*>EPA
EPA/600/R-18/053 | February 2018 | www.epa.gov/research
United States
Environmental Protection
Agency
Engineering Issue:
Soil Vapor Extraction (SVE Technology
Office of Research and Development
National Risk Management Research Laboratory
Land and Materials Management Division
-------
ngineering Issue
v>EPA
United States
Environmental Protection
Agency
Soil Vapor Extraction(SVE)
TABLE OF CONTENTS
1 INTRODUCTION 1
1.1 Purpose.... 3
1.2 Background.............. 3
1.3 Overview of SVE........................................4
2 SVE APPLICABILITY........ ...6
2.1 Contaminant Type and Phase ...6
2.2 Site Characteristics.. 7
2.3 Performance Objectives .11
3 SITE CHARACTERIZATION AND
CONCEPTUAL SITE MODEL
DEVELOPMENT ......................................... .....13
3.1 Site Investigation .13
3.2 Phased Approach .14
4 DESIGN AND INSTALLATION................. .16
4.1 Pilot Testing and SVE Design Basis. 16
4.2 Total Gas Extraction Rate........ 18
4.3 Well Layout and Screening 20
4.4 System Sizing and Vapor Treatment.. 22
4.5 Site Access, System Selection, Layout,
Piping, and Instrumentation..... .23
4.6 Health and Safety Issues 24
5 OPERATIONS, MONITORING, AND
PERFORMANCE EVALUATION....... 25
5.1 Operations and Monitoring 25
5.2 Data Evaluation 27
5.3 System Optimization 29
5.4 System Transitions and Vapor
Intrusion Mitigation... 32
6 SYSTEM SHUTDOWN AND SITE CLOSURE .34
6.1 Framework for Assessing SVE
Termination or Technology Transition .....35
6.2 Summary of Methods for Evaluating
Attainment of SVE Endpoints 37
6.3 Models for Evaluating Impacts of
Residual Mass in the Vadose Zone 38
7 COST CONSIDERATIONS FOR SVE
SYSTEMS 41
7.1 System Design 41
7.2 Cost Components 41
7.3 Operation and Maintenance 41
7.4 Cost Estimating Tools 42
8 SVE ENHANCEMENTS AND
COMPLEMENTARY AND PASSIVE
TECHNOLOGIES 42
8.1 Potential Enhancements to an Existing
SVE System 43
8.2 Passive SVE 44
9 CASE STUDIES................................................46
9.1 Case Study #1: Assessment of a Small
Persistent TCE Source............................46
9.2 Case Study #2: Evaluation of Mass
Transfer across the Capillary Fringe
from SVE Operational Data ...............47
10 ACKNOWLEDGEMENTS.................................49
11 REFERENCES..................................................49
APPENDIX A. Case Study #1, Assessment of a
Small Persistent TCE Source............................ 55
APPENDIX B. Case Study #2, Assessment of
Mass Transfer across the Capillary Fringe
from Contaminated Groundwater to Vadose
Zone Soils [[[63
1 INTRODUCTION
The U.S. Environmental Protection Agency (EPA)
Engineering Issue Papers (EIPs) are a series of
technology transfer documents that summarize the
latest information on selected waste treatment and
site remediation technologies and related issues. The
information is presented in a conveniently accessible
manner to the user community. EIPs are designed to
help remedial project managers (RPMs), on-scene
coordinators (OSCs), contractors, and other
practitioners understand the type of data and site
characteristics needed to evaluate a technology for a
specific site, as well as ways to design and optimize a
technology for a particular application. Each EPA
HIP is developed in conjunction with a small group of
engineers and scientists from inside EPA and outside
consultants, with a reliance on peer-reviewed
literature, EPA reports, Web sources, current ongoing
research, and other pertinent, available, and verifiable
information.
This HI I' assembles, organizes, and summarizes the
current knowledge on soil vapor extraction (SVE)
-------
above the water table. As a technical support
document, it describes SVE technologies with a focus
on remedial scoping needs, but it does not represent
EPA policy or guidance.
2
Soil Vapor Extraction (SVE) Technology
-------
Key Updates to the SVE EIP
• SVE testing as a component of site characterization at VOC-
contaminated sites (Section 3)
• Pilot testing and adaptive SVE implementation (Section 4.1)
• Rebound evaluation for assessing cleanup progress
(Section 5.3)
• Transitioning and implementing SVE for vapor intrusion
mitigation (Section 5.4)
• Relating mass transfer constraints quantified during SVE to
vapor intrusion and groundwater impacts (Section 6)
1.1 Purpose
This document summarizes the state-of-the-science
regarding the widespread use of SVE as a major
treatment technology for removing VOCs from soil.
SVE can be applied alone or as an integral
component of more complex remedial technologies
that volatilize subsurface contaminants (e.g., thermal
remediation, air sparging). This EIP provides updated
information since the issuance of the original
Engineering Bulletin (U.S. EPA, 1991a) and two
Engineering Forum Issue Papers (U.S. EPA, 1996a,
1997b) on SVE. It provides information describing
SVE and its applicability and limitations; site
characterization; design and construction;
performance monitoring, evaluation, optimization,
and shutdown; complementary technologies; costs;
case studies; and references for further information.
Key updates are described in the text box above.
To stay concise, this EIP summarizes relevant
information and provides references and Web links
for more in-depth material. These Web links,
although verified as accurate at the time of
publication, are subject to change.
1.2 Background
SVE is an in situ technology widely used in
commercial remediation for more than 25 years. SVE
is generally a cost-effective remediation process for
gasoline, solvents, and other relatively volatile
compounds. SVE has been the industry default
remedy for VOCs in soils for more than 20 years
(U.S. EPA, 1993, 1996b) and the technology has been
chosen as a component of the remediation plan at
Soil Vapor Extraction (SVE) Technology
more than 285 Superfund sites (U.S. EPA, 2012a).
SVE can be used alone to physically remove VOCs
from unsaturated (vadose zone) soils, but it is often
used with other technologies that enhance
biodegradation (for biodegradable VOCs),
volatilization, or both. For example, air injection into
the underlying groundwater accompanied by SVE is
known as air sparging (U.S. EPA, 2001). During
thermal remediation, SVE is a key component for
capturing volatilized contaminants (U.S. ACE, 2014).
Typical petroleum hydrocarbon releases occur at retail
gasoline stations and bulk fuel transfer facilities
through leaking underground storage tanks and
piping. In these situations, the contaminant usually
exists as a lighter-than-water nonaqueous phase liquid
(LNAPL) that disperses in the vadose zone according
to geology. If sufficient fuel is released, pools of
LNAPL can collect atop low permeability strata and
the water table. In this scenario, SVE is integral to
multiphase extraction that simultaneously performs
vacuum-enhanced extraction/recovery to remove
free-phase LNAPL from the water table, vapor
extraction to remove volatile vapors, and aeration
(bioventing) to stimulate biodegradation (U.S. ACE,
1999; U.S. EPA, 1996c). SVE alone can be effective
at removing significant contaminant mass but can
take a long time. As the most volatile compounds
such as benzene are preferentially stripped from the
LNAPL, the overall LNAPL volatilization rate
diminishes exponentially over time, slowing the
progress towards cleanup goals. Wien this occurs,
transitioning from SVE to a more cost-effective
technology such as bioventing or biosparging is a
common optimization step.
Chlorinated solvents are typically released into the
subsurface through solvent spills, poor storage
practices, or accidental releases of water contaminated
with dissolved solvents (e.g., overflow of a NAPL/
water separator). Chlorinated solvents differ from
petroleum hydrocarbons in that they do not readily
biodegrade in the subsurface (U.S. EPA, 2012b).
Because they do not easily biodegrade, they can travel
through the subsurface with water and by vaporous
-------
diffusion over extended times and distances, resulting
in long plumes and a significant mass of dissolved and
adsorbed chlorinated compounds sequestered in fine-
grained soils. In this scenario, VOC mass extraction
during SVE rapidly becomes limited by the rate of
contaminant migration out of the fine-grained soils
and into the path of soil gas flowing to the extraction
point. Even as the rate of mass removal is largely
diminished, subsurface transport of the remaining
mass may pose an unacceptable source of
contamination for vapor intrusion into buildings or
underlying groundwater under ambient conditions.
The existence of chlorinated solvents as a denser-
than-water NAPL (DNAPL) in the subsurface
generally requires a more aggressive source reduction
technology than SVE alone.
Basic descriptions of SVE can be found in various
online sources (e.g., Federal Remediation
Technologies Roundtable [FRTR], 2008; U.S. EPA,
2012a). U.S. EPA (2010) outlines green remediation
best management practices for SVE and air sparging
technologies. The SVE information in this EIP draws
on numerous publications containing details on the
design, implementation, and evaluation of SVE since
the issuance of the Engineering Forum Issue Papers
(U.S. EPA, 1996a, 1997b). These publications include
U.S. EPA (2001), Ar Force Center for Environ-
mental Excellence (AFCEE, 2001), U.S. Army Corps
of Engineers (U.S. ACE, 2002), and U.S. Department
of Energy (U.S. DOE, 2013). Additional references
for details on various topics are provided throughout
this document.
1.3 Overview of SVE
SVE is an attractive treatment technology for VOCs
such as gasoline and chlorinated solvents because the
soil is treated in place, sophisticated equipment is not
required, and the cost is typically lower than other
remedial options. Favorable conditions for SVE
include:
• Low to moderate soil organic matter content
• Moderate to high permeability soils
• Low to moderate soil moisture
• Modest heterogeneity (e.g., sandy soils
interlayered with thin fine-grained units)
• Moderate to deep depth to groundwater (e.g.,
5-100 ft).
SVE applies a vacuum to vertical or horizontal wells
screened in unsaturated (vadose) zone soils to induce
airflow through the contaminated soil. Figure 1
illustrates a general schematic of the in situ SVE
process and its components. Clean air enters soil
where contaminants reside and carries volatilized
contaminants to the extraction wells. In this way, the
gases in the soil are flushed, or exchanged,
continuously to increase the volatilization rate of
contaminants. Without the application of a vacuum,
the soil gas is relatively stagnant and volatilized
contaminants migrate slowly into groundwater or to
the ground surface as vapors in soil gas. After
entering groundwater, VOC contaminants pose a risk
to drinking water supplies. Where buildings overlie
contaminated soil and groundwater, VOC vapors in
soil gas can collect beneath them, enter the buildings,
and pose an inhalation hazard.
The introduction of clean air into VOC-contaminated
soil volatilizes contaminants for extraction, thereby
reducing the overall contaminant mass in soils. A
vacuum blower supplies the motive force by inducing
gas flow throughout the contaminant vapor plume as
illustrated in Figure 1. Extracted VOC vapors are
separated from condensed water in an air/water
separator before or after the blower. The blower
discharge then carries the contaminated vapors to a
surface off-gas treatment system prior to atmospheric
discharge. Typical SVE off-gas treatment consists of
contaminant adsorption onto granular activated
carbon when concentrations are low or destruction by
catalytic or thermal oxidation for higher
concentrations (AFCEE, 2001; U.S. ACE, 2014; U.S.
EPA, 2006).
4
Soil Vapor Extraction (SVE) Technology
-------
Surface Structures
Subslab Probes
TT
VaporTreatment
Vacuum Blower
Clean Air
*¦ Process Residual
Vapor/Liquid
Separator
LiquidTreatment
Clean Water
Process Residual
Water Sump;
Passive or Active
Extraction Wells
Extraction Wells
Passive or Active
AirVent
i
\ §
\
\
VaporProbes a
WaterTable
Extraction Wells
JUL
\
\
QroundSurface
Figure 1. Process schematic of SVE
The continual flushing of air through the
contaminated soil matrix occurs primarily in
permeable soils and the initial rate of contaminant
removal diminishes rapidly with the reduction of
VOC mass in more permeable soils. Slower rates of
VOC transport in less permeable soils (where soil gas
does not flow) then result in a transition from
advection- to diffusion-limited removal.
The primary factors in designing a cost-effective off-
gas vapor treatment system are the initial contaminant
mass in more permeable and less permeable soils, the
rate of contaminant removal, and the contaminant's
subsequent decay over time. These parameters can be
difficult to predict a priori, and introduce
uncertainties to system design economics (e.g., Will
off-gas treatment by thermal oxidation be required for
a month, a year, or longer? Or is activated carbon
with several frequent exchanges upfront more cost
effective?). Further complications are posed by
whether contaminants exist as LNAPL, such as
gasoline, or DNAPL, such as pure dry-cleaning
solvents. Changing conditions in mass extraction
during SVE are usually addressed in periodic efforts
to optimize the system with respect to costs and
attainment of cleanup goals (AFCEE, 2001).
In addition, determining the endpoint for SVE
operations is complex when considering the potential
impact of residual contaminant mass on underlying
groundwater and overlying buildings or
recontamination of the vadose zone by underlying
contaminated groundwater (U.S. DOE, 2013). The
potential and conditions for transitioning SVE
systems from vadose zone remedies to vapor
intrusion mitigation is a topic of current EPA study
Soil Vapor Extraction (SVE) Technology
5
-------
(Lutes et al., 2017; Schumacher et al., 2017; Truesdale
et al., 2016) and is discussed in Section 5. Recently
developed methodologies for evaluating the impact of
residual mass in the vadose zone with respect to site-
specific cleanup objectives are reviewed in Section 6.
2 SVE APPLICABILITY
For decades, SVE has effectively been used to
remove VOCs from the vadose zone in many
settings; however, this does not guarantee SVE will
achieve acceptable VOC removal at all sites. The
performance of SVE technology depends on site-
specific characteristics (e.g., soil gas permeability,
initial contaminant mass, contaminant mass
distribution) that are often difficult to predict from
conventional site characterization data alone.
However, three general categories of information can
be used to assess the applicability of SVE to a
particular site:
• Contaminant type and phase: How
amenable are the specific contaminants for
removal by SVE, aerobic degradation, or both?
Do contaminants exist as an LNAPL or a
DNAPL?
• Site characteristics: How well can an SVE
system induce significant air flow through the
contaminated soil matrix given a site's
hydrogeologic properties (e.g., permeability,
moisture content, heterogeneity), depth to
contaminant and to groundwater, and surface
obstructions to equipment placement?
• Performance objectives: What are the site
remediation goals, expressed as concentration
or mass flux reductions in specific volumes
and within specific time frames?
The following sections describe how to answer these
questions and decide whether SVE is applicable at a
specific site.
2.1 Contaminant Type and Phase
In general, SVE is applicable to compounds that
volatilize (from water or NAPL) into soil gas at
concentrations that yield significant mass removal
rates relative to the mass in the soil. Important
contaminant properties that relate to assessing the
suitability of SVE at a given site include vapor
pressure, solubility, Henry's law constant, octanol-
water partition coefficient, air and water diffusivities,
and aerobic biodegradation rates. U.S. EPA (1996d)
describes most1 of these parameters and provides
values for most VOCs of interest at SVE sites. How
these properties affect contaminant fate and transport
and SVE performance, is described briefly below.
• Vapor pressure provides information on a
compound's volatility. With respect to SVE,
higher vapor pressures (and low solubility)
allow for faster extraction of contaminant
mass. For compounds with lower vapor
pressures, and adequate aerobic biodegradation
rates, bioventing may be more suitable than
SVE.
• Solubility is the tendency of compounds to
dissolve in porewater. With higher solubility,
moist vadose zone soils retain a higher
dissolved mass in the soil porewater.
• Henry's law constant (H) is defined as the
ratio of a chemical's equilibrium partial vapor
pressure in the gas phase to its dissolved
(aqueous-phase) concentration. SVE
applicability increases with increasing Henry's
constant as the compound increasingly prefers
the gas phase, and Henry's constant is more
important than vapor pressure for evaluating
SVE suitability. For example, methyl tertiary
butyl ether (MTBE) is volatile (high vapor
pressure) but has a very low Henry's constant,
indicating a high solubility and a diminished
SVE mass removal rate.
1 Aerobic biodegradation rates can be found in Aronson et al.
(1999) and Lawrence (2006).
6
So/7 Vapor Extraction (SVE) Technology
-------
• Octanol-water partition coefficient (Kow) is
the equilibrium ratio of the concentration of a
compound in octanol to its concentration in
water and is readily available from laboratory
measurements. Kow strongly correlates to the
partitioning of a compound dissolved in water
and adsorbed to the organic matter in soil
solids (U.S. EPA, 1996d). With a higher
partitioning coefficient, more of a compound
is adsorbed to moist soil solids. This slows
mass removal by SVE because the mass must
dissolve into the water layer surrounding the
soil particles and then volatilize into soil gas for
removal.
• Air and water diffusion coefficients are used
to estimate diffusion rates in soil gas (air) and
soil moisture (water). Water diffusivities are
orders of magnitude lower than air diffusivities,
resulting in much slower diffusive transport
through wet soil than dry.
• Aerobic biodegradation rates determine
whether certain ancillary technologies that
aerate the subsurface, like bioventing or air
sparging, will increase contaminant removal by
supplying oxygen to enhance biodegradation
when used in conjunction with SVE. In
general, petroleum hydrocarbons have high
aerobic biodegradation rates and are amenable
to these technologies while chlorinated VOCs
do not aerobically biodegrade readily and are
not effectively remediated by bioventing.
SVE as a soil treatment technology is effective for
most VOCs and possibly effective for some
semivolatile organic compounds depending on site
characteristics and performance objectives such as a
deadline to achieve remedial action objectives. SVE is
considered ineffective for low volatility compounds
such as polychlorinated biphenyls (PCBs), pesticides,
dioxins/furans, cyanides, corrosives, metals, asbestos,
radionuclides, and explosives. As described in Section
1, SVE is often applicable to fuel hydrocarbons, such
as gasoline, that exist as an LNAPL. However, if
chlorinated solvents are present as a DNAPL, SVE
alone is generally inadequate to address the mass
because of the duration of operation.
2.2 Site Characteristics
Given a contaminant in soil with a volatility yielding
an appreciable vapor concentration, the success of
SVE generally depends on effective soil gas flow and
contaminant mass transfer in the subsurface. The
primary site properties governing soil gas (air) flow
and contaminant transport are:
• Soil gas permeability
• Moisture content
• Organic carbon content of the soil
• Heterogeneity in the above properties.
Dry soil gas permeability can be related to soil grain
size (Domenico and Schwartz, 1990) and native soil
permeability can be estimated from laboratory tests
on undisturbed soil cores; however, field extraction
tests are directly applicable to SVE and yield bulk soil
estimates of permeability. Air (soil gas) flow in the
subsurface is generated by applying a vacuum to
extraction wells screened in the contaminated soil.
From Darcy's law, the extraction rate is proportional
to the soil permeability to air flow and the applied
vacuum (Johnson et al., 1990b). The total system
extraction rate depends on the number of wells, the
length of well screens, and, to a lesser extent, the well
spacing. Johnson et al. (1990b) provide a simple
expression for flow from a single well assuming a
homogeneous, uniform soil permeability. Based on
their work, the following engineering relationship can
be used to estimate the maximum practical flow from
a single well, Q (scfm), per unit length of screen, H
(ft), or given measures of those parameters, an
estimate of the absolute dry soil gas permeability, k
(Darcy):
Q_ _ nkkrPw [1 - (Patm/Pw)2]
H ln(Rw/Ri)
This relationship includes the absolute pressure at the
well, Pw, and the atmospheric pressure, Patm (inches
Soil Vapor Extraction (SVE) Technology
7
-------
of H20), the ambient viscosity of air, // (g/cm/sec),
the radius of the extraction well, Rw (ft), and an
assumed radius of influence for the SVE vacuum, Ri
(ft).
In this relationship, kr is the unitless relative
permeability. The relative permeability is a function of
the fraction of pore space occupied by water, known
as saturation, S. Higher water contents reduce the
permeability to gas flow by occupying more pore
space and reducing the air-filled porosity. The water
saturation is readily available from straightforward
measures of the soil total porosity and moisture
content. Numerous correlations exist for relative
permeability based on moisture content (as fractional
saturation). The following common correlation is a
modified Brooks-Corey relation (Brooks and Corey,
1964):
kr = (1 -S)3
When the water saturation is high (on the order of
70% to 80% of the pore space), the gas phase is
mostly disconnected into bubbles and gas flow is
orders of magnitude less than the flow through dry
soil. Under these conditions, very little if any gas flow
occurs under practical vacuums at the extraction well.
To utilize the above Johnson equation, the vacuum at
the well is input as the absolute pressure at the well,
Pw, assumed in this example to be 100 inches H20
less than the atmospheric pressure, Patm, which is
assumed to be standard pressure at sea level of 407.5
inches H20. The expression also includes the ambient
viscosity of air, //, at 0.00018 g/cm/sec. Values for
Rw and Ri. vary by well construction and site
conditions but a typical well radius is 0.5 ft and an
example radius of vacuum influence is 100 ft
(Johnson et al., 1990b). Entering these values into the
Johnson equation above yields:
Q
H
scfm
/t
= k[Darcy]( 1 — S)3( 1.42 )
For a fine sand with a permeability of 1.0 Darcy and a
low porewater saturation of 0.15, we find Q/His
approximately 0.9 standard cubic feet per minute per
foot (scfm/ft) while a medium sand of permeability
5.0 Darcy with a high water saturation of 0.5 yields
the same rate. A high permeability sand of 10 Darcy
with a low porewater saturation and a 20-ft screen
yields 180 scfm. Consider a silt (0.1 Darcy) with a
moderate saturation (0.4) to find a maximum flow of
only 0.03 scfm/ft. For a 20-ft screen, the silt yields
less than 1 scfm and would necessitate many closely
spaced extraction wells to recover appreciable
contaminant mass under conditions typical of a
gasoline station. SVE alone may not be suitable in this
last scenario. Contaminated vadose zones with high
moisture contents also pose numerous process
challenges for implementing SVE, including the
generation of large quantities of wastewater and
fouling of activated carbon.
For sites where appreciable gas flow can be generated,
the contaminant mass extraction rate is governed by
the transfer of mass from soil matrices to the flowing
soil gas. The initial mass extraction rate can be
estimated from measures or estimates of vapor
concentration at the site multiplied by the anticipated
total extraction rate. However, this rate diminishes
over time, sometimes rapidly, as contaminant vapors
in permeable pathways are swept out. This initial rate
of diminishment is proportional to the contaminated
soil volume and the extraction rate (i.e., the time
required to sweep the contaminated vapor pore
space) as discussed in Section 4.1.
After the initial decay in extracted concentration
stabilizes, mass transfer constraints emerge that
govern long-term trends and time to attain remedial
action objectives. Mass transfer limitations exist on
multiple scales, from pore-level partitioning between
solid/liquid/gas phases to vapor diffusion on the
scale of feet associated with geologic heterogeneity
(e.g., permeability contrasts such as soil layering) and
varying moisture content (Li and Brusseau, 2000).
8
Soil Vapor Extraction (SVE) Technology
-------
Partitioning among phases on the pore scale is
generally much faster than mass transfer constraints
posed by geologic heterogeneities (Brusseau, 1991).
Pore-level partitioning is generally assumed to be at
equilibrium with chemical and soil properties defining
the distribution of contaminant between the mass
volatilized in vapor, mass dissolved in adjacent
porewater, and mass adsorbed to soil solids. The soil
properties include absolute (total) porosity (
-------
has a high water content and the air-filled pore space
is disconnected and separated by lenses of water, the
contaminant must diffuse through this water before it
can be volatilized and extracted. The slow process of
diffusion is the most common mass transfer
constraint limiting cleanup at SVE sites where
LNAPL or DNAPL does not exist. Consider again
the example of the perched water lens; the extracted
vapor concentrations may fall rapidly during initial
SVE operations as contaminants are removed in soil
gas from the more permeable zones, but once the
SVE operations cease, contaminant vapor
concentrations build slowly back to the initial
condition by diffusion from the clay and water lens as
only a small fraction of the initial contaminant mass
was removed.
The timescale for diffusive transport (and mass
transfer limitations) and the time to attain remediation
goals can be approximated from mathematical models
of contaminant transport. Such models are available
from the U.S. ACE (2002, Appendix F) and the U.S.
