&EPA Ground Water Issue
United States
Environmental Protection
Agency
Assessment and Delineation of DNAPL
Source Zones at Hazardous Waste Sites
Bernard H. Kueper* and Kathryn L. Davies**
1.0 - Introduction
Groundwater contamination from classes of chemicals such as
chlorinated solvents, polychlorinatedbiphenyls (PCBs), creosote.
and coal tar is frequently encountered at hazardous waste sites
(40, 43). These types of contaminants have low solubilities in
water and have densities greater than that of water. Therefore.
they can exist in the subsurface as Dense, Non-Aqueous Phase
Liquids (DNAPLs) and have the potential to migrate as a sepa-
rate liquid phase to significant distances below the water table in
bothunconsolidated materials and fractured bedrock. Because of
the physicochemical properties associated with DNAPLs, they
migrate through the subsurface in a very selective and tortuous
manner (13, 27, 29). Thus, the majority of DNAPL present in
the subsurface may not be found immediately below the entry
location and directly encountering DNAPLs with conventional
drilling techniques may be difficult.
Determining the presence or absence of a DNAPL is an impor-
tant component of the conceptual site model and is critical to the
proper selection of the remediation approach. Subsurface DNAPL
acts as a long-term source for dissolved-phase contamination and
determines the spatial distribution and persistence of contaminant
concentrations within the dissolved-phase plume. Once it hasbeen
determined that DNAPL exists within the subsurface, subsequent
characterization activities are typically conducted to better de-
lineate the boundaries of the DNAPL source zone. The DNAPL
source zone is the overall volume of the subsurface containing
residual and/orpooledDNAPL. It shouldbe recognized that there
will be uncertainty associated with the delineation of the DNAPL
source zone. In addition to the DNAPL, there may be significant
amounts of contaminant mass that have diffused into low perme-
ability zones. Back diffusion of contaminant mass from these
zones may sustain dissolved-phase plumes for significant periods
of time, even after DNAPL has been removed. Establishing the
presence and locations of such non-DNAPL sources is beyond
the scope of this document.
In January 1992, EPA published a Fact Sheet entitled 'Estimat-
ing Potential for Occurrence of DNAPL at Superfund Sites' (42)
with the goal to help site personnel determine if DNAPL-based
characterization strategies shouldbe employed at a particular site.
In September 1994, EPA issued a subsequent Fact Sheet entitled
'DNAPL Site Characterization' (3 9) discussing direct and indirect
methods to assess the presence of DNAPL in the subsurface. Since
Queen's University, Kingston, Ontario CANADA
U.S. EPA, Region 3
the publication of the initial fact sheets, there have been advance-
ments incharacterizationtools, site investigation approaches (14)
and knowledge of DNAPL source zone architecture within the
subsurface. This document builds on information from the previ-
ous fact sheets to provide a framework for not only assessing the
presence of DNAPL, but also for delineating the spatial extent
of the DNAPL source zone, a priority at many sites due to the
more prevalent use ofin-situ remediation technologies (38). The
strategy described in the present document utilizes converging
lines of evidence that incorporate the scientific advancements in
the field and expands the applicability of the document to include
both unconsolidated deposits and fractured bedrock. An iterative.
flexible site investigation approach (7) is encouraged.
2.0 - Nature of the DNAPL Source Zone
Uponrelease to the subsurface, DNAPL will distribute itself in the
form of disconnected blobs and ganglia of organic liquid referred
to as residual DNAPL, and in connected distributions referred to
as pooled DNAPL (Figure 1). Residual DNAPL is found both
above and below the water table within the pathways of DNAPL
migration, and typically occupies between 5% and 30% of pore
space in porous media (6,27,44) and in rock fractures (21). Re-
sidual DNAPL is trapped by capillary forces, and typically will
not enter an adjacent monitoring well, even under the influence
of aggressive groundwater pumping (6, 27).
Pooling of DNAPL can occur above capillary barriers, which are
typically layers and lenses of slightly less permeable material
(Figure 1). Pooling can therefore occur at any elevation in the
subsurface, and not just at the base of permeable zones. Absence
of pooling above clay aquitards and bedrock may be due to the
presence of dipping fractures, bedding planes, joints and faults
which may allow the continued downward migration of the
DNAPL. Pools represent a continuous distribution of DNAPL.
and typically correspond to DNAPL saturations of between
30% and 80% of pore space in both porous media and fractures.
The frequency of pool occurrence and the thickness of pools
are increased by the presence of horizontal capillary barriers,
lower DNAPL density, higher interfacial tension, and an upward
component to groundwater flow (17,22). The thickness of pools
typically ranges from fractions of an inch to a few feet, depending
on fluid and media properties (36) as well as the volume released.
Because pools represent a connected distribution of DNAPL, the
pooled DNAPL is susceptible to mobilization through drilling
activities and can short-circuit along existing monitoring wells
and piezometers. In addition, pools may also be mobilized in
response to changes in hydraulic gradient. The gradient required
to mobilize a pool is a function of the DNAPL-water interfacial
-------
MW1
DNAPL
Entry
Dissolved-Phase
Plume
DNAPL Pool
in Fracture
Dissolved-Phase Plume
in Fracture and
Diffused into Matrix
Residual DNAPL
in Fracture
Figure 1 - Schematic illustration of contamination associated with a DNAPL release. Note that DNAPL migrates in three dimensions,
and that residual DNAPL accumulated above bedrock is the result of the release at ground surface. The reader is referred
to Figure 2 for a depiction of matrix diffusion. Figure is not to scale.
tension, the pool length, and the permeability of the surrounding
material (6, 27). Pumping groundwater from beneath DNAPL
pools, for example, can lead to an increase in capillary pressure
and subsequent downward DNAPL mobilization.
The spatial distribution of residual and pooled DNAPL is strongly
influenced by geology, and also by DNAPL properties and release
history (frequency, intensity, duration, volume and location).
DNAPL migration can occur through lenses and laminations of
porous media at the scale of inches or less (17,29). ForDNAPLs
that are non-wetting (see wettability in glossary) with respect to
water (which is usually the case), migration below the water table
is typically through the largerpores (and hence higherpermeability
regions) in unconsolidated media and larger aperture fractures in
bedrock. The orientation of stratigraphic and structural features
will largely determine the degree of lateral and vertical DNAPL
spreading. DNAPL migration from the release location can occur
in any direction, and is typically not greatly influenced by low
ambient hydraulic gradients except for creosotes and coal tars
which have densities close to that of water.
The overall region of the subsurface containing residual and
pooled DNAPL is referred to as the DNAPL source zone. For
high density and low viscosity DNAPLs (such as chlorinated
solvents), migration in relatively permeable media can cease as
soon as a few months to a few years following the time of release
(3, 17, 27, 29). Some geological conditions, such as horizontal
to sub-horizontal fractures, gently dipping strata and sand seams
in low permeability media can give rise to longer time scales
for migration of chlorinated solvent DNAPLs, particularly for
large volume DNAPL sources. For low density and high viscos-
ity DNAPLs (such as creosote and coal tar), migration has the
potential to continue for many decades (12). The overall depth
of DNAPL migration is dependent not only on the presence or
absence of capillary barriers, but also on the volume released, the
interfacial tension, the degree of lateral spreading, and the bulk
retention capacity (see glossary) of the medium. Because frac-
tured rock has very low bulk retention capacity, small volumes of
DNAPL can migrate greater distances in bedrock in comparison
to the same volume released into unconsolidated deposits (18).
Groundwater flowing past residual and pooled DNAPL will result
in dissolved-phase plumes of contamination. Complete dissolution
of allDNAPLasaresult of natural groundwaterflow is expected to
take from several decades to hundreds of years for most DNAPLs.
For multi-component DNAPLs, the presence of more than one
component typically suppresses the aqueous solubility of the
other components in the DNAPL (6, 27). Exceptions to this can
occur, however, when co-solvents such as alcohols are present in
the DNAPL. In the absence of co-solvents, the concentration of
any particular component dissolving into groundwater can often
be approximated using Raoult's Law (2, 6, 27). Early in the dis-
solution process, the plume chemistry will be dominated by the
higher effective solubility components which tend to be those
present in the largest mass fraction within the DNAPL, and those
-------
with the highest single-component (handbook) solubility values
(24). The concentration of any or all components in groundwa-
ter downgradient of a multi-component-DNAPL source zone
will typically be lower than expected using a single component
solubility limit. With time, both the DNAPL composition and
the plume composition will change in response to the dissolution
process. The dissolved components that comprise the plume will
migrate in groundwater subject to advection, dispersion, sorption.
volatilization, and degradation processes.
