v>EPA
United States
Environmental Protection
Agency
EPA/540/S5-90/002
August 1990
SUPERFUND INNOVATIVE
TECHNOLOGY EVALUATION
Technology Demonstration
Summary
CF Systems Organics Extraction
System, New Bedford Harbor,
Massachusetts
The Site Program demonstration of
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 pilot dredging
studies managed by the U.S. Army
Corps of Engineers at the New
Bedford Harbor Superfund site in
Massachuetts. Contaminated sed-
iments were treated by CF Systems'
Pit Cleanup Unit (PCU) that used
liquified propane/butane as the
extraction solvent. The PCU was a
trailer-mounted system with a design
capacity of 1.5 gpm (20 bbl/day). CF
Systems claimed that the PCU would
extract organics from contaminated
soils based on solubility of organics
in liquified propane/butane.
The objectives included an
evaluation of (1) the unit's
performance, (2) system operating
conditions, (3) health and safety
conditions, and (4) equipment and
system materials handling problems.
Extensive sampling and analyses
were performed showing that
polychlorinated biphenyl (PCS)
extraction efficiencies of 90 percent
were achieved for sediments
containing PCBs ranging from 350 to
2,575 ppm. In Test 2, sediments
containing 350 ppm were reduced to
40 ppm after 10 passes, or recycles,
through the PCU. In Test 3, a 288 ppm
feed was reduced to 82 ppm after 3
passes. In Test 4, a 2,575 ppm feed
was reduced to 200 ppm after 6
passes. Some operating problems
occurred, such as the intermittent
retention of solids in system
hardware and foaming in the treated
sediment collection tanks. These
problems did not affect extraction
efficiency but could affect operation
of a full-scale unit. Corrective
measures will be addressed by the
developer and EPA. A mass balance
established over the entire
demonstration showed excellent
accountability for 96 percent of the
total mass. Operation of the unit did
not present any threats to the health
and safety of the operators or the
local community.
This Summary was developed by
EPA's Risk Reduction Engineering
Laboratory, Cincinnati, OH, to
announce key findings of the SITE
program demonstration that is fully
documented in two separate reports
(see ordering information at back).
-------
Introduction
CF Systems Corp., developer of an
organlcs extraction technology, was
selected to demonstrate their "system at
the New Bedford Harbor, Massachusetts,
Suporfund site. The system demon-
strated was CF Systems' Pit Cleanup
Unit (PCU), a trailer-mounted system with
a design capacity of 1.5 gpm (20
bbl/day). Successful application of the
technology depends on the ability of
organic pollutants to sclubilize in the
process solvent, a liquified gas. The
process used a mixture of liquified pro-
pane and butane, at 240 psi and 69
degrees F, as a solvent for extracting
organics from soils. As liquified solvent
was mixed with the waste, organics were
extracted into the solvent. The sblvent-
organics mixture was then decanted from
the separated solids and water. The
pressure of the solvent-organics mixture
was reduced slightly to vaporize the
solvent which allowed separation from the
organics. The solvent was recovered by
the system and compressed to a liquid
for reuse.
The site is located on the Acushnet
River Estuary north of Buzzard's Bay in
the city of New Bedford, Massachusetts,
where sediments contain pollutants
discharged to the harbor from a variety of
industrial sources. The pollutants include
polychlorinated biphenyls (RGBs), poly-
nuclear aromatic hydrocarbons, copper,
chromium, zinc, and lead. PCBs present
the greatest toxic threat and concentra-
tions range up to 30,000 ppm.
The following technical criteria were
used to evaluate the effectiveness of the
CF Systems process for extracting PCBs
from New Bedford Harbor sediments:
1. System Performance
» Evaluate PCB concentration in
sediments before and after
treatment.
Evaluate PCB extraction efficiency
with each pass, or recycle, of
sediments through the unit.
* Evaluate mass balances established
for total mass, solids and PCBs.
