United States
Environmental Protection
Agency
Research and
Development
(RD681)
EPA/540/A5-89/003
July 1989
EPA
Terra Vac In Situ
Vacuum Extraction
System
Applications Analysis Report
SUPERFUND INNOVATIVE
TECHNOLOGY EVALUATION
-------
-------
EPA/540/A5-89/003
July 1989
Terra Vac In Situ Vacuum
Extraction System
Applications Analysis Report
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Notice
The information in this document has been funded'wholly or in part by the U.S.
Environmental Protection Agency under the auspices of the Superfund Innovative
Technology Evaluation (SITE) Program and under Contract No. 68-03-3255 to Foster
Wheeler Enviresponse, Inc. It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute an endorsement or
recommendation for use.
11
-------
Foreword
The Superfund Innovative Technology Evaluation (SITE) program was
authorized in the 1986 Superfund amendments. The program is a joint effort
between EPA's Office of Research and Development and Office of Solid Waste
and Emergency Response. The purpose of the program is to assist the develop-
ment of hazardous waste treatment technologies necessary to implement new
cleanup standards which require greater reliance on permanent remedies.
This is accomplished .through technology demonstrations which ctre designed
to provide engineering and cost data on selected technologies.
This project consists of an analysis of Terra Vac's in situ vacuum
extraction process, which was demonstrated on the property of an operating
machine shop. The property is part of the Groveland Wells Superfund Site
in Groveland, Massachusetts. The demonstration effort was directed at
obtaining performance and cost information which would be useful for assess-
ments at other sites. Documentation will consist of two reports. The
Technology Evaluation Report (EPA/540/5-89-003a) describes the field activities
and laboratory results. This Applications Analysis provides an interpretation
of available data and discusses the potential applicability of the technology.
Additional copies of this report may be obtained at no charge from
EPA's Center for Environmental Research Information, 26 West Martin Luther
King Drive, Cincinnati, Ohio, 45268, using the EPA document number found ort
the report's front cover. Once this supply is exhausted, copies can be
purchased from the National Technical Information Service, Ravensworth
Bldg., Springfield, VA, 22161, (702) 487-4600. Reference copies will be
available at EPA libraries in their Hazardous Waste Collection, You can
also call the SITE Clearinghouse hotline at 1-800-424-9346 or 382-3000 in
Washington, D.C. to inquire about the availability of other reports
Margare^ M. Kelly, Dire
Technology Staff, Offij
Program Management and
Technology
Director,
rironmental Engineering
anci-Technology Demonstration
iu
-------
Abstract
This document is an evaluation of the Terra Vac in situ vacuum extraction system and
its applicability as a treatment method for waste site cleanup.
This report analyzes the results from the Superfund Innovative Technology
Evaluation (SITE) Program's 56-day demonstration at the Valley Manufactured
Product Company's site in Groveland, Massachusetts and data from other
applications. Conclusions were reached concerning the technology's suitability for use
in remediations involving both similar and different materials at other sites.
Operational data and sampling and analysis information were monitored carefully to
establish a database against which vendor's claims for the technology could be
evaluated.
The conclusions from the results of the Groveland demonstration test and from other
available data are: (1) the process can be used to remediate a site contaminated with
VOCs; (2) the process can remove VOCs from soils with permeabilities 'as low as 10-8
cm/s; (3) the process operates well in all weather conditions; and (4) the process
implementation costs can be as low as $10/ton, depending on various site-specific
conditions.
IV
-------
Contents
Page
Foreword iii
Abstract iv
Tables vi
Figures vii
Abbreviations and Symbols viii
Conversions xi
Acknowledgments xii
1. Executive Summary 1
2. Introduction 3
The SITE Program 3
Site Program Reports 3
Purpose of the Applications Analysis Report 4
Key Contacts 4
3. Technology Applications Analysis 5
Introduction 5
Conclusions 5
Application of Vacuum Extraction Technology 7
Environmental Regulation Requirements for the Technology 8
Materials Handling Required by the Technology 9
Personnel Issues '. 10
Testing Procedures 10
4. Economic Analysis 11
Introduction 11
Results of Economic Analysis 11
Site-Specific Factors Affecting Cost 12
Basis of Economic Analysis 13
Appendices
A. Process Description 17 ,
B. Vendor's Claims for the Technology 21
C. SITE Demonstration Results 27
Bibliography 34
D. Case Studies 35
-------
Tables
Number Page
4-1 Estimated Cost 12
A-l Groveland Site Equipment List 18
A-2 System Variables 19
A-3 Dimensionless Henry's Law Constants for Typical VOCs at 10°C ... 20
B-l Terra Vac's Estimated Cost for Complete Remediation of
Valley Property at Groveland 22
B-2 Partial List of Terra Vac Projects 24
C-l Analytical Methods 28
C-2 Reduction of Weighted Average TCE Levels in Soil .. 29
C-3 Extraction Well 4: TCE Reduction in Soil Strata 31
C-4 Monitoring Well 3: TCE Reduction in Soil Strata 31
C-5 Comparison of Wellhead Gas VOC Concentration and
Soil VOC Concentration 32
D-l Chronology of Events at the Bellview, Florida, UST Site 40
D-2 Maximum and Frequency Detected for Organic Compounds
Analyzed in Subsurface Soil Samples from the Former
Lagoons, Collected by Baker/TSA (Onsite RI) 51
VI
-------
Figures
Number
A-l Schematic Diagram of Equipment Lay out ....
A-2 Schematic Diagram of an Extraction Well ....
C-l Pretreatment Shallow Soil-Gas Concentration
C-2 Posttreatment Shallow Soil-Gas Concentration
C-3 Wellhead TCE Concentration vs. Time ..-.-
D-l Soil Formations at Bellview, FL
D-2 Process Schematic at Bellview, FL ..:
D-3 Vacuum Extraction Well Location Map ......
D-4 Schematic of VES at Verona, MI
D-5 Relative Extraction Rates
D-6 Total Ib VOCs Extracted '.
Page
18
19
29
29
32
40
41
42
47
49
49
vn
-------
Abbreviations and Symbols
API
ARAR
ASTM
BDAT
BNA
BTEX
CAA
CERCLA
CFR
cm/s
CDM
CWA
DCA
DCE
DCS
DMF
ECD
EPA
EPTox
EW
FDER
FID
ft3
GC/MS
g/ml
gmol-
HSWA
kPa
American Petroleum Institute
Applicable or Relevant and Appropriate Requirements
American Society for Testing and Materials
Best Demonstrated Available Technology
base neutral/acid (extractable)
benzene, toluene, ethylbenzene, and xylenes
Clean Air Act
Comprehensive Environmental Response, Compensation, and
Liability Act of 1980
Code of Federal Regulations
centimeters per second
Camp, Dresser, McKee, Inc.
Clean Water Act
1,2-dichloroethane
trans 1,2-dichloroethylene
dichlorobenzene
dimethyl formamide
electron capture detector
Environmental Protection Agency
Extraction Procedure Toxicity Test
extraction well
Florida Dept. of Environmental Regulation
flame ionization detector
cubic feet
gas chromatograph/mass spectrometer
grams per milliliter
gram mole
Hazardous and Solid Waste Amendments to RCRA -1984
kilopascal (s)
vm
-------
Abbreviations and Symbols (Continued)
kW
LDR
m/s
MEK
mg/kg
mg/L
MIBK
MW
NCP
N/m
NPDES
NPL
O&G
ORD
OSHA
OSWER
PAHs
PCE
PL
ppb
ppm
ppmv
ppmw
PRP
psi
PVC
RCRA
RI/FS
ROD
RREL
SARA
scfh
kilowatt(s)
land disposal restriction
meters per second
methyl ethyl ketone
milligrams per kilogram
milligrams per liter
methyl isobutyl ketone
monitoring well
National Contingency Plan
Newtons per meter
National Pollutant Discharge Elimination System
National Priorities List
oil and grease
Office of Research and Development
Occupational Safety and Health Act
Office of Solid Waste and Emergency Response
polycyclic aromatic hydrocarbons
tetrachloroethylene
Pub.lic Law
parts per billion
perts per million
parts per million by volume
parts per million by weight
potentially responsible party
pounds per square inch
polyvinyl chloride
Resource Conservation and Recovery Act of 1976
Remedial Investigation/Feasibility Study
Record of Decision
Risk Reduction Engineering Laboratory
Superfund Amendments and Reauthorization Act of 1986
standard cubic feet per hour
IX
-------
Abbreviations and Symbols (Continued)
SITE Superfund Innovative Technology Evaluation
TCA 1,1,1-trichloroethane
TCE trichloroethylene
TCLP Toxicity Characteristic Leaching Procedure
TCP 1,2,3-trichloropropane
THF tetrahydrofuran
TOG total organic carbon
TRI 1,1,1-trichloroethane
TSCA Toxic Substances Control Act of 1976
UCS unconfined compressive strength
um micrometer
ug/L micrograms per liter
UST underground storage tank
VES vacuum extraction system
VOC volatile organic compound
yd3 cubic yard
x
-------
Conversions
English (US)
Area: 1ft2
Iin2
Flow Rate': 1 gal/min
1 gal/min
1 Mgal/d
1 Mgal/d
1 Mgal/d
Length: 1 ft
lin
lyd
Mass: 1 Ib
lib
Volume: 1 ft3
Ift3
Igal
Igal
ft = foot, ft;2 = square foot, ft3 = cubic foot
in = inch, in2 — square inch
yd = yard
Ib = pound
gal = gallon
gal/min = gallons per minute
Mgal/d = million gallons per day
m = meter, in2 = square meter, m3 = cubic meter
cm = centimeter, cm2 = square centimeter
L = liter
g = gram
kg = kilogram
m3/s = cubic meters per second
L/s = liters/sec
m3/d = cubic meters per day
Metric (SI)
9.2903 x 10-3 m2
6.4516 cm2
6.3090 x 10:5 m3/s
6.3090 x lO-2 L/s
43.8126 L/s
3.7854 x 103 m3/d
4.3813 x 10-2 m3/s
0.3048 m
2.54cm
0.9144m
4.5359 x 10-2 g
0.4536 kg
28:3168 L
2.8317 x 10-2 m3
3.7854 L
3.7854 x 10-3 m3
XI
-------
Acknowledgments
This report was prepared under the direction and coordination of Mary Stinson, EPA
SITE Program Manager in the Risk Reduction Engineering Laboratory - Cincinnati,
Ohio. Contributors and reviewers for this report were Mr. James Ciriello of EPA
Region I - Remedial Project Manager for the Groveland Superfund site at the time of
the project; John Kingscott from the Office of Solid Waste and Emergency Response;
Greg Ondich from the Office of Research and Development; James Malot, Ed
Malmanis, and Neil Janis from Terra Vac Corp.; and Patrick Ford and James Thomas
from Alliance Technologies, Inc.
This report was prepared for EPA's Superfund Innovative Technology Evaluation
(SITE) Program by Peter A. Michaels of Foster Wheeler Enviresponse, Inc. for the U.S.
Environmental Protection Agency under Contract No. 68-03-3255.
XII
-------
Section 1
Executive Summary
Terra Vac, Inc.'s in situ vacuum extraction process
has been employed at several Superfund and non-
Superfund sites. Available data from four sites
where field activity has occurred, including three
Superfund and one UST site, were reviewed and are
summarized in Appendix D of this report.
Conclusions
The following conclusions, regarding applications of
the technology, were drawn from reviewing the data
on the Terra Vac in situ vacuum extraction process,
both from the SITE Demonstration Test and other
available data (Appendices C and D):
• The process represents a viable technology to
fully remediate a site contaminated with
volatile organic compounds (VOCs).
• The major considerations in applying this
technology are the contaminant compound's
volatility, site soil porosity, and the site-specific
cleanup level.
• The process demonstrated good performance in
removing VOCs from soil with measured
permeability ranging between 10-4 and 1O8
cm/s. In practical terms, the process works well
with most soil types. It was determined that air-
filled porosity of a soil is a more important
factor than permeability in the application of
this technology.
• It is important where soils of low permeability
and high moisture content, i.e. low air-filled
porosity, are encountered that a pilot
demonstration test be considered to determine
the feasibility of dewatering the soil.
• The process operated well in all weather
conditions. There had been concerns raised on
its applicability during extreme winter
conditions. The technology is relatively simple
and should be considered reliable.
• Chemicals with Henry's Constant greater than
0.001 (dimensionless), see Table A-3, have been
successfully extracted by the Terra Vac process.
The process successfully extracted not only
VOCs but also less-volatile hydrocarbons such
as gasoline, diesel fuel, kerosene, and heavy
naphthas. Henry's Constant is the parameter
that can be used to determine whether a
particular contaminant has sufficient volatility
to be extracted by the process.
• The economics of this process strongly depend
on whether off-gas treatment is required and
whether any waste water is generated at a site.
This latter cost element can add as much as
20% to the total cost.
• Based on available data, the treatment costs
; are typically near $50/ton. Costs can be as low
as $10/ton and for a small urban spill site as
high as $150/ton. For a large remediation
project, when no off-gas treatment is required
and no wastewater is generated, the
remediation cost can be less than $10/ton.
The vacuum extraction technology offers an
economical option to remediate sites contaminated
with volatile organics. Even when a contaminated
site contains semivolatile organics and heavy metals
in addition to VOCs, it may still be economical to
first remove the volatile organics using vacuum
extraction technology and then use other
remediation technologies to remove, immobilize, or
destroy the remaining contaminants after they have
been excavated.
Discussion of Conclusions
The Terra Vac in situ vacuum extraction process was
tested and evaluated under the Superfund
Innovative Technology Evaluation (SITE) Program
at the Valley Manufactured Products Co.'s machine
shop in Groveland, Massachusetts. The site is part of
the Groveland Wells Superfund site and was
apparently contaminated by a leaking storage tank
and by previous improper practices in the storage
and handling of waste oils and solvents. The SITE
Demonstration was conducted from December 1,
1987 to April 26, 1988. The major objectives of the
SITE project were to collect data on performance,
cost, and reliability to help support the above
-------
conclusions, and to investigate procedures which
could be used to predict the time required to achieve
site remediation by attempting to cleanup a
relatively small part of the total site during a 56-day
period.
The demonstration project primarily consisted of
sampling around four extraction wells which were
constructed at the site. Three of the wells were
designed as a barrier to isolate the fourth well by
intercepting contamination from the larger source
area. Testing consisted of extensive wellhead gas
monitoring and soil sampling which occurred before,
during, and at the end of the vacuum extraction
activity. The data show wellhead gas concentrations
decrease exponentially with time. The rate of
decrease varied for the different wells. Due to the
pattern of contamination, one of the barrier wells
actually became isolated from the larger area of
contamination and the adjacent soil was cleaned to a
far greater extent than the other three wells.
Calculations were made to check a theoretical
relationship between wellhead concentrations and
surrounding soil levels. The calculated values were
lower than actual soil concentrations by at least an
order of magnitude. Thus, although wellhead gas
levels are easier to measure, care should be exercised
when they are used as a surrogate for the actual soil
concentration.
Before one can attempt to make a rough estimation
of the remediation time for a site, a target value for
the particular contaminant in the remediated soil
must be known. For the demonstration this target
concentration is calculated by using two
mathematical models, the Vertical and Horizontal
Spread Model and the Organic Leachate Model.
These mathematical models are used by EPA to
evaluate petitions to the Agency for delisting wastes
contaminated with VOCs. The mathematical models
allow the use of a regulatory standard for drinking
water, in order to arrive at a target soil concen-
tration.
Once the target soil concentration is determined, a
rough estimation of the remediation time can be
made by taking a site-specific empirical ratio of soil
concentration to wellhead gas concentration, and
extrapolating to obtain a wellhead gas concentration
at the target soil concentration. As an example, the
calculated target soil concentration for the
Groveland site is 500 ppbw. This corresponds to an
approximate wellhead gas concentration of 89 ppb
for one of the extraction wells (EW1S). The equation
correlating wellhead gas concentration with time is
then solved to give an approximate cleanup time of
150 days.
After this time period, the vacuum extraction system
can be run intermittently to see if significant
increases in gas concentrations occur on restarting,
after at least a two-day stoppage. If there are no
appreciable increases in gas concentration, the soil
has reached its residual equilibrium contaminant
concentration, and the system may be stopped and it
is recommended that soil borings be taken and
analyzed.
An example of a full-scale cleanup is a site
remediation by Terra Vac at the Upjohn Facility
Superfund site in Barceloneta, Puerto Rico, where
the VOC concentrations in soils were reduced to
levels below the detection limit of 5 ppb of CC14 from
initial concentrations as high as 2,200 ppm. The site
had a 300-ft-deep vadose zone, consisting mostly of
clay soils. The site was considered remediated by
EPA when nondetectable levels of CCl^ in the
wellhead gas were demonstrated for a period of three
consecutive months. The vacuum extraction
operations at this site remediated a total of 7,000,000
yd3 in 3 years. If incineration were used to complete
this remediation in the same timeframe it would
require a capacity of 7,0,00 ton per day and would be
impractical and uneconomical.
Operation of the process during the winter poses no
problems as long as care is taken to prevent freezing
in the system piping and valves by the application of
electrical tracing and insulation: The actual process
of vacuum extraction is unaffected by outside
ambient temperatures, since subsurface
temperatures generally do not vary by more than
10°F from summer to winter. Water extraction rates
at Groveland were high during the winter because of
thawing of the snow cover.
Terra Vac treatment costs generally range between
$10 and $50 per ton with isolated instances of small
urban spill sites as high as $150 per ton. The
Groveland 56-day demonstration costs, modified to
eliminate sampling and analytical costs not
normally incurred by Terra Vac, amounted to $47
per ton. An estimate was made on a complete
remediation of the portion of the site occupied by the
demonstration. The complete remediation time was
projected to be 150 days and the resulting cost
amounted to $66/ton. Terra Vac's own estimate of the
complete remediation of the entire Valley
Manufactured Products Co. portion of the Groveland
site was just under $39 per ton. The Terra Vac
estimate was not that far removed from the EPA
estimated costs and reflects the economy of scale.
The quantity at the entire site is approximately
20,000 tons, whereas the demonstration treated
6,000 tons.
-------
Section 2
Introduction
The SITE Program
In 1986, the EPA's Office of Solid Waste and
Emergency Response (OSWER) and Office of
Research and Development (ORD) established the
Superfund Innovative Technology Evaluation (SITE)
Program to promote the development and use of
innovative technologies to clean up Superfund sites
across the country. Now in its third year, the SITE
Program is helping to provide the treatment
technologies necessary to implement new federal and
state cleanup standards aimed at permanent
remedies, rather than quick fixes. The SITE
Program is composed of two major elements: the
Demonstration Program and the Emerging
Technologies Program.
