United States
Environmental Protection
Agency
Office of Research
and Development
Washington, D.C. 20460
EPA/540/A5-90/002
August 1990
oEPA
CF Systems Organics
Extraction Process
New Bedford Harbor, MA
Applications Analysis Report
SUPERFUND INNOVATIVE
TECHNOLOGY EVALUATION
-------
EPA/540/A5-90/002
August 1990
CF Systems Organics Extraction Process
New Bedford Harbor, Massachusetts
Applications Analysis Report
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
Notice
The information in this document has been funded by the U.S. Environmental
Protection Agency under Contract No. 68-03-3485 and the Supcrfund Innovative
Technology Evaluation (SITE) Program. It has been subjected to the Agency's peer
review and administrative review and it has been approved for publication as a
USEPA document. Mention of trade names or commercial products does not
constitute an endorsement or recommendation for use.
-------
Foreword
The 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 development of hazardous waste treatment technologies necessary to
implement new cleanup standards that require greater reliance on permanent
remedies. This is accomplished through technology demonstrations that are
designed to provide engineering and cost data on selected technologies.
This project consists of an analysis of CF Systems' proprietary organics
extraction process. The technology demonstration took place at the New Bedford
Harbor Superfund site, where harbor sediments are contaminated with polychlori-
nated biphenyls and other toxics. The demonstration effort was directed at
obtaining information on the performance and cost of the process for use in
assessments at other sites. Documentation will consist of two reports. A Technol-
ogy Evaluation Report described the field activities and laboratory results. The
Applications Analysis provides an interpretation of the data and conclusions on the
results and potential applicability of the technology including a projection of costs
from the demonstrated pilot unit to a full-scale commercial unit.
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 on 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, (703) 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, DC, to inquire about the availability
of other reports.
garet M. Kelly, Director / AlfreoW. Lindsey, Acting Director
Technology S taff, Office Office of Environmental
of Program Management Engineering and Technology
and Technology OSWER Demonstration
in
-------
Abstract
The SITE Program Demonstration of the CF Systems organics extraction
technology was conducted to obtain specific operating and cost information that
could be used in evaluating the potential applicability of the technology to
Superfund sites. The demonstration was conducted concurrently with dredging
studies managed by the U.S. Army Corps of Engineers at the New Bedford Harbor
Superfund site in Massachusetts. Contaminated sediments were treated by CF
Systems' Pit Cleanup Unit (PCU) that used a liquefied propane and butane mixture
as the extraction solvent. The PCU was a trailer-mounted system with a design
capacity of 1.5 gallons per minute (gpm), or 20 barrels per day (bbl/day). The
technology extracts organics from contaminated soils based on solubility of
organics in a mixture of liquefied propane and butane.
The objectives included an evaluation of (1) the unit's performance, (2) system
operating conditions, (3) health and safety considerations, (4) equipment and
system materials handling problems, and (5) projected system economics. The
conclusions drawn from the test results and other available data are:
• Polychlorinated biphenyl (PCB) extraction efficiencies of 90 percent
were achieved for New Bedford Harbor sediments containing PCBs
-ranging from 3 50 to 2,575 parts per million (ppm). Concentrations
of PCBs in the clean sediment were as low as 8 ppm.
• Extractionefficienciesof95percentaredemonstratedin the laboratory
for volatile and semivolatile organics contained in aqueous and
semisolid waste matrices.
Some operating problems occurred during the SITE tests, such as
intermittant retention of solids in system hardware and foaming in the
treated sediment collection tanks. Corrective measures were
identified, and will be incorporated in the full-scale commercial
unit.
Operation of the PCU at New Bedford did not present any threats to
the health and safety of operators or the local community.
• The projected cost of applying the technology to a full-scale cleanup
at New Bedford Harbor ranges from $148 to $447 per ton. These
projections include pre- and post-treatment costs, material handling
costs, and costs for a specialized process configuration designed to
remediate sediments, however the post-treatment cost did not include
the final destruction of the concentrated extract.
• Site specific pre- and post-treatment costs account for approximately
one-third of the estimated costs.
• The predicted onsiream factor for the full-scale commercial unit is
the variable that introduces the greatest uncertainty to the cost
estimates.
IV
-------
Contents
Page
Foreword iii
Abstract iv
Figures vi
Tables vi
Acknowledgments vii
Abbreviations and Symbols ix
1. EXECUTIVE SUMMARY 1
1.1 SUMMARY 1
1.2 CONCLUSIONS 1
1.3 APPLICATIONS ANALYSIS 2
1.4 RESULTS 3
2. INTRODUCTION 7
2.1 THE SITE PROGRAM 7
2.2 SITE PROGRAM REPORTS 7
2.3 KEY CONTACTS 8
3. TECHNOLOGY APPLICATIONS ANALYSIS 9
3.1 OVERALL TECHNOLOGY APPROACH 9
3.2 TECHNOLOGY EVALUATION 10
3.3 WASTE CHARACTERISTICS AND OPERATING REQUIREMENTS 14
3.4 MATERIAL HANDLING REQUIREMENTS 17
3.5 HEALTH AND SAFETY ISSUES 18
3.6 TESTING PROCEDURES 18
4. ECONOMIC ANALYSIS 21
4.1 INTRODUCTION 21
4.2 BASIS FOR PROCESS DESIGN, SIZING, AND COSTING 21
4.3 DEVELOPER'S ESTIMATEFOR A NEW BEDFORD HARBOR CLEANUP 24
4.4 EVALUATION OF THE DEVELOPER'S ESTIMATE 27
4.5 EXTRAPOLATION OF CF SYSTEMS' SLUDGE TREATMENT COSTS
TO OTHER SITES 29
4.6 CONCLUSIONS AND RECOMMENDATIONS 29
APPENDICES
A. PROCESS DESCRIPTION 31
B. DEVELOPER (VENDOR) COMMENTS 45
C. SITE DEMONSTRATION RESULTS 55
-------
Contents (Continued)
Figures
Number Page
3-1 New Bedford Harbor Application How Diagram 19
Tables
Number Page
1-1 CF Systems'Soils Treatment Extraction Unit Designs 3
3-1 Bench-Scale Test Results for Wastewaters and Groundwaters 11
3-2 Bench-Scale Test Results for Sludges and Soils 12
3-3 Texaco, Port Arthur Performance Data 15
3-4 United Creosote Superfund Site Performance Data 16
3-5 Sludge and Soil Feed Requirements 16
4-1 Base Case and Hot Spot Case Summary 26
4-2 Estimated Cost .....". 28
VI
-------
Acknowledgments
This report was prepared under the direction and coordination of Richard
Valentinetti, EPA SITE Program Manager in the Office of Environmental Engi-
neering & Technology Development in Washington, D.C. Contributors and re-
viewers for this report were Frank Ciavattieri of EPA Region I, Remedial Project
Manager for the New Bedford Harbor Superfund Site; Jim Cummings from the
Office of Solid Waste and Emergency Response; Paul de Percin, Gordon Evans,
Diana Guzman, and Laurel Staley from the Office of Research and Development;
Christopher Shallice and Thomas Cody, Jr. from CF Systems Corporation; and
Alan Fowler of EBASCO Services, Inc.
This report was prepared for EPA's SITE Program by Science Applications
International Corporation (SAIC), McLean, VA, for the U.S. Environmental
Protection Agency under Contract No. 68-03-3485, by Don Davidson, John
Bonacci, Richard Hergenroeder, and Omer Kitaplioglu. Laboratory analyses were
conducted by E.G. Jordan, Inc., Portland, ME, and Radian Corp, Austin, TX.
Vll
-------
Abbreviations and Symbols
amps amperes
ASTM American Society for
Testing and Materials
bbl/day barrels per day
BDAT best demonstrated
available technology
BNAs base/neutral and acid
extractable compounds
Cd cadmium
COE U.S. Army Corps of Engineers
cP centipoise
Cr chromium
CR column reboiler
Cu copper
CWA Clean Water Act
dPa.s decapascal.seconds
ECD electron capture detector
EPA Environmental Protection Agency
EPT extract product tank
EP Tox Extraction Procedure Toxicity
Test - leach test
F Fahrenheit
FK feed kettle
g grams
GC gas chromotography
gpd gallons per day
gpm gallons per minute
kw-hr kilowatt hours
Ibs pounds
Ib/gal pounds per gallon
Ib/min pounds per minute
max maximum
MBAS methylene blue active substances
mg milligrams
mg/kg milligrams per kilogram
min minimum
ms mass spectrometry
MSA method of standard additions
MS/MSD matrix spike/matrix spike duplicate
ND not detected
NIOSH National Institute of Occupational
Safety and Health
NR not reported
ORD Office of Research and Development
OSWER Office of Solid Waste and
Emergency Response
OVA organic vapor analyzer
oz ounces
PAHs polyaromatic hydrocarbons
Pb lead
PCBs polychlorinated biphenyls
PCU Pit Cleanup Unit
PNAs polynuclear aromatics
ppm parts per million
psig pounds per square inch gauge
QA quality assurance
QC quality control
RCRA Resource Conservation and
Recovery Act of 1976
RPD relative percent difference
RREL Risk Reduction Engineering
Laboratory
RSD relative standard deviation
SARA Superfund Amendments and
Reaulhorization Act of 1986
SET still bottoms tank
SITE Superfund Innovative
Technology Evaluation Program
SRC solvent recovery column
TDS total dissolved solids
TS total solids
TSD treatment, storage, and disposal
TSS total suspended solids
VAC volts, alternating current
VOAs volatile organic analytes
Zn zinc
< less than
IX
-------
Section 1
Executive Summary
1.1 Summary
The CF Systems Corporation pilot-scale soil treat-
ment technology was tested and evaluated under EPA's
Superfund Innovative Technology Evaluation (SITE)
Program. The technology uses a mixture of liquefied
propane and butane as a solvent to extract organics from
harbor sediments. Successful application of the technol-
ogy depended on the ability of the organic pollutants to be
solubilized by the solvent. Mixing solvent with waste
seeks to achieve intimate contact between the solvent and
the contaminants. Variables include solvent-to-feed ra-
tios, mixing energy, and residence time in the reactor.
Following decanting of the solvent-organics mixture from
the solids and water, pressure reduction vaporizes the
solvent and separates it from the organics. The solvent is
recovered and compressed to a liquid for reuse. The
separated organics are collected for disposal. Treated soil
may requiredewatering. Soil that meets cleanup standards
can be returned to the site. Water that meets applicable
standards can be discharged directly to a stream or to a
Publicly Owned Treatment Works.
Solvent extraction technology relies upon the prefer-
ential solubility of organics in certain solvents versus the
soil and water in which the contaminants are found in
environmental matrices. Application of sol vent extraction
at Superfund sites is essentially a pretreatment step, result-
ing in significant reductions in the amount of material that
must be subjected to further treatment, e.g., incineration.
Thus, in soils contaminated with oil and grease at 1,000
ppm (0.1 percent), the amount of material requiring incin-
eration would be reduced by a factor of 1,000 (assuming a
removal efficiency of 99 percent).
Removal efficiency depends on a number of factors
including the ability of the technology to bring the solvent
into proximity with the contaminant(s) and the degree to
which the contaminants prefer the solvent to the medium in
which they are located.
CF Systems also markets a wastewater treatment
system that uses liquefied carbon dioxide as the solvent.
This system was not tested under the SITE Program.
1.2 Conclusions
The conclusions drawn from reviewing limited opera-
tional data on the CF Systems technology, both from the
SITE evaluation tests and from the information supplied
by the developer, are:
• The soils treatment system was tested on sedi-
ments obtained from New Bedford Harbor in
Massachusetts that contained PCBs at 350 and
2,575 ppm concentration levels. A pilot-scale
mobile unit was used for this test. This unit
required recycle of product to simulate the
operation of a full-scale, four-stage unit. This
mode of operation caused material handling
problems which in turn restricted process through-
put. The multiple-pass mode of the demon-
stration limits our ability to extrapolate to full-
scale units intended to be operated in a once-
through mode.
• The technology can separate organics from har-
bor sediments, sludges, and soils. PCB extrac-
tion efficiencies greater than 90 percent were
achieved for New Bedford Harbor sediments and
attained levels as low as 8 ppm for PCBs. CF
Systems' pilot-scale unit has also been success
fully demonstrated at petroleum refineries, pet-
rochemical plants, and hazardous waste treat-
ment storage and disposal facilities.
• The technology can also separate organics from
wastewater; however, the mixing equipment
and solvent used are different from that used for
sludges and soils. Although the SITE program
tests were conducted only on the soils treating
unit, some information is presented in this report
on the wastewater treatment unit.
• Operational control was difficult to maintain
during the New Bedford Harbor tests. Solvent
flow fluctuated widely and caused the solvent-to-
feed ratio to fall below specifications. Solids
were retained in process hardware, solids were
observed in organic extracts, and foaming of
treated sediments also occurred. The vendor
-------
believes that these problems are correctable by
equipment design changes and by operating in
a once-through mode instead of a recycle mode.
Bench-scale test results show the potential for
extraction of a broad range of organics from
wastewater, ground water, and semi solids. These
tests were useful for determining whether or not
organics contained in a waste matrix will be ex-
tracted by a liquefied solvent such as carbon
dioxide or propane. Laboratory results indicated
that equilibrium conditions did not limit reduc-
tion of solid PCB content to levels of 10 ppm.
Pretreatment technology may be necessary to
condition feed materials. Coarse solids removal
may be required to maintain feed sediment par-
ticle sizes below three-sixteenths inch; water
must be added to viscous sludges or dry soils and
heat must be supplied to feeds less than 60
degrees Fahrenheit. Post-treatment technology
also may be necessary such as thermal destruc-
tion of the concentrated extract. In some cases,
the cleaned material could be subjected to fur-
ther treatment. -
Water addition during the SITE Demonstration
to achieve required viscosities increased the
mass of waste by about 33 percent. Such water
addition may result in a requirement for post-
treatment dewatering.
Costs for implementing the CF Systems' tech-
nology at New Bedford Harbor were projected
based on an economic model. Since a full-scale
unit has been placed in the field only recently, op-
erating and cost data for a full-scale system were
not available. The estimated cost for removing
90 percent of the PCBs from New Bedford Har-
bor sediments containing 580 ppm is S148 per
ton, which includespre-and post-treatmentcosts.
The cost for removing 99.9 percent of the PCBs
from New Bedford Harbor "hot spots" contain-
ing 10,000 ppm is $447 per ton including pre-
and post-treatment costs. These costs represent
a range of costs anticipated for full-scale appli-
cation of the technology at New Bedford Harbor.
Approximately one-third of the estimated costs
are pre- and post-treatment costs. These cost es-
timates did not include the final destruction of the
concentrated extract.
• CF Systems has designed and fabricated a 50-
ton-per-day (200 barrels/day) soils treatment
unit for Star Enterprises, Inc.facility (Texaco) in
Port Arthur, Texas, to treat API separator sludge.
Oils extracted from the sludge will be recovered
for reuse. CF Systems has agreed to allow the
SITE program to monitor the operation of this
unit to demonstrate that operability parameters
associated with materials handling and on-stream
factors are within commercial design claims.
1.3 Applications Analysis
Applications of the CF Systems organics extraction
technology depend on waste characteristics, waste vol-
ume, and degree of pollutant removal required. Waste
characteristics determine the type of solvent to be used and
the need for pre- and post-treatment. If a waste contains
organics, such as PCBs and PAHs, that are not very soluble
in water then a hydrocarbon solvent, such as propane or a
propane and butane mixture, is used. Less soluble organics
are typically sorbed to soil particles found in sludges,
therefore propane is commonly used to extract organics
from soils and sludges.
Pre- and post-treatment must be considered if feed
materials (1) contain gravel or cobbles, (2) are below 60
degrees F, or (3) are not pumpable. For wastewaters and
groundwaters that are relatively free of solids, liquefied
carbon dioxide is the preferred extraction solvent since this
solvent seeks polar materials in water, is nontoxic, and has
favorable thermodynamic properties. CF Systems ini-
tially assesses feed materials by conducting bench-scale
tests in the laboratory. If bench-scale tests are successful,
pilot-scale tests are run with either a laboratory-based pilot
scale unit or a mobile, trailer-mounted unit. Only the
propane-based unit was evaluated during the SITE tests
and is therefore the primary subject of this report.
CF Systems offers standard modular systems for
different markets and applications. For sludge and solids
treatment the capacity range is about 10 to 1,000 tons per
day per unit and liquefied propane or a butane and
propane mixture is the extraction solvent. The capacity
range for wastewater treatment is about 5 to 150 gpm and
liquefied carbon dioxide is used as the extraction solvent.
Systems of these size ranges, constructed of carbon or
stainless steel, can be modularized, shipped, and field
assembled economically. As a result of this approach,
several unit sizes have been developed and designed. The
units can be configured in parallel if high throughput
capacities are required. If high extraction efficiencies are
necessary, the units can be arranged in series.
The soils treatment unit evaluated during the SITE
tests at New Bedford Harbor was the PCU-20, which is
rated at a 5-ton-per-day nominal capacity. The unit is often
used for pilot-scale tests, but is also used for remediating
small volumes of contaminated sludges or soils. The PCU-
20 has a 1 -1/2 foot diameter, two-stage extractor that used
a mixture of propane and butane as the extraction solvent.
-------
During the SITE tests, treated sediments were recycled
through the unit to simulate the design and operation of a
full-scale, four-stage unit that has 6-foot diameter extrac-
tors.
CF Systems has proposed two types of systems for
New Bedford. The first system, the base case, applies to the
treatment of 880,000 tons (695,000 cubic yards) of harbor
sediments containing approximately 580 ppm of PCB. The
second system applies to the treatment of 63,000 tons
(50,000 cubic yards) of sediments from harbor "hot spots"
that contain approximately 10,000 ppm. Each system
differs from the PCU-20 tested at New Bedford. For the
base case, two PCU-lOOOs, each rated at 250 tons per Jay,
would be placed in parallel to accommodate the large
volume of waste to be treated. For the hot spot, four PCU-
500s each rated at 125 tons per day would be used and these
would be configured in pairs so that two parallel trains are
available with each train providing a total of 8 stages of
treatment for the contaminated sediments.
Key extraction system elements of generic, prede-
signed soils treatment units offered by CF Systems are
shown in Table 1-1. These designs apply to the remedia-
tion of soils and sludges at any site but do not include any
site-specific support facilities or pre- and post-treatment
equipment. All components of the various units can be
obtained from "off-the-shelf sources and no custom fab-
rications are required.
1.4 Results
Performance
The most extensive evaluation of the CF Systems
technology was performed as part of the SITE tests at New
Bedford. Qualitative results are also reported by CF
Systems for three field demonstrations of a pilot-scale unit
and for numerous bench-scale laboratory tests. CF Sys-
tems achieved an overall PCB concentration reduction of
over 90 percent for New Bedford Harbor sediment samples
that contained 350 ppm and 2,575 ppm during the SITE
tests. The unit generally operated within specified condi-
tions for flowrates, pressure, temperature, pH, and viscos-
ity. Deviations from operating specifications could not be
correlated to changes in extraction efficiency. No signifi-
cant releases of pollutants to the atmosphere or surround-
ing area soils occurred. Results of the demonstration tests
show that the CF Systems technology is capable of reduc-
ing the PCB content of contaminated sediment by greater
than 90 percent without a risk to operating personnel or the
surrounding community.
CF Systems reports the following field demonstration
results for its pilot-scale units:
• Texaco: A unit was run September and October
of 1987. Different feed types were run through
the system including material from a clay pit,
ditch skimmer sludge, tank bottoms, and other
miscellaneous waste streams found at the Port
Arthur refinery site. The system consistently
Table 1-1. CF Systems' Soils Treatment Extraction Unit Designs
Unit
Designation
Nominal Throughput
(Tons Per DavWn
Extractor
Diameter (feet)
Number of
Stages
Site
Preparation (2)
Cost (Dollars)
PCU-20
12
1.5
N/A
PCU-200
PCU-500
50
125
4 $350,000
4 $700.000
PCU-1000
250
6.5
$2,000,000
NOTES:
(1) 1.26 tons is equivalent to 1 cubic yard of New Bedford Harbor sediments.
(2) Site preparation costs include clearing, grading, constructing a foundation, and providing access for utilities.
Costs are applicable to any site.
-------
achieved high removals of total oil and grease to
less than 1 percent residual of the dry solids.
Levels of individual components, including
benzene, ethylbenzene, toluene, xylene, naph-
thalene, phenanthracene, pyrene, and other poly-
nuclear aromatics (PNAs), met or bettered the
existing best demonstrated available technology
(BOAT) standards. In many cases, these levels
were found to be below detection limits. Fol-
lowing the demonstration,Star(Texaco)awarded
CF Systems a contract to provide a 50-ton-per-
day commercial unit to remediate 20,000 cubic
yards of API separator sludges and ditch skim-
mer wastes. This unit was fully operational in
July 1989.
Petro-Canada: A unit was operated at the Petro-
Canada refinery in Montreal for a six-week
period. During this time, the unit successfully
processed 14 different feed types, ranging from
API separator sludges to contaminated solids.
The unit consistently achieved organic removal
levels better than existing BDAT standards.
Tricil: A unit was used to run a series of demon-
stration tests at Tricil Canada's treatment, stor-
age, and disposal (TSD) facility in the Province
of Ontario. The feeds processed included API
separator sludge, paint wastes, synthetic rubber
process waste, and coal tar wastes. The level of
volatile organics was reduced such that disposal
of the material in a local Canadian landfill was
acceptable and volumes for disposal were sig-
nificantly reduced.
BASF: A mobile treatment system was run at
the BASF Kearny, New Jersey, plant site. One
of the waste streams from this plant is an emul-
sified stream containing di-octyl phthalate (DOP),
water, and other organic materials. The system
successfully separated the emulsion into a re-
coverable DOP stream and a wastewater suitable
for discharge to the wastewater treatment facil-
ity.
Unocal: The unit completed a series of demon-
strations at Unocal's Parachute Creek, Colorado,
facility. Among the wastes successfully run
were samples of shale-oil wastes, drilling muds,
and other process and refinery wastes. High re-
covery of good-quality oil was obtained from
shale-oil wastes. Drilling mud wastes were
treated to the standards required for land dis-
posal.
United Cresote NPL Site: A field treatability
study was completed for the Texas Water Com-
mission, aSuperfund Site in Conroe,Texas. The
objective of this study was to evaluate the effec-
tiveness of solvent extraction for remediation of
soil contaminated with creosote. PAH concen-
trations in the soil obtained from the capped area
were reduced from 2,879 ppm to 122 ppm, dem-
onstrating that 95-percent reductions were pos-
sible.