Both residual and pooled DNAPL, and dissolved-phase plumes
that are in direct contact with clays, silts, or a porous bedrock
matrix, can diffuse into the low permeability media (forward dif-
fusion). If concentrations outside of the low permeability zone
become lower than those inside, diffusion will occur back into
the higher permeability zone (back diffusion) and can result in
plume persistence (5, 33). The forward and back diffusion pro-
cesses are collectively referred to as matrix diffusion (Figure 2).
The persistence of DNAPL in fractures in bedrock, saprolite and
clay can be shortened by the matrix diffusion process (19, 28).
In addition, the rate of advance of a dissolved-phase plume in
fractured rock with a porous matrix can be strongly attenuated
by the matrix diffusion process (20, 35). The influence of matrix
diffusion on dissolved-phase plume migration in fractured rock
and clay relative to other processes such as advection, dispersion.
sorption, and possible degradationprocesses will vary depending
on site specific geological conditions and contaminant properties.
In general, matrix diffusion has a greater influence on dissolved-
phase plume migration in the case of wider fracture spacing.
smaller fracture aperture, lower hydraulic gradient, higher matrix
porosity, and higher matrix organic carbon.
Above the water table, volatile DNAPL can vaporize into air
filled pore spaces (Figure 1). For DNAPLs with significant
vapor pressure, this can lead to expanded vapor-phase plumes
in the unsaturated zone. The concentration of contaminants in
the vapor phase will be governed by the vapor pressure, and for
a multi-component DNAPL can often be approximated using
Raoult's Law. In relatively warm and dry environments, the
persistence of some DNAPLs (e.g., chlorinated solvents) can
be relatively short (on the order of months to a few years) in
unsaturated media. The absence of residual and pooled DNAPL
in the unsaturated zone may not, therefore, be sufficient evidence
to conclude that DNAPL has not migrated below the water table
at the site of interest.
3.0 -Types of DNAPLs
Coal Tar is a complex mixture of hydrocarbons produced through
the gasification of coal that was produced as a by-product of
manufactured gas operations as early as 1816 in the United States.
It is still produced as a by-product of blast furnace coke produc-
tion. Coal tar contains hundreds of hydrocarbons, including light
oil fractions, middle oil fractions, heavy oil fractions, anthracene
oil, and pitch. The low density (typically 1.01 g/cc to 1.10 g/cc
Groundwater
Flow
1
Diffusion Front
./ in Matrix
T
Dissolved-Phase
-Plume in Fracture
-1.
Residual
DNAPL
Dissolution of Residual
DNAPL Into Rock
Matrix and Fracture
Distance from DNAPL
Time
Figure 2 - Matrix diffusion of dissolved-phase contaminants adjacent to DNAPL and along length of plume in fracture. Matrix diffu-
sion can attenuate the rate of plume advance in fractured rock (bottom ten concentration vs distance plot), and can result
in delayed breakthrough curves (bottom right concentration vs time figure). These factors need to be considered when
relying upon groundwater concentration data to assess DNAPL presence.
-------
compared to 1.00 g/cc of water [at 4°C]) and high viscosity (up
to 200 to 300 times, or more, than that of water) facilitate long
time-scales of migration, with the possibility of movement con-
tinuing for many decades following initial release. Due to the
lengthy list of compounds present in coal tar, many investigators
select a sub-set of coal tar compounds based on mobility and
toxicity to assess water quality. These compounds may include
benzene, toluene, ethylbenzene, xylenes (BTEX),benzo[a]pyrene,
naphthalene, and phenanthrene. Depending on the age of the
DNAPL and groundwater velocity, some of the lower molecular
weight and more soluble compounds of the coal tar may have
been leached out of the DNAPL by the time a site investigation
is initiated. Naphthalene is often the dominant compound in
present day coal tar (9). In addition, the various components in
the plume will migrate at different velocities because of varying
degrees of sorption and degradation (often aerobic conditions).
The lowermolecularweight, less sorbingcompounds (e.g., BTEX)
can migrate significantly further in groundwater than the higher
molecular weight, more sorbing compounds (e.g., PAHs).
Creosote is composed of various coal tar fractions and was
commonly used to treat wood products. It is still used today in
certain wood treating operations and as a component of roof-
ing and road tars. Creosote is a multi-component DNAPL that
contains many hydrocarbons, primarily polycyclic aromatic hy-
drocarbons (PAHs), phenolic compounds, and carrier fluids such
as diesel. The low density (typically 1.01 g/cc to 1.13 g/cc) and
high viscosity (typically 20 to 50 times that of water) of creosote
facilitate long time-scales of migration, with the possibility of
movement continuing for many decades following initial release.
Most investigators select a sub-set of creosote compounds, based
on mobility and toxicity to characterize water quality, such as
naphthalene, benzo(a)pyrene, and phenanthrene.
Polychlorinated Biphenyls (PCBs) are a class of 209 chemical
compounds referred to as congeners, inwhichbetweenone and ten
chlorine atoms are attached to abiphenyl molecule. The majority
of PCBs were manufactured between 1930 and 1977 under the
trade-name Aroclor for use in capacitors, transformers, printing
inks, paints, pesticides, and other applications. Aroclors differ
based on the amount and types of congeners present. PCBs by
themselves are DNAPLs, and were often blended with carrier
fluids suchas chlorobenzenes and mineral oilpriorto distribution.
The density of most PCB oils ranges from 1.10 g/cc to 1.50 g/cc,
while the viscosity ranges from 10 to 50 times that of water. Most
congeners are very hydrophobic and their transport canbe retarded
strongly relative to the rate of groundwater migration. In some
cases, however, PCB transport in groundwater canbe facilitated
through the formation of emulsions or the presence of colloids.
Chlorinated Solvents suchas trichloroethene (TCE), tetrachloro-
ethene (PCE) and carbontetrachloride (CT) have been produced in
large quantities since the mid 1900's. Some chlorinated solvents
contain trace amounts of stabilizers, preservatives and impuri-
ties. Typical uses vary widely and include dry cleaning, metal
degreasing, pharmaceutical production, pesticide formulation,
and chemical intermediates. Chlorinated solvents canbe encoun-
tered as single component DNAPLs (e.g., as primarily PCE at a
dry cleaning facility, or as primarily TCE at a vapor degreasing
facility), or as part of a multi-component DNAPL containing
other organic compounds. The relatively high density (typically
1.10 g/cc to 2.20 g/cc) and low viscosity (typically ranging from
half to twice that of water) of chlorinated solvents can result in
a relatively short time-scale of migration following release com-
pared to coal tar and creosote. In a dissolved-phase plume, most
chlorinated solvents are not retarded strongly relative to the rate
of groundwater flow.
Mixed DNAPLs ADNAPL that contains two or more compounds
is referred to as a multi-component DNAPL (e.g., creosote). A
mixed DNAPL is a multi-component DNAPL that contains a wide
variety of organic compounds as a result of blending and mixing
prior to disposal operations, or as a result of cotemporaneous dis-
posal. Examples include DNAPLs encountered at former solvent
recycling facilities and industrial disposal sites. SuchDNAPLs can
contain aromatic compounds normally associated with LNAPLs
(e.g., toluene) along with chlorinated solvents, PCBs, alcohols,
ketones, and tetrahydrofuran. The density of mixed DNAPLs
typically ranges from 1.01 g/cc to 1.60 g/cc, and the dissolved-
phase plumes associated with mixed DNAPLs usually contain a
wide variety of compounds with varying mobility.
4.0 - DNAPL Source Zone Investigation Methods
This section presents various site investigation methods and related
interpretation techniques that can be useful when characterizing
a DNAPL source zone. These methods and techniques will be
relied upon in Sections 5 (Assessing DNAPL Presence) and 6
(Delineation of the DNAPL Source Zone). Additional informa-
tion is provided in (6, 26, 37).
Visual Observation
DNAPL obtained from the bottom of a monitoring well
or as an emulsion from a pumped water sample is con-
clusive evidence of DNAPL presence (pooled DNAPL).