The major focus has been on the Demonstration
Program, which is designed to provide engineering
and cost data on selected technologies. To date, the
demonstration projects have not involved funding for
technology developers. EPA and developers
participating in the program share the cost of the
demonstration. Developers are responsible for
demonstrating their innovative systems at chosen
sites, usually Superfund sites. EPA is responsible for
sampling, analyzing, and evaluating all test results.
The result is an assessment of the technology's
performance, reliability, and cost. This information
will be used in conjunction with other data to select
the most appropriate technologies for the cleanup of
Superfund sites.
Developers of innovative technologies apply to the
Demonstration Program by responding to EPA's
annual solicitation. To qualify for the program, a
new technology must have a pilot or full scale unit
and offer some advantage over existing technologies.
Mobile technologies are of particular interest to
EPA.
Once EPA has accepted a proposal, EPA and the
developer work with the EPA regional offices and
state agencies to identify a site containing wastes
suitable for testing the capabilities of the technology.
EPA prepares a detailed sampling and analysis plan
designed to thoroughly evaluate the technology and
to ensure that the resulting data are reliable. The
duration of a demonstration varies from a few days to
several months, depending on the type of process and
quantity of waste needed to assess the technology.
While it is possible to obtain meaningful results in a
demonstration lasting one week using an
incineration process, where contaminants are
destroyed in a matter of seconds, this is not the case
for a physical treatment such as the Terra Vac
process where contaminants are extracted from the
earth over long periods of time. In order to be able to
make good extrapolations of the data, it was
necessary that the Terra Vac demonstration test last
for a duration of two months. After the completion of
a technology demonstration, EPA prepares two
reports, which are explained in more detail below.
Ultimately, the Demonstration Program leads to an
analysis of the technology's overall applicability to
Superfund problems.
SITE Program Reports
The results of the SITE Demonstration program are
incorporated in two basic documents, the Technology
Evaluation Report and the Applications Analysis
Report. The former provides a comprehensive
description of the demonstration and its results. The
likely audience will be engineers responsible for a
detailed evaluation of the technology relative to a
specific site and waste situation. These technical
evaluators will want to understand thoroughly the
performance of the technology during the
demonstration, and the advantages, risks, and costs
of the technology for the given application.
The Applications Analysis Report is directed at
decision-makers responsible for selecting and
implementing specific remedial actions. This report
provides information in order to determine if the
technology merits further consideration as an option
in cleaning up specific sites. If the candidate
technology described in the Applications Analysis
appears to meet the needs of the site engineers, a
more thorough analysis of the technology based on
the Technology Evaluation Report and information
from remedial investigations for the specific site will
be made. In summary, the Applications Analysis will
assist in the determination of whether the specific
-------
technology should be considered further as an option
for a particular cleanup situation.
Purpose of the Applications Analysis
Report
Bach SITE demonstration will evaluate the
performance of a technology while treating the
particular waste found at the demonstration site.
Additional data from other projects will also be
presented.
Usually the waste at other sites will differ in some
way from the waste tested. Waste characteristic
differences could affect waste treatability and use of
the demonstrated technology at other sites. Thus,
successful demonstration of a technology at one site
does not assure that a technology will work equally
well at other sites. The determination of the total
operating range over which the technology performs
satisfactorily will be made by examining, not only
the demonstration test data, but other data available
from field applications of the technology.
To enable and encourage the general use of
demonstrated technologies, EPA will evaluate the
applicability of each technology to sites and wastes
in addition to those tested, and will study the
technology's likely costs in these applications. The
results of these analyses will be distributed through
the Applications Analysis Report.
Key Contacts
Information useful to potential technology users can
be provided by the following sources:
Terra Vac
356 Fortaleza St.
Box 1591
San Juan, Puerto Rico 00902
James J. Malot, President
(809) 723-9171
Terra Vac
4897-J West Waters Ave.
Tampa, FL 33634
Ed Malmanis, Vice President
(813) 885-5374
U.S. EPA - ORD
Releases Control Branch
MS-104
Edison, NJ 08837
Mary K. Stinson, Demonstration Project Manager
(201) 321-6683
-------
Section 3
Technology Applications Analysis
Introduction
This section of the report addresses the applicability
of the Terra Vac In Situ Vacuum Extraction Process
to remediate various sites contaminated with
various volatile organic compounds, based on the
results of the SITE demonstration test and other
Terra Vac test results data.
Since the results of the demonstration test provide
the most extensive database, conclusions on the
effectiveness and its applicability to other potential
cleanups will be strongly influenced by those results,
which are presented in detail in the Demonstration
Report. Additional information on the Terra Vac
technology, including a process description, vendor
claims, a summary of the Demonstration Test
results, and case studies of other applications are
provided in Appendices B-D.
Conclusions
Following are the overall conclusions on the Terra
Vac technology. These conclusions were drawn from
a review of the existing data on the Terra Vac
process:
• The process removes VOCs from soils of low, as
well as high, permeability.
• The process represents a viable technology for
fully remediating a site contaminated with
VOCs.
• The major factors to be considered in applying
the technology are contaminant volatility, site
soil porosity, and the site-specific cleanup level.
• Declining VOC recovery rates correlated well
as an exponential function of time of treatment.
This makes it possible to estimate cleanup
times.
• The process operates well in all weather
conditions.
* The process costs are typically near $50/ton but
can be as low as $10 per ton of soil treated.
Discussion of Conclusions
Soil Permeability
Terra Vac claims that the vacuum extraction
technology will remove volatile organic compounds
from virtually all hydrogeological settings including
clays, silts, sand and gravel, alluvium, colluvium
and glacial till, wetlands, and fractured rock and
karst (see Appendix B). The SITE program
demonstration of the Terra Vac process in
Groveland, Mass, showed that it was possible to
remove VOCs from clays having a permeability of
lO-8 cm/s. The Groveland results (see Appendix C)
showed considerable reductions in TCE levels in
certain clay strata; at a certain location (MW 3) they
were reduced from 200-1,600 ppm TCE before the
test to ND-60 ppm TCE after the test. The largest
site to date at which Terra Vac has completed
operations is the Upjohn Facility Superfund Site in
Barceloneta, Puerto Rico (see Appendix D-l). This
site has a 300-ft-deep vadose zone with soil
formations that are mostly clay having hydraulic
conductivities of 10-7 cm/s. The action involved the
vacuum extraction from the vadose zone of a carbon
tetrachloride spill. Terra Vac had a portion of this
site to remediate, and almost all soil samples that
were taken at the completion of a 2 1/2-year-run
were below the detection limit of 5 ppb CCl4. Terra
Vac is in the process of remediating the Tyson's
Dump Superfund Site in Upper Merion Twp., Pa.,
which has a hydrogeology of clay and silty sands. At
Tyson's Dump, Terra Vac installed more than 100
wells at varying depths and is using four vacuum
blowers for the job - two 700-hp machines, and two
500-hp machines. They are using an onsite
activated-carbon regeneration-system employing
steam for regeneration, and packed-tower air-water
stripper for the treatment of contaminated
groundwater that is extracted by the vacuum
system.
The pore size for clays is small, offering a large
resistance to the flow of liquids and a much lower
resistance to the flow of the smaller air molecules.
For vacuum extraction to work well, the soil must
have a reasonable value for the air filled porosity so
that the induced stripping air may have a low
-------
resistance to flow. It was found that the total
porosities of the soils at the Groveland site, whether
they were sands, silty sands or clays, were
approximately the same. As long as the water
content of the soil is not too high, there would be
enough air-filled porosity. As an example, if the total
porosity is measured at 50% and the moisture
content is 20%, the calculation of air-filled porosity,
using a bulk density of the soil of 1,761 kg/m3, would
be:
Total porosity = 0.5 = fraction of total volume
occupied by pores
For 1 m3 of soil, 0.5 m3 is occupied by pores
For 1 m3 of soil, there is 0.2x1,761 = 352.2 kg H2O
content
Volume ofH2O = 352.2 kgx 1 m3/l,000 kg = 0.3522
m3
Volume of air = 0.5 - 0.3522 = 0.1478 m3
Air-filled porosity = 14.78% .
The above calculated value of air-filled porosity is for
the Groveland site and the value obtained was high
enough to achieve good vacuum extraction. If, for the
same total porosity and bulk density, the moisture
content was 28.4%, the air-filled porosity would be
zero and the soil matrix would have to be dewatered
in order to start the extraction process for VOCs.
This type of dewatering can actually be done by the
Terra Vac system, but it may extend the time
required by the remediation.
Contaminant Volatility
Theory tells us that the vacuum extraction process
can remediate a site contaminated with VOCs, and
that it is just a matter of time before this can be
accomplished. The basis for successful application of
vacuum extraction to clean up soils and associated
interstitial water is Henry's Law. Simply stated,
Henry's Constant is the ratio of a compound's
concentration in air divided by the concentration in
water. Compounds with higher Henry's Constants
will clean up faster. Effective recovery of chemicals
with Henry's Constants greater than 0.001
(dimensionless) have been demonstrated by Terra
Vac.
Site-Specific Cleanup Levels
For the contaminant trichloroethylene (TCE),
successful petitions for delisting have been made to
EPA, where the Agency has applied the
mathematical models for delisting to come up with a
level of 0.59 ppm. For the purpose of this study it was
assumed that the Groveland Superfund site could be
considered remediated if the soil residual TCE
concentration is 0.59 ppm of lower. The actual ROD
for the Groveland site, which selected a combination
of vacuum extraction and groundwater stripping as
the remedial technology for the site, did not set,a soil
concentration target. It is reasonable to assume,
given the results of the demonstration (see Appendix
C) that this level may be reached for this site if there
were enough extraction wells and enough vacuum
pumps installed to do the job. The results in the
small demonstration area indicate that the test area
may be remediated within about 150 days.
Terra Vac has remediated the Upjohn Facility
Superfund Site, and has accomplished this in a 2 1/2-
year time frame on a site with clay soils. The
Groveland site data show a steady decline in the
recovery rate over time, and were correlated using
an exponential equation of the form y = ae-kx, where
y = concentration of contaminant in extracted
vapors and x = time. From these data it is possible to
estimate the time required to achieve a desired
contaminant concentration in the extracted vapors.
Estimation of Remediation Time
The relationship between the contaminant
concentration in the extracted vapors and the soil
concentration value is not known with any
reasonable degree of certainty and should be
assumed as site-specific. This is important because
although measurements of the extracted vapor
contaminant levels are relatively easy to make, the
SITE demonstration showed that they cannot
assume to represent the soil concentrations. One
must analyze a series of soil borings (this has been
done in the Groveland SITE program), and then
determine by analysis what the concentration in the
wellhead extracted vapors was when the soil borings
were taken. Then a gross extrapolation may be made
to see what the wellhead concentration would be at
the desired remediated soil concentration (see
Appendix C).
This will only be a rough estimation of the time
required for remediation. It is important to know
how long the process is to continue running before it
may be stopped and the soils tested. The vacuum
extraction process exhibits higher contaminant
concentration levels of the extracted vapors when
starting up after a shutdown of a few days. When the
soil is close to reaching its residual equilibrium
concentration, or when the vacuum extraction
process is effectively not removing any more
contamination, this increase in concentration will
not occur upon startup after a prolonged shutdown. It
is recommended that the vacuum extraction process
be run intermittently after reaching the time
calculated by the foregoing rough estimation
method. The final soil borings may be taken if the
residual equilibrium concentration is reached.
Reliability
The Terra Vac system in Groveland operated in the
winter during severe weather conditions. The entire
-------
piping system from the extraction wells to the
equipment was electrically heat traced to prevent
freeze-ups, which would otherwise occur, especially
during brief shutdowns. The snow cover on the site
offered no interference with air induction in the
Terra Vac process, even though Terra Vac does not
use air injection wells. During the winter at
Groveland more groundwater was extracted by the
process because of thawing of the snow cover.
Siltation caused a blockage and cessation of the
vacuum extraction process in one of the deep wells
(E W4D) during the second half of the demonstration
period. Attempts to clear the blockage by pulse-
pumping water down the well during the mid-test
break were unsuccessful. When siltation of a well
occurs causing a blockage it is usually necessary to
drill a new well adjacent to the blocked one.
The Terra Vac vacuum extraction system may be left
unattended for long periods of time except when
activated-carbon canister replacement becomes
necessary. The degree of operator attention required
is a function of the size and the complexity of the
project. A project the size of the Groveland
demonstration may require no-one for long periods of
time whereas the Tyson's Superfund site, which has
an activated carbon regeneration facility and a water
treating facility in addition to about 100 extraction
wells requires 20 people during the daytime.
Costs
Typical costs for vacuum extraction systems depend
on a number of site-specific factors including: the size
of the site, the nature and amount of the
contamination and the hydrogeological setting (see
Section 4). For a large site, for example 100,000 ton
with sand soil and a contaminant such as TCE the
implementation costs may be $10 per ton. For small
sites, such as 10,000 yd3 with clay soils and TCE
contaminant, the cost might be $50/ton.
Application of Vacuum Extraction
Technology
Vacuum extraction of volatiles and semivolatiles is
favorable for a wide variety of applications
including:
Source control
Assessment of source areas
Emergency response
Liquid-phase recovery
Enhanced groundwater treatment
In situ or ex situ processes
As a safer, more-cost-effective, and permanent
alternative to excavation and disposal, vacuum
extraction provides effective source control of
contaminants in soils. Also, whereas excavation is
disruptive and may be impractical due to physical
conditions at the site - such as depth, location of
buildings, roadways or utilities, and the proximity to
residents — vacuum extraction has been successfully
implemented under buildings, industrial tank farms,
gas stations, and beneath large diameter (150 ft)
above-ground storage tanks.
Studies have shown that the process of soil
excavation can release 60% to 90% of the volatile
contaminants into the atmosphere, even where
engineering controls are in place. Accordingly,
volatiles rapidly released from the process of
excavation, could violate air emissions regulations,
cause unnecessary health risks to workers and to
people in nearby residences, and cause nuisance
odors. One significant advantage of the vacuum
extraction process is that soils and groundwater are
treated in situ, without excavation.
However, if excavation is required, treatment of
volatiles can be handled effectively using a process
known as "heap vacuum extraction". In this process,
contaminated soils are placed on a treatment pad
consisting of a horizontal, covered vacuum system.
All of the volatiles remaining in the soil are
extracted by essentially air stripping the soils in a
controlled system. Thus, volatiles in soils are
removed and treated.
Vacuum extraction technology may be used to
delineate the extent and magnitude of VOC sources.
Vacuum extraction wells can be installed and vapors
extracted during site assessment. Since the extracted
vapors represent a volume of soil that is much larger
than that indicated by soil sampling, the extent and
magnitude of the source area is thereby better
quantified than by soil sampling alone. In addition,
remediation can begin immediately if any
contaminants are extracted.
In emergency response situations, rapid deployment
of vacuum extraction can remove VOCs from soils
before contaminants reach the aquifer. Rapid
elimination of the residual contaminants in the
vadose zone using vacuum extraction will minimize
the impact of contaminants on groundwater quality.
Treatment of vapors produced by the process are
typically handled in one of three ways: dispersion,
carbon adsorption, or thermal destruction. Other
methods — such as condensation, biological
degradation, ultraviolet oxidation, and others - have
been applied, but only to a limited extent. Dispersion
often renders the vapors harmless. Quite often, the
health risks posed by resultant concentrations in the
air are much less severe than those posed by
groundwater or contaminated soils. Methods that
destroy the contaminants are; preferable, but they do
increase costs. Destruction by thermal incineration
-------
or catalytic oxidation is quite effective, especially for
hydrocarbons. Chlorinated hydrocarbons can be
effectively treated by thermal destruction or
adsorbed by activated carbon. Regeneration of the
spent carbon on site is preferred to landfilling, so
that chemicals adsorbed on the carbon can either be
recycled or destroyed.
Vacuum extraction is a safe process with minimal
permitting required to implement a system. Since all
contaminants are under vacuum until treated, the
possibility of a release is virtually eliminated. Many
states require an air permit for operation of the
vacuum extraction process. However, compliance
with even the strictest regulations is easily within
the capability of the technology.
Equipment used in the Terra Vac process can be
either mobile or fixed equipment. Typically, the
leasing of mobile extraction units is more cost
effective than capitalizing onsite equipment, since
cleanup is so rapid compared to common equipment
amortization schedules. Mobilization of Terra Vac's
portable equipment can be accomplished within
about one week, with startup and full-scale
operations within about two weeks. Operation and
maintenance costs are minimal and the Terra Vac
system has demonstrated its ability for safe,
unattended, continuous operation.
Environmental Regulation Requirements
for the Technology
This section discusses briefly the environmental
regulations that pertain to hazardous waste
remediations. The discussion focuses on the
restriction on land disposal of hazardous wastes and
the use of the in situ vacuum extraction technology
for Superfund actions.
The Resource Conservation and Recovery Act
(RCRA)
The Resource Conservation and Recovery Act
(RCRA) was enacted on October 21, 1976, and was
amended in 1980 and 1984. The amendments of 1984
are called the Hazardous and Solid Waste
Amendments (HSWA).
There are nine subtitles that comprise RCRA, and
the subtitles that pertain to the cleanup of hazardous
wastes are:
• Subtitle C - Hazardous Waste Management
System
• Subtitle D - State and Regional Solid Waste
Plan
• Subtitle I-UST Regulations
Congress took a position in the HSWA to ban the
land disposal of untreated hazardous wastes.
Deadlines were established by Section 3004 of
HSWA covering the restriction of untreated
hazardous-waste disposal into landfills. EPA
developed treatment standards for the first deadline
of November 8, 1986 covering both spent solvent
waste listed by EPA and listed dioxin wastes.
Treatment standards for the second deadline of July
8, 1987 covered a group of mostly liquid hazardous
wastes identified by the California Department of
Health Services or "California list wastes".
"California list wastes" include corrosive wastes,
wastes containing metals and cyanides, and
halogenated organics. EPA developed treatment
standards for the first third of its list of hazardous
wastes on August 6,1988. EPA has deadlines of June
8, 1989, for the second third of that list and May 8,
1990, for the final third.
