EPA/540/5-90/002
                                    January  1990
               VOLUME I
     TECHNOLOGY EVALUATION REPORT
 CF SYSTEMS  ORGANICS  EXTRACTION  SYSTEM
      NEW BEDFORD, MASSACHUSETTS
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 Superfund 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.
                                      ii

-------
                                   FOREWORD
     The Superfund  Innovative  Technology  Evaluation  (SITE)  program  was
 authorized  in  the 1986 Superfund  amendments.  The program is  a joint effort
 between EPA's  Office  of  Research  and  Development and Office of Solid Waste and
 Emergency Response.   The purpose  of the program is to assist  the 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
 polychlorinated 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.  This Technology Evaluation Report describes the field activities and
 laboratory results.  An  Applications Analysis will follow and provide an
 interpretation of the data and conclusions on the results and potential
 applicability  of the technology.

     Additional copies of this report may be obtained at no charge from EPA's
 Center for Environmental  Research Information,  26 West Martin Luther King
 Drive,  Cincinnati, Ohio  45268,  using the EPA document number found on the
 front cover of the report.  Once this supply is exhausted, copies can be
purchased from the National  Technical  Information Service, Ravensworth Bldq
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.
                                       E.  Timothy Oppelt,  Director
                                       Risk Reduction Engineering Laboratory

-------
                                   ABSTRACT

     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 Massachusetts.  Contaminated sediments were
treated by CF Systems' Pit Cleanup Unit (PCU) that used liquefied 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
1n liquefied propane/butane.

     The objectives  included an evaluation of  (1) the unit's performance, (2)
system operating conditions, (3) health and safety considerations, and (4)
equipment and system materials handling problems.  Extensive sampling and
analyses were performed showing that polychlorinated .biphenyl  (PCB) 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 operators  or  the local community.
                                       iv

-------
                              VOLUME I CONTENTS*
                                                       '                 Page

 Foreword	 .  .  .  .  .  .....        iii
 Abstract  	 ......                       iv
 Figures	.  .	•....;•!.' .  .  '''•'•••••    v_
 Tables  .  .	 .; .  . .  ....  .  ._.  ..'.'.'.   viii
 Acknowledgements  		         ix
 Abbreviations and Symbols.	  .  .    .  .  .^.  .  .  .  .      x

 1.   Introduction	      i

     1.1   Background	      i
     1.2   Program Objectives	      2
     1.3   Technology Evaluation Criteria	      2
     1.4   Description of Operations	  .  .  .      3
     1.5   Project Organization.	'.'.'.'.'.      4

 2.   Summary of Results	      6

     2.1   System Performance.  	      6
     2.2   Operating Conditions	  .  .          7
     2.3   Health and Safety Considerations.  .  .	      8
     2.4   Equipment and Material Handling Problems	  .  .      9
     2.5   Lessons Learned ..... 	    10
          \
 3.   Proces^ Design	        13
           \              \              "'      ••••••  .•
     3.1   Process Description	        13
     3.2   Equipment Specifications		  .  .  .        13
     3.3   Process Flow Diagram  	  .........    IB

 4.   Demonstration Site Description .  .  .  ....  .  .  ....  	    22

     4.1   Site Characteristics	          22
     4.2  Predemonstration Samples	-.',-'.'.'.'.'.'.'.    24

 5.   Field Activities	   	    29

    5.1  Bench-Scale Tests	'....".."	    . .    29
    5.2  Operations  Summary	    31
    5.3  Operations  Chronology	  .........    34

6.  Sampling and Analytical Program	-	    41

    6.1  Sampling Locations	    41
    6.2  Sampling Schedule 	  .............      43
    6.3  Analytical  Methods and Physical Tests	 . .  . .    45
    6.4  Process Control  and  Field Measurement Devices  	    48
*Volume II contains  Sampling and Analytical Reports and Operating Log Data.

-------
                        VOLUME I CONTENTS* (Continued)
7.  Results and Discussion
    7.1  System Performance	
    7.2  Operating Conditions	
    7.3  Developer's Goals 	
    7.4  Health and Safety Monitoring	
    7.5  Equipment and Material Handling Problems.
    7.6  Data Quality Assurance	
8.  References 	

APPENDIX A - MASS INVENTORIES FOR TESTS 2, 3, AND 4

APPENDIX B - MASS BALANCES FOR DECONTAMINATION EFFLUENTS
Page

,  51

  51
  63
  70
  71
  72
  73

  81

  82

  92
*Volume II contains Sampling and Analytical Reports and Operating Log Data.
                                       vi

-------
                                    FIGURES
Number

1.1   SIMPLIFIED FLOW CHART

3.1   PIT CLEANUP UNIT

3.2   CF SYSTEMS PROCESS SCHEMATIC

4.1   DEMONSTRATION LOCATION

4.2   PREDEMONSTRATION SAMPLES

5.1   PLAN SKETCH FOR NEW BEDFORD HARBOR DEMONSTRATION

6.1   SYSTEM FLOW DIAGRAM

7.1   TEST 2 PCB REDUCTION

7.2   TEST 3 PCB REDUCTION

7.3   TEST 4 PCB REDUCTION

7.4   POTENTIAL PIT CLEANUP UNIT PCB REDUCTION

7.5   ILLUSTRATIVE INVENTORY SHEET         :

7.6   SOLIDS THROUGHPUT PER PASS, TEST 2

7.7   SOLIDS THROUGHPUT PER PASS, TEST 3

7.8   SOLIDS THROUGHPUT PER PASS, TEST 4

7.9   MEAN SOLVENT FLOWRATE, TEST 2

7.10  MEAN SOLVENT/FEED RATIO, TEST 2

7.11  MEAN SOLVENT FLOWRATE, TEST 3

7.12  MEAN SOLVENT/FEED RATIO, TEST 3

7.13  MEAN SOLVENT FLOWRATE PLUS ALTERNATE SOLVENT FLOWRATE, TEST 4

7.14  MEAN SOLVENT/FEED RATIO PLUS ALTERNATE SOLVENT/FEED RATIO
      TEST 4

7.15  FEED/TREATED SEDIMENTS SOLIDS CONCENTRATION, TEST 2

7.16  FEED/TREATED SEDIMENTS SOLIDS CONCENTRATION, TEST 3

7.17  FEED/TREATED SEDIMENTS SOLIDS CONCENTRATION, TEST 4
5

14

19

23

25

32

41

53

53

53

55

57

61

61

61

67

67

67

67

67


67

69

69

69

-------
                                    TABLES
Number
3.1
3.2
PROCESS EQUIPMENT DESCRIPTION
RANGE OF OPERATING CONDITIONS FOR TESTING
4.1   PREDEMONSTRATION RESULTS OF TOTAL SOLIDS, OIL AND GREASE,  PCS,
      AND pH ANALYSES
4.2   PREDEMONSTRATION RESULTS OF METALS ANALYSES
5.1   BENCH-SCALE TEST DATA
6.1   SAMPLES FOR CF SYSTEMS NEW BEDFORD TESTS 2, 3, AND 4
6.2   METHOD 680 ANALYTICAL RESULTS FOR TEST 4
6.3   PROCESS CONTROL AND FIELD MEASUREMENTS
7.1   PASS-BY-PASS PCB CONCENTRATIONS AND REDUCTION EFFICIENCIES
7.2   MASS ACCUMULATION AND LOSS IN THE SYSTEM
7.3   METALS CONTENT OF FEED, TREATED SEDIMENT, AND EXTRACT
7.4   EP TOXICITY CHARACTERISTIC OF TREATED AND UNTREATED SEDIMENTS
Page
 15
 16

 27
 28
 30
 44
 47
 49
 52
 58
 64
 65
                                     vili

-------
                                ACKNOWLEDGMENTS

     This  report was prepared under the direction and coordination  of  Richard
Valentinetti, EPA SITE Program Manager in! the  Risk Reduction Engineering
Laboratory, Cincinnati, Ohio.  Contributors and  reviewers for this,  report were
Frank Ciavattieri of EPA Region I, Remedial Project Manager for the New
Bedford Harbor Superfund site; Jim Cummings and  Linda Galer from the Office of
Solid Waste and Emergency Response; Paul Desrosiers, Diana Guzman, Paul de
Percin, and Laurel Staley from the Office of Research and Development;
Christopher Shallice and Thomas Cody, Jr. from CF Systems Corporation.
Logistics at the site was coordinated by Mark Otis, U.S. Army Corps of
Engineers; and Alan Fowler and Siegfried Stockinger of EBASCO Services, Inc.

     This report was prepared for EPA's Superfund Innovative Technology
Evaluation (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, Richard Hergenroeder,  Kim Gotwals,  Jorge
McPherson ,and Fernando Padilla.   Laboratory analyses were conducted by
E.C. Jordan,  Inc.,  Portland,  ME, and Radian Corporation, Austin, TX.
                                     ix

-------
                          ABBREVIATIONS AND SYMBOLS
amps          amperes
ASTM          American Society for Testing and Materials
bbl/day       barrels per day
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
 NIOSH
 NR
 ORD
 OSWER
 OVA
 OZ
 PAHs
 Pb
 PCBs
 PCU
 ppm
 psig
 QA
 QC
 RCRA
 RPD
 RREL
 RSD
 SARA
 SBT
 SITE
 SRC
 TDS
TS
TSS
VAC
VOAs
Zn
 not  detected
 National  Institute of Occupational Safety and Health
 not  reported
 Office of Research and Development
 Office of Solid Waste and Emergency Response
 organic vapor analyzer
 ounces
 polyaromatic hydrocarbons
 lead
 polychlorinated biphenyls
 Pit Cleanup Unit
 parts per million
 pounds per square inch gauge
 quality assurance
 quality control
 Resource Conservation and Recovery Act of 1976
 relative percent difference
 Risk Reduction Engineering Laboratory
 relative standard deviation
 Superfund Amendments and Reauthorization Act of 1986
 still bottoms tank
 Superfund Innovative Technology Evaluation Program
 solvent recovery column
 total dissolved solids
 total solids
 total suspended solids
 volts,  alternating current
volatile organic analytes
zinc
 less than

-------

-------
                                   SECTION 1

                                 INTRODUCTION
1.1  BACKGROUND
     In response to the Superfund Amendments and Reauthorization Act of 1986
(SARA), the Environmental Protection Agency's Offices of Research and Develop-
ment (ORD) and Solid Waste and Emergency Response (OSWER) have established a
formal program to accelerate the development, demonstration, and use of new or
innovative technologies as alternatives to current containment systems for
hazardous wates.  This program is called Superfund Innovative Technology
Evaluation or SITE.

     The major objective of a Demonstration Program is to develop reliable
cost and performance information on innovative alternative technologies so
that they can be adequately considered in; Superfund decision making.  Common
measurement, monitoring, and evaluation guidelines and protocols were
developed by ORD and used to collect the data and information from the
demonstration.
     CF Systems Corp., developer of an organics extraction technology, was
selected to demonstrate their system at the New Bedford Harbor, Massachusetts
Superfund site.  The system demonstrated 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 solubilize in the process solvent, a liquefied gas.  The
process used a mixture of liquefied propane and butane, at 240 psi and 69
degrees F, as a solvent for extracting organics from soils.  As liquefied
solvent was mixed with the waste, organics were extracted into the solvent.
The solvent-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 harbor sediments contain

-------
pollutants discharged to the harbor from various industrial  sources.   The
pollutants include polychlorinated biphenyls (PCBs),  polynuclear aromatic
hydrocarbons, copper, chromium, zinc, and lead.   PCBs present the greatest
toxic threat and concentrations as high as 30,000 ppm have been observed.  The
estimated volumes of harbor sediments containing various concentrations of

PCBs are:
         PCB Concentration (ppm)

               0-50
               50-500
               500-5,000
               >5,000
Volume (cubic yards)

       878,000
       236,000
        91,000
        16,000
Samples of harbor sediments were dredged by the U.S. Army Corps of Engineers
and stored in 55-gallon drums for use in this demonstration.


1.2  PROGRAM OBJECTIVES

     The objectives of this SITE demonstration of the CF Systems organics
extraction technology at the New Bedford Superfund site were to evaluate the

following:


     1.  Performance of the process  in terms of PCB extraction efficiency and
         a mass  balance.

     2.  Variations in process  operating conditions and possible effects on
         performance.

     3.  Potential health  and safety impacts resulting from system operation.

     4.  Equipment and material handling problems.

     5.  Projected system  economics.


1.3  TECHNOLOGY  EVALUATION CRITERIA

     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

          a.  Evaluate PCB concentration in sediments before and after
              treatment.

-------
          b.  Evaluate PCB extraction efficiency with each pass, or recycle  of
              sediments through the unit.
          c.  Evaluate mass balances established for total mass, solids, and
              PCBs.
      2.  Operating Conditions
          a.  Compare operating conditions to operating specifications for
              flow,  temperature, pressure, and physical  sediment character-
              istics of the sediment and assess the effect on extraction rate.
      3.  Health and Safety Considerations
          a.  Determine if significant amounts of propane/butane or PCBs are
              emitted to the air by the process.
          b.  Determine if staging area soils are contaminated by system spills
              or malfunctions.
          c.  Decontaminate the unit with toluene to  levels  less than  50 ppm in
              decontamination  residues.                                    H
      4.   Equipment  and Material  Handling Problems
          a.   Observe  equipment and material  handling problems that would
              affect the performance of  a full-scale site  cleanup.
1.4   DESCRIPTION OF OPERATIONS
      Contaminated sediments from  five harbor  locations were processed by  the
PCU.  The Corps of  Engineers dredged sediments from the harbor  and  stored  them
in 55-gallon  drums  for processing  by the PCU. , Sediments were obtained  from
locations H-20, H-21, H-22, H-23,  and I-ll .shown in Figure 4.2.  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 as follows:
Test No.
   1

   2

   3
                         Feed Stock
A 50-gallon mixture of sediments taken from locations H-20
H-21, and H-23.  PCB concentration was 360 ppm.           '
A 50-gallon mixture of sediments taken from locations H-20
H-21, and H-23.  PCB concentration was 350 ppm.           '
A 50-gallon mixture of sediments taken from locations H-20
H-21, and H-23.  PCB concentration was 288 ppm.

-------
   4          A 50-gallon mixture of sediments taken from locations H-22 and
              1-11.  PCB concentration was 2,575 ppm.

Test 1 was a system shakedown run to set flow rates and operating pressures
and to provide samples for laboratory evaluation of sample matrices.  Samples
were collected 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.

     The process steps included extraction, phase separation and solvent
recovery.  A simplified flowchart is shown in Figure 1.1.  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 1n  pressure  caused  the solvent 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  from the separator  and pumped to storage  (step six).

     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 three passes and Test 4 involved six passes.   Samples  were taken from
the feed kettle,  extract collection tank, and treated  sediment  tank.

1.5  PROJECT ORGANIZATION
      Through a Cooperative Agreement between EPA and CF Systems Corp., CF
 Systems was responsible for operating their  equipment  and EPA prepared the
 demonstration plan,  prepared the test site,  arranged for the sampling plan
 analyses,  conducted sampling, evaluated the  data,  and  prepared  the Technology
 Evaluation Report.  The evaluation included the following activities:

      o  Preparation of staging area to support testing, equipment setup, and
         health and safety orientation of field staff

-------
  •   Pre- and post-demonstration  sampling and analysis of  staginq  area
      soils

  •   Public  information meetings  held  to  review CF Systems  technoloav and
      tests on New Bedford  Harbor  sediments

  t   A shakedown  test,  three  extraction tests,  and  equipment
      decontamination

  •   Staging  area  closure  and  disposal  of  waste  materials.
Simplified Flow Chart
 Here
 from
 1. Solid or liquid waste fed Into
 top of extractor.
_  _,	unit operating cycle, for extracting
or solid waste:

                  4. Mixture of solvent gas and
                  organles leaves extractor,  '
                  passes to separator through ,
                  valve where pressure Is
                  partially reduced.
                Extractor
     1.
       Wastewater
         or Sludge
2. Condensed by compression
at 70° F, solvent gas flows
upwards through extractor,
making non-reactive contact
with waste. Solvent typically
dissolves out up to 99+* of
organlcs.

3. Clean water or water/solids
mixture then removed from
extractor.
                       3.
                      Water
                  and/or Solids
                                              OrganScs
                                                                            Compressor
5. In separator, extraction gas
vaporized and recycled as fresh
solvent.

6. Organlcs drawn off from
separator, recovered for
disposal or recycling as
feedstocks or fuel.
                          Figure 1.1   Simplified Flow Chart

-------
                                  SECTION 2

                              SUMMARY  OF  RESULTS


     The program obtained a large amount of analytical  and operating data and

was able to evaluate the four criteria stated in  Section 1.3.   A summary of

the results, which correspond to the program objectives, is presented below.


2.1  SYSTEM PERFORMANCE

     The performance of the treatment unit was evaluated in terms of extrac-

tion 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.  Results of these evaluations are

summarized  as follows:


     •  PCB removal efficiencies of 90 percent were achieved for sediments
        containing  PCBs  ranging  from 350 to  2,575 ppm.  A  high  removal
        efficiency  was achieved  after several passes, or  recycles,  of  treated
        sediments through the unit.