EPA (2001, Chapter 13). The rate of vapor diffusion
is governed by the retardation coefficient in the
diffusive soil, the thickness of the fine-grained unit
(Z3), the diffusion coefficient of the contaminant in
free air (Dair), the gas tortuosity of the soil (t) (Jury et
al., 1991), and the differential in the contaminant
vapor concentrations as contaminants move from the
diffusion-limited high moisture soil into the more
permeable sandy soil where advective conditions
predominate (U.S. ACE, 2002). An estimate for the
optimal cleanup time, tc, for a diffusive source can be
calculated from the following equation, assuming the
advective soil is kept at a very low concentration by
the sweep of SVE (U.S. EPA, 2001):
C,
endpoint
Cinitial
= exp
TD • TEZ
luair11
~RZf
Tortuosity,r =
-------
This relationship displays the exponential decay in
concentration characteristic of long-term SVE and
the tortuosity relationship illustrates the strong
dependence of vapor diffusion on soil water
saturation (5). The thickness of the fine-grained soils
harboring contaminants is clearly a dominant
parameter. If the saturation is high, diffusion through
water governs the process, and this diffusion will be
about three orders of magnitude slower than diffusion
through air. This makes SVE likely impractical but
also contains the contaminant until the moisture
content decreases. Rearranging the equation yields an
expression for the optimal cleanup time given the
initial concentration and a desired endpoint
concentration:
^ _ RZg ^ / C initial \
T-Da(rTr2 \CendpointJ
As an example, assume sufficient flow is maintained
to yield a low advective concentration (i.e., maximum
concentration differential), a total porosity of 0.35,
and a typical retardation coefficient for TCE of 3
along with a free air diffusion coefficient of 8 ft2/day.
The time required is represented as:
0.15 / Cjnitiai \ day
(1-S)3-3 n \CendpoinJ ft2
For a silt interval with a thickness of 3 ft and a water
saturation of 0.60, we find the minimum timescale for
cleanup effecting a reduction of two orders of
magnitude in concentration within the silt to be on
the order of 120 days, suggesting SVE will be
effective. Whereas a thickness of 10 ft yields a best-
case timescale exceeding 4 years despite best efforts
to optimize operating conditions. Hence, sites with
thick, moist silt or clay horizons in the vadose zone
that have experienced decades of exposure to
contaminants may not respond to SVE alone in a
timely manner and are strong candidates for the
application of enhancements such as heating or
fracturing (U.S. EPA, 1997b).
Soil Vapor Extraction (SVE) Technology
Figure 2 illustrates the removal of contaminant mass
from the vadose zone above groundwater. At sites
with contaminated groundwater, this water can also
serve as a long-term source of vapors for extraction
or volatilization and transport to the surface, a subject
of interest for the mitigation of vapor intrusion.
Determining the point of cessation for active SVE is
complicated by the presence of contaminated
groundwater: Will residual mass in the vadose zone
impact groundwater in the future or will groundwater
recontaminate the vadose zone negating much of the
SVE effort and pose an unacceptable vapor intrusion
risk? These questions are related to current EPA
research efforts (Lutes et al., 2017; Schumacher et al.,
2017; Truesdale et al., 2016) on the operational
parameters and cost effectiveness of transitioning
SVE systems from vadose zone cleanup to vapor
intrusion mitigation. If contaminated groundwater is
the primary source of contaminants for vapor
intrusion, SVE has the potential to capture such
vapors over a wide area, protecting multiple buildings
on the scale of a city block with a minimal number of
wells.
SVE is also applicable to LNAPL contamination,
such as gasoline, that is amenable to aerobic
biodegradation. Vapor extraction induces direct
volatilization of the lighter-end components in
LNAPL, and, thus, can be an optimal approach at
sites with small to moderate levels of LNAPL because
it simultaneously promotes contaminant removal and
the introduction of oxygen for biological degradation.
Sites with extensive masses of LNAPL would likely
benefit from direct LNAPL recovery (e.g., dual-phase
vacuum extraction along the groundwater smear
zone) before considering SVE (U.S. EPA, 1997b).
LNAPLs pose additional constraints on volatilization
through the limited surface area between the LNAPL
and flowing gas and the reduced vapor concentrations
from a multicomponent LNAPL according to
Raoult's law (Carroll et al., 2009).
These major subsurface conditions and contaminant
attributes greatly affect the potential for SVE to be
successful at a given site, although this list is certainly
-------
not exhaustive (also see AFCEE, 2001; FRTR, 2008;
MPCA, 1993). Additionally, it is important to note
that SVE has been successful in suboptimal
conditions, such as finer, wetter soils (e.g., clays), but
at much slower removal rates (U.S. ACE, 2002).
2.3 Performance Objectives
Regulatory compliance, exposure pathway risk, and
post-cleanup site use (e.g., residential, commercial,
industrial) generally lead to quantifiable goals for
contaminant concentrations or fluxes and a desired
attainment timeframe that are independent of
remedial technology selection. These are usually
documented in a Record of Decision (ROD) for
Superfund sites or a Corrective Measure Decision at
Resource Conservation and Recovery Act (RCRA)-
regulated corrective action sites. If SVE is determined
to be applicable and the goals are technically
achievable, these goals provide the basis for SVE
design, operation, optimization, transition, and
closure. The order-of-magnitude calculations in
Section 2.2 illustrate the first steps in setting the
design basis.
The most common method for determining SVE site
closure criteria in the past has been the assessment of
potential aquifer degradation resulting from the
transport of residual contaminants in the vadose zone
to underlying groundwater. However, vapor intrusion
into buildings has recently gained an equal footing for
setting SVE site closure criteria. Both pathways are
governed by the diffusive transport of vapors from a
source in the vadose zone to an interface such as the
foundation of an overlying building (for vapor
intrusion) or the underlying water table surface (for
the groundwater pathway). Mass transfer can be
assessed across such interfaces to estimate the impact
of the diffusive transport.
A typical remedial goal for the groundwater pathway
requires that groundwater concentration meet a
drinking water standard at a compliance location.
Quantification of the VOC flux into the groundwater
and subsequent calculation of estimated VOC
concentrations in groundwater using a mixing
approach is one method to address this requirement.
Similar flux calculations and mixing models can be
applied to building basements or floor slabs to
address the vapor intrusion pathway. Such methods
are described by U.S. DOE (2013). Requirements to
perform modeling based on these approaches can be
reviewed before site characterization and SVE design
to ensure that adequate data have been collected and
to assess the likelihood of success in meeting remedial
goals.
Assessing the rate of VOC mass removal by an SVE
system is a closure approach often suggested by SVE
operators. A strong desire arises to discontinue SVE
when the VOC removal rate decays very slowly or
approaches a perceived asymptote at low
concentrations; however, this behavior alone is not a
sufficient basis for termination. Asymptotic behavior
is a supporting line of evidence for observations
during rebound testing and for input to models
assessing impacts to groundwater and vapor intrusion.
This topic is discussed further in Section 6.
After a long application of SVE, significant VOC
mass may remain in less permeable soils and
concentrations may rebound after SVE is stopped. In
such cases, removal can often be improved by altering
the extraction strategy (e.g., operating different or
new wells, transitioning to a lower extraction rate,
pulsing SVE operation). Transitioning an SVE system
from vadose zone cleanup to vapor intrusion
mitigation is an area of active EPA research (Lutes et
al., 2017; Schumacher et al., 2017, Truesdale et al.,
2016). Sites deemed impractical for cleanup by SVE
(e.g., DNAPL) can consider SVE for mitigation of
vapor intrusion in lieu of subslab or basement
systems. Design of systems for this objective follow
the same procedures outlined in this EIP for vadose
zone cleanup.
Soil Vapor Extraction (SVE) Technology
-------
3 SITE CHARACTERIZATION AND
CONCEPTUAL SITE MODEL
DEVELOPMENT
The primary objective of site characterization is to
inform the development and maintenance of a
conceptual site model that includes information to
design, install, operate, and evaluate an SVE system to
meet performance objectives. This information can
be expressed as the responses to several questions:
• What are the contaminants and in what phase
(LNAPL or DNAPL or dissolved) were they
released?
• What volume of soil is contaminated and where
are the suspected source areas?
• What are the properties of the soil volume for
transmitting soil gas? What fraction of the volume
effectively transmits soil gas?
• Do major variations in permeability (e.g., soil
layering) occur and if so, how thick and what are
the physical properties of each layer? Are such
intervals extensive across the site or are they
limited lenses?
• What is the depth to groundwater? Do unusually
high soil moisture contents occur anywhere in the
vadose zone? Does the water table fluctuate
seasonally or with changes in local water usage?
• Is the groundwater contaminated? What
concentrations are found?
• What is the surface boundary condition? Is the
contaminated soil volume under a building,
parking lot, open field? Do preferential utility
pathways or fill materials exist in the shallow
subsurface?
• Where are potential receptors located and what
are the current and future uses anticipated for the
site? What are the expected endpoints and
timeframes for the remediation based on the
performance objectives?
The answers to these questions form the basis of a
robust conceptual site model for the implementation
of an SVE remedy at a VOC-contaminated site. As
with any conceptual site model, it can be continually
refined as the SVE process moves through its various
phases and additional information becomes available.
3.1 Site Investigation
Data collected during all site investigation activities
can be matched to one or more of the remedial
objectives for the site. The scoping and applicability
equations in Section 2 provide a template for the
parameters to be measured and their specific purpose.
For example, drilling and installing a single well can
include the collection and analyses of soil cores from
each soil type encountered while logging the thickness
of differing stratigraphic intervals (or soil layers) in
the well. These soil samples can be screened in the
field for VOCs with handheld instruments, shipped to
a geophysical laboratory for physical property analyses
(e.g., porosity, moisture content), or shipped to an
analytical laboratory for detailed chemical analyses.
Drilling can also include the periodic collection of soil
gas samples ahead of the drill depth to assess the
vertical extent of contamination and a grab water
sample if the boring extends into the groundwater.
However, chemical analyses of very small samples of
soil, water, and soil gas recovered from borings
provide limited representativeness of the larger
system. The location of this first and additional
borings can be based on historical site usage, prior
investigations, or new soil gas samples collected from
temporary or permanent probes installed in the
shallow vadose zone to better define the areal
footprint of the vapor plume. These various
investigative methods are described in detail in
AFCEE (2001) and U.S. ACE (2002).
While the data described to this point are static in
nature and reflect near equilibrium conditions, SVE is
an active and dynamic process based on creating
disequilibrium in concentration gradients. Hence, the
site characterization can include steps to induce and
measure disequilibrium early in the process before a
full-scale system design is performed. For example,
upon completing the first soil gas extraction well and
a few soil gas sampling probes, the well can be used
Soil Vapor Extraction (SVE) Technology
13
-------
for short-term extraction tests to assess soil
permeability, the response of subsurface VOC vapor
concentrations during an initial sweep of permeable
soils, and the rate of concentration rebound after this
initial extraction. Such testing represents the initial
SVE pilot test described below and in Section 4.1.
Detailed descriptions of demonstrations combining
site characterization with SVE pilot testing at multiple
Air Force sites are available (Praxis, 2000).
3.2 Phased Approach
Once the conceptual site model is sufficiently
advanced to recognize the general nature and extent
of VOC contamination, extraction testing can be
conducted in a well near the suspected center of the
source, or at least the center of the VOC vapor
plume. The vapor plume will likely provide a cloud of
contamination that masks the locations of much
smaller source volumes. These can be uncovered by
operating in pilot mode for a sufficient period to
sweep a pore volume of soil gas from the initial
plume volume while monitoring vapor concentration
reductions (or increases) at surrounding soil gas
monitoring points. SVE and bioventing system
designers can collect site-specific venting
performance data during this phase to support the
design process. The pilot test can collect air
permeability data from vacuum responses in
monitoring probes (including subslab probes if
applicable) as well as VOC concentrations in soil gas
to assess the radius of effective remediation from the
single extraction well.
For permeable sites, vacuum responses may be low or
undetectable in surrounding probes because of
instrument limitations or variations in atmospheric
pressure even though significant soil gas flow is
occurring. In this case, monitoring of concentration
responses is the most valuable metric for assessing
radius of influence. This approach is also true of more
distant soil gas probes at sites where lateral flow is
dominant; vacuum response may be very low or
undetected but concentration responses are observed.
Without such data, more wells than necessary might
be installed and operated inefficiently.
For these reasons, beyond pilot testing, a phased
approach to SVE system installation and operation
has advantages over a single full-scale SVE design
event. Much can be learned about the way a site
behaves during remediation each time the operator
goes to the field to collect data on system
performance. However, these opportunities do
require time and resources for which their judicious
use balances against the extra expense and time of
suboptimal operation. Because each new phase of the
SVE remediation system is predicated upon the
knowledge gained from the previous phases, design
flexibility is essential because it accommodates design
and operational changes as more information is
gained about the site.
An initial remediation phase or a pilot test (discussed
in Section 4) can further the understanding of the site
and the applicability of SVE to remediate the site
(U.S. ACE, 2002). In comparison to more traditional
engineering projects (e.g., bridge design), the basis for
design for subsurface environmental remediation is
quite weak and often the very execution of a
remediation design (e.g., installation of SVE injection
and extraction wells) dramatically increases the
understanding of site characteristics and confidence in
the conceptual site model. Considerable time and
expense can be saved if SVE is assumed to be the
remedy for sites contaminated by VOCs in the vadose
zone and the SVE is then designed and implemented
using a phased approach. If SVE alone is found to be
inadequate in an early phase, a robust conceptual site
model is available to assess more aggressive
technologies.
In a phased approach to design and implementation,
site characterization and SVE pilot testing merge to
facilitate deployment of cost-effective vapor
treatment systems, minimize oversized equipment,
minimize the number of wells operated (and therefore
the number of "dead zones" between wells where
remediation is inefficient), and identify and address
14
Soil Vapor Extraction (SVE) Technology
-------
persistent sources early in the process to reduce the
remediation timeframe. The most efficient SVE
system utilizes one extraction well in the center of the
source volume, as long as it reduces the furthest
lateral extent of the vapor plume of concern within
the same timeframe as reductions in concentrations in
the source volume. Consider the simultaneous startup
of multiple extraction wells; the wells effecting
changes at various locations cannot be discerned and
dead zones, also known as stagnation zones, may
provide false indicators of a persistent source volume
only because flow is locally stagnant despite a
substantial vacuum response.
The primary SVE design parameters are (1) the nature
and extent of contamination in the soil, (2) the
permeability distribution (i.e., heterogeneities) in the
soil, and (3) contaminant concentrations in extracted
soil gas. In historical practice, the nature and extent of
contamination in the soil were determined with
expensive chemical analyses of very small samples of
soil, water, and soil gas recovered from exploratory
borings. Because of their limited representativeness of
the larger system, such results are not amenable to
predicting SVE performance. Data representative of
remediation conditions are needed to develop an
appropriate design. In recent practice, field testing as
described above is used to gather air permeability data
and extracted VOC concentrations in soil gas. Many
guidance documents (e.g., DePaoli et al., 1991;
Johnson et al., 1990a, b; Pederson and Curtis, 1991)
describe this test. However, the extraction testing has
a deficiency because it does not account for the soil
heterogeneities that control cleanup (Armstrong et al.,
1994; McClellan and Gillham, 1990). Site
characterization techniques to quantify the impact of
heterogeneities on SVE can also be employed.
The U.S. Air Force conducted field demonstrations of
advanced characterization tools for sites with soil
layers of varying permeability (AFCEE, 2001). The
tools included direct push with Geoprobe's
membrane interface probe (MIP) technology and
vertical profiling of extraction well screens using the
PneuLog® technology. PneuLog provides
Soil Vapor Extraction (SVE) Technology
simultaneous, continuous logs of VOC
concentrations and flow entering along an installed
well screen. Interpretation of the data provides
continuous vertical profiles of air permeability and
adjacent soil gas concentrations. Logs from multiple
operating SVE wells effectively define the extent of
mass transfer limitations at a site. Additional
information and descriptions of the MIP and
PneuLog technologies can be found in Parsons (2001)
and U.S. ACE (2002).
A second method to assess the mass transfer
limitations in general, or in conjunction with the
profiling tools, is to conduct the pilot test over a time
sufficiently long enough to sweep the permeable soils
and observe VOC concentration reductions in soil
gas, and then observe the rebound in concentrations
in soil gas probes once the SVE system is shut down.
The rate of rebound and the reduction in the
equilibrated rebound concentration provide valuable
data for assessing the time of operation to achieve
performance objectives as well as a benchmark for
assessing progress toward cleanup during subsequent
operations (Brusseau et al., 2010, 2015).
In summary, a phased approach to SVE system
design and installation identifies and addresses
persistent sources early in the process to allow course
corrections and to reduce the remediation timeframe.
Rapid reduction of a vapor plume may provide
mitigation, but it does not necessarily indicate
significant progress toward cleanup if the system is
not extracting soil gas efficiently from low
permeability areas, dead zones, or unknown sources
where significant VOC mass may reside. With a
phased approach, as the conceptual site model is
refined and the SVE process moves forward, such
problem areas in the subsurface can be identified early
and subsequent efforts can be focused, including
possible SVE enhancements described in Section 8.
-------
4 DESIGN AND INSTALLATION
The implementation of SVE appears deceptively
straightforward as only limited parameters form the
basis for design of technology infrastructure:
• Total gas extraction rate, Q
• Number and location of extraction well screens
• Vapor treatment technology for system off-gas.
The primary determinants for these SVE design
parameters are the (1) nature and extent of
contamination in the soil, (2) permeability distribution
(i.e., heterogeneities) in the soil, and (3) contaminant
concentrations in extracted soil gas. This information
is expected to be available from an evolving
conceptual site model as described in Section 3 and
the pilot testing described in Section 4.1. As described
in this section, a myriad of straightforward details
arises in the design and implementation of SVE after
specifying the vapor extraction rate and well designs
and layout.
4.1 Pilot Testing and SVE Design Basis
Conventional site characterization data are important
for evaluating SVE; however, these data are relatively
static and do not provide adequate data for full-scale
design. In particular, the dynamic behavior of the
contaminant mass extraction rate is difficult to predict
without performing a pilot test of SVE. The
extraction behavior is governed largely by the volume
of contaminated soil, the fractions of the soil volume
characterized as advective versus diffusive, the mass
transfer characteristics of the diffusion-limited source
zones, the location of extraction screens relative to
sources, and the existence of a NAPL. The following
discussion does not consider a NAPL, although a
zone of persistent concentration that returns to near
identical equilibrium concentration after multiple
periods of extraction is an indicator of NAPL.
The earlier pilot testing occurs in the remedial
planning process (preferably as a component of site
characterization), the less likely that design
modifications will be needed after system startup.
16
Pilot testing is especially recommended at larger,
more complex sites.
Designing the pilot test requires specifying a desirable
total gas extraction rate or duration of extraction.
Ideally, the pilot test will extract the equivalent of one
or more (three in the example below) full pore
volumes of soil gas from the contaminated soil. The
purpose of this flushing is to operate the system long
enough to observe the initial decay in the extracted
VOC concentration and concentration reductions in
soil gas probes at varying distances. This will provide
a first estimate for mass transfer constraints and the
radius of effective remediation from a single well
(DiGiulio and Varahan, 2001a). As a rule of thumb,
the rate and duration for the pilot test can be based
on the total volume ( V) of contaminated soil in the
conceptual site model, the soil porosity, and the soil
moisture content as follows:
Qt = vSOii
-------
also that previous vapor sampling suggested a
maximum TCE vapor concentration of 600 ppmv
(3,300 mg/m3), yielding an initial TCE mass
extraction rate of:
k[Darcy] >
40 scfm
initial mass extraction rate
lbs~\
= QC°
= (40 scfm) (3,300
v m6'
lbs
=°'5 hr
Assuming 200 lbs of granular activated carbon
(capable of adsorbing 10% of its weight in TCE) is
procured as a primary vapor treatment system (with a
second drum for polishing before atmospheric
discharge), a minimum extraction period of 40 hours
is available (200 lbs x 0.1 / 0.5 lbs/hr). However, the
concentration is expected to decay after the initial
16 hours of extraction, allowing for a longer period
(e.g., up to 72 hours).
TCE vapor concentrations during 3 days of extraction
at 40 scfm in a well placed near the center of a
suspected source zone for TCE vapors are shown in
Figure 3. The extracted concentration decayed
rapidly during the initial hours of extraction in
accordance with the estimated soil gas extraction and
exchange rate. The TCE concentration then followed
a much slower decay during subsequent extraction
that is associated with diffusive mass transfer
constraints in a confining clay unit in the middle of
the vadose zone. These observations indicate that the
pilot system was adequate to serve as the full-scale
system at this small site. Use of activated carbon for
off-gas vapor treatment was also demonstrated to be
cost effective.
Note that the desired flow rate could be achieved
with extraction through a single well only if the soil
permeability satisfied the previously described flow
relationship:
Q
H
scfm
/t
= k[Darcy]( 1 — S)3( 1.42 )
Soil Vapor Extraction (SVE) Technology
(20 /t)(l — 0.3)3(1.42 )
= 4 Darcy (a medium sand)
For lower soil permeability, a second well may have
been required to achieve the desired flow or a longer
flushing period may have been necessary to identify
the mass transfer constraints. As described later,
additional information on the mass transfer
constraints was obtained by measuring the rebound in
the TCE vapor concentration at the well after
extraction ceased. In addition, if the TCE vapor
concentration had been higher initially and persisted
at a substantially higher value after the initial decay,
suggesting the existence of a DNAPL, carbon
adsorption may not have been cost effective for the
higher mass extraction rate.
Monitoring points can also be installed at multiple
depths, including subslab if applicable, and within the
radius of influence range (e.g., 10—50 ft) of a pilot
extraction well, if not already available from previous
site characterization activities. Each monitoring
location could have multiple nested points across the
vertical extent of the vadose zone depending on the
depth to groundwater and the geologic layering. As
illustrated in Figure 4, points can be placed above,
below, and within suspected sources. During pilot
testing, these locations are used to measure both
vapor concentration and vacuum responses.
The utility of vacuum data is highly dependent on the
permeability of the soils and the data cannot be relied
upon to assess the radius of influence for SVE. Of
more importance is the vapor concentration response.
In permeable sands, a very small vacuum response
may be associated with a relatively high flow of air,
whereas a significant vacuum response in a clay
provides no evidence that appreciable flow is
associated with the vacuum. However, the vacuum
monitoring data can be used to assess the lateral
versus vertical extent of flow and the impact of
surface conditions (e.g., low permeability leakage
across a slab or a soil surface open to atmosphere) on
the flushing of the surface soil volumes.
-------
— 700
>
i. 600
o
a>
<_>
o
u
500
400
300
200
o
5 100
0
.'nit
ial Flush
3"
O
O
°o
0
0 0
Mass Trai
isfer Limite
'd
oc
0
O
0
0.5 1 1.5 2 2.5 3
Elapsed Days from the Start of Extraction
3.5
Figure 3. Example vapor concentration data from an SVE pilot test
During pilot testing, a robust monitoring program for
VOC vapor concentration is recommended to
identify trends in soil gas monitoring points. These
trends can be correlated with the pore volume of soil
swept during the pilot test to provide a basis for the
spacing of extraction wells in the full-scale design
based on the desired flushing frequency (i.e., pore
volume exchange rate), as discussed in the next
section. Use of a field gas chromatograph by an
experienced operator is encouraged to cost effectively
increase the size of the soil gas VOC dataset.
Often, the direct discharge of off-gasses without
treatment is unacceptable because of health, safety, or
public concerns. If conditions indicate it is necessary,
off-gas treatment technologies such as activated
carbon, thermal oxidization, or other relevant
technologies can be implemented to improve the off-
gas quality for release to the atmosphere (U.S. EPA,
2006).
4.2 Total Soil Gas Extraction Rate
and the exchange rate for that pore volume. The
VOC-impacted pore volume is determined from the
total volume of contaminated soil in the conceptual
site model, the soil total porosity, and the soil
moisture content. Conceptually, the pore volume
exchange rate needs to be frequent enough to
maintain a low extraction concentration relative to the
source vapor concentration to optimize mass transfer
from diffusive source zones to soil gas flowing
through advective soils.