Monitoring wells canbe sampled for DNAPL using bot-
tom loading bailers lowered to the bottom of the well or
pumping from the bottom of the well. If an interface probe
indicates DNAPL presence, then the sample should be
retrieved and it shouldbe confirmed (visually, or through
laboratory analysis) that the substance is DNAPL. If
DNAPL is visually observed in drill cuttings or in a soil
sample for the first time, then a sample should be sent to
the laboratory for confirmatory evidence. This line of
evidence is applicable in both unconsolidated deposits
and fractured rock, but it should be noted that visual
observation of DNAPL in rock core is rare because of
the aggressive flushing nature of the drilling process.
Because of the typically sparse and tortuous nature of
DNAPL distribution in the subsurface, DNAPL is not
encountered and visually observed within many DNAPL
source zones.
Chemical Concentrations in Soil Above Threshold
DNAPL Saturation
Chemical concentrations in soil exceeding the value
correspondingtoathresholdDNAPLsaturationare con-
clusive evidence of DNAPL presence (see Calculation 1).
The thresholdDNAPLsaturationforuse in Calculation 1
should be set to be between 5% and 10% of pore space
for all DNAPL types. The particular threshold satura-
-------
tion chosen should result in a chemical concentration
in soil that is an order of magnitude higher than that
determined in line of evidence C. It follows that high
organic carbon content soils and highly hydrophobic
chemicals may require the use of threshold saturations
toward the higher end of the above range. This method
is applicable to unconsolidated media both above and
below the water table, but is not applicable in fractured
rock. The calculationrequiresknowledgeof site-specific
parameters and a quantitative chemical analysis of the
soil. Care should be taken to sample soil horizons in
core exhibiting the highest headspace readings and the
strongest visual indication of DNAPL presence. The use
of fixed depth intervals or compositingfrom several depth
intervals is discouraged when collecting soil samples to
evaluate the presence of DNAPL. Methanol preservation
or a similar technique to reduce VOC losses during han-
dling and transport of soil samples should be employed.
Chemical Concentrations in Soil Above Partitioning
Threshold
Chemical concentrations in soil exceeding the value cor-
respondingto equilibrium partitioning relationships (see
Calculation2) are consistent withDNAPL presence (11).
The composition of the DNAPL need not be known (see
Calculation 4). The calculation is applicable to uncon-
solidated media both above and below the water table,
but is not applicable in fractured rock. The calculation
requires knowledge of site-specific parameters and a
quantitative chemical analysis of the soil. Measured
concentrations that only marginally exceed the calculated
partitioning threshold may be false positives primarily
because of uncertainty associated with estimating the
soil-water partition coefficient.
Site Use/Site History
Investigations during the past 30 years have shown that
the subsurface occurrence of DNAPL is often associated
with the industries, practices, and processes outlined in
Table 1. Site Use/Site History can be ascertained using
methods suchas employee interviews, company purchase
and sale records, aerial photographs, and building plans.
Former lagoons, underground tanks, floor drains and
leach fields are sometimes coincident with the location
of DNAPL source areas.
Vapor Concentrations
The location of a vapor-phase plume may be coincident
with the current or former presence of DNAPL in the
vadose zone. Mapping the vapor-phase plume may
be useful in deciding where to collect additional data.
Because some DNAPLs can completely vaporize in
relatively short time periods (yet the vapors will persist
much longer), the presence of vapors and the mapping
of a vapor-phase plume should generally not be used in
isolationto conclude thatDNAPL is present in the vadose
zone, or to delineate the spatial extent of the DNAPL
source. Care should also be taken to avoid mistaking
vapors derived from off-gassing of a groundwaterplume
with vapors derived from DNAPL sources. In-situ
vapor concentrations can be sampled using invasive
techniques (soil vapor surveys), and can be monitored
during drilling. This line of evidence is not applicable
to DNAPLs lacking a significant vapor pressure (e.g.,
coal tar, creosote, PCBs).
Hydrophobic Dye Testing
Hydrophobic dyes such as Oil Red O will partition into
DNAPL, imparting a red colorto the organic liquid. Dye
techniques are particularly useful when encountering a
colorless DNAPL. Hydrophobic dye techniques include
the jar shake test in which a soil or water sample is placed
into ajar with a small amount of dye (6), and down-hole
samplers that force a dye-impregnated absorbent ribbon
against the borehole wall in either fractured rock or a
direct push borehole (30). It should also be noted that
the absence of staining on a down-hole ribbon sampler
is not evidence of the absence of DNAPL, since only
pooledDNAPLcan migrate towards the sampler (residual
DNAPL may be present in the formation adjacent to the
sampling interval, and remain undetected).
Table 1 - Industries and Industrial Processes Historically Associated With DNAPL Presence (modified after USEPA, 1992).
Industry
Industrial Process
Manufactured gas plant, Wood preservation (creosote),
Electronics manufacturing, Solvent production/recycling,
Pesticide/Herbicide manufacturing, Dry cleaning, Instrument
manufacturing, Metal product manufacturing, Engine
manufacturing, Steel industry coking operations (coal tar),
Chemical production, Airplane maintenance, Transformer oil
production
Storage of solvents in uncontained drum storage areas, Metal
cleaning/degreasing, Metal machining, Tool and die operations,
Paint stripping, Use of vapor and liquid degreasers, Storage
and transfer of solvents in above and below ground tanks and
piping, Burning waste liquids, Storage and treatment of waste
liquids in lagoons, Use of on-site disposal wells, Loading and
unloading of solvents, Transformer reprocessing, Disposal of
solvents in unlined pits.
-------
The following lines of evidence Gl through G6 all make use of
groundwater quality data and can be evaluated every sampling
round.
Magnitude of Groundwater Concentrations
Sampled groundwater concentrations in excess of 1%
effective solubility (see Calculation 3) indicate that the
sampled groundwater may have come in contact with
DNAPL. IfthecompositionoftheDNAPLisnotknown,
Calculation 6 can be used. The distance to the possible
DNAPL locations cannot be determined from the mag-
nitude of the concentration alone. Sampled groundwater
concentrations downgradient of a DNAPL source zone
can be significantly less than the effective solubility
because of hydrodynamic dispersion, wellbore dilution,
non-optimal monitoring well placement, and degrada-
tion processes. In cases where significant degradation
is occurring in the dissolved-phase plume, daughter
product concentrations can be converted to equivalent
parent product concentrations before comparing to the
1% effective solubility threshold (see Calculation 8).
However, it should be noted that daughter product com-
pounds may also be part of a multi-component DNAPL.
Monitoring well points where groundwater concentra-
tions exceed 1% effective solubility can also be useful
in locating additional sampling points potentially nearer
to the possible DNAPL source zones. The interpretation
of groundwater concentrations exceeding 1% effective
solubility is discussed further in (27).
I Persistent Plume
The presence of a contiguous and persistent plume
extending from suspected release locations in the
downgradient direction is evidence of a continuing
source (e.g., DNAPL). If 'sufficient time' has passed
since the last possible introduction of contaminant to the
subsurface and the plume has not 'detached' itself from
the suspected release locations, a DNAPL source may
be present. The 'sufficient time' is dependent on site-
specific conditions such as groundwater velocity and the
amount of sorption occurring (see Calculation 7). This
line of evidence is applicable to both unconsolidated
deposits and fractured rock, but can be inconclusive
in environments subject to significant amounts of back
diffusion (e.g., fractured bedrock with a porous matrix,
fractured clay). Significant amounts of back diffusion
can be the source of a persistent plume even if DNAPL
is not present. This line of evidence is therefore most
applicable to high permeability settings.
I Presence of Contamination in Apparently
Anomalous Locations
The presence of contaminated groundwater in locations
that are not downgradient of known or suspected sources
may be evidence of DNAPL presence hydraulically
upgradient of the monitoring point in question. An
example includes the presence of dissolved-phase con-
tamination in groundwaterthat is older than the potential
(Gil
(G6J
contaminant release (using age dating) or in groundwater
on the other side of a flow divide located between the
monitoring location and suspected release locations. In
Figure 1, for example, the presence of contamination
in the illustrated monitoring well cannot be explained
without the upgradient presence of DNAPL. This line of
evidence is not contingent on any concentration threshold.
Temporal changes in hydraulic heads and groundwater
flow directions, as well as changes in historic pumping
patterns shouldbe considered at sites where groundwater
extraction has, or is, occurring. Consideration should
also be given to the presence of unknown or off-site
sourcesthatmayaccountfortheobservedcontamination.
Groundwater Concentration Trends with Depth
Abrupt reversals of groundwater contaminant concen-
tration levels with depth or increasing concentrations
with depth can be associated with DNAPL presence.