     •  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  that ranged from 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 peformarice  of  a
         full-scale  commercial system.   CF  Systems'  full-scale  designs  do  not
         include recycling, since additional  extraction  stages  and  longer
         processing  times are involved.   In addition,  only 50  to 150 gpd were
         run through the  unit, which was designed  to handle up  to 2,160 gpd.
         Therefore,  some  solids  may have been retained in  equipment dead spaces
         and intermittently discharged during subsequent passes.

      •  A mass  balance  was not  established for  PCBs.   A total  of 157 grams was
         fed to  the unit during  system shakedown  and Tests 2,  3, and 4.  Of the
         total,  80 grams were accounted for in system  effluents.  Decontamina-
         tion 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  imbalance may be the result of 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 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.  CF Systems
         did not decontaminate the unit with toluene prior  to the tests at New
         Bedford.   CF Systems' standard operating procedures now incorporate
         decontamination with toluene.

      •  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.

      •  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,  although high concentra-
         tions of  metals were  present.   Copper  and zinc typically exceeded
         1,000 ppm.   Chromium  and  lead concentrations  ranged from 500 to
         1,000 ppm.

      •  The decontamination procedure showed  that PCBs were separated  from  the
         sediment  during the tests  since  nearly all  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.   Subsequent decontamination of the  PCU with 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.

      e   A QA/QC review  showed that  analysis data  of PCBs in sediments  for
         Tests  1 through 5 were sufficiently accurate and precise  for an
         engineering assessment of the efficiency  of this demonstration.


2.2   OPERATING CONDITIONS


     Operating conditions essential to the efficient performance of the PCU
were manually controlled and monitored during Tests 2, 3, and 4.   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 predicted by the developer,
although some deviations from the planned specifications  occurred.  An

evaluation of operating conditions is summarized as follows:
     t  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 from 190 to 290 psig.

-------
    •  During Test 2, feed temperatures for the last 4 passes were 10 degrees
       F lower than the minimum specification, 60 degrees F.  Decreased
       extraction efficiency, which was apparent during this test, may have
       been related to low feed temperatures.  Sustained low temperatures
       could have the effect of seriously reducing extraction efficiency in a
       full-scale commercial system.

    t  Solvent flow fluctuated as much as 75 percent above and below the
       nominal flow rate, 12 Ib/min.  In Test 2, Pass 1 this caused the
       sol vent-to-feed ratio to fall below specifications.  This could affect
       the extraction efficiency in full-scale system, since 1 ess 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 one-eighth  inch.

    •  specifications for maximum viscosity,  1,000 centipoise, were met  by
       adding water  to form  a  pumpable  feed  mixture.  Feed  viscositiestranged
       from 25 to 180 centipoise.  The  mass  of waste  increased by about  33
       percent,  because unit operators  arbitrarily added  water.

    •  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.

     •   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.


2.3  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.  Some

observations,  based on an evaluation of health and safety monitoring conducted

during the demonstration,  were:


     •  Combustible gas meters indicated that the unit did not leak signifi-
        cant amounts of propane.  Therefore, operation of the unit does not
        present  an explosion threat much different than that associated with
        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 and butane or
        liquids.  Only minor  leaks occurred and staging area soils were not
        affected.
                                      *8

-------
     •  The unit was not completely decontaminated before it left the site.
        Toluene was fed to the unit as a decontamination wash.   Toluene wash
        was collected at the extract product tank and the two treated sediment
        product tanks.  The final toluene wash from the treated sediment tank
        contained less than 34 ppm of PCBs.  The final wash from the extract
        tank contained 60 ppm of PCBs.


2.4  EQUIPMENT AND MATERIAL HANDLING PROBLEMS

     Equipment and system material handling problems occurred during the
tests,  although some problems were anticipated.   Problems included:
     •  Internal surfaces of extractor hardware and piping collected PCBs as
        evidenced by mass balances 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 accumu-
        lations of organics are believed to be the result of the short dura-
        tion of the tests and the small volume of  organics contained in the
        feed sediment, relative to the volume of the extraction system
        hardware.   PCBs are soluble in oil  and grease,  which is believed to
        coat the internal surfaces of system hardware.   Continuous operation
        of the unit has resulted in continuous discharge of extracted organics
        during 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.   Therefore,
        some solids may have been retained  in equipment dead spaces and
        intermittently discharged during subsequent passes.

     •  Solids were observed in extract samples, which  were  expected to be
        free of solids.   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:contained organic  sludge  from
        prior  use  of  the unit at a petroleum refinery.   Presence of the
        petroleum  residuals prevented  complete interpretation of data
        collected  for oil  and grease and semivojatile organics.

     •  Low-pressure  dissolved  propane caused  foaming 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.

-------
2.5  LESSONS LEARNED

     Lessons learned that could be useful in planning future technology

demonstrations are discussed in this section.  The lessons learned center on

demonstration preparation and logistics, treatment goals, and sampling and

analysis for PCBs.


Demonstration Preparation and Logistics

     •  The developer's equipment must be decontaminated prior to being
        brought on site.  Quality Assurance Project Plans (QAPPs) should
        include predemonstration sampling and analysis to ensure that the
        technology is free of contaminants.  The PCU contained residue from a
        prior demonstration at a petroleum refinery.  In addition, the system
        carbon canister, which receives  vented low pressure gas, had not been
        changed.

     t  The developer must commit time needed to decontaminate the unit before
        and after the demonstration, to  conduct preventive maintenance, and to
        respond to required sampling protocol.  No equipment failures occurred
        at New Bedford; however, a week  was  added to the schedule while
        preventive maintenance was conducted.

     •  The generation  of trash  (e.g., gloves and Tyvek  suits) and process
        residue can  be  substantial.  Adequate provisions should  be made for
        waste  residue and trash  management.  Program personnel and the
        developer should develop realistic projections of the amount of
        process residues and  trash that  will be produced as a result of the
        demonstration.  At New Bedford,  87 drums of waste were produced,
        compared  with 3 drums of harbor  sediments fed  to the unit.  The  large
        volume of waste resulted from  water  added to the feed; water used to
        decontaminate personnel  and  sampling gear;  and trash, such as  gloves
        and Tyvek suits, that were not compacted.

 Developer's Input

      •  The developer should  provide specific  protocol  for  conducting  shake-
        down  tests.   This  protocol  should distinguish  process control  from
        onsite process  optimization.

      •  The  developer  should  provide a basis or  a  design procedure  for scaling
        up bench- or pilot-scale test results  to  a  commercial-scale  system  and
         for  correlating batch test results to  operation  of  a  continuously fed
         unit.


 Sampling  and  Analysis Methods

      t  Interpretation  of results from any PCB treatability study should
         include a discussion of the precision of the analysis method.
                                       10

-------
     •  A portable gas chromatograph GC and a chemist should be available
        onsite to allow a rapid response to changes in feed composition 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.

     e  The unit did not selectively extract one class of PCBs.  Reviewers
        suggested the use of EPA Method 680, since the CF Systems technology
        could have selectively extracted higher molecular weight PCB
        congeners.  Method 680 would reveal any selective extraction, since
        this method 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 it analyzes
        for the classes of congeners that constitute the majority of PCB
        contaminants (Aroclors 1242 and 1254) in the harbor sediments.

     s  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 sedi-
        ment PCB concentration of 8,700 ppm; while Method 8080 showed 2,575
        ppm.  Data quality objectives were met for each measurement.  There-
        fore,  regulatory or engineering interpretation of future PCB analyses
        should include consideration of the,analysis methods used.

Conclusions and Recommendations

     •  Even though solids retention caused cross-contamination of treated
        sediments, significant PCB removals occurred.

     •  System decontamination procedures showed that PCBs Were separated from
        the sediment since nearly all 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.

     •  Benctvscale tests are useful  for determining whether organics
        contained in a soil  will  be  extracted by a liquefied solvent such as a
        propane-butane mixture.   Bench-scale tests may also be used to
        determine if a liquefied  solvent selectively extracts  specific classes
        of organics,  such as high or  low molecular weight PCBs.   Bench-scale
        tests,  however,  do not yield  information relating to operational  and
        material  handling issues,  such as pumpability,  foaming,  and
        temperature.

     •  Commercial-scale designs  for  application of the technology  should
        ensure  that operating specifications are maintained.   Wide  fluctua-
        tions  in  the feed-to-solvent  ratio  should be minimized,  since
        extraction efficiency is  directly related to the  amount of  solvent
        available for  solubilizing organics contained  in  the feed.   However,
        the technology does  accommodate wide ranges in operating conditions.
        Solid  wastes with viscosities up to 1,000 centipoise and solids
        contents  of 60 percent can be fed to the unit.
                                      11

-------

*  Feed materials are likely to be well below 60 degrees F throughout
   winter months, which could affect system performance.  Therefore, heat
   must be added to sediments fed to a commercial-scale unit.

•  Pretreatment technology will be required to condition feed materials.
   Coarse solids removal will be required to maintain feed sediment
   particle sizes below one-eighth inch and water must be added to ensure
   pumpability.

*  Health and safety monitoring showed that OSHA Level B protection will
   be necessary for personnel handling system input and output.  However,
   only OSHA Level C protection will be adequate for unit operators.

•  Operations, materials handling and health and safety issues are
   addressed in the Application Analysis Report.  Costs are estimated for
   several case studies 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 through-
   puts (60 gpm) than the PCU (0.9 gpm).  Full-scale extractors have 4 to
   6 foot diameters as compared with the 18 inch diameter of the PCU
   extractors.  Recommended pretreatment technology includes conveyors,
   screening, heat and water addition, and 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.
                                  12

-------
                                    SECTION;3

                                 PROCESS DESIGN
 3.1  PROCESS DESCRIPTION
      CF Systems Pit Cleanup Unit (PCU), shown in Figure 3-1, is a continuous
 processing unit that used a liquefied 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 unit at a nominal, rate of 0.9 gpm.  Feed viscosity
 was maintained below 1,000 cP, by adding water 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.   Liquefied 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 liquefied solvent,
 were extracted.   After extraction,  treated sediments were  decanted and
 separated from the liquefied 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  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'."*

 3.2  EQUIPMENT SPECIFICATIONS
     The major pieces of equipment and their function are described in Table
3-1.  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
                                      13

-------
Figure 3-1. Pit Cleanup Unit
            14

-------
                    TABLE 3-1.   PROCESS EQUIPMENT DESCRIPTION
 Process Equipment     Designation
                       Function  in System
 Feed Kettle



 Basket Strainer


 Extractor l


 Decanter 1
 FK



 S-l


 E-l


 D-l
 Extractor  2
 Decanter  2
 Cartridge  Filter
Solvent Recovery
 Column
Column Reboiler
 E-2



 D-2


 F-2


 SRC
CR
Treated Sediment          RPT-l
(Raffinate) Product Tank  RPT-2
Extract Product Tank
EPT
  Holds approximately  100 gallons  of
  strained,  slurried feed.  Counter-rotating
  agitators  homogenize feed.

  Prevents oversized (>l/8  inch) feed
  material from entering the system.

  Extracts organics from water-solids feed
  mixture with sol vent  from D-2.

  Allows! separation of  solvent-organic
  mixture from water-solids layer.   Sends
  water-solids layer to Extractor 2 (E-2)
  and sblvent-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-l).

 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-l overflow.

Receives extracted organics effluent from
the Column Reboiler (CR).   Recovers
 residual  propane  via flash pressure
reduction and heat from the water  jacket.
                                     15

-------
            TABLE 3-1.  PROCESS EQUIPMENT DESCRIPTION (Continued)
Process Equipment     Designation
                       Function in System
Main Compressor
C-l
Low Pressure Solvent
 Compressor
C-2
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-l).
             TABLE 3-2.   RANGE OF OPERATING CONDITIONS FOR TESTING

Extractor Pressure (PSIG)
Extractor Temp, (degrees F)
Feed Temp, (degrees F)
Solvent Flow (Ib/min)
Feed Flowrate (GPM)
Solvent/Feed Ratio
Feed Solids (percent by weight)
Solids Size (maximum)
pH (standard units)
Viscosity (cP)
Minimum
180
60
60
8
0.2
1
10
—
6
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
                                       16

-------
 exceed normal  operating conditions.   To guard against sudden overpressure,
 each vessel  had a relief valve that  vented to a header system that discharged
 to the pollution control  system.

      The utility and process materials  requirements  that  were necessary  to
 operate the  PCU at New Bedford Harbor were:
      •
      •
      •
      t
      •
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
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 equipment consumed much more energy per gallon of
throughput.

     The operating conditions listed in Table 3-2 are essential 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
performance.  The operating parameters were set during the shakedown portion
of the demonstration.  CF Systems claimed that minor fluctuations would not
affect performance.

     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.
                                      17

-------
     The viscosity and solids content must be 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 a RCRA-approved facility.

     The feed flow rate represents the rate at which material 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.

3.3  PROCESS FLOW DIAGRAM
     The PCU process  flow  diagram is shown  in  Figure 3-2.  The  extraction
portion  of  this unit  consisted of two  stages of  counter-current extraction
with solid-liquid separation 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
 iiiiscible  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.
                                       18

-------
H
                                                                             Pressure Letdown Valves
                                                                       Treated
                                                                      Sediment
                                                                       Product
                                                                        Tank
                                                                         #1
                                                                      V
 Treated
Sediment
 Product
  Tank
   #2
Extract
Product
 Tank
                                                                          T
                                                                                                        Cartridge
                                                                                                          Filter
                                Solvent
                               Recovery
                                Column
                                                          Figure 3-2. CF Systems Process Schematic
                                                                                                                                   Legend
                                                         Feed
                                                         Propane-Butane Solvent
                                                         Propane/Organics Mixture
                                                         Extracted Organics
                                                         Processed Sediments

-------
     The pressure difference between the first decanter  and the second
extractor moved the solid-liquid stream into the second  extractor for second-
stage extraction.  Fresh liquefied 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 solvent-organics mixture flowed to decanter 2.   At decanter 2, two
Immiscible layers were formed.  The organics-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-solvent layer recycled to the first extractor
for final stage extraction.

     The organic-solvent stream from the first stage extractor 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  remained as  a  separate  liquid.
The mixture of organics containing dissolved  propane  gathered  in the column
reboiler 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
necessary 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 opened a  relief
valve,  material  was vented to a relief header,  which  directed the material to
 a blowdown tank where solids and liquids were removed from the vented stream.
 The  gases from the blowdown tank passed through a 42-gallon activated carbon
                                       20

-------
filter to remove contaminants in the propane gas.  The gas then passed through
a flame arrestor and was vented to the atmosphere.  This system was used only
once during the demonstration, at the conclusion of PCU decontamination.
                                     21

-------
                  SECTION  4.   DEMONSTRATION  SITE  DESCRIPTION

4.1  SITE CHARACTERISTICS
     The site selected for demonstration of CF Systems'  process was New
Bedford Harbor, located in New Bedford, Massachusetts.   During the 1970s,  PCBs
and other contaminants were identified in the sediments and marine life of New
Bedford Harbor and parts of Buzzard's Bay.  Studies conducted by EPA in 1980
led to New Bedford Harbor being proposed in 1982 for EPA's National Priorities
List.   The main areas of New Bedford Harbor under EPA investigation are the
Acushnet River Estuary and the harbor.  The estuary is the area of the site
north of the Coggeshall Street Bridge shown in Figure 4-1.  Areas of extremely
high PCB contamination are located at the northern tip of the estuary.  The
lower harbor includes Buzzard's Bay and the waters below the Coggeshall Street
Bridge.  The demonstration took place on a parcel of city-owned property
adjacent to the cove north of the Coggeshall Street Bridge, as shown in Figure
4-1.  Area within the dashed  line was also the site for a pilot dredging and
disposal study conducted  jointly by EPA, the Massachusetts Department  of
Environmental  Quality Engineering, and  the U.S.  Army Corps of  Engineers (COE).

     PCB concentrations  in the harbor  range  from a  few ppm to  more  than 30,000
ppiB and elevated  levels  of copper, chromium,  zinc,  and lead also  are present
at the site  (COE, 1987).  Most organic  compounds detected  in  sediments are
co-located with PCBs  and  occur at concentrations less  than or  equal to the
observed PCB  concentrations  (Ebasco,  1987).   An  evaluation of the relative
toxlcitles of PCBs  indicates  that the environmental and  public health  risks
 from sediment exposure  will  be dominated by the  PCB constituents.  However,
 some polynuclear  aromatic hydrocarbon (PAH)  compounds  occur  at certain
 locations,  and at concentrations that may significantly  contribute to  the
 overall  environmental  and public health risk associated  with  exposure  to  the
 sediments.

      PCBs  consist of a mixture of chlorinated biphenyls,  which contain a
 varying number of substituted chlorine atoms on aromatic rings.  The  persis-
 tence of PCBs in the environment and their toxicity increases as the  chlorine
 content increases.   The commercial  products of the complex chlorobiphenyls
                                       22

-------
                       Confined
                       Disposal
                        Facility
                        (CDF)
                                   (not to scale)
Figure 4-1. Demonstration Location
              23

-------
were registered and manufactured under the trademark "Aroclor."  Fifty percent
of the congeners in the New Bedford Harbor area of interest have been identi-
fied as Aroclors 1242 and 1254.  Therefore, PCBs were measured using the
Aroclor 1242 and 1254 standards.

