and process sludges that had
been buried at the site, indicating that these materials were the
likely source of the groundwater contamination. Groundwater
contamination was higher in samples collected adjacent to the
burn pit area, and lower or below the detection limit near the
downgradient site perimeter. Site related contamination were not
detected in samples collected from the off-site wells.
The Remedial Investigation for the site identified the need for
two operable units at the site. Operable Unit 1 (OU-1) addressed
the stabilization and remediation of contaminated groundwater at
the site. The Feasibility Study and Record of Decision for OU-1
were issued in 198S. The OU-1 Remedial Design and Remedial
Action, which are the primary topics of this paper, were
completed in 1989.
Operable Unit 2 addressed the remediation of the source materials
and contaminated stream sediments. The remedy consisted of
excavation and treatment of the burn pit materials, process
sludges, and-stream sediments. The burn pit materials and stream
sediments were solidified with Portland cement. The process
sludges were incinerated on-site, and the residual ash
solidified. All stabilized materials were then buried back in
the original excavation. The Operable Unit 2 Remedial Action was
completed in September, 1992.
SYSTEM DESIGN AND CONSTRUCTION
The objectives for the OU-1 remediation as defined in the
Remedial Design Report were to (1) control further migration of
the contaminated groundwater toward the site perimeter, and (2)
remove contaminated groundwater for subsequent treatment and
discharge. To accomplish this, two separate extraction and
treatment systems were designed. The first system, called the
outer tier, was designed to capture and treat the groundwater at

-------
CERCLA Groundwater Remediation Project
Glenn G. Boylan, P.E.
Everett W. Glover, P.E.
Brian P. Anders, E.I.T.
ABSTRACT
A CERCLA groundwater extraction and treatment project has been
completed at an industrial facility in Region IV. The two
primary objectives of the project were stabilization and
remediation of the contaminated groundwater.	The CERCLA
Five-Year Review has just been completed for the site, and shows
that the groundwater at the site has been stabilized and the
treatment system is effectively removing organic compounds from
the groundwater.
The project involved the design and construction of two
extraction and treatment systems. The first system, called the""
outer tier, was designed to stabilize the groundwater at the
downgradient property line to prevent migration off-site. The'
groundwater captured by the outer rJ =>r has relatively low
concentrations of contaminants and is treated by air stripping
and carbon adsorption.
The second system, or inner tier system, was designed to capture
the more highly contaminated groundwater just downgradient of the
buried source materials. The treatment for the inner tier water
consists of a combination of physical, chemical, and biological
processes.
The systems have proven to be reliable. The only significant
problem encountered to date has been recurring biomass upsets in
the biological reactor. This problem was solved by minor changes
to the sludge wasting rate.
Physical and chemical data monitored for both systems show that
the remedial goals are being met. The outer tier wells are
creating a hydraulic barrier to groundwater flow, keeping the
contamination on-site. The inner tier wells are capturing the
highly contaminated groundwater at the source, and the inner tier

-------
THOMAS F. SCIPLE is Environmental Supervisor with Ciba-Geigy
Corporation (Ciba) in Mcintosh, Alabama, where he is responsible
for permit and permit compliance in air, water and solid waste
areas. Mr. Sciple has been employed with Ciba for 25 years; of
which 22 years have been in the Environmental Field with
experience in Research, Development, Operations, Regulatory and
Legal. Mr. Sciple is a member of various Air, Water and Waste
Management Associations including Chemical Manufacturing
Association's (CMA) Ground Water Task Group, Synthetic Organic
Chemical Manufacturing Association (SOCMA), AlaChem Association,
The Business Council of Alabama, Alabama's Air Toxic Advisory
Committee, Alabama's Waste Minimization Advisory Committee,
Alabama's Clean Air Act Implementation Committee, Alabama's
Comprehensive Ground Water Program Advisory Committee and The
Alabama Emergency Management Commission.
PATRICK D. HALLETT is an Environmental Associate with Ciba in
Mcintosh, Alabama, where he provides technical and management
support for the RCRA groundwater management program and other
environmental management issues. Mr. Hallett has worked in the
environmental field since 1981 and has been involved with RCRA

-------
GROUND WATER REMEDIATION ACTIVITIES AT
THE CIBA MCINTOSH, ALABAMA SITE
Presented by Patrick D. Hallett
Environmental Associate, Mcintosh Site
Facility History
The Mcintosh facility, formerly owned by Geigy Chemical Corporation, began
operations in October 1952, with the manufacture of one product. Through 1970,
Geigy expanded its Mcintosh facilities by adding the production of herbicides,
insecticides, agricultural chelating agents, sequestering agents for industry and
fluorescent brighteners used in laundry products.
In 1970, Geigy merged with Ciba (Chemical Industry in Basel, Switzerland),
forming the CIBA-GEIGY Corporation (Ciba). Since then Ciba has continued to
expand its operations with the production of resins and additives used in the
plastics industry and small volume specialty chemical products (i.e. water
treatment chemicals and fire fighting foams). The present facility occupies
approximately 1,500 acres and employs around 1,200 workers.
Site Location and Topography
The Ciba plant is located approximately 50 miles north of Mobile in southern
Washington County northeast of Mcintosh, Alabama. The developed plant site is
situated between the Southern Railroad nght-of way on the west and extends
nearly to the escarpment separating the upland terrace from the floodplain of the
Tombigbee River. The property boundary's »?xtend beyond the railroad westward
toward U.S. Highway 43. The northern ~jiVjo of the property merges into a pine
forest and to the south, the property is hounded by Olin Corporation. The
southeastern portion of the Ciba property extends to the banks of the Tombigbee
River. The nearest population center to the site is the Town of Mcintosh, which is
located approximately two miles to the southwest.
The Ciba facility is located in the Southern Pine Hills District of the East Gulf
Coastal Plain Physiographic Province and s north of and on the western margin of
the Mobile-Tensaw River Delta. The Ciba facility is situated on a low terrace
adjacent to the Tombigbee River Floodplam. Trie well-developed floodplain in the
vicinity of the site is characterized by broad meanders and oxbow lakes.
Corrective Action Background
The EPA Region IV conducted an investigation in the late 1970's and early
1980's of the Ciba site. As part of the investigation, EPA sampled water wells on
the Ciba property. The sampling indicated the presence of Lindane in one well

-------
sample. As a result of this information and supplemental data from surface water
and soils samples, the Ciba Mcintosh site was included on the National Priorities
List (NPL) in September 1983.
Ciba established a ground water monitoring program in 1981 in response to
the then, new, RCRA regulations. The monitoring indicated the presence of ground
water contamination in the alluvial aquifer. Ciba contracted the services of a
consultant to conduct a series of studies investigating ground water quality and
aquifer characteristics. Those efforts resulted in the formulation of the ground
water corrective action program which focuses on hydraulic control of the alluvial
aquifer and source removal or control. In October 1985, EPA issued Ciba a RCRA
Permit, which included a Corrective Action Plan prepared by Ciba proposing to
install a series of pumping wells to form a hydraulic barrier at the site. The
interceptor well system began operating in late 1987.
The Corrective Action Plan also included provisions for a Site Investigation
(i.e., Remedial Investigation/Feasibility Study) to identify and characterize past
disposal areas that may act as sources of ground water contamination. Other
elements of corrective action included closing RCRA permitted hazardous waste
management facilities that may have contributed to environmental contamination.
These RCRA facilities included surface water impoundments used to treat or store
wastewater or sludge, as well as two landfills. The RI/FS, driven by both RCRA and
CERCLA authority, subsequently identified fcleven additional past disposal sites in
the upland area of the Ciba property. The contaminants associated with these past
disposal sites include pesticide residues, by-products and intermediates from
pesticide manufacturing. One objective of the RI/FS is source removal or control to
eliminate the continuing contribution of contaminants to the alluvial aquifer. This
task is currently in the Remedial Design phase and is being handled as a CERCLA
activity.
Geology and Hvdroaeoloav
Southwestern Washington County is underlain by recent alluvium,
Pleistocene age low terrace deposits and sediments comprising the Miocene Series
Undifferentiated. These strata consist of alternating deposits of sand, clay, silt and
gravelly sand. A nearly continuous surficial clay layer is underlain by deposits of
silt, sand, gravel and clay. The surficial clay layer ranges in thickness from a few
feet to over 50 feet. The contact with the underlying sand is characterized by
sandy clay, sand and small caliber gravel. These Pleistocene age deposits range in
thickness from.60 to approximately 100 feet. The Pleistocene deposits
uncomformably overlie more than 700 feet of alternating layers of Miocene age
sand, gravel and clay. The Upper Miocene clay, with a thickness ranging from
approximately 30 feet to over 100 feet at the site, acts as an effective aquitard
between the alluvial aquifer and the Miocene aquifer system.
Ground water is found in the semi-confined alluvial aquifer approximately 35
feet below land surface. Ground water is available from three aquifer systems:

-------
(1)	- The Recent and Pleistocene terrace and alluvial deposits present under
semi-confined conditions;
(2)	- The underlying sand and gravel deposits of Upper Miocene age (Upper
Miocene aquifer) under locally confined conditions; and,
(3)	- The highly mineralized Lower Miocene aquifer (Miocene aquifer) under
confined conditions.
The shallow alluvial aquifer is separated from the Upper Miocene aquifer by a
clay aquitard with a measured permeability of 10"8 cm/sec. This aquitard has been
shown through pumping tests to be an effective barrier to the contaminated alluvial
aquifer at the site. Periodic sampling of the Miocene aquifer has not indicated
contamination.
The water table in the Alluvial aquifer normally slopes gently south-southeast
towards the Tombigbee River. This natural pattern is modified in some portions of
the plant site by the pumping of plant water supplies and recharge from the river
water reservoir. In addition, the Ground Water Corrective Action Program has
produced a dynamic water surface which has reversed the natural ground water
flow patterns along the southern portion of the site.
Interceptor Well System
A groundwater interceptor well system was installed at the site following
extensive hydrogeologic investigations and numerical modelling. The interceptor
well system consists of ten extraction wells located along the southern tier of the
Ciba property. The wells were initially designed to pump approximately 2.3 MM
gallons of groundwater daily, with individual pumping rates ranging from 50 gallons
per minute (gpm) to 450 gpm. The well system currently removes approximately
1.4 MM gallons~per day with individual pumping rates ranging from 50 gpm to
approximately 185 gpm. The ground water is then conveyed to the on-site
wastewater treatment facility where it is treated in the activated sludge system
prior to being discharged to the Tombigbee River through the site's permitted
NPDES discharge point.
The natural flow direction of the ailuvial aquifer is to the south-south east, as
is illustrated in the water surface map constructed in 1987, prior to initiating
pumping. Utilizing data derived from pumping tests, our consultants (P.E.
LaMoreaux and Associates) produced a numerical model of the configuration of the
alluvial aquifer at various pumping rates and simulated the areal extent of draw
down. A water surface map was produced after the interceptor system was in
operation. The map clearly illustrates the effectiveness of the system in capturing
the ground water in the alluvial aquifer in the portion of the property south of the
waste management areas. Surface water maps depicting the reversal of the
hydraulic gradient are produced semi-annually and provided to the State and EPA.

-------
As noted above, the pumping rates of the individual wells and the overall
system have been adjusted since the system began operating in late 1987. The
reduced pumping is the result of adjusting the system to minimize pumping volume
while ensuring the reversal of the hydraulic gradient. In several instances, such as
well PW-9, the pumping capacity is impacted by the presence of iron precipitating
bacteria. The predominant iron bacteria is the genera of Leptothrix . Persistent bio-
clogging reduces the yield and has resulted in the need for frequent well
maintenance to ensure sufficient yield to maintain a reversal of the hydraulic
gradient.
A well treatment involves the addition of acids and recirculating the acidified
column water. A treatment may take two or three days. The treatment typically
includes the following steps:
-	Removing the pump and noting the column pipe condition;
-	Adding muriatic acid (hydrochloric acid), (approximately 60 gallons);
-	Adding hydroxyacetic acid, (approximately 2 gallons);
-	Circulating the acid mixture in the well with a submersible pump for
approximately 2 hours;
-	Repeating the process in the afternoon;
-	Allowing the acid mixture to sit undisturbed overnight;
-	Repeating the process a second day and re-installing the pump on the third
day.
The treatment schedule varies by wen. Two wells require an abbreviated
one-day treatment quarterly, while the remaining eight wells require treatment on a
semi-annual basis.
Annual maintenance costs for the well treatment currently are approximately
$250,000.00.
Ground Water Quality
The Ground Water Protection Standard for the Ciba site was negotiated in
1985 based on ground water quality data gathered since 1981. The Standard is
composed of the "Primary Drinking Water Standards", "indicator parameters" and
25 additional compounds identified through analysis of 40 CFR 261, Appendix VIII
constituents. Of this list, 14 constituents comprise the predominant set of alluvial
aquifer contaminants. Since the ground water interceptor system began operation
in late 1987, the concentrations of all the listed contaminants have been

-------
substantially reduced. The initial list of 29 constituents detected above background
in the wells has been reduced to less than ten. Nonetheless, the Ground Water
Protection Standard has not yet been achieved in the Compliance wells located
downgradient of the waste management units that have most likely contributed to
the ground water contamination.
Ciba prepares time-trend graphs of selected constituents to track the
progress of the corrective action on ground water quality. It is clear from these
graphs of the indicator parameters (i.e., specific conductance and TOX) and alpha-
BHC that the concentrations in the ground water have been substantially reduced
since the surface impoundments were closed in 1988 and 1989. In general, the
trend continues to indicate a reduction in contaminant concentrations but at a
much more gentle slope. The on-going remedial activities, focusing on source
removal or control, is expected to result in further reductions in contaminant
concentrations. These activities are scheduled to move into the Remedial Action
phase in 1996.

-------
Figure 14
Time Series Plots
Wells Illustrating Change in 2 ug/l Contour
Mercury Concentrations, ug/i
Monitoring Well WE-3
Jan-87 Jan-89 Jarv-91 Jan-93

Monitoring Well E-3
8

6


4
1

2~


r\
i

Jan-87
Jan-89 Jarv91 Jan-93
Monitoring Well E-1
250

-------

-------
RADIATOR SPECIALTY COMPANY
GROUNDWATER ASSESSMENT and INTERIM REMEDIATION PROGRAM
by
Richard L. Harmon, P.G., and Chris Cutler
ABSTRACT
The Radiator Specialty Company (RSC) Indian Trail, North Carolina Facility is located in the
contact zone between the Carolina Slate Belt, and the Charlotte Belt, within the Piedmont
Physiographic Province in which groundwater flow is fracture controlled. RSC has recently
completed Phase V of an on-going groundwater assessment program, and submitted Addendum
8 to the Facility's Post Closure Permit Application. Concurrent to these activities, RSC is in
the process of implementing an interim groundwater recovery and pretreatment system. This
system includes four (4) recovery wells, an extensive header pipe system, and packed air
stripping tower. Effluent from the system is discharged to the Union County POTW.
INTRODUCTION
At the Indian Trail, North Carolina facility, Radiator Specialty Company (RSC) manages three
(3) distinct operational divisions. These Divisions consist of the manufacturing of household
and automotive chemical products, hose packaging and the distribution of RSC product lines
worldwide. The Chemical Production Division blends aftermarket automotive chemicals in a
variety of consumer packages. The CRP Hose division packages a myriad of hoses for
numerous applications. The Distribution Division ships finished goods to customers
worldwide.
This facility was constructed in 1971 on approximately 180 acres of wooded farm land within
a rural community south of Charlotte, North Carolina. Since it's opening, this facility has
grown to currently employ over 270 people.
Regulated Units
Resource Conservation and Recovery Act (RCRA)
In August 1987, the North Carolina Department of Human Resources (since combined
into the North Carolina Department of Environment, Health and Natural Resources)
approved a revised Closure Plan for closing the two (2) surface impoundments located
on the east side of the site. This plan as implemented included removing the liquid and
sludge from these impoundments for transport via railroad tank cars to the DuPont
Chamber Works in Deepwater, New Jersey for disposal. Once this material was
removed from the site the impoundments were capped with clay and seeded with grass.
In January 1989, RSC submitted a Post-Closure Permit Application to the State
regulatory authority. In conjunction with the RCRA permitting activities, RSC initiated
a Groundwater Assessment Program to delineate the extent of the contaminated
groundwater on-site.

-------
Solid Waste Management Units (SWMUs)
On October 17, 1989, NUS Corporation under contract to the EPA issued the Final
Environmental Priorities Initiative Preliminary Assessment of Radiator Specialty
Company Indian Trail facility. In this report, NUS defined eight (8) Solid Waste
Management Units (SWMUs). These units included: the two (2) RCRA regulated
closed impoundments; the product mixing/recycling units; the tank farm; the product
transfer line pad; the solid waste container; the solid waste compactor; and the product
transfer pump pad. Of these eight (8) units, two (2) are currently regulated by the
Resources Conservation and Recovery Act (RCRA), and six (6) were recommended for
no further action. This NUS report specified additional assessment activities at only the
following two (2) of these units; the tank farm and the transfer line pad. While the tank
farm is listed as one (1) unit by NUS, this unit consists of two (2) above ground tank
farms (Tank Farms A and B), and one (1) former underground storage tank (UST) farm
(Tank Farm C).
Based on this report, RSC proceeded with facility improvements including installation
of a covered dumpster area, a covered product transfer pump pad, installation of a vapor
extraction system within Tank Farm B and excavation and removal of contaminated soil
adjacent to the transfer line pad. Additionally, RSC has replaced all the in-ground
piping with overhead systems, and constructed a covered truck unloading area.
As a result of these activities, the EPA has removed the dumpster area (i.e the solid
waste container and the solid waste compactor) and the product transfer pump pad from
the SWMU list. Due to the close proximity of the five (5) remaining units, RSC is
currently assessing and remediating these areas simultaneously.
Regional Geology
The RSC site is located along the boundary between two geologic belts within the Piedmont
Physiographic Province. These two (2) units consist of the Carolina Slate Belt (CSB) to the
east; and the Charlotte Belt to the west. The Piedmont Province generally consists of a
northeast/southwest trending region extending southward from the Hudson River in New York
to Alabama. The Piedmont topography is typically described as a slightly elevated region with
low to moderate relief, generally dissected with valleys created by streams flowing on rocks
of varying erosional resistance. Ridges and up-lands have been developed by slower
weathering of more resistant rock types. Most rock types in this region are covered by thick
residual soil; however, thin weathered zones are not uncommon. Soil typical of this region was
generally developed through the in-situ weathering of fractured and faulty rock structures
(Thornburg, 1965).

-------
Carolina Slate Belt
The CSB is generally described as consisting of low-rank metamorphosed volcanic and
sedimentary rocks, situated in the east-central portion of the Piedmont of North
Carolina. Where Coastal Plain sediments overlap it to the east and the Charlotte Belt,
consisting of higher rank metamorphic and igneous rocks, borders it to the west.
Rocks comprising the CSB consist primarily of laminated and non-laminated
metamorphosed pelitic rocks. Rock color varies from bluish-gray when fresh, to brown
and reddish-orange when weathered. When present, laminated bedding planes are
typically well developed and exhibit bedding plane cleavage. Igneous intrusions have
been observed throughout the Belt. These intrusions are typically comprised of diabase
and metagabbro. The diabase occurs most commonly as dikes. General mineralogy is
plagioclase, clinopyroxene, and chlorite, with biotite and quartz occurring as void
fillings. The metagabbro is generally xenoblastic, but abundant sericite and epidote
obscure structure and mineralogy (Randazzo, 1972).
Quartz veins are present throughout the CSB. These veins (i.e. dikes and sills) are
typically intruded into fractures in the country rock. Large muscovite flakes are
commonly seen along these intrusions, possible indicating recrystallization of country
rock during time of intrusion. Quartz is usually milky in appearance (Randazzo, 1972).
A major structural feature of the CSB is the Gold Hill Shear Zone. The Gold Hill Shear
Zone is thought to be a thrust fault generally described as trending north fifteen (15)
degrees east, and extending southward from near Southmont, Davidson County, North
Carolina to the eastern edge of Indian Trail, North Carolina. Previous studies indicate
the rocks along the western margin of this Shear Zone are characterized by higher
ranking metamorphism, including the occurrence of both slate and phyllite. Local fault
planes, quartz veins and minor joints are also commonly associated with this major
Shear Zone trend (Randazzo, 1972; The Geologic Map of North Carolina, 1985).
Charlotte Belt
The neighboring Charlotte Belt is described as metamorphosed diorite, grandodiorite,
and tonalite intruded by gabbro, granite, diabase, syenite,and quartz batholiths, dikes
and sills. Each rock unit is generally massive with weak foliation. Most units grade
from one rock type to another over distances often ten feet to a few miles. Sharp
contacts are rare except along perimeters of intrusions where banding and contact
metamorphism may be observed (Randazzo, 1972). The Charlotte Belt is generally
described as a granite/diorite complex. Relationships of the granite and diorite in
complexes of this nature are obscure. Many areas in the Charlotte Belt suggest granite
has intruded the diorite, while at others the diorite appears to have intruded the granite.
The mineralogy of granite and diorite are essentially the same. They include three (3)
-»

-------
major rock-forming components: quartz (silica), feldspar (potassium, sodium, calcium
and aluminum), hornblende, and/or biotite (iron and magnesium). The different rock
types are identified by the percentages of quartz relative to other felsic minerals.
Granite contains 20 percent to 60 percent quartz while diorite contains less than 10
percent. Due to the similar mineralogical make-up of the parent rock, soil development
for the two are nearly identical (Stucky and Steel, 1953).
Site Specific Geology
Numerous subsurface investigations have been performed at the facility. These investigations
indicate the study area lies near the contact between the CSB and the Charlotte Belt. Test
borings indicate the site is generally underlain by two distinct separate rock types, phyllite and
diorite. Data suggest that the phyllite was intruded by the diorite, with evidence of possible
contact metamorphism exhibited in rock cores from various wells on-site. Similar orientation
of the fractures in both the phyllite and the diorite suggest a tectonic relationship with major
known structure trends (i.e. the Gold Hill Shear Zone).
Due to the hydrogeologic similarities between the phyllite and the diorite, the site has been
classified into three (3) hydrogeologic units. These units consist of saprolite. partially weather
rock (PWR), and bedrock.
The saprolite has been defined as material which could be bored with conventional hollow-stem
augers and exhibited less than 100 blows per foot, using the Standard Penetration Resistance
Test as described in ASTM D 1586. The saprolite consisted primarily of silty clay and ranged
in thickness from approximately 2.5-feet to approximately 25-feet across the site. Some silty
fine sand lenses were encountered within this material; however, they appear to be non-
continuous and merely an in-situ weathered product of mineralogical or grain size variations
in the parent rock.
The material encountered beneath the soil horizons was typically referred to as partially
weathered rock (PWR). PWR is a geotechnical engineering classification for material which
exceeds 100 blows per foot using Standard Penetration Testing procedure as described in
ASTM D 1586. For the purposes of this investigation, the base of this material has been
defined as hard, fresh to slightly weathered competent rock. The PWR encountered on-site is
mineralogically similar to the soil mantle; yet generally rock fragments, boulders, rock layers,
etc. are more prevalent due to lesser exposure to chemical and mechanical weathering effects.
The thickness of PWR observed from 8-feet to approximately 43-feet.
For the purposes of this investigation, bedrock has been defined as competent rock, described
as hard, fresh to slightly weathered. The bedrock underlying the PWR, is a hard, slightly
weathered to fresh laminated phyllite and diorite. The phyllite encountered typically exhibits
bedding planes of approximately 0.125-inches in thickness with a dip of approximately 75°.
Phyllite is typically defined as argillaceous rock intermediate in the metamorphic grade

-------
between a slate and a schist. Phyllite is typically differentiated from other argillites in the field
by its silky sheen observed along the cleavage planes imparted by the mica crystals. This unit
also exhibited two (2) sets of fractured joints: which include a slightly more predominant
subvertical set with dips of approximately 60° to 75°, and a subhorizontal set of fractures with
dips of approximately 0° to 20°. The subvertical fracture set appeared to parallel the bedding
planes of the formation.
The diorite component of the bedrock is a hard, slightly weathered aphanitic to phaneritic
diorite, typically possessing three (3) poorly developed fracture joint sets. These fracture joint
sets consist of: 1) a closely spaced, sub-horizontal fracture set with dips of approximately 10°
to 20°; 2) a closely spaced, sub-vertical fracture set with dips of approximately 30° to 60°; and
3) a less predominant vertical fracture set with dips of approximately 80°.
A fracture trend analysis conducted at the site indicates that both the phyllite and diorite units
mapped on-site are highly fractured. Figures 2-1 and 2-2 present both a revised Joint Rose and
a Contoured Stereo Net of 165 fracture sets mapped in the study area. Steep angle fracturing
was observed in both geologic units with less than 2 percent of the total of 165 points
expressing dips of less than 28°. Five (5) minor joint sets were observed in the stereo-net.
These sets each contained greater than 8 percent of the total. These five (5) sets are as
follows:
N10E78NW	N20E80NW
N40W80SW	N24W82NE
N36W84NE
These joint sets appear to correlate fairly well with two of the three fracture set dip angles
observed in the core samples collected across the study area.
AQUIFER DESCRIPTION
The aquifer underlying the study area is typical of the composite weathered residuum-
crystalline fractured rock aquifers located within the Piedmont region of North Carolina. The
aquifer is unconfined, existing under phreatic, or water table conditions. Under these
conditions, the water table surface is in equilibrium with atmospheric pressure and is not
confined by low-permeable layers between the surface of the water table and the surface of the
ground. Groundwater under water table conditions is typically recharged by direct infiltration
from precipitation. This water flows through the pore spaces of the geologic media to its
ultimate discharge in topographic low regions where the water table surface intersects the
surface of the ground. As a result of these typical flow patterns, the potentiometric (water
table) surface expressed by water table aquifers typically appears as a subdued replica of the
land surface topography.