Concentration trends can be best detected using small
interval sampling techniques [e.g., direct push sampling
devices; short well screens; multilevel completions;
cone penetrometer equipped with measurement probes
(16, 26)]. Multilevel monitoring completions can be
incorporated into open holes in bedrock to provide
concentration as a function of depth. Other methods in
bedrock include the use of temporary straddle-packer
assemblies to sample specific depth intervals, and the
use of diffusion bag samplers placed at specific depths.
Use of these latter methodologies should be made only
when intraborehole flow conditions have been adequately
characterized.
Groundwater Concentration Trends with Time
Groundwater downgradient of a multi-component
DNAPL may exhibit a temporal decline in the concentra-
tion of the higher effective solubility compounds and a
stable or increasing trend in time of the lower effective
solubility compounds. Highly soluble and mobile com-
pounds, such as low molecular weight alcohols, furans,
ketones and some solvents such as methylene chloride
may show a decreasing concentration versus time sig-
nature downgradient of a DNAPL source zone while
at the same time higher molecular weight alcohols and
semi-volatile compounds may show a stable concentra-
tion trend. This line of evidence is primarily applicable
to mixed DNAPLs. Consideration should be given to
compound specific biodegradation, which may result in
the concentration of certain compounds decreasing and
others (such as low molecular weight daughterproducts)
increasing within the plume. Dissolved-phase concen-
trations downgradient of a single component DNAPL
may decline due to removal of some of the source mass
during dissolution; a declining concentration versus time
signature does not preclude the presence of DNAPL.
Detection of Highly Sorbing Compounds in
Groundwater
The detection of highly sorbing and low solubility com-
pounds which have low mobility in groundwater may be
-------
associated with a nearby DNAPL source. This line of
evidence can be useful in delineating the extent of the
DNAPL in the downgradient direction. Examples of
compounds that have very low mobility in groundwater
(absent transport facilitated by colloids, cosolvents, or
emulsions) include PCBs and high molecular weight
PAHs.
Other Types of Methods
Partitioning interwell tracer tests (PITTs) [1, 4, 15]
involve the injection and withdrawal of a tracer that
has the ability to partition into the DNAPL. While the
method can be used to detect the presence of DNAPL.
given the significant effort involved in conducting tracer
tests, PITTs are typically employed after some level
of source zone characterization has been completed.
Literature sources suggest (for certain sites with appro-
priate geologic conditions and contaminant properties)
measuring a depletion of Radon-222 in groundwater (3 4).
Direct push platforms can be used to deploy a variety of
probes to vertically profile contaminant concentrations.
These probes include laser induced fluorescence (LIF)
measurement devices (6, 31, 32) such as ROST (rapid
optical screening tool) and TarGOST (tar-specific green
optical screeningtool), whichis specifically designedfor
detecting the presence of coal tar and creosote (32); and
probes employ ing Raman methods (31). LIF techniques
respond well to the presence of NAPLs containing aro-
matic hydrocarbons, but may not be suitable for many
chlorinated solvent DNAPLs. Direct push platforms
can also be used to deploy a membrane interface probe
(MIP) or a hydrosparge probe (8), both of which transfer
contaminants to a flowing gas stream for analysis at the
surface. Another measurement probe is the precision
injection/extraction (FIX) device (23). The use of mea-
surement probes with direct push platforms is becoming
increasingly popular, but care should be taken in inter-
preting results with respect to DNAPL presence given
that most of these devices provide a relative measure of
total concentration. Consideration of the potential for.
and consequences of, false positives should be given to
each of these methods.
5.0 -Assessing DNAPL Presence
Determining the presence or absence of DNAPL is an important
component of the site characterization process and subsequent
development of a conceptual site model. The length of time and
degree of effort required to determine the presence or absence of
DNAPL will vary from site to site. Once it has been determined
that DNAPL resides in the subsurface, the objectives for further
investigation and potential remediation strategies can be estab-
lished. This section focuses on methods to assess the presence
of DNAPL; Section 6 of this document focuses on methods to
delineate the DNAPL source zone.
Converging lines of evidence can be used to determine whether
or not DNAPL is present in the subsurface. Figure 3 presents a
graphical summary of the converging lines of evidence approach.
Example calculation procedures are contained in Appendix A. All
lines of evidence are discussed in Section 4, and are applicable
to both unconsolidated deposits and fractured rock, unless noted
otherwise. As indicated in Figure 3, either line of evidence A or
B will lead to the conclusion that DNAPL is present. If A and B
are both found to be negative, then the determination of whether
DNAPL is present must be made on the basis of a weight of
evidence approach, with multiple converging lines of evidence
Inconclusive
Rely on Other
Lines of Evidence
NO
No DNAPL
Source Zone
See text for explanation on
the use of multiple lines of
evidence required for YES
or NO determination.
H
Visual
Observation
DNAPL
Saturation
Soil
Partitioning
Site Use
Site History
Vapors
Dye
Testing
Groundwater
(anyofGl-G6)
Others
YES
YES
DNAPL
Source Zone
Figure 3 - Converging lines of evidence approach to assessing DNAPL presence. Methods B and C are not applicable to fractured
rock.
-------
combining to form either a positive or negative determination.
Note that it is not likely that all of C through H will be satisfied
at any one particular site, and that neither A nor B are neces-
sary requirements to conclude that DNAPL is present. Most
confirmed DNAPL source zones will have some of A through H
determined to be negative. Because conditions vary from site to
site, this document does not prescribe a specific number of lines
of evidence that must be satisfied to arrive at cither a positive or
negative determination.
If the various lines of evidence contradict each other, it may be
necessary to collect more data. It is possible that a minority of
positive determinations can outweigh a majority of negative de-
terminations if the positive lines of evidence cannot be explained
without the presence of DNAPL. It should also be noted that
not all sites lend themselves to collecting all of the types of data
outlined here. In fractured rock, for example, soil vapor data and
partitioning calculations would not be relied upon.
Evaluating the presence of DNAPL is an iterative process that
incorporates new data as they are obtained. It is recognized here
that certain types of data are more likely to be collected in the
early stages of site investigation, while others (e.g., groundwater
concentrations) can be collected on a routine basis throughout the
investigation process. The fact that a number of lines of evidence
are outlined in Figure 3 does not suggest that they should all be
pursued at any one particular site. Site specific conditions will
dictate what lines of evidence should be pursued. Care should be
taken, however, to ensure that a negative response to the various
lines of evidence is not simply attributable to inadequate charac-
terization and an insufficient amount of data.
6.0 - Delineation of the DNAPL Source Zone
Depending on the spatial density of sampling points installed
during initial investigation efforts, the general area within which
the DNAPL resides may have been identified. Once it has been
determined that DNAPL is present in the subsurface, the objec-
tives for delineation of the source zone can be established. These
objectives can vary from site to site, but typically involve one or
more of the following:
* Delineation of the DNAPL source zone to ensure that the
flow paths and quality of the groundwater downgradient of
the source zone are monitored for the presence of dissolved-
phase contaminants to assess protection of current and
potential receptors.
* Delineation of the DNAPL source zone to facilitate proper
design of containment systems involving groundwater ex-
traction and/or physical barriers.
* Delineation of the DNAPL source zone to facilitate imple-
mentation of DNAPL mass removal technologies.
* Delineation of the DNAPL source zone as part of establish-
ing boundaries for institutional controls.
« Delineation of the DNAPL source zone as part of Technical
Impracticability assessments (41).
Given the selective nature of DNAPL migration, it is not feasible
to determine the exact location and extent of individual DNAPL
migration pathways within the overall confines of the source zone
in either unconsolidated deposits, or fractured bedrock. Because
data collection efforts typically involve a finite number of local-
scale measurements taken at discrete locations (e.g., water quality
samples, soil samples, etc.), some uncertainty will exist regarding
the delineated spatial extent of the source zone.
To address the issue of uncertainty, it is recommended that both
a 'Confirmed/Probable' DNAPL source zone be delineated, as
well as a 'Potential' DNAPL source zone (see Figure 4). The
Confirmed/Probable source zone is the volume within which
compelling and multiple lines of evidence indicate thatDNAPL is
present. Note that what may be a compelling line of evidence at
one site may not be so at another site (e.g., G2 Persistent Plume,
is a stronger line of evidence in a high permeability setting than at
a site where back-diffusion may dominate). The Potential source
zone is of larger spatial extent, and is defined as that volume of
the subsurface within which some lines of evidence indicate
that DNAPL may be present, but the lines of evidence are not
as numerous, consistent, or compelling as within the Confirmed/
Probable source zone. Defining a Potential source zone outside
of the Confirmed/Probable source zone addresses the uncertainty
associated with finite amounts of data. This can be particularly
useful in the hydraulicalfy downgradient direction where it is
often difficult to determine the distance to the edge of the DNAPL
source zone based on groundwater quality data (e.g., using lines
of evidence Gl through G6).