     Based on the available literature (COE 1987, Ebasco 1987), .the pollutants
of concern were determined to be PCBs, semivolatile organics (which includes
PAHs), cadmium, chromium, copper, zinc, and lead.  CF Systems  requested
analyses for additional parameters to evaluate the overall system performance.
011 and grease was  used by CF Systems as  a process control parameter  since the
majority of organic pollutants  are extracted by  the process.   Total solids,
pH, and viscosity can  affect equipment performance.  Therefore,  these
parameters were chosen for monitoring in  addition to the pollutants of
concern.

4.2   PREDEMONSTRATION  SAMPLES
      COE  dredged  collected  sediment  from  the  Acushnet  Estuary  for  use in this
 demonstration project.  An  attempt was  made  to gather  sediments from areas
 with differing levels  of contamination  so that a range of concentrations would
 be available for  the demonstration.   Collected sediments were  stored in drums
 that were assigned identification numbers.  Drum Numbers H-20, H-21,  H-22,
 H-23, and 1-11 were later used in the demonstration.   The drum identification
 numbers correspond to a grid system used by COE in previous harbor characteri-
 zation work.  This grid system shown in Figure 4-2 enabled COE to estimate
 probable concentrations from particular areas to ensure that an appropriate
 range of samples was collected.

      The drummed sediments were conditioned for predemonstration sampling by
 removing solids greater than 1/8 inch, adding water to produce  a slurry,  and
 stirring.  Contents of each drum were hand-shoveled and sieved  through  a
 1/8-inch screen into  a second  drum.  More than  99 percent of  the solids were
 passed through the screen.  As the  second drum  was filled, harbor water was
 added to produce a stirrable  slurry.  Slurried  contents were  manually  stirred
 with a shovel and  three  samples  were drawn after 15 minutes of mixing.
 Samples  from each  drum were analyzed for PCBs,  oil and  grease,  pH,  total
                                        24

-------
NJ
Ul
D VMM 0-50ppm

c i—HI 5o-500ppm

  EZjQ 500-5000ppm

  Ex^?j >5000ppm

 1.  a  I '' 14'.  ' .  «.  7  .  '.. .»  I «  . " .  *
                                                                                                            Demonstration
                                                                                                                 Site
                                                    ,    ,
14  . 15 |  16 t  17 | IB ( 19 |  20  | 21   22   23   24   25   26
                                                                                                                27  28  29  30   31   32   33   34
                                                                                                                                Note: Depths are 0-12"
                                                       Figure 4-2.  Predemonstration Samples

-------
solids, moisture content, semivolatile organlcs, and sediment particle size
distribution.  The methods used for the sampling and analysis of predemon-
stratlon samples were the same as those used during the tests (see Section 6.3
for a list of analytical methods).

     Predemonstration results of total solids, oil and grease, PCS, and pH
analyses are summarized  in Table 4-1.  Total solids ranged from 29 to 45
percent solids.  Oil and grease contents  ranged from 1.5 to 7.8 percent.  No
correlation was apparent among total  solids, oil and grease, and PCB concen-
trations.  As expected,  the  sediments from  location I-ll contained PCBs well
over 5,000 ppm.  The other samples  ranged from  160 to  640 ppm of PCB.  The
range  of pH  was 7.3 to  8.2 standard units among the samples.

      Data  reported for  particle  size distribution  and  semivolatile organics
were  similar among the  drummed  sediments.  Particle size distributions were
approximately 37  percent sand,  41 percent silt, and 22 percent  clay  for each
drum.   PAHs  were  the  predominant class of semivolatiles, as  expected, based on
the literature (COE 1987,  Ebasco 1987).   PAHs detected in  each  of  the 5 drums
 included napthalene,  acenapthalene, dibenzofuran,  fluorene,  phenanthrene,
 anthracene,  d1-n-butylphthalate, fluoranthene,  pyrene, benzo(a)anthracene,
 chrysene,  bis(2-ethylhexyl)phthalate, benzo(a)pyrene,  1ndeno(l,2,3-cd)pyrene,
 dibenz(a,h)anthracene and benzo(g,h,i)perylene.

      A composite sample and a composite  sample replicate were obtained for
 overall waste characterization, from the drummed sediments selected for the
 demonstration.  Also included in these composites were aliquots from four
 drums not used in the demonstration.  The  composites were analyzed for 23
 metals, cyanide, volatile organics,  and  extraction procedure (EP) toxicity
 metals.  Analysis results for 23 metals  and cyanide are shown in Table 4-2.
 Metals and  cyanide analysis results  compare well with previous studies that
 showed high  levels of  copper, chromium,  lead,  and zinc (COE, 1987).  EP  tox
 metals data  are discussed in Section 7.1,  and  were less than RCRA regulatory
 limits.  Volatile organics  analyses  were run for 34 parameters and  the fol-
 lowing were detected at levels  greater  than  1  mg/Kg;  vinylchloride,  methylene
 chloride, acetone, toluene, 1,2-dichloroethene, 2-Butanone,  and ethylbenzene.
 The  volatile organics  data  are  not critical  to this demonstration since  the
 measured  concentrations are low relative to  semivolatile  organics data.
                                        26

-------
     TABLE 4-1.  PREDEMONSTRATION RESULTS OF TOTAL SOLIDS, OIL AND GREASE
                              PCB,  AND  pH  ANALYSES
Sample (a)
H-20
H-21
H-22
H-23
1-11
Concentration
Total Solids
(percent) (b)
29
34
30
43
35
Concentration
Oil and Grease
(percent) (b)
3.3
1.5
7.8
1.2
7.1
Concentration
PCB
ppm (b, c)
402
333
. 640
162
32,333
PH
(Range in
Standard Units)
7.3-8.2
7.6-8.1
7.7-7.8
7.8-8.2
7.3-7.6
                                                                    C°E
(b)  Represents the average of 3 grab samples taken from each drum.

(c)  Polychlorinated biphenyl  (PCB) reported as the sum of Aroclors 1242 and
                                     27.

-------
Parameter
           TABLE 4-2.  PREDEMONSTRATION RESULTS OF METALS ANALYSES
Concentration
 (mg/kg) (1)
Aluminum
Antimony
Arsenic
Barium
Berylium
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Potassium
Selenium
Silver
Sodium
Thallium
Vanadium
Zinc
Cyanide
    10,000
         8
        12
       140
         4
        29
     2,550
       425
         9
       900
     16,500
       590
     5,300
       185
          1.4
       130
     2,200
     ND  (2)
          4
     11,000
         ND
         70
      1,850
Less than 6
 Notes:  1.  Data reported are the mean of 2 grabs from a composite of
             predemonstration samples.
         2.  ND indicates not detected.
                                       28

-------
                          SECTIONS.  FIELD ACTIVITIES

      SITE Program activities were conducted for the CF Systems New Bedford
 Harbor demonstration program from June to October, 1988.  In July, the U.S.
 Army Corps of Engineers (COE) obtained drummed samples of harbor sediments.
 CF Systems conducted bench-scale tests on archived sediment samples in June
 and later, in August, conducted bench-scale tests using the COE samples.
 Public meetings were held and a Fact Sheet was distributed throughout June,
 July,  and August.  The demonstration was conducted in September.

      The bench-scale tests are described in Section 5.1.   The plans for the
 demonstration are described in Section 5.2.   A summary of the actual
 operations is described in Section 5.3.   The few operational  deviations from
 the Demonstration Plan are also discussed in Section 5.3  along with some
 additions to  the analytical  testing program that were incorporated into the
 Program while at the site.

 5.1 BENCH-SCALE TESTS

     Bench-scale tests  were  conducted, prior to  the  demonstration,  in  order  to
 determine the best operating conditions  for  the  PCU.   Bench Test 1  consisted
 of  extracting PCBs from three different  sediment samples.  In  Bench Tests 2
 and 3,  a  fourth  and  fifth sample were  run.  Each sample was divided into 3
 portions  and  each portion was extracted at different sol vent-to-feed ratios.

     The  bench-scale extraction apparatus consisted of a stainless steel,
 single-stage,  counter-current extractor with a mechanical stirrer and a
 solvent delivery system.  A nominal amount of material (between 180-190g) was
 placed in the extractor (about 0.8 liters capacity) at the beginning of a run.
 A portion of the remaining volume in the extractor was filled with solvent.
 The system was then pressurized to 150-200 ,psig at an ambient temperature of
 70-80 degrees  F:   The stirrer activated and the contents were mixed.   After  a
suitable residence time, the stirrer was switched off and  the contents  allowed
to settle.  The system was depressurized in order to permit collection  of
raffinate and  extract samples.
                                     29

-------
     Test data are shown in Table 5-1 that include the test number,  sample
number, feed PCB concentration,  and treated sediment PCB concentration for
respective sol vent-to-feed ratios.  The data generally show that higher ratios
of sol vent-to-feed used in the extraction result in higher extraction
efficiencies.  However, data for sample 2 (ratio 20:1) and sample 4 (ratio
2:1) do not show this trend.  These differences may be attributed to the range
of accuracy associated with the screening method used to analyze PCBs in each
sample.

     CF Systems used the data in Table 5-1 to roughly estimate the number of
passes required to treat wastes to specific levels in the demonstration.
For example,  the single-stage bench test results for sample 4 showed that 210
ppm could be  reduced to 17 ppm with a 20-to-l feed-to-solvent ratio.
Therefore, CF Systems estimated that the two-stage PCU would require 10
passes, or recycles, at a  1.5-to-l feed-to-solvent ratio to achieve a similar
reduction.

      Bench-scale tests were useful for confirming  the solubility of PCBs  in
the  liquefied solvent.  However,  these tests were  only  of  limited use  for
setting PCU  operating  conditions.  In  addition,  the  bench-scale tests  did not
provide data relating  to  operational and  material  handling  issues such  as
temperature  requirements,  foaming, and pumpability.

                       TABLE 5-1.  BENCH-SCALE TEST DATA
Test Number
Sample Number
1
1
1
2
1
3
2
4
3
5
 Samp!e
Sol vent/Feed
   Ratio
                                                   PCB Concentration (ppm)
Feed
Treated Sediment
Treated Sediment
Treated Sediment
(Not Applicable) 6,200
2:1
10:1
20:1
4,600
3,800
3,300
210
170
23
68
5,300
4,000
1,600
430
210
340
57
17
11,000
5,800
940
220
                                       30

-------
  5.2  OPERATIONS SUMMARY
       CF Systems'  pilot unit occupied a 60- by 100-foot staging area adjacent
  to Sawyer  Street,  as shown in Figure 5-1.  ; The staging area,  covered with an
  aggregate  base  of crushed  3/4-inch  stone  wjth a 6-inch depth,  was amenable to
  both  tractor  trailer and hand-cart  traffic.   An 8-foot high,  chain-link fence
  topped  with 2 feet of barbed  wire cordoned off the  staging  area  from the
  remainder  of  the  EPA-COE controlled  location.   An 18-foot wide gate provided
  access  to  the site.   Access to the demonstration test  staging  area  was
  provided through a hard-packed earth  roadway  running through the  EPA-COE area
  This  roadway  exited  onto Sawyer Street.  Security for  the demonstration  test '
  staging area  was provided 24 hours per day.

      The overall procedure started with preparing sediments for feeding  to the
 PCU for each test.  Sediments previously collected by COE and stored in  drums
 were sieved to remove particles greater than 1/8 inch and other debris,  such
 as leaves and sticks.  The sieving apparatus consisted of a custom-made  steel
 rim, a removable 1/2-Inch screen,  and a removable 1/8-inch screen.  Initially
 both screens were being used,  but  after several days of sieving,  it was
 determined  that  only the l/2-inch  screen was necessary  because the 1/8-inch
 screen was  not catching any additional material not  already  caught by the
 1/2-inch screen.   Later,  during test  l,  CF Systems expressed concerns that
 oversized feed material  was interfering with the unit's pumps.  In order to
 determine whether  oversized materials  had  not  been sieved  out  of  the feed
 material, the  feed  material  from the  test  was  sieved through an aluminum
 window screen  (1/8  inch) placed on top of  the  l/8-inch  sieve originally  used
 on  the drum samples.   A small  amount of humus  material  was caught  between the
 two  screens.   Sediments were shoveled  onto  the  screens,  which were placed over
 an empty  drum.  A trowel was used to push the  sediments  through the  screens
 and  harbor  water was  used to wash the  sediments  into the drums.  Feed prepared
 for  each test  was pumped into the feed kettle.

     Each of the four tests were run similarly except that number of passe-s
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
                                     31

-------
U)
ro
              o
              in
                  I ( I Jl L I I ^ I	
                 O O f«— Alternate Drum
                           Storage Area
Exclusion Zone
                                                                Propane Tanks
                                                             QQ99
                                                          Unit
                       Tarps
                               Alternate Drum
                               Storage Area
      Exclusion Zone
                                         Exclusion
                                        Zone Lines
                                                       Power
                                                        Pole
                                                                                                                      Do
                                                                          o
                                                                           North

                           Fence - 8' high topped with 2' barbed wire
                                                                                     Gate No. 1
                                                                                   Vehicle Access
                                                                         0 Fire Hydrant
                                       Figure 5-1. Plan Sketch for New Bedford Harbor Demonstration
                                                          Site Demonstration
                                                                Layout
                                                              CF Systems
                                                          Extraction of PCBs
                                                          New Bedford Harbor

                                                              Plan Sketch

-------
  sediment through the PCU constituted an additional  pass.   Recycling was

  conducted to simulate the 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:


       *'   Ihf^,,WaST£UVS/ shakedown  test to  set  pressure and flowrates in
           drL  nlh ThVond uao,a 5°-9allon composite of sediments  taken from
           nf 5m  ±PS ?h20' H~21' and H'23'   The feed had  a PCB concentration
           hancmn            passes  were  run to  gain experience  with  materials
2.
3.
          Test 2 was a 10 pass test.  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-Bfficiency process'and to   '
          achieve treated sediment levels less than 10 ppm.  A 350 ppm
          concentration was chosen for this test since this represents an
          average, or typical, PCB concentration in the harbor

          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 purpose of this test was to reproduce the results of the first
          three passes of Test 2.
          nnmnn'                  ",  Ttl8 feed- W3S 8 2,575 PPm,  299-pound
          composite of sediments taken from drum numbers I-ll and H-22   The
          purpose  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 toluene 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.


     Information was logged in a notebook to report the sampling operations
and the overall  operations at the site.  This included the following:
                                      33

-------
     •   Operating  conditions,  gage  reading, mass,and volume determinations
         I
     •   Notes  on daily  preparations of  CF  Systems  and  EPA
     •  Problems
     •  Health and Safety related procedures,  meetings,  and  concerns
     •  Chronology and summary of daily activities,  including check-in and
        check-out of all personnel
     •  Weather conditions.

5.3  OPERATIONS CHRONOLOGY
     The demonstration commenced at the New Bedford site on September 6 and
continued until September 29.  Test 1, the system shake-down, was conducted on
September 6.  Test 2 was conducted from September 8 through 18.  Test 3 was
conducted from  September 18 through 20.  Test 4 was conducted from September
20 through 29.  System  decontamination took place from September 30 through
October 4.  Highlights  of  the  site preparation, operation,  and decontamination
are  described below.
                                                                        Samples
Test 1
     The planned system shakedown, Test 1, consisted of three passes.
were taken at the following locations for PCB analyses:

     t  Feed kettle
     •  Treated sediment product  tank
     •  Extract product tank.

     The  feed was a  composite  of  sediments taken  from  COE drums nos. H-20,
 H-21, and H-23.  The feed  had  a PCB  concentration of 360 ppm.  Field personnel
 reported  that Pass 1 treated sediments  contained  some  foam  and appeared  to
 have a  very  low  solids content relative to the  feed.   A layer of  foam  several
 inches  thick formed  in the treated sediment  collection drum during Passes 2
 and 3   Since  the  foam hindered sampling, the treated  sediment was allowed  to
 degas in collection  drums overnight.  A Pass 1 extract sample was collected
 several days after the Pass 1 raffinate sample.  This  delay ocurred over a
 holiday weekend.
                                        34

-------
  Test  2                                    :

       Test  2  was  a  10-pass  test  conducted  on  a  low  PCB  concentration  (350  ppm)
  sediment.  The source  of sediments  was  the same  as for  test  1  (COE drums  nos.
  H-20, H-21,  and  H-23).  Samples were  taken at  the  following  locations:

       •  Cartridge  filter
       •  Basket strainer
       •  Feed kettle
       •  Treated  sediment product tank
       •  Extract  product tank.