Recall the previous discussion of optimal timescale to
attain cleanup in Section 2.2. This timescale is based
on a characteristic time for contaminants to diffuse
into permeable soils and be extracted with the soil
gas. The time constant (or characteristic time) in the
exponential decay for mass transfer provides a first
estimate for the frequency of sweeping the soil pore
volume to maintain a near maximum concentration
gradient (U.S. ACE, 2002):
Mass Transfer by Diffusion
RZd
timescale, tdiff ~ ^13(1 _ ^3.3^.^2
As illustrated in the pilot test example, the total soil
gas extraction rate for design of an SVE system is
related to the pore volume of the contaminated zone
18
Soil Vapor Extraction (SVE) Technology
-------
Upper Boundary is Site Specific (eg. Atmosphere, Gap, or Basement) ;
Cd Diffusive Source
¦
Ca Advective Soil
U ^ Groundwater
Aquitard
r
^gw
J I
Upper Boundary is Site Specific (eg. Atmosphere, Gap, or Basement)
B
HI
Diffusive Sources
Ca Advective Soil
y ^ Groundwater
c
^gw
Aquitard
Figure 4. Conceptualized scenarios for diffusion-limited mass transfer and typical soil gas monitoring points
Using this timescale to define the pore volume
exchange frequency yields the following estimate for
the minimum total soil gas extraction rate to maintain
a low extracted VOC vapor concentration and
relatively high mass transfer rate:
Flushing Timescale
« Mass Transfer Timescale
t =
VsoiMi-s)
RZ2d
(p13(l-Sy3Dairn2
Q =
^soil^air
Z2R
n2
-------
single, thick fine-grained interval governs the mass
transfer, although the rate is expected to be slow and
the interfacial area with flowing air is small. However,
cleanup of the more complex scenario on the right
will proceed faster for the same volume of
contaminated soil and mass of contaminant because
of the shorter diffusion lengths and increased
interfacial area between the flowing gas and sources.
For the more complex scenario illustrated on the
right, an alternative, more empirical approach is
recommended. Fitting the concentration during
extraction to a sum of exponential decays,
representing the early sweep of the site and later mass
transfer limited mass removal, yields an estimate for
the site-averaged rate of mass transfer. These periods
are illustrated in Figure 5 and the fitting equation and
procedures are described in Section 5, U.S. ACE
guidance (2002, Appendix F), and U.S. DOE
guidance (2013, Addendum to Appendix A). A
detailed discussion of other similar design models is
available in Nyer et al. (2001). Figure 5 also illustrates
the reduction in equilibrated VOC concentration
observed with each subsequent rebound period as
source VOC mass is removed and allows an
assessment in progress toward cleanup as described
by Brusseau et al. (2010, 2015).
4.3 Well Layout and Screening
The layout and screening of extraction wells can be
based on optimizing the exchange rate of the pore
volume in the contaminated soil as described above.
The radius of influence assessed from vacuum
responses away from the extraction wells is a
secondary consideration used to refine the well layout
(DiGiulio and Varahan, 2001a). The objective of
extraction well spacing is to create subsurface flow
sufficient to maintain mass transfer near maximum
rates throughout the contaminated soil volume.
Hence, the number of wells required is based on the
total soil gas extraction rate and the anticipated soil
gas extraction rate from a single well. Single-well
extraction is determined from pilot testing and
O Measured Concentration
Model Fit
Rebound 1
Extended Extraction
Rebound 2
Elapsed Days from the Start of Phase I Extraction
Figure 5. Parameter fit to SVE operations and periods of rebound
20
Soil Vapor Extraction (SVE) Technology
-------
estimated from site permeability and the vacuum
applied to the well screen. The minimum number of
extraction wells (or screens) required is simply the
total gas extraction rate divided by the flow per well
(or screen). This number of wells and the spacing are
refined based on the pilot test results, the locations of
sources, and the necessity to rotate extraction among
wells to mitigate the impact of stagnation zones
between extraction wells. During pilot testing, vapor
concentration monitoring is expected to identify
zones of short-term response as compared to those
with little response. A lack of concentration decay
during the pilot test extraction has two likely causes.
First, if installed within permeable soils, the
monitoring location is too distant from the extraction
point and therefore beyond the radius of effective
remediation. Second, if located close enough for
locations, the extraction point is installed within a low
permeability soil that will ultimately be associated with
mass transfer constraints limiting the time of
remediation.
The layout of wells is expected to be more closely
spaced near the center of source zones and less dense
toward the outer boundaries. As mentioned
previously, the optimum SVE layout is a single well in
the center of the contaminant source if the radius of
effective remediation for that well is acceptable.
Other well design components include determining
extraction well depth and screened interval, lateral
extraction well placement and distribution, the use of
nested wells, horizontal wells or existing wells, the use
of vent wells (air injection wells), and well
construction parameters, are more fully discussed in
U.S. ACE (2002) and summarized below.
• Nested wells: At sites with a deep vadose
zone and multiple lenses of contamination, a
nested well design incorporating two or more
wells in the same boring at different depth
intervals may be ideal (e.g., a shallow,
intermediate, and deep well nest would include
three physically separate wells located within a
few feet of each other vertically). Nested wells
are also recommended because of the pressure
drop experienced along the well screen; for
long screens, it is common for most gas flow
to emerge near the top of the screen rendering
the deeper portion of the screen ineffective.
• Horizontal wells: Horizontal vacuum
extraction wells or trenches are effective at
sites with shallow water tables, especially if
surface seals are used to reduce short
circuiting, or under surface infrastructure (e.g.,
buildings and roads). With a shallower depth to
water, it is more likely that an extraction well
screen will extend into the saturated zone and
vacuum strength from the SVE system will
cause additional water to rise further up the
screen and create air blockage (Suthersan,
1999a). Additionally, if the vadose zone depth
is less than 10 ft and the area of the site is quite
large, a horizontal piping system or trenches
may be more economical than conventional
wells. Horizontal wells are also useful to reach
VOC-contaminated soils under roads or
occupied buildings.
• Existing wells: Existing monitoring or
sparging wells can be used if they are screened
properly and meet other system specification
(e.g., well diameter, material compatibility).
• Extraction well material: Wells may be
constructed of polyvinyl chloride (PVC) or
stainless steel pipe with slotted PVC screen or
stainless steel wire wrapped screen through the
zone of contamination. PVC is cheaper and
lighter to install than stainless steel. The choice
of material should also consider compatibility
with potential future enhancements (e.g.,
thermal remediation, in situ chemical
oxidation).
• Vacuum monitoring points: If a monitoring
well network is not already in place within the
treatment area, an appropriate number and
distribution of vacuum monitoring points will
need to be installed. These points can be
screened in the same vertical extent as the
Soil Vapor Extraction (SVE) Technology
21
-------
extraction wells, with at least two wells being
present within the treatment zone and four
wells delineating the outside of the source area
or treatment zone.
• Air injection or air vent wells: Some SVE
systems are installed with air injection wells in
expected dead zone areas within the active
SVE area to minimize no flow zones. These
wells may either passively take in atmospheric
air or actively use forced air injection. Passive
injection is unlikely to provide sufficient flow
and, if rotation of extraction wells is
insufficient to mitigate dead zones, active
injection can be considered. However, the
system should be designed so that the air
injected into the system does not cause an
escape of VOCs to the atmosphere. Proper
design of the system can also prevent offsite
contamination from entering the area being
extracted (AFCEE, 2001; U.S. ACE, 2002).
Horizontal barriers to flow, such as surface covers
and the water table, also can create dead zones. A
detailed discussion on the operation of multiple
extraction wells and the mitigation of dead zones is
provided in Section 5.1. Additional well design
information can be found in U.S. ACE (2002).
4.4 System Sizing and Vapor Treatment
Pilot testing results are highly useful for determining
the total extraction rate, the applied vacuum required
to achieve the flow rate, and the size and type of full-
scale vapor treatment required. Oversizing vapor
treatment for short-term pilot testing (e.g., doubling
the estimated activated carbon requirement), as
opposed to long-term full-scale operation, may be a
good investment to avoid costly delays and multiple
field mobilizations to complete the field testing. Many
pilot tests fail to achieve the anticipated sweep of the
contaminated soil volume because the off-gas vapor
treatment system is exhausted prematurely (physically
or financially). For example, if granular activated
carbon (GAC) is employed for off-gas treatment, the
quantity available for pilot testing should anticipate
initial concentrations remaining near maximum values
until a full sweep of the active soil pore volume is
completed rather than anticipating a rapid exponential
decay of concentration. Full-scale design data will not
be adequate when a pilot test is terminated before
contaminant trends and mass transfer limitations can
be observed. Once the desired total extraction rate,
manifold vacuum, and trend for the initial vapor
extraction concentration behavior are established, the
blower type (e.g., variable frequency drive, positive
displacement, regenerative) can be selected and a
cost-benefit analysis performed to select a matching
off-gas vapor treatment system as described in U.S.
ACE (2002).
For off-gas treatment, activated carbon units are
relatively cost effective and GAC is used frequently at
low mass removal, diffusion-limited sites with VOC
removal rates up to approximately 5 lbs/day.
Activated carbon units usually require one-third of
the maintenance required for thermal systems and are
strongly preferred for chlorinated solvent sites. More
options are available for petroleum hydrocarbon sites
where extracted vapor concentrations are typically
much higher than solvent sites. GAC adsorption
capacity is based on VOC type, concentration, vapor
temperature, and relative humidity. Water vapor sorbs
to the GAC and leaves less capacity for the VOCs,
and GAC capacity decreases with increasing
temperature. Because the incoming air stream
temperature can be elevated due to pumping and
compression, there may be a need for off-gas cooling
prior to GAC adsorption. Other types of air emission
control devices used in SVE systems include catalytic
and thermal oxidation, incineration, cavitation, photo-
oxidation, ultraviolet oxidation, titanium dioxide,
internal combustion, biofilters, and direct discharge.
Additional information on air emission control
devices for SVE systems can be found in U.S. EPA
(2006).
22
Soil Vapor Extraction (SVE) Technology
-------
4.5 Site Access, System Selection, Layout,
Piping, and Instrumentation
Site access needs to be adequate to bring mobile
drilling rigs onsite for construction of SVE wells and
for trucks to deliver the equipment required for the
SVE system (e.g., vacuum blowers, vapor-liquid
separator, emission control devices, GAC canisters).
SVE systems vary in size and complexity depending
on the capacity of the system and the treatment
requirements for vapor (off-gas) and liquid (produced
water) effluents. SVE system components are
typically transported by vehicles ranging from trucks
to specifically adapted flatbed semitrailers; a proper
staging area for such vehicles can be incorporated
into design plans and plans for future site access for
system modifications.
A small to medium commercial-size SVE system (e.g.,
15 wells or less) requires about 100 ft2 of ground area
for the equipment skid and a height clearance of 10 ft.
This area does not include space for the vapor
treatment system, which is typically of similar area
and height. Space may also be needed for a forklift
truck to occasionally exchange skid-mounted GAC
canisters when regeneration is required. Large systems
with integrated vapor and liquid treatment systems
will need extra area based on vendor-specific
requirements. Power availability typically includes 3-
phase 230 V or 3-phase 480 V, which meet the most
common electrical service needs. For some SVE
applications, water may be required at the site. The
quantity of water needed is vendor- and site-specific
(AFCEE, 2001; Nyer et al., 2001; U.S. ACE, 2002)
but usually is small. Packaged SVE systems can
reduce design and construction costs, but can be less
effective if the system does not mesh well with site-
specific characteristics.
Major system design considerations are highlighted in
the following selection checklist by system
component:
Vacuum pump or blower equipment selection
• Use pilot test or modeling results to specify the
wellhead vacuum at individual extraction wells
Soil Vapor Extraction (SVE) Technology
necessary to achieve the desired extraction
rates.
• Perform detailed calculations for pressure
drops through piping connecting extraction
wells to the manifold entering the SVE
vacuum pump or blower.
• Select a vacuum pump or blower sufficient to
generate the vacuum at the manifold to achieve
the desired extraction rates.
• For low vacuum and high flow, use
regenerative blowers. For medium vacuum, use
positive displacement blowers. For high
vacuum resistance, use liquid ring pumps and
install a dilution valve.
• Typical well vacuums range from 10 to 60
inches H20.
• Higher air flow rates require larger equipment
size and increased power, and higher operation
and maintenance and emission control costs.
• At sites with a non-homogenous vadose zone,
higher airflow rates may not remediate the site
more quickly.
Air/water separator and air filter
• Soil gas extracted by the SVE wells first enters
an air/water separator to remove moisture,
followed by an air filter to remove particulates,
and then are usually pumped to the air
treatment system.
• Moisture and particulate removal devices
protect system equipment.
Off-gas treatment technology
• Use pilot test data to determine whether the
SVE system will likely exceed the Clean Air
Act (CAA) thresholds for air treatment.
• Determine whether an air treatment system
will be required or selected based on best
management practices.
• Carbon adsorption using GAC is the most
frequently implemented technology, especially
for smaller sites with lower VOC
concentrations.
-------
• Conduct a cost analysis to consider all options.
• At sites with higher levels of contamination,
larger scale equipment or more than one
technology may be necessary.
Flow, vacuum, and concentration measurement
devices
• Install flow, vacuum, and concentration
measurement devices in the appropriate
locations to monitor that the system is
operating correctly and evaluate its overall
effectiveness over different temporal scales.
Piping
• High density polyethylene (HDPE) is
recommended for subgrade conveyance piping.
It is easy to install, as it comes in large spools
that can be rolled out along the trenches. It can
be "swept" to make turns and bends, which
avoids the use of elbows that create frictional
losses. HDPE is also fuse welded rather than
using glued fittings that tend to crack or break
subgrade due to temperature fluctuations.
• Schedule 80 PVC can be used for above-grade
piping on the vacuum side of the extraction
blower. It is ideal for the manifold as cutting
and gluing PVC pipe and fittings is generally
much faster than assembling threaded steel
pipe and fittings
• Blowers generally heat the air to a temperature
greater than PVC can withstand, therefore
metal pipe should be used for the discharge
side of the blower. Generally, carbon steel or
aluminum can be used, but the metal should be
compatible with the constituents in the
discharged air stream.
• Piping to extraction wells should include
sumps or similar features to remove water
from the piping, especially at low spots where
water can accumulate and restrict air flow.
• Make sure to test underground piping for
vacuum, pressure, and heat resistance prior to
implementation.
Sensory considerations
• In population-dense areas, minimize noise by
placing SVE equipment in a better soundproof
building and consider visual appeal when
selecting equipment or housing that will be
viewable by the public.
• Oversized equipment is likely to result in
excessive noise.
• Heat exchangers can be noisy, so allow for hot
weather operation when evaluating noise
impacts.
4.6 Health and Safety Issues
Given the uncertainties and potential exposure to
explosive or toxic vapors, it is critical to define and
address health and safety issues, along with regulatory
concerns and objectives, prior to implementing and
operating an SVE system. VOC contaminants
typically present in SVE off-gases are usually
hazardous because of their toxicity, ignitability, or
other reasons (U.S. ACE, 2002). Therefore, practices
such as proper selection of equipment components,
monitoring of system off-gas, and evaluating what
off-gas treatment is needed should be done early to
ensure safety of personnel and the facilities.
Blowers and other electrical motor driven equipment
(including wiring) must be designed and constructed
in accordance with applicable National Fire
Protection Association (NFPA) code, with proper
consideration given to environmental conditions such
as moisture, dirt, corrosive agents, and hazardous area
classification. Hazardous area classification should
follow practices outlined in the applicable NFPA
code, and taking into consideration process
equipment in the area, characteristics of hazardous
liquids/gases, the amount of ventilation in the area,
and the presence of equipment such as piping with
valves, fittings, flanges, or meters. The classification
of the area will determine the potential need for
explosion-proof motors or other system components
(U.S. ACE, 2002).
Monitoring of system off-gases may be conducted
using multiple approaches to evaluate the potential
Soil Vapor Extraction (SVE) Technology
-------
exposure for explosive or toxic vapor risks, as well as
permit requirements for CAA rules. Monitoring may
be conducted using an explosimeter (or combustible
gas indicator) or appropriate VOC monitoring
equipment, such as a flame or photoionization
detector, as well as intermittent sample collection and
analysis for specific VOC compounds of concern
(U.S. ACE, 2002). When using on-site monitoring
equipment, care should be taken to ensure that
equipment is properly calibrated and all operating
instructions and potential limitations as indicated by
the manufacturer are understood. Sampling and
laboratory analysis of system off-gases can also aid in
determining contaminant levels for safety and
regulatory (permitting) purposes.
5 OPERATIONS, MONITORING, AND
PERFORMANCE EVALUATION
The design parameters developed during site
characterization and pilot testing form the initial
conceptual site model for evaluating SVE
performance. This model evolves and is refined as
field operations progress and more data become
available. In particular, depending upon the range and
frequency of monitoring and data collection,
persistent sources that govern attainment of remedial
goals become increasingly visible.
At the start of operations, a "cloud" of somewhat
uniform contaminant vapors usually envelops the
contaminated soil volume. Referring to Figure 4, this
cloud hides the sources (i.e., the primary
accumulations of contaminant mass resulting from
the release) because most of the soil gas monitoring
points in the figure would initially yield soil gas with
high VOC concentrations.
With the new dataset from the start of SVE, a review
of geologic logs may substantiate observations about
source areas (e.g., identification of a thin slit or clay
layer considered unimportant before startup). As the
high-concentration cloud is swept away by initial SVE
operation, the locations of sources become better
defined as illustrated in Figure 6 where only the
monitoring points within or near the source lack a
significant concentration decay. Figure 6 also shows
two scenarios for persistent sources that may yield
nearly identical initial concentration decays in the
monitoring points, as most practical monitoring
networks are relatively sparse.
As described in this section, analysis of changes in the
mass extraction rate over time and vapor
concentrations during periods of rebound provide
data for differentiating and characterizing the VOC
sources. In this manner, SVE provides a wealth of
data to generate a more accurate site model of
contaminant transport that is directly applicable to
assessing closure of a site. The physics of mass
transfer during SVE are identical to those of transport
during ambient conditions, when the vacuum-induced
flow through the vadose zone is not present. Hence,
the SVE mass extraction rate provides a worst-case
estimate for a site-averaged contaminant mass flux
toward the surface for vapor intrusion or downward
mass flux entering groundwater. However, as
discussed in this section, inefficient operation of
multiple extraction wells or the existence of vertical
barriers can complicate such assessments.
5.1 Operations and Monitoring
Details for operating, maintaining, and monitoring
SVE systems are available from numerous guidance
documents (e.g., U.S. ACE, 2002). Beyond
maintaining and troubleshooting vapor extraction and
treatment system operation (e.g., troubleshooting a
frequent shutdown of a thermal oxidizer), a common
operating issue is the production of water from vapor
extraction wells. Water that collects in vapor
conveyance piping can restrict flow from the
subsurface and the introduction of water to a vapor
treatment system reduces treatment efficiency. This
problem tends to be seasonal, occurring with rain and
cold weather (an opportune time to schedule rebound
evaluations). Planning for the production of water
and being prepared to predict and handle the excess
water increases the efficiency of the SVE operation.
Soil Vapor Extraction (SVE) Technology
25
-------
Upper Boundary Is.'Site
Specific
Diffusive Source
Advective Soil
Groundwater C
Aquitard
Q AC
1
I I.
; Upper Boundary-is Site Speeific- |
Diffusive Sources
Advective Soil
Groundwater
Aquitard:
Figure 6. Conceptualized scenarios for diffusion-limited mass transfer during SVE
As described in Section 4.5, the piping to extraction
wells should include sumps or similar access for
removing water from the piping. Piping in a trench,
for example, under a road, provides a low spot for the
accumulation of water that must be removed through
a sump to prevent the severe restriction of gas flow in
the piping.
Other waste streams generated by SVE include those
from VOC vapor treatment (e.g., spent GAC),
wastewater produced from the SVE system, and
potentially soil cuttings and personal protective
equipment. Compared to ex situ treatment
technologies and other soil remedies such as soil
excavation, SVE systems generally produce less waste
(U.S. EPA, 2010).
Standard remedial system monitoring for SVE
includes temporal logging of the following
parameters:
• Vacuum/pressure in monitoring and extraction
wells
• Gas flow rate in extraction wells and manifolds
• VOC concentrations in monitoring and
extraction wells (field instruments such as
photoionization detectors, flame ionization
detectors, portable gas chromatographs; fixed
laboratory analyses)
• Vapor treatment system compliance data (flow,
destruction/reduction efficiency)
• Utility usage
• Water levels in the knockout tank
• Waste quantities for disposal
• Recordkeeping of system data and a log of
SVE on/off times
• Recordkeeping of system maintenance (e.g., oil
changes, water transfer from knockout tank).
Frequency of sampling VOC concentrations using
field measurements can range from regular site visits
(e.g., weekly) to a continuous data logger that
electronically reports soil gas flow and VOC
concentration measurements to a remote remedial
technician. VOC samples are typically collected for
laboratory analysis monthly to satisfy permit vapor
treatment requirements although such sampling can
be extended to individual wells or monitoring points
of interest for correlating field instrumentation
readings. Total mass of contaminants removed by the
26
il Vapor Extraction (SVE) Technology
-------
SVE system can be calculated by multiplying the flow
rate by VOC concentrations over time. The trends in
mass removal from the site and individual wells are
primary components of both system optimization and
site closeout.
VOC concentrations at monitoring points are also
typically measured by sampling and fixed laboratory
analysis at least quarterly. The sampling program can
include monitoring points located within the source
zone and outside of the contaminated area (perimeter
of compliance wells). Depending on the vertical
extent of the plume, soil gas monitoring points may
be located at different depths in well "nests"
(AFCEE, 2001; Matzke et al., 2010; Suthersan,
1999a).
5.2 Data Evaluation
Common SVE operation practice is to generate a
monthly or quarterly report from operating and
monitoring activities that includes system operating
time, maintenance performed, operational problems
encountered and solutions, ongoing operational
issues, data collected and graphed as a function of
time, periodic and cumulative masses removed, and
plans for the next reporting period. Assessments of
progress toward cleanup goals can be made annually
or as needed, and can be general statements until
VOC concentrations are near or below a
predetermined cleanup threshold or appear to plateau
at levels well above a threshold. A straightforward,
ongoing data evaluation that refines the conceptual
site model and ultimately serves to support SVE
cessation and site closure is ideal. Such data
evaluations can be included in a progress report (as
appropriate) and can include a "forecast" of system
performance for the next reporting period to help
identify anomalies or deviations from the conceptual
site model that may require attention.
Observed trends in extracted soil gas VOC
concentrations, monitored soil gas VOC
concentrations, and VOC mass removal can be
evaluated on a site-averaged scale by comparing the
trends with anticipated trends representative of
Soil Vapor Extraction (SVE) Technology
subsurface conditions. For example, extracted soil gas
VOC concentrations that plateau at elevated levels
would be consistent with the existence of a layer of
gasoline floating on the water table below the vadose
zone, as other site data may suggest. Comparing these
concentrations with equilibrium concentrations (i.e.,
VOC concentrations just above gasoline) also
provides a rough estimate for the extraction
"efficiency" governed by mass transfer between the
fuel and flowing air. A detailed discussion of such a
scenario and its evaluation are provided by Johnson et
al. (1990a). SVE can also cause compositional changes
in a petroleum product as a result of preferential
volatilization, as described by Carroll et al. (2013).
For SVE at chlorinated solvent sites, the most
common mass transfer constraint is between
permeable soils where air flows (advective soils) and
adjacent lesser permeable soils (diffusive soils) that
hold contaminants dissolved in the porewater (see
Figure 6). The mass transfer limitations of diffusive
soils do not appear until the initial sweep of
permeable, advective soils is complete. In its simplest
form, this mass transfer limitation can be modeled by
a bulk mass transfer coefficient, ad (Goltz and Oxley,
1994), averaged over the volume of the source zone:
dCd
Rd ~~^r = ~ad(Cd — Ca)
Use of the bulk mass transfer coefficient is a
simplistic method to quantify the physics of diffusion
in the source zone. The first order estimate for the
cleanup time provided in Section 2 is based on linear
diffusion and the same physics yield a first order
estimate for the mass transfer coefficient based on the
same parameters (U.S. ACE, 2002; U.S. EPA, 2001):
TdDairK2
The impact of mass transfer on the extracted
concentration, Ca, equivalent to a well-mixed average
concentration in the advective soils, is described by
U.S. ACE (2002, Appendix F),
-------
Rr
dCg
dt
+ ad
VsoilVCL ~ S)fa
cn
(f )W,-C„)
The site-averaged vapor concentration in the source
(diffusive) soil is represented by Cd. The first term on
the right describes the sweep of advective soils while
the second term describes mass transfer contributions
from the source soils. The retardation coefficient and
other parameters are as defined previously. Note this
simplistic representation of the vadose zone soils
requires specifying the soil volume as either advective
or diffusive expressed as a fraction of one (fa + fd =
I). Geologic profiles or other vertical logging
techniques such as PneuLog can be used to estimate
the fractions.