CF Systems has collected bench-scale test data for a
wide range of organic pollutants contained in wastewaters,
sludges, and soils. Carbon dioxide was used as a solvent to
remove volatile and semivolatile organics from wastewa-
ters and groundwaters. Extraction efficiencies ranged
from 95 to 99.99 percent for 24 pollutants that ranged in
concentrations from 0.4 ppm to 520 ppm, as shown in
Table 3-1. Propane was used as a solvent to extract
polyaromatic hydrocarbons (PAHs) and benzene, ethylben-
zene, toluene, and xylene from refinery sludges, API
separator sludges, and contaminated soils. Extraction
efficiencies ranged from 80 to 99 percent for concentra-
tions that ranged from 0.3 ppm to 1930 ppm, as shown in
Table 3-2.
Economics
The cost of installing and operating a commercial-
scale system depends primarily on (1) waste characteris-
tics that affect the need for pre- and post -treatment, (2) the
amount of waste to be treated, (3) the degree of treatment
required, and (4) the percentage of time that the system is
actually operational. Soil pretreatment includes water ad-
dition, large solids removal, and possibly heat addition,
while post-trealment includes dewatering. The amount of
waste at a site affects equipment sizing, the total amount of
time required to clean up a site, and life-cycle costs. The
degree of treatment required affects operating costs since
longer residence times of the waste in the equipment are
needed to achieve higher pollutant removals. The percent-
age of time that the unit is fully operational can have a
significant effect on the unit treatment costs, in terms of
cost per ton.
CF Systems sized and costcd two soil treatment units
for PCB removal from New Bedford Harbor sediments.
The objective was to estimate cleanup costs at New Bedford
using the CF Systems technology and to illustrate the
design approach used to scale-up the technology fora com-
mercial application. The estimate was based on data
obtained for PCB extraction from New Bedford Harbor
sludge using the pilot-scale unit and on a commercial
design of the unit at the Texaco, Port Arthur facility. The
base case addressed a large mass of sediment (880,000
cubic yards) at a 580 ppm PCB concentration. Treatment
would be conducted over an eight-year period to produce
sediment concentrations of 50 ppm at a rate of 500 tons/day
-------
raw feed. Pre- and post-treatment are required to reduce
viscosity and raw feed solids content. Total treatment cost
is estimated at $148/ton (1989 dollars) of which one-third
is associated with pre-and post-treatment. The hot spot
corresponds to treatment of a small mass (63,000 tons) at
10,000 ppm PCB concentration over a one-year time
frame. The treated sediment concentration goal is 10 ppm
PCB. Total treatment costs are $447/ton (1989 dollars), of
which approximately one-third are pre- and post-treatment
costs.
The economic analysis addressed costs directly re-
lated to the extraction unit as well as site-specific costs.
Costs were categorized as fixed facility, extraction unit,
pre- and post-treatment, contingency, and project manage-
ment costs. CF Systems assigned an accuracy of plus or
minus 20 percent to their cost estimates. However, indus-
try experience with innovative technologies has shown
that costs could range from plus 50 percent to minus 30
percent. The uncertainly associated with the estimated
costs is believed to be low since CF Systems incorporates
"off-the-shelf equipment into their designs. CF Systems
based their estimates on a unit construction for use at a
Texaco refinery and on designs specific to New Bedford.
The greatest source of uncertainty associated with CF
Systems' cost estimates is their assumption of the percent
of time that the unit will be on-stream. CF Systems
assumed an on-stream factor of 85 percent; however, this
was not demonstrated by operating the PCU-20 at New
Bedford. CF Systems claimed that material-handling
problems associated with the operation of a pilot unit
would be minimized with a commercial unit. CF Systems
must demonstrate that an 85-percent on-stream factor is
achievable for a commercial unit. EPA intends to evaluate
the vendors' claim for the 85-percent on-stream factor by
observ ing the performance of a commercial unit at a future
date. EPA will also observe and evaluate materials han-
dling associated with the operation of a full-scale unit to
verify mitigation of the problems experienced with the
pilot unit.
CF Systems also offers a wastewater treatment unit
that differs from the soils treatment unit in the types of
solvent and equipment used. Liquefied carbon dioxide is
used instead of propane or butane. CF Systems has
delivered a wastewater treatment unit to a Clean Harbors,
Inc., facility in Baltimore. Although CF Systems reports
typical wastewater treatment costs of 5 to 15 cents per
gallon, the cost for treating wastewater at the Baltimore
facility is 15 cents per gallon, which is approximately $35
per ton. Costs associated with CF Systems' wastewater
treatment unit are lower than those associated with the soils
treatment unit for two reasons. First, the equipment used
in the design differs. Second, no pre- or post-treatment is
required since solids content and viscosity are low and
temperatures are moderate.
-------
Section 2
Introduction
2.1 The SITE Program
In 1986, the EPA's Office of Solid Waste and Emer-
gency Response (OSWER) and Office of Research and
Development (ORD) established the SITE Program to
promote the development and use of innovative technolo-
gies to clean up Superfund sites across the country. Now in
its third year, SITE is helping to provide the treatment
technologies necessary to implement new Federal and
Slate cleanup standards aimed at permanent remedies. The
SITE Program is composed of three major elements: the
Demonstration Program, the Emerging Technologies
Program, and the Measurement and Monitoring Technolo-
gies Program.
The major focus has been on the Demonstration Pro-
gram, which is designed to provide engineering and cost
data on selected technologies. EPA and developers par-
ticipating in the program share the cost of the demonstra-
tion. Developers are responsible for demonstrating their
innovative systems at chosen sites, usually Superfund
sites. EPA is responsible for sampling, analyzing, and
evaluating all testresults. The result is an assessment of the
technology's performance, reliability, and cost. This infor-
mation will be used in conjunction with other data to select
the most appropriate technologies for ihe cleanupof Super-
fund sites.
Developers of innovative technologies apply to the
Demonstration Program by responding to EPA's annual
solicitation. EPA also will accept proposals at any time
when a developer has a treatment project scheduled with
Superfund waste. To qualify for the program, a new
technology must be at the pilot or full scale and offer some
advantage over existing technologies. Mobile technolo-
gies are of particular interest to EPA.
Once EPA has accepted a proposal, EPA and the
developer work with ihe 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 thor-
oughly 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
length of time and quantity of waste needed to assess the
technology. After the completion of a technology demon-
stration, EPA prepares two reports, which are explained in
more detail below. Ultimately, the Demonstration Pro-
gram leads to an analysis of the technology's overall appli-
cability to Superfund problems.
The second principal element of the SITE Program is
the Emerging Technologies Program, which fosters the
further investigation and development of treatment tech-
nologies that are still at the laboratory scale. Successful
validation of these technologies could lead to the develop-
ment of a system ready for field demonstration. The third
component of the SITE Program, the Measurement and
Monitoring Technologies Program, provides assistance in
the development and demonstration of innovative meas-
urement technologies to better characterize Superfund
sites.
2.2 SITE Program Reports
The analysis of technologies participating in the
Demonstration Program is contained in two documents,
the Technology Evaluation Report and the Applications
Analysis Report. The Technology Evaluation Report
contains a comprehensivedescription of the demonstration
sponsored by the SITE program and its results. This report
gives a detailed description of the technology, the site and
waste used for the demonstration, sampling and analysis
during the test, and the data generated.
The purpose of the Applications Analysis Report is to
estimate the Superfund applications and costs of a technol-
ogy based on all available data. This report compiles and
summarizes the results of the SITE demonstration, the
vendor's design and test data, and other laboratory and field
applications of the technology. It discusses the advan-
tages, disadvantages, and limitations of the technology.
Costs of the technology for different applications are
estimated based on available data on pilot- and full-scale
applications. The report discusses the factors, such as site
and waste characteristics, that have a major impact on costs
and performance.
The amount of available data for the evaluation of an
innovative technology varies widely. Data may be limited
to laboratory tests on synthetic wastes, or may include
performance data on actual wastes treated at the pilot or full
scale. In addition, there are limits to conclusions regarding
-------
Superfund applications that can be drawn from a single
field demonstration. A successful field demonstration
does not necessarily ensure that a technology will be
widely applicable or fully developed to the commercial
scale. The Applications Analysis attempts to synthesize
whatever information is available and draw reasonable
conclusions. This document will be very useful to those
considering the technology for Superfund cleanups and
represents a critical step in the development and commer-
cialization of the treatment technology.
2.3 Key Contacts
For more information on the demonstration of the CF
Systems technology, please contact:
1. Regional contact concerning the New Bedford
Harbor, MA, site:
Mr. Frank Ciavattieri
Waste Division (HPLEAN1)
USEPA, Region 1
John F. Kennedy Building
Room 2203
Boston, MA 02203.
617-565-3715
2. EPA project manager concerning the SITE demon-
stration:
Laurel Staley
USEPA
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Drive
Cincinnati, OH 45268
513-569-7863
3. Vendor concerning the process:
CF Systems Corporation
Mr. Christopher Shallice, x 158
Mr. Thomas C. Cody, Jr., x 162
140 Second Avenue
Waltham,MA 02154-0100
617-890-1200
-------
Section 3
Technology Applications Analysis
3.1 Overall Technology Approach
CF Systems' organics extraction technology physi-
cally separates organic contaminants from the inorganic
components of a waste matrix. This separation and volume
reduction technology allows the organic contaminants to
be ultimately disposed in a more cost-effective manner.
For example, the cost of incinerating a large volume of oil-
laden soils can be minimized by separating the oils from the
soils, then incinerating only the small volume of oils. Any
inorganic contaminants, such as heavy metals, that remain
in the treated product may require additional treatment.
The SITE Program showed, however, that the organics
extraction process did not affect the physical or chemical
characteristics of the metals contained in the sediments.
Metals were not extracted by the solvent and remained with
the treated sediments. The presence of metals in the
sediments did not affect the extraction of organics. Fur-
thermore, the metal leaching characteristics, as determined
by the EP Tox procedure, were not affected by the process.
The extraction process is not an ultimate disposal method,
but it is a significant organics separation technique that can
make ultimate disposal more economic.
Applications of the CF Systems organics extraction
technology depend on the physical/chemical characteris-
tics of the waste, its volume, and the degree of pollutant
removal required. Waste characteristics determine the
type of solvent to be used. For example, liquefied propane,
or a mixture of propane and butane, is used to extract
organics that are not very soluble in water, such as PCBs
and PAHs. These hydrophobic organics tend to sorb to
particulate matter present in soils and sludges. CFSystems
has shown in the laboratory and in pilot-scale demonstra-
tions that propane and butane are effective extraction
solvents for removing these organics from soilsand sludges.
Carbon dioxide is used by CF Systems to extract water-
soluble organics from wastewater and groundwatcr since
carbon dioxide seeks polar materials in water, is nontoxic,
and has favorable engineering properties. Carbon dioxide
used in the process can be maintained near its thermody-
namic critical point, the operating region where the liquid
makes a phase transition to a gas. At this point, carbon
dioxide has the viscosity of a gas, mixes easily with waste,
and has the solvent properties of a liquid. Waste character-
istics also determine the nature and extent of pre- and post-
treatment that may be required. For example, dry solids
require water addition to create a pumpable slurry, and the
ultimate disposal of treated wastes with water added may
require de watering.
Waste characteristics, waste volume, and the degree
of pollutantremoval significantly affect system design. CF
Systems has designed standard modular systems for differ-
ent markets and applications. For sludge treatment units,
the capacity range is about 12 to 250 tons per day. For
wastewater treatment units, the capacity range is about 5 to
150 gallons per minute. The units are constructed of
carbon or stainless steel and can be modularized, shipped,
and field assembled. If high throughput capacities are
required, the modular units can be placed in parallel. If
high extraction efficiencies are necessary, several units can
be arranged in series. As a result of this approach, a number
of specific units have been developed and designed.
The soils treatment units are designed to process high
solids sludge feeds and contaminated soils. They contain
extractors and separators designed to facilitate the treat-
ment of oily solids typical of petroleum sludges and waste
materials found in refinery impoundments requiring reme-
diation. The systems included in this product series are:
• PCU-50: This system, designed to process a
maximum of about 12 tons per day, is a standard
product for refinery sludges regulated by EPA's
RCRA land disposal ban and sludges found at pit
bottoms, as well as oil- and PCS-contaminated
soils and silts. The system is modular and will be
designed for installation into confined spaces so
as to be readily integrated into existing opera-
tions.
• PCU-200: This system, designed to process a
maximum of about 50 tons per day, is a standard
product for refinery sludges regulated by EPA's
RCRA land disposal ban and sludges found at pit
bottoms, as well as oil- and PCB-contaminated
soils and silts. The system is mounted on two flat
bed trailers, and can be demobilized-remobil-
izcd at a new location in several days.
• PCU-500: The PCU-500 is similar to the PCU-
200 design, but with increased extractor capacity
-------
to provide for throughputs up to about 125 tons
per day. Although the cost increment over the
PCU-200 is relatively small, the PCU-500 re-
quires somewhat longer time for mobilization
and demobilization vs. transportable modules. It
is designed for remediation of fixed base use
where site relocation is infrequent.
• PCU-1000: This system, with a 250-ton-per-day
nominal capacity, is intended for large remedia-
tion jobs where onsite time is projected to be one
year or more at a single location. Modular and
transportable, but with multiple modules, this
system requires several weeks for mobilization
and demobilization.
The LL series is designed for the extraction of dis-
solved or emulsified organics in water streams. Solids are
usually not present at a significant level in these streams.
If present, solids must be reduced to the 2 to 3 percent level
by pretreatment. Organics content of the feed can range as
high as 30 to 50 percent and removal efficiencies can
exceed 99.9percent. Applications for the LL series include
a wide range of organic was_tewaters.
3.2 Technology Evaluation
The most extensive evaluation of the CF Systems
technology was conducted for a soils treatment unit as part
of the SITE tests at New Bedford. Qualitative evaluations
are also available for similar units tested by CF S ystems at
other locations. CF Systems has reported the results from
extensive bench-scale tests conducted with either propane
or carbon dioxide used as the extraction solvent.
CF Systems initially assesses a clients' waste by con-
ducting bench-scale tests in the laboratory to determine if
the organic constituents will solubilize in the liquefied
solvent. CF Systems is also able to use rules-of- thumb to
roughly estimate the number of processing stages that
might be required to achieve a desired extraction effi-
ciency. In the laboratory, the waste can be observed to
determine if large particles are present that could clog
system hardware and to determine if water should be added
to make the waste pumpable. If the bench-scale tests show
that the organic constituents can be separated from the
waste, then pilot-scale tests are run. Wastewaters contain-
ing organics that are amenable to extraction by liquefied
carbon dioxide are tested with a pilot-scale unit located in
CF Systems' Massachusetts laboratory. Soils, sludges, and
other semisolids that are effectively treated by liquefied
propane or a propane/butane mix are tested in the field with
CF Systems' trailer-mounted unit. Based on successful
field demonstration results, clients have placed orders for
soils and wastewater units.
Bench-Scale Tests
CF Systems has conducted numerous bench-scale
tests for contaminated wastewaters.groundwaters, sludges,
and soils. Table 3-1 shows extraction efficiencies achieved
for removing various pollutants from wastewaters and
groundwaters. Liquefied carbon dioxide was used to
reduce contaminant concentrations that ranged from 0.4 to
520 ppm by more than 95 percent for 23 volatile and
semivolatile organics. Liquefied propane was used to
extract organics from samples of refinery sludges, separa-
tor sludges, and contaminated soils. Table 3-2 shows
extraction efficiencies for the separation of volatile and
semivolatile organics that range in concentration from 0.3
to 1,930 ppm. Extraction efficiencies ranged from 80 to
99.9 percent with a median of 97 percent. The bench-scale
data demonstrate that a wide range of organics can be
separated from aqueous and semisolid wastes; however,
the extraction of organics from semisolids is somewhat
less efficient than that of aqueous wastes.
Pilot-Scale Tests
The SITE program tests on the soils treatment unit in
New Bedford produced analytical and operating data used
for the evaluation system performance, operating condi-
tions, and equipment and material handling problems. The
performance of the unit was evaluated in terms of extrac-
tion efficiency and a mass balance. Extraction efficiency
per pass was defined as the input PCB concentration minus
the output PCB concentration divided by the input PCB
concentration (multipliedby lOOpercent). Aninventoryof
system inputs and outputs was established and evaluated
for total mass, total solids, and total mass of PCBs. Five
tests were run. Test 1 was a shakedown test and Test 5 was
a decontamination test. Results of these tests and evalu-
ations are summarized as follows:
• PCB removal efficiencies of 90 percent and
greater, were achieved for sediments containing
PCBs ranging from 350 to 2,575 ppm. A high
removal efficiency was achieved after several
passes, orrecycles,of treated sediments through
the unit. The low concentration for PCBs that
was achieved was 8 ppm.
• Extraction efficiencies greater than 60 percent
were achieved on the first pass of each test. Later
passes of treated sediments through the unit
resulted in efficiencies ranging from zero to 84
percent. This wide range was due to solids re-
tention in the system. Solids retained in the
system cross-contaminated treated sediments that
were recycled. (Recycling was necessary to
simulate the performance of a full-scale com-
mercial system. CF Systems' full-scale designs
do not include recycling since additional extrac-
10
-------
Table 3-1. Bench-Scale Test Results For Wastewaters and Groundwaters
Compound
Acetone
Acetonitrile
Acrylonitrile
Benzene
bis (2-ethylhexyl) phthalate
2-Butanone
Chloroform
1, 2-Dichloroethane
2, 4-Dimethylphenol
Dimethyl Phthalate
Isophrone
Methylene Chloride
2-Methylphenol
Napthalene
Nitrobenzene
PCB-1242
Phenol
Tetrachloroethane
Tetrahydro Furan
Trichloroethane
1,1,1-Trichloroethane
Toluene
Raw Waste
Concentration
(parts per million)
82
355
275
22
4
520
180
180
0.9
0.400
2.9
35
0.4
0.400
52
3.1
4
20
6
77
22
44
Extraction
Efficiency (1)
99.7
99.0
99.9
99.9
97
99.96
99.99
99.99
97.7
95
99.3
99.98
95
95
99.96
95
95
99.97
96.1
99.99
99.97
99.98
Notes:
(1) Bench-scale measurements, not necessarily an equilibrium limitation. Extraction efficiency calculated as percent of
pollutant removed.
tion stages and longer processing times are in-
volved.) Some solids appear to have been re-
tained in equipment dead spaces and intermit-
tently discharged during subsequent passes.
A mass balance was not established for PCBs. A
total of 157 grams of PCBs were fed to the unit.
Of the total, 80 grams were accounted for in
system effluents. Decontamination washes pro-
duced an additional 169 grams. The sum of
effluents and decontamination washes was,
therefore, 101 grams greater than that fed to the
unit.
This large difference may be due, in part, to
limitations of the analytical method. PCB ana-
lytical Method 8080 precision criteria estab-
lished for this project were plus or minus 20
percent and accuracy criteria were plus or minus
50 percent. In addition the mass balance calcu-
lation was dominated by the Test 4 feed concen-
tration. Therefore, error associated with the
Test 4 feed sample could also be a source of the
PCB mass imbalance. Another possibility is
contamination of the PCU from prior use at other
sites. However, CF Systems has not p r e v i -
ously fed materials to the unit that were known
to contain PCBs.
A good mass balance was established for total
mass and solids through the system. A total of
3-1/2 tons of solids and water were fed to the unit
during Tests 2, 3, and 4; of the total, 96 percent
was accounted for in effluent streams. A total of
789 pounds of solids was processed. Of the total,
93 percent was accounted for in effluent streams.
The slight imbalances, 4 and 7 percent, are at-
tributed to the inaccuracy of the weighing device
(1 percent), sample error, and accumulation of
mass in system hardware.
Metals were notexpected to be removed from the
sediments, and were not removed during the
extraction. Extraction Procedure Toxicity (EP
Tox) test results indicate that metals did not
leach from either treated or untreated sediments.
Characteristics of the sediments, with respect to
the EP Tox test, were not al tered by the treatment
11
-------
Table 3-2. Bench-Scale Test Results for Sludges and Soil
Average Feed Concentration
(PPM)
Oil and Grease
Volatiles
Benzene
Ethyl Benzene
Toluene
Xylenes (Total)
Semivolatiles
Acenophthylene
Acenaphthene
Anthracene
Benzo(A)pyrene
Bis(2-Ethylhexyl)phthalate
Chrysene
Fluoranthene
Fluorene
Naphthalene
Phenanthrene
Pyrene
Refinery
Sludge
32.2
370
<0.3
390
1160
714
1930
667
<35
889
<35
1360
<35
Separator
Sludge
5.68
23.8
25.0
13.4
106.3
27.7
1.9
4.7
6.4
13.9
41.5
27.7
6.9
Soil
10.5
Typical
Percent
Reduction
93 to 98
99
80 to 99
99
99
95
32
143
—
-
-
34.0
—
—
56.0
38.0
91 to 99
96 to 99
90 to 99
82
85
93 to 97
94 to 98
97
97 to 99
97 to 99
90 to 95
process. Copper and zinc concentrations were
typically greater than 1,000 ppm, while chro-
mium and lead ranged from 500 to 1,000 ppm.
• The decontamination procedure showed that
PCBs were separated from the sediment. Most
of the PCBs were contained in extract subsys-
tem hardware. Of the 81 grams of PCB fed to the
unit during Tests 2, 3, and 4, only 4 grams re-
mained in the final treated sediments. Subse-
quent decontamination of the PCU with a tolu-
ene wash showed that some PCB had accumu-
lated in system hardware. However, 91 percent
of the PCBs contained in decontamination resi-
dues were found in extract subsystem hard-
ware.
•A quality assurance/quality control (QA/QC)
review showed that analysis data of PCBs in
sediments for Tests 1 through 5 were sufficiently
accurate and precise for an engineering assess-
ment of the efficiency of this demonstration.
Operating conditions essential to the efficient per-
formance of the PCU were manually controlled and moni-
tored during Tests 2, 3, and 4. The operating conditions
included feed temperature, particle size, flow rate, pH, and
solids content; solvent flow rate and solvent/feed mass
ratio; and extractor pressure and temperature. The unit
generally performed as CF Systems predicted, although
some deviations from the planned specifications did occur.
An evaluation of operating conditions is summarized as
follows:
• Feed flow rates and extractor pressures were
controlled throughout the tests within specified
ranges. Feed flow rates ranged from 0.6 to 1.4
gpm. Extractor pressures ranged between 190
and 290 pounds per square inch guage (psig).