With respect to the various criteria for assessing DNAPL presence
outlined inSection4. lines of evidence AandB will bothfall within
the Confirmed/Probable source zone. All other lines of evidence
(C through H) could fall within either the Confirmed/Probable
source zone, or the Potential source zone. Note also that positive
determinations for lines of evidence A and B are not necessary to
define a Confirmed/Probable source zone. The defining feature
of the Confirmed/Probable source zone is that multiple lines of
evidence indicate that DNAPL is present. In practice, this will
manifest itself as various lines of evidence all plotting within the
same general spatial area on plan view and cross-section figures
(see Figure 4 for plan view example). Within the Potential source
zone, there will be fewer lines of evidence, and their occurrence
may not be as contiguous as within the Confirmed/Probable source
zone. Consideration should be given to known DNAPL release
locations and structural aspects of the geology (e.g., dipping beds,
dipping fractures) when delineating both the Confirmed/Probable
and Potential source zones.
There is no prescriptive number of lines of evidence that separate
the two source zone delineations. The individual lines of evidence
cannot be weighted cither, as the strength of the uncertainty/cer-
tainty determination is dependent on how often more than one
line of evidence occurs at a particular location and how many
contiguous locations have multiple lines of evidence; assigning a
weighting factor to each line would negate this obj ectivity. Further-
more, many factors influence the transport of the DNAPL and the
associated concentration of the dissolved-phase constituents such
that a weighting factor could not be fairly assigned for all types of
hydrogeologic environments and types of DNAPL contaminants.
The amount of acceptable uncertainty in delineating the source
zone boundaries is likely to be dependent on the remedial actions
considered. If hydraulic or physical containment of the DNAPL
source zone were a component of the remedial actions, for example,
an accurate delineation of the Potential source zone would be war-
-------
Groundwater
Flow
Dissolved-Phase Plume
Potential DNAPL
Source Zone
Confirmed/Probable
DNAPL Source Zone
LEGEND
• DNAPL Entry Location
O Visual Observation of DNAPL (A)
x Soil Samples Above CD (B)
A Soil Samples Above Cr(C)
® Non-Detect Groundwater Concentration
• Groundwater Concentration above 1% (G1)
A Groundwater Concentration in Apparently Anomalous Location (G3)
0 Other (H)
Figure 4 - Example of plan view schematic illustrating confirmed/probable and potential DNAPL source zones. Note that not all lines
of evidence are depicted. Types and distribution of lines of evidence will vary from site to site.
ranted (the likely target for hydraulic containment) and accurate
delineation of the Confirmed/Probable source zone may not be
necessary. If the remedial actions included implementation of
a DNAPL mass removal technology, however, then an accurate
delineation of the Confirmed/Probable DNAPL source zone (the
likely target for mass removal) would be warranted. A similar
approach may be appropriate for designating a zone of technical
impracticability (TI). Overestimating the size of the Confirmed/
Probable source zone could overstate costs for technology appli-
cation and may result in a particular technology being screened
out. Underestimating the size of the Confirmed/Probable source
zone, on the other hand, could lead to underestimation of costs
and the perception of poor performance following completion of
technology application. Monitoring points outside of an under-
estimated source zone may provide data showing little, if any,
benefit resulting from source zone removal or treatment.
Typically, to refine the locations of the boundaries, additional
drilling and sampling may be required between the Confirmed/
Probable and Potential DNAPL areas. Figure 5 depicts an itera-
tive process of data collection. Usually the degree of uncertainty
in delineating these two zones will be greater in a more complex
hydrogeologic environment. Although additional sampling points
may be easily installed in shallow, unconsolidated materials, the
same level of effort may not be feasible or may be cost prohibitive
in deep fractured rock. Care must also be taken to ensure that
drilling and sampling activities do not mobilize DNAPL deeper in
to the subsurface. Strategies in place of extensive drilling to depth
within the source zone include drilling adjacent to the suspected
source zone and using lines of evidence such as Gl through G6
to infer DNAPL presence in the upgradient direction.
In all environments, the risks of potentially mobilizing the DNAPL
and the associated incremental costs of additional sampling points
should be compared to the benefits of increased ability to evaluate
the spatial extent of the DNAPL. Additionally, site investigators
should have a DNAPL Contingency Plan on hand in the field to
address actions to be takenif pooledDNAPL is encountered during
drilling. At some sites, it may be desirable to adopt an 'outside
in' approach to reduce the number of invasive borings that need
to be placed within the DNAPL source zone.
In addition to delineating the spatial extent of the source zone,
investigators may need to assess whether or not DNAPL is still
migrating within the subsurface. The assessment of mobility can
be carried out using screening calculations (27) and observations
such as an expanding area of lines of evidence indicating DNAPL
presence. Otherfeatures of the source zone that may be of interest
include the mass of DNAPL present, the mass flux downgradient
of the source zone, and the relative proportions of residual versus
pooledDNAPL. Calculation 1 canbeusedto distinguishbetween
residual and pooledDNAPL in soil samples by selecting a saturated
threshold above whichDNAPLis considered pooled. Also of note
is the fact that residual DNAPL will not enter monitoring wells,
implying that the accumulation of DNAPL in a well indicates the
presence of pooled DNAPL in the formation. Details regarding
how to estimate the mass of DNAPL present in a source zone or
the distribution of mass flux downgradient of the source zone,
however, are beyond the scope of this document.
-------
Data Collection
and Interpretation
(see Section 4)
NO
Has enough data been
collected to assess
DNAPL presence?
(see Figure 3)
YES
Does a DNAPL
source zone exist?
YES
Is there a need to
delineate the
DNAPL source zone?
YES
Map the
Confirmed/Probable
and Potential DNAPL
source zones
Collect more data
(see Section 4)
YES
Is more accurate
delineation of either
source zone required?
NO
No further data
collection required
in the context of
DNAPL source assessment
and delineation.
NO
NO
Figure 5 - Flowchart depicting iterative data collection process used in refining the DNAPL source zone boundaries.
10
-------
7.0 - Glossary
Bulk Retention Capacity is defined as the total volume of DN APL
that lias been retained as residual and pooled DNAPL in a unit
volume of the subsurface. The bulk retention capacity accounts
for the fact that not all lenses, laminations and geological units
within a source zone contain DNAPL (27), and it is a function of
the release history, geology and DNAPL properties. In uncon-
solidated media, the bulk retention capacity can be in the range
from 0.005 to 0.03 (36). In fractured media, the bulk retention
capacity can be in the range of 0.0002 to 0.002 (36). Fractured
rock and clay cannot retain as much DNAPL per unit volume as
unconsolidated deposits.
Capillary Barriers are fine grained lenses, layers and laminations
upon which lateral spreading and pooling of DNAPL can occur.
Even if the capillary barrier is penetrated by the DNAPL, it is
likely that lateral spreading will have occurred along the top surface
of the barrier prior to the capillary pressure having exceeded the
entry pressure of the barrier. The finer grained the capillary bar-
rier, the higher the pool height of DNAPL that it can support (17).
Capillary Pressure is the pressure difference between two im-
miscible liquids and arises because of interfacial tension. It is
calculated as the non-wetting phase pressure minus the wetting
phase pressure. If the DNAPL is the non-wetting phase and water
is the wetting phase, for example, the capillary pressure would
be the DNAPL pressure minus the water pressure.
DNAPL (Dense, Non-Aqueous Phase Liquid) is an organic liquid
that is more dense than water and does not mix freely with water.
A single-component DNAPL is composed of only one chemi-
cal. A multi-component DNAPL is composed of two or more
chemical components.
DNAPL Source Zone The DNAPL source zone is the overall
volume of the subsurface containing residual and/or pooled
DNAPL. Not all portions (e.g., lenses, laminations, or fractures)
of the source zone will contain residual and/or pooled DNAPL.
The Confirmed/Probable DNAPL Source Zone is the part of
the source zone within which it is known or highly likely that
DNAPL exists. The Potential DNAPL Source Zone is the part
of the source zone within which it is possible that DNAPL exists,
but the lines of evidence indicating DNAPL presence are either
fewer or arc not as strong as those associated with the Confirmed/
Probable DNAPL Source Zone.