In practice, an apparently uniform site likely has a
maximum advective fraction of roughly 0.8 as
moisture content varies. For a predominantly low
permeability site, the advective fraction may be only
0.1 or 0.2 and SVE may not be applicable or will
require years of operation. Given initial vapor
concentrations for the soils roughly equal to the peak
vapor concentration at the start of SVE (Cd0 «
Q[,o = c o), a simple solution to describe the mass
transfer limited behavior during extraction that is
provided in U.S. ACE (2002, Appendix F). The
solution is easily fit to field data to forecast future
SVE performance and track progress. During the
initial extraction period, mass transfer provides little
contribution until the extracted concentration
decreases significantly as the advective soil volume is
swept. For this initial extraction period, the extraction
concentration is approximately:
C,
a,initial
C0 exp
Qt
Vsoii(p(l - S)faRa
Fitting extracted concentration data with this
expression using measured concentrations, flow rate,
and soil properties provides an estimate for the
volume of advective soil (Ja I Tsod). At later times, the
extracted concentration changes very slowly and can
be maintained much less than the average diffusive
source concentration. Under this assumption, the
diffusive concentration is approximately (Goltz and
Oxley, 1994):
C,
d.late
C0 exp
adt
Rr
After substituting this expression, solving the
governing equation yields for later, diffusion-limited
conditions is approximately:
C,
a.late
C0ad
Vsoil
-------
_1000
>
E
Q.
Q.
C
o
Ca,initial C^exp
0)
u
c
O
U
o
Q.
2
100
ja.late
10
e~ o
Qt
Vsoii
-------
The VOC mass transfer coefficient at the water table
interface between the bottom of the vadose zone and
water table surface is difficult to predict but can be
estimated near the end of SVE operations when
contaminated groundwater becomes the primary
source of VOC mass and is a potential source for
vapor intrusion after SVE ceases. This topic is
explored in Appendix B of this paper.
5.3 System Optimization
AFCEE (2001) recommends review of SVE
operations from three perspectives:
• Evaluation and optimization of the operation
of the existing system with the goal of
maximizing the rate of contaminant mass
removal to achieve the greatest reductions in
contaminant concentrations and to minimize
operating costs.
• Re-evaluation of the system components (e.g.,
wells, blowers, off-gas treatment system) to
determine if changing or adding to the system
will improve performance or to determine
whether a wholly new technology is necessary.
• Re-examination of the remedial goals in light
of new regulations, risk thresholds, or changes
in exposure scenarios.
With respect to maximizing the rate of contaminant
mass removal, the guiding principle is to extract soil
gas as close as possible to the center of remaining
sources of mass with the fewest number of extraction
wells that site conditions will allow. This method
minimizes stagnation zones and the introduction of
dilution air.
When operating multiple extraction wells, the
operating wells may be rotated to assess the impact of
stagnation zones and source locations. A common
misstep in attempting to optimize the configuration
of extraction wells is to rotate wells too often, before
the site has re-equilibrated to the changes in flow. The
re-equilibration time is, at a minimum, a single pore
volume exchange rate that may span days to weeks,
negating the utility of measuring changes over hours
or a few days. A legitimate reason to add or subtract
one or two extraction wells on a monthly or quarterly
basis is to address the potential of a stagnation zone
overlapping a source zone and prolonging
remediation. Extraction wells close to a source zone
shown to be near cleanup goals can be turned off to
focus flow on more persistent source zones and to
evaluate rebound effects. However, the potential for
stagnation zones when operating multiple extraction
wells requires detailed evaluation to assess whether a
persistent concentration is the result of stagnation
rather than a diffusive source. Plots of contaminant
mass extraction rates from individual wells over time
and application of the site-wide two-region mass
transfer model described above can help identify the
benefits or inefficiencies of various combinations of
the extraction wells.
Stagnation zones are readily calculated from potential
flow models, identical to those used to calculate
capture zones in groundwater with zero groundwater
velocity. Consider a system of three extraction wells,
screened over identical depth intervals (Figure 8).
When all three wells are operated, two stagnation
zones exist at the elevation of the extraction screens
as indicated on the left-hand side of the figure where
the blue lines represent streamlines of air flow toward
the extraction wells.
Figure 8 is oriented to show a site plan. If the source
zone is predominantly in the middle of the three
wells, very inefficient flushing occurs as most of the
flow and flushing occurs outside the triangle of wells
as illustrated by the low flow near the vapor
monitoring point. Hence, more efficient approaches,
assuming a single well has insufficient radius of
effectiveness, is to rotate operation among two of the
wells at a time to mitigate the stagnation zone impact
as illustrated on the right-hand side of Figure 8 and
the much higher flushing rate through the probe
location. Another approach would be to rotate among
the wells operating one at a time to eliminate all
stagnation zones during operation.
30
Soil Vapor Extraction (SVE) Technology
-------
To illustrate the impact of stagnation zones, the vapor
concentration history at the vapor monitoring point
shown in Figure 8 is provided in Figure 9. The
vapor point has a nest of three probes (5, 25, and 35
ft below the surface) and the extraction wells are
screened from 35 to 50 ft below the surface. The
initial operation utilized the two-well configuration
shown on the right-hand side of Figure 8 and
flushing reduced the vapor concentration rapidly
("Outside Stagnation") at all three depths as
illustrated in Figure 9. However, the third well was
added and flow occurred at an elevation of 35 ft
below surface as illustrated on the left side of Figure
8. The deep probe increased in concentration ("Inside
Stagnation") to near initial levels. The shallower
probes were flushed in both configurations primarily
by vertical flow downward from the surface to the.
deeper extraction wells. Without these concentration
histories, if all three wells were operated without
rotation, the probe at 35 ft would falsely suggest a
persistent source area.
A second approach to mitigating stagnation zones is
to utilize vent wells. However, passive injection is
unlikely to provide sufficient flow at most sites. If
rotation of extraction wells is insufficient for flushing
persistent source zones, active air injection can be
considered. However, the system should be designed
so that any air injected into the system does not
pressurize the subsurface and cause VOCs to escape
to the atmosphere. The flow rate of injected air is
expected to be relatively low compared to the gas
extraction rate in the closest extraction wells. The
most common use of air injection is to introduce
oxygen into the vadose zone when aerobic
degradation is oxygen limited.
The previous discussion involves horizontal
stagnation zones but an often-overlooked source of
stagnation zones are vertical barriers such as surface
covers and the water table. If contamination exists
under a concrete slab, deeper SYB can create a
significant vacuum beneath the slab, but flow will not
occur and the VOC vapors will persist. This condition
can be alleviated with; the installation of vents just
beneath the slab. Similar conditions exist at the
capillary fringe; no upward flow occurs for flushing
unless air sparging is applied in the groundwater.
Stagnant zone
'O Former stagnantzone
• Active extraction well
O Dormantwell
~ VaporMonitoringPoint
Figure 8. Conceptual streamlines and stagnation zones during SVE
Soil Vapor Extraction (SVE) Tech nolo
m
U
31
-------
Inside
Stagnation
Stagnation
1,000,000
CO
CUD
3 100,000
c
o
'¦M
ro
Outside
10,000
1,000
4/1/16 4/11/16 4/21/16 5/1/16 5/11/16 5/21/16 5/31/16
Figure 9. Vapor concentration histories during SVE well rotation
System optimization also involves re-evaluating
system components such as the vapor treatment
system. The most common consideration at
hydrocarbon sites is transitioning from a thermal to a
catalytic oxidizer to save on supplemental fuel and
then from an oxidizer to GAC to treat low VOC
concentrations in off-gas. These transitions are
discussed in detail in U.S. ACE (2002) and U.S. EPA
(2006).
A second area of system optimization is the use of
new extraction or monitoring wells. These can be
installed after drilling into suspected persistent source
areas during a supplemental characterization effort or
after confirmation borings reveal a previously
unknown source. New extraction wells may also be
justified if existing extraction wells have very long
screen intervals that draw primarily dilution air from
shallow soils while diminishing flow through deeper,
contaminated soils. The pressure drop associated with
gas flow through the well screen and upward to the
surface is routinely neglected; however, for typical
extraction rates the pressure drop along the screen as
flow enters the well can limit the utility of installing
longer screens. Vertical profiling of gas flow in
extraction wells with PneuLog rarely measures
uniform flow along the well screen (Lloyd "Bo"
Stewart, pers. comm. 2018)).
For sites that are capped, insufficient flow may occur
in the shallow contaminated soil above the top of the
screen interval as identified by monitoring points or
vertical profiling. In this case, the cap could be
removed or perforated, or passive vent wells can be
installed. Extraction wells that are no longer in service
are also good candidates to act as passive or active
vent wells to mitigate stagnation in source areas.
5.4 System Transitions and Vapor Intrusion
Mitigation
When evaluating SVE performance for completion or
a transition to another technology or objective,
multiple lines of evidence are considered. A first step
is to perform rebound testing. Rebound testing can
Soil Vapor Extraction (SVE) Technology
-------
identify persistent source zones to focus future
extraction configurations, assess progress toward
completion, and refine the conceptual site model. The
two-domain SVE model equations presented
previously are applicable with the gas extraction rate
set to zero for a defined period. When the extraction
is restarted, observed increases in extracted VOC
concentration and mass removal rate are indicative of
the remaining sources (U.S. ACE, 2002) as are the
rate and magnitude of rebound in contaminant
concentrations at monitoring points. The
interpretation and utility of periodic rebound testing
is also described in detail in Brusseau et al. (2010,
2015).
An example SVE rebound data set is illustrated in
Figure 5. The pilot test data plotted in Figure 3
suggested a small persistent mass of TCE existed in a
low permeability soil horizon near a building. The
mass was deemed to be small but sufficient to pose a
vapor intrusion risk. After the initial 3-day pilot test, a
17-day pause (first rebound period) was followed by
extended extraction over 60 days and then a second
rebound period as shown in Figure 5. The second
rebound period was evaluated with a 6-hour
extraction period to mimic extraction and sampling
performed at the start of both the pilot test and the
extended extraction period. The model described in
the previous section was fit to the concentration data
and provided an excellent fit throughout both the
extraction and subsequent rebound periods. The
results show the extraction achieved a reduction in
TCE mass and concentration of more than 90%
within the source zone. The results from that site also
illustrate the efficacy of using an SVE well proximate
to a contaminant vapor source in the vadose zone for
vapor intrusion mitigation. The blower could be
operated periodically, as indicated by monitoring data,
for short periods to sweep the zone or a smaller
blower could be operated continuously to prevent
vapors from approaching the building subslab.
Conceptually, vapor intrusion mitigation—where a
negative pressure is maintained below a subslab—is a
special case of SVE. However, the performance
Soil Vapor Extraction (SVE) Technology
objective differs. Preventing vadose zone vapors from
entering a building is not the same as extracting
contaminant mass to attain cleanup, although mass
removal during SVE may achieve the mitigation
objective. The use of SVE within the deeper vadose
zone for vapor intrusion mitigation emphasizes
capturing vapors prior to approaching the subslab
rather than the maintenance of subslab
depressurization. If contaminant vapors emanate
from sources deeper in the vadose zone, maintenance
of subslab depressurization can result in contaminants
being drawn toward the subslab for capture.
In evaluating SVE performance, operators may
conclude that SVE alone is inadequate to achieve
vadose zone cleanup objectives in an acceptable
timeframe. However, the location and rate of
contaminant mass removal may successfully meet
mitigation objectives with no modification to nearby
buildings. The efficacy of this type of application or
even the initial use of SVE solely for vapor intrusion
mitigation is evolving (Lutes et al., 2017; Schumacher
et al., 2017, Truesdale et al., 2016).
The design of an SVE system solely for vapor
intrusion mitigation or the transition of an existing
system to meet these alternative objectives follows the
identical design concepts as those described in
Section 4,
• Installation and operation of a pilot extraction
screen proximate to sources of contaminant
vapors to determine site properties and the
radius of effective contaminant vapor
reduction (subslab vacuum responses may be
negligible even though concentration reduction
is achieved)
• Evaluation of vapor concentration rebound to
assess the frequency of sweeping the vadose
zone pore volume to maintain a relatively low
vapor concentration (compared to the source
concentration)
• Assessment of the surface area to be protected
and the total gas extraction rate requirement to
achieve the frequency of pore volume sweeps
-------
• Placement of extraction screens within or near
the vapor sources
• Selection of vapor treatment and system
installation (based on a cost analysis, the
preferred approach may be a fixed system or
periodic use of a mobile system)
• Operation of the system with particular
attention to minimizing the impact of
stagnation zones (substantial subslab vacuums
may be created with no effective flow through
the zone; vent wells may be beneficial).
If transitioning an existing system, the infrastructure
is likely in place and the design will consist of
determining the optimal extraction strategy.
The design and performance evaluation criteria in this
EIP were applied to a site with contaminated
groundwater acting as the sole source of vapors
migrating upward through the vadose zone as
described in Section 9.2 and Appendix B. The
evaluation provides a robust basis to determine the
most cost-effective, site-specific implementation of
vapor extraction in a former groundwater monitoring
well to meet vapor intrusion mitigation objectives for
a downgradient dissolved plume under buildings.
6 SYSTEM SHUTDOWN AND SITE
CLOSURE
As an SVE system begins to show evidence of the
diminishing contaminant removal rate illustrated in
previous sections, SVE performance needs to be
evaluated against the site-specific performance
objectives. In particular, this evaluation may
determine whether the system can be terminated,
optimized, enhanced, or transitioned to another
technology to replace or augment SVE.
Recent guidance from the U.S. DOE (2013)
specifically addresses the elements of this type of
performance assessment. The guidance summarized
and built on previous guidance for SVE design,
operation, optimization, and closure from the U.S.
EPA (2001), U.S. ACE (2002), and AFCEE (2001).
The framework and steps for making SVE remedial
decisions described in the guidance are applicable to
all SVE sites. The framework is summarized in this
section. A spreadsheet model (SVEET) associated
with the U.S. DOE guidance is available to the public
and treats both source scenarios illustrated in Figure
6 as a single monolithic layer in the vadose zone that
does not decay.
Periodic SVE shutdowns to evaluate rebound will
provide information necessary to assess progress
toward remedial goals of absolute vapor
concentrations or mass fluxes. An initial rebound test
performed after the initial decay in extracted
concentration will provide an indicator of source
zones and a baseline for later comparison. The
periodic evaluations illustrated in Figure 5
demonstrate a reduction in average source
concentration of more than one order of magnitude.
Over time, the cases of local rebound concentrations
above applicable cleanup criteria are expected to
decrease. Rebound testing will help identify locations
where persistent high VOC levels warrant additional
investigation for potential augmentation or locally
aggressive treatment with an alternative technology
(e.g., chemical oxidation, electrical resistance heating).
Additionally, transport and exposure models can be
used with rebound test results to estimate whether
system closure may pose a risk to potential receptors.
Methods for evaluating the rebound data are provided
in U.S. ACE (2002) as summarized in Section 5.2.
Additional details on the interpretation and use of
periodic rebound testing are described by Brusseau et
al. (2010, 2015). The key results from the rebound
testing are reductions in equilibrium VOC vapor
concentrations and VOC mass flux rates from the
residual VOC sources. The use of these results as
inputs for fate and transport modeling after SVE
ceases is described at the end of this section.
34
Soil Vapor Extraction (SVE) Technology
-------
6.1 Framework for Assessing SVE Termination
or Technology Transition
The U.S. DOE guidance (2013) presents a logic
process for decision makers to determine if (1) the
site is ready for SVE termination and closure, (2) the
existing SVE system should be optimized to improve
performance, or (3) alternative technologies should be
considered to meet remediation goals. The primary
focus is to identify if and when SVE can be
terminated based on assessments of the residual VOC
mass remaining in the vadose zone, the expected
VOC mass flux towards groundwater and towards the
land surface (vapor intrusion), and the longevity of
the source under natural attenuation. If the
remediation goal is unlikely to be attained through
optimization or a reasonable period of continued
SVE, then potential alternative approaches and
enhancements can be considered as described in
Section 8. These alternatives are introduced in the
context of enhancing, augmenting, or replacing SVE
technology at specific locations at a site.
Before assessing SVE termination or transition, the
components and phases of the SVE operation can be
reviewed for completeness and appropriateness as
follows:
Site characterization and conceptual site model
• Adequate characterization of the original
source mass and location
• Determination of LNAPL or DNAPL
presence, age, amount, etc.
• Adequate characterization of the subsurface
soils/geology including heterogeneity,
preferential flow paths, any vapor confining
units, etc.
SVE system design
• Appropriate well spacing and flows given the
characterization
• Appropriate monitoring well and probe
distribution (horizontal and vertical)
• Adequate monitoring instrumentation and data
collection methods
Operations and monitoring
• Appropriate and adequate data collection
frequency to define trends
• Avoided prolonged stagnation zones and
performed rebound tests
• Relatively accurate and comparable calculation
of mass removal rates to estimate original
VOC mass
• Diminishing mass removal rates, caused by
mass depletion and mass transfer constraints
and not a result of poor design or
uncharacterized sources
Performance assessment
• Diminishing mass removal rate that is
consistent with diffusion rate-limited vapor
transport
• Models of vapor diffusion predict
commensurate values to the observed mass
removal rate during operations using the
maximum observed concentrations during
rebound and a characteristic length scale for
diffusion determined from the site geology
• Relatively accurate forecasts of system
performance using models commensurate with
those of Section 5.2 with only modest
adjustment to fitted parameters over time
• Rebound concentrations that can be correlated
to VOC concentrations in fine-grained units or
underlying contaminated groundwater
Based on the analysis of the SVE phases described
above, the determination of appropriate site goals
described in Section 2.3, and likely impacts from
remaining VOC sources, the following three-step
decision logic summarizes the determination of
appropriate future actions at a site detailed in Figures
10a and 10b (see U.S. DOE, 2013 for the chart):
1. If SVE is shut down, will remediation goals be
met, based on cleanup levels and lines of
evidence? Could residual contamination cause
groundwater or indoor air goals to be
exceeded? Could residual contamination
necessitate the installation and operation of a
Soil Vapor Extraction (SVE) Technology
35
-------
vapor intrusion mitigation system for indoor
air exposures? If the answer to these
questions is no, proceed to Step 2.
2. Has the existing SV.P. system been optimized
to remove subsurface contaminant mass more
effectively and allow for ideal closure
conditions? Is significant VOC mass
remaining in permeable zones? Is SVE
reducing VOC contaminant source strength?
Will the current VOC removal rate allow for
remediation goals to be met within a
reasonable timeframe? If reasonable
optimization options for SVE have been
exhausted, proceed to Step 3.
3. If optimization of the current SVE system is
complete and performance objectives for
active SVE have been met, SVE termination
and possibly site closure can proceed if
quantification of remaining sources and
impacts are acceptable. U.S. DOE guidance
(2013) provides background and a
recommended approach for this
quantification building on previous work.
Methods for quantifying residual sources and
impacts on groundwater and vapor intrusion
are described in Section 6.2.
I f the assessments in Section 6.2 of remaining sources
indicate attaining performance objectives is not
feasible with conventional SVE alone, SVE
enhancement or augmentation may be possible as
described m Section 8.
Begin the process of information
collection and data evaluation to
support SVE remedy decisions
Environmental
impact pathways
Cumulative risk
Site remediation
goals
Revisitthe Conceptual Site
Model (CSM) to reflect new
monitoring/operationsdata
o
Assess the environmental
impact pathways & regulatory
compliance context
Quantify the impacts of
remaining source material
Apply the Decision Approach
for SVE optimization,
termination, or transition
/ Consider Aspects of the CSM
C.
What is the nature of
the remaining sources?
Will the remaining
contamination cause
groundwater goals to
be exceeded?
Consider SVE optimization,
SVE enhancement, or
other remedial options
Seek site closure, pending
evaluation of vapor
intrusion (if relevant)
Collect Recent Data
What are the dominant
transport processes?
What transformation
processes are relevant?
Site data
SVE system data
What receptors (human
or ecological) exist?
Are there complicating
factors?
Information on
contaminant sources
Site type
categorization
Quantify vadose zone
source (strength / location)
Estimate impact to ground-
water (Type I & II sites)
Estimate impact to
vapor intrusion
Estimate impact of source
decay and attenuation
[Continue to Figure 10b]
Source: U.S. DOE, 2013
Figure 10a. Decision flowchart for SVE system optimization, transition., and closure
36
Soil Vapor Extraction (SVE) Technology
-------
[Continued from Figure 10a]
Assess potential to
optimizethe SVE system
Is there accessible
contaminant mass in
permeable units in the
vadosezone?
Yes
Will SVE reduce
contamination enough
to reach remediation goals
in a reasonable time?
No
Consider optimization
approaches
Are optimization
^approaches not applicable
.and/ortoo uncertain?
No
Applythe optimization
approach
Continue
operating SVE
system and
assess closure
at a later date
Consider SVE
enhancements or
alternative remediation
technologies
Control mass
flux from
vadosezone
source to
groundwater
Control
mass flux
to the
ground
surface
Apply a
more
aggressive
remediation
technology
Re-do the full evaluation of SVE for optimization,
termination, or transition at a later date
Source: U.S. DOE, 2013
Figure 10b. Decision flowchart for SVE system optimization, transition, and closure
6.2 Summary of Methods for Evaluating
Attainment of SVE Endpoints
SVE termination and site closure criteria must
consider general YOC mass transfer limitations and
site-specific performance limitations of SVE in the
context of remedial action objectives (RAOs) based
on potential human health or environmental risks
(Switzer, 2004; U.S. ACE, 2002, U.S. DOE, 2013),
The: most common method for determining SVE site
closure criteria is the assessment of potential aquifer
degradation from the transport of contaminants from
the vadose zone to the groundwater. JL typical
requirement is the attainment of specified soil
concentrations or vapor concentrations based on the
premise that mass flux from the vadose zone to
groundwater not result in levels exceeding MCLs (or
other regulatory limits or RAOs).
The: simplest interpretation of this requirement is the
attainment of equilibrated rebound concentrations
that do not exceed MCL-equivalent vapor
Soil Vapor Extraction (SVE) Technology
concentrations (i.e., vadose zone porewater at or
below MCL), Bor example, TCE at about 5 ppb in
water equilibrates to about 350 ppb in soil vapor,
providing a target value independent of SVE trends.
However, this approach could result in soil
remediation efforts beyond what is necessary for
protection of groundwater or other remediation goals.
For these reasons, the evaluation of aquifer
degradation could also include consideration of mass
fluxes after SVE ceases and attenuation processes that
reduce the concentration of VOCs in groundwater
and the vadose zone over time and with distance
from the source.
These more complex evaluations involve modeling
the fate and transport of residual contaminants in die
vadose zone into underlying groundwater with a
cleanup goal in the vadose zone back-calculated to a
residual source concentration in the vadose zone that
does not result in a groundwater exceedance of MCL
at a specified aquifer location. One of the first such
-------
modeling assessments resulted in the U.S. EPA
production of the widely used, but overly simplistic,
VLEACH model (Rosenbloom et al., 1993).2 Of note,
SVE operations continue as of the date of this
publication at the first example site for VLEACH
from 1993. Updated models for assessing the impact
to groundwater are described in Section 6.3.
Vapor intrusion into buildings, especially ground-level
and sub-ground floors, is of particular importance for
VOCs in the subsurface. The current version of
EPA's Technical Guide for Assessing and Mitigating the
I rapor Intrusion Pathway from Subsurface I ^apor Sources to
Indoor Air (U.S. EPA, 2015) provides current technical
and policy recommendations on determining if the
vapor intrusion pathway poses an unacceptable risk to
human health at cleanup sites. This guide presents a
framework for assessing the potential for harmful
concentrations of VOCs in buildings.
The EPA guide leads the decision maker through an
evaluation process based on multiple lines of evidence
of the potential for harmful vapor intrusion. If the
soil gas VOC concentrations exceed indoor air vapor
intrusion screening levels, additional sampling may be
warranted. If VOC concentration in the indoor air is
above action levels, and the source is likely
subsurface, then vapor intrusion mitigation may be
warranted. Because the vapor intrusion exposure
pathway can be strongly influenced by the specific
structures and conditions in the subsurface, at the
ground surface, and within buildings, it can be
difficult to model and assess accurately. Existing
vapor intrusion analyses, therefore, often rely on
surface-based measurements and analyses as lines of
evidence. However, this is an area of active research
and recent tools have been published by U.S. DOE
(2016) for the assessment of site-specific vapor
intrusion based on vapor-phase diffusion from
residual vadose zone source mass.