      The cartridge filter was sampled as planned.  The sampling plan was
 modified to include basket strainer sample collection at the end of each pass
 rather than at the end of each test.  The basket strainer filled with
 particulate matter more rapidly than anticipated.  This procedure was repeated
 for all  passes except Pass  9 when  the sample  volume was too small.   Extract
 samples  were  taken only at  the end of Passes  1, 3,  and 10 because the volumes
 generated  with each pass were so small.   Treated sediment samples were taken
 at the end  of each of the 10 passes.   Grab samples  of tap water and  harbor
 water  also  were taken.   Field measurements included pH,  weight, and  viscosity.

     in  addition  to changes in the  sampling plan  mentioned for  extract and
 basket strainer sampling, other  deviations occurred.  For example, the feed
 material was  sampled  three  times.   The second feed  sample was obtained after
 operating personnel  added tap water  to the feed slurry.   This was done to
 determine the  effect  water  content would have on  analyses for PCBs.   A third
 feed sample,  "Pass  4  feed"  was a mixture of\ Pass  3  treated sediments  taken
 from product tank no. l and overflow collected  from product tank no.  2.  The
 second tank needed  draining to allow continued  operation.  The  drainage was
 mostly foam, which  was thought to contain material from Passes  1 and 2, as
 well as Pass 3.   Therefore,  treated sediment  collected from both tanks'was
 composited as a representative sample of the   Pass 4 feed.

     Sampling logistics were modified at Pass  4 for all  future passes because
of the foaming problem.   Since foaming in both treated sediment tanks was
                                      35

-------
significant, collection drums were positioned at each.   Treated sediments

collected from each tank would be composited in a single drum after degassing

for each pass.  Samples then would be taken from the composite of both tanks.

Foaming was a continuous problem throughout Test 2, so treated sediment was

left in collection drums overnight to degas for Passes 1, 2, 3, 5, and 8.

Saraples for Passes 4, 6, 7, 9, and 10 were able to be collected on the same

day the run occurred since several daylight hours were available to allow

degassing to occur before samples were acquired.


     Sampling personnel made the following observations on the physical

characteristics of the  various wastestreams:


     •  Treated Sediment.  Foaming caused  two  sampling problems:   (1)  treated
        sediment  would  splatter  out  of the collection drums  and   2)  represent-
        ative samples of  liquids and solids  could  not be  drawn with  a  samp  e
        beaker  Analyses  for methylene blue  active substances  (MBAS),  total
        dissolved solids  (TSS),  and  total  suspended solids  (TSS)  subsequently
        were  requested  for the  tap water,  harbor water,  and  Test  3 reed.
        Results  from these analyses  indicated  no significant levels  of MBAS,
        TSS,  or  TOS.

      •   Extract.   Only  very  small  volumes  were available for sampling.
         Laboratory personnel  observed parti culates in  each  extract sample.

      «   Basket Strainer.   With each  pass the strainer  collected  one gallon of
         liquid that contained fibrous solids and large particles.

      •  Carbon Canister.   The carbon canister had not been  changed since prior
         use of the unit at a petroleum refinery.
         Tap Water and Harbor Water.  Field personnel conducted an
         field experiment to determine if either the tap or the harbor could be
         the source of foam in the raffinate.  Aeration of a iquots of tap and
         harbor water produced no foaming.  Subsequent methylene blue active
         substance analyses indicated that neither tap nor harbor water would
         be a source of foaming agents.

         Cartridge Filter.  The cartridge filter was coated with a dark parti-
         culate as anticipated.
 Test 3
       Test  3 was  designed  to  duplicate  the  first  three passes of Test 2 and the

  test was conducted  as  such.   However,  two  events occurred that caused some
  concern about sample integrity:   (1) communications  among the CF  Systems  and
  EPA revealed  that the  unit had not been decontaminated  prior to departing from
                                        36

-------
 a prior demonstration at a petroleum refinery; and (2) potentially contami-

 nated wash water inadvertently got mixed with Pass 1 treated sediments.


      The source of sediments was the same as used for Tests 1 and 2 (COE drums

 nos.  H-20, H-21, and H-23).   Samples were taken at the following locations:


      •  Cartridge filter

      •  Basket strainer

      •  Feed kettle

      •  Raffinate product tank

      *  Extract product tank.


      The cartridge  filter was  sampled at  the end  of the  test.   The  basket

 strainer and raffinate  were  sampled  after  each of the three  passes.  Extract

 was sampled  at Passes 1 and  3.   Extract sample volumes were  small and  their

 flow  was erratic at each pass.   A  sample of  feed  was  collected  for  methylene

 blue  active  substances  analyses  to determine  if the harbor sediment, contained

 agents that  caused  foaming.  Foaming problems  persisted  through Test 3,  so

 samples  were  degassed overnight  and collected  the  following  day.  Field
 measurements  included pH,  volume,  weight,  and  viscosity.


      Several  additional  samples  were taken to  assess  the impact of  residuals

 left  in  the unit from a prior demonstration.   Site personnel attempted to wash

out these residuals after Test 3, Pass 1 with  a refined naphtha fuel.   Another

sample was taken to determine the effect of the inadvertent addition of wash

water to the Pass 2 feed.  Each of these events is discussed below.


     •  Fuel Wash - Test 3 was interrupted to clean the extract product tank
         (EPT) and the column reboiler (CR).  The CR was opened and drained for
        one hour.  The drainage,  at first,  looked like black water,  then some
        oil appeared before a yellow-orange,  frozen material  was drained   The
        mixture was sampled for PCB analysis.  The EPT top  was then  opened and
        one gallon of Coleman brand fuel,  which is a naptha-based product  was
        added.  The EPT  sample valve  was opened and a sample  of the  fuel  wash
        was taken.   Then 18 gallons of fuel,were added to the EPT.   The
        IB-gallon fuel  wash was released from the  EPT and mixed with the
        initial wash for sampling and PCB  analysis.  Subsequent sampling of
        Pass 3 extract showed that not all  of the  fuel wash had drained from
        the EPT, since biphasic characteristics were observed in the Pass 3
        extract.  Laboratory  analyses of the  fuel  wash and  column  reboiler
                                     37

-------
        drainage showed that the unit had  been  accumulating  PCBs  in  the
        extract product tank and still  bottoms  tank.   Most of  the PCBs  fed  to
        the unit apparently were not being transferred from  the  reboiler  to
        the extract tank with each pass.   Therefore,  closure of  a PCB mass
        balance for each pass based on feed,  raffinate,  and  extract  samples is
        not possible.   Similarly, a total  mass  balance closure for each  test
        is not possible.
        Wash Water - During Pass 2 the hose of  the feed pump was placed  in  a
        wash water drum and 20 to 30 gallons of wash water were  accidentally
        pumped into the feed kettle.  The  wash  water contained residue  from
        previous samples of treated sediment and feed material.   The wash
        water was mixed with Pass 1 treated sediment and the mixture was
        sampled as feed to Pass 2.  Laboratory  analyses showed that,  this
        incident had little effect on the  demonstration.  The  Pass 1 treated
        sediment and Pass 2 feed PCB analyses differed by less than  35
        percent.
Test 4
     Test 4 was conducted, as planned, on a high PCB concentration (2575 ppm)
sediment.  Six passes were run in order to ensure that the final treated
sediment PCB concentration would be equal to or less than the Tests 2 and 3
feed concentrations.  Treatment beyond those concentrations would be inferred
by Test 2 and 3 results.  Test 4 feed was a composite of sediments taken from
COE drums nos. 1-11 (one-third) and H-22 (two-thirds).  Samples were taken at
the following locations:

     •  Cartridge filter
     •  Basket strainer
     •  Feed kettle
     •  Treated sediment  product tank
     •  Extract product tank.

     The  cartridge  filter was  sampled  at the end  of  the test.   The basket
strainer  and  treated  sediment  were  sampled  at  each of the  six  passes.   Extract
was sampled at Passes 4 and  6.  Treated  sediment  foaming problems persisted
throughout the test and extract volumes  continued to be small.   Additional
samples  were  shipped  as part of the effort  to  compare analytical  results from
analysis  methods  8080 and 680.  Field measurements  included  pH,  volume,
weight,  and viscosity.
                                       38

-------
      Test 4 also included air sampling and analysis to determine the release
 of PCBs from sample collection containers to the atmosphere.  The potential
 release of PCBs to the air from the low temperature process was assumed to be
 insignificant based on the very low vapor.pressure of PCBs.  Personal air
 sampling pumps were mounted on the side of the collection containers.  During
 mixing of collection container contents,  the container was covered with an
 aluminum foil  and gases evolving from the container were pumped with the
 personal air sampler through a sorbent tube.   This procedure was conducted
 during the degassing of Passes 1,  2,  3, arid 4 treated sediment and Pass 4
 extract.  Sorbent tubes were shipped  for  PCB analysis.

 Decontamination

      System  decontamination was  conducted with toluene  used as the feed.
 About 100  gallons of toluene were  added to  the feed kettle,  then collected in
 the  treated  sediment and extract product  tanks.   The developer defined  a  feed
 volume  of  30 gallons,  or 10 percent of the  extraction system capacity,  as
 sufficient to  run a  pass through the  extractor.   Therefore,  the  equivalent of
 three single passes  of fresh  toluene  was  run  through the  system  (100  gallons/
 30 gallons per  pass  =  3.3  passes).  Toluene wash  was collected from the
 primary and  overflow treated  sediment  tanks.   Extract was collected in  two
 batches and  contained  mostly  toluene.   All toluene  wash samples  were  analyzed
 for  the purpose of determining the  mass of. PCBs that had  accumulated  in the
 treated sediment  and extract  product tanks and related equipment.  At the  end
 of tank draining, grab  samples were taken at  the  drain valves  to establish
 that  the unit had been  decontaminated.

      The demonstration  staging area shown in  Figure 5-1 was not contaminated
 with  PCBs  that may have been  released during  testing and  sampling.  On August
 11, 1988, soil  samples were taken from 10 locations  in the staging area.
 Samples of soil were taken at zero to six inches of depth.  Analytical results
 showed  PCB levels to range from less than 5 to 37 parts per million in the
soil. The 10 locations were resampled on October 6, 1988,  after CF Systems
 removed their equipment from the site and  SAIC removed debris from the site.
Analytical  results showed PCB levels to range from less  than 0.5 to 5.4 parts
per million in  the soil.  At nine of ten locations, post-demonstration values
                                     39

-------
were less than predemonstration values.  This difference may be the result of

variability associated with sampling and analysis.


     Debris was collected in 55-gallon drums over the course of the
demonstration.  The following materials were produced during the demonstration

and removed from the staging area by Clean Harbors, Inc.
     Number of Drums

            6
            6
            2
           15
            8
           20
            8
           22

    Total  = 87
          Drum Contents

Toluene (unit decontamination residue)
Toluene rinsewater
Naphtha-based fuel product and unit residue
Sediments
Sediments and water
Decontamination water
Tyvek suits and water
Clothing and gloves
 In addition to these, 14 drums containing harbor sediments obtained from

 previous COE activities were also removed.
                                        40

-------
                  SECTION  6.   SAMPLING AND 'ANALYTICAL PROGRAM  •

6.1  SAMPLING LOCATIONS
     A detailed description of major process equipment and the process flow
diagram were presented in Section 3.  As discussed,  feed sediment is treated
for organics removal, then the treated sediment is discharged.  PCBs and other
organics separated from the feed sediments are discharged as extract.   Through
a vapor recompression cycle solvent is reused.  Minor solvent losses occur due
to venting.  Vented emissions pass through a carbon  canister after the uni't is
shut down.  Other material losses occur  as a result  of the use of particulate
filters that are used throughout the system to protect operating equipment.
Materials also exit the unit during system shutdown  and cleanout.   These
materials are residues that have adhered to internal  surfaces of tanks,
piping, and equipment and that are removed by use of a washing medium,  such  as
toluene or another nonpolar solvent.  Figure 6-1 is  a simplified diagram of
system inputs and outputs.  Each system  influent,  effluent,  or loss  is  defined
and described below.
              Feed Sediments
                                      Carbon Canister
Treatment System
                                       1
                                     Treated
                                   Sediments
                  I
               Extract
                             Figure 6-1. System Flow Diagram
                                     41

-------
Feed Kettle (FK) -- The feed kettle holds a maximum of approximately 80
gallons of slurried sediments.  Contents were continuously agitated to provide
a fairly homogenous feed.  Feed drum contents were continuously agitated and
pumped into the feed kettle.  Grab samples were taken, during agitation, with
a stainless steel ladle.  Feed weight was recorded for Tests 2 through 5.
Analyses were conducted during Tests 2 through 4 for PCBs, semivolatiles,
cadmium, chromium, copper, zinc, lead, oil and grease, total solids, pH, and
viscosity.  Only PCBs were analyzed during Test 1.  Test 5 used only toluene
as  feed for purposes of decontamination.

Treated Sediment Product Tank  (RPT) — The treated sediment, or raffinate,
product tank was accessible at  its outlet, a point underneath  the  trailer.
During the demonstration, it  was drained  and product  was  collected in  either  a
55-gallon or an 85-gallon drum.  At the conclusion of Tests  2  through  5, the
following were  determined;  PCBs, semivolatiles, cadmium,  chromium,  copper,
zinc,  lead, oil and  grease, total  solids,  and  pH.  Only PCBs were  analyzed
during Test 1.   During  Test 4,  gases  evolving  from the sampling containers
were sampled using National Institute of  Occupational Safety and Health
 (NIOSH)  physical  and chemical  analytical  method 253.   Sorbent  tubes were used
to collect analytes contained in gases emitted from  the unit.   Decontamination
 residue  was also collected  at this point.

 Extract Product Tank (EPT)  -- The extract product tank was accessible at its
 outlet,  a point underneath  the trailer.   Extract  drained  and collected in a
 covered 5-gallon,  stainless steel  pail during the demonstration.   At the
 conclusion of Tests 2 through 5, the following were determined:    PCBs,
 semivolatiles, cadmium, chromium, copper, zinc,  lead, oil and grease, total
 solids, and pH.  Only PCBs were analyzed during Test 1.   During Test 4, gases
 evolving from  the sampling containers were sampled using National  Institute of
 Occupational Safety and Health (NIOSH) physical and  chemical analytical method
 253.  Sorbent  tubes were used to collect analytes contained in gases  emitted
 from the unit.  Decontamination residue  was also collected  at this point.

 Basket Strainer (S-l) — The basket  strainer  prevented oversized  feed  material
 from  entering  the system.  For each  pass during Tests 2  (except Pass  9)
 through 4, CF  Systems  removed  the basket strainer.   Test 2, Pass  9,  was not
                                        42

-------
  sampled because sample volume was not sufficient.   Solids were scraped with a
  stainless steel spoon or spatula from the basket strainer and drained the
  strainer casing into a stainless steel  pail  for a weight determination.
  Solids  were shipped for PCB and total  solids analyses.

  Cartridge Filter  (F-2)  - The cartridge filter  removed  fine  particles in  the
  submicron range from the solvent-organics 'mixture.   CF  Systems  installed
  cartridge filters/and  at the conclusion  of  each pass removed the  filters  and
  scraped  solids  for  a mass determination.

  Carbon Canister (CC)  -  Propane  blowndown from  the PCU  at the conclusion of
  the tests and the decontamination procedure was vented  through a 20-gallon
  carbon canister.  Following decontamination, CF Systems  removed the carbon
  canister  from the treatment system.  The carbon canister was emptied with a
  shovel.   At 16 equally spaced distances along the canister height, aliquots
 were obtained with a stainless steel beaker.   The beaker was used to mix the
 surface  of the carbon bed as each aliquot was obtained.   The 16 aliquots were
 composited in a stainless steel pail and then mixed prior to filling sample
 containers.   Sample weight, total solids,  and PCB concentration were
 determined.

      Additional  planned sampling was conducted as  part of the Health and
 Safety Plan  to  ensure worker safety  and  to determine  if  releases  to the
 environment  from process equipment,  open tanks,  and sample collection
 containers occurred.