Treatment standards are based on the performance
of the BOAT to treat the waste. A technology is
considered to be demonstrated for a particular waste
if the technology is in full-scale commercial
operation for treatment of that waste. Treatment
standards can be established either as a specific
technology or as performance standard based on a
BOAT. When treatment standards are fixed at a
performance level, the regulated community may
use any technology not otherwise prohibited to treat
the waste so that it meets the treatment standard.
Since the Terra Vac technology is an in situ
treatment, it does not involve the placement of
hazardous waste. Therefore, land ban requirements
do not apply. However, residuals which are
generated from the treatment of restricted RCRA
hazardous waste are themselves considered as
restricted RCRA hazardous waste. Therefore,
residuals such as activated carbon and recovered
water may be subject to the land ban requirements in
the future.
The Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA)
Statutory Requirements
The Comprehensive Environmental Response,
Compensation, and Liability Act (CERCLA) of 1980,
as amended by the Superfund Amendments and
Reauthorization Act (SARA) of 1986, provides for
federal funding to respond to releases of hazardous
substances to air, water, and land. CERCLA
authorized EPA to prepare the National
Contingency Plan (NCP) for hazardous substance
response. The NCP defines methods and criteria for
determining the appropriate .extent of removal,
remedy, and other measures. Specific techniques
mentioned in the NCP for remedial action at
hazardous waste sites include in situ treatment as a
-------
cost-effective technology for handling contaminated
soil and sediment.
Section 121 of SARA, entitled Cleanup Standards,
strongly recommends remedial actions using
treatment that "...permanently and significantly
reduces the volume, toxicity, or mobility of
hazardous substances." The actions must assure
protection of human health and the environment,
meet ARARs, be in accordance with the NCP, and be
cost-effective. This means that when selecting an
appropriate remedial action, the first step is to
determine the level of cleanup that is necessary to
protect the environment, and the second step is to
choose the most cost-efficient means of achieving
that goal. SARA further states that "offsite transport
and disposal...without such treatment should be the
least favored alternative remedial action where
practical treatment technologies are available."
SARA also added a new criterion for assessing
cleanups that includes consideration of potential
contamination of the ambient air. This is in addition
to general criteria requiring that remedies be
protective of human health and the environment.
Since vacuum extraction involves the collection of
contaminated vapors, all these criteria must be
considered in assessing appropriate levels of air
treatment.
Superfund Response Actions
Superfund response actions must meet "applicable or
relevant and appropriate requirements" (ARARs) for
remediation. Land disposal restrictions (LDRs) may
be "applicable or relevant and appropriate" to
Superfund actions. LDRs are applicable when
existing federal or state laws can be used to have
direct authority over placement of restricted
hazardous wastes in, or on, the land. LDRs may be
relevant and appropriate when Superfund hazardous
substances are sufficiently similar to restricted
industrial hazardous wastes such that use of LDRs is
suited to the circumstances of the releases.
In addition to industrial process waste, the HSWA
also addresses soil and debris that result from
CERCLA response actions and RCRA corrective
actions. Effective August 8, 1988, EPA issued a
national capacity variance through November 8,
1990 for all CERCLA/RCRA soil and debris, which
are contaminated with hazardous wastes whose
BOAT standards are based on incineration.
In the meantime, the EPA intends to develop
separate BOAT treatment standards for soil and
debris, because the BOAT standards were developed
for industrial waste processes, which are often
different from the soil or debris waste matrices in
terms of chemical/physical composition,
concentrations, and media within and among sites.
Until standards are developed for soil and debris,
remedies will continue to be selected on a site-
specific basis. Since these remedies are not likely to
conform to the BOAT standards for industrial
process waste, a variance is often required.
Since vacuum extraction is an in situ process (no
excavation and placement is involved) the land ban
restrictions are not applicable. Individual states may
however have rules or regulations affecting cleanup
levels for a particular VOC contaminant in soil.,
Cleanup levels will be established on a site-specific
basis.
In addition there are contaminated residuals that
may result from the application of this technology.
These include recovered groundwater and spent
activated carbon from offgas treatment.
Contaminated water requires treatment in
accordance .with the State National Pollution
Discharge Elimination System permit levels prior to
surface water discharge, or in accordance with
pretreatment requirements prior to sewage
discharge. Spent activated carbon should be disposed
in accordance with the Agency's Offsite Disposal
Policy.
Materials Handling Required by the
Technology
Since this is an in situ process, the soil itself is not
excavated for treatment. There are, however, various
related activities that do involve the handling of
material.
The construction of the extraction and monitoring
wells requires the mobilization of a portable drill rig.
If soil borings are required for analysis, they are
obtained during the construction of the wells by
hammering a split spoon into the ground, generally
every two feet, before the soil is augered out. Soil
tailings from the drilling operation are drummed
and are usually disposed of offsite by incineration.
The grout mix and bentonite for the construction of
the wells are in bags and are mixed and poured by
hand into the boring holes are supplied by the
drilling contractor.
When offgas treatment is required and activated
carbon canisters are used, the canisters employed are
skid mounted so that they can be moved by using a
fork lift truck. When contaminated groundwater is
recovered by the vacuum extraction process, it can
usually be disposed of by treating with carbon
adsorption or with an air stripper and discharging on
site. If this is not permitted, the contaminated water
can be pumped into a holding tank. This holding
tank can then be emptied by a tank truck that will
-------
periodically haul the contaminated wastewater to a
disposal facility.
Personnel Issues
The number of people involved in the operations of a
Terra Vac installation is dependent on the size of the
installation. For the Groveland site one person,
working only during the day, was employed. During
construction at Groveland, Terra Vac employed
three people during the day for a period of almost two
weeks. A large scale project, such as Tyson's (see
Appendix D) would require a full-scale construction
force while the project is being built. Tyson's
required three people per shift during the normal
operating period. All personnel must pass the
appropriate physical examinations and have
certificates of completion of an approved 40-hour
hazardous materials training course as per the
requirements of 29 CFR1910.
Testing Procedures
The description of the sampling and analytical
procedures used in the demonstration test is in
Appendix C. For most jobs, it will not be necessary to
employ all the tests 'specified therein. The most
important tests are:
• Pretreatment soil boring VOCs by purge-and-
trap GC/MS.
• Posttreatment soil boring VOCs by purge-and-
trap GC/MS.
• Periodic VOC analysis in wellhead gas and
activated carbon offgas by GC/FID or ECD.
10
-------
Section 4
Economic Analysis
Introduction
A primary purpose of this economic analysis is to
attempt to estimate costs (not including profits) for a
commercial-size remediation using Terra Vac tech-
nology. It was expected that, where applicable,
vacuum extraction technologies would be less
expensive than most other technologies, such as
incineration. Terra Vac technology has already been
applied successfully at numerous Superfund and
non-Superfund sites. Complete cost data could not be
obtained from any of these applications. Therefore,
this analysis is based on the partial remediation of
part of the Groveland Superfund site during the
SITE Demonstration. The demonstration test run
lasted 56 days and the total amount of contaminated
soil treated was estimated to be 6,000 tons. Thus, the
base case (Case 1) in Table 4-1 represents the actual
costs incurred during the demonstration.
Case 2 in Table 4-1 represents the estimated costs for
complete remediation of the 6,000-ton portion of the
Groveland site. The time necessary to achieve
complete remediation was estimated from the use of
actual data taken during the demonstration in the
theoretical models that have been developed by
Terra Vac and discussed in detail in Appendix C of
this report. It was estimated that it would take 150
days (as per Appendix C) to accomplish the
remediation. Since the combined offgas treatment
and wastewater disposal costs represent the largest
fraction of the total cost, assumptions had to be made
in estimating these for Case 2 calculations.
Many costs were incurred at the Groveland site
during the demonstration that do not typically occur
in a commercial application. Therefore, such costs
were not included in the analysis presented here.
Permitting and regulatory costs were not included.
Site preparation costs were primarily limited to the,
costs associated with drilling wells. In gener.a.1, 'all-
cost elements that are normally associated with the
use of this technology are addressed and an attempt
is made to develop all the costs to be incurred by
those responsible for the remediation of a site.
Results of Economic Analysis
The result of the analysis shows a cost-per-ton range
of $27 to $66. The lower value is based on the
Groveland site demonstratic>n costs, but assuming
that the technology is applied at a site where no
offgas treatment is required and no wastewater is
generated for disposal. Terra Vac has applied their
technology at a number of both Superfund and non-
Superfund sites where that has been the case. At the
Barceloneta Superfund site, no offgas treatment was
required and after the initial operation, no
wastewater was generated. At the Groveland site,
this cost element represented approximately 43% of
the total costs.
The second major cost element associated with this
technology is the cost of amalyses. Pretest and
posttest soil analyses should be performed at all
sites. In addition, periodic wellhead-gas and stack-
gas analyses are required. The costs for wellhead and
stack-gas analyses are directly proportional to the
operating time at a site, and thus may be site-
specific.
At the Groveland site, the analytical costs were
estimated to be 38% of the total cost for the
demonstration, and 31% of the total cost for the full
remediation case. Unfortunately, similar cost
information was not available from any of the other
sites at which the Terra Vac system has been used.
All the remaining costs associated with the use of the
Terra Vac technology constitute between 20-25% of
the total costs. In actual terms, based on the
Groveland site analysis, these ranged from $9/ton to
$15/ton. Clearly, these costs are strongly (but not
linearly) dependent on the operating time required
at a site.
The costs estimated in this report are higher than
those normally claimed by Terra Vac. It should be
noted that a number of cost elements, such as the
maintenance of a field laboratory trailer and the
adherence to strict QA/QC procedures, are either not
11
-------
Table 4-1. Estimated Cost, $/Ton
Case 1 Case 2
Demonstration Complete remediation
56 days 1 50 days
Site preparation
Permitting and regulatory
Equipment
Terra Vac, $ (b)
Contingency (1 0% of direct costs)
Startup and fixed cost
Operations procedures/training
Mobilization and shakedown
Depreciation (10% of direct costs)
Insurance & taxes (10% of direct costs)
Labor costs
Supplies - Raw materials
Supplies - Utilities
Electricity
Effluent treatment
Residual disposal
Analytical
During operation
Pretest and posttest analyses
Facility modifications
(10% of direct costs)
Site demobilization
TOTALS
2.83
—
50,000
0.28
0.40
0.84
0.28
0.28
2.77
—
0.40
12.83
7.50
1.44
16.66
0.28
0.33
47.12
2.83
—
50,000
0.56
0.40
0.84
0.56
0.56
6.93
...
1.08
19.25
11.25
3.86
' 16.66
0.56
0.33
65.67
(a) This cost analysis does not include profits of the contractors involved.
-------
the less expensive construction of fewer, but deeper
wells.
Type of Soil
It is generally believed that sandy soils will take less
time to remediate than clay soils. Whether the soil be
sand or clay, sufficient air-filled porosity must exist
in the soil matrix in order to effect a reasonable flow
of extraction air.
Nature of the Contamination
The volatility and the Henry's Law Constant have a
great effect on the time required to remediate a site.
A material with a high value of Henry's Law
Constant will take a shorter time to extract than will
one with a low value. For the same air flow, the
contaminant material with a higher Henry's Law
Constant affords a higher partial pressure of this
contaminant in the air stream than one with a lower
Henry's Law Constant. This means more
contaminant is removed per unit time. Remediation
times for the materials with high and low Henry's
Law Constants can be affected by employing
different vacuum equipment, i.e., higher air flow and
higher vacuum for the contaminant with the lower
Henry's Law Constant.
Amount of Contamination
It is logical that a more-contaminated site will take
longer to remediate. In addition, if offgas treatment
is required (e.g., activated carbon), the higher carbon
costs for the more contaminated site will increase the
overall remediation cost.
Requirements for Offgas Treatment
Some sites will require only dispersion stacks for the
offgas, depending on the toxicity of the contaminant.
If offgas treatment is required (e.g., in the form of
activated carbon), this can amount to as much as
30% of the cost/ton of the remediation. Large sites
can have .a. system for regeneration of the activated
carbon onsite, which can cut the cost of carbon
almost in half.
Wastewater Generation Possibilities
If little or no water is recovered from the vadose zone,
the costs/ton may be up to 20% lower than for a site
with large amounts of water. If water is recovered it
is usually cost-effective to run the collected
groundwater through a carbon adsorption or air
stripping unit prior to discharge on site, rather than
disposal offsite. This would be significantly less
expensive than the cost of offsite disposal of the
waste water.
Basis of Economic Analysis
The cost analysis was prepared by breaking the costs
into twelve groups. These will be described in detail
as they apply to the Terra Vac process. The
categories, some of which do not have costs
associated with them for this technology, are as
follows:
• Site preparation costs -- including site design
and layout, surveys arid site investigations,
legal searches, access rights and roads,
preparations for support facilities,
decontamination facilities, utility connections,
and auxiliary buildings.
• Permitting and regulatory costs — including
permit(s), system monitoring requirements,
and development of monitoring and analytical
protocols and procedures.
• Equipment costs — broken out by subsystems,
including all major equipment items: process
equipment, and residual handling equipment.
• Startup and fixed costs -- broken out by
categories, including mobilization, shakedown,
testing, depreciation, taxes, and initiation of
environmental monitoring programs.
• Labor costs — including supervisory and
administrative staff, professional and technical
staff, maintenance personnel, and clerical
support.
• Supply costs — includes raw materials.
• Supplies and consumables costs — includes the
utilities such as fuel and electricity.
• Effluent treatment costs ~ includes offgas and
wastewater treatment.
• Residual wastewater disposal costs — including
the preparation for shipping and the actual
wastewater disposal charges.
• Analytical costs -- including laboratory
analyses for operations and environmental
monitoring.
• Facility modification, repair, and replacement
costs — including design adjustments, facility
modifications, scheduled maintenance, working
capital, and equipment replacement.
• Site demobilization costs - including shutdown,
mobile equipment decontamination and
13
-------
demobilization, site cleanup and restoration,
permanent storage costs, and site security.
Some general assumptions defining the basis of the
estimates are as follows:
• A total of 6,000 yd3 of soil are processed in the
two cases estimated.
• Case 1 is based on the actual system operation
costs for 56 days under the SITE program. Only
partial remediation of 6,000 yd3 of the
Groveland Superfund site was achieved during
the SITE demonstration. Case 2 is based on a
complete remediation of a 6,000-yd3 portion of
the Groveland site. As discussed in Appendix C,
the total remediation time is estimated to be
150 days.
• There is a prime contractor on site responsible
for the complete site cleanup, who will provide
certain functions for the Terra Vac unit (such as
site preparation), whose costs are not included.
These costs are expected to be site-specific, but
minimal.
The twelve cost factors, along with the assumptions
utilized, each are described below:
Site Preparation Costs
Site preparation consisted of the setup and outfitting
of a trailer, which was used as a base of operations;
the provision of electrical service for the trailer, the
vacuum pump skid and water pump; the minor
grubbing and cleaning of the site; and installation of
extraction and monitoring wells. The vacuum pump
skid required the extension of an existing 440-V, 3-
phase, 60-Hz line in the Valley Manufactured
Products plant back to the demonstration area, with
the installation of a circuit breaker and electrical
power meter. The trailer power line was run from a
street pole and meter installed. Various outside
lights and switches were installed to make the area
more secure. During the drilling operations, the
services of a health and safety officer were required
to set forth the personnel protection requirements
based on monitoring of VOC emissions. These costs
are included as part of the site preparation costs.
Site preparation costs are highly site-dependent.
Some sites may require the construction of access
roads; however, no such expenses were incurred at
the Groveland site.
Permitting and Regulatory Costs
Since Groveland is a Superfund site, it was assumed
that no permits were required, either Federal nor
State. The need for developing analytical protocols or
monitoring records is assumed not to exist. On non-
Superfund sites, this activity could be expensive and
very time consuming.
Equipment Costs
Based on the information provided by Terra Vac, the
capital cost of the equipment utilized at the
Groveland site is estimated to be $50,000. This
includes the costs for the design, engineering,
materials and equipment procurement, and
fabrication and installation of the Terra Vac's
transportable unit. The cost of all subsystems and
components installed on their respective skids and
trailers is included; however, the cost of tractors for
the transport of trailers is not included.
A contingency cost, approximately 10% of the
equipment cost (per annum prorated to the actual
time on site), is allowed for unforeseen costs. For
Case 12, total use time of four months is assumed,
whereas total use time of eight months is assumed
for Case 2.
Startup and Fixed Costs
The costs included in this group are operating
procedures and operator training, initial shakedown
of the equipment, equipment depreciation, and
insurance and taxes.
In order to ensure a safe, economical, and efficient
operation of the unit, a program to train operators
and operating procedures are necessary. The
associated costs will accrue: the preparation of
health-and-safety and operating manuals; the
development and implementation of an operator
training program; and equipment decontamination
procedures. For this analysis, it was assumed that
trained operators from Terra Vac were available,
thus requiring no training costs. The costs shown in
this category represent the costs to prepare the
manual mentioned above.
Mobilization and shakedown costs include the
transportation of the unit to the site, initial setup,
onsite checkout, construction supervision, and
working capital. Personnel travel costs to the site are
not included. These costs are site-specific and may
vary depending on the nature and location of the site.
The depreciation costs are based on a 10-yr life for all
the equipment. Therefore, the costs are based on
writing off $50,000 worth of equipment over ten
years.
Insurance and taxes are lumped together and are
assumed, for purposes of this estimate, as 10% of
direct costs taken on an annual basis. Again, the
effective duration of four months is assumed for Case
1, and eight months for Case 2.
14
-------
Labor Costs
A major advantage of the Terra Vac technology is
that it requires very few operating personnel, even at
a very large remediation project. At the Grovelahd
site, in the initial period, two operators were
required, but once the system operation stabilized,
only one was required, which was the case during the
demonstration period. The main tasks for the
operators were to take daily gas samples and replace
carbon canisters. Assuming $25/h for the operator
and one 8-h-shift/d coverage, the labor cost for 56
days of operation is $11,200. The living cost for the
same period was estimated to be $5,400. In Case 2, in
which the operation would last for 150 days to
achieve full remediation, the living costs could be
reduced a little by renting accommodations rather
than using hotels. A 20% reduction from the prorated
living expenses is assumed for Case 2.
Supply Costs
There'are no supplies required in typical Terra Vac
system operations.
Utility Costs
The only utility required for this technology is the
electric power needed for the motors that operate the
pumps, fans, vacuum pump, and the small separator
water pump. Auxiliary electrical requirements to
power the two field trailers and other site lighting
are minimal and are included in the total utility
costs. The cost of electricity used is $0.08/kWh.