6.3 Models for Evaluating Impacts of Residual
Mass in the Vadose Zone
Termination of SVE requires an assessment of
impacts from residual mass in the vadose zone as
described in Section 6.2. Such assessments are based
on the quantification of the VOC mass flux into
underlying groundwater or into buildings on the
surface. The flux into groundwater yields a calculated
VOC concentration in groundwater using a mixing
approach that varies among models and generally
ignores existing groundwater contamination. The
calculated groundwater concentration is the primary
method to assess groundwater impact.
Mathematical calculation of vapor intrusion impacts is
less developed; however, the same concepts of mass
flux across a vertical barrier (e.g., concrete slab)
followed by mixing inside a surface enclosure (e.g.,
basement) undergoing an exchange with the
surrounding atmosphere suggest the modeling can be
very similar. In general, lateral transport is
conservatively neglected, leaving vertical transport
from residual sources in the vadose with upper
(surface) and lower (water table) boundary conditions
as the basis for mathematical modeling of vapor
concentrations and fluxes.
As mentioned above, an early mathematical model
available from U.S. EPA is VLEACH (Varadhan and
Johnson, 1997) and it is widely utilized. However, the
underlying assumptions are overly simplistic and
result in predictions that are inconsistent with
observations during SVE and in post-SVE
monitoring. The VLEACH model assumes the
vadose zone soil is homogeneous despite the
observed mass transfer constraints observed in
essentially all SVE applications. Typically, soil
property measurements are averaged for VLEACH
and result in higher diffusion rates from source soils
than reality. Measured vapor concentration profiles
are calculated to dissipate quickly, on occasion faster
than observed during active SVE, and the model
2 Available for download from https://www.epa.g~ov/water-
re se arch / vadose-zone-le aching--vie ach.
38
Soil Vapor Extraction (SVE) Technology
-------
over-predicts the flux entering groundwater. In
addition, two boundary conditions are available at the
water table. A zero vapor concentration (i.e.,
immediate dissolution into groundwater) can be
specified that over-predicts the entering flux.
Secondly, a zero concentration gradient can be
specified such that mass only enters groundwater via
infiltration of contaminated porewater; however, the
infiltration rate is very difficult to predict. Sites in arid
climates may experience extensive evapotranspiration
yielding a negative infiltration rate. Finally, VLEACH
is needlessly numerical and can be solved analytically
as derived by DiGiulio et al. (1998). DiGiulio et al.
(1999) and DiGiulio and Varahan (2001b) also
extended the choice of boundary conditions and
methods of assessment.
SESOIL (Hetrick and Scott, 1994) is also an early
model occasionally used to calculate the fate and
transport of VOCs in the vadose zone; however, the
model is not applicable to this scenario. In the
SESOIL model, downward movement of pollutant
occurs only with the soil moisture, while upward
movement can occur only by vapor phase diffusion.
For sites applicable to SVE, downward vapor
diffusion generally contributes much more
significantly to contaminant transport than downward
soil moisture migration and cannot be neglected.
The difficulties in modeling the interaction of the
vadose zone and groundwater are well described by
Truex et al. (2009):
The use of analysis techniques needs to be considered
in the context of the dominant mode of contaminant
transport within the vadose ~one ... As vapor-phase
transport becomes more important (e.g., for arid sites
with low aqueous recharge), three-dimensional
contaminant movement in the vadose ~one may be
more inportant, the contact area of vadose ~one
contamination on the water table is more difficult to
estimate, and transport of contaminants across the
water table includes a mass transfer resistance. Thus,
when vapor-phase transport is significant, these issues
should be considered in terms of computing
contaminant flux to the ground water and the
resultant ground water contaminant concentrations.
As a result of the limitations in the early models and
the complexity of the boundary condition at the water
table, researchers continue to develop methods to
examine the interaction of vadose zone contaminants
with the groundwater in the context of SVE
performance (Carroll et al., 2012; Oostrom et al.,
2010; Truex et al., 2009). These approaches and those
in U.S. ACE (2002) estimate the residual mass as well
as the mass transfer coefficient (or utilize a one-
dimensional diffusion model) to assess the mass flux
from a residual source and then assume various
interactions with the underlying groundwater.
Numerous specific scenarios can also be calculated
using methods and spreadsheets provided in U.S.
DOE (2013). This approach can also be integrated
into the framework of VLEACH, where a numerical
formulation exists, by allowing soil properties and
water saturation to vary with depth and by
implementing more realistic boundary conditions at
the water table. These variable properties can be
correlated to the site-average mass transfer coefficient
between the advective and diffusive soils (U.S. ACE,
2002) as described in Section 5.2. The SVE estimated
fractions of advective versus diffusive soils can also
be utilized to validate more realistic models of source
soils. In addition, more complex water table
interactions are necessary to provide realistic
estimates of the vadose zone mass flux entering
groundwater, or potentially, the mass flux of
contaminant entering the vadose zone from
contaminated groundwater. Recall from Section 5.2,
SVE data can also be used to estimate a mass transfer
coefficient for the interface between the vadose zone
and groundwater, if groundwater is contaminated at
the start of SVE.
Example results utilizing VLEACH modified to
include a layered vadose zone with varying physical
properties and water saturation and a more realistic
interaction with contaminated groundwater are
provided in Figure 11. The plots on top illustrate
vapor concentration profiles calculated after SVE
ends. The plots on the bottom provide a mass
balance over time for calculated mass remaining in
Soil Vapor Extraction (SVE) Technology
39
-------
the vadose zone, cumulative mass entering
groundwater, and cumulative mass lost to the
atmosphere through the soil surface. The layered soil
model results on the left were calculated by varying
soil properties in VLKAGH with depth, particularly a
tighter soil interval in the deep vadose zone with a
high moisture content. The uniform vadose zone soil
model results on the right were calculated with
VI .K \( :i 1 as downloaded from the ETA Web site
and averaging the measured soil properties.
The layered model profiles are observed to maintain
nearly the same shape for all the times providing a
limited validation of using a layered model, whereas
the uniform soil model rapidly approaches an
Layered Soil Model
20
40
60
80
100
120
140
0 years
5 years
20 years
100 years
2000 4000 6000 8000 10000
TCE Vapor Concentration (ppbv)
12000 14000
70
60
_ 50
to
"O
c
0 40
Q.
1 30
LU
20
10
0
Vadose Zone
Entered GW
f *
/ /
20
40 60
Elapsed Years
80
100
unrealistic uniformity that is not consistent with initial
profiles or the results observed during SYK at the site.
The layered model was further validated by
comparing calculated mass flux rates from tight layers
with those calculated using the mass transfer
coefficient estimated from SVE, operations, as
described in Section 5.2, and a concentration gradient
between permeable and tight soils. The mass balances
between the two models were significantly different
as the uniform soil model suggested most of the
residual mass in the vadose zone at the end of SVE
would enter groundwater within a couple of decades,
with smaller losses to the atmosphere, and a relatively
clean vadose zone within about 50 years.
Uniform Soil Model
20
40
60
80
100
120 '
140
0 years
5 years
20 years
100 years
2000 4000 6000 8000 10000
TCE Vapor Concentration (ppbv)
12000 14000
60
50
vo 40
30
20
10
0 L-
Vadose Zone
----Lost to Atm
Entered GW
[
20
40 60
Elapsed Years
80
Figure 11. Comparison of modeled vapor concentration profiles and mass balances for layered and uniform soil models
40
Soil Vapor Extraction (SVE) Technology
-------
Conversely, the layered model, correlated to SVE
operational results, indicates the mass will remain in
the vadose zone for much longer with significantly
slower entrance to groundwater and a higher
proportion of the mass lost to the atmosphere at the
surface. A mixing model was used to calculate
groundwater concentrations resulting from the
calculated mass flux at the water table and the results
compared very well with historical groundwater
concentrations observed in a nearby, downgradient
monitoring well, including a trend of decreasing
groundwater concentration observed during several
years of SVE.
7 COST CONSIDERATIONS FOR SVE
SYSTEMS
As described in Section 2, SVE, either alone or in
conjunction with other treatment technologies, has
been selected as the remedial action at more than 285
Superfund sites (U.S. EPA, 2012a), and has likely
been used for thousands of non-Superfund cleanups,
such as SVE treatment of leaking underground
petroleum storage tank sites. Cases of successful
implementation of SVE treatment technology have
been documented at both the federal and state level,
and more than two decades ago EPA designated SVE
as a presumptive remedy for VOCs in soils at
Superfund sites (U.S. EPA, 1993). At the state level, a
study conducted by the California State Department
of Toxic Substances Control indicated that SVE was
the most frequently selected cleanup alternative for
carcinogenic VOCs in vadose zone soils (CalEPA,
2010).
Thus, there is much information and experience
available from federal, state, and local authorities on
actual SVE system costs and long-term performance.
However, SVE system design, installation, operation,
and maintenance costs vary significantly based on site
conditions. Some of the more important
considerations are listed below, but the reader is also
referred to the cited references in this section for
more detailed information on costing an SVE system,
Soil Vapor Extraction (SVE) Technology
and to a qualified remediation engineer for designing
and costing an actual SVE system for a specific site.
7.1 System Design
SVE systems may be designed from the ground up or
a packaged system may be available for rent or
purchase from various vendors. SVE systems
designed from the ground up may be more expensive
to design and construct than an off-the-shelf
packaged system. While using a packaged system may
reduce design and construction costs (Goldstein and
Ritterling, 2001; U.S. EPA, 1997b), it could increase
the operational cost if the extraction blower or
treatment train are not appropriately sized for site-
specific conditions.
7.2 Cost Components
SVE system capital costs typically include extraction
and monitoring well construction; vacuum blowers
(e.g., regenerative, positive displacement, or
centrifugal); vapor and liquid treatment systems
piping, valves, and fittings (usually plastic); and
instrumentation to support system operation,
monitoring, and maintenance (U.S. EPA, 1997b).
Most of these SVE system components are readily
available off the shelf. However, the exact system
design, configuration, process components, layout,
operation, and maintenance of the extraction and
monitoring system is strongly influenced by site-
specific factors including the properties and
concentration of contaminants; the areal and vertical
extent of contamination; system size (number of
extraction wells and blower size); the need for and
complexity of off-gas treatment; and the length of
time the system must be operated to reach cleanup
targets (U.S. EPA, 2001). System design modifications
may also be required after installation depending on
the actual operating conditions and system
performance.
7.3 Operation and Maintenance
Operations and maintenance costs for SVE systems
include labor, electrical power, maintenance,
monitoring, and air and groundwater treatment
-------
activities. SVE system components are highly reliable
and are capable of continuous operation for the
duration of the cleanup without constant operator
oversight. Controls are often integrated to allow for
the automatic shutdown of the system if conditions
indicate component failure. If the system shuts down,
the controls can be designed to immediately notify
the operator so that component failure can be
identified and repairs made quickly for minimal
downtime. In terms of off-gas and groundwater
effluent, cleanup costs can significantly increase if
both require treatment (FRTR, 2008). Air emissions
treatment and testing is often the most costly
component of an SVE system (AFCEE, 2001).
Electric power costs can vary significantly by location
(i.e., local utility rates and site conditions).
7.4 Cost Estimating Tools
Because most SVE system components are available
off the shelf, RSMeans (2017) or other standard
construction cost estimating tools may be used as a
resource for costing various SVE system components
such as parts and labor for installation of subsurface
piping, electrical equipment, and treatment system
components. For costs specific to SVE systems (as
well as ancillary remediation technologies such as
bioventing and air sparging), use FRTR's Remedial
Action Cost Engineering and Requirements (RACER)
software (FRTR, 2016). RACER is a PC-based system
originally developed in 1992 by the U.S. Air Force
that uses a methodology for generating location-
specific cost estimates for remediation technologies.
RACER allows the user to select the desired models
from a list of available technologies (including SVE),
define the required parameters in the selected
technology, and tailor the estimate by verifying and
editing secondary parameters. RACER uses current
multiagency pricing data, and is researched and
updated annually to ensure accuracy. In addition to
users within the federal government, RACER is used
by a variety of state regulatory agencies, engineering
consultants, facility owners and operators, financial
institutions, and law firms.
FRTR (2008) includes various RACER outputs, links
to ancillary technologies, references, and typical
system design diagrams for SVE systems (Figure 12).
FRTR (2008) also provides an example detailed cost
analysis developed in 2006 using RACER for "easy"
and "difficult" SVE remedial actions at "small" and
"large" sites, with total costs ranging from $80,295 for
an "easy" and "small" site, to $368,465 for a
"difficult" and "large" site. Note that it is not clear
what level of off-gas or collected groundwater
treatment is included in the operating and
maintenance (O&M) line item, which may be a
significant cost at some sites.
Caution is recommended in using these costs and
references in the context of an actual site cleanup
because the base year of the estimates varies and
because the contaminant properties, site
characteristics, and performance objectives may differ
significantly from the conditions at the specific site of
interest. In short, SVE costs are highly variable due to
site-specific variation in critical parameters that can
impact the SVE process and how these characteristics
can change over time, and this variability must be
considered to get adequate cost estimates for site-
specific applications of SVE.
8 SVE ENHANCEMENTS AND
COMPLEMENTARY AND PASSIVE
TECHNOLOGIES
When evaluating the termination of active SVE, a
number of SVE system enhancements can be
undertaken to hasten cleanup by focusing on the
more persistent VOC sources revealed during SVE
operation (U.S. EPA, 1997a). If SVE enhancements
are unlikely to complete the cleanup in the timescale
desired, transitioning to another complementary
technology may be needed. Alternatively, at sites
where cleanup timescales are less critical or cleanup
activities are nearing an end, passive SVE systems can
be a cost-effective and low-impact alternative to
maintaining active systems. Finally, a flux control
approach can be implemented by sequestering the
remaining VOC mass rather than extracting or
42
Soil Vapor Extraction (SVE) Technology
-------
Vacuum Relief Valve
ir Filter
Moisture Separator Inlet
Manual Starter for
Hazardous Locations
Gas Discharge
Moisture
Separator -
Fume Incineration
High Level
Inlet flir
Shut-Off Float
Catalytic Oxidation
Carbon Treatment
To Off-Gas Treatment ;•
Moisture Drain
Steel Skid
Vacuum Blower
Contaminated Zone
Source: FRTR, 2008
Figure 12. Typical in situ soil vapor extraction system
destroying the mass in place. However, flux control is
the least desirable of these alternative approaches as it
does not reduce the source of YOCs.
8.1 Potential Enhancements to an Existing SVE
System
If SVE alone is not successful in meeting remediation
objectives, other cleanup technologies can enhance
SVE to hasten attainment of cleanup goals. In these
cases, SVE. is still part of the technology, but other
mass removal mechanisms are added or enhanced to
address ^Ones ot persistent source concentrations.
These technologies can be applied to small mass
fluxes after SVE has largely removed the source, or
where SVE alone is unlikely to attain further
significant source reduction. Preference is given to
technologies that make use of the existing SVE
infrastructure (wells, piping, blowers, etc.) and
leverage the capital investment for SVE,
Bioveiiling. At sites where the primary contaminants
are aerobically biodegradable (or cometabolically
degradable3), replacing active extraction with air
injection provides oxygen to the native bacteria and
stimulates additional contaminant removal without
the cost of off-gas treatment. This approach is widely
applied at petroleum hydrocarbon sites. Air injection
can be pulsed, with the pulse frequency and duration
based on observed Oxygen uptake rates. Existing
wells, piping, and blowers can often be used. The
addition of gaseous nutrients (e.g., nitrous oxide,
triethyl phosphate) can be used to maximize
3 Cometabolism is the simultaneous degradation of two
compounds, in which the degradation of the second
compound depends on the presence of the first compound.
Soil Vapor Extraction (SVE) Technoh
43
-------
degradation rates, although many sites have been
addressed without nutrient addition (Leeson and
Hinchee, 1996; U.S. EPA, 1995).
Multiphase Extraction. Multiphase extraction
simultaneously extracts vapors and liquids from the
same well, using either a single vacuum pump or
separate pumps for the separate phases. The liquid
extraction may enhance the removal of mass from the
location of the smear zone/capillary fringe by
lowering both the water table and levels of soil
saturation. The application of vacuum can also
enhance the removal of liquids from soils with
modest permeabilities for simultaneous recovery of
dissolved mass or LNAPL (bioslurping) from the
source areas (U.S. ACE, 1999).
In Situ Thermal Treatment. The application of heat
to contaminated soil increases the vapor pressures of
most organic contaminants and is accompanied by
changes in the solubility, viscosity, surface tension,
and density of NAPLs. Thermal enhancements are
relatively expensive and costs depend on the size of
the subsurface zone being treated. Therefore, it is
prudent to apply thermal technologies only after SVE
reveals zones of persistent contaminants.
Bioremediation rates and hydrolysis (for chlorinated
ethanes) may be significantly enhanced at modestly
elevated temperatures due to faster reaction kinetics
at higher temperatures or the enhancement of
biodegradation by robust thermophilic bacteria.
Soil heating is more tolerant of soil heterogeneity than
most other in situ technologies. Heat can be
introduced through electrical resistivity heating
(passing currents between electrodes placed into the
soils to be treated), thermal conduction heating
(where heat propagates through conduction from
heaters placed in wells), or steam injection. The
vapors generated by the process are typically collected
via vapor extraction wells (Kingston et al., 2010; U.S.
ACE, 2014). Recent work has extended the
applicability of SVE to emerging contaminants, such
as 1,4-dioxane, that are resistant to volatilization
because of high water solubility and to biological
degradation with the use of heated air injection to
reduce porewater saturation (Rob Hinchee, pers.
comm. 2016).
In Situ Air Sparging. Air sparging involves the
injection of air into wells with short screen intervals
below the water table. The injected air moves
outwards and upwards from the well based on
buoyancy and air-entry pressures of the soil strata,
which are closely related to typical pore size and
connectedness. If the remaining sources are
concentrated near the water table and capillary fringe,
the mass may not be easily accessible to SVE. In situ
air sparging may allow vertical air passage through
these zones for either subsequent capture of the
contaminant vapors with the existing SVE system or
discharge to the vadose zone and ultimately to the
atmosphere. The air also agitates moisture in the
capillary fringe and may provide a source of dissolved
oxygen to promote aerobic degradation of some
compounds in both the groundwater and vadose zone
(U.S. ACE, 2013).
Hydraulic/Pneumatic Fracturing. Hydraulic and
pneumatic fracturing are two technologies that induce
fractures in the subsurface to enhance the
remediation of contaminants by increasing the
effective (interconnected) porosity of subsurface
materials. These technologies are particularly useful
and cost-effective at contaminated sites with low-
permeability soil and geologic media, such as clays,
shales, and tight sandstones, where remediation is
difficult without some sort of permeability
enhancement. However, the usefulness of fracturing
technology is not limited to low-permeability sites.
Beneficial effects can also be achieved by creating
new or enlarging existing fractures in the subsurface,
which improves air flow to encourage degradation
and removal of contaminants (Riha et al., 2008;
Suthersan, 1999b).
8.2 Passive SVE
Typical SVE operations rely on "active" SVE
technology, where active vacuum or blower pressure
is applied to the subsurface with standard SVE
Soil Vapor Extraction (SVE) Technology
-------
equipment to induce advective air flow into vapor
extraction wells. Passive SW,B capitalizes on the use of
natural pressure gradients between the subsurface and
atmosphere to effect an advective flush in the
subsurface (Kuang etal., 2013; SRNL, 2010; U.S.
ACEj 2002), A one-way passive mechanical valve is
located at the wellhead that allows air to exit the
subsurface during diurnal or weather-related pressure
outflow periods without allowing ambient air to enter
the well otherwise. Passive SVE is often beneficial as
a polishing strategy prior to monitored natural
attenuation or formal site closure because of its
relative effectiveness in removing low levels of
residual contamination, its low operation and
monitoring cost, and minimal site disruption.
Figure 13 illustrates how temporal fluctuations in
barometric pressure (based on natural diurnal change
or multiday weather phenomena) results in
intermittent removal of soil vapors from the vadose
Temporal Fluctuations in Barometric Pressure
TIME
BAROMETRIC PRESSURE
TARGET ZONE PRESSURE
Resulting Exchange in Air Between Atmosphere and
Subsurface Enviromnent
INHALATION
' ...
P <
r Target zone
p
r Barometric
J-
m
TARGET
ZONE
EXHALATION
P >
' Target zone
D
Barometric
ri
TARGET
ZONE
Reproduced from SRNL, 2010
Figure 13. Conceptual model for passive 'inhalation' and 'exhalation'
Soil Vapor Extraction (SVE) Technology
45
-------
zone (SRNL, 2010). If these natural variations do not
produce sufficient flow, microblowers may also be
implemented. These small direct current blowers can
be powered by renewable energy such as one or two
typical solar panels and provide a small but
measurable boost in vacuum levels, while still
lowering capital costs. Similar to active SVE, passive
SVE works more effectively in coarse, high
permeability soils. However, passive SVE can work
more effectively than active SVE in removing residual
contamination from low permeability soils that are
dependent on diffusion rather than advective flow for
contaminant removal.
Consider using passive SVE when
• Asymptotic conditions are reached at low
concentrations with active SVE
• A large number of active SVE wells can be
converted to a passive system
• Lower permeability source soils require
treatment
• A site is remote and has long timescales for
cleanup
• An active site requires minimal surface
disturbance and has long timescales for
cleanup.
Passive SVE is likely not applicable under the
following conditions
• High soil moisture
• Preferential pathways
• Minimal contaminant zone stratification
• High contaminant concentrations causing
aboveground treatment needs
• Short cleanup timeframe
• High extraction rates
• Large number of new wells.
Similarly, bioventing can be implemented passively as
described above but with a reversal in the flow
direction. NFESC (2000) evaluated the applicability
of passive bioventing at 15 U.S. Department of
Defense sites.
9 CASE STUDIES
Two case studies from actual SVE-remediated sites
are provided to illustrate the application of the
concepts and methods described in this EIP. Case
Study #1 illustrates how one can address a newly
emerged, small, but persistent source at an existing
SVE site by using limited SVE testing, past data, and
modeling to assess past and future mass removal and
optimize a system to close a site. Case Study #2
provides an example of how one can assess the
potential for groundwater to recontaminate a site
after SVE cleanup by using SVE operational data and
modeling at a site with potential vapor intrusion
exposures. Each case study is summarized below with
more detailed site descriptions in Appendices A and B
for Case Study #1 and #2, respectively.
9.1 Case Study #1: Assessment of a Small
Persistent TCE Source
In the first case study, after years of SVE and
attainment of remedial goals in nearly all areas of the
site, a small persistent source of TCE emerged during
drought conditions as the high moisture content
separating the source from permeable pathways
dissipated. SVE operations began in 1997 and
operated through 2009 extracting over 1,600 lbs of
TCE. Several years of monitoring under ambient
conditions supported closure until 2014 when the
small source emerged and reopened the site for
further investigation.
A review of historical boring logs and soil sampling
for physical analyses indicated a clayey silt interval
existed from a depth of about 25 to 35 ft below
ground and likely created some degree of perched
water at 25 ft. The lateral extent of such water was
not known. The conceptual site model was revised to
account for this water and its disappearance during
the drought conditions, exposing the interface of the
clayey silt to overlying silty sand. In effect, the barrier
to volatilization was removed. The contamination in
the clayey silt was the result of overflow and leaks of
TCE-laden water from oil-water separators. The
water migrated downward and seeped into this fine-
Soil Vapor Extraction (SVE) Technology
-------
grained interval and remained relatively pristine with
respect to TCE concentration as long as it was
covered by the high moisture content layer.
Rather than drill numerous borings to attempt to
define the lateral extent of the contamination, a
decision was made to perform the equivalent of a
pilot SVE test and evaluate the initial decay and
subsequent rebound to define the volume of
contaminated soil, the mass of TCE in this residual
source, and the mass transfer coefficient for
calculating a mass flux toward the surface. A small,
one horsepower regenerative blower powered by a
standard single phase, 120V, 20A wall outlet was used
for the pilot SVE with GAC for vapor treatment,
simplifying the testing.