 6.2  SAMPLING SCHEDULE

     The  sampling frequency  scheme planned for Tests 2,3, and 4 is shown in
 Table 6-1.  Sampling  was  designed to allow determination  of:   (1) a PCB
 extraction efficiency  with each pass and (2) a mass balance with each pass.
 Sampling planned for decontamination was similar to that  planned for Tests 2,
3, and 4, except that no chemical analyses were conducted for the feed that
was composed of toluene.
                                     43

-------
                TABLE  6-1.   SAMPLES  FOR  CF  SYSTEMS NEW BEDFORD
                              TESTS 2, 3, AND 4
Sample Location
  Sample
Frequency(l)
Cartridge Filter
Basket Strainer
Feed Kettle
Final Pass


Each Pass


First Pass
 Treated  Sediment Product Tank   Each Pass
                                 Final  Pass
 Extract Product Tank
 Each Pass
                                 Final Pass
                                                  Parameter
PCBs
Total Solids

PCBs
Total Solids

Volume and Weight
PCBs
Semivolatiles
Cd,  Cr, Cu, Zn, Pb
Oil  and Grease
Total Solids, pH
Viscosity

PCBs
Volume  and  Weight

PCBs
Volume  and  Weight
Semivolatiles
Cd, Cr,  Cu, Zn,  Pb
Oil and Grease
Total  Solids, pH

 Volume and Weight.
 PCBs
 Total  Solids

 PCBs
 Volume and Weight
 Semivolatiles
 Cd, Cr, Cu, Zn, Pb
 Oil and Grease
 Total Solids, pH
  Note  1-   Test  2  involved  10  passes,  Test  3  involved 6 passes, and Test 3
           involved  3  passes through  the  PCU.
                                        44

-------
e
a
0
  6.3  ANALYTICAL METHODS AND PHYSICAL TESTS
       The analytical methods selected for the demonstration at New Bedford
  Harbor included:

       •  PCB by EPA Method 3550/8080 (EPA 1986)
       •  PCB by EPA Method 3550/680 (EPA 1983, EPA 1986)

       *  1984)Xtra°ti0n  Modificat1on by  Spittler  Screen  (Fowler  1987,  Spittler

       •  Waste  Dilution  by EPA Method 3580; (EPA 1986)
       •  Semivolatiles by  EPA Method 3550/8270 (EPA 1986)
       •  Trace  Metals by EPA Method  3050/6010 (EPA 1986)
       •   Particle Size by  ASTM Method C-136-84A (ASTM 1984)
          Total Recoverable Oil and Grease by EPA Method 9071 (EPA 1986)
         Percent Solids by EPA Method 160.3 (EPA 1983)
         EP Toxicity (metals) by EPA Method 1310/6010 (EPA 1986)
      •  pH in Calcareous and Noncalcareous Soils  by EPA Method 9045 (EPA 1986)
      •  Total Metals by  EPA Method 3010  (EPA 1986).

      All  methods were intended  to be performed as published,   in the event
 that modifications  to the  analytical program were required due to matrix
 constraints,  only approved methods were  substituted.  Any  deviations from
 approved  protocol for  these  methods  are  discussed in  Section 7.   The Spittler
 Screen and  a  comparison  between Analytical  Methods 8080 and 680  are  discussed
 below.

 Spittler  Screen

     The  Spittler Screen was used  to  identify the approximate  level  of PCBs in
 a sample  to aid analysts in choosing dilution factors (Spittler, 1984)
 Subsequent analyses by EPA Methods were then carried out more efficiently.
 The Spittler Screen provided rapid analytical turnaround for large volumes of
samples to optimize those subject to further analyses.  Rapid turnaround of
analytical results allowed CF systems to  more closely control  operating
                               45

-------
conditions.  This screening procedure was not be used for any critical
measurements.

Comparison of Analytical Methods 8080 and 680 for PCBs
     Method 8080 is a Resource Conservation and Recovery Act (RCRA) analysis
method for determining PCBs as Aroclors.  The analysis is conducted by gas
chromatography with an electron capture detector (GC/ECD), and quantitation is
performed  by external standard calibration against one or more Aroclor stan-
dards.  Pesticide compounds are used  as surrogate standards  for estimating
analytical accuracy.

     Method  680  is  an analysis method for  determining  PCBs as  congeners
 (homologous  classes based  on  degree of chlorination) under the Clean  Water  Act
 (CWA).   The  analysis  is conducted  by gas  chromatography  with detection by mass
 spectroscopy (GC/MS), and  quantitation is-performed  by internal  standard
 calibration.  Carbon  13-labeled  PCB compounds (one  from each congener) are
 used as surrogate standards for  estimating analytical  accuracy.

      Method 8080 was favored as the analysis method for the  SITE demonstration
 at New Bedford Harbor because:

      o  Method 8080 is a RCRA method and Method 680 is not
      o  Procedures for extraction  of PCBs from sediments have been developed
         and validlted for Method 8080, while Method 680 only  addresses  water
         samples
      o  Because the sediments from New Bedford Harbor were  known  to  contain
         PCBs, and  the PCB content  has been thoroughly character zed  by  COE,
         the added  specificity of GC/MS  (Method 680) was not required
      o  Method  8080 costs  less than Method 680.

      Method 8080 was finally chosen after  a  review  of the analytical results
  for Test  4  samples.  In Table 6-2, Method 680  analytical  results  are shown for
  Test 4  feed  and Pass  4 treated  sediment samples,  and are  reported as congener
  groups rather than the 209 individual congeners.   The percent of  PCB removed
  from  the  feed after  the fourth  pass is calculated  for each  congener group.
  The removal data show  that the extraction technology removes less chlorinated
                                        46

-------
              TABLE  6-2.   METHOD  680 ANALYTICAL RESULTS FOR TEST 4
 PCB  Congener    Feed
   Group        (ppm)
                Pass 4
                Treated
                Sediment
                  (ppm)
            Removal
           (percent)
                 Congener  Group
                as  a  Percentage
               of the Total  Feed
                   (percent)
Mono-

Di-

Tri-

Tetra-

Penta

Hexa-

Hepta-



TOTAL
      39

   1,150

   2,800

   3,000

   1,400

     260

Not detected



   8,700
  0.58

 30

 98

130

 69

 18

 12




350
    99

    97

    97

    96

    95

    93

Not applicable



    96
Less than 1

13

32

34

16

3

Less than 1
                                                         Not applicable
                                     47

-------
PCBs (e.g., mono- and dl-) more efficiently than more chlorinated PCBs (e.g.,
penta- and hexa-).  However, this difference between congener groups is not
significant since all calculated removals differ from the mean removal by less
than five percent.  Furthermore, the monochlorobiphenyl group composes less
than one percent of the total mass.  Method 8080 was used to measure Arodors
1242 and 1254, since these  represented nearly 70 percent of total PCBs
contained  in the sediments  (COE, 1987).

6.4  PROCESS CONTROL AND  FIELD  MEASUREMENT DEVICES
     Process control and  field  measurement devices  used during the  demonstra-
tion are listed  in Table  6-3.   Also  shown  in Table  6-3 are measurement
locations,  rationale for  measurement,  measurement frequency, measurement
units,  and precision and  accuracy.   The  developer's extractor pressure gages
were calibrated  against nitrogen gas regulated  by an Airco model  0-4000
pressure  regulator/gage at  145  and 261 psig.  The Airco  pressure  gage was
subsequently  calibrated by  Calibration Central, Inc.,  of Herndon, Virginia.
The extractor  temperature gage  was a Sybron  Corporation,  Bi-Therm model.   The
 feed slurry and  cooling water temperature gages were manufactured by Universal
 Enterprises.   The temperature gages were calibrated in an equilibrated mixture
 of ice and water at 32 degrees F and electrically  heated water  at 100 degrees
 F.  Feed kettle slurry volume was estimated using  a calibration curve
 developed by CF Systems.   A Micro Motion, high-pressure Model  D Meter measured
 solvent flow.   The mass flowmeter employed magnetic fields in its operation
 and was calibrated electronically.  Cooling water  flow from the system was
 measured via the classic stopwatch-and-bucket  method.

      Feed slurry viscosity was  measured with a Haake Model Vt~02 cup-and-
 spindle viscometer.  Feed  slurry pH was measured by a Corning Model pH 106 pH
 electrode.  Feed, treated  sediment, extract, strainer solids, filter residues,
 and other process materials were weighed on one of  two scales.   A  Howe-
 Richardson HCR  international XL 1000  Ibs. x 1/2 Ib. scale was used to weigh
 materials  greater than 25  Ibs.,  and a Hanson 25 Ibs.  x  1 oz. General Household
 Scale  was used  to weigh  materials  less  than 25 Ibs.   The electric  power was
 metered by standard residential-use unit supplied  by  the local electric
 utility.
                                        48

-------
TABLE 6-3.  PROCESS CONTROL AND FIELD MEASUREMENTS
Critical
Measurement
PROCESS CONTROL
MEASUREMENTS
Extractor
Pressure
Extractor
Pressure
Extractor
Temperature
Feed
Slurry Volume
Feed Slurry
Temp.
Solvent Flow
Cooling Water
Viscosity
Feed Slurry pH
Measurement
Location
E-l Press.
Gage
E-2 Press.
Gage
E-l Temp.
Gage
FK
FK
Control
Room Readout
Outlet Line
Feed Drum
Immediately
Prior to Use
FK
Rationale for
Measurement
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Verify Operating
Conditions
Measurement
Frequency
10 min.
10 min.
10 min.
10 min.
Start of
Pass
10 min.
10 min.
Start of
Test
Start of
Pass
Measurement
Units Precision Accuracy
PSig 1% (1) 5% (2)
Psig 1% (1) 5% (2)
degrees F 0% (i) n% (2)
gal/min. (3) (4)
degrees F 5% (1) 2% (2)
Ibs/min. (3) o.2% (5)
GPM (3) 0.25% (2)
dPa.S (3) 5% (2)
PH 0.01 units (6) 0.01 units

-------
                            TABLE 6-3.   PROCESS  CONTROL AND  FIELD MEASUREMENTS (Continued)
Critical
Measurement
MASS INVENTORY
MEASUREMENTS
Feed, Treated
Sediment Weight
Extract Weight

o strainer

Filter Residues

UTILITIES INPUT
MEASUREMENTS
Electric Power
Cooling Water
Temperature
Measurement
Location


Feed Weight
Scale
Low
Weight Scale
Low Weight
Scale
Low Weight
Scale


Input Line
to Pilot Unit
Input and
Output Streams
Rationale for
Measurement


Mass Inventory

Mass Inventory

Mass Inventory

Mass Inventory



Utilities
Measurement
Utilities
Measurement
Measurement
Frequency


Start of
Pass
End of Test

End of Pass

End of Test



During Pass
10 min.
Measurement
Units Precision


Ib 1% (1)

Ib 1% (1)

Ib 1% (1)

Ib 1% (l)



kw-hr (6)
degrees F 5% (1)
Accuracy


1% (2)

1% (2)

1% (2)

1% (2)



(5)
2% (2)
NOTES:  (1)  Precision as Relative Standard Deviation (RSD),
        (2)  Accuracy as Relative Percent Difference (RPD).
        (3)  Precision data not available.
        (4)  Accuracy data not available.
        (5)  Accuracy as reported by manufacturer.
        (6)  Precision as reported by manufacturer.


-------
                                    SECTION 7
                              RESULTS  AND  DISCUSSION

  7.1   SYSTEM  PERFORMANCE

       The  evaluation  criteria established in  Section  1.3  for  an  evaluation of
  system performance were:

       *  PCB  concentration  in sediments before and after  treatment
       •  PCB  extraction efficiency with each  pass of  sediments through the PCU
       t  Mass balances established for total  mass, solids, and PCBs.

 These criteria are discussed with respect to analytical results below.

 PCB Concentration Reductions

      PCB analyses for feed sediments and treated sediment, conducted for
 samples collected at each pass,  are shown in Table 7.1.  The data are
 displayed graphically in  Figures 7.1, 7.2, and 7.3.   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:
Test
 2
 3
 4
 Percent Reduction
in PCB Concentration
          72%
          92%
Number of
 Passes
    10
     3
     6
     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
                                      51

-------
   TABLE 7-1. PASS-BY-PASS PCB CONCENTRATIONS AND REDUCTION EFFICIENCIES
Test
Number
2
2
2
£^
2
o

-------
                 Mean PCB Concentration, (ppm)
Mean PCB Concentration, (ppm)
Mean PCB Concentration, (ppm)
u>
0
-

"°_

—

s*-
tu
a
o
3.
T3
D>
en
en
2 0)
O



"
«.
-
_^
o
08t§8§§8§8§8§
i i i i i t i i i i i i j^
	 	 "
*>"'""
i
i
i
-

-
_4
S X 4^
n5 3-
i1* 51 "
1" lo,

o" °
3


».
-
.
O-
°OOOOOOOOOOi
1 1 I i I 1 1 1 1 1 1
1
t
t
V
t
V












o
§
o
-

l».

-

(0 § .Ct-
•+• 3
U 0
Q.^ Z 0>.
c .
s
o
3
00-
-

/"* —
en -L ~+ ro ro co (
3 S O Ol O 01 O i
3 ° 0 0 0 0 0 <
1 1 1 1 I 1 h
^^•*— * *"
r"""""
/
j£.
/
/
V
\
1
jt
1
1
1
r
/
\
\

J
fl
3






s1
a
Sa
si
?5
8-
o
s.
o
s






-------
at Test 3, passes 2 and 3.   These anomalies are not related to the extraction
process.  Instead, they reflect cross contamination within system hardware 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 PCBs 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 concentrations

coincide, then a PCB reduction curve can be plotted.  Data are displayed

below, side-by-side, so that similar concentrations overlap.


                        Pass-by-Pass  PCB  Concentrations

                      Test 4         Test 3         Test 2

                       2,575
                       1,000
                         990
                         670
                         325
                         240
                         200
288
 47
 72
 82
                                                     350
                                                      77
                                                      52
                                                      20
                                                      66
                                                      59
                                                      41
                                                      36
                                                      29
                                                       8
                                                      40

Based on the presentation of the data in Figure 7-4,  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 efficiencies are shown in Table

7-1 and are calculated as PCB extracted divided by concentration at the
                                       54

-------
2600
2400 —

2200 —

2000 —
1800 —
^^^^
f[ 1600 —
a.
c 1400 —
g
£ 1200 —
§
c 1000 —
£j
CO 800 —
2
600 —
400 —
200 —

0 —


i Legend
| - - D Testa
1 * Test3
\ - - • • m Test 4
\
\
\
\
\
\
\
\
V \^
\ •••••-.
\
\
x .„''„,'
X
---s^ *• -
a """""IT- — » 	 ,

I I | 	 T" ~H" ""
                                                  10
                                                          12
                                                                  14
                                   Pass Number
                       Figure 7.4. Potential Pit Cleanup Unit PCS Reduction

beginning of the pass  (multiplied  by  100 percent).   For each test, the first
pass results in efficiencies greater  than 60 percent.   However, 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.                       .  r   	,

     Data show that  the system  irregularly retained and discharged treated
sediments.  For some passes, as much  as  50 percent  of  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 efficiency 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.
                                      55

-------
     The occurrence of cross contamination affects interpretation 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 7-5 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  residuals  remaining on the surface
of the  feed drums.  Outputs from the system included  samples, spills, con-
tainer  residuals,  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.   Appendix A  shows mass
Inventories for each  test.  The  items  listed in Appendix  A correspond to the
depicted  in Figure 7-5.  Total  mass and  total  solids  are  shown  in  the tables
as  pounds.  The mass  of  PCB is  shown  as  grams.

      The  amount of material  accumulated  or the amount of  accumulated  material
 discharged are shown  for each pass in Table 7-2,  "Accumulation  or  Losses."
 This term was calculated as total  feed minus output.   The unit irregularly
 retained and  discharged material  throughout Tests 2,  3, and  4.   No correlation
 could be established between the mass inventories and extraction efficiency
 for each pass.  The mass balances for PCB and total  solids are discussed in
 more detail below.

 PCB Balance
      Table 7-2 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
                                       56

-------
               Inventory  Sheet
          Test	           pass
          1. Feed Material
                                   6. Water
 2. Sampling••
  3. Strainer
   4. Spills
5. Residuals
                     S.Treated Sediment          Accumulations and Other Losses

         Figure 7.5 Illustrative Inventory Sheet
                         57

-------
            TABLE  7-2.   MASS  ACCUMULATION  AND  LOSS IN THE SYSTEM
Accumulation (Loss) in the
Test
2
o
£+
2
2
2
«*
2
2
2
2

3
w
3


4
4
~
4
4
*T
4

Pass
1
2
3
4
*T
5
6
7
8
9
10
Subtotal
1
2
3
Subtotal
i
•L
2
3
4
5
6
Subtotal
Total Mass
(Pounds)
122
55
(25)
78
22
68
(51)
(7)
(16)
9
254
24
58
29
in
5
(83)
74
(80)
106
(53)
(31)
Total Solids
(Pounds)
39
6
(16)
32
(6)
3
(1)
(11)
(3)
(3)
"40"
(13)
6
6
T
10
(12)
9
4
6
(3)
IT
System
Total PCBs"
(Grams)
14.21
0.70
0.50
(0.22)
(0.07)
0.3
0.04
(0.07)
0.29
(0.54)
15 . 14
6.28
1.42
(0.99)
6 . 71
37.79
(5.25)
8.72
2.55
1.63
(3.33)
42 . 11
Note:  Parentheses indicate a loss or discharge from the system.
                                      58

-------
 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).