Effluent Treatment
At the Groveland site, offgas from the vacuum
extraction system was treated by the use of activated
carbon. A total of 15,200 Ibs of activated carbon was
used during the 56 days of SITE demonstration. The
cost of this activated carbon included the cost of
regeneration.
The offgas treatment cost for Case 2 cannot be
determined by prorating in the ratio of the total
operating time. As the operation goes on, the amount
of organics extracted by the system decrease
exponentially. For this analysis, it was assumed that
the cost of activated carbon required for full
remediation would be approximately 50% higher
than the cost during the demonstration. Since this
extrapolation is based solely on engineering
judgement, there is a significant uncertainty
associated with it.
It must be emphasized that offgas treatment may not
be required at all sites. For example, the remediation
at the Upjohn Superfund site in Puerto Rico and the
Bellview site in Florida required no offgas
treatment.
Residual Disposal
At the Groveland site, 17,000 gal of wastewater was
extracted during the demonstration. This
contaminated water may be treated onsite or
disposed of to a fully-permitted treatment facility.
For the demonstration (Case 1), the wastewater was
disposed of off site and the actual disposal costs are
included in these costs. Again, it is assumed that for
Case 2, the cost of wastewa.ter disposal would be
approximately 50% higher than the cost during the
demonstration. It should be noted that there is a
significant uncertainty associated with this cost for
Case 2.
It is estimated that onsite treatment of the
wastewater using a packed- column air stripper or
activated carbon would reduce this cost by
approximately 50%.
Again, this cost element is also very site-specific. At
the Upjohn Superfund site, a separate ground water
treatment system was used since the contamination
had also reached the aquifer. The small amount of
contaminated water extracted from the vacuum
extraction system was treated by the groundwater
treatment system without any significant impact on
the cost. At other sites, the sjrstem may not generate
any wastewater at all. ' .
Analytical Costs
These costs include the cost for pretest and posttest
soil sampling and analyses, and the periodic
sampling and analyses of wellhead gas and stack
gas. The frequency of wellhead-gas and stack-gas
analyses are site-dependent. Typically, Terra Vac
sends one person to the site once a week with a
portable gas chromatograph to perform these
analyses. These costs for the SITE demonstration
were much higher than estimated herein due to the
nature of the testing. The costs included here are
more representative of how a typical commercial
operation would be remediated.
Facility Modification Costs
The costs accrued under this category include
maintenance and working capital. Maintenance
materials and labor costs are difficult to estimate
and cannot be predicted as functions of preliminary
design concepts. Therefore,, annual maintenance
costs are assumed to be 10% of capital costs. Working
capital costs are assumed to be negligible, as all
supplies purchased to have on hand are assumed to
be fully consumed by the project's completion. The
cost of using money early in the project is neglected.
Site Demobilization Costs
The only costs estimated for this category for the
Groveland site were the decontamination and
15
-------
demobilization costs. The cost of providing site
security was not included. Since the site was not
fully remediated, the costs associated with site
restoration — such as refilling and capping the wells,
etc. — was also not included.
16
-------
Appendix A
Process Description
The vacuum extraction process is a technique for the
removal of volatile organic compounds (VOCs) from
the vadose (or unsaturated) zone of soils. This is the
subsurface soil zone located between the surface soil
and the groundwater. In general, VOCs are present
in these soils in one of these ways: as dissolved
constituents in the aqueous phase; as constituents
adsorbed in the solid soil material; or ,as free
constituents in the liquid and vapor phases in the
void space of the soil. Once a contaminated area is
completely defined; an extraction well or wells,
depending on the extent of contamination, will be
installed. The extraction well is connected by piping
to a separator device.
Vacuum extraction wells are designed with a
vacuum-tight seal near the surface and an extraction
zone (screen) corresponding to the profile of
subsurface contamination. Extraction systems may
be vertical (wells) or horizontal (screens installed in
trenches or horizontal borings). Horizontal systems
are effective in areas where groundwater and
contamination is very shallow (i.e., less than 10 ft)
and removal of groundwater is to be minimized.
A vacuum pump or blower induces air flow through
the soil, stripping and volatilizing the VOCs from the
soil matrix into the air stream. Liquid water is
generally extracted along with the contamination.
Vacuum is applied to the vacuum extraction well via
a manifold system. The vacuum at the wellhead is
directly related to the radius of influence of the well
and the cleanup rate that can be achieved; the higher
the vacuum, the faster the cleanup. When vacuum is
exerted on the well, subsurface vacuum propagates
laterally, volatilizing contaminants in place.
Subsurface air and vapors migrate toward the
vacuum extraction well in response to the negative
pressure gradient around the well.
The two-phase flow of contaminated air and water
flows to a vapor/liquid separator where con-
taminated water is removed. The contaminated air
stream then flows through activated carbon
canisters arranged in parallel. Primary, or main,
adsorbing canisters are followed by secondary, or
backup, adsorbers to ensure that no contamination
reaches the atmosphere. Table A-l presents the
required equipment for the site, Figure A-l
illustrates the layout of wells and equipment, and
Figure A-2 shows a schematic diagram of an
extraction well.
In order for a vacuum extraction system to be
successful, the system design would have to consider
a number of important variable parameters. These
parameters are listed in Table A-2. The control
variables are the parameter that one has control over
in the design and operation of the vacuum extraction
system. The response variables are the parameters
that change in response to change in the control
variables. The initial variables are the parameters
that exist initially at the site. The spacing of vacuum
extraction wells is critical to efficient remediation.
Depending on the depth to groundwater and the soil
type, the radius of influence of an extraction well can
range from tens to hundreds of feet. Soil
permeability, porosity, moisture content,
stratigraphy, and depth of groundwater are
important factors in determining the radius of
influence. Terra Vac's vapor flow models are often
calibrated to site conditions to determine design
parameters and sensitivity before pilot testing or
full-scale cleanup is implemented.
Site conditions, soil properties, and the contaminant
chemical properties are the important considerations
in determining the success of a vacuum extraction
system. The depth of the vadose zone should be at
least 10 feet for cost effectiveness, since beyond this
depth excavation costs become very expensive and
far outstrip the costs of installing a vacuum
extraction system. The soil should have a sufficient
air-filled porosity to allow the vacuum and the
extraction air to do its job of in situ stripping of the
VOCs from the soil matrix. Water is a deterrent to
this stripping action as the water reduces the air-
filled porosity. The contaminant should have a
Henry's Constant of 0.001 or more for it to be
removed effectively in a vacuum extraction system.
Table A-3 lists the Henry's Constant at a
temperature of 10°C for typical VOCs. The
compounds are easier to remove by the vacuum
extraction process as one goes down the list.
17
-------
Table A-1. Groveland Site Equipment List
Equipment Number required
Description
Extraction wellsO.
Monitoring wells
Vapor-liquid separator
Activated carbon canisters
Vacuum pump skid
Holding tank
4 (2 sections each)
4 (2 sections each)
1
Primary: 2 units in parallel
Secondary: 1 unit
1
1
2 in. Sched. 40 PVC, 24-ft total depth
2 in. Sched. 40 PVC, 24-ft total depth
1,000-gal capacity, steel
Canisters with 1,200 Ib of carbon in each
canister - 304 SS
4 in. inlet and outlet nozzles
Terra Vac recovery unit model 17 25 HP motor
2,000-gal capacity, steel
N
MW4
Figure A-1. Schematic diagram of equipment layout.
18
-------
2"PVC Pipe
-
fS
=
_
=
=
-------
Table A-3. DImensionless Henry's Law Constants for
Typical VOCs @ 10°C
1 . methyl ethyl ketone
2. ethylene dibromide
3. 1,1,2,2-tetrachloroethane
4. dibromochloromethane
5. 1,1,2-lrichtoroethane
6. methyl isobutyl ketone
7. telralin
8. 1 ,2-dichloroethane
9. 1 ,2-dichloropropane
10. 1 ,2,4-trichlorobenzene
11. methylene chloride
12. 1 ,2-dichtorobenzene
13. chloroform
14. 1 ,4-dichlorobenzene
15. 1,3-dichtorobenzene
16. chlorobenzene
17. cis-1,2-dichloroethylene
18. o-xylene
19. ethylbenzene
20. benzene
21. methyl ethyl benzene
22. 1,1-dichloroethane
23. toluene
24. 1 ,3,5-trimelhylbenzene
25. m-xylene
26. p-xylene
27. trtchloroethylene
28. propylbenzene
29. lrans-l,2-dichloroethy!ene
30. chloroethane
31. 2,4-dimethylphenol
32. tetrachloroethylene
33. 1,1.1-tnchloroethane
34. carbon tetrachloride
35. vinyl chloride
36. 1,1-dichloroethylene
37. methyl cellosolve
38. tnchlorolluoromethane
39. decalin
40. cyclohexane
41. 1,1.2-tnchlorotrifluoroethane
42. n-hoxana
43. nonane
44. 2-methylpentane
0.0121
0.0129
0.0142
0.0164
0.0168
0.0284
0.0323
0.0504
0.0525
0.0555
0.0603
0.0702
0.0740
0.0702
0.0951
0.1050
0.1162
0.1227
0.1403
0.1420
0.1511
0.1584
0.1640
0.1734
0.1769
0.1808
0.2315
0.2445
0.2539
0.3267
0.3568
0.3641
0.4153
0.6370
0.6456
0.6628
1.8980
2.3068
3.0127
4.4329
6.6279
10.2430
17.2152
29.9975
20
-------
Appendix B
Vendor's Claims for the Technology
(NOTE: The following material was provided by
Terra Vac directly.)
Introduction
Vacuum extraction technology effectively removes
volatile and semivolatile compounds from soils and
groundwater. Removal of liquid-phase hydrocarbons
floating on the water table using vacuum extraction
technology is faster and more effective than
traditional approaches. Vacuum extraction is
typically implemented in situ; however, treatment of
excavated soils onsite using vacuum extraction
technology is also effective. Groundwater can be
removed simultaneously from vacuum extraction
wells to further enhance recovery of groundwater
contaminants and reduce the time frame for total
cleanup.
Vacuum extraction technology was originally
developed by Terra Vac. Since its inception more
than five years ago, the technology has been widely
used to clean up soil and groundwater contaminated
with volatile organic compounds (VOC). Two broadly
applicable process patents exist that relate directly
to vacuum extraction and the process of removing
volatile contaminants from the vadose zone using
vacuum (US Patent Nos. 4,593,760 and.4,660,639).
Most applications of the technology require a license
to use vacuum extraction to remove VOCs from the
vadose zone.
Cleanup of contaminated soils is often the most
important aspect of rapid, cost-effective remediation
at sites contaminated with volatile organics. A leaky
tank or pipeline, surface spill, storage lagoon,
landfill, or other release can quickly contaminate a
large volume of subsurface soil and rock. Vacuum
extraction removes contaminants directly from the
source area, eliminating further migration.
Implementation can be immediate, without risk of
complicating future cleanup efforts.
The high incidence of soil contamination observed at
leaky underground storage systems, industrial sites,
and Superfund sites requires that efficient and cost-
effective solutions be found to achieve cleanup
activities. Industrial experience with groundwater
remediation has shown that where soils are
contaminated and allowed to remain in place
untreated, the groundwater cleanup process is costly
and lengthy. Furthermore, without effective source
control, remedial efforts may never restore the
groundwater for use as a potable supply.
Potential Application
Waste Compatible with Technology
Vacuum extraction technology is effective in
treating soils containing virtually any chemical with
a volatile character. All of the volatile priority
pollutants and many of the semivolatiles have been
successfully extracted with the vacuum process.
However, metals (except mercury), heavy oils, and
PCBs will remain in place as the volatile compounds
are extracted by the process.
Terra Vac has applied vacuum extraction at more
than 60 sites across the country. Applications range
from small gas stations to large Superfund sites.
Large volumes (up to 7 million yd3) of contaminated
soil have been treated effectively with the process.
Remediation is rapid due to the high recovery rates
that have been achieved by Terra Vac: up to 4,000
Ib/d from a single well.
Contaminants Removed--
Terra Vac has successfully extracted the following
chemicals: >
Volatiles:
benzene
toluene
xylenes
ethylbenzene
hexane
chloroform
methylene chloride
tetrachloroethylene (PCE)
trichloroethylene (TCE)
dichloroethylene (DCE)
ethyl acetate
cyclohexane
methyl ethyl ketone (MEK)
methyl isobutyl ketone (MIBK)
21
-------
methanol
acetone
pyridine
tetrahydrofuran
dimethylfuran
carbon tetrachloride
trichloroethane (TCE)
Semivolatiles:
chlorobenzene
dichlorobenzene (DCB)
trichloropropane
Hydrocarbons:
gasoline
jet fuel
Stoddard solvent
diesel
kerosene
heavy naphthas
For those sites with numerous types of compounds
(i.e., VOCs, PCBs, pesticides, and metals) a phased
approach is often required. In these cases, it is
prudent to remove VOCs first using vacuum
extraction so that other technologies can then be
applied more cost-effectively and safely. For
example, in the chemical treatment or incineration
of soil, which requires excavation, the health risk of
excavation is minimized if the majority of VOCs are
removed first, in situ, by vacuum extraction. Not
only is the health risk minimized but also the
excavation is accomplished faster and more
economically because of the lower level of protection
required. Many methods used to chemically stabilize
metals are more effective after vacuum extraction
has removed VOCs.
Concentration Umits--
Vacuum extraction of VOCs has been applied at
numerous sites by Terra Vac to reduce soil and
groundwater contaminant levels from saturated
conditions down to nondetectable. Therefore,
concentration limitations are virtually eliminated
with this technology. Dual vacuum extraction of
groundwater and vapors has been effective at
restoring groundwater quality to drinking water
standards within short periods of time.
Favorable Conditions for Vacuum Extraction
Technology
Vacuum extraction has been demonstrated by Terra
Vac to be effective in virtually all hydrogeologic
settings:
• clays
• silts
• sands and gravel
• alluvium, colluvium, and glacial till
• wetlands
• fractured rock and karst
Advantages of vacuum extraction systems are that
cleanup of contaminated soil, free product, and
groundwater is in situ, rapid, and low cost. (See
Table B-l for a breakdown of costs for remediation
estimated for the Groveland site.) Vacuum
extraction systems are not limited by depth to
groundwater, with successful application
demonstrated by Terra Vac at sites with
groundwater as deep as 300 ft and as shallow as 3 in.
Table B-1. Terra Vac's Estimated Cost for Complete
Remediation of Valley Property at
Groveland, S/Ton
Site preparation and design 0.70
Permitting and regulatory
Equipment:
Rental 4.20
Consumables (piping and materials) 0.40
Contingency (10% of direct costs) 2.67
Startup and fixed cost
Installation 2.50
Startup 0.60
Mobilization and shakedown 0.30
Depreciation (10% of direct costs) 2.67
Insurance and taxes (10% of direct costs) 2.67
Labor costs 5.82
Supplies - Raw materials 5.82
Supplies - Utilities
Electricity 3.30
Effluent treatment, liquid and vapor 6.00
Residual disposal (included in effluent
treatment)
Analytical 3.90
Facility modifications
(10% of direct costs) , 2.67
Site demobilization 0.20
TOTAL 38.60
Pumping and treating groundwater is usually
conducted at all sites where groundwater quality is
significantly impacted. However, pump and treat
systems alone do not treat the soils and source areas
directly. Initial capital costs may be low to moderate,
but high operations and maintenance costs are
required as contaminants continue to leach from and
to soils. Pumping and treating groundwater alone is
22
-------
very slow, taking decades to restore aquifers to
common cleanup goals. ^
Free Product Removal--
Vacuum extraction is an effective means of removing
hydrocarbons floating on the water table. Compared
to typical double-pump systems, skimmer pumps, or
air-displacement product-recovery systems, vacuum
extraction is faster, more effective, and low in cost
per gallon of product removed. Liquid-phase
hydrocarbons are removed without pumping
groundwater so that separation and treatment of
large volumes of contaminated water is eliminated.
Where free product is present, traditional pump-and-
treat systems can cause groundwater quality to
become worse. As residual hydrocarbons are spread
throughout the cone of depression of a pumping well,
groundwater concentrations rise significantly
whenever pumps are turned off. However, if
combined with vacuum extraction, complete
remediation of soils and groundwater can be
achieved.
Dual Extraction—
Where contaminants have already reached the
groundwater, a "dual extraction" approach is
effective. Dual extraction is a term used to describe
the process of simultaneously extracting
groundwater and vapors under vacuum, using the
same well. In the simplest form, operating a
submersible pump within a vacuum extraction well
will lower the water table and increase the effective
unsaturated zone in which the vacuum extraction
process will vaporize contaminants.
Simultaneous extraction of groundwater and vapors
under vacuum has several benefits that enhance the
rate of groundwater cleanup. First, the rate of
contaminant removal increases compared to
groundwater extraction alone since contaminants
have two pathways for removal — aqueous phase and
vapor. Even in areas were there have been no sources
of soil contamination other than that provided by
groundwater movement beneath the water table, the
dual extraction process often yields the same mass
flux (i.e., Ib/d) from the vapor phase as the aqueous
phase. In medium-to-low permeability aquifers the
maximum rate at which groundwater can be
extracted from a given well increases two-to-three
fold. The net effect of these two phenomena can yield
a 6-fold increase in the overall contaminant removal
rate, and hence, a 6-fold reduction in the time
required to reach cleanup objectives.
Case Studies
The Terra Vac system was first demonstrated at a
Superfund site in Puerto Rico - the Upjohn facility
in Barceloneta — where carbon tetrachloride leaked
from an underground storage tank. The first aquifer
was the sole source of drinking water 300 feet below.
Although groundwater contamination occurred
rapidly, most of the pollutant was in the soil. Terra
Vac has remediated its assigned portion of this
Superfund site.
Since the first application in Puerto Rico, Terra Vac
has applied the vacuum extraction process to more
than 40 sites across the country. The process for the
removal of volatile contaminants from the vadose
zone of contaminated ground using vacuum
extraction is patented (U.S. Patent Nos. 4593760 and
4660639).
Following is a list of Terra Vac projects that are
either completed or are in various stages of
completion (see Table B-2):
1. Superfund Site, Puerto Rico
The cleanup of a large area contaminated with
carbon tetrachloride resulted in the development
of Terra Vac's vacuum extraction process. The
first three months of the project demanded daily
interaction with EPA officials, local agencies,
and regulators, since several municipal wells
were impacted. The vacuum extraction process
was developed in stages to recover solvents from
depths up to 300 ft, over an area of nearly
600,000 ft2. The process recovered up to 800 Ib/d
of carbon tetrachloride. As a result of Terra Vac's
process, more than 80% of the spill volume has
been recovered, and groundwater concentrations
have been reduced by 95%-99%.