To analyze the test, a two-region model of SVE
developed by Praxis Environmental Technologies and
published as Appendix F of the U.S. ACE (2002)
SVE and Bioventing Engineers Manual was utilized.
The extracted vapor concentration data and the
cumulative mass extracted were used to calibrate the
two-region SVE model. The two-region model of
SVE provides an estimate for the residual mass and
quantifies mass transfer constraints. In short, the
model volume-averages vapor concentrations over
the contaminated soil volume (conceptualized as
partially mobile/advective and partially immobile/
diffusive) and is based on an overall mass balance.
The results of the field testing and data evaluation
were:
• The characteristic volume of contaminated soil
was estimated to be 2,700 yd3
• The bulk mass transfer coefficient between soil
types was estimated to be 0.0854 day"1
• Estimated initial mass of TCE within the
influence of extraction was 13.4 lbs
• Estimated mass of TCE extracted during the
testing was 12.6 lbs
• 3 days of initial extraction removed 40% of the
original TCE mass
• 57 days of additional extraction removed 54%
of the original TCE mass
• 6% of the original TCE mass remained after
testing
• Estimated residual mass of TCE of about
0.8 lbs
• At the end of the extraction testing, the TCE
mass removal rate was 0.03 lbs/day
• Maximum theoretical vapor concentration
remaining in the small low permeability source
soils is 35 ppmv. reduced from an initial
measured concentration over 600 ppmv, a 95%
reduction after 77 days including 60 days of
active SVE.
In summary, years of SVE at this site extracted more
than 1,600 lbs of SVE; yet, a residual mass of about
13 lbs in a clayey silt layer required additional SVE to
complete the cleanup after the moisture content at
the site decreased.
9.2 Case Study #2: Evaluation of Mass Transfer
across the Capillary Fringe from SVE
Operational Data
In the second case study, SVE was applied in a single
groundwater monitoring well with 20 ft of screen
exposed to the bottom of the vadose zone as a result
of a falling table over the past 20 years, making the
well amenable to SVE. The well was about 500 ft
downgradient from the original source area of TCE
releases and appeared to reside near the centerline of
the resulting dissolved TCE groundwater plume. The
action provided an assessment for the potential mass
extraction rate of contaminant volatilized from
contaminated groundwater and the use of SVE to
mitigate vapor intrusion from contaminated
groundwater.
SVE operated at well MW-07 from November 30,
2011 until March 18, 2014, with an average gas
extraction rate of about 200 scfim. The purpose was
to assess the benefits of TCE mass removal from
underlying groundwater and surrounding vadose zone
soils. The total mass of TCE extracted from the well
Soil Vapor Extraction (SVE) Technology
47
-------
was estimated to be 322 lbs during the initial 20-
month extraction period with an initial rate of
1.4 lbs/day and a final rate of about 0.22 lbs/day.
A conceptual model for contaminant mass available
for vapor extraction around the well consists of
dissolved mass transported with groundwater from
the original source zone that subsequently volatilized
and diffused upward. The available mass for upward
diffusion increased as the water table fell, leaving soil
partially saturated with immobile contaminated water
(formerly groundwater) at the bottom of the vadose
zone. In addition, a relatively low permeability unit in
the middle of the vadose zone acted as a cap that
constrains the rate of diffusion upward to the surface.
Straightforward two-domain modeling as described in
Section 5.2, modified to include contaminated
groundwater as source for extracted mass, was
matched to the extracted vapor concentration history
and mass extraction to assess the activity. The total
mass of TCE extracted was about 320 lbs and
exceeded the estimate of 234 lbs for the initial mass in
the vadose zone. The balance was the result of TCE
volatilization from contaminated groundwater over a
large area after sweeping the permeable soils and
creating a driving concentration gradient. At the end
of the extraction in March 2014, the mass extraction
rate was nearing a balance with the volatilization rate
from groundwater at a rate of about 0.16 lbs/day and
an estimated 117 lbs of TCE was volatilized from
underlying groundwater and extracted during the 20
months of extraction. Restart of the extraction in
November 2015 after a 20-month rebound period
and subsequent operation through April 2016 yielded
an additional 45 lbs of TCE.
The data fit yielded a bulk (volumetric) mass transfer
coefficient for volatilization across the capillary fringe
of 0.0002 day4 with an average groundwater TCE
concentration of 1,400 (Jg/L. These values, assuming
the average groundwater concentration is constantly
recharged, yield an asymptotic TCE vapor extraction
mass rate of about 0.15 lbs/day. Hence, over a
longer-term steady operation, vapor extraction is
expected to level off at about 0.15 lbs/day, with
underlying groundwater as the source that is
somewhat independent of the extraction rate if the
extracted concentration is maintained at a low value
compared to the groundwater concentration.
Design of additional SVE or pulsing can be based on
the extraction rate or exchange rate in an effort to
balance the mass transfer processes. Consider the
governing equation for the extracted concentration
provided in Section 5.2,
[0a(l -Sa)FaRa]
dC9„
dt
= -YtCa+"d(CH-CS)
+ a (HCW — Cg>\
' ^gw v gw ^ a )
Scaling arguments yield estimates for extraction rates
by comparing the extraction term with the
groundwater volatilization while neglecting the faster
mobile/immobile domain mass transfer,
— Cg » a (HCW — Cg}
y ^ ^gwy11 W?w LaJ
Q » agwVt
' h rw r®N
n L.gW L.a
k 5? ,
ttgw Vt(10 - 1) ~ 9 aQW Vt
gw
Q» 9 (0.0002 day'1) (650,000 m3)
3
m
= 1,100 —— = 29 scfm
day
In this estimate, the driving concentration gradient
for mass transfer was assumed to be held at 90% of
the maximum (Cextmct ~ 17 mg/m3) and yielded an
extraction rate significantly lower than the flow
sustained during the extraction test.
This first order estimate suggests an extraction rate as
low as 30 scfm may be sufficient to keep
concentrations low over a large area and could be
used to mitigate vapor intrusion for buildings on the
surface. The low extraction rate is a direct result of
the low volatilization rate from groundwater. The
Soil Vapor Extraction (SVE) Technology
-------
estimated bulk mass transfer coefficient between
groundwater and the bottom of the vadose zone is
0.0002 day4 and suggests a re-equilibration timescale
of about 10 years. These results have implications for
designing low-level SVE to mitigate vapor intrusion
by preventing contaminant mass from diffusing
upward toward the surface. The testing described in
this case study has applicability for design at any site
where contaminated groundwater is a potential source
for vapor intrusion.
In a current EPA research effort to assess vapor
intrusion mitigation with SVE (Lutes et al., 2017;
Schumacher et al., 2017, Truesdale et al., 2016), this
same interaction between contaminated groundwater
and mass extracted during SVE is being studied to
develop field test methods for quantifying the mass
flux potential from contaminated groundwater as the
source for vapor intrusion.
10 ACKNOWLEDGEMENTS
This Engineering Issue Paper was prepared for the
U.S. EPA Office of Research and Development
(ORD), National Risk Management Research
Laboratory (NRMRL) by RTI International under
Contract No. EP-C-11-036, Task Order 025. John
McKernan served as the EPA Task Order Manager.
Ed Barth of EPA's ORD NRMRL, served as EPA's
technical lead. Robert Truesdale managed and
technically directed this work and Dr. Lloyd "Bo"
Stewart of PRAXIS Environmental Technologies was
the primary author and technical lead. Jennifer
Redmon provided technical input to an earlier version
of the document, and Coleen Northeim was the RTI
Contract Manager.
This EIP is intended as an overview of SVE for EPA
staff, regional program offices, RPMs, and state
governmental environmental staff. Because SVE is an
evolving remediation technology, interested parties
can further consult the body of literature and
experience that constitutes the state-of-the-science for
SVE treatment. As of the date of this publication,
questions may be addressed to Mr. McKernan or
Mr. Barth, EPA ORD NRMRL
(mckernan.john@epa.gov. 513-569-7415;
bartli.ed@epa.gov. 513-569-7669).
For additional information, please contact:
John McKernan, Director
U.S. EPA ORD Engineering Technical Support
Center (ETSC)
26 W. Martin Luther Kng Drive MLK-489
Cincinnati, OH 45268
513-569-7415
mckernan.) ohn@epa.gov
Reference herein to any specific commercial products,
process, or service by trade name, trademark,
manufacturer, or otherwise, does not necessarily
constitute or imply its endorsement, recommendation
or favor by the United States Government. The views
and opinions of the authors expressed herein do not
necessarily state or reflect those of the United States
Government, and shall not be used for advertising or
product endorsement purposes.
11 REFERENCES
AFCEE (Air Force Center for Engineering and the
Environment). 2001. Guidance on Soil I Tapor
Extraction Optimisation. United States Air Force
Environmental Restoration Program, Air Force
Center for Environmental Excellence, Brooks
AFB, TX. June.
Armstrong, J. E., E.O. Frind, and R.D. McClellan.
1994. Nonequilibrium mass transfer between the
vapor, aqueous, and solid phases in unsaturated
soils during vapor extraction. Water Resources
Research 30(2): 355-368.
Aronson, D., M. Citra, K. Shuler, H. Printup, and
P.H. Howard. 1999. Aerobic Biodegradation of
Otganic Chemicals in Environmental Media: A
Summary of Field and Laboratory Studies. Final Report
SRC TR 99-002, prepared by Syracuse Research
Corporation for U.S. EPA Office of Research and
Development, Athens, GA. January.
http: / / energytomorrow.org/- / media/files / ehs /
clean water/gw other/aerobicbiodegrateconstant
rptl999pdf.pdf.
Soil Vapor Extraction (SVE) Technology
49
-------
Brooks, R.H., and A.T. Corey. 1964. Hydraulic
Properties of Porous Media. Hydrology Papers, No. 3,
Colorado State U., Fort Collins, Colorado.
Brusseau, M.L. 1991. Transport of organic chemicals
by gas advection in structured or heterogeneous
porous media: development of a model and
application to column experiments. IVater Resources
Research 27(12): 3189-3199.
Brusseau, M.L., J. Mainhagu, C. Morrison, and K.C.
Carroll. 2015. The vapor-phase multi-stage CMD
test for characterizing contaminant mass
discharge associated with VOC sources in the
vadose zone: application to three sites in different
lifecycle stages of SVE operations. Journal of
Contaminant Hydrology 179: 55-64.
Brusseau, M.L., V. Rohay, and M.J. Truex. 2010.
Analysis of soil vapor extraction data to evaluate
mass-transfer constraints and estimate source-
zone mass flux. Ground Water Monitoring
Remediation 30(3): 57-64.
CalEPA (California Environmental Protection
Agency). 2010. Proven Technologies and Remedies
Guidance: Remediation of Chlorinated I Tolatile Organic
Compounds in I Tadose Zone Soil. California
Department of Toxic Substances Control,
Sacramento, CA. April.
http:/ /www.dtsc.ca.gov/SiteCleanup/upload/ cV
PC 040110.pdf.
Carroll, K.C., R. Taylor, E. Gray, and M.L. Brusseau.
2009. The impact of composition on the physical
properties and evaporative mass transfer of a
PCE-diesel immiscible liquid. Journal of Hazardous
Materials 164(2-3): 1074-1081.
doi:10.1016/j.jhazmat.2008.09.003.
Carroll, K.C., M. Oostrom, M.J. Truex, V.J. Rohay,
and M.L. Brusseau. 2012. Assessing performance
and closure for soil vapor extraction: Integrating
vapor discharge and impact to groundwater
quality. Journal of Contaminant Hydrology 128: 71-82.
Carroll, K.C., M.J. Truex, M.L. Brusseau, K.R. Parker,
R.D. Mackley, and V.J. Rohay. 2013.
Characterization of persistent volatile
contaminant sources in the vadose zone. Ground
Water Monitoring and Remediation 33(2):68-84.
DePaoli, D.W., S.E. Herbes, and M.G. Elliott. 1991.
Performance of in situ soil venting system at jet
fuel spill site. P. 260-272 in Pederson, T.A. and
Curtis, J.T. Soil I ^apor Extraction Technology, Reference
Handbook. Camp Dresser & McKee, Inc. for U.S.
EPA, Cincinnati, Ohio. EPA/540/2-91/003.
DiGiulio, D.C., M.L. Brusseau, and V. Ravi. 1998.
Use of diffusion modeling to aid assessment of
rate-limited vapor transport for SVE. P. 155-160
in Wickramanayake, G.B., and R.E. Hinchee
(eds.) Physical, Chemical, and Thermal Technologies:
Remediation of Chlorinated and Recalcitrant Compounds,
Vol. 1(5). Battelle Press, Columbus, OH.
DiGiulio, D.C., V. Ravi, and M.L. Brusseau. 1999.
Evaluation of mass flux to and from ground water
using a vertical flux model (VFLUX): application
to the soil vacuum extraction closure problem.
Ground Water Monitoring Remediation 19(2): 96-
104.
DiGiulio, D.C., and R. Varahan. 2001a. Limitations of
ROI testing for venting design: description of an
alternative approach based on attainment of a
critical pore-gas velocity in contaminated media.
Ground Water Monitoring Remediation 21(1): 97-
114.
DiGiulio, D.C., and R. Varahan. 2001b. Analysis of
water saturation, NAPL content, degradation half-
life, and lower boundary conditions on VOC
transport modeling: Implications for closure of
soil venting systems. Ground Water Monitoring
Remediation 21(4): 83-95.
Domenico, P.A., and F.W. Schwartz. 1990. Physical
and Chemical Hydrogeology, John Wiley & Sons, New
York, 824 p.
FRTR (Federal Remediation Technologies
Roundtable). 2008. Soil\TaporExtraction. 4.8.
Remediation Technologies Screening Matrix, I Tersion 4.0.
U.S. Army Environmental Center, MD.
http: / /www.frtr.gov/matrix2/section4/4-7.html.
FRTR (Federal Remediation Technologies
Roundtable). 2016. Remedial Action Cost Engineering
Requirements (RACER®J System. U.S. Army
Environmental Center, MD.
https: / / frtr.gov/ ec2 / ecracersvstem.htm.
Goldstein, M.L., and J.M. Ritterling. 2001. A Practical
Guide to Estimating Cleanup Costs. U.S. EPA. Paper
30.http: / / digitalcommons.unl.edu / usepapapers / 3
0.
Soil Vapor Extraction (SVE) Technology
-------
Goltz, M.N., and M.E. Oxley. 1994. Technical note:
An analytical solution to equations describing
rate-limited soil vapor extraction of contaminants
in the vadose zone. Water Resources Research 30(10):
2691-2698.
Hetrick, D.M., and S.J. Scott. 1994. The New SESOIL
User's Guide (Revision 1.6). PUBL-SW-200-94
(Rev). Wisconsin Department of Natural
Resources, Madison, WI. August.
Johnson, P.C., M.W. Kemblowski, and J.D. Colthart.
1990a. Quantitative analysis for the cleanup of
hydrocarbon-contaminated soils by in-situ soil
venting. Groundwater 28 (3) :413-429.
Johnson, P.C., C.C. Stanley, M.W. Kemblowski, D.L.
Byers, and J.D. Colthart. 1990b. A practical
approach to the design, operation, and
monitoring of in situ soil-venting systems. Ground
Water Monitoring Review 10(2): 159-178.
Jury, W. A., W.R. Gardner, and W.H. Gardner. 1991.
Soil Physics, John Wiley, New York, 328p.
Kingston J.T., P.R. Dahlen, P.C.Johnson, E. Foote,
and S. Williams. 2010. Critical Evaluation ofState-of-
the-Artln Situ Thermal Treatment Technologies for
DNAPL, Source Zone Treatment, Final Report.
ESTCP Project ER-0314, Environmental Security
Technology Certification Program, Alexandria,
VA. http://www.serdp.org/content/download/4
576/67301 /file /ER-0314-FRpdf.
Kuang, X., J.J. Jiao, and H. Li. 2013. Review on
airflow in unsaturated zones induced by natural
forcings. Water Resources Research 49: 6137-6165,
doi:10.1002/wrcr.20416.
Lawrence, S.J. 2006. Description, Properties, and
Degradation of Selected I Tolatile Organic Compounds
Detected in Ground Water—A Review of Selected
literature. USGS Open-File Report 2006—1338. 62
p. http://pubs.usgs.gov/ofr/2006/1338/.
Leeson, A., and R.E. Hinchee. 1996. Principles and
Practices of Bioventing, I Tolnme II: Bioventing Design.
U.S. Air Force Center for Environmental
Excellence, Brooks Air Force Base, TX.
Li, Z., and M.L. Brusseau. 2000. Nonideal transport
of reactive solutes in heterogeneous porous media
6. Microscopic and macroscopic approaches for
incorporating heterogeneous rate-limited mass
transfer. Water Resources Research 36(10): 2853-
2867.
Lutes, C., J. Lowe, R. Truesdale, B. Schumacher, J. H.
Zimmerman, R. Connell, L. Stewart, J. De Loera.
2017. Detailed field evaluation of soil vapor
extraction to prevent vapor intrusion. Battelle
Fourth International Symposium on Bioremediation and
Sustainable Environmental Technologies. Miami, FL.
May.
Matzke, B.D., J.E. Wilson, L.L. Nuffer, S.T. Dowson,
J.E. Hathaway, N.L. Hassig, L.H. Sego, C.J.
Murray, B.A. Pulsipher, B. Roberts, and S.
McKenna. 2010. I Tisnal Sartple Plan I Tersion 6.0
User's Guide. PNNL-19915, Pacific Northwest
National Laboratory, Richland, WA.
http://vsp.pnnl.gov/documentation.stm.
McCarthy, K.A., and R.L.Johnson. 1993. Transport
of volatile organic compounds across the capillary
fringe. W^ater Resources Research 29(6): 1675-1683.
McClellan, R.D., and R.W. Gillham. 1990. Vapour
extraction of trichloroethylene under controlled
field conditions. Proceedings of the IAH Conference on
Subsurface Contamination by Immiscible Fluids, April
18-20, Calgary, Alberta, Canada.
MPCA (Minnesota Pollution Control Agency). 1993.
Design Criteria and Reporting Requirements (I Tapor
Extraction, Air Sparging). MPCA Voluntary
Investigation and Cleanup Guidance Document
#17. December, https:/ /www.pca.state.mn.us/sit
es/default/files/vic-gdl 7.pdf.
NFESC (Naval Facilities Engineering Service Center).
2000. Natural Pressure-Driven Passive Bioventing, Final
Report. ESTCP Project ER-199715,
Environmental Security Technology Certification
Program, Alexandria, VA. https://www.serdp-
estcp.org/Program-Areas/Environmental-
Restoration / Contaminated-
Groundwater / Monitoring /ER-199715 /ER-
199715.
Nyer, E.K, P.L. Palmer, E.P. Carman, G. Boettcher,
J. Bedessem, D.F. I-Cidd, F. Lenzo, G.J. Rorech,
and T.L. Crossman. 2001. In Situ Treatment
Technology, 2nd Edition. CRC Press, Boca Raton,
FL, 536p.
Oostrom, M., M.J. Truex, G.D. Tartakovsky, and
T.W. Wietsma. 2010. Three-dimensional
simulation of volatile organic compound mass
flux from the vadose zone to groundwater. Ground
Water Monitoring Remediation 30(3): 45-56.
Soil Vapor Extraction (SVE) Technology
51
-------
Parsons (Parsons Corporation). 2001. Final Res/tits
Report for S oil I ^apor Extraction Optimisation Tools
Tested at: Site S-1, Kelly Air Force Base, Texas; Site IC-
1, McClellan Air Force Base, California; and Site 57,
Mather Air Force Base, California. Prepared for: Air
Force Center for Environmental Excellence,
Technology Transfer Division, Brooks Air Force
Base, San Antonio, TX.
Praxis Environmental Technologies, Inc. 2000.
Demonstration of Characterisation Progressing to
Remediation (CPR): Beale, Griffis, and Nellis Air Force
Installations and Demonstration of Soil I Tapor
Extraction Optimisation: Beale, Castle, George, Mather,
McClellan and Norton Air Force Installations.
Prepared for Air Force Center for Environmental
Excellence.
Pederson, T.A., and J.T. Curtis. 1991. Soil I Tapor
Extraction Technology Reference Handbook.
EPA/540/2-91/003. U.S. EPA, Office of
Research and Development, Washington, DC.
February, https: / /nepis.epa.gov/Exe/ZvPURL.c
gi?Dockey=200085H6.txt.
Riha, B.D., W.K. Hyde, and R. Hall. 2008. Fracture
Enhanced Soil I TaporExtraction Technology
Demonstration at the A-014 Onfall. WSRC-STI-
2008-00141. Washington Savannah River
Company, Aiken, SC. March.
Rosenbloom, J., P. Mock, P. Lawson, J. Brown, and
H.J. Turin. 1993. Application of VLEACH to
vadose zone transport of VOCs at an Arizona
Superfund site. Ground Water Monitoring
Remediation, 13(3):159-169.
RSMeans. 2017. Construction CostEstimatingData.
https:/ /www.rsmeans.com/.
Schumacher, B., J.H. Zimmerman, R. Connell, R.
Truesdale, L. Stewart, C. Lutes, J. Lowe, J. De
Loera. 2017. Updates from ORD's research on
vapor control technologies. U.S. EPA Workshop:
Finding Practical Solutions for the Chlorinated I rapor
Intrusion Pathway: Helping RCRA Facilities Meet
Significant Challenges. AEHS Foundation 27th Annual
International Conference on Soil, Water, Industry, and
Air. San Diego, CA. March.
https:/ /iavi.rti.org/WorkshopsAndConferences.c
fmPPagelD=documentDetails&AttachID=1014.
SRNL (Savannah River National Laboratory). 2010.
Passive Soil I TaporExtraction. SRNL-STI-2009-
00571. Aiken, SC. March 15.
Suthersan, S.S. 1999a. Soil vapor extraction. In
Suthersan, S.S. (ed.), Remediation Engineering. Design
Concepts. CRC Press, Boca Raton, FL.
Suthersan, S.S. 1999b. Hydraulic and Pneumatic
Fracturing, in Suthersan, S.S. (ed.), Remediation
Engineering. Design Concpts. CRC Press, Boca
Raton, FL.
Switzer C, T Slagle, D Hunter, and DS Kosson. 2004.
Use of rebound testing for evaluation of soil
vapor extraction performance at the Savannah
River site. Ground Water Monitoring Remediation
24(4): 106-117.
Truesdale, R.S., B.A. Schumacher, J.H. Zimmerman,
L.B. Stewart, C. Lutes, J. Lowe, and J. De Loera.
2016. Soil vapor extraction pilot to prevent vapor
intrusion. Presented at AWMA Specialty Conference
on I rapor Intrusion, Remediation and Site Closure. San
Diego, CA. December.
https:/ /www.researchgate.net/publication/31665
5113 Soil vapor extraction pilot to prevent va
por intrusion - Preliminary results.
Truex, M.J., M. Oostrom and M.L. Brusseau, 2009.
Estimating persistent mass flux of volatile
contaminants from the vadose zone to ground
water. Ground Water Monitoring Remediation,
29(1): 63-72.
U.S. ACE (Army Corps of Engineers). 1999. Multi-
Phase Extraction - Engineering and Design. Engineer
Manual EM 1110-1-4010. U.S. ACE, Washington,
DC. June 9.
U.S. ACE (Army Corps of Engineers). 2002.
Engineering and Design: Soil I TaporExtraction and
Bioventing. Engineer Manual EM 1110-1-4001.
U.S. ACE, Washington, DC. June 3.
http:/ /www.publications.usace.army.mil/Portals/
7 6 /Publications /EngineerManuals /EM 1110-1-
4001.pdf.
U.S. ACE (Army Corps of Engineers). 2013.
Environmental Quality: In-Situ Air Sparging. Engineer
Manual EM 200-1-19. Washington, DC.
December 31.
http:/ /www.publications.usace.army.mil/Portals/
76 /Publications /EngineerManuals /EM 200-1 -
19.pdf?ver=2014-01-06-104233-540.
52
Soil Vapor Extraction (SVE) Technology
-------
U.S. ACE (Army Corps of Engineers). 2014. Design:
In Situ Thermal Remediation. Engineer Manual EM
200-1-21. U.S. ACE, Washington, DC. May 30.
http:/ /www.publications.usace.army.mil/Portals/
76 /Publications /EngineerManuals /EM 200-1 -
21.pdf?ver=2014-05-08-155746-393.