• During Test 2, feed temperatures for the last 4
passes were 10 degrees F lower than the mini-
mum specification, 60 degrees F. Decreased
extraction efficiency, which was apparent dur-
ing this test, could have been related to low
feed temperatures. Sustained low temperatures
could have the effect of seriously reducing ex-
traction efficiency in a full-scale commercial
system.
• Solvent flow fluctuated as much as 75 percent
above and below the nominal flow rate, 12 lb/
min. In Test 2, Pass 1, this caused the solvent-
to-feed ratio to fall below specifications. The
solvent flow fluctuations could affect the extrac-
tion efficiency in a full-scale system, since less
12
-------
solvent would be available to extract organic
pollutants from the feed soil.
• Specifications for maximum particle size, one-
eighth inch, were met by sieving sediments
through a screen. This was necessary to pre-
vent damage to system valves. Less than 1
percent of the sediment particles were greater
than one-eighth inch.
• Specifications for maximum viscosity, 1,000
centipoise (cP), were met by adding water to-
form a pumpable feed mixture. Feed viscosities
ranged from 25 to 180 cP. However, added water
increased the mass of waste by about 33 percent.
• Solids contents ranged from 6 to 23 percent and
fell below the minimum specification, 10 per-
cent, after the fourth pass of Tests 2 and 4. A 10-
percent minimum specification was set merely
to ensure that the technology would be demon-
strated for high solids content feeds.
• EPA and the developer will address corrective
measures for operational controls and material
handling issues. However, these measures are
not the subject of this report.
Equipment and system material handling problems
occurred, although some problems were anticipated.
Problems included the following:
• Internal surfaces of extractor hard ware and pip-
ing collected PCBs as evidenced by mass bal-
ances for PCBs and subsequent washes of the
unit with a refined naphtha fuel and later with
toluene. The washes recovered accumulated
PCBs as well as oil and grease.
These accumulations of organics are believed to
be the result of the short duration of the tests and
the small volume of organics contained in the
feed sediment, relative to the volume of the ex-
traction system hardware. PCBs are soluble in
oil and grease, which is believed to coat the
internal surfaces of system hardware. Continu-
ous operation of the unit has resulted in continu-
ous discharge of extracted organics at other
demonstrations of the technology.
• The unit intermittently retained and discharged
feed material solids. This is the result of the
relatively small volumes that were batch fed to
the unit. The unit was designed for continuous
operation, not short-term tests. In addition, only
50 to 150 gpd were run through the PCU, which
was designed to handle up to 2,160 gpd. There-
fore, some solids may have been retained in
equipment dead spaces and intermittently dis-
charged during subsequent passes.
• Solids were observed in extract samples, which
were expected to be solids free. This indicates
poor performance or failure of the cartridge
filter. An alternative type of filter should be
investigated by the developer.
• Extractor and treated sediment hardware con-
tained organic sludge from prior use of the unit
at a petroleum refinery. Presence of the petro-
leum residuals prevented complete interpreta-
tion of data collected for oil and grease and semi-
volatile organics.
• Low-pressure dissolved propane caused foam-
ing to occur in the treated sediment product
tanks. This hindered sample collection and
caused frequent overflow of treated sediment to
a secondary treated sediment product tank. CF
Systems states that design of a commercial-scale
unit will allow release of propane entrained in
the treated sediment and eliminate the foaming.
However, EPA cannot verify the claims on this
issue until it evaluates system operability for a
full-scale commercial unit.
CF Systems reports successful demonstration of its
mobile soils treatment unit at petroleum refineries, petro-
chemical, and TSD facilities throughout North America,
including:
• Texaco,Port Arthur, Texas
• Tricil, Toronto, Canada ,
• Chevron, Salt Lake City, Utah
• Exxon, Baton Rouge, Louisiana
• Chevron, Perth Amboy, New Jersey
• Unocal, Parachute Creek, Colorado
• BASF, Kearny, New Jersey
• United Creosote, Conroe, Texas
• Petro-Canada, Montreal, Canada
The unit had its initial startup at Texaco's Port Arthur
refinery in September 1987. Feeds run through the unit
included material from a clay pit, ditch skimmer sludge,
and tank bottoms. The resulting treated solids product
streams were analyzed by Texaco, and representative re-
sults are shown in Table 3-3. Levels of individual compo-
nents, including benzene, ethylbenzene, toluene, xylene,
and phenanthracene bettered the existing BDAT stan-
dards. In many cases these levels were found to be below
detection limits. Following the demonstration, Texaco
awarded CF Systems a contract to provide a commercial
unit to remediate 20,000 cubic yards of API separator
sludges and ditch skimmer wastes.
13
-------
The unit also operated at the Petro-Canada refinery in
Montreal for a six-week period. During this time, the unit
successfully processed 14 different feed types, ranging
from API separator sludges to contaminated solids. The
unit consistently achieved organic removal levels better
than existing BDAT standards.
A series of demonstration tests was run at Tricil
Canada's TSD facility in the Province of Ontario. The
feeds processed included API incinerator sludge, paint
wastes, synthetic rubber process waste, and coal tar wastes.
The unit affected a large-volume reduction of the material
processed and the level of volatile organics was reduced
such that disposal of the material in a local Canadian
landfill was acceptable.
Generic solvent extraction and incineration technolo-
gies were named by EPA as BDAT for listed petroleum
refinery hazardous wastes (40 CFR 261.32 K048-K052).
CF Systems' and other developer's performance data, from
pilot-scale tests, were included in the basis for setting
performance specifications for treatment of these wastes.
The MDU completed a treatability study for the Texas
Water Commission in conjunction with Roy F. Weston at
the United Cresoting Superfund Site in Conroe, Texas.
The objective of this study was to evaluate the effective-
ness of solvent extraction for remediation of soil contami-
nated with creosote. PAH concentrations in the soil
obtained from the capped area were reduced from 2,879
ppm to 122 ppm, demonstrating that reductions greater
than 95 percent were possible. Representative results from
this study are shown in Table 3-4.
Full-Scale Applications
Operating, performance, and cost data are not avail-
able for a full-scale system. EPA intends to collect these
data at a later date. Over the past 18 months, CF Systems'
commercial activity has consisted of the following major
efforts:
• In March 1989 the first of the Company's 50-
tons-per-day unit was shipped to Star's Port
Arthur, Texas, refinery (Texaco) for a 14-month
full-scale commercial cleanup of oily sludge
wastes. Under this contract, the soils treatment
unit will treat about 20,000 tons of sludge to
produce cleaned solids, treatable water, and oil
for recycle. This unit became operational in
July 1989.
• A custom-built, 60-tons-per-day soils treatment
unit was shipped to ENSCO's El Dorado, Arkan-
sas, incineratorfacility in November 1988. Since
ENSCO is reorganizing their El Dorado opera-
tion the unit has not been placed on line; how-
ever, the unit will extract organic liquids from a
broad range of hazardous waste feeds sent to the
site for incineration. The extracted liquids will
be used as incinerator secondary combustion
fuel, while the residues, reduced in heat content,
will allow higher incinerator throughputs for
ENSCO.
• A 20-gpm wastewater treatment unit was sold to
Clean Harbor, Inc. It is expected to be installed
at a TSD facility in 1989 in Baltimore, MD.
• CF Systems has established performance speci-
fications for the LL-series wastewater treatment
unit. A 99-percent extraction efficiency is speci-
fied for 2,000 ppm of trichloroethylene in waste
waters. A 97-percent extraction efficiency is
specified for 12,000 ppm of methyl isobutyl
ketone in wastewater.
3.3 Waste Characteristics and Operating
Requirements
The SITE program tests provided waste characteriza-
tion and system operating data for the propane-based pilot
unit, which is designed for the treatment of soils and
sludges. CF Systems' wastewater treatment unit was not a
subject of the SITE tests; therefore, no discussion of that
unit appears in the sections that follow. However, some
aspects of system operation and economics for the two
technologies are similar. Details on the two technologies
are presented in Appendix A - Process Description.
Feed Material Specifications
Physical characteristics of wastes fed to CF Systems'
sludge and soils treatment technology must fall within the
ranges shown in Table 3-5. Solids greater than 3/16 inch
may clog process valves and piping. Feed pH must be
maintained between 6 and 10 to protect process equipment
from corrosion. The feed must be pumpable in order to
flow through the system under pressure; therefore, a
maximum viscosity of 5,000 cP is established. Viscous or
dry materials are typically slurried with water, although
this practice increases the volume of waste and may require
dewatering. If the feed is less than 60 degrees F, freezing
may occur in the extractor. Conversely, feeds greater than
120 degrees F may cause solvent vaporization. CF Sys-
tems' experience has shown that extraction efficiencies are
high when feed solids and water contents fall within the
wide ranges shown in Table 3-5.
If the technology was considered for a full-scale
cleanup at New Bedford Harbor, pretreatment would be
required to bring the sediments within the required physi-
14
-------
Table 3-3. Texaco, Port Arthur Performance Data
BOAT
Parameter Levels
(mg/Kg)
CLAY PIT AREA(1) SLUDGE(1) SLUDGE(1)
Feed Treated Water Feed Treated TCLP Feed Treated
Solids Solids Solids
(mg/Kg) (mg/Kg) (mg/L) (mg/Kg) (mg/Kg) (mg/L) (mg/Kg) (mg/Kg)
DITCH SKIMMER(1)
Feed Treated TCLP
Solids
(mg/Kg) (mg/Kg) (mg/L)
Water(1)
Solids(1)
60.5
22.3
17.2
62
32
Total Oil
&Grease(1)
Benzene(S)
Ethylenzene
Toluene
Xylenes
Fluorene
Naphthalene
2-Methyl
Napthalene
Phenanthrene
--
9.5
67
9.5
63
..
--
--
7.7
--
9.6
13
16
<0.1
—
210
300
—
1.9
<0.1(3) <0.01 <2.0
<0.1 <0.01 <2.0
<0.1 <0.01 <2.0
<0.01 <2.0 <2.0
—
<5.3 <50
<5.3
31
3.6
<2.0
<0.01
<0.01
<0.01
<0.01
Notes:
<3.3
<3.3
(1) Water, solids, oil, and total oil and grease reported as percent by weight.
(2) TCLP is the Toxicity Characteristic Leaching Procedure.
(3) < indicates less than the detection limit shown.
57
33
10
13.7
20.2
54.4
75.9
45
30
1.0
<2.0
<3.3
<3.3
53
35
12
18.6
0.7
5.1
13
52
71
9.3
16.5
<0.1 <0.01
<0.1 <0.01
<0.1 <0.01
<0.1 <0.01
<0.20
<0.20
<0.20
-------
Table 3-4. United Creosote Superfund SITE Performance Data
Feed
Treated
Compound
Acenaphthene
Acenaphthylene
Anthracene
Benzo(A)anthracene
Benzo(A)pyrene
Benzo(B)fluoranthene
Benzo(G,H,I)perylene
Benzo(K)fluoranthene
Chrysene
Dibenzo(A,H)anthracene
Fluoranthene
Fluorene
Indeno(l,2,3-CD)pyrene
Naphthalene
Phenanthrene
Pyrene
Total (MG/KG)
Soil
Soil
360
15
330
100
48
51
20
50
110
ND
360
380
19
140
590
360
2879
3.4
3.0
8.9
7.9
12
9.7
12
17
9.1
4.3
11
3.8
11
1.5
13
_LL
122.6
Notes: Mg/Kg on a dry weight basis. ND indicates not detected.
Table 3-5. Sludge and Soil Feed Requirements
Solids Size
pH
Viscosity (centipoise)
Feed Temp, (degrees Fahrenheit)
Feed Solids (percent by weight)
Water (percent by weight)
Organics (percent by weight)
Minimum
6
0.5
60
0
20
1
Nominal
1/8
10
70
30
40
20
Maximum
3/16 inch
10
5,000
120
50
90
90
16
-------
cal specifications. Less than one percent of the New
Bedford Harbor sediments were greater than 3/16 inch;
nonetheless, sieving through screens was required to remove
oversize particles. The sediments were also viscous;
therefore, water was added to ensure pumpability. On one
day during the SITE tests, ambient temperatures fell and
caused the feed to drop below 60 degrees F, which may
have affected extraction efficiency. A full-scale applica-
tion would require sieving of untreated sediments, water
addition, and heat addition. Cost-effective disposal of
treated sediments would require dewatering to minimize
disposal volumes. In turn, dewatering effluent would
require treatment at a publicly owned treatment works or
by an onsite waste water treatment system. CF Systems'
experience has shown that oversized solids removal is
sometimes required and that water addition is necessary for
dry solids.
The technology is capable of treating the wide range of
waste matrices found in most waste handling situations.
The ranges specified for viscosity, feed solids water con-
tent, and organics content are very broad. The ranges
specified for solids size, pH, and temperature are also
broad but these parameters are more likely to exceed input
specifications. However, off-the-shelf technology isavail-
able to bring off-spec feeds within CF Systems' required
operating ranges.
Utilities and Labor
Utility requirements for the technology include (1)
electricity, (2) cooling water maintained at 60 to 80
degrees F, (3) commercial-grade propane and/or butane,
and (4) nitrogen to pressure test the equipment during
startup. The amounts of electricity, propane, and butane
used during the SITE tests were not significant. In addi-
tion, the unit was satisfactorily pressure-tested. Theamount
of noncontact cooling water, 5 gpm, was significant on this
site and should be considered in the design of any future
application.
All CF Systems' units are mobile and can be trans-
ported on public roads. The modular design of the units
mitigates the need for field fabrication. The sizes of CF
Systems treatment modules are limited by the need for
transportability on public roads. A firm, level foundation
is needed and the area required for the 200-ton-pcr-day
commercial-sized unit, including ancillary skids, is about
4,000 square feet. An estimated 2,000 hours of labor are
required to install the system. A site engineer, a site
manager, and additional labor and safety equipment would
also be required. Space for safe storage of the liquefied
solvent is also necessary. A large commercial-scale unit
can be operated continuously by four or five people per
shift (two or three unit operators, one supervisor, and a pre-
/post-treatment operator). Three such teams, each operat-
ing on an 8-hour shift, can be used to operate the unit on a
24-hours-per-day, 7-days-per-week basis.
3.4 Materials Handling Requirements
Pre- and Post-Treatment
Requirements for pre- and post-treatment of wastes
are site specific. The SITE program experience, at New
Bedford, provided an example of the types of material
handling needs that must be addressed. Pretreatment of
New Bedford sediments would be required to remove par-
ticles greater than 3/16 inch, to decrease viscosity, and to
maintain feed temperature. In addition, the feed consis-
tency should be homogeneous to minimize process uncer-
tainty and to improve control of flowrates. Hence, solids
removal, water addition, mixing, and storage are important
pretrcatment steps. The addition of water and heat can be
incorporated into either the solids removal or the mixing
operations. Sufficient storage capacity is also important
for those operating days w hen treatment goals are not being
met because of equipment failure or slug loads of high
concentration wastes.
A sieving and screening method is the most appropri-
ate pretreatment method for New Bedford Harbor sedi-
ments, based on experience during the demonstration test,
and was thus selected for this application. Vibrating
screens are more widely used than any other screen types
because of their larger capacity per unit screen area and
their higher efficiency. However, wet or sticky materials
tend to blind the screen; therefore, wet screening with
sprays can be used to discourage blinding.
Manual or automated high-pressure water spraying is
assumed adequate to treat oversized solids. These coarse
solids would be disposed of with fine-grained sediments
treated by the CF Systems technology. Spray water would
be collected and reused. Common mixing equipment and
storage tanks are adequate to provide a homogeneous
source of feed for the CF Systems technology. Heat can be
provided by steam addition.
Post-treatment must be considered for the two product
streams generated by this process. The ex tract contains the
concentrated organics and the treated sediments contain
the water and solids. Provisions for extract containment,
handling, storage, and transport off site would have to be
made. The volume of treated sediments would be greater
than that of the untreated if water is added during prelreat-
mcnt; dewatering could be necessary. However, dewater-
ing effluent could be reused in the pretreatment operation.
Thus, wastewatcr treatment costs would be minimized.
Treated sediments would be disposed of in either a Re-
source Conservation and Recovery Act (RCRA) approved
landfill or a confined disposal facility located in the harbor.
17
-------
Materials handling requirements would be integrated
with the CF Systems technology for applying the technol-
ogy to a New Bedford Harbor cleanup. The overall process
would consist of the following steps:
Step 1. Dredging
Step 2. Untreated Sediment Storage
Step 3. Untreated Sediment Handling
Step 4. Coarse Solids Separation, Water, and Heat
Addition
Step 5. Extraction
Step 6. Extract Collection
Step 7. Treated Sediment Dewatering
Step 8. Transportation of Treated Sediments
Step 9. Offsite Disposal of Extracted Organics
Step 10. Disposal of Treated Sediments.
Each of these steps is shown in the flow diagram,
Figure 3-1.
Process Operability
Foaming in the treated sediments and extract product
tanks was evident throughout the SITE tests. This is
suspected to be caused by propane entrainment in the
treated sediments propane mixture, and has two adverse
effects. First, extracted PCBs may be present in the foam.
Second, foaming increases the volume of material that
must be handled in the product stream, thereby increasing
the probability of PCB migration and decreasing the feed
throughput. Foaming can be mitigated by using oversized
tank volumes, which have lower surface-to-volume ratios;
thus, nominal throughputs can be maintained by use of
large treated sediment collection tanks. CF Systems has
addressed these issues in their scaled-up design and in the
latest unit to be built. The commercial designs also contain
an additional pressure relief step to more gradually de-
crease the pressure and thereby decrease foaming.
Solids and oil retention in process hardware also
affected interpretation of SITE test data. The pilot-unit
was operated in a recycle mode to simulate m ultiple stages,
which caused cross-contamination of the recycled treated
sediments. In addition, very small volumes were run
through the unit during each day of testing. CF Systems'
full-scale units do not incorporate recycling, operate in a
once-through mode, and are expected to be on-line 20
hours per day. Therefore, solids and oil retention is not
expected to be a significant problem, although some oil
will coat internal hardware surfaces and should be re-
moved with an organic solvent if the character of the feed
changes substantially, to prevent cross contamination.
Solids were observed in extracted oils; however, this
minor problem can be corrected by more frequent changes
of filter elements or by selection of different filters.
3.5 Health and Safety Issues
The SITE tests indicated that no acute threats to
operator health and safety are associated with operation of
the unit. Combustible gas meters indicated that the unit did
not leak significant amounts of propane. Therefore, opera-
tion of the unit does not present an explosion threat much
different from that associated with domestic propane us-
age. Background air sampling and personnel monitoring
results indicate that organic vapors and PCB levels were
present at levels below the detection limit for the analytical
methods. The unit did not cause a sudden release of
propane/butane or liquids. Only minor leaks occurred and
staging area soils were not affected. Gases vented from the
system at the conclusion of the tests were passed through
a carbon canister. Analysis results showed that the gases
contained minor amounts of PCB. The greatest threat
presented by handling of the New Bedford Harbor sedi-
ments was dermal exposure. OSHA Level B protection is
recommended for personnel who handle treated and un-
treated New Bedford Harbor sediment. Level C protection
is recommended for extraction process operators.
All electrical equipment is explosion proof and all
potential sources of ignition are restricted for a 20-foot
perimeter around the unit. Spark-proof tools are also used.
The solvent recovery hardware, which involves major
phase changes for propane, is very similar to commercial
refinery depropanizers, used safely throughout the world.
3.6 Testing Procedures
A portable GC and a chemist should be available
onsite to allow a rapid response to changes in feed compo-
sition or operational control. The Spittler Method was used
at New Bedford as a more timely alternative to EPA
methods. However, even with this method, 24 hours were
required for sample shipment and subsequent analysis.
Reviewers suggested the use of EPA Method 680,
since the CF Systems technology could have selectively
extracted higher molecular weight PCB congeners as
opposed to lower weight PCB congeners. Method 680
would reveal any selective extraction, since Method 680
is used to analyze individual PCB congeners. Method
8080, a less expensive analysis method, would not reveal
selective extraction since it is used to analyze mixtures of
PCBs called Aroclors, instead of individual congeners.
EPA Method 8080 was chosen over Method 680 since
selective extraction was minor and since it analyzes for the
classes of congeners that compose the majority of PCB
18
-------
Dredging Ship
.0
±1
STEP
1
Harbor
Bub Temporary
Storage
ConcMM rated
PC a/Organic*
Conveyors
and
Other Solids
Handling
To On or I STEP
Orf-Slla V 10
Confined
Olspotal
Facility
Figure 3-1. New Bedford Harbor Application Flow Diagram.
19
-------
contaminants (Aroclors 1242 and 1254) in the harbor sediment PCB concentration of 8,700 ppm while Method
sediments. 8080 showed 2,575 ppm. Dataquality objectives were met
, . fn „„„„ for each measurement. Therefore, regulatory or engineer-
Methods 680 and 8080 produced similar relative } interpretation of future PCB analyses should include
results, but very different absolute results. Use of Method consideralion of ^ analysis methods used. interpretation
680 in Test 4 showed a PCB extracuon efficiency of 96 of ^^ from a pCB ^^lity stud should include
percent and Method 8080 showed a similar efficiency, 87 a discussion of the ision of ^ ^ is melhod ^ wdl
percent. However, Method 680 showed an untreated as ^ accuracv
20
-------
Section 4
Economic Analysis
4.1 Introduction
The objective of the economic analysis was to esti-
mate costs for a commercial-size site remediation using the
CF Systems technology. This evaluation illustrates how
variations in process conditions, such as volume to be
treated, treatment time, water dilution of the raw feed, and
reduction in outlet PCB concentration can impact system
design and pre- and post-treatment costs. Five treatment
cases were evaluated for PCB removal from New Bedford
Harbor sediments to illustrate the cost methodology.
CF Systems developed costs for a base case and a hot
spot case, then extrapolated the costs to three other cases.
The base case refers to the treatment of 880,000 tons
(695,000 cubic yards) of sediments containing 580 ppm of
PCB. The hot spot case refers to the treatment of 63,000
tons (50,000 cubic yards) of sediments containing 10,000
ppm of PCB. The three additional cases were developed
that represent variations of both the base and hotspotcases.
These variations include changes of the on-slream factor,
elimination of the need for adding water to reduce solids
content, and higher PCB removal goals.