      Appendix B shows mass  balances for the naptha-based fuel  and the toluene
 wash, respectively.  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 demonstration  indicate the
 possibility of sampling  and  analytical  error.  For example,  laboratory
 precision  and accuracy criteria were 20 and 50 relative  percent  difference,
 respectively.  In  addition,  quadruplicate grab samples were  collected of the
 Test 3  feed, the Test 4  feed, and the Test 3  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.
                                     59

-------
     In spite of the calculated PCB imbalance,  a positive  separation  of  PCB
from the harbor sediments was accomplished.   The mass balances in Appendix A
show 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 (Appendix B) show that some PCB accumulated in system hardware.   However,
91 percent of the PCBs contained in decontamination residues were contained  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  generate  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  approximately  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 significant.   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 decontamination procedures.   The 285 pounds of activated
 carbon contained in the canister  collected less than 1 gram of PCB.   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 intermittantly throughout
 the tests.  This behavior is demonstrated by tracking the sediment solids.
 Figures 7-6,  7-7, and 7-8 show the pass-by-pass throughput of solids for Tests
 2,  3, and  4.  The mass of solids  accumulated on a pass-by-pass basis is
 significant.  The  flow of solids  per pass  ranges  from 55  percent accumulated
  to  150 percent discharged.   There is no  consistent correlation  between  solids
  retention  and PCB  concentration reduction.
                                        60

-------
      Percent Throughput, Output/Input
Percent Throughput, Output/Input
                                                                                                     Percent Throughput, Output/Input
 ro
m
&
3
a
o
.3
TJ

M
(A

O
o -
                     CO  SO O -l
                                              o
                                              3
                                              tn
                                              Z
                                                 03 -
                                                 o -
                                                                                                   oo o o o oo o oo oooooo


                                                                                                                   I—1	I  I   I  i  I
                                         £

                                         3
                                         o_
                                         o'
                                         3
                                        03
                                        W
                                        W
                                                                                                 w-
              tndqBnojm
         indu.Bnoju.i spnos
                                                                                                     ssej JOd indijBnojm sp

-------
     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
Semi volatile Organics
      System  feed,  final  treated  sediment,  and  final extract  were  sampled for
base/neutral and acid extractable organics (semivolatiles)  during each  test
 for the  purpose  of (1)  characterizing  materials  for disposal  and  (2)  observing
 any extraction  of semivolatiles.   Interpretation of the  semivolatiles data,
 shown in Volume  II,  is  limited for two reasons:   (1)  the unit contained
 sludges  from a  previous demonstration 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  and 2-methylphenol were found in treated sediments and extracts
         but were not measured in  feed drums, the feed kettle, or toluene
         washes.
      •  Test 4  resulted in a  reduction of  1,4-dichlorobenzene and pyrene, but
         chrysene  and bis(2-ethylhexyl phthalate) were increased.  Similar
         inconsistencies occur for Test 2  and 3.
      •  2-Chlorophenol, 1,3-dichlorobenzene, and benzo(k)fluoranthene were  fed
         to  the  unit but not  detected  in any system effluents.
                                        62

-------
  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 7-3 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 extracts were one to two orders of, .magnitude less than treated
  sediments.

  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 characteristic of
  EP Toxicity is shown in Table 7-4.  Also shown are analytical  results for
  (1) two samples taken from a composite of drummed harbor  sediment collected  by
 COE during the waste presampling and (2) a :sample of demonstration Test 4
 Pass 6 treated sediment.   Concentrations for each sample shown are less than
 the regulatory maximum for the definition of the EP toxicity characteristic.

 7.2  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  operating  data  are  shown
 as well  as  the  range  of data  recorded  for each mean value.  Generally   the
 technology  accommodated wide  ranges  of operating  conditions,  although'precise
 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
•mmmum pressure, 190 psig, was above the 180 psig minimum specification
                                     63

-------
                                  TOBLE 7-3.  t€T«JS CONTENT OF FEED, TREATED SEDIMENT, HO EXTRACT
Parameter Units
Cadmium, ppm
Chromium, ppm
Copper, ppm
Lead, ppm
Zinc, ppm
Total Residue, %
lest 2
Feed
35.7
596
1790
619
2150
23.3
Test 2
PassS
Treated
Sediment
32.5
581
1£50
537
2220
1S.2
Test 2
Pass 4
Feed
44.0
761
1S30
792
2680
15.0
Test 2
Pass ID
Treated
Sediment
42.8
816
1740
892
2610
9.4
Test 2
Pass ID
Extract
NR(1)
3
5(2)
NR(1)
5(2)
NR(3)
• —
Tests
Feed
32.0
525
1320
520
1900
19.4
.11 ii 	 — *
Tests
PassS
Treated
Sediment
62.3
1020
2570
1100
3550
10.3
TestS
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
1SCO
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 sanple volume for analysis method.

-------
    TABLE 7-4.   EP TOXICITY CHARACTERISTIC  OF TREATED  AND  UNTREATED  SEDIMENTS

                            UNITS  (PARTS  PER  MILLION)
               Composite Sample of
             Waste Presampling Drums
             Sample 1       Sample 2
  Treated
  Sediment
Test 4, Pass 6
Maximum Concentration
   Allowable for
 Characteristics of
    EP Toxicity
Arsenic
Barium
Cadmium
Chromium
Lead
Mercury
Selenium
Silver
.. 	
0.011
0.16
0.11
0.18
0.34
<0.0002
<0.005
<0.015
— — — — — — — _
0.008
0.15
0.12
0.098
0.23
<0.0002
<0.005
<0.015
—————— _______
<0.005
0.36
0.30
0.053
0.16
<0.0002
<0.02
0.015
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.
                                    65

-------
Because pressures were so stable, no relationship between extraction
efficiency and extractor pressure was apparent.

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 minimum 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 extraction efficiencies
 calculated  for  Test 2 to the  low feed temperatures,  although  other  factors
 were probably important.   These factors include cross  contamination in  the
 treated sediments collection  tank.  In  addition,  reentrainment  of  solvent in
 decanter underflows may have  caused disproportionately  large  effects on low
 concentration sediments.  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,  throughout the 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  as shown
  1n Figures 7-9 through 7-14.  Because  of  this wide  variation,  it  was suspected
  the flow meter was malfunctioning.   In Test 4, an alternative  measuring device
  was used and flow measurements continued  to show wide  variations,  as seen in
  Figures 7-13 and 7-14.

       The variable solvent flows caused the solvent/feed ratio also to
  fluctuate widely.  This ratio was  calculated as solvent (lb/min)/feed
  (gpD/feed density  (Ib/gal).   The  minimum solvent-to-feed ratio specification,
                                         66

-------
                           Figure 7-9
                    Mean Solvent Flowrate
                             Test 2
                  Figure 7-10
           Mean Solvent/Feed Ratio
          -i—I—i—I—I	i    i    i   i   i
5   1.0
              23456789  10
                    Extraction Pass Number
                         Figure 7-12
                  Mean Solvent/Feed Ratio
                        "  Tests
                                               — Norn
      0123
                    Extraction Pass Number

       Notes: Flowmeter manufacturer reports 0.2% accuracy.
            Mean reported with range of recorded data.

2.5
2.0


0.5
0.0




I-

1






1










1



Tes






it 2















-
i









1







                                                                                                                      Max
0123456789   10
             Extraction Pass Number


                    Figure 7-13
           Mean Solvent Flowrate Plus

22
20
18
16
| 14
*£ 12
E
I 10
8
6
4
2
0
Test 4
-

-
-
-












i


-
I


•














     12345
            Extraction Pass Number
x - Alternate Solvent Flowrate
                       Figure 7-11
                 Mean Solvent Flowrate
                          Tests
            1            2            3
               Extraction Pass Number

                   Figure 7-14
          Mean Solvent/Feed Ratio Plus


2.5
I
o
1
0,5
0.0
Alternate solvent/Feed Ratio 	
Test 4









!
i

X




1


1



0123456
            Extraction Pass Number

x . Alternate SorventfFeed Ratio

-------
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 correlation 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 liquefied
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 as shown in
Figures  7-15 through 7-17.   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
volumes 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 minimitn specification,  10 percent.  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 untreated 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  and  8.5  standard  units.   This  narrow band
                                        68

-------
                Figure 7.15                                   Figure 7 16
Feed/Treated Sediments Solids Concentration   Feed/Treated Sediments Solids Concentration
                                                               Test #3
Test #2
                Figure 7.17
Feed/Treated Sediments Solids Concentration
                 Test #4



X
CO
5












S 5*
3 C
•» C






• c
• : . : 2
i
f

-





I
I 1 1 I I j
1
1
I"
1
1
l"

. :


• ••


1


•

S 8s
3 C
_ w

-


- co

_
o
Z
-*> 9
CO
Q.
C
,0
tj
« 2
™ 4^
111
CM
O
>
>







X
CO
2







§ 1
•• u
T 	
1.
1
1
1
1 '
I
1
1 " .
j !









i i i 	 1 — • —
J s£ vp sP «^P v
d^ 0s 0s 0s- o
> O 0 0 O C
5 W •* P5 CM T













1

I t
_. \^r
T—

-co

o'
Z
- (D 0)

-------
fell within the 6 to 12 specification range.   The developer established this
range to prevent corrosion to PCU hardware.

7.3  DEVELOPER'S GOALS
     Since a feasibility study had not been completed for the New Bedford
Superfund site, the following treatment goals were set by the developer:

     •  Test 1 - Shakedown test only.
     •  Test 2 - Reduce PCB content of a 350 ppm feed by at least 90 percent
        or to a concentration below 5 ppm after the tenth pass.
     •  Test 3 - Reduce PCB content of a 288 ppm feed by at least 50 percent
        or to a concentration below 50 ppm after the third pass.
     .  Test 4 - Reduce PCB content of a 2,575 ppm  feed to the  J^ed levels
        observed in Test  2 and 3  after 6 passes.   (288 ppm was  the lower  feed
        of Tests 2  and 3).
     .  Test 5 - Decontamination  process.  Achieve  effluent concentrations
        less than 50  ppm  in  final  toluene washes  collected from the extract
        and treated sediment  subsystems.

     Goals for  Tests  3 and  4 were achieved.   In  Test  3,  the Pass 3 treated
 sediment  PCB  concentration  was  72 percent  less  than the  feed.   Thus,  the  50
 percent reduction  goal was  met.   In Test 4,  the  pass  6 treated sediment PCB
 concentration  was  200 ppm.   Thus, the 288  ppm concentration  goal was  met.  The
 final  treated  sediment concentration, at pass 10 for  Test 2,  was 40  ppm,  which
 meets  neither  the 5 ppm concentration nor  the 90 percent reduction  goals.
 However,  a 20 ppm level  was observed at pass 3 and 8 ppm was  observed at pass
 9.  Results for Test 5 also fell slightly short of goals for  decontamination
 as discussed in Section 7.4.

      These goals were achieved even though (1) operational problems such as
 foaming and solids retention occurred and (2) the analysis method provided
 results with a precision of plus or minus 20 percent (measured as relative
 percent difference).
                                        70

-------
 7.4  HEALTH AND SAFETY MONITORING
      During the demonstration of CF Systems'  process unit,  personnel  were
 potentially exposed to the contaminated harbor sediments.   A monitoring pro-
 gram was conducted to determine potential  exposures and provide a basis for
 selection of proper personal  protective equipment.   Several  types of  portable
 monitoring equipment were  used during the  various  phases of the field investi-
 gations,  including:

      •   Portable Organic Vapor Analyzer (Century OVA)
      •   Portable Photoionization Meter  (HNu)
      •   Combustable gas/oxygen/hydrogen sulfide meters  (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 containing 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  encountered 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 equipment.  The OVA also  was used as a survey meter  on the
 process equipment to search for possible fugitive  emissions from the
 equipment.  All  measurements indicated that organic vapor levels remained in
 the range of background 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
                                      71

-------
two integral combustible gas detectors located on either end of the unit.
During the normal extraction process, combustible gas readings remained at
background levels.  However, 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 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 decontaminated  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 demonstration  as  discussed  in Section  5.3.

 7.5  EQUIPMENT AND MATERIAL HANDLING PROBLEMS
      Equipment and material handling problems occurred throughout the  dem-
 onstration, as described in Section 5.   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 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.
                                       72

-------
      Previous use of the unit affected interpretation of semivolatiles  data
 and may have  contributed to imbalance  of the, PCB inventory.   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.   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  performance 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  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 solvent entrained  in the  treated sediment  and  elimination of the
 foaming problem.

 7.6  DATA QUALITY ASSURANCE
     All  field sampling and  laboratory analyses were accompanied by a program
 of quality assurance/quality control  (QA/QC) checks designed to assess the
 validity of the  sampling and analysis effort.  This program was designed  to
 ensure that the  samples are representative,,  and that the analytical  data
 accurately describes the characteristics and concentrations  of constituents in
 the samples. This QA/QC program is outlined  in the approved  Quality Assurance
Project Plan (QAPP).
                                      73

-------
Technical Systems Reviews
     EPA conducted a technical systems review of the field activities during
the dress-rehearsal for the demonstration.  This field audit verified that the
sampling procedures used would result in samples that met the requirements of
the QAPP, and resulted in a satisfactory rating, the highest rating
achievable.

     EPA also conducted a technical systems review of each of the laboratories
Involved in the analysis of samples from the CF Systems Demonstration project,
examining the laboratory's procedures for sample storage, preparation,
analysis, and documentation.  While some concerns were identified during each
of these reviews,  these concerns were addressed to the satisfaction of the
reviewer before analyses were continued.

Data Review and Validation
     In  order to  assess the validity of the measurement data, the QC  data
generated with the environmental data were evaluated with respect to  the
project  quality assurance objectives defined  in the QAPP, as well as  the
specific quality  control  requirements of  the  analysis methods used.   These
laboratory QC data are presented in Volume II.  The evaluation  of the QC  data
was  accomplished  through  review of:

      •  Completeness  of  analytical  reporting
      •  Analytical holding  times
      •  Target  analyte identification  criteria
      •  Tentatively identified  compound  (TIC) identification  criteria
      •  Calibration frequency and  acceptance  criteria
      •  Presence and  contamination of method  blanks
      i  Surrogate compound recovery
      •  Presence and  conformance  with acceptance  criteria of matrix spike,
         matrix spike  duplicate,  and duplicate analyses.
                                       74

-------
 Evaluation and Qualification of Measurement Data
      The following sections present brief summaries of the findings of the
 data review and validation, by analyte class.

 Polychlorinated Biphenyls
 Method 8080 Analyses
      Fifty-three sediment samples were analyzed for PCBs from Tests 1 through
 5,  either as feed stock,  treated sediments,  or decontamination residues.
 Ninety-five percent of the recoveries calculated from the matrix spike/matrix
 spike duplicate (MS/MSD)  sample pairs analyzed with these samples were within
 the project quality assurance objectives  defined in the  QAPP  indicating
 acceptable accuracy for the analysis  of these  samples.

      Forty-eight percent  of the surrogate recoveries  calculated  from  these
 samples  exceeded the acceptance criteria.  Because  the surrogate compound used
 for these samples (dibutylchlorendate)  is not  a polychlorinated  biphenyl
 (PCB), the the  matrix spike data are  more Indicative  of  the accuracy  of
 analysis for  PCBs in these  samples.

      Only twenty-one percent  of the relative percent, differences (RPDs)
 calculated  between  the  MS/MSD sample  pairs were within the defined  aceptance
 criteria,  indicating that the established precision criteria  (plus  or  minus 20
 percent)  were too ambitious.  The average RPD  calculated  between MS/MSD pairs
 was  47 percent.   EPA analysis method  8080 QC acceptanqe criteria show  RPDs
 greater  than 60  percent for PCB  1242  and  PCB 1254 (EPA, 1986).

     None of the  method blanks  associated with  these samples were contaminated
 with PCBs.

     The  results  of  analyses' of PCBs  in extract samples were not used  as part
 of the demonstration evaluation and have nojt been included as part of  this QA
 evaluation.                                      ,

     Collectively, these data show that the;analysis of PCBs in sediments for
Tests 1 through 5 are sufficiently accurate and precise to allow for engineer-
 ing assessment of the efficacy of this Demonstration.
                                      75

-------
Method 680 Analyses
     The recoveries of congeners spiked into samples analyzed by method 680
were outside of the interim acceptance criteria for monochlorobiphenyl through
tetrachlorobiphenyl isomer groups in one MS/MSD sample pair (Test 1). The RPDs
calculated for this sample pair were outside of the interim acceptance
criteria for mono- and dichlorobiphenyl.  The surrogate recoveries for this
Test 1 analysis were also outside of the acceptance criteria for the
monochlorobiphenyl isomer group.  In Test 4 analyses, the accuracy and
precision problems were not the case.  Therefore, the data presented are
insufficient for evaluation of the accuracy and precision of analysis for the
less chlorinated congeners.  Based on  this information, the analytical data
for the more highly chlorinated isomers should, however, be of acceptable
accuracy and precision.

Spittler Analyses
     The data  from the Spittler screening analyses  were used for decisions in
the field,  but were not the basis  for  any conclusions drawn for the  final
report.  These data were,  therefore, not  included  in the data  review and
validation  process.

Base/Neutral  and  Acid  Extractable  Compounds  (BNAs)
      Both  sediment and oil  samples were  analyzed  for BNAs.  The recovery  of one
or more compounds in  the  each  of  the MS/MSD  samples analyzed with  the sediment
samples were outside  of the established  acceptance criteria. In  addition,  each
of the sediment  samples was spiked with  surrogate compounds.   The  recovery of
 these surrogate  compounds were within  acceptance  criteria  for  20  of the  26
 samples.  While these  data are  in  apparent conflict, they  cast  doubt over the
 acceptability of the  analytical  accuracy.   The relative percent difference
 (RPD) calculated between analyses of at least one compound in  each of the
 MS/MSD pairs exceeded the acceptance criteria specified in the analytical
 method, indicating a generally unacceptable degree of analytical  precision.