U.S. DOE (Department of Energy). 2013. Soil I Tapor
Extraction System Optimisation, Transition, and Closure
Guidance. PNNL-21843, RPT-DVZ-AFRI-006.
Pacific Northwest National Laboratory, Richland,
WA. February, http://bioprocess.pnnl.gov/public
tn/PNNL-21843 SVE Guide.pdf.
U.S. DOE (Department of Energy). 2016. I rapor
Intrusion Estimation Toolfor Unsatnrated-Zone
Contaminant Sources. PNNL-23381. Pacific
Northwest National Laboratory, Richland, WA.
January, http://bioprocess.pnnl.gov/VIETUS R
equest.htm.
U.S. EPA (Environmental Protection Agency). 1991a.
Engineering Bulletin: In Situ Soil I TaporExtraction
Treatment. EPA/540/2-91/006. Office of
Emergency and Remedial Response, Washington,
DC. Office of Research and Development,
Cincinnati, OH. https://nepis.epa.gov/Exe/ZyP
URL.cgi?Dockey= 10001JJ4.txt.
U.S. EPA (Environmental Protection Agency). 1991b.
Guide for Conducting Treatability Studies under
CERCLA: Soil I rapor Extraction. Quick Reference
Fact Sheet. EPA/540/2-91/019B. Office of Solid
Waste and Emergency Response, Washington,
DC. September, https://nepis.epa.gov/Exe/ZyP
URL.cgi?Dockev= 10001K82.txt.
U.S. EPA (Environmental Protection Agency). 1993.
Presumptive Remedies: Site Characterisation and
Technology Selection for CERCLA Sites n-'ith I Tolatile
Otganic Compounds in Soils. Quick Reference Fact
Sheet. EPA 540-F-93-048. Directive: 9355.0-
48FS. Office of Solid Waste and Emergency
Response, Washington, DC.
https:/ / semspub.epa.gov/work/05/75775.pdf.
U.S. EPA (Environmental Protection Agency). 1995.
Manual: Bioventing Principles and Practice.
EPA/540/R-95/534a. Washington, DC.
U.S. EPA (Environmental Protection Agency). 1996a.
Engineering Fontm Issue Paper. Soil I ^apor Extraction
Implementation Experiences. Quick Reference Fact
Sheet. EPA/540/F-95/030. Publication 9200.5-
223FS. Office of Solid Waste and Emergency
Response, Washington, DC. April, https://clu-
in.org/download/remed/sveissue.pdf.
U.S. EPA (Environmental Protection Agency). 1996b.
User's Guide to the I'OCs in Soils Presumptive Remedy.
EPA/540/F-96/008. Directive No. 9355.0-6FS!
Office of Solid Waste and Emergency Response,
Washington, DC. July.
https: / / nepis.epa.gov/Exe / ZyPURL.cgiPDockey
=2000L28L.txt.
U.S. EPA (Environmental Protection Agency). 1996c.
Assessing UST Corrective Action Technologies: In Situ
SI E-Based Systems for Free Product Recovery and
Residual Hydrocarbon Removal. EPA/600/R-96/042.
Office of Solid Waste and Emergency Response,
Washington, DC.
U.S. EPA (Environmental Protection Agency). 1996d.
Soil Screening Guidance Technical Background Document.
EPA/540/R-95/128. Office of Solid Waste and
Emergency Response, Washington, DC.
https:/ /www.epa.gov/superfund/superfund-soil-
screening-guidance.
U.S. EPA (Environmental Protection Agency). 1997a.
Analysis of Selected Enhancements for Soil I Tapor
Extraction. EPA/542/R-97/007. Office of Solid
Waste and Emergency Response. Washington,
DC. http://clu-
in.org/download/remed/sveenhmt.pdf.
U.S. EPA (Environmental Protection Agency). 1997b.
Engineering Fontm Issue Paper. Soil I ^apor Extraction
Implementation Experiences. Quick Reference Fact
Sheet. EPA/540/F-95/030. Publication 9200.5-
223FS. Office of Solid Waste and Emergency
Response, Washington, DC. January, https://clu-
in.org/download/contaminantfocus / dnapl/Treat
ment Technologies /SVE-issue-97.pdf.
U.S. EPA (Environmental Protection Agency). 2001.
Development of Recommendations and Methods to Stpport
Assessment of Soil I Tenting Performance and Closure.
EPA/600/R-01/070. Office of Research and
Development. Ada, OK.
https://cfpub.epa.gov/si/si public record Repo
rt. c fm ? dirEn trvID=74386.
Soil Vapor Extraction (SVE) Technology
53
-------
U.S. EPA (Environmental Protection Agency). 2006.
Off-Gas Treatment Technologies for Soil I Tapor
Extraction Systems. EPA-542-R-05-028. Office of
Solid Waste and Emergency Response,
Washington, DC. March.
https: / / nepis.epa.gov/Exe / ZyPURL.cgiPDockey
=2000DlDW.txt.
U.S. EPA (Environmental Protection Agency). 2010.
Green Remediation Best Management Practices: Soil
I TaporTixtraction Air Sparging. EPA 542-F-10-
007. Office of Solid Waste and Emergency
Response, Washington, DC. March.
https: / /www, epa.gov/remedytech/green-
remediation-best-management-practices-soil-
vapor-extraction-air-sparging.
U.S. EPA (Environmental Protection Agency). 2012a.
A Citizen's Guide to Soil I TaporTixtraction and Air
Sparging EPA 542-F-12-018. Office of Solid
Waste and Emergency Response, Washington,
DC. September.
https:/ /www.epa.gov/remedvtech/citizens-guide-
soil-vapor-extraction-and-air-sparging.
U.S. EPA (Environmental Protection Agency). 2012b.
Petroleum Hydrocarbons and Chlorinated Solvents Differ
in TheirPotentialfor I ^aporIntrusion. Office of
Underground Storage Tanks, Washington, DC.
March, https://www.epa.gov/ust/petroleum-
hydrocarbons-and-chlorinated-solvents-differ-
their-potential-vapor-intrusion.
U.S. EPA (Environmental Protection Agency). 2015.
Technical Guide for Assessing and Mitigating the I rapor
Intrusion Pathway from Subsurface I ^apor Sources to
Indoor Air. OSWER Publication 9200.2-154.
Office of Solid Waste and Emergency Response,
Washington, DC. June.
Varadhan, R., and JA. Johnson. 1997. 11 .BACH: A
one-dimensionalfinite difference vadose ^vne leaching
model., v2.2. Center for Subsurface Modeling
Support, U.S. EPA, Ada, OK.
http: / /www.epa.gov/ada/csmos / models / vleach.
html.
54
Soil Vapor Extraction (SVE) Technology
-------
Appendix A.
Case Study #1, Assessment of a Small Persistent TCE Source
A.1 Background
Site 32 is in California and includes facilities for
aircraft maintenance and repair. Historical activities at
Site 32 and adjacent sites have resulted in soil, soil
gas, and groundwater contamination. Historical
operations included use and storage of
trichloroethylene (TCE). Infrastructure of interest at
Site 32 includes 2 underground storage tanks (USTs)
containing TCE that were removed in 1997, 13
former oil/water separators (OWSs) subsequently
removed, 2 vehicle wash racks, and 1 aircraft wash
pad. To address soil gas contamination at Site 32, a
soil vapor extraction (SVE) system was installed and
began operation in April 1998. The system consisted
of two vapor extraction wells (VE 2 and VE 3) and
five nested vapor monitoring points. The conceptual
site model, circa 1999, is illustrated in Figure A-l.
Until the year 2000, when two OWSs were removed
and VE 4 was installed, the SVE system operated
inefficiently with a stagnation zone located within a
source zone and pulled mass from a stronger,
unidentified source zone associated with the two
OWSs as illustrated in Figure A-l. Despite these
limitations, the system removed more than 1,000 lbs
of TCE.
The expanded SVE system was operated until 2008
with a total of five rebound periods spaced
throughout the operating period. The number of
wells operated was reduced and the operations
changed to pulsing in 2009 when rebound
concentrations were below cleanup goals in all
locations except the shallow screen of VE 4. The
extracted TCE concentration history at the system
manifold and the cumulative TCE mass removed are
plotted in Figures A-2 and A-3, respectively. As
shown in Figure A-2, the extracted TCE
concentration dropped below the maximum
contaminant level (MCL)-equivalent vapor
concentration of 350 ppbv in 2007, more than two
orders of magnitude below the initial concentration of
about 100 ppmv. The concentration decay resulted in
a diminishing mass extraction rate as illustrated in
Figure A-3 for the cumulative mass removed. The
addition of VE 4 in 2000 generated most of the mass
of TCE extracted after that date. Extraction at this
location eliminated the migration of TCE from this
area toward VE 2 and reduced the mass extraction
rate from VE 2. A decision to pursue permanent
shutdown of the SVE system was made in 2009 based
on the low volatile organic compound (VOC)
concentrations in rebound soil gas samples, very low
mass removal rates, and escalating cost per pound of
VOCs removed. All extraction was ceased in June
2009 and the SVE system was later dismantled.
Initial review of the SVE data and quarterly
monitoring of vapor concentrations through 2014
suggested the site could be closed. However,
California experienced an extended period of
drought-like conditions starting in 2013 and the
moisture content in the shallow vadose zone at Site
32 began to decrease. A soil gas sampling event in
early 2014 in the shallow screen of VE 4 yielded a
TCE concentration more than an order of magnitude
over the previous quarter and the regulatory
community rescinded a previous verbal agreement to
close the site pending an assessment of the remaining
source mass.
A review of the historical boring log for VE 4 and soil
sampling for physical analyses indicated a clayey silt
interval existed at VE 4 from a depth of about 25 to
35 ft below ground and likely created some degree of
perched water at 25 ft. The lateral extent of such
water was not known. The conceptual site model was
revised to account for this water and its disappearance
during the drought conditions exposing the interface
of the clayey silt to overlying silty sand. In effect, the
Soil Vapor Extraction (SVE) Technology
55
-------
barrier to volatilization was removed. The
contamination in the clayey silt was the result of
Bldg 1086
~~m
a
SVE system
^ f~
lit
Washrack
TCE Mass Removed
'97-1,100 lbs
'99-100 lbs
IGF
ows
BldgslO
r4/1076
DDn
VMP5
TCE Concentrations
in mg/m3
vmp 3
Silty Sand
= 363
Permeable Sands
_XZ
2 OWSs
3T |T) pi)
VE 4
1430
~
Figure A-1. Conceptual Site Model of Site 32 in 1999
•Cleanup Level (350 ppbv)
01/98 07/98 01/99 07/99 01/00 07/00 01/01 07/01 01/02 07/02 01/03 07/03 01/04 07/04 01/05 07/05 01/06 07/06 01/07 07/07 01/08 07/08 01/09
Figure A-2. Extracted soil gas TCE concentration history, Site 32
56
Soil Vapor Extraction (SVE) Technology
-------
Total VOCS
TCE
Mass primarily from VE-4 (-500 lbs)
01/98
01/99
01/02
01/03
01/04
01/05
01/06
01/07
01/08
01/09
Figure A-3 Cumulative Contaminant Mass Extracted, Site 32
overflow and leaks of TCE-laden water from the
OWSs that migrated downward and seeped into this
fine-grained interval and remained relatively pristine
while covered by a layer of high moisture content.
A.2 Field Test and Data Collection
Rather than drill a large number of borings to attempt
to define the lateral extent of the contamination, the
equivalent of a pilot vapor extraction test was
performed and initial decay and subsequent rebound
were evaluated to define the volume of contaminated
soil, the mass of TCE in this residual source, and the
mass transfer coefficient for calculating a mass flux
toward the surface. A small, 1 horsepower
regenerative blower powered by a standard single
phase, 120V, 20A wall outlet was used for the
extraction in the shallow interval of VE 4, screened
from 10 to 30 ft below the surface. Extracted vapors
were routed through granulated activated carbon
(GAC) prior to atmospheric discharge. The testing
was performed in four phases:
• Soil vapor extraction from October 1, 2014 to
October 4, 2014 (Phase I extraction)
• Rebound from October 4, 2014 to October 21,
2014
• Soil vapor extraction from October 21, 2014 to
December 17, 2014 (Phase II extraction)
• Rebound from December 17, 2014 to January
21,2015.
The rebound following the Phase II extraction
included a 6-hour extraction on January 14, 2015 at
well VE 4s with the collection and analyses of soil gas
samples for comparison with the previous extraction
periods following rebound.
SVE was not conducted at VE 4s between June 2009
and October 2014. It is assumed this long period of
dormancy before starting the mass transfer testing
allowed the site to re-equilibrate fully to ambient
conditions. A plot of the TCE vapor concentrations
measured in VE 4s as a function of the elapsed days
from the start of extraction is provided in Figure A-
Soil Vapor Extraction (SVE) Technology
57
-------
4. The concentration is plotted on a log scale to
illustrate the rapid initial decay associated with
sweeping vapors from the permeable soils followed
by a second, slower decay associated with mass
transfer limitations during the longer extraction
period of Phase II. These two trends are described in
more detail in the next section.
Comparing the initial TCE concentration in
Figure A-4 of about 600 ppmv with the historical
concentrations plotted in Figure A-2 demonstrates
the TCE concentration in 2014 was higher than that
measured before years of SVE at the site (-100
ppmv). Clearly the high moisture content atop the
clayey silt interval effectively sequestered the
dissolved TCE at the top of this interval. Comparing
the peak TCE concentration and decay during Phase I
extraction with that in Phase II reveals a rebound in
concentration between the two phases and a
decreased peak value. This behavior is repeated
between the Phase II extraction and the 6-hour
sampling event for rebound. In addition, the longer
period of extraction in Phase II yielded a much lower
peak TCE concentration for rebound that was more
than an order of magnitude less than the Phase I
peak. These observed trends display the classic decay
and rebound expected from the operation of SVE
when a significant fraction of the TCE mass is
removed from the soils.
The calculated cumulative TCE mass extracted
through VE 4s is plotted in Figure A-5. At the end
of the 3-day extraction in Phase I, the trend in the
cumulative mass extracted from VE 4s was clearly
increasing and led to the decision to implement the
longer-term Phase II extraction. The initial extraction
rate in VE 4s during Phase II was lower than Phase I
and the lesser rate is evident in Figure A-5. After the
rate was increased and over time, the mass extraction
rate from VE 4s approached an asymptote. At the
end of the Phase II extraction, the TCE mass
extraction rate was only 0.03 lbs/day. As indicated,
the total mass extracted was about 12.6 lbs.
A.3 Data Evaluation and Implications
To analyze the classic scenario of this test, a two-
region model of SVE developed by Praxis
Environmental Technologies and published as
Appendix F of the U.S. Army Corps of Engineers
SVE and Bioventing Engineers Manual is available.4
The vapor concentration data illustrated in
Figure A-4 and the cumulative mass extracted
illustrated in Figure A-5 at VE 4s were used to
calibrate the two-region SVE model. Application of
the model to this testing is illustrated conceptually in
Figure A-6.
The two-region model of SVE provides an estimate
for the residual mass, quantifies mass transfer
constraints, and evaluates possible future extraction
strategies. In short, the model volume-averages vapor
concentrations over the contaminated soil volume
(conceptualized as partially mobile/advective and
partially immobile/diffusive) and is based on an
overall mass balance. The following parameters are
assumed measured or otherwise available and
employed as input data by the model:
• h (m), depth interval of the contaminated soil
volume
• T (K), soil temperature
• H (kPa m3/mol), Henry's constant
• Kj,,„ (L/kg), the distribution coefficient in
mobile region
• Kja (L/kg), distribution coefficient in the
immobile region
• Dr (m2/day), pure component free air diffusion
coefficient for TCE
• f-pm, porosity of mobile region
• <$, porosity of immobile region
4 Reference: U.S. Army Corps of Engineers, 2002. Engineering
and Design: Soil Vapor Extraction and Bioventing. EM 1110-
1-4001.
58
Soil Vapor Extraction (SVE) Technology
-------
Rebound
Rebound
Elapsed Days from the Start of Phase I Extraction (0 = l-Oct-14)
Figure A-4. TCE vapor concentrations in extraction well VE 4s, Site 32
14
12.6 lbs
~ Calculated Mass Extracted
8
Model Extracted Mass
lu a
l ) o
Model Residual Mass
4
2
0.8 lbs
0
0 20 40 60 80 100 120
Elapsed Days from the Start of Phase I Extraction (0 = l-Oct-14)
Figure A-5. Cumulative TCE mass extracted from well VE 4s, Site 32
Soil Vapor Extraction (SVE) Technology
59
-------
Q|Ca
i Building on a Slab
b—I— *d *
t
. !
Zt~ 30 ft
1
Cd Diffusive Source
Clay Confining Layer
Advective Soil C=0
:
3
I
Figure A-6. Conceptual mode! of extraction from well VE 4s, Site 32
• (§/cm3)5 solid density of soil grains in
mobile region
• Si (g/cm3), solid density of soil grains in
immobile region
• S:„, water saturation in the mobile zone
• Si, water saturation in the immobile zone
• Q (scfm), total soil vapor extraction rate,
allowed to be a step-wise transient.
The model is calibrated by varying the following
parameters until finding the best match (minimum
error) with field measures of extracted vapor
concentration and cumulative mass removed:
• (jjjas (mg/m3), initial vapor concentration in the
mobile zone
• Cm (mg/m3), initial vapor concentration in the
immobile zone
• (m2)3 equivalent area of soil contamination
• f, fraction of the contaminated volume that is
characterized as mobile
• a (day4), bulk mass transfer coefficient
between mobile/immobile: zones.
60
In general, if the site has been dormant, the initial
mobile and immobile zone concentrations are in
equilibrium and are approximated by the early
extraction concentration. The area of soil
contamination with the depth interval represents the
total soil volume used in the volume averaging. The
mass transfer constraints for contaminant removal are
lumped (averaged) into a bulk mass transfer
coefficient, a. When molecular diffusion alone from
fine-grained soils (immobile soils) provides the
constraint, a, is roughly related to diffusion by:
^p^rci-s.r3^2
Ri Li
JU is the characteristic path length for vapor diffusion
and R is the vapor retardation coefficient for TGK in
the immobile region. The inclusion of the porosity
and water saturation in tins coefficient represents the
tortuosity for diffusion through the soil. The
characteristic length is most often correlated with the
halt-thickness ot fine-grained layers such as clays.
Soil Vapor Extraction (SVE) Technology
-------
Calibration of the model is achieved by finding best
fits of the model parameters to the concentration and
mass extraction data. Fitting is performed by varying
the soil volume, fraction of mobile soil, and mass
transfer coefficient in a downhill simplex optimization
routine. The fit to the TCE vapor concentrations
measured during the mass transfer testing in VE 4s is
shown in Figure A-7. The model fit to the extracted
mass of TCE and the estimated residual mass of TCE
remaining in the soil is plotted in Figure A-5. The
specified soil properties and the resulting best-fit
model parameters are listed in Table A-l.
Average initial TCE vapor concentrations in both
zones were assumed to be in equilibrium at 636 ppmv
over a total soil volume of 2,700 yd3 yielding an initial
total mass estimate of 13.4 lbs of TCE in the vadose
zone at the start of the mass transfer test. The model
fit yielded 40% for the fraction of the vadose zone
characterized as mobile, leaving 60% as immobile.
Hence, the initial mass of TCE was 5.0 lbs in the
mobile soils and 8.4 lbs in the immobile soils. In total,
the mass transfer testing removed about 12.6 lbs of
TCE, leaving an estimated 0.8 lbs of residual TCE
mass as indicated in Figure A-5. During the three
days of extraction in Phase I, just over 5 lbs of TCE
were extracted, which corresponds roughly to the
initial mass in the mobile soils and suggesting Phase I
provided a single flush of the mobile soil pore
volume. Recall, years of SVE extracted over 1,600 lbs
of SVE; yet, a residual mass of about 13 lbs required
additional extraction to complete the cleanup after the
moisture content decreased.
The characteristic volume of contaminated soil was
calculated to be 2,700 yd3. If the vadose zone
thickness is assumed equal to the VE 4s screen length
of 20 ft, the effective area was 3,660 ft2 with a
corresponding circular radius of 34 ft. Using the
depth from the ground surface to 30 ft, yields a
circular radius of 28 ft. This calculated effective area
is much smaller than the original source zone for Site
32 North but consistent with a small residual hot
spot.
The bulk mass transfer coefficient of 0.0854 day"1 is
associated with a characteristic diffusion path length
of 0.91 ft and indicates a re-equilibration timescale of
about 30 days. Vapor retardation coefficients in the
mobile and immobile zones were calculated to be 2.6
and 3.6, respectively.
The results of the field testing and data evaluation are
in summary:
• Estimated initial mass of TCE within the
influence of extraction at VE 4s was 13.4 lbs
• Estimated mass of TCE extracted from VE 4s
was 12.6 lbs
• 3 days of extraction in Phase I removed 40%
of the original TCE mass
• 57 days of extraction in Phase II removed 54%
of the original TCE mass
• 6% of the original TCE mass remains
• Estimated residual mass of TCE proximate to
VE 4s was about 0.8 lbs
• At the end of Phase II extraction, the TCE
mass removal rate was 0.03 lbs/day and
continuing to decline
• Maximum theoretical vapor concentration near
VE 4s after mass extraction was 35 ppmv,
down from the initial concentration of 635
ppmv, a 95% reduction.
Closure of this site remains under review by
regulators.
Soil Vapor Extraction (SVE) Technology
61
-------
O Measured Concentration
Model Fit
Elapsed Days from the Start of Phase I Extraction (0 = l-Oct-14)
Figure A-7. Model fit to TCE vapor concentrations in well VE4s, Site 32
Table A-1. Parameters for Two-Region Modeling of TCE Extraction from VE 4s
Property
Units3
VE 4s
Measured or Assumed Soil Properties
Porosity, mobile
nd
0.40
Porosity, immobile
nd
0.40
Water Saturation, mobile (vol/vol)
nd
0.25
Water Saturation, immobile (vol/vol)
nd
0.40
Fraction of organic carbon in soil solids (foc)
nd
0.0004
Kd, mobile
Ukg
0.05
Kd, immobile
Ukg
0.05
Temperature
°C
20
Initial Vapor Concentration, mobile
ppmv
636
Property
Units3
VE 4s
Initial Vapor Concentration, immobile
ppmv
636
TCE Properties
Henry's Constant
nd
0.38
Octanol-Water Partition Coefficient
nd
200
Diffusion Coefficient in Air
m2/day
0.68
Model Best-Fit Parameters
Effective Volume of Contaminated Soil
m3
2,073
Fraction of Soil Characterized as Mobile
nd
0.40
Fraction of Soil Characterized as Immobile
nd
0.60
Bulk Mass Transfer Coefficient
1/day
0.0854
a nd = dimensionless
62
Soil Vapor Extraction (SVE) Technology
-------
Appendix B.
Case Study #2, Assessment of Mass Transfer
across the Capillary Fringe from
Contaminated Groundwater to Vadose Zone Soils
B.1 Background
At a site contaminated with a chlorinated solvent
(trichloroethylene or TCE), soil vapor extraction
(SVE) was selected and applied to clean up the
contaminated soil. After about 2 years of SVE,
rebound monitoring was used to assess the TCE mass
volatilized into the vadose zone and to assess the
potential mass extraction rate of TCE volatilized from
contaminated groundwater. This case study also
illustrates the use of pneumatic logging to characterize
the permeability of different vadose zone layers of
interest to the assessment.
Rebound monitoring was applied in a single
groundwater monitoring well (MW-07) with exposed
screen at the bottom of the vadose zone. The water
table near MW-07 (screened from 77.5 to 127.5 feet
below ground surface [ft bgs]) was historically
shallower and has fallen about 20 ft over the past 20
years to roughly 97 ft bgs. The deeper water table
exposes about 20 ft of the upper screen in MW-07 to
unsaturated soils making the well amenable to vapor
extraction. MW-07 is located about 500 ft
downgradient of the source zone where TCE releases
occurred and along the centerline of a dissolved TCE
plume emanating from this source zone. Nested soil
vapor monitoring probes are near the well.
The conceptual model for contaminant mass available
for vapor extraction around this well consists of
dissolved mass transported with groundwater from
the original source zone that subsequently volatilized
and diffused upward, resulting in TCE mass in the
vadose zone that has the potential to pose an
exposure risk via vapor intrusion. The available mass
for upward diffusion increased as the water table fell,
leaving soil partially saturated with immobile
contaminated water (formerly groundwater) at the
bottom of the vadose zone. In addition, a relatively
low permeability unit in the middle of the vadose
zone acts as a cap that constrains the rate of TCE
diffusion upward to the surface.