Standard process design sizing and costing algo-
rithms were used by CF Systems. This consisted of: using
off-the-shelf equipment of standardized size; obtaining
total treatment capacity by adding units in parallel; and
obtaining increased reduction of PCB outlet concentration
by adding units in series. CF Systems assignedan accuracy
of plus or minus 20 percent to its cost estimates. This is a
reasonable estimate given the fact that off-the-shelf equip-
ment is incorporated into CF Systems' designs. This
accuracy goal falls within the order-of-magnitude esti-
mates of plus 50 to minus 30 percent defined by the
American Association of Cost Engineers. Results of the
analysis and apparent trends are as follows:
• The estimated base case cost, including pre-and
post- treatment, is $148/ton of raw solids feed,
with an accuracy range of $104/ton to$222/ton.
Sediment excavation and pre- and post-treat-
ment costs are 41 percent of the total cost. This
post-treatment estimate does not include the fi-
nal destruction of the concentrated extract.
• The above costs are based on a system design
using two PCU-1000 units, each with nominal
capacities of 250 tons/day. This design uses two
extraction units in parallel with one solvent re-
covery section in series. One extraction unit
consists of two mixer/settler units. A total
treatment time of 8.3 years is projected.
• The hot spot cost, including pre- and post- treat-
ment, is $447/ton of raw solids feed, with an
accuracy range of $3l3/ton to $671/ ton, with
sediment excavation and pre- and post-treatment
costs being 32 percent of the total.
• The design for the hot spot case is based on the
PCU-500, with a nominal capacity of 100 tons/
day. This design utilizes two modules in series,
with each module consisting of an extraction and
a solvent recovery unit in series. A total
treatment time of 1 year is projected.
• Key to all designs is the assumption of 85 percent
on-stream factor. This was not demonstrated by
operating the PCU-20 unit at New Bedford Har-
bor because of significant materials handling
problems associated with recycle of treated
solids. This recycle was required to evaluate
PCB extraction using more than two extraction
stages. Since recycle is not a unit operation for
a commercial-size unit.CFSystemsclaimsthat
material handling problems would be minimized
with a commercial unit. CF Systems must
demonstrate an 85 percent on-strcam factor on
commercial unit.
• To attain a total treatment cost less than $100/
ton, the solids feed content to extraction unit
must be greater than 26 percent to minimize pre-
and post-treatment costs.
4.2 Basis for Process Design, Sizing, and
Costing
In general, soil remediation projects encompass exca-
vation, treatment, containment, and/or removal of con-
taminated soils and sludges. Depending upon the types of
contamination and the level of cleanup required, further
processing of sediments treated by CF Systems' extraction
system may be necessary. This may include fixation for
heavy metals and incineration of the extracted organics;
21
-------
however, these costs are not addressed in this study. A
typical remediation project may consist of any combina-
tion the following steps, for which equipment sizing and
costing is required to meet a specific treatment plan for
total tonnage, treatment time, and reduction in contami-
nant concentration:
1 The excavated material may have to be slurried
with water to create a pumpable mixture.
2 The slurry is passed through a shaker screen to
remove material larger than 1/8-inch diameter.
Oversized material may be crushed and recycled
to the screens or separately washed.
3 The pH of the sieved slurry is monitored and, if
required, lime is added to the mixture to maintain
a pH between 6 and 8.
4 The slurry may require thickening prior to the
slurry being pumped to the CF Systems Extrac-
tion Unit.
5 The slurry is processed in modular extraction
units to reduce the PCB content of the solids.
6 Two product streams exit the extraction unit: a
solids/water stream and a liquid organic stream.
The organic stream will generally be returned to
the client for reuse or disposal.
7 The solids/water stream is dewatered through
the use of a gravity thickener, filter press, or
centrifuge. The water from the dewatering step
may be reused to slurry dry feed solids. Excess
dewatering effluent could be discharged to a
POTW or treated and discharged onsite.
CF Systems has developed a proprietary model for
estimating site remediation costs. Outputs of the model are
only intended for use in planning, scoping, and the inviting
of firm bids. The Agency based this economic analysis on
estimates prepared by the developer. No attempt was made
to mirror the developer's work since this would involve a
substantial effort to design and cost a facsimile of CF
Systems' technology. Some features of the technology are
unique to CF Systems' design approach. These features
include unit modularity, process component integration,
safety instrumentation, relief system backup, and auto-
matic shutdown.
A cost analysis was prepared by breaking the costs
into 12 groupings. These will be described in detail as they
apply to the CF Systems technology. The categories, some
of which do not have costs associated with them for this
technology, are as follows:
• Site preparation costs
• Permitting and regulatory costs
• Equipment costs
• Startup and fixed costs
• Labor costs
• Supply costs
• Supplies and consumables costs
• Effluent treatment and disposal costs
• Residuals and waste shipping, handling, and
transport costs
• Analytical costs
• Facility modification, repair, and replacement
costs
• Site demobilization costs.
The 12 cost factors, along with the assumptions
utilized by CF Systems in their proprietary cost model, are
described below with respect to the soils treatment tech-
nology.
SITE Preparation Costs
Approximately 20 weeks are required to mobilize and
demobilize the extraction unit and pre- and post-treatment
equipment. The cost of ancillary service, such as construc-
tion of concrete pads and rental of construction equipment,
increases the site preparation costs by about 50 percent. No
land costs are assumed for the New Bedford site.
Permitting and Regulatory Costs
Since New Bedford Harbor is a Superfund site, it is
assumed that no permits will be required, neither Federal
nor State. The need for developing analytical protocols or
monitoring records is assumed not to exist based on SITE
program tests.
Equipment Costs
Capital costs include equipment, maintenance and
technical service, engineering, procurement, fabrication,
permitting, startup and operating assistance, and facility
modification, repair, and replacement. Provisions for pre-
and post-treatment of New Bedford sediment would in-
volve solids handling and feed treatment equipment. Each
cost element is described below.
The solids handling equipment is provided to move
New Bedford Harbor sediments from the stockpile to the
CF Systems treatment site. Contaminated dry soils will be
excavated through the use of equipment such as frontend
loaders, backhoes, or bulldozers. These soils will then be
fed into a preliminary screening device to remove any
materials larger than four-inches in diameter. Solids cap-
tured in the screens will be collected, washed, and disposed
of in an appropriate manner. Screened material will be
transported on a conveyer belt to a pug mill where size
reduction is effected. The pug mill will combine the dry
22
-------
solids with water to produce a solids/water exlrudate. This
paste will then travel via a second conveyer belt to a tank
or pump where additional water will be added to produce
slurried solids. Solids handling equipment costs are based
on 10 hours of daily operation for the duration of the
remediation.
Feed pretreatment equipment is provided to screen
and slurry the feed prior to the solvent extraction system.
Slurried solids from either the pug mill or the dredge will
pass through a multilayered shaker screen similar to those
used in the oil drilling industry. The objective will be to
screen out solids larger than 3/16 inch in diameter. Solids
captured by the screen will be collected, washed, and
recycled to the pug mill or crusher/grinder for size reduc-
tion. Sludge passing through the screen will be collected
in a storage tank equipped with mixers. If required, lime
will be added at this point to maintain a pH between 6 and
8. The slurry will then be pumped from this tank either to
the extraction unit or to a thickener. If pumped to a
thickener, the slurry will be thickened to approximately 50-
percent solids. This is accomplished through the use of
either a moving screen or a decantation system, depending
on the water solubility of the waste. Water extracted by the
thickener will be returned to the dredge area or to another
approved discharge point. The thickened solids slurry will
be pumped to another holding tank and then fed to the
Extraction Unit.
The product handling equipment is provided to re-
ceive the product streams from the extraction system and
to deliver these product(s) to the client for disposal. The
de-oiled solids and water produced from the extraction
process will be dewatered. This stream will be run through
a belt filter press, where a combination of pressure and
conditioning flocculents, if required, will remove excess
water, leaving a cake with approximately 40- to 45-percent
solids. Water separated from the slurry will be returned to
the dredge area or to the water treatment system. De-oiled
solids in the form of a cake will move via conveyer from the
belt filter press to a small blending mill.
Startup and Fixed Costs
Various facilities would be required to support the
operation and maintenance of the CF Systems technology
or any other onsite remediation technology. Those facili-
ties would include office, laboratories, laboratory analy-
ses, security, sanitary facilities, power generation, and a
cooling water supply. Most of these facility costs are fixed
for a given site. However, some costs, such as power
generation and cooling water supply, vary in proportion to
the capacity of the extraction unit.
Labor Costs
The extraction unit would operate 24 hours a day, 7
days a week. Fulltime operating staff would include 2
operators and a shift supervisor. A site engineer and a site
manager would be onsite 8 hours per day. Pre- and post-
treatment would require 2 operators 24 hours a day, 7 days
a week. Safety equipment for all site personnel is estimated
to cost $40 a day per man, which includes disposal of
contaminated gear.
Supplies and Consumable Costs
No supply costs are incurred.
Utilities Cost
Actual equipment to generate and deliver utilities is
accounted for in the Startup and Fixed Cost Group. Utili-
ties include electrical power and propane. Unit costs used
in the cost estimates for electricity were 6 cents per kilowatt
hour and 20 cents per pound of propane.
Effluent Treatment and Disposal Costs
The only continuous wastewater effluent associated
with this technology is once-through, noncontact cooling
water. If noclosed loop system is available, water from the
post-treatment solids dewatering step would beused to
slurry dry feed solids. Excess dewatering effluent would
bereturned to the dredge area or intermittently discharged
to the harbor.The cost for monitoring these discharges is
included in the Analytical Cost Group.
Residual and Waste Shipping, Handling, and
Treatment Costs
No costs are estimated here for residuals shipping.
The costs associated with treated solids dewatering and
storage and extract storage are estimated under the Site
Preparation, Equipment, Labor, and S upplies Cost Catgeo-
ries. Solids would be returned to the harbor or would be
treated by fixation for metals. Extracted oils would be
transported and incinerated at minimal cost since the
extract could serve as a fuel supplement.
Analytical Costs
In the absence of a site sampling and analysis plan,
analytical costs are estimated at $500 per day and are
included in the Startup and Fixed Cost Group.
Repair, and
Facility Modification,
Replacement Costs
These costs are borne by the developer since the
equipment is marketed though lease agreements. There-
fore, the developer has included these costs in the Equip-
ment Cost Group.
23
-------
SITE Demobilization Costs
Demobilization costs are included in the Site Prepa-
ration Cost Group.
4.3 Developer's Estimate for a New Bedford
Harbor Cleanup
CF Systems prepared cost estimates using their pro-
prietary model for two cases: a large mass (695,000 cubic
yards of sediment) of low PCB concentration (580 ppm)
referred to as the "base case;" and a small mass (50,000
cubic yards) of high PCB concentration (10,000 ppm)
referred to as "hot spot." Each is described below:
Base Case: The quantity of material to be treated, for
the base case, is 695,000 cubic yards of PCB-contaminated
soil. This quantity of material represents removal and
treatment of all the contaminated soil in the New Bedford
Harbor estuary. The level of PCBs in this material is
assumed to average 580 ppm on a dry solids weight basis.
The PCBs in this material w ill be reduced to a 50-ppm level
via solvent extraction. The time schedule for processing
this material is about eight years.
For this case, which involves "a large tonnage removal
for m ultiple years on site, CF Systems recommends the use
of two PCU-1000s but only one solvent recovery section.
This system will process about 500 tons/day in the follow-
ing configuration:
HotSpotCase: Thequantity of material to becleaned,
for the hot spot case, is 50,000 cubic yards of PCB-
contaminated soil. This quantity of material represents
removal and treatment of the high concentration spots in
New Bedford Harbor. The level of PCBs in this material
is assumed to be 10,000 ppm on a dry solids weight basis.
The PCBs in this material will be reduced to 10 ppm on a
dry solids weight basis via solvent extraction technology.
This represents a 99.9-percent removal of PCBs. The time
schedule for processing this quantity of material is ap-
proximately one year.
For this case, CF Systems recommends the use of four
PCU-500s, which would complete the remediation in
about 1.2 years. These are 100 ton/day units, each having
its own extraction and solvent recovery sections. The
configuration of these units is shown below.
The selection of this size unit and the paired configu-
ration is made to reduce onsite time and the units can be
deployed to other c ustomers at the end of the job. Two units
in series are required to achieve an extraction efficiency of
Pretreated
Waste
PCU-1000
Extraction
l~" Section
- PCU-1000
Extraction
Section
—i
Solvent
Recovery
Section
i
To
Product
Handling
24
-------
Pretreated_
Waste
JPCU-500 \-
JPCU-500J--
JPCU-500J--
I J
JPCU-500 [_
i I
I
To
^Product
"Handling
i
99.9 percent. The parallel configuration is required to
handle the total volumetric throughput.
Process conditions and costs developed by CF Sys-
tems are summarized for each case in Table 4-1. The base
case involved removing 91 percent of the PCB from a large
volume of sediment. The total average cost over the 8-year
duration of the project is $148 per ton treated. Pre- and
post-treatment costs represent about 41 percent of the total
cost. The hot spot case involves removing 99.9 of the PCB
from a somewhat smaller volume of sediment. The total
average cost for treating hot spot sediments is $497 per ton
over a project life of approximately one year. The pre- and
post-treatment costs account for 32 percent of the total
cost. Cost differences between base and hot spot cases are
due to the significantly different PCB removals required,
as well as the different project lives.
Variations of the base and hot spot cases were evalu-
ated to determine the cost impacts of different removal
efficiencies, pretreatment requirements, and on-stream
factors. These various cases are listed below and are
compared to the base and hot spot cases:
Case 1A The base case: The base case in-
volves the exraction of 91 percent PCBs con-
tained in 695,000 cubic yards of harbor sedi-
ments. An 8-year project life, and an 85-percent
on-stream factor were assumed. The base case
includes pre-treatment for the reduction of the
solids content.
Case IB The base with a 70-percent on-stream
factor: This case is similar to the base case
except that an on-stream factor of 70 percent is
assumed instead of 85 percent. A less optimis-
tic on-stream factor would result if material-
handling problems or equipment breakdowns
ocurred. A lower on-stream factor would require
equipment with higher capacities in order to keep
the project within an 8-year project life.
Case 1C The base case without solids content
reduction: This case is similar to the base case
except that no solids content reduction would be
required. Harbor sediments contain approxi-
mately 40 percent solids; however, the SITE
program tests showed that solids concent reduc-
tion was necessary to improve pumpability.
This involved adding water to the sediment to
reduce the solids content to 17 percent. The
consequences of water addition include increased
throughput and increased equipment sizes. With
more experience at the New Bedford site, CF
Systems may be able to modify their equipment
and operating procedures to accommodate sedi-
ments with 40 percent solids. Thus the need for
water addition would be eliminated, throughput
would be decreased, and equipment sizes would
also be decreased.
Case ID The base case with increased extrac-
tion efficiency: This case is similar to the base
case except that an extraction efficiency of 98
percent, instead of 91 percent, is assumed. This
change would result in a PCB outlet concentra-
tion of 50 ppm instead of 100 ppm. Increased
extraction efficiency requires an increased
number of extraction units that would be aligned
in a series flow configuration.
Case 2 The hot spot case: The hot spot case in-
volves treating 50,000 cubicyardsof sediments
containing 10,000 ppm of PCB. An extraction
efficiency of 99.9 percent, an on-stream factorof
85 percent,and a 1-yearprojectlifeareassumed.
The hot spot case includes pretreatment for the
reduction of the solids content.
25
-------
Table 4-1. Base Case and Hot Spot Case Summary
Capacity Base Case Hot Spot Case
Raw sludge (40% solids): cubic yards 695,000 50,000
tons 880,000 63,000
Processing Time: years 8.35 1.19
Operating Days 2,591 369
Raw sludge feed rate (at 40% solids)
tons/operating day 339.5 171.5
Extractor Feed: % Solids 26.7 26.7
total tons processed 1,319,414 94,922
nominal system size (tons/day) 500 250
feed rate (tons/operating day) 509.2 257
Inlet PCS Concentration: ppm 580 10,000
Outlet PCB Concentration: ppm 50 10
PCB Reduction: percent 91 99.9
Configuration* (1) (2)
Processing Fee (1989 $)
Facilities $ 5,170,676 $ 762,496
Extraction $62,109,781 $15,857,695
Pre-/Post-Treatment $46,172,028 $7,993,608
Contingency $ 11,345,248 $ 2,461,380
Project Management $ 5.672.624 $1.230.690
TOTAL $130,470,358 $28,305,869
Total Life Cycle Unit Cost ($/ton):
Extraction only $ 71 $251
Total $148 $447
NOTES:
'Configuration: 1 - Two extraction sections connected in parallel feeding one solvent recovery section connected in
series.
2 - An extraction and solvent recovery section in series connected parallel with a second identical
extraction and solvent recovery section.
26
-------
Costs for all of the cases were developed by CF
Systems and are shown in Table 4-2. Process differences
among the cases are also shown, as are costs for each cost
category. The differences among the cases provide the
following conclusions:
• A decrease in the on-stream factor from 85 to 70
percent increases all costs by approximately 20
percent. This is the result of increased equip-
ment capacities and sizes required.
• Elimination of the pretreatment step to decrease
the solids content can result in a 30-percent cost
savings. This savings occurs as a result of re-
duced volumetric throughputs, reduced equip-
ment sizes, and elimination of some pre- and
post-treatment steps.
• Changing the base case PCB removal goal from
91 to 98 percent increases total costs by approxi-
mately one-fifth. Although not shown in Table
4-2, an additional case was evaluated to ob-
serve the effect of reducing the PCB removal
efficiency from 99.9 to 99 percent for the hot
spot case. This resulted in a cost decrease of
approximately one-quarter. Therefore, increas-
ing or decreasing the removal efficiency by an
order of magnitude results in corresponding
increase of decrease of approximately 25 per-
cent.
• Startup and fixed and analytical costs account for
4 to 6 percent of remediation costs.
• Costs specific to the extraction unit account for
53 to 68 percent of remediation costs.
• Sediment excavation and pre-and post-treatment
costs account for 28 to 41 percent of remediation
costs.
• Eliminating the need for decreased solids con-
tent in the feed affects costs more than any other
variable. However, the greatest uncertainty lies
with the assumptions for the on-stream factor
since EPA has not evaluated this variable and
CF Systems has no long-term operational dala
available.
4.4 Evaluation of the Developer's Estimate
CF Systems has designed and built a 50-ton/day single
train system, which was shipped to a customer in the first
half of 1989 and was scheduled for startup in 1989. They
have also designed larger systems of 100- and 200-tons/
day throughput, but have not built these to date. The
system designed for the base case is called a PCU-2000
and is configured as two 200 ton/day extraction sections
connected in parallel and one propane solvent recovery
section connected in series. The complete extraction
system provides a total capacity of about 500 tons/day,
which, in combination with an 85-percent on-stream fac-
tor, results in an 8-year treatment time for 695,000 cubic
yards of sediment. The 92-percent reduction in solids PCB
concentration and 26-percent solids feed to the extraction
unit are based on data obtained from testing the PCU-20 at
New Bedford Harbor.
The total life-cycle cost for the extraction unit was not
developed from an explicit capital cost investment (equip-
ment list) or specific operating and maintenance cost
assessments. The developer's proprietary estimates were
used in combinations with cost-capacity curves and ratios
based on literature values and general experience. The
greatest uncertainty associated with this estimate is related
to the assumption of an 85-percent on-stream factor. The
reasons for this are:
• Sizing and costing equipment to handle five
times the capacity of a first commercial unit is
not expected to involve major uncertainties
because CF Systems utilizes industrial standard-
ized off-the-shelf equipment.
• An on-stream factor could not be measured dur-
ing the demonstration test at New Bedford with
the PCU-20 due to materials handling problems
associated with recycling processed feed. Recy-
cling is not a commercial design operation.
• Commercial operating data are not currently
available for the PCU-200, which has been
installed and is in a startup phase at a refinery in
Texas.
• If a commercial on-stream factor lower than 85-
percent results, then a larger system design for
tons/day would be required for the base case to
maintain the 8-year treatment time.
Asa means of accounting for the uncertainty in the on-
stream factor it is recommended that the cost range of plus
or minus 20 percent for a budget estimate be downgraded
to an order-of-magnitude estimate of plus 50 percent and
minus 30 percent as defined by the American Association
of Cost Engineers. This level of estimate is associated with
no preliminary design work using cost-capacity curves
and ratios. This results in an accuracy range of Sl04/ton to
$222/ton for the base case, and S317/ton to S67 I/ton for the
hot spot.
The developer's extraction unit design and capital
costs cannot be verified without a significant effort. Any
attempt to duplicate the proprietary design must include
provisions for the unit's modularity, the integration of
process components, safety-related instrumentation, pres-
sure relief system backups, and automatic shutdown.
However, some elements of the remedial design and esti-
27
-------
Table 4-2. Estimated Cost
Case
Description^
Total Waste Volume (Tons) 880,000
PCB Reduction (Percent) 91
Solids Content (Percent) 27
On-Stream Factor (Percent) 85
Remediation Duration (Weeks) 434
Site Preparation
Extraction Unit 3.02
Pre/Post Treatment 1.95
Excavation 21.44
43.94
Permitting and Regulatory
Equipment
Extraction Unit 48.39
Pre/Post Treatment 23.86
48.91
Startup and Fixed Costs 6.76
Labor
Extraction Unit 10.72
Pre/Post Treatment 10.80
Supply
Supplies and Consumables
Extraction Unit Utilities 17.06
Pre/Post Treatment Utilities 2.29
Effluent Treatment
Residual Transport
Analytical 1.98
Facility Modifications
Site Demobilization
TOTALS, $/ton 148.27
1A
Base Case
Factor
1B
Base Case
With Reduced
On-Stream
Reduction
1C
Base Case
Without Solids
Content
880,000 880,000
91 91
27 40
70 85
527 280
Estimated Cost, $/ton
2.96 3.02
1.95 1.95
26.03 13.97
58.52 31.53
28.98 15.50
8.21 4.39
13.02 6.97
13.11 7.01
19.51 11.08
2.78 1.48
2.41 1.29
177.48 98.19
1D
Base Case
With Increased
PCB Removal
Efficiency
880,000
98
27
85
347
5.86
1.58
17.14
77.81
19.08
5.40
10.86
9.01
23.91
1.83
1.59
174.09
2
Hot Spot
63,000
99.9
27
85
64
47.53
23.57
173.57
13.73
33.68
24.09
29.20
4.68
4.07
446.97
Notes: 1) A narrative description of the cases appears in the text.
2) These estimates are only intended for use in plannning, scoping, and the inviting of
firm bids. The American Association of Cost Engineers has established an accuracy goal
of plus 50 to minus 30 percent for preliminary estimates such as these.
3) The costs shown are based on a proprietary model developed by CF Systems, Inc. Cost
model outputs are presented in Appendix B for the Base Case and the Hotspot Case.