-------
NUMBER OF LINEATIONS = 165
AZIMUTH INTERVAL = 10°
ORIENTATION OF JOINTS IN OUTCROP NEAR RSC
eject:
Radiator Specialty Co,
Indian Trail, NC
Title:
Join":
\C s e
Job:
Figure:
2-1
_L
ca.e:
MTS
PO Ba* *>087
Rtieign. nC
Triangle
Environmental goo-wt-siis

-------
LEGEND
CONTOURED STEREO NET
0% - 1.5%
1.5% - 5%
[7
/
5% - 8%
>8%
TOTAL 165 LINEATIONS
oject:
Radiator Specialty Co.
Indian Trail, NC
Title:
Steieo Net
Job: 1
Figure:
2-2
Scale:
NTS
?0 Bo* *<087
_	Biiexjn. nC 27629
Triangle	9i9-476-$us
EiwIROHMSHTAL 300-3*9-5115
!nc	
-------
Figure 2-3 presents a map of the potentiometrifc surface underlying the study area. This map
was generated from water table elevations collected on September 14, 1993 from monitoring
wells in which the well screen interval intersects or very nearly intersects the water table
surface. This map suggests that the shallow groundwater within the study area flows radially
away from the location of the former impouncjments.
Assessment Activities to Date
Since initiating assessment activities in 1989 REC has completed five (5) extensive phases of
the groundwater investigation. These combined! assessment activities included the construction
of 60 monitoring and/or interim recovery wells tanging in depth from 15 feet to 500 feet below
grade. During these activities, approximately 90 discrete aquifer zones were sampled to
characterize the groundwater chemistry.
Numerous innovative methods were employed in sampling these zones including discrete zone
packer testing, temporary well point construction and permanent Type II and Type III
monitoring wells. Specialty Type III monitorimg wells were designed and constructed using
as many as six (6) strings of casing installed in|a telescopic manner. Auger, core, mud-rotary
and air rotary drilling technologies were utilized [throughout various phases of this groundwater
assessment.
All drill cuttings, drilling muds and formation Waters generated during assessment activities
were contained and sampled. Based on analytical results RSC either recycled these materials
into existing product lines or shipped them off-site for proper disposal.
Remediation Activities To Date
To date RSC has implemented four (4) ongoing remedial systems. Two (2) vapor extraction
systems were installed to remediate SWMU regulated soil contamination. RSC is also
recovering non-aqueous phase liquid (NAPL) from Tank Farm C and recycling this material
into various product lines.
Additionally, RSC has constructed and is in the process of implementing an interim
groundwater recovery and treatment system. This system consists of four (4) interim recovery
wells located along the primary fracture set. Pumps within these wells are connected via
double-walled header piping to a pretreatment system constructed on-site. Due to potential
variations inherent to operation of an interim system, RSC has designed a piping network
which may be readily altered. This system consists of a series of manholes and vaults
connected with large diameter PVC pipe. Numerous smaller diameter header pipes may be fed
through these conduits to accommodate increased flow from existing or new recovery wells.
-------
x>
\
Sewer Line
l. MP"
^
-t-
\\
v
\ ,
/
o6^
.X\ * i
V- \ ' ,664.82
' •)
i
v
*
Legend
Direction of Shallow Component of
Groundwater Flow
Water Table Potentiometric Contour
(Sept. 14, 1993)
$ Monitoring Well
& Recovery Well
ojoct:	j^e: Water Table
Radiator Specialty Co. f 1 T>ot-p>jitjLQjnetric Map
Indian Trail, NC	! Job:
Figure:
2-3
Scale:
1"=200'
Tkt/wcta
PO. So* *108J
gjjngft. NC 27639
91+476-SIIS
aEwoonauntal gooat+siis
1U !nc
-------
The header pipes from the entire recovery system feed an equalization tank where the
recovered water is homogenized prior to pretreatment. From the equalization tank the
homogenized water is pumped through a packed air stripping tower and discharged to the
municipal sewer system.
The recovery/pretreatment system is operated from a central control panel. From this location
RSC personnel can monitor the entire system. In the event of an inner header piping leak, low
water level in any recovery well, or pretreatment system failure, the entire system automatically
shuts down and RSC personnel are notified by visual and audible alarms. In addition to
automated failsafes RSC personnel visually inspects the entire system daily.
Discharge Permitting
After pretreatment, the water is discharged to Union County's Publicly Owned Treatment
Works (POTW). In 1991, RSC in association with Union County, and the North Carolina
Division of Environmental Management (DEM) negotiated a pretreatment permit for the
discharge of 90,000 gallons of pretreated water per day with effluent monitoring, and control
limitations.
Regulatory Status
RSC has recently completed Phase V of the groundwater assessment program and Addendum
8 to the facility's Post Closure Permit Application. Pending approval of the Permit, RSC will
continue with the groundwater remediation associated with the RCRA regulated groundwater
plume in addition to assessing and remediating the SWMU related groundwater contamination.
-------
REFERENCES
ASTM, 1990, Annual Book of ASTM Standards. Vol. 04.08, American Society of Testing and
Materials
North Carolina Geologic Survey, 1985, "Geologic Map of North Carolina".
Randazzo, F., 1972, Petrography and Stratigraphy of the Carolina Slate Belt, Union
County, North Carolina, pp. 2-32.
Stuckey and Steel, 1953, Geology and Mineral Resources of North Carolina, 35 p.
-------
Christopher C. Cutler
Radiator Specialty Company
600 Radiator Road
Matthews, NC 28105
Chris Cutler is currently employed with Radiator Specialty Company as Environmental
Supervisor. Chris received a B.S. from East Carolina University in Greenville, North
Carolina and is currently pursuing a M.B.A. at Queens College in Charlotte, North
Carolina. While at RSC Chris has assisted in the development and implementation of
numerous industrial compliance activities.
Richard L. Harmon, P.G.
Triangle Environmental, Inc.
4201-J Stuart Andrew Blvd.
Charlotte, NC 28217
Rick Harmon is currently employed with Triangle Environmental, Inc. as a Senior
Hydrogeologist. Rick received a B.A. in Geology from Humboldt State University in
California, and is currently practicing as a licensed geologist in the State of North Carolina
and Delaware with over 9 years experience in hydrogeological site assessments in
conjunction with RCRA, CERCLA and UST regulations. Specific tasks include the design
and implementation of numerous site assessment projects including the use of fracture trace
studies, specialized monitoring well designs, various methods of discrete zone sampling,
and aquifer testing including constant head permeability tests, constant discharge pumping
-------
-------
ABSTRACT
Title: "Sanders Lead Company, Inc., Recovery/Re-use Program for
Inorganic Contaminated Ground Water"
Presented To: EPA/Region IV Conference
In 1988 Sanders Lead Company initiated actions due to RCRA
requirements to identify any constituents of concern (COC's).
at Sanders7 Troy secondary lead smelter. Lead is recovered at
Sanders' secondary smelter from lead-acid batteries and sold
back into lead consuming industries.
Due to past battery breaking methodologies, acid, contaminated
with lead and cadmium leached into soil at the Troy location and
created a plume which was confined to an upper sandy/clay. This
geological area (upper zone) was not a ground water use zone, but
the plume did present a future concern to other sources if
migration was allowed to continue. All waste units were closed,
and through RCRA Permit Conditions, a well recovery program with
monitoring well network was designed, placed, and the plume was
defined. Recovery wells were installed and in 1990 ground water
was recovered at a yield rate of one gallon per minute per ten
(10) units. This water is then re-used in non-potable areas,
contained, treated and discharged to the local POTW.
Data from 1989 through 1993 demonstrates plume management and
control. Details of the automatic system are presented in the
-------
NOTE
•uw	Atajrr it
7j*vcr IT i»« WUtfi' v«
aCOQaTIJ.
-LCfiEUQ-
0 C*tCn*«*0* *01
0 OmCCTM ACTIO) MCLL
— «^fnoHk>n M»o*orr uxi
CONCOtTKATKW llOKflH - TO TAJ. IM l»M
CONCCKflUftON L«0P\XTH - TOTAL Lt«l tH)
CONCENTRATION ISOPLETH MAP
UPPER AQUIFER TOTAL LEAD
COMPARISON OF FIRST QUARTER
1989 AND FIRST QUARTER 1993
(SmIm)
SANDERS LEAD CO. INC.
mOY. ALABAMA
SCALE. 1" =¦ 400'	|**ALW
PREPARED BY:
otmmM mommcG
ym nwn mu «nc
HASfitXi. Vi J721I
(Ij-JIHIUD
-------

NOTE
pui* ao^hd fwom Auouri it ino
«£*Uk. Mifr f uoi iwrr mO
A-S30CUTU.
ifcewo
TTUUIII MIL *
CM OMKMBT
e ortcnvoma xu
Ooowncuvc actoi «u.
¦ > — —	A^MMkATt MOWY LMI
°-0l	. eo«aDmuTHH ooktw - potm. e
®-0i -+A —ooHConiunaK mopitim - tot* cmmuh ihj
CONCENTRATION ISOPLETH MAP
UPPER AQUIFER TOTAL CADMIUM
COMPARISON OF FIRST QUARTER
1989 AND FIRST QUARTER 1993
SANDERS LEAD CO.. INC.
TROY, ALABAMA
SCALE. 1" - 400'
'Atwl-fratl'
PREPARED BY:
-------
(2 f[ DM i urt
—	RcJjcT.t^	x
-------
LlrAD
-------

.J'
• .-.; V'; : .••.	v .•'/
»*	tK« 4 y v	*~» *> . ¦. *.	• .1'.
V ?: ¦; . V*
• v,> '*£¦.¦: *


-; v ^
".'Is- ./' ^ '
ffeV	.	,y
,7^ 7A A
i'-.'- ./¦#?*
$>0h .
• . f V*!		
>V*. .*^4*.
*>, \ '
t '	» *	/,« ^ *

,w#v,
> o	>•
\ ."*, 4*
v

re
V 1L
-------
MANHOLE VAULT
WALL
SEE DETAIL 8
DETAIL A
.1 :|-v. v	•> i' ' vl '
PVC LINER
©
1/4 X 2' CLOSED
CELL NEOPRFNE SPONGE
SC 42-M
NEOPRENE
AOHESIVE
1/4" X 2" STAINLESS
STEEL BATTEN
WASHER
3/8" x 2 1/2
ANCHOR BOLT
STAINLESS ST
BUTYL MASTIC
TAPE 1779
GUNOLE
PVC LINER
DETAIL B
LINER CONNECTION DETAIL
-------
©ELECTRICAL TRENCH DETAIL
FOR CONCRETE SUB
-------
-EXISTING GROUND
SURFACE
ELECTRICAL CONDUIT
(NUMBER AND SIZE AS REQUIRED
FOR 230V ELECTRICAL SERVICE
TO WELL VAULTS)
CONCRETE
CONOUIT
SUPPORT
|-a 4"—j | ¦ 4*—j | ¦ 4*—»j
ELECTRICAL CONDUIT
i\TRENCH CONSTRUCTION DETAILS
TO FOR NON—TRAFFIC AREAS
-------
TICHT JOINT
/Tnpipe connection at detention tank
V37 AND EMERGENCY HOLDING TANK
-------
to wwm «*n HCfcfiff* UM.
PLAN
07 owtwc auemtn sue
TO fWU» «MTO HCUM Tan*.
0PIPE CONNECTION DETAIL AT EDGE
OF EXISTING CONCRETE SLAB
-------
TRENCH CONSTRUCTION DETAIL FOR
DISCHARGE FROM CORRECTIVE ACTION WELLS
CA-1, CA—3, CA-4
-------
-EXISTING GROUND SURFACE

4" PVC SECONDARY CONTAINMENT-
1PE (WHERE REQUIRED)
2" PVC DISCHARGE PIPE-
COMPACTCD'
• BACKFILL
. V CLEAN BEDOiNG.
... . . . SANO...
18* MIN.
\ i i
£ /ITntrench construction detail
3 W FOR NON-TRAFFIC ARE«S
-------
-------
-------
B,
anchor oan (n?.>-
230V ClCCTRlC
PO*TR SUPPLY
r PVC DISCHARGE PIPE
TO PROCESS WATER HOLDING TANK
41' WAitfTW. *• PfttCAir CONCRETE
CCCCMTWC UAM40LT OONC tUPPMTCO
•V 4*' 0 w.
t MIOft (WM TOiM UlUNBPOt ra
- • **¦
X tout Of TM( IKMS MOW W YHCIC OCTAL!
am cmmm out* lug ra* guMTC.
actum rueoan tfT? khminu
-------
i_
-------
/Tnpump placement detail for
V^J CORRECTIVE ACTION WELL
-------
SANDERS LEAD COMPANY, INC.
GROUND WATER RECOVERY PRESENTATION
Presenters: 1) E. Roy Baggett, B.S., M.S., Manager of Environmental
Affairs, Sanders Lead Company, Wiley Sanders Truck Lines, KW
Plastics, Troy, Alabama and KW Plastics, Bakersfield, California
Mr. Baggett has since 1970 been employed in the Environmental
Management field. From 1970 through 1972 Mr. Baggett worked under
an EPA Region IV, Solid Waste Management Act Grant; 1972-1978, Mr.
Baggett worked with the State of Georgia EPA as an inspector and
emergency response coordinator. 1978-1982 - Mr. Baggett worked as
a consultant for various hazardous waste projects. 1982-1988 -
Mr. Baggett was employed in his own private consulting business.
In 1988 to present he came to Sanders Lead as a RCRA Part B
Coordinator and in 1989 was made Manager of Environmental Affairs.
In his twenty three year career, Mr. Baggett has worked with over
250 emergency responses. He has worked with numerous industrial
processes and has sought and obtained numerous operating permits.
At Sanders, a Part B has been issued, along with HSWA attachments
and the corrective action program with over 60 monitoring wells.
Since 1988 data from Sanders Corrective Action Program has
demonstrated plume capture and management.
2) J. Chris Rutherford, Project Coordinator, Sanders Lead Company
Mr. Rutherford graduated in 1987 with a Bachelor of Science degree
while concentrating his studies in the fields of Geology and
Biology. Since 1988 Mr. Rutherford has been employed in the
environmental management and consulting area. From 1988 to 1990,
Mr. Rutherford worked for OGDEN Environmental and Energy Services
Company as~ a site project manager for several environmental
investigations and remediation projects. From 1990 to the
present Mr. Rutherford has worked for Sanders Lead Company as an
environmental project coordinator and hydrogeologist.
3) Mark Hobbs, Branch Manager, OGDEN Environmental and Energy
Services Company
Mr. Hobbs obtained a Bachelor of Science degree from the
University of South Carolina in 1985. Since 1985 Mr. Hobbs has
been employed in the environmental management and consulting
field. From 1985 to the present Mr. Hobbs has worked for OGDEN
Environmental and Energy Services Company as an Environmental
Branch Manager directing the work efforts of 20 plus engineers,
geologists, and technicians in environmental investigations and
remediation projects
STATti PROGRAMS SECTION
WAC,TF P-CGRAMS BRANCH
V.. a.-' . . vi ^ /
¦ PA-i'*v''
-------

Cardinal,
-------
ThermalKEM Inc., Rock Hill, SC - ThermalKEM, SCD 044 442 333
PRESENTATION TITLE: A Case Study of Groundwater Remediation at the ThermalKEM,
Rock Hill, South Carolina Facility.
ABSTRACT: ThermalKEM Inc. is the owner and operator of a hazardous waste treatment
and storage facility located in Rock Hill, SC. On June 28, 1988, EPA issued a Corrective
Action permit pursuant to the Hazardous and Solid Waste Amendments of 1984. The permit
identified three solid waste management units (SWMUs) and ThermalKEM has since
identified three additional areas of concern (AOCs). The contamination at the facility
resulted from the waste and product handling practices of the previous owners. A bedrock
groundwater remediation well (EW-1) has been in operation since 1988 to contain and
remediate dissolved groundwater contamination. This well has also served to contain a fuel
oil pure product plume existing on the site. The pumping regime of EW-1 has been
modified several times to improve remediation capabilities. Several remediation methods
have been investigated or pilot tested to recover fuel oil from the fuel oil product plume. A
second bedrock well (PW-1) was pumped from 1988 until 1992 to remediate an additional
area of groundwater contamination. An additional extraction well is being designed to act
with EW-1 to remediate and prevent potential off-site migration of dissolved or pure phase
groundwater contamination.
PRESENTER: Alice B. Clark, Senior Hydrogeologist, ENSR Consulting and Engineering,
Acton, MA.
(LO^
PRESENTER BIOGRAPHY: Alice B. Clark received her B.A. in Engineering Sciences from
Harvard University and her M.S. in CivtJ Engineering from the Water Resources and
Environmental Engineering Division, M.l.T She has eight years of experience in the
hazardous waste field and is currently employed by ENSR Consulting and Engineering in
Acton, MA. Her major areas of expertise include hydrogeologic characterization,
hydrogeologic and contaminant modeling, and spatial and temporal statistical applications
to hazardous waste investigations. Ms. Clarfc may be contacted at ENSR Consulting and
-------
A Case Study of Groundwater Remediation
at the ThermalKEM, Rock Hill, South Carolina Facility
by
Alice B. Clark1, Lisa J. Wolf1, John Bierschenk 1, Ralph S. Baker1, Arthur Lazarus
Linda Eickhoff1, Janice Baker2, Luis DeAndino3 and William Zeigler3
The purpose of this paper is to present the groundwater remediation technologies pilot tested
and implemented at the ThermalKEM, Rock Hill, South Carolina Facility. Groundwater
contamination at the site is believed to be the result of poor waste storage and disposal practices
of previous site owners.
I. Description of Facility
ThermalKEM, Inc. (ThermalKEM) is the owner and operator of a hazardous waste treatment and
storage facility located in Rock Hill, SC (Figures 1 and 2). The facility consists of a RCRA-
permitted hazardous waste incinerator, hazardous waste tanks and container storage units. All
waste receipt, mixing and storage, and incineration activities are conducted in roofed buildings
with concrete floors and walls. All processing areas have fugitive emission controls. Secondary
containment is employed for all hazardous waste operations with the exception of ash storage.
Ash roll-off containers are not used to store waste with free liquids, therefore, secondary
containment is not required. Ash roll-off containers are kept covered and are inspected daily.
All tanks are above ground and nitrogen blanketed, and a leak detection al- n system is in place
for tanks and feed lines. The current operations are not thought to represent potential sources
of contamination to the environment.
The current owners have operated the site since 1983. Quality Drum Company and Industrial
Chemical Company owned and operated the site from 1966 to 1983. Prior to 1966, the site was
undeveloped. During its period of operation, Quality Drum/Industrial Chemical received spent
solvents from off-site facilities and recovered these solvents by distillation. Still bottoms
generated by the solvent recovery operations were sent off-site to a nearby landfill. In 1980,
shortly after the landfill was closed, Quality Drum/Industrial Chemical installed a hazardous waste
incinerator at the site for still bottoms treatment. In addition to this incinerator, Quality
Drum/Industrial Chemical also staged and stored drums containing hazardous waste on the
ground and stored hazardous waste in above-ground tanks. These tanks were not designed to
Subpart J design standards, but consisted of any salvage tanks that Industrial Chemical could
find (i.e. wrecked tank wagons, etc.). In addition to these units, stormwater runoff from tank and
container storage areas, and non-contact cooling water from stills, drained into an unlined ditch
1	ENSR Consulting and Engineering, Acton, MA
2	ENSR Consulting and Engineering, Raleigh, SC
3	ThermalKEM Inc., Rock Hill, SC
-------
on the site, referred to herein as the Solvent Recovery Containment Ditch. Also, fuel oil was
stored in two underground tanks.
In May 1983, Stablex Inc. acquired the site. At that time there were some 26,000 drums of waste
staged directly on the ground on the site and 200,000 gallons of waste stored in above-ground
tanks, which may have been structurally deficient. The Solvent Recovery Containment Ditch was
also still in use.
In December 1985, Stablex Inc. was purchased by the American NuKEM Corporation and
changed its name to ThermalKEM Inc.
II. Site Hydrogeology
Geology
The ThermalKEM facility is underlain by two primary hydrologic units: gabbroic bedrock, and
surficial soils derived from the gabbro and/or alluvium (Figure 3). Weathering of the gabbro has
yielded a dense, cohesionless subsoil, saprolite, which immediately overlies the gabbro.
Advanced, in-situ weathering of the saprolite has, in turn, produced a cohesive silty clay or
clayey fine sand. The transition between the cohesive soil and saprolite, and between the
saprolite and bedrock is gradational. Sand and gravel inclusions evident within the cohesive
subsoil may reflect localized heterogeneity of the gabbro or reworking of the cohesive soils with
alluvial deposits.
Fill material, topsoil and the cohesive soils overlying the saprolite, classified herein as residual
soils, range from 0 to approximately 21 feet in thickness. The saprolite layer is comprised of an
upper and slightly denser lower unit, and ranges in thickness from less than 10 feet to
approximately 22 feet. Degree of bedrock fracturing is variable across the site, with primary
structural strike and dip direction being north-northeast and south-southeast, respectively, and
fractures reportedly extending to a depth of 135 feet below grade (NTH, 1987). Across the site,
the water table (10 to 20 feet deep) may intersect the cohesive residual soils, the saprolite or the
bedrock.
Aquifer Hydraulic Properties and Characteristics
The fine, siity clay and siity sand soils which characterize the unconsolidated material also define
its hydraulic properties. The saprolite has a high porosity (27 to 40 percent) and, therefore, high
capacity for water storage. Estimates of hydraulic conductivity (K) for the saprolite are
approximately 3 to 4 x 10"* centimeters/second (cm/s)(NTH, 1984; SEC, 1988).
The bedrock is approximately an order of magnitude more permeable than the saprolite, exhibits
porosities ranging from 10 to 30 percent and is highly fractured. Fractures are likely to be the
primary water-conducting voids in the bedrock. Estimates of K for the bedrock are approximately
2 to 2.5 x 10"3 cm/s (NTH, 1984; ENSR, 1992b). In general, fracturing of the bedrock in the
vicinity of the ThermalKEM site allows for water production of 5 to 10 gpm in shallow wells (ie.,
50-100 feet deep) and 25 to 50 gpm in deeper wells (i.e. 400-700 feet deep)(NTH, 1984).
However, individual well production is highly dependent upon well placement relative to fractures
-------
(i.e. EW-1, open in the bedrock from approximately 30 to 64 feet deep, can produce 74 gpm
(SEC, 1988)).
Groundwater Flow Patterns
The ThermalKEM site is located near the confluence of Fishing Creek and Wildcat Creek. The
groundwater flow patterns at the ThermalKEM site can be described on both the regional and
local scale. Surface topography indicates that regional groundwater flow is to the south, toward
and along Wildcat and Fishing Creeks. Where unaffected by on-site pumping, localized
groundwater flow in the saprolite is primarily from west to east towards Wildcat Creek (Figure 4).
Pumping from a bedrock remediation well (EW-1), located approximately at the middle of the
active portion of the site, at approximately 15 gpm caused a capture zone in the saprolite which
laterally extended from MW-118 to the southwest toward MW-103 to the northeast. The
downgradient stagnation point of the capture zone in the saprolite extended from wells BP-1,
MW-113, and MW-114 (Figure 5).
Where unaffected by EW-1 pumping, bedrock groundwater flow is generally to the east toward
Wildcat Creek and upward; pumping from the bedrock has reversed the natural vertical gradient
within the well's zone of capture. With EW-1 pumping at approximately 15 gpm, the cone of
depression extended from as far south as MW-105, potentially as far north as MW-100, and
potentially as far downgradient as well MW-115B, but certainly to well MW-113B. The observed
north-south orientation of the cone reflects preferred fracture strike direction (Figure 6). Under
recent elevated pumping rates of approximately 20 to 25 gpm at EW-1, data indicate that the
capture zone likely extends downgradient to MW-115B in the bedrock.
Groundwater/Surface Water Interaction
The nature of groundwater/surface water interaction along Wildcat Creek was evaluated using
stream gauging, water level monitoring, piezometric maps, and cross-sectional flownets.
Analysis of stream gauging data suggests that groundwater discharges to the creek. Connection
between the unconsolidated aquifer and Wildcat Creek is further indicated by unconsolidated
piezometric surface maps and flownets which show that shallow groundwater discharges to
Wildcat Creek. -Deeper groundwater may flow beneath and along the creek for some distance
before eventually discharging to the creek, in any case, the lateral vicinity of the creek is a
groundwater divide and groundwater from the ThermalKEM site does not migrate significantly
to the far side of Wildcat Creek.
III. Description of Solid Waste Management Units and Areas of Concern
On June 28, 1988 the EPA issued a Corrective Action permit pursuant to the Hazardous and
Solid Waste Amendments of 1984. The permit identified three solid waste management units
(SWMUs) requiring investigation and possible corrective action. Also, through the course of
investigation conducted by ThermalKEM at the facility since issuance of the Corrective Action
permit, ThermalKEM has identified three additional areas of concern (AOCs) that were
investigated. These SWMUs and AOCs (Figures 7, 8 and 9) are referred to herein as:
•	Incinerator Building Sump (SWMU 8)
•	Truck Wash Station and Sump (SWMU 19)
-------
•	Container Storage Area (SWMU 11)
•	Solvent Recovery Containment Ditch (AOC)
•	Fuel Oil Area (AOC)
•	Old Burn Pits (AOC)
Only the Solvent Recovery Containment Ditch, the Fuel Oil Area, and the Old Burn Pits represent
areas of potential groundwater contamination.
Solvent Recovery Containment Ditch
The Solvent Recovery Containment Ditch was located just south of the existing drum process
and blend tank buildings. This ditch had been created by Industrial Chemical, presumably to
contain spills or leaks from storage tanks used in their solvent recovery operation. The ditch was
approximately 2 feet wide, 2 feet deep and 100 feet long. In 1983, visibly contaminated soils
were removed from the ditch and the ditch was filled with clean soil. Groundwater investigations
in the vicinity of the ditch confirm the presence of volatile organic constituents (VOCs) in the
groundwater as well as the possibility of light, non-aqueous phase liquids (LNAPLs).
Fuel Oil Area
Fuel oil contamination was first detected in the groundwater in piezometer P-2 during a routine
water level reading in June 1990. The source of this fuel oil could have been a leaking fuel line
which was promptly detected and repaired the previous year, or could be associated with two
old underground fuel oil tanks operated under Industrial Chemical's ownership of the site. The
tanks were in the vicinity of the existing fuel oil plume but were removed soon after Stablex's
purchase of the site.
Old Burn Pits
In the 1960's or early 1970's, Industrial Chemical disposed of still bottoms by open pit burning
in what is herein referred to as the Old Burn Pits. In 1985, via a sitewide magnetometer survey,
Stablex discovered buried remnants of drums and materials in the Old Bum Pit area. Stablex
immediately remediated the Old Bum Pits, under S.C. DHEC supervision, and back-filled the pits
with clean soils.
In April, 1985, the first groundwater samples from a newly installed monitoring well (MW-100),
upgradient of the Old Burn Pits, in the vicinity of the office building, began showing elevated
levels of trichloroethene (TCE), tetrachloroethene (PCE) and other solvents. Industrial Chemical
had operated Production Well 1 (PW-1), also located near the office building, at a rate of
approximately 25 gpm. The contamination at MW-100 may have been caused by Industrial
Chemical's pumping of groundwater at Production Well 1 (PW-1) prior to Stablex's purchase of
the site and remediation of the Old Bum Pits. In 1988, ThermalKEM converted PW-1 into a
remediation well and PW-1 continued to pump at low rates (0 to 5 gpm) to remediate the
contamination at MW-100. Pumping at PW-1 was discontinued in 1992 when concentrations in
MW-100 showed a significant decrease. The discontinuation of pumping at PW-1 also allowed
for increased pumping at the other site remediation well (EW-1), as discussed below.
-------
By June 1990, constituents similar to those detected in MW-100 appeared in downgradient wells
MW-103 and BP-1. The similarity of constituents detected in wells BP-1, MW-103 and MW-100
suggests that the contamination in all three wells may have come from the Old Burn Pits.
IV. Scope and Extent of Groundwater Contamination
The discovery of groundwater contamination on the site resulted from a series of subsurface
hydrogeologic investigations which Stablex voluntarily initiated upon acquisition of the site. Since
1983, groundwater monitoring has helped to identify the extent of contamination and constituents
present at each area of concern, as outlined below.
Solvent Recovery Containment Ditch
Wells OB-110A, 110B, 8A, MW-104, 113B and 115B show volatile constituents expected to be
associated with the Solvent Recovery Containment Ditch. As the history of solvent disposal at
the Solvent Recovery Containment Ditch would indicate, the constituents discovered in this area
include several volatile organic compounds (VOC's), the class of compound which characterizes
most solvents. The highest concentration levels and frequency of detection occur at OB-110A,
the well nearest the old ditch. OB-8A shows slightly lower levels of the same constituents found
at OB-110A. Metals were also detected in the Solvent Recovery Containment Ditch area.
Given all of the data, it is likely that pure phase constituents exist in the Solvent Recovery
Containment Ditch area. Light Non-Aqueous Phase Liquid (LNAPL) is likely present at OB-110A,
but is not expected to extend to EW-1 based on EW-1 recovery data.
The constituent levels in the wells provide an indication of the extent of the dissolved
contaminant plume. The presence of the more mobile VOC's at MW-104 and recently at MW-
113B and MW-115B, indicates that the edge of the plume may have migrated downgradient to
MW-104 in the saproiite portion of the aquifer and to MW-115B in the shallow bedrock system.
Recent total volatiles concentrations at the three monitoring wells are similar (0.1 to 0.8 parts per
million (ppm)). No site-related contamination nas been detected in the creek. Figures 10 and
11 shows total volatile organics through time at welte OB-8A, OB-110A, MW-104,113B and 115B.
The remediation operation of EW-1, whtcn « expected to contain the Solvent Recovery
Containment Ditch plume, is discussed in more deuut betow.
Fuel Oil Area
Monitoring of the fuel oil area has focused on defining the extent and thickness of the pure
product plume. To date, pure product has been recorded in wells P-2, TW-1, PW-1A, PW-2A,
OB-11, 12 and 13, and OB-21, 22 and 23.
Based on water and product level measurements and product baildown tests performed in the
wells containing floating product, the minimum and maximum lateral extent of the floating
product has been estimated, as shown in Figure 12.
At downgradient wells MW-111 through 114, screening level analyses of groundwater samples
are performed regularly to ensure the dissolved plume has not migrated this far downgradient.
-------
At present, the pure product fuel oil and likely the dissolved plume are effectively contained by
the pumping of extraction well EW-1. The remediation of the pure product fuel oil is discussed
in more detail below.
Old Burn Pits
The primary VOC plume constituents detected at monitoring wells MW-100, MW-103, BP-1 and
BP-1B, near the Old Burn Pits, are 1,1,1-TCA, PCE, and their degradation products (1,1-DCE,
1,1-DCA, trans-1,2-DCE, TCE, methylene chloride and vinyl chloride). TCE has shown the
highest concentration, appearing in MW-100 as high as 41 parts per million (ppm) (greater than
1% of its pure-phase solubility in water), possibly indicative of DNAPL. However, TCE
concentrations in MW-100 have declined to below 1% of solubility since April 1987, and have
consistently declined further since then. Figure 13 shows the change in TCE concentrations with
time at MW-100. Due to the decline in concentration levels at MW-100, pumping from PW-1,
which was remediating the MW-100 area, was discontinued in 1992. This also allowed for
increased pumping at remediation well EW-1, as discussed below.
While BP-1 and BP-1B represent the most downgradient monitoring points for this source,
constituent levels at BP-1 and BP-1 B are generally less than 100 and 600 ppb, respectively. No
site-related contamination has been detected in the creek.
V. Groundwater Remediation Systems
Initial Remediation
Upon acquisition of the site in May 1983, Stablex cleaned and closed the Solvent Recovery
Containment Ditch by October 1983. Underground fuel oil tanks were removed by June 1984.
Waste in other above ground tanks was cleaned out by September 1984. Surface soil,
contaminated by poor waste storage practices, was removed by October 1983. RCRA facility
standards were implemented, such as appropriate operating records, a waste analysis plan, an
inspection program and proper training. Additionally, the incinerator was upgraded to 40 CFR
Subpart O standards.
A subsurface investigation, which consisted of a comprehensive magnotometer survey over the
entire site, was performed in 1985 (NTH, 1985a). This survey revealed two areas of buried
materials. A total of eight intact buried drums were found and excavated in an area near MW-
103, and four old burn pits were excavated, cleaned to background soils, and filled with clean
soil. Both operations were witnessed and approved by SC DHEC.
Extraction Well EW-1
In 1986 and 1987 studies were conducted to design an extraction well (EW-1) to contain and
remediate the groundwater plume emanating from the Solvent Recovery Containment Ditch area.
Extraction well EW-1 was placed on-line in July, 1988. EW-1 is a 64 foot deep open-hole
bedrock well, located downgradient of the Solvent Recovery Containment Ditch as shown in
Figure 8. Investigations of the extraction well's hydraulic effects were conducted in late 1988.
A pump test at 20 gpm showed EW-1 to have a capture zone including all groundwater known
-------
at that time to be affected by the Solvent Recovery Containment- Ditch and likely a portion of
Wildcat Creek (Figure 14).
EW-1 was pumped at between 5 and 18 gpm from April, 1989 until March, 1992, when its
pumping rate was adjusted to a fairly constant 14 gpm. Between May, 1991 and March, 1992,
EW-1 was operated consistently in the 7 to 10 gpm range. Water levels throughout the site were
observed on June 25, 1991 to determine the cone of influence of EW-1 at the 7 to 10 gpm
pumping rate (Figure 15).
Groundwater extracted from both EW-1 and PW-1 (described below) is treated in the on-site
groundwater treatment plant by chemical precipitation of iron and manganese, sand filtration of
inorganics, and removal of organics by three activated carbon canisters connected in series.
The treated groundwater is then diverted to the facility wastewater treatment process unit where
it is used for scrubber water. The volatile removal efficiency is greater than 99%.
Production Well PW-1
PW-1, an open-hole bedrock well about 150 feet deep, was installed in 1979 and was operated
as a production well with an estimated pumping rate of 25 gpm until 1988 when a city water
supply line was connected to the site (Figure 9). When contamination was discovered in nearby
well MW-100, PW-1 was used to withdraw contaminated groundwater from the vicinity for
diversion to the on-site wastewater treatment plant. During use as an extraction well, PW-1 was
pumped at a rate varying from 0 to 5 gpm. PW-1 was taken off-line in 1992, because
concentrations at MW-100 had decreased to near or below detection levels. This also allowed
the pumping rate at EW-1 to be increased to 25 gpm, to extend its zone of capture, as described
below.
Fuel Oil Area
Initial remediation efforts in the fuel oil area focused on an extraction well system located in the
saprolite formation where the fuel oil was believed primarily to reside. In February. 1991, a pilot
purge well study was conducted to investigate the feasibility of removing the product via shallow
pumping wells (Kunkle, 1991). Two purge wells were installed, PW-1A and PW-2A (Figure 9),
and screened at depths of 10 to 20 feet and 20 to 30 feet, respectively, and pumped at between
4 and 8 gpm. Capture zones of 5 and 20 feet, respectively, were produced at the two purge
wells. It was therefore concluded that more effective methods of removing product from the
formation might be available, and alternative strategies were pursued, as described below.
VI. Modified Remediation Systems
Fuel Oil Area Remediation
Since the pilot purge wells alone did not remove significant quantities of fuel oil or induce large
capture zones, enhanced methods were pilot tested. In August through October, 1992, three
phases of recovery were explored (Figure 16) (ENSR, 1993). In Phase I, a total fluids pump was
employed at well VE-1, an extraction well specifically installed for this pilot study and screened
for 25 feet across the water table. The objective of Phase I was to measure the pre-
enhancement rate of total fluid (product and water) recovery from the extraction well.
-------
In Phase II, VE-1 was sealed and vacuum was applied to it providing a combined soil vacuum
extraction (SVE) and total fluid recovery system. The objectives were to assess whether vapor
extraction would enhance liquid recovery, and to evaluate the effectiveness of a soil vapor
extraction (SVE) system to reduce soil gas concentrations.
In Phase III, a multiple well recovery approach was attempted. Existing observation wells having
different diameters, screened at different intervals and located in different places within the area
were employed singly and in combination for fluids recovery. The objective was to use the
existing wells to show empirically how well diameter, location and screened interval might affect
product recovery, and to predict total liquid flow rate from these wells for treatment system
design.
Based upon the objectives and results of the pilot test, ENSR concluded that application of a
vacuum via a soil vapor extraction (SVE) system had a measurable positive effect on total fluid
recovery but would not significantly enhance the fuel oil recovery operation. Comparison of
liquid recovery before vaccum enhancement (< 1 gallon per day (gpd)) and after (average of 1.75
gpd) reflects a significant (42%) increase. However, product recovery was greatest from the
multiple well test (average of 12 gpd).
Effectiveness of soil vapor extraction (SVE) to reduce soil gas concentrations was demonstrated,
but limited by the uneven influence of applied vacuum within the saprolite formation.
As stated above, significant fuel oil recovery was obtained from the multi-well test, representing
one day of continuous pumping from each well. The majority of the product and water
recovered during the pilot test was probably removed from those wells which intersect secondary
fractures. During long-term pumping, a portion of the free product remaining in fractures and
larger pores of the saprolite could migrate slowly into the larger fractures and might ultimately
be recovered.
Effectiveness of vacuum on enhancement of both fuel oil recovery and soil vapor extraction is
influenced strongly by the fracture and pore size distribution within the saprolite, which control
its permeability and liquid retention properties.
EW-1 Operation and Dissolved Constituent Remediation
As of August, 1992, when ThermalKEM submitted a revised RFI Work Plan to EPA (ENSR,
1992b), EW-1 (pumping at approximately 14 gpm) was thought to be successfully preventing off-
site migration of constituents from the Solvent Recovery Containment Ditch area and also
remediating the constituent levels. EW-1 's cone of depression was also preventing the fuel oil
product plume from migrating downgradient, and four monitoring wells (approximately 200 feet
upgradient of Wildcat Creek) monitored monthly showed that the dissolved fuel oil plume had
not migrated beyond them.
In June, 1992 three shallow bedrock wells (30 to 45 feet deep) and one saprolite well (10 feet
deep) were installed near Wildcat Creek to further delineate the extent of dissolved contamination
and define more fully the three-dimensional groundwater flow pattern near the creek. Bedrock
monitoring well BP-1B was installed, adjacent to previously existing saprolite well BP-1, in the
area downgradient of the old burn pits. Bedrock monitoring well MW-113B was installed,
-------
adjacent to existing saprolite well MW-113, downgradient of the Solvent Recovery Containment
Ditch area approximately 250 feet upgradient of Wildcat Creek. Bedrock monitoring well MW-
115B and saprolite well MW-115 were installed downgradient of the Solvent Recovery
Containment Ditch area approximately 75 feet upgradient of Wildcat Creek (Figure 8). Quarterly
monitoring of these wells in July, 1992, October, 1992, and January, 1993 revealed positive
results for organics in all three bedrock wells and the previously existing saprolite well BP-1, but
not in saprolite wells MW-113 and MW-115. Excluding one value of 1.4 ppm for trans-1,2-
Dichloroethene at BP-1B in July, 1992, concentrations at BP-1B were all below 600 ppb;
concentrations at BP-1 were all below 100 ppb; concentrations at MW-113B and MW-115B were
all below 200 ppb. The EPA was notified in writing of these results in February, 1993. The
constituent levels detected in these downgradient wells are one to two orders of magnitude lower
than the levels recently detected near the Solvent Recovery Containment Ditch. Though the
downgradient extent of the Solvent Recovery Containment Ditch plume is yet undetermined, no
contamination which is expected to have migrated from the site has been detected in the creek
(ENSR, 1992b).
In response to the observation of dissolved organic constituents in the groundwater near Wildcat
Creek, ThermalKEM increased the pumping rate at EW-1 to approximately 35 gpm in April, 1993,
in order to increase the capture zone and prevent migration to the creek. Water level data
collected at this pumping rate and compared to data at lower rates indicate that MW-115B and
BP-1B were likely included in EW-1's capture zone at this elevated pumping rate.
In July 1993,10 to 15 feet of fuel oil floating product from the Fuel Oil Area were detected in EW-
1 and its pumping rate was reduced to approximately 20 to 25 gpm. At this time, a decrease
in product entering the treatment system was observed ar.-i the well continued to pump with
satisfactory operation of the treatment system. Water level data indicate that EW-1's capture
zone at 20 to 25 gpm likely encompasses MW-115B and BP-1B.
integrated Site-wide Remediation
Given the desire for certainty of complete dissolved and pure product constituent capture and
for an efficient remediation system, a site-wide remediation program with components which
interact well withfeach other has been developed. The conceptual plan includes an additional
extraction well to be installed in the shallow bedrock near the Old Bum Pits. This extraction well
would be operated along with EW-1 to protect the creek from the migration of dissolved
constituents from all of the areas of potential groundwater contamination discussed above
(Figure 17). EW-1 would also continue to prevent the fuel oil product plume from migrating to
the creek.
VII. Results of Remediation
Extraction Well EW-1
Since commencement of operation, EW-1 's effectiveness has been monitored through analysis
of water samples taken from the influent to the groundwater treatment system. Influent
concentrations to the on-site wastewater treatment system from PW-1 and EW-1 show that the
system is effectively recovering contaminated groundwater. Even while PW-1 was in operation,
-------
the majority of the influent concentration was attributed to EW-1, because of its higher pumping
rate.
Figure 18 shows the treatment system's influent concentrations of total volatiles. Total volatiles
have been recovered at three ppm, on average. The recovered concentration levels do not
appear to be decreasing through time, which is attributed to the high contaminant concentrations
which still exist immediately upgradient of the pumping well (EW-1).
Concentrations of total volatiles at the wells impacted by the Solvent Recovery Containment
Ditch (OB-8A, 110A, MW-104, 113B and 115B) are shown in Figures 10 and 11. Figure 10
indicates that total volatile levels have decreased slightly over time near the source (OB-8A, OB-
110A). MW-104, slightly farther downgradient, had shown the same trend, but has not show any
consistent decline within the last four quarters of monitoring. Figure 11 shows that contaminants
have reached MW-113B and 115B at relatively low levels.
Production Well PW-1
As described above, PW-1 was operated until contaminant levels in well MW-100 were near or
below detection limits. At that point PW-1 was taken out of operation. A time history of TCE
concentrations at MW-100 is shown in Figure 13, demonstrating the steady decline of the
contaminant of highest concentration and greatest concern.
Fuel Oil Release
It is believed that operation of extraction well EW-1 is effectively halting migration of both pure
and dissolved fuel oil away from the Fuel Oil Area. Pilot purge wells PW-1 A and PW-2A did not
exhibit large capture zones. Phases I through III of the ensuing fuel oil recovery pilot study,
described above, removed almost 300 gallons of fuel oil over a three week period. ThermalKEM
is currently pursuing more effective remedial alternatives for removing larger quantities of the
resident product, as described above, while continuing to pump EW-1 to contain the product,
and continuing to monitor for potential off-site migration (Figure 19).
VIII. Conclusion
In summary, EW-1 has been successfully preventing off-site migration of constituents from the
Solvent Recovery Containment Ditch area travelling m me upper, saprolite aquifer (Figure 5), and,
as discussed above, it is likely that current EW-1 operation is also preventing off-site migration
of the Solvent Recovery Containment Ditch contaminant plume from the bedrock aquifer (Figure
6). EW-1 has also been successfully removing constituents at a consistent rate (Figure 18).
Finally, it is believed that operation of extraction wefl EW-1 is effectively halting migration of both
pure and dissolved fuel oil away from the Fuel Oil Area.
DC Costs of Remediation
It is important for industrial site managers, agency personnel and consultants to observe the
relative costs of various remedial techniques and various remediation approaches. When the
costs, benefits and consequences of the remedial activities at a site are presented together, all
-------
parties gain important information and experience. Table 1 shows the incurred remediation costs
at the site.
X.	Acknowledgements
The authors wish to acknowledge Mark B. Sweatman, G.R. Kunkle and Associates, who
managed most of the site investigation and remediation prior to 1988. Also, we thank Jeff Munic,
ENSR Consulting and Engineering, for his contribution to the 1992 Fuel Oil Recovery Pilot Study.
XI.	References
NTH, 1983. Report on Geotechnical Investigation. Neyer, Tiseo & Hindo, Ltd., Farmington Hills,
Michigan.
NTH, 1984. Report on Preliminary Hvdrooeologic Investigation. Neyer, Tiseo & Hindo, Ltd.,
Farmington Hills, Michigan.
NTH, 1985a. Report on Geophysical Investigation. Neyer, Tiseo & Hindo, Ltd., Farmington Hills,
Michigan.
NTH, 1985b. Report on Installation. Neyer, Tiseo & Hindo, Ltd., Farmington Hills, Michigan.
NTH, 1986. Groundwater Recovery System-Conceptual Design Field Investigation. Proposal,
Neyer, Tiseo & Hindo, Ltd., Farmington Hills, Michigan.
NTH, 1987. Conceptual Purge Well Design Study. Neyer, Tiseo & Hindo, Ltd., Farmington Hills,
Michigan.
SEC, 1988a. Monitoring Well Construction Report. Sirrine Environmental, Greenville, S.C.
SEC, 1988b. Initial Evaluation of Purge Well and Aguifer Conditions. Sirrine Environmental,
Greenville, S.C.
GRK&A, 1991a. Hvdrooeolooic Study Report (Preliminary RFh of Fuel Oil Contamination. G.R.
Kunkle and Associates, Inc., Brighton, Michigan
GRK&A, 1991b. Report on Pilot Purge Study Well Installation and Groundwater Quality. G.R.
Kunkle and Associates, Inc., Brighton, Michigan.
GRK&A, 1991c. Interim Corrective Measures Study (ICMS). G.R. Kunkle and Associates, Inc.,
Brighton, Michigan.
GRK&A, 1992. Letter Report on Lineament Survey for the Area Surrounding the ThermalKEM.
Inc. Facility. G.R. Kunkle and Associates, Inc., Brighton, Michigan.
ENSR, 1991. Soil Sampling Analysis and Evaluation at the Container Storage Area. Rock Hill.
SC. ENSR Consulting and Engineering, Acton, Massachusetts.
-------
ENSR, 1992a. Letter Report on Characterization of Background Soil for Investigation of
Container Storage Area. ThermalKEM. Rock Hill. SC. ENSR Consulting and Engineering, Acton,
Massachusetts.
ENSR, 1992b. RCRA Facility Investigation WorkPlan. ENSR Consulting and Engineering, Acton,
Massachusetts.
ENSR, 1993. Fuel Oil Area Remediation Pilot Study. ENSR Consulting and Engineering, Acton,
Massachusetts.
Baker, R.S. and J. Bierschenk, 1993. Vacuum-enhanced Recovery of Water and NAPL: Concept
and Field Test. Presented at Eighth Annual Conference on Contaminated Soils, Univ. of MA,
Amherst, MA, Sept. 20-23, 1993 (manuscript in review).
Cherry, et al., 1991. Identification of DNAPL Sites: An Eleven Point Approach.
US EPA, 1983. Soils Levels.
USGS 7.5 Minute Topographic Map, Edgemoor Quadrangle, South Carolina.
USGS 7.5 Minute Topographic Map, Rock Hill West Quadrangle, South Carolina.
-------
8L09156
A Case Study of Groundwater
Remediation/Stabilization at the
ThermalKEM, Rock Hill, SC Facility
Presented By: Alice Clark
in
ENSR Consulting and Engineering
-------
8193164
2
-------
SUB 156
3
Figure 2. ThermalKEM Site and
Vicinity, Rock Hill, SC
Seureo:US.QS 7S nemos artastfaoognpfitc) quad at
RockHBJWoa-Soulli ^anitta-YerttCa.
Ia--. : '-hr-ad-
1 mito