2. Operating Conditions
Compare operating conditions to
operating specifications for flow,
temperature, pressure, and physical
sediment characteristics of the
sediment and assess the effect on
extraction rate.
3. Health and Safety Considerations
Determine if significant amounts of
propane/butane or PCBs are emitted
to the air by the process.
« Determine if staging" area soils, are
contaminated by, system spills or
malfunctions.
Decontaminate the unit with toluene
to levels less than 50 ppm in
decontamination residues.
4. Equipment anc( Material Handling"
Problems j
Observe equipment and material
handling problems that would affect
the performance of a full-scale site
cleanup. ; ' . ; '
' ' i
. i /
Facility and Process
Description |
Contaminated j sediments from five
harbor locations jwere processed by the
PCU; The U.S. Army Corps of Engineers
dredged sediments from the harbor and
stored them \n\ 55-gallon drums for
processing by the PCU. The drummed
sediments were blended to provide
feedstocks for four tests. Each of the four
tests was run similarly except that the
number of [passes and PCB
concentrations were varied for each test.
A pass was defined as one cycle of the
feed through the PCU. A pass of feed
results in a treated sediment product and
an extract product. Collecting and
recycling the treated sediment through
the PCU constituted an additional pass.
Recycling was cpnducted to simulate the
operation of a full-scale commercial
system. The PCU is only a two-stage
system, whereas commercial designs
include four stages, longer extractor
residence times, and longer phase
separation times:
I.Test 1 was run as a shakedown test
to set pressure and flowrates in the
PCU. The; feed had a PCB
concentration of 360 ppm. Three
passes were run to gain experience
with materials handling.
2.Ten passes, were run in Test 2 to
simulate a high-efficiency process
and to achieve treated sediment
levels less than 10 ppm. The feed
had a PCB concentration of 350
ppm. A 350 ppm concentration was
chosen forj this test since this
represents an average, or typical,
PCB concentration in the harbor.
3. Test 3 was a 3-pass test that used a
288 ppm feed. The purpose of Test
3 was to replroduce the results of the
first three passes of Test 2.
4. Test 4 was a 6-pass test. The
purpose of this test was to reduce a
high-level waste (2,575 ppm) 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. ,
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.
Decontamination of the system involved
running toluene through the PCU as a
solvent wash.
The PCU 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. The PCU process flow
diagram is shown in Figure 1. For each of
the 3 demonstration tests, a batch of
approximately 50 gallons of sediments
was fed to the unit at 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.
The PCU was not designed for large-
scale remedial actions. Therefore, treated
sediments were recycled, or passed
through the unit to simulate operation of a
commercial-scale unit. CF Systems'
commercial-scale designs do not include
recycling. These designs feature 60 gpm
flowrates, several extraction stages, and
longer processing times.
The process steps included extraction,
phase separation and solvent recovery. A
simplified flowchart is shown in Figure 1.
In step one, sediments were fed into the
top of an extractor at a rate of 0.9 gpm. In
step two, solvent was compressed to a
liquid state and allowed to flow through
the same extractor. In the extractor, the
solvent was thoroughly mixed with the
waste at a pressure of 240 psig.
Following this extraction procedure, the
residual mixture of water/solids was
removed from the base of the extractor
(step three). In step four, the mixture of
solvent and organics left the top of the
extractor and was expanded across a
valve prior to passing to a separator. The
reduction in pressure caused the solvent
i 2
-------
Simplified Flow Chart
Here is the CF Systems unit operating cycle, for extracting and separating organics
from liquid or solid waste:
1. Solid or liquid waste fed into
top of extractor.
4. Mixture of solvent gas and
organics leaves extractor,
passes to separator through
valve where pressure is
partially reduced.
Extractor
Wastewater or Sludge
2. Condensed by compression
at 70°, solvent gas flows
upwards through extractor,
making non-reactive contact
with waste. Solvent typically
dissolves out up to 99 + % of
organics.