28
-------
mated costs, which are not directly related to the extraction
unit design, can be checked against available construction
cost data. Excavation, pre- and post-treatment equipment,
and labor costs are the most significant cost elements aside
from the extraction unit. Each of these elements is com-
pared below to costs reported in the literature:
• Excavation-The base case excavation cost is
$21.44/ton, which compares well with a$19.60/
ton cost reported for bulk excavation in a cof-
fer dam with a clamshell (Means, 1986).
• Pre- and Post-treatment~The base case cost is
$23.86/ton, and is composed primarily of unit
costs for leased equipment services. The unit
costs compare with costs reported in the litera-
ture (Means, 1986).
• Labor-The base case labor cost is $21.52/ton or
$43,635/week for 952 labor hours/week. These
costs are equivalent to an average labor rate of
$46/hour. This hourly rate is not unreasonable
since it is a composite rate that includes engineer-
ing and management, costs for safety gear, and
employee benefits and overhead.
4.5 Extrapolation of CF Systems' Sludge
Treatment Costs to Other Sites
A generic cost model was developed to provide a
method for end users to estimate the remediation costs at a
specific site for the CF Systems sludge treatment technol-
ogy. A total system cost consists of an extraction system
cost(E) and a pre-and post-treatment cost(P). A procedure
has been developed to estimate E as a function of: total
mass of sediments to be treated; total treatment time;
percent reduction in PCB solids content; and level of
dilution of raw feed by waterprior toextraction. A generic
method of estimating P is not provided because this cost is
highly site-specific.
Development of the extraction system cost involves
three steps:
• Defining the following elements: the basic PCU
processing unit to be used (PCU-50, P C U -
200, PCU-500, PCU-1000); the number of units
in parallel (NP); and the number of units in
series (NS).
• Estimating unit capital and operating costs for
one processing unit and then multiplying these
unit costs by NP plus NS to obtain the total ex-
traction cost.
- Select the total mass of solids to be treated (T)
and total treatment time (t).
• Calculating the total system throughput of raw
feed by dividing T by t.
- Calculate the total extration feed rate by ad-
justing the raw feed rate to account for water
dilution to condition the solids to the required
solids content required by the extraction sys-
tem.
- Select a PCU series module of given capacity.
- The number of parallel units is obtained by
dividing the total extraction feed rate by the
capacity of the PCU module selected.
- Select the fractional reduction in inlet PCB
concentration that is required.
- Some of the significant variables that affect
extrac lability performance are: size and na-
ture of solids; source and nature of organic
contamination; relativeconcentrationsofvola-
tiles and semivolatiles; age of feedstock; and
initial concentration. For purposes of scoping
costs, the reduction in concentration that can
be projected per module is a function of the
organic content of the feed: inlet concentra-
tion in excess of 10 percent, 99.5-percent re-
duction; for inlet concentration of 1 percent,
95-percent to 99-percent reduction; and for
inlet concentration less than 500 ppm, 95-
percent reduction. If greater percentage re-
moval of organics is required at correspond-
ing inlet concentrations, then additional
modules can be added in series or extra
stages can be added per module.
4.6 Conclusions and Recommendation
The cost estimates developed by CF Systems for the
base case (estuary) and hot spot case involve two key
assumptions: that capital and operating costs can be scaled
up to a commercial capacity based on pilot-scale testing of
the PCU-20; and that an 85-percent on-stream factor ap-
plies to a commercial unit. Assumptions regarding scale-
up of equipment costs are considered to be less critical
because CF Systems designs arc based on purchase of off-
the-shelf equipment, and field tests at New Bedford dem-
onstrated that outlet PCB concentrations of 50 ppm and 20
ppm were obtained using mixer/settler equipment. How-
ever, the number of modules required to meet a total
throughput capacity is dependent on the value of the on-
stream factor. CFSystemsmust demonstrate an on-stream
factor of 85 percent for a commercial operation in order to
reduce the uncertainty associated with projecting a cost of
S148/ton for treating a large mass of New Bedford Harbor
sediments using CF Systems technology. This will also
increase the confidence in using the generic model to
estimate costs for waste treatment at other sites. Based on
this discussion it is recommended that EPA verify the
credibility of the use of the 85-percent on-stream factor.
29
-------
Appendix A
Process Description
A.I Introduction
CF Systems technology uses a liquefied gas, such as
propane or carbon dioxide, as a solvent to extract organics
from soils, sludges, and waste waters. The solvent is mixed
with the waste, then the solvent-organics mixture, (after
extraction) which is not soluble in the solids and water, is
separated from the solids and water. The pressure of the
separated solvent-organics mixture is then reduced to
vaporize the solvent and separate it from the organics. The
solvent is then recovered and compressed to a liquid for
reuse. Separated organics are collected for disposal or use
in fuel blends.
CF Systems currently offers two treatment systems.
• The soils and sludge system uses liquefied pro-
pane to extract organics contained in a solid
waste using a series of mixer/settler units oper-
ated at pressures below the critical point of pro-
pane.
• The wastewater system uses liquefied carbon
dioxide to extract organics from a water stream
using a series of sieve extraction trays con-
tained in vertical column which is operated at a
pressure at or near the critical point of carbon
dioxide.
Combinations of the above systems could be used to
remove organics contained in both water and solids after
segregating the water and solids phases.
A.2 Process Design Sludge Extraction System
Process Description
CF Systems Pit Cleanup Unit (PCU), shown in Figure
A-l, is a continuous processing unit that used a liquified
propane/butane mix as the extraction solvent. The solvent
mix was 70 percent propane and 30 percent butane. For
each of the 3 demonstration tests, a batch of approximately
50 gallons of sediments was fed to the unitat a nominal rate
of 0.9 gpm. Feed viscosity was maintained below 1,000 cP,
by adding water in order to produce a pumpable slurry.
Particles greater than one-eighth inch were screened from
the feed to prevent damage to valves. Sediments were
pumped to the extractors, which were typically operated at
240 psig and 70 degrees F. Liquified solvent was also
pumped to the extractors at a rate of 2.3 gpm (10 Ib/min)
and mixed with the sediments. Organics, such as PCBs that
are soluble in the liquified solvent were extracted. After
extraction, treated sediments were decanted and separated
from the liquified solvent and organics mixture. The
mixture flowed from the extractor and passed to a separator
through a valve that partially reduced the pressure. The
pressure reduction caused the solvent to vaporize and
separate from the extracted organics. The solvent was
recycled and compressed to a liquid for reuse in the system.
The PCU-20 was not designed for large-scale reme-
dial actions. Therefore, treated sediments were recycled,
or passed through the unit to simulate operation of a
commercial-scale unit. CF Systems' commercial-scale de-
signs do not include recycling. These designs feature 60
gpm flowrates, several extraction stages, and longer proc-
essing times.
Equipment Specifications
The major pieces of equipment and their function are
described in Table A-1. Process equipment that contacted
the solvent or feed materials were constructed of 316
stainless steel. All process pumps were constructed of
stainless steel, and both compressors were made of carbon
steel. All of the process equipment was designed to
withstand temperatures and pressures that exceed normal
operating conditions. To guard against sudden overpres-
sure, each vessel had a relief valve that vented to a header
system that discharged to the pollution control system.
Table A-l outlines the major equipment items and the
function of each piece of equipment in the process.
The utility and process materials requirements that
were necessary to operate the PCU at New Bedford Harbor
were:
• Electricity~480 VAC 3 Phase, 100 amps
• Process Water-5 GPM, 60-80 degrees F
inlet, 30-90 psi
• Potable Water-Available
• Propane-four, 100 gallon bullets, 95-97
percent purity
• Butane-As needed, for Propane/Butane (70/
30) solvent mix
31
-------
to
CF SYSTEMS
Figure A-1. Pit Cleanup Unit.
-------
Table A-l. Process Equipment Description
Process Equipment
Feed Kettle
Basket Strainer
Extractor 1
Decanter 1
Designation
FK
S-1
E-1
D-1
Extractor 2
Decanter 2
Cartridge Filter
Solvent Recovery
Column
Column Reboiler
Treated Sediment
(Raffinate) Product Tank
Extract Product Tank
Main Compressor
E-2
D-2
F-2
SRC
CR
RPT-1
RPT-2
EPT
C-1
Low Pressure Solvent
Compressor
C-2
Function in System
Holds approximately 100 gallons of
strained, slurried feed. Counter-rotating
agitators homogenize feed.
Prevents oversized (>1/8 inch) feed
material from entering the system.
Extracts organics from water-solids feed
mixture with solvent from D-2.
Allows separation of solvent-organic
mixture from water-solids layer. Sends
water-solids layer to Extractor 2 (E-2)
and solvent-organics layer to the solvent
recovery system.
Extracts organics from water-solids
mixture with fresh propane from the
solvent recovery process.
Allows separation of solvent-organics
layer from water-solids mixture.
Filters residual solid fines from solvent-
organics stream leaving Decanter 1 (D-1).
Separates propane solvent from organics
via pressure reduction and heat from the
Column Reboiler (CR). Solvent vapor flows
out the overhead while organics are
deposited in the CR.
Provides both holdup for the recovered
organics and heat for the Solvent Recovery
Column (SRC) via a tube bundle heat
exchanger.
Receives treated sediments (raffinate)
from Decanter 2 (D-2). Recovers residual
propane via flash pressure reduction and
heat from water jacket. RPT-2 receives
RPT-1 overflow.
Receives extracted organics effluent from
the Column Reboiler (CR). Recovers
residual propane via flash pressure
reduction and heat from the water jacket.
Compresses both Low Pressure Solvent
Compressor (C-2) outlet solvent and
Solvent Recovery Column (SRC) overhead
solvent. Outlet sent to Column Reboiler
(CR) for heat exchange before returning
to Extractor 2 (E-2).
Compresses scavenged propane from Extract
and Raffinate Product Tanks (EPT, RPT-1,
and RPT-2). Sends compressed solvent to
Main Compressor (C-1).
33
-------
Table A-2. Range of Operating Conditions for Testing
Minimum
Extractor Pressure (PSIG) 180
Extractor Temp, (degrees F) 60
Feed Temp, (degrees F) 60
Solvent Flow (Ib/min) 8
Feed Flowrate (GPM) 0.2
Solvent/Feed Ratio 1
Feed Solids (percent by weight) 10
Solids Size (maximum)
pH (standard units) 6
Viscosity (cP) 0.5
Nominal
240
100-110
70
12
0.2-0.5
1.5
30
7
10
Maximum
300
120
100
15
1.5
2
60
1/8 inch
12
1,000
• Nitrogen (for pressure testing during shakedown
period)-(2) 1A size cylinders.
Utility usage for a commercial-scale unit cannot be
easily compared with the PCU because pilot-scale equip-
ment consumed much more energy per gallon of through-
put.
The operating conditions listed in Table A-2 are essen-
tial to the efficient operation of CF Systems' pilot-scale
unit. Failure to operate the unit within the specified
operating ranges can result in decreased extraction per-
formance. The operating parameters were set during the
shakedown portion of the demonstration. CF Systems
claimed that minor fluctuations would not affect perform-
ance.
The feed temperature is that of the material piped into
the feed kettle. The feed must be maintained above 60
degrees F to avoid freezing, which could interfere with the
extraction process. The feed must be maintained below
120 degrees F to prevent vaporization of the solvent.
The extractor pressure, measured at the gauges on
extractors 1 and 2, is controlled by the main compressor
and at the extract discharge from the extraction segment of
the unit.
The viscosity and solids content mustbe such that the
feed material is pumpable. Pretest sampling determines
the viscosity of the potential feed. Any potential feed with
a viscosity above the listed range is slurried with water to
yield a pumpable mixture.
In order to prevent damage to the process equipment,
the pilot-scale unit has a maximum limit for solids size.
Basket strainers, located between the feed pump and the
first extractor, prevent larger-than-allowable size solids
from entering the system. Oversized solids removed from
the feed were hauled to an RCRA-approved facility.
The feed flow rate represents the rate at which mate-
rial is pumped from the feed kettle into the extraction
system. Operational flow rates above the listed maximum
can force segments of the system, such as decanters and
control valves, beyond their effective hydraulic capacity.
The feed flow rate is manually controlled through the feed
pump controller located beneath the feed kettle. Average
detention time of throughput is about one hour.
Process Flow Diagram
The PCU process flow diagram is shown in Figure A-
2. The extraction portion of this unit consisted of two
stages of counter-current extraction with solid-liquid sepa-
ration between the extractors. The feed was transferred
from a feed preparation drum to the feed kettle with a
pump. In the feed kettle, slurry solids were kept suspended
while in the feed kettle by two counter-rotating agitators.
During this process, feed was pumped from the feed kettle
through a basket strainer, which removed any particles
greater than 1/8 inch in diameter. Then feed flowed to the
first extractor, where feed was mixed with the liquid
propane/butane solvent. An agitator (not shown in the
figure) provided mixing action before the solvent-organics
mixture flowed to decanter 1. At decanter 1, the mixture
separated into two immiscible layers. The solids and water
settled into the underflow to the second extractor. The
decanter overflow, which contained extracted organics,
propane/butane, and fine solids, flowed through a filter and
then to a solvent recovery column.
The pressure difference between the firstdecanter and
the second extractor moved the solid-liquid stream into the
34
-------
Legend
««*»«*»«•• Feed
•—••••« Propane-Butane Solvent
^—^— Propane/Organics Mixture
—— Extracted Organics
«-*»—" Processed Sediments
Figure A-2. CF Systems Process Schematic.
35
-------
second extractor for second-stage extraction. Fresh liq-
uified solvent (propane/butane mixture) from the solvent
recovery process then mixed with the solids/water stream
and further extracted the organic components. An agitator
(not shown in the figure), which was located above the
second extractor, provided mixing action before the sol-
vent-organics mixture flowed to decanter 2. At decanter 2,
two immiscible layers were formed. Theorganics-solvent
layer floated to the top while the solids sank into the
underlying water layer. The lower water-solids layer
flowed from the bottom of the decanter to the treated
sediment product tanks, while the upper organics-solvcnt
layer recycled to the first extractor for final stage extrac-
tion.
The organic-solvent stream from the first stage extrac-
tor passed through a filter cartridge that collected fine
solids and went to the solvent recovery column. In the
solvent recovery column, the solvent vaporized and was
removed from the column overhead, while the organics re-
mained as a separate liquid. The mixture of organics
containing dissolved propane gathered in the column re-
boiler and subsequently passed to the extract product tank.
Solvent from the column overhead flowed to the main
compressor. The compressed solvent passed through the
column reboiler heat exchanger to provide the heat neces-
sary to boil off residual solvent remaining in the organic
mixture. The condensed solvent left the reboiler and re-
entered the extraction system via the second extractor.
The residual solvent that vaporized off the system
products in the extract or the treated sediment tanks flowed
to the low-pressure solvent compressor. The outlet stream
of the low-pressure solvent compressor fed to the main
compressor, where it was compressed along with vapors
from the column overhead.
During system shutdown or if overpressure within a
vessel opens a relief valve, material is vented to a relief
header, which directs the material to a blowdown tank
where solids and liquids are removed from the vented
stream. The gases from the blowdown tank pass through
a 42-gallon activated carbon filter to remove contaminants
in the propane gas. The gas then passes through a flame
arrester and is vented to the atmosphere. This system was
used only once during the demonstration, at the conclusion
of PCU decontamination.
A.3 Process Design Waste Water Extraction System
Process Description
The CF Systems wastewater extraction process is a
solvent extraction technique which, instead of using a
typical solvent such as methylene chloride, toluene or
hexane, uses liquefied carbon dioxide (CO2) gas. This
solvent has high solubilities for most hazardous organic
compounds. In addition, CO2 is inexpensive, non-toxic,
and can be easily separated from the extracted compounds.
In contrast to the sludge unit the wastewater unit uses a
sieve tray extraction instead of mixer/settler extraction
units.
Figure A-3 provides a simplified flow diagram of the
CF Systems extraction process using liquid CO2 as a
solvent to extract organics from wastewater. As shown in
the figure, organic-bearing wastewater is continuously fed
into the top of the extractor and flows down the column
through a series of sieve tray downspouts. Simultaneously,
liquid CO2 is fed into the bottom of the extractor, and jets
upward through perforations in the sieve trays because
liquid CO2 has a lower density than water. During this
countercurrent contact between CO2 and wastewater, or-
ganics are dissolved out of the water phase to form a CCy
organics phase, or extract, which continuously exits from
the top of the extractor. As the extract stream flows from
the extractor to the separator vessel, it passes through a
pressure reducing valve that allows some of the CO2 to va-
porize and exit from the top of the separator. The CO2
vapor leaving the top of the separator vessel is continu-
ously fed to a compressor, recompressed, liquefied and
then reused as fresh solvent, resulting in a totally enclosed
recycle system. In the separator vessel, as CO2 changes
phase from liquid to vapor, the liquid organics are released
and flow to the bottom of the separator, where they are
collected and removed as a concentrated stream usually
containing less than five percent water. Both the concen-
trated organic stream and the water effluent from the
bottom of the extractor vessel are reduced in pressure prior
to being pumped off-skid.
• Equipment Specification
CF Systems Organics Extraction Unit Model
LL20CO-1 is designed such that the extraction
process will not be interrupted by component
failure. To accomplish this, each component has
design parameters appropriate to its function in
the process. Most components are designed for
use in environments with temperatures and pres-
sures of 350°Fand ISOpsig, respectively. Table
A-3 lists all major components and their func-
tions. Utility and process water requirements
are given in Table A-4.
• Process Flow Diagram
This section describes the functioning of the
major operating units shown in a process flow
diagram, Figure A-4.
• Extraction
In the extraction process, the wastewater feed is
pumped from a storage tank through a strainer,
a heat exchanger, and then to a surge drum. The
36
-------
Extract
Vapor CO2
Makeup
Compressor
Water
Figure A-3. Simplified Flow Diagram.
37
-------
Table A-3. Major Components and Functions
Component # Name
T-1 Extractor
T-2
D-1
D-2
D-3
D-4
D-5
D-6
D-7
D-9
D-ll
E-l
E-2
E-3
(continued)
Solvent Recovery
Column
Feed Drum
Medium Pressure
Extract Flash Drum
Low Pressure Extract
Flash Drum
Medium Pressure
Raffinate Flash Drum
water.
Low Pressure
Raffinate Flash Drum
C-2 Interstage
Knockout Drum
C-3 Interstage
Knockout Drum
C-1 Suction
Knockout Drum
CO2 Storage Drum
Column Reboiler
After condenser
Solvent Subcooler
Function
Provides contact between the water-
organics feed stream and the liquid
CO2 solvent.
Separates most of the CO2 solvent
from the organics in the extract
stream.
Intermediate surge for the wastewa
ter feed between the low pressure
pump (GP-102) and the high pres
sure feed pump (P-1).
Provides intermediate pressure
separation of CO2 solvent from ex
tracted organics and cooling of CO2
vapor discharged from the second
stage of the Low Pressure Compres
sor (C-3).
Provides low pressure separation of
CO2 from the organics.
Provides initial separation of CO2
fromextracted water.
Provides final separation of the CO2
from extracted water.
Provides removal of any organics
condensed from the CO2vapor going
to the second stage of the Medium
Pressure Compressor (C-2).
Provides removal of organics con
densed from CO2 vapor before en
tering the second stage of the Low
Pressure Compressor (C-3).
Provides removal of any liquid con
densed from the CO2 vapor from the
Solvent Recovery Column (T-2) and
the C-1 Recycle Cooler (E-5) before
the vapor enters the Main Compres-
sor (C-1).
Provides location for CO2 storage.
Heat exchanger used to transfer the
heat of vaporization to the liquid
CO2 in the Solvent Recovery Col-
umn (T-2).
Heat exchanger used to condense the
CO2 vapor/liquid mixture coming
from the Column Reboiler (E-l).
Cools the liquid CO2 from the
Aftercondenser (E-2)
38
-------
Table A-3. (Continued)
Component #
E-4
E-5
E-6
E-7
E-8
E-10
E-12
E-13
E-14
C-1
C-2
C-3
P-1
P-2
P-3
P-4
P-5
Name
Feed Precooler
C-1 Recycle Cooler
C-2 Intercooler
C-3 Intercooler
C-1 Lube Oil Cooler
Reflux Cooler
C-2 Recycle Cooler
C-3 Recycle Cooler
D-2 Reboiler
Main Compressor
Medium Pressure
Compressor
Extract Flash Drum (D-2).
Low Pressure
Feed Pump
Solvent Charge
Pump
Extract Pump
Raffinate Pump
C-1 Auxiliary
Lube Oil Pump
Function
Cools the incoming wastewater feed
from storage.
Cools the CO2 vapor recycled through
the Main Compressor (C-1).
Cools the hot compressed CO2 va-
porfrom thefirststageoftheMedium
Pressure Compressor (C-2) before it
enters the second stage of C-2.
Cools the hot compressed CO2 va-
por from the first stage of the Low
Pressure Compressor (C-3) before it
enters the second stage of C-3.
Cools the lube oil for the Main
Compressor (C-1).
Cools the liquid CO2 from the Solvent
Subcooler (E-3) before it enters the
Solvent Recovery Column (T-2) to
aid in CO,, vapor and organics sepa-
ration.
Cools the CO2 vapor recycled from
the Medium Pressure Compressor
(C-2) back into the Medium Pres-
sure Extract Flash Drum (D-2).
Cools the CO2 vapor recycled from
the Low Pressure Compressor (C-3),
back to the Low Pressure Extract
Flash Drum (D-3).
Supplies heat to the liquid in the
bottom of the Medium Pressure
Extract Flash Drum (D-2).
Compresses the CO2 vapor from the
Solvent Recovery Column (T-2).
Compresses the CO2 vapor collected
in the Medium Pressure
Compresses the CO2 vapor collected
in Compressor the Low Pressure
Extract Flash Drum (D-3).
High pressure pump taking wastewa
ter from the Feed Drum (D-1) and the
Extractor (T-1).
Pumps CO2 makeup into the Solvent
Recovery Column (T-2).
Takes suction on the Low Pressure
Extract Flash Drum (D-3) discharg-
ing to storage.
Takes suction on the Low Pressure
Raffinate Flash Drum (D-5) discharg-
ing extracted water to storage.
Provides initial lube oil pressure to
start Main Compressor (C-1).