2. Superfund Site, California
Terra Vac designed,, built, and operated a series
of vacuum extraction pilot tests for the recovery
of a wide range of solvents (including DCE, TCE,
and DCB) from clayey subsoil beneath an active
microchip manufacturing facility. Operation of
the extraction system required close monitoring,
due to the political climate in the area and the
close proximity of several resident communities.
The activated-carbon treatment system was
designed and operated for compliance with Bay
Area Air Quality standards while still
maintaining high carbon-usage efficiencies.
3. Verona Well Field Superfund Site, Michigan
The Verona Well Field was the first Superfund
site where EPA specified in the ROD that
vacuum extraction be used to clean up the
subsoils. Terra Vac was awarded the contract
based on its technical proposals and costs. The
pilot test for this site has been completed and the
full-scale vacuum-extraction-system (VES) is
23
-------
Table B-2. Partial List of Terra Vac Projects
Site
Tyson's
Barceloneta
Spill
Verona
Michigan
South Carolina
Michigan
Puerto Rico
Puerto Rico
Belleview
Florida
Manati
Puerto Rico
Sites Demo
Massachusetts
Type
Superfund
Superfund
Superfund
Industrial
Industrial
Major Oil
Refinery
Gas station
Gas station
Industrial
underground
tanks
Superfund
Date
Start/End
Nov 86/May 87
Nov 82/Nov 86
Sep 87/On
going
Oct 86/On
Going
Dec 86/Oct 87
Jul 85/Aug 86
June 85/Aug
85
Jan 86/Aug 87
Dec 84/Feb 85
Dec 87/Apr 88
Depth to
Groundwater Soil Length of
Compounds (ft) Type Operation
TCP, PCE
TCE
Xylenes
Toluene
Benzene
CCL4
PCE, TCE
MEK
MIBK
BTEX
PCE
TCE
PCE
Gasoline
Benzene
Benzene
Toluene .
Xylenes
Combustibles
Benzene
Toluene
Xylene
Gasoline
Hexane
DMF
THF
Acetone
TCE
18 Clay 30 days
300 1 . Clay 2.5 years
2. Rock
25 Sand/Silt 1 month
12-20 Clay 1 year
35 Sand 3 months
1 1 Sand/Silt 1 year
2-1 1 Clay/Silt 2 months
50 Clay/Sand 4 months
Rock
13 Sand 3 months
25 Sand/Clay 50 days
Cleanup
Goal
Pilot Test
No
1 ppm
No
Drinking water
5 ppb benzene
No
Demonstrate
technology
1 gal/day
extraction rate
Demonstrate
technology
Initial
Recovery Rate
(Ib/d) Status
20-100 Full-scale system
under
construction
250 Achieved goals
2,000 On going cleanup
100 On going 90%
cleanup
30-80 Achieved goals
2,000 Goal achieved
300 Goal achieved
2,000 Demonstration
completed
400 Cleanup complete
on going
monitoring
60 Demonstration
completed
-------
being designed and installed. The project is
expected to last two years.
4. Tyson's Superfund Site, Pennsylvania
A pilot vacuum extraction system (VES) was
installed to remove residual hydrocarbons and
solvents from old disposal areas. The system used
activated carbon to eliminate discharge of
hazardous components. The recovery rates from
the system totalled 150 Ib/d of benzene, toluene,
xylenes, TCE, PCE, and trichloropropane. Based
upon the pilot test results, EPA has modified the
previous ROD to specify vacuum extraction as
the preferred remedial technology. Terra Vac
worked with the client and the client's
consultant on the design of a VES to clean the
entire site. Terra Vac was responsible for
successful cleanup of the site. Total estimated.
cost of the cleanup was $4M.
5. Industrial Client, New Jersey
The site of a former chemicals-manufacturing
facility contained abandoned lagoons
contaminated with volatile organics. Terra Vac
has designed a pilot test in this hydrogeologically
complex setting that will use a dual extraction
system to recover contaminants deep within
clayey sediments. The initial groundwater level
was within six in. of the surface. The pilot test is
underway to determine whether vacuum
extraction can be used to successfully remove
volatile organics from subsoils at the site.
6. EPA, Massachusetts
Terra Vac was selected by EPA to demonstrate
vacuum extraction for the Superfund Innovative
Technology Evaluation (SITE) Program. The
purpose of the program is to document the
effectiveness of proven technologies.
7. Industrial Client, New Hampshire
Terra Vac is performing a pilot test of the
vacuum extraction process at this Superfund Site
for the Potentially Responsible Party (PRP)
Committee. The pilot test will demonstrate the
efficacy of the Terra Vac extraction process as
part of the Feasibility Study. The site contains
various chlorinated and aromatic solvents.
8. Industrial Client, Wisconsin
Terra Vac completed a cleanup of approximately
250 gal of high-boiling-point heavy-naphthas
from a former tank location. The cleanup was
accomplished in one month.
9. State Environmental Ag;ency, Florida
Terra Vac conducted a pilot test for the recovery
of gasoline from an underground pipeline leak at
a service station. The pilot test was successful
and continued for full cleanup of the site. The
cleanup achieved soil concentrations below the
proposed Florida Dept. of Environmental
Regulations (DER) limits for the indicator
compounds (including benzene) at 1 ppb.
10. Industrial Client, Michigan
Terra Vac installed a pilot VES to recover TCE
and PCE in soils that were contaminated from
old housekeeping problems. The pilot test was
extended into a full cleanup operation. The
cleanup goal of 5 ppb was achieved by early 1988.
11. Department of Energy, South Carolina
Terra Vac performed a pilot test of soil
contaminated with solvent from a leaking
pipeline. Successfully removed more than 250
Ib/d of PCE and TCE from highly stratified
sediments to a depth of 120 ft. Further testing
will follow to determine design parameters for a
full-scale cleanup.
12. Industrial Client, North Carolina
Vacuum extraction is being used to remove PCE
and TCE from soil and lagoon sludge. Recovery
rates of 100 to 300 Ib/d have been achieved. Also,
Terra Vac successfully recovered solvents from
under a building where sump leaks had occurred.
Full cleanup is expected by mid-1988.
13. Petroleum Spill, Puerto Rico
Efforts by conventional methods to remove
floating hydrocarbons from perched groundwater
after a pipeline leak were futile. Terra Vac
installed a vacuum recovery process that
removed residual hydrocarbons and about 1 in. of
free product within ten days of operation.
Floating product was perched above clayey soils
in gravel backfill. All tra.ces of free product were
eliminated from monitoring wells.
14. Petroleum Client, Puerto Rico
A leaky tank caused gasoline to seep into
surrounding soils. The contaminated soil
provided a source of hydrocarbon vapors to a
nearby underground utility line. During
installation of the electrical cables, explosive
conditions in the underground line interrupted
operations immediately. Up to 7 in. of free
25
-------
product was observed in the vacuum extraction
wells installed at the site.
15. Industrial Client, Puerto Rico
The soil around eight underground tanks is
being monitored for hydrocarbons by using the
Terra Vac vacuum extraction process. Monthly
testing for leaks has successfully detected
several leaks and spills. The same vacuum
system is being used for cleanup of the site.
16. Petroleum Client, Puerto Rico
A tank leak at a service station caused gasoline
contamination of a sewer and creek outfall. Terra
Vac operated a three-phase recovery system for
vapors, free product, and contaminated
groundwater. Complete cleanup of the area took
place in 14 weeks.
17. Petroleum Client, Puerto Rico
A leak at a service station reportedly released
800 gal of gasoline. After 13 weeks of vacuum
extraction, Terra Vac had reduced hydrocarbon
levels in soils from 600 ppm to nondetectable.
The equivalent of more than 1,400 gaLof gasoline
were recovered.
Cost Effectiveness
Vacuum extraction is typically more cost-effective
than any Nother treatment technology for soils
contaminated with VOCs. Typical implementation
costs range from $10/yd3 to $50/yd3. For extraction of
liquid-phase hydrocarbons floating on the water
table, vacuum extraction costs "about $10/gal of
gasoline to $35/gal for complete .onsite~ destruction of
hydrocarbons and compliance with rigorous air-
emissions controls. This compares to about $50/gal
for product recovery using typical dual-pump, total-
fluids or skimmer-type systems.
26
-------
Appendix C
Site Demonstration Results
Introduction
This SITE program demonstration test was planned
to determine the effectiveness of Terra Vac Inc.'s
vacuum extraction technology in the removal of
volatile organic compounds from the vadose zone.
The location of the test was on the property of an
operating machine shop. The property is part of a
Superfund site and is contaminated by degreasing
solvents, mainly trichloroethylene.
Objectives
The main objectives of this project were:
• The quantification of the contaminants
removed by the process.
• The correlation of the recovery rate of
contaminants with time.
• The prediction of operating time required
before achieving site remediation.
• The effectiveness of the process in removing
contamination from different soil strata.
Approach
The objectives of the project were achieved by
following a Demonstration Test Plan, which included
a Sampling and Analytical Plan. The Sampling and
Analytical Plan contained a Quality Assurance
Project Plan (QAPP). This QAPP assured that the
data collected during the course of this project would
be of adequate quality to support the objectives.
The sampling and analytical program for the test
consisted of four periods: pretest (or pretreatment);
active; midtreatment; and posttreatment.
The pretreatment sampling program consisted of:
• soil boring samples taken with split spoons
• soil boring samples taken with Shelby tubes
• soil gas samples taken with punch bar probes
Soil borings taken by split spoon sampling were
analyzed for volatile organic compounds (VOCs)
using headspace screening techniques, purge and
trap, GC/MS procedures, and the EPA-TCLP
procedure. Additional properties of the soil were
determined by sampling using a Shelby tube, which
was pressed hydraulically into the soil by a drill rig
to a total depth of 24 ft. These Shelby tube samples
were analyzed to determine physical characteristics
of the subsurface stratigraphy such as bulk density,
particle density, porosity., pH, grain size, and
moisture. These parameters were used to define the
basic soil characteristics.
Shallow soil-gas concentrations were collected
during pre-, mid-, and posttreatment activities. Four
shallow vacuum-monitoring-wells and twelve
shallow punch-bar-tubes were used at sample
locations. The punch bar samples were collected from
hollow stainless steel probes that had been driven to
a depth of 3 to 5 ft. Soil gas was drawn up the punch
bar probes with a low-volume personal pump and
tygon tubing. Gas-tight 50-mL syringes were used to
collect the sample out of the tygon tubing.
The active treatment period consisted of collecting
samples of:
• wellhead gas
• separator outlet gas
• primary carbon outlet gas
• secondary carbon outlet gas
• separator drain water
All samples with the exception of the separator drain
water were analyzed on site. Onsite gas analysis
consisted of gas chromatography with a flame
ionization detector (FID) or an electron capture
detector (BCD). The FID was used generally to
quantify the trichloroethylene (TCE) and trans 1,2-
dichloroethylene (DCE) values, while the BCD was
used to quantify the 1,1,1-trichloroethane (TRI) and
the tetrachloroethylene (PCE) values. The use of two
detectors, FID and ECD, was necessitated by high
27
-------
concentrations of TCE in the extracted wellhead gas.
Owing to the high TCE concentrations, most of the
samples injected on the ECD had to be diluted. Even
with dilution factors of 333 to 1, the TCE
concentration on the ECD would exceed the linear
range of the detector, thus necessitating the use of
two detectors.
The separator drain water was analyzed for VOC
content using SW846 8010. Moisture content of the
separator inlet gas from the wells was analyzed
using EPA Modified Method 4. This method is good
for the two-phase flow regime that existed in the gas
emanating from the wellhead. Table C-l lists
analytical methods used for this project.
Table C-1. Analytical Methods
Parameter
Grain size
pH
Moisture (110°C)
Particle density
Oil and grease
EPA-TCLP
TOG
Headspace VOC
VOC
VOC
VOC
VOC
VOC
VOC
Analytical method
ASTM D422-63
SW846' 9040
ASTM D2216-80
ASTM D698-78
SW846" 9071
F.R. 11/7/86,
Vol. 51, No. 216,
SW846" 8240
SW846' 9060
SW846"3810
GC/FID or ECD
GC/FID or ECD
SW846'8010
SW846'8010
Modified P&CAM 127
SW846" 8240
Sample source
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil gas
Process gas
Separator liquid
Ground water
Activated carbon
Soil borings
•Test Methods for Evaluating Solid Waste, 3rd Ed., Nov. 1986.
The posttreatment sampling essentially consisted of
repeating pretreatment sampling procedures at
locations as close as possible to the pretreatment
sampling locations.
The activated carbon canisters were sampled, as
close to the center of the canister as possible, and
these samples were analyzed for VOC content as a
check on the material balance for the process. The
method used was P&CAM 127, which consisted of
desorption of the carbon with CS2 and subsequent
gas chromatographic analysis.
f?esu/ts
VOC Removal from the Vadose Zone
The permeable vadose zone at the Groveland site is
divided into two layers by a horizontal clay lens,
which is* relatively impermeable. Each extraction
well had separate shallow and deep sections to
enable VOCs to be extracted from sections of the
vadose zone above and below the clay lens. The
quantification of VOCs removed was achieved by
measuring:
• gas volumetric flow rate by rotameter, and
wellhead gas VOC concentration by gas
chromatography
• the amount of VOCs adsorbed by the activated
carbon canisters by desorption into CS2
followed by gas chromatography.
VOC flow rates were measured and tabulated for
each well section separately. The results of gas
sampling by syringe and gas chromatographic
analysis indicate a total of 1,297 Ib of VOCs were
extracted over a 56-day period, 95% of which was
TCE. A very good check on this total was made by
the activated-carbon VOC analysis, the results of
which indicated a VOC recovery of 1,353 Ib; virtually
the same result was obtained by these two very
different methods.
One view of the reduction in VOC concentrations in
the vadose zone can be seen from examining the
three-dimensional shallow soil-gas plots. Soil gas
was collected during pretreatment, midtreatment,
and posttreatment from punch bar probes and
shallow vacuum monitoring-wells. The collection
points were located on a coordinate system with
Extraction Well 1 as the origin (0,0).
Each collection point has an x and y coordinate, and
TCE concentrations are plotted on a "Z" scale, which
gives a three-dimensional plot across the grid.
Values of "Z" between data points and around the
grid are generated by the Kriging technique, which
uses given data points and a regional variable theory
to generate values between and around sample
locations. Kriging is the name given to the least
squares prediction of spatial processes and is used in
surface fitting, trend surface analysis, and
contouring of sparse spatial data.
A total of 12 shallow punch-bar-tubes were used,
along with the 4 shallow vacuum monitoring-wells.
The punch bars were driven to a depth of 3 to 5 ft,
and, as with the vacuum wells, soil gas was drawn up
the punch bar probes with a low-volume personal
pump and tygon tubing. Gas-tight syringes (50 mL)
were used to collect the sample out of the tygon
tubing. The gas samples were analyzed in the field
trailer using gas chromatographs with flame
ionization detectors and electron capture detectors.
The soil gas results show a considerable reduction in
concentration over the course of the 56-day
demonstration period, as can be seen from Figures C-
28
-------
1 and C-2. This is to be expected, since soil gas is the
vapor halo existing around the contamination and
should be relatively easy to remove by vacuum
methods.
EW2
EW3
RAW
MW2
1200 EW3
EW4
Map View
W4
Figure C-1. Pretreatment shallow soil-gas concentration.
A more modest reduction can be seen in the results
obtained for soil VOC concentrations by GC/MS
purge-and-trap analytical techniques. Soil
concentrations include not only levels seen in the
vapor halo, but also interstitial liquid contamination
that is either dissolved in the moisture in the soil or
that exists with the moisture as a two-phase liquid.
Table C-2 shows the reduction of the weighted-
average TCE levels in the soil during the course of
the 56-day demonstration test. The weighted-
average TCE level was obtained by averaging soil
concentrations obtained every two feet by split-spoon
sampling methods over the entire 24-ft depth of the
wells. The largest reduction in soil TCE
concentration occurred in EW4, which had the
highest initial level of contamination. EWI, which
was expected to achieve the greatest concentration
reduction, exhibited only a minor decrease over the
course of the test. Undoubtedly this was because of
Figure C-2. Posttreatment shallow soil-gas concentration.
the greater-than^expected level of contamination
that existed in the area around MW3 that was drawn
into the soil around EWI. The decrease in the TCE
level around MW3 tends to bear this out.
Table C-2.
Well
Reduction of Weighted Average TCE
Levels in Soil (TCE concentration in mg/kg)
Pretreatment Posttreatment
Reduction
EW1
EW2
EW3
EW4
MW1
MW2
MW3
MW4
33.98
3.38
6.89
96.10
1.10
14.75
227.31
0.87
29.31
2.36
6.30
4.19
0.34
8.98
134.50
1.05
13.74
30.18
8.56
95.64
69.09
39.12
62.83
-
Extraction wells 2, 3, and 4 were designed to be
barrier wells for EWI. This means that they were
designed to intercept contamination from the highest
contaminated soils, which were under the building.
Since the area around MW3 was highly
contaminated and well within the radius of influence
29
-------
of EW1, the extraction wells EW1, EW2, and EW3 in
effect became barrier wells for EW4. The soil
adjacent to EW4 thus was cleaned to a far greater
extent that the other three wells.
Effectiveness of the Technology in Various Soil
Types
The soil strata at the Groveland site can be
characterized generally as consisting of the following
types in order of increa'sing depth:
• medium to very-fine silty sands
• stiff and wet clays
• sand and gravel
Soil porosity, which is the percentage of total soil
volume occupied by pores, was relatively the same
for both the clays and the sands. Typically, porosity
over the 24-ft depth of the wells would range between
40% and 50%. Permeabilities, or more accurately
hydraulic conductivities, ranged from 10-4 cm/s for
the sands to 10-8 cm/s for the clays, with
corresponding grain sizes in the range to 10-1 to 10-3
mm.