3. Clean water or water/solids
mixture then removed from
extractor.
Figure 1. CF Systems organics extraction simplified flow chart.
Organics
Compressor
5. In separator, extraction gas
vaporized and recycled as fresh
solvent.
6. Organics drawn off from
separator, recovered for
disposal or recycling as
feedstocks or fuel.
to vaporize through the top of the
separator. It was then collected and
recycled through the compressor as fresh
solvent (step five). The liquid organics left
behind were drawn off the separator and
pumped to storage (step six). About 1 to
2 hours were required to run a feedstock
through the PCU.
Results and Discussion
The program obtained a large amount
of analytical and operating data for
evaluating the effectiveness of the PCU
for extracting PCBs from New Bedford
Harbor sediments. The results are
summarized beiow.
System Performance
The performance of the treatment unit
was evaluated in terms of extraction
efficiency and a mass balance. Extraction
efficiency per pass is defined as the input
PCB concentration minus the output PCB
concentration divided by the input PCB
concentration (multiplied by 100 percent).
An inventory of system ..inputs and
outputs was established and evaluated
for total mass, total solids, and the total
mass of PCBs.
PCB analyses for feed sediments and
treated sediment, conducted for samples
collected at each pass, are shown in
Figures 2, 3, and 4. The data show that
treated sediment concentrations below 10
ppm are achievable and that as much as
84 percent of the PCB contained .a
sediment that can be removed in a single
pass. In Test 2, feed containing 350 ppm
of PCB was reduced to 40 ppm after 10
passes through the PCU. In Test 3, a 288
ppm feed was reduced to 82 ppm after 3
passes. In Test 4, a 2,575 ppm feed was
reduced to 200 ppm after 6 passes.
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. Instead, they reflect
cross->contamination within system
hardware. Only 50 to 150 gallons per
day were run through the unit, which was
designed to handle up to 2,160 gallons
per day. Therefore, some solids may
have been retained in equipment dead
spaces and intermittently discharged at
later passes. Since the treated sediment
collection tanks were under pressure, it
was not possible to clean out collection
hardware and piping.
Extraction efficiencies greater than 60
percent were achieved on the first pass
of each test. Later passes, or recycles, of
treated sediments through the unit
resulted in efficiencies that ranged from
-------
1
,Q.
§
'a
'£
c
1
8
Q.
§
-------
zero to 84 percent. This wide range was
due to solids retention in the system.
Solids retained in the system cross-
contaminated treated sediments that
were recycled. Recycling was necessary
to simulate the peformance of a full-scale
commercial system. CF Systems' full-
scale designs do not include recycling
since more extraction stages and longer
processing times are involved.
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 were processed
during Tests 2, 3, and 4. Of the total, 93
percent was accounted for in effluent
streams. The slight imbalances, 4 and 7
percent, are attributed to the inaccuracy
of the weighing device (1 percent),
sample error, and accumulation of mass
in system hardware.
A mass balance was not established for
RGBs. A total of 157 grams were fed to
the unit during system shakedown and
Tests 2, 3, and 4. Of the total, 80 grams
were accounted for in system effluents.
Decontamination washes produced 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 analytical Method
8080 precision criteria established for this
project were plus or minus 20 percent
and accuracy criteria were plus or minus
50 percent. In addition, the mass balance
calculation was dominated by the Test 4
feed concentration. 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.
Metals were not expected to be
removed from the sediments, and were
not removed during the extraction. 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 changed by the treatment
process.
The decontamination procedure
showed that PCBs were separated from
the sediment during the tests since
nearly all, 88 percent, of the PCBs were
contained in extract subsystem hardware.
Of the 81 grams of PCB fed to the unit
during Tests 2, 3, and 4, only 4 grams
remained in the final treated sediments.
This indicated an overall PCB separation
efficiency of 95 percent. Subsequent
decontamination of the PCU wjth a
toluene wash showed that some PCB had
accumulated in system hardware.