39
-------
Table A-4. Utility and Process Water Requirements
Electric Power Criteria
460V, 3 phase, 60 Hz AC electric power is distributed to the following loads:
Component (item no.) Motor HP
Main Compressor (C-1) 125
Medium Pressure Compressor (C-2) 60
Low Pressure Compressor (C-3) 25
Feed Pump (P-1) 20
Extract Pump (P-3) 5
Raffinate Pump (P-4) 5
C-1 Auxiliary Lube Oil Pump (P-5) 1
Feed Mixer (MX-1) 0.5
Total Connected 241.5
Hot Water Criteria
Requirement Design Value
Supply Temperature 180°F
Supply Pressure By Client
Design Flow Rate 25 GPM
Return Temperature 150°F
Design Pressure Drop Across Skids 20 psi
Thermal Relief Valve Setpoint 100 psig
Individual component hot water requirements are listed below:
Individual Component Flow Rate (GP
Feed Drum (D-1) 13.3
Medium Pressure Extract Flash Drum (D-2) 9.1
Low Pressure Extract Flash Drum (D-3) 1.5
Total Hot Water 23.9
Refrigerated Water Criteria
Refrigerated water is supplied at the following conditions:
Design Value
Supply Temperature 55 °F
Supply Pressure By Client
Design Flow Rate 110 GPM
Return Temperature 70 °F
Pressure Drop Across Skid 30 psi
Thermal Relief Valve Setpoint 100 psig
Individual component refrigerated water requirements are as follows:
Refrigerated Water Requirements Design Value
Component Flow Rate
Aftercondenser (E-2) 30
Solvent Subcooler ((E-3) 30
Feed Precooler (E-4) 30
C-1 Recycle Cooler (E-5) 5
C-2 Inter and Recycle Coolers (E-6/E-12) 6
C-3 Inter and Recycle Coolers (E-7/E-13) 4
C-1 Lube Oil Cooler (E-8) 1
Reflux Cooler (E-10) 3
Total Refrigeration Requirement 109
40
-------
strainer removes any solid particles larger than
60 mesh to prevent these solids from entering the
high pressure feed pump and the extractor. The
resulting strained feed stream has no more than
two percent suspended solids. The Feed Pre-
cooler (E-4) cools the wastewater feed to a tem-
perature below 70°F. The jacket on the feed
drum heats the wastewater to above 65 °F. The
Feed Drum (D-l) provides five minutes of re-
serve feed for the system process.
From the feed drum, wastewater is pumped by
the Feed Pump (P-l) to the Extractor (T-l) at
about 950 psig. Wastewater enters at the top of
the extractor and flows downward countercur-
rent to the CO2 solvent stream. The CO2 solvent
enters at the bottom of the extractor and flows
up through the column as the dispersed phase.
The column internals consist of sieve trays and
downcomers. CO2 passes through the sieve
tray holes as it flows upward from stage to stage.
The aqueous solution, coming from the top,
flows across a sieve tray before passing through
a downcomer to the stage below. The arrange-
ment allows for optimum mixing and contact
between the fluids to accomplish extraction of
the organics into the CO2 solvent.
The extract stream from the top of the extractor
consists of liquid C02, organics extracted from
the feed stream, and a small quantity of dissolved
water. The organics concentration in the extract
stream is over 98 weight percent
on a CO2-free basis. The CO2-organics extract
stream flows through a pressure-reducing valve
and is partially flashed. By reducing the
stream pressure from 935 psia to 750 psia, about
15 percent of the CO2 vaporizes. During the
flash evaporation process, the liquid loses
sensible heat to vaporize the CO2 and is cooled
from 70 °F to 60 ° F. At the lower temperature,
the solubility of water in CO2 decreases and a
water phase is formed.
Pressure Letdown and Carbon Dioxide Distilla-
tion
An important feature of the CO2 distillation
and stripping is that the bulk of the CO2 in the
extract stream is separated from the product
organics at a pressure that is very near the
extractor pressure. This minimizes the
compression work needed to recycle the CO2
back to the extractor.
The CO2 separation is performed in the
Solvent Recovery Column (T-2).
Vaporization heat is supplied from two
sources. First, the hot compressed CO2
vapor discharged from the Main Compressor
(C-l) gives up its sensible heat of cooling and
latent heat of condensation as it cools from the
vapor phase and condenses into the liquid phase.
A kettle type heat exchanger, Column Reboiler,
(E-l) is provided perform the heat exchange.
Second, superheated CO2 from the second stage
discharge of the Medium Pressure Compressor
(C-2) is injected into T-2 for direct heat ex-
change.
The column reboiler kettle is equipped with a
boot for collecting the water layer formed in the
pressure letdown and CO2 distillation. The water
layer, which will contain some organics, collects
in the boot and is recycled back to the feed drum
for organics recovery.
Extract Stream Flash Evaporation
The overhead CO2 vapor from T-2 is fed to the
suction of compressor C-l. The still bottoms
stream contains about fifty percent of the CO2
originally present in the extract stream, is fed into
a cascaded flash evaporation stage for further
CO2 removal.
This organics-rich bottoms stream is first flashed
across a control valve to about 125 psia. This
stream flows into the Medium Pressure Extract
Flash Drum (D-2) where the vaporized CO2 is
vented to compressor C-2 for recompression.
The liquid is removed from D-2 on level control
and flashed to a final pressure of 20 psia in D-3,
the Low Pressure Extract Flash Drum, for nearly
complete removal of CO2. The CO2 vapor from
D-3 goes to the first stage of the Low Pressure
Compressor (C-3). The organics extract stream
is withdrawn from D-3 on level control and is
pumped to storage by the Extract Pump (P-3).
Some of the heat required to vaporize the CO2 in
drum D-2 is supplied by the superheated CO2
coming from the second stage of C-3, which is
injected directly into D-2. D-2 is equipped with
a heating coil, and D-3 with a heating jacket, to
heat the organics and to help remove the CO2.
CO2 vapor from the Raffinate Flash Drums (D-4
and D-5) is also injected directly into the extract
flash drums.
Raffinate Stream Flash Evaporation
The vapors from the Low Pressure Raffinate
Flash Drum (D-5) are combined with the vapors
from D-3 before recompression by C-3 and in-
jection into D-2. Similarly, the vapors from the
Medium Pressure Raffinate Flash Drum (D-4)
41
-------
Ratfinate Flash Drums
Extract Flash Drums
Figure A-4. CF Systems Organics Extraction Unit Simplified One-Line Diagram.
42
-------
are combined with the vapors from D'2 before
recompression by C-2 and injection into T-2.
The pressure in the raffinate flash drums is con-
trolled by back pressure controllers located in the
vapor outlet lines on the two extract flash drums.
Both D-4 and D-5 are equipped with level con-
trollers for maintaining proper liquid flow from
the drums. From D-5, the raffinate stream is
pumped to a storage tank by Raffinate Pump
(P-4).
Carbon Dioxide Recompression, Condensing and
Recycling
The last process involves recompressing, con-
densing, and recycling the CO2. The CO2 vapor
coming from D-3 is compressed from 18 psia to
128 psia by compressor, C-3. The CO2 vapor dis-
charged from C-3 is injected into drum D-2 for
direct heat exchange with the liquid in this drum.
The CO2 vapor coming from D-2 is compressed
from 124 psia to 753 psia by compressor, C-2.
The CO2 vapor discharged from the second stage
of C-2 is injected into the Solvent Recovery
Column (T-2) for direct heat exchange to help
vaporize CO2 and to cool the hot compressed CO2
stream. The overhead vapor from T-2 is fed to
compressor C-l for final recompression from
745 psia to a final pressure of 980 psia. Both two
stage compressors, C-2 and C-3, are equipped
with intercoolers and knockout (KO) drums.
The vapor flowing to the first stage of the com-
pressors contains a small amount of organic
vapor. When the hot stream from the first stage is
cooled, the organics condense. This liquid has to
be removed in a knockout drum before the vapor
goes to the compressor second stage. Liquid or-
ganics are removed from the knockout drums on
level control and are drained to their respective
extract flash drums.
From the main compressor (C-l) discharge, the
CO2 flows through three heat exchangers in se-
ries. In the heat exchangers, the CO2 is cooled,
condensed, and subcooled to 70 °F so that it can
be recycled to the extractor tower. The CO2 is
cooled and partially condensed by supplying the
heat of vaporization of CO2 in the kettle. The
subcooled liquid CO2 from the last exchanger
then flows to the Extractor (T-l).
43
-------
Appendix B
Developer (Vendor) Comments
1. Introduction
CF Systems Corporation (CF Systems) is a technol-
ogy-based company in the hazardous-waste-treatment and
resource-recovery business, offering services and equip-
ment based on a proprietary extraction technology. CF
Systems' solvent-extraction units are designed forremoval
of organics from soils, sludges, and aqueous streams,
concentrating the extracted organics for recovery or final
disposal. The result is the minimization of waste volumes,
reduction of treatment and disposal costs, and recovery of
materials such as oil products, solvents, and chemicals.
In August 1988, the U.S. EPA designated solvent
extraction as Best Demonstrated Available Technology
(BOAT) for petroleum refinery wastes (K048-K052).
Performance data from CF Systems were incorporated in
the evaluation used to set these standards for RCRA
refinery waste treatment.
CF Systems has demonstrated the effectiveness of its
extraction technology through the operation of its Mobile
Demonstration Unit (MDU) for nearly two years. To date,
the MDU has operated at eleven locations, including
refineries, chemical plants, and treatment, storage and
disposal (TSD) facilities in the United States and Canada.
In general terms, use of CF Systems' extraction tech-
nology provides important benefits for remediation and
treatment for land disposal:
• By-product credits for the recovered organics;
• Significant volume reduction of the treated sol-
ids;
• Effluent water acceptable for conventional was-
tewater treatment;
• An environmentally acceptable extraction sol-
vent with low residues.
2. Commercial Activity
CF Systems' major commercial activity has consisted
of the following:
• Texaco awarded CF Systems a 14 month reme-
diation contract to clean-up 20,000 cubic yards
of First-Third refinery wastes at its Port Arthur
Refinery in Texas. This is the first commercial
application of any type of solvent extraction
technology to treat hazardous waste in the petro-
leum industry. The Texaco project includes feed
pretreatment and material handling services, as
well as the solvent extraction system. CFS has
installed a full-scale PCU-200 solvent extrac-
tion unit at the site and start-up of the system is
scheduled for mid-June, 1989.
In November, 1988, Clean Harbors Inc. pur-
chased a commercial-scale LL-20 system to
process 20 gallons per minute (GPM) of organic
wastewaters at their Baltimore facility. The unit
will be shipped for start-up in July, 1989.
Clean Harbors is a rapidly-growing company in
the commercial waste treatment, storage, and
disposal (TSD) business, with multiple loca-
tions nationally. The LL-20 unit will treat a wide
range of organic wastewaters to produce dis-
posable water and an organics fraction generally
suitable for fuel use.
A custom 60 ton-per-day system was purchased
by ENSCO's El Dorado, Arkansas incinerator
facility. The unit is designed to extract organic
liquids from a broad range of hazardous-waste
feeds sent to the site for incineration. The treated
solids produced by the system will be processed
by the incinerator at a much faster rate than un-
treated solids due to their reduced fuel value and
the extracted liquids can be used as incinerator
secondary combustion fuel.
The 20 barrel-per-day mobile unit (MDU) has
been in operation for test and demonstration
purposes under client funding since September,
1987. Operating experience includes six petro-
leum refineries, U.S. and Canadian; aU.S. chemi-
cal plant; a Canadian TSD site; two Superfund
sites, a PCB clean-up under this EPA SITE
program sponsorship, and the other a
woodtreating waste impoundment. In April,
1988, a commercial clean-up job was performed
at a major chemical company in New Jersey.
45
-------
3. The Technology
Critical-fluid solvents, the basis for the CF Systems'
technology, are condensed gases or supercritical fluids,
such as carbon dioxide, freon, propane, ethylene, ammonia
and others, in the vicinity of their critical points. Above the
critical point, the transition from gas to liquid is continuous
rather than abrupt. At or near such conditions, fluids have
very favorable solventproperties. They behave like liquids
in that they are capable of dissolving significant amounts
of oil or other substances. They behave like gases in that the
rates of extraction are extraordinarily high compared to
liquid solvents.
In the CF Systems process, a liquid feed such as an
organic-containing hazardous waste is admitted to an ex-
tractor, along with the solvent. At or near the solvent
critical point (usually ambient temperature and several
hundred psi), the organics in the waste dissolve into the
solvent. The two phases are separated, extracted organics
are removed with the solvent, while clean water and solids
are removed through an underflow. The extract then goes
to a second vessel, where the temperature and pressure are
decreased, causing the organics to separate from the sol-
vent. Clean solvent is recycled to the extractor, and
concentrated organics are recovered and removed.
Examples of organic pollutants that can be extracted
economically using the CF Systems unit include a wide
range of aromatic and aliphatic hydrocarbons, chlorinated
hydrocarbons, phenols, alcohols, ketones, ethers, and
organic acids. The CF Systems technology can be applied
to sludges and solids as well as liquid wastes.
4. CF Systems Equipment Systems
CF Systems has developed a series of standard modu-
lar equipment systems. For sludge and solids treatment, the
capacity range is about 10-500 tons per day; for liquids,
about 5-30 gallons per minute (GPM). Treatment systems
are assembled as skid-mounted modules to facilitate ship-
ping and field assembly. Production of standard modules
also allows high quality, low-cost fabrication.
4.1 The PCU-Series
The PCU-series systems are designed to process high-
solids sludge feeds and contaminated solids such as soils.
They contain specially-designed extractors and separators
to facilitate the treatment of oily solids typical of petroleum
sludges and waste materials found in refinery impound-
ments requiring remediation. Organics removal can be as
high as 99.9% or better. At this time thePCU series design
has been economically optimized to meet the projected
requirements for remediation of both refinery and Super-
fund applications. Where high levels of clean up are
required the economics of the technology will favor opera-
tion of the systems at higher temperature and pressures.
The systems included in this product series are:
• PCU-50 - This system, designed to process a
maximum of about 12 tons per day, is a standard
product for refinery sludges regulated by EPA's
RCRA land-disposal ban, impoundment sludges,
and oil- and PCB-contaminated soils and silts.
The system is skid-mounted and designed for
installation into confined spaces and ready inte-
gration into existing operations.
• PCU-200 - This system, designed to process a
maximum of about 50 tons per day, is a larger-
scale product for oily sludges and contaminated
soils. The system is mounted on two trailers,
and can be mobilized and demobilized in 10-15
days.
• PCU-500 - The PCU-500 is a modified PCU-200
design, with the same solvent-recovery subsys-
tem, but increased extractor capacity to pro-
vide for throughputs up to about 100 tons per
day. Depending on location and cleanout re-
quirement, mobilization-demobilization may
require 4-8 weeks.
• PCU-1000 - This system, with a 200 ton-per-day
nominal capacity, is intended for large remedia-
tion jobs and relatively long terms (one year or
more) at a single location. It is skid-mounted and
transportable, but with multiple modules, requir-
ing 2-3 months for mobilization and demobili-
zation.
4.2 TheLL-Series
The LL-series systems are designed for the extraction
of dissolved or emulsified organics in water streams.
Solids are usually not present at a significant level in these
streams, or must be removed to the 2-3% level as a
pretreatment. Organics content of the feed can range as
highas 30-50% andremovalefficiencies can exceed99.9%.
The market for the LL-series includes a wide range of
organics wastewaters.
The systems are skid-mounted and transportable;
however, the extractor is a column which is field-erected.
In contrast to carbon steel in the PCU systems, stainless
steel is required for this series because of the corrosion
potential of the feeds.
46
-------
5. The Product and the Application
5.1 Applications
The technology is applicable to any solid or liquid
feedstock which contains organics water. Depending on
the feedstock type and organics various solvent systems
are available to meet the product specifications.
For organics contaminated waste water the solvent of
choice is carbon dioxide which enables most semi soluble
organics to be extracted from the water. Even highly
soluble organics such as alcohols may be extracted by
correct design of the extraction system.
For sludges where petroleum hydrocarbons or chlo-
rinated hydrocarbons are present together with solids plus
water, propane is the solvent of choice.
Finally in these circumstances where non-flammable
solvents are required to be used for extraction of organics
from sludges environmentally safe chloro-floro-carbons
may be used.
6. Application/Market Characteristics
The Company is positioned in the segments of the
hazardous waste treatment market where removal of or-
ganic material from liquids, soils, and sludges is required.
Among the benefits to the user are:
• reduction of wastes to small residual volumes
suitable for land disposal;
• elimination of the legal and financial liability of
off-site disposal in many cases;
• recovery of organic material with value as a
product or fuel; and
• a cost-effective alternative to the next best tech-
nical option, incineration.
A major additional advantage is the absence of RCRA
permitting requirements in most markets. Under RCRA,
treatment systems usually operate under permits to ensure
that the treatment itself will not represent a hazard. Incin-
eration, for example, requires permitting and the concomi-
tant requirement for public hearings is delaying incinera-
tion capacity by 3-5 years.
CF Systems' permit exemptions fall under three cate-
gories:
• the recycling exclusion, where a useful by-prod-
uct is produced (e.g., petroleum refineries);
• wastewater pretreatment exemption, where the
CF Systems unit is a pretreater to final wastewa-
ter treatment, as in chemical plants;
• the "totally enclosed" treatment system exclu-
sion, for contained units such as the Company's
products.
6.1 Superfund Sites
The EPA's inventory of potentially hazardous sites
throughout the United States has been stated as greater than
25,000 [2] in number. By the spring of 1989, about 900 of
these sites had progressed through the evaluation stage to
the point where they were on the National Priorities List
(NPL) and subject to enforcement action under CERCL A.
In another report, the General Accounting Office estimated
the universe of potential hazardous waste sites at some-
where between 130,000 and 425,000 [3]. Whichever
figure is used, it is clear that substantial resources will be
dedicated to the clean up of old hazardous waste disposal
sites for some time to come.
Using refinery experience as a reference, it is esti-
mated that the average Superfund site might contain 50,000-
100,000 tons of material and that about 150 applicable sites
will be remediated by 1993. On this basis, the Superfund
tonnage treated in the period 1989-1993 might be 7.5 to 15
million tons of material.
6.2 Petroleum Refining Wastes
Of the approximately 180 petroleum refinery sites in
the United States, about 100 are active and 80 have shut
down. Canada has about 50 refineries (active plus inac-
tive).
Refineries have two primary categories of waste treat-
able by this technology:
• oily sludges produced from current operation,
(ongoing wastes) such as API separator sludges
(40 CFR 261.32 K048-K052).
• oily sludges and solids from past operations
stored in pits, ponds, and lagoons (surface im-
poundments).
Ongoing wastes are subject to the RCRA land ban in
August, 1990. While some refineries may gain further
delays, current environmental control activity clearly indi-
cates that many refineries are preparing to have treatment
capacity in place by that time.
The average active refinery in the U.S. is estimated to
produce 3-5,000 tons per year of listed hazardous wastes
(predominantly API separator sludge) subject to the 1990
land ban; as well as additional wastes listed by states such
as California, and contaminated soils, which will total 2.5
million tons over five years.
47
-------
Closure plans for impoundments will be implemented
in the refining industry over the next 10 years at active as
well as inactive sites. Most refineries have pits or surface
impoundments containing waste sludges generated in the
past. In the U.S. and Canada, it is estimated that the total
of these impoundments exceed 10 million tons.
6.3 Incinerator Pretreatment
Less than 10 RCRA-permitted incinerators for de-
struction of solid hazardous wastes presently exist in the
United States. The demand for such incinerators far out-
strips the supply, but are slim because of public opposition
to permitting, the prospects for closure of that gap soon.
Thus, any means for increasing the throughput of existing
incinerators is of obvious value.
A significant fraction of solid hazardous wastes con-
tains organic liquids that can be extracted with a CF
Systems unit. This extraction, as a pretreatment, produces
an incinerator feed with reduced heat content (BTU per
pound of feed). Incinerator capacity is limited by the heat
that can be removed (BTU), not the pounds of feed flowing
through. Thus, a controlled evolution of the heat content
allows more pounds of hazardous waste to be destroyed.
Moreover, the extracted liquid fraction can be used as fuel
in the so-called secondary burn, which would otherwise
require purchased fuel. As a result, the incinerator operator
gets a double benefit from CF Systems' pretreatment.
7. Remediation Experience Histories
CF Systems has generated process and equipment
design information for applications in the hazardous waste
treatment and remediation industry. A substantial data-
base has evolved for a wide range of extractions from
sponsored research conducted at our bench-scale and pilot-
plant facilities. Continued growth of that database and
process correlations is on-going in our research facilities.
As noted earlier, CF Systems' extraction technology
has been successfully demonstrated in the field. The 20
barrel-per-day MDU successfully started-up in 1987 at the
Texaco refinery in Port Arthur, Texas. Since then, the unit
has processed a wide variety of wastes at several refineries,
chemical plants, and TSD facilities in North America.
These include:
Texaco, Port Arthur, Texas
Chevron, Salt Lake City, Utah
Chevron, Perth Amboy, New Jersey
BASF, Kearny, New Jersey
Petro-Canada, Montreal, Canada
Tricil, Toronto, Canada
New Bedford Harbor Superfund NPL Site,
New Bedford, Massachusetts
Exxon, Baton Rouge, Louisiana
Unocal, Parachute Creek, Colorado
United Creosote Superfund NPL Site,
Conroe, Texas
1. Texaco.Port Arthur. October. 1987-Januarv. 1988
The MDU had its initial operation at Texaco's
Port Arthur refinery in late 1987. A range of
different feed types were run through the system,
including spent oily clay, primary separator
sludge, and tank bottoms. The resulting treated
solids product streams were analyzed by Texaco,
and representative results are shown in Table
7-1. Performance consistently met what later
became BDAT standards for RCRA first- third
(K048-K052) refinery wastes. The results of this
demonstration led Texaco to award CF Systems
a contract to provide a commercial unit to
remediate 20,000 cubic yards of primary separa-
tor sludges.
2. BASF. Kearnv. New Jersey
A mobile treatment system was run at the BASF
Kearny, New Jersey, plantsite. Oneof the waste
streams from this plant is an emulsified stream
containing di-octyl phthalate (DOP), water, and
other organic materials. The system success-
fully separated the emulsion into a recoverable
DOP stream and a wastewater suitable for dis-
charge to the wastewater treatment facility.
3. Petro-Canada. Montreal
The MDU operated at Petro- Canada's Montreal
refinery for a six-week period. During this time,
the unit successfully processed 14 different feed
types ranging from API separator sludges to
contaminated soils. The unit achieved organic
removal levels better than existing BDAT stan-
dards. In some cases, the levels of residual
organics, both volatile and semivolatile, were
better than those obtained with incineration.