Pretest soil boring analyses indicated in general that
most of the contamination was in the stratum above
the clay lens, with a considerable quantity perched
on top of the clay lens. This was the case for EW4,
which showed an excellent reduction of TCE
concentration in the medium-to-fine sandy soils
existing above the clay layer, with no TCE detected
in the clay in either the pretest or posttest borings
(see Table C-3). One of the wells, however, was an
exception. This was MW3, which contained the
highest contamination levels of any of the wells, and
was exceptional in that most of the contamination
was in a wet clay stratum (see Table C-4). The levels
of contamination were in the 200-1,600 ppm range
before the test. After the test, analyses of the soil
boring adjacent to MW3 showed levels in the range of
ND-60 ppm in the same clay stratum. The data, as
shown in Table C-5, suggest that the technology can
desorb or otherwise mobilize VOCs out of certain
clays.
From the results of this demonstration, it appears
that the permeability of a soil need not be a
consideration in applying the vacuum extraction
technology. This may be explained by the fact that
the porosities were approximately the same for all
soil strata, so that the total flow area for stripping air
was the same in all soil strata. It will take a long
time for a liquid contaminant to percolate and
permeate through clay with its small pore size and
consequent low permeability. However, the much
smaller air molecules of vapor contaminant have a
lower resistance in passing through the same
pores.This may explain why contamination was
generally not present in the clay strata, but when it
was, it was not difficult to remove. Further testing
should be done in order to confirm this finding.
Correlation of Declining VOC Recovery Rates to
Time
The vacuum extraction of volatile organic
constituents from the soil may be viewed as an
unsteady state process taking place in a
nonhomogeneous. environment acted upon by the
combined convective forces of induced stripping air
and by the diffusion of volatiles from a dissolved or
sorbed state. As such it is a very complicated process,
even though the equipment required is very simple.
Unsteady-state diffusion processes in general
correlate well with time by plotting the logarithm of
the rate of diffusion versus time. Although the
representation of the vacuum extraction process
presented here might be somewhat simplistic, the
correlation obtained by plotting the logarithm of the
concentration of contaminant in the wellhead gas
versus time and obtaining a least-squares best-fit
line was reasonably good. This type of plot, shown in
Figure C-3, represents the data very well and is more
valid than both a linear graph or one plotting
concentration versus log time, in which a best fit
curve would actually predict gas concentrations of
zero or less.
Looking at the plots for EW1, shallow and deep,
equations are given for the least-squares best-fit line
for the data points. If the vacuum extraction process
is run long enough so that the detection limit for
TCE on the BCD, which is 1 ppbv, is reached, the
length of time required to reach that concentration
would be approximately 250 days on the shallow well
and approximately 300 days on the deep well.
Prediction of Time Required for Site Remediation
The soil concentration that would be calculated from
the wellhead gas concentration using Henry's Law is
included in the last column of Table C-5.
Calculations for the predicted soil concentrations
were made assuming a bulk density of the soil of
1,761 kg/m3, a total porosity of 50%, and a moisture
content of 20%. The calculated air-filled porosity of
the soil is approximately 15%. Henry's Constant was
taken to be 0.492 kPa/m3-gmol at 40°F
Given the nonhomogeneous nature of the subsurface
contamination and interactions of TCE with organic
matter in the soil, it was not possible to obtain a good
correlation between VOC concentrations in wellhead
gas and soil in order to predict site remediation
times. Henry's Law Constants were used to calculate
soil concentrations from wellhead gas
concentrations; and the calculated values obtained,
30
-------
Table C-3. Extraction Well 4: TCE Reduction in Soil Strata
TCE cone., ppm
l^OfJU 1
(ft)
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
ufjOlst I|JUUI 1
of strata '
Med. sand w/gravel
Lt brown fine sand
Med. stiff It. brown fine sand
Soft dk. brown fine sand
Med. stiff brown sand
V. stiff It. brown med. sand
V. stiff brown fine sand w/silt
M. stiff grn-brn clay w/silt
Soft wet clay
Soft wet clay
V. stiff brn med-coarse sand
V. stiff brn med-coarse w/gravel
r GI i iiGaiuMuiy -
(crn/s)
10-4
10'4
10-s
TO'5
10-4
10-4
10-4
10-8
TO'8
ID'8
10-4
TO'3
pre
2.94
29.90
260.0
303.0
351.0
195.0
3.14
ND
ND
ND
ND
6.71
post
ND
ND
39.0
9.0
ND
ND
2.3
ND
ND
ND
ND
ND
ND - Non-detectable level.
Table C-4. Monitoring Well 3: TCE Reduction in Soil Strata
TCE cone., ppm
L^vsf^ll 1
(ft)
0-2,
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
is*j&\ji i|jiiui i
of strata
M. stiff brn. fine sand
M. stiff grey fine sand
Soft It. brn. fine sand
Lt. brn. fine sand
Stiff V. fine brn. silty sand
Silty sand
Soft brown silt
Wet green-brown silty clay
Wet green-brown silty clay
Wet green-brown silty clay
Silt, gravel, and rock frag.
M. stiff It. brn. med. sand
r GI iMoexuuiiy
(cm/s)
10-s
10-5
10-4
ID'4
10"*
10-4
10-4
10-8
10-8
10-8
10-4
10-4
pre
10.30
8.33
80.0
160.0
ND
NR
316.0
195.0
218.0
1570.0
106.0
64.1
post
ND
800.0
84.0
. ND :
63.0
2.3
ND
ND
62.0
2.4
ND
ND
ND - Non-detectable level.
NR - No recovery of sample.
correcting for air-filled porosity, were lower than
actual soil concentrations by at least an order of
magnitude (see Table C-5).
Before one can attempt to make a rough estimation
of the remediation time, a target value for the
particular contaminant in the remediated soil must
be calculated. This target concentration is calculated
by using two mathematical models, the Vertical and
Horizontal Spread Model and the Organic Leachate
Model. (Federal Register, Vol. 50, No. 229,
Wednesday, November 27, 1985, pp. 48886-48910.)
The mathematical models allow the use of a
regulatory standard for drinking water in order to
arrive at a target soil concentration.
31
-------
Extraction Well #1
Shallow
>
I
o
HI
•
•
^
— x
*t"^*^
""k**vs*
.• •• -^<-
•^*v^'
^"•v^
^"-«s^
\^
^\
0 20 40 60 80 10
Day of Active Treatment
Y = 159.33 »EXP(-°-°Sxj
Curve Coefficient: R2 = 0 62
Figure C-3. Wellhead TCE concentration vs. time.
Table C-5. Comparison of Wellhead Gas VOC
Concentration and Soil VOC Concentration
cy =
Extraction
well
1S
1D
2S
20
3S
3D
4S
TCE
concentration
in wellhead gas
(ppmv)
9.7
5.6
16.4
14.4
125.0
58.7
1.095.6
TCE
concentration
in soil
(ppmw)
54.5
7.2
ND
20.4
20.9
18.0
9.1
Soil
concentration
predicted by
Henry's Law
(ppmw)
0.11
0.07
0.20
0.17
1.53
0.74
12.49
C0 =
erf =
z -
Y =
X =
at =
The VHS model is expressed as the following
equation:
Cy = C0erf(Z/(2(azY)0.5))erf(X/(atY)0.5)
where:
concentration of VOC at compliance point
(mg/L)
concentration of VOC in groundwater
(mg/L)
error function (dimensionless)
penetration depth of leachate into the
aquifer
distance from site to compliance point (m)
length of site measured perpendicular to
the direction of ground water flow (m)
lateral transverse dispersivity (m)
vertical dispersivity (m)
A simplified version of the VHS model is most often
used, which reduces the above equation to:
Cy = C0Cf
where:
32
-------
Cf = erf (Z/(2(azY)0.5)) erf (X/(atY)0.5), which is
reduced to a conversion factor
corresponding to the amount of
contaminated soil
The Organic Leachate Model (OLM) is written as:
C0 = 0.00211 CS0-6Y8S0.373
where:
C0 = concentration of VOC in groundwater
(mg/L)
Cs = concentration of VOC in soil (mg/L)
S = solubility of VOC in water (mg/L)
The regulatory standard for TCE in drinking water
is 3.2 ppb. This regulatory limit is used in the VHS
model as the compliance point concentration in order
to solve for a value of the groundwater concentration.
This value of groundwater concentration is then used
in the OLM model to solve for the target soil
concentration.
Once the target soil concentration is determined, a
rough estimation of the remediation time can be
made by taking the ratio of soil concentration to
wellhead gas concentration and extrapolating in
order to arrive at a wellhead gas concentration at the
target soil concentration. The calculated target soil
concentration for this site is 500 ppbw. This
corresponds to an approximate wellhead gas
concentration of 89 ppb for EW1S. The equation
correlating wellhead gas concentration with time
(see Figure C-3) is then solved to give 150 days
running time.
After 150 days the vacuum extraction system can be
run intermittently to see if isignificant increases in
gas concentrations occur on restarting, after at least
a two-day stoppage. If there are no appreciable
increases in gas concentration, the soil has reached
its residual equilibrium contaminant concentration
and the system may be stopped and soil borings
taken and analyzed.
33
-------
Bibliography
1. American Petroleum Institute, 1972. The
Migration of Petroleum Products in Soil and
Groundwater. Pub. No. 4149. Washington, D.G.
2. Eklund, B. 1985. Detection of Hydrocarbons in
Groundwater by Analysis of Shallow Soil
Gas/Vapor. API Publication 4394.
3. Goring, C.A.I., J.W. Hamaker, and J.M.
Thompson. 1972. Organic Chemicals in the Soil
Environment. Marcel Dekker, Inc., New York
City.
4. Hougen, O.A., K.M. Watson, and R.A. Ragatz.
1954. Chemical Process Principles. John Wiley
and Sons, Inc., New York City.
5. Hutzler, N.J., J.S. Gierke, and L.C. Krause.
1988. Movement of Volatile Organic Chemicals
in Soils in Reactions and Movement of Organic
Chemicals in Soils. Shawney, B.L., Ed. Soil
Science Soc. of America, Madison, Wis.
6. Kerfoot, H.B., and C.L. Mayer. 1986. The Use of
Industrial Hygiene Samplers for Soil Gas
Measurements. Groundwater Monitoring
Review.
7. Klute, A., Ed. 1986. Methods of Soil Analysis,
Part 1, Physical and Mineralogical Methods,
2nd ed. American Soc. of Agronomy, Inc., Soil
Science Soc. of America Inc., Madison, Wis.
8. Malot, J.J. 1985. Unsaturated Zone Monitoring
and Recovery of Underground Contamination.
Fifth National Symposium on Aquifer
Restoration and Groundwater Monitoring,
Columbus, Ohio.
9. Werner, M.D. 1985. The Effects of Relative
Humidity on the Vapor Phase Adsorption of
Trichloroethylene by Activated Carbon.
Journal American Industrial Hygiene
Association.
10. USEPA. 1986. A Primer on Kriging. U.S.
Environmental Protection Agency, Statistical
Policy Branch, Office of Policy, Planning, and
Evaluation, Washington, D.C.
11. Federal Register, Volume 50, No. 229,
Wednesday, November 27, 1985, pp. 48886-
48910.
12. USEPA. 1988. State of Technology Review Soil,
Vapor Extraction Systems. U.S. Environmental
Protection Agency, Office of Research and
Development, Cincinnati, Ohio.
34
-------
Appendix D
Case Study D-1
Upjohn Superfund Site at Barceloneta, Puerto Rico
Introduction
The Upjohn facility is an active pharmaceutical
manufacturing plant located on the north coast of the
island of Puerto Rico on the boundary between the
municipalities of Arecibo and Barceloneta. The
Atlantic Ocean lies about 3.7 miles to the north,
A number of chemical feedstocks are stored in
underground tanks at the Upjohn facility. These
chemicals are processed and warehoused in plant
buildings. The waste chemicals from processing are
temporarily stored in underground tanks in the tank
farm. It is from one of these tanks that
approximately 15,300 gal of organic chemical wastes
spilled in August 1982. This mixture was believed to
contain 65% carbon tetrachloride (CCU) and 35%
acetonitrile [D-1].
In January 1983, Upjohn began drilling more than
50 boreholes, ranging in depth from 30 to 200 ft, in
the unsaturated zone in and around the tank farm to
delineate the zones of highest contamination. Carbon
tetrachloride concentrations were measured in the
headspace of soil samples collected from each
borehole. The investigation revealed that the CC14
concentrations in each borehole varied widely but, in
general, tended to decrease with depth. At the
soil/limestone interface, the concentrations
measured by the headspace method dropped
dramatically to very low levels. The exception to this
pattern was for soils sampled in very deep deposits of
the blanket sands at depths ranging from 100 to 210
ft below the ground surface, in which elevated levels
of CCU were evident.
A pilot vacuum-extraction-system (VES) was
implemented at the site by Terra Vac Corp. in
January 1983. The pilot system basically consisted of
three vacuum extraction wells (VE), four vacuum
monitoring wells (VM), one vacuum pump (VP), a
cold water condenser, and a collection tank.
The primary extraction well (VE-1) was installed to
a depth of 75 ft. In May 1983, three more wells and
two vacuum pumps were installed. By the time the
complete system was fully constructed, late in 1984,
a total of 19 vacuum extraction wells had been used
in conjunction with six vacuum pumps.
The initial operation of the VES used the cold water
condenser to condense the liquids out of the soil gas.
The collected liquids consisted of primarily water
contaminated with CCU and traces of acetonitrile
[D-2]. The stack gas, also contaminated (mainly with
CCU), was vented through the exhaust stack without
any offgas treatment. Comparison between the
amount of CCU condensed from the cold water
condenser and the exhaust stack indicated that the
rate of CCU removal through the stack was more
than 1,000 times higher than collected by the
condenser [D-2]. Therefore, the condenser was
eliminated from the process for subsequent
operations.
Most of the vacuum extraction operations relied on
the air dispersion exhaust stacks. The impact of the
exhausted CCU on the quality of air in the tank farm
was evaluated by taking vapor samples at various
distances from the vent stack under different wind
conditions. The field data indicated that the
threshold limit value (TLV) imposed by OSHA was
reached within about 30 ft of the 10-ft stack under
worst conditions (i.e., low wind and high
concentration). The area around the stack was roped
off up to 50 ft and designated a restricted area.
Consequently, studies were undertaken by Upjohn to
improve the conditions. The immediate solution was
to increase the stack height to 30 ft. Air quality
modeling was also conducted to predict the impact of
the complete (expanded) VES on air quality. These
analyses showed that there was no need to
incorporate any offgas treatment system prior to
exhaust stacks.
For the pilot system operation., distances between the
monitoring wells and the primary vacuum
extraction well (VE-1) were based on preliminary
calculations of the radius of influence of the vacuum.
As a result, monitoring wells were spaced at 3, 5, 10,
and 30 ft from VE-1 in locations accessible by a drill
rig. Based on the continuous pressure monitoring at
these wells, it was found that the radius of influence
of the vacuum extraction wells deve-loped to only 3 ft
35
-------
after 3 weeks of operation and finally stabilized at 10
ft after about 90 days.
Testing and Closure Protocols
Vacuum monitoring wells were used for vapor
samples and for monitoring subsurface vacuum. The
vapor samples from these monitoring wells were
analyzed for CC14 two or three times per week,
whereas subterranean pressures were recorded
daily. Gas samples from the vent stacks were
analyzed three times a day for CCU [D-l]. Most of
these samples were analyzed using the field gas
chromatograph. Daily measurements of the volume
of liquids collected and volumetric flow rate of vapors
passing through the vacuum pump were used to
calculate the daily volumes of CCU extracted.
Qualitative analyses of the vented fumes and
collected liquids from the vacuum extraction systems
indicated the presence of CCU, and traces of
acetonitrile, acetone, chloroform, methylene chloride
and methane. Since CCU was the major constituent
of the waste stream, and because it is the most toxic
compound of the chemicals detected, it is the only
chemical that was quantitatively analyzed.
In addition to the above, soil samples were taken
from test borings periodically. The "Standard
Penetration Test" was used to collect the soil
samples. The soil samples were analyzed in the field
using the "headspace method" of analysis using a
portable gas chromatograph with a flame ionization
detector.
The question of when the VES had completed its job
was very difficult to answer. In June 1984, Geotec, a
consulting engineering company in Puerto Rico that
prepared an RI/FS for the Upjohn facility, presented
a criterion for the termination of the VES [D-3]. This
closure criterion was based on the economic
relationship between extraction from the aquifer
versus extraction using the vacuum system. The
allowable drinking water limit for CCU at the time
was 50 ppb. Based on this limit, the cost of pumping
sufficient quantity of groundwater out to extract one
pound of CCU was estimated. Determination of the
equivalent removal rate by the vacuum system at
the same operating cost was proposed as the
economic termination point for the vacuum system.
In simple words, there is a minimum concentration
in the vacuum exhaust that is as economical in
removing CCU from the soil as compared to pumping
water contaminated with CCU from the aquifer.
Such an analysis showed that in the tank farm area,
the concentration of CCU in the exhaust stack must
go down to 2.2 ug/L before the vacuum system should
be terminated. For the analytical procedure used,
this translated to a non-detect level of CCU in the
exhaust gases.
The VES continued to operate for another two years
without showing non-detectable contamination in
the exhaust gauges. Therefore, Upjohn
Manufacturing Co. drilled four boreholes in the area
of originally high contamination [D-4]. The
"Standard Penetration Test" was used for these
boreholes, which varied from 40 to 105 ft, measured
from grade elevation. The results showed CCU to be
non-detectable in all samples. Based on these results,
Upjohn argued that the vacuum extraction must be
accepted as complete.
Since the above proposition was based on only four
boreholes, the regulatory agency did not accept it.
The closure criterion established by the agency
required non-detect levels of CCU in all the exhaust
stacks for three consecutive months. This was finally
achieved in March 1988 when the VES was termi-
nated.
Major Conclusions
The following conclusions are made from this case
study.
- The VES proved effective in removing CCU
from the clayey soils and limestone beneath the
tank farm. The action significantly reduced the
amount of CCU entering the aquifer from the
vadose zone.
- The offgas from the vacuum extraction process,
in this case, did not require any treatment
before atmospheric venting.
- The concentration of CCU in the offgas after
any system shutdown period was found to be
higher than that measured before the system
shutdown.
- The demonstration of non-detect indication of
CCU in the offgas for three consecutive months
was an acceptable criterion (by the regulatory
agency) for termination of the vacuum
extraction system operation.
- The VES seemed to perform well with the soil
with measured permeability in the range 10-4-
10-7 cm/s.
Data Summary
Data from the Upjohn site is presented in detail in
References [D-l] through [D-5]. A summary of some
key data is provided below.
36
-------
Soil data before treatment:
concentrations as high as 2,200 ppm
measured.