However, 91 percent of the PCBs
contained in decontamination residues
were found in extract subsystem
hardware.
Operating Conditions
Operating conditions that were
essential to the efficient performance of
the PCU were manually controlled and
monitored during Tests 2, 3, and 4.
These included (1) feed temperature,
particle size, flow rate, pH, and solids
content, (2) solvent flow rate and
solvent/feed mass ratio, and (3) extractor
pressure and temperature. The unit
generally performed as predicted by the
developer, although some deviations
from the planned specifications occurred.
During Test 2, feed temperatures for
the last 4 passes were 10 degrees F
lower than the minimum specification, 60
degrees F. This may have contributed to
decreased extraction efficiency that was
apparent during this test. Sustained low
temperatures could have the effect of
seriously reducing extraction efficiency of
a full-scale commercial system.
Solvent flow fluctuated as much.as 75
percent above and below the nominal
flow rate, 12 Ib/min. This may have
affected the solvent-to-feed ratio in Test 2
Pass 1. Low solvent-to-feed ratios could
directly affect extraction efficiency in a
full-scale system, since less 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 prevent damage to
system valves. Less than 1 percent of the
sediment particles were greater than the
one-eighth inch.
Specifications for maximum viscosity,
1,000 centipoise, were met by adding
water to form a pumpable feed mixture.
Feed viscosities ranged from 25 to 180
centipoise. 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 percent, after the fourth
pass of Tests 2 and 4. A 10-percent
minimum spec was set merely to ensure
that the technology would be
demonstrated for high solids content
feeds.
Health and Safety
Considerations
The Health and Safety Plan established
procedures and policies to protect
workers and the public from potential
hazards during the demonstration. Air
emissions from the unit did not affect
operating personnel or the local
community. Combustible gas meters
indicated that the unit did not leak
significant amounts of propane.
Therefore, operation of the unit does not
present an explosion threat much
different than that associated wfth
domestic propane usage. 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. The
treated sediment subsystem was
successfully decontaminated before
leaving the site. The extract subsystem
was decontaminated with toluene,
however, the decontamination goal of 50
ppm was not achieved since the final
wash contained 60 ppm of PCB.
Equipment and Material
Handling Problems
Equipment and material handling
problems occurred throughout the dem-
onstration. 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 anticipated since relatively small
volumes of sediments were batch-fed to
a unit that was designed for continuous
operation. The nominal capacity of the
unit is 2,160 gallons per day, but only 50
to 100 gallons per day were batch-fed
during shakedown and Tests 2, 3, and 4.
Consequently, the unit intermittently
discharged and retained solids with each
pass.
Previous use of the unit affected
interpretation of semivolatiles data.
Internal surfaces of extract collection
hardware collected PCBs as evidenced
by mass balances. 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. As a result of this
demonstration, CF Systems now requires
-------
more rigorous decontamination
procedures for the PCD.
Solids were observed in extract
samples that were expected to be solids-
free. This indicates poor performance or
failure of the pleated paper cartridge
filter. An alternative type of filter should
be investigated by the developer.
Low-pressure dissolved propane
caused foaming 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 elimination of the foaming
problem.
Two analysis methods for PCBs were
used and results were compared.
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
Alaclors, 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 contaminants (Aroclors
1242 and 1254) in the harbor sediments.
Methods 680 and 8080 produced
similar relative results, but very different
absolute results. Use of Method 680 in
Test 4 showed a PCB extraction
efficiency of 96 percent and Method 8080
showed a similar efficiency, 87 percent.
However, Method 680 showed an
untreated sediment PCB concentration of
8,700 ppm while Method 8080 showed
2,575 ppm. Data quality objectives were
met for each measurement.
Conclusions and
Recommendations
Based on 'the above data and
discussions, the following conclusions
and recommendations can be made
concerning the operation and
performance of the CF Systems organics
extraction process.