Dissolved-phase Plume The zone of contamination containing
dissolved-phase constituents resulting from groundwater flowing
past residual and pooled DNAPL. The contaminants present in
the plume arc subject to advection, dispersion, and possibly sorp-
tion, decay, and matrix diffusion. Dissolved-phase plumes can
be sustained by back diffusion from low permeability regions in
the absence of DNAPL.
Effective Solubility For a multi-component DNAPL, the equi-
librium solubility in water of any component of the DNAPL is
referred to as the component's effective solubility. In general, the
various components of a DNAPL suppress each other's aqueous
solubility implying that effective solubilities are typically less
than single-component (handbook) solubilities. For structurally
similar compounds, the effective solubility can be estimated us-
ing Raoult's Law (2).
Interfacial Tension (1FT) is a tensile force that exists in the
interface separating DNAPL and water. Because of interfacial
tension, DNAPLs do not mix freely with water and exist in the
subsurface as a separate liquid phase. IFT is a site-specific value
that can be assessed with a simple laboratory test if a sample of
DNAPL can be obtained. Literature values tend to overestimate
the IFT encountered at sites. In general, higher IFT leads to more
lateral spreading of DNAPL in horizontally bedded deposits,
stronger capillary trapping forces, and a greater tendency for
DNAPL pooling.
Mole Fraction refers to the proportion of a component, on the
basis of moles, in a multi-component DNAPL. The sum of all the
mole fractions is unity. Mass fractions, as provided by laboratory
analysis, can be converted to mole fractions using the molecular
weight of each component (see calculation 5).
1% Rule of Thumb is a generality that sampled groundwater
concentrations in excess of 1% effective solubility (see Calcula-
tion 3) indicate that DNAPL may be present in the vicinity of
(any direction) the monitoring point of interest. The distance
between the monitoring point in question and the DNAPL source
zone varies from site to site and is generally difficult to quantify
with a high degree of accuracy.
Pooled DNAPL refers to local, continuous distributions of DNAPL
that accumulate above capillary barriers. The capillary barriers
are typically lower permeability horizons, and they can occur at
any elevation in the subsurface. Within the pool, the DNAPL
saturation is typically between 30% and 80% of pore space inboth
porous media and fractures (27). Because pools are contiguous
through the pore structure they are potentially mobile and can
migrate into monitoring wells, and can be mobilized by increases
in the hydraulic gradient or lowering of IFT.
Raoult's Law is given by C. = mS. where C. is the effective
solubility (rng/1) of component i, m. is the mole fraction (unitless)
of component /' in the DNAPL, and S. is the single-component
(handbook) solubility of component/ (2). This expression assumes
ideal partitioning behavior and is used to estimate the maximum
concentrations in groundwater immediately adjacent to residual
and pooled DNAPL.
Residual DNAPL refers to disconnected blobs and ganglia of the
DNAPL, trapped by capillary forces in the pore space of both
porous media and fractures (21, 27, 44). The blobs and ganglia
are typically from 1 to 10 grain diameters in size in unconsolidated
deposits (44), and are left behind in the pathways that DNAPL
has migrated through.
Residual Saturation refers to the volume of residual DNAPL
present in a unit volume of pore space. Residual DNAPL satura-
tions typically vary between 5% and 30% of pore space in both
porous media and fractures (21, 27. 44).
Source Zone Architecture refers to (i) the overall shape and
dimensions of the source zone, (ii) the ratio of residual to pooled
DNAPL (also referred to as the ganglia to pool ratio), (iii) the
lateral continuity of zones of residual DNAPL and DNAPL pools,
(iv) the thickness of zones of residual DNAPL and DNAPL pools,
and (v) the portion of lenses and layers containing DNAPL versus
those void of DNAPL. The source zone architecture influences
the downgradient dissolved-phase plume concentrations and mass
flux distribution.
11
-------
Wettability refers to the affinity of the DNAPL for a solid surface
in the presence of water (6,27). Many DNAPLs are non-wetting,
implying that they will preferentially occupy the pore spaces within
coarser grained lenses and laminations, and larger aperture frac-
tures. Some DNAPLs are wetting with respect to water, however,
implying that they will preferentially coat the aquifer materials
and thereby occupy the pore spaces of the finer grained media.
Coarser grained horizons and larger aperture fractures represent
capillary barriers to DNAPLs lhalare welting with respect to water.
Acknowledgements
TheU.S. EPA Office of Research and Development (ORD) wishes
to express their appreciation to the U.S. EPA Ground Water Fo-
rum. The Ground Water Forum was helpful in the development
and review of this document along with ORD scientist Dr. David
Burden.
Notice
The U.S. Environmental Protection Agency through its Office
of Research and Development and the Office of Superfund Re-
mediation and Technology Innovation funded and collaborated
on the document under Contract No. 68-C-02-092 to Dynamac
Corporation. Mentionof trade names or commercial products does
not constitute endorsement or recommendation for use.
References
1) Amiable, M.D., Rao, P.S.C., Graham, W.D., Hatfield, K.
and Wood, A.L., 1998. Use of partitioning tracers for mea-
suring residual NAPL: Results from a field-scale test. J.
Environmental Engineering, 124(6), pp. 498-503.
2) Bancrjcc, S., 1984. Solubility of organic mixtures in Water.
Environmental Science & Technology, 18, 587-591.
3) Brewster, M.L., Annaa A.P., Greenhouse. J.P., Kueper,
B.H., Olhoeft, G.R., Redman, J.D., and Sander, K.A., 1995.
Observed migration of a controlled DNAPL release by geo-
physical methods. J. Ground miter, 33(6), 977-987.
4) Brooks, M.C., Amiable, M.D., Rao, P.S.C.. Hatfield, K.,
Jawit.z, J.W, Wise, W.R., Wood, A.L. and Enfield, C.G.,
2002. Controlledrelease,blindtestsofDNAPLcharacteriza-
tion using partitioning tracers. J. Contaminant Hydrology,
59, pp. 187-210.
5) Chapman, S.W. and Parker, B.L.. 2005. Plume persistence
due to aquitard back diffusion following dense nonaqueous
phase liquid source removal or isolation. Water Resources
Research, 4LW12411.
6) Cohen.R.M. andMercer, J. W, 1993. DNAPL Site Evaluation.
C.K. Smoley, CRC Press.
7) Crumbling. D.M., Lynch, K., Howe, R... Groenjes. C.,
Shockley, J., Keith, L., Lesnik, B., Van E, J. and McKenna,
J., 2001. Managing uncertainty in environmental deci-
sions. Environmental Science & Technology, 35(19),
pp. 404A-409A.
8) Davis, W.M., Wise, M.B., Furey, J.S. and Thompson, C. V.,
1998. Rapid detection of volatile organic compounds in
groundwater by in situ purge and direct-sampling ion-
trap mass spectrometry. Field Analytical Chemistry &
Technology, 2(2), pp. 89-96.
9) Electric Power Research Institute, 2008. MGP Site
Characterization Best Practices Manual for Bedrock. EPRI,
Palo Alto, CA; PSEG, Newark, NJ; NYSEG, Binghamton,
NY; and FirstEnergy, Madison, NJ; 1018276.
10) Falta, R.W., P.S. Rao, and N. Basu, 2005. Assessing the
impacts of partial mass depletion in DNAPL source zones.
I. Analytical modeling of source strength functions and
plume response. ,/. Contaminant Hydrology, 78, 259-280.
11) Feenstra, S.,Mackay,D.M.,and Cherry, J.A., 1991.Presence
of residua 1 NAPL based on organic chemical concentrations
in soil samples. Ground Water Monitoring & Remediation,
11(2), 128-136.
12) Gerhard, J.I.. Pang. T.andKueper,B.H., 2007. Time scales of
DNAPL inigralionin sandy aquifers examined via numerical
simulation. ./. Ground Water, Vol. 45, No. 2, pp. 147-157.
13) Huling, S.G. and J.W. Weaver, 1991. Dense Nonaqueous
Phase Liquids, USEPA Ground Water Issue Paper,
EPA/540/4-91/002.
14) Interstate Technology andRegulatory Council (ITRC), 2003.
An Introduction to Characterizing Sites Contaminated with
DNAPLs. Washington, DC: ITRC Dense Nonaqueous Phase
Liquids Team.