- Generally, concentration of CCl^ decreased
with depth.
- Very high concentrations found around 100 ft
deep in deposits of blanket sands.
in 1982. This could be explained by the following
possibilities;
- Other storage tanks may have ruptured and
leaked CC14.
concentration higher in clay as compared
to that in limestone.
Soil data after treatment (from Reference D-4):
Bore
No.
102
102
102
Depth of
sample, ft
30
50
105
CCI4
ND
ND
ND
CHCLg
ND
ND
ND
CH2CL2, ng/kg
11.2
4.0
13.0
Offgas data:
During initial operation, concentration of CCL.4 in
the offgas was as high as 70 mg/L or 10,500 ppm in
air.
CC14 removal:
During the initial pilot operation from January 12,
1983 to April 25, 1984, approximately 8,000 gal of
CC14 were reported to be extracted from the soil [D-
2]. A cumulative amount of 17,781 gal was finally
removed from the soil [D-l]. Since the initial spill
was reported to be 15,300 gal, of which 65% was
CC14, only 9,945 gal were pure CC14. These values
indicate that the amount of CCLt removed by the
VES was almost double what was reported as spilled
- Leakage of CC^ may have occurred prior to
1982, but was not detected.
- The original mixture of the waste material
originally spilled contained a higher percentage
of CCU than the reported value of 65%.
References
D-l. Feasibility Study, Upjohn Manufacturing Co.,
Barceloneta, Puerto Rico. Final report
prepared by CDM Federal Programs Corp.
under U.S. EPA Contract No. 68-01-7331,
June 1988.
D-2 Remedial Investigation Report and Feasibility
Study for Upjohn Manufacturing Co.,
Barceloneta, Puerto Rico. Prepared by Geotec
for Upjohn Manufacturing Co., May 18, 1984.
D-3 Evaluation of Closure Criteria for Vacuum
Extraction at Tank Farm, Upjohn
Manufacturing Company, Barceloneta, Puerto
Rico. Prepared by Geotec, June 12, 1984.
D-4 Vacuum Assessment Report, TFI Vacuum
Assessment Investigation. Prepared by Soil
Tech for Upjohn Manufacturing Co., March
23, 1987.
D-5 Malot, J.J. Unsaturated Zone Monitoring and
Recovery of Underground Contamination.
Terra Vac, Inc., P.O. Box 550, Dorado, Puerto
Rico. Presented at the Fifth National
Symposium on Aquifer Restoration and
Ground Water Monitoring, May 21-24, 1985.
37
-------
-------
Case Study D-2
Assessment of Vacuum Extraction Technology Application Bellview
(Florida) UST Site
Introduction
The Bellview gasoline contamination site is located
in Marion County in the north central portion of
Florida. This site first came to the attention of the
Florida Department of Environmental Regulation
(FDER) in late 1982. The chronology of events at this
site is shown in Table D-l.
Nature of Contamination
Previous investigations directed by the FDER
indicated that an underground storage tank at the
Union 76 gas station in Bellview, Fla., was a source
of subsurface hydrocarbons. In December 1976, four
6-1/4 in. soil borings were made at this site, and soil
samples were obtained every 2-1/2 ft. The maximum
depths sampled were from 52 to 60 ft. In addition,
soil samples were obtained with a hand auger at
three locations in the storage tank area. All samples
were analyzed onsite for gasoline components.
Benzene, toluene, ethylbenzene, and xylenes
(BTEX), and total volatile hydrocarbons were
quantified by gas chromatography. Initial soil
concentrations of BTEX were as high as 97 mg/kg,
and total hydrocarbons were as high as 230 mg/kg.
The highest concentrations of benzene, toluene, and
xylene were observed at approximate depths of 10 to
20 ft, above a clay layer and perched water table.
Soil/site Conditions
The soil borings revealed four distinct stratigraphic
zones. Clayey sands were observed from the surface
to depths of 18 to 21 ft, where gumbo (a plastic clay)
was encountered. The thickness of this clay layer
varied from 5 to 13 ft below the surface. The silty
sand layer was underlain by a weathered limestone
that consisted of sand, shells, and cavities in the
upper portion. Geologists noted that this limestone
layer is probably part of the Upper Eocene Ocala
formation. Groundwater was encountered at depths
between 48 and 53 ft below the surface, while
perched groundwater was observed above the clay
layer in wells VE-1, ME-1, and ME-2 (Figure D-i).
Experimental/System Design
The soil VES consisted of: six extraction wells; a
vacuum pump; a gas flow meter; various plumbing,
valves, gauges, and sampling ports; a gasoline/water
separator; and monitoring wells (Figure D-2).
Pavement, which was already in place, was used as a
cap. A vacuum extraction/monitoring well was
installed in each borehole (see Figure D-3 for
locations). The wells VE-1 and VE-2 were used
primarily for subsurface hydrocarbon vacuum-
extraction. Multi-level, dual-purpose wells, which
could monitor the subsurface vacuum as well as
extract hydrocarbons from two to three
hydrogeologic zones, were installed at the other two
boreholes (ME-1 and ME-2). Well ME-1, which
consisted of three monitoring wells, was capable of
monitoring the subsoil at depths of 13, 35, and 50 ft.
Well ME-2 actually consisted of two monitoring
wells, at depths of 16 and 58 ft. Each of the wells was
connected to the vacuum extraction unit by way of a
manifold system. So as not to interfere with the
continuous operation of the service station,
wellheads were installed in underground valve
boxes, and the vacuum extraction manifold was
covered by concrete. The system was modified to
include an inline air/water separator to separate
small quantities of gasoline product and water that
were being extracted from the subsoils along with
the hydrocarbon vapor.
Status of Experiment/Cleanup
A pilot test of the vacuum extraction operation
started on January 29, 1987. By the end of February
1987, the pilot test results indicated that it was
necessary to operate the system continuously in
order to estimate the time required for cleanup. Due
to power outages and numerous administrative prob-
lems (for example, approval of a permit to discharge
extracted water), the system experienced limited
operation during the months of March, April, and
May. During June and July, the system was
operating on a nearly continuous basis. Initial
extraction rates for gasoline hydrocarbons ranged
39
-------
Table D-1. Chronology of Events at the Bellview, Florida, UST Site
Date Events
Oct 1979-
MaM980
Aug 1982
Oct 1982-
Jun 1984
Dec 1984
Sep1986
Dec 1986
Jan 1987
Aug 1987
Sep 1987-
Oct 1987
Sep 1988-
Oct 1988
Unleaded gas (10,000 gal) leaked from buried pipeline.
Contamination discovered in city water-supply wells.
City abandoned water-supply wells; developed new wells outside city limits
State conducts a contamination assessment study.
State negotiates with Terra Vac to evaluate technology at this site.
Terra Vac installs a series of extraction wells.
Terra Vac VES begins operation.
Terra Vac VES shut down after State contract expired.
EPA conducts vacuum extraction technology evaluation with existing system.
Startup and first two weeks of operation of the Terra Vac VES.
West
East
10-
20-.
30-
40-
50-
60 J
VE2
ME1
13' 50'
VE1
ME2
16' 58'
VE7 VE5
0 5' 10'
Scale: Horizontal = Vertical
Figure D-1. Soil formations at Bellview, FL.
40
-------
ME-2-58
ME-1-50
Legend
PI - Pressure Indicator
VN1 - Venturi Meter at Exhaust
VN2 - Venturi Meter from Wells
AB1A.AB2A - 2"Annubars
AB1,AB2,AB3 - 4"Annubars
TE - Temperature Elements
WS - Water Separator
SP - Sampling Point
Figure D-2. Process schematic at Bellview, FL.
from 295 Ib (39 gal) per day in VE-1, to 1,950 Ib (260
gal) per day in ME-1-50. During this pilot test,
wellhead concentrations decreased with time, which
would indicate the subsoils were being cleaned up. A
total of 22,027 Ib (2,937 gal) of gasoline hydrocarbons
had been extracted from the site as of August 1987.
An independent evaluation of this system was
conducted in September through October 1987 for
EPA by Camp, Dresser, McKee, Inc. (COM).
Additional soil borings, soil vapor samples, and
groundwater samples were collected. During this 25-
day evaluation period, 22 additional pounds of BTEX
41
-------
Union 76
Service Station
Underground
Tanks
CTJ
ME-1
Underground
Pipe Line
Pavement
S E Abshier Avenue
Figure 0-3. Vacuum extraction well location map.
were removed. Taking into account other volatile
gasoline components removed, an additional 200 Ib
of hydrocarbons were removed. No significant
changes in the soil, soil vapor, or groundwater sam-
ples were noted.
Results and Conclusions
Based on the Terra Vac work under contract from
FDER and the evaluation done by Camp, Dresser,
and McKee, Inc. (under EPA contract), the latter
reported the following results and conclusions.
1. Benzene, toluene, ethylbenzene, and xylenes
were removed in significant quantities by the
application of the vacuum extraction system to
the subsurface zones. In this system, where
initial soil concentrations of total BTEX ranged
from 0.2 to 12 ug/g, a total of approximately 10
kg of BTEX was removed during the 25-day test
period of this study.
In addition to the BTEX removed, other major
volatile gasoline components were removed.
The total hydrocarbons, removed was
approximately 90 kg.
Overall removal of BTEX from the soil matrix
could not be determined on an overall mass
balance, based on the initial and final soil
sampling, because of the heterogeneity of the
BTEX contamination and the relatively small
number of soil samples analyzed.
The extracted soil gas concentrations
demonstrated a decrease in the daily quantity
of BTEX extracted over the study period. The
decrease in the BTEX extraction rate indicates
a relative decrease in the concentration of the
source material. Decreased soil gas
concentrations resulted from mass transfer
between the following zones: from the surface of
the liquid to the bulk surface of the soil to the
internal pore space. The rate limiting step is
the diffusion of the contaminants from the
pores.
The groundwater quality was not affected by
the vacuum extraction system, even though a
42
-------
measurable air flow was observed within the
vicinity of high ground water contamination.
The variations in groundwater quality over the
vacuum-extraction system operation can be
attributed to seasonal fluctuations in the water
table and migration of contaminants away from 9.
the source area.
Based on a statistical comparison to the soil gas
data collected in the initial and final soil gas
surveys, the mean concentrations of total BTEX
were within the same concentration range. The
mean soil concentrations for total BTEX from
the initial and the final soil gas survey were in
the range of <0.2 to 0.3 ppmv. The results of
the surveys demonstrate no statistical
difference between the two data sets. However,
it is clear that soil gas concentration levels were
near or below detectable limits for BTEX.
A statistical comparison of the initial and final
soil-sampling analyses demonstrate no
difference between the two data sets for 10.
toluene, ethylbenzene and total xylenes. The
benzene concentrations demonstrated a real
difference between the initial and final data
sets, in the increased levels of the final soil-
sampling data. The mean concentration for
each individual BTEX component was in the
range 0.2 to 1.3 ug/g. Concentrations were not
observed above 6.0 ug/g.
Monitoring wells MW-2, MW-6, and MW-7 all
had measurable concentrations of total BTEX
for both groundwater sampling events. All
other monitoring wells were near or below the
detection limit for each BTEX component of 1.0
ug/L. MW-6 contained the highest BTEX
concentration of about 10,400 ug/L in the initial
sampling program and about 10,700 ug/L in the
final sampling program. A comparison between
each BTEX component showed no statistical
difference between the two data sets for either
MW-6 or MW-7 groundwater quality.
The quantity of contaminants extracted from
each extraction well varied greatly over the n.
spatial distribution of the site. In general, the
shallow extraction wells located in the clayey
sands exhibited significantly higher extracted
soil gas concentrations than observed from the
deep extraction wells located in the weathered
limestone formation. However, the deep
extraction wells exhibited significantly greater
air flow rates than the shallow extraction wells
when the same vacuum was applied at the
wellhead. The net result of this combination of 12.
soil/gas concentration and the air flow rate, the
BTEX extraction rate, demonstrated that the
deep extraction wells extracted more mass of
contaminants (approximately 70% of the total
mass extracted) than the shallow extraction
wells did, even though the relative contam-
ination was greater in the clayey sands.
The quantity of air extracted from the
extraction wells varied greatly with the soil
type under the same applied vacuum. In
general, each shallow extraction well yielded
approximately 570 L/m (20 cfm) of air at
applied wellhead vacuums of about 130 cm (50
in.) of water. These wells were screened at the
interface of clayey sand and gumbo clay
formations.
Two of the deep extraction wells each yielded
approximately 8,500-11,300 L/m (300-400 cfm)
of air at applied vacuums of about 80-100 cm
(30-40 in.) of water. These wells were screened
in a weathered limestone formation where
significant cavities were observed.
The transmission of the applied wellhead
vacuum through the soil matrix was also
affected by the soil type. The cone of influence
or convective air-flow boundary was
significantly less in the clayey sands than the
weathered limestone. The vacuum influence in
the clayey sands extended approximately 6 m
where about 1% of the wellhead vacuum was
observed. The vacuum influence in the
weathered limestone due to the deep extraction
wells extended about 30 m in the northwest
direction (where about 3% of the wellhead
vacuum was measured) and more than 30 m in
the southeast direction (where about 17% of the
wellhead vacuum was observed).
When a vacuum was applied at Well VE-1 in
the clayey sands, an induced vacuum was also
observed in the weathered limestone (where
about 8% of the wellhead vacuum at VE-1 was
observed) more than 60 meters away in .the
southeast direction. This influence between the
two geologic formations confirms the
communication between the two formations.
The contaminant extraction rates from each
well, as well as the combined gas flow,
decreased over the 25-d study period for each
BTEX component and total hydrocarbons. The
hydrocarbons extraction rate from all the
extraction wells was about 6.0 kg/d at the
startup of the field studjr and quickly decreased
within days, to an average extraction rate of
about 3.0 kg/d.
The deep extraction wells located in the
limestone extracted a greater fraction of total
hydrocarbons as compared with the total BTEX
components than the shallow extraction wells
43
-------
located in the clayey sands. The deep extraction
wells extracted a total mass ratio of
hydrocarbons to BTEX of about 17 to 1, whereas
the shallow extraction wells extracted a total
mass ratio of hydrocarbons to BTEX of about 6
tol.
13. The fraction of BTEX concentrations compared
to the total hydrocarbon concentrations in the
clayey sands more closely matches the relative
BTEX fraction in pure gasoline than those
observed in the weathered limestone. The
residual BTEX in the weathered limestone
seems to contain a disproportional amount of
xylenes.
14. The concentration ratio between each BTEX
component varied systematically over the time
period of the EPA field investigation of the
vacuum extraction system operation. Benzene
and toluene, the more volatile components of
the BTEX fraction, decreased in the extracted
gas concentration at a faster rate than the less
volatile components, ethylbenzene and xylenes,
from several of the shallow extraction wells.
This indicates that the relative amounts of the
more volatile components are being removed
from the soil matrix to a greater degree than
the less volatile components.
Cost Evaluation
The cost figures reported by CDM include capital
costs of $106,100 and operating and maintenance
costs of $68,000, for a total of $174,100.
No attempt has been made by CDM or any other
contractor to determine the amount of soil that was
contaminated by this leaking gasoline underground-
storage tank. However, an order of magnitude
estimate of the amount of contaminated soils at the
Bellview site could be between 3,000 and 10,000 yd3.
Based on these estimates, the total cost, i.e., capital
and operating costs, to remediate these soils in a one-
year period would be in the range of $20 to $60/yd3.
However, these costs depend on the required cleanup
goals of the site and will increase if longer treatment
periods are required to further reduce the
contamination level.
The above costs do not include extraction air
treatment as this was not required at the Bellview
site. However, if air treatment was required and it
was assumed that a carbon adsorption system would
be used, the costs of a carbon adsorption system
would more than double the total vacuum extraction
system costs. Actual costs of the carbon adsorption
system would depend on the carbon usage rates of
the extracted air flow.
Bibliography
1. Applegate, J., J.K. Gentry, and J.J. Malot.
Vacuum Extraction of Hydrocarbons from
Subsurface Soils at a Gasoline Contamination
Site. Volatile Organics Monitoring Remedi-
ation, pp. 273-279,1987.
2. Camp, Dresser, and McKee, Inc. Assessment of
Vacuum Extraction Technology Application:
Bellview, Florida, LUST Site. Prepared for U.S.
Environmental Protection Agency, Edison, N.J.
Contract 68-03-3409. Prepared by Camp,
Dresser, and McKee, Inc., One Center Plaza,
Boston, Mass. 02108 (preliminary draft). 1988.
3. Hutzler, N.J., B.E. Murphy, and J.S. Gierke.
State of Technology Review - Soil Vapor
Extraction Systems. Report No. CR-814318-01-
1. U.S. Environmental Protection Agency,
Hazardous Waste Engineering Research
Laboratory, Cincinnati, Ohio 45268,1988.
4. Terra Vac Corporation. Union 76 Gas Station
Clean-Up, Bellview, Florida. Tampa, Fla. 1987.
5. Terra Vac Corp. Initial Report - Vacuum
Extraction System Construction, Startup, and
Operations. October 7, 1988.
6. Terra Vac Corp. Union 76 Gas Station Cleanup,
Bellview, Florida. Prepared for the Florida
Department of Environmental Regulation,
1987.
44
-------
Case Study D-3
Verona Well Field (Michigan) Superfund Site Soil Remediation Using
Large Scale Vacuum Extraction
Introduction
As part of the USEPA's total site remediation
strategy, a soil vacuum extraction system (VES) was
selected to remediate soils at the Verona Well
Field/Thomas Solvents Raymond Road Facility
Superfund site. In 1987 Terra Vac Corp. was
contracted to design, install, and operate the VES in
conjunction with other already ongoing work at the
site, which included the recovery and treatment of
groundwater.
The well field is located in Battle Creek, Michigan
and supplies potable water to over 35,000 consumers.
In 1981 VOCs were discovered to have contaminated
both private and city wells, and the well field was
placed on the National Priorities List during 1982.
Subsequent USEPA investigations revealed a former
solvent storage and transfer facility was one of the
sources of the groundwater contamination. The site
included twenty-one underground storage tanks,
some of which had previously leaked chlorinated and
nonchlorinated solvents.
A Record of Decision was signed in 1985 specifying
soil vacuum extraction as the Remedial Measure for
treatment of soils at the former solvent facility, to
eliminate one source of groundwater contamination.
The underground storage tanks were scheduled for
removal after completion of soil treatment by
vacuum extraction. Design and construction of the
VES began in September 1987. Construction was
completed in March 1988, with full-scale operations
commencing thereafter.