      No MS/MSD samples were analyzed with the oil samples analyzed for BNAs,
 therefore, the recovery of matrix spike compounds cannot be used to evaluate
 the analytical accuracy.   While the samples were spiked with surrogate
                                       76

-------
 compounds,  the  recoveries  cannot be  calculated in most of the samples because
 of dilution effects.  There is,  therefore,  an  inadequate basis for  evaluation
 of analytical accuracy  and precision.

      The  only target  analytes detected  in  the method  blanks  associated with
 these analyses  are  common  plasticizers  (phthalic  acid esters),  which  may have
 been  introduced into  the samples during field or  laboratory  activities.

      The  BNA data should be  considered  as  qualitative,  rather  than
 quantitative, based on  conflicting or inadequate  information  available  to
 evaluate  analytical accuracy, as well as poor precision  of analysis.

 Volatile  Organic Analytes  (VOAs)
      The  recoveries calculated  from  the M^S/MSD sample  analyzed  with the
 sediment  samples are  within  acceptance  criteria,  as are  the  recoveries
 calculated  from the surrogate compound  spikes.  These  data indicate an
 acceptable  degree of  accuracy for this  noncritical measurement.

      The  RPDs calculated between  the spike compounds  in  the MS/MSD pair are
 within the  acceptance criteria  specified in the analysis method, indicating
 acceptable  precision  for the analysis of VOAs in  sediment samples.

      The  method blank associated with these samples show that the samples may
 have  been contaminated with  low  concentrations of methylene chloride, a common
 laboratory  contaminant.   The data from analysis of sediment samples for VOAs
 are not qualified in any way, and should be regarded as quantitative.

     Metals analyses were performed on sediment grab samples, composite
 samples and oil  samples.

     The  recoveries calculated from one of the two MS/MSD sample pairs
 analyzed with the sediment grab samples are outside of acceptance criteria for
many of the target analytes; however, those calculated from the other MS/MSD
pair are acceptable.  These conflicting data do not provide sufficient
 information for  the evaluation of analytical  accuracy.  The RPDs calculated
                                      77

-------
between one MS/MSD sample pair, and duplicate sample pairs, are within
acceptance criteria, indicating an acceptable level of analytical precision
for these analyses.  The method blanks associated with these samples
demonstrate that the samples were not contaminated in the laboratory with the
target analytes.  Because of the lack of information for evaluation of
accuracy, these data should be regarded as qualitative.

     The recoveries calculated from the MS/MSD samples analyzed with the
composite sediment samples, and the RPDs calculated between the MS/MSD sample
pairs and duplicate sample pairs, indicate an acceptable level of analytical
accuracy and precision for these analyses.  The method blanks associated with
these samples demonstrate that the samples were not contaminated in the
laboratory with the target analytes.

     All oil samples were analyzed for metals by  the method of standard
additions  (MSA).   Significant  matrix  interferences were  reported for the
majority of these  samples.  The RPDs  reported between  the  analysis of a
duplicate  sample pair were within acceptance criteria  for  two of the five
analytes of interest.  The,method blank associated with  these samples
demonstrates that  the samples  were not contaminated in the laboratory with  the
target  analytes.   Together, these data indicate  that  the results from metals
analyses in oil samples  have an unacceptable level of  accuracy and precision,
and should be  regarded as  qualitative.

      In addition  to the  total  metals  analyses, sediment  samples  were analyzed
for the characteristic of  Extraction  Procedure Toxicity  (EP Tox). The
 recoveries calculated  from the MS  sample  prepared with these  samples are
within  acceptance  criteria for six  of the eight  target analytes.  The RPDs
 calculated between two duplicate  sample  pairs  are within acceptance criteria
 for some target analytes,  outside  of  acceptance  criteria for  others, and
 cannot  be  evaluated for  the  remaining analytes due to  nondetectable
 concentrations in the  samples.  The method blanks associated  with these
 samples demonstrate that the  samples  were not  contaminated in the  laboratory
 with the target analytes.   Together,  these data  indicate an acceptable  degree
 of accuracy  and an unacceptable degree of precision for these analyses.   The
 data should,  therefore,  be regarded as qualitative.
                                       78

-------
 Total Residue
      The recoveries calculated from control samples analyzed for total
 residue, along with the sediment samples, are within acceptance criteria, as
 are the RPDs calculated between 99 percent of the duplicate samples analyzed.
 These data indicate acceptable accuracy and precision for these analyses.

      The data from analysis of sediment samples for Total Residue are not
 qualified in any way,  and should be regarded as quantitative.

 Total Oil and Grease
      Twenty-five percent of the recoveries calculated from the analyses of
 control  samples for total  oil  and grease,  and fifteen percent  of the RPDs
 calculated  between replicate analyses,  were outside of the acceptance
 criteria.   About one third of the method blanks associated with these samples
 were found  to have detectable  concentrations of oil  and grease, demonstrating
 the potential for low.-level  contamination  of-the samples in the laboratory.

 Cyanide

      The recovery calculated from a control  sample  analyzed along with sedi-
 ment samples  for  cyanide is  within acceptance  criteria.   The recovery  of
 cyanide  from  a  post-digestion  spike sample  was,  however,  outside of  acceptance
 criteria, providing conflicting information  for  the  evaluation  of analytical
 accuracy.   The  RPD  calculated  from the  analysis  of a replicate  pair  is within
 acceptance  criteria, indicating acceptable  analytical precision.

      The data from  analysis  of sediment  samples  for  cyanide  should be  regarded
 as  qualitative, based on conflicting evidence  for the evaluation of  accuracy.

 Total Dissolved Solids
     The recovery from a spiked control  sample analyzed with water sample for
 total dissolved solids is within acceptance criteria, indicating acceptable
 accuracy. Because no replicate pairs of  samples were analyzed for total
dissolved solids, the analytical precision of this noncritical  measurement
cannot be evaluated.
                                      79

-------
     The IDS data should be considered as qualitative, rather than
quantitative, based on inadequate information to evaluate the analytical
precision.

Total Suspended Solids
     No quality control check sample or replicate sample was analyzed with the
water samples analyzed for total suspended solids, therefore the analytical
accuracy and precision of this measurement cannot be evaluated.

     The TSS data should be considered as qualitative, rather than
quantitative, based on inadequate information to evaluate the analytical
accuracy and precision.

PH
     No quality control check samples were prepared for the measurement of pH
1n sediment samples, therefore the analytical accuracy of this measurement
cannot be evaluated. The RPD calculated between pH measurements made on a
replicate sample pair are within acceptance criteria; however, the frequency
of replicate pair analysis was insufficient.

     The pH data should be considered as qualitative, rather than
quantitative, based on inadequate information to evaluate the analytical
accuracy and precision.
                                      80

-------
                                   SECTION 8

                                  REFERENCES
ASTM, 1984.  Standard Method  for Sieve Analysis of Fine and Coarse Aggregates
ASTM 0-4059.  American Society for Testing and Materials.  Philadelphia  PA
December 1984.

COE, 1987.  Pilot Study of Dredging and Dredged Material Disposal Alterna-
tives.  U.S. Army Corps of Engineers.  New England Division.  Waltham  MA
September 1987.

Ebasco, 1987.  Draft New Bedford Harbor Feasibility Study Test 22, Environ-
mental Evaluation Activity 22.4 - Selection of Additional Contaminants for
Inclusion in the Risk Assessment and Feasibility Study.  Boston, MA.  July
1987.                                    ;,

Fowler, Bruce A. and Bennett, Joseph T., 1987.  Screening for Characterization
of PCB-containing Soils and Sediments.  Proceedings of the National Conference
on Hazardous Wastes and Hazardous Materials.   Washington, DC.  March 1987.

Spittler, T.M., 1984.  "Field Measurement of Polychlorinated Biphenyls in Soil
and Sediment Using a Portable Gas Chromatograph," Environmental Sampling for
Hazardous-Wastes.  American Chemical.Society.   1984.   37-42.

EPA, 1983.   Methods for Chemical  Analysis of Water and Wastes.  U.S.  Environ-
mental Protection Agency.   Environmental Monitoring and Support Laboratory,
Cincinnati,  OH.  600/4-79-020.  Revised March 1983.

EPA, 1986.   Test Methods for Evaluating Solid Waste.   U.S.  Environmental
Protection Agency.   SW-846.   U.S.  Government  Printing Office.   Washington,
D.C.  Third Edition.   November 1986.
                                     81

-------
              APPENDIX A



MASS INVENTORIES FOR TESTS 2, 3, AND 4
                   82

-------
                                                                           TOTAL MASS  BALANCE  (Lbs)
                                                                                   TEST #2
            ITEM
PASS #
   1
                                                                                                                                          10
                                                                                                                                                    TOTALS
1 Feed Material
2 (Saving,
3 (Strainer)
4 (Spills)
5 (Residuals)
6 Water
7 Total Feed
| 483.50 |
| 0.00 |
I 4-63 |
| 0.00 |
| 88.00 |
| 120.12 |
| 510.99 |
367.15 |
3.09 |
0.00 |
0.00 |
2.88 |
41.75 |
402.93 |
328.46 | 351
0.00 | 1
0.00 | 0
0.00 | 0
1.33 | 2
20.57 | 41
347.70 | 389
.54 |
.13 |
.00 |
.00 |
.86 |
.88 |
.43 |
296.50 |
0.00 |
0.00 |
0.00 |
1.56 |
49.06 |
344.00 |
305.00 |
0.00 |
0.00 |
0.00 |
0.31 |
31.32 |
336.01 |
259.00 |
5.13 |
0.00 |
0.00 |
0.18 |
29.19 |
282.88 |
321.25 |
0.00 |
0.00 |
0.00 |
0.43 |
34.36 |
355.18 |
347.00 |
0.00 |
0.00 |
0.00 |
0.76 |
22.25 |
368.49 |
378.00
0.00
0.00
0.00
0.76
22.31 |
399.55 |
| 3437.40
| 9.35
| 4.63
| 0.00
| 99.07
| 412.81
3737.16
00
8 Raffinate |
9 Sampling |
10 Spills |
11 Residuals |
12 Basket Strainer |
13 (H20) |
14 Cartridge Filter |
15 Extract |
16 Spills |
Output |
367.15 |
0.00 |
0.00 |
0.00 |
22.00 |
0.00 |
0.00 |
0.06 |
0.00 |
389.21 |
328.46 |
2.74 |
0.00 |
0.00 |
17.01 |
0.00 |
0.00 |
0.05 |
0.00 |
348.26 |
351.54 |
9.53 |
0.00 |
2.75 |
8.75 |
0.00 |
0.00 |
0.19 |
0.00 |
372.76 |
296.50 |
5.19 |
0.00 |
1.50 |
8.50 |
0.00 |
0.00 |
0.00 |
0.00 |
311.69 |
305.00 |
6.09 |
0.00 |
2.25 |
8.38 |
0.00 |
0.00 |
0.00 |
0.00 |
321.72 |
259.00 |
0.00 |
0.00 |
0.75 |
8.19 |
0.00 |
0.00 |
0.00 |
0.00 |
267.94 |
321.25 |
4.44 |
0.00 |
0.00 |
8.25 |
0.00 |
0.00 |
0.00 |
0.00 |
333.94 |
347.00 |
5.77 |
0.00 |
1.25 |
8.25 |
0.00 |
0.00 |
0.00 |
0.00 |
362.27 |
378.00 |
5.77 |
0. 00 |
1.00 |
0.00 |
,;0.00 |
0.00 |
0.00 |
0.00 |
384.77 |
346.75 |
31.81 |
0.00 |
0.00 |
8.94 |
0.00 |
2.81 |
0.31 |
0.00 |
390.62 |
3300.65
71.34
0.00
9.50
98.27
0.00
2.81
0.61
0.00
3483.18
   Accumulation &
    Other Losses
                           121.78
                                        54.67
                                                   -25.06
                                                                77.74
                                                                            22.28
                                                                                        68.07

                                                                                I            I
                                                                       -51.06   |    -7.09   |   -16.28
                                                                                                                                         8.93
                                                                                                                                                   253.98


-------
                     TOTAL SOLIDS BALAJtCC (U*>
                              TEST «Z
ITEM
1 Feed Material |
2 (Sampling) |
3 (Strainer) |
4 (Spills) |
5 (Residuals)
6 Water
7 Total Feed
8 Raffinate
9 Sampling
10 Spills
11 Residuals
12 Basket Strainer
13 (H20)
14 Cartridge Filter
15 Extract
16 Spills
Output
PASS #
1
111.21 |
0.00 |
1.06 |
0.00 |
12.32 |
0.00 |
97.82 |
55.07 |
0.00 |
| 0.00 |
0.00 |
I 4-18 I
| 0.00 |
| 0.00 |
I o.oo j
| 0.00 |
| 59.25 |
2
55.07
0.46
0.00
0.00
0.43
0.00
54.18
45.98
0.38
0.00
0.00
1.36
0.00
0.00
0.00
0.00
47.73
3
| 45.98 |
| 0.00 !
| 0.00 |
| 0.00 |
1 0.19 |
| 0.00 |
| 45.80 I
| 58.00 |
| 1.57 |
| 0.00 |
I 0.45 |
I 1-66 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
| 61 .69 |
4
58.00 |
0.19 |
0.00 |
0.00 I
0.47 |
0.00 |
57.35 |
23.72 |
0.42 |
0.00 |
0.12 |
1.11 |
0.00 |
0.00 |
0.00 |
0.00 |
25.36 |
5
23.72
0.00
0.00
0.00
0.12
0.00
23.60
27.45
0.55
0.00
0.20
1.42
0.00
0.00
0.00
0.00
29.63
6
I 27.45 I
| 0.00 |
| 0.00 |
| 0.00 |
| 0.03 |
| 0.00 |
| 27.42 |
I 23.31 |
| 0.00 I
| 0.00 |
| 0.07 |
| 1.06 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
| 24.44 |
7
23.31 |
0.46 |
0.00 |
0.00 |
0.02 |
0.00 |
22.83 |
==============
22.17 |
0.31 I
0.00 |
0.00 |
1.16 |
0.00 |
0.00 |
0.00 |
0.00 |
23.63 |
8
22.17 |
0.00 |
0.00 |
0.00 |
0.03 I
0.00 |
22.14 |
==========
31.58 |
0.53 |
0.00 |
0.11 |
0.91 |
0.00 |
0.00 |
0.00 |
0.00 |
33.12 |
9
31.58 |
0.00 |
0.00 |
0.00 |
0.07 |
0.00 |
31.51 |
=== ===========
34.02 |
0.52 |
0.00 |
0.09 |
0.00 |
0.00 |
0.00 |
0.00 I
0.00 |
34.63 |
10
34.02
0.00
0.00
0.00
0.07
0.00
33.95
=====:
31.21
2.86
0.00
0.00
1.25
0.00
1.60
0.00
0.00
36.92
TOTALS
| 432.51
| 1.11
| 1.06
| 0.00
| 13.75
| 0.00
| 416.59
| 352.51
I 7.13
| 0.00
| 1.05
| 14.11
| 0.00
| 1.60
| 0.00
| 0.00
| 376.41
Accumulation &
 Other Losses
I     38.57  I
6.45
 -15.89  |     31.99
====================
                          -6.03
                                       2.98
                                                  -0.80
                                                            -10.99
                                                                                I            I
                                                                         -3.12  I     -2.97  |    40.19

-------
TOTAL PCB BALANCE (Grams)
         TEST #2
I TEH
1 Feed Material
2 (Sampling)
3 (Strainer)
4 (Spills)
5 (Residuals)
6 Water
7 Total Feed
§} a Raffinate
9 Sailing
10 Spills
11 Residuals
12 tasket Strainer
13 (H20)
14 Cartridge Filter
15 Extract
16 Spills
Output
Accumulation &
Other Losses
PASS #
1
| 17.67 |
| 0.00 |
1 0.17 |
| 0.00 |
1 0.95 |
| 0.00 |
1 16.55 |
1 1.» I
I 0.00 |
| 0.00 |
| 0.00 |
| 0.41 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
1 2.34 |
I SS S S ESS 5 SSS SS JSS S S'.
\ 1
2
1.93 |
0.02 |
0.00 |
0.00 |
0.02 |
0.00 |
1.89 |
mnnsnun*
1.09 |
0.01 |
0.00 |
0.00 |
0.10 |
0.00 |
0.00 |
0.00 |
0.00 |
1.20 |
1
0.70 1
3
1.09
0.00
0.00
0.00
0.00
0.00
1.08
0.53
0.01
0.00
0.00
0.02
0.00
0.00
0.00
0.00
0.57
0.51
4
1 0.53 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
1 0.52 |
1 0.71 |
| 0.01 |
| 0.00 |
| 0.00 |
| 0.01 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
1 0.74 |
==================
1 1
1 -0.22 1
5
0.71 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.71 |
0.74 |
0.01 |
0.00 |
0.01 |
0.02 |
0.00 |
0.00 |
0.00 |
0.00 |
0.77 |
=========1
1
0.07 1
6
0.74 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.73 |
0.43 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.44 |
1
n ™ i
7
0.43
0.01
0.00
0.00
0.00
0.00
0.42
0.36
0.01
0.00
0.00
0.01
0.00
0.00
0.00
0.00
0.38
n lu
8
| 0.36 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
j 0.00 |
| 0.36 |
1 0.42 |
1 o.oi |
| 0.00 |
| 0.00 |
1 0.01 |
1 0.00 |
| 0.00 |
1 0.00 |
| 0.00 |
| 0.44 |
1 1
1 .n n? i
9
0.42 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 }
0.41 |
0.12 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.12 |
:============
1
n -3rt i
10
0.12
0.00
0.00
0.00
0.00
0.00
0.12
0.57
0.05
0.00
0.00
0.01
0.00
0.00
0.03
0.00
0.66
TOTALS
| 23.99
| 0.03
| 0.17
| 0.00
| 0.98
| 0.00
| 22.81
| *.•»
| 0.12
| 0.00
| 0.02
1 0.61
| 0.00
| 0.00
| 0.03
| 0.00
| 7.66
1
                                                  0.2
                                                             -0.54
                                                                         15.15