4. Tricil. Toronto. Canada
A series of demonstration tests were run at Tricil
Canada's TSD facility in Missasauga, Ontario.
The system de-oiled a majority of the organic
feed materials arriving at this facility. The wastes
processed included API separator sludge, paint
wastes, synthetic rubber process waste, and coal
tar wastes.
48
-------
Table 7.1. Texaco Port Arthur Performance Data.
Boat
Levels
(mg/Kg)
Water (WT. %)
Solids (WT. %)
Oil (WT. %)
Total Oil
& Grease
(WT. %, Dry)
Benzene
Ethylbenzene
Toluene
Xylenes
Fluorene
Naphthalene
9.5
67
9.5
Reserved
Reserved
Ditch Skimmer(LAB) Clay Pit Area (MDU)
Feed Treated TCLP Feed Treated Water
Solids Solids
(mG/Kg) (mg/Kg) (mg/L) (mg/Kg) (mg/Kg) (mg/L)
60.5
22.3
3.1 17.2
5.1
13
52
71
50
0
0.
0
0
0
0
.052
.06
.13
.44
.59
.1
<0.
<0
,0005
.001
0.0027
<0,
0,
.003
.0005
2-Methyl Naphthalene
Phenanthrene
Ohromium
Lead
7.7
20
400
1100
0
560
1300
.16
0,
0.
31
,0015
.02
1.9
9.6 <0.1
13 <0.1
16 <0.1
63 <0.1
210 <5.3
300 <5.3
<0.01
<0.01
<0.01
<0.01
SITE 127 Sludge (MDU) SITE 143 Sludge (MDU) Ditch Skimmer (MDU)
Feed Treated TCLP Feed Treated Feed Treated TCLP
Solids Solids Solids
(mg/Kg)( mg/Kg) (mg/L) (mg/Kg) (mg/Kg) (mg/Kg) (mg/Kg) (mg/L)
62 57 53
32
6
<2.0
<2.0
<2.0
<2.0
<50
31
3.6
<2.0
<3.3
<3.3
<0.01
<0.01
<0.01
<0.01
33
10
13.7
20.2
54.4
75.9
45
30
35
12
1.0
<2.0 5.1
<0.1 13
<0.1 52
<0.1 71
9.3
<3.3 1 6.5
<3.3 18.6
400
1100
0.7
<0.1 <0.01
<0.1 <0.01
<0.1 <0.01
<0.1 <0.01
<0.20
<0.20
<0.20
560 0.02
1300 31
-------
Two specific Tricil requirements were achieved:
• a large volume reduction of the wastes proc-
essed;
• reduction of the level of volatile organics such
that land disposal of the residual solids was
acceptable.
5. New Bedford Harbor Superfund NPL Site. New
Bedford. Massachusetts
CF Systems participated in the EPA Superfund
Innovative Technology Evaluation (SITE) pro-
gram at New Bedford Harbor in Massachusetts,
a location which is heavily contaminated with
PCB's. Data obtained during the program indi-
cated that it is feasible to obtain PCB removal
to levels in excess of 99.9% at economic costs.
6. Unocal. Parachute Creek. Colorado
The MDU completed a series of demonstrations
at Unocal's Parachute Creek, Colorado facility.
Among the wastes successfully run were
samples of shale-oil wastes, drilling muds, and
other process and refinery wastes.
High recovery of good-quality oil was obtained
from shale-oil wastes. Drilling mud wastes were
treated to the standards required for land dis-
posal.
7. United Creosote Superfund NPL Site.
Conroe. Texas
The MDU completed a treatability study for the
Texas Water Commission in conjunction with
Roy F. Weston at a Superfund Site in Conroe,
Texas. The objective of this study was to evalu-
ate the effectiveness of solvent extraction for
remediation of soil contaminated with creosote.
PAH concentractions in the soil obtained from
the capped area were reduced from 2879 ppm to
122 ppm, demonstrating 95+% reductions were
possible. Representative results from this study
are shown in Table 7-2.
Table 7.2. Conroe Performance Data
FEED
DRY SOIL (MG/KG)
COMPOUND
Acenapthene
Acenaphthylene
Anthracene
Benzo(A)anthracene
Benzo(A)pyrene
Benzo(B)fluoranthene
Benzo(GHI)perylene
Benzo(K)fluoranthene
Chrysene
Dibenzo(A,H)anthracene
Fluoranthene
Fluorene
lndeno(1,2,3-CD)pyrene
Naphthalene
Phenanthrene
Pyrene
Total Pah Cone. (MG/KG)
360
15
330
100
48
51
20
50
110
ND
360
380
19
140
590
360
2879
RAFFINATE
DRY SOIL (MG/KG)
3.4
3.0
8.9
7.9
12
9.7
12
17
9.1
4.3
11
3.8
11
1.5
13
11
122.6
50
-------
8. Remediation Services
In general, remediation projects encompass excava-
tion, treatment, and removal of contaminated soils and
sludges (Figure B-3). Depending upon the types of con-
tamination and the level of cleanup required, further proc-
essing downstream of CF Systems' extraction system may
be necessary. This further processing may include fixation
for heavy metals and incineration of the extracted organics.
A typical remediation project may consist of the fol-
lowing steps or any combination of these steps:
1. The soil is excavated and/or the slurry is dredged.
2. If necessary, the excavated material is slurried
with water to create a pumpable mixture.
3. The slurry is passed through a mul tilayered shaker
screen to remove material larger than 1/8-inch
diameter. Oversized material may be crushed
and recycled to the screens.
4. The pH of the screened slurry is monitored and,
if required, lime is added to the mixture to main-
tain a pH between 6 and 8.
5. The slurry may require thickening prior to the
slurry being pumped to the CF Systems Extrac-
tion Unit.
6. Two product streams exit the Extraction Unit; a
solids/water stream and a liquid organic stream.
The organic stream will generally be returned to
the Client for reuse or disposal.
7. The solids/water stream is dewatered through
the use of a belt filter press or a centrifuge. The
water from the dewatering step may be used to
slurry dry feed solids. Any excess water is clean
enough to be disposed of in domestic sewers or
in a waste water treatment plant.
8. The dewatered solids may require chemical
fixation, if there are significant quantities of
leachable solids, such as heavy metals, so that the
treated solids may be disposed of in a non-
hazardous landfill.
9. The treated solids must then be transported to
and disposed of in a landfill or other suitable site.
8.1 Excavation and Feed Pretreatment
8.1.1 Dry Soils
Contaminated dry soils will be excavated through the
use of equipment such as front end loaders, backhoes, or
bulldozers. These soils will then be fed into a preliminary
screening device to remove any materials larger than 4
inches in diameter. Solids captured in the screens will be
collected, washed, and disposed of in an appropriate manner.
Screened material will be transported on a conveyor
belt to a pug mill where size reduction is effected. The pug
mill will combine the dry solids with water to produce a
solids/water extricate. This paste will then travel via a
second conveyor belt to a tank or pump where additional
water will be added to produce slurried solids.
8.1.2 Sludges
A diesel engine powered, auger head dredge will
slurry the waste sludges for pumping with water either
presenter added. Water addition is required in areas where
the waste is partially solidified. If water addition is
necessary, enough water will be added to float the dredge,
creating a pond area that the dredging operation will
expand.
8.1.3 Solids Screening and Thickening
Slurried solids from either the pug mill or the dredge
will pass through a multilayered shaker screen similar to
those used in the oil drilling industry. The objective will be
to screen out solids larger than 1/8 inch in diameter. Solids
captured by the screen will be collected, washed, and
recycled to the pug mill or crusher/grinder for size reduc-
tion.
Sludge passing through the screen will be collected in
a storage tank equipped with mixers. If required, lime will
be added at this point to maintain a pH between 6 and 8.
The slurry will then be pumped from this tank to either the
extraction unit or to a thickener.
If pumped to a thickener, the slurry will be thickened
to approximately 50-percent solids. This is accomplished
through the use of either a moving screen or a decantation
system depending on the water solubility of the waste.
Water extracted by the thickener will be returned to the
dredge area or to another approved discharge point. The
thickened solids slurry will be pumped to another holding
tank and then fed to the Extraction Unit.
8.2 Product Disposal
The de-oiled solids and water produced from the ex-
traction process will be dewatered. This stream will be run
through a belt filter press where a combination of pressure
and conditioning flocculents, if required, will remove
excess water, leaving a cake with approximately 40- to 45-
percent solids. Water separated from the slurry will be
returned to the dredge area or to the water treatment
system. De-oiled solids in the form of a cake will move via
conveyer from the belt filter press to a small blending mill.
9. Cost Estimate for a Specific Superfund
Application
Cost estimates provided here are CF Systems' stan-
dard budget estimates quoted to commercial customers for
use in planning, scoping, and inviting of firm bids. The
51
-------
estimates' accuracy basis is +/- 20% of the expected final
quotations given the same basis and assumptions. Cf
Systems would utilize subcontractors to do portions of the
described work, specifically solids handling before and
after the key extraction step. However, CF Systems is
willing to provide the services under a user/project man-
ager or under a prime contractor/project manager.
9.1 New Bedford Harbor, Massachusetts Clean Up
Case Study
CF Systems has prepared cost estimates for two cases
to provide solvent extraction technology to cleanup PCB-
contaminated silt at New Bedford Harbor, Massachusetts.
The following comprises the Scope of Work and the basis
for the cost estimate for the two cases.
9.1.1 Estuary Case Description
The quantity of material to be cleaned in Case 2 is
695,000 cubic yeards of PCB-contaminated soil. This
quantity of material represents removal and treatment of all
contaminated soil in the New Bedford Harbor estuary. The
level of PCBs in this material is assumed to average 580
ppm on a dry solids weight basis. The PCBs in this material
will be reduced to 50 ppm via solvent extraction. The time
schedule for processing this material is about five years.
9.1.2 Case 1: Hot Spot Case Description
The quantity of material to be cleaned in Case 1 is
50,000 cubic yards of PCB-contaminated soil. This quan-
tity of material represents removal and treatment of the hot
spots in New Bedford Harbor. The level of PCBs in this
material is assumed to be 10,000 ppm on a dry solids
weight basis. The PCBs in this material will be reduced to
10 ppm on a dry solids weight basis via solvent extraction
technology. This represents a 99.9% removal of PCBs.
The time schedule for processing this quantity of material
is approximately one year.
9.2 Scope of Work
The scope of work for both cases described above is as
follows:
l.The PCB-contaminated material has been re-
moved from New Bedford Harbor and stock-
piled, by others, at an appropriate site on land.
2.CF Systems will move the contaminated soils
from the stockpile to its processing site, using
typical heavy-duty earth moving equipmentsuch
as backhoes and bulldozers.
3.CF systems will screen the material to remove
oversize particles. Solids larger than 1/8" will be
retained on the shaker screens, then be sent to a
crusher/grinder for size reduction. Solids that
cannot be reduced in size will be rejected and
returned to the client, who will be responsible for
their disposal.
4. CF Systems will slurry and pretreat the screened
material before it is processed in the solvent
extraction system.
5. CF System will process the pretreated feed in its
solvent extraction units to remove the PCBs from
the material.
6. Exiting the extraction system will be two product
streams: a PCB-rich extracted organics stream
and a PCB-free solids/water stream. The PCB
stream will be returned to the client for disposal.
CF Systems will dewater the solids/water stream
and stockpile the solids for disposal by the
client.
9.3 Basis for Costing
CF Systems will provide the solventextraction system
for the cleanup of New Bedford Harbor. All auxiliary and
support equipment required will be supplied by CF Sys-
tems through subcontractors. It should be noted that the pre
and post treatment equipment will be assumed to operate
only 10 hours per day. Sufficient storage capacity both in
the front and back will be assumed so that the extraction
system can operate uninterrupted, 24 hours per day. The
current cost estimate for the two cases assumed 1989 costs
for the subcontracted services. A breakdown of the sub-
contracted items and the cost basis for these items are as
follows:
Solids Handling Equipment:
The solids handling equipment is provided to move
the PCB contaminated material from the stockpile to the
CF Systems treatment site. Equipment required for this
operation is assumed to include:
One(l)Frontend Loader S 500/day
One(l)D6 Bulldozer Sl.lOO/day
Two(2)25Ton Trucks $2,000/day
Safety equipment for the operators is included in the
above costs. The above equipment is required for the
duration of the job and is assumed to be operating 10 hours
per day.
Solids/Sludge Delivery Equipment:
The solids/sludge delivery equipment is provided for
size reduction and delivery of the solids to the feed pretreat-
ment system. Equipment required for this operation is
assumed to include:
One(l) Frontend Loader $ 500/day
52
-------
Two (2) Pug Mills $l,000/day
One (1) Crusher/Grinder $ 750/day
Four (4) Conveyors $ 200/day
Safety Equipment $ 80/day
The above equipment is required for the duration of
the job and is assumed to be operating 10 hours per day.
Feed Pretreatment Equipment:
The feed pretreatment equipment is provided to screen
and slurry the feed prior to the solvent extraction system.
Equipment required for this operation is assumed to in-
clude:
Two (2) Shaker Screens $ 40/day
Two(2) Clarifiers $ 74/day
Two (2) Feed Pumps $ 52/day
Two (2) Mud Tanks $ 90/day
Two (2) Fractionating Tanks $ 200/day
The above equipment is required for the duration of
the job.
Product Handling Equipment:
The product handling equipment is provided to re-
ceive the product streams from the extraction system and
deliver these product(s) to the client for disposal. Equip-
ment required for this operation is assumed to include:
Two (2) Filter Systems $ 800/day
Two (2) Liquid PCB Storage
& Transfer System $ 410/day
Two (2) Solids/Water Storage
& Transfer Systems $ 160/day
Two (2) Conveyors $ 30/day
One (1) Front-End Load $ 500/day
One (1) D6 Bulldozer $l,100/day
The above equipment is required for the duration of the job.
Facilities:
The following site facilities are provided to support
the site personnel and equipment.
One (1) Sanitary/Office Trailer $ 80/day
One (1) Laboratory Trailer $ 50/day
Site Security $ 300/day
Analytical Services $ 500/day
Two (2) Electrical Generator Sets $ 600/day
Two (2) Packaged Cooling Towers $ 200/day
Safety Clothing for Personnel
(per man cost) $ 40/day
Utilities:
It is estimated that all the equipment on-site (extrac-
tion systems and auxiliaries) will consume approximately
2750 kwh/hr giving a cost of $3,960 per day for electrical
consumption at $0.06 per kwh for the estuary base case.
Labor:
The following labor and current (1989) rates for super-
vising and operating the various operations have been
included in the cost estimate:
Supervisors for CF Systems
Extraction System $ 720/day
CF Systems' Extraction
System Operators $l,800/day
Pre/Post Treatment Operators $ 600/day
Site Engineer $ 300/day
Site Manager $ 400/day
Other Labor $ 200/day
Safety Equipment for Above Personnel $ 720/day
9.4 Actual Cost Estimates
The specific costs for the two cases are tabulated in
Table B-l. CF Systems utilized its proprietary in-house
cost model and generated costs for each of the steps listed
in the scopeof work. The extraction only related costs were
broken out and tabulated according to the 12 cost elements
defined by the EPA. Pre- and post-treatment costs involv-
ing most of the rental equipment for solids handling were
lumped together and some details provided on a confiden-
tial basis to allow total system analysis. The contingency
and project management fees are self-explanatory.
53
-------
Table B-l. CF Systems Budget Cost Estimates
Facilities (Including utilities)
Facilities
CFS Extraction Costs
Site Preparation(1)
PCU Capital Charges
Labor
Utilities
Analytical
Total Extraction Costs
Pre/Post Treatment
Site Preparation(1)
Excavation/Solids Handling
Solids Delivery
Feed Pretreatment
Product Handling/Post Treatment
Utilities
Labor
Total Other Costs
Total Job Costs
Contingency (10%)
Project Management (5%)
Base Case
Estuary
$ 5,170,676
$
2,307,849
37,027,058
8,202,600
$ 13,053,273
$ 1,519,000
$ 62,109,781
$ 1,495,200
$ 16,405,200
$ 9,326,660
$ 1,974,700
$ 6,957,020
$ 1,749,888
8,263,360
$ 46,172,028
$113,452,485
$ 11,345,248
S 5,672,624
$
Hot Spot
$ 762,496
$ 2,616,261
$ 9,555,141
$ 1,854,720
$ 1,607,573
$ 224,000
$15,587,695
Overall Budget Cost of Remediation S130,470,358
(1) Includes mobilization, startup and demobilization
$ 1,297,800
$ 2,419,200
1,375,360
291,200
1,025,920
258,048
$ 1,326,080
$ 7,993,608
$24,613,799
S 2,461,380
S 1,230,690
S28.305.869
9.5 Description of Extraction System as Costed
9.5.1 Estuary Case 2
For this case, which involves a large tonnage removal
for multiple years on-site, CF Systems recommends the use
of a custom made PCU-2000 system which will process
about 500 tons/day.
The total time on-site will be 8.35 years to remove the
PCBs in 695,000 cubic yards of waste from 580 ppm to the
50 ppm level (91.4% removal efficiency).
9.5.2 Hot Spot Case 1
For this case, CF Systems recommends the use of 4
identical modular systems called PCU-500s which would
complete the remediation in about 1.2 years. These are
approximately 125 ton/day units each having its own set of
extraction stages and a solvent recovery section.
The selection of this size unit and system configura-
tion is to minimize total time on-site and total job cost. Two
units in series are required to meet the required efficiency
(99.9%) i.e., PCB removal from 10,000 ppm to 10 ppm.
Two sets in parallel are required to handle the total volu-
metric throughput.
54
-------
Appendix C
SITE Demonstration Results
Introduction
Sediments were dredged from five New Bedford
Harbor locations and stored in 55-gallon drums for proc-
essing by the PCU-20. Drummed sediments were sieved
to remove particles greater than one-eighth inch that could
damage system valves. Water was also added to produce
a pumpable slurry. The drummed sediments were blended
to provide feedstocks for four tests.
Test 1 was a system shakedown run to set flow rates
and operating pressures and to provide samples for labora-
tory evaluation of sample matrices. Samples were col-
lected during Tests 2,3, and 4 to provide data for evaluating
the system's performance. A fifth test was run with toluene
used as a feedstock for decontaminating the PCU. About
1 to 2 hours were required to run a feedstock through the
PCU. Test 2 involved passing, or recycling, the feedstock
10 times. Test 3 involved 3 passes and Test 4 involved 6
passes. Recycling was conducted to simulate the design
operation of a full-scale commercial system. The PCU is
only a two-stage system, whereas commercial designs
include four or more stages, longer extractor residence
times, and longer phase separation times. Conditions that
varied for each test were:
1 .Test 1 was run as a shakedown test to set pressure
and flowrates in the PCU. The feed was a 50-
gallon composite of sediments taken from drum
numbers H-20,H-21, and H-23. The feed
had a PCB concentration of 360 p p m .
Three passes were run to gain experience with
materials handling.
2.Test 2 was a 10 pass lest. The feed was a 350
ppm, 511 pound composite of sediments taken
from drum numbers H-20, H-21, and H-23. Ten
passes were run to simulate a high-efficiency
process and to achieve treated sediment levels
less than 10 ppm. A 350 ppm concentration was
chosen for this test since this represents an aver-
age, or typical, PCB concentration in the harbor.
3.Test 3 was a 3 pass test. The feed was a 288 ppm,
508-pound composite of sediments taken from
drum numbers H-20, H-21, and H-23. The pur-
pose of this test was to reproduce the results of
the first three passes of Test 2.
4.Test 4 was a 6 pass test. The feed was a 2,575
ppm, 299-pound composite of sediments taken
from drum numbers 1-11 and H-22. The pur-
pose of this test was to reduce a high-level waste
to a lower level waste such as that used in Tests
1, 2, and 3. High-level wastes are found at
several "hot spots" in the harbor.
Decontamination of the system involved running tolu-
ene through the PCU as a solvent wash. Samples were
taken of the feed at the commencement of each test.
Treated sediment products and extracts were planned for
sampling at each pass. Additional samples were taken of
system filters and strainers, although the amount of PCB
contained in these miscellaneous samples later proved to
be small. PCU operating pressures, temperatures, and
flow-rates were monitored throughout the tests. Field tests
were conducted for feed viscosity, pH, and temperature.
Results
A large amount of analytical and operating data was
obtained, and it was sufficient to meet the program objec-
tives. The detailed results and operating summaries are in
the Technology Evaluation Report. The objectives indi-
cated an evaluation of (1) the unit's performance, (2)
system operating conditions, (3) health and safety consid-
erations, and (4) equipment and system material handling
problems.
System Performance
The evaluation criteria established for system per-
formance were:
• PCB concentration in sediments before and after
treatment
• PCB extraction efficiency with each pass of
sediments through the PCU
• Mass balances established for total mass, solids,
and PCBs.
55
-------
These criteria are discussed with respect to analytical
results below.
PCB Concentration Reductions
PCB analyses for feed sediments and treated sedi-
ment, conducted for samples collected at each pass, are
shown in Table C-1. The data are displayed graphically in
Figures C.I, C.2, and C.3. The data show that treated
sediment concentrations of 8 ppm are achievable and that
as much as 84 percent of the PCB contained in sediment
can be removed in a single pass. In Test 2, feed containing
350 ppm of PCB was reduced to 8 ppm after 9 passes
through the PCU. In Test 3, a 288 ppm feed was reduced
to 47 ppm after just one pass. In Test 4, a 2,575 ppm feed
was reduced to 200 ppm after 6 passes. The percent
reductions in PCB concentration, based on a comparison of
untreated feed to the final pass, for each test were:
Percent Reduction
Test in PCB Concentration
2 89%
3 72%
4 92%
Number of
Passes
10
3
The data for each test show general reduction trends
based on differences between initial feed and final treated
sediment concentrations. However, these trends are not
consistent on a pass-by-pass basis. For example, PCB
concentrations in treated sediments increase at Test 2,
passes 4 and 10, and at Test 3, passes 2 and 3. These
anomolies are not related to the extraction process. In-
stead, they reflect cross contamination within system hard-
ware or limited analytical precision and accuracy. Since
the treated sediment collection tanks were under pressure,
it was not possible to clean out collection hardware and
piping. Therefore, a pass-by-pass mass balance could not
be established.
Data for each test can be used to construct a curve that
shows the potential number of passes required to reduce
PCB s in harbor sediments to specific concentrations using
the Pit Cleanup Unit (PCU). If data from Test 2, 3, and 4
are displayed side-by-side, such that similar concentra-
tions coincide, then a PCB reduction curve can be plotted.
Data are displayed below, side-by-side, so thatsimilar
concentrations overlap.