LEven though solids retention caused
cross-contamination of treated
sediments, significant PCB removal
occurred. For example, in Test 2
after Pass 9, treated sediments
contained 8 ppm of PCB.Compared
with a Test 2 feed concentration of
350 ppm, this represents an
extraction efficiency of 98 percent.
2. System decontaminated procedures
showed that PCBs were separated
from the sediment since nearly all,
88 percent, of the PCBs were
contained 'in extract subsystem
hardware. Of the 81 grams of PCB
fed to the Unit duing Tests 2, 3, and
4, only 4 grams remained in the final
treated sediments. This indicates an
overall PCB separation efficiency of
95 percent. |
3. Bench-scale tests are useful for
determining whether or not organics
contained in a soil will be extracted
by a liquified solvent such as a
propane-butane mixture. Bench-scale
tests may also be used to determine
if a liquified solvent selectively
extracts specific classes of organica
such as high or low molecular weight
PCBs. Bench-scale tests, however,
do not yield information relating to
operational and material handling
issues such a pumpability, foaming,
and temperature, for example.
4. Commercial-scale designs for appli-
cation of |the technology should
ensure that operating specifications
are maintained. Wide fluctuations in
the feed-to-solvent ratio should be
minimized, since extraction
efficiency may be directly related to
this parameter.
5. Feed materials are likely to be well
below 60 [degrees F throughout
winter months and this could affect
performance. Therefore, heat must
be added; to sediments fed to a
commercial-scale unit.
6. Pretreatment technology will be
required to. condition feed materials.
Coarse splids removal will be
required to maintain feed sediment
particle sizes below one-eighth inch
and water jnust be added to ensure
pumpability.
7.Health and!safety monitoring showed
that OSHAj level B protection will be
necessary: for personnel that will
handle input and output. However,
only OSHA level C protection will be
necessary for unit operators.
S.Regulatory or engineering
interpretation of PGB analysis should
include consideration of the analysis
methods used.
9. Operations, materials handling, and
safety issues are addressed in the
Application Analysis Report. Costs
are estimated for several cases
involving the New Bedford Harbor
Superfund site. A significant cost
element for a full-scale system is
extraction process equipment which
must be scaled to handle much
higher throughputs (60 gpm) than the
PCU (0.9pgm). Full-scale extractors
have 4 to 6 foot diameters as
compared with the 18 inch diameter
of the PCU extractors. Recom-
mended treatment technology
includes conveyors, screening, heat
and water addition, mixing and
holding tanks. Post treatment
technology includes treated
sediment dewatering, wastewater
treatment and reuse, holding tanks,
conveyors and disposal of treated
sediments and extracted organics.
Onsite analytical capabilities and
health and safety program
implementation are additional cost
elements.
10. 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.
-------
-------
The EPA Project Manager, Richard Valentlnetti, was with' the Risk Reduction
Engineering Laboratory, Cincinnati, OH 45268 (see\below).
The complete report consists of two volumes, entitled "Technology Evaluation
Report, SITE Program Demonstration Test, OF Systems Organics
Extraction System, New Bedford Harbor, Massachusetts."
"Volume I (Order No. PS 90-786 495/AS; Cost: $21.95, subject to change)
discusses the results of the SITE demonstration \
"Volume II (Order No. PB 90-116 024/AS;Cost: $42.95, subject to change)
contains the technical operating data logs, the sampling and analytical
report, and the quality assurance project plan/test plan
These two reports will be available only from: \
National Technical Information Service ;
5285 Port Royal Road \
Springfield, VA 22161 <
Telephone: 703-487-4650 i
A related report, entitled "Applications Analysis Report; CF Systems Organics
Extraction System" which discusses application and costs, is under
development. \
For further information, Laurel Staley can be contacted at: \
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency ;
Cincinnati, OH 45268 '
United States
Environmental Protection
Agency
Center for Environmental Research
Information !
Cincinnati OH 45268 !
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/540/S5-90/002
------- |