15) Jin, M., Delshad, M., Dwarakanath, V, McKinney, D.C.,
Pope, G.A., Sepehrnoori, K., Tilburg, C.E., and Jackson,
R.E.. 1995. Partitioning tracer tests for detection, estima-
tion and remediation performance assessment of subsurface
non-aqueous phase liquids. Water Resources Research,
31(5), 1201-1211.
16) Kram, M.L., Keller, A.A., Rossabi, J., and Everett L.G.,
2001. DNAPL characterization methods and approaches,
Parti, Performance comparisons. Ground Water Monitor ing
& Remediation, Fall, 109-123.
17) Kueper, B.H., Redman, J.D., Starr, R.C., Reitsma, S. and
Mah, M.. 1993. A field experiment to study the behavior
of tetrachloroethylene below the watertable: Spatial distri-
bution of residual and pooled DNAPL. J. Ground Water,
31(5), 756-766.
18) Kueper, B.H. and McWhorter, D.B., 1991. The behavior of
dense, non-aqueous phase liquids in fractured clay and rock.
J. Ground Water, 29(5), 716-728.
19) Lenczewski, M., McKay, A., Pitner, A., Driese, S., and V
Vulava, 2006. Pure-Phase transport and dissolution of TCE in
sedimentary rock saprolite.,/. Ground Water, 44(3), 406-414.
12
-------
20) Lipson, D., Kueper, B.H. and Gefeii, M.J., 2005. Matrix
diffusion-derived plume attenuation in fractured bedrock.
J. Ground Water, 43(1), 30-39.
21) Longino, B.L. and Kueper, B.H., 1999. Non-wetting phase
retention and mobilization in rock fractures. Water Resources
Research, 35(7), 2085-2093.
22) Longino, B.L. and Kueper, B.H., 1995. The use of upward
gradients to arrest downward DN APL migration in the pres-
ence of solubilizing surfactants. Canadian Geotechnical
Journal, 32(2), 296-308.
23) Looney, B.B., Jerome, K.M., and Davey, C., 1998. Singe
well DNAPL characterization using alcohol injection/extrac-
tion. Proc. First Int. Conf. on Remediation of Chlorinated
and Recalcitrant Compounds. Battelle Press, Columbus,
OH, 113-118.
24) Mackay, D., SMu, W.Y. and Ma, K.C., 1992. Illustrated
handbookofphysical-chemical properties andenvironmental
fate for organic chemicals. Volumes 1 - 3. Lewis Publishers.
25) Mariner, P.E., Jin, M., and Jackson, R.E., 1997. An algorithm
for the estimation of NAPL saturation and compositionfrom
typical soil chemical analyses. Ground Water Monitoring
& Remediation, 17(2), 122-129.
26) Nielsen, D. (editor), 2006. Practical Handbook of
Environmental Site Characterization. CRC Press, Boca
Raton, FL.
27) Pankow, J.F. and Cherry, J.A., Editors, 1996. Dense
Chlorinated Solvents and other DNAPLs in Ground Water.
Waterloo Press, Portland, OR.
28) Parker,B.L.,Gillham,R.W.,andCherry,J.A., 1994.Diffusive
disappearance of dense immiscible phase organic liquids in
fractured geologic media. ./. Ground Water, 32, 805-820.
29) Poulsen, M. and Kueper, B.H., 1992. A field experiment
to study the behavior of tetrachloroethylene in unsaturated
porous media. Environmental Science & Technology. 26(5),
889-895.
30) Riha, B.D., Rossabi, J., Eddy-Dilek, C.A., Jackson, D., and
Keller, C., 2000. DNAPL characterization using the ribbon
NAPL sampler: Methods and results. Proc. Second Int. Conf.
on Remediation of Chlorinated and Recalcitrant Compounds.
Battelle Press, Columbus, OH, 33-40.
31) Rossabi, J., Riha, B.D., Eddy-Dilek, C.A., Lustig, A.,
Carrabba, M., Hyde, W.K., andBello, J., 2000. Field tests of
a DNAPL characterization system using cone penetrometer
- base Raman spectroscopy. Ground Water Monitoring &
Remediation, 20(4), 72-81.
32) St. Germain, R., Adamek, S. and Rudolph, T., 2006. In
situ characterization of NAPL with TarGOST at MGP sites.
Land Contamination and Reclamation, 14(2), pp. 573-578.
33) Sale, T., Newell, C., Stroo, H., Hinchee, R. and Johnson, P.,
2008. Frequently asked questions regarding management of
chlorinated solvents in soils and groundwater. Environmental
Security Technology Certification Program (ESTCP), U.S.
Department of Defense, Washington, DC.
34) Semprini, L.M., Cantaloub, M., Gottipati, S., Hopkins, O.,
and Istok, J., 1998. Radon-222 as a tracer for quantifying
and monitoring NAPL remediation. Proc. Firs tint. Conf. on
Remediation of Chlorinated and Recalcitrant Compounds.
Battelle Press,' Columbus, OH. 137-142.
35) Sudicky, E.A. andFrind, E.O., 1982. Contaminant transport
infractured porous media: analytical solutions for a system of
parallel fractures. Water Resources Research, 18,1634-1642.
36) United Kingdom Environment Agency, 2003. An Illustrated
Handbook of DNAPL Transport and Fate in the Subsurface.
R&D Publication 133, ISBN 1844320669.
37) USEPA, 2004. Site characterization technologies for DNAPL
investigations. EPA/542/R-04/017.
38) USEPA, 2003. The DNAPL Remediation Challenge: Is There
a Case for Source Depletion, EPA/600/R-03/143.
39) USEPA, 1994. DNAPL Site Characterization, OSWER
Publication 9355.4-16FS, EPA/540/F-94/049.
40) USEPA, 1993. Evaluation of the likelihood of DNAPL pres-
ence at NPL sites, national results. OSWER Publication
9355.4-13, EPA/540-R-93-073.
41) USEPA, 1993. Guidance for Evaluating Technical
Impracticability of Ground-Water Restoration. OSWER
Directive 9234.2-25. EPA/540-R-93-080.
42) USEPA, 1992. Estimating Potential for Occurrence
of DNAPL at Super/and Sites, OSWER Publication
9355.4-07FS.
43) USEPA, 1992. Dense Nonaqueous Phase Liquids - A
Workshop Summary, Dallas, TX. April 17-18, 1991,
EPA/600/R-92/030.
44) Wilson, J.L., Conrad, S.H., Mason, W.R., Peplinski, W. and
Hagcn, E., 1990. Laboratory Investigation of Residual Liquid
0rga;«c.y.'uSEPA/600/6-90/004, R.S. KerrLab., Ada, OK.
13
-------
Appendix A - Example Calculations
Note that the following calculations are generally subject to uncertainty because of input parameter variability. This variability may
stem from spatial or temporal variation in site-specific conditions, or variation in textbook parameters such as contaminant chemical
properties. The investigator is advised to make conservative choices with respect to input parameters and consider using a range of
either measured or estimated values when performing calculations.
Calculation 1 - Chemical Concentration in Soil Corresponding to Threshold DNAPL Saturation
CD = soil concentration (mg/kg) corresponding to threshold
Pb
DNAPL saturation [calculated],
Sr = threshold DNAPL saturation [set between 0.05 and 0.10],
4> = effective porosity (unitless) [site specific measurement],
pN = DNAPL density (g/cc) [site specific measurement],
pb = dry soil bulk density (g/cc) [site specific measurement],
CT = amount of contaminant (mg/kg) present in the soil sample
in the aqueous, vapor, and sorbed phases [see Calculation 2
to evaluate CT].
Example Calculation
PCE DNAPL (pN = 1.62 g/cc) in a soil sample with Sr = 0.05, § = 0.25 and pb = 2.0 g/cc corresponds to (ignoring the CT fraction)
CD = 10,125 mg/kg. Note that the quantity CT is typically negligible compared to the DNAPL saturation term. The above equation
is applicable to single-component DNAPLs in unconsolidated porous media. See reference (25) for the relationship between CD and
DNAPL saturation for a multi-component DNAPL. It should be noted that 0.05
-------
Example Calculation
Consider a single-component DNAPL composed of TCE (C. = 1100 mg/1, Koc =126 ml/g, H' = 0.31) in a soil sample having 0w = 0.15,
0a = 0.10, pb = 2.0 g/cc, and ^c = 0.003. The corresponding value of CT is 515 mg/kg. For a multi-component DNAPL, a separate
value of C.r would be calculated using the above equation for each component detected in the soil sample.