l-h'i	rT
Sceto
-------
8103156
Figure 3. Geologic Profile B - B'
4
-------
950812A
IfCFXP
yw-i07
UNCONSOUOATED ¥CLL
UW-106
BEDROCK ttELL
\ MttMOl
HYBRID
W-1
©	UNKNOKN
1 	SSO	 P1EZOUETWC ELEVATION CONTOUft (ft US.)
CONTOUR INTERVAL - I FOOT
200 	0	200
APPROXIMATE SCALE IN FEET
1" = 200-0"
400
FIGURE 4
PIEZOMETRIC SURFACE
5-3-85 BACKGROUND
THERMALKEM FACILITY
ROCK HILL, SC
-------
EXPIANAT10N
tPSOCJA) CREEK SURFACE WATER ELEVATION ON 7/8/02
—O GROUNDWATER now DUTCCTtON
EQUP01EN1KMAL CONTOUR
natM GROUNDWATER ELEVATION (FT USL) MEASURED ON
tauv 7/8^2, 12-1100, IN SAPRCUTE UONITORMC WEIL
CONTOUR M1ERVA1. - 0.25 FT
• FREE PRODUCT PRESENT M MOM TORINO WEU
. WATER l£\«. (FT US.) MEASURED AT HUE OTHER
THAN WITHIN 12-1106 ON 7/«/82
'	CAPTURE ZONE UMT
»-4JJ dk CREEK DISCHARGE MEASURES ALONO STREAM OAUOtNO TRANSECTS
B - BORING
BP - BURN PfT
UW - MONITORING WEU
OB - OBSERVATION WEU
P - PIEZOMETER
PW - PURGE WELL
JW - TEST WELL
W - WELL
WP - WEU POINT
~ - BEDROCK WEU
• - UNCONSOLIDATED SAPROUTE WELL
0 - HYBRID OR UNCLASSIFIED WELLS,
m - ABANDONED BORING LOCATION
SCALE IN FEET
1" = 200'-0"
\1 // FIGURE 5
\' PIEZOMETRIC SURFACE IN
UNCONSOLIDATED SAPROLITE
IBM—OBPiP *p~ ON WP \ !, WELLS ON 7/6/92
""	#/tHERMALKEM FACILITY
ma * im oomcsmcm
sor of arttnc.
fUV. - 527 §7
-------
950814A
Mr-tat A
(ana)
0MMO7

EWLANATMN
CREEK SURFACE WATER ELEVATION ON 7/I/V2
	— GROUNDWATER aOB DIRECTION
wn egupotenhonal contour
GROUNDWATER ELEVATION (FT US.) MEASURED ON
isuuv j/f/ti. 12-13:00. M SAPHOUTt U0MT0RMG «EII
CONTOUR WTEHVAL - 0.2S FT
• FHEE PRODUCT PRESENT M M0MT0RM0 «EU
. WATER LE\CL (FT USU) MEASURED AT HUE OTHER
THAN WTHM 12-13:00 ON 7/B/B2
i CAPTURE ZONE UttT
»-*« dk CREEK DISCHARGE MEASURED ALCNO STREAM OAUONO TRANSECTS
200
MSStE.
B - BORING
BP - BURN PIT
UW - UONITORING WELL
OB - OBSERVATION WEIL
P - PIEZOMETER
PW - PURGE WELL
TW - JEST WELL
W - WELL
WP - WELL POINT
A - BEDROCK WELL
• - UNCONSOLIDATED SAPROUTE WELL
6 - HYBRID OR UNCLASSIFIED WELLS
¦ - ABANDONED BORING LOCATION
Juxri) •
(no a)
Mr-ioi
LSTICAM CAUCUS
IHAKSCCT I
0
200
400
SCALE IN FEET
1" = 200-0"
nv-oous "q~ on r»
wa - lot oomam
cuv. - u;i7
\ // FIGURE 6
PPIEZOMETRIC SURFACE IN
1 ON BEDROCK WELLS
ON 7/6/92
MALKEM FACILITY
OCK HILL, SC
-------
S1931M
Solid Waste Management Units (SWMUs)
•	Incinerator Building Sump (SWMU 8)
•Truck Wash Station and Sump (SWMU 19)
-------
Areas of Concern (AOCs)
•	Solvent Recovery Containment Ditch
•	Fuel Oil Area
-------
LEGEND
B - BORING
BP - BURN PIT
UN - MONITORING WELL
OB - OBSERVATION WELL
p - ptczoucrm
PW - PURGE wax
TW - TEST WELL
W - WELL
WP - WELL POINT
A - BEDROCK WELL
• - UNCONSOUMTEO SAPROUTE WEU
9 - HYBRID Off UNCLASSIFIED WELLS
¦ - ABANDONED BORING LOCATION

y
IBM-CHISELED ON TOP
WALL - l£FT DOWNSTREAM
9DC OF BRIOGE
ELEV. - 127.87
200
0
200
400
SCALE IN FEET
1" = 200'—0"
FIGURE 7
SITE PLAN
THERMALKEM FACILITY
-------
950807A
LEGEND
B - BORING
BP - BURN PIT
UW - UONITORtNC WELL
OB - OBSERVATION WEIL
P - PIEZOMETER
PW - PURGE wax
TW - TEST mm
W - WELL
WP - WELL POINT
A - BEDROCK WELL
• - UNCONSOLCATED 5APROUTE WELL
9 - HYBRID OR UNCLASSIFIED WELLS
m - AP TD BORING LOCATION


200
0
200
400
SCALE IN FEET
1" = 200-0"
FIGURE 8
SWMUe
THERMALK£M FACILITY
-------
95080BA
0 - BORING
BP - BURN PIT
Mr - yOMTOfSNG WELL
OB - OBSERVATION KtU
P - PIEZOMETER
PW - rVRCE WELL
TW - TEST WELL
w - t*ru
WP - WELL POINT
A - BEDROCK WELL
• - UHCONSOUtMTEV SAPROUTE WELL
$ - HYBRID OR UNCLASSflED WELLS
¦ - ABANDONED BORING LOCATION
200
0
200
400
SCALE IN FEET
1" = 200-0"
FIGURE 9
AOCs
THERMALKEM FACILITY
-------
Extent of Contamination
•	Solvent Recovery Containment Ditch
•	Fuel Oil Area
-------
8U9tM
Figure 10. Total Volatile Organics Near Solvent
Recovery Containment Ditch
0B-8A Tolal VotatlUs
OB-llOA Tolal Volatile*
00
C
Q
C
0)
u
a
o
L>
tti
C
o
Zj
ta
u
4J
c
0)
o
C
O
a
200000
6/3/88
(391.160:
l/l/fl*
($06,200)
0/80/86
(414.300)
«/t&/ao
(3I7.M0)
\ ?
±,u.
* * I I
UW- 104 Total VolatUes
\
00
a
o
£
*
u
C
V
a
c
o
-------
15
81931M
Figure 11. Total Volatile Organics Downgradient of
Solvent Recovery Containment Ditch
MW-113B Total Volatiles
MW-115B Total Volatiles
tsfl
C
o
• M
-*->
a
u
c
0)
o
a
o
c_>
700
600 -
500
400 -
300 -
200
100
on
G
O
(0
u
+->
a
0)
o
a
o
u
700
600 -
500 -
400
300 -
200 -
-------
B - BORING
BP - BURN PIT
UW - UOMTOMNG WELL
OB - OBSERVATION WELL
P - PIEZOUETER
PW - PURGE well
TW - TEST WEIL
W - WELL
WP - WELL POINT
A - BEDROCK WELL
9 - UNCONSOLIDATED SAPROUTE WELL
• - HYBRID OR UNCLASSIFIED WELLS
B - ABANDONED BORING LOCATION

y
TBM-CHISELED ON TOP
WALL - LEFT DOWNSTREAM
Si DC OF BRIDGE
ELCV - 327.67
200
200
SCALE IN FEET
1" = 200-0"
400
FIGURE 12
EXTENT OF FUEL OIL
THERMALKEM FACILITY
-------
S193160
Figure 13: TCE and PCE at MW-100
if&R
MW-100 PCE
DO
a
o
¦ rH
-*-»
(0
u
*->

o
C
o
CJ
80
60
40
20
8/3/88
(83)
Below Detection Limit
ettoooe • e e • ®
i
_L_
_1_
13 3 3 3 1
DO
3
G
O
(0
C
V
o
a
o
u
MW-100 TCE
40000
30000
20000
10000
3/1/85
(41.000)
Detection Limit < 10 ug/l
8/3/88
(30,000)
-------
Groundwater Remediation Systems
• Initial Remediation
° Extraction Well (EW-1)
° Production Well (PW-1)
-------
SL03156
19
ENR
Initial Remediation
(June 1983- 19135)
® Cleaned and Closed Solvent Recovery
Containment Ditch by October 1983
° Removed Contaminated Surface Soil by
October 1983
»Removed Underground Fuel Oil Tanks by
June 1984
° Cleaned Out Above Ground Tank Waste by
-------
sun 1&6
20
ENhR
Initial Remediation
(June 1983- 1385) (Cont'd)
•	Implemented RCRA Facility Standard
•	Upgraded Incinerator to Subpart O Standards
•	Performed Magnetometer Survey; Identified
Buried Drums and Burn Pits in 1985
•	Removed Buried Drums Near MW-103;
Remediation Approved by DHEC
•	Excavated Burn Pits; Remediation Approved
-------
8L931M
Figure 14. Drawdown in Saprolite with EW-1
-------
950B15A
iSsW-too
i rerun
MW-107
V
HV-10S
V
MWrlOl
~
*1-1
UMCONSOUDA1ED HELL
BEDROCK WELL
HYBWO
UMOMNM
5MO	 pczoumoc ELEVATION CONTOUR (ft MSL)
CONTOUR MTDIVAL - I FOOT
200	0	200
APPROXIMATE SCALE IN FEET
1" = 200-0"
400
FIGURE 15
PIEZOMETRIC SURFACE, 6/25/91
THERMALKEM FACILITY
-------
SL69156
Fuel Oil Area Pilot Purge Wells
• Two Shallow Purge Wells
-	Screened 10 to 20 ft. deep; across product level and
product/water interface
-	Screened 20 to 30 ft. deep; below product level, but
across product/water interface
•	Pumped at 4 to 8 gpm
-------
6L931t^
Fuel Oil Recovery Pilot Study
•	Phase I: Total Fluid Recovery From One Well
•	Phase II: Vacuum Extraction and Fluid
Recovery From One Well
•	Phase III: Multiple Well Recovery - Six Wells
-------
= VAPOR LXmrtv/iio,.
EXISTING
CONCRETE
SLAB
= EXISTING GROUNDWATER MONITORING WELL
= EXISTING GROUNDWATER PUMPING WELL
(PW-2A IS A FORMER PUMPING WELL
EW-1 IS ACTIVELY PUMPING)
OB—13 A
OB—11
APPROXIMA SCALE: 1"=30'
FIGURE 16
PILOT STUDY
THERMALKEM FACIl
ROC VIILL, SC
-------
Phase 1 Liquid Extraction System
10
WAS ft WA FIR
IREA1MINI
Pi AMI
Lf Cf NO
EXTRACTION WELLS
PI
-
PRESSURE ft D
-------
Vt-I
VAPOR tXTRAClKW
ncti
VAPOR
PHASE
CARBON
i *
' NOeUMLt
I aosto
I VM.VI
I
I
U-til-
chrbow ar-p»ss
J£L
BLOtCR
WLET
filter
-tNLEI
SILENCER
MO

LfCfHD
FP - now MEASUREMENT POR!
ts - um smich
PI	- PRESSURE INDICATOR
PS	- PRESSURE SWITCH
SP	- SAMPLING POR I
It	- UMPCRATURE Etcutfatf . ,
II	- TEMPERATURE tNOtCAlOR
TS	- TtUPERATURE SM1CH
dpi	- differential pressure woicaior
• MANUAL VALVt 
-------
Phase 3 Multiple Well Groundwater
Extraction System
SftER
FOOT
VALVt «/
STRAMCft
IFGIHO
-	PRESSURE MOICAfOR
. PRESSURE SWITCH
-	PRESSURE RELAY
-	F10W 10TAL12ER MOlCAlOR
-	VALVt
•	OIAPHRAM VAL\ff
•	SOUNOC VAIVE
-------
Fuel Oil Recovery Pilot Study Results
•	Vacuum Had Measurable Positive Effect, But
Not Significant
•	Satisfactory Recovery From Multiple Wells
•	Most Recovery From Secondary Fractures
•	Long Term Pumping Might Recover Remaining
-------
SL991S6
30
EN3R
EW-1 Remediation
August 1992:
February 1993:
• April 1993:
July 1993:
EW-1 Pumping at 14 gpm;
Preventing Off-site Migration
EPA Notified That Newly Installed
Downgradient Wells Show Low
Level VOCs
EW-1 Raised to 35 gpm;
Preventing Off-site Migration
Product Found in EW-1; EW-1
Lowered to 25 gpm; Preventing
-------
Conceptual Sitewide Remediation Plan
•	One Additional Extraction Well
•	Two Wells Pumped to Prevent Off-site Migration
-------
950811A
legend
B - BORING
BP - BURN PtT
UW - UONWOMNC WELL
08 - OBSERVATION WELL
P - PIEZOMETER
PW - PURGE WELL
TW - TtST MEU.
w - mi
WP - WELL POINT
A - BEDROCK WELL
« - UNCONSOUDATED SAPROUTF WELL
% - HYBRID OR UNCLASSIFIED WELLS
¦ - ABANDONED BORING LOCATION
A - ADDITIONAL EXTRACTION WELL
	POTENTIAL GRAVITY DRAINS
—— - APPROXtmn CAPTURE ZONE