Table C-l. Pass-by-Pass PCB Concentrations and Reduction Efficiencies
Test
Number
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
4
4
4
4
4
4
4
Pass
_Number
Feed
1
2
3
4
5
6
7
8
9
10
Feed
1
2
3
Feed
1
2
3
4
5
6
PCB Concentration*
350
77
52
20
66
59
41
36
29
8
40
288
47
72
82
2,575
1000
990
670
325
240
200
Pass-by-Pass Concentration
Reduction Efficiency
(percent
Not Applicable
78
32
62
No Reduction
11
31
12
19
72
No Reduction
Not Applicable
84
No Reduction
No Reduction
Not Applicable
61
1
32
52
26
17
*PCB data represent feed and treated sediment concentrations.
56
-------
«00
Test 2 PCB Reduction
I
B?
300
J
3.a
3.e
5.*
j.j
2
i.e
1.6
1.4
1.3
I
0.8
O.S
0.«
0.3
0
246
Extraction Past Na.
Figure C-1.
Test 3 PCS Reduction
Extraction POM No.
Figure C-2.
Tat « PC8 Reduction
Citmcllon POM No.
Figure C-3.
57
-------
Test 4
2,575
1,000
990
670
325
240
200
Pass-by-Pass PCB Concentrations
Test 3 Test 2
288 350
47 77
72 52
82 20
66
59
41
36
29
8
40
Based on the presentation of the data in Figure C.4, it
can be construed that harbor sediments containing 2,500
ppm of PCB could be reduced to 100 ppm after 6 passes
through the PCU. A level less than 10 ppm may be
achievable after 13 passes.
Extraction Efficiency
Pass-by-pass PCB concentration extraction efficien-
cies are shown in Table C-l and are calculated as PCB
extracted divided by concentration at the beginning of the
pass (multiplied by 100 percent). For each test, the first
pass results in efficiencies greater than 60 percent. How-
ever, at later passes efficiencies range from negative
values to 72 percent. This wide range is the result of cross-
contamination of solids retained in the treated sediment
subsystem.
Data show that the system irregularly retained and
discharged treated sediments. For some passes, as much as
50 percentof the feed was retained in the system. That feed
was treated sediment that clung to internal piping and tank
surfaces. If discharged with a later pass, the combined
discharge could have a higher concentration than feed for
the later pass. For example, assume an extraction effi-
ciency of 60 percent, a feed concentration of 350 ppm, and
a carry-over of solids from the first pass to the second pass
of 25 percent. Then, the treated sediment would contain
77 ppm, instead of 56 ppm if no cross contamination
occurred.
The occurrence of cross contamination affects inter-
pretation of each test, but it does not invalidate the fact that
treated sediment concentrations as low as 8 ppm were
produced. Furthermore, the decontamination procedure,
showed that PCB, which accumulated in system hardware,
was contained in the extract subsystem, not the treated
sediment subsystem.
Mass Balances
Total mass, total solids, and total mass of PCBs were
determined for various system inputs and outputs for the
purpose of establishing a mass balance. Figure C.5 depicts
the inventory sheet used to account for system input and
output. Input included feed material and water, although
some feed was lost to sampling, sieving, spills, and residu-
als remaining on the surface of the feed drums. Outputs
from the system included samples, spills, container residu-
als, treated sediment, and residue collected on the basket
strainer and cartridge filter. The difference between input
and output resulted in either accumulations within the
system or unaccounted-for discharges of accumulated
material from the system. Mass inventories were devel-
oped for each test.
PCB Balance
Table C-2 illustrates the fate of PCB on a pass-by-pass
inventory basis. The system accumulated 15.15 grams
during Test 2,6.71 grams during Test 3, and 42.11 grams
during Test 4. Only an approximate PCB balance is
possible for Test 1, since Test 1 was a shakedown test only.
Approximately 21 grams of PCB accumulated within the
system during Test 1. Thus, total accumulation within the
system from Test 1 through Test 4 was about 85 grams
(where 84.96 = 15.14 + 6.71 + 42.11 + 21).
The fuel wash, which occurred immediately after the
first pass of Test 3, flushed 35 grams of PCB from the
extract subsystem. Final system decontamination with
toluene wash delivered an additional 151 grams. Total
wash output was 35 plus 151, or 186 grams. Ideally, the
amount of PCB washed from the system should equal
amount accumulated, or
Accumulation - Wash = 0
However, in this case,
85 grams - 186 grams = 101 grams
The amount of PCB washed from the system is shown
above to be greater than the amount fed, which raises the
possibility that (1) sampling and analytical errors occurred,
or (2) the system was contaminated from a previous CF
Systems demonstration.
Quality control samples collected during the demon-
stration indicate the possibility of sampling and analytical
error. For example, laboratory precision and accuracy
criteria were 20 and 50 relative percent difference, respec-
tively. In addition, quadruplicate grab samples were col-
lected of the Test 3 feed, the Test 4 feed, and the Test 3
58
-------
Table C-2. Mass Accumulation and Loss in the System
Test
2
2
2
2
2
2
2
2
2
2
3
3
3
4
4
4
4
4
4
1
2
3
4
5
6
7
8
9
10
Subtotal
1
2
3
Subtotal
1
2
3
4
5
6
Subtotal
TOTAL
Accumulation (Loss) in the System
Total Mass Total Solids Total PCBs
(Pounds) (Pounds) (Grams)
122 39 14.21
55 6 0.70
(25) (16) 0.50
78 32 (0.22)
22 (6) (0.07)
68 3 0.3
(51) (1) 0.04
(7) (11) (0.07)
(16) (3) 0.29
& !21 ffi£41
254 40 15.14
24 (13) 6.28
58 6 1.42
22 .£ (0.99)
111 1 6.71
5 10 37.79
(83) (12) (5.25)
74 9 8.72
(80) 4 2.55
106 6 1.63
(53) JS1 (3.33)
(31) 14 42.11
Note: Parentheses indicate a loss or discharge from the system.
59
-------
6
a.
a.
c
« 3
-
PCB Concentration Reduction Model
All Tests Combined
Extraction Pass No,
+ Test 3
10 11
Test 2
12
13
U
Figure C-4. Potential Pit Cleanup Unit PCB Reduction.
60
-------
Inventory Sheet
Test Pass
1. Feed Material
6. Water
2. Sampling •
3. Strainer
4. Spills
5. Residuals
8.Treated Sediment
Accumulations and Other Losses
Figure C-5. Illustrative Inventory Sheet.
61
-------
treated sediment and the RPD calculated for each set
ranged from 12 to 47 percent. In particular, the Test 4 feed
had a mean concentration of 2,575 ppm, which dominates
all other measurements used in the balance, and it had an
RPD of 22. Another possible source of the PCB imbalance
was contamination of the PCU from prior use at another
site. CF Systems did not decontaminate the unit with
toluene prior to this demonstration. CF Systems' standard
operating procedures now incorporate decontamination
with toluene.
In spite of the calculated PCB imbalance, a positive
separation of PCB from the harbor sediments was accom-
plished. The mass balances that 81 grams of PCB were
contained in sediments fed to the PCU in Tests 2,3, and 4.
Resulting treated sediments contained 4 grams of PCB,
which indicates a mass removal efficiency of 95 percent.
Decontamination residue data show that some PCB accu-
mulated in system hardware. However, 91 percent of the
PCBs contained in decontamination residues were con-
tained in the extract subsystem. The remaining 9 percent
was contained in the treated sediment subsystem hardware.
Basket Strainer, Cartridge Filter, and Carbon
Canister
The basket strainer and cartridge filter, which gener-
ate residuals that are normally discarded as a waste stream
separate from extract and raffinate, did not accumulate a
significant PCB mass. The mass balances, shown in
Appendix A, show that the accumulation was approxi-
mately 2 percent of the PCB mass fed to the system. When
compared to PCB removals of 90 percent, this indicates
that PCB removal by the basket strainer was not signifi-
cant. In addition, chemical analysis of the PCB content of
filtered solids indicate that the concentration of filtered
solids associated with each pass roughly correlated with
the treated sediments from the previous pass.
Low pressure propane/butane was vented through the
PCU carbon canister at the conclusion of the decontamina-
tion procedures. The 285 pounds of activated carbon
contained in the canister collected less than 1 gramofPCB.
This indicates that air emissions are not significant and
PCBs are separated from the solvent when expanded in the
PCU.
Total Mass of Solids
The PCU retained and discharged feed material inter-
mittantly throughout the tests. This behavior is demon-
strated by tracking the sediment solids. The mass of solids
accumulated on a pass-by-pass basis is significant. The
flow of solids perpass ranges from 55 percent accumulated
to 150 percent discharged. There is no consistent correla-
tion between solids retention and PCB concentration re-
duction.
During Tests 2,3,4, and 5 the system accumulated 302
pounds total mass and 53 pounds total solids. Total mass
accumulation represents approximately 4 percent of total
mass fed to system during Tests 2 through 5, and total
solids accumulation represents about 7 percent of total
solids fed to the system.
A total of 3-1/2 tons of solids and water were fed to the
unit over the course of 19 passes throughout Test 2,3, and
4. Of the total, 96 percent was accounted for in the system
outputs. Of 789 pounds of solids fed to the system, 93
percent was accounted for in system outputs.
Other Data
Semivolatile Organics
System feed, final treated sediment, and final extract
were sampled for base/neutral and acid extractable organ-
ics (semivolatiles) during each test for the purpose of (1)
characterizing materials for disposal and (2) observing any
extraction of semivolatiles. Interpretation of the semivola-
tiles data, shown in Volume II, is limited for two reasons:
(1) the unit contained sludges from a previous demonstra-
tion at a petroleum refinery, and (2) a naphtha-based fuel
product was added to the unit during Test 3 to clean out the
still, extract product tank and related hardware. The
following conclusions can be drawn:
• Semivolatiles detected in the toluene wash were
also detected in the feed drums, the source being
New Bedford Harbor sludge.
• Phenol and2-methylphenol were found in treated
sediments and extracts but were not measured in
feed drums, the feed kettle, or toluene washes.
• Test4 resulted in a reduction of 1,4-dichloroben-
zene and pyrene, but chrysene and bis(2-eth-
ylhexyl phthalate) were increased. Similar in-
consistencies occur for Test 2 and 3.
• 2-ChIorophenol, 1,3-dichlorobenzene, and
benzo(k)fluoranthene were fed to the unit but
not detected in any system effluents.
Fate of Metals
A firm conclusion cannot be drawn concerning the
fate of metals after each test, since the unit tends to
accumulate solids. However, the data in Table C-3 show
that treated sediments metals concentrations generally
equal or exceed feed metals concentrations. The data also
show that metals were not extracted and discharged in the
organics effluent. Metals concentrations in organic ex-
tracts were one to two orders of magn itude less than treated
sediments.
62
-------
Table C-3. Metals Content of Feed, Treated Sediment, and Extract
Parameter Units
Cadmium, ppm
Chromium, ppm
Copper, ppm
Lead, ppm
Zinc, ppm
Total Residue, %
Test 2
Feed
35.7
596
1790
619
2150
23.3
Test 2
Pass 3
Treated
Sediment
32.5
581
1650
587
2220
18.2
Test 2
Pass 4
Feed
44.0
761
1990
792
2680
15.0
Test 2
Pass 10
Treated
Sediment .
42.8
816
1740
892
2610
9.4
Test 2
Pass 10
Extract
NR(1)
3
5(2)
NR(1)
5(2)
NR(3)
Test3
Feed
32.0
525
1320
520
1900
19.4
Test3
Pass 3
Treated
Sediment
62.3
1020
2570
1100
3550
10.3
Test3
Pass 3
Extract
6(2)
20
6(2)
NR(1)
8(2)
NR(3)
Test 4
Feed
87.5
1480
2650
1300
5370
16.4
Test 4
Pass 6
Treated
Sediment
120.0
1790
3700
1800
7260
5.6
Test 4
Pass 4
Extract
5
26
5
35
15
NR(3)
Test 4
Pass 6
Extract
5
31
4
40
15
NR(3)
Notes: 1. Not reported, severe matrix effects.
2. Matrix effects indicated.
3. Not reported, insufficient sample volume for analysis method.
-------
EP Toxicity
RCRA regulations at 40 CFR 261.24 specify test
methods for determining if a solid waste exhibits the
characteristic of EP (extraction procedure) toxicity. The
maximum concentration of contaminants for the character-
istic of EP Toxicity is shown in Table C-4. Also shown are
analytical results for (1) two samples taken from a com-
posite of drummed harbor sediment collected by COE
during the waste presampling and (2) a sample of demon-
stration Test4, Pass 6 treated sediment. Concentrations for
each sample shown are less than the regulatory maximum
for the definition of the EP toxicity characteristic.
Feed and Extraction Temperature
Feed and extraction temperatures were stable for Tests
3 and 4. Feed temperatures ranged between 60 and 70
degrees F while extraction temperatures ranged between
60 and 80 degrees F. However, data for Test 2 indicate that
feed temperatures fell about 15 degrees F below the mini-
mum specification after pass 5. This caused extraction
temperatures to drop, with pass 9 falling 4 degrees F below
the minimum specification, 60 degrees F.
The developer attributes much of the fluctuating ex-
traction efficiencies calculated for Test 2 to the low feed
Table C-4. EP Toxicity Characteristics of Treated and Untreated Sediments
Units (Parts Per Million)
Sample 1
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
0.011
0.16
0.11
0.18
0.34
<0.0002
<0.005
<0.015
Composite Sample of
Waste Presampling Drums
Sample 2
0.008
0.15
0.12
0.098
0.23
<0.0002
<0.005
<0.015
Treated
Sediment
Test 4. Pass 6
<0.005
0.36
0.30
0.053
0.16
<0.0002
<0.02
0.015
Maximum Concentration
Allowable for
Characteristics of
EP Toxicitv
5.0
100.0
1.0
5.0
5.0
0.2
1.0
5.0
Note: < indicates detected less than the detection limit shown.
Operating Conditions
The system specifications that CF Systems requires
for normal operation were discussed in Section 3. In this
section, observed operating conditions are summarized
and operating data are interpreted with respect to treatment
efficiency. In tables throughout this section, mean operat-
ing data are shown as well as the range of data recorded for
each mean value. Generally, the technology accommo-
dated wide ranges of operating conditions, although pre-
cise operational control was limited since all controls were
manual rather than automatic.
Extraction Pressure
Pressures in both extractors used in the system were
fairly stable for all tests. Pressure levels were close to the
nominal level of 240 psig. The maximum pressure, 285
psig, was below the 300 psig maximum specification. The
minimum pressure, 190 psig, was above the 180 psig
minimum specification. Because pressures were so stable,
no relationship between extraction efficiency and extractor
pressure was apparent.
temperatures, although other factors were probably impor-
tant. These factors include cross contamination in the
treated sediments collection tank. In addition, reentrain-
ment of solvent in decanter underflows may have caused
disproportionately large effects on low concentration sedi-
ments. Each factor must be addressed by the developer in
the design of a full-scale system.
Feed Flow Rate
The feed flow rate ranged consistently, throughoutthe
tests, from 0.6 to 1.4 gpm. This range compares well with
the 0.2 gpm minimum specification and the 1.5 gpm
maximum specification.
Solvent Flow Rate
The solvent flow fluctuated outside the minimum
specification, 8 Ib/min, and the maximum specification, 15
Ib/min throughout Tests 2,3, and 4. Because of this wide
64
-------
variation, it was suspected the flow meter was malfunc-
tioning. In Test 4, an alternative measuring device was
used and flow measurements continued to show wide
variations.
The variable solvent flows caused the solvent/feed
ratio also to fluctuate widely. This ratio was calculated as
solvent (lb/min)/feed (gpm)/feed density (Ib/gal). The
minimum solvent-to-feed ratio specification, 1.0, was not
met on Pass 2 of Test 4 based on mean data. Individual
readings frequently exceeded the 1.0 to 2.0 specification
range. A pass-by-pass comparison of solvent/feed ratios to
extraction efficiencies was attempted but no direct corre-
lation or trend was apparent.
Nonetheless, it is believed that the solvent/feed ratio is
a significant factor in process design since the solubility of
an organic in liquified propane-butane is the fundamental
basis for the extraction. With higher solvent/feed ratios,
the feed is exposed to a larger amount of solvent and
extraction efficiency should increase. However, these
relationships were not observed, given the available data.
Feed Solids
Feed solids content steadily declined during each test.
Initial feeds had solids contents ranging from 15 to 22
percent. Final treated sediments ranged from 6 to 11
percent solids. This change is primarily a result of water
added to the feed kettle by operating personnel, during
each pass. This unnecessary practice caused waste vol-
umes to increase by 33 percent over the course of the
demonstration program. Another, but less significant,
factor that affected solids content was accumulation of
solids in system hardware. The solids mass balance
showed that 7 percent of the solids accumulated in the
system and were not washed out during decontamination.
Treated sediments that were fed to the unit after Pass
3 of each test, had solids contents below the minimim
specification, lOpercent. This dilution of the feed material
is believed to affect system performance.
Viscosity and pH
Feed viscosity and pH fell within specifications and
did not affect system performance. Viscosities for un-
treated feed and recycled sediments ranged from 20 to 170
centipoise, well below the 1,000 centipoise maximum
specification. This specification was set by the developer
only to ensure that the feed would be pumpable. Untreated
and recycled sediments had pH values that ranged between
7.3 and8.5 standard units. This narrow band fell within the
6 to 12 specification range. The developer established this
range to prevent corrosion to PCU hardware.
Health and Safety Monitoring
During the demonstration of CF Systems' process
unit, personnel were potentially exposed to the contami-
nated harbor sediments. A monitoring program was con-
ducted to determine potential exposures and provide a
basis for selection of proper personal protective equip-
ment. Several types of portable monitoring equipment
were used during the various phases of the field investiga-
tions, including:
• Portable Organic Vapor Analyzer (Century OVA)
• Portable Photoionization Meter (HNu)
• Combustible gas/oxygen/hydrogen sulfide me-
ters (MSA and Enmet-Tritector)
• Detector tubes and sampling pump (Sensidyne-
Gastec)
• Personal air sampling pumps (Dupont-P200).
It was suspected that some level of organic vapors
would be encountered, particularly when drums contain-
ing contaminated sediments were first opened during the
feed preparation phase. Continuous monitoring using both
the OVA and HNu instruments was conducted while the
drums were being opened. These instruments detected a
slight elevation above background levels of organic vapor
immediately upon opening the drums. The levels returned
to background levels within a few seconds. No measurable
levels of hydrogen sulfide or combustible gas were en-
countered while opening the drums or handling the sedi-
ments during the feed preparation phase.
During the various test runs of the extraction unit at the
New Bedford site, organic vapors, PCBs, combustible
gases, and hydrogen sulfide were monitored. The OVA
and HNu meters were used to monitor for organic vapors
at all work stations on the extraction unit, while CF
Systems and SITE personnel monitored process equip-
ment. The OVA also was used as a survey meter on the
process equipment to search for possible fugitive emis-
sions from the equipment. All measurements indicated
that organic vapor levels remained in the range of back-
ground levels. Two portable combustible gas meters were
used to check for elevated levels of propane during the
equipment shakedown period and for spot testing during
the demonstration. The pilot unit also contained two
integral combustible gas detectors located on either end of
the unit. During the normal extraction process, combus-
tible gas readings remained at background levels. How-
ever, while treated sediment and extract samples were
collected, the combustible gas meters indicated that levels
exceeding only 20 percent of the lower explosive limit for
propane were encountered. These episodes of elevated
propane levels generally lasted for less than 60 seconds and
65
-------
subsided rapidly depending on the length of time sampling
occurred and the strength of the wind at the time.
Sampling was conducted using personal sampling
pumps and 150-mg charcoal tubes and florosil tubes to
determine personal exposures to organic vapors and PCBs,
respectively. All air sample results indicated that, if
present, organic vapors and PCB levels were present only
at levels below the detection limits for the analytical
methods. No measurable levels of hydrogen sulfide were
detected using either detector tubes or portable monitoring
devices.
Treated sediment and extract subsystems were decon-
taminated with toluene. The final concentration of PCB
contained in the treated sediment subsystem toluene wash
was 34 ppm, which was below the decontamination goal of
50 ppm. The final concentration of PCB contained in
extract subsystem toluene wash was 60 ppm .which slightly
exceeded the decontamination goal of 50 ppm. Staging
area soils were not affected by any leaks or emissions that
may have occurred during the duration of the demonstra-
tion.
Equipment and Material Handling Problems
Equipment and material handling problems occurred
throughout the demonstration. While these problems did
not impede achievement of the developer's treatment
goals, they could impact the economic performance of a
full-scale commercial system. Some problems were an-
ticipated since relatively small volumes of sediments were
batch-fed to a unit that was designed for continuous opera-
tion. The nominal capacity of the unit is 700 gallons per
day, but only 50 to 100 gallons per day were batch-fed
during shakedown on tests 2,3, and 4. Consequently, the
unit irregularly discharged and retained solids with each
pass.
Previous use of the unit affected interpretation of
semivolatiles data and may have contributed to imbalance
of the PCB inventory. Internal surfaces of extract collec-
tion hardware collected PCBs as evidenced by mass bal-
ances. In addition, Test 3 was interrupted and viscous oils
were found accumulating in extract subsystem hardware.
PCBs are soluble in oil, which coated the internal surfaces
of system hardware. The amount of oil that can coat
internal piping and collection tanks could be significant.
For example, assume (1) a hardware surface area of 10
square meters, (2) a coating thickness of 0.1 millimeters,
and (3) an oil density of 1.0 grams/cubic centimeter. This
is equivalent to 100 grams of oils that cling to the internal
surfaces of extract subsystem hardware. As a result of this
demonstration, CF Systems now requires more rigorous
decontamination procedures for the PCU.
Solids were observed in extract samples that were
expected to be solids-free. This indicates poor perform-
ance or failure of the cartridge filter. An alternative type of
filter should be investigated by the developer.
Low-pressure dissolved propane and butane caused
foaming to occur in the treated sediment product tanks.
This hindered sample collection and caused frequent over-
flow of treated sediment to a secondary treated sediment
product tank. CF Systems states that design of a commer-
cial-scale unit will allow release of solvent entrained in the
treated sediment and elimination of the foaming problem.
t, U.S. Government Printing Office: 1990 — 751-267
66
-------
Environmental Protection
Agency
Information
Cincinnati OH 45268
ouurv MM 11
POSTAGE & FEES PAID
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 D,
detach, or copy this cover, and return to the address in the
upper left-hand corner
EPA/540/A5-90/002
------- |