-------
                                                                     TOW KASS BALANCE (Lbs)
                                                                              TESTtt
I TEH
1 Feed Material
Z (Sampling)
3 (Strainer)
4 (Spills)
5 (Residuals)
6 Water
7 Total Feed
g> 8 Raffinate
9 sampling
10 Spills
11 Residuals
12 Basket Strainer
13 (H20)
14 Cartridge Filter
15 Extract
16 Spills
Output
— =— —sis ss SS===S===S=
Accumulation &

PASS*
1
| 460.75
| 40.74
| 0.00
| 0.00
| 0.87
| 88.86
| 508.00
| 465.00
| 5.95
| 0.00
| 2.75
| 9.75
| 0.00
| 0.00
| 0.19
| 0.00
| 483.64
I
I 24 36

Z
| 563.00 |
I 4.64 |
| 0.00 |
| 0.00 |
| 3.26 |
| 0.00 |
| 555.10 |
| 482.30 |
I 5.81 |
| 0.00 |
| 0.80 |
| 8.44 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.00 |
| 497.35 |
1 1
1 57.75 1

3 4 5 6 7 8 9 10
482.30 | | I I I I I
0.00 | | I I I I I
0.00 | | I I I I I
0.00 | | | I I I I
130.55 | | |l I I I
0.00 | | I I I I I
351.75 | | 1 1 1 1 1
310.75 | | I I I' '
0.00 | I I I I I I
0.00 | | I I I I '
0.00 | | I I I I '
9.37 | I I I I I '
0.00 | | I I I I '
1.37 | I I I l-l I
0.94 | | I I I I '
0.00 | I II I ! '
322.43 | I I I I I '
:==========================================^=====^=~=========~========================
i i i i j | '
29.32 | | | I I I '

TOTALS
| 1506.05
| 45.38
| 0.00
| 0.00
| 134.68
| 88.86
| 1414.85
| 1258.05
| 11.76
| 0.00
| 3.55
| 27.56
| 0.00
| 1.37
| 1.13
| 0.00
} 1303.42
:=============
1
| 111.43
„.___—————
=============================:

-------
TOTAL SOLIDS BALANCE (Lbs)
         TEST #3
I TEH
1 Feed Material |
2 (Sampling) |
3 (Strainer) |
4 (Spills) |
5 (Residuals) |
6 Water |
7 Total Feed |
S3 8 Raffinate |
T9 Sampling |
10 Spills |
,11 Residuals |
• 12 Basket Strainer |
13 (H20) |
14 Cartridge Filter |
15 Extract |
16 Spills |
Output |
Accumulation & |
Other Losses |
PASS #
1
72.94
6.45
0.00
0.00
0.14
0.00
66.35
69.75
0.89
0.00
0.41
3.90
0.00
0.00
0.00
0.00
74.96
-8.61
23 4 56 7 8
| 67.56 | 57.88 | III | |
| 0.56 | 0.00 | II | | |
| 0.00 | 0.00 | | | | | |
| 0.00 | 0.00 | | | | | |
| 0.39 [ 15.67 || | | | |
| 0.00 | 0.00 | || || |
| 66.61 | 42.21 | II | | |
| 57.88 | 34.18 | | | II |
| 0.70 | 0.00 | | | '| 'j I
| 0.00 | 0.00 | II | | |
1 0.10 | 0.00 | | | III
I 1-52 I 1.69 | I | | I I
I 0.00 | 0.00 | | | II |
I 0.00 | 0.74 | | || II
I 0.00 | 0.00 || | | II
| 0.00 | 0.00 | | | | | I
| 60.19 | 36.61 | | | | | |
1 1 I I 1 II 1
1 6-42 | 5.60 | | 1 | | |
9 10 TOTALS
| | 198.37
1 1 7.01
| | 0.00
I | o.oo
I | 16.19
| | 0.00
| | 175.17
| | 161.81
| 1 1.59
| | 0.00
1 1 0.51
1 1 7.11
| | 0.00
1 1 0.74
I I o.oo
I | o.oo
| | 171.75
I I
1 1 3.42

-------
=====================
 K'JL     I           I

I   00'I-     I  Zri     I   Z6'9     I        sasson
I            I           I           I       »
                    =======================
Sl'9 1111
OO'O I | 1 1
U-L | I 1 1
OO'O 1 1 1 1
OO'O ill)
l£'0 1 1 1 1
lO'O 1 1 1 1
OO'O | I 1 I
WTO 1 1 1 1
1.9", I 1 1 1
9>'£l 1 1 1 1
OO'O 1 I I I
SS'O I 1 1 1
OO'O 1 1 1 I
OO'O 1 1 1 1
88-0 i 1 1 I
06-H 1 1 1 1
S1V101 01 6 8
1 1 1 1 8£'Z
| I i i OO'O
i i i i son
| | 1 1 OO'O
| I 1 1 OO'O
| | 1 1 90'0
| I 1 1 OO'O
| 1 1 1 OO'O
| I 1 1 OO'O
1 1 1 I 6Z't
I 1 1 1 8£'L
III! oo'o
1 1 1 1 is'o
1 1 1 1 OO'O
| 1 1 1 OO'O
1 I 1 1 oo-o
{ I 1 I 68'1
/ 9 S * £
1 U'l
\ OO'O
i oo'o
1 OO'O
1 OO'O
1 SO'O
1 OO'O
1 OO'O
1 ZO'O
i 68' I
1 6£'£
i OO'O
1 ZO'O
1 OO'O
i OO'O
1 £0'0
1 WE
Z
| 08' I 1 V*»0
\ OO'O 1 slHdS 91
i SO'O ! J3SJJX3 SI
I OO'O 1 JailH a6piJJJB3 n
| OO'O 1 (OZH) £1
, oz-o 1 WQSWZI
| lO'O 1 sienpisaa u
| OO'O 1 sll!dS 01
| EQ'O 1 Bujvtaes 6
00
| (,^'1 | aiEUl^jea g OO
-—- .^-.— — —- *• *-B=====sss:si=sss==^T:==gas
1 ZZ'8 1 P33^ lel°l L
\ oo'o i JS*Bn 9
1 lO'O 1 (sienpjsaa) S
| OO'O 1 (si lids) f
| OO'O 1 CJSUIBJJS) £
| 59-0 I (6u$ldws> 2
1 SS'6 I 1BJJ91BHPMJ I
I H31I
#SSVd
                                                       (SMJ9> 33WftV8 83d TVIOI

-------
          I TEH
PASS #
   1
                                                                        TOTAL MASS BALANCE (Lbs)
                                                                                 TEST #4
                                                                                                                8           9          10        TOTALS
1
2
3
4
5
6
7
Feed Material
(Sampling)
(Strainer)
(Spills)
(Residuals)
Water
Total Feed
| 333.25 |
| 33.86 |
| 0.00 |
| 0.00 |
| 0.56 |
| 0.00 |
| 298.83 |
279,
0.
0.
0.
0.
39.
318.
.75 |
.00 |
,00 |
00 |
68 |
32 |
39 |
348.75 |
0.00 |
0.00 |
0.00 |
0.94 |
39.14 |
386.95 |
298.00 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
298.00 |
360,
0.
0.
0.
1.
16.
375.
.20 |
,00 |
,00 |
00 |
20 |
70 |
70 |
256
0
0
0
0
22,
278,
.00 | |
.00 | |
.00 | |
.00 | |
.26 | |
.42 | |
•16 | |
I | | 1875.95
I I | 33.86
I | | 0.00
I | | 0.00
I I | 3.64
I | | 117.58
I | | 1956.03
                                                                     s=sss==sss=
a Raffinate |
9 Sampling |
10 Spills |
11 Residuals |
12 Basket Strainer |
13 (H20) |
14 Cartridge Filter |
15 Extract |
16 Spills |
Output |
279.75 |
5.48 |
0.00 |
0.00 |
8.25 |
0.00 |
0.00 |
0.00 |
0.00 |
293.48 |
348.75 |
6.57 |
0.00 |
37.75 |
8.32 |
0.00 |
0.00 |
0.00 |
0.00 |
401.39 |
297.50 |
6.76 |
0.00 |
1.00 |
8.06 |
0.00 |
0.00 |
0.00 |
0.00 |
313.32 |
360.20 |
4.76 J
0.00 |
3.75 |
8.25 |
0.00 |
0.00 |
1.00 |
0.00 |
377.96 |
256.00 |
4.23 |
0.00 |
1.75 |
8.00 |
0.00 |
0.00 |
0.00 |
0.00 |
269.98 1
319.25 | II
o.oo | 	 | |
0.00 | | |
0.50 | ||
7.81 | | |
0.00 | | |
2.00 | | |
1.38 | ||
0.00 | | |
330.94 I 1 I
I | 1861.45
| ... |._.. 27.80
|| 0.00
I | 44.75
I | 48.69
| | 0.00
I | 2.00
1 1 2-38
| | 0.00
1 1 1987.07
Accumulation &
 Other Losses
                           5.35
                                     -83.00
                                                 73.63
                                               =========
                                   -79.96
                                               105.72
                                                           -52.78
                                                                                                                       -31.04


-------
TOTAL SOLIDS BALANCE (Lbs)
         T6ST«
I
ITEM
1 Feed Material |
2 (Sampling) |
3 (Strainer) |
4 (Spills) |
5 (Residuals) |
6 Water ' 1
7 Total Feed |
8 Raff inate |
9 Saapling |
10 Spills |
11 Residuals |
12 Basket Strainer |
13 (H20) |
14 Cartridge Filter |
15 Extract |
16 Spills |
Output |
Accumulation & |
Other Losses |
>ASS*
1
53.99 |
5.49 |
0.00 |
0.00 |
0.09 |
0.00 |
48.41 |
36.37 |
0.71 |
0.00 |
0.00 |
1.40 |
0.00 |
0.00 |
0.00 |
0.00 |
38.48 |
I
9.93 |
2
36.37 |
0.00 |
0.00 |
0.00 |
0.09 |
0.00 |
36.28 |
========:
41.85 |
0.79 |
0.00 |
4.53 |
0.83 |
0.00 |
0.00 |
0.00 |
0.00 |
48.00 |
I
-11.72 |
3
41.85 |
0.00 |
0.00 |
0.00 |
0.11 |
0.00 |
41.74 |
31.24 |
0.71 j
0.00 |
0.11 |
0.81 |
0.00 |
0.00 |
0.00 |
0.00 |
32.86 i
1
8.88 |
4
31.29 |
0.00 |
0.00 |
0.00 |
0.00 |
0.00 |
31.29 |
26.11 |
0.35 |
0.00 |
0.27 |
1.03 |
0.00 |
0.00 |
0.00 |
0.00 |
27.76 |
I
3.53 |
5
26.11
0.00
0.00
0.00
0.09
0.00
26.02
18.43
0.30
0.00
0.13
0.88
0.00
0.00
0.00
0.00
19.74
6.28
6
| 18.43 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.02 |
| 0.00 |
| 18.41 |
I 19.95 |
| 0.00 |
| 0.00 |
| 0.03 |
I 0.55 |
| 0.00 |
I 0.38 |
| 0.00 |
| 0.00 |
| 20.91 |
:========:
1 1
1 -2.50 |
7 8 9 10 TOTALS
, | | | 208.04
| I I I 5M
| | I | 0.00
II I I o.oo
| | I I 0.40
| | I | 0.00
| | I I 202.15
, | | | 173.95
| I 1 | 2.86
| | I | 0.00
| I 1 | 5.06
,||| 5.50
,||| 0.00
,||| 0.38
| | I 1 0.00
, || | 0.00
1 1 i i 1S7-76
1 II j
	 =====-================== 	

-------
                                                                            TOTAL PCB BALANCE (Grams)
                                                                                     TEST #4

1
2
3
4
5
6
7
==;
ITEM
Feed Material
(Sampling)
(Strainer)
(Spills)
(Residuals)
Water
Total Feed
===============
PASS
1
I 62
I 6
I o,
I o.
I o.
I o.
I 56.
#
.!>6
.36
.00
.00
,11
00
10
2
| 16.51 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.04 |
| 0.00 |
| 16.47 |
3
18
0.
0.
0.
0.
0.
18.

.81 |
.00 |
.00 |
,00 |
05 |
00 |
76 |
4
9.54
0.00
0.00
0.00
0.00
0.00
9.54
s======:::
5
| 3.85 |
| 0.00 |
| 0.00 |
| 0.00 |
| 0.0133 |
| 0.00 |
1 3.84 |
=================:
6
2
0,
0.
0.
0.
0.
2.

.01 |
.00 |
.00 |
00 |
00 |
00 |
01 |
7 8 9 10 TOTALS
I I I | 113.28
I I I | 6.36
I I || 0.00
I I I | 0.00
I I I | 0.2l"
I I I | 0.00
jl|| 106.71

-------
                APPENDIX B
MASS BALANCES FOR DECONTAMINATION EFFLUENTS
                     92

-------
                                                     Mass Balance for Test 3 Fuel  Wash
                                                                                                       Analyte Data
CO
Test 3 Pass 1 09-17-88
EPT-W1----- — -- — Coleman Fuel (gal), est
* Specific Gravity
* Density of Water (Ibs/gal)
Wash Amount (Ibs)
EPT-U20 	 Coleman Fuel (gal), est
* Specific Gravity
* Density of Water (Ibs/gal)
Wash Amount (Ibs)
Still Bottoms 	 Residuals (gal), est
* Specific Gravity
* Density of Water (Ibs/gal)
Wash Amount (Ibs)
EPT-U20
Still Bottoms
Total Output (Ibs)
Date
I
Raw
Data
1.00
0.72
8.35
6.00
==========
18.00
0.72
8.35
108.07
0.47
0.96
8.35
3.75
6.00
108.07
3.75
117.82
(Correctec
| Data

| Std Dev
(Ibs)
0.0000
0.0000
0.0000
0.0293
========
0.0077
0.0000
0.0000
0.5277
=========
0.0015
0.0000
o.dooo
0.3125
========
0.0293
0.5277
0.3125
0.9325

Cone.
(ppm)
560
570
:
3400
Std Dev
(ppm)
95.2
96.9
578
Mass
(g)
2.12
38.78
8.03
2.12
38.78
8.03
48.93
Std Dev
(9)
0.05
I
0.52
I
I
0.86
0.05
0.52
0.86
1.20

-------
                              Analyte Data
Hass Balance for Test 5, Toluene Mash
Test 5 Passes 1-3 10-2-88 Date

* Specific Gravity
* Density of Water (Ibs/gal)
Net Feed


- Drum |9/29
Toluene Wash
Toluene Wash
+ Sample Losses
vo
-P* Total Toluene Wash
* -
in

1 - orun 19/29
3 Toluene Wash
1 Toluene Wash
g + Sample Losses
9 Total Toluene Wash



- Drum A 19/29
Toluene Wash
1
Raw |
Data
100.00
0.87
8.35
723.11
=========
226.00
38.50
187.50
187.50
0.52
	
188.02


74.00
38.75
35.25
35.25
0.52
35.77


323.50
39.50
284.00
1
:orrected|
(Ibs)
100.00
0.87
8.35
723.11
=======
226.00
38.50
187.50
187.50
0.52
	
188.02
========


74.00
38.75
35.25
35.25
0.52
35.77
————————


323.50
39.50
284.00
/ariance
(lbs)"2
2.08
0.00
0.00
145.25
========
0.00
0.00
0.00
0.00
0.00
0.00


0.00
0.00
0.00
0.00
0.00
0.00
=ssssss=


0.00
0.00
0.00
Cone. |<
(ppa)
0.00

185.00



34.00



842.50
	 PC
/arionce
(ppn»A2
0.00

989.10



33.41



(20513.40
3 	
Mass |\
(9)
0.00

18.16
18.16
0.05
	
18.21


0.63
0.63
0.01
0.64


125.26
/arionce
0.00

7.17
0.00
0.00
0.00
=========


0.01
0.01
0.04
0.04
=========
-

341.02

-------