TBM-CW SB-ED ON TOP
WALL - l£FT OOVMSmCAM
SIDE OF BRUXZ
ELEV. - 327.67
200
0
200
400
SCALE IN FEET
1" = 200'—0"
FIGURE 17
CONCEPTUAL REMEDIATION PLAN
THERMALKEM FACILITY
-------
84.03156
33
ENS
Results of Remediation
•	EW-1 Continues to Remove Mass
•	Total Volatiles Near the Source Decreased
•	EW-1 Prevents Off-site Migration of Dissolved
Constituents and Pure Product Fuel Oil
-------
6L03
34
Figure 18. Total Volatile Recovery in Treatment
System from PW-1 and EW-1
TOTAL VOLATILES
6000
5000
4000
3000
i r
25.499 I ppb
2/91
&0
2
c
o
(0
u.
e
01
c 2000
o
u
1000
o
a>
Z
<
o
o>
K
%
0
01
J
p
0
01
£-
CJ
O
z
CT>
(Z
%
-J
o> o>
E-
U
O
CM
CT>
<
CM
CT>
K
0*
<
13.108 ppb
11/82
6.679 ppb_
1/93
C\J
Ol
C\1
CO
t-
CJ
O
n
en
z
<
i	r
6.M3 ppb
5/93
CO
O,
<
CO
J
CO
CT)
t-
U
-------
95081OA
B - BORING
BP - BURN PIT
UW - UOMTOfUNC WELL
OB - OBSERVATION WEIL
P - PIEZOUETER
PW - PURGE WEIL
TW - TEST WELL
W - WEU
WP - WEIL POINT
A - BEDROCK WELL
• - UNCONSOUMTEO SAPROUTE WELL
e - HYBRID OR UNCLASStfTtD WELLS
m - AB/ T> BORING LOCATION
200
200
SCALE IN FEET
1" = 200'-0"
400
FIGURE 19
EW-1 CONTAINMENT
OF FUEL OIL
THERMALKEM FACIUTY
-------
TABLE 1
Approximate Remediation Costs - 1983 to Present
ThermalKEM, Rock Hiil Facility, SC
REMEDIAL TASK
COST
Initial Remedial Programs - Magnetometer Study and
Remediation of Old Burn Pits [1985]
$1,200,000
Design and Installation of Remediation Wells
EW-1 and PW-1 [1988]
$37,000
Fuel Oil Pilot Purge Wells [1991]
$80,000
Fuel Oil Recovery Pilot Study [Aug - Oct, 1992]
$130,000
Site Characteristics Investigations [1988 • 1993]
$260,000
Groundwater Monitoring [Annually]
$35,000 for
years; now
$12,000
RFI WorkPlan Preparation and Regulatory Response
-------
p
-------
RCRA Ground Water Remediation Conference
December 1-3, 1993
-------
-------
The Torrington Company
Clinton Bearings Plant
© Roller Bearing Manufacturer.
@ 550,000 sq. ft., 156 acres.
Employing 1200+ People.
© Clinton, South Carolina.
Between Greenville and Columbia, off I-26.
® Mostly Rural Housing and Farmland.
-------
HAZARDOUS WASTE
-------
Hazardous Waste Management Units
•	5 Surface Impoundments (R)
A.	Copper Cyanide and Floating Oil
B.	Copper Cyanide and Floating Oil
C.	Copper Hydroxides
D.	Chemical Milling Solution
E.	Chemical Milling Solution
•	Cyanide Destruction Units 1 and 2
•	Sludge Drying Beds 1 and 2 (R)
•	5 Industrial Wastewater Lagoons
Clay-lined, Earthen
•	Several Industrial Waste Landfills
-------
LJ
-------
Hydrogeology Beneath Facility
Consists of three interconnected water
bearing zones.
•	Zone 1 - Soil Zone
Sandy Silts and Clays near surface.
•	Zone 2 - Partially Weathered Rock (PRW) Zone
Transition between soil and bedrock.
Variable quantities of both sandy silts
and highly fractured rock.
Preferential flow path for ground water.
•	Zone 3 - Weathered, Fractured, and/or Jointed
Bedrock Zone
Extends from base of PWR to bottom of uppermost
-------
Hyrogeology Beneath Facility
con't
•	Verticle limit of uppermost aquifer has been defined
to be below the uppermost 10 feet of unweathered
competent bedrock.
•	Presence of iron staining on joint surfaces is indicative
of ground water flow through the fractured bedrock.
•	Low hydraulic conductivities suggest that movement is
very slow.
•	Based on potentiometric surface contour maps, ground
water flow in the aquifer system is from north to south-
-------
-------
Scope & Extent of Ground Water Contamination
Prior to Ground Water Remediation Efforts
BACKGROUND
•	To comply with ground water monitoring requirements,
we conducted a hydrogeologic study in 1981.
•	Six wells were installed near the regulated units.
•	Sampling and analysis was done quarterly until mid 1985.
•	Monitoring activity indicated that ground water had been
impacted with indicator and several specific parameters.
•	Three phases of ground water assesment were conducted
between Jan. 1986 and Aug. 1987.
•	These studies concluded that ground water had been impacted
-------
Scope & Extent of Ground Water Contamination
Prior to Ground Water Remediation Efforts
•	Phase II data suggested that the horizontal extent
appeared to be limited to Torrington property and
was discharging into North Creek and North Creek Pond.
® Vertical extent is limited to the uppermost aquifer.
•	Phase III was conducted to better define the horizontal
extent.
•	Data collected in the North Creek area further defined
the horizontal and vertical extent to the south-southeast
with point sources in the vicinity of the lagoons.
•	Phase III concluded with a conceptual corrective action
plan (CCAP) to remediate ground water through the
-------
-------
Ground Water Remediation System
•	The Ground Water Recovery (GWR) system is intended to
produce and maintain a hydraulic barrier for ground
water-flow.
•	The GWR system consists of 15 recovery wells and
corresponding observation wells.
•	Recovery Wells R-1 through R-11 were installed in a
linear network along the southern property boundary.
Each well pumps into an aboveground pipe manifold which
carries the water to the wastewater pretreatment plant.
The water is collected in a 7,500 gallon tank and then
pumped through twin air stripper towers for VOC reduction.
•	Recovery Wells R-12 through R-15 are located in zones of
specific contamination. Each well (except R-12) has a
dedicated discharge line and is treated separately at the
-------
Ground Water Remediation System
•	Each well is constructed of a 6-inch diameter
stainless steel casing and screen.
® Each well has a submersible pump that is rated
for the flow necessary to achieve the desired
ground water table drawdown.
•	Flow controls and electronic components are
located at each well.
•	A main control/status panel is located in the
pretreatment plant which monitors operational
conditions of each well and provides ON/OFF
-------
Ground Water Remediation System
Operational Problems, Difficulties, Changes
•	Casing Shift in R-4. This recovery well descends
through an old debris landfill. Evidently, some settling
has occurred which has bent the casing more than 4 inches.
The pump could not be removed for repair, so a Jet pump
was installed above the well to ensure operation and
barrier integrity. Well repair is still being considered.
•	Modification of R-13. Source area recovery well R-13 is
utilized to remove light, non-aqueous phase liquids (LNAPL)
from the ground water surface. The skimmer pump never
ran because the layer of LNAPL was not thick enough. A
recovery pump has been installed below the skimmer pump to
generate a hydraulic gradient that will pull more of the
-------
Ground Water Remediation System
Operational Problems, Difficulties, Changes
•	Colonizing Iron Bacteria. We have had to replace more
than 5 submersible pumps due to failures related to the
bacteria problems.
The bacteria growth has affected the ground water
recovery rate. At times, it has come close to jeopardizing
to integrity of the barrier systen:.
We are about to undertake a project of rehabilitation for
the system. This is to include a pH adjustment and a UV
Oxidation step to kill the bacteria.
•	Pressure Sustaining Regulators. Due to periods of drought
and flood, we have installed the regulators to adjust the
-------
-------
4 12 3 4 12 3 4 12 3 4
87
2 d 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3
oo
(39
90
92
93

-------
PPM
"3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3
-------
PPM
O
O
4 12 3
7| 88
2 3 4 1 2 3 4 1 2 3 4 1
91 |
-------
PPM
3 4 1 2 3 4 1 2 3 4 1 2 3
1 2 3 4 1 2 3 4 1 2 3
1 2
86 | 87 | 88 | 89 | 90
-------
-------
-------
Progress In Ground Water Remediation
•	Since startup of GWR system, there has been a continual
decrease of VOC and inorganic constituent concentrations.
•	Plume size has not greatly decreased.
•	Wells near the recovery system that previously exhibited
artesian conditions, no longer do so.
•	Additional off-site assesment, including additional well
installations, have confirmed that North Creek acts as a
hydraulic barrier and R-12 is successfully remediating
-------
-------
Costs of Ground Water Remediation System
• One Time Costs
Logistics and Planning	$ 15,000
Pilot Recovery System	175,000
Barrier System Implementation	410,000
Source Recovery Wells	200,000
Total	$ 800,000
Annual Costs
Repair	32,000
Modification/Upgrade	20,000
Utility	36,000
Personnel	22,000
Total
-------
David Carl Cromwell Age: 27
BS Industrial Chemistry	Western Carolina University
Position: Project Engineer A Title: Plant Chemist/Environmental
Coordinator
Joined Torrington in 1990.
Two years Active Duty - Army
-------
Cardinal.
-------
WSRC-MS-93-511
Groundwater Cleanup: An Update of the
Savannah River Site Experience (U)
J.G. Horvath
Westinghouse Remediation Services, Inc.
C.L. Bergren
Westinghouse Savannah River Company
ABSTRACT
A full-scale pump-and-treat groundwater remediation program that addresses a large plume of volatile organics has
been ongoing at the Savannah River Site (SRS) since 1985. The system has recovered more than 135,000 kilograms
of solvent and is containing the center of the plume. While overall protection is being achieved, reducing the
concentration of contaminants to regulatory acceptable levels is a problem.
INTRODUCTION
As presented at ER '91', the Westinghouse Savannah River Company (WSRC) and Department of Energy (DOE)
experience at the Savannah River Site (SRS) suggests that meeting cleanup standards is a challenge in light of
technical realities.
The purpose of this paper is to update the reader regarding the corrective action program that is addressing a large
plume of volatile organics beneath the A/M Area of SRS. At ER "91, we described the history and status of the
program, costs, measures of performance, lessons learned, and challenges faced.1 Below, we briefly review these
areas and provide a list of recently completed actions.
HISTORY
SRS, which has been in operation since 1953, is a 780-square kilometer site that produced special isotopes for the
national defense program. As a result of past waste disposal practices, groundwater at several locations within the
Site has become contaminated with solvents, metals, and radionuclides. In 1981, the groundwater located beneath the
Site's fuel and target fabrication facility was found to be contaminated with volatile, organic degreasing solvents,
primarily trichloroethylene (TCE) and tetrachloroethylene (PCE). The sources of contamination were a settling basin
and sewer line (now closed)2, a solvent storage area, and other release points located near the fabrication facility. In
response, DOE voluntarily initiated a groundwater corrective action program, including an extensive groundwater
monitoring system. Groundwater remediation in the A/M Area began in 1982 with the startup of an experimental air
stripper. To date, more than 300 monitoring wells have been installed to characterize the plume.
Full-scale groundwater recovery with treatment by air stripping has been ongoing in A/M Area since 1985. The
remedial system now comprises 12 recovery wells and two air strippers. Since the beginning of remediation, more
than 135,000 kilograms of degreaser solvents have been removed from more than 6,700,000,000 liters of
groundwater. The system is reducing the mass in the central plume region effectively and is serving to contain the
contamination present there. SRS has realized through continuing evaluations that the current system will require
augmentation to address other areas as discussed below.
Ongoing investigations have determined that a significant amount of solvent remains in the vadose zone. More
recently, a separate phase of solvents was detected in a monitoring well located near the closed settling basin. Both
occurrences continue to influence the amount of solvent in the aqueous phase. WSRC has completed a program to
characterize more fully the vadose zone beneath source areas. Contamination in the vadose zone has been and will be
addressed by vacuum extraction. The presence of dense nonaqueous phase liquids (DNAPLs) is the focus of a
current characterization effort. DNAPLs represent an even greater challenge in characterization and remediation
-------
WSRC has completed some enhancements and plans to expand the corrective action program. In 1989, packing in
the full-scale air stripper was replaced, the system was tested,3 and flow from the recovery wells to the air stripper
was increased from 1500 to 1900 liters per minute, a 25% increase. Additional recovery capacity was added in 1991;
SRS installed a recovery well and relocated a prototype air stripper to a source area near the Site's northern
boundary, outside the influence of the original recovery system.
In addition to relocating the prototype air stripper, WSRC and DOE also plan to install additional groundwater
recovery capacity near the Site's northern boundary. The DOE-WSRC team also initiated projects that will address
contamination remaining in the vadose zone. A 1987 pilot system recovered more than 450 kilograms from the
vadose zone in a three-week period. More recently, experimental horizontal well systems, tested by the Savannah
River Technology Center under the auspices of the Office of Technology Development, have recovered more than
7200 kilograms from the subsurface by using vacuum extraction and in situ air stripping of the aquifer near a source
area (the closed sewer line). Based on the success of these efforts, WSRC plans to pursue installing four additional
systems that will address source sites.
Completed actions since ER '91 are presented below:
WSRC obtained SRS's first CERCLA records of decision (RODs) in 1992. These RODs are serving as a model
for implementing EPA's fundamental study area concept. The concept was developed to efficiently manage sites
with several operable units. It is expected that integrating CERCLA requirements into the program will
encourage the use of risk and cost-benefit analyses to determine the degree of and need for future enhancements
to the program.
The design of a vadose zone remediation system has been completed. WSRC plans to install the system for DOE
during FY 94. The system will use vacuum extraction and will be skid mounted.
The design of a 250 GPM groundwater recovery system to be installed near the Site's northern boundary has
been completed. This system will complement the 70 GPM unit completed in 1991.
DOE's integrated demonstration project has been incorporated into the actual cleanup operation. The hardware
(horizontal wells, treatment systems, etc.) from the successful technology development projects are considered to
be a part of the total cleanup program.
•	Forum for national focus on DNAPLs was established. The discovery of DNAPLs in the subsurface near the M-
Area Settling Basin prompted WSRC/DOE to lead a DNAPL national task team. The results of this task have
been visible in recent DOE reports.4
Risk assessments are ongoing for the multiple operable units in the fundamental study area. The risk assessments
are complicated because of the multiple contributions of the various operable units and specifically how the units
influence groundwater contamination.
•	Cone penetration and resistivity tools were used successfully to screen areas for additional well placement and to
characterize DNAPL contamination. The dilute plume south of the closed M-Area Settling Basin was
characterized further using cone penetration to minimize the addition of monitoring wells. Closer to the basin
cone, penetration combined with resistivity techniques and other methods were used to determine the nature and
presence of DNAPLs in the subsurface.5
The remediation program at SRS is providing knowledge and experience on how well cleanup programs work.
WSRC is conducting extensive risk assessments associated with contaminated sites for DOE. A better understanding
of the limits of remediation, combined with the amount of risk posed by the impact, eventually should lead to
improved decisions regarding cleanup goals.
PROGRAM COSTS
The original 11 recovery wells and air stipping system cost $4,800,000 to design and construct. The system costs
approximately $100,000 per year to operate and maintain. The system has performed very well with an operating
utility of greater than 95%. The major maintenance concern is lightning strikes.
Other costs associated with the corrective action program are the expenses of groundwater monitoring, interpreting
and managing data, modeling, reporting, and continuing investigations and project development.
REMEDIATION REQUIREMENTS, GOALS
The A/M-Area groundwater corrective action program is permitted under the postclosure requirements of the
Resource Conservation and Recovery Act (RCRA). The facility became fully permitted in September 1987. Exact
conditions for the program are stated.in the Site's hazardous waste permit6 and in the associated permit application7.
-------
The initial goal of the program was to remediate the most contaminated groundwater by maximizing mass removal
from the center of the plume. It has been SRS's position that the remediation would take place in stages. The
corrective action program calls for groundwater remediation to be continued until "the groundwater concentration ...
no longer poses a threat to human health and the environment..." However, concentration limits listed in the permit
and permit application for the primary constituents of concern are very near analytical detection limits. Although
concentrations of a few other constituents are above the limits described in the permit and associated application,
they do not exceed these limits in the influent stream to the air stripper. These other constituents are of limited extent
and are detected in only a few monitoring wells near the closed settling basin.
WSRC and DOE have used the following tools to evaluate performance of the recovery system:
Zone of capture modeling - The volumetric extent of groundwater is recovered during a specified time interval.
Estimated changes in subsurface inventory - The original subsurface inventory is compared to current inventory
using data from monitoring wells.
Mass removal - This is the mass balance of contaminants across the groundwater treatment system. The total
mass removed is compared to a hypothetical 99% mass removal curve.
These methods are explained in detail in the ER '91 paper.1
CONCLUSION
SRS has developed a groundwater program intended to detect, characterize, and remediate groundwater impacts.
When evaluated against original goals, the pump-and-treat system has been successful. A large amount of mass has
been removed, and significant progress has been made regarding plume containment. However, low concentrations
probably are not achievable through a standard approach. Although new technologies offer hope, they still must
overcome the difficulties presented by a complex hydrogeologic system. In addition, other issues such as vadose
zone Contamination and uncertainties associated with DNAPLs need to be addressed. It is imperative that the sources
of contamination be addressed to prevent further degradation of the groundwater resource. However, the limits of
any type of remediation must be faced such that reasonable cleanup targets are chosen.
DOE and WSRC are committed to restoration, while assessing risk and continuing to search for less-expensive
methods. Risk-based approaches to cleanup may lead to more logically deduced cleanup goals and standards. It is
essential that the benefit of such efforts be assessed so that resources are applied wisely.
REFERENCES
1.	J.G. HORVATH and SCOTT SUROVCHAK, Groundwater Cleanup: The Savannah River Site Experience,
presented at Environmental Remediation '91, Pasco, Washington.
2.	S.R. MCMULLIN and J.G. HORVATH, M-Area Basin Closure: Savannah River Site, Westinghouse Savannah
River Company, presented at Environmental Remediation '91, Pasco, Washington.
3.	S.T. MCKILLIP, K.L. SIBLEY and J.G. HORVATH, Air Stipping of Volatile Organic Chlorocarbons: System
Development, Performance and Lessons Learned, Westinghouse Savannah River Company, presented at
Environmental Remediation '91, Pasco, Washington.
4.	TIE Quarterly, Vol 2, No. 3, Summer 1993, U.S. DOE, Office of Environmental "Restoration.
5.	B.B. LOONEY, et al, Characterizing DNAPL Contamination: A/M-Area Savannah River Site, Presented at
Environmental Remediation '93, Augusta, Georgia (unpublished).
6.	Hazardous Waste Permit No. SCI-890-008-989, Part IV, South Carolina Department of Health and
Environmental Control, Office of Environmental Quality Control, Bureau of Solid and Hazardous Waste,
Columbia, S.C. (September 1987).
7.	Application for a Post Closure Care Permit, M-Area Hazardous Waste Management Facility, Volume III, E.l.
-------

L LAKE
I
H
RI v	SOUTH
ER	CAROLINA
-------
BIOGRAPHICAL SKETCH
Chris Bergren is a geologist with Westinghouse Savannah River Company. He is a
manager within the Environmental Restoration Department at the Department of Energy's
Savannah River Site (Aiken, South Carolina) and is responsible for monitoring and
remediation efforts within the Site's A/M Area. Previous assignments over the past eight
years at the Savannah River Site involved well drilling and sampling, monitoring and
-------
-------
Evaluation of Pump & Treat System Performance
Robert C. Borden
North Carolina State University
Pump and treat systems have been designed to prevent contaminant migration,
recover both dissolved and separate phase contaminants from the subsurface, and to
restore ground water to a quality suitable for other beneficial uses. Typical pump and
treat systems consist of recovery wells and/or trenches, an above ground treatment
system (oil/water separator, air stripper or carbon adsorber), and some method for
disposal of the final effluent (infiltration gallery, injection wells, surface discharge or
POTW discharge).
Pump and treat systems can be effective in preventing contaminant migration
but are often less effective at restoring ground water quality. This is often due to the
presence of non-aqueous phase liquids or NAPLs. These NAPLs may be present in
the subsurface as single discontinuous droplets or large elongated blobs. Because of
the chemical properties of the contaminants, most petroleum compounds will partition
into the NAPL and dissolve very slowly. To evaluate the process, a series of laboratory
column experiments were conducted using sand which had been previously flooded
with gasoline. Water was then pumped through the contaminated aquifer material and
monitored for dissolved BTEX (benzene, toluene, ethylbenzene and xylene isomers).
The concentration of most compounds in the column effluent declined steadily and then
reached a plateau where little additional change in concentration occurred. The less
soluble compounds take longer to decline and then reach the same plateau.
A statistical study was conducted of the long term performance of pump and
treat systems installed at underground storage tank (UST) sites in North Carolina.
Monitoring histories from 48 wells at 13 sites were evaluated to detect trends in the
concentration of BTEX over time. For the 48 wells analyzed, 12 to 29% reported a
statistically significant negative slope (95% confidence) in the concentration versus
time plot. Cleanup times were estimated by extrapolating the regression line until it
reached the appropriate groundwater standard for each BTEX compound. The cleanup
times were generally in the range of zero to 15 years, or enormously long, indicating
the would not reach standards at the current trend. These results indicate that the
existing pump and treat systems will not be effect at meeting ground water standards
within a reasonable time period.
Monitoring data were also used to estimate the amount of benzene recovered
from the ground by pump and treat systems. Benzene was selected for analysis
because it has the lowest allowable concentration in drinking water and is the least
biodegradable under the anaerobic conditions common in many petroleum
-------
equivalent gasoline volume based on an estimate of the average amount of benzene in
gasoline. Results from this analysis show that many pump and treat systems are very
effective in removing contaminants from the ground. In three of the four systems
evaluated, recoveries ranged from 1200 to 2800 gallons of equivalent gasoline. One
system was poorly designed and recovered little or no contamination after the initial
startup, even though free product was still present in some wells.
Summary and Recommendations
Pump and treat systems can be very effect at removing contaminants from the
ground but are typically much less effective at meeting ground water standards. The
design of these systems should be modified to maximize contaminant recovery by
pumping more water through the contaminated interval. Once the readily recoverable
contaminants are removed, other alternatives should be considered for managing any
-------
Evaluation of Pump & Treat
System Performance
Robert C. Borden
North Carolina State University
PUMP & TREAT OBJECTIVES
Q Prevent migration.
Q Recover contaminants from
subsurface.
© Restore CW quality.
-------
-------
STATISTICAL EVALUATION OF
PUMP & TREAT PERFORMANCE
9 Collect data on NC P&T systems
over 3 yrs old.
O Evaluate change in BTEX in MW
over time.
Q Estimate time to cleanup.
O Estimate benzene removed from
ground.
SITE CHARACT1STICS



Sito
Rogion

Monitor
Raoov
System
Pmp


(yrs)
Walls
Well*

Rat*





<9Pm)
1
Mount.
5
8
2
O/W
1.25
2
Coast.
4
11
2
Strip.
3.5
3
Cout.
4
16
2
Strip.

4
Plednxt.
4
7
1
Strip.
1.
5
Coast.
4
5
2
Strip.
4.
6
Coast.
4
9
1
O/W

7
Cout.
5
7
3
Strip.
1.3
8
Piodmt.
4
12
2
Strip.
3.5
9
Mount.
5
5
1
O/W

10
Coast.
5
8
1


11
Fiadmt.
3
5
1
Strip.
1.
12
Coast.
6
6
3
Strip.
3.
13
Mount.
6
11
3
Strip.
1.25
WELLS WHERE BTEX IS DECREASING
(95% Confidence)
Q Benzene	21%
O Toluene	21%
9 Ethylbenzene	20%
-------
TIME TO MEET STANDARD AT A WELL(%)
(95% confidence concentration is decreasing)
Tim* Benxene Toluene 1-Benx. Xylene BTEX
(yr*)
0-5 11	24	18	24	B
5-10 3
10-15 5
15-20
20-25
25-30
30-50
50-100
TIME TO MEET STANDARD A T A WELL(%)
(no significance test)
Time Benzene Toluene I-Ben*, xylene BTEX
(yre)





0-5
21
84
48
55
18
5-10
16
3
12
5
8
10-15
13

3
3
10
15-20
5


5
5
20-25



8

25-30
3



3
30-50
S
3
3

5
50-100
5

3

5
>100
29
10
31
24
46
TIME TO MEET STANDARD AT A SITEC/o)
(no significance test)
Time Benzene Toluene B-Benx. Xylene BTEX
(yr»)