This project is ongoing. The following is a summary
of the progress made thus far and preliminary
conclusions.
Vacuum Extraction System Design
The major system design criteria are the following:
Hydrogeologic Considerations-
The geology at this site is composed of
unconsolidated material derived from glacial
outwash and floodwater channel deposits, overlying
the Marshall Formation, a sandstone bedrock. The
soils at the site consist of fine-to-coarse grained sand
with localized lenses of gravel and silt. Groundwater
fluctuates between 20 and 30 ft below the surface
and a localized cone of depression is present due to
groundwater extraction wells both on and off site.
Nature of Contamination-
Previous investigations indicated the presence of
VOCs, mainly chlorinated hydrocarbons, aromatics,
and ketones. Soil concentrations as high as 1,800
mg/kg of specific contaminants were reported.
Contaminants included TCE, PCE, TCA, methylene
chloride, xylenes, DCA, acetone, toluene, and
ethylbenzene. Contamination was indicated
throughout the unsaturated zone, with the
possibility of nonaqueous phase liquid (NAPL). An
area of approximately 35,000 ft3 as addressed by the
VES design.
Cleanup Criteria-
Achievement of cleanup criteria will be verified by
posttreatment soil sampling and analysis. The
specified.cleanup criteria requires all soil samples to
be less than 10 mg/kg total VOCs, with no more than
15% above 1 mg/kg total VOCs.
VES Emissions Controls-
Limits placed on a number of VOCs present required
the design and operation of an activated carbon
system. Allowable concentration at the VES
discharge stack were:
tetrachloroethylene
trichloroethylene
methylene chloride
chloroform
carbon tetrachloride
vinyl chloride
benzene
0.0024 mg/L
0.0073 mg/L
0.0406 mg/L
0.0008 mg/L
0.0016 mg/L
0.0162 mg/L
0.0057 mg/L
Underground Tank Impacts-
Twenty-one underground tanks at the site are not
scheduled for removal until soil treatment is
complete. The impact of the tanks on the subsurface
45
-------
flow regime, and on subsurface vacuum levels was
evaluated, and is reflected in the VES design.
Vacuum Extraction System
Implementation
A two-phase approach was used for the
implementation of the VES design. The first phase
was a pilot phase during which the preliminary
design was confirmed. The second phase of
implementation, full-scale system design and
construction, started immediately after data from
the pilot phase were evaluated.
Four vacuum extraction wells were installed in
October of 1987 to serve as a pilot phase for the soil
remediation. During the installation of the vacuum
extraction wells, soils were sampled and analyzed
using an onsite gas chromatograph. and the
headspace method to quantify the distribution of
VOCs in soils in the area of the pilot wells. Residual
VOC concentrations in the area of the four pilot test
wells ranged as high as 1,380 mg/kg. The soil
concentration data was used to confirm the design
basis of the vacuum extraction system, to determine
screened intervals, and was later used to correlate
monitoring results with soil concentrations. In
addition, a soil gas survey was performed throughout
the site, including support areas, to further delineate
the real extent of soils to be treated by the VES.
Each well consisted of a 4-in. PVC slotted-well screen
and riser, a silica sand pack in the annular space,
and bentonite and grout seals to make the wells
suitable for vacuum service. Wells later installed as
part of the full-scale system included both 4-in. wells
with slotted screen and 2-in. continuous wire
wrapped screens. An aboveground PVC piping
manifold was connected to the equipment used
during the pilot phase, which consisted of an
air/water separator, a 30-hp vacuum extraction unit,
and an emissions control system containing four
1,000-lb canisters of vapor-phase activated-carbon.
(Two in primary service, and two in backup service.)
A 30-ft-high discharge stack was constructed.
After startup of the pilot phase VES, each well was
developed individually. In general, during the
development period of a vacuum extraction well, the
soils surrounding the well are dried, air flow paths
are developed, a well's maximum radius of influence
is reached, and steady state flows are established.
For the wells at this site, the development period for
each well was very short, on the order of 1 to 4 hours.
During all phases of operation of the VES, individual
wellhead VOC concentrations and other vapor
stream concentrations throughout the VES were
determined by onsite gas chromatography. Air flow
rates were measured using self-averaging pitot
tubes, or rotameters, depending on flow rates. The
radius of influence for each well was determined by
measuring subsurface vacuums using piezometers
and other VES wells at different distances from the
vacuum extraction well being developed.
The radius of influence for the pilot phase VES wells
was determined to be in excess of 75 ft. As an
example, a subsurface vacuum of 1/4 in. of water was
measured 95 ft away from VE-1, one of the first wells
to be developed, shortly after startup.
After intermittent operation of the pilot phase VES
wells for approximately 70 h over a period of 15 d,
construction of the full scale system was started. The
full-scale VES began operations in March 1987. A
total of 23 vacuum extraction wells were installed at
the site. The location and number of wells reflected
pilot-phase experience, accounted for the effect on
subsurface airflows of the underground tanks, and
provided operating flexibility to ensure all affected
areas would be treated.
The vacuum extraction system consisted' of two
vacuum units. The vapor-phase activated-carbon
system was enlarged to 8 canisters with 4 in primary
service, and 4 in backup service. Carbon adsorption
efficiency was determined to be equivalent under
positive pressure and vacuum during the pilot phase,
therefore the full-scale system was constructed with
activated carbon under, vacuum to minimize leaks,
and to eliminate possible emissions of contaminant-
laden air to the atmosphere.
Although the need to change frequently and
regenerate activated carbon offsite during the first
10 days of operation dictated attended operation, the
VES was designed and constructed for unattended
operation. Instrumentation and controls installed
included: pressure, flow and temperature indicators;
a high-water-level shutdown in the air/water
separator, a carbon monoxide monitor and shutdown
in the activated carbon system, high temperature
shutdowns, and an online PID VOC monitor for
detecting primary-carbon system breakthrough.
Vapor samples were analyzed with the onsite gas
chromatograph at various VOC levels to determine
the PID monitor's response to specific compounds.
Whe'n VOC concentrations entering the backup
carbon system reached a predetermined setpoint on
the PID monitor, the VES system was automatically
shut down. The vacuum extraction system is shown
schematically in Figure D-4.
To date, the vacuum extraction system at this
Superfund Site has operated for approximately 55
days. Onsite gas chromatography has been used to
monitor wellhead VOC concentrations and
extraction rates. The logistics of changing,
transporting, and regenerating activated carbon
46
-------
Discharge
to
Atmosphere
I -_o I
voc-
Contaminated (P I) (Tl) ( S
Air
Legend
(FT) Pressure Indicator
(rj) Temperature Indicator
(s) Sampling Port
C F) Flowmeter
Vacuum Extraction
Unit
Figure D-4. Schematric of VES at Verona, Ml.
offsite have been the limiting factor for VES
operations.
Field Analytical Program
An onsite laboratory was established to provide a
rapid screening of both soil and extracted vapor
samples. A Hewlett-Packard 5890A gas
chromatograph with dual flame ionization detectors
and capillary columns was coupled with a personal
computer and chromatography software. Twenty-two
compounds were identified as substances of concern
and methodology was developed to analyze for these
compounds. Chlorinated hydrocarbons (including
TCA, TCE, PCE, and vinyl chloride), aromatics,
(including benzene, toluene, and xylenes) and
ketones (acetone, MEK, and MIBK) were the three
major groups of compounds analyzed for.
The minimum detection limits (MDLs). for each
compound of interest was determined. The MDLs
ranged from 0.0001 to 0.0080 mg/L and normally
showed more sensitivity for aromatic compounds
than chlorinated compounds. Calibration was
accomplished by injection of a certified standard gas.
A QA/QC program was implemented using
standards, replicates, duplicates, and blanks.
During well installation, more than 200 soil sainples
were screened by the onsite gas chromatograph,
providing field data to confirm VES well design. The
onsite laboratory routinely analyzed vapor samples
for the purposes of tracking VOC extraction rates,
verifying activated carbon breakthrough,
quantifying stack VOC discharge rates and
monitoring the progress of soil treatment.
Vacuum Extraction System Performance
During the operation of the vacuum extraction
system, extracted airflows from individual VES
wells have ranged from 60 to 165 cfm, with wellhead
vacuums ranging from 2 to 5 in. of mercury.
Individual well extraction rates were determined for
all VES wells using measured flow rates and VOC
concentrations determined by onsite gas
chromatography. A total of VES extraction rate was
also routinely determined. Total VES extraction
47
-------
rates were confirmed by offsite analysis of spent
carbon.
Individual initial VES well extraction-rates ranged
from 4,400 Ib/d to 23 Ib/d. The highest individual
well extraction rate, 4,400 Ib/d, was measured during
the pilot phase. TCA was extracted at a rate of 1,316
Ib/d, with TCE, PCE, toluene, methylene chloride,
and xylenes all being extracted at rates in excess of
100 Ib/d. Figure D-5 shows the relative VOC
extraction rates in Ib/d of one well, VE-2.
Individual wellhead VOC concentrations declined
during the operation of the VES. Since the
concentration of VOCs in the airstream extracted
from a VES well is representative of the aggregate
soil-gas concentration within a well's radius of
influence, the wellhead concentrations provide an
indication of the degree of cleanup being achieved.
Well VE-2 had the highest initial wellhead
concentrations, which were measured to be in excess
of 250 mg/L total VOCs during well development.
After approximately 55 days of VES operations,
wellhead concentrations decreased to below 10 mg/L.
At other sites where vacuum extraction has been
applied, the wellhead concentration vs. time data
follows a characteristic curve. Preliminary
evaluation of the data from this Superfund Site
indicates that soil cleanup objectives will be attained
in approximately 100 days of VES operation.
To date, more than 28,000 Ib of VOCs have been
extracted by the vacuum extraction system,
representing approximately 55 days of operating
time for the VES, as shown in Figure D-6. The total
amount of VOCs extracted is based on monitoring of
the system using onsite gas chromatography. Offsite
analysis of spent carbon confirms that the onsite
monitoring is accurate to within approximately 5%.
Conclusions
Since this project is still in progress, very specific and
definitive conclusions cannot yet be drawn. However,
based on the evaluation of operating data from the
application of a large-scale vacuum extraction
system to the VOC contaminated soils at this
Superfund Site, vacuum extraction has been
successful in significantly reducing VOG
concentrations in the soil. Although factors not
associated with performance of the VES have
resulted in approximately one year of activity at the
site, the short VES operating timeframe
(approximately two months) has resulted in the safe
recovery of VOCs that would take many years to
recover using groundwater recovery and treatment
only.
Cost information on this project was unavailable.
References
1. CH2M-Hill, Inc., Remedial Planning/Field
Investigation Team, Verona Well Field -
Thomas Solvent Co., Battle Creek, Michigan,
Operable Unit Feasibility Study, Contract No.
68-01-6692, June 17,1985.
2. CH2M-HU1, Inc., Operable Unit Remedial
Action, Soil Vapor Extraction at Thomas
Solvents, Raymond Road Facility, Battle Creek,
Mich., Quality Assurance Project Plan,
October, 1987a.
3. CH2M-HU1, Inc., Appendix B - Sampling Plan,
Operable Unit Remedial Action; Soil Vapor
Extraction at Thomas Solvents Raymond Road
Facility, Battle Creek, Mich., October, 1987b.
4. Malmanis, E., D. Fuerst, R. Piniewski,
Superfund Site Soil Remediation Using Large
Scale Vacuum Extraction, Terra Vac Corp.,
Tampa, Fla. , 6th National RCRA Superfund
Conference and Exhibition, April 1989.
5. Hutzler, N.J., B.E. Murphy, J.S. Gierke, State
of Technology Review - Soil Vapor Extraction
Systems, Report No. CR-814318-01-1,
Hazardous Waste Engineering Research
Laboratory, USEPA, Cincinnati, Ohio, 1988.
48
-------
to
o
a>
< <
O m
O co
Pounds VOCs Removed
01 o
o o
o o
o o o
-» to ro u
01 o ai o
o o o o
Sooo
o o o
-it
(B
9
en
I
st
I
I
3
s
o
i
(D
Extraction Rate (Ib/d)
c?
I
CD
CD
O
CO
A
s
3
3
-------
-------
Case Study D-4
Terra Vac Vacuum Extraction Technology Application at Tyson's
Superfund Site Upper Merion Township, Pennsylvania
Introduction
Tyson's site is an abandoned septic waste and
chemical waste disposal site within a sandstone
quarry that was operated from 1962 to 1970. The
site, located in Upper Merion Township,
Montgomery County, Pennsylvania, is
approximately four acres and consists of unlined
lagoons. It is bordered on the east and west by
tributaries of the Schuylkill River, on the south by a
steep quarry highwall, and on the north by a Cohrail
railroad switching yard. The group of PRPs for the
site is headed up by the Ciba-Geigy Corporation.
Site Contamination
The soils in the lagoons are contaminated with
volatile organic compounds and with lesser levels of
contamination of semi-volatile compounds. Table D-
2 shows the analysis of subsurface soil samples from
the lagoons collected by Baker/TSA for the onsite
remedial investigation (RI) in 1984.
Chronology of Events
From 1960 to 1970 the site was owned by companies
owned by Franklin P. Tyson and by Fast Pollution
Treatment, Inc. The current owner of the land,
General Devices, Inc., had owned the stock in Fast
Pollution Treatment, Inc. along with Franklin P.
Tyson. In 1969 General Devices purchased the
property from Fast. The Pennsylvania Department
of Environmental Regulation (DER) ordered the site
closed in 1973 and the lagoons were emptied of
standing water, backfilled, vegetated, and the
standing waste was transported offsite. In 1983, the
EPA Emergency Response Team conducted a
preliminary investigation of the site in response to
an anonymous citizen complaint. In March 1983
EPA instituted immediate removal actions,
including a leachate collection system, an air-
stripping leachate-treatment system, an activated-
carbon offgas treatment-system for the air stripper,
and installation of monitoring wells.
Table D-2. Maximum and Frequency Detected for Organic
Compounds Analyzed in Subsurface Soil
Samples from the Former Lagoons, Collected
by Baker/TSA (Onsite RI) (mg/kg, dry weight
basis)
Compound
Maximum
Frequency
detected
Semivolatiles
2-Chlorophenol 7
2,4-Dimethylphenol 7
Phenol 240
2-Methylphenol 22
4-Methylphenol 18
l,2;4-Trichlorobenzene 210
2-Chloronaphthalene 0.57
1,2-Dichlorobenzene 140
1,4-Dichlorobenzene 1.1
Naphthalene 0.92
Nitrobenzene 23
N -Nitrosodiphenylamine 1.3
Bis(2-ethylhexyl)phthalate 14
Di-n-butyl Phthalate 130
Di-n-octyl Phthalate 11
2-Methylnaphthalene 12
Volatiles
1,2,3-Trichloropropane 25,000*
Ethylbenzene 13,000
Tetrachloroethene 13,000
Toluene 8,800
Trichloroethene 0.82
o-Xylene 27,000
Chlorobenzene 6.4
2
5
6
6
6
14
1
14
3
2
,4
2
4
14
1
1
23
23
23
23
23
23
• 23
23
23
23
23
23
23
23
23
23
18
18
4
12
1
20
1
23
23
23
23
23
23
23
Tentatively identified compound, concentration estimated.
Various studies were initiated, including an RI/FS
by Michael Baker, Jr., Inc. under subcontract to NUS
Corp. in 1984.
In January of 1985 EPA issued a ROD for the site
using excavation and removal of surficial soils to a
RCRA landfill. EPA however did not study the deep
aquifer and as a result, was not able to identify the
principal pathway of contaminant migration. A
subsequent comprehensive Feasibility Study was
prepared for Ciba-Geigy by ERM, Inc. and as a result
of this study and pilot demonstrations done by Terra
51
-------
Vac in the field in 1986 and 1987, Terra Vac was
awarded the contract for the complete remediation of
the Tyson's Superfund site.
Terra Vac's Pilot Demonstration
Terra Vac installed 4 vacuum-extraction wells into
the surficial materials to the top of the bedrock
surface in the east lagoon area. A fifth area (SB-2)
was installed in a bedrock area, 40 ft north of VE-1.
Bedrock is encountered 11 ft below the surface at this
location. The ground water table at SB2 was 28 ft
below the surface. Vapors were extracted from the
bedrock formation between the top of rock at 11 ft
and the water table at 28 ft below the surface. The
results and conclusions of Terra Vac's 21-day pilot
test were:
1. Plowrates from vacuum extraction wells
increased significantly as the wells were
developed with the vacuum extraction process,
up to 165 cfm from an individual well.
2. The radius of influence of the vacuum extraction
wells was developed to about 40 ft in the subsoils
of the east lagoon area.
3. VOCs detected in extracted vapors include
benzene, toluene, xylenes, TCE, PCE, and TCP,
at rates up to 154 Ib/d.
4. About 1,330 Ib of VOCs were extracted from the
subsoils and treated with activated carbon
during the pilot test.
5. The highest recovery rates were observed for
xylenes, toluene, and TCP.
6. Vacuum extraction of VOCs from the bedrock
formation beneath the contaminated subsoils
successfully recovered about 15 Ib/d and induced
a radius of influence of about 100 ft within the
bedrock and into the subsoils of the lagoon area.
Terra Vac's Full-Scale Remediation of
Tyson's
Terra Vac has installed and has been operating,
since November 1988, a full-scale remediation
facility for the Tyson's Superfund site. Almost 100
extraction wells have been installed, consisting of 81
shallow-soil wells, 9 deep-rock wells, and 7 shallow-
rock wells. The system can handle an air flow rate of
9,000 scfm in two parallel branches. There are 4
primary activated-carbon canisters and 2 secondary
or guard canisters. The spent activated carbon is
regenerated onsite by steam in a batchwise
operation.
The vacuum system has a total of 4 blowers, 2 with
700-hp motors and 2 with 250-hp motors. The
activated carbon canisters are on the discharge side
of the blowers. Water treating facilities are located
on the site and consist of an, air stripper with an
activated carbon canister for the offgas. Laboratory
analyses are performed in a permanent onsite
laboratory. The system is operated continuously and
is manned by 3 people per shift, 24 h/d. A total of 20
people are working at the site during the daytime
hours performing various monitoring functions.
Cost information on this project was unavailable.
52
-------
-------
Agency
Cincinnati OH 45268
3 AID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use, $300
Please make all necessary changes on the above label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish to receive these reports CHECK HERE n;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
EPA/540/A5-89/003
------- |