SVE operation and rebound testing were performed
from 2011 through 2016. The SVE system operated
at groundwater monitoring well MW-07 from
November 30, 2011 until March 18, 2014 when SVE
was terminated to allow the observation of rebound
in the vadose zone. Rebound monitoring continued
through October 28, 2015 when the SVE system was
restarted and continued until April 26, 2016 when
SVE was terminated. Monitoring of rebound
continues at the site as of the date of this report
(December 31, 2016).
B.2 Vertical Profiling along the SVE Well
Screen
Before starting long-term SVE in MW-07, PneuLog
profiles were measured on September 14, 2011. A
temporary SVE system operated for 71 minutes to
obtain the profiles. Logging was initiated after about
25 minutes of extraction at 90 standard cubic feet per
minute (scfm) with an applied vacuum of 17 inches
H20. The interpreted vertical profiles for soil
permeability to vapor flow and TCE soil vapor
concentration are shown in Figure B-l.
Figure B-l illustrates mostly uniform and permeable
soils from above the top of the screen down to a
depth of 91 ft bgs with a narrow, less permeable zone
from about 81 to 83 ft bgs. The effective
permeability suggests the entire screen interval is
sandy. No flow was detected below 91 ft bgs down to
the water table, although the exact depth of the water
table was not measured. Uncertainty in the profiles at
the bottom of the well are indicated by the dashed
lines.
Soil Vapor Extraction (SVE) Technology
63
-------
75
80
85
CJj
•o
n.
4)
O
90
95
100
t
0.001 0.01 0.1 1
Effective Permeability (scfm/ft/inH20)
J
J
1
50 100 150 200 250
Soil Gas TCE Concentration (mg/m3)
Effective vapor permeability and interpreted soil gas profiles in MW-07
Figure B-1.
The measured TCE vapor concentrations were
relatively uniform across the screen interval although
a vapor sample from 81.5 ft bgs at the transition to a
lesser permeable soil was low, suggesting a lesser
vapor concentration in the less permeable soils from
81.5 to 87 ft bgs. The permeability of this interval was
roughly half that of the adjacent soil layers indicating
this layer was not acting as a source of residual TCE
mass for the surrounding, more permeable soil. SVE
had not been performed in this well other than brief
testing and this contributed to the relative uniformity
of the vapor concentrations. The vapor sample
collected just above the water table yielded a TCE
concentration of 174 mg/m3. If this vapor sample
were in equilibrium with porewater (e.g., the capillary
fringe and underlying groundwater), the water
concentration would be on the order of 0.45 mg/L.
The geologic log for this well indicates gravel-
dominated soil from 75 to 95 ft bgs except for a silty
sand from 80 to 85 ft bgs. The PneuLog effective
permeability profile is in general agreement. Below 95
ft bgs, the geologic log recorded mostly silt-
dominated soils with some sand intervals down to the
bottom of the boring at 135 ft bgs. These
observations suggest that a continued decline in the
water table is not expected to increase vapor flow
from the well as the already exposed gravel intervals
will control the flow with little appreciable addition
from deeper silty soils. This log also suggests lesser
permeable clays and silts provide an upper cap at 45
to 50 ft bgs for SVE applied in the deeper sand and
gravel zones.
The falling water table over the past 20 years and the
interpreted TCE concentration profile in MW-07
suggest contaminated groundwater was stratified in its
transport through the aquifer before the water table
declined. Historically, contaminated groundwater
likely flowed through sandy gravels above and below
the silty sands logged from 80 to 85 ft bgs. During
that time, the slow rate of liquid diffusion from gravel
Soil Vapor Extraction (SVE) Technology
-------
intervals into the silty sands may have left the lesser
TCE concentration observed from 81.5 to 87 ft bgs
in the PneuLog profile.
B.3 Field Operations for SVE and Data
Collection
SVE was initiated in MW-07 on November 30, 2011
and continued until March 18, 2014. During this
operation, vapor concentrations and flows were
measured in MW-07 and vapor concentrations were
measured in nearby soil vapor monitoring points. The
measured vapor concentrations of TCE in vapors
extracted from MW-07 are plotted in Figure B-2.
During the first few months of extraction, the
concentration decayed by about two-thirds, from an
initial 170 mg/m3 to roughly 60 mg/m3. This range
corresponds roughly to the TCE concentrations
observed in the PneuLog profile for productive sandy
gravels (-200 mg/m3) and in the silty sand interval
(-60 mg/m3). Over the subsequent 20-month
extraction period, the concentration decreased at a
much slower rate, approaching a decaying asymptote
around 20 mg/m3.
The calculated cumulative TCE mass extracted
through MW-07 is plotted in Figure B-3. At the end
of the extraction, the trend in the cumulative mass
extracted was decaying but significant mass removal
continued. The total mass of TCE extracted from
MW-07 was estimated to be 320 lbs during the 20-
month extraction period with an initial rate of
1.4 lbs/day and a final rate of about 0.22 lbs/day.
B.4 SVE Data Evaluation
Modeling of the SVE results utilized the two-region
model of SVE developed by Praxis Environmental
Technologies5 but the model was modified to include
volatilization from contaminated groundwater as a
source for extracted mass. The initial response to
SVE in MW-07 is consistent with the two-region
model as TCE vapors emanating from contaminated
groundwater had decades to diffuse upward and into
less permeable (immobile) soils. However, the long-
term, asymptotic behavior at MW-07 is consistent
with underlying groundwater as a significant source of
extracted mass given the lack of other historical
sources in the vicinity.
To accommodate contaminated groundwater as a
source of extracted mass during SVE at MW-07, the
two-region (mobile-immobile) soil model was
extended to include mass transfer across the water
table as a boundary condition. The expanded model
including contaminated groundwater is described in
Section 5.2 of the main EIP as represented by
d,Ca Q
Ra~dt=~Vsoil(p{l-S)faCa
+ «d(^)(Q-ca)
\/a'
+ Otgw(HCgw ~ Ca)
The mass transfer coefficient at the interface between
the bottom of the vadose zone and groundwater is
difficult to predict but can be estimated from SVE
operations when contaminated groundwater becomes
the primary source of extracted mass. This model
formulation is identical to the U.S. Army Corps of
Engineers mobile/immobile model if the mass
transfer coefficient at the water table is set to zero.
The model concept is illustrated in Figure B-4.
Extracted vapors sweep through more permeable
(mobile) soils first yielding a relatively rapid decay in
extracted concentration (Cm) and then a slow decay in
diffusion-limited mass transfer from groundwater.
5 Published as Appendix F of U.S. ACE (Army Corps of
Engineers). 2002. "Engineering and Design: Soil Vapor Extraction
g. Engineer Manual EM 1110-1-4001. U.S. ACE,
Washington, DC. June 3. http://www.publications.usace.army
Soil Vapor Extraction (SVE) Technology
.mil / Portals / 76 / Publications / EnpineerManuals / EM 1110-1-
4001.pdf
65
-------
200
f 180
ao
E_
c
o
LU
rv
o
160
140
120
100
80
60
40
20
o
\
* o
\
\
\
\
\
¥
~ \
\
\
\
o
4^
~
O 4
. o
A
-Ol-.
* * V
o|
1
1
1
Oi
1 ^
o
o
o
11/1/11 3/1/12 7/1/12 10/31/12 3/1/13 7/1/13 10/31/13 3/1/14
Figure B-2. MW-07 TCE vapor concentrations during SVE
350
~ Extracted Mass
"5 300
Model Extraction
Q.
250
200
150
100
50
o
11/1/11 3/1/12 7/1/12 10/31/12 3/1/13 7/1/13 10/31/13 3/1/14
Figure B-3. Cumulative TCE mass extracted from MW-07 during SVE
66
Soil Vapor Extraction (SVE) Technology
-------
Homes and Office Buildings
WaterTable
Fluctuation
Open Space
Sandy Soil
Silt & Clay Layer
Sand/Gravel Soil
C ~ 0.7-6 mg/L
U Groundwater
Aquitard
Source Area
Figure B-4. Conceptual model of extraction from MW-07
The mass transfer characteristics for TCE extracted
from MW-07 were estimated initially by fitting the
extended SVE two-region model to the measured
flow rates and TCE concentration data through
March 18, 2014. Additional parameters to include
contaminated groundwater as a TCE source are a
mass transfer coefficient for volatilization across the
water table and the average concentration of TCE
dissolved in the groundwater. As described above, the
early rapid decay in concentration corresponds to the
sweep of the permeable soil volume holding vapors
volatilized previously from the groundwater over a
long period as well as vapors from residual porewater
left by the retreating water table. The subsequent
longer decay in concentration is associated with the
volatilization of TCE from contaminated
groundwater and from porewater in vadose zone soils
with high moisture contents. The volatilization rate
from groundwater is governed by the mass transfer
coefficient, the groundwater TCE concentration, the
interfacial area between groundwater and vadose
zone, and the vapor extraction rate. The interfacial
area is estimated as the volume of impacted vadose
zone soil divided by the depth interval of impact in
the vadose zone. Based on the geologic log and
Soil Vapor Extraction (SVE) Technology
concentrations in vapor monitoring points, this
interval is assumed to extend from the water table up
to the bottom of the confining clays and silts in the
middle of the vadose zone (-50 ft = 97 ft bgs to 47 ft
bgs) illustrated in Figure B-4.
Best fits of the model parameters to the initial
extraction and concentration data were achieved using
an optimization routine that converged to the fit
shown in Figure B-5. The specified soil properties
and the resulting best-fit parameters including the
representative soil volume, initial contaminant
concentrations, and volumetric mass transfer
coefficients are provided in Table B-l.
The fraction of soil characterized as mobile or
permeable around MW-07 was 0.56, consistent with
the geologic log and PneuLog permeability profile.
The groundwater concentration was assumed to be
constant during SVE, that is, the extracted mass from
the groundwater was considered negligible compared
to the total mass in the groundwater traveling through
the zone. The average of 10 monthly groundwater
samples collected in 2011 before SVE was 2,800
Mg/L, and in 2014, after SVE ceased the average was
-------
&o
E
<*
V .
XI c
o
o
o
o <
o
o
O
~
11/1/11 3/1/12 7/1/12 10/31/12 3/1/13 7/1/13 10/31/13 3/1/14
Figure B-5. MW-07 TCE vapor concentrations measured and modeled during extraction
Table B-1. Parameters for Two-Region Modeling of TCE Extraction at MW-07
Property
Units
2011 Source Zone
2014 MW-07
Measured and Assumed Soil Properties
Porosity, mobile
0.45
0.45
Porosity, immobile
0.45
0.45
Water Saturation, mobile (vol/vol)
0.25
0.25
Water Saturation, immobile (vol/vol)
0.40
0.50
Kd, mobile
L/kg
0.05
0.05
Kd, immobile
L/kg
0.05
0.05
Temperature
°C
20
20
TCE Properties
Henry's Constant
—
0.38
Octanol-Water Partition
—
200
Diffusion Coefficient in Air
m2/day
0.68
Measured/Assumed TCE Concentrations
Initial Vapor Concentration, mobile
mg/m3
840
180
Initial Vapor Concentration, immobile
mg/m3
840
180
Average Groundwater Concentration
mq/l
—
1,400
Model Best-Fit Parameters
Effective Volume of Contaminated Soil
m3
1,221,833
650,000
Fraction of Soil Characterized as Mobile
nd
0.24
0.56
Bulk Mass Transfer Coefficient Mobile/Immobile Soils
1/day
0.00073
0.002
Bulk Mass Transfer Coefficient with Groundwater
1/day
—
0.0002
2,600 (Jg/L. However, assuming an average
groundwater TCE concentration of 2,700 (Jg/L as the
source over-predicted the volatilization rate. A good
fit to the data was achieved with a lower groundwater
concentration of 1,400 (Jg/L and indicates the
measured groundwater concentrations may be
representative of conditions closer to the bottom of
the screen interval.
Soil Vapor Extraction (SVE) Technology
-------
Average initial TCE vapor concentrations were
assessed to be 180 mg/m3 and in equilibrium between
the mobile and immobile soils at the start of
extraction. These vapor concentrations are consistent
with the values measured during the PneuLog test,
the initial vapor extraction in MW-07, and in the deep
screens of vapor monitoring points. Hence, the
primary fitted parameters are the total vadose zone
volume impacted by TCE vapors, the fraction of the
volume through which vapors flow (mobile), the mass
transfer coefficient between mobile and immobile
soils, and the mass transfer between the vadose zone
and underlying groundwater. The data fit with the
model minimized the difference between measured
and modeled extraction concentrations at MW-07 as
well as matching the total mass extracted.
The model extracted mass of TCE during SVE in
MW-07 is plotted in Figure B-6 and compared with
the calculated mass based on measured
concentrations and extraction rates. The model also
estimates the initial mass of TCE in the impacted
vadose zone surrounding MW-07. Calculated from
initial vapor concentrations and soil properties, the
estimate was 234 lbs. As indicated, the total mass of
TCE extracted was about 320 lbs and exceeded the
initial estimate. The balance was the result of TCE
volatilization from contaminated groundwater. At the
end of the extraction in March 2014, the mass
extraction rate was nearing a balance with the
volatilization rate from groundwater at a rate of about
0.16 lbs/day and an estimated 117 lbs of TCE had
been volatilized and extracted from underlying
groundwater leaving about 30 lbs in the vadose zone
influenced by SVE in MW-07.
B.5 Rebound Data Evaluation
After vapor extraction was terminated in MW-07 on
March 18, 2014, vapor samples were periodically
collected and analyzed from MW-07 and surrounding
soil vapor monitoring points. This activity continued
until vapor extraction was restarted on October 27,
2015 and operated continuously through April 26,
2016. The rebound data are reflective of TCE vapor
concentrations in the vadose zone immediately
around MW-07 while the second extraction period
provides data directly comparable to the initial
extraction effort. The model calibrated to the initial
extraction data in the previous section was run to
include the subsequent rebound and second
extraction periods for validation. During rebound
with no extraction, mass transfer between the vadose
zone domains and groundwater seek a new
equilibrium using the disequilibrium in concentrations
at the end of SVE as the initial condition. Similarly,
the conditions at the end of the rebound period
provide the initial conditions for the second
extraction period. TCE vapor concentrations
measured during the initial extraction, rebound, and
second extraction in MW-07 are plotted in
Figure B-7 along with model forecasts for conditions
after March 18, 2014.
The start of the rebound is marked by the upward
inflection in the model mobile soil concentration. In
general, vapor concentrations in MW-07 decreased
immediately after shutdown but then started to
increase after a lag period of about 5 months. This
concentration was expected to continue climbing
slowly with volatilization of TCE from underlying
groundwater and was measured to be 83 mg/m3 at
the start of the second extraction period after 20
months of rebound consistent with the model
prediction. The lag time of 5 months in rebound may
be associated with higher moisture contents found in
deeper vadose zone soils above the water table.
Notice the model diffusive soil volume acts as a sink
for volatilized TCE from the underlying groundwater
that enters the permeable soil and diffuses upward
throughout the deep vadose zone. The start of the
second extraction period is marked in Figure B-7 by
the sudden downward inflection in the extracted TCE
concentration on October 27, 2015. The measured
concentrations in MW-07 decayed more rapidly than
predicted by the model. As described previously, this
initial decay during extraction is governed primarily by
the volume of permeable soil holding contaminated
vapors within the influence of the SVE well. Hence,
Soil Vapor Extraction (SVE) Technology
69
-------
350
300
250
200
150
100
50
I
~ Extracted Mass
Model Extraction
Model Vadose Zone
Model GW Volatilization
11/1/11 3/1/12 7/1/12 10/31/12 3/1/13 7/1/13 10/31/13 3/1/14
Figure B-6. MW-07 TCE mass extraction and mass balance, Case Study #2
200
~ Measured Extraction
O Measured Rebound
Model Extraction - Mobile
Model Diffusive - Immobile
EbO
E 160
.2 140
11/1/11 7/1/12 3/2/13 10/31/13 7/2/14 3/2/15 11/1/15 7/1/16
Figure B-7. MW-07 TCE vapor concentrations measured and modeled during rebound
the volume of TCE impacted soil during rebound was
less than the original volume. The TCE vapor
concentrations measured during extractions and
rebound in the three deepest points of nearby soil
vapor monitoring point S\M 16 at depths of 56, 68
and 80 ft bgs are plotted m Figure B-8 along with
MW-07 and the model forecast for the mobile soil.
The concentrations are plotted on a logarithmic scale
to illustrate the trends. All three points in SVM-16
displayed a steep decay with the initial extraction in
Soil Vapor Extraction (SVE) Technology
-------
MW-07 and remained low. Shortly after ceasing
extraction, the shallower depths of 56 ft bgs and 68 ft
bgs began to rebound while the deeper point at 80 ft
bgs did not show a response until extraction was
restarted. All three depths remained at least one order
of magnitude below the initial concentrations
measured before SVE was started in the more deeply
screened MW-07. The low, rapid responses observed
at 56 and 68 ft bgs may be influenced by continued
mass transfer from the immobile soils or they may
reside in more permeable soil (lower moisture
content) that communicates more rapidly with the
groundwater. All the concentrations are expected to
continue climbing slowly without SVE as a result of
the slow volatilization of TCE from underlying
contaminated groundwater and could approach the
initial values of 2011 over decades.
The model mass for TCE in the vadose zone
increases during rebound as a result of the
volatilization from groundwater as the site seeks a
new equilibrium. This increasing mass is indicated in
Figure B-9 and coincides with the increased
cumulative mass of TCE volatilized from
groundwater during rebound. The volatilized mass
entering the vadose zone during rebound was
estimated to be 75 lbs or about 0.13 lbs/day. The
initial mass in the vadose zone at the start of the
second extraction period was modeled to be 108 lbs.
The measured and modeled extracted masses of TCE
during the post-rebound SVE application in MW-07
are also plotted in Figure B-9. As indicated, the
measured mass of TCE extracted during the second
period was about 45 lbs and was about half of the
model estimate of 85 lbs. The discrepancy is the
direct result of the measured extraction concentration
decaying more rapidly than predicted by the model as
was shown in Figure B-7. However, toward the end
of the extraction in April 2016, the mass extraction
rate was approaching the previously observed balance
with the volatilization rate from groundwater at a rate
of about 0.16 lbs/day.
At the start of the second extraction period on
October 27, 2015, the measured concentrations in
MW-07 decayed more rapidly than predicted by the
model. This initial decay was governed primarily by
the volume of permeable soil holding contaminated
vapors within the influence of the SVE well. Hence,
the volume of TCE impacted soil during rebound was
less than the original volume. The lesser volume is the
result of incomplete re-equilibration with the
underlying groundwater. For the two-domain model,
the mobile and immobile domains are somewhat
uniformly distributed and the vapor extraction
exchanges the pore volume in the mobile domain
rapidly. However, the underlying contaminated
groundwater is not uniformly distributed but rather
lies along one boundary of the domain and therefore
its impact is not uniformly observed until near re-
equilibration occurs. In other words, only the deeper
vadose zone soils above the water table were
impacted during the 20-month dormancy, while
shallower vadose zone soils originally impacted were
not yet equilibrated.
B.6 Implications for Future Operation of SVE
at MW-07
Including contaminated groundwater as a source of
mass extracted in the vapor phase yields the near
asymptotic value in vapor extraction concentration
observed in MW-07. The data fit yielded a bulk
(volumetric) mass transfer coefficient for
volatilization across the capillary fringe of 0.0002 day"1
with an average groundwater TCE concentration of
1,400 (Jg/L. These values, assuming the average
groundwater concentration is constantly recharged,
yield an asymptotic TCE mass volatilization rate of
about 0.15 lbs/day that is somewhat independent of
the extraction rate if the extracted concentration is
maintained at a low value compared to the
groundwater contamination.
Soil Vapor Extraction (SVE) Technology
-------
11/1/11 7/1/12 3/2/13 10/31/13 7/2/14 3/2/15 11/1/15 7/1/16
1000
¦Model Extraction
• MW-07
Figure B-8. TCE vapor concentrations measured and modeled during rebound
450
~ Extracted Mass
~ 400
Model Extraction
0 350
Q.
1 300
Model Vadose Zone
Model GW Volatilization
jj 250
l*s
2 200
150
100
50
11/1/11 7/1/12 3/2/13 10/31/13 7/2/14 3/2/15 11/1/15 7/1/16
Figure B-9. MW-07 TCE mass extraction and mass balance during rebound
72
Soil Vapor Extraction (SVE) Technology
-------
Design of additional SVE or pulsing can be based on
the extraction rate or exchange rate in an effort to
balance the mass transfer processes. Consider the
governing equation for the extracted concentration,
W>„n -.v,j/¦;,«„]
d£l
dt
=+ ad(c£ - Cf)
+ a (HCW — C3}
' ^gw v ^gw )
Scaling arguments yield estimates for extraction rates
by comparing the extraction term with the
groundwater volatilization while neglecting the faster
mobile/immobile domain mass transfer,
— Cg » a (HCW — Cg}
y ^ ^gwy11 W?w LaJ
Q ^
f LJ/~*W
n ugw
CL
^ rt
CL
^ n
&gw
- 1) ~ 9 agw Vt
Q » 9 (0.0002 day'1) (650,000 m3)
3
m
= 1,100 —— = 29 scfm
day
In this estimate, the driving concentration gradient
for mass transfer was assumed to be held at 90% of
the maximum (Cextmct ~ 17 mg/m3) and yielded an
extraction rate significantly lower than the flow
sustained during the extraction test.
This first order estimate suggests an extraction rate as
low as 30 scfm may be sufficient to keep
concentrations low over a large area. The low rate is a
direct result of the low volatilization rate from
groundwater.
B.7 Implications for Future Operation of SVE
at MW-07
Including contaminated groundwater as a source of
mass extracted in the vapor phase yields the near
asymptotic value in vapor extraction concentration
observed in MW-07. The data fit yielded a bulk
(volumetric) mass transfer coefficient for
volatilization across the capillary fringe of 0.0002 day"1
with an average groundwater TCE concentration of
1,400 (Jg/L. These values, assuming the average
groundwater concentration is constantly recharged,
yield an asymptotic TCE mass volatilization rate of
about 0.15 lbs/day that is somewhat independent of
the extraction rate if the extracted concentration is
maintained at a low value compared to the
groundwater contamination.
Design of additional SVE or pulsing can be based on
the extraction rate or exchange rate to balance the
mass transfer processes. Consider the governing
equation for the extracted concentration provided in
Section 5.2,
W>„n -
d£l
dt
= + ai(e" -
+ a (HCW — Cg>\
' ^gw v ^gw )
Scaling arguments yield estimates for extraction rates
by comparing the extraction term with the
groundwater volatilization while neglecting the faster
mobile/immobile domain mass transfer,
— C3»a (HCW — C3}
y La " "-gwy'^gw La J
Q » agwVt
&gw
-------
This first order estimate suggests an extraction rate as
low as 30 scfm may be sufficient to keep
concentrations low over a large area. The low rate is a
direct result of the low volatilization rate from
groundwater. The estimated bulk mass transfer
coefficient of 0.0002 day-1 suggests a re-equilibration
timescale of about 10 years. These results have
implications for designing low-level SVE to mitigate
vapor intrusion by capturing contaminant mass
before it diffuses upward toward the surface. The
testing described in this case study has applicability
for design at any site where contaminated
groundwater is a potential source for vapor intrusion.
For example, even as vadose zone remediation by
SVE is terminated, keeping a portion of the
infrastructure in place is recommended if
groundwater contamination remains. Mitigation of
vapor intrusion over a large area encompassing
multiple buildings may be as simple as a brief,
periodic (e.g., annual) flush of the vadose zone vapor
pore volume using a small, portable blower on a well
installed previously for SVE (i.e., free for mitigation
purposes).
74
Soil Vapor Extraction (SVE) Technology
-------
vvEPA
United States
Environmental Protection
Agency
PRESORTED
STANDARD
POSTAGE & FEES
PAID EPA
pcpmit Mi-1 n.
Office of Research and Development
(8101R) Washington, DC 20460
Offal Business
Penalty for Private Use$300
Soil Vapor Extraction (SVE) Technology
75
------- |