0-5
8
75
20
33
8
5-10
17
8
8
8
8
10-15
17

8
8
8
15-20
8


8
8
20-25





25-30





30-50





50-100





>100
50
17
60
43
-------
BENZENE RECOVERY ANAL YSIS
9 Multiply concentration x flow rate
in recovery welL
9 Plot cumulative recovery versus
time.
9 Convert to equivalent gallons
gasoline.
EQUIVALENT GASOLINE RECOVERY
(based on Kg Benzene)
EQUIVALENT GASOLINE RECOVERY
-------
EQUIVALENT GASOLINE RECO VERY
Site Flowrate	Time Benzene	Gas
(gpm)	(yrs) (Kg)	(gal)
A 3.5	4 67	1196
B 4.0	4 118	2106
C 2.0	5 50	893
D 3.0	6 156	2793
CONCLUSIONS
9 Pump and Treat is not effective at
meeting groundwater quality
standards.
Q Pump and Treat is effective at
removing significant amounts of
BTEX from the ground.
O Most systems are designed for
containment only, not remediation.
RECOMMENDA TIONS
O Maximize pumping rates to remove
more contaminant
O Use mass recovery analysis to determine
when system modifications are needed.
-------
-------
SUCCESSES AND FAILURES USING ALTERNATIVE TECHNOLOGIES AT
GROUNDWATER CONTAMINATED SITES IN NORTH CAROLINA
Burrie vD. Boshoff
North Carolina Department of Environment, Health and
Natural Resources
Division of Environmental Management - Groundwater Section
Conventional pump and treat systems have been installed and
operated at many UST sites in North Carolina. These systems have
had varied success although none of the sites have been closed
out. This can be contributed to several factors such as pump and
treat systems are inherently less effective at restoring
groundwater guality but more effective in preventing contaminant
migration; existing systems have been either improperly used or
were designed with a limited ability to remediate groundwater;
groundwater cleanup requirements in North Carolina are relatively
more stringent than in many other states; and the time since
operation was started at these systems is fairly short.
Alternative treatment technologies have been introduced with
high expectations and claims by consultants that sites will be
cleaned up much quicker and at significantly lower cost.
Innovative systems have been put into operation, but experience
from these identified some pitfalls which prompted the development
of guidelines to be used by field personnel. Some of the
successes and problems are illustrated by the following three
examples:
Examples of Alternative Technologies
i) Pump and Treat with Bioremediation
The former Dennis Equipment Rental had seven USTs with
gasoline arid diesel releases. The site is located in the Triassic
Basin of central North Carolina, which is primarily composed of
mudstones and siltstones. The soils encountered consisted mainly
of fine sandy clays and silty clays, with occasional clay lenses
and rock fragments. The site hydrology is interpreted as an
unconfined water table aquifer with an estimated saturated
thickness of approximately 15 ft. Petroleum hydrocarbon (PHC)
concentrations detected in the soil amounted to 12,000 ppm
gasoline and 3,000 ppm diesel fuel. The groundwater contained
7 7.1 ppm BTEX and 215.5 ppm TPHCs. A pump and treat system was
installed consisting of a bioremediation system to treat the
groundwater using a closed-loop treatment program with recovery
wells and infiltration galleries. Recovery well and biotreatment
system influent and effluent samples were collected and analyzed
on a weekly basis for the first several months of the project.
Sampling was then reduced to biweekly of bioreactor influent and
effluent, and monthly sampling of recovery wells and selected
monitoring wells. Over a period of two years, this system reduced
the PHC concentration in the soil to 2,900 ppm gasoline and less
than 10 ppm diesel fuel. The concentration of BTEX in the
-------
groundwater was reduced to 24.9 ppm. The total cost of the
cleanup operation performed to date amounts to $578,627 which
includes construction and maintenance costs. Operation and
maintenance and lease costs per month are about $12,000. Concern
about the system at this stage is mainly due to the high cost
especially for operation and maintenance. The question that comes
to mind is whether the same cleanup could not have been achieved
if all contaminated soil was excavated and hauled away during the
first month. In which case, costs may have been reduced to about
$150,000 and the same results could have been achieved in a much
shorter time.
A major concern has developed with innovative technologies
where consultants have patent rights on a system or part of a
system. The system is then installed and leased at a monthly rate
which turns out to be quite substantial. This has prompted a
policy whereby systems will only be leased to an amount equal to
the purchase price of such a system. A further complication of
systems where patents are pending is that the patent holder does
not want any information about the system to be released or
published until the patent has been granted. This is in conflict
with the 40 CFR 280.67 federal rules and state rules. Corrective
action plans cannot be approved with any stipulation that prevents
the public notice or information being distributed to interested
parties.
ii) Vacuum Air Sparging System
Dodge Store, Fayetteville, North Carolina is a former retail
gasoline outlet and convenience store located in central North
Carolina on flat land with a slight slope from the front toward
the back. Groundwater flow was determined to be in a
northwesterly direction and parallel to the road in front of the
store. A gasoline release was associated with a five-tank
complex. The surficial aquifer is a heterogeneous mixture of
sand, silt and clay layers and the water table is generally just
below 10-11 feet in most monitoring wells. There is a thick and
hard clay layer starting about 10 to 13 feet deep beneath the
southeastern part of the site, but it is absent on the
northwestern part, being replaced by sand with some silt. A
vacuum air sparging system was installed using sparging wells and
vacuum lines on the same wells. One of the main concerns about
this system was that there is a patent pending on the design of
the system. Further questions were raised on how effective the
system was because all but one of the monitoring wells were
eventually used for sparging wells and the sparging wells were
used to sample and determine the groundwater cleanup status.
Information is still needed to determine the radius of influence
of the system and to what extent the system is actually cleaning
up the groundwater. Several requests to the consultant to provide
this information were unsuccessful and approval of this system for
any other sites is being withheld subject to submittal of the
requested information.
-------
iii) Conventional Vacuum Air Sparging System
The Corner Store located on the southwest corner of
intersection of N.C. Highway 71 and State Road 1312, Robeson
County, had five USTs removed. Initial assessment indicated a
release of petroleum hydrocarbons had impacted soil and
groundwater in the area. Five residential potable water supply
wells were impacted. Eight county water supply wells are located
within one mile of the site, pumping a combined total of over 8.5
million gallons per day. The soils encountered varied from fine
through course sand with underlying clayey sand and streaks of
coarse sand and sandy clay. Four monitoring wells were installed
initially and sampled with nearby county supply wells. Total
contaminant concentrations detected ranged from 6 ppb to 60,330
ppb. MTBE was the only compound detected in one of the county
supply wells at a concentration of 5 ppb. An air sparging system
consisting of four sparging points were installed around the west
and south sides of the former UST pit and vary in depth between 28
and 35.5 feet. The points were constructed identically to 2-inch
monitoring wells with a 2-foot screen at the bottom and were
terminated from 7 to 12 feet below the water table. Individual
PVC air lines were trenched in to each well and connected below
grade. A soil vapor extraction system was installed to operate in
conjunction with the air sparging system. It was recognized that
a pump and treat system may be necessary as an additional
treatment system because of (1) the depth of confirmed
contamination in groundwater, (2) the depth limitations of the air
sparging system, and (3) the known effects of pumping rate at one
of the county supply wells. The pump and treat system would serve
to begin treatment of the highly contaminated groundwater in the
vicinity of the former USTs, as well as assist in counteracting
the pumping effects of the county supply well to limit to the
extent practical any continued plume migration. The air sparging
system is limited to the treatment of the upper 20 feet of the
shallow aquifer, due to pressure restrictions associated with
oil-less blowers required by the State of North Carolina. The
vacuum air sparging system was operated for approximately eight
months and significant results have been obtained in that MTBE
concentration was reduced from 8,600 ppb to 270 ppb, benzene from
12,000 to 100 ppb, toluene from 21,000 to 940 ppb and xylenes from
17,000 to 3,200 ppb.
Guidelines on the Use of Innovative Technologies
New and innovative assessment and remediation technologies
are continually being developed and proposed for sites throughout
North Carolina. The question is frequently raised about how they
should be dealt with from regulatory and trust fund reimbursement
standpoints. In general, the Groundwater Section encourages the
use of new and innovative approaches and technologies for
assessment and remediation. However, the State Trust Funds should
not be used to finance research and development of untried and
untested methods and technologies.
-------
The following guidelines were formulated to address the use
of new and innovative technologies for site assessment and
remediation activities. It treats the approval and cost
reimbursement of innovative technologies with a documented history
of successful application or testing differently than technologies
which are untried or untested.
Assessment
Current assessment technologies and investigatory methods are
effective, but substantive improvements may be possible. The use
of new technologies and methods may increase the amount of data
available for the money spent or may provide a more accurate
result. Because the assessment is the basis for all subsequent
actions taken at a site, the use of new and innovative approaches
and technologies that could provide better or more complete
information should be encouraged. At the same time, these new
methods should be approached cautiously.
A consultant, vendor or tank owner proposing a new site
assessment technology should demonstrate that it has a reasonable
probability of providing additional or more reliable information
than conventional methods or will provide equivalent information
at equal or lesser cost. The following information should be
provided:
1.	A description of the tecnnology's operation.
2.	Detailed results of previous site testing (either full
or pilot scale) which describe the characteristics of
site(s) with respect to geology, hydrogeology, and
contaminants present.
3.	A summary of the technology's applicabilities and
limitations due to site and contaminant characteristics,
and onsite or offsite interferences.
4.	Identification of the financial or technical benefits.
If based upon the above information the consultant
demonstrates that the new technology has been successfully used or
tested on other sites with similar characteristics and
contaminants, and that the technology is applicable for the
proposed site, the new method may be approved for use as a
stand-alone method. As long as the technology is used in an
efficient manner, meeting the criteria of "reasonable and
necessary" as applied to other assessment technologies, the State
Trust Fund would reimburse the activity.
If an approach appears to have merit, but there has not been
adequate testing conducted to make the above demonstration (e.g.
it has had limited testing or use, but in different geologic
settings or with different contaminants), it should only be
permitted as a supplemental method or technology. The use of
untested assessment methods and technologies should be relegated
to a duplicative role only. The cost of using an untested method
-------
in this role should be the risk of the consultant or responsible
party.
Remediation
New technologies or methods which can be demonstrated to
increase the effectiveness of a remedial action or may reduce the
cost should be encouraged, as long as they do not reduce the
ability to control the threats to human health or the environment.
In essence, new methods and technologies should only be
approved on a trial basis and only if the consultant can
demonstrate that it has a high probability of being as effective
as other methods and at a cost equal to or less than conventional
methods. All cost comparisons should assume an equal duration of
cleanup activities, unless substantial information to the contrary
is provided. The demonstration must include use at other sites or
pilot scale testing and the consultant must provide the following
information:
1.	A description of the technology's operation.
2.	Detailed results of previous site testing (either full
or pilot scale) which describe the characteristics of
site(s) with respect to geology, hydrogeology, and
contaminants present.
3.	A summary of the technology's applicabilities and
limitations due to site and contaminant characteristics,
and onsite or offsite interferences.
4.	Identification of the financial or technical benefits as
compared to conventional methods.
Any approval of a new approach or technology should be as an
"interim" remedial measure only until it has been shown to be
effective on that particular site. Such interim measures should
be reviewed^ and approved (or denied) on a site-by-site basis for
limited periods. The length of this interim approval will vary
based on the technology and site specific conditions; generally a
12-month operational/monitoring period under interim approval with
quarterly review of the data will be appropriate.
The number of sites for which interim approval of a specific
innovative method or technology is given should be limited until
the Department has some real knowledge that this approach will
have the benefits at least equal to those already offered by
existing methods and technologies; generally, this interim
approval should be limited to three (3) sites statewide until
monitoring data show that the technology is working. A remedial
action technology which obtains "interim" approval would be
eligible for reimbursement of reasonable and necessary costs
through the end of the trial period and beyond if it proves to be
successful and cost effective.
Before any methods or technologies are incorporated into an
-------
review of the effectiveness and applicability of the proposed
method or technology to the specific site. In addition, no
approach or technology (new or existing) should be assumed to be
universally applicable. The technical staff should use their
scientific and engineering knowledge and experience to evaluate
the methods and technologies proposed for each site. A given
technology should be approved in a CAP only where it was used as
an interim remedial measure or where the conditions are obviously
equivalent to such a site in all critical parameters.
Because the Department's primary mandate is to protect human
health and the environment, care should be exercised to minimize
the potential for additional threats when approving any
technology. Untried remedial approaches and technologies should
only be allowed at sites where a delay in effectively cleaning up
does not significantly increase the threat to human health. The
cost of using an untested remediation technology should be at the
risk of the consultant or responsible party. The State Trust Fund
will not reimburse costs associated with use of this technology
unless it proves to be successful. To be considered successful ,
the innovative technology must be shown to remediate the site to,
at least, the extent that an appropriate conventional technology
would have, at equal or lesser cost.
-------
-------
DEVELOPING A CONCEPTUAL FRAMEWORK AND RATIONAL GOALS FOR
GROUNDWATER REMEDIATION AT DNAPL SITES
J.A. Cherry, S. Feenstra, and D.M. Mackay
Waterloo Centre for Groundwater Research
University of Waterloo
Waterloo, Ontario
N2L 3G1 Canada
For Publication in Proceedings of:
Subsurface Restoration Conference
Third International Conference on
Ground Water Quality Research
Dallas, Texas, June 21-24, 1992
DRAFT
-------
ABSTRACT
Many tens of thousands of sites with DNAPL in groundwater exist across North America
including most Superfund sites, many RCRA sites, and many military and municipal lands.
Chlorinated solvents are the most common DNAPL problem. Few, if any, DNAPL sites have
been permanendy restored although many attempts have been made using pump-and-treat.
DNAPL-contaminated aquifers have two segments: the source zone where immiscible-phase
liquid is the primary cause of contamination, and the plume where contamination is dissolved
and sorbed and where immiscible-phase liquid is absent. This paper categorizes technologies in
the context of source zones and plumes. It considers technologies that are proven and those that
are experimental, for containment and also for restoration. A conceptual framework is presented
for remediation tasks and goals. Emphasis if directed at the potential role for relatively passive
technologies for containing source zones and plumes without need for pump-and-treat. Except
for cutoff wall enclosures, passive technologies are new, including permeable treatment curtains
and funnel-and-gate systems that can be placed across plumes/5r downgradient of source zones.
The challenge facing these new technologies for containment of dissolved contamination is
considerable but not near that required for source-zone restoration, which is an elusive goal
because essentially all of the immiscible-phase must be removed to restore the source-zone part
of the aquifer to drinking water use. The development of cost-effective source removal
technologies will be slow, arduous, and expensive. Thus, we are transferring into the next
century the task of operating and maintaining numerous active containment systems in order to
prevent further growth or spread of contaminants from subsurface source zones. Although the
transfer of this responsibility is presently unavoidable, today we have an opportunity to enhance
-------
possibilities for future remedial efforts by a systematic advance of scientific and engineering
knowledge of the processes and performance of remediation technologies.
-------
INTRODUCTION
Remediation or remedial measures conducted at .sites of subsurface contamination are
engineered activities designed to protect present or future human health and maintain or improve
environmental quality. Subsurface contamination may occur as soil contamination (ie.,
contamination in the vadose zone) or groundwater contamination (ie., contamination below the
water table) or both. Subsurface contamination may present a variety of types of risks including
risks via groundwater usage, surface water usage, soil contact, vapour inhalation, or food
consumption. Any of these risks may provide cause for subsurface remediation. However, the
nature of the subsurface problem and the strategy for remediation depends greatly on which of
these risks are significant and what degree of risk reduction or environmental protection is
required.
This paper provides a conceptual development of subsurface remediation goals, that are
rational and consistent with our present understanding of subsurface contamination caused by
industrial organic liquids, particularly organic chemicals thayare immiscible and heavier-than-
water. These liquids are commonly known as DNAPLs (dense non-aqueous phase liquids) and
commercial, industrial, and waste disposal sites where DNAPL is the major cause of
groundwater contamination are referred to as DNAPL sites. In the 1980's, it became evident in
North America and Europe that portions of many aquifers had become contaminated by a variety
of chemicals of industrial, municipal, or agricultural origin, particularly nitrate, chloride and
sulphate salts, metals, and industrial organic chemicals. Most sites that have received, or will
soon receive, major remediation efforts are afflicted with DNAPLs, the most common of which
are chlorinated solvents. Other important DNAPLs are: creosote, coal tap, PCB oils, some types
-------
of mixed organic industrial wastes and some pesticides in pure-product form. Chlorinated
solvents (such as trichloroethylene, tetrachloroethylene, trichloroethene, and methylene chloride)
have been used by nearly all manufacturing industries in North America and continue to be used
by many. Also, chlorinated solvents have been used for many decades in most dry cleaning
facilities. They are much more common than was thought to be the case a few years ago. Most
Superfund sites, most military facilities, and many RCRA sites are in this category. Many of the
U.S. Department of Eftergy's sites have DNAPL in the subsurface along with other types of
contaminants. Realization that DNAPLs are common at industrial and military sites came
gradually over a decade, evolving from the interpretations of site conditions by a few
groundwater scientists in the earl$f 1980's to broad recognition by the groundwater profession
by the early 1990's. Although the occurrence of DNAPLs at many sites across the continent is
now recognized, there is currently no consensus on how to deal with this immense groundwater
problem. DNAPL sites pose much greater technical problems and financial burdens than other
types of sites. Meeting regulatory cleanup criteria is elusive at most DNAPL sites. This paper
is intended to provide a structural scientific framework for conduct of debate on this issue.
Figure 1 shows schematic diagrams for three different conceptual models for groundwater
contamination caused by industrial chemicals. In the simplest model (Figure la), plumes of
contaminated groundwater are caused by leaching of contaminants from the soil or waste situated
above the water table. Permanent restoration of such a site can be relatively simple, though not
necessarily inexpensive: remove the contaminated soil or waste by excavation or remove the
contaminants from the soil or waste by in situ methods, and then extract the plume by
groundwater pump-and-treat. It is now clear that this model fits only a very small proportion of
-------
those industrial sites where organic contaminants exist in groundwater. In the second model
(Figure lb), petroleum products such as gasoline, fuel oil, or kerosene (ie., LNAPLs - light non-
aqueous phase liquids or immiscible liquids less dense than water), exist above the water table
or in the zone of water table fluctuation. This subsurface mass of immiscible liquid emits
dissolved contaminants to groundwater and soil water, causing a plume to form. Contaminant
plumes from LNAPLs are usually not large mainly because natural microbiological degradation
of LNAPL-derived constituents in the plumes usually restricts growth of the plumes. Also, the
fact that LNAPLs do not penetrate much below the water table tends to limit LNAPL-derived
plumes to relatively shallow depths and therefore to restrict spread of plumes because shallow
groundwater usually encounters local hydrologic boundaries. This is not to say that some plumes
from LNAPL sources do not cause major problems, but rather that the problems are generally
much more limited in geographic extent than other types of plumes, particularly plumes at
DNAPL sites. To permanently clean an LNAPL site, the mass of immiscible-phase liquid in the
subsurface, above and below the water table, must be removal by excavation, or by extraction
processes and in- situ destruction and the plume must be removed by pump-and-treat or in situ
degradation.
The most difficult site to remediate is represented in Figure lc. The primary cause of
contaminant at DNAPL sites is slow dissolution of the mass of immiscible organic liquid situated
below the water table. In Germany, Frederich Schwille recognized the importance of this type
of groundwater contamination in the 1970's. Schwille (1982, 1984) conducted intensive
laboratory research on chlorinated solvents in groundwater. Schwille's work was not known in
North America until 1984. Villaume (1983) published the first North American paper describing
-------
DNAPL (creosote) in groundwater, but it was not until the mid- to late-1980's that the
conceptual model displayed in Figure lc and its extension into a DNAPL paradigm became
known amongst groundwater scientists and engineers in North America (Cherry, 1984; Mackay
et al., 1985; Mackay and Cherry, 1989; Feenstra, 1984). Specific consideration of DNAPL sites
was not included in the late 1980's edition of the U.S. EPA guidance document for remedial
investigations of Superfund sites (U.S. EPA, 1988). It was not until the early 1990's that the
general implications of DNAPL for site monitoring and remediation were recognized by the U.S.
Environmental Protection Agency (U.S. EPA, 1991). Such official recognition has not yet been
made by regulatory agencies in Canada, Mexico, or Europe. The fact that it took about a decade
for the new paradigm to move faim creation by a small segment of the scientific community to
widespread recognition is not surprising. Paradigm shifts in science and engineering often take
a decade or more. Unfortunately, the large and costly effort of site investigation and subsurface
remediation undertaken in the United States in the 1980's and continuing in the 1990's is based
primarily on concepts for site conditions represented by non-DNAPL causes of contamination
(Figures la and lb). DNAPL sites typically have groundwater contamination that is deeper and
much more widespread than is common at other sites. The main hazardous organic contaminants
in groundwater at DNAPL sites tend to be more persistent than those at LNAPL sites.
Some industry; sites have very complex conditions of subsun. ce contamination better
represented by a combination of the three conceptual models shown in Figure 1 rather than any
one of the models. At these sites, the DNAPL part of the problem is usually deeper in the
subsurface and most difficult to remediate permanently.
-------
REMEDIATION CONCEPTS AND TERMINOLOGY
Groundwater remediation technologies can be grouped in two general categories: proven or
experimental. To be classified as a proven technology, a technology must have been applied
previously at a sufficient number of sites so that a base of general knowledge has been developed
to allow detailed design and reliable estimation of the cost and time required to achieve a
specific remedial goal. Experimental technologies are those for which there is an expectation of
effectiveness based on conceptual models, computer models, laboratory studies, or field trails,
but for which performance cannot yet be predicted reliably.
Another useful categorization of remediation technologies distinguishes between those that
are active or passive. Active technologies are those that require a considerable amount of
ongoing engineering activity (operations, monitoring, and maintenance) to maintain adequate
performance of the technology. Conventional groundwater pump-and-treat is an example of an
active technology. In contrast, passive technologies are those that are put in place and
subsequently require little or no ongoing energy inputs or maintenance other than performance
monitoring. Low-permeability caps and covers to minimize infiltration and plume generation are
examples of passive technologies. Active technologies have the disadvantage of transferring
significant operating costs to future generations, whereas passive technologies minimize such
transfer.
Subsurface chemical masses that emit persistent groundwater plumes are referred to as
sources or source zones. Commonly, the most important sources at industrial sites are masses
of heavier-than-water immiscible industrial liquids (DNAPLs) located below the water table
(Figure 2). However, old drums or tanks containing chemicals or near-surface soils or sludges
-------
may also be active sources of groundwater contamination. At many sites, even when the drums,
tanks, sludges, and soils are removed, the DNAPL that remains in the subsurface represents a
significant continuing source of groundwater contamination. Until this subsurface source is
removed, the plume will continue to receive contaminants leached from the DNAPL.
Considering the relatively low solubilities of DNAPL chemicals and low drinking water
standards for many DNAPL chemicals, plumes may be generated in this fashion for very long
periods of time, even when the mass of DNAPL in the subsurface source zones is relatively
small. Removal of the subsurface source zones to a degree that prevents continued formation of
the plume is by far the most difficult task in aquifer restoration. The DNAPL mass that exists
below the water table typically represents a much greater groundwater problem than any buried
drums or buried sludge or soil that may exist at the site. The concept of the source of
groundwater contamination at a DNAPL site represents a dichotomy: for groundwater scientists
and engineers focused on the technical matters of groundwater remediation, the most significant
'source' is generally the mass of immiscible liquid residing below the water table, whereas for
the regulatory or the legal community, the 'significant source' on the property is commonly the
buried drums or sludge that still remains. The buried drums and sludge are typically all above
the water table. The regulatory or legal viewpoint is what governs most remedial actions at
industrial sites. Thus, at many DNAPL sites, particularly Superfund and RCRA sites, large
remedial effort is commonly directed at the drum or sludge 'source' and little effort at the
below-water table DNAPL source. Thus, at many DNAPL sites, the dichotomy in the concept
of 'source' results in 'source remediation' for which benefits in risk reduction and improvement
in groundwater quality may be minimal even after large expenditures have been made.
-------
In the groundwater context, risk can be reduced or eliminated by controlling the migration
of the aqueous-phase and vapour-phase plumes to achieve plume containment. A fully contained
plume leads to no further expansion of contamination in the groundwater zone. Thus, around a
contained plume, no additional portion of the groundwater resource is becoming contaminated.
At many, if not most, DNAPL sites, full containment is not achievable without engineered
controls such as groundwater pump-and-treat (Figure 3a). In order to eliminate or reduce risk
to an acceptable level, remedial measures for plume containment must continue for as long as
the subsurface source zone persists. This typically requires long-term application of technologies
such as groundwater pump-and-treat. In some cases, the condition offull containment is achieved
by natural site circumstances. An example would be a plume that emanates from the DNAPL
source and discharges into a nearby stream (Figure 3b). Although the plume may cause surface
water contamination, the portion of the contaminated aquifer does not expand, and therefore in
the context of the groundwater environment, the plume is contained naturally.
Aquifer restoration is a task much different from plume containment. Aquifer restoration
requires source~zone containment or source removal as well as plume removal to a degree
sufficient to provide for the original beneficial use of the aquifer. This beneficial use is
commonly drinking water supply. A fully restored aquifer is one that requires no further source
zone containment or plume containment (Figure 4). Of the thousands of contaminated aquifers
in North America, some have ongoing successful measures for plume containment, but very few
have been fully restored to drinking water use because of difficulties with source-zone
restoration. We know of no significant DNAPL site where full aquifer restoration has been
achieved and substantiated with detailed monitoring data. Claims of full aquifer restoration have
-------
generally not been confirmed by adequate testing, or the cleanup standard has been lenient, not
requiring restoration to drinking water use.
At the present time, the only remedial technologies that should be regarded as proven
technologies for most organic industrial chemicals in most hydrogeologic settings are: (i)
groundwater pump-and-treat for containment of groundwater plumes and source zones (but not
source removal); (ii) vertical cutoff walls coupled with hydraulic control inside the enclosure for
source zone containment, and, in some cases, coupled with caps for source containment; and (iii)
soil vapour extraction for containment of vapour plumes in the vadose zone, but not for complete
removal of all significant contaminant mass from the vadose zone.
On-going elimination of risk isCall areas outside of the source area and plume and prevention
of further degradation of environmental quality can generally be achieved by plume containment
without the necessity of significant reduction in contaminant mass in the source zone. It is
possible to perform cost-benefit analyses for pump-and-treat for plume containment if the risks
associated with the plume are known. In contrast, the lack of proven in situ technologies for
source removal, and hence for full aquifer restoration, renders cost-benefit analysis for these
technologies impossible because the ultimate cost of remediation cannot be defined. The difficult
and unresolved issue in groundwater remediation therefore pertains to the degree of effort and
the level of financial resources that should be directed towards the presently elusive goal of full
aquifer restoration. This goal is elusive because of the extreme uncertainty in the performance
of remediation technologies to achieve common regulatory cleanup criteria. This uncertainty
derives from the limitations inherent in the technologies and from the complications for
technology performance imparted by subsurface geologic heterogeneity. A technology that
-------
performs well in a simple geologic site situation can be entirely inadequate in a complex
geologic setting. There are many more complex sites than simple sites.
A partially restorea aquifer is an aquifer in which the segment of the plume between the
plume front and the downgradient periphery of the source zone (zone between A and B, Figure
5) has been removed, while the source zone is contained by engineered means such as cutoff
walls or pump-and-treat to prevent dissolved contaminants emitted from the source from causing
regrowth of the plume. Thus, partial restoration of the aquifer depends on long-term operation
of the source-zone containment system. Partial aquifer restoration makes available for use, the
portion of the aquifer in the area beyond the source-zone control system. Partial aquifer
restoration can be accomplished by one of two approaches: (1) plume removal by pump-and-treat
(Figure 5), or (2) plume renovation by natural-gradient groundwater flushing (Figure 6). Plume
removal by pump-and-treat enables the plume to be removed more rapidly (Figure 6a), whereas
plume renovation relies on natural groundwater flow to flush the plume from the aquifer. Once
the source zone containment system is functioning, the pluitfe exists only temporarily if there
is an engineered or natural discharge point. After the plume separates from the source zone, it
is flushed along the natural groundwater flowpaths toward points of groundwater discharge at
streams, rivers, wetlands, or water supply wells. Under these constraints; natural-gradient plume
flushing results in a flux of contamination to the river, lake, or water supply well for a finite
period of time.
The cumulative contaminant flux is the total initial contaminant mass in the plume (aqueous
and sorted phases) emitted from the groundwater flow system at the points of discharge over
the length of time of the emission. This emission at points of discharge is typically small. Once
-------
separated from the source zone, plumes normally have a small total aqueous and sorbed mass.
Mackay and Cherry (1989) provide examples of estimates of total aqueous phase contaminant
mass in several large plumes in sand or gravel aquifers. In some cases, part or all of the plume
mass may disappear due to biodegradation or abiotic degradation prior to arrival at groundwater
discharge points. In many cases, the point-of-discharge emission causes insignificant
environmental impact, and/or insignificant risk to actual or potential receptors. Plumes from
many LNAPLs and DNAPLs, particularly chlorinated solvents, only rarely cause detectable or
significant impact on rivers or lakes. Therefore, isolation of the source zone followed by natural
flushing for plume removal may be a reasonable option in many cases when decision-making can
be governed solely by issues of science and economics. In some cases, the emission involves
significant risk to receptors or causes a significant loss in use of water resources, surface water,
groundwater, or both. In such cases, these risks can be prevented by plume containment or
plume removal.
Full-scale plume-removal is the higher cost remedial option when compared to natural-
gradient plume renovation or plume containment. In full-scale plume removal, expenses accrue
from capital costs and operating and maintenance costs for the network of pumping wells and
water treatment and for performance monitoring. For natural-gradient plume renovation, the
main expense is monitoring to ensure that the aquifer renovation rate and impact at discharge
points such as rivers or lakes are within acceptable limits; the required source-zone containment
can usually be accomplished using only a small number of wells and a relatively small water-
treatment system or cutoff wall enclosure around the source zone with little or no pumping.
-------
At most sites where industrial organic contaminants are a groundwater problem, the majority
of the expense for groundwater pump-and-treat is due .to capital and operating costs for the
portion of the system needed for full-scale plume removal rather than the portion needed for
source containment. Therefore, selection of the full-scale plume removal option rather than the
natural-gradient plume renovation option coupled with source containment is a selection of a
higher-cost, lower-risk plan over one with a much lower cost and higher temporary risk. For
the selection to be done in a cost-benefit context, it is necessary that remedial designs specific
to each of these options be compared. Some pump-and-treat systems put into operation during
the 1980's were compromises that accomplished neither full containment of the source zone nor
removal of the plume.
A third option, intermediate between the two described above, is natural flushing with
capture. In this option, the source zone is contained (Figure 6a) and the front of the plume is
captured. The plume, thus detached from the source zone, is allowed to flush under only slightly
accelerated above natural gradient conditions to the downgradi^ftt capture system. This is a lower
cost option than the full-plume removal option because of less pumping and lower water
treatment costs. This option accomplishes the same goal as the full-plume removal option but
it takes longer for the plume to disappear.
TASKS IN PLUME CONTAINMENT AND AQUIFER RESTORATION
Remediation of a DNAPL or LNAPL site can be considered on the basis of four
fundamental tasks: plume-front containment, source-zone containment, plume removal, and
source removal. The first three of these tasks can be undertaken to accomplish partial aquifer
-------
restoration but all four must be applied for an attempt at full aquifer restoration (Figure 2).
Conventional pump-and-treat is a proven technology for accomplishing any of the first three of
these tasks. Pump-and-treat generally accomplishes plume front or source zone containment with
minimal difficulty if applied properly. Pump-and-treat is a viable technology for plume removal,
but at some sites, subsurface complexity causes large uncertainty in the time and cost to achieve
the specified degree of plume removal. Thus, long-term cost estimates for this task are generally
inaccurate or unreliable.'Groundwater pump-and-treat is generally inadequate to accomplish the
fourth task, source removal. Although conventional pump-and-treat can perform effectively for
three of the tasks, it is not inexpensive and involves substantial ongoing costs because of energy
use, maintenance, and monitoring.-^Therefore, there is incentive to develop less expensive and
generally more passive means of accomplishing these tasks. Experimental technologies aimed
at this goal are described briefly below.
Plume Front Capture
Most dissolved plumes are too large to be enclosed practically by cutoff wall barriers. The
goal of plume-front containment is to capture the entire mass flux at the front of the: plume
during the entire duration of the flux. Contaminant fluxes at the plume front are commonly
small, requiring a low annual rate of contaminant mass removal.Three potential engineering
options for capture or containment of the plume-front are 1) conventional pump-and-treat, 2) in
situ treatment curtains, and 3) in situ funnel-and-gates with in situ or ex-situ treatment (Figure
7). The options for plume-front containment are nearly the same as those for source-zone
-------
containment with the exception of cutoff wall barriers that can be used effectively at the source
zone in many cases but generally not at the plume front.
An in situ treatment curtain is a reactive permeable material placed as a wall vertically
across the plume (Figure 7b). As the plume passes through the curtain, physical, chemical, or
microbial processes remove the contaminants from the plume. An effective treatment curtain
avoids the need for removing water from the aquifer to achieve plume cutoff. The in situ
treatment curtain is currently an unproven or experimental technology. Various versions of the
technology for many contaminant species are being subjected to research and development.
Bums and Cherry (1992) describe the status of this research. Gillham and Burns (this volume)
provide a more recent review. Recently, prototypes of the technology have been installed at a
few field sites.
A funnel-and-gate system, also known as a wall-and-gate system, uses segments of
impervious wall to direct groundwater flow to engineered gaps in the wall, referred to as gates,
through which the plume is channelled (Figure 7c). As the #ater flows through the gate, the
contaminants are removed by comparable types of in situ treatment processes used in the in situ
treatment curtains (Figure 7d). Plume containment may be achieved by a single funnel-and-gate
placed across the entire plume, or if the plume is relatively wide, by a series of funnel-and-gates
across the plume. Intensive research on in situ treatment curtains and funnel-and-gate systems
began only recently. Starr and Cherry (1993) describe the hydraulic aspects of funnel-and-gate
systems. The first prototype funnel-and-gate system was installed at a field site in Canada in late
1992. The funnel-and-gate configuration for in situ treatment has an advantage over the curtain:
ease of installation and possibility for repair or rejuvenation.
-------
There are three general categories of in situ treatment zones, regardless of whether they are
positioned as curtains or in gates: 1) physical systems (e.g. air sparging); 2) chemical (abiotic)
systems, and 3) microbial systems (microbial filters). There are various types of air sparge
systems for containing or cutting off plumes, three of which involve pumping air into or through
a line of wells positioned in a row across the plume. In the other air sparge system, air is
sparged upward through the plume as it passes through gates in a funnel-and-gate system
(Pankow et al., 1993). In this type of funnel-and-gate, the gate is an open water column with
no porous medium packed in it. The injected air strips volatile organic contaminants from the
open water column in the gate. In each of the air sparge systems, air is used to strip volatile
contaminants from the groundwater in situ. The contaminants are swept to the vadose zone
where they are degraded or to ground surface where they are released to the atmosphere, or
collected and treated. In direct contrast to air sparge systems, which generally bring the
contaminants to land surface, the chemical or microbial treatment zones are intended to destroy
the contaminants in the treatment zone as the plume passes through it.
Source-Zone Containment
Source-zone containment may be accomplished by: (1) conventional pump-and-treat; (2)
cutoff wall enclosure with or without interior pump-and-treat; (3) in situ treatment curtain or
funnel-and-gate system immediately downgradient of the source zone; or (4) cutoff wall
enclosure with in situ treatment gates as outlets from the enclosure (Figure 8). The first two
approaches (Figures 8a and b) indicated above arc proven technologies for many sites, whereas
the others have components that are experimental. When conventional pump-and-treat is used
-------
for source zone containment, the pumping system need only capture the contaminant mass flux
emanating from the source. It is not necessary to remove mass at a rate that depletes the source
mass at a significantly higher rate than what is needed for source zone control. For source-zone
containment, as is also the case for plume-front containment, the required mass-removal rate
needed may be relatively small. In some circumstances, hydrogeologic complexities and
contaminant mass distributions are such that the only viable approach for source-zone
containment is pump-and-treat. In other situations, cutoff wall enclosures are a proven
technology. The choice can depend on total cost,-cash-flow factors, and future remedial
intentions. Pump-and-treat involves lower capital cost (design and construction of the water
treatment facility) and higher annual operating and maintenance costs than cutoff wall enclosure.
Cutoff wall enclosures will have a higher capital cost but may provide the opportunity for educed
operating costs resulting from lower groundwater pumping rates. The construction of a cutoff
wall enclosure may also facilitate further application of in situ remedial technologies such as
surfactant flushing.
Most DNAPLs are much more dense than water and, at many DNAPL sites, the bottom of
the immiscible-source mass is much deeper than non-DNAPL chemical substances would have
penetrated. At some sites, the bottom of the immiscible source mass is hundreds of metres deep,
in some cases deep into bedrock. Conventional cutoff walls generally cannot be installed to such
depths. At many sites, the bottom of the DNAPL has not yet been determined. In some cases,
shallower cutoff wall enclosures can be used to isolate the upper part of the source mass with
minimal pump-and-treat, leaving the deeper part to be contained by pump-and-treat.
-------
Cutoff wall enclosures can be installed in fractured bedrock at most sites and to greatest
depth by injection grouting. This technique involves injection of sealant (ie., grout) into closely
spaced boreholes to achieve sealing of fractures. The technique has been used for many decades
in the geotechnical industry to reduce groundwater seepage at dam sites and other large-project
construction sites. There is no doubt that considerable reduction in bulk hydraulic conductivity
is achievable in nearly all types of fractured rock, but often at very high cost. The degree of
hydraulic-conductivity reduction is generally much more difficult to predict and verify than is
the case for conventional cutoff wall enclosures in overburden or at shallow depth in soft
bedrock. As is the case for cutoff wall enclosures in overburden, decisions regarding selection
of cutoff wall enclosures by borehofe grouting should be founded on issues of cost and benefits.
There is no specific hydraulic conductivity that must be achieved in the cutoff wall enclosure
in order to be effective. Lower bulk hydraulic conductivity of the wall provides lower leakage
through the wall. Consequently a lower pumping rate inside the enclosure is required to
maintain a hydraulic gradient directed inward across the wall. The inward gradient prevents
outward lateral advection of dissolved contaminants. Lower pumping from the interior results
in lower long-term cost of water treatment. The long-term cost of water treatment "can be
reduced further by capping or diverting surface runoff in the area inside the enclosure to
minimize groundwater recharge. In some cases, much of the water entering the enclosure is
upward seepage from below. Upward seepage at the bottom of the enclosure is generally
desirable to prevent leakage of dissolved contaminants downward into deeper strata. However,
in some cases, natural downward hydraulic gradients may be too large to allow gradient reversal
to be imposed by pumping.
-------
There are several technologies available for construction of low-permeability cutoff-wall
enclosures (described by Mitchell, this volume). These technologies are proven technologies in
the sense that they can achieve specific, relatively low permeability values (less than 10"6 or 10"7
cm/s). There is very little published information pertaining to how long these barriers will
maintain low permeability. There is a need for research in the longevity of low-permeability
barriers. Many decades or centuries are generally expected. In the event that the permeability
of the wall increases, the consequence would be an increase in the rate of pumping from the
interior of the enclosure. If the increase in permeability of the wall is large, pump-and-treat
inside the enclosure evolves towards the conventional pump-and-treat option without a barrier.
The alternative would be to replace or repair the barrier. Repair of most vertical types of
barriers is generally a much easier task than the initial construction of the barrier. At sites where
no NAPL occurs, or where LNAPL but no DNAPL occurs, the depth to which the cutoff wall
must be installed is generally not great. The hydrogeologic nature of the bottom of the
subsurface source zone at these sites is less important than at DNAPL sites. For cutoff wall
enclosures to beJully effective at DNAPL sites, it is generally necessary for the bottom of the
wall to extend somewhat below the deepest zone of DNAPL. If the input of DNAPL to the
groundwater zone has ceased prior to construction of the wall, it is expected that the DNAPL
has already gone as deep as it will go under the existing conditions. It is important that the wall
be installed and the hydraulics of the enclosure be operated in a manner that DNAPL
remobilization is avoided. Pumping within the enclosure that dewaters the DNAPL zones may
cause further migration of the DNAPL downward.
-------
For source-zone containment, an alternative to combining cutoff-wall enclosures with
interior pumping is to place in situ treatment gates in the cutoff wall and thereby avoid interior
pumping. The water that enters the enclosure by infiltration of rain or snowmelt and by seepage
into the enclosure has contaminants removed when it flows through the gates at the downgradient
side of the enclosure. If the enclosure has a sufficiently low permeability , and if surface runoff
from the source zone area is diverted, the flux through the gates will be small because
groundwater recharge would be small. For most treatment processes within the gates, treatment
efficiencies will be greater flow low water flux rates through the gate. This fully passive
approach to source-zone containment is less suitable for sites where there is downward
contaminant migration out of the bfcttom of the enclosure.
Of the many experimental technologies being developed for groundwater remediation, those
that will accomplish plume containment and source-zone containment most passively and at
lower cost than conventional pump-and-treat offer the best prospects for advancing in the next
few years to the status of proven technologies. The technical challenge associated with
restoration of source zones at DNAPL sites is much greater than that of source-zone containment
or plume capture. The technical limitations imposed by geologic complexity are severe for
source-zone restoration, but commonly much less severe for source-zone containment or plume
containment. Several experimental technologies now in the research stage, such as surfactant
flush, alcohol flush, and chemical oxidant flush, are directed at removal of or in situ destruction
of contaminant mass from source zones. A secondary advantage provided by cutoff wall
enclosures is enhancement of the potential for more effective and environmentally safe
application of the above-mentioned in situ flush technologies. The enclosure can effectively
-------
prevent lateral escape of chemical reagents or aqueous contaminants mobilized during
immiscible-phase removal. Also, the enclosure minimizes the mass of chemical reagent that must
be added for effective enhancement of the mass removal or mass destruction processes. The
source zone situated inside the enclosure can be manipulated in a manner analogous to a very
large batch reactor rather than a groundwater flow system.
Plume Removal
Plumes can be removed using arrays of pumping wells positioned according to the well-
established principles of aquifer hydraulics and solute transport. The initial mass of aqueous
phase contamination in the plume can be removed by pumping but requires removal of much
more groundwater than that equivalent to the volume of the plume. Additional plume volumes
must be pumped to remove the mass of adsorbed-phase contamination and the mass of dissolved
and adsorbed phase stored in lower permeability zones in the plume and in aquitards above or
below the aquifer. Mathematical models can be used to determine optimal well positioning and
pumping rates to remove the plume in the shortest possible time. However, predictions of the
time required to achieve regulatory cleanup criteria are commonly unreliable due to processes
which generally cannot be accounted for rigorously by the models (Figure 10), such as
desorption effects, delayed yield from the capillary fringe, and diffusion out of low permeability
zones.
Possibilities exist for enhancing plume removal by combining pump-and-treat with aquifer
heating, chemical flushing, or microbial processes. However, we can expect that plume removal
is a task for which these new technologies may not provide much improvement in effectiveness
-------
or efficiency over conventional pump-and-treat. These enhancement processes for plume removal
are intended to stimulate removal or destruction of the initially adsorbed portion of the plume
mass. Efficiency of release of the adsorbed phase is expected to be inhibited at many sites by
geologic complexity and slow diffusion from low permeability zones.
Source Zone Restoration
Nearly all of the contaminant mass in the source zone is normally immiscible-phase organic
liquids relative to the other forms of contaminant mass: the dissolved and sorbed mass. If the
entire plume mass, excluding the source mass, is removed and if control of the source zone is
not maintained thereafter, the jtfume will reestablish itself due to continued dissolution of
immiscible phase into the groundwater. Conventional pump-and-treat is generally ineffective for
removing much of the immiscible phase mass because the mass removal rate is insufficient due
to dissolution and solubility constraints. Hydraulic systems such as pumping wells, and drainage
trenches, with or without water injection, can in some situations remove considerable
immiscible-phase mass, however, most of the immiscible-phase mass remains in the aquifer after
these systems have reached their effectiveness limit. Thus, to accomplish source mass reduction
beyond what can be accomplished by hydraulics alone, enhancement of the removal rate is
required by means of experimental technologies such as chemically enhanced pump-and-treat,
or other approaches such as steam flush or in situ chemical destruction of the NAPL.
Chemically enhanced pump-and-treat for removal of the immiscible-phase involves flushing
the DNAPL zone in the aquifer with water containing chemical additives or reagents (Figure
11). Research currently focuses on two categories of chemical additives: surfactants and alcohol.
-------
Surfactants added to water enhance the effective solubility of the DNAPL so that the contaminant
mass-removal rate by the surfactant-enhanced pump-and-treat is much greater than that of
conventional pump-and-treat (Wunderlich et al., 1992). For cost-effectiveness, it is generally
necessary to remove the contaminants from the mixture of water, surfactants, and contaminants
in above-ground facilities so that the surfactant solution can be used for repeated aquifer flushes.
The use of surfactants for removing DNAPL has recently progressed from laboratory studies to
small-scale field prototype tests of flush systems.
For alcohol flushing to be effective for removal of immiscible phase DNAPL, flushing must
be done with alcohol mixed with only a minor fraction of water. The cost of the alcohol is an
important factor and therefore the efficiency of each flush and potential for reuse of the alcohol
are critical factors. Alcohol flush for DNAPL removal is currently being assessed in laboratory
studies and has not yet reached the field prototype stage.
Another technology for flushing DNAPL from aquifers relies on the use of steam. Steam
injected into the source zone drives the contaminant in the inyniscible phase as well as in the
vapour phase to withdrawal wells. Udell and co-workers have advanced the understanding of the
physics of steam flush, using theoretical and laboratory studies, as well as small-scale field
trials. The effectiveness of steam flush for removing NAPL from below the water table is
currently being assessed by field prototype trials.
The flush technologies described above have in common the process of moving the DNAPL
mass from the aquifer via wells to above-ground facilities for treatment or disposal. An
alternative approach is to flush the source zone with chemicals that destroy the DNAPL in situ.
Schnarr and Farquhar (1992) conducted laboratory and prototype field trials at the Borden site
-------
in Canada of in situ chemical destruction of immiscible-phase chlorinated solvents. Additional
field trails are in progress. The destruction is accomplished using a strong chemical oxidant:
potassium permanganate.
There are several factors that can severely limit the effectiveness of mass removal from
source zones using surfactants, alcohol, steam, or chemical oxidants. Firstly, the injectant must
be very efficient for removal or destruction of the immiscible phase chemicals when brought in
contact with these chemicals. Secondly, if pump-and-treat is used as part of the flush system,
there must be effective facilities above ground for treating, recycling, or disposing of the
injectant and reaction products. Thirdly, the injectant chemicals and/or hazardous reaction
products from the in situ processes^must be prevented from escaping from the remediation zone
into other parts of the aquifer. Finally, the injectant must be brought sufficiently in contact with
the chemicals to enable the removal or destruction processes to be effective.
These latter two problems generally pose exceptional problems. At some sites, geologic
complexity and the nature and amount of DNAPL make the task of complete source-zone
restoration impossible because the injected solution cannot be made to contact adequately the
immiscible-phase contaminant mass. For chemical flush or steam flush technologies to be
effective, the spatial distribution of the immiscible phase DNAPL and the spatial distribution of
hydraulic conductivity in the groundwater zone must be favourable for contact between the flush
fluid and the immiscible phase. The contaminant-mass removal rates or in situ destruction rates
must be sufficient to produce source-zone restoration in a practical time period, a period of
months or years rather than decades. The spatial distributions of the immiscible phase and
hydraulic conductivity derive their character primarily from geologic heterogeneity or
-------
complexity. Current capability to characterize DNAPL source zones adequately to allow
prediction of the performance of technologies for in situ source removal or destruction is
minima]. The problem of contact between the flush fluid and the immiscible phase mass is
exacerbated by the fact that DNAPL presence lowers considerably the hydraulic conductivity of
the porous medium with respect to the aqueous solution in the portion of the aquifer where free-
product DNAPL resides. A proposed advantage of the steam flush method is that, if pressures
are released after initial heating, the NAPL caught in lower permeability zones may be made to
volatilize and advect into the more permeable strata (K. Udell, pers. comm.)
All technologies for in situ source removal or destruction of DNAPL have the potential to
make the site conditions worse rather than better. A particularly adverse possibility is
remobilization of DNAPLs, causing some of the DNAPL mass to sink deeper into the aquifer
or to sink through an aquitard into an underlying aquifer not previously contaminated.
Unfortunately, the state of knowledge of technologies for source mass-removal is poor pertaining
to their potential to cause DNAPL remobilization. Therefore, decisions to use in situ mass
removal or destruction technologies in source zones should not be taken lightly. Poorly
controlled experimentation at real sites can be hazardous to groundwater resources.
In the context of environmental safety, the ideal form of in situ source-zone remediation is
one that destroys the immiscible phase to an adequate degree, produces insignificant hazardous
reaction products in situ, and causes no adverse mobilization of the immiscible phase. For
chlorinated solvent DNAPLs, in situ destruction by chemical oxidation is the technology that
currently comes closest to meeting these performance criteria. However, knowledge of the
-------
degree of the effectiveness of this technology is currently limited to a few chlorinated solvents
in relatively pure form.-No trials at actual, industrial sites have yet been undertaken.
In some types of porous fractured rock such as sandstone and fractured clayey or silty
overburden, all or much of the mass of immiscible-phase organic liquid in the source zone may
have transferred from the immiscible phase in the fractures to the aqueous and sorbed phase in
the porous matrix over the years or decades since the immiscible mass entered the subsurface.
Rather than a favourable change in state, this transfer has a major negative effect. The
contaminants enter the matrix by molecular diffusion. The mass flux and time-scales of the
diffusion process back out of the porous matrix are such that no technology for source mass
removal (other than excavation),-Whether it be conventional pump-and-treat or any of the
chemically-enhanced, heat-enhanced, or steam-enhanced expermc-'al technologies when fully
developed, can be expected to accomplish effective source mass removal except over extended
periods of time (Parker et al., 1993).
DISCUSSION
Mass Removal
At DNAPL and LNAPL sites, nearly all of the contaminant mass occurs in the source zone
as immiscible phase organic liquid (DNAPL and LNAPL). Therefore, when full aquifer
restoration is the goal, essentially all of the immiscible phase mass must be removed.
Conventional pump-and-treat accomplishes source mass removal too slowly and therefore it is
not a practical technology for source mass removal.
-------
Some degree of mass removal is a consequence of most remedial measures. Containment
of source zones or plumes does not require a specific degree of mass removal. Containment
requires only that necessary changes in the groundwater flow pattern be imposed. However,
removal of a source zone or a plume clearly requires a specific degree of mass removal. At
DNAPL sites, it will seldom be possible to predict or even determine when a sufficient
percentage of the total mass has been removed to achieve full source zone restoration. The
cumulative mass removed by a remedial technology such as groundwater pump-and-treat or soil
vapour extraction can generally be measured reliably. However, at most sites, the total initial
mass in the plumes or the source zones is not known. Mass estimates for source zones,
particularly the part of the source zone below the water table, commonly have uncertainties of
one or more orders of magnitude. Consequently, it is not possible to compare reliably the mass
removed to the mass remaining (Figure 12a) and therefore the percentage of mass removed
cannot be determined.
For the most common type of DNAPL sites (chlorinated solvent sites), restoration of source
zones requires reduction of bulk soil concentrations in the source zone from hundreds to
thousands of parts per million to a part per million or lower. Restoration of the source zone to
such low levels may often require ultimate mass removal effectiveness greater than 99.9%. This
is a truly exceptional challenge for all in situ mass removal technologies. The best that the
petroleum industry is able to do when applying advanced technologies for oil recovery is 40 or
50% of oil removal from the geologic stratum. The difficulty of this challenge generally has
gone unrecognized. It may be viewed that any degree of mass removal is a beneficial step
towards eventual achievement of source zone restoration. However, from a-practical perspective,
-------
the benefits cannot be quantified unless the mass removed can be quantitatively related to the
degree of restoration achieved, or unless the mass .removed allows more cost-effective
implementation or elimination of other remedial measures that might otherwise be required.
Without such information, which is typically the case, the cost-benefit comparison for the mass
removal action cannot be made.
There are currently great expectations in government and industry for technologies such as
surfactant and steam flushing for restoration of DNAPL source zones. These technologies appear
capable of removing DNAPL mass much more rapidly than conventional pump-and-treat, as
illustrated in Figure 12b. The DNAPL mass removal graphs (Figures 12a and b) illustrate that
source restoration is characterized by diminishing returns. As the restoration proceeds, the
DNAPL mass remaining in the subsurface decreases, unfortunately, so does the efficiency of the
extraction technology. Considering that the remaining DNAPL mass must be reduced to a small
percentage of the original mass, the ultimate practicability of all enhanced removal technologies
is in doubt and will remain in doubt for some time because the necessary research will be slow,
difficult, and very expensive.
To select and compare rational options for remedial measures we must determine the
appropriate conceptual model for the subsurface contamination given the type of chemicals, the
nature of chemical release to the subsurface, and the hydrogeological setting. In simplest terms
this would be the determination of whether the site has or likely has immiscible-phase organic
liquids (LNAPL or DNAPL) in subsurface source zones and the approximate spatial extent and
character of such zones. Also, we must acquire sufficient site-specific information on the
subsurface conditions and distribution of contamination to define or estimate the lateral and
-------
vertical extent of the migrating aqueous-phase plumes. The two tasks indicated above need to
be accomplished only to the level of detail appropriate for consideration of the performance of
those technologies that appear, based on general knowledge of site use and conditions, to have
potential applicability. It should be clearly recognized which technologies are considered proven
technologies and which are experimental technologies in the site-specific context. A proven
technology for one type of site may be unproven or experimental for other types of sites.
Experience with groundwater pump-and-treat for common industrial organic contaminants
has shown that restoration of plumes (but not source zones) in sandy aquifers to achieve drinking
water standards is possible in a practical timeframe at some sites, but only after the source zones
are adequately identified and contained. Full restoration of significant subsurface source zones
has not yet been demonstrated. No technology or groups of technologies for in situ removal or
in situ destruction of chemical mass has yet advanced to the stage where it is possible to predict
when, or even if, sufficient restoration of DNAPL source zones can be achieved.
Issues of Time and Benefits
Until source zone restoration is achieved, it will be necessary to maintain a source-zone
control systems such as pump-and-treat or cutoff wall enclosures to prevent plume growth. With
no special source-mass removal activity, the pump-and-treat system at a typical DNAPL site may
need to function for centuries or longer before the source mass will be dissipated sufficiently to
meet an aquifer-restoration condition. When source mass removal is pursued by active
engineering means such as chemical flush or steam flush or in situ destruction (ie., enhanced
source mass removal), the number of years that might be required to .achieve restoration is
-------
reduced. Thus, if enhanced source-mass removal achieves only partial source-mass removal,
some benefit is achieved in terms of the duration of pu;np-and-treat. However, this benefit is
hypothetical only because its magnitude cannot be predicted and it accrues at some undefinable
future time. This hypothetical relationship of source mass removal to time-of-achievement of full
aquifer restoration is displayed schematically in Figure 13. For example, with no enhanced
source mass removal, the length of time that pump-and-treat would be necessary to achieve full
aquifer restoration might be 1,000 years. If enhanced source mass removal removes half of the
source mass and is then discontinued, the length of time necessary for pump-and-treat would be
some fraction of 1000 years, perhaps 500 years. A direct proportion assumption on time is
optimistic; a more plausible expectation is a much more adverse time relationship. The key issue
pertaining to partial source mass removal centres on the value of the ultimate benefits achieved
in the distant future by such removal relative to the cost of the removal. If the benefit does not
accrue for decades, then it has very little present worth and the source removal may not be
justified on solely economic grounds.
A major tenet of modem environmental ethics is the desirability of choosing courses of
action that will lead most rapidly and/or assuredly to a sustainable environmental and economic
condition. In a groundwater context, a sustainable environmental condition may be considered
to be a condition where groundwater resources are not significantly diminished in quality due
to continuing growth of contaminant plumes. Across North America, the ever-increasing number
of pump-and-treat systems for source-zone control represents large-scale action for risk control
and groundwater protection that cannot be considered sustainable in the long-term because pump-
and-treat is an excessively active (and expensive) technology. Pump-and-treat requires continuing
-------
commitment by future generations of their societal resources to maintain the pump-and-treat
facilities and for performance monitoring and energy use. Given the ephemeral nature of
governmental policies and the transient existence of some of the corporations that carry some
of the cost of the "in perpetuity" pump-and-treat burden, it is likely unreasonable to assume that
the pump-and-treat facilities put into operation in this century will continue to be maintained
properly or even maintained at all through the next century. Similarly, it may be unreasonable
to expect that sufficient societal resources will be available to evaluate ever-increasing reams of
monitoring data to determine if corrections and adjustments in the pump-and-treat systems are
required.
Full aquifer restoration is generally not achievable at DNAPL sites and, therefore,
unavoidably, our society must transfer responsibilities and risks associated with these sites to the
future. Given this transferred burden, we should ask what more positive legacy we may be able
to offer to future generations. Our best answer is that technologies for in situ aquifer remediation
will be improved considerably in the next few decades, but probably not in the next few years.
Therefore, onexif the goals of current aquifer remediation should be to pass to future generations
site conditions that are better suited for application of new technologies than the present
condition. Conventional pump-and-treat systems will likely contribute little towards this goal,
particularly if their performance and effects are monitored as casually as is typically the case at
present.
Conceptually, an array of remediation options are available for application to DNAPL sites
(Figure 13), ranging from the most 'active' option, which is aggressive, large-scale pump-and-
treat to the most passive, which has cutoff wall enclosures and in situ treatment systems without
-------
pump-and-treat. In situ treatment curtains and funnel-and-gates are new technologies still in the
prototype field trial stage. However, we expect that these technologies will advance rapidly in
the next few years. The in situ contaminant mass flux that these technologies must accommodate
in most plumes is not large and therefore the technical challenge is not particularly difficult.
However, even when these new technologies are fully developed, hydrogeologic conditions at
many sites, particularly fractured rock sties, are such that they will not be suitable, leaving only
pump-and-treat as the technically and economically viable option for groundwater remediation.
The Cap Dilemma
A common dilemma in decisionmaking for remediation of waste disposal sites is whether
or not to put a cap on the ground surface over the source zone. The construction of highly
engineered low-permeability surface covers (ie., caps) is commonly undertaken at waste disposal
sites. In many cases, much of the expense of the cap derives from a specification that the cap
prevent infiltration of rainwater and snowmelt through the waste. Prevention of infiltration
through the waste is intended to prevent groundwater contamination. This intention makes sense
if the site condition fits the conceptualization shown in Figure la where the entire source mass
is situated above the water table. However, if the site is a DNAPL site (Figure lc), allocation
of large financial resources to construct a cap to achieve low infiltration to the water table is
typically unwarranted on scientific grounds. At sites where appreciable DNAPL exists below the
water table, it is generally not possible to achieve any significant reduction in plume growth or
plume control by means of a cap. At DNAPL sites, the contribution of water to the plume from
-------
infiltration through the waste to the water table is typically very small in comparison to lateral
groundwater flow through the DNAPL zone (Figure 15).
At sites where chlorinated solvents exist above the water table, even if only in small
amounts, it is reasonable to expect that plumes of solvent vapour will occupy the vadose zone
beneath the site and will move by diffusion away from the site. Solvent vapour in the vadose
zone causes groundwater contamination by transfer of some vapour mass to the groundwater
zone. Mendoza and McAlary (1989) and Hughes and Gillham (1992) describe solvent vapour
migration in the vadose zone and its impact on groundwater. Placement of a low permeability
cap over the land area where immiscible-phase solvents occur in the vadose zone does not
necessarily diminish impacts of the solvents on groundwater. On the contrary, in many cases it
is likely that the cap increases the tendency of vapour to contaminate groundwater by causing
greater lateral spread of the vapour beneath and beyond the cap (Figure 16). This spread can be
prevented by vacuum extraction of vapour, however, this imposes a need for long-term active
engineering because full cleanup of vadose zone DNAPL by vacuum extraction typically takes
decades. Another alternative is to connect the cap to vertical low-permeability walls (ie.,
enclosures) that extend below the water table, with active or passive vapour venting.
In circumstances where source zone containment is implemented, the principal objective of
a cap may be to prevent direct exposure to the contaminated soil and to resist erosion and
weathering. These objectives may result in a cap design that is considerably different, and
perhaps less costly, than one intended to reduce infiltration.
The main issue addressed here is that a remedial action such as a cap, which is a beneficial
action at sites that have no solvent DNAPL, can be an beneficial option at solvent DNAPL sites
-------
unless the cap design is adapted to the special demands of the DNAPL problem. We expect that
a detailed post mortenvof the many caps that have been constructed at Superfund and RCRA
sites would show, because of DNAPL issues, that little or no benefit was achieved by the
groundwater environment, contrary to the expectation at the time that the cap was built. In fact,
many caps have probably been detrimental to groundwater remediation.
In Situ Bioremediatiod
Of the various types of experimental technologies for in situ remediation at industrial sites,
the one that has had the most research is in situ bioremediation. Unfortunately, the prognosis
for most types of in situ bioremediation at DNAPL sites is not good.
Nearly all DNAPLs in the immiscible phase are toxic to microbes. Also, the groundwater
in and near DNAPL source zones commonly has sufficiently high aqueous concentrations
leached from the DNAPL to be toxic to microbes. Therefore, in situ microbial technologies are
particularly ill-suited for removal of DNAPL mass from source-rone. There is some hope for
in situ microbial technologies to assist in the restoration of source zones at some stage after
chemical or steam flush technologies or after in situ chemical destruction technologies have
removed nearly all of the immiscible-phase mass from the source zone. This assumes, however,
that the primary technologies do not create a biological or geochemical environment that is
inhospitable to in situ bioremediation.
The potential for in situ microbial methods to be useful is greatest in the plume at some
distance from the DNAPL source zone where the aqueous-phase contaminant concentrations are
lower than levels toxic to microbes. However, the most common approaches suggested for in
-------
situ bioremediation of plumes involve injection of nutrients and electron acceptors (ie., oxygen)
into the plume to stimulate activity of natural microbes for destruction of contaminants that
desorb from the porous medium. This approach suffers from various limitations related to
delivery of the injectants to all zones where contamination is present. These limitations typically
derive from geologic complexity or heterogeneity. An alternative approach is to stimulate
microbial activity for in situ contaminant destruction in a permeable microbial curtain or in
funnel-and-gates placed across the plume. In this in situ use of bacteria, the contaminants are
bought to the microbial zone by natural groundwater flow and therefore much of the injectant
delivery problem is avoided. However, research on this type of in situ bioremediation is in its
infancy. The main challenge is to develop relatively passive life support systems for the in situ
microbial processes.
CONCLUDING STATEMENT
To date, when remedial measures are implemented to achieve source containment and/or
plume containment at DNAPL sites, the measures nearly always include pump-and-treat. In most
cases, these pump-and-treat systems will require operation "in perpetuity" because of our
inability to achieve adequate removal or restoration of DNAPL source zones. This approach does
not meet reasonable criteria for long-term environmental sustainability because pump-and-treat
is an excessively active technology. A more passive approach for source zone containment would
be the use of low-permeability cutoff-wall enclosures, coupled with minimal pump-and-treat or,
in some cases, no pump-and-treat. This approach involves higher up front capital expenditures
for construction but lower average annual long-term operations and maintenance costs. Since low
-------
permeability cutoff wall enclosures as well as pump-and-treat systems are proven technologies
for source zone containment, the choice is between one or the other of these two proven
technologies. Issues of long-term cost and environmental sustainability should govern the
technology choice. Cutoff wall enclosures may facilitate future use of promising in situ
technologies for source mass removal that are more difficult or risky to use without such
enclosures. There are many sites where significant mass of DNAPL exists too deep for use of
enclosure or in fractured bedrock. Cutoff wall technologies are much less well developed for
fractured rock than overburden, particularly hard rock. In hard rock such as granite, cutoff walls
in the conventional sense cannot be constructed. Other types of walls made by injecting sealants
into fractures via boreholes can be&onstructed but reliable prediction of the performance prior
to construction is generally not possible. In this situation, pump-and-treat is usually the only
proven technology available for source zone or plume containment.
Apparent failures in the arena of subsurface restoration abound because of the lack of clear
recognition of the difference between: (i) plume containment, (ii) plume removal, and (iii) full
aquifer restoration. Many attempts at aquifer restoration have been made or are in progress.
Remediation aimed directly at the goal of full aquifer restoration in this decade or even in the
next decade or two should be recognized clearly as experimentation rather than actual
remediation. In the United States, the Superfund process is intended to be primarily an aquifer
restoration program, but it cannot perform as such because of the nature of the subsurface
contamination at DNAPL sites and the inherent limitations of existing remedial technologies
relative to what must be accomplished to achieve full aquifer restoration. In a groundwater
context, most remedial efforts of the Superfund program be regarded as numerous and costly
-------
experiments; unfortunately, in scientific terms, they are relatively unmonitored experiments with
poorly defined hypotheses. In retrospect, the progress of these remedial efforts is best described
as "trial and error". Opportunities to acquire information that would ultimately provide better
remedial choices and designs and commensurate cost savings are being lost on a grand scale.
To derive benefits from careful site monitoring and experimentation, the experimentation
must be designed specifically to track and assess in detail the progress of the restoration or
plume containment attempts and to develop general scientific and engineering knowledge on the
remediation technologies. This is rarely done. Rather, in the United States and Canada
monitoring of remedial activities is generally implemented simply to satisfy basic regulatory
requirements. Even the data acquired from such basic monitoring are rarely studied to derive
guidance that will improve future efforts.
The acquisition of carefully-obtained scientific and engineering knowledge from large-scale
site experimentation will allow better design and implementation of remedial actions at other
sites. However, the benefits may accrue mainly to future activities at other sites and to future
generations of cleanup. Eventually, this class of large-scale experimentation is likely to pave the
way to an era of practical groundwater restoration where the stated goals are in fact achievable
in a practical time scale.
Large-scale experimentation with technologies for subsurface source removal or aquifer
restoration should normally be preceded by appropriate bench-scale studies, prototype technology
trials, and then pilot-scale technology assessments (Figure 17) in a variety of hydrogeological
settings. In the rush towards aquifer restoration that is driven by public opinion and legislation,
careful completion and critical review of each of these stages in technology development is
-------
rarely accomplished and hence large-scale site experimentation is hampered by a weak base of
fundamental knowledge even when large-scale field experimentation is accompanied by detailed
monitoring. Rigorous methods or protocols for technology trials and pilot-scale trials are not
well developed because of underemphasis of these essential steps in technology development.
It is truly unfortunate that, for most sites, today's technologies are inadequate for the task
of source mass removal, and hence inadequate for full aquifer restoration. The current use of
pump-and-treat or alternative technologies having the same intention, which can perform well
for plume or source zone containment but not for source removal, must therefore continue until
such time as more cost-effective mass removal and containment technologies are developed, or
until source zone restoration through mass removal can be accomplished. The development of
cost-effective source removal technologies will be slow, arduous, and expensive. Thus, we are
transferring into the next century the task of operating and maintaining numerous active
containment systems in order to prevent further growth or spread of contaminants from
subsurface source zones. Although the transfer of this responsibility is presently unavoidable;
today we have an opportunity to enhance possibilities for future remedial efforts by a systematic
advance of scientific and engineering knowledge of the processes and performance of
remediation technologies.
-------
ACKNOWLEDGEMENTS
This paper is a contribution from the University Solvents-In-Groundwater Research
Program, Phase n, supported by funds from the Natural Sciences and Engineering Research
Council (NSERC), the Ontario University Research Incentive Fund (URIF), and the following
corporations: Boeing, Ciba-Geigy, General Electric, Eastman Kodak, and Laidlaw
Environmental Services.
-------
REFERENCES
Burris, D.R. and Cherry, J.A., 1992. Emerging plume management technologies: In situ
treatment zones. Paper presented at Subsurface Restoration Conference, 3rd International
Conference on Ground Water Quality Research, Dallas, Texas, June 21-24.
Cherry, J.A., 1984. Contaminant migration in groundwater: Processes and problems. In:
Proceedings of the Second National Water Conference - The Fate of Toxics in Surface and
Groundwater, January 24-25.
Feenstra, S., 1984. Groundwater contamination from dense non-aqueous phase liquids (DNAPL)
chemicals. Paper presented at the Geological Association of Canada Annual Meeting, May 14-
16, London, Ontario (published abstract and distributed paper).
Feenstra, S. and Cherry, J.A., 1988. Subsurface contamination by Dense Non-Aqueous Phase
Liquid (DNAPL) chemicals. Presented at: International Groundwater Symposium, International
Association of Hydrogeologists, May 1-4, Halifax, Nova Scotia.
Gillham, R.W., and Burris, D.R^, 1992. In situ treatment walls - Chemical dehalogenation,
denitrification, and bioaugmentation. Paper presented at Subsurface Restoration Conference, 3rd
International Conference on Ground Water Quality Research, Dallas, Texas, June 21-24.
Hughes, B.M., Gillham, R.W., and Mendoza, C.A., 1993. Transport of trichloroethylene
vapours in the unsaturated zone: A field experiment. IAH Conference on Subsurface
Contamination by Immiscible Fluids, Calgary, Alberta, April 18-20, 1990, Weyer, K.U. (ed.),
pp. 81-88.
Mackay, D.M., Roberts, P.V., and Cherry, J.A., 1985. Transport of organic contaminants in
groundwater. Env7 Sci. Technol., 19(5):384-392.
Mackay, D.M. and Cherry, J.A., 1989. Limitations of pump and treat remediation. Env. Sci.
Technol., 23(6):630-637.
Mendoza, C.A. and McAlary, T.A., 1989. Modelling of groundwater contamination caused by
organic solvent vapours. Ground Water, 28(2): 199-206.
Pankow,, J.F., Johnson, R.L., and Cherry, J.A., 1993. Air sparging in gate walls in cutoff
walls and trenches for control of plumes of volatile organic compounds (VOCs). Ground Water,
31(4).
Parker, B.L., Gillham, R.W., and Cherry, J.A., 1993. Diffusive disappearance of immiscible
phase organic liquids in fractured geologic media. Ground Water (submitted July).
-------
Schnarr, M.J. and Farquhar, G.J., 1992. An in situ oxidation technique to destroy residual
DNAPL from soil. Paper presented at Subsurface Restoration Conference, 3rd International
Conference on Ground Water Quality Research, Dallas, Texas, June 21-24.
Schwille, F., 1984. Volatile chlorinated hydrocarbons in porous and fractured media (in
German). Spec. Contrib. Germany Water Studies Annual Report No. 46, Federal Institute for
Water Studies, Koblenz, Germany.
Schwille, F., 1981. Groundwater pollution in porous media by fluids immiscible in water. The
Science of the Total Environment, 21:173-185.
Starr, R.C. and Cherry, J.A., 1993. In situ remediation of contaminated groundwater: The
funnel-and-gate system. Ground Water (submitted May).
U.S. Environmental Protection Agency, 1988. Guidance on remedial actions for remediation
actions for contaminated ground water at Superfund sites. EPA/540/-88-003.
U.S. Environmental Protection Agency, 1992. Dense nonaqueous phase liquids: A workshop
summary. Dallas, Texas, April 16-18.
Villaume, J.F., 1983. Recovery of coal gasification wastes: An innovative approach.
Proceedings Third National Symposium on Aquifer Restoration and Groundwater Monitoring,
National Water Well Association, Worthington, Ohio, 434-445.
Wunderlich, R.W., Fountain, J.C., and Jackson, R.E., 1992. In situ remediation of aquifers
contaminated with dense non-aqueous phase liquids by chemically enhanced solubilization. J.
Soil Contam., l(4):361-378.
-------
LIST OF FIGURES
Figure 1: Three different conceptual models for groundwater contamination: (a) non-
NAPL case; (b) LNAPL case; and (c) DNAPL case.
Figure 2: Anatomy of a DNAPL site: (a) plan view of source zone and plume; and (b)
vertical section of source zone. Note contaminant sources (sludge, drums,
contaminated soil, residual) above the water table and immiscible-phase
DNAPL (residual, lenses, pool) source below the water table.
Figure 3: Containment of plumes at DNAPL sites: (a) pump-and-treat for containment
of plume; and (b) natural hydrologic boundary (river bed).
Figure 4: Conceptual representation of aquifer restoration: source and plume removal
sufficient to meet cleanup standard: (a) before aquifer restoration; and (b)
conditions when fully restored; flux from source insufficient to produce
plume.
Figure 5: Conceptual representation of partial aquifer restoration using pump-and-
treat: (a) initial condition; (b) performance of plume purge wells and
contaminant well; and (c) partially restored aquifer with source zone
containment to prevent regrowth of plume.
Figure 6: Three approaches for achieving partial aquifer restoration: (a) aggressive
pump-and-treat; (b) flushing to containment well; and (c) natural flushing of
plume to hydrologic discharge zone (contaminant flux to river).
figure 7: An in situ treatment curtain for plume control: (a) plan view; (b) vertical
cross-section, both showing plume entering the treatment curtain and treated
plume water exiting downgradient side.
Figure 8: Options for source zone containment: (a) pump-and-treat; (b) cutoff wall
with interior pumping; (c) in situ treatment curtain; and (d) funnel-and-gate.
Solid dots are extraction wells, thick lines are impermeable walls, and areas
with small circles are permeable media.
Figure 9: Cutoff wall enclosures around DNAPL source zones: (a) natural gradient
case (b) inward flow everywhere induced by pump-and-treat (c) downward
flow from bottom cannot be reversed by pump-and-treat.
Figure 10: Influence of delay processes on time to completion of partial aquifer
restoration by pump-and-treat.
-------
Figure 11: Restoration of DNAPL source zone using chemically-enhanced flushing with
injection and withdrawal wells: (a) plan, injection and withdrawal wells for
chemical flush across source zone and purge wells to remove plume; and (b)
section; and (c) comparison of mass removal with and without chemical
enhancement.
Figure 12: Schematic representation of cumulative mass removal for restoration of
DNAPL source zones: (a) total mass in source zone is uncertain; (b)
technology performance graph contaminant mass in source zone assumed
known; and (c) discrepancy between desired performance and actual
performance caused by heterogeneity, etc.
Figure 13: Relationship of percent contaminant mass removed from source zone during
period of source zone restoration to time necessary to achieve full restoration of
source zone by conventional pump-and-treat.
Figure 14: Variety of options for full aquifer restoration: ranging from fully active
technology at top fully passive at bottom.
Figure 15: Contaminant contribution to groundwater plume from: A-A', zone of leaching
of contaminants from vadose zone to groundwater zone; and B-B', zone of
volatile from dissolution of DNAPL residual and pool causing main contaminant
flux to plume. A' flux is typically small relative to B' flux.
Figure 16: Comparison of the effect of caps on two source zones: (a) source materials exist
only above the water table and emit no contaminant vapour; (b) volatile source
materials exist above the water table causing contaminant vapour to contaminate
groundwater.
Figure 17: Stages in the development of groundwater remediation technology from concept
-------
NON NAPL
TANK LEAKAGE
LNAPL
	rut
FLOATING LNAPL
L5.-. PLUME;'—^
•	«	•	a	,	* 	•
RESIDUAL LNAPL
Figure 1: Three different conceptual models for groundwater cc.ntamination: (a) non-
-------
ANATOMY OF A DNAPL SITE
(a)
DNAPL ENTRY
LOCATIONS
DISSOLVED PHASE PLUME
GROUND
WATER
FLOW
S\
PROPERTY
BOUNDARY
DNAPL ZONE
RESIDUAL^UENSES
AND POOLS BELOW
WATER TABLE
(i.8. AREA OF
SUBSURFACE
DNAPL SOURCES)
PLUME OUTLINE
BASED ON SOME LIMIT
SUCH AS 0RINKIN6 WATER
MPC OR DETECTION LIMIT
PLUME
ADVANCE
TOWARDS
RECEPTORS
SUCH AS WELLS,
WETLANDS
OR STREAMS
(b)
VADOSE
ZONE
SOURCES
NFILTRATION
	2_
- SOURCE ZONE
CONTAMINATED
SLUDGE
DRUMS
7
FILL
CONTAMINATED
SOIL
GROUND-
WATER
ZONE i
SOURCES
->—'—7—T: "j ^—T 1 ' T—r—r
IMPERVIOUS LAYER
PLUME
TRAVELS
TOWARDS
RECEPTORS
Figure 2: Anatomy of a DNAPL site: (a) plan view of source zone and plume; and (b)
vertical section of source zone. Note contaminant sources (sludge, drums,
contaminated soil, residual) ab*>ve the water table and immiscible-phase
-------
PLUME CONTAINMENT
(a )
PUMP AND TREAT
FLOW
PLUME
PUMPING WELL
FOR PLUME
CONTAINMENT
/ '.'-JS
(b) NATURAL BOUNDARY
FLOW
Figure 3: Containment of plumes at DNAPL sites: (a) pump-and-treat for containment
-------
AQUIFER RESTORATION
(a) BEFORE AQUIFER RESTORATION
FLOW
-i—i r*
SOURCE-
ZONE
-L-LX
PLUME-
(b) FULLY RESTORED AQUIFER
F LOW
/"FORMERS
SOURCE
v ZONE /
"S	^"

FORMER PLUME
Figure 4: Conceptual representation of aquifer restoration: source and plume removal
sufficient to meet cleanup standard: (a) before aquifer restoration; and (b)
conditions when fully restored; flux from source insufficient to produce
-------
PARTIAL AQUIFER RESTORATION
(a)
(b)
SOURCE ZONE .
CONTAINMENT 1
WELL
PLUME PURGE
WELLS
"6*. *• o • '"o*' 'n.'-'
"N
SOURCE ZONE CONTAINMENT WELL
PLUME PURGE WELLS
—	MCL
(C)
PORTION OF RESTORED
AQUIFER
\
I
/
PERSISTANT
CONTAMINATION
Figure 5: Conceptual repres?"'.ation of partial aquifer restoration using pump-and-
treat: (a) initial - eition; (o) performance of plume purge wells and
contaminant wel ."-d (c) partially restored aquifer with source zone
-------
PARTIAL AQUIFER RE
(a )
THREE APPROACHES
(1) AGGRESSIVE PU MP - A ND - TREAT
SOURCE ZONE
CONTAIMENT
(2) FLUSHING TO CONTAIMENT WELL
CONTAINMENT
WELL
2 3 4 5 * »l
.v V \ > °n|
	*	*	is. _

(3) NATURAL FLUSHING OF PLUME
(b)


PROGRESS OF PARTIAL AQUIFER RESTORATION
LU
LU
<2
z o

-1 ^ > n
a. o 5 O
TIME
Figure 6: Three approaches for achieving partial aquifer restoration: (a) aggressive
pump-and-treat; (b) flushing to containment well; and (c) natural flushing of
-------
IN- SITU TREATMENT CURTAIN
(a)
FLOW
PLUME
IN - SITU TREATMENT
CURTAIN

a;
' L_
jgj.	^
A'
j
i#—
(b)
FLOW
FILL
BY CHEMICAL OR
MICROBIAL PROCESSES
V
X
A X


* V

•' • PLUME
• " >
• • •
• »
Jt' •
ZONE WITH NO
—~ CRITICAL CONTAMINANTS
REMOVAL OF //
0 • •
0 O9
^ PERMEABLE
MEDIUM
Figure 7:
An in situ treatment curtain for plume control: (a) plan view; $) vertical
cross-section, both showing plume entering the treatment curtain and treated
-------
SOURCE ZONE CONTAINMENT
PUMP a TREAT
CUTOFF WALL
ENCLOSURE
TREATMENT
CURTAIN

FUNNEL 8 GATE
Figure 8: Options for source zone containment: (a) pump-and-treat; (b) cutoff wall
with interior pumping; (c) in situ treatment curtain; and (d) funnel-and-gate.
Solid dots are extraction wells, thick lines are impermeable walls, and areas
-------
CUTOFF WALL SCENARIOS
CUTOFF WALL
FLOW UP
OR DOWN
AQUIFER
HEAD
PUMP 6 TREAT
/1 t t \
~
PUMP 8 TREAT
- ::Vi
¦ : V	' UNSATURATED ZONE
Figure 9:
Cutoff wall enclosures around DNAPL source zones: (a) natural gradient
case (b) inward flow everywhere induced by pump-and-treat (c) downward
-------
RESPONSE OF
PURGE WELLS
A	RAPID PLUME REMOVAL WHERE DELAY
PROCESSES ARE INSIGNIFICANT
B DELAY CAUSED BY PROCESSES SUCH AS
SLOW DESORPTION, DIFFUSION FROM LOW
PERMEABILITY ZONES IN THE AQUIFER
AND FROM AQUITARDS
MCL STANDARD TO WHICH CONCENTRATION
MUST DECLINE FOR SHUT DOWN
OF PUMP AND TREAT
Figure 10: Influence of deiay processes on lime to completion of partial aquifer
-------
SOURCE RESTORATION BY CHEMICAL FLUSH
(a) CHEMICAL FLUSH: INJECTION AND WITHDRAWAL WELLS
-v.rvjM. USING
PVJRGE
yjEU-S
SOURCE ZONE
^RESTORATION
CHEMICAL FLUSH			
, CONVENTIONAL
^ ^ ^ PUMP 8 TREAT
Tl ME
Figure 11: Restoration of DNAPL source zone using chemically-enhanced flushing with
injection and withdrawal wells: (a) plan, injection and withdrawal wells for
chemical fl,-sh across source zone and purge wells to remove plume; and (b)
section; and (c) comparison of mass removal with and without chemical
-------
~ ^UNCERTAINTY OF
^ J CONTAMINANT MASS IN SITU
/
'NEW
TECHNOLOGIES FOR
CONTAMINANT MASS REMOVAL
///^^		CONVENTIONAL
—	 p a t _
TIME
MASS REMOVAL FOR SOURCE
ZONE TO MEET CLEANUP STANDARD
"V PERFORMANCE OF MASS
REMOVAL TECHNOLOGY
TIME
T-
TECHNOLOGY LIMITATION :
EFFECT OF HETEROGENEITY
Schematic representation of cumulative mass removal for restoration of
DNAPL source zones: (a) total mass in source zone is uncertain; (b)
technology performance grapi" contaminant mass in source zone assumed
-------
NATURAL
FLOW —*
<
1
<
£
-J
Li.
SOURCE
ZONE '
&
s. J>>	p.
CONTAMINANT
MASS FLUX
FROM SOURCE
L /
A'
PERCENT SOURCE MASS
^INITIALLY REMOVED
0%
_L
DRINKING WATER —
, STANDARD
10 100 1000
YEARS
Figure 13:
Relationship of percent contaminant mass removed from source zone during
period of source zone restoration to time necessary to achieve full restoration of
-------
APPROACHES FOR SOURCE
CONTROL AND PLUME REMOVAL
UJ
> >
D O
Ll. <
FULL SCALE P and T
LlI
>
°3
CUTOFF WALL ENCLOSURE WITH
INTERIOR/EXTERIOR P and T
TREATMENT ZONE WITH PLUME/P and T
CUTOFF WALL ENCLOSURE WITH TWO
TREATMENT ZONES

Figure 14: Variety of options for full aquifer restoration: ranging from fully active
-------
AREA OF INPUT FROM
- RAIN/SNOW 	
VADOSE
ZONE
GROUNDWATER
ZONE
5
o
i
z
^	A CONTAMINATED SOIL
iVii.fYi
« . i J	'	.* '' ' Ttifc'
B
B'
^ RESIDUAL

POOL. .
Figure 15: Contaminant contribution to groundwater plume from: A-A', zone of leaching
of contaminants from vadose zone to groundwater zone; and B-B', zone of
volatile from dissolution of DNAPL residual and pool causing main contaminant
-------
CASE A WASTE WITH NON - VOLATILE CONTAMINANTS
CAP
CASE B WASTE WITH VOLATILE CONTAMINANTS
Figure 16: Comparison of the effect of caps on two source zones: (a) source materials exist
only above the water table and emit no contaminant vapour; (b) volatile source
materials exist above the water table causing contaminant vapour to contaminate
-------
STAGES IN TECHNOLOGY DEVELOPMENT
PROVEN
TECHNOLOG
EMERGING
TECHNOLOGY.
EXPERIMENTAL
TECHNOLOGY
COSTisjEFFECTIVE TECHNOLOGY
LARGE TRIALS
PROTOTYPE ASSESSMENT
FIELD STUDIES WITH RIGOROUS
MASS BALANCE CONSTRAINTS
SCREENING OF NUMEROUS
TECHNOLOGY CONCEPTS
CONCEPTS
Figure 17: Stages in the development of groundwater remediation technology from concept
-------
Reprinted from ENVIRONMENTAL SCIENCE & TECHNOLOGY, Vol. 23, Page 630, June 1989
-------
Groundwater
Pump-add-tr
Second of a me-part series
Douglas M. Mackay
University of California
Los Angeles, CA 90024
John A. Cherry
University of Waterloo
Waterloo, ON, Canada
Almost all remediation of groundwater
at contaminated sites is based on
groundwater extraction by wells or
drains, usually accompanied by treat-
ment of the extracted water prior to dis-
posal. This often causes an initial de-
crease in contaminant concentrations in
the extracted water, followed by a lev-
eling of concentration, and sometimes a
gradual decline that is generally ex-
pected to continue over decades. In
such cases, the goal of reaching strin-
gent health-based cleanup standards is
very remote and the ultimate cost of
cleanup very high (I).
The purpose of this paper is to ex-
plore reasons for the observed difficulty
of groundwater cleanup and note some
implications that become clear during
this process. Our discussion is limited
to organic contaminants because they
are the most common health-threaten-
ing chemicals detected in groundwater
and because the greatest difficulties in
groundwater remediation have been en-~
countered at organic contamination
sues.
Organic contaminant plumes
Prior to the passage of the Compre-
hensive Environmental Response,
Compensation, and Liability Act
(CERCLA) and revised Resource Con-
servation and Recovery Act (RCRA)
legislation in 1980 and 1984, respec-
tively, detailed monitoring of ground-
water at industrial and waste disposal
sites was rare, particularly for trace or-
ganic contaminants. Now, as the end of
the decade approaches, knowledge of
the nature of organic contamination of
groundwater has advanced considera-
bly because of the expenditure of more?
than a billion dollars on site investiga-
tions and cleanup activities. Hundreds
of plumes of organic contaminants have
now been delineated by networks of!
monitoring wells.
Several examples of organic plumes
in the United States and Canada are
given in Table I. Table I pertains to
plumes in sand and gravel aquifers; al-
though there are many plumes in frac-
tured rock, summary data such as we
630 Environ Sci. Technol . Vol 23, No. S, 1989
-------
present in this table are not readily
available because the outer boundaries
of such plumes and the fracture poros-
ity are very difficult to determine.
Plumes such as those in Table 1, which
extend 0.5-10 km from the source and
which generally have formed over dec-
ades, are common in North America
and Europe. Each of the sites listed in
the table represents a major plume in
the sense that millions of dollars have
been spent on plume characterization,
feasibility studies for alternative reme-
dial actions, and attempts at remedia-
tion. Cleanup programs are now
planned or underway for some of the
plumes in the table at ultimate costs for
each plume estimated at tens of millions
of dollars or more. The CERCLA and
RCRA programs address hundreds of
plumes such as those in the table. There
are probably thousands of other such
plumes in North America that are not
addressed by these programs.
Of the various organic contaminants
found in groundwater, the widely used
industrial solvents and aromatic hydro-
carbons from petroleum products are
most common (2-4). Much if not most
of groundwater contamination of this
type is caused by leakage, spillage, or
disposal of organic liquids immiscible
with' water (nonaqueous-phase liquids
[NAPLs]) into the ground. Dissolution
of the NAPL and subsequent transport
of the dissolved constituents by ground-
water is thought to generate many
plumes (5), although there is mounting
evidence that migration within the va-
por phase of the unsaturated zone with
subsequent transfer of vapor-phase con-
tamination to soil water and ground-
water may also cause formation of
groundwater plumes (6, 7).
The plumes in Table 1 are listed in
decreasing order of the mass estimated
to be present in the dissolved form (ex-
pressed as equivalent volume of NAPL
in the right column). Note that there are
documented examples of plumes that
encompass much larger volumes of
groundwater and greater masses of dis-
solved contaminants than listed in this
short table. Also, not all types of con-
TABLE 1
Relatively well-documented organic contaminant plumes in sand-gravel aquifers'
S'tB location
«... a plume map
0
Presumed
sources
5 km
Flow
Predominant
contaminants"
Plums volume
(liters)0
Contaminant mas9
dissolved In plume (as
equivalent NAPL volume In liters
or 55-gal drums)"
chemical
plant
electronics
plants
sewage
infiltration
beds
TCE
TCA
PER
TCE
TCA
TCE
PER
Detergents
5,700.000,000
6,000,0003000
40,000,000,000
15,000 (72 drums)
9800 (47 drums)
1500 (7 drums)"
Traverse C Ity, Ml
aviation
fuel
storage
Toluene
Xylene
Benzene
400,000,000
1000 (5 drums)
Gloucester. ON
Canada
special
waste
landfill
1,4 Dioxane
Freon 113
DEE, THF
102,000,000
190 (0.9 drum)
San Jose CA
electronics
plant
trainyard,
airport
TCA
Freon 113
1, 1 DCE
TCE
TCA
DBCP
5,000,000.000
4,500,000,000
130(0.6 drum)
80 (0.4 drum)
•	Readers aware of othe' well-documented cases (or which reliable estimates of contaminant mass distribution aad organic carbon content (foe) of
the aquifer solids are available are encouraged to contact the authors, who plan to expand this compendium.
0 TCE «trichioroethyien ; TCA -1. 1 trichloroethane; PER - per-, i.e., tetrachloroethyl?ne; 1,1 DCE^ 1, 1 dichloroethylene; CHCL3-chloroform;
DEE =» diethyl ether, TH =• tetrahydr "uran; DBCP = dibromochloropropane.
" Approximate estimate., denved frc plume length, groundwater velocity, contaminant i-oncuntration distributions, etc., provided for illustrative
purposes only Estimated contaminant mass accounts only for the dissolved phase (I.e., c'oes not account for contaminant sorbed to the aquifer
media throughout the plume or for NAPL contaminant, if any, from the sources). Most ot oasic data is trom unpublished sources; data on three
plumes arB published (13, 27, 28, 2S).
*	This mass estimate is for the hafcx mated contaminants only (i.e., detergents are exclrded).
-------
FIGURE 1
Schematic of granular subsurface environment*
•Illustrates phases in which organic contaminants may be present or migrate. Note deflection of NAPLs by large
day strata and fine lenses of less permeable material such as day or silt withm the predominantly sand-gravel
aqurfet Note that sorted phase may be assoaated with the exterior of the particles or with in tenor sites.
*lKcNoroethylene.
taminants encountered in groundwater
plumes are represented, though most of
those listed are quite commonly de-
tected. Nevertheless, the table is in-
structive. It is clear, for example, that
the contaminant load in an aquifer can-
not be judged from the magnitude of
the contaminated area; two of the ap-
parently larger plumes have the least
contaminant mass dissolved in them.
This, of course, reflects the differences
in the concentrations within the various
plumes as well as their total volume.
All but two of the plumes in the table
contain more than a billion gallons of
contaminated water. Yet the two small-
est plumes contain relatively large
amounts of contaminants, in part the
result of their relatively high solubility
compared with the halogenated com-
pounds in the other plumes.
The table also illustrates that plumes
often contain less mass in the dissolved
form than would be present in a few
drums of NAPL. This may be a minis-
cule fraction of the total NAPL mass
that entered the subsurface at many
sites where the total amount of organic
chemicals used or disposed of would
have been extremely large, often mea-
sured in hundreds of tanker truck loads.
These examples ignore the contami-
nant mass that would be sorbed to the
aquifer media contacted by the plume,
which may be on the same order as or
significantly greater than the dissolved
mass (8). Nevertheless, it is clear that
the primary challenge in groundwater
cleanup is to remove the organics
masses that serve, in effect, as subsur-
face sources and cause the plumes to
grow and persist, rather than simply to
remove the mass of dissolved contami-
nants that defines the plume. This is-
sue, so fundamental to the proper diag-
nosis and efficient solution of the
problem, is explored further below.
Cleanup of sand and gravel aquifers
Ideally, a remediation project for
contaminated sand and gravel aquifers
would be designed based on a solid un-
derstanding of the mass and types of
pollutants released, the current location
of all the mass remaining in the subsur-
face, and the processes controlling the
removal of the mass from the subsur-
face (or its destruction in situ). Unfor-
tunately, none of these requirements is
generally met in practical investiga-
tions, leading to a considerable amount
of guesswork in developing a cleanup
plan.
Figure 1 illustrates a few of the many
known complexities of granular subsur-
face environments as well as the phases
in which organic contaminants may be
present and migrate through the porous
media that comprise the unsaturated
(above the water table) and saturated
(below the water table) zones. Upon re-
lease to the subsurface, the total mass
of each pollutant will be distributed
among the various phases by the move-
ment of vapors and liquids and diffu-
sion of the pollutants within thjjfi.
NAPL may be present in pools at near
saturation, having displaced most of the
pore water, or in "residuals" at "resid-
ual saturation," on the order of 1-10%
of the pore volume (6, 9) remaining
from contact with a migrating NAPL
slug. The mass of organic pollutant
present in a given volume of NAPL-
contaminated soil may be many orders
of magnitude greater than the mass
present if the soil volume were contam-
inated only with vapor, dissolved and
sorbed phases. We thus refer to the
pools and residuals as subsurface con-
taminant sources.
With regard to cleanup efforts, it is
clearly advisable to first remove the
NAPL sources from the subsurface if
all possible, because they may contaH
most or nearly all of the total mass cf
fugitive contaminants. For NAPL s
such as benzene and other petroleum
products, which tend to floa" '
groundwater, the p. have been succ ss
in pumping a sigr:5cant fraction o." th.
NAPL to the surface (JO). Yet for oth-
ers more dense than water (e.g., chlo-
rinated solvents, crersots, -mc T"
rich oils), very little success has been
achieved in even locating the subsur-
face NAPL sources, let alone removing
them (11).
For NAPLs that cannot be removed
directly or that remain in residual satu-
ration, the pollutant mass they contain
will generally have to be removed in a
much more dilute form, such as by va-
porization into the soil gas or by disso-
lution in groundwater. These removal
methods require the extraction of con-
siderable volumes of gas or water.
Practical experience indicates that the
cleanup process is lengthy and expen-
sive, especially for contaminants in the
saturated zone.
The problem of groundwater cleanup
is exacerbated by the desire to return
aquifers to drinking-water quality,
which for many important organic con-
taminants requires concentrations less
than 100 parts per billion (ppb) and in
some cases less than 5 ppb. However,
the concentrations in groundwater
withdrawn by wells are controlled in
part by transfer of contaminant mass to
Jie flowing water from other phases
acting as contaminant reservoirs: con-
taminant sorbed by the aquifer solids,
contaminant present in immobile pock-
ets of contaminated groundwater in less
permeable but porous strata or lenses,
vapor spreading from residual or
-------
pooled NAPL in the vadose zone, or
dissolution of residual or pooled NAPL
in the saturated zone. As discussed be-
low, such transfers of contaminant
mass can cause the extracted ground-
water to fail to meet drinking-water
standards for prolonged periods of
time, a problem compounded by the
slow rate at which these transfers often
occur.
Effects of contaminant desorption
from solids
Dissolved organic contaminants gen-
erally move more slowly through gran-
ular aquifers than the groundwater it-
self because of sorptive interactions
with the aquifer solids (12-16). Al-
though there have been only a few field
studies that have yielded quantitative
understanding of the relative mobility
of organic contaminants, a review of
them indicates—as expected from labo-
ratory studies and basic geologic
knowledge—that field retardation var-
ies among contaminants for a given site
and among sites for a given contami-
nant (8).
The greater the retardation, the more
time will be required to remove the
contaminants for a given pumping rate.
Furthermore, the removal of dissolved
and sorbed contaminants by pumping
requires the extraction of more water
than is contaminated at the onset of re-
mediation. Figure 2a illustrates this in
an idealized case for a contaminant
with a retardation factor of two (i.e, in
which the sorption/desorption interac-
tions cause the contaminant to move at
a constant fraction, 1/2, of the ground-
water velocity). The retardation factors
observed in sand-gravel aquifers for
contaminants such as those listed in Ta-
ble 1 vary from 1 to 33 (8); in other
media or for other contaminants, the
retardation may be even greater. Thus,
unless injection wells are used to supply
the clean water, which is rarely the
case, the "pump-and-treat" approach
may utilize a considerable volume of
uncontaminated groundwater surround-
ing the site to flush the contaminants
from the polluted area.
However, kinetic limitations to de-
sorption can occur during groundwater
extraction programs, as has been ob-
served in field studies (17, 18) and im-
plied by laboratory investigations (19-
22). The practical effect of these kinetic
FIGURE 2
Hypothetical examples of contaminant removal ;
(a) Uniform sand-gravel aquifer"
to
Contaminant concentration in extracted
water
t1
to
nsr
(b) Stratified sand-gravel aquifer
to
t1
	
	
		—i; »¦







tO t1
(c) Clay lens in uniform sand-gravel aquifer
to
to t1
(d) Uniform sand-gravel aquifer
¦
2. •/ •
to
t2
: IVf."..'.1

•Dens® color Indicates NAPL contaminant, sttppilng indicates corrtcmT^ni tn dissolved and corfcrd phesci*
(assumed untformty distributed rnio&Uy), and anwj indicate relaijv© vedoaiy ot grounzwaiz? flow. The
groundwater Is assumed to be extracted from the waJI el the samo rats in the four cases.
^Dotted fines cjvJqss toiaJ vc!umo o7 ivetsr tfiat uculd bs pumped to removo contzminsnt with rennfcion factor
0(2.
limitations is to slow the removal of the
contaminants from the aquifer, thereby
increasing both the time required to
achieve cleanup and the total volume of
water that must be extracted to flush the
¦¦contaminated zone. Furthermore, if
pumping is ceased before all of the con-
taminant is removed, the contaminant
concentrations in the groundwater will
rise as desorption continues (1). Whif-
fin and Bahr (17) observed such results.
Effects of geologic complexity
Although the processes that affect or-
ganic contaminant transport are essen-
tially the same in various sites, the opti-
mal design of a remediation program is
very site specific, mainly because of
geologic complexity. Sand and gravel
aquifers typically have silty or clayey
strata above, within, or beneath them
(Figure 1). These strata normally are
less permeable than the aquifer by a
factor of 1000 to 10 million. Ground-
water flow and contaminant migration
are distorted because of these strata.
The positions of NAPL sources in the
aquifer are commonly complex and un-
predictable because exact locations and
volumes of spills or leakages are un-
known, and, as illustrated in Figure 1,
the positions of subsurface NAPL
sources are determined by deflections
and pooling caused by stratification.
Current research suggests that even
slight heterogeneities can influence
NAPL penetration into porous media,
particularly in the saturated zone (23).
As plumes spread through aquifers,
the dissolved contaminants move
quickly through more permeable zones
while they slowly invade the less per-
meable ones by flow or diffusion (24).
Over the years and decades, this inva-
sion can cause the plume to occupy
large volumes of low permeability ma-
terial. To obtain clean water from
wells, it is generally necessary for the
lower permeability parts of the aquifer
system to be cleaned as well as the high
permeability zones.
Figure 2b illustrates an idealized case
in which the aquifer is composed of two
relatively distinct and horizontally con-
tinuous sand-gravel strata, the upper
stratum having a somewhat higher hy-
draulic conductivity than the lower. The
contaminant is assumed present in dis-
solved and sorbed phases, uniformly
spread in the volume indicated. If
groundwater is extracted from a fully
penetrating well or a well with most of
its screen in the more permeable zone,
as is common practice, the bulk of the
water will be moving through the upper
stratum. Thus, even in sand-gravel
aquifers, some strata may be flushed of
contaminants long before others. Fig-
ure 2b assumes that the retardation fac-
tor of the contaminant is two in both
-------
FIGURE 3
Schematic of subsurface environment composed of fractured rock
under the overburden*
Diffused
into and sorbed
onto rock matrix
Dissolved
The fracture system may lead to the appearance erf NAPL Of dissolved contaminant in unpredictable locations.
Dense ookx indicates NAPL contaminant: stippling indicates contsminnnt m the dissotrod phase, either In water
in the fractures or diffused into water held in the porous rock matrix.
THdiloroethytene.
strata. If the retardation is greater in the
lower stratum and desorption is kineti-
cally limited, then the flushing will be
even more inefficient (25).
Figure 2c illustrates another situa-
tion: a very low permeability clayey
stratum in the middle of an aquifer that
has been contaminated for decades.
Dissolved contaminants have perme-
ated the clayey stratum during this per-
iod primarily by molecular diffusion.
Although the permeability of a clayey
stratum is low, the porosity is usually as
large as or larger than that of the adja-
cent aquifer, thereby facilitating diffu-
sion into the clay. Furthermore, the ca-
pacity of clayey strata to sorb
contaminants may be much greater per
unit volume than that of the aquifer.
When the aquifer is flushed by clean
water, the only significant process for
release of the contaminant from the
clay will be a reversal of the diffusion
direction. The relatively slow rate of
release of contaminants from the clay
by diffusion and the potentially appre-
ciable contaminant mass contained in
dissolved and sorbed fo m in the clay
causes a long-term bleed of contami-
nants into the aquifer during remedia-
tion (1). In many aquifers there are nu-
merous thin beds of silt and clay that
634 Environ. Sci. Technol.. Vol. 23. No. 6, 1989
compound this problem of delayed dif-
fusive release. Such processes may
maintain contaminant concentrations in
the extracted water above typically low
cleanup criteria for very long periods of
time, gready increasing the duration
and cost of cleanup. In most investiga-
tions, however, the details of site geol-
ogy are not well enough understood to
allow prediction of the effect of delayed
diffusive release.
Effects of fugitive NAPL
As described above, organic NAPLs
are often suspected to be present in the
subsurface, but reliable estimates of the
volume of NAPL spilled or disposed of
exist for very few groundwater contam-
ination sites. Furthermore, even after
exceptionally detailed site investiga-
tions are conducted, it is normally not
possible to predict reliably where these
NAPL pools are. Not knowing the size
and location of NAPL pools and zones
of residual NAPL makes it impossible
to predict how long a pump-and-treat
program must operate in order to clean
the aquifer. Figure 1 shows how clayey
lenses in the aquifer can cause the
NAPL to be deflected laterally so that
most of the NAPL mass exists as iso-
lated pools away from the spill origin.
Perched pools of NAPL can jeopardize
site investigations because it is very
easy to unknowingly drill through the
pool and the bed it sits on, causing the
-------
deeper part of the aquifer or into a dif-
ferent aquifer. When pumping causes
lowering of the water table from a posi-
tion above perched NAPL pools or
NAPL residual zones to below them,
remobilization of the pools and resid-
uals may occur, allowing them to drain
deeper into the aquifer.
Figure 2d illustrates the effect that
even relatively small quantities of
NAPL might have on a typical cleanup
program based on groundwater extrac-
tion. In the general case the extraction
well is located at some distance from
the NAPL that has penetrated the satu-
rated zone. The groundwater extraction
might quickly remove the bulk of the
dissolved and sorbed contaminant that
had migrated from its NAPL source(s)
and thereby achieve an initial decrease
in concentration in the extracted water.
However, the NAPL present in pools or
residual saturation may be dissolved
slowly by the groundwater flowing
around or through them, maintaining
contaminant concentrations at signifi-
cant levels for a long time, even in rela-
tively uniform aquifers. When zones of
lower permeability are contaminated
with NAPL, the cleanup time would be
expected to increase considerably.
Feenstra and Cherry (11) present a
more complete review of these and
other issues regarding NAPL behavior
and cleanup of NAPLs denser than wa-
ter.
Cleanup of fractured rock aquifers
Generally, rock aquifers contain a
myriad of cracks (fractures) of various
lengths, widths, and apertures. Most
rock aquifers in North America are
permeable primarily because of the ef-
fective porosity provided by these frac-
tures rather than that of the rock ma-
trix, which is relatively impervious.
The effective fracture porosity of frac-
tured-rock aquifers is generally in the
range of 0.001-0.1%, which is much
smaller than the porosities of typical
granular aquifers (20-40%). For exam-
ple, a rock mass with one fracture per
linear meter with fracture apertures of
500 nm would be very permeable but at
saturation would have a very small
storage volume of mobile groundwater
(only about one-half liter per cubic me-
ter of rock).
When NAPL enters such aquifers, it
flows mainly through the intercon-
nected fractures and settles out in dead-
end segments of the fracture system
(Figure 3). Relatively small volumes of
NAPL can move deep and far into the
rock because the retention capacity of-
fered by the dead-end fractures and the
immobile filaments and globules in the
larger fractures is so small—much less
than the percentage given above for wa-
ter in saturated fractured rock. Al-
though the rock matrix typically has a
relatively small intergranular porosity,
it is commonly large enough to allow
dissolved contaminants from the frac-
tures to enter the matrix by diffusion
and be stored there by adsorption, as
shown in Figure 3.
The prognosis for cleanup of frac-
tured rock aquifers, particularly those
containing NAPL contaminants, is
worse than for sand and gravel aqui-
fers. Even if the location of the spill is
known exactly, the location of the
NAPL is typically difficult or impos-
sible to determine from site investiga-
tions. This is because NAPL pathways
"...cleanup of groundwater
contamination by organic
chemicals typically pro-
ceeds slowly using the
common pump-and-treat
approach"
At
through the fracture system are excep-
tionally complex and distribute the
NAPL into many small and scattered
amounts (Figure 3). When attempts are
made to clean such fractured rock aqui-
fers by pumping water, major improve-
ments in water quality are exceedingly
slow because little or no water flushes
through dead-end fracture segments or
through the porous but impervious rock
matrix, both of which are likely tcyfe-
tain the bulk of the contaminated mass.
Such has been the experience at an or-
ganic liquids disposal site in Ville Mer-
rier, Quebec, where the effectiveness of
pump-and-treat remediation of the
large plume has been severly hampered
by the penetration of the NAPLs into
the fractured bedrock (26).
Summary and implications
We have explored, via simplified ex-
amples and illustrations, many of the
reasons cleanup of groundwater con-
tamination by organic chemicals typi-
cally proceeds slowly using the com-
mon pump-and-treat approach. At
many sites of significance, a relatively
large mass of contaminants has been
leaked, spilled, or disposed into the
subsurface, and in comparison the rate
of contaminant mass removal by pump-
ing wells is exceedingly slow. In such
cases the pump-and-tre:it option is best
thought of as a management tool to pre-
vent, by hydraulic manipulation of the
aquifer, continuation of contaminant
migration.
This option often effectively shrinks
the plume toward its source(s), but for
the shrinkage to persist it is necessary
for the pumping to continue. Even if
contaminants remain in some portions
of the aquifer, it is often possible by
such -hydraulic influence to eliminate
real, potential, or perceived risks to
public health. However, the long-term
cost of such pumping with treatment of
the extracted water is often high
whether measured in dollars spent for
system operation and maintenance or in
gallons of previously uncontaminated
groundwater used to flush out the con-
taminants. The mass of NAPL at or be-
low the water table is not known with
sufficient detail at most sites to make
reliable predictions of the time neces-
sary for cleanup by pump-and-treat
programs. In general, it is appropriate
to view such approaches as remediation
in perpetuity.
At quite a few sites, the dissolved
mass of organic contaminants in rela-
tively large plumes is quite small, less
than that present in one drum of the
pure chemical. This suggests that seem-
ingly innocuous and often unnoticed
leakages or spills of a few gallons per
day or less, so common at many indus-
trial and military sites, may pose a ma-
jor threat to groundwater. Considering
the extremely high cost per equivalent
NAPL gallon of removing the contami-
nant from aquifers, it is clear that the
economic advantage of preventing such
small leaks or spills is immense.
A number of new technologies are
under development for groundwater re-
mediation, as explained in later articles
in this series, which may accelerate
contaminant removal from the subsur-
face (e.g., injection of steam, surfac-
tants) or destroy the contaminant in situ
(e.g., bioreclamation). For many if not
most sites, it is important to recognize
that all of these technologies will be
severely hampered by geological com-
plexities and the difficulty of locating
the subsurface contaminant sources.
Laboratory studies and small-scale field
prototype trials are likely to yield over-
optimistic expectations for the applica-
tion and efficiency of these technolo-
gies.
Site characterization programs often
seem to the public and regulators to be
inefficient and excessively lengthy.
Shortening site characterization efforts,
a currently popular demand, may be
appropriate for pump-and-treat pro-
grams intended to prevent contaminant
migration, but cannot be expected to be
satisfactory for permanent aquifer
c' -anup. At sites where permanent
c. anup is the goal, detailed and accu-
rate site characterization is a prerequi-
site for a reasonable probability for suc-
cess (1). As reviewed in other papers in
: lis series and elsewhere (1, 8), there is
-------
much potential for improvement in site
characterization and remediation,
through both the development of new
tools and the continued training of ev-
eryone involved: site owners, consult-
ing engineers and hydrogeologists, reg-
ulators, and the public.
Acknowledgements
Preparation of this paper was supported in
part by the National Science Foundation
through its sponsorship of the U.C.L.A.
Engineering Research Center for Hazard-
ous Substance Control, and by the Univer-
sity of Waterloo part of the University
Consortium Solvents-in-Groundwater Re-
search Program.
We are grateful for comments and assist-
ance from B. Kueper (University of Water-
loo), R. Jackson (Environment Canada),
S. Feenstra (Applied Groundwater Re-
search, Ltd.), and D. McWhorter (Colo-
rado State University). Numerous people
assisted in the preparation of the table, in-
cluding J. Anderson (IBM), M. Brother
(Eckenfelder, Inc.), G. Patrick (Golder As-
sociates), K. O'Brien (Pacific Environ-
mental Services), C. Haddox (Ebasco
Services, Inc.), T. Hunt (Harding-Lawson
Assoc.), J. Armstrong (Traverse Group),
and L. Barber (U.S. Geological Survey).
References
(1)	Robert S Kerr Environmental Research
Laboratory "Practical Limits to Pump
and Treat Technology for Aquifer Reme-
diation", paper prepared for (he ; I S En-
vironmental Protection Agun<-> Ada,
OK, 1987
(2)	Westrick, J. J ; Mello, W J.; Thomas.
R F J Amer. Water Works Assoc 1985,
6, 52-59.
(3)	Mackay, D. M , Gold, M , Leson, G In
Proceedings of the I6lh Biennial Confer-
ence on Ground Water, Devries, J. J ,
Ed , University of California Water Re-
sources Center Report No 66, Davis,
CA, 1988, pp 97-110
(4)	Barbash, J E , Roberts, P. V. J. Water
Pollut Control Fed 1986, 58(5), 343-
48.
(5)	Mackay, D. M.; Roberts, P V, Cherry.
J. A Environ. Sci. Technol 1985, 19(5),
384-92.
(6)	Schwille, F. Dense Chlorinated Solvents
in Porous and Fractured Media. Model
Experiments (Engl transl.); Lewis Pub-
lishers' Ann Arbor, MI, 1988.
(7)	Mendoza, C. A., McAlary, T. A.
Groundwater, in press.
(8)	Mackay, D M. Presented at the National
Research Council colloquium on ground-
water and soil contamination remedia-
tion Washington, DC, April 1989.
(9)	Hoag, G. E.; Marley, M. C. J. Environ.
Eng Div. (Am. Soc. Civ. Eng.) 1986,
112(3), 586-604.
(10)	Roy F. Weston, Inc., University of Mas-
sachusetts, Environmental Science Pro-
gram, Division of Public Health; Reme-
dial Technologies for Leaking
Underground Storage Tanks; Lewis Pub-
lishers: Chelsea, Ml, 1988.
(11)	Feenstra, S., Cherry, J. A In Proceed-
ings of International Groundwater Sym-
posium; International Association of Hy-
drogeologists. Halifax, Nova Scotia,
May 1988; pp 61-69.
(12)	Roberts, R V; Schreiner, J.. Hopkins.
G D. Water Res. 1982, 16. 1025-35
(13)	Patterson. R. J. et al. Water Sci. Technol.
1985, 17. 57-69
(14)	Schwarzenbach, R. P. et al. Environ. Sci.
Technol 1983, 17. 472-79
(15)	Mackay, D. M et al. Water Resour Res.
1986, 22(13). 2017-30
(16)	Roberts, P V, Goltz, M. N ; Mackay,
D M Water Resour. Res. 1986, 22(13),
2047-58.
(17)	Whiffin, R. B , Bahr, J M In Proceed-
ings of the Fourth National Symposium on
Aquifer Restoration and Ground Water
Monitoring; National Water Well Associ-
ation Worthington, OH, 1985, pp. 75-
81.
(18)	Bahr, J. M. J. Contam Hydrol., in press.
(19)	Karickhoff, S J Hydraul. Eng. 1984,
110(6), 707-35.
(20)	Curtis, G L., Reinhard, M., Roberts.
P V. Water Resour Res. 1986, 22(13),
2059-67
(21)	Wu, S-C , Gschwend, P M Environ.
Sci. Technol 1986, 20. 717-25
(22)	Ball, W P Ph D Dissertation, Stanford
University, Stanford, CA, 1989
(23)	Kueper, B H , McWhorter, D B., Frind,
E O In Proceedings of the International
Symposium on Contaminant Transport in
Groundwater; International Association
for Hydraulic Research, Stuttgart, FRG,
April 1989.
(24)	Gillham. R W Cherry, J. A. In Recent
Trends in Hydrogeology. T. N Nori-
simhan (Ed.), Geological Society of
America (Special Publication), 1982,
189. pp. 31-62
(25)	Mackay, D M . Michelsen, C , Thorb-
jarnarson, K. Abstracts of Papers; Amer-
ican Geophysical Union Conference, San
Francisco, EOS 1988, 69(4). p. 1197
(26)	Martel, R "Groundwater Contamination
by Organic Compounds in Ville Mercier.
New Developments", report to the
NATO/CCMS Pilot Study of Remedial
Action and Technologies for Contami-
nated Land and Groundwater. Bilthoven,
The Netherlands, Nov 1988
(27)	Barber. L. E II et al. Environ. Sci Tech-
nol. 1988, 22. 205-11
(28)	Armstrong, J. M , Sammons, J. H. Pro-
ceedings of the 1986 Hazardous Materi-
als Conference. May 1988, St. Louis,
MO
(29)	Rifai, H. S et al. J Environ. Eng 1988,
114(5), 1007-29
Douglas M. Mackay (I) is an assistant pro-
fessor in the Environmental Sciences and
Engineering Program of the UCLA School
of Public Health, with graduate degrees in
civil engineering. Since 1981 his research
has focused on integrated field and labora-
tory investigations of the transport and re-
mediation of organic contaminants in the
subsurface.
John A. Cherry (r) has degrees in geologi-
cal engineering ¦ >'.d hydrology and, since
1971, has been a professor at the Univer-
sity of Waterloo where he is a member of
the Waterloo Centre for Groundwater Re-
search. For the past 20 years, his research
has focused on processes and monitoring
of groundwater conuimination.
-------
GOALS AND EFFECTIVENESS OF PUMP AND TREAT REMEDIATION
VOLUME I
A Review of Selected Case Studies of Large Plumes
of Chlorinated Solvents or Pesticides in Sandy Aquifers
J. Harman, D. M. Mackay and J. A. Cherry
Waterloo Centre for Groundwater Research
University of Waterloo
Waterloo, Ontario, Canada N2L 3G1
-------
EXECUTIVE SUMMARY
The majority of'groundwater remediation programs in the United States has used the groundwater
remediation method known as "pump and treat" (hereafter abbreviated as P&T) in which contaminated
groundwater is extracted from the aquifer, treated to prescribed contaminant levels and then disposed. In
the early years of its application, this method was expected to permanently remediate the contaminated
zones and thereby restore the aquifers to drinking water standards within reasonable time frames (1 to 10
years). However, remedial performance data indicates that actual progress in aquifer restoration is
generally much slower than originally expected.
This study provides an analysis of past P&T performance at selected sites that overlie sand and
gravel aquifers. The primary goals are to better understand the capabilities and limitations of the P&T
approach and to propose procedures for performance assessments that may be readily applied at other
sites.
The eleven contaminant plumes evaluated in this study were selected qualitatively based on size,
long histories of site investigations, and abundance of site monitoring information. This review is restricted
to plumes in sand and gravel aquifers because, in comparison to plumes in other types of hydrogeologic
settings such as fractured rock, they generally are defined more completely and offer better prospects for
full restoration. At 10 of the 11 sites, the predominant groundwater contaminants were chlorinated organic
chemicals typically used as industrial solvents (at the remaining site, the contaminants are halogenSted
alkane pesticide&-with expected groundwater transport properties similar to those of chlorinated solvents).
At 7 sites, P&T remediation has been underway for seven to thirteen years. At 1 site, P&T began as this
report was being finalized. At the other 3 sites, detailed characterization was complete or nearly so and
P&T systems were being planned or designed. Of the 11 sites, 7 are regulated under the Superfund
program. The other 4 are regulated by state or federal agencies under other programs.
Our review indicates that, with one exception, only recent remedial designs have recognized that
subsurface contamination by chlorinated solvents often is comprised of two conceptually different zones:
1) the subsurface source zone, in which dense nonaqueous phase organic liquids (a.k.a. DNAPL's) may
exist, and 2) the plume, in which only dissolved and sorbed contaminants are present. If the source zone
is significant, it will often be the case that far more contaminant mass is present there than in the plume
itself, where contaminant concentrations are often relatively low. It is our conclusion that the progress of
many remedial programs has appeared slow because the expectations for the programs were incorrectly
based on the assumption that only the plume existed in the subsurface. In other words, until recently it has
been the rare case in which the significance of subsurface source zones has been recognized in the
-------
remedial design process for chlorinated solvent contamination.
None of the 7 operating P&T systems reviewed in this study was found to have completely restored
the aquifer. Here we define complete restoration as removal of contaminants from both the concentrated
subsurface source zone and the more dilute but generally much more extensive plume. This observation
is consistent with anecdotal information obtained from a wide variety of sources by the authors over the
last several years; except for petroleum contamination, there appear to be no well-documented cases in
which significant VOC plumes as well as the subsurface source zones have been fully remediated by P&T.
In three of the cases, partial restoration of the aquifer was achieved by reducing the contaminant
concentrations within the dissolved plume to satisfy the regulatory agency or proposed cleanup criteria.
This degree of success seemed to require that the contaminant mass in the subsurface source zone either
be insignificant, removed or isolated from the remainder of the plume by hydraulic or physical means.
Since, in general, the contaminant mass in the subsurface source zones is unlikely to be insignificant if the
plumes are significant, our review^uggests that partial restoration of the aquifer can be expected only if
the subsurface source zone is isolated to prevent regeneration of the plume. On the other hand, this review
confirms that P&T can, under certain circumstances, sufficiently restore considerable portions of an aquifer
to satisfy regulatory agencies or meet cleanup criteria and prevent the spread of contaminants from source
zones. This condition of partial aquifer restoration can be maintained as long as P&T or physical isolation
continues as a source-zone control measure.
In several.of the cases examined, plume containment (i.e. capture) was specified as the remedial
goal, with gradual mass removal as a secondary goal and automatic consequence. Success in achieving
containment, as defined by these goals, varied from site to site. Only two P&T systems with plume
containment as the goal were designed in a manner to capture the entire plume. At one site which is not
yet operating, the pumping wells are to be positioned at the front of the plume. At the second site the wells
were located within the plume and pumped at rate predicted to attain full capture; however, the system has
not yet been in operation long enough for capture to be confirmed via monitoring data.
At other sites, the remedial goal appeared to be partial plume containment or plume cutoff rather
than plume containment. Generally, these older extraction systems were not installed at the front of the
plumes but rather to varying degrees within them, often at site boundaries. In these cases, the wells do
not appear to capture the entire plume and some of the plume may have been lost to migration beyond
the influence of the extraction system. However, once installed, the P&T systems appear successful in
preventing further migration of contaminants beyond therr.. Phis suggests that, if installed in the path of
a migrating plume, P&T may under many ;oncitions sue sjd in plume capture. Thus, P&T can be an
-------
effective means of limiting or removing risk of exposure to contaminated groundwater.
Free-phase, dense non-aqueous phase liquid (DNAPL) was confirmed to be present (detected) in
the subsurface at only 1 of the 11 sites. However, our review indicated that, at most of the sites, DNAPLs
were very likely present in subsurface source zones below the water table, acting as significant long-term
subsurface contaminant sources. In many cases, other contaminants may be present as well (e.g.
nonchlorinated solvents and/or fuels). In one case, it appears that an LNAPL (non-aqueous phase liquid
less dense than water) was present which had chlorinated solvents dissolved within it. The importance or
even existence of DNAPL sources was not generally recognized at the time that most of the remedial
measures went into operation at these sites. Among the sites examined in this study with operating P&T
systems for plume restoration, the presence of a significant subsurface source of contaminant mass was
explicitly considered in initial remedial designs in only one. We believe that progress towards partial aquifer
restoration was most rapid at that site, primarily because the source zone was isolated early using a cutoff
wall enclosure with pumping inside the enclosure, preventing continued regeneration of the plume.
However, it is important to note th^t though progress has been made toward cleanup of the plume,
remediation of the isolated source zone continues with no certainty of the completion date.
The initial volumes of the studied plumes ranged from 115,000 to 230,000,000 m3. Prior to P&T,
the flux of groundwater and contaminants through these plumes ranged from 3.4 to 46 Us and 5 to 1700
kg/yr, respectively. The contaminant mass estimated to be dissolved within the plumes prior to P&T
generally ranged from 44 to 9,000 kg (0.133 to 33 drums), although one site had an estimated dissolved
mass of over 80,000 kg (275 drums). At the 11 study sites, it was not possible to estimate the total mass
of contaminants in the subsurface, since neither the sorbed or DNAPL mass in the subsurface source
zones nor the sorbed mass in the plumes are known or reliably estimated.
P&T systems achieved mass removal rates from 100 to 3,000 kg/yr (0.34 to 10 drums per year).
Although the mass removal rates are significant at some sites, the actual progress towards complete clean-
up cannot be assessed at any of the sites, because the total initial mass is unknown. The discrepancy
between the minimum mass removal (100 kg/yr) and the minimum dissolved mass in place (44 kg)
suggests that additional mass is contributed to the dissolved plume during the remediation. This additional
mass is likely from dissolution of source zones or desorption from the aquifer media. Despite the fact that
the total mass in the subsurface cannot be estimated, these other types of calculations proved useful for
structuring our review and highlighting apparent inconsistencies in data or their interpretation.
There is evidence for many of the plumes evaluated herein that natural in-situ transformation
processes are converting some of the parent contaminants to other contaminants. Although apparently not
-------
the case at the sites reviewed in this study, such transformations could yield nontoxic daughter products
and proceed at rapid enough rates to either accelerate the progress of pump and treat remediation or
render it unnecessary (i^e. no action required to prevent risk). In contrast, at several of the sites reviewed
in this study, anaerobic biotransformation of the solvents apparently yielded chlorinated daughter products
which were of regulatory concern. In two of the cases reviewed in this study, natural abiotic transformation
of 1,1,1-TCA appears to have produced sufficient mass of 1,1-DCE such that the daughter product now
is the constituent controlling the achievement of the negotiated remediation goal (note that the drinking
water criterion for 1,1-DCE is significantly lower than for 1,1,1-TCA , i.e., 6 ^xg/L vs 200 (a.g/L). In one of
these cases there is apparently a continuing source of 1,1,1-TCA to maintain the groundwater plume.
Therefore, there is a continuing source for in-situ production of 1,1-DCE. It is likely that plume restoration
in this case will not be possible without source removal or isolation.
In summary, our review indicates that well-executed P&T technology, despite its limitations, has
been and can be successful for partial aquifer restoration. However, P&T or other source control or
removal measures must be continue^ in order to maintain the condition of the partially restored aquifer.
In other cases, P&T can be used effectively to contain plume migration and reduce or eliminate risk even
without restoring the existing plume.
The primary factor preventing P&T from achieving complete aquifer restoration appears to be the
continued presence of DNAPL zones below the water table. These source zones are likely to be present
and significant at most sites contaminated with chlorinated organic chemicals, yet they have rarely been
addressed directlyjn remedial design until recently. In addition, evidence suggests that such source zones
should be expected even if no direct observations can be made to locate them precisely. This conclusion
implies that better remedial designs would be likely if such source zones were simply assumed to exist and
then roughly located so that some form of source containment could be accomplished.
Finally, it is important to realize that the same factors that limit the effectiveness of P&T systems
for source zone restoration are also likely to limit most, if not all, of the alternative technologies that require
active manipulation of groundwater or air in the subsurface as replacements of or enhancements to P&T.
This includes chemically enhanced pump-and-treat, such as surfactant or alcohol flushing. If this is true,
then although these new technologies may initially offer advantages in rate of approach to the cleanup
standard, they too may be unable to reach the required goals in reasonable time frames. This study
suggests that a realistic approach for groundwater remediation at sites with appreciable quantities of
chlorinated solvent contaminants in sand and gravel aquifers is subsurface source-zone isolation or control
using cutoff-wall enclosures, hydraulic manipulation of the groundwater zone, or passive in situ treatment
systems. In addition where conditions allow, the dissolved groundwater plume outside the isolation or
-------
control boundary can be removed, or at least captured, by P&T.
-------
-------