Calculation 3 - Effective Solubility Calculated Using Raoult's Law (see Ref. 2)
C. = effective solubility (mg/1) of component /' [calculated],
m. = mole fraction (unitless) of component /' in the DNAPL
[site specific measurement],
S. = single-component solubility (mg/1) of component /'
[handbook].
Example Calculation
Consider a 3-component DNAPL composed (by mass) of 25% TCE (S. = 1100 mg/1), 35% PCE (S. = 200 mg/1), and 40% toluene
(S. = 500 mg/1); the corresponding mole fractions (see Calculation 5) are 0.23, 0.25, and 0.52 respectively, and the corresponding
effective solubilities are 250 mg/1, 50 mg/1, and 260 mg/1 respectively. Sampled groundwater concentrations in excess of 1% of any
of these effective solubilities are evidence of possible DNAPL presence in the vicinity of the monitoring point. The distance to the
DNAPL cannot be determined on the basis of the magnitude of the groundwater concentration alone. In cases where some of the
components of the DNAPL are not known, the unknown mass fraction can be assigned an estimated molecular weight, or the aver-
age of the molecular weights of the known components.
Calculation 4 - Threshold Chemical Concentration in Soil Based on Partitioning Relationships Where
Composition of DNAPL is Not Known
. -
-
CTobs, = reported concentration (mg/kg) of component /' [site specific
measurement],
CTS . = single component soil partitioning concentration (mg/kg) of
component /' (see C[in Calculation 2),
n = number of components observed in the soil sample [site
specific measurement].
For a multi-component DNAPL of unknown composition, the sum of the mole fractions must equal unity. DNAPL will therefore
CT
be present in a soil sample if sum of °^.' exceeds unity.
Cs,t
Note that CTS . is calculated for each component in the summation using Calculation 2 with the single-component solubility as input.
The presented technique can be prone to false negatives in cases where the soil sample was not analyzed for some of the components
of the DNAPL. Because of this, it may be prudent in some cases to only use the calculation for demonstrating that DNAPL was
present in a soil sample and not rely upon it to demonstrate that DNAPL was absent from a soil sample.
Example Calculation
The table below provides an example calculation for a soil sample in which 5 components have been detected. The sample is char-
acterized by a porosity of 25%, a fraction organic carbon of 0.003, and a dry bulk density of 1.99 g/cc. The last column of the table
sums to greater than 1.0, indicating that DNAPL was present in the soil sample.
15
-------
Compound
Trichloroethylene
Tetrachloroethylene
Carbon Tetrachloride
Chlorobenzene
1,1 ,1-Trichloroethane
C7
ofa, i
(mg/kg)
145
155
200
177
213
KOC
(I/kg)
126
364
439
330
152
Handbook Solubility
(mg/l)
1100
200
790
500
1320
r7"
^S,i
(mg/kg)
554
244
1140
558
768
SUM =
CT
^obs.i
CT
*-S,i
0.262
0.636
0.175
0.317
0.277
1.668
Calculation 5 - Mole Fraction (n-component DNAPL)
m = •
ms
mw,
r ,
ms,.
mwi mwj+l mwn
m. = mole fraction of component /' (unitless) in the DNAPL
[calculated],
ms. = mass fraction of component /' (unitless) in the DNAPL
[measured],
mw. = molecular weight (g/mol) of component /' [handbook].
Example Calculation
Consider a 3-component DNAPL composed by mass of 25% TCE (mw = 131.5 g/mol), 35% PCE (mw = 165.8 g/mol), and 40%
toluene (mw = 92.1 g/mol). The corresponding mole fractions are 0.23, 0.25, and 0.52 respectively. In cases where some of the
components of the DNAPL are not known, the unknown mass fraction can be assigned an estimated molecular weight, or the aver-
age of the molecular weights of the known components.
Calculation 6-1% Effective Solubility Threshold Not Knowing DNAPL Composition
C°bs = sampled groundwater concentration (mg/l) of component /'
[site specific measurement],
S. = single-component solubility (mg/l) of component /'
[handbook],
= cumulative mole fraction of the sample [set],
• = a
a
= number of components in groundwater sample.
Calculation assumes that the degree of borehole dilution, dispersion, and degradation is identical for each component of interest
in an obtained groundwater sample. If the 1% rule-of-thumb is used, DNAPL may be present in the vicinity of a monitoring well
if a > 0.01. The procedure can be applied on a sample-by-sample basis without having to make the assumption that the DNAPL
composition is spatially uniform in the subsurface. If it is believed that a value other than 1% effective solubility indicates DNAPL
presence, a can be set to the corresponding value. The presented technique can be prone to false negatives where the groundwater
sample was not analyzed for some of the components of the DNAPL. Because of this, it may be prudent in some cases to only use
the calculation for demonstrating that a has been exceeded in a sample and not rely upon it to demonstrate that a was not exceeded
in a sample.
16
-------
Example Calculation
The table below presents an example calculation for 5 components. Although each component has been detected at a concentration
less than 1% of S., the cumulative mole fractions sum to 3.4%, providing evidence of possible DNAPL presence in the vicinity of
the monitoring location. If the groundwater sample is not analyzed for all components present in the DNAPL, or if any compounds
are degrading in the aqueous phase, the calculation procedure will underestimate the likelihood of DNAPL presence.
Compound
Trichloroethene
Tetrachloroethene
Toluene
Chlorobenzene
Trichloromethane
/~iobs
y L,
* s.
C°bs
(mg/l)
4.4
1.8
3.5
4.0
48.0
S.
(mg/l)
1100
200
500
500
8000
ct
st
0.004
0.009
0.007
0.008
0.006
0.034
Calculation 7 - Plume Detachment Time
t = time (yrs) required for contaminants to migrate through
source zone of length! in the direction of groundwater flow,
v = average linear groundwater velocity (m/yr) [site specific],
R = retardation factor (unitless) for the contaminant of interest
[site specific measurement - see calculation below],
L = length (m) of source zone in direction of flow [site specific
measurement].
Calculation assumes unidirectional, steady-state flow conditions subj ect to advection and sorption only (dispersion and matrix diffusion
are ignored). The calculation assumes that contaminant mass is not being added to the saturated flow system from any unsaturated
zone sources (e.g., leaching and desorption). Note that R is often approximated in unconsolidated media by
where pb is the dry bulk density (g/cc), $ is the porosity (unitless), Koc is the organic-carbon partition coefficient (ml/g), and^c is the
fraction organic carbon (unitless). Calculations considering dispersion and degradation can be found in (10).
Example Calculation
Using L = 50 m, v = 25 m/yr, and R = 5, the source zone should be flushed of dissolved and sorbed contaminants in approximately
10 years following the last release of contaminants. Dispersion, which always occurs, will lengthen this time as will back-diffusion, if
it is occurring. In cases where complicated flow conditions exist and where it is desired to account for dispersion and back-diffusion,
numerical models can be used to perform the assessment.
Calculation 8 - Conversion to Parent Compound
Daughter product concentrations can be converted to equivalent parent product concentrations by converting the daughter mass/
volume concentrations to moles/volume, attributing that number of moles to the parent, and then converting the parent concentra-
tion to mass/volume.
17
-------
Example Calculation
Consider a ground-water sample containing 500 ppb PCE, 400 ppb TCE, 1300 ppb cis-1.2 DCE and 44 ppb VC at a site where it is
known that only PCE was released to the subsurface. It is assumed that biodegradation has not progressed beyond VC. The PCE
concentration of 500 ppb is less than 1% of the PCE solubility (1% PCE solubility is 2000 ppb). Given TCE, cis-1,2 DCE and VC
molecular weights of 131.5, 97.0 and 62.5 g/mol, respectively, the groundwater concentrations of these compounds are equal to
3.042E-06 mol/1,1.340E-05 mol/1 and 7.040E-07 mo 1/1, respectively. Assuming that each mole of daughter product derives from one
mole of parent product, the equivalent total concentration of parent product is 2.016E-05 mol/1. This corresponds to an equivalent
parent (PCE) concentration of 3343 ppb (PCE molecular weight 165.8 g/mol). which exceeds the 1% solubility value of 2000 ppb.
18
-------
-------
&EPA
United States
Environmental Protection
Agency
National Risk Management
Research Laboratory
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
EPA/600/R-09/119
September 2009
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERED;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
POSTAGES, FEES PAID
EPA
PERMIT No. G-35
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |