vvEPA
           United States
           Environmental Protection
           Agency
           Risk Reduction
           Engineering Laboratory
           Cincinnati, OH 45268
EPA/600/9-90/006
Feb. 1990
           Research and Development
Remedial Action,
Treatment and Disposal
of  Hazardous Waste

Proceedings of the
Fifteenth Annual
Research Symposium

-------

-------
                                                     EPA/600/9-90/006
                                                     Feb. 1990
         REMEDIAL ACTION, TREATMENT, AND DISPOSAL
                    OF HAZARDOUS WASTE
  Proceedings of the Fifteenth Annual Research Symposium
             Cincinnati, OH, April 10-12, 1989
Sponsored by the U.S. EPA, Office of Research & Development
           Risk Reduction Engineering Laboratory
                   Cincinnati, OH  45268
                      Coordinated by:

                        JACA Corp.
                Fort Washington, PA  19034

                            and

                      PEI Associates
                   Cincinnati, OH  45246
                     Project Officers:
                     Eugene F. Harris
                        John Glaser
                       Teri Shearer
                   Cincinnati, OH  45268
           RISK REDUCTION ENGINEERING LABORATORY
            OFFICE OF RESEARCH AND DEVELOPMENT
           U.S. ENVIRONMENTAL PROTECTION AGENCY
                   CINCINNATI, OH  45268

-------
                                  NOTICE
      These Proceedings have been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication.  Mention of trade names  or
commercial products does not constitute endorsement or recommendation for
use.
                                    -11-

-------
                                 FOREWORD
      Today's rapidly developing technologies and industrial practices fre-
quently carry with them the increased generation of materials, that if
improperly dealt with, can threaten both public health and the environment.
The U.S. Environmental Protection Agency is charged by Congress with pro-
tecting the Nation's land, air, and water resources.  Under a mandate of
national environmental laws, the agency strives to formulate and implement
actions leading to a compatible balance between human activities and the
ability of natural systems to support and nurture life.  These laws direct
the EPA to perform research to define our environmental problems, measure
the impacts, and search for solutions.

      The Risk Reduction Engineering Laboratory is responsible for
planning, implementing, and managing research, development, and demonstra-
tion programs.  These provide an authoritative, defensible engineering
basis in support of the policies, programs, and regulations of the EPA with
respect to drinking water, wastewater, pesticides, toxic substances, solid
and hazardous wastes, and Superfund-related activities.  This publication
is one of the products of that research and provides a vital communication
link between researchers and users.

      These Proceedings from the 1989 Symposium provide the results of proj-
ects recently completed by RREL and current information on projects pre-
sently underway.  Those wishing additional  information on these projects
are urged to contact the author or the EPA Project Officer.

      RREL sponsors a symposium each year in order to assure that the
results of its research efforts are rapidly transmitted to the user com-
munity.  The 1989 symposium attracted over 900 attendees from industry,
Federal and State agencies, consulting firms, and universities.  The 1990
symposium is planned for April  3, 4, and 5 in Cincinnati, OH.

                        E. Timothy Oppelt,  Director
                   Risk Reduction Engineering Laboratory
                                   -iii-

-------
                                 ABSTRACT


      The Fifteenth Annual Research Symposium on Remedial  Action,  Treatment,
and Disposal of Hazardous Waste was held in Cincinnati, OH, April  10-12,
1989.  The purpose of this Symposium was to present the latest significant
research findings from ongoing and recently completed projects funded by
the Risk Reduction Engineering Laboratory (RREL).

      These Proceedings are organized in four sections: Sessions A, B, and
A/B consist of paper presentations.  Session C contains the poster
abstracts.  Subjects include remedial action treatment and control tech-
nologies for waste disposal, landfill liner and cover systems, personnel
protection, underground storage tanks, and demonstration and development of
Innovative/alternative treatment technologies for hazardous waste.  Alter-
native technology subjects include thermal destruction of hazardous wastes,
field evaluations, existing treatment options, emerging treatment pro-
cesses, waste minimization, and biosystems for hazardous waste destruction.

-------
                                 CONTENTS


                                 SESSION A

                                                                     Page

Development and Validation of a Surrogate Metals Mixture
  R.G. Barton, Energy and Environmental Research Corporation.	  1-10

Incinerating Ethylene Dibromide and Dinoseb Stocks
  Donald A. Oberacker, U.S. Environmental Protection Agency..	  11-20

Pyrolytic Thermal Degradation of a Hazardous Waste Incinerability
Surrogate Mixture
  D.A. Tirey, University of Dayton Research Institute	  21-31

Incinerability Ranking of Hazardous Organic Compounds
  Robert E. Mournighan, U.S. Environmental Protection Agency*	  32-42

A Prototype Baghouse/Dilution Tunnel System for Particulate
Sampling of Hazardous and Municipal Waste Incinerators
  P.M. Lemieux, U.S. Environmental Protection Agency	  43-49

Evaluation of Alternative Treatment Technologies for Hazardous
Wastes from Acrylonitrile Production
  E. Radha Krishnan, PEI Associates, Inc	  50-63

The Role of Site Investigation in the Selection of Corrective
Actions for Leaking Underground Storage Tanks
  Myron S. Rosenberg, Camp Dresser & McKee Inc	  64-82

Summary of the Results of EPA's Evaluation of Volumetric
Leak Detection Methods
  Joseph W. Maresca, Jr., Vista Research, Inc	  83-98

An Outreach Process:  Case Histories of Underground Storage
Tank Corrective Actions
  William M. Kaschak, COM Federal Programs Corporation.	  99-108

Considerations of Underground Storage Tank Residuals at Closure
  Warren J. Lyman, Camp Dresser & McKee, Inc	  109-123

Laboratory Studies of Vacuum-Assisted Steam Stripping of Organic
Contaminants from Soil
  Arthur E. Lord, Jr., Drexel University	  124-136

Low Temperature Thermal Desorption for Treatment of Contaminated
Soils Phase II Results
  Richard P. Lauch, U.S. Environmental  Protection Agency....	  137-150
                                    -v-

-------
                           SESSION A (Continued)
                                                                     Page
Detection of Macro Defects in Soil-Bentonite Cutoff Walls
  Andrew Bodocsi, University of Cincinnati	..	   151-163
A Field Test of Hydraulic Fracturing in Glacial Till
  L.C. Murdoch, University of Cincinnati..	   164-174
Computer-Based Methods of Assessing Contaminated Sites:  A Case
History                                                           •    ,c ,„
  W.G. Harrar, University of Cincinnati	   1/5-185
Results and Preliminary Economic Analysis of an APEG Treatment
System for Degrading PCBs in Soil
  John A. Wentz, PEI Associates, Inc	   186-200
Destruction of Cyanides in Electroplating Wastewaters Using
Wet Air Oxidation
  H. Paul Warner, U.S. Environmental Protection Agency	   201-208
Determining Cost Effective Approaches to  the Environmental
Control of Electroplating Operations
  John 0. Burckle, U.S. Environmental Protection Agency.....	   209-227
                                    -vi-

-------
                                 SESSION B

                                                                     Page

PCB Degradation:  Status and Directions
  P.R. Sferra, U.S. Environmental Protection Agency	  228-240

The Development of Recombinant Bacteria for Polychlorinated
Biphenyl Degradation
  Frank J. Monde!lo, GE Research and Development Center	  241-250

Treatment of Wood Preserving Soil Contaminants by White Rot Fungus
  John A. Glaser, U.S. Environmental Protection Agency	  251-263

Biological Treatment of Petrochemical Sludges
  Stephen D. Field, Louisiana State University	  264-272

The Determination of Biodegradability and Biodegradation
Kinetics of Organic Pollutant Compounds with the Use of
Electrolytic Respirometry
  Henry H. Tabak, U.S. Environmental Protection Agency	..273-296

Prediction and Modeling of "Biodegradation Kinetics of Hazardous
Waste Constituents
  Rakesh Govind, University of Cincinnati.......	  297-311

Preliminary Results on the Anaerobic/Aerobic Biochemical Reactor
for the Mineralization of Organic Contaminants Bound on Soil Fines
  Robert C. Ahlert, Rutgers University	  312-329

Fate and Effects of RCRA and CERCLA Toxics in Anaerobic Digestion
of Primary and Secondary Sludge
  Richard A. Dobbs, U.S. Environmental Protection Agency	  330-339

Fate and Effects of Selected RCRA and CERCLA Compounds in
Activated Sludge Systems
  San joy K. Bhattacharya, University of Cincinnati	  340-349

Compatibility of Flexible Membrane Liners and Municipal Solid
Waste Leachates
  Henry E. Haxo, Jr., Matrecon,  Inc	  350-368

Geosynthetic Concerns in Landfill Liner and Collection Systems
  Robert M. Koerner, Drexel University	  369-378

Attenuation of Priority Pollutants Codisposed with MSW in
Simulated Landfills
  Frederick G. Pohland, University of Pittsburgh	  379-395

Site Demonstration of Hazcon Process
  Paul R. de Percin, U.S. Environmental Protection Agency	  396-408
                                   -VT1-

-------
                           SESSION B (Continued)
Site Demonstration of the Terra Vac In Situ Vacuum Extraction
Technology
  Peter A. Michaels, Foster Wheeler Enviresponse, Inc	   409-426

The Office of Research & Development WRITE Program
  Ivars J. Licis, U.S. Environmental Protection Agency	   427-436

Solidification/Stabilization as a Best Demonstrated Available
Technology for Resource Conservation and Recovery Act Wastes
  R. Mark Bricka, Department of the Army	   437-447

Volatile Emissions from Stabilized Waste
  Leo Weitzman, Acurex Corporation	   448-458

Technologies Applicable for the Remediation of Superfund
Radiation Sites
  Ramjee Raghavan, Foster Wheeler Enviresponse, Inc	   459-469
                                  -viii-

-------
                                SESSION A/B
                                                                     Page
RRELevant QA Diagnostics:  Major Findings in FY 88
  Guy F. Simes, U.S. Environmental Protection Agency	   470-479

Effect of Feed Characteristics on the Performance of EPA's
Mobile Incineration System
  James P. Stumbar, Foster Wheeler Enviresponse, Inc..	   480-498

Long-Term Field Demonstration of the Linde® Oxygen Combustion
System Installed on the EPA Mobile Incinerator
  Min-Da Ho, Union Carbide Industrial Gases Inc	   499-514

In Place Treatment of Contaminated Soil at Superfund Sites:  A
Review
  M. Roulier* U.S. Environmental Protection Agency	   515-525

RREL Expert Systems Project:  Developing Tools for Hazardous
Waste Management
  Jay E. Clements	   526-534

Assessment of Chemical and Physical Methods for Decontaminating
Buildings and Debris at Superfund Sites
  Michael L. Taylor, PEI Associates, Inc	   535-556
                                    -IX-

-------
                                 SESSION C
                                                                     Page
Immobilization Mechanisms in Solidification/Stabilization Using
Cement/Silicate Fixing Agents
  L.G. Butler, Louisiana State University	  557

Waste Reduction Evaluations at Federal Sites
  James S. Bridges, U.S. Environmental Protection Agency	  558

Demonstrate Computer Assisted Engineering (CAE) Techniques
for Remedial Action Assessment
  P.R. Cluxton, University of Cincinnati.	  559

Hydraulic Mechanisms of a Multiple Soil Layer Cover
  Richard C. Warner, University of Kentucky	  560

Evaluation of Solidification/Stabilization Treatability Studies
at the United States Environmental Protection Agency Center
Hill  Facility
  Edwin F. Barth, U.S. Environmental  Protection Agency	  561

The EPA Manual for Waste Minimization Opportunity Assessments
  Mary Ann Curran, U.S. Environmental Protection Agency.....	  562

The U.S.  EPA Combustion Research  Facility
  Johannes W. Lee, Acurex Corporation	  563

Evaluating the Cost  Effectiveness of  SITE Technologies
  Gordon  M. Evans, U.S. Environmental Protection Agency.....	  564

Soliditech Site  Demonstration
  Walter  E. Grube, Jr., U.S. Environmental Protection Agency	  565-566

BioTrol Soil Washing System
  Steven  B. Valine,  BioTrol, Inc	  567

Separation  of Hazardous Organics  by  Low Pressure Reverse  Osmosis
Membranes
  M.E. Williams,  University of  Kentucky	  568-582

Testing of  a  Leachate  Treatment System  Based on a Wood
Degrading Fungus
  John A. Glaser, U.S. Environmental  Protection Agency	  583

Assessment  of KPEG Treatment for PCB  Contaminated Soils
  Alfred  Kernel,  U.S.  Environmental  Protection Agency	  584
 State-of-the-Art Field Hydraulic Conductivity Testing  of
 Compacted Soils
   Joseph 0.  Sai, K.W.  Brown & Associates,  Inc	
585
                                     -x-

-------
                           SESSION C (Continued)
                                                                     Page
Review of Soil Washing Technologies for Soils Contaminated
with Heavy Metals
  Carl Gutterman, Foster Wheeler Enviresponse, Inc	   586

Rotary Kiln Incineration
  V.A. Cundy, Louisiana State University	   587-588

Biological Treatment of Chiorophenol-Contaminated Groundwater
  Thomas J. Chresand, BioTrol, Inc....	   589

Computerized Management and Dissemination of Information for
Research and Development Operations at the Technical Information
Exchange (TIX) Edison, NJ
  May Smith, Enviresponse, Inc	   590-591

The EPA Treatability Database
  Stepahanie A. Hansen, Radian Corporation	   592-593

Chemical Treatment of Metals in Wastewaters and Sludges at
the T&E Facility
  Douglas W. Grosse, U.S.  Environmental Protection Agency	   594

Biological Degradation of  Chlorinated Phenoxy Acids
  R.A. Haugland, University of Illinois at Chicago.....	   595

Use of Foam  Technology for Control  of Toxic Fumes During
Excavation at Superfund Sites
  Ramjee  Raghavan, Foster  Wheeler  Enviresponse, Inc.........	  596

Soil  Washing—Removal of Semivolatile Organics Using Aqueous
Surfactant Solutions
  Edward  Coles,  Foster Wheeler Enviresponse,  Inc	  597

Preparation  for  SITE  Demonstration of a Powdered Activated Carbon
Treatment (PACT) Unit
  John F.  Martin,  U.S. Environmental  Protection Agency	  598
                                    -XI-

-------

-------
           DEVELOPMENT AND VALIDATION OF A SURROGATE METALS MIXTURE
                 R. G. Barton, W. D. Clark, and W. R. Seeker
                Energy and Environmental Research Corporation
                              Irvine,  California
                                  C. C. Lee
                     Risk Reduction  Engineering Laboratory
                    U. S. Environmental Protection Agency
                               Cincinnati,  Ohio
                                   ABSTRACT
     The U.S. Environmental Protection Agency (EPA) is developing regulations
to   control   the   burning   of  metal-bearing  wastes   in   hazardous  waste
incinerators.   These regulations affect the majority  of the incinerators in
the U.S. since  nearly  every waste contains at least trace quantities of some
toxic  metals.     Recent  research   indicates  that   metal   emissions  from
incinerators are controlled  by a number of parameters  including metals type,
incinerator temperature and  waste chlorine content.  This makes it difficult
to  define  appropriate permit  conditions  which  will  guarantee  acceptable
emissions of all toxic metals given the data obtained from current trial burn
procedures.  A  surrogate  metals mixture was developed  to provide a coherent,
defensible method for evaluating  the behavior of metals  in incinerators.  The
mixture  consists of four  surrogate  metals whose behavior  can  be  used  to
predict  the  behavior of  toxic metals.   Components of  the  surrogate metals
mixture  were  selected  based  on  five  criteria  -  volatility,  toxicity,
abundance, chemical  species, and  cost.

     A  test  series  was  planned to  verify the  appropriateness  of  using  a
surrogate mixture and determine if the  metals  selected were  acceptable.  The
test  program involved  spiking  a synthetic waste with  both  the  surrogate
metals  and toxic  metals,  burning the  waste  under a variety of  operating
conditions and  comparing  the surrogate metals behavior  with  the  behavior of
toxic metals.

     A  validation  test was carried  out  at  the  EPA's Combustion  Research
Facility (CRF).   The results  from this test were analyzed  and the validity
of the surrogate metals approach was assessed.

-------
                                 INTRODUCTION
     The  EPA  is  developing  regulations to  control  the  burning  of  metal
bearing  wastes  and   has   issued   guidance   on   metals  related  permitting
procedures for  hazardous waste incinerators  (1).   The guidance  is  based on
dispersion modeling  and worst  case  risk  assessment.    Federal  and  State
officials are implementing the guidance.  In addition, States are instituting
their own regulations restricting metals emissions.  This body of regulations
is likely to affect the majority of the incinerators in the U.S. since nearly
every waste contains at least trace quantities of some toxic metals.

     Recent research  indicates  that toxic metals emissions from incinerators
are controlled  by  a number of  parameters,  including incinerator temperature
and waste chlorine  content (2).  In the light of these findings,  it will be
difficult  to  define  appropriate   permit  conditions   which  will  guarantee
acceptable metals  emissions from waste  incinerators given  the  data  obtained
from current trial  burn procedures.

CURRENT REGULATORY APPROACH

     The  approach   that will   be  used  to   regulate metals  emissions  from
combustion  devices  burning metal-bearing  wastes  has  not  yet been  fully
defined.   Current  guidance  calls  for  the  establishment of  three  different
evaluation criteria  or "tiers"  (1).  A  facility which  can meet the  criteria
associated with any one  of the  tiers will  be  allowed to  operate.    The
criteria  are  based  on  dispersion  modeling  and  worst  case  risk assessment.
Under  Tier  III,  site  specific dispersion  modeling  is  used   to  determine
allowable  emissions  based on  predicted ambient concentrations.    Tier II
establishes  allowable   emissions   limits   based  on  reasonable  worst-case
dispersion modeling.   Tier I sets  limits on metals feed rates.  The Tier I
rates are back-calculated from Tier II emission limits by assuming all of the
metals in the waste are  emitted.

     In general, if  a facility can  qualify  under any one of the three tiers,
it will  be permitted to burn the waste  in question.   It will not be allowed
to burn a waste with  higher metals concentrations.  The current guidance does
address the potential impact of operating conditions.

RATIONALE FOR MODIFYING  CURRENT APPROACH

     Tier I is adequate  but very conservative.   In many cases, operators will
be unable  to meet the  criteria set forth and  will  be  forced  to use either
Tier II or Tier III.  However the procedures associated with Tiers II and III
are  subject  to a  number  of   limitations  which  make   it  difficult to  set
appropriate permit  conditions  given the data to be  gathered.  The principal
limitations are as follows:

     t    Waste sampling is often  difficult  and unreliable  due  to  the non-
          homogeneous nature and low metals content of most wastes.

     •    Behavior of many metals is not well characterized.

-------
     •    Emission  of toxic  metals  may  present  a  danger during  the  trial
          burn.                                                 '

     •    Tests involving a large number of metals are costly.


PROPOSED MODIFICATION

     Because of the significant  disadvantages associated with monitoring all
of  the  toxic metals  present  in  a  hazardous waste  during a  trial  burn,  an
alternative  approach  was developed.    This  approach is centered on  use of a
mixture  of  non-hazardous  surrogate  metals.   The  mixture consists  of four
non-toxic metals  which  are added to  wastes  prior to  incineration  during a
trial burn.   The  emissions  of these  four  metals are  representative of the
emissions of all metals.

     Use of  a  surrogate mixture  has  a number of  advantages.   The emissions
are  not dangerous  because non-hazardous  metals  are   used  in the  mixture.
Metals  whose behaviors are well  defined  are introduced into  the  waste in a
known physical  and  chemical  form.  Thus  the  interactions  experienced by the
metals will  be known  and the  relationship between the  behavior of the metals
during  the  trial   burn  and  during  actual  operation  will  be known.   In
addition,  use  of  the  same   metals   in  all  trial  burns  will  allow  the
establishment of a  database on metals  behavior.   Since only four metals need
be spiked and monitored, the costs associated with using the mixture would be
only  slightly  greater  than  those associated with  running  a  current  trial
burn.

     There are disadvantages  associated with  the  surrogate mixture approach.
The   most   significant  is   the  fact   that  the  concept   has   not   been
experimentally verified.  While  in theory it  is  possible to use a few metals
to  indicate  the behavior  of  all metals, in  practice  each metal  may behave
differently.   Unanticipated  interactions  and kinetic  limitations  can have
different  effects  on  different metals.     A  second  disadvantage  is  the
significant  amount  of  engineering  analysis  which  is   required to determine
permit conditions based  on data  obtained  from trial  burns  in which surrogate
metals  are  used.    A final disadvantage  is that  the technique neglects the
actual  physical  form  of a metal in  the  waste.   Metals  in the mixture are
always  added in the same physical and chemical form while the metals in the
waste may be present  in  a  different  form.   The physical and chemical form of
the metals may have an important impact on their behavior.
                        SELECTION OF SURROGATE METALS
APPROACH
     Components of  the surrogate metal mixtures  were selected based  on the
following criteria:

          •    Metals volatility

          •    Toxicity

-------
           •    Abundance

           •    Chemical  Species

           •    Cost
 SELECTION  OF  METALS

 Volatility

      The   particles  which  have  the greatest  potential  to  escape capture
 contain  relatively  high  concentrations  of metals  which  vaporize  in  the
 incinerator.   The quantity of  a  metal  emitted  by an incinerator is directly
 related to the volatility of the metal.  The surrogate metals should reflect
 the  range  of  volatilities.   Table 1 lists  toxic metals  ranked  based on the
 temperature  at which  their  effective vapor pressure  is  1 x  10'6  atm.   The
 effective  vapor pressure is  the combined vapor pressures of all  of  the metal
 species that  would be  present in  the  incinerator at equilibrium.

 Toxicitv

      It  is desirable  to  use  nontoxic surrogate metals  in  trial  burns.   For
 this  analysis,  a metal is classified as toxic if it is listed in 40 CFR 261,
 Appendix  VIII.   All   other metals are  considered  nontoxic.   Table 2 lists
 selected nontoxic metals ranked  according to the temperature  at which their
 effective  vapor pressure is 1  x  10'6 atm.   Surrogate  metals can be selected
 from  this  list.   One surrogate  metal  should be volatile  at  all  reasonable
 temperatures  and  one should be  nonvolatile at all reasonable temperatures.  A
 metal  which  is  volatile  at  low  kiln  temperatures  and  a  metal   which  is
 volatile at high  kiln  temperatures would also be selected.

 Abundance

      In  order  for   surrogate  metals  to  negate  the   effects   of  waste
 variability,  the  amount  added  must  be  much  greater  than  the  quantity
 originally in  the waste.   Thus, metals commonly found  at high concentrations
 in the waste  should  not  be used as surrogate metals.  If they were used, the
 cost  of doping  would  be  prohibitively high and  the  metal and its media would
 alter the  characteristics of the waste.  It was assumed  that the  surrogate
 metals should  be doped at a level  ten  times greater than the expected metal
 concentration  in the  waste.   An  analysis of the impact  of  various mixture
 media  (such  as  water and #2   fuel  oil) on  the  characteristics of typical
 wastes and the solubility of metal compounds in  the media indicates that no
 more than  5000 ppm of  any one metal should be added.  Thus metals selected as
 surrogates  should be  those which are  not  typically present  in wastes  at
 concentrations higher  than 500  ppm.  However, since  the cost of adding metals
 increases  with  the  amount added  it is desirable to choose metals  which are
 present in concentrations much  less than 500 ppm.

 Chemical  Species

     One of the key assumptions  associated  with  the selection  of  surrogate
metals  is  that  all   metals  related  reactions  are very   fast.    Any  metal

-------
Table 1.  Volatility of toxic metals (waste chlorine concentration = 10 %)
Metal
Lead
Mercury
Arsenic
Osmi urn
Thai 1 i urn
Cadmium
Selenium
Silver
Antimony
Nickel
Barium
Beryl 1 i urn
Vanadium
Chromium
Temperature*, K
260
290
310
310
410
490
590
900
930
970
1180
1330
1610
1880
Temperature*, °F
5
60
90
110
280
420
610
1160
1220
1280
1660
1930
2430
2930
Principal Vapor Species
PbCl4
Hg
As203
0564
T10H
Cd
Se02
AgCl
Sb203
NiCl2
Bad 2
Be(OH)2
V02
Cr02/Cr03
     Temperature at which the effective  vapor  pressure  is  1  x 10'6 atm.  The
     effective  vapor  pressure  is  the combined  vapor  pressures  of all  the
     species involving the metal of interest which are present at equilibrium
     under given conditions.
Table 2.  Volatility  of  nontoxic  metals   (waste  chlorine  concentration  =
          10 %)
Metal
Boron
Copper
Cesium
Lithium
Potassium
Sodium
Bismuth
Iron
Zinc
Calcium
Strontium
Tin
Magnesium
Silicon
Titanium
Tantalum
Thorium
Zirconium
Temperature*, K
< 250
< 400
770
780
840
870
890
890
1050
1130
1190
1320
1410
> 2000
> 2000
> 2000
> 2000
> 2000
Temperature*, °F
< -10
< 260
930
950
1050
1100
1150
1150
1430
1580
1690
1920
2070
> 3140
> 3140
> 3140
> 3140
> 3140
Principal Vapor Species
H3B306
CuoCls
CsCl2/Cs2Cl2
Li2CT2/LiCl
KC1/K2C12
NaCl/Ra2Cl2
Bi
FeCT2
Zn/ZnO
CaCl2
SrCl2
SnO
MgCl2
SiO/Si02
Ti02
Ta02
Th02
ZrO
     Temperature at which the effective vapor pressure  is  1  x  10"6 atm.   The
     effective vapor  pressure is  the combined  vapor pressures  of all  the
     species involving the metal  of interest which are present  at equilibrium
     under given conditions.

-------
present  in the  waste  will  quickly  react to  form  the  species  favored  by
thermodynamic equilibrium.  Thus the species in which each surrogate metal  is
introduced into the waste is generally unimportant.

     However, care must be  taken  not  to introduce elements which could alter
the species favored  by equilibrium or form other inorganic  pollutants.   For
example, the presence  of chlorine  leads  to the formation of metal chlorides.
Metal  chlorides  are generally  more volatile  than  other metal  species.   In
addition, chlorine leads  to  the formation  HC1.   Thus, if a metal were intro-
duced  as a  metal  chloride in a waste which does not normally contain chlor-
ine, metals  and  HC1  emissions would  be  greater than they  would  be  under
normal  operating conditions.

     Organometallic compounds warrant special  consideration.   When a metal  is
present as an organometallic compound it is generally much more volatile than
when present in inorganic form.  Thus if a waste contains an organometal  then
the surrogate mixture should also contain an appropriate organometal.

Cost

     The  cost  of  each  surrogate metal  species   is  also  an  important
consideration.   The costs  of  the chemicals   vary  greatly depending on  the
particular species of  interest.  In general,   inorganic  salts are relatively
inexpensive  ranging  from $  9.00  to  $  136.00  per  pound.    Organometallic
compounds  are usually  more expensive  ranging  from around  $  9.00  to  over
$ 114,000.00 per pound.

POTENTIAL MEDIA

     The medium used to introduce the metals  into  the  waste  must meet  three
criteria:
          a    It must be capable of dissolving the metal  species.
          •    It must be relatively safe and  easy to handle.
          t    It  should   not   alter  the  characteristics   of  the  waste
               significantly.

Water  and  dilute  aqueous acids are good  media for  introducing  a surrogate
mixture to  a  solid waste or an aqueous based waste.   Aqueous  solutions  can
not be used in conjunction  with  organic  liquid wastes which are  burned  iri
liquid  injection  incinerators.     The  water   would   interfere  with   the
atomization and flame patterns.   In those  cases an organic fuel, like number
2 fuel  oil, can be used.   The fuel oil  would  be soluble in the organic  waste
and thus  would not  interfere  with the  combustion processes.   Wastes  which
contain organometallic  compounds will  also require the use of  organic  media
since organometallic compounds are usually insoluble in water.

RECOMMENDED SOUPS

     The selection of surrogate metals based on the  criteria  discussed  above
is summarized in Table  3.   The  costs  shown in  the tables are based on retail
prices  for  small  lots.   The  price for the quantities  required  for  a  trial
burn will probably be less.
                                      ,6

-------
     The nontoxic metals chosen for use in trial burns are:

               Copper
               Bismuth
               Strontium
               Magnesium

     Copper is very volatile under common combustion conditions.  Bismuth and
strontium exhibit intermediate volatility and are usually only encountered in
very low  concentrations  in wastes.  Magnesium  is  not  volatile at reasonable
incinerator conditions.  All of the  metals  are available as soluble nitrates
which range in cost from $ 14.00 to $ 18.00 per pound.
                              VERIFICATION  TESTS
OBJECTIVES
     A series of tests are needed to verify the ability of surrogate mixtures
to represent the behavior  of many different metals.   A number of assumptions
are associated with the development of surrogate mixtures.  These assumptions
are listed  in  Table 4.   The verification tests must  assess  the validity of
these  assumptions.    In  addition,  the  tests  should  examine   the  specific
mixture proposed and determine if it is appropriate.

PRELIMINARY RESULTS

     A series of tests using the surrogate metals mixture were carried out at
the EPA's  Combustion  Research Facility  (CRF).  The tests  involved  burning a
synthetic waste spiked with both hazardous metals and the non-toxic surrogate
metals in a pilot scale  rotary  kiln.   The impacts of chlorine concentration,
kiln  temperature  and  afterburner  temperature  on   metals  emissions  were
examined.   Metals  emissions  data  from  the  tests  are  not  yet  available.
However, information  on the  composition  of the  residual  ash  is  available.
Figure 1 shows  the  impact of chlorine concentration  and  kiln temperature on
the normalized enrichment of each metal in the ash.  Enrichment is defined as
the ratio of the concentration of a metal  in the ash to the concentration in
the  feed.    Magnesium  is  not  expected  to  vaporize  under  the  conditions
encountered in  the  kiln.   Thus,  the enrichment  of each metal  was normalized
by the enrichment of  magnesium to  account  for  dilution effects.  Normalized
enrichments close to 1.0 indicate that little vaporization has occurred while
values close to 0.0 indicate that most of the metal has vaporized.

     The surrogate  metals cover the  range of toxic  metals  behavior  well.
Copper and bismuth  are  both  very volatile.  Strontium exhibits intermediate
behavior, while  magnesium  is non-volatile.   However,  it must  be  emphasized
that these are  preliminary results.  Additional data  will  be  analyzed as it
becomes available.

                                   SUMMARY
     A surrogate metals mixture can be  used  to  aid in the analysis of metals
behavior  in  waste  incinerators.     The  surrogate  mixture  can  be used  to
facilitate the following activities:

-------
Table 3.  Summary of non-toxic surrogate metals selection.
Metal
Compound
Metal Volatility Cost**
Temperature , °F $/g
Abundance
of metal , ppm (3)
Inorganic: Medium - Dilute Nitric Acid
Copper
Bismuth
Strontium
Magnesium
Cu(N03)2
Bi(N03)3
Sr(N03)2
Mg(N03)2
< 260 0.03
1150 0.04
1690 0.01
2070 0.02
1.5 - 550 (4)
ND
1.08
5.0 - 590
Organometallic: Medium - Number 2 Fuel Oil
Metal
Copper
Bismuth
Strontium
Magnesium
Compound
Copper(II) 2,4 -
pentanedionate,
CuC10H14°4
Triphenyl Bismuth,
Strontium 2,4 -
pentanedionate,
Magnesium 2,4 -
pentadionate,
Mg(C5H702)2
Compound
Boiling Point, °F
ND
M.P. = 172
M.P. = 428
ND
Cost**
$/g
0.14
1.08
0.52
0.25
ND - No Data

     effective  vapor pressure  is  the  combined  vapor pressures  of  all  the
     species involving the metal of interest which are present at equilibrium
     under given conditions.
**
     1987  retail  price for small  lots.
     could be significantly less.
The actual  price  for a trial  burn
Table 4.  Assumptions made during surrogate metals mixture development.

     •    The  initial   form  of  a  metal  is  important  only when  extremely
          volatile metal, compounds are present.

     §    Dispersed metals will vaporize to a larger extent than agglomerated
          metals.

     •    Metal interactions will have a negligible impact on behavior.

     •    The use of a limited number of metals is sufficient to indicate the
          behavior of all toxic metals.
                                       8

-------
          LU
( .U

0.7!
t
NORMALIZED ENRICHMEI
p o
9 ro en
in et
L
rc
„ t^
^B
^m
•••
i
] T * 1600°F, 3X C1
I T » 1700°F, 3X C1
1

Wi
I




V/////////////////////A

••
I






1
1
'/////////////////////////A




mm
Y/////////////////////////A
-
As
cd
Ba   Sr
                                                       Cr
Figure 1. Preliminary  analysis  of  results  from  the  Combustion  Research
          Facility  tests  involving  surrogate  metals  mixture.    Surrogate
          metals are underlined.

-------
          •    Determination  of  the  maximum  allowable  concentrations  of
               toxic metals in incinerated wastes.

          •    Establishment of permit conditions

          t    Evaluation of APCD capabilities.

          e    Establishment of a common base of information  on  the behavior
               of metals in incinerators.


                                 REFERENCES
(1)   Versar  Inc.,  "Guidance  on Metals  and Hydrogen  Chloride Controls  for
     Hazardous Waste Incinerators,"   U.S.  EPA,  Office of Solid Waste,  March
     1988.

(2)   Barton,  R.G.,  et  al., "Prediction of the Fate of Toxic  Metals in  Waste
     Incinerators," 13th National Waste  Processing Conference. Philadelphia,
     May 1-4, 1988.

(3)   Fennelly,  P.,  et  al./'Environmental   Characterization  of  Waste  Oil
     Combustion     in  Small  Boilers," Hazardous  Waste.  1 (4), 1984, p 489.

(4)   Kristensen, A.,"Operating  the  Rotary Kiln  Incinerators  at Kommunekemi,"
     Hazardous Waste and Hazardous Materials. 2  (1), 1985,  p  7.
                                     10

-------
               INCINERATING  ETHYLENE  DIBROMIDE  AND  DINOSEB  STOCKS

                   by:   Donald A. Oberacker
                        USEPA, RREL
                        Cincinnati, OH   45268

                        and

                        Carol Stangel
                        USEPA, OPP
                        Washington, DC   20460

                                   ABSTRACT


      Pursuant to  the Federal Insecticide, Fungicide, and  Rodenticide Act
 (FIFRA), pesticides for which registrations have been suspended due to
 imminent hazard and then finally cancelled based on findings of unreasonable
 adverse effects on human health and  the environment can no longer be mar-
 keted and used in  the U.S.  for their intended  purposes.  In several cases to
 date, after initiating  such actions to terminate the use of a pesticide, the
 U.S. Environmental Protection Agency (EPA) has been obligated under Section
 19 of FIFRA to indemnify owners of the suspended and cancelled pesticide and
 to accept products for  interim storage and/or  safe disposal.  The 1988
 amendments to FIFRA will shift the responsibility for and the cost of
 disposal of suspended and cancelled pesticides to their manufacturers, in
 the future.  EPA will continue to be involved  in the recall, storage and
 disposal process but will assume an oversight  role.  However, at present,
 EPA must fulfill its pre-FIFRA '88 obligation  to complete the disposal of
 ethylene dibromide (EDB), dinoseb and 2,4,5-T  and silvex pesticides.

      This paper summarizes EPA's progress in  achieving the proper destruc-
 tion of two pesticide inventories,  ethylene dibromide (EDB) and dinoseb, at
 commercial  hazardous waste incineration facilities.  Over three hundred
 thousand gallons of EDB were incinerated during 1988 and in early 1989
 using a special technique involving the presence of sulfur during
 incineration that prevented the release of bromine to the atmosphere.
 EPA is currently initiating a similar program that will  result in the
 incineration of two to four million gallons of dinoseb.   In this case,
the major technical concern is the  potential  for NOx generation due to
the nitrogen content of dinoseb.

      The characteristics of EDB and dinoseb,  the disposal  options con-
sidered for treatment and disposal  of each pesticide,  the results of
                                    11

-------
pilot or field-scale testing, the incineration method ultimately
selected, and incinerator performance testing results are also addressed
in this paper.

                              INTRODUCTION


      After EPA suspended and cancelled the registrations of the
pesticides ethylene dibromide (EDB) and dinoseb, and was required by
FIFRA to accept stocks of these pesticides for disposal, the Agency
essentially was placed in the position of "waste generator" of these
materials, which represented a large quantity of RCRA hazardous wastes.
While this was by no means the first occasion that the government or EPA
became a waste generator, the chemical characteristics of EDB and
dinoseb did present interesting challenges in terms of seeking
environmentally safe, viable, and economic disposal methods.  Neither of
these pesticides had ever been commercially disposed of in any
significant quantity or by any commercial scale, carefully evaluated and
proven hazardous waste treatment methodology.  Except for
unsubstantiated claims that these or similar materials had been
incinerated in small amounts by the commercial incineration or chemical
manufacturing industry, essentially no meaningful data existed to guide
EPA  in disposing of the EDB and dinoseb stocks for which the Agency was
responsible.

      Besides incineration, EPA carefully considered a host of other
treatment options including chemical detoxification, materials recovery,
deep-well injection, and distillation techniques.  These other options
all  held some potential for treating EDB and dinoseb, but each was
deemed to require extensive sub-scale development and evaluation to
demonstrate its effectiveness.  All of the non-incineration options
appeared to need performance testing, scale-up, design, and permitting
activities most of which would be very time and cost intensive.  Deep
well  injection was rejected for both pesticides (except for weak
rinsates from EDB containers) due to uncertainties in predicting long-
term environmental effects, though  it appeared to be the least costly
disposal solution.  The various chemical detoxification and distillation
schemes  studied for both EDB and dinoseb were ultimately ruled out for
reasons  of long process equipment development, projections of
uncertainties in the permitting process, and the likelihood that any
recovered chemicals of value would  not be of assured purity and
marketability.

      High temperature incineration held a seemingly more  immediate,
safe, and economically competitive  potential provided answers were found
for  EDB's bromine control  issue and for dinoseb's tendency to generate
nitrogen  oxides or NOx.  This paper describes EPA's pilot- or field-
scale incineration tests which  investigated these particular performance
issues  and also generated  data  on destruction and removal  efficiency
(ORE)  and particulate emission  performance on these two pesticides.
Following these tests which  proved  successful,  EPA proceeded with
                                     12

-------
incineration as the selected treatment/disposal technology for both EDB
and dinoseb.

                           EDB  CHARACTERISTICS
      Ethylene dibromide or EDB is a liquid halogenated hydrocarbon
which was registered as a pesticide in 1948 and suspended in 1983 and
1984.  It was used largely in agriculture as a pre-plant soil fumigant
and to fumigate stored grain.  Human exposure to EDB resulted primarily
from grain-based food products, and mounting evidence beginning as early
as 1975 regarding EDB's carcinogenic, mutagenic, and adverse
reproductive effects ultimately led to suspension, cancellation, and
indemnification actions under FIFRA by 1985.  Between 1985 and 1988, EPA
made indemnity payments for and/or accepted for disposal a total of
approximately 329,000 gallons (3.7 million pounds) of EDB pesticides.

      As with most pesticides, there existed a variety of individual EDB
products formulations with a range of concentrations of active and inert
ingredients (associated solvents or vehicles, etc.) as well as a variety
of sizes and types of containers, both pressurized and non-pressurized.
Major constituents of EDB pesticide products included ethylene dibromide
(1.5 to 50 percent or more), ethylene dichloride (up to 45 to 60%),
carbon tetrachloride (16 to 80%), carbon disulfide (0 to 16%), sulfur
dioxide (dissolved, 0 to 3%), chloropicrin (0 to 38%), and small amounts
of diesel oil, naphtha, and pentane, etc., all expressed in terms of
individual component weight percentages.

      The most critical incineration characteristic of EDB was its
bromine content.  The EDB molecule itself (C2H4Br2) is approximately 85%
bromine by weight.  Unlike chlorine which readily combines with hydrogen
to form scrubbable HC1, past experience with brominated compounds shows
that thermal destruction will normally result in significant (and
visible) bromine gas (Br2)  emissions from an incinerator stack.

                         DINOSEB  CHARACTERISTICS
      Dinoseb pesticides have been used for several decades primarily as
contact herbicides to control broadleaf weeds, but also as desiccants to
dry vegetation on food crops in the fields and facilitate harvesting of
vegetable and seed crops, etc.  The active ingredient dinoseb is an
organo-nitrogen compound (2-sec-butyl-4, 6-dinitrophenol) manufactured
and formulated into over two dozen varieties of water and/or oil diluted
forms, all liquid in nature, except for the "technical" or "parent acid"
forms which are dry solids.  Some of dinoseb's challenging
characteristics relative to disposal, beyond it's high nitrogen content,
are the explosive nature of the dry solids, the tendency for certain of
the water-mixed formulations to precipitate solid salts of dinoseb upon
exposure to sub-40°F ambient temperatures,  and the tendency for
evaporation of alcohol or other low boiling point vehicles if handled in
open containers.  However, water or oil dilution readily controls the
                                     13

-------
explosion hazard issue for solids.  In addition, many of the
formulations contain significant amounts of sodium, calcium, and inert
material, characterizing dinoseb as a "salt waste" and raising
incineration issues of refractory attack and drawing attention to
residue and particulates.  No heavy or toxic metals are involved, nor do
dinoseb products (now wastes) contain chlorine.

      The human health risks associated with dinoseb are of similar
concern to the Agency as those of EDB.  Dinoseb was suspended by EPA
under an emergency order in October, 1986 and then cancelled in June of
1988 because of evidence indicating that it may cause birth defects,
male sterility, cataracts, damage to the immune system, and ecological
problems.  Limited evidence also establishes dinoseb as a possible
carcinogen.  To date, EPA has received requests for disposal of some 2.7
million gallons of liquid dinoseb and about 50,000 pounds of solid
materials, and EPA anticipates ultimately receiving requests for
disposal of a total of as many as 4 million gallons of liquids.

      Returning to the nitrogen issue, dinoseb's active ingredient
molecule contains nearly 11% nitrogen.by weight.  Various diluted or
formulated stocks contain from 1 to 6% nitrogen by weight, although the
average,nitrogen concentration of all stocks combined is closer to 1%.
Host of the stocks are of the water-based types.
                       ERA'S INCINERATION PROGRAMS
EDB
       Initially,  EPA  considered  incineration of EDB to be a low
feasibility disposal  option due  to the potential for bromine gas
release,  as noted above.   Instead, the Agency first selected the option
of chemical treatment for  the detoxification of EDB's major constituents
and recovery of chemical feedstocks.  Process development (including
decanning of EDB  stocks) was pursued under an EPA contract in the  1985-
87 time frame.  Except for the decanning  activity, this effort proceeded
unsatisfactorily  and  the projected completion time and costs were  soon
deemed intolerable due to  unforeseen process equipment scale-up
problems.

       In  the fall of  1987, the EDB incineration option was revisited as
a direct  result of an unsolicited proposal EPA received from a major
commercial hazardous  waste incineration firm.  Proposed was the concept
that the  incineration of EDB along with adequate concentrations of
sulfur (as S,  S02,  etc.) in the  hot zone  of the combustion chamber would
encourage virtually complete chemical conversion of Br2 to hydrogen
bromide or HBr, which then should be scrubbed at high efficiency in the
incinerator's  emission control system.  The effectiveness of the sulfur
process,  it was assured, had been functionally demonstrated on EDB
materials in one  of the proposer's incinerators in the past as verified
by plume  opacity  observations, but no detailed stack gas verification
analyses  had been performed.  On the issue of how sulfur enters the
                                     14

-------
reactions to promote HBr, EPA found that published literature existed on
older laboratory studies which measured reaction rates, but these were
conducted at temperatures much lower than those in an incinerator's
environment.  The literature also offered no answer as to what chemical
reactions take place, and therafore offered no verification of the
proposer's bromine control solution.

      EPA elected to evaluate the sulfur/bromine conversion process at
field-scale by conducting a detailed test burn using test quantities of
EDB, co-fired with a sulfur source in the form of 10% dilute sulfuric
acid.  This test burn, conducted in December, 1987, was a complete
success and the results are described below.

      During the week of December 7, 1987, EPA conducted a detailed
field-scale test burn of ethylene dibromide (EDB) at a permitted
RCRA/TSCA commercial incineration facility owned by Rollins
Environmental Services (RES), Incorporated in Deer Park (Houston),
Texas.

      The three objectives for the test burn were:

      1.  To confirm the ability of the incinerator to achieve the
          required levels of destruction and removal efficiency (ORE)
          for the principal hazardous components of the EDB materials.

      2.  To verify the effectiveness of sulfur addition to the
          combustion chamber to force the formation of hydrogen bromide
          (HBr).

      3.  Assess the compatibility of EDB co-firing with normal waste
          disposal operations at the facility.

      The pesticide test burned at the Rollins site consisted of 20,000
gallons of an EDB/ethylene dichloride (EDC)/and carbon tetrachloride
(CCl^) mixture and 5,000 gallons of an EDB/chloropicrin formulation.
Scoping tests, which were brief initial test firings of the EDB
pesticides, were included as part of the test burn program to
immediately test the sulfur concept and to select pesticide flow rates
to be used in the more detailed test burn.  Scoping tests showed good
performance with various combinations of the two types of formulations,
but the test burn itself involved only non-chloropicrin material.   The
approximately 22,000 gallons of material remaining after completion of
the scoping and test burn program (including chloropicrin) was also
incinerated at Rollins during the several days following the test burn.

      The following is a summary of the test conditions used during the
EDB test burn (1):

      EDB Waste Stream Composition (by weight)

          EDB     10.8 percent
          EDC     44.5 percent
                                     15

-------
          CC1,
j     42.8 percent

 pesticide flow rate into incinerator -  49.7-50.7  Ib/min.
 (fed to rotary kiln)

 net waste flow to incinerator - nominally 300 Ib/min.
   including a sulfur stream as noted below,  PCB and RCRA
   solid and liquid waste streams
              kiln temperature
              afterburner temperature
              stack Oo level
              stack CD level
              stack C02 level
              stack NOX level*
              stack S02 level
              stack flow rates
                             1780 to 2000°F
                             2230 to 2250°F
                             10 percent
                             16-19 ppm
                             8-9 percent
                             47-63 ppm (primarily NO)
                             42-46 ppm
                             39,800 to 43,000 dscf
*During scoping tests, the chloropicrin material (38-40% EDB) was
 briefly fired at up to 40 Ib/min. and NOX increased to 70-90 ppm
 in the stack.

      During the EDB test burn, the incinerator achieved destruction and
removal efficiencies and satisfied other regulatory standards as follows
(as determined by VOST and M-5 methods):
      DRE:
 EDB
 EDC
 ecu
      Particulate Emissions:
      Bromine Level in Stack:
      Sulfur Feed:
in excess of 99.9999 percent
in excess of 99.99999 percent
99.999 to 99.99999 percent

0.0081 to 0.0123 grains per dscf
@ 7% C02

non-detectable (detection limit
4-5 micrograms
per dscf)

10-25 Ib/min. of a 10% sulfuric acid
solution (fired into the kiln next to
the EDB gun)
      During the  scoping runs, the  sulfur stream was intentionally
 stopped  several times  for  brief periods  (each stoppage lasted a few
 seconds  only).  A visible  brownish  plume resembling typical bromine
 fumes would issue forth from  the  stack whenever the sulfur stream was
 stopped.  These momentary  emissions, coupled with the reliable lack of
 visible  or detectable  bromine emissions with the sulfur present,
 demonstrated the  effectiveness of the reaction in which Br2 is converted
 to  HBr by sulfur  and then  scrubbed  in the air pollution control device.
 Mass balance calculations  to  account for bromine at all entrance and
 exit and transient mass accumulation points associated with the
                                     16

-------
 incinerator showed that essentially all  bromine was captured in the
 scrubber water streams.

 DINOSEB
       As  EPA considered disposal  options  for dinoseb,  incineration
 immediately emerged  as  a priority option  for this  RCRA P-020  type of
 waste.  Still,  EPA felt it  necessary to conduct  pilot-scale tests to
 evaluate  and quantify the ORE,  NOx and particulate generation and
 control issues.   A secondary option,  distillation  for  recovery of
 potentially marketable  chemicals  was  considered  but not pursued due  to
 the  uncertain marketability of  recovered  materials plus the questionably
 long facility design, construction,  and permitting time which would  be
 required.

       EPA conducted  two pilot-scale  incineration test  programs on
 dinoseb formulations during 1987,  one at  EPA's Research Triangle Park
 (RTP)  facility  in North Carolina  and  one  at  a contractor's facility  in
 Tulsa, OK operated by the John  Zink Company.  Sampling  and analyses  and
 reporting for both studies  were conducted by an  EPA contractor,  Acurex
 Incorporated.  The results  of these pilot-scale  studies are summarized
 below:

 Tests  at  RTP

      Objectives:

              to  determine  ORE, NOx,  and particulate emissions  for the
              "Dynamyte 5"  formulation of dinoseb,  one  of formulations
              with the  highest  concentration  of dinoseb

              to  obtain  data with  and without NOx  control via  special
              burning techniques

      The Dynamyte 5 pesticide  had the following characteristics:
              dinoseb
              diesel #2 oil
              xylene
              inerts
              heating value (HHV)
              nitrogen content
54.4%
 4.04%
32.5%
 9.06%
13,076 BTU/lb.
 6.63% by weight
      The tests at RTP utilized a low-NOx burner/package boiler
simulator device with a nominal heat release capacity of 3 million
BTU/hr.  Dynamyte 5 was fired as received at approximately 16.7 gallons
per hour in all tests.  The firing techniques consisted of:

          Dynamyte 5 as the only fuel, with no special  techniques to
          reduce NOx
                                     17

-------
         Dynamyte  5  as  the  only  fuel, with  air  staging  to  reduce  NOx

         Dynamyte  5  fired alone  in  the  primary  chamber, with  both air
         staging and natural  gas-fired  reburning  in  secondary chamber
         to  reduce NOx.
      The  results  showed the following  (2):

                         Temp,  (dea.  F)      Temp,  fdea.  F)
                                     NOx

conventional
firing
firing with
air staging
air staging
primary
chamber
1306
2450
2104
secondary
chamber
1122
1171
1205
ppm
(corrected
to 7% 02)
2998
85
88
      and reburning

      ORE of dinoseb
greater than 99.99% in all tests
      particulate emissions .024 to.045 gr. dscf (or 54 to 101 mg/dscm)
@ 7% 02 in all tests

Tests at John Zink

      Objectives:

              to incinerate a mixture of dinoseb blended from all
              inventories of different types of formulations
              proportioned approximately according to known volumes
              awaiting disposal

              to confirm ORE performance and quantify particulate
              emissions

              to quantify NOx emissions under typical RCRA and TSCA
              operating temperatures

              to experience the handling,  blending, and feeding
              characteristics of  injecting a blend of dinoseb
              pesticides into a typical incinerator burner nozzle

      The  dinoseb mixture used for  the Zink test had the following
 overall characteristics as fired:

              percents by weight:   82% Dyanap;  16% Dynamyte 3;
              and  1.6% Dynamyte 5
                                     18

-------
               net dinoseb and dinoseb salts:   21.2% by weight
               sodium ananap (Naptalam):   19.0% by weight
               sodium hydroxide:   5.1%
               water (primarily)  and other inerts:   40.6%
               xylene,  alcohols,  diesel oil, etc.:   13.5%
               overall  heating value (HHV):  4473  BTU/lb.
               nitrogen content:   1% approximately,  by weight
       The  testing results at  the Zink facility resulted in  the  following
 findings  (3):
               dinoseb  mixture firing rates:   4 to  47  gallons  per  hr.
               ORE of dinoseb  POHC:   in excess of 99.999%
               NOx emissions from natural  gas  alone:   92-150 ppm @ 7% 02
               NOx emissions when firing dinoseb:   112-836 ppm @ 7% 02 at
               1750F  from  low  to  high flow rates
               NOx emissions when firing dinoseb:   274-307 ppm @ 7% 09 at
               2200F  with  low  flow rate
               NOx emissions when firing dinoseb utilizing patented
               "Noxidizer"  system:   40 ppm
               particulate  emissions  before scrubber:   .014 to .305
               grains/dscf  @ 7% 02
               particulate  emissions  after scrubber:   .0025 to .0079
               grains/dscf  @ 7% 02
Status of pesticide  disposal
      Based upon  the field-scale  EDB and the pilot-scale dinoseb
incineration work described above, the Agency has been proceeding with
disposal of both  pesticides through  incineration.   The current status as
of early 1989  is  that almost all  of the EDB inventories have been
incinerated and the  entire task  should be completed in early spring.
Regarding dinoseb, EPA awarded incineration disposal contracts to two
commercial incineration firms, Chemical Waste  Management, Inc. and
Rollins Environmental Services,  Inc., at the end of 1988.  Dinoseb
incineration demonstration tests  are anticipated shortly at up to five
different sites.  The demonstration tests will involve firing
representative quantities of dinoseb pesticide waste coupled with
observations of routine performance parameters plus NOx measurements.
                                    19

-------
             The demonstration  burns,  if  successful,  are  to  be  followed  by routine
             incineration disposal.   EPA  estimates  that at least  12 months time will
             be required to  treat  all  of  the  dinoseb  inventories  for which the Agency
             has disposal responsibility.

                                            REFERENCES
              1.  White,  M.O.,  et  al.   RES(Tx)  EDB  Test  Burn  Program  Emission  Test
                 Results,  Final Reports,  Vol.  I  and  II,  Alliance  Project  No.  5-879-
                 999, Alliance Technologies  Corporation,  Bedford,  Massachusetts,
                 June,  1988.   Vol.  I  (199 pp)  and  Vol.  II  (296  pp).

              2.  Linak,  W.   Results to Date  of AEERL Dinoseb Tests,  unpublished.   EPA
                 memorandum from  William  Linak,  AEERL/CRB-Research Triangle Park,
                 North  Carolina to  E.  T.  Oppelt, RREL/OD-Cincinnati,  April 4,  1988.
                 12  pp.

              3.  Wool,  H.,  Villa, F.  and  Mason,  H.   Test Report for  the Trial  Burn of
                 Dinoseb in a  Pilot-Scale Incinerator.   Acurex  Corporation, Mountain
                 View,  California,  September 20, 1988 (EPA and  NTIS  report numbers to
                 be  assigned). 19  pp.
                                                  20
_

-------
   PYROLYTIC THERMAL DEGRADATION OF  A HAZARDOUS WASTE
                INCINERABILITY SURROGATE MIXTURE


    Presented at the 15th Annual EPA Research Symposium on Land Disposal,
        Remedial Action, Incineration, and Treatment of Hazardous Waste

                              Cincinnati, Ohio
                              April .1.0-12, 1989
                      D.A. Tirey, B. Dellinger, P.M. Taylor
                       Environmental Sciences Group
                    University of Dayton Research Institute
                             300 College Park
                         Dayton, Ohio 45469-0001


                                   and


                                 C.C. Lee
                    U.S.  Environmental Protection Agency
                    Risk Reduction Engineering Laboratory
                        26 W. Martin Luther King Dr.
                          Cincinnati, Ohio  45268


                               ABSTRACT

      This report presents preliminary data from a study  whose objective is to
develop a single mixture of organic compounds that can be consistently used to
demonstrate incinerator performance and compliance with  EPA regulations  The
proposed mixture consists of six compounds (sulfur hexafluoride,  trifluoro-
chloromethane,  chlorobenzene, pentachlorobenzene,  acetonitrile,  and tetra-
chloroethylene) which are expected to be very difficult to destroy under various
incineration failure conditions. Our initial studies have been conducted  in the
absence of oxygen, thus simulating a localized waste/oxygen mixing failure mode
of an incinerator. With the notable exception of sulfur hexafluoride,  agreement
between predicted and observed relative POHC stability was observed. Benzene
hydrogen cyanide, benzonitrile, and naphthalene have been identified as possible
surrogates for PIC formation.
                                 21

-------
                            INTRODUCTION

      Due to the public concern  over possible emissions of toxic organic
compounds,  there  has  been considerable  debate concerning  how to best
determine  the performance  of  a hazardous waste  incinerator.  The current
regulatory approach requires demonstration that a facility can destroy or remove
principal organic hazardous  constituents (POHCs) in the waste feed with a
destruction and removal efficiency  (ORE) of  99.99% (or 99.9999%  for
polychlorinated  biphenyls and  polychlorinated dibenzo-p-dioxins). Since it is
impractical to test the performance of an incinerator for every possible organic
component of the waste, considerable effort has been  expended to develop an
"incinerability11 index. Several such indices have been developed and applied to
predicting the relative incinerability  of toxic  organics  with varying degrees of
success.1

      Unfortunately there is only limited  data to confirm that any of these indices
are working successfully, although currently available comparisons indicate that a
thermal  stability-based  approach appears to have considerable promise.1'2
Furthermore,  these indices  have  not  always been  uniformly applied  in  the
permitting process,  at times causing delays in  the facility being certified for
operation. In addition, any time there is a special type of waste being burned, there
is concern that some toxic component (which  may not be on EPA's Appendix VIII
list)  is more difficult to incinerate than the POHCs selected for ORE performance
testing during the trial burn.

      As a result,  there has  been  considerable interest  in  developing a
standardized ORE  surrogate mixture  that  could be used to determine  the
performance of any incinerator.3 Each of the components of this mixture would test
the capabilities of the incinerator by testing the incinerator's failure modes. A failure
mode of an incinerator may be defined as an  operating  condition where a fraction
of the material experiences a departure from the spatially and temporally averaged
optimum operating conditions that results in an increase in toxic  emissions. In
general, modes of failure can be thermal (e.g., quenching, temperature gradients,
heat transfer limitations), temporal  (e.g.,  uncertainties  in  residence  time
distributions due to plug flow assumptions), or fuel-oxidant mixing (e.g., poor micro-
 mixing resulting in "pyrolysis pockets") related. The fuel-oxidant mixing failure has
 been suggested as the most important, although all failure modes likely contribute
to increased emissions of undestroyed POHCs as well  as products of incomplete
 combustion (PICs),2-5

       One type of temporal or residence time failure is  the possibility of slow
 vaporization of condensed phase materials. However, it has been shown that this
 phenomenon would have minimal impact on the relative incinerability of POHCs.6
 What would be more crucial in this case are the thermal failure modes that  may
 exist by virtue of quenching of hot gases in the post flame zone by cool secondary
 air  It is conceivable that sufficiently low temperatures may be  reached such that
 oxidation reactions  involving  POHCs are slow enough to result in  measurable
 emissions.  This has  never  been conclusively demonstrated  in  full-scale
                                   22

-------
      Acetonitrile was selected as an example of a non-chlorinated compound that
has been shown to be stable under both oxidative and pyrolytic conditions.2'7 This
compound  was also chosen as a model compound of the generally very stable
group of nitrile-containing  compounds.2 Chlorobenzene was selected as an
example of a chlorinated compound that has been shown to be very stable under
pyrolytic conditions,2 thus representing a mixing failure test of an incinerator.
Although there was some concern over PIC formation distorting its apparent ORE,
this compound was  included in the mixture at a feed concentration thought to
minimize this possibility. Pentachlorobenzene is a good example of a chlorinated
compound which should exhibit high thermal stability under oxidative conditions.7
It was thus included in the surrogate mixture to test the thermal failure mode of an
incinerator. Tetrachloroethylene was selected as  an example of a  perchlorinated
compound which is stable under pyrolytic conditions and moderately stable under
oxidative conditions.2-7 Trifluorochloromethane  and sulfur hexafluoride  were
included to exclusively test the thermal quenching failure mode of an incinerator.
These compounds should represent the most stable components of the mixture
with destruction occurring above 1000°C.8'9 Toluene was included as a suitable
solvent material of moderate thermal stability for this complex organic mixture.
                      EXPERIMENTAL  APPROACH

      Experiments were conducted using the Thermal Decomposition Analytical
System (TDAS). The TDAS is a closed, in-line thermal instrumentation system
capable of accepting a solid, liquid, or gas-phase sample, exposing this sample to
a highly controlled thermal environment and then performing an analysis of the
effluents  resulting from  this exposure.  Its design and operation have  been
described in previous reports and publications.10'11

      Due to the low vapor pressure  of  pentachlorobenzene  under  ambient
conditions, a liquid  phase  organic mixture  was  considered  the  most viable
approach  to produce a relatively stable, homogeneous sample within the target
concentrations for each component. The sample was thus prepared by dissolving a
specific quantity of solid into the liquid phase and purging the liquid with the two
gases for 30  minutes. Samples  were  ultimately prepared with  little or  no
headspace, a prerequisite for keeping  the integrity of the gases dissolved in the
liquid intact.  Experiments  performed before beginning data acquisition  revealed
that samples prepared in this manner  and stored at room temperature could be
used  as long as  48 hours without  appreciable losses of sulfur hexafluoride or
trifluorochloromethane detected. High purity  (> 98%) samples of each component
were used without additional purification in preparing the mixture.

      The gas-phase mixture concentration in the tubular fused silica flow reactor
was maintained at a concentration of ~1 x 10'3 moles/liter using a manual  liquid
injection technique. For each experiment, a volume of 0.2 \i\ was injected into the
system at  a rate of ~0.05 (xl/s using a small sub-microliter syringe. All components
                                  23

-------
Incinerators, but it seems wise to address this possibility in the development of a
surrogate mixture.

      Sensitivity analyses indicate that relative POHC thermal stability is most
sensitive to oxygen concentration.2 This leads us to believe that an appropriate
surrogate  mixture should include components of high thermal stability for both
oxidative and pyrolytic conditions. Although generally less sensitive, the relative
thermal stability of POHCs also varies as a function of temperature. Thus, the
surrogate mixture should also contain components of high relative thermal stability
when based on both low temperature (~700°C) and high temperature (~100u°C)
destruction efficiency measurements. From a kinetic viewpoint, this  means we
should select mixture components whose primary mechanism of decomposition
covers all known reaction classes and which are examples of slow reactions within
each class.

      Table 1 presents the components of a low chlorine content surrogate mixture
that were  selected based upon the  above criteria and which were tested in this
study.3

                                  Table  1
               Proposed  Incinerability Surrogate Mixture
Component
Acetonitrile
Chlorobenzene
T99 (2)a
1000b
990b
Destruction
Mechanism
C-H bond rup.,
H abstraction
C-CI bond rup.,
Conc.d
(wt. %)
11.4
19.3
Failure Mode
Represented
Mixing and
Thermal
Quenching
Mixing
Pentachlorobenzene        940b


Sulfur Hexafluoride         1090°

Tetrachloroethylene         895b


Trifluorochloromethane     1010°

Toluene (solvent)           890b
                                  Cl displacement
                                  G-CI bond rup.,
                                  Cl displacement

                                  S-F bond rup.

                                  Cl displacement


                                  C-CI bond rup.
 3.5


 0.1

13.2


 1.0

51.5
Mixing and
Thermal
Quenching
Thermal
Quenching
Mixing and
Thermal
Quenching
Thermal
Quenching
Mixing
                                   C-H bond rup.,
                                   H abstraction
 Footnotes:
      a.  For mixtures of compounds with elemental composition of
          CaHsCli and waste/oxygen equivalence ratio of 3.0.
      b.  Based on experimental data.
      c.  Estimated using unimolecular reaction rate theory.
      d.  Relative concentrations based on waste mixture with an elemental
          composition of C3H3 2Cl0.3N0i15F0 015S0.oo3-
                                   24

-------
of the sample were simultaneously vaporized in the insertion  region which was
held isothermally at 275°C for all experiments.

       Precise two-second thermal exposures were conducted in helium carrier
(99.99+% purity) over a temperature range of 300-1000°C. Duplicate experiments
were conducted along the temperature profiles to determine data reproducibility.
Blank runs were conducted daily to determine detection limits.

      The effluents resulting from thermal exposure were cryogenically focussed
at the head of a 30  m x 0.25 mm ID fused silica capillary column (0.25|i DB-5
stationary phase) held at  -60°C. The GC oven temperature was  programmed to
290°C at 10°C/min. Helium was used as the GC carrier gas and a mass selective
detector in the full scan mode (35 to 400 m/e) was used for effluent quantitation.
The lightest of these effluents (35 ^ m/e <, 132) were not amenable to this technique
and were  analyzed using "on the fly" mass spectrometry. Mass  spectral response
curves were generated for each POHC. Calibration of mass spectral response for
the  myriad  of PICs  generated  has  not yet been  completed.  Tentative PIC
identifications and their semi-quantitative formation and destruction curves were
determined by mass spectral data  analysis.
                                RESULTS

      The individual thermal decomposition profiles of the six POHCs are
presented in Figure 1. The thermal stability index (Tgg (2)) for each of the POHCs
and its theoretically predicted versus observed  relative stability are displayed in
Table 2. Except for the unexpected low stability of sulfur hexafluoride, the predicted
and observed relative stabilities are in agreement. The differences in absolute
thermal stabilities in Tables 1 and 2 reflect the differences in test conditions, i.e., 0 •
3.0 versus oxygen-free and H/CI ratios of 3:1  versus 10:1, respectively.

      As  illustrated in Figure  1, for greater  than 99% destruction, the relative
stability of tetrachloroethylene and pentachlorobenzene reversed.  This  rare
thermal decomposition behavior was not observed for the other components in the
mixture. Despite the fact that the decomposition  profiles of these two compounds
crossed at approximately 875°C,  their thermal decomposition behavior was the
most reproducible (±5%) of all components in the mixture. The thermal behavior of
the light gases was less reproducible (±20%), due to the lack of  controlled,
cryogenically trapping during the data acquisition procedure. Similarly,  acetonitrile
was somewhat difficult to quantitate with the GC  column employed because of the
polarity of this compound. The extreme tailing behavior of this compound made
accurate quantitation of the peak area difficult. Because of this, the detection limit
for this compound was necessarily high and  a  fairly high level  of uncertainty (±10-
20%) in data was observed.
                                   25

-------
   10
                 Sulfur Hexafluoride
                 Trlfluorochloromethane
                 Acetonitrile
                 Tetrachloroethylene
                 Chlorobenzene
                 Pentachlorobenzene
      500
600          700          800
          TEMPERATURE  (°C)
                                                         900
                            1000
Figure 1. Thermal decomposition behavior of proposed surrogate mixture in
         flowing nitrogen. Mean, gas-phase residence time = 2.0 s.
                                TABLE 2
        Experimental  versus  Predicted POHC Thermal  Stability
Compound
     Predicted
Experimental
Exp. T99 (2)
Sulfur Hexafluoride
Trifluorochloromethane
Chlorobenzene
Acetonitrile
Pentachlorobenzene
Tetrachloroethylene
1
2
3
4
5
6
3
1
2
4
5
6
-920
1010
975
-910
870
860
      Table 3 lists  the  tentative PIC  identifications and their relative semi-
quantitative concentrations for temperatures greater than 700°C. Inspection of this
table shows that numerous complex PICs were produced from the mixture of seven
simple starting materials. Several general observations were evident. First, the PIC
with the greatest frequency of occurrence  and concentration was benzene.
Second, the trend in  product  formation  was toward unsaturated, conjugated
compounds which were thermodynamically favored because they are resonance
stabilized, e.g., polynuclear aromatic hydrocarbons, (PNAs). Almost exclusively,
these products were the only PICs detected at high reactor temperatures other than
stable lighter gases such as hydrogen chloride and methane. This trend toward
unsaturated, conjugated  products with increasing reactor temperature was also
one observed  in previous studies. The relative  paucity of halogenated compounds,
e.g., benzyl chloride, trichlorobenzene, tetrachlorobenzene was consistent with the
large H/CI ratio of this mixture. 100+20% of  the chlorine was accounted for as
                                  26

-------
hydrogen chloride at 1000°C. The remainder of the observed PICs can be
generally characterized as substituted benzenes. The substituent groups included
methyl, ethyl, ethenyl, ethynyl, propynyl, phenyl, chlorophenyl, and nitrite.

                                  TABLE 3
                       Tentative PIC Identifications
             and Relative Semi-Quantitative Concentrations
Tentative
Identification 700
Hydrogen Chloride XX
Methane
Hydrogen Cyanide
Benzenepropanenitrile
Benzene XX
Ethylbenzene
Ethenylbenzene
Ethynylbenzene
Isocyanobenzene or Benzonitrile
Propynylbenzene
Methylbenzonitrile
Benzyl Chloride X
Trichlorobenzene
Naphthalene or Azulene
Quinoline or Isoquinoline
Tetrachlorobenzene
1-Pheny|naphthalene X
1,1'-Biphenyl or
Ethenylnaphthalene
Methyl-1,1'-biphenyl isomers
1 , 1 '- ( 1 , 2-Ethanediyl) bis-benzene
Acenaphthalene br Biphenylene
Naphthalenecarbonitrile
Chlorobiphenyl
9H-Fluorene
Chloropentafluorobenzene
Phenanthrene or
9-Methylene-9H-fluorene
2-Phenylnaphthalene
1 -Phenyimethylene-1 H-lndene
Pyrene or Fluoranthene
Ci3HsCl4 isomers
1 1 H-Benzofluorene (a or b isomers)
Triphenylene or Chrysene or Naphthacene
or Benzophenanthrene
Benzofluoranthene or Benzopyrene
or Benzacephenanthrylene
750
XX


XX
XXX
XX
XX


XX

X

XX
X

XX
XX

XX
XX

X



XX

XX

XX
XX





Temperature (*C)
800 850 900
XX
X

XX
XXX

XX

XX
XX



XXX
XX
XX

XX

XX

XX
XX

XX
X
XXX

XXX
XX
XX
XX
XX
XX



XX
XX
X
XX
XXX

XX

XX
XX
XX


XXX
X
XX

XXX

XX

XX
XX
X
XXX
X
XXX

XXX
XX
XXX

XX
XX

XX

XX
NA
NA

xxxx


XX
XXX
XX


XX
XXX

X

-XXX



XX
XX
X
XX

XXX

XX

XXX


XX

XX

950
XXX
NA
NA

xxxx


XX
XXX
XX



XXX



XXX



XX
XX
XX
XX

XXX

XX

XXX


XX

XX

1000
XXX
NA
NA

xxxx


XX
XXX




XXX



XXX



XX
X


,
XXX



XXX


XX

XX

Legend:
          NA: GC/MS data was not obtained at these temperatures.
          Concentration based on GC/MSD integrated response:
             0.1% <. X < 0.25% of total mass injected.
             0.25% <; XX < 2.5% of total mass injected.
             2.5% ^ XXX < 25.0% of total mass injected.
             XXXX ^ 25.0% of total mass injected.
                                     27

-------
                                           DISCUSSION

                   Perhaps the most surprising result of this study was the apparent instability
             of sulfur hexafluoride, being undetectable at a temperature of 900°C. This result
             contradicts previous studies where this compound was not destroyed (at the 99%
             level) below  temperatures of ~1100°C.8'9-12  Experiments were performed to
             determine whether its apparent instability was possibly due to reactions in the ion
             source of the MSD. Pure  sulfur hexafluoride was  exposed to 300°C (a non-
             degradative temperature) and 900°C (the temperature at which sulfur hexafluoride
             was below detection limits in  the mixture  data)  and  no  decomposition  was
             observed. Sulfur hexafluoride in the presence of helium was similarly tested  with
             no observed reaction. The experiments were repeated with successive additions of
             carbon dioxide, hydrogen chloride, water  vapor,  and  methane without  any
             observed decomposition.

                   The only fluorine-containing product detected was a very small peak
             tentatively identified as pentafluorochlorobenzene. Two other products observed
             as PICs in a  previous study of sulfur hexafluoride mixture decomposition,8 sulfur
             tetrafluoride and hydrogen fluoride, were not detected. However, the concentration
             of these compounds may have  been below analytical detection limits due of the
             low sulfur hexafluoride concentration in the feed (see Table 1). Because of the  lack
             of mass balances for fluorine, we are continuing to assess this anomalous result.
             We plan to study simple binary sulfur hexafluoride/organic mixtures to determine if
             bimolecular thermal decomposition reactions in the flow reactor can account for the
             observed  profile. We also plan to verify our  results  using different  detectors
             including an electron capture and flame photometric detector.

                   The relative stability of  the remaining POHCs  accurately reflect their
             predicted  mechanism  of destruction. Trifluorochloromethane  decomposes by
             unimolecular C-CI bond rupture, which is a relatively  slow reaction due to  its  high
             bond dissociation energy (A = 3.16 x 1014 s'1, Ea = 85 kcal/mole).8 Under pyrolytic
             conditions, chlorobenzene likely decomposes by Cl atom displacement by H,13 the
             latter being the dominant reactive species in the system as illustrated in Table 4.
             The observation that pentachlorobenzene was more  fragile is due, in part, to the
             larger number of chlorines available for displacement.  The rate of the displacement
             reaction at 1085 K has been measured as ~4.0 x 108 liter/mole-s per Cl atom.13

                                              Table 4
                           Reactive Species Equilibrium  Mole  Fraction
                                    at  Elevated  Temperatures14
Specie
Cl atoms
H atoms
700°C
6.99E-1 1
3.59E-10
800°C
8.68E-10
4.65E-09
900°C
7.04E-09
3.91 E-08
1000°C
4.13E-08
2.37E-07
                   All bonds  are quite strong in acetonitrile and the most likely pathways of
             destruction were methyl radical displacement  by H  atoms  forming  hydrogen
             cyanide and H abstraction by methyl radicals yielding methane.15 The stability
             reported here  compares favorably  with that reported in the literature when one
                                               28
-

-------
accounts for the initiation of radical  chain  reactions  by other more  fragife
compounds which generate reactive H and Cl atoms.15 Tetrachloroethylene has
recently been studied as a pure  compound in this laboratory.16 The most likely
mechanism  of destruction in this mixture  was Cl displacement by H  atoms,
although trichloroethylene was not observed as a PIC. The lower C-CI bond (85
kcal/mole versus 95 kcal/mole for chlorobenzene) facilitates Cl displacement,
resulting in its being the most fragile compound in the mixture.  There was also
evidence that toluene reformed at elevated temperatures.  Initial decomposition by
C-H bond rupture of the side chain results in the formation of stable benzyl radicals.
Their recombination with high concentrations of H atoms  may be responsible for
the observed reformation.

      Figure 2 presents a plot of integrated response for all detectable organic
PICs  and  POHCs as a function  of temperature. This graph indicates that PIC
emissions are greater than POHC emissions at temperatures above 800°C. It also
displays how  the total PIC response  correlates  with benzene formation and
destruction, suggesting that benzene may be an appropriate surrogate for organic
PIC emissions in full-scale incinerators.
 HI
      500
600          700         800
          TEMPERATURE  (°C)
900
1000
 Figure 2. Comparison of POHC decay curve versus semi-quantitative PIC (and
         benzene) formation and destruction curves.

       At 700°C, the major organic PICs are benzene and benzenepropanitrile.
 The former was probably produced by electrophilic displacement of the toluene
 methyl substituent  by  H atoms,  the latter  by recombination of  benzyl  and
 ethanenitrile radicals. Each of these two species was probably formed by C-H bond
 homolysis  in the initial decomposition of the parent molecules. The ethanenitrile
                                  29

-------
species was apparently also present at higher temperatures as evidenced by the
detection of naphthalenecarbonitrile. At temperatures above 700°C, the presence
of  nitrile  radicals was  evidenced  by  the  formation  of  benzonitrile and
methylbenzonitrile. Benzyl chloride  was  a relatively  low-yield  recombination
product as a result of the low chlorine concentration in this mixture.

      As the temperature was raised to 750°C, more substituted benzenes began
to dominate. One potential route to the formation of ethyl-, ethenyl-, and propynyl-
benzene likely involved displacement of the methyl substituent from toluene. It is
Interesting to note the formation of ethynylbenzene at temperatures greater than
900°C. This was likely  due to the higher temperature required for ethynyl radical
formation from the decomposition of a suitable aromatic precursor. At temperatures
of 800-900°G, aromatic substitution products, e.g.,  biphenyl and chlorobiphenyl,
were produced from  toluene  displacement  reactions involving phenyl and
chlorophenyl radical (produced from toluene  and chlorobenzene, respectively).
Tetrachlorobenzene  was formed  as a dechlorinated product  of penta-
chlorobenzene at  800-850°C  and  this  molecule, in turn, dechlorinated  to
trichlorobenzene at higher temperatures.

      The  remaining products at the highest temperatures were benzene and
numerous PNAs. A mechanism suggested in the literature for the formation of
PNAs involves vinyl radical chain displacement reactions followed by cyclization.17
At higher temperatures, increasingly unsaturated PNAs  were formed, which may
ultimately lead to soot formation. Besides benzene, the most notable PIC was toxic
hydrogen cyanide, produced from methyl radical displacement by H atoms. This
compound appeared at  850°C and persisted at higher temperatures.


                               SUMMARY


      We plan to  continue our study of this  mixture under other failure  mode
conditions.  It is encouraging that the predicted and observed POHC stability  are in
agreement under pyrolytic conditions, with  the notable  exception of sulfur-
hexafluoride. Benzene, hydrogen cyanide, benzonitrile, and naphthalene have
been identified as  appropriate PIC surrogates.  However, the relative  absence of
chlorinated PICs suggests that the  mixture elemental composition should  be
modified to increase chlorinated PIC formation. These compounds are desirable to
provide surrogates for  testing incinerators  burning highly chlorinated hazardous
waste.


                         ACKNOWLEDGMENTS

      We gratefully acknowledge  M.  Tissandier for assisting in data acquisition
and J. Kasner for performing the equilibrium calculations. This research was
partially supported by the US-EPA under cooperative agreement CR-813938.
                                  30

-------
                             REFERENCES
1.   Dellinger, B., Graham, M. and Tirey, D. Hazard. Waste Hazard. Mater., 3,
    293 (1986).
2.   Taylor, P. H., Dellinger, B. and Lee, C.C. Environ. Sci. Technol., in review,
    1989.
3.   Dellinger, B., Taylor, P.H. and Lee, C.C. "Development of Hazardous Waste
    Incinerability Surrogate Mixtures," Presented at the 2nd Annual National
    Symposium on Incineration of Industrial Wastes, San Diego, CA, 1988.
4.   Tsang, W. "Fundamental Aspects of Key Issues in Hazardous Waste
    Incineration," ASME Publication 86-WA/HT-27, 1986.
5.   Lee, K.C.JAPCA.SS, 1542(1988).
6.   Chang, D., et al. "Relationships Between Laboratory and Pilot-Scale
    Combustion of Some Chlorinated Hydrocarbons," Proceedings of AlChE
    Summer National Meeting, Denver, CO, 1988.
7.   Dellinger, B., Torres, J., Rubey, W.A., Hall, D.L., Graham, J.L., and Carnes, R.A.
    Hazard.  Waste Hazard. Mater., 1, 137 (1984)
8.   Taylor, P.H. and Chadbourne, J.F. JAPCA, 2Z, 729 (1987).
9.   Tsang, W. and Shaub,  W.  in Proceedings  of the 2nd Conference on
    Management of  Municipal,  Hazardous and  Coal  Wastes, p.241,  1983.
1 0. Rubey, W.A. "Design Considerations for a Thermal Decomposition Analytical
    System," EPA-600/2-80-098,  U.S. Environmental Protection Agency,
    Cincinnati, OH, 1980.
11. Rubey, W.A. and Carnes, R.A. Rev.  Sci. Instrum. 56. 1795 (1985).
12. England, W., et al. in Proceedings of the  79th  APCA Annual Meeting,
    paper 86-61.1, 1986.
13. Tsang, W. "High-Temperature Chemical and Thermal Stability of Chlorinated
    Benzenes," Presented at the International Flame Research Committee
    Symposium on the Incineration of Hazardous , Municipal, and Other Wastes,
    Palm Springs, CA, 1987.
14. Reynolds, W.C. "STANJAN Equilibrium Program, Version 3.0," Department of
    Mechanical Engineering, Stanford  University, Stanford, CA, 1986.
15. Lifshitz, A., Moran, A., and Bidani,  S. Int. J. Chem. Kin., 1£,  61  (1987).
16. Taylor, P.H. and Dejlinger, B. "Development of a Thermal Stability Based Index
    of Hazardous Waste Incinerability," Fiscal Year 1988 Report prepared for EPA
    Cooperative Agreement CR-813938,  C.C. Lee, Project Officer, December
    1988.
17. Cole, J.A., et al. Combust.  Flame, 56, 51  (1984).
                                  31

-------
             INCINERABILITY  RANKING  OF  HAZARDOUS  ORGANIC  COMPOUNDS

                by: Robert E. Mournighan
                    Marta K. Richards
                    Howard 0. Wall

                    Technology Research Section
                    Thermal Destruction Branch
                    Waste Minimization Destruction and
                      Disposal Research Division
                    U.S. Evironmental Protection Agency
                    Cincinnati, Ohio 45268
                                   ABSTRACT

      Since EPA became involved with  the regulation and permitting of hazardous
waste incinerators, developing a reliable measure of incinerator performance has
been one of its goals.  The use of a thermal stability index to rank Principal
Organic  Hazardous  Constituents   (POHCs)  has  facilitated  the  evaluation  of
incinerators.  The development of data to support the  thermal stability rankings
has been an ongoing effort.

      This paper discusses  the results of a parametric study in which temperature
and oxygen  concentrations  were evaluated relative to  destruction  and removal
efficiency  (ORE)  of the test  mixture.   The  compounds used in  the test mixture
were: toluene, chlorobenzene,  tetrachloroethylene, tetrachlorobenzene and sulfur
hexafluoride.
                                      32

-------
                                 INTRODUCTION

      Since the regulation of hazardous waste by  the USEPA, both the Agency and
the regulated community have been  searching for means of evaluating incinerator
performance.   The concept of  a  "trial burn," a series  of tests  designed to
evaluate the ability of an  incinerator to  process  waste  to certain standards,
was developed; and surrogates were introduced (1) for hazardous and solid waste
and for Principal  Organic Hazardous Constituents, or POHCs.   In  essence,  the
incinerator's ability to destroy waste was based on its performance in destroying
the POHCs.

      Destruction  and  Removal  Efficiency   (DRE)  is  used  by  EPA  to  express
incinerator performance.  The DRE of a POHC is calculated as follows:
      DRE = (Wln_-Wput) x 100
              	  u-
                       win
(2)
            Where  in and  out are the mass flow rates of the POHC
            input and output (at the stack) respectively.

      As regulations  and  policies regarding trial  burns  developed,  it became
apparent to all concerned that DRE determinations were rapidly becoming extremely
expensive  and time  consuming.    It  is not  unusual to  see trial  burn  costs
approaching 2% of the capital  costs  of the facility.  Largely responsible for
these  costs was the requirement  that  several types  of POHCs  be evaluated,
resulting  in  the use of multiple  sampling trains.  The  process of obtaining
representative samples  could take a  week, or longer.   Sample  processing and
analysis,  also expensive, would  drag out  the time-span between  trial burn and
results to over three months.  The objective  for regulators  and regulated alike
was to search for a  less expensive approach.  The approach was similar to that
taken for  POHCs  and  solid waste:   find a  surrogate for the waste being burned
and develop a cheap, quick analytical method  for the analysis of  the surrogate.
Several compounds, e.g., Sulfur hexafluoride  (SFg),  Freons^, and  fluorocarbons,
were put forth as candidates.(3)

      Because of its  high thermal  and chemical  stability and the fact that it
is inexpensive and non-toxic, SFg received  much attention as a surrogate.  Quick,
relatively  inexpensive, and  reliable methods were developed for its analysis,
and initial evaluation began  in 1984.(4) A system employing a gas  chromatograph,
equipped with  an electron capture detector  (6C/ECD),  was  used  to obtain SFg
concentrations in stack gas  every  2 to 4  minutes, a much shorter time than the
sampling time required  for  sampling  trains.   Equipment set-up,  operation, and
calculation of results  could be done in  1-1/2 days  (4),  a  far  cry from three
months.

      Since that time,  a  great deal  of effort, time and money has been spent,
both in the United  States  and Canada  (5),  to exhaustively evaluate SFg as a POHC
or waste surrogate.

      During the same time period, the concept  of  using a standard  POHC mixture
(a "POHC Soup") for trial  burns, by employing a group of  compounds  with a range
of chemical and physical characteristics,  was developed.  A paper  describing the
                                      33

-------
rationale and results of that effort is being given at this conference.(6)

      Both of these efforts have been researched extensively at the laboratory
stage (5, 6, 7).  An evaluation  program  at pilot  scale for  the USEPA Combustion
Research Facility  (CRF) was proposed in 1987.  The CRF has two pilot-scale
(3 x  10° Btu/hr)  incinerators,  a  liquid  injection system and  a  rotary kiln.
The SFg/"POHC Soup" program was  performed  in the  liquid injection system.  This
paper describes the ORE results and discusses the implication of this work and
of other researchers.

      Interest  in  assuring  and  demonstrating  compliance with  RCRA  permit
conditions has  also  concerned the  hazardous waste community.   Trial burns are
prohibitively expensive to  carry out on  a  routine or even annual  basis, with
not much gained in the process.   This  approach still  leaves "gaps in coverage,"
where incinerator performance may be unknown.  It became apparent, after initial
research at EPA,  that  residence time, temperature and  turbulence  were not the
only factors to consider in the evolution of incineration  criteria.  The Toxic
Substance  Control   regulations   (8)  for  PCB   incineration   reflected  that
realization, and  stipulated  not only  time and  temperature, but also a minimum
stack Op concentration as well as a minimum combustion efficiency of 99.9% are
required.

     This paper concentrates  on  the  evaluations  of SF6  and the "POHC Soup" as
trial burn surrogates.

                            EXPERIMENTAL  BACKGROUND

     The experimental  program was executed at  the  USEPA  Combustion Research
Facility's Liquid  Injection  System (LIS).  Figure 1 is  a  sketch  of the unit,
showing each of the elements that make up  the incinerator and the air pollution
control system.

      The LIS  is  fired  with liquid waste, and  propane is used  to maintain
temperature control.  Combustion air  is supplied by forced-draft fan.  The waste
mixture containing the "POHC  Soup" was  stored  in a stainless  steel  vessel and
pumped continuously to the LIS.(9)  SFg was injected  into  the liquid feed stream
as a gas, where it dissolved into the liquid phase just  prior to incineration.

      Sampling of the stack gas  for both SFg and  the  POHCs  was done at the exit
of the air pollution control systems,  just  after the ionizing wet scrubber (IWS).
Even though  the gas passes  through  additional  gas  cleanup,  sampling further
downstream was not conducted.  The final gas cleanup  is unique to this facility
and is not standard for commercial  and private incinerators.  Table 1 shows the
experimental conditions and the  ORE results  for the SFg and  the POHC components.

      Oxygen and  incinerator temperature were  varied as  part of  a designed
experiment.   All other  variables, such as waste composition  and flow rates, were
kept constant.

     The compounds used  in  the  test  mixture were toluene  (70%), chlorobenzene
(10%), pentachlorobenzene (10%)  and tetrachloroethylene (10%).  This combination
of compounds resulted in a mixture  which contained both volatile components and
                                      34

-------
                                           •4-J
                                            I/)
                                            U
                                            0)
                                            •o
                                            •r-
                                            3
                                             O)


                                             O)
35

-------
 semivolatiles.   Stack  sampling  of these  components  was  accomplished with the
 VOST (USEPA Method 0030) and Modified Method  5 (USEPA Method 0010).   The  SF6
 sampling method was proportional  gas sampling, with the sample being fed directly
 to 6C/ECD.

                Table 1.  Data Table - Temperature, Oxygen and
                                  DREs  (number  of nines)
          Afterburner

             Exit
           Tempera-  Exit
Experiment ture, °C Oxygen %
                      ORE
     Tetra-
     chloro-            Chloro-   Pentachloro-
SFg* ethylene* Toluene* benzene*    benzene*
1
2
3
4
5
6
7
8
9
10
*Calcu

1030
1114
943
1274
1091
945
1310
1077
1175
1105
lated ORE by

2.3
3.3
5.6
1.3
5.1
8.2
4.5
9.2
8.0
4.7
the formu

4.35
3.59
3.33
7.00
5.31
4.17
5.33
3.44
5.70
3.96
la ORE

5.40
5.52
6.14
5.51
5.74
6.66
5.34
5.85
6.70
5.16
= -Log

6.19
6.21
7.10
6.11
5.85
6.68
5.59
5.74
7.55
6.39
1-DRE
100
6.11
6.04
6.38
6.40
4.14
6.66
5.80
6.28
6.54
5.68


7.80
7.52
7.34 '
7.30
6.98
7.41 •
7.49
7.35
7.43
7.44


                          ANALYSIS OF RESULTS - SFg

      To gain as much information about the behavior of the POHC/SFg DREs with
respect to the two variables, a regression  analysis  was performed for each data
set,  one  for  each  compound.   Table 2  lists the  results  of  the  regression
analysis, the model used and the correlation coefficient.

      Figure 2  is  a contour plot with  independent  variables,  temperature and
afterburner exit oxygen concentration, on  the Y  and X  axis  respectively.   The
SFg ORE  (expressed in the  number  of nines)  is  displayed as contours  of the
regression surface.  This figure demonstrates why  it is  so  hard to  make solid
judgments about  the relationship between  ORE and  a single variable.   As the
figure clearly  illustrates, whether  ORE  goes up  or down with  increasing  Op
depends on the operating temperature.  At 950°C,  ORE increases from about
3-nines to  nearly  4-nines  as  oxygen increases.    At 1350°C,  the same  change
results in the opposite effect, dropping the  ORE  from  greater than  7-nines  to
less than 5-nines ORE as oxygen increases.

      At this point, it is stressed that the regression analysis and the
resulting figures developed from it should only be used in general,  not
                                      36

-------
      SF6 ORE  VS T,  O2
  1400
T

^1300H
P
E
R 1200
A
T
U 1100-
R
E
  1000
  900
     01
               x
                                +
i	1	1	1	1	r
   23456
                 i	r
              7   8   9  10
O2 CONCENTRATION
 3-9's

 6-9's
       ORE

     +  4-9's

     x  7-9's
                         *•  5-9's
             Figure 2. SF6 ORE vs. T, 02.
                 37

-------
predicting absolute ORE figures for specific compounds.  Until additional data
are developed, these relationships should only be applied to the CRF incinerator
and can only  be used  to describe  the  relationship between dependent (ORE) and
independent  (Temp.-  and" Op)  variables  within  the  range of the' experiment.
Extrapolation of the results outside the range is risky.
                  Compound
                             Table 2.   Data  Table
Model
                  SFg                 1st Order
                  Toluene             1st Order
                  Tetrachloroethylene 1st Order
                  Chlorobenzene       2nd Order
                  Pentachlorobenzene  2nd Order
Correlation
Coefficient
                    0.77
                    0.45
                    0.75
                    0.76
                    0.68
      In the  laboratory  work  done on the thermal decomposition  of  SFg at the
University of Dayton Research Institute (UDRI) (7),  it was determined that SFg
destruction was independent of  oxygen  concentration,  indicating  the mechanism
to be unimolecular decomposition.  Although  Figure  2  shows  an effect of 02 on
SFg ORE, it may not mean that Q£ specifically causes the changes.  What is not
depicted is  that  as the 02 concentration increases,  air to  fuel  ratios and
incinerator residence times change concurrently.  The variation in SF6 ORE may
be related to those effects and is  not necessarily  inconsistent  with the UDRI
results.

     Figure 3  is  a plot of the SFg  DREs  versus temperature  with 3 data sets
illustrated.    The  data  shown  as  diamonds  were  the  SFg  DRE-temperature
relationship calculated from the data supplied  in References  o and 7.  This was
calculated for  a  2-second residence time and  represented  the DRE temperature
relationship for SFg exclusive  of any other physical or chemical  effects.

     The data plotted  as Xs 'were data from  the  work  conducted at the Alberta
Environmental Centre (5)  in which  the  effect of the presence of refractory on
SFg DRE was evaluated.  As one can see,  the presence of refractory  in the reactor
had a marked effect on the DRE-temperature relationship, reducing the required
temperature by some 200°C.

      The  third data set plotted in Figure 3 was the  CRF  data,  and shows the
regression line for those data  points.  Here, as with the Canadian data, there
is a  marked  difference  in  SFg  DRE-temperature  behavior.   While some  of the
difference could be attributed to the presence  of refractory  and residence time
changes in  the incinerator,  the  authors feel that there is  more to  it than that.
Since the laboratory  data (UDRI) were taken with non-flame  systems, in which the
substances underwent an extremely narrow temperature distribution,  SFg DRE should
not be  expected to' be  similar,  since  in flame  conditions,  SFg would "see" or
experience a wide range of temperatures in the  incinerator environment.  At any
rate, the behavior is not the same, and more investigation is. warranted.
                                      38

-------
s
7
6
5
4
3
2
   SF6 THERMAL  DECOMPOSITION
       PILOT/LAB  COMPARISON
 ORE (-LOG(1-DRE/100))
               D
         D
n
     D
n
800   900    1000   1100   1200    1300
          TEMPERATURE, DEG  C
                            1400
      REF 5
      CRF REGRESSION
             REF 5 and 7
             CRF DATA
     Figure 3.  SF6 Thermal decomposition pilot/lab comparison.
                   39

-------
                                 POHC RANKING

     Figure 4 is a contour plot of the relationship between  conditions necessary
to achieve 6-nines SFg ORE and that of the individual POHCs.   It  also illustrates
why  looking  at  one  variable   at  a  time  can  be  confusing.    At   low  Oo
concentrations, tetrachloroethylene requires only 900°C for  6-nines while at 4.5%
Oo> 1250°C is required,  surpassing  that for chlorobenzene, which was more stable
at the lower Q£ concentrations.  Pentachlorobenzene DREs were  never below
6-nines and therefore are not plotted.

      Ranking  the POHCs is shown  in Figure  4.   This  figure  illustrates that
rankings  can  change, depending  on  combustion  conditions.    This  is  also true
depending on the  nature of the experimental device.

      Table 3 is a list of rankings  of POHCs derived from Figure 4  and from data
supplied by  UDRI  in Reference 9.   Although  not  definitive, the  research into
POHCs and POHC rankings have produced results which seem to make surrogate use
even more questionable than when it  was first suggested.

                       Table 3.  POHC Ranking'With 02  Present9

                                                  LIS Incineration
                                 UDRI     Low 0?  (1.5%)      High 09 (5%)
SFg
Tetrach 1 oroethy 1 ene
Toluene
Pentachlorobenzene
Chlorobenzene
1
2
5
3
4
1
4
2
5
4
2
3
1
5
3
               being most stable; 5 being  least stable.
                                  CONCLUSIONS

      o     SFg is a limited surrogate.  Data showing reactivity with refractory
            and difficulty  with  using  it may have  reduced  its  apparent value
      o     POHC ranking Js;not  absolute and depends on  combustion conditions
      o     Toluene  is themost  stable1'of the orgahics at  elevated temperature
            and 02  levels,  while SFg is most stable at  intermediate and  lower
            levels.

                                 RECOMMENDATIONS

      o     Modify the use  of  surrogates for trial  burns,  and  specify
            minimum  combustion cbnditions
      o     Choose a minimum number  of  POHCs for trial burns
      o     Use the most stable POHCs at low oxygen values, since that condition
            is present in most incinerator failures
      o     Develop  method  for  continuous monitoring  of toluene and benzene,
            for performance monitoring.
                                      40

-------
RELATIVE POHC STABILITY

T
E
M
P
E
R
A
T

U
R
E


C
l,^UU

1,300-


1,200-

1,100-



1,000-


900-
Qf\f\
* • . ' '
*.'
„,*'
A*
.••"i*

• ' " " ^** ++
NJ/T^ j_~r
++"f
4++
+
4-
i
Dn ++ .'..••'
>DnapDn . ' . ....
+• DDDQ
+ DnnDaannaaDDannnnaD •

    1   2  3  4  5  6   7  8  9  10
       O2 CONCENTRATION
           6-9'S POHC ORE
      SF6  + TCE   * TOL   ° CLBZ
       Figure 4.  Relative POHC stability.
            41

-------
                                  REFERENCES

1.    Oppelt, E.T.   Incineration of Hazardous Waste - A Critical Review.  JAPCA.
      37:  558, 1987.

2.    40 CFR Part 264.343.

3.    Tsang, W.M. and Shaub, W.M.  Surrogates  as  Substitutes  for Principal
      Organic Hazardous Constituent Validation  of Incinerator Operation.   In:
      Proceedings of the Second Conference  on  Municipal,  Hazardous  and'Coal
      Wastes, Miami, FL 1983.  p. 241.

4.    England, W.G., Rappolt, T.J., Teuscher, L.H., Kerrin, S.L. and Mournighan,
      R.E.  Measurement of Hazardous Waste Incineration Destruction and Removal
      Efficiencies  Using  Sulfur  Hexafluoride as a Chemical Surrogate.   In:
      Proceedings of the 79th Annual  APCA Meeting, 1986.   106:162097W.  .

5.    Pandompatam,  B., Liem, A.J., Frenette,  R. and Wilson, M.A.   Effect of
      Refractory on the Thermal Stability of SF6.   JAPCA.  39:  310, 1989.

6.    Del linger, B.  Testing and  Evaluation of a  POHC/PIC Incinerability
      Mixture.  U.S. EPA 15th Annual Research Symposium, Cincinnati, Ohio, April
      April 10-12,  1989.

7.    Taylor, P.H.  and Chadbourne, J.   Sulfur  Hexafluoride as a  Surrogate.
      JAPCA.  37:   729 1987.

8.    40 CFR 761.70a.

9.    Waterland, I.E. and  Staley, L.J.   Pilot  Scale Listing of SFg As A
      Hazardous Waste Incinerator Surrogate.   To  be Presented at the 82nd
      National Meeting of the Air and  Waste  Management  Association, Dallas, TX,
      1989.  Paper  89-23B.4.
                                      42

-------
                 A PROTOTYPE BAGHOUSE/DILUTION TUNNEL SYSTEM
                           FOR PARTICULATE SAMPLING      ".". -
                 OF HAZARDOUS AND MUNICIPAL WASTE INCINERATORS

                                  '   'by  -'•  .-'• ,  ,-"   .'  , •-;,.
                                P.M. Lemieux,
                                J.A. McSorley,
                               .'   W.P.  Linak., . "   ".,.  '.',.'..

                 United States  Environmental Protection Agency
                Air and Energy Engineering Research' Laboratory
                       Research Triangle Park,  NC 27711
                                  , ABSTRACT                           .

      EPA's  Air  and Energy Engineering Research Laboratory  (AEERL)  has
developed a prototype baghouse/dilution tunnel sampling system.  This  system
was designed originally  for  the  sampling of flue gas particulate from  fossil
fuel combustors, but has been modified to obtain samples of  particulate  matter
from hazardous  and municipal waste incinerators. _.- Samples  collected by this
sampling  system are to  be used  for  health  effects, studies.  , The sampling
system simulates the flue gas quenching processes 'occurring'upon emission from
stack to  the atmosphere.  A nominal  10:1  dilution-• with ambient air promotes
nucleation  of   vapor-phase organic compounds  and condensation  on existing
particulate matter.   This unit is  able to sample  2.8  dscm/min (100 cfm)  of
effluent.  At the allowable particulate loading rate of  180  mg/dscm  stipulated
by RCRA  regulations,  this sampler  is able to  capture  approximately 20 g of
sample in  1  hour.  At this  rate, it  is  feasible  to generate  kilogram-sized
particulate  samples that  are adequate  for  bioassay directed fractionation
and/or mouse skin painting carcinogenicity tests.   Replicate samples can also
be obtained,  so that duplicate health  effects  tests can  be performed, a  luxury
not normally available.   It is also  possible to sample semi-volatiles using
XAD-2 either upstream or downstream of the baghouse.

-------
                                INTRODUCTION

      Exhaust gases from combustion sources  typically  consist  of  a mixture of
Nar H2O,  COs,  O2f CO,  NOX,  acid gases,  particulate  matter,  and organic
compounds.   When the hot  stack gases contact the  atmosphere and cool  from
approximately 200°C (400°F) to  ambient  conditions, the particulate can  act as
nucleation sites  for  the condensation of semi-volatile hydrocarbons.   These
particles, when respired, can potentially deposit toxic,  carcinogenic, or
mutagenic material into the lungs  of humans.

      A  substantial  database  has been  accumulated on the  mutagenicity of
various, combustion  emissions, including  coal-fired power  plants,  diesel
engines,  and woodstoves.  It  is  desirable  to  extend  the  database to include
hazardous  and  municipal waste incinerators in order to  place their  health
effects into perspective with those of  other combustion emissions.

      The difficulty in acquiring  representative samples from incinerators for
the purpose  of  performing health effects research is compounded by  the need
for large samples.   Stack-based particulate  samples  from incinerators  may
typically contain approximately  1%  extractable  organic material  by  mass.
Whereas  chemical analyses require microgram-sized quantities  of  extractable
organic  material, microbial bioassays require milligram-sized  samples,  and
animal  cancer  studies require gram-sized samples.   In  other words, 1  kg of
particulate  is  needed to provide sufficient extractable  organic material to
perform  animal  studies.  At typical  stack  particulate loading rates  of 180
mg/m3, it is obvious that the  small sampling systems currently in use are not
practical.

                                  BACKGROUND

      In  1982,  EPA designed and constructed a large  particulate  sampler for
use in its synfuels program.   The sampler, a combination baghouse and dilution
tunnel system, was designed to pull 2.8 dscm (100  cfm) of  sample from a stack,
dilute  it with  28  dscm  (1000 cfm)  of filtered  ambient  air  to  simulate
atmospheric  quenching,  and pass the  diluted effluent  through a fabric filter
to capture the particulate.    The particulate  could then  be  removed from the
filter bags  by  mechanical  rapping or pulse  jet.   The sampler was designed and
constructed,  but never  used,  because the  synfuels program  was  discontinued
soon  after  the  unit's  construction.   When the  question of  health effects
sampling of incinerators  arose  in  1987,  this unit  was  resurrected  and
modified.

                                  APPARATUS

      The apparatus,  shown in  Figure 1,  consists  of five main components: the
ambient filter  housing, the  sample  probe and delivery  line,  the  dilution
tunnel,  the  baghouse,  and the blower.   All  parts  of the system except the
blower  are made  from  304  or 316  stainless  steel.  The entire system can be
placed  on a  48 ft  (14.6 m)  long  flatbed trailer  for  transportation to field
test  sites.    The  system can also  be oriented in  a number of  different
configurations  to allow for  space limitations when  the  unit  is to be mounted
on the  ground.
                                      44

-------
       The  ambient  filter housing contains,  in series,  a particulate filter,  a
 charcoal filter, a High Efficiency Particulate Air (HEPA)  filter,  and a 10 kW
 air preheater.   This filtering system removes particulate  matter and organics
 from  the  ambient  dilution  air,   and raises  the temperature so that  the
 resulting  diluted  gas mixture is 5-10°C  (10-20°F)  above the dew point.

       The  sample itself is withdrawn  from  the stack using a  stainless steel
 probe  and  a flexible 7.62  cm  (3  in.) Outside diameter (OD)   stainless steel
 sample line.   A 3 in.  NPT  nipple is required on the  stack  to be  sampled.
 Flanges  are  used to connect the  sample line to the dilution tunnel.

       The  dilution tunnel is  made  from 20 cm (8 in.) OD  x  3.35 m (11  ft)  long
 tubing,  and  contains a  number of ports for  external sampling.   Thermocouples
 are positioned before; and  after dilution.   A 3-way valve  provides  for
 continuous emission monitor  (CEM)  sampling  before or  after dilution.   There
 are also four 3-in. NPT ports  that can  be  used for XAD sampling,  pitot  tube
 measurements, or sling psychrometry.

       The  baghouse is 71 cm (28  in.) OD and 106 cm (42  in.)  tall.   It  contains
 a Gore-Tex filter  cartridge (model #38293-2).   An air  jet  above the cartridge
 provides a reverse pulse of  high  pressure air or nitrogen.  A Pyrex bulb at
 the bottom of  the  baghouse  catches the particulate matter that is  pulsed off
 the filter.   The dilution tunnel connects with the baghouse tangentially, so
 that a cyclonic effect is produced, driving some of the large particles  to the
 bottom of  the baghouse without impinging on the filter.

       The  sample and the  dilution air  are drawn  through the  system, by an
 induced  draft blower, capable of providing 1.49 kPa (60 in. water) of  static
 head pressure.   A  gate  valve  isolates the blower from, the rest of  the'system
 so  that the blower does not need to be shut  off during  pulsing.

                                   OPERATION

       The  baghouse/dilution  tunnel  sampler  has fairly  substantial  power
 requirements.   A 208 V,  3-phase,  90 A power supply is required  to run the
 blower and the  preheater.   Aside from the electricity requirements, the  unit
 is  self-contained.   It  can  be operated from an 80 kW generator if  sufficient
 local  power is unavailable.

       For  the proper operation of this unit  in cold weather, it is  absolutely
 critical to  ensure  that  the temperatures  of  no surface exposed to the  sample
 gas falls  below the dew point.   Incinerator  stack gas typically contains  HC1,
 which,  when dissolved in water, reacts with the stainless steel  sample system.
 In the case of adverse ambient weather conditions,  it is  necessary to insulate
 some or all of the exposed metal pieces, especially the  sample delivery line.
 Prior  to taking a  sample, it  is necessary to flush the system with preheated
 ambient air in order to  raise  the temperature of all exposed parts to at least
 45°C (110°F) .  Heated tape is  also  used to preheat the sample delivery line to
 100°C  (212°F).                                            .        .

      A pitot tube  can  be inserted into the dilution  tunnel  to periodically
monitor  the  gas  velocity.    The  dilution  ratio can  be  calculated either
 directly  by  CO2  ratios before  and after  dilution  or  indirectly  by  a
 calculation  based  on the temperatures before ,and  after dilution.   The gas
velocity, along with the dilution ratio,  can be used to calculate sample flow
 and particulate loading  rates.
                                      45

-------
      Periodic monitoring of the dew point .of the stack gases in the dilution
tunnel by sling psychrometry is necessary.  It is quite simple to monitor wet
bulb  temperature in  the dilution  tunnel due  to  the high  velocity of  the
diluted gas.   The air preheater  should  be adjusted so that  the  diluted gas
temperature is. always slightly "above the  dew point.

      The dilution tunnel vacuum and the  pressure drop across the baghouse are
monitored continuously.  By adjusting'the blower gate valve so that the system
vacuum is approximately  6'.07  kPa  (3 in.  of'water),  the flow rate through the
system can be held constant.  The initial pressure drop is approximately 0.12
kPa  (5 in.  of water)  with a new  filter, -and  rises  to  approximately 1.12 kPa
(45  in. of  water)  when the filter  becomes loaded with particulate.   Pulsing
the  system  will  knock particulate off the filter,  arid the pressure drop will
fall  back to  approximately 0.5  kPa  (20 in. of'water)." After the particulate
drops into  the  bulb,  the sample  is removed and ' isolated,  and the  bulb is
replaced.              ,'             ,     .                   ....'...'..

                                    RESULTS

      To date,'the EPA  prototype'baghouse/dilution tunnel  sampler  has-been
used at several different facilities,"with varying degrees of success.  First,
the  sampler was  operated on  the Rotary  Kiln  Incinerator  Simulator at the
Environmental Research Center in Research "Triangle Park, NC.   Experiments were
performed examining transient puffs resulting from batch charging of plastics.
Two  samples,  each weighing approximately 100 g, were acquired, and the EPA's
Health Effects Research Laboratory (HERL)  performed Salmonella mutagenicity
 (Ames) bioassays on the organic  material  extracted from  the particulate.  A
typical dose-response curve is shown in Figure 2.   HERL  is  performing animal
cancer studies on the  organic extracts at the present time.

      The  sampler  was  then-tested  at  several  full-scale  municipal waste
incinerators.    During these  tests,   water   condensation  problems^ were
discovered.  Because  the  flue  gas  from municipal waste  combustion ^typically
contains HC1,  the condensation  of water in the stainless steel sampling system
was  detrimental  to the usefulness of the samples.  When water condensed on^the
walls, it would  absorb HC1  from the gas. stream,  and create an acidic  solution.
Aqueous  HC1  can leach chromium  (Cr)  and other metals out  of the  stainless
steel.  Cr  present in bioassay  samples ruins the bioassay  because the Cr is  so
toxic that it kills the bacteria before they can revert  (mutate).   Since the
sampler was  initially  designed  for sampling fossil-fuel  combustors, the
extremely  high water content present  in  effluents  from  municipal  waste
 combustion (15-35% water by  weight)  was unforeseen.   In the presence  of low
 ambient temperatures, the only  way to  ensure that the dilution air  would
 remain above the dew point was  to provide  external heating to the  dilution
 air.  A 10  kW air preheater was installed  after these tests.

       After modifications, the  baghouse/dilution  tunnel was  then used  to
 sample another  full-scale municipal waste incinerator.   Although  the  unit
 operated properly,  and  water condensation was minimized, very  little sample
 was  collected.   Simultaneous  filter  samples  from  the EPA's  Source  Dilution
 Sampling System (SDSS) recovered no material either.   It  is possible that the
 particulate matter was so small that it  passed  through the filter, or that the
 particulate matter  was  an inorganic  fume rather  than flyash.   The Gore-Tex
 filters are rated to  capture  100%  of the particulate  matter greater than 0.6
 Urn.  Scanning electron microscopy  indicated that the residue deposited on the
 stack sample probe was mostly composed of particles smaller than 1 pm.
                                       46

-------
      The  system was also used  to sample fugitive emissions  at an aluminum
smelter.  In one of-the process buildings at  the smelter,  a  high  concentration
of  vapor phase  organic material  and a  considerable amount  of particulate
matter  were present.   The  system was  able  to collect  a  35.5  g  sample of
particulate matter,  which  was  within 10% of  the calculated capture, based on
reported particulate loading and estimated flow rates.  The  analyses for these
final two tests are being performed at the current  time.

                                 CONCLUSIONS

      Earlier successful tests  indicate that  this system has potential for use
as^a high-volume particulate sampler for health  effects studies.  Problems do
exist, but they appear to be surmountable.  This unit may not be  suitable for
sampling at facilities with  very small particulate . «1 |Jm)  or with very high
water content  (>30% by weight) .   It  also appears to be more  attractive to
install  a  baghouse/dilution tunnel  system in a permanent  or semi-permanent
fashion at a facility.   Even at extremely high sample volumes, several days or
even weeks of sampling may be necessary to ensure that sufficient particulate
matter is collected for subsequent health effects studies.

                               ACKNOWLEDGEMENTS

      Special thanks to David  M.  De Marini of HERL- for providing the sample
mutagenicity data.
      This paper has been  reviewed in accordance with the US  EPA's peer and
administrative review policies and approved  for presentation and publication.
                                     47,

-------
                 AMBIENTFILXERHOUSING
Fartkulate Filter
                                                                   DM— Sample Gate Valve
                                                                       - Thermocouple
           Charcoal Filter
     BLOWER
                 Blower Gate Valve
                                                               o
                                                                                SAMPLE DELIVERY
                                                                                            \
                                                                                 Sample Line'
                                                         DILUTION TUNNEL
                                                               O
                                                                                     Probe •
                                                                     • Thermocouple
              Figure  1: Diagram of Prototype Baghouse/Dilution
                              Tunnel Sampling System
                                            48

-------
           1000 r
                        20        40        60       80
                             Dose   (gg/plate)
100
 Figure 2:  Dose-Response Curve Derived from Salmonella TA98 Bioassay of
Particulate Matter Extract from Prototype Baghouse/Dilution Tunnel System
 During the Combustion of Polyethylene in EPA's Rotary Kiln Incinerator
                              Simulator                      "  '
                                 49

-------
                     EVALUATION OF ALTERNATIVE TREATMENT
                      TECHNOLOGIES FOR HAZARDOUS WASTES
                        FROM ACRYLONITRILE PRODUCTION
                             E. Radha Krishnan
                            PEI Associates, Inc.
                              Cincinnati, Ohio
                              Ronald J. Turner
                    U.S. Environmental Protection Agency
                    Risk Reduction Engineering Laboratory
                              Cincinnati, Ohio
                                    ABSTRACT

     Three aqueous waste streams from acrylonitrile production have been
included in the list of hazardous wastes under the U.S. Environmental Protec-
tion Agency's land disposal restrictions program:  1) the bottom stream from
the wastewater stripper (EPA Hazardous Waste No. K011); 2) the bottom stream
from the acetonitrile column (EPA Hazardous Waste No. K013); and 3) bottoms
from the acetonitrile purification column (EPA Hazardous Waste No. K014).
The listing constituents for K011, K013, and K014 include acrylonitrile,
acetonitrile, acrylamide, and hydrocyanic acid.

     The waste streams contain suspended solids consisting primarily of spent
metallic oxide catalyst particles and acrylonitrile polymers.  Current prac-
tice in industry is to mix the aqueous waste streams in settling ponds/ tanks
where the suspended solids are separated as a sludge, and to dispose the
liquid stream by deep well injection.  Because the sludge is derived from
hazardous wastes K011, K013, and K014, it has the same listing constituents
as the aqueous waste streams.  The sludge is generally disposed in offsite
landfills.

     This paper presents the results from: 1) a bench-scale evaluation of wet
air oxidation as a treatment technology for the mixed K011/K013/K014 waste
stream, and 2) a field investigation to evaluate the performance of rotary
kiln incineration as a treatment technology for the sludge.  The wet air-
oxidation tests were conducted at Zimpro Passavant's research facility in
Rothschild, Wisconsin, and the incineration tests were conducted at the John
Zink Company's test facility in Tulsa, Oklahoma.
                                       50

-------
 INTRODUCTION

      The Risk Reduction Engineering Laboratory (RREL)  of the U.S.  Environ-
 mental  Protection Agency (EPA)  is providing the Office of Solid Waste (OSW)
 with  data on various  hazardous  waste treatment technologies  to assist in  the
 development of land disposal  restriction standards  under the Resource Conser-
 vation  and Recovery Act (RCRA).  Three aqueous wastes  from acrylonitrile
 pr2d™Jl?n (which are 1dentified by EPA Hazardous Waste Numbers K011, K013,
 and K014) result from the purification of product streams in the acryloni-
 trile production process and  are defined as follows:

           K011:   Bottom stream  from the wastewater  stripper.
                                             "
           K013:   Bottom stream  from the acetonitrile column.

           K014:   Bottoms from the acetonitrile purification  column.

      In  addition to these wastes,  a K011/K013/K014  sludge is  also generated
 at acrylonitrile production plants  by  separation of the  suspended solids  from
 the mixed aqueous  wastes.

      The  EPA  is  required to'set treatment  standards for -acrylonitrile  produc-
 tion  wastes based  on  the best demonstrated available technology  (BOAT), as a
 prerequisite  for the  placement  of treatment  residues in  land  disposal  facil-
 ities..  The effective date of the  treatment  standard is  June  8,  1989.  In
 1987-88,  EPA:RREL  conducted a program  consisting of 1) sampling  and analysis
 to characterize  K011/K013/K014  aqueous  and sludge waste  streams  from  several
 acrylonitrile producers,  and  2)  treatability testing of  wet air  oxidation and
 incineration on  the mixed K011/K013/K014 aqueous waste stream and sludge
 respectively, from one  acrylonitrile production facility.  This  paper  pre-
 sents the  results of these tests.                     -         »    •


WASTE CHARACTERIZATION                                        -    ,

     There are six facilities in the United States that are involved in the
production of acrylonitrile and could generate K011, K013, and K014 listed :
wastes.   A few facilities do  not purify the crude acetonitrile stream pro-
duced in the process and, hence, do not generate K014.   The listing con-
stituents for K011, K013, and K014 include acrylonftrile, acetonitrile,
acrylamide, and hydrocyanic acid.                   ,                  ,  ?
                                     51

-------
     The K011 waste stream represents the largest single pollutant load
within the acrylonitrile production process.  Typically, the K011 waste
stream contains about 100 to 4,000 ppm of cyanide, 40 to 3,000 ppm of
acetonitrile, 0.2 to 8,000 ppm of acrylonitrile, 1,000 to 2,000 ppm of
acrylamide, and less than 200 ppm of acrolein.  In addition to the primary
contaminants listed above, this stream also contains approximately 4 percent
suspended solids.  The suspended solids consist largely of spent metallic
oxide catalyst particles and polymeric acrylonitrile.  Also, the K011 stream
contains about 10 percent dissolved sulfates.  Typical generation rates at
acrylonitrile plants for the K011 waste stream vary from 100 to 200 gallons
per minute.

     The K013 waste stream typically contains about 26 to 60 ppm of cyanide,
less than 35 ppm of acetonitrile, less than 10 ppm of acrylonitrile, less
than 120 ppm of acrylamide, and less than 1 ppm of acrolein.  This waste
stream constitutes greater than 99 percent water.  Typical generation rates
for the K013 waste stream at acrylonitrile plants vary from 100 to 200 gal-
lons per minute.

     Primary pollutants  in the K014 waste stream  are acetonitrile and cya-
nide.  Typically, the  K014 waste stream  contains  1,000  to 60,000 ppm of
acetonitrile,  and 5 to  5,000 ppm of cyanide.  Typical generation rates for
the K014 waste stream  at acrylonitrile plants vary from  5,000  to 20,000
gallons per day.

      In a  typical acrylonitrile production  facility,  the aqueous waste
streams  (K011, K013,  and K014) are  comingled  prior to their ultimate dis-
posal.  The mixed waste is  sent to  settling  ponds/tanks  where  the  suspended
solids  are removed  as  an underflow  sludge.   The  liquid  effluent  is  disposed
of by deep well  injection.   The sludge  generation rate  per  plant varies  from
 100  to  250 tons  per year.   The predominant  practice  in  industry  is  to  peri-
odically  dispose of the K011/K013/K014  sludge in offsite landfills.  The
 approximate concentrations  of  the major constituents in the  K011/K013/  K014
 sludge are as  follows:
                Constituent

      Silicon, molydenum, iron, and aluminum oxides
      Acrylonitrile polymers
      Inert shell, dirt, gravel, and kiln dust
      Water
      Acrylonitrile
      Acetonitrile
      Acrylamide
      Cyani de
Concentration
    range

   2 to 50%
   2 to 25%
  10 to 40%
  10 to 50%
 0.4 to 1 ppm
 0.7 to 3 ppm
   2 to 3 ppm
 900 to 2000 ppm
      The sludge has a heating value of 2,400 to 3,000 Btu/lb.
                                      52

-------
TEST FACILITY DESCRIPTIONS

     Several technologies, or combinations of technologies, may be applicable
for treatment of the mixed K011/K013/K014 aqueous wastes and sludge from the
acrylonitrile production process.  Some applicable technologies for treatment
of the listed organic constituents and cyanide present in the mixed K011/
K013/K014 aqueous wastes include (1) wet air oxidation, (2) wet air oxidation
followed by biological treatment, and (3) critical fluid extraction followed
by additional cyanide treatment of the aqueous phase and recycling or incin-
eration of the solvent phase.  An applicable technology for treatment of the
cyanide or organic components in the K011/K013/K014 sludge is rotary kiln
incineration.  The incineration technology may be combined with stabilization
for further treatment of the metals present in the incinerator ash.*

     In 1988, EPA-RREL conducted a bench-scale evaluation of wet air oxida-
tion technology for treatment of the mixed K011/K013/K014 aqueous wastes and
a pilot-scale evaluation of rotary kiln incineration technology for treatment
of the K011/K013/K014 sludge.  The mixed aqueous wastes and the sludge for
the tests were obtained from the same acrylonitrile production facility.  The
wet air oxidation tests were conducted at Zimpro Passavant's research facili-
ty in Rothschild, Wisconsin, and the incineration tests were conducted at the
John Zink Company's test facility in Tulsa, Oklahoma.

Zimpro Wet Air Oxidation Treatment System

     Wet air oxidation is a liquid-phase, oxidation process in which an
aqueous solution or suspension of organic and/or oxidizable inorganic com-
pounds are oxidized to carbon dioxide and other innocuous end products.  The
waste stream is thoroughly mixed with a gaseous source of oxygen (usually
air) at temperatures of 175 to 327°C (347-621°F) and pressures of 300 to
3,000 psig.  Elevated temperatures enhance the solubility of oxygen in aque-
ous waste thus providing a strong driving force for oxidation.  Elevated
pressures are required to control evaporation by maintaining water in the
liquid state.

     The bench-scale wet air oxidation tests on the combined K011/K013/K014
aqueous waste stream were carried out in 0.5-liter capacity titanium auto-
claves at Zimpro's research facility.  During each test, 125 ml of raw waste
was charged into the autoclave.  The autoclave was pressurized to 810 psig
with air to provide about 120 percent of the sample chemical oxygen demand
(COD) air requirements.  The autoclave was then placed in a heater-shaker
mechanism, heated to the desired temperature, and maintained at the tempera-
ture for the designated residence time.   The autoclave was cooled with tap
water, immediately after the completion of the residence time, and the off-
gases were analyzed by gas chromatography for total hydrocarbons and methane.
A series of three autoclave oxidations were performed at temperatures of
200°, 240°, and 280°C (392% 464°, and 536°F) and a liquid residence time of
60 minutes.  A total of four autoclave "runs were made at each condition and
the oxidized material combined before analyses.
  U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
  D.C.  Best Demonstrated Available Technology (BOAT) Background Document
  for K011, K013, and K014.   December 1988.
                                      53

-------
John Zink Rotary Kiln Incineration System

     Figure 1 presents a schematic of the John Zink pilot-scale rotary kiln
incineration system used for conducting tests on the K011/K013/K014 sludge.
The test system consisted of a rotary kiln and kiln afterburner for combus-
tion, a ram feeder for feeding waste contained in fiber packs into the kiln,
a water quench for kiln ash, and a venturi-scrubbing system for flue gas
treatment.  In addition, the Oklahoma State Department of Health required
that an additional afterburner, a fume incinerator, -be used during all tests
involving hazardous waste to provide further thermal treatment of the primary
combustion system's flue gas. , The minimum heat input to the rotary kiln
incineration system (including waste and natural gas supplementary fuel) was
2 million Btu/h and the maximum was 3 million Btu/h.  The kiln provided a
1-hour solids' residence time at 0.25 rpm.  The kiln afterburner provided a
flue gas residence time of 2.2 seconds at 2000°F.  Temperatures were con-
trolled by manually changing the supplementary fuel feed rate and/or the
combustion air flow rate.

     Continuous monitoring and recording of C02, CO, 02, and temperature were
conducted at the exit of the kiln afterburner at a point prior to the solids
separator.

     The hot combustion gases leaving the kiln afterburner entered a cyclonic
solids separator, which removed large pieces of fly ash that might damage
downstream equipment.  The gases then entered an adjustable-throat venturi
scrubber, which was followed by a droplet separator consisting of a cy-
clonic-flow knockout chamber and a vent stack.  A 15 percent sodium carbonate
solution was periodically added to the scrubber water at a rate of 0.8 gal-
lons per minute for a total of 10 to 15 minutes during each of the incinera-
tion tests to maintain the pH of the scrubber recycle water between 6 and 7.
The scrubber did not require any blowdown during the tests.;  Makeup water,
however, was continuously added to the scrubbing system to compensate for
evaporation losses.

     A total of three incineration tests, each of 2 to 2\ hours .duration,
were conducted on the K011/K013/K014 sludge.  The waste feed rate for all the
tests was 400 Ib/h; this feed rate corresponded to^6ne,: 10-1 b fiber pack with
K011/K013/K014 sludge being fed to the kiln every 90 seconds'.  Each sludge
fiber pack was spiked with 2 weight percent of chlorobenzene to evaluate
destruction removal efficiency (ORE) for the incineration tests.  Chloro-
benzene was selected as a surrogate principal organic hazardous constituent
(POHC) for ORE evaluation because of 1) stack gas sampling arid analytical
problems associated with the least incinerable POHC in the sludge (i.e.,
acetonitrile), and 2) the close ranking of incinerability parameters (i.e.,
heat of combustion and thermal stability) for chlorobenzene and acetonitrile.
During the three tests, the rotary kiln incineration test system operated
under near steady state conditions at average kiln operating temperatures of
1800°F and kiln afterburner temperatures of 2000°F.  The pressure drop across
the venturi scrubber averaged about 25 inches of water.  Stack gas oxygen
concentrations averaged about 4.2 percent, and CO concentrations were less
than 10 ppm.
                                     54

-------
                                                                            (O
                                                                           •o  10
                                                                            E  -(->
                                                                            (O  (/I
                                                                               0}
                                                                           •O  4J
                                                                            O)
                                                                            o;  c
                                                                           M-  O
                                                                               •r~
                                                                           TD  •»->
                                                                            C  res
                                                                            >••-
                                                                            to
                                                                               0)
                                                                           +-> I—
                                                                             S-
                                                                            o  o
                                                                            S- 
-------
                 Following the incineration tests, stabilization testing was conducted at
            the U.S. Army Waterways Experiment Station (WES), Vicksburg, Mississippi, on
            a composite kiln ash sample from the three tests.


            TEST PROTOCOLS

            Wet Air Oxidation Tests

                 The following four aqueous streams were sampled for the combined K011/
            K013/K014 bench-scale wet air oxidation tests:
                      K011/K013/K014 aqueous waste feed
                      Product from oxidation at 200°C (392°F) and 810 psig
                      Product from oxidation at 240°C (464°F) and 810 psig
                      Product from oxidation at 280°C (536°F) and 810 psig
                 The raw waste and the treated waste samples from each of the three tests
            were analyzed for the organic listing constituents of K011/K013/K014 as well
            as total and amenable cyanide content.  Chemical analyses for the listed
            organic compounds (with the exception of acrylamide) and cyanide were per-
            formed according to methods, from U.S. EPA's Test Methods for Evaluating Solid
            Wastes, Third Edition, SW-846, November 1986.  Acrylamide was quantitated by
            high pressure liquid chromatography with ultraviolet detection.   Additional
            analyses were also conducted on all four samples for sixteen metals and
            volatile organic priority pollutant compounds.

            Rotary Kiln Incineration Tests

                 Figure 1 identifies the five streams that were sampled for  the rotary
            kiln incineration tests of the K011/K013/K014 sludge:
                      Sample Point

                           A
                           B
                           C
                           D
                           E
          Description

K011/K013/K014 sludge waste feed
Kiln bottom ash
Fly ash from solids separator
Scrubber water before waste feed
Scrubber water durina waste feed
                 Eight different K011/K013/K014 sludge feed samples (A)  were submitted
            for analysis, two from the first test and three from each of the subsequent
            two tests.  The samples represented feed to the kiln at 1-hour intervals.
            Six different kiln ash samples (B) were collected:   two from each test.   The
            samples were collected at 1 to li hour intervals to enable the waste repre-
            sentative of the ash sample to be treated for 1 to  U residence times before
            ash sample collection.  One fly ash sample (C) was  collected after the com-
            pletion of the three tests.  One sample of the pretest scrubber water (D)  was
            taken immediately prior to initiation of the first  K011/K013/K014 sludge
            incineration test.  Six scrubber water samples (E)  were collected:  two
            during each test.  One sample was obtained near the midpoint of the test,  and
                                                 56
_

-------
one near the end of the test period.  Comprehensive analyses were performed
on the samples for the following parameters:  volatiles, semivolatiles,
metals, water quality parameters, organochlorine pesticides, phenoxyacetic
herbicides, organophosphorous herbicides, polychlorobiphenyls (PCBs), dioxins
and furans.  Table 1 identifies the specific constituents that were detected
above the practical auantitation limit (PQL) in any of the sludge feed, ash,
or scrubber water samples.  Samples were prepared and analyzed in accordance
with SW-846 methods.  Acrylamide, one of the organic listing constituents of
the K011/K013/K014 sludge, could not be analyzed by standard SW-846 methods;
high pressure liquid chromatography was used for acrylamide analysis.

     Stack sampling was conducted during each of the three tests at a loca-
tion after the venturi scrubber control system, but prior to the fume in-
cinerator, to determine the concentrations and mass emission rates of the
following:

     0    Volatile organics, including surrogate POHC (chlorobenzene) for ORE
          evaluation

     0    Particulate matter, hydrochloric acid (HC1) and total cyanide (CM)

          Metals

     Volatile organic emissions were measured in accordance with the Volatile
Organic Sampling Train (VOST) protocol (SW-846, Methods 0030 and 5040).
Particulate, HC1, and CN concentrations and mass rates were measured with a
modified EPA Method 5 sample train.  Metal emissions were collected with a
modified EPA Method 12 sampling train.

     Three different samples from stabilization testing on the composite kiln
ash were submitted for metals analysis.


TEST RESULTS

Wet Air Oxidation Tests

     Table 2 presents the analytical results for the organic listing con-
stituents and cyanide for each of the three wet air oxidation tests.  The
combined K011/K013/K014 aqueous waste feed for the test contained 575 ppm of
acetonitrile, 24.4 ppm of acrylonitrile, and 270 ppm of acrylamide.  The
amenable and total cyanide contents of the waste feed were 1,011 and 1,277
ppm, respectively.  The data from the tests showed that appreciable oxidation
of acetonitrile only occurred at 280°C (536°F), corresponding to an acetoni-
trile concentration of 90 ppm in the oxidized liquor.  This represents an 84
percent reduction relative to the feed.  Acrylonitrile was detected at 0.47
ppm in the liquor from oxidation at 200°C (392°F), but nondetectable at the
higher temperatures.  The acrylamide results indicated reductions of 96 to 98
percent in the oxidized samples, corresponding to residual concentrations of
11 to 5 ppm.  Amenable and total cyanide destruction efficiencies approached
                                     57

-------
         TABLE 1.  ANALYTES DETECTED IN SAMPLES FROM K011/K013/K014
                          SLUDGE INCINERATION TESTS
     Volatiles
 Metals
  Water quality
    parameters
Organophosphorous
    pesticides
Acetonitrile
Acrylonitrile
Chloroform
Methylene Chloride
1,1,1-Trichloroethane
Trichloroethylene
Acetone
Benzene
Styrene
Silver,,
Aluminum'
Arsenic
Barium
Beryllium
Cadmium
Chromium
Copper
Iron
Molybdenum
Nickel
Lead
Antimony
Silicon
Vanadium
Zinc
Hydrogen sulfide
Sulfate
Total organic carbon
Total cyanide
Chloride
Fluoride
  Phorate
  Ethyl parathion
            TABLE 2.  ANALYTICAL RESULTS FROM BENCH-SCALE WET AIR
               OXIDATION TESTS ON K011/K013/K014 AQUEOUS WASTE


Analyte
Acetonitrile
Acrylonitrile
Acryl amide

Feed
ppm
575
24
270
Cyanide (amenable) 1011
Cyanide (total)
a Product from
Product from
0 Product from
Not detected.
1277
oxidation
oxidation
oxidation

Oxidized
, Liquor No. 1 ,
ppm
690
.4 0.47
11
37
39
at 200°C and 810
at 240°C and 810
at 280°C and 810

Oxidized .
Liquor No. 2 ,
ppm
560
NDd
10
11.4
13.6
psig.
psig.
psig.

Oxidized
Liquor No. 3,
ppm
90
NDd
5
12.9
15.5




                                     58

-------
 99  percent at the two higher oxidation  temperatures,  corresponding  to  cyanide
 residual  levels  of 11.4 to  15.5  ppm.  There  was  no  significant  difference  in
 the residual  cyanide levels at 240°C  (464°F)  and 280°C  (536°F).   Higher
 operating pressures  than those used  in  the bench-scale  wet  air  oxidation
 tests  should  result  in further reductions of residual cyanide levels.

     Molybdenum  was  the metal  present at the  highest  concentration  (23.6 ppm)
 in  the waste  feed.   Other metals present in  the  feed  at concentrations above
 1 ppm  include barium (15.4  ppm), zinc (1.8 ppm), and  nickel  (1.08 ppm).  The
 products  from oxidation contained approximately  22  ppm  of molybdenum,  ap-
 proximately 3 ppm of barium,  up  to 54 ppm of  zinc,  approximately  3  ppm of
 nickel, and up to 391 ppm of copper.  The higher concentrations of  zinc,
 nickel, and copper in the oxidized products  relative  to the  feed  are attri-
 buted  to  possible solubilization of metals from  the solids  in the feed.

     Off-gases from  the tests  contained about 25 ppm  of total hydrocarbons
 and a  trace of methane.   The  oxygen concentration in  the off-gases  was main-
 tained in excess  of  6 percent  for all three  tests,  and  CO concentrations were
 below  the detectable level  of  100 ppm.

 Rotary Kiln Incineration Tests

     Table 3  presents the analyses for the detected volatile organic com-
 pounds and total  cyanide in K011/K013/K014 sludge feed  samples from each of
 the three incineration  tests.  The K011/K013/K014 sludge used for the tests
 contained low levels of acetonitrile  (up to 2.7  pg/g) and acrylonitrile (0.95
 ug/gh characteristic of sludges  generated from  physical treatment  of aqueous
 wastes from acrylonitrile production.  Acrylamide was not detected  in the
 sludge feed above its PQL of 6.5  yg/g.  Other volatiles  detected  at high
 levels in the K011/K013/K014 sludge feed were benzene and styrene.

     Tables 4 and 5  present the  incinerator ash  and scrubber recycle water
 analyses,  respectively,  for the  volatile organic  compounds and total" cyanide
 in  samples  from each of  the three  incineration tests.   No volatiles were
 detected  in the ash  and  scrubber  recycle water samples.   Lower detection
 limits  could  not  be  achieved for  the  volatile analytes  in the ash samples
 because of  analytical  interferences.  Total  cyanide concentrations were
 reduced from  levels  as high as 2,000 pg/g (ppm)  in  the  sludge feed to levels
 of  less than  22 pg/g in  the ash.    No cyanide  was  detected in the  scrubber
 recycle water samples.   As  with  the feed, no  acrylamide was detected in the
 ash and scrubber  recycle  water samples above  the  PQL of 6.5 pg/g and 1125-
 mg/liter,  respectively.

     The major metals found in the K011/K013/K014 sludge feed samples were
 aluminum  (up  to 1100  pg/g), barium (up to 200 pg/g), chromium (up to 200
 ug/g),  Iron (up to 4000  pg/g), nickel (up to 470 pg/g),  and zinc  (up to 210
 pg/g).  Molybdenum was also detected in the feed  samples at levels as high as
 17,000 pg/g.   The major metals  in the kiln ash were molybdenum (up to 45,000
 ug/g),  iron (up to 32,000 pg/g),  aluminum (up to 6,700 pg/g), and nickel  (up
 to 2,700 pg/g).  The  ash  Toxicity Characteristic Leaching Procedure,(TCLP)
 extracts contained the following metals  above their respective PQLs:  ar-
 senic, barium, chromium,  copper,  lead, nickel, selenium, and zinc.^Nickel
was the metal  with the highest ash TCLP  value at concentrations  of up to 14.0
mg/liter  (ppm).  For  the  scrubber recycle water samples, the major metals

                                      59

-------
r
                    TABLE 3.  VOLATILES AND CYANIDE ANALYSES OF K011/K013/K014
                          SLUDGE FEED FOR ROTARY KILN INCINERATION TESTS
                Analyte
Test No. 1,      Test No. 2,
    yg/g             yg/g
              Test; No., 3,
                  yg/g
           Volatiles-
             Acetonitrile
             Acrylonitrile
             Chloroform
             Methylene chloride
             1,1,1-Trichloroethane
             Trichloroethylene
             Acetone
             Benzene
             Styrene

           Total cyanide
  0.870
  0.410
  0.032
  0.034
  0.045
  0.016
    NDa
    57
    16

  1,200
1.200
0.540
0.030
0.240
0.029
0.014
0.095
  48
  15

2,000
2.700
0.950
0.042
0.210
0..032
0.018.
  NDa
  55
  18

1,500
             Not detected.
                                                 60

-------
         TABLE 4.   VOLATILES AND CYANIDE ANALYSES OF K011/K013/K014
             INCINERATOR ASH FOR ROTARY KILN INCINERATION TESTS
     Analyte
Test No. 1,
    yg/g
Test No. 2,
    yg/g
                                                              Test No.  3,
Volatiles
  Acetonitrile
  Acrylonitrile
  Chloroform
  Methylene chloride
  1,1,1-Trichloroethane
  Trichloroethylene
  Acetone
  Benzene
  Styrene

Total cyanide
  <500
  <500
  < 10
  <250
  < 10
  < 10
  <250
  < 10
  < 10

   10
  <500
  <500
  < 10
  <250
  < 10
  < 10
  <250
  < 10
  < 10

   12
<500
<500
< 10
<250
< 10
< 10
<250
< 10
< 10

 22
                                      61

-------
         TABLE 5.  VOLATILES AND CYANIDE ANALYSES OF K011/K013/K014
          SCRUBBER RECYCLE WATER FOR ROTARY KILN INCINERATION TESTS
     Analyte
Test No. 1,
 mg/liter
Test No. 2,
 mg/liter
                               Test No.  3,
                                mg/liter
Volatiles
  Acetonitrile
  Acrylonitrile .
  Chloroform
  Methylene chloride
  1,1,1-Tri chloroethane
  Trichloroethylene
  Acetone
  Benzene
  Styrene

Total cyanide
<0.5
<0.5
<0.01
<0.25
<0.01
<0.01
<0.25
<0.01
<0.01
                   <0.5
                   <0.5
                   <0.01
                   <0.25
                   <0.01
                   <0.01
                   <0.25
                   <0.01
                   <0.01
                   <0.5
                   <0.5
                   <0.01
                   <0.25
                   <0.01
                   <0.01
                   <0.25
                   <0.01
                   <0.01
                                      62

-------
present were iron (up to 99 rug/liter), molybdenum (up to 260 rag/liter),
silicon (up to 16 mg/liter), and zinc (up to 4 nig/liter).

     With one exception, no organophosphorous pesticides, organochlorine
pesticides/PCBs, organochlorine herbicides, dioxins or furans were detected
above the PQL in any of the-,K011/K013/K014 sludge feed, ash, or scrubber
water samples.  One sludge feed sample contained 0.11 pg/g of phorate and
0.27 yg/g of ethyl parathion.

     Stack emissions data showed a ORE greater than 99.99 percent for the
surrogate POHC, chlorobenzene.  Particulate concentrations corrected to 7
percent oxygen ranged between 0.10 and 0.31 grains per dry standard cubic
foot (gr/dscf) for the three tests, with an average of 0.21 gr/dscf; these
results exceed the required RCRA limit of 0.08 gr/dscf at 7 percent oxygen.
The failure to meet the RCRA particulate limit was attributed to poor per-
formance of the scrubber on the pilot-scale rotary kiln incineration test
system.  Full-scale air pollution control systems should be able to meet the
RCRA particulate limit.  HC1 emission rates ranged between 0.01 and 0.02
Ib/h, considerably below the 4.0 Ib/h RCRA limit.  Cyanide was not detected
in the stack emission samples.  Iron and molybdenum were the metals that
exhibited the highest concentrations in the stack gas, corresponding to mass
emission rates of up to 0.5 Ib/h and 0.1 Ib/h, respectively.

     Complete results from the kiln ash stabilization tests are currently not
available.
CONCLUSIONS

     Wet air oxidation and incineration appear to be effective treatment
methods for the K011/K013/K014 aqueous wastes and sludge, respectively, based
on the results of the tests conducted by EPA-RREL.  The wet air oxidation
tests demonstrated significant reduction of the K011/K013/K014 organic list-
ing constituents and cyanide at the highest oxidation temperature of 280°C
(536°F).  It is believed that residual organic constituents from wet air
oxidation would be amenable to biological treatment.  Residual metals in the
oxidized liquor may be treated by chemical  precipitation.  Operating data
collected during incineration tests of the K011/K013/K014 sludge show that
rotary kiln incineration is an appropriate treatment technology for the
sludge.  The ORE performance standard of 99.99 percent was achieved for the
surrogate POHC, chlorobenzene.  The incineration tests demonstrated signifi-
cant reductions of organic constituents and cyanide in the ash.  Metals in
the residual ash will require subsequent treatment by stabilization pro-
cesses.
                                      63

-------
                      The Role of Site Investigation
                   in the Selection  of  Corrective Actions
                   for Leaking Underground  Storage Tanks

                     by:  Myron  S.  Rosenberg
                          David  C.  Noonan
                          Camp Dresser & McKee Inc.
                          One Center Plaza
                          Boston, MA  02108

                                    and

                          Anthony N. Tafuri
                          Chi-Yuan Fan
                          U.S. Environmental Protection Agency
                          Risk Reduction Engineering Laboratory
                          Woodbridge Avenue
                          Edison, NJ  08837

                                    and

                          Iris Goodman
                          U.S. Environmental Protection Agency
                          Office of Underground Storage Tanks
                          401 M  Street, S.W.
                          Washington, DC  20460
                                 ABSTRACT

     There are numerous sites across the nation where soil treatment
technologies are being applied to clean up soil contaminated with petroleum
hydrocarbons from leaking underground storage tanks (USTs).  Developing an
accurate understanding of subsurface conditions at a site (i.e., a site
investigation) increases the likelihood that a given soil treatment
technology will be effective at a given site.  This paper presents an
approach for conducting a site investigation including identifying what
information about the subsurface environment and the released petroleum
product is needed and how it can be used.

     Worksheets are provided, to help the reader make a preliminary
determination as to where in the unsaturated zone most of the petroleum
hydrocarbons are likely to be and where they are likely to move.  Critical
success factors, or factors for determining how effective a given soil
treatment technology will be in cleaning up a particular site, are
provided.  Worksheets that contain the critical success factors for  two
soil treatment technologies (soil venting and biorestoration) are provided
in this paper so that a cross-comparison of site conditions can be
reviewed.
                                     64

-------
INTRODUCTION
     Cleaning up petroleum hydrocarbons in the unsaturated zone is a real
world problem at every site where releases have occurred.  The science
associated with soil treatment technologies is not as well documented or
understood as that associated with water treatment technologies.  Despite
this uncertainty, however, soil contamination must always be addressed by
engineers on behalf of the owners and operators at an UST site, and some
kind of soil cleanup is typically necessary when a release occurs.
Furthermore, the proposed approach for cleaning up the unsaturated zone
must be approved by the regulatory agencies charged with monitoring the
cleanup.

     The question at hand is how to move ahead with cleaning up the
unsaturated zone by selecting a treatment technology that is likely to be
effective given the uncertainty surrounding the removal of petroleum
hydrocarbons from soil.  The purpose of this paper is to provide an
approach for evaluating the likely effectiveness of soil treatment
technologies in the face of incomplete or uncertain data.  Our approach
builds on the present state of knowledge of soil treatment with a focus on
utility;  what technologies will likely work given our (limited)
understanding of conditions at the site.
     Intuitively, certain basic information is known about the site.
Information can be used to develop what we have termed a "site
investigation".
This
     A good understanding of the conditions in the unsaturated zone is
essential to selecting an appropriate soil treatment technology.  A good
site investigation includes enough information to answer the following
questions:

     •   What was released? Where? When?

     •   Currently, where in the unsaturated zone is most of the petroleum
         likely  to be?

     •   How much petroleum product is likely to be present in different
         locations and phases?

     «   How mobile are  the constituents of the contaminant, and where are
         they likely  to  travel and at what rate?

     Answers to  these questions—or at least "ballpark estimates" of  the
answers — can sometimes be easily developed.  Often an estimate is
sufficient to gain a  relative understanding of the potential effectiveness
of a given technology at a given site.  The need for site-specific measure-
ments of important parameters can never be eliminated, because there  is no
substitute for accurate, site-specific, field data.  However, it is
possible to combine actual measurements with literature values for other
                                     65

-------
             parameters  (for which no  field data are presently available) to make at
             least qualitative assessments of site conditions and to judge which
             corrective  action technologies are likely  to work.

             OVERVIEW OF APPROACH

                  Figure 1 shows the three basic components of the approach:  1) site
             investigation, 2) technology selection, and 3) monitoring and follow-up
             measurements.  Figure 1 also shows which questions are answered in each
             step.  Critical success factors, or CSFs as shown on Figure 1, are
             parameters  that are likely to determine the effectiveness of a given
             technology  in a given situation.  They are discussed in detail later in
             this paper.

             Site Investigation

                  A site investigation begins with basic information about the release
             itself.   Basic information about the release may be obtained by asking:

                  •   What contaminants were released?

                  •   How much was released?

                  •   Was the release slow or instantaneous?

                  •   How long ago did the release stop?

                  •   When was the release detected?


                  Information about  the site is  gathered next,  and usually involves
             searches of records  and  use of professional expertise.   Typical  information
             sought about the site includes:

                      Soil  Temperature
                      Soil  pH
                      Rainfall, Runoff, and Infiltration Rate
                      Soil  Surface Area
                      Organic Content
                      Soil  Porosity
                      Particle density
                     Bulk  density
                     Hydraulic Conductivity
                     Permeability                                       ,
                     Field Capacity
                     Soil Water Content
                     Local Depth to GW

                 If  information regarding  these critical parameters is unavailable  for
             the site of  interest, there are various sources of data that can provide
             default values for making  ballpark estimates of these critical parameters.
                                               66
.

-------
    KEY
 QUESTION
  WHAT WAS
  RELEASED?
 WHERE IN THE
 ,UN3ATURATED
 '  ZONE IS IT?'1
  HOW MUCH IS
    THERE?
  WHERE IS IT
    GOING?
 WHAT ARE THE
 TECHNOLOGIES?
 WHAT ARE EACH
 TECHNOLOGY'S
 ADVANTAGES?
 WHAT ARE THE
CRITICAL SUCCESS
   FACTORS?
.WHAT WILL. WORK
  'AT MY SITE?  ,
 EVALUATION OF
   SUCCESS*
   SELECTION
                                        ACTION
                                   SITE INVESTIGATION
  EVALUATE PHASE(S) OF
  CONTAMINANT(S) IN SOIL
    SELECT TECHNOLOGY
      BASED ON YOUR
SITE AND HOW IT MATCHES THE
    TECHNOLOGIEtfCSFs
                                             CE
PERFORMANCE MONITORING
                                   SAMPLING & MEASUREMENT
                                       OFSITECONDTIONS
                                   COMPARE TO PRE-ESTABLISHED
                                        CLEAN-UP GOALS
                                    YES
                                                          NO
                                    « CONTINUE TO MONITOR
                          Figure 1 .  An Overview of the Approach
                                           67

-------
     In addition  to release-related and site-related information, a site
investigation should include an understanding of the physical and chemical
nature of the contaminants released.  By drawing the relevant physical and
chemical characteristics, it is possible to estimate how the contaminant
may partition in  the subsurface (what phase it is likely to be in), how
readily it will move away from the site as a vapor or liquid, and whether
it is likely to degrade significantly over time.  Contaminant-specific
parameters, which can be found in many chemical handbooks, include:

         Unweathered Composition
         Pore Vapor Pressure
         Water Solubility
         Liquid Viscosity
         Liquid Density
         Vapor Density
         Soil Sorption Coefficient
     With this information in hand, some basic determinations can be made
as to "which "phase" most of the contaminants are in.

     For simplicity, it was assumed that petroleum hydrocarbons in the
unsaturated zone can exist only in three phases (as shown on Figure 2):  as
contaminant vapors in the pore spaces (vapor phase), as residual liquid
trapped between soil particles (liquid phase), or as liquid dissolved in
the pore water that surrounds soil particles (dissolved phase).  Note that
for recent petroleum releases (less than 1 year ago), most of the petroleum
is likely to be in the residual saturation phase with somewhat smaller
portions existing in the vapor and dissolved phases.  Although only 3
phases are considered in this paper, petroleum hydrocarbons can be found in
other phases as well.  Recent research done by EPA's RREL Office in Edison,
New Jersey identified as many as 14 different conditions under which
petroleum could be found in the subsurface.

     Finally, critical parameters for evaluating the mobility of contami-
nants exist.  Mobility is used here as a general term to indicate how
readily a contaminant moves into air and water.  Vapor pressure and
solubility are the primary indicators used in determining product mobility.
Vapor pressure is an indicator of how easily contaminants will volatilize
into air.  Solubility is an indicator of a contaminant's affinity for
water.  The mobility of the contaminants in the unsaturated zone is very
important.  Mobility directly affects the choice of corrective action
technology.

Technology Selection

     The second step of the approach involves selecting a technology based
on the collected information.  Critical success factors (CSFs), or factors
that are likely to determine the effectiveness of a given technology in
cleaning up the unsaturated zone in a given situation, can be identified
for each technology.  Critical success factors for soil venting, for
example, include the volatility of the contaminants, soil temperature,' and
moisture content of the soil.  Conditions at any site can be compared with
these factors to determine which soil treatment technologies are most
                                   68

-------
               GROUNDWATER'
DRY SOIL
PARTICLES
PORE
SPACES
WET SOIL
PARTICLES
                                                           UNSATURATED
                                                           ZONE
                CAPILARY2ONE

                SATURATED
                ZONE
                                    BEDROCK
                                       PETROLEUM
                                       VAPORS
                                       WPORE
                                       SPACES
                        RESDUAL
                      PETROLEUM
                        TRAPPED
                        BETWEEN
                       PARTICLES
                                                                  PETROLEUM
                                                                   DISSOLVED
                                                                      WSOIL
                                                                   MOISTURE
                CLEAN SOIL
SOIL CONTAMINATED BY
 PETROLEUM RELEASE
          Figure 2.  Representation of Three Different Phases in which
                    Petroleum can be Found in Unsaturated Zone
                                      69

-------
likely to be effective.  For this paper, CSFs are presented for soil
venting and biorestoration, two soil treatment technologies used at leaking
UST sites.  Note that the CSFs are not equally important.  Depending on
site conditions, it is likely that some CSFs will be much more important
than others.  It is not possible to prioritize the CSFs for all conditions
at all sites.

     Final selection of soil treatment technology will likely be based on
other criteria, however, such as what cleanup criteria are being used and
how much time is available to clean up the site.  In urgent situations,
such as where a municipal water supply well is threatened, the soil would
likely be excavated; excavation can be undertaken quickly.  If the cleanup
is not urgent, then other technologies such as biorestoration (which can
take as long as a year to be effective) might be preferred.  Certain soil
treatment technologies may not be able to achieve the cleanup goals if the
concentrations being used for cleanup criteria are set too low.

Performance Monitoring and Follow-up Measurements

     After a technology (or technologies) is selected and installed, it is
important to track its performance and monitor its effectiveness during
cleanup activities.  Monitoring and follow-up are essential because of the
uncertainty surrounding subsurface conditions.  If a technology's
performance is poor, it may be necessary to reexamine the data collected
during the site investigation and their inherent assumptions.  Poor cleanup
performance could be attributed to a misinterpreted or incomplete site
investigation which resulted in an incorrect selection of the technology.
This feedback loop is an important step in the entire cleanup process.

RESULTS OF EFFORT

Contaminant Location and Migration

     The rate and degree of partitioning of the residual liquid into the
vapor phase and dissolved phase depend on site-specific and
contaminant-specific factors, as well as time.  A contaminant's volatility
(as measured by vapor pressure) is an indicator of how easily it will move
into the air.  A contaminant's solubility is a measure of how easily it
will dissolve in water (including pore water).

     Vapor analysis and soil sampling can be undertaken to determine in
which phase most of the contamination resides.  Typically, soil is analyzed
for hydrocarbon concentrations of bulk liquid, while soil gas is sampled
for evidence of hydrocarbon vapors.  General "rules-of-thumb" for
determining the phase which contains the contamination are presented in
Table 1.  Vapor sampling at the site provides information into these rules
of thumb, and later, information that can be used to evaluate the soil
treatment technologies.

     The final step in a site investigation focuses on the mobility of the
contaminants.  Knowing where contaminants are likely to move, and how
likely they are to move, provides insight into how mobile the contaminants
are.  The success of a corrective action plan depends on the contaminants'
mobilization.   Mobilizing contaminants means being able to move
contaminants from one phase into another that can be more directly removed
                                   70

-------
                                  TABLE 1
                     RULES-OF-THUMB* FOR DETERMINING  IN
                   WHICH PHASE CONTAMINATION CAN BE FOUND
              Evidence of residual liquid contamination;

                  High concentrations (>1% by weight) of contaminants in
                  several soil analyses; (i.e., petroleum makes up >1% of
                  the weight of soil sample);

                  High concentrations (>10% by volume) of pure chemical
                  vapor density in several soil gas analyses (i.e.,
                  contaminant vapors are above 100,000 parts per million).
              Evidence of contaminant vapors;

              -   Presence of residual liquid contaminants;
                  Significant concentrations in several soil gas analyses.
              Evidence of pore-water contamination;

                  Significant concentrations of contaminants in several
                  analyses of pore water or groundwater;

                  Presence of residual liquid contaminants and a
                  significant soil moisture content.
^Sampling is required to apply these rules of thumb.
                                  71

-------
 with a given treatment technology.   For example,  soil venting works by
 causing a disequilibrium between phases.   An equilibrium is typically
 established among all three phases  when a petroleum release occurs.
 Removing contaminant-saturated air  with a soil venting system causes an
 equilibrium shift.  As clean (non-contaminated) air replaces the
 contaminant-saturated vapors that are removed, contaminants remaining as
 residual liquid will volatilize into the fresh air,  seeking to establish
 equilibrium.   As the process continues,  more and  more contaminant  in the
 residual liquid will be "mobilized" into the vapor phase,  where it  can be
 captured by the soil venting system.

      The migration of petroleum hydrocarbons into and out  of the two
 dominant phases,  as residual liquid and as vapors,  is governed by specific
 chemical and  environmental  factors.   Chemical and environmental factors
 that  can influence a liquid's mobility in the unsaturated  zone are  listed
 in Table 2 (grouped as soil-related and contaminant-related).   In Table 2,
 items are listed with both  qualitative descriptors  (high,  medium, and low),
 and  corresponding quantitative ranges of  values.  Although the quantifica-
 tion  ranges are somewhat  subjective,  they have been  arranged so that the
 right-hand column indicates "high mobility"  and the  left-hand  column "low
 mobility."

      Using Table 2 to evaluate conditions at a site  of  interest, it is
 possible to get an understanding of  the relative  mobility  of liquid
 contaminants  at the site.   If the preponderance of  factors at  a site fall
 in the  right-hand columns of Table  2,  liquid contaminants  would likely be
 more  mobile and likely to migrate than if most  factors  matched  those in the
 left  hand  column.   Where site-specific values  are unavailable,  default
 values  can be substituted as appropriate.  Additional work is  needed to
 prioritize or rank the CSFs because  in same  cases one CSF  could  override a
 preponderance of  other CSFs on one side of  the  table or  the  other.    Work
 continues  to  refine and improve the  tables and  to make  them  as  usable as
 possible.
     Vapors are generally mobile in
mobility greatly depends on the air-
porosity less that portion filled by
Several other factors also influence
zone as listed in Table 3.  Table 3
i.e., by comparing parameters from a
listed in the table, it is possible
mobility of vapor phase contaminants
the unsaturated zone.  The degree of
filled porosity of the soil (the total
 water or liquid contaminants).
 vapor transport in the unsaturated
is structured similarly to Table 2,
 given site of interest with those
to get an understanding of the relative
 at that site.
     Contaminant vapors may be mobilized (and subsequently removed) by
several natural or induced processes or driving forces.  These include:

     •   Bulk transport due to pressure gradients (e.g., from vacuum
         extraction wells);

     •   Bulk transport due to vapor density gradients (which could result,
         for example, if the contaminant vapor has a significantly
         different density than air or from temperature gradients);
                                   72

-------
TABLE 2
FACTORS TO EVALUATE THE EXTENT
OF MIGRATION OF LIQUID CONTAMINANTS
FACTOR
UNITS
SITE OF
INTEREST
INCREASING MIGRATION ^

RELEASE RELATED
• Tim* Sine* Last Release
Months

SITE- RELATED
• Hydraulic Conductivity
• Soil Porosity
• Soil Surfsos AIM
• Liquid Contaminant Content
• Sell Temperature
• Roek Fractures
• Water Content
CONTAMINANT- RELATED
• Uquld Viscosity
• Uquld Density
cm/sec
%
ma/g
%
«fc
—
%

cPofce
3
g/cfn










Long
(>12)
o
Medium
(1-12)
O
Short
(<1 )
O

Low
(<10's)
O
Low
(<10)
O
High
(>50)
O
Low
(<10)
0
Lew
(«10)
O
Absent
O
Wgh
(>30)
O
Medium
(10'5-10'9)
O
Medium
(10-30)
0
Medium
(5-50)
O
Medium
(10-30)
O
Medium
(10-20)
O
O
Medium
(10-30)
O

High
(>20)
O
Lew
<«D
O
Medium
(2-20)
O
Medium
(1-2)
O
High
(>10'»)
O
High
(>30)
0
Low
(«5)
0
High
(>30)
O
High
(>20)
O
Present
O
Low
(<10)
O

Low
(<2)
0
High
(>2)
O

73

-------
TABLE 3. FACTORS TO EVALUATE THE EXTENT OF
MIGRATION OF CONTAMINANT VAPORS
FACTOR
UNITS
SITE OF
INTEREST
INCREASING MIGRATION ^
•
'SITE -RELATED
• Air Filled PoroaKy
• Total Poreatty*
• Water Content*
• Depth Below Surface
%
%
%
nwtorc




Low
(<10)
o
Low
(<10)
O
High
(>30)
O
Deep
(>10)
O
CONTAMINANT- RELATED
• Vapor Density
g/m

Low
(<50)
O
* the total porosity bsc that fraction rifled wtth water equals th« *lr flltod poroclty


Medium
(10-30)
0
Medium
(10-30)
O
Medium
(10-30)
O
Medium
(2-10)
O

Medium
(50-500)
O


High
(>30)
O
High
(>30)
0
Low
(<10)
O
ShaBow
(<2)
O

High
(>500)
0

74

-------
     •   Sweep flow due to the in-situ generation of gases or vapors (e.g.,
         vapors volatilizing from liquid contaminant or gases generated by
         microbial biodegradation of contaminants);

     •   Molecular diffusion due to concentration gradients.

     The current scientific understanding of these processes spans the
range from well-documented to rudimentary and hypothetical.  Molecular
diffusion is, perhaps, the best understood and the easiest to address
experimentally and, hence empirically.  However, in some circumstances it
may not be the most important process governing vapor mobility.

     A soil with medium to high water-filled porosity will tend to
immobilize vapors; however, if one applies a vacuum extraction system,—
inducing soil gas flow — the soil will tend to dry out and this will
increase vapor mobility over the treatment period.

Technology Selection

     The focus of the approach thus far is developing an understanding of
what was released from the UST, where most of the contamination resides
(i.e., in which phase most of the contamination exists), and how likely the
contamination is to migrate (i.e., move into and out of other phases).
With the site investigation completed, the next step becomes one of
selecting a treatment technology that will be effective given the
conclusions drawn from the site investigation.  Many of the parameters
determined during the site investigation serve as indicators as to how well
a given soil treatment technology will perform.  These indicators of likely
performance are called CSFs.

     Tables 4 and 5 show the critical success factors for soil venting and
biorestoration.  The three right-hand columns of each table provide values
for the CSFs that suggest whether a technology is "less likely," "somewhat
likely," or "more likely" to be effective.  A column is provided for the
user to write down the values for the CSFs at the site of interest.  These
values can then be compared with the values in the right hand column, and
the corresponding circles checked off or darkened.  When completed, these
tables provide valuable insight as to which technologies are likely to be
effective and, more important, which are not.  A brief description of each
of the technologies and their related CSFs is provided below.

Soil Venting

     Soil venting is a general term that refers to any technique that
removes contaminant vapors from the unsaturated zone.  Venting may occur
passively (with no energy input) or actively.  Passive venting, which is
often used at sanitary landfills for methane gas removal, consists of
perforated pipes sunk into the contaminated area that provide an easy path
to the atmosphere.  These vents sometimes have a wind-driven turbine at the
outlet to provide a slight draft.

     More effective is active venting, where a pressure gradient is induced
to move vapors through the soil.  Most common is vacuum extraction
technology, where extraction wells are placed near the release site and a
                                   75

-------
TABLE 4.
WORKSHEET FOR EVALUATING THE FEASIBILITY OF
SOIL VENTING BEING EFFECTIVE AT YOUR SITE
CRITICAL SUCCESS
FACTOR
UNITS
SITE OF
INTEREST
SUCCESS
LESS
LIKELY
SUCCESS
SOMEWHAT
LIKELY
SITE RELATED
• Dominant
Contaminant Phase
• Soil Temperature
• Soil Hydraulic
Conductivity
• Moisture Content
* Geological
Conditions
• Soil Sorptlon Capacity
• Surface Area
• Depth to around water

Phase
C
cm/tec.
%
—
cma/g
nt









Sorbed to soil
o
Low
(<10)
0
Low
(< ID"5)
O
Moist
(>0.3)
0
Heterogeneous
O
High
(>50)
O
Low
(<3)
O

Uquld
0
Medium
(1(T- 20)
O
Medium
(10-4-10^
O
Moderate
(0.1 to 0.3)
O
o
o
Medium
(3-15)
O

SUCCESS
MORE
UKELY

Vapor or
Liquid
0
high
(>20)
O
high
(>10^
O
Dry
(<0.1)
o
Homogeneous
O
Low
(1S)
0

CONTAMINANT- RELATED
• Vapor Pressure
• Solubility
mmHg
mg/L
OTHER CONSIDERATIONS
• Co«t to tram S 1 5 to 160 p*r cubic yird.
• dp** 01 nmeving rouMnoi or oKtom.
• Air «nuiar* wii lk*y n*M to M mitt wBi QAC.


Low
(<10)
O
HlQn
(>1000)
O
Medium
(1010100)
O
Medium
(100-1000)
O
a tr— i-
r*Qn
(>100)
O
Low
(<100)
O
• •TiMmnt«nl»dar*on-ifti
• CM tnuH b« «Mn e MBU «viaHeni MOM* ««pan
• ClMnjp MM «M n m t» Mcmnogy * net
^propn»«<»n«n«m»fB»neyf»«ar»«l»n»«o«a
76

-------
TABLE 5.
WORKSHEET FOR EVALUATING THE FEASIBILITY
OF BIORESTORATION BEING EFFECTIVE AT YOUR SITE
CRITICAL SUCCESS
FACTOR
RELEASE • RELATED
• Tim* Sine* Release
SITE RELATED
• Dominant
Contaminant Phasa
• Soil Temperature
• Soil Hydraulic
Conductivity
• SollpH
• Molatura Content
CONTAMINANT- RELATED
• Solubility
• Biodagradablllty
• Raf rectory Index
• Fual Typa
UNITS

Months
SITE OF
INTEREST
SUCCESS
LESS
LIKELY


Short
<8)
O
o«y
«y>

Low
«100,
Low
(<001)
No. 6 Fual OH
(Heavy)
O
Vapor
O
Medium
(10». 20*C)
O
Medium
(10-5.10^
O
O
Moderate
(0.1 to 0.3)

Medkim
(100 to 1000)
O
(0.01 to 0.1)
No. 2 Fuel Oil
(Medium)
O
Long
(>12)
O

Dissolved
O
Ngh
(>20*C)
O
Hgh
(>io-»)
0
(6-8)
O
Moist
<1)3)

ttgh
(>1000)
Hgh
(>ai,
Gasoline
(Light)
O
OTHER CONSIDERATIONS
• Cost is from $60 to $125 per cubic yard.
• Completely destroys contaminants under optimal conditions
• Effectiveness varies depending on subsurface conditions
• Biologic systems subject to upset
• Public opinion sometimes against putting more chemicals in ground
• Difficult to monitor effectiveness
• Minimizes heath risk by keeping contaminarte in ground and on ste
• Takes long time to wortt— not for emergency response
77

-------
vacuum is applied to the wells.  The soil gas is drawn through  the soil to
the extraction well and brought to the surface.

     Venting removes contaminants in the vapor phase, but also  affects to
a limited extent residual liquid contaminants.  Dissolved contaminants are
unaffected.  Hydrocarbons typically are found in all three phases, and an
equilibrium is established, with a certain fraction existing in each phase.
The portion found in each phase depends on both the particular  compound and
the local conditions.  If conditions change, the equilibrium will shift,
and contaminants will transfer between phases to re-establish an
equilibrium.                          ".

     The success of a vacuum extraction program depends both on the
properties of the contaminants and the properties of the soil.  Compounds
with high Henry's Law constants will be removed to a greater degree by
vacuum extraction than the others, so this technology is more effective for
volatile compounds.

     The water solubility of each contaminant will also affect  the success
of venting, although this factor is relatively less important than those
listed above.  Highly soluble compounds will tend to exist predominantly
dissolved in pore water, with less in the vapor phase.  Vacuum  extraction
tends to dry out the soil, however, and over time dissolved contaminants
will likely volatilize and be removed.  Work continues on ranking the CSFs
to improve the usability of the tables.

     Soil properties also greatly influence the success of soil venting.
Table 4 lists several soil properties to consider.  Most important is the
soil hydraulic conductivity.  Soil with low permeability, such  as clay,
restrict the movement of vapors through the soil and towards wells.
Contaminants can still be removed from low permeability soil by soil
venting, but the process requires more closely spaced wells or  a greater
vacuum.  Most soils have preferential flow paths that are responsible for
most of the soil's permeability.  These flow paths, which result from
things such as root intrusions, prevent the vapors from coming  into
intimate contact with all of the contaminated soil, thus decreasing the
effectiveness of the technique.

     Other important properties include soil temperature and moisture
content.  The ambient temperature of the soil has a strong effect on the
volatility of the contaminant.  As temperature rises, vapor pressure and
Henry's Law constant rise dramatically.  For this reason, soil  venting
would be expected to be more successful in areas where soil temperature is
high.  In some cases, air is heated prior to injection to raise the
temperature of the soil and increase volatilization.

     The moisture content has two effects on the soil.  First,  because soil
with a high water content has relatively less air-filled porosity, higher
water content leads to a lower air permeability and therefore a lower
removal rate.  Second, pore water can absorb (dissolve) contaminants from
the vapor phase, which serves to retard the removal of contaminant vapors.
This is especially true of contaminants with low vapor pressures and low
Henry's Law constants.  Dry soil is thus better suited to in-situ stripping
than wet soil.  The vacuum extraction process tends to dry out  the soilj
                                  78

-------
 over time the air permeability will increase and the dissolved contaminants
 will volatilize, both of which tend to increase the degree of removal.

      Carbon content is related to vapor pressure and can serve as a
 substitute indicator.  Compounds with lower carbon content (C, to C •) can
 be vented more effectively than compounds with high carbon content (C0 to
 C16).                                                                9

      Once Table 4 has been completed, other factors, such as cost, must be
 included in the evaluation before making a final selection of a technology.
 In general,  soil venting is a relatively inexpensive technique compared
 with other alternatives, especially when large volumes of soil must be
 treated.  The capital costs of venting consist basically of the extraction
 and monitoring well construction, one or more blowers and housing, pipes,
 valves,  fitting, and other hardware, and electrical instrumentation.
 Operations and maintenance costs consist of labor,  power, maintenance, and
 monitoring.   Venting wells constructed of two-inch  diameter slotted PVC
 pipe cost approximately $20 per linear foot (0.3 m) for a twenty foot
 depth.   Vacuum pump sizing depends on local soil conditions and the volume
 to be treated.   Operating costs vary depending on time of operation and
 local utility rates.  The actual costs of soil venting at a Florida site
 were estimated to be $106,000 (capital) and $68,000 (annual O&M).   Air
 treatment,  if necessary,  would have more than double costs.  These figures
 correspond to roughly $20 to $60/yd  ($25 to $78/m3).   Air emission control
 via GAC  is usually assumed to double the total capital cost of the cleanup.
 Soil venting programs are relatively easy to implement and may be  installed
 and started  in two to four weeks.   This time is devoted to determining the
 extent of contamination,  designing the system,  acquiring pumps and piping,
 and installing the equipment.

      A venting program will typically be operated for  six to  twelve months.
 The removal  rate is  usually highest at the beginning of the program (once
 the vacuum is established)  and falls off after the  most volatile
 contaminants  are removed.   Volatilization from dissolved and  sprbed
 contaminants  then becomes  rate-limiting,  and the system's effectiveness  may
 decline  dramatically.

 Biorestoration

      Biorestoration  is  the  process  of  adding nutrients  to enhance  the
 natural  biodegradation  processes of soil  microbes.   It  may also involve  the
 addition of specially-adapted  microbes to  the subsurface,  but  this  is  not a
 common procedure.  Experience  has shown  that at  sites where the release
 occurred long ago, native soil microbe populations  are  very large,  and the
 use  of introduced microorganisms should not  be necessary.   Indigenous
 microbes are expected  to be as efficient at  degradation as
 specially-acclimated microorganisms.   Acclimated microorganisms should be
 thought  of only as a convenient method of quickly increasing microbial
 populations.  Specially-acclimated  microorganisms must  compete with native
 populations and be able to move from point of injection  to  the location of
 contamination, and maintain selectivity for  the contaminant of interest.

     Oxygen is the single most important ingredient in a successful
biorestoration program.  Although biodegradation may continue to occur
anaerobically, the lack of oxygen severely limits the rate of cleanup.
                                   79

-------
Oxygen may be introduced by pumping air into the unsaturated zone, e.g.,
soil venting.  In addition to oxygen, soil microbes also require
macronutrients (nitrogen and phosphorous) to survive and prosper.  These
nutrients are typically added in order to facilitate biodegradation.
                                                                         by
     Slightly alkaline soil pH is optimum for biodegradation, but anything
in the range of 6.0-8.0 is considered acceptable.  Redox potential is
important for various anaerobic microbes.  The ratio of oxidized materials
to reduced materials in soil establishes an electrical potential, called
the redox potential.  The temperature of the soil environment will also
affect the rate of degradation.  Warmer temperatures generally result in
higher rates of degradation.  While biodegradation has been shown to occur
over a wide temperature range, the range of 20° to 35 °C seems optimal.
Also, microbes generally have a low tolerance for severe temperature
changes .

     A knowledge of the hydraulic conductivity of the soil and the site
hydrogeology in general is extremely important in assessing the feasibility
of biorestoration for a particular site.  Even when all other factors are
positive, in-situ biorestoration will not be successful if a low hydraulic
conductivity prevents the added nutrients and oxygen from contacting the
zone of contamination.  The residence time should be short enough so that
the oxygen concentration is sufficient throughout the site for microbes to
degrade all of the organic compounds.  Also, the geochemistry of the
subsurface could inhibit adequate mixing if reactions (such as metal oxide
precipitation) clog the soil.  The soil microbes themselves may clog the
soil and decrease the hydraulic conductivity.

     Table 5 lists the critical success factors for biorestoration.  By
comparing the parameters at the site of interest with those in this table,
a general understanding can be obtained of the suitability of
biorestoration at that site.  A preponderance of CSFs that match the
right-most column would indicate that biorestoration is likely to be
effective at that site.

     As with soil venting, several other considerations must be included in
the evaluation before a final selection can be made.

     The costs of biorestoration for the unsaturated zone vary widely and
are difficult to quantify and compare.  Also, most reported costs refer to
cleaning up groundwater rather than the unsaturated zone.  One estimate
gave costs of $60 to $123 per cubic yard ($78 to $160 per cubic meter).
Unit costs for larger volumes are generally lower due to economies of
scale.

     A biorestoration program can be set up relatively quickly, but it may
take several months for the microbes to become adjusted and start
significant degradation if the contaminant release is recent, or if
non-indigenous microorganisms are used.  The system may need to be
"fine-tuned" (i.e., varying the levels of oxygen and nutrients added) to
operate efficiently.  The start of a biorestoration program may be delayed
due to the drilling of injection and extraction wells, the design and
procurement of construction of the oxygenation equipment, and the need for
injection permits.  Other important factors are listed at the bottom of
Table 5.
                                    80

-------
 SUMMARY

      Table 6 presents a summary of the critical success factors for several
 treatment methods and excavation.   Table 6 includes only the objective
 (scientific) factors that could affect the choice of technologies at a
 specific site.   Economic, political,  regulatory,  and other potentially
 controversial factors often are as important as these objective factors.
 Therefore,  this summary table is useful for direct comparison of the
 technologies on the grounds of technical feasibility only.   When potential
 treatment technologies have been narrowed to those that appear technically
 feasible, other considerations (cost,  public perception,  etc.) will likely
 affect  the final selection.

      The intent has been to provide an approach for selecting a soil
 treatment that  is based on scientific  understanding to the extent possible
 yet  recognizes  the uncertainties associated with  cleaning up the
 subsurface.   Several limitations should be noted  when using the approach
 proposed herein:

      •    Not for Emergency Response -  It is assumed that  all necessary
          emergency responses have  been taken,  that the source of the
          release (e.g.,  tank or supply line)  has  been identified and
          repaired,  and that  proper notification of government agencies
          (local,  state and federal) has taken place.

      •    Unsaturated Zone Coverage Only - This paper addresses site
          investigation and corrective  action  for  contamination in the
          unsaturated zone only.  Guidance is  not  provided  for contamination
          of  the saturated zone.  In some instances,  an assessment of the
          unsaturated zone might  proceed regardless of potential or actual
          groundwater contamination, but it  should  not be assumed that  the
          groundwater is  uncontaminated.   The  presence of a  floating
          contaminant layer on  the  water table,  or  a contaminant  plume  in
          the groundwater,  may  ultimately affect the selection of the
          unsaturated zone corrective action or actions.  Integrated
          guidance  for site assessment,  corrective  action, and evaluation
          for both  the unsaturated  and  saturated zones contamination is not
          addressed  in this document.

     •    Focus  on Petroleum  Hydrocarbons  as Contaminants - Because
          petroleum  products  comprise by far the greatest percentage of
          materials  stored  in USTs,   most  assumptions made in  this  paper are
          biased  toward  these materials.   Special focus  is given  to
          gasoline.

     The  proposed approach outlined in  this paper  shows promise  for  the
future.   EPA, recognizing  some of  the  inherent weaknesses in  the  state of
knowledge regarding  soil  treatment, continues  to pursue research  aimed at
removing  some of these weaknesses and uncertainties.  EPA's Office  of
Underground  Storage Tanks  and Office of Research and  Development are moving
forward with research projects and studies  that focus on increasing  the
understanding of how soil  treatment can be made more effective and
efficient.
                                     81

-------
Z
O
t-
>
u
X
UJ


O (/)
Ij EC
zffl
SOIL
FLUSHING
BIO-
STORATION
c


A

tn


1 CRITICAL SUCCESS
FACTORS




















1 RELEASE -RELATED



1





I

1
f?
1



1


1

1 TIME SINCE LAST RELEASE


8
•j-
H




|

1
1



1


i

1 VOLUME OF SPILL


8
t
1




1

1
1



*»,
I


"E

1 VOLUME SOIL CONTAMINATED




















e
L
t



i





|

o
2,
i
o
2.
I



o
2,
I


1

1 HYDRAULIC CONDUCTIVITY



1





1

1
i



i


p

I SOIL TEMPERATURE



1





1

!
i



§
t


§

I SOIL SORPTKJN CAPACITY



1





1

1
Mgh(>30)



S
v
1


*

I SOIL MOISTURE CONTENT



1





1

1
p
p



1


5

i









i

i
i



i
o
\

1
Is
IJ
UJ
i









"v
1

t
1



t


*

1 SOIL CARBON CONTENT

J_l
V

«



	


ff




2.


1

i
?



i





i











M
UJ
£
§
J
: 5
'



I





1











1 PHOXtMTiY TO LANDFILL









1

Htgh(>1000)
1



i
(
}


i

1 CONTAMINANT- RELATED
CONTAMINANT SOLUBILITY



1





1

6-
^
1



I
I


i

i



1





1

1
WBh(>0.1)






^

i
\
i



1





v
I

(7
^
1






i





1





1

(7
*
i
i



*
*
i


•





i





i

i
i



i











UJ
o
z
DC
2
CC
S FOR OPTIMUM
SS FACTOR!
UJ
8
F CRITICAL SI
O
c
<
s
1 a
i
*
82

-------
                  SUMMARY OF THE RESULTS OF EPA'S EVALUATION
                     OF VOLUMETRIC LEAK DETECTION METHODS
               by: Joseph W. Maresca, Jr., James W. Starr, and Robert D. Roach

                                     Vista Research, Inc.
                                       100 View Street
                                       P.O. Box 998
                                 Mountain View, CA 94042
                                            and                                 ,     :
                            John S. Farlow and Robert W. Hillger

                            U.S.  Environmental Protection Agency
                                  Releases Control Branch
                           Risk Reduction Engineering Laboratory
                                     Edison, NJ 08837                              "

                                        ABSTRACT

      A United States Environmental Protection Agency (EPA) research program evaluated the
 current performance of commercially available volumetric test methods for the detection of   '
 small leaks in underground gasoline storage tanks. The evaluations were performed at the EPA
 Risk Reduction Engineering Laboratory's Underground Storage Tank Test Apparatus in Edison,
 New Jersey.

      The methodology used for evaluation made it possible to determine and resolve most of
 the technological and engineering  issues associated with volumetric leak detection, as well as to
 define the current practice of commercially available  test methods. The approach used
 (1) experimentally validated models of the  important sources of ambient noise  that affect vol-
 ume changes in nonleaking and leaking tanks, (2) a large database  of product-temperature
 changes  that result from the delivery of product to a tank at a different temperature than the
 product  in the tank, and (3) a mathematical model of each test method to estimate the perform-
 ance of that method.  The test-method model includes the instrumentation noise, the configura-
 tion of the temperature sensors, the test protocol, the  data analysis algorithms,  and  the detection
 criterion.

      Twenty-five commercially available volumetric leak detection systems were evaluated.
 The leak rate measurable by these systems ranged from 0.26 to 6.97 L/h (0.07 to L84 gal/h),
 with a probability of detection of 0.95 and a probability of false alarm of 6.05. Five methods
 achieved a performance less than 0.57 L/h  (0.15 gal/h). Only one  method was able to detect
 leaks less than 0.57 L/h (0.15 gal/h) if the probability of detection was increased to 0.99 and
 the probability of false alarm was decreased to 0.01.  The measurable leak rates ranged from
 0.47 to 12.95 L/h (0.12 to 3.42  gal/h) with  these more stringent detection and false alarm
 parameters.

     The performance of the methods evaluated was primarily limited by test  protocol, opera-
 tional sensor configuration, data analysis, and calibration, rather than by hardware.   The exper-
imental analysis and model calculations suggested that substantial performance improvements
could be realized by making only procedural changes.  With modifications, it is estimated that
the majority of the methods should be able to achieve a probability of detection of 0.99 and a
probability of false alarm of 0.01 for leak rates between 0.19 L/h and 0.57 L/h (0.15 gal/h).
                                           83

-------
                                     INTRODUCTION
      Leaking underground storage tank systems represent a serious environmental threat.
There are over two million underground storage tank systems in the United States.  Estimates of
the fraction of these tank systems that may be leaking range from 10 to 25%. Records from
past release incidents indicate that, without the use of release detection, a release can become
substantial before it is detected (1,2).
      There are many commercially available  methods for detecting leaks in underground stor-
age tanks.  Those which are the most widely used in the petroleum industry are in a category
called volumetric tank tests (also known as "precision," "tank tightness," or "tank integrity" tests).
The premise of a volumetric tank test, and hence its name,  is that any change in the volume of
fluid  within a tank can be interpreted as a leak.  Detection  of these leaks is difficult because
there are many physical mechanisms which produce volume changes that can be mistaken for
leaks. Most of the volumetric tanks tests on the market today claim the ability to detect leaks
as small as 0.19 L/h (0.05 gal/h). (This is the "practice" recommended in the National Fire
Protection Association (NFPA) Pamphlet 329  (3) for volumetric tests  in tanks less than 47,316 L
(12,499 gal) in capacity.) These volumetric tank tests, however, do not specify the reliability of
their test results in terms of probability of detection (PD) and probability of false alarm (PFA)
against this 0.19-L/h leak rate.   The 1984 Hazardous and Solid  Waste Amendments to  the
Resource Conservation and Recovery Act of  1976 charged the United States Environmental Pro-
tection Agency (EPA) with developing regulations for the detection of releases from under-
ground storage tanks. The new  regulations, released in September 1988, state that all  volumetric
tank test methods must have the capability of detecting leaks as small as 0.38 L/h (0.1 gal/h)
with a PD of 0.95 and a PFA of  0.05.  These are only minimum standards, and the tank
owner/operator may want better protection against the possibility of a testing mistake.
      Development of technically sound and defensible regulations required that both the threat
 to the environment and the technological limits of release detection be known.  The threat to
 the environment is extremely difficult to define because the amount of petroleum that is haz-
 ardous to the environment is site specific.
       A performance standard that is based on the current  technology will minimize the uncon-
 trolled release of petroleum product.  Unfortunately, the data required to formulate a realistic
 regulatory policy were incomplete or nonexistent before the study described in this paper was
 undertaken.  While many commercial leak detection methods are available  and can be used to
 detect small releases, the performance  of these methods was unknown.  Very little evidence,
 theoretical or experimental, had been provided by the manufacturers to support their perform-
 ance claims. The limited evidence available  prior to these  evaluations suggested that  the major-
 ity of these methods were not reliably meeting these claims.  In 1986, the  Risk Reduction
 Engineering Laboratory (RREL) (formerly the Hazardous Waste Engineering Research
 Laboratory) of the EPA initiated an experimental research program to evaluate the current
 practices.  Participation in the program was voluntary, and the manufacturers of 25 commer-
 cially available methods elected to participate.
                                            84

-------
     The specific objectives of the program were to produce the technical data necessary to
support the development of release detection regulations; to define the current practice of com-
mercially available systems; to make specific recommendations to improve the current practice;
and to provide technical information that would help users select suitable methods for testing
the integrity of underground storage tanks. This paper summarizes the results of the EPA
research program. A detailed description of the Edison evaluations is presented in a two-
volume EPA  technical report (4).  Volume I of the report contains the objectives and the chro-
nology of the experiments, a thorough explanation of the engineering principles underlying the
experiments,  and a comprehensive analysis of the results.  Volume II includes the individual
evaluation  reports written for each of the 25 test methods.
                              VOLUMETRIC TEST METHODS
      A volumetric test method measures the change in product volume that results from a leak
in the tank; a leak may manifest itself as a release of product from the tank or an inflow of
ground water into the tank.  Figure 1 gives an overview of the test procedure used by most test
methods.  These methods measure product level and product temperature. The product-level
and temperature data are converted to product volumes.  The volume changes produced by
thermal changes of the produced are then subtracted from the product volume data. The result-
ing temperature-compensated flow rate is then compared to a predetermined flow rate for that
method called the threshold.  If the flow rate exceeds the threshold, the tank is declared
leaking.  If not, the tank is declared tight.  The test procedure usually  incorporates one or more
waiting periods after any addition of product to the tank to minimize the effects of structural
deformation and horizontal inhomogeneities in the product-temperature field. While the details
of the actual instrumentation, measurement protocols, data  reduction and analysis algorithms,
and detection criteria differ from method to method, the testing approach is essentially the same
for all methods.
      Volumetric tank tests can be divided into two categories. In the first, the tank is filled to
capacity, and in the second, the tank is partially filled.  In filling a tank to capacity the opera-
tor does not stop until the level of the product reaches the  fill tube (or a standpipe located
above grade); hence, the term "overfilled" is applied to these tests.  Overfilled-tank tests can be
further categorized according to those conducted  under nearly constant hydrostatic pressure (i.e.,
those with a nearly constant product level) and those conducted under  variable hydrostatic pres-
sure (i.e., those with a fluctuating product level).
      In a constant-level test product is added or removed  in order to maintain a constant fluid
level in the tank's fill tube or standpipe. To conduct a successful test, it is  necessary, once a
tank has been filled and then again after it has been topped off prior to testing, to observe a
waiting period long enough to ensure that the tank has expanded to its maximum capacity and
the temperature fluctuations have subsided. Then, if the fluid level is kept constant, the tank
will neither  expand nor contract during the test, and measured volume changes will accurately
represent actual volume changes.  Partially-filled-tank tests are generally considered constant-
level tests.  Because the surface area of  the product is spread across the width and  length of the
tank, any level changes  will be quite small, regardless of the size of the associated volume

                                           85

-------
change.  Unless product is added or removed during the test, causing the tank to deform, a
partially filled tank behaves in the same manner as an overfilled tank given that the appropriate
waiting periods are observed.
                                 Grois Adjustment ol Product Level
                                        (Product Delivery)
                                      Wait for Temperature
                                    and Structural Deformation
                                        to Become Small
                                        Equipment Setup
                                 Fine Adjustmant of Product Level
                                 	  (Topping the Tank)
                             Product Level and Temperature Measurement
Figure 1.  Overview of volumetric tank test procedure.
      In a variable-level test, the fluid level is allowed to fluctuate. When such a test is con-
ducted in an overfilled tank, the surface area of the product is extremely small ~ it is usually
limited to  one or more small diameter openings (e.g., fill tube or vent pipe).  Any volume
changes  will thus be seen as significant height changes.  Unless the deformation characteristics
of the tank being tested are known (as well as those of the backfill and surrounding soil) it is
not possible to distinguish between the volume changes due to a leak and those that normally
occur in a nonleaking tank. These deformation characteristics are not known at the time of the
test, and it is impractical to measure them. There is, consequently, a high risk of large errors
in variable-level tests (5).

                             HOW PERFORMANCE IS DEFINED
      The  accuracy of a  test method in ascertaining whether the tank is leaking or not is what is
meant by the performance of the method.  Performance is defined by the test method's PD and
Ppx for  each  leak rate that the method claims to be able to detect. The PD refers to the test's
chances  of correctly identifying a leaking  tank compared to its chances of failing to detect a
leak that is actually present.  The  PFA refers to a test's chances of declaring a leak when in fact
                                             86

-------
none exists. The key to how well a test method performs is its ability to discriminate between
the volume changes produced by a leak (i.e., the signal) and other volume changes that normally
occur in a tank (i.e., the noise).  The latter can either mask a leak, or mimic a leak and thus be
confused with one. The noise  is the sum of  the product-level or product-volume changes pro-
duced by the measurement system itself, by the environment, and by the operational practice.
The major sources of environmental noise are produced by product temperature changes, vapor
pockets, structural deformation of the tank system, evaporation and condensation  from the
product surface and tank walls, and internal  and surface waves.  The noise field is particularly
large immediately after delivery of product to the tank or  topping of the tank during a leak
detection test.
     The performance of a detection system can only be determined once the fluctuation level
(product-level or product-volume changes) at the output of the measurement system  is known
with and without the signal present.  For any test method, the statistical fluctuation of the noise
is observed in the histogram of the volume-rate results created by plotting the measured volume
rates from a large number of tests conducted (1) over a wide  range of conditions, (2) with many
systems on one or more nonleaking tanks, and (3) by many different  operators. The  histogram
indicates the probability that a particular volume rate will  result from a test  on a  nonleaking
tank.  The histogram  of the noise is developed experimentally. The histogram of  the signal-
plus-noise is usually developed from a model that indicates how the signal adds to the noise.
     A  complete specification of system performance requires a description of the PD and the
PFA at a defined leak rate and  an estimate of the uncertainty  of the PD and PFA.  If, in addi-
tion, a frequency of testing is specified, then the limits of the threat  to the environment, the
confidence with which these limits can be met, and the costs associated with mistakes in testing
can be defined.
     The threshold is often confused with the leak rate to be detected.  The EPA release
detection regulation requires volumetric test methods to have a minimum detectable leak rate of
0.38 L/h (0.1 gal/h).  In order  for a test method to meet this requirement, its threshold must be
less than 0.38 L/h. If the threshold is equal  to the  leak rate to be detected, the PD is only 0.50.
     Choosing the right balance between the PD and PFA is a very difficult  task.  Missed
detections result in the release  of product into the ground and the consequential contamination
of the nation's major source of potable water.  False alarms lead to the  expense of additional
testing and/or the repair or replacement of tanks that are not leaking. It is fair to expect  tank
owners to interpret this balance in terms of financial considerations.  The clean-up costs result-
ing from a missed detection must be weighed against the cost of unnecessary testing  and repairs
resulting from a false alarm.

                                EVALUATION APPROACH
     A  three-step procedure was used to conduct the  evaluations (9-12). The first step was to
develop and experimentally confirm models of the  important sources  of noise that control  the
performance of each test method.  If the total noise field is accurately modeled, the sum of the
volume contributions  from each noise source will be equal to the product-level changes in a
nonleaking tank.  As  part of the modeling effort, two large databases, reflecting the  different
product  temperature conditions which could  be experienced during field testing, was obtained to

                                            87

-------
simulate a test performed after a delivery of approximately 15,000 L (4,000 gal) of product at
one temperature to a 30,000-L (8,000 gal) storage tank half-filled with product at another tem-
perature. Temperature data were collected for a range of temperature differences of -jJO°C
between new and existing product.  One  database corresponds to tests in a tank as the product
temperature naturally attempts to come into equilibrium with the surrounding backfill and
native soil, and the second to tests that deliberately mix or circulate the product in the tank.
      The second step was to develop and validate, for each leak detection method, a model that
mathematically described it.  The test-method model includes the precision and accuracy of the
instruments; the test protocol; the data collection, analysis and compensation algorithms; and the
detection criterion. The model, in turn,  was validated in two steps. First, each manufacturer
was required to review the model for accuracy and to concur that it accurately represented the
method before the evaluation was allowed to continue; and second, the manufacturer was
required to participate in a three-day program of tank-test and calibration experiments at the
UST Test Apparatus. The manufacturer used his own testing crews and test equipment for the
three days of testing. Methods that were not operational at the time of the tests, or that were
different from those with which their respective manufacturers had concurred, were not evalu-
ated.
      Finally, a performance estimate for each method was made by combining, in a simulation,
the test-method model approved by the manufacturer, the product-level measurements estimated
from the noise models, and the temperature database. The performance of a test method was
evaluated by repeatedly simulating  the conduct of a tank test in order to develop a histogram of
the noise.  Operational effects and deviations from the prescribed protocols during the three-
day field testing program were also examined and discussed.
      0.2
  I
  I  0.1
  8T
 u_
      0.0
       -400
        R3
        1
                0
Volume Rate - ml/h
400
Figure 2. Histogram of the noise generated for the five-thermistor test method.
      Examples of the noise histogram and performance curves illustrative of the output gener-
 ated for each test method evaluated are shown in Figures 2 and 3 for a hypothetical test method
 which tests when product is near the top of the tank, using an array of five equally spaced,
                                            88

-------
volumetrically weighted thermistors.  It was assumed that the temperature and product-level
sensors used by this method had sufficient precision to measure ambient product-volume
changes that were less than 0.04 L/h (0.01 gal/h).  The data were sampled once per minute and
the duration of the test was 1 h.  The only source of noise considered in the simulation was
thermal expansion or contraction of the product.
                                                    I  e-'°
                                  1000   2000   MOO
                        Threshold • mtfh
•UOO   -2000   -IOM    0   1000   2000   MOO

           Threshold • ml/h
                                                    S.
                      o.oc      o.io     o.u
                     Probability ol False Alarm
        O.OS     0.10     016
       Probability at False Alarm
 Figure 3. Examples of performance curves for the five-thermistor test method.  (A) Pj> vs.
 Threshold, (B) PPA vs. Threshold, (C) PD vs. PPAfor flow out of the  tank and (D) PD vs. PPAfor
 flow into the  tank.
       The performance is presented in three displays, the first of which is  a plot (Figure 3A) of
 the probability of detection versus detection threshold for a family of leak rates with flow into
 and out of the tank (positive and negative volume rates).  The second display is a plot (Figure
 3B) of the probability of false alarm versus threshold.  The third display shows the probability
 of detection versus the probability of false alarm for a family of leak rates.  Note that the third
 display is separated into two  plots, one for outflows (Figure 3C) and one for inflows (Figure
 3D).
                                             89

-------
      Performance curves were generated that were based on the simulated noise and signal-
plus-noise histograms. For high levels of performance, the PD and PFA are estimated from the
tails of the histogram. With limited data, good estimates of the PD and PFA are sometimes
difficult to make. In this study, the performance estimates were typically based on 50 to 200
independent realizations of the manufacturer's test. To reduce the uncertainty in  the perform-
ance estimates at the higher PDs and lower PFAs, an exponential curve was fit to the tails of the
histogram. The PD and the PFA obtained from the curve were used to estimate performance; an
estimate of the uncertainty of the PD  and the PFA was also made.
               UNDERGROUND STORAGE TANK (UST) TEST APPARATUS
      The evaluations were performed by the Risk Reduction Engineering Laboratory (RREL)
at the  EPA's UST Test Apparatus located in Edison, New Jersey.   The Test Apparatus is envi-
ronmentally safe and was designed and built to evaluate the performance of in-tank leak detec-
tion systems.  Construction was completed in August 1986. The Test Apparatus consists of two
2.43-m (8-ft)-diameter, 30,000-L (8,000-gal) underground storage tanks installed  in a
pea-gravel backfill material; one is  a steel tank coated with plastic, and  the other  is a fiberglass
tank. -Two above-ground tanks are used to heat or cool product for simulation  of a delivery to
the underground tanks.  With this combined apparatus, different product temperatures, product
levels, and leak rates can be generated and accurately measured. The apparatus can be and was
configured to investigate each source of noise experimentally. In  November 1987, a pressurized
pipeline system was added to the Test Apparatus, but was not used in this evaluation.
      To address the overall project objective, a set of data quality objectives was established at
the beginning of the program and was adhered to throughout the data collection (11).  The data
quality objectives for this project were established to evaluate the 0.19-L/h (0.05-gal/h) per-
formance claim.  The precision and accuracy of the product-level and temperature data col-
lected  at the UST Test Apparatus were specified so as to evaluate the performance of each test
method at a leak rate of 0.19 L/h with a probability of detection of 0.95 and  a  probability of
false alarm of 0.001, a more stringent requirement  than either the draft (PD of 0.99, PFA of
0.01) or the final (PD of 0.95, PFA of 0.05) EPA release detection  standard. This requires that
the precison of the instruments used to measure temperature and product level and the accuracy
of the  constants used to convert temperature and product level to volume must have a total
uncertainty of less than 0.04 L/h (0.01  gal/h) when the data are combined to estimate the
temperature-compensated volume rate. The UST Test Apparatus instrumentation, calibration
procedures, and data quality analyses after each test were designed to verify that the data were
meeting the data  quality objectives.

                        RESULTS OF THE EDISON EVALUATION
     The names  of the 25 commercially  available volumetric test  methods that were evaluated
by the EPA at the UST Test Apparatus are presented, along with their manufacturers,  in
Table 1. The performance of 19 of the 25 test  methods is summarized in Table 2.  Proper
interpretation of the quantitative performance estimates given in Table 2 for overfilled-tank-
test methods requires the use of the last column in  Table 2.  Table 3 lists the names of the
methods for which no quantitative estimates could be made and the reason.
                                           90

-------
Table 1.  PARTICIPANTS COMPLETING THE EPA VOLUMETRIC TEST METHOD EVAL-
UATION PROGRAM
         Test Method Name
           Test Method
           Manufacturer
    Telephone
     Number
AES/Brockman Leak Detecting System
Ainlay Tank 'Tegrity Tester
Automatic Tank Monitor and Tester
 (AUTAMAT)
Computerized VPLT Tank Leak Test-
ing
 System
DWY Leak Sensor
EZYCHEK
Gasoline Tank Monitor (GTM)
Gilbarco Tank Monitor
Inductive Leak Detector 3100
INSTA-TEST
Leak Computer
Leak-O-Meter
LiquidManager
LMS-750
MCG-1100
Mooney Leak Detection System
OTEC Leak Sensor
PACE Leak Tester
 Petro Tite
 Portable Small Leak Detector (PSLD)
 S.M.A.R.T.

 Tank Auditor
 Tank Monitoring Device (TMD-1)
 Tank Sentry II
 TLS-250 Tank Level Sensing System
Associated Environmental Systems
Soiltest, Inc.
Exxon Research and Engineering Co.

NDE Technology, Inc.
DWY Corp.
Horner Creative Products, Inc.
Tidel Systems
Gilbarco, Inc.
Sarasota Automations, Inc.
EASI, Inc.
Tank Audit, Inc.
Fluid Components, Inc.
Colt Industries
Pneumercator Co., Inc.
L & J Engineering, Inc.
The Mooney Equipment Co., Inc.
OTEC, Inc.
PACE (Petroleum Association for
Conservation of the Canadian Envi-
ronment)
Heath Consultants, Inc.
TankTech, Inc.
Michael & Associates of Columbia,
Inc.
Leak Detection Systems, Inc.
Pandel Instruments, Inc.
Core Laboratories, Inc.
Veeder-Root Co.
(805) 393-2212
(312) 869-5500
(201) 765-3786

(213) 212-5244
(715)
(517)
(214)
(919)
(813)
(219)
(619)
(619)
(813)
(516)
(312)
(504)
(715)
(416)
735-9520
684-7180
416-8222
292-3011
366-8770
239-7003
693-8277
744-6950
882-0663
293-8450
396-2600
282-6959
735-9520
298-1144
 (617) 344-1400
 (303) 757-7876
 (803) 786-4192

 (617) 740-1717
 (214)660-1106
 (512) 289-2673
 (203) 527-7201
      In Table 2, test methods are arranged by alphabetical order in three categories:
 (1) partially filled-tank-test methods, (2) constant-level overfilled-tank-test methods, and
 (3) variable-level overfilled-tank-test methods.  The reason for this arrangement is that
 performance is largely controlled by characteristics particular to each of these categories.  For
 example, vapor may be trapped in an overfilled-tank-test but not in a partially-filled-tank test;
 evaporation and condensation will be an important source of error in a partially-filled-tank test
 but not in an overfilled-tank test; in overfilled tank tests, the  tank must be topped prior to the
 test, an action that carries with it the risk of degrading the performance, while in a
 partially-filled-tank test topping  is not necessary.

      The first column in Table 2 lists the name of each method. The second, third, and fourth
 columns show the mean, standard deviation, and number of the simulated temperature-
 compensated tank test data that were used to estimate performance. Using these data, perform-
 ance curves were generated for each method.  Two interesting results, shown in column 5, and
 columns 6 and 7, were derived from these curves.  The fifth column presents the
                                             91

-------
Table 2. ESTIMATES OF THE PERFORMANCE OF 19 VOLUMETRIC TEST METHODS
PD & PFA Smallest Smallest
to Detect Detectable Detectable
Number a 0.38 L/h Leak Rate Leak Rate Waiting
of (0.10 for for Period
Standard Simulat gal/h) PD=0.9S PD=0.99 After
Test Method Name

Mean
(L/h)
Deviation ed Tests
(L/h)

Partially-Filled-Tank
Gasoline Tank Moni-
tor
(GTM)i
Gilbarco Tank Moni-
tor1
Inductive Leak
Detector
3100
Tank Sentry II
TLS-2501
0.105


0.016

0.055


-0.093
-0.016
0.408


0.075

1.012


0.154
0.142
Overfilled-Tank Test
Leak Computer
MCG-110
Petro Tite
0.005
0.206
0.002
0.096
0.119
0.209
13


59

45


23
46
Leak Rate PFA=0.05
(PD,PFA)
(L/h)
PFA=0.01
(L/h)
Topping
(h)
Test Methods
0.73,0.21


0.96,0.003

0.72,0.33


0.89,0.16
0.15,0.001
1.35


0.26

4.23


0.58
0.51
1.91


0.47

9.54


0.89
0.90
N/A


N/A

N/A


N/A
N/A
Methods/Nearly Constant Level
132
97
25
0.97,0.04
0.97,0.09
0.79,0.21
Overfilled-Tank Test Methods/Variable
AES/Brockman Leak
Detecting System2
Ainlay Tank 'Tegrity
Tester
Computerized VPLT
Tank Leak Testing
System
EZY CHEK
Leak-O-Meter
LiquidManager
Mooney Leak Detec-
tion
System
PACE Leak Tester
Portable Small Leak
Detector (PSLD)
S.M.A.R.T.
Tank Auditor
-0.167

0.076

0.023


0.048
-1.060
0.307
-0.266


0.143
-0.192

-0.033
1.048
0.910

0.470

0.230


0.184
2.072
0.168
0.551


0.810
0.871

0.366
1.107
112

284

99


399
231
79
196


245
135

81
207
0.45,0.34

0.50,0.31

0.66,0.19


0.86,0.15
0.57,0.49
0.80,0.14
0.47,0.38


0.37,0.32
0.63,0.32

0.58,0.32
0.57,0.43
0.32
0.36
0.80
Level
6.79

2.97

1.08


0.62
6.96
0.75
3.13


6.97
3.05

2.25
6.27
0.65
0.86
1.11

12.95

3.93

1.84


0.93
10.80
1.25
4.58


11.12
5.60

3.43
•12.52
Variable
Variable
Variable 3

0

2

0


1
0
0.75
Variable


0
0

0 or 12
0
1


2


S
These test methods were employed in a special precision test mode rather than in their
normal operating mode as automatic tank gauging systems (ATGS)
Data analysis algorithms for this method had to be modified in order to determine per-
formance.
5 to 8 min/1,000 gal of product in tank at high level and 0 h when product level is
dropped to low level for testing.
                                          92

-------
performance that the method should be able to achieve in actual tank tests.  This performance is
expressed in terms of the probability of detection and probability of false alarm for leak rates
as small as 0.38 L/h (0.1 gal/h), using  the manufacturer's detection criterion as defined at the
time of the evaluation.  (The majority  of the manufactuers used a detection threshold of 0.19
L/h (0.05 gal/h) to determine whether the tank was leaking or not.) The sixth and seventh
columns present the potential performance of the method in terms of the smallest detectable
leak rate that might be  achieved for a  PD of 0.95 and a PFA of 0.05, and a PD of 0.99 and a  PFA
of 0.01, respectively. This performance estimate does not employ the manufacturer's detection
threshold; instead, a threshold was selected which yields  a probability of false alarm of 0.05 and
0.01, respectively.  These quantitative  performance estimates should not be used unless the
reader first understands what was used to generate  them.
      There were no quantitative performance estimates  made for 6 of the 25 methods in the
evaluation program; they are listed  in  Table 3, along  with the reason why no performance  esti-
mate was made.  The reason was one of three. In two cases, the manufacturer's test crew  could
not perform a satisfactory tank test during the 72-hour period allotted to them during the  field
tests at the UST Test Apparatus. In three cases, the  data obtained during the field tests clearly
indicated that the method was not behaving as the  manufacturer had said it would.  In general,
these methods used an  integrating temperature-compensation measurement approach, which had not
been experimentally validated adequately by the manufacturer prior to the evaluation. In the
final case, the Test Apparatus was  not properly configured for all of the field tests, preventing
an adequate field test of the method; in this latter  case,  however, the temperature-compensation
scheme of this method  had also not been experimentally validated prior  to the evaluation.
Table 3. LIST OF TEST METHODS  NOT EVALUATED
       Test Method Name
               Reason
 Autamat
 DWY Leak Sensor
 Insta-Test
 LMS-750
 OTEC Leak Sensor
 TMD-1
Operational principles could not be verified
Operational principles could not be verified
Did not successfully conduct a tank test
Operational principles could not be verified
Improper configuration of Test Apparatus
Did not successfully conduct a tank test
                      INTERPRETATION OF EVALUATION RESULTS
       Three general remarks apply to nearly all of the methods evaluated.  First, the majority of
 methods exhibited a bias; the magnitude of this bias is evidenced in the size of the mean of the
 data in Table 2. In general, the performance of methods with a bias can not be accurately
 estimated, because unless the physical mechanisms producing the bias are known and can be
 quantified (so that the bias can be removed), performance can change from test to test. The
 bias was arbitrarily removed for the estimates presented in Table 2.  If the bias is  large (i.e., if
 it represents a large percentage of the leak rate to be detected), the method should be consid-
 ered suspect.
                                              93

-------
      Second, experimental estimates of the precision and accuracy of each method's instrumen-
tation were derived from data obtained from a calibration of the level and temperature mea-
surement systems; this calibration was done as part of the evaluation.  It is assumed that the
instruments used in actual practice are accurately and routinely calibrated, and that the
precision and accuracy of these instruments are equal to or better than the precision and accu-
racy used in the performance estimate.  It was observed during the evaluation field tests that the
calibration procedures that many of the  manufacturers used or claimed to use were generally
inadequate, and that in many cases the instruments had never been satisfactorily calibrated at
all.
      Third, deviations from the protocol alter performance; during these evaluations, perform-
ance was seen to improve as well as to suffer as  a result of changes in protocol.  In order  to
make the performance estimates, therefore, it was assumed that the test protocol as given by the
manufacturer is always followed precisely.  This implicitly  assumes that only the best test  crews
are used to  execute a test, that is, the type of crew that participated in the evaluation program.
      The remaining remarks apply specifically to test methods in certain categories. Not  all
sources of error were included in the performance estimates presented in Table 2, and as a
consequence, the actual performance of a method may be poorer than the performance shown
here.
      The effects of evaporation and condensation were  not included in the estimates for meth-
ods that test in partially filled tanks. In general, these effects  are small, but in some circum-
stances  they can be large enough to cause testing errors.
      The effects of a product delivery are included in all of the performance estimates for
methods that overfill the tank and maintain a constant level of product. However, the degrada-
tion in  performance that results from topping the tank during an overfilled-tank test were not
included in the performance estimates, nor were the effects of any product-level changes that
are required before starting a test.  The  effects of topping are  to produce spatial inhomogenei-
ties in the product-temperature field due to the addition  of product at a different temperature
than exists in the tank and changes in tank volume due to structural deformation.  Waiting
periods for  temperature  fluctuations and structural deformation to subside are usually incorpo-
rated into test protocols, but in general these waiting periods were found to be too short.  The
short waiting periods shown in Table 2 suggest the magnitude of the problem. Thus, in actual
practice, performance could be  significantly reduced in comparison to that presented in Table 2.
      In all overfilled-tank tests the potential exists for trapping vapor in the top of the tank or
in its associated piping.  The effects of trapped vapor were not included in the  performance
estimates in Table 2 for either one of two reasons: most  manufacturers claim to be able to
remove vapor before a test begins and without experimental evidence to the contrary,  this claim
was accepted as true; and even if trapped vapor were included in the estimates, it would be
difficult to do this  accurately, because the distribution of the volume of trapped vapor is
unknown.  That vapor will be trapped, however, is almost inevitable, and the performance esti-
mates shown in Table 2 will be reduced  if this vapor is not removed before a test.
                                           94

-------
     The same effects that have the potential to degrade the performance of constant-level
tests also impact the performance of variable-tests. In addition, when the product level within
the fill tube or standpipe is allowed to fluctuate during a test, it is nearly impossible to convert
product-level measurements to volume.  Test methods that allow this situation to occur have a
basic flaw in protocol and should be considered suspect (4,5).

                STATISTICAL SUMMARY OF PERFORMANCE RESULTS
     The estimates of potential performance from Table 2 are summarized  in Tables 4 and 5.
Table 4 summarizes the methods1 actual performance using the manufacturers' detection criteria.
Table 5 summarizes potential performance using a detection criterion that produces a probability
of false alarm of either 0.05 or 0.01. This performance is expressed as the smallest leak rate
than can be detected with a probability  of 0.95  or 0.99. Table 5 suggests that the methods are
divided into two distinct levels of performance.
     An estimate of potential performance based on the experimental and theoretical work per-
formed during the  program is presented in Table 6.  For many methods, as  discussed in [4], a
significant increase in performance can  be achieved by means of protocol changes alone. Of
course, the actual performance improvement would depend on the specific changes made by the
manufacturer.
Table 4. ESTIMATES OF MANUFACTURERS' PERFORMANCE IN TERMS OF PD and PFA
FOR DETECTION OF A LEAK RATE OF 0.38 L/h (0.1 gal/h) USING THE MANUFAC-
TURERS' DETECTION CRITERION

0.90 *
0.65 -
0.35 <
0.10-
PD

-------
      The temptation to use only those methods that were ranked highest in this evaluation
should be avoided for two reasons.  First, Table 6 suggests that with modifications most of the
methods should be able significantly increase their performance. Since more than a year has
elapsed since the evaluations were performed, many methods have made changes  that should
improve their performance.  Second, the quantitative estimates presented in Table 2 alone are
not sufficient to assess the performance of the method.

Table 6. ESTIMATE OF THE PERFORMANCE OF VOLUMETRIC TEST METHOD EVAL-
UATED AT THE UST TEST APPARATUS AFTER TWO LEVELS OF MODIFICATIONS
EXPRESSED IN TERMS OF THE SMALLEST LEAK RATE THAT CAN BE DETECTED
WITH A PD OF 0.99 AND A PFA OF 0.01

                                               Number of Methods Having
                                                This Detectable Leak Rate
Detectable Leak Rate
(L/h)
0.19 < LR < 0.57
0.57 < LR < 0.95
0.95 < LR < 1.32
Evaluation
Results
1
5
2
After Minor
Modification
6
13
After Protocol and
Equipment Modifi-
cations
12
7
                                       SUMMARY
      An important EPA-sponsored research program has been completed that has evaluated and
made estimates of the performance of commercially available volumetric leak detection methods
as they existed in the period March through July 1987.  The performance estimates assumed that
the test procedures were implemented by competent test crews using calibrated equipment.  For
each method evaluated, recommendations were made, as required, to improve performance.
This two-year project has determined and resolved key technological and engineering issues
associated with this general type of leak detection. The following objectives were accomplished:
(1) evaluation of the performance of 25 currently available volumetric systems for detection of
leaks in underground gasoline storage tanks; (2) generation of technical information important in
the development of EPA's underground storage tank  regulations; (3) development of specific
recommendations that will allow manufacturers to improve the current practice of each method;
and (4) development of basic information to assist the test users in selecting a method that meets
EPA's  new regulatory requirements for  underground  storage  tanks. A  summary of some of the
key conclusions of this research project are provided below.
     Current performance is significantly less than claimed  by most manufacturers. Of the 25
commercially available volumetric leak detection systems evaluated, most presently perform at a
level that is considerably  lower than the common industry claim of 0.19 L/h (0.05 gal/h).  There
are two reasons for this discrepancy between vendor  claims and actual performance. First, in
almost all instances, the measurements made by EPA during  this project  appears to be the first
systematic evaluation of the test method. Second, the performance estimates were presented in
terms of a probability of  detection and a probability  of false alarm, a format that most man-
ufacturers have not previously used to quantitatively  describe performance.
                                          96

-------
      EPA regulatory requirements for performance are met by five of the methods evaluated.
Many more should meet the requirements after modifications.  By and large, the leak detection
systems evaluated were limited by protocol and operation practice rather than by hardware. In
general, such limitations can be overcome by rather modest modifications to testing procedures,
such as waiting 3 h after topping the tank and before starting a test; major equipment redesign
is not necessarily required.
      Volumetric test methods should not be selected solely on the ranking contained in this
paper.  Because, as discussed in [4], the performance of many test methods can be significantly
improved with only minor modifications, some low-performance  methods are expected to meet
or exceed the level of high-performance methods in the immediate future.  As a consequence,
and because many methods have already incorporated many of the recommended changes, the
current ranking of test method performance implied by this paper is not particularly  significant
in the selection of reliable methods.
      Tank testing is complex, but a high level of performance can be achieved if several key
principles are followed. Those systems that did well in this evaluation had adequate spatial
sampling of the vertical temperature profiles of the product in the tank; incorporated adequate
waiting periods after product delivery and/or topping the tank (in tests that overfill tanks) to
allow the tank deformation and the spatial inhomogeneities in the product temperature field to
become negligible; maintained a nearly constant hydrostatic pressure head during the test; made
an  experimental estimate of the height-to-volume conversion factor; and used sound data analy-
sis  algorithms and detection criteria. Performance  of a test method suffered significantly when-
ever one of these aspects of testing was ignored or poorly implemented. In general, any method
will perform poorly and provide results that are difficult to interpret if it: (1) fails to maintain
a constant hydrostatic head during the test; (2) does not accurately estimate the  height-to-
volume conversion factor;  (3) tops the tank and begins to test almost immediately, or (4) waits
an  insufficient period of time after product delivery before beginning the test.  Most
manufacturers recognized  the need to wait after product delivery, but they did  not appear to
fully appreciate the magnitude of the degradation that occurs when the waiting period after
topping (in the overfilled-tank test methods) is not long enough.
      Reliable tank testing takes time.  The total time required for the methods evaluated at the
UST Test Apparatus to complete a reliable  tank test, from delivery of product to removal of the
equipment from the testing site, is generally 12 to  24 h.  The time required is controlled  by the
waiting periods after product delivery or after topping the tank.   The waiting periods can be
minimized by incorporating data analysis algorithms into the test protocol which identify the
point at which these two effects have become negligible.
      Two inevitable outcomes of this research project will be (1) rapid improvements in per-
formance based on changes implemented by manufacturers and (2) increases in test users' and
regulators' expectations concerning verification of  future performance claims for all volumetric
methods.  Manufacturers of many of the methods that were  evaluated by RREL have already
begun to make the changes necessary to improve their systems' performance and to verify the
new performance claims.

                                             97

-------
                                      REFERENCES
1.


2.



3.



4.




5.
6.
7.
8.
9.
10.
11.
12.
U.S. Environmental Protection Agency.  Part 280 —- Underground Storage Tanks; Pro-
posed Rules. Federal Register, Vol. 52, No. 74, 1987.
U.S. Environmental Protection Agency.  Part 280 — Technical Standards and Corrective
Action Requirements for Owners and Operators of Underground Storage Tanks. Federal
Register, Vol. 53, No. 185, 1988.

National Fire Protection Association. Underground Leakage of Flammable and Combus-
tible Liquids.  NFPA Pamphlet 329, National Fire Protection Association, Quincy, Massa-
chusetts, 1987.

Roach, Robert D., James W. Starr, and Joseph W. Maresca, Jr.  Evaluation of Volumetric
Leak Detection Methods for Underground Fuel Storage Tanks.  Final Report, Vol. I,
EPA/600/2-88/068a and Vol. II, EPA/600/2-88/068b.  U. S. Environmental Protection
Agency, Cincinnati, Ohio, 1988.

Roach, Robert D., James W. Starr, Christopher P. Wilson, Daniel Naar, Joseph W.
Maresca, Jr., and John S. Farlow.  Discovery of a New Source of Error in Tightness Tests
on an Overfilled Tank.  In:  Proceedings of the Fourteenth Annual Research Symposium.
Hazardous Waste Engineering Research Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Cincinnati, Ohio, 1988.

Maresca, Joseph W., Jr. A method of determining the accuracy of underground gasoline
storage tank leak detection devices. In:  Proceedings of the  Underground Tank Testing
Symposium, Petroleum Association for Conservation of the Canadian Environment,
Toronto, Ontario, 1982.

Maresca, Joseph W., Jr., Christopher P. Wilson, and Noel L. Change, Jr. Detection per-
formance and detection criteria analysis of the tank test data collected on the U.S. Envi-
ronmental Protection Agency national survey of underground storage tanks. Final Report,
Vista Research Project 2013.  Vista Research, Inc., Palo Alto, California, 1985.

Maresca, Joseph W., Jr., Noel L. Chang, Jr., and Peter J. Gleckler. A leak detection
performance evaluation of automatic tank gauging systems and product line leak detectors
at retail stations.  Final Report.  American Petroleum Institute, Vista Research Project
2022, Vista  Research, Inc., Mountain View, California, 1988.

Wilson, Christopher P., Joseph W. Maresca, Jr., Harold  Guthart, John A. Broscious,
Shahzad Niaki, and Douglas E. Spitstone.  A Program Plan to Evaluate Underground Stor-
age Tank Test Methods.  Vista Research Project 2011, Vista Research, Inc., Palo Alto,
California, 1985.

Starr, James W., John A. Broscious, Shahzad Niaki, John S. Farlow, and Richard Field.
An  approach to evaluating leak detection methods in underground storage tanks.  In:
Proceedings of the 1986 Hazardous Material Spills Conference.  U.S. Environmental Pro-
tection Agency, St. Louis, Missouri, 1986.

Starr, James W. and Joseph W. Maresca, Jr. Protocol for Evaluating Volumetric Leak
Detection Methods for Underground Storage Tanks. Technical Report, Contract No.
68-03-3244, Enviresponse, Inc.,  Livingston, New Jersey, and Vista Research, Inc., Palo
Alto, California, 1986.

Maresca, Joseph W. Jr., Robert D.  Roach, James W. Starr, and  John S. Farlow.  U.S. EPA
evaluation of volumetric UST leak detection methods.  In:  Proceedings of the Thirteenth
Annual Research Symposium.  Hazardous Waste Engineering Laboratory, Office of
Research and Development, U.S. Environmental Protection Agency, Cincinnati, Ohio,
1987.
                                           98

-------
              AN OUTREACH PROCESS;  CASE HISTORIES OF UNDERGROUND
                        STORAGE TANK CORRECTIVE ACTIONS

                   by:   William M. Kaschak, P.E.
                         Harold E. Lindenhofen
                         COM Federal Programs Corporation
                         Fairfax, Virginia  22033

                         Robert W. Hillger
                         Richard A. Griffiths
                         U.S. Environmental Protection Agency
                         Risk Reduction Engineering Laboratory
                         Edison, New Jersey  08837
                                 ABSTRACT

    The U.S. Environmental Protection Agency's (EPA) regulations for
underground storage tanks (USTs) require corrective action to be taken in
response to leaking USTs.  However, the level of experience of personnel in
the EPA regions, the states, and the local environmental agencies vary
widely.  EPA is expanding its previously developed Case History File
database to facilitate technology transfer among the personnel involved in
UST corrective action.  This information will allow UST personnel to obtain
the immediate benefit of the experiences of other people involved in UST
and other cleanup activities.

    The original Case History File contains reports filed by On-Scene
Coordinators (OSCs) and Remedial Project Managers (RPMs) about the
technical, administrative, financial, and other aspects of spill and/or
waste site cleanups that they have handled.  The File consists of a
database section and a narrative section.  The database section allows
menu-controlled computerized searches to be made in any of 27 categories.
The narrative section contains detailed reports on the response actions
taken at the site.  The narrative section is organized into 10 subsections:
General Information, Chemical Information, Effects of the Incident, Site
Characteristics, Containment Actions, Removal/Cleanup Actions, Treatment
Actions, Disposal Actions, and Operational Considerations.

    The original File has been modified to incorporate additional data
relevant to USTs, such as methods of detection, causes of UST leaks,
tank/piping construction, etc.  New reports are being added as the EPA's
Edison office receives them from the states and EPA regions.

    This paper provides an overview of the UST Case History File, describes
how the data were collected, analyzes the initial data including discussion
of various technologies used for UST site cleanups, and provides a sample
scenario of a database search for an UST-related incident.
                                    99

-------
                                 INTRODUCTION
     After  December  22,  1988,  the  U.S. Environmental Protection Agency  (EPA)
 regulations  (40  CFR 280)  require  corrective action for leaking underground
 storage  tanks  (USTs).   However, the level of experience among EPA, state,
 and local  response  personnel  vary considerably in this new fieldj and what
 constitutes  appropriate corrective action is not always clear.  Presently,
 no mechanism is  in  place  to facilitate  the transfer of technological
 information  among the federal, state and private communities concerning UST
 corrective actions.  To clarify this matter and to improve technology
 transfer among response personnel, EPA  expanded its previously developed
 Case History File (File)  database for hazardous material spills and waste
 site remedial actions to  include  information on UST corrective actions.

     The File is a component of EPA's Computerized On-Line Information
 System (COLIS), which is  maintained by  the Environmental Emergency Response
 Unit-Technical Information Exchange (EERU-TIX) contractor,  Enviresponse,
 Inc.^at EPA's Edison, New Jersey facility.   The intent of the File is to
 facilitate technology transfer among response personnel who need to select
 site-specific corrective actions.   The File is an easy-to-use informational
 tool that eventually will include a significant amount of case history
 data.

s    The UST File is an on-line computerized  system with a database section
 and a narrative section.  The database section allows  searches to  be  made
 using any combination of 27 different  criteria,  such as EPA Region, state,
 hazardous substance, hydrology,  UST construction,  corrective action
 technologies, etc.   The narrative  section of  the  File  contains detailed
 information in a plain-text format.  The on-line  system is  very easy  to use
 and files can be created in only a few minutes.

     Most  of the data in the database and the narrative sections  have  been
 obtained  from after-action reports submitted by federal/state On-Scene
 Coordinators  (OSCs)  or Remedial  Project  Managers  (RPMs).  The Case History
 File will allow the  review of  technical  information, such as  past  perform-
 ance,  cost, practicality,  reliability, and other  factors, to  facilitate  the
 selection of  cost-effective corrective actions.  The objectives  of the
 project were  to  expand  the existing database to accommodate information
 relevant  to UST  corrective actions, and  to collect case history data  on  UST
 corrective  actions involving a range of  site conditions, locations, and
 technologies.

                             CASE  HISTORY FILE

    The Case  History File's data are organized into two sections:  a
database section  and a narrative section.  The database section allows
searches by using key words.   The  user selects one of 27 available search
criteria shown in Table  1.  Upon selection of a search criterion, either a
                                    100

-------
new menu of key words appears  for  further selection or  the user is required
to input certain information.  An  example of  this would be for search
criteria 12, detection method.  The user could choose external detector,
internal detector, inventory records, sight,  smell, tank tightness,  taste,
or others.  In the case of criterion 19, depth to groundwater, the user
would be prompted for a range  such as 15 to 20 feet.

           TABLE 1.  DATABASE  CRITERIA OF THE CASE HISTORY FILE
          1 - Incident number
          2 - Date of incident
          3 - Date of report
          4 - Type of incident*
          5 - U.S. EPA Region
          6 - State
          7 - NPL rank
          8 - Site name
          9 - Chemicals*
         10 - Quantity
         11 - Origin*
         12 - Detection method*
         13 - Main effects
14 - Resources affected
15 - Area affected
16 - Population affected
17 - Topography
18 - Hydrology*
19 - Depth to groundwater*
20 - Annual precipitation*
21 - Ground materials
22 - UST construction*
23 - Site uses
24 - Containment*
25 - Removal/cleanup*
26 - Site treatment*
27 - Disposal*	
         *Search criteria of specific interest to UST OSCs and RPMs


    The narrative section of the Case History File contains detailed
reports in a text format and is organized into 10 subsections:  General
Information, Chemical Information, Effects of the Incident, Site
Characteristics, Containment Actions, Removal/Cleanup Actions, Treatment
Actions, Disposal Actions, and Operational Considerations.

    A user starts a search in the database section.  When the user
specifies -a search category and a key word or numeric value for that       e
category, the system creates a user's file that contains the incidents that
match the user's criterion.  The user may create up to 10 files that have
10 different criteria and may view the data for the incidents in any one of
these 10 files at any time.

                              DATA COLLECTION

    Data collection began with the 10 U.S. EPA regional UST coordinators,
who recommended states in which to begin the initial data collection
efforts.  A checklist was developed and sent to 28 states to collect
initial information on sites, and was used as a screening tool to reduce
the sites to a manageable number for the initial data collection efforts.
The checklist was tailored to ensure that the initial sites selected would
provide a good representation of diverse hydrogeological conditions,
environmental settings, geographical areas, and corrective action tech-
nologies.  Information on tank locations, leak quantities, hydrogeological
data, etc. was also contained in the checklist.
                                      101

-------
     After-action reports were sent to 24 states and the District of
 Columbia.  Site visits were made to 14 ongoing corrective actions in 7
 states.  A total of 50 after-action reports were collected from 16 states,
 the District of Columbia, and 7 local offices.

                    ANALYSIS OF DATA SUBMITTED BY STATES

     This section presents an analysis of the data collected during the
 initial project activities.  The objectives of the project were to focus on
 petroleum USTs and completed or ongoing corrective actions.  The infor-
 mation presented is limited to the narrative sections of the Case History
 File that emphasize UST technology transfer:  site characteristics,
 immediate corrective actions,  long-term corrective actions, free product
 removal, and operational considerations.

 SITE CHARACTERISTICS

     The site characteristics subsection includes  information that describes
 the site geology/hydrogeology and the various site investigation techniques
 that have been undertaken to quantify subsurface  conditions.   The site
 characteristics are critical to the corrective action selection process.
 The results of the survey pertaining to site characteristics are limited to
 a  discussion of the topic of field  sampling and analysis techniques.

     Hydrogeological studies of varying degrees of complexity were performed
 at 40 of the 50 sites.   The most  common technique used to  define the  extent
 of contamination consisted of  soil  borings,  installing monitoring wells,
 and groundwater sampling.   Modeling was used at only  4 sites  to assist  with
 plume identification.   Soil gas surveys were the  most  common  field
 screening technique that  was used at  7 sites as part  of the overall study.
 The complexity of  the  field studies ranged  from having 6 shallow soil
 borings to  installing  100 monitoring  wells.   Indicator compounds were used
 at 38 sites,  and most  of  the sites  used more than  one.   The predominant
 indicator compounds used  were  benzene,  toluene, xylene (BTX)  and benzene,
 toluene,  ethyl  benzene, and  xylene  (BTEX).

 IMMEDIATE CORRECTIVE ACTIONS

    The section on  immediate corrective actions contains information  on
 initial actions to  mitigate  the impact of a  sudden or  newly detected
 release.  Such  actions are usually  initiated within a  few hours  to a  few
 days  from the  time  the release is discovered, but may  take  from  a  few hours
 to  several months to complete.  Immediate corrective actions  focus on
 source  control.

    Immediate response actions were reported at 46 out of 50 sites.  The
predominant technologies used were contaminated soil removal  (at 17 sites)
and emptying the tank (at 11 sites).  Removal of free product, replacing
the piping, and removing the tank were implemented at 9 sites.  Additional
technologies used at several sites were tank testing,   tank excavation,
installation of monitoring wells,  pipe repair, explosivity monitoring, and
tank replacement.  The distribution of immediate corrective action
                                    102

-------
technologies reported  for  the 46 sites is presented in Table 2.  Immediate
corrective actions at  each site usually consisted of several technologies
used in combination.   Therefore, the  total number of incidents (115) is
greater than the  total number of sites reporting immediate corrective
actions (46).

            TABLE 2.   IMMEDIATE CORRECTIVE ACTION TECHNOLOGIES
           Technology
  Incidents Reported
           Remove Soil
           Empty Tank
           Remove Free Product
           Remove Tank
           Replace Pipe
           Test Tank
           Excavate Tank
           Repair Pipe System
           Install Monitoring Wells
           Explosivity Monitoring
           Replace Tank
           Other

           TOTAL
          17
          11
           9
           9
           9
           8
           6
           5
           5
           4
           3
          29

         115
LONG-TERM CORRECTIVE ACTIONS

    Long-term corrective actions are those actions undertaken to mitigate
the more extensive effects of contamination on public health and the
environment.  They are not clearly separable from the immediate actions
taken or the actions taken to recover free product that has been lost.
Long-term corrective actions are applied to both the groundwater and the
soil and have been reported for 41 of the 50 sites.

Groundwater

    Groundwater treatment was reported at 41 sites and involved 4 primary
techniques.  Each technology and the occurrence of that technology are pre-
sented in Table 3.  Long-term corrective actions for groundwater vary
greatly in complexity.

              TABLE 3.   GROUNDWATER TREATMENT TECHNOLOGIES
         Technology
Incidents Reported
         Air Stripping
         Activated Carbon
         Interceptor Trench
         Gravity Separator

         TOTAL
          23
          19
          11
           8
          61
                                      103

-------
     In certain instances, the groundwater was extensively treated with a
 well-planned series of steps.  Preliminary separation was accomplished
 either with a dual pump system or immediately after the recovery with an
 oil/water separator.  The effluent water was then treated with air
 stripping,  followed by carbon adsorption, and eventually discharged to the
 sewer system.  The applications,  as a general rule,  used a combination of
 technologies rather than a single technology.  A discussion of the various
 technologies is presented below.

 Air Strippers—Air stripping is a method of removing dissolved volatile
 organics,  such as BTEX compounds, from the water.   Air is forced over thin
 layers of water within a column packed with material to encourage the
 volatilization of contaminants from the water into the air.  Air strippers
 are the most commonly used technology.   Nearly half (23) of the unit
 operations  used air stripping.

 Activated  Carbon Treatment—Activated carbon treatment entails adsorption
 of  the organics present in the water or in the air in a stationary carbon
 bed.   The  carbon will eventually  become saturated  with organics and stop
 adsorbing any additional contaminants.   The carbon is then replaced.   The
 spent carbon is either regenerated or treated as a hazardous waste and
 disposed of.

    Activated carbon units were used at 19 sites.  At 8 of these sites,
 activated  carbon treatment followed air stripping  as either a  step to
 achieve maximum cleanup levels or to remove pollutants from the air
 discharge.

 Interceptor Trenches—Interceptor trenches are used  primarily  for  free
 product recovery.   The trench intercepts  the groundwater and free  product
 in  lieu of  pumping from wells for recovery.   The use of interceptor
 trenches is limited,  however,  by  groundwater depth.

 Gravity Separators—A gravity separator is a coarse  method of  separating
 free  product  from  groundwater.  The lighter free product  phase  floats  on
 the water and can  be removed  for  independent  disposal.

 Soil

    Like groundwater treatment, soil  treatment  technologies vary in com-
 plexity.  The major  difference appears  in  the  relative  number of techniques
 employed for  the corrective action.  Overall,  29 of  the 50 sites had some
 sort of  corrective action  listed  for soil.  All but  2  of  these used
 excavation  of contaminated soil as  either  the  primary  or secondary
 remediation tool.  Other corrective actions utilized  in  the soil treatment
were aeration, biological  treatment, and incineration.  The number of
 incidents using the  various technologies for soil treatment is shown in
Table 4.
                                      104

-------
              TABLE 4.  SOIL TREATMENT/DISPOSAL TECHNOLOGIES
          Technology
Incidents Reported
Excavation
Aeration
Biological
Incineration
TOTAL
25
10
3
2
40
FREE PRODUCT REMOVAL

    Free product removal generally refers to the recovery of product that
is located beneath the surface in large enough quantities for recovery by
mechanical methods.  Free product removal was reported at 42 of the 50
sites.  At 18 sites, free product removal consisted of several technologies
(2-3) used in combination.  Pump systems were utilized at 26 sites.
Recovery systems were reported at 14 sites, while trenches were reported at
5.  Additional devices used for free product removal include the vacuum
truck, skimmer, gravity separation tank, sorbents, and incineration.  The
length of time required to complete the actions ranged from as little as 6
hours to as long as 3 years.

OPERATIONAL CONSIDERATIONS

    This section of the report discusses operational considerations such as
permits, public involvement, administrative issues, and cost information.
The subtopics of major interest that will be discussed a.re permits and
costs.

    Twenty-six after-action reports indicated that some type of permit was
required.  The majority of required permits were National Pollutant Dis-
charge Elimination System (NPDES) or groundwater permits.  At 6 sites, air
quality permits were required.  There were no reported project slippages
due to the inability to obtain the required permit(s).

    Of the 50 after-action reports, only 19 provided any information on
site remediation costs.  The most expensive cleanup cost was $1.2 million,
and involved the installation of 14 wells in overlapping capture zones
where water was withdrawn and treated by an air stripper to remove volatile
organic compounds.  The payment of cleanup cost was predominantly borne by
the site owner (38 out of 39 reported instances).

                 SAMPLE CASE HISTORY FILE DATABASE SEARCH

    To illustrate the operation of the Case History File, the following is
a sample database search to demonstrate how an OSC would create files to
review the actions that other UST personnel have taken when confronted with
a similar incident.  The scenario is for a gasoline tank that has ruptured
in a drinking water aquifer recharge area and released an unknown quantity
of gasoline into the environment.  The OSC is interested in reviewing
                                     105

-------
 information on sites that have used either of two cleanup
 technologies—bioremediation or air stripping—for remediation.

     The OSC initiates the search by selecting UST incidents to create
 fileO.  He then selects the chemical, gasoline, to search and create filel.
 Files 0 and 1 are then combined to create file2, which is UST incidents
 with gasoline.  Figure 1 provides a graphical presentation of what is in
 file2 and how the OSC would complete a search of the Case History File and
 create specific files based upon the criteria selected.   The OSC begins to
 search by the primary areas of interest:  treatment technologies and
 hydrology of the site.   These searches result in files 3, 4, and 5 for air
 stripping, bioremediation, and aquifer recharge, respectively.   To narrow
 the field of files to review, the OSC would combine files using "AND/OR"
 logic.  In this scenario,  the OSC is interested in sites that have used
 either air stripping or bioremediation and combines files 3 or 4 to create
 file6.  The OSC further refines the field by combining fileS and file6 to
 create file?,  which would  be a site in an aquifer recharge area that has
 either air stripping or bioremediation as the treatment  technology.   In
 this scenario, only one site meets all of the search criteria,  the City of
 Farmington,  New Mexico  site, which actually employed both treatment
 technologies.

     At this point of the file search,  the OSC is ready to review the files.
 Upon the review of file?,  the OSC could review any of the files created and
 may want to review some or all of file6 to see where air stripping or bio-
 remediation had been previously used.   The OSC may also  desire  to review
 other sites  in fileS to see what other technologies may  have been used.
 The OSC could  also search  by additional criteria to create new  files to
 review.

     For any  of the files,  the complete after-action reports could be
 reviewed on  the system.  The initial  display  on  the computer screen  for the
 City of Farmington incident  would  be  a one-page  Abstract  of Incident as
 presented in Figure  2.   The  Abstract  of Incident  is  a summary of  the
 database information for the site.  After the Abstract of Incident is
 reviewed,  the  user may  then  review the narrative  sections,  which  provide
 more specific  details of the site  and  actions taken.

 CONCLUSIONS

     The  level  of  detail  and  the  questions  answered among  the 50 Case History
 File After-Action Reports varied widely.   Many of  the questions could not  be
 answered  from  the file material  alone.   Follow-up discussions on  the Case
 History File After-Action Reports were  required in numerous  instances.

     The responses  to the Case History  File After-Action Reports indicated
 that  the  information requested is  thorough in capturing the  information that
would  be useful in responding to leaking USTs.  The state personnel also
acknowledged their willingness to work with a computerized database system
and  indicated a need for such a system  to promote technology transfer and  to
educate new staff.  The Case History File was demonstrated at a national UST
                                    106

-------
SAMPLE DATA BASE SEARCH FOR A SPECIFIC UST
                INCIDENT
File 2
UST incidents with Gasoline
Circle Mobil, SC
Cftyof Farrnlngton, NM
City of Cypress, CA
Chevron, CA
Thompson Grocery, LA
Stadium Mobil, MA
Molan Oil Co., MO
Exxon Site, MD
Caltrans, CA
Week Brothers Mobil, IA
Wans Service Station, MA
Crossroads USA, IW
Boll GAS, MO
Texaco USA., FL
Southland, TX
Super America, MN
Armour Oil, CA
D.C. DES, WA DC
W.W. Service Center, AZ
Boron Gas Farmington, Ml
White Oil, CO
Total number of Sites
associated with a specific
file
/
/
4
4
4
4
4
/
4
/
4
4
4
4
•f
4
/
•f
4
4
4
21

Fil^3
Air
Stripping

/
4






4
4


4

4
4

4

4
9

File 4
Bio-
remediation

/
4



4

4







4




5

FileS
Aquifer
Recharge
4
4

















4

3

File 6
File 3 or
File 4

4
4



4

4
4
4


4

4
4

4

4
11

File?
File 5 and
FileS

4



















1

                  FIGURE 2
                    107

-------
 CASE HISTORY FILE
                                          Abstract of Incident
 Incident
 Region
 Site name

 Substance
 Quantity

 Origin
 Effects
 Resources
 UST const

 Geography
 Hydrology
 Ground
 Site uses

 Containmt
 Removal
 Treatment
 Disposal
     102                   Type :  UST     Inc.  date
       6                   State:  NM      Rpt.  date
     City of Farmington                   NPL rank

     Gasoline                             CAS #
     30,000 gallons                       DOT #

     UST installation,  error              Detection
     Soil and water contamination          Area
     Groundwater,  drinking water          Popula
     steel
15 May 85
02 Feb 88
N/A

71-43-2
1114

inventory
100 acres
5
     valley                               Precip    :  8.2  in.
     aquifer recharge                     GW depth  :  10 ft.
     sand,  gravel,  bedrock,  asphalt  surface,  concrete surfaced
     commercial,  residential,  rural,  parkland

     groundwater  control
     empty  tank,  replace  tank,  remove tank,  excavate  soil, exc
     biodegradation,  air  stripping,  groundwater extraction, vo
     land farming,  evaporation,  treatment, soil washing,  injec
               FIGURE 2.   SCREEN DISPLAY OF AN ABSTRACT OF INCIDENT
                          FOR THE CASE  HISTORY FILE FOR INCIDENT  102
workshop held  in  Santa  Fe, New Mexico  in November 1988, as one of several
tools presented for use by state personnel in managing their UST programs.

    The major  conclusions drawn from the analysis of the data and
information reported in the narrative  section of the Case History File are:
    o
    o
    o
    o

    o
    o
Time for cleanup varied from a few hours to years.
Indicator compounds are used to quantify extent of contamination.
Immediate corrective actions were implemented at 46 of the 50 sites.
Predominant groundwater treatment technologies were air stripping and
activated carbon.
Permits were required, but did not slow down the response.
Contaminated soils were usually excavated from the site.
    The Case History File provides an excellent forum for technology
transfer.  The initial data collection efforts of the UST incidents have
been well received by the state personnel.
                                     108

-------
      CONSIDERATIONS OF UNDERGROUND STORAGE TANK RESIDUALS AT CLOSURE
    by:  ¥arren J. Lyman
         Camp Dresser & McKee Inc.
         Boston, MA  02108
and    Anthony Tafuri
       U.S. Environmental
         Protection Agency
       Edison, NJ  08837
                                 ABSTRACT

    The U.S. Environmental Protective Agency (EPA) is currently evaluating
several technical and regulatory aspects of underground storage tank (UST)
closures.  A key concern is the manner and extent of tank cleaning that is
appropriate and feasible when a tank is removed from service.  The
objective of this work was to obtain a thorough technical/scientific
understanding of UST residuals at closure:  their origins,
physical/chemical properties and ease of removal by different cleaning
methods.  The information generated will be used as an aid in the
regulatory process and will be useful to those implementing/overseeing
closure activities.

    Information was obtained via phone contacts with knowledgeable
individuals including tank cleaning companies, published and unpublished
literature, site visits and worksheets completed by state UST program
representatives.  The investigation was limited to underground storage
tanks containing gasoline-and diesel.

    Gasoline and diesel USTs were found to have significant quantities of
residuals in them at closure, typically tens to a few hundreds of gallons.
However, although there is little explicit guidance available, tank
cleaning and removal companies are apparently capable of removing most of
these residuals with fairly simple cleaning techniques.  Objectives and
quantitative evaluations of the effectiveness of the cleaning techniques
are not possible at present because of the nearly complete lack of anything
but anecdotal data.
                               INTRODUCTION

    The magnitude of the environmental contamination problem - actual and
future potential - presented by underground storage tanks (USTs) has only
become clear in recent years.  Hidden leaks and unreported spills have
probably occurred at tens of thousands of USTs and would continue at
thousands more without intervention.  Federally mandated intervention will
                                     109

-------
come as a result of EPA's Final Rule published in the Federal Register on
September 23, 1988.  Many states and local communities have also passed
their own set of UST regulations.  The direct or indirect effect of these
rules will be the closure and removal - or abandonment in place - of
hundreds of thousands of USTs over the next 10-20 years.

    A key question related to UST closure involves the quantity and compo-
sition of residual liquids in the UST and the dangers to human health and
the environment that would result from their improper handling or release.

    In August, 1988, EPA's Risk Reduction Engineering Laboratory (RREL)
initiated a study to obtain basic scientific and technical information on
UST closure activities, with special attention on UST residual and cleaning
procedures used to remove them.  Specific tasks included:  (1) characteriza-
tion of UST residuals (e.g., quantities and composition); (2) an evaluation
of removal and disposal alternatives for UST residuals; (3) an events study
of UST closure activities using the Darning management approach; and (4) a
sampling and analysis of USTs before and after cleaning operations.

    This paper covers only work carried out in support of the first two
tasks.  The approach taken to obtain desired information included:

    o  Phone contacts with over seventy knowledgeable individuals
       including tank cleaning companies

    o  Observations of four tank cleaning/closure operations

    o  Review of the limited amount of published information available

    o  Review of the answers to a worksheet/questionnaire filled out by
       state UST program representatives (and others) at an UST workshop
       (total of 23 responses)

    o  Carrying out a few simple engineering calculations related to
       residuals generation and closure costs


                      INVESTIGATION OF UST RESIDUALS

QUANTITY

Overview

    Underground storage tanks containing either gasoline or diesel can
usually be emptied to. within 4-6 inches (10-15 cm) of the tank bottom.
This distance usually dictates residual quantity; e.g., for a 10,000-gallon
tank, this distance translates into about 100-200 gallons* of materials
*Because gallons  is a more common measure of volume and UST capacity,  these
 units are used in this paper.  For conversion  to metric:   1 gal = 3.785
 liters.
                                     110

-------
remaining on the tank bottom.  Much larger amounts are frequently found,
usually associated with abandoned tanks or with leaking tanks into which
groundwater has flowed.

Information from Phone Survey

    During the phone survey  , a total of 18 estimates was obtained
regarding the volume of residuals for an average 10,000-gallon tank,
containing either gasoline or diesel.  Additional estimates of fuel oil
residuals, for comparison purposes, and rinseates (resulting from cleaning
operations) were also obtained.

    Gasoline — The volume of residuals found in gasoline tanks at any one
site can vary significantly, from 0 to 10,000 gallons.  While these ranges
are in fact possible, they are not typical.  On the average, the twelve
sources reporting volumes of gasoline UST residuals quoted residual
quantities from 0 to 1,000 gallons as typical.  The mean of the values
reported was 160 gallons, while the median of the values reported was 75
gallons.

    Diesel — Most respondents qualitatively agreed that diesel tanks
contained more residuals than gasoline tanks, all other things being equal.
However, the six quantitative responses of volume of diesel residuals
obtained during the survey ranged from 0 to 200 gallons, with a mean value
of 58 gallons.  The median value was slightly higher and similar to that
estimated for gasoline tanks: around 75 gallons.

    Fuel Oil — Most respondents qualitatively agreed"that fuel oil tanks
produced the greatest amount of residuals, in comparison with gasoline and
diesel oil tanks.  The two quantitative responses of volume of fuel oil
residuals obtained were 500  and 1,000 gallons, averaging to 750 gallons,
significantly higher than gasoline and diesel.

    Rinseates — As one would imagine, the volume of rinseates generated
during the cleaning procedures can vary widely with the type of cleaning
procedure used.  The three phone-survey values of rinseate volumes ranged
from 100 to 3,300 gallons, with an average of 1,200 gallons.  Again, as one
would expect, these volumes  are significantly higher, in fact one order of
magnitude higher, than the residuals themselves.  As is noted below, the
American Petroleum Institute's (API) Recommended Practice 1604 calls for
the tank to be filled nearly to the  top for cleaning and/or vapor removal
purposes.   This practice would generate much greater volumes of rinseate.

Detailed Information from One Company

    In connection with an UST sediment characterization project for the
State of Minnesota, Delta Environmental Consultants, Inc., examined the
files of one  tank cleaning and removal company that kept detailed records
of the depth  and volume of residuals in each UST removed.   The range of
residuals volumes found for  gasoline, fuel oil and waste oil tanks is shown
in Figure 1.  A statistical  summary  is provided in Table 1.
                                     Ill

-------
        Ill
        g
        Q
        CO
        Ul
        CC
        Q
        LU
        Q

        j

        2
        U.
        O
        CC
        LU
        CO
                6   20 " 40  60 "  80 " 100 "  140  180  250  400  600 800   >1000
                                               200  300            1,000
0   20   40   60  80  100  140  180  250  400  600  800   >1000
                               200  300            1,000
              10
                 0   20  40   60   SO   100  140  180  250  400  600  800   >1000
                                               200   300           1,000

                      VOLUME OF RESIDUALS FOUND IN TANK (gal)
                          (NOTE CHANGES IN SCALE AT 100,200,300 AND 1000 gal)
Source: Delia Environmental Consultants (1988)
      FIGURE 1.  QUANTITY OF RESIDUALS FOUND IN UST's BY ONE
                 MINNESOTA COMPANY (9-1-87 TO 8-30-88)
                                     112

-------
            TABLE 1.  QUANTITY OF RESIDUALS FOUND IN USTs BY ONE
                 MINNESOTA COMPANY (9-1-87 to 8-30-88)*


Number of Tanks
Avg. Tank Capacity (gal)
Avg. Residuals Volume (gal
Median Residuals Volume (gal)
Gasoline
214
5,800
49
-20
Fuel
Oil
221
5,900
81
-40
Waste Oil
151
3,600
162 (93**)
-50
 information derived from raw data in reference 3.
**Excluding one tank with 9,375 gallons of residuals.

ORIGIN, NATURE AND COMPOSITION

Overview

    Based primarily on anecdotal data and some rough calculations, it is
estimated that 70-90% of the gasoline and diesel residual consists of the
product itself, probably of somewhat diminished purity.  The remaining
10-30% consists mostly of water (with numerous dissolved constituents);
product-related residuals (e.g., gum, sediment, tars); rust and scale (in
steel tanks); dirt and other foreign objects; and a small, but
disproportionately- important mass of microorganisms.  The importance of
the microorganisms comes from the significant internal corrosion that can
be due to the action of sulfate-reducing bacteria.

Location in Tank

    The typical location of residuals within an UST is diagrammed in Figure
2.  While most of the residuals will reside on the bottom of the tank, the
presence of some side-wall scale and gum has also been observed.  The
bottom residuals, while containing some gum, scale and grit, are usually
pumpable liquids, but might properly be considered sludges with low solids
content.  The settleable solids are frequently found pushed slightly to the
side of the UST (e.g., to the 5- and 7-o'clock positions as viewed end on)
due presumably to turbulence during filling operations.  Finally, in tanks
that are installed with a slight end-to-end tilt, as is proper, a greater
fraction of the residuals will be found at the low end of the tank.

Composition of Solids/Sludge Fraction

    Excluding larger debris (e.g., broken dip sticks, beverage cans, rubber
hoses), the solids/sludge content of UST residuals is probably mostly dirt,
rust (tank scale), and high molecular weight organic material (e.g., tars
and gums).  Visual observation of samples scraped from the bottoms of USTs
during manual cleaning - after all easily pumpable material had been
removed - indicated about a 50-percent solids content.  Some data from
analyses of such samples for benzene, toluene, ethyl benzene, xylene and
total hydrocarbons are shown in Table 2.
                                    113

-------
                  Underground Storage Tank with most Product
                  Removed in Preparation for Closure
                                          Residual Fuel
                                           (approx. 4 to 6 inches)
Sediment, grit, gum
(dirt, rust particles, or fuel sediment)    Water Layer
Thickness of sludge may be enhanced   (probably < 1 inch)
at 5- and 7-o'clock positions in vicinity
of fill tube.
    FIGURE 2. SCHEMATIC OF UST RESIDUALS
                            114

-------
      TABLE 2.  BENZENE, TOLUENE, ETHYL BENZENE, XYLENE, AND
                TOTAL HYDROCARBONS IN THE  SOLID/SLUDGE  PORTION
                         OF UST RESIDUALS3

SarnpL
Numbei
MDLC
1
2
2d
3
4
4d
5
6
7
7d
8
Parent
* Material .
c of Sample B<

Gasoline
Mixed Oil
Mixed Oil
Mixed Oil
Mixed Sludge
Mixed Sludge
Mixed Sludge
tt6 Fuel Oil
tt2 Fuel Oil
#2 Fuel Oil
Dried Residual
Concentration (rag/kg)
snzene
0.12
110
5.1
4.9
39
1.9
1.5
190
1.0
3.8
4.0
7.1
Toluene
0.12
270
11
10
100
12
8.4
310
2.4
11
10
160
Ethyl
Benzene
0.12
30
1.8
1.8
13
6.2
4.4
44
1.0
1.9
1.7
41
Total
Xylene Hydrocarbons
0.12
140
8.8
8.9
67
19
13
210
6.0
9.4
8.3
210
1.0
1700
120
110
800
270
180
2400
86
109
110
1800
 Data from reference 3.

 Sample Descriptions;

        1.  Sludge at bottom of gasoline storge tank
   2. & 3.  Drying mixed oil residuals in two different tanks
   4. & 5.  From mixed sludge drums
        6.  From drum of #6 fuel oil residuals
        7.  From drum of #2 fuel oil residuals
        8.  Composite of dried residual from a number of open tanks

CMDL = Method Detection Limit

 Duplicate analysis                                      ,
                              115.

-------
Water Content

    The phone survey  also provided significant evidence for the presence
of a layer of waiter at the bottom of many, if not most, USTs.  It is a
common practice for USTs in service, for example, to check for the presence
of water (and sediment) with a dip stick prior to refilling the UST.  The
end of the dip stick is coated with a special paste that changes color on
contact with water.  A rule of thumb for some gas stations is to limit the
depth of water to about one inchj greater amounts are pumped out prior to
refilling the tank.  (One inch of liquid in a 10,000 gallon tank represents
about 12-18 gallons.)  At stations where frequent monitoring was not
undertaken, or where water input rates were high, the volume of water at
the bottom of the tank could clearly be much higher.  Also, abandoned tanks
are frequently found to have large volumes of water due to precipitation
runoff into open fill tubes or groundwater leakage in through holes formed
by corrosion.

    Residual water would contain a significant amount of dissolved-
hydrocarbons (~100-300 mg/L), dissolved salts (e.g., Na , Cl  , Fe   , HC03~,
Pb  ) and other soluble components or additives present in the fuels (e.g.,
ethanol, methyl-t-butylether (MTBE), detergents).  The composition  could
lead to such water being classified as a hazardous waste and  to a
requirement for pretreatment prior to discharge to any sewer.

    There is one mechanism by which water can accumulate in USTs solely by
inputs dissolved in  the product delivered to the  site.  The water present
in solution in the product delivered to an UST is likely to be near the
solubility limit and, in summer at least, warmer  than ground  temperatures.*
As the fuel enters the UST and cools down, the solubility limit is  lowered
causing some water to come out of solution and form, or add  to, a separate
aqueous phase.

    The order of magnitude of  the contribution of this phase  separation
process is estimated to be one gallon of water per  tank refill for
gasoline USTs.  Water is also  suspected  to enter  USTs via entry of  moist
air through  the open fill  tube followed  by condensation.

    This UST water may play  a  significant role in the  internal corrosion of
steel  tanks.  Several surveys  have  shown that  internal corrosion of steel
tanks  is a fairly  common occurrence  although external  corrosion is  roughly
three  times more important.   Significantly different mechansims may be at
work in internal and external  corrosion,  although the  presence of water  is
probably necessary in both cases.   For  internal  corrosion,  this water may
be present as a  condensate on  the  tank walls or  as  a layer  on the bottom.
 *The solubility limits for water in gasoline and in diesel are not known
  precisely but are probably on the order of 1,000 mg/L and 100 mg/L,
  respectively.  Significantly larger amounts might be present in solution
  if the fuels have hydrophilic additives such as ethanol or MTBE.
                                     116

-------
Tank Rust or Scale

    The phone survey and a limited literature review provided evidence that
steel tanks are likely, over time, to shed rust particles (iron oxide,
Fe2°3^ an<* *ron scale on the ir»side.  This internal corrosion may be caused
by galvanic action or bacterial action (see subsection on Microorganisms).

    Some of the rust and scale may remain on the tank walls while portions
will drop and accumulate on the bottom.  The total volume of side and bottom
scale is thought to be relatively small, perhaps no more than one liter.

    A rough estimate of the amount of rust that might accumulate in an UST
can be derived from a calculation in which 0.1% of the mass of the steel
tank is assumed to be converted from Fe to Fe^O,,.  This leads to an
estimate of about 9 kg (20 Ib) of rust in a 10,000-gallon tank.

Microorganisms     <                                                .

    Like water, microorganisms appear to be fairly ubiquitous in petroleum
storage and distribution_systems.  While they may appear to be present in
large numbers (10  to 10  organisms per liter), their mass is small.  At
times, however, large floes can be formed which can clog fuel lines and
fuel filters.

    The microorganisms involved include several varieties of bacteria and
fungi.  Of special import are the class of sulfate-reducing bacteria that
can cause significant corrosion to iron and steel products.  These bacteria
are strict anaerobes that perform anaerobic respiration by oxidizing
certain organic compounds or H2, and reducing sulfate, and often other
reduced sulfur compounds, to hydrogen sulfide.  The sulfide can then react
with iron to form an iron sulfide precipitate (FeS) which can contribute to
the solid portion of UST residuals.

    Corrosion is usually evidenced as pits below microbial mats.  There are
numerous theories as to the biochemical and chemical basis for the
corrosion, but no current agreement on any one.  No data were found on
rates of microbial corrosion to be expected in USTs.

    Microorganisms do need water to thrive and, in storage tanks, are
usually found at the fuel-water interface.  The mix of hydrocarbons, water
content, oxygen content (low for anaerobes), nutrient content and pH are
all important factors in the growth of these microorganisms.  They
apparently thrive better in kerosene and fuel oils than in gasoline.

Ignitability
                                                                   o
    Delta Environmental measured the flash points of UST residuals.   In
general, their results indicated that gasoline sludge, scale and dried
residuals, and mixed sludge had flash points near 50-60°F and can be
flammable.  Residuals from higher boiling fuels (e.g., sample nos. 2, 6 and
7 in Table 2) had flash points above 140°F, the point above which materials
are not considered hazardous because of ignitability.
                                     117

-------
EP Toxicity

    Delta Environmental also evaluated the EP Toxicity of the UST
residuals.   In this procedure, concentrations of eight heavy metals in an
aqueous extract of the waste are compared with specified limits.  EP
Toxicity analyses were below regulatory limits for all metals except for
one lead sample.
                           REMOVAL AND DISPOSAL
CLOSURE PRACTICES
    Phone survey and analysis of worksheet questionnaires (from the EPA/UST
Workshop ) showed that tank removal, involving some form of tank cleaning,
was most commonly practiced.  Second was closure in place, which involves
filling the tank with an inert material (e.g., clean sand) after removing
the residuals.  In a few instances, tank closure involved a change in
service for the tank (i.e., a change in the liquids stored in the UST).
The following sections cover only closure in which tank removal is a part,
and focus on the cleaning practices.

CLEANING PROCEDURES

Procedures Used

    A variety of tank cleaning and removal procedures appear to be in use,
although many are variations of a simple, logical theme.  Many of the steps
in these procedures are dictated by safety considerations* and by state and
local regulations, rather than by a direct concern for strict tank
cleanliness.  Most procedures involve an initial pumping of residuals with
a suction line and a subsequent rinse with water followed by rinseate
removal.  The water rinse may involve:  (1) filling the tank with water;
(2) rinsing with spray from a "garden" hose [low pressure]; (3) rinsing
with high pressure water; (4) steam hosing; and (5) possible use of a
detergent.  The American Petroleum Institute's recommended procedures (API
1604)  call for filling the tank with water followed by sequential removal
of floating product and water.

    Several tank cleaning companies cut a manhole into the UST, allowing a
man to enter and physically remove bottom grit and (with a "squeegee")
liquids adhering to the side walls.  Some companies consider this procedure
too dangerous, especially for gasoline tanks; the practice is prohibited in
some areas.  A summary of the basic UST cleaning and removal steps in the
API 1604 procedure, and in other procedures described in the phone survey,
is provided in Table 3.
*Prevention of human exposure to  toxic chemicals, fires and explosions, and
 spillage.
                                     118

-------
     TABLE 3.  SUMMARY OF UST CLEANING AND CLOSURE PROCECURES
SOURCE*

A,B,C
API 1604**
                       COMMENT
  E
   H
 1.   Prepare workers and area for safe operations.
 2.   Drain product piping into tank; also cap or
     remove product piping.
 3.   Remove liquids and residues from tank.
 4.   Excavate to top of tank.
 5.   Remove tank piping, pumps and other fixtures.
 6.   Purge tank of flammable vapors.
 7.   Fill tank with water until floating product
     nears the fill opening; remove floating product.
 8.   Pump out water.
 9.   Test tank atmosphere for flammable or
     combustible vapor concentrations.
10.   Plug or cap all accessible holes except 1/8"
     vent hole.
11.   Excavate and remove tank.

 1.   Empty tank as much as possible.
 2.   Triple rinse, with high pressure water (gasoline
     tank) or detergent (diesel tank).
 3.   Inert tank with NZ or C02-
 4.   If necessary, enter tank and physically remove
     sludge.                 2
 5.   Punch 6 holes, each 1 ft , to render tank
     useless.
 6.   Remove tank from ground.

 1.   Empty tank as much as possible.
 2.   Purge tank.
 3.   Cut opening(s) in tank.
 4.   Rinse, pump out rinseate.
 5.   Remove tank from ground.

 Proprietary process involving pumping fuel out of
 tank, filtering through vacuum, spraying fuel back
 into tank through nozzle, pumping, filtering, etc.
 May take numerous cycles  to clean  tank.

 Warm water and detegent used as rinse agent;
 high-pressure not used because of  safety hazard.

 1.  Empty tank as much as possible.
 2.  Inert with C02.
 3.  Cut opening in tank.
 4.  Worker enters  tank, physically removes any
     sludge or scum.
 5.  Remove tank from ground.
                               119

-------
 TABLE 3.   SUMMARY OF  UST  CLEANING AND CLOSURE  PROeEDURES (Cont'd)
SOURCE*
                       COMMENT
  K
  H
               1.  Empty  tank as much  as  possible.
               2.  Purge  tank.
3.
4.

5.
6.

1.
2.
3.

4.
5.

1.
2.
3.
4.
5.
6.

1.
2.
3.
4.

1.
2.
3.
4.
                  Cut 2-ft  opening  in  tank.
                  Worker enters  tank:   squeegees  side and  bottoms;
                  scrapes sides  and  bottoms; washes with water.
                  Rinseate  is  pumped out.
                  Tank  is removed  from  the ground.
Empty tank as much as possible.
Purge tank.
If tank has a manhole:  rinse with caustic
pH) detergent.
Pump out residuals and rinseate.
Remove from ground, lay on its side.

Empty tank as much as possible.
Inert with C02*
Cut opening in tank.
Physically clean residuals.
Inert tank.
Remove from ground.

Empty tank as much as possible.
Triple rinse with high pressure steam.
Inert with C02.
Remove from ground.

Empty tank as much as possible.
Remove from ground.
Cut manhole in tank.
Worker physically removes residuals.
                                                              (high
 *Each letter (A, B, C...) represents a different source that was
  interviewed in the phone survey.  Most of the soures are tank
  cleaning and removal companies who are describing their own
  standard procedures.  In two instances, the procedures are those
  specified by a county agency.
                                               P
**Basic steps in API Recommended Practice 1604.   Several details
  relating to safety and regulatory compliance have been omitted
  for brevity.
                               120

-------
    With some companies, it is common to put both initially-pumped
residuals and aqueous rinseate into the same tank truck (for off-site
treatment and disposal).  Other companies segregate the residuals from the
rinseate thus facilitating subsequent product recovery and/or treatment.
The volume of rinseate generated appears to range from a low of 100-200
gallons per tank to about one third of the tank's volume, except for the
API 1604 procedure, which calls for filling the tank with water.

Cleaning Effectiveness

    There are no data which provide objective evidence of the degree of
cleanliness achieved by the procedures used.  For tanks that are
subsequently reused as scrap metal (i.e., crushed or cut-up and then
remelted), a modest amount of retained residuals may be environmentally
acceptable.  Worker protection may be the more stringent basis for
regulation.  For tanks that are filled in place or landfilled, the
residuals remaining after cleaning operations (retained residuals) are
likely to pose only a small-to-negligible risk of adverse environmental
impact.  This would be related to the small volume of retained residuals,
limited environmental mobility for most constituents, and limited
toxicological significance for the bulk of the constituents.

    In practice, tank cleanliness (i.e., the absence of sludge, scale,
sediment and liquid product or rinse water) is "determined" by a number of
methods including:

    o  Visual inspection of the tank

    o  Visual inspection of a wipe sample

    o  Analysis of the cleaning rinseate

    o  Adherence to a standard cleaning procedure (e.g., API 1604 or the
       cleaning company's own standard procedures)

    The frequency of inspection (of cleaned tanks) by state or local
officials appears to be quite low.

TREATMENT AND DISPOSAL OF SECONDARY WASTES

    Secondary wastes from UST cleaning operations fall into several groups:

    o  "Pure" residuals removed from the UST (and stored separately) before
       any cleaning is undertaken; this is expected to be 70-90% fuel
       product (i.e., gasoline or diesel)

    o  Liquid/sludge residuals;  What is left after "pure" product is
       removed; may also have enhanced solids concentration if tank walls
       were manually scraped

    o  Rinseate;  usually resulting from a simple water wash sprayed into
       the UST.  Detergents may be used
                                      121

-------
     °  Combined wastes;   Some cleaning companies pump the pure residuals
        and rinseate into the same tank truck at the UST site

     o  Debris;   including large items found in tank as well as discarded
        cleaning rags and protective clothing used during cleaning

     Very little data were obtained on the quantity of secondary wastes
 beyond  what is  known about the amounts of "pure" residuals usually found in
 USTs (approx. 100 gal for 10,000-gal tank).   The amount of wash water used
 varies  widely from as little as 100 gal to amounts approximating one  third
 of  the  tank volume.   If  the API Recommended Practice 1604 is adhered  to
 strictly,  a nearly full  tank of wastewater would be generated.

     No  reliable data were obtained on the composition of these wastes
 beyond  what was provided in Table 2.   It  is fairly easy to speculate,
 however,  given  the general understanding  of the composition of "pure"
 residuals and the extent of dilution by water in the cleaning process.   A
 "combined waste",  for example,  might roughly be a 50/50 mixture of fuel  and
 water;  detergent would be present if used in the cleaning process.  Without
 scraping and thorough pumping,  a significant fraction of the side scale  and
 gum,  and bottom sediment,  might not be transferred from the UST to the
 secondary wastes.  All waste groups are likely to have emulsion
 characteristics,  i.e., small droplets of  one phase dispersed throughout  the
 other phase.  The use of detergents would increase the degree of
 emulsification.

     Very little information has been obtained on actual treatment  and
 disposal of UST secondary wastes.   However,  an initial separation  of
 hydrocarbon liquids  from aqueous phases (in  a large settling tank)  is
 common.  The aqueous  phase often ends up,  after varying degrees  of
 pretreatment, in a sanitary sewer leading to a municipal biological
 treatment  plant.   The hydrocarbon phase is  often treated as  a waste oil
 (i.e.,  shipped  to  an  oil refiner),  burned for its heat  content,  incinerated
 or,  less  likely,  drummed and landfilled.

     Most states  providing information on  the worksheets  at  the EPA/UST
Workshop  indicated that  UST residues  and  cleaning by-products were
 considered  hazardous  wastes  and thus  had  to  be handled  according to
 appropriate state  and Federal rules  for such material.


                                 CONCLUSIONS

    This study has shown that there are significant amounts  of residuals
left  in USTs at  the time closure  is  initiated. The handling  and  removal of
 these residuals during closure  -  and  the  resulting  tank  cleaning - are
being carried out  by  a wide  variety of  procedures often  dictated more by
preferences of local  officials  or  the selected contractor  then by objective
guidelines.  The cleaning  procedures  typically generate  a significant
volume of aqueous  rinseate  that also  presents disposal problems.
                                     122

-------
    A significant fraction of the residuals consists of recoverable fuels,
and fuel recovery is practiced by a number of tank cleaning contractors.
The residuals (before or after removal of recoverable fuels) are
potentially hazardous as a result of both ignitability and toxicity
characteristics, yet in some states the material is classified as a waste
oil and not a hazardous waste.

    The tank cleaning methods currently in use appear to be able to
satisfactorily clean most gasoline and light oil USTs.  However, no data
exist indicating just how clean the tanks do get, no standard or generally
accepted method of tank cleaning has been identified, nor have any criteria
been developed for tank cleanliness for any of the various tank disposal
alternatives.

    Monitoring of selected tank cleaning methods, quantitative measurements
of the amounts of residuals left in the tank after cleaning, and a
characterization of the rinseate generated are planned for the next
portions of this study.
                                REFERENCES
1.  The phone survey data cited in  this paper are documented in  the
    following report:   "Evaluation  of  the Technical Aspects of UST
    Closure," Interim Report, EPA Contract No. 68-03-3409, Work  Assignment
    No. 16,  September,  1988.

2.  "Removal and Disposal of Used Underground Petroleum  Storage  Tanks," API
    Recommended Practice 1604, second  edition, American  Petroleum
    Institute, Washington,  D.C., 1987.

3.  Delta Environmental Consultants, Inc. (St. Paul, MN),  "Tank  Sediment
    Characterization and Disposal Report," report to Minnesota Pollution
    Control  Agency, St. Paul, MN, November,  1988.

4.  "Making  it Work:  Workshop for  State Tank Program  Managers,"
    Conference/workshop sponsored by the U.S. EPA's Office of Underground
    Storage  Tanks,  held in  Santa Fe, NM, November 15-17,  1988.
                                     123

-------
 LABORATORY STUDIES OF VACUUM-ASSISTED STEAM STRIPPING OF ORGANIC CONTAMINANTS
                                  FROM SOIL

        by: Arthur E. Lord, Jr., Robert M. Koerner,  and  Donald E. Hullings
           Geosynthetic Research Institute
           Drexel University
           Philadelphia, PA  19104
           and
           John E. Brugger
           Risk Reduction Engineering Laboratory
           Superfund Technology Demonstration Division
           U.S. Environmental Protection Agency
           Edison, NJ  08837
                                  ABSTRACT

     A  long-term  research   and  demonstration  project   is  underway  to
investigate vacuum-assisted,  steam  stripping  of  organic  chemicals  from
contaminated soil and to develop a field unit to steam strip and collect such
pollutants.  In previous work, an analytical  model  was developed  for steam
stripping  in the field. The model  involved  steam flow to the surface from
pipes  embedded below the ground  surface.  The  data  needed to implement  the-
model  were obtained  from experiments  that  involved vacuum-assisted steam
stripping  of kerosene from a  variety  of soil types  (from  sands to  silts)  in
small  scale  laboratory  experiments. This  approach was used  to determine  the
time  to decontaminate  a  given kerosene spill in a particular soil.  Small
scale  pilot studies were  also made  of the  field  unit  employing a  unique
geosynthetic,  vacuum cap assembly. The results  of  this first phase  showed
that kerosene  (and also gasoline)  could be quite  effectively removed  from a
wide variety of soils with the vacuum-assisted, steam stripping technique.
Also,  the  geosynthetic  cap  (geotextile plus  geomembrane)  anchored  in  the
soil,  performed  quite  well   in  confining  and collecting the  steam  and
contaminant.

     The chemical  analysis method  in the first  phase of the  work used  simple
volume separation  of  the  collected kerosene  and  water fractions.  Gas
chromatography  (GC)  is  presently employed to analyze the residual chemical
content in the soil after steam stripping.   The efficiency of vacuum-assisted,
steam  stripping of alkanes (octane,  decane  and dodecane)  from  a  range  of
soils  from sands  to silts,  was determined. Octane,  which has a relatively
high vapor pressure  (b.p. = 126°C) could be reduced from an initial content
of  5-10%,  to  less  than 10  ppm in  a few  hours.  Dodecane,  which  has  a
                                    124

-------
relatively low vapor pressure (b.p.  =  216°C) was reduced to about 1000 ppm in
the same time frame. Decane which has  an intermediate vapor pressure (b.p.  =
174°C)  was  reduced  to  80  ppm.  Octanol,   which  has  vapor   pressure
characteristics between  decane  and dodecane (b.p. =  195°C) and  is  somewhat
polar,  was   reduced  to  10  ppm.  Butancl,   which  has  vapor   pressure
characteristics similar  to  octane (b.p. = 118°C)  and is  strongly polar  was
reduced to lower than 1 ppm.

    As  a  result  of  this  work,  the  utility  of vacuum-assisted,  steam
stripping methods to decontaminate soils at Superfund  (and other) sites is on
a somewhat firmer technical  basis  than  before.  This  paper identifies future
research work,  such  as  determining  the  exact nature  of  the relation between
compound  vapor pressure and polarity  on the  ability to  steam strip  the
compound from the soil.

                          INTRODUCTION AND  OVERVIEW

    With  contaminated soils at Superfund (and  other)  sites, it is  important
that  the chemicals be prevented from reaching the groundwater.  Fortunately,
in many locations the  partially  saturated  or vadose  zone  exists and acts as
temporary containment  retarding  the downward movement of  the  pollutant.  The
remediation options are:

     • Excavation and off-site disposal
     • Excavation and on-site treatment
     • In-situ treatment (via a  number of possible methods, e.g., biological,
       physical or chemical)

    A number of these techniques  (and  others)  have  been  reviewed in recent
articles  (1,2). The in-situ techniques have  been discussed  by.the authors
 (3) .

    The present study  falls in the in-situ treatment  category wherein  the
authors  propose  to  have pipes  inject  steam into  the  soil beneath  the
contaminated zone. Steam stripping of  the  chemical  occurs  and when aided by a
vacuum  at  the ground surface brings the contaminants to  a collection point
where they can be  properly treated.  A  unique  aspect of the study is  a
geosynthetic cap assembly consisting of a high transmissivity geotextile and
a  flexible membrane  liner (geomembrane).  The  vacuum is applied to  the
underside of this liner  and  the  contaminated gas and/or liquid moves beneath
the liner  in the geotextile to  the outlet ports. A  schematic  diagram of a
proposed system is given in Figure 1 for reference  purposes.

    There have been three  steam stripping,  soil decontamination studies
reported  in  the literature, as far as  the authors are  aware  (4,5,6).    [One
company mentions steam  stripping  in their advertisements for hazardous waste
site  remediation  (7).]  The field works  are  site  specific  projects,  and
understandably make  no attempt  to look  at  the general problem  of  steam
stripping a wide variety of chemicals  from a wide  variety  of soils.

     The present work involves  a long term study to determine  the ability of
vacuum-assisted,  steam stripping to  decontaminate general organic  chemical
                                    125

-------
 species from a variety of soil types.

                         PREVIOUS WORK ON THIS PROJECT

     The  work performed in the first phase of the project  has  been reported
 elsewhere (3,8,9). Only  a  brief  review of the results of  this work will  be
 given here.  Among the tasks undertaken, in a wide variety of soils were:

     •  Observations  were made of the transient steam  front movements in two
        dimensional flow.  (The sample size was 2 ft x  2 ft  x  1  in.  and steam
        at about 5  psi pressure was used.)

     •  The   ste,am permeabilities  were  determined   in  conventional  one-
        dimensional flow.  Results  are  given in Table 1. (The  sample  size was
        2.5 in. diameter X 6 in. high.)

                   TABLE 1.   STEAM AND WATER PERMEABILITIES

 Soil                 Standard Water  Steam Permeability  Steam Permeability
                      Permeability                        Water Permeability
 Sand (%)  Silt (%)

   100         0

    75       25

    50       50

    25       75

     0      100
  (cm/sec)

1.38 X10-3

2.06 X 10-4

8.82 X 10-5

9.61 X 10-5

3.6 X 10-6
  (cm/sec)                 —

1.70 X 10-4               Q.13

3.14 X 10-5               0.15

2.78 X 10-5               o.31

2.15 X 10-5               0.22

possible steam fracture
     •  The efficiency!  of steam  stripping of kerosene  (and gasoline)  from
       various soil types was  determined in the same cells  where the steam
       permeabilities  were determined. Results are shown in Figures 2 and 3.

     •  A. steady  state analytical model  was  developed  where steam  flowed
       upward to  the  collection  cap  from pipes embedded in the  soil  (see
       Figure  4  for  the  model) .  The  model,  used   together  with  the
       permeability  and stripping efficiency  data  described  previously,
       allowed a determination  of  the  time to  decontaminate a given kerosene
       spill.  The  results using the model are given in detail in reference  3.

     •  A small scale model  of the  geosynthetic cap was used to determine its
       feasibility  as a cover assembly during  steam  stripping.  The  cap
       consists  of a geotextile and geomembrane which is anchor-trenched  to
       the  soil.   A  schematic diagram of  the  experimental setup  is  shown  in
Efficiency means the rate  and degree of removal of the contaminant.
                                    126

-------
      Figure  5.  Results are  shown in Figure  6 for  vacuum-assisted steam
      stripping  in  beach sand  for various levels  of kerosene saturation.
      (Dimensions  given  on diagram.)  No  noticeable escape  of  steam or
      kerosene  was detected  during  the experiments,  indicating  the cap
      performed well in containing and collecting the  steam and kerosene.

    The work in Phase  I pointed  out at  least three definite needs:

    • Our- "chemical analytical  technique"  involved separation of the water
      and  the kerosene  (or gasoline)  and measuring the amounts  of each
      volumetrically.  Material lost to outgassing and that condensed  in the
      lines  is  not counted as  output  and hence a  more  precise analytical
      technique  is needed.  A  major effort  is  now underway to determine the
      amount  of  material  remaining in the soil after vacuum-assisted steam
      stripping. Gas chromatography (GC)  is  the  technique being used.

    • Kerosene  and  gasoline are mixtures of very  many compounds  of widely
      varying properties. Therefore, in order to understand the process more
      completely,   it  is  desirable  to  determine the   efficiency  of
      vacuum-assisted, steam stripping of individual  compounds. This  is now
      being  pursued with  a  series of alkanes of differing vapor pressures,
      and  also  a series of polar organics of  differing  polarities.  It is
      felt  that vapor  pressure  and polarity are two  very  important
      parameters in determining  the ability  to steam strip  a particular
      compound.

    • The  method shown in  Figures 1 and  4 is only  one  possible  means of
      injecting  the steam.  Also to be considered (more in keeping with the
      field  work performed  so far)  is  the  flow  of  steam from vertical pipes
      with perforations  over  part  of their lengths.  It  is  possible that a
      vertical  vacuum tube(s)  will be  placed in  the soil  for  steam and
      contaminant  collection.  Here the cap will be a secondary collection
      system and mainly  function  to reduce  any  unwanted Vaporizations tinto  the
      ambient.

    The  first two  points  (GC  and simple compound studies) are addressed in
this paper.

                                PRESENT WORK

RESULTS  WITH  INDIVIDUAL COMPOUNDS

Volumetric Measurement  of  Steam  Stripping Efficiency

    Volumetric  analysis  of the steam  stripping ability  of various organic
compounds  (and kerosene)  were determined. Typical results  are  shown  in  Figure
7, Here  is shown the  apparent percentage removal in  a 50%  sand - 50%  silt
soil for dodecane, decane, octane,  octanol  and  kerosene.   It  is obvious  from
Figure 7 (and similar  results  for the other soil types) that  a more rigorous
method  of  determining stripping  efficiency  is  needed.  For  example, the
stripping  in octane  is essentially complete after  one hour, therefore the
                                    127

-------
  apparent  removal efficiency  of  40% is certainly artificially low.

      Rather than  continuing  on  with  detailed accounting for all the chemical
  steam stripped,  a more  simple,  reliable, precise method was sought with which
  to   determine  the  stripping  efficiency.  The  method  chosen  was  gas
  chromatography  (GC) analysis of the  chemical remaining  in  the  soil  after
  steam stripping.  Results using  GC are discussed in the next Section.

  Gas  Chromatocrraphic Determination of Residual Chemicals in the Sn-i 1

      Standard gas chromatographic  (GC)  techniques were employed.  Three-fold
  extraction  (from  the steam-stripped soil samples)  with methylene chloride was
  used.  Some  actual GC results for the residual chemical are shown  in Table 2
  for  decane. The  results and those to be shown in Figure 8 show that vacuum-
  assisted  steam stripping is  a  very rapid and efficient  means for  removing
  organics from soils. The amount of decane remaining in the  soil unfortunately
 does  not  follow any trend  as a function of silt content. This observation
 also  applies  for all the  other chemicals  (dodecane,  octane and  octanol).1
 Therefore, for want of a better approach, the overall results  for the various
 chemicals  will  be  presented as the  average  (over  the  five  soil types)
 remaining in the soil after some particular  time period of  steam stripping.

      Figure  8  gives  such a presentation.  Here is  plotted the average amount
 of chemical remaining in the  soil after  five  hours of  steam stripping versus
 the vapor pressure of the-pure chemical at 100°C.  For the alkanes,  it is  seen
 that the  removal is a very  strong  function of the  vapor  pressure. The low
 vapor pressure dodecane is relatively much more difficult to remove than the
 high vapor  pressure octane.  However even dodecane,  with  a very  low vapor
 pressure is still reduced to only 0.1% of its initial  value in  the  soil, via
 vacuum-assisted steam stripping.

      It is interesting to compare our results for  the alkane steam stripping
 with the  theoretical  value  for the  simple steam  distillation  of a  two
 component  immiscible liquid mixture  (10). The theoretical alkane-water ratio
 in the steam-distilled mixtures  is
                     alkane
                    W  „
                     water
                M     P
                  alkane  alkane
                 M     P
                  water  water
(1)
where
    W
     alkane
    W
     water
    M
     alkane
    M
     water
    "alkane
ratio of weights stripped (distilled)


molecular weight of the alkane

molecular weight of the water

vapor pressure of the pure alkane at the  combined boiling point
 It may be that more chemical remains in the finer grain soils,  but  it may be
more  difficult to remove with the GC extraction solvent.
                                     128

-------













rH
0
IS
H
Oi
H
B
CO

J2
w *»^
EH f?
W ft
* 2
g &
is
£ o
g CO
nj
s! °
H -Q
g CO
Bfi T|
O £,
g|
Q 5.
H
CO
§

CM
H

CQ
£














h_
?*
-H
G
-H
g
d)
PS
0\°


CO
EH

CO ft
§


-H
O
CO

t7>
O
in
H
3

w
§j3
O


,_fl
0) -H
M O
O CO
J
rH
-H
o
CO


a>
i— i
i*
CO






^ (O CO ^ C3 rH ^d1 C3 C^CM CM
OrHrHrHrHHOrHrHOO








cQooSScM
i-H rH






vorovovor>CMC~c3^'<0o
rH-*C~r~-^inrHin<^rHrH


cocoooooooomm

"* ^ -ST rH rH i-H CM




CMt~rOCMCMOOCOint~COOO
CN]'CvIrocnrororo'a<'<3J
T-irHr^c^-LOLOLOr^r^




rHCMC-OOCO^rHCT^OlO^O
rH rH





























»r-4
•rl
o

o
CO
g
m
Cn
a
o
•H
i-H
rH
"2
(D
a
o

M
a)
ft

0)
G
n)
o
0)
•c

o

CO
(d
M
tn
*
>l CO 0
M +J -P

fi (1) 'D
-rl M 1>
H 3 a>
fl> CO C
ft S «
g "H

goo
_« n s

o n
*O O O
s ^*
M1 ri
* G 0
-S-H-.H
,_1 III
0 rH 0
o m
M M
0) CM O
C*^ ' i)
• id
o "
">• g; * -g
•y -H co x!
"^ M (d 4J
•jj g 4J O
3 id ^^
T3 0) /-) t,
0 +J S 3
M CO 5
gj IH -p M-I
o ^
0) 3 >i
X! <1> 0 0
.p o o (d
G (d M
tu m 3
—t (i) rO (d
.3 cu w o>
4-^ rH 4J
(0 rj
O Q) co >i
-H g d) H
O *rH* |-| fl)
G 4J (0
•H 0 -H
m U O
G S 0)
(d M M
^ 3 ft
w ° „,
>°^ §
t_ (~{ £j
o **
f^ -H a) qj
•P +J -H

rH -rH -H Ti
129

-------
      "water
              =  vapor pressure of pure water at the combined boiling point
 The very simple argument  will be  tested here, that  the amount  of  alkane
 remaining  in the  soil  will be  inversely proportional  to the  theoretical
 weight ratio  of alkane  to water  (that  is,  the larger ratio  favors quicker
 removal). Table 3 gives  this  theoretical  ratio  (Eqn. 1), the inverse ratio of
 the theoretical alkane weights  (based on octane) and  the experimental ratio
 of remaining chemical in the soil. The trend (comparing the second and third
 columns of Table 3) is certainly  in  the correct  direction, but relative sizes
 are considerably  off,  i.e.,  the  strict ideas of steam  distillation  of
 immiscible liquid mixtures does not apply to steam stipping of  alkanes from
 soil.  This is certainly not  surprising or unexpected  as  the steam stripping
 in soils is a much more  complicated  process.

     In the case of the  alkane-based alcohols  (which have some solubility in
 water),  it is seen that  low  vapor pressure compounds  (e.g. octanol)  can be
 very effectively steam stripped from a wide variety of soils. The butanol has
 been reduced below detection  limits  (1 ppm) by the steam stripping technique.
 The reason for  the very  efficient removal  of the  alcohols  (versus  the
 alkanes)  is not understood at present and needs to  be  investigated further.
 Octanol is mildly polar  (dielectric constant  = 3.4)   and butanol  is  quite
 strongly polar  (diel. const.  = 20). The  degree  of polarity  is certainly one
 of the major  differences between the alcohols and the alkanes. ^

                                 CONCLUSIONS
     It appears that vacuum-assisted, steam stripping of  a  number  of organic
 chemicals,  in  particular  alkanes and  alkane-based alcohols from a wide
 variety of  soils is quite feasible. The residual chemical (in most  cases)  can
 be  reduced  to below 100 ppm in a relatively short treatment  time.  It  is  hoped
 that the work presented here leads to a better understanding of the mechanism
 and future  potential of the  process  in the  very  important area of  in-situ
 soil  decontamination.  This  research has  showed the  importance  of  vapor
 pressure and polarity in determining the steam stripping ability.

     The unique geosynthetic cap comprised  of  a geotextile,  for rapid lateral
 flow  of steam  and  contaminant, and  a  geomembrane,  for  containment of the
 steam  and contaminant,  appears  to act quite effectively. When anchored into
 the soil at  the edges  and  connected to  a   vacuum,  the  geosynthetic cap
 performed well in the steam stripping of kerosene from sand.

    More research work remains  to be  done. In particular:

     •  More  chemicals  need to  be  investigated to determine the  general
       ability  of steam  stripping.

     •  A clay fraction should  be added to the  soils - clays  will bind certain
       chemicals more than the coarser grained materials.
•••It may be that  the alcohols are  more difficult  to  extract  (with  the GC
solvent) from the soils than the  alkanes -  and hence appear  (artificially) to
be more effectively removed.
                                     130

-------
     •  The GC methods 's reliability  in  regard to the steam-stripped samples
       must be pursued further.

     •  Work should proceed in developing the  unique geosynthetic cap assembly
       - and testing it on a larger scale.

     •  The use of the technique  on excavated  soils should be considered.

                                 REFERENCES

1. Kovalic, J. M. and  Klucsik, J. F., "Loathing for Landfills Sets Stage for
   Innovative Hazardous Waste Treatment Technology," Hazard.  Mat. and Waste
   Manag. £, 1987, 17-18.
2. Cheremisinoff ,  P. N., "Update:
   M,  Feb.  1987,  42-49.
                                    Hazardous Waste Treatment," Pollut. Eng.
3. Lord, A. E., Jr., Koerner,  R.  M.  and Murphy, V. P.,  "Laboratory Studies of
   Vacuum-Assisted Steam Stripping or Organic  Contaminants from Soil, " Proc.
   14th  Annual _ Conf.  on Land Disposal.  Remedial Action  and Treatment  of
   Hazardous  Waste, Cincinnati,  Ohio, April, 1988,  sponsored by  the  Risk
   Reduction  Engineering Laboratory, U.S. Environmental Protection Agency,
   Cincinnati, Ohio.

4. Hilberts,  B.  (1985),  "In-Situ Steam Stripping,"  Assink, J. W. ,  Van Den
   Brink, W.  J., Eds., Contaminated  Soil,  Proc.  of  1st Intern.  TNO Conf.  on
   Contaminated Soil,  Utrecht,  The  Netherlands,   Nov.  11-15,  1985,  pp.
   680-687.

5. Baker,  R., Steinke,  J.,  Manchak,  F.,  Jr.,  and  Ghassemi,  M.,  "In-Situ
   Treatment  for  Site  Remediation, "  Proc.  Third  Annual Conference  on
   Hazardous  Waste  Law and Management,  Seattle,  Wash., October 27  and 30,
   1986, and Portland,  Oregon,  October 31  and  November  1, 1986.

6. Baum, R.,  short  article  describing  process  appearing  in  Chemical  and
   Engineering News, December  12,  1988,  pp. 24-25.  Work done by K.  Udell,  J.
   Hunt, and N. Sitar at University of California,  Berkeley  and A.  Nagtel  of
   Solvent  Services, Inc.,  San Jose,  CA.  Also described by  Kelley,  K. P.  in
   Haz. Mat. World,  January 1989,  pp.  12-14.

7. Advertisement of  GeoCon Inc., Pittsburgh, PA,  -  appearing in  Hazardous
   Material Control,  Volume 1,  #4, July-August 1988.

8. Lord, A.  E., Jr.,  Koerner, R. M.,  Murphy,  V.  P.  and  Brugger,  J. E.,
   "In-Situ,  Vacuum-Assisted,  Steam  Stripping of  Contaminants from  Soil, "
   Proc. - af — Superfund _ '87r   8th National   Conference on Management  of
   Uncontrolled Hazardous  Waste Sites, November  1987, Washington, DC, pp.
   390-385.  Sponsored  by the Hazardous Materials Control Research Institute,
   Silver Spring, MD.
                                   131

-------
9. Murphy, Vincent,  P.,  "In-Situ, Vacuum-Assisted, Steam-Stripping to Remove
   Volatile  Pollutant  from  Contaminated  Soils," Masters Thesis in  Civil
   Engineering, Drexel University, Phildelphia,  PA, June  1988.

lO.Prutton,  C.  F. and Maron,  S. H., Physical Chemistry,  MacMillan Co.,  NY
   (1975), pp. 175-177.
                               ACKNOWLEDGEMENTS

     This project is funded by the U.S. Environmental Protection  Agency under
Cooperative  Agreement  CR-813022-01.  The Drexel  authors  offer  our  sincere
appreciation to the Agency for their support.   Thanks ar  due  Dr.  Frank Davis,
Ping Chen  and  Bang Chi Chen of the  Chemistry Department at  Drexel  for their
involvement with the GC work.
TABLE 3.  THEORETICAL WEIGHT RATIOS OF ALKANE  TO WATER IN  STEAM DISTILLATION
          COMPARED TO RESIDUAL ALKANE IN  SOIL  AFTER  STEAM  STRIPPING
           W
            alkane
                          W
                           octane
            W
             water
                          W
alkane j
                                         W
                                          alkane
                                         W
                                          octane j
                  theor.
                                 theor.
                    expt'1.
                 Boiling Point of
               (Alkane/Water) Mixture
octane       2.83

decane       0.86

dodecane     0.39
 1

 3.30

 7.25
  1

  8

100
  90°C

97.5°C

99.5°C
                                      132

-------
                                               «HINJECTION PIPCS
                            PIPE MANIFOLD SYSTEM —a




                                    PLAN VIEW
                                                          VALVES
                                              VACUUM COIUCTIOB
                             MCHDBTRCKH


                                 riOUlU HCMIMM LIMU
                                   4 MIOUB WINWD»E« CUICXTIIC
                                   C—    81 tUCOHmiTC


                                       t -A  t
                                     HUM BltTILlATIUH
                                                        IDUICIION PlPta
                                               . CRQUIM WA1ER TMLE
                                  ELEVATION VIEW




Figure 1 - Schematic  Diagram  of  Proposed  In-Situ  Vacuum-Assisted Steam

            Stripping Field Apparatus
                     100
                 z
                 Ui
                 tn
                 o
                                200
400      600





   TIME (minutes)
                                                         eoo     1000
Figure 2 - Steam Stripping Efficiencies  of Kerosene from the  Various  Soils  in

            the  One-Dimensional  Cells  (Volumetric  Analysis)
                                        133

-------
                            STEAM STRIPPING FOR GASOLINE
                    • 0
                             50       100      ISO


                                   TIME (min)
200
Figure 3 - Steam  Stripping  Efficiencies of Gasoline from  Two  Soil  Types
          (Volumetric Analysis),
                                  ground  surface
                                              flow lines
                                             equi-pressure
                                               contours
           steam pipe
          Figure 4 -Model Used in Steam Stripping Calculations
                               134

-------
               Contaminated
                  Soil
To Condenser
 and Vacuum
             Clean Soil
  Geosynthetic Cap
'(Geotextile and Geomembrane)
                                           m
                                           ••
                         14"xl4"
                         Soil Box
                                               Steam in
  Figure 5 - Schematic Diagram of Pilot  Scale Experiment  for Vacuum-Assisted,
             Steam Stripping Using the Geosynthetic Cap
              100
          I
          O
          I
          K

          I
          e
           10% Saturation
           25% Saturation
           50% Saturation
                                     1000

                                   TIME(min)
                                                           2000
Figure 6 - Results for Steam Stripping Kerosene Contaminated Beach Sand Using
           Pilot Scale Geosynthetic  Cap of Figure 5 (Volumetric Analysis)
                                     135

-------
                        REMOVAL EFFICIENCY FOR 50% SAND
                                  TIME (hours)
 Figure 7 -Results  for Steam Stripping Various  Chemicals  from  50% Sand/50%

            Silt Mixture  (Volumetric Analysis)
                                     SOIL STEAM STRIPPED

                                       FOR 5 HOURS
                        •§• 1000

                        CL
                        Q.
                       HI



                       O
                       O
                          100-
                       O

                       E
                       ui
                       HI
                       cc

                       LU

                       8
                       cc
                       UI
Dodocane
                                                   Butanol
                             0    100   200   300   400   500



                               VAPOR PRESSURE AT 100 C (mm Hg)





Figure 8  - Results  of Average Residual Chemical in  Soil After Steam Stripping

           for 5 Hours (Gas Chromatography Analysis)
                                       136

-------
                   LOW TEMPERATURE THERMAL DESORPTION FOR
                       TREATMENT OF CONTAMINATED SOILS
                              PHASE II RESULTS
                   Richard P. Lauch and Robert C. Thurnau
                    Risk Reduction Engineering Laboratory
                    U.S. Environmental Protection Agency
                            Cincinnati, OH  45268

                          Ed AT perin and Arie Groen
                               IT Corporation
                            Knoxville, TN  37923

                 Barbara B. Locke and Catherine D. Chambers
                            PEI Associates, Inc.
                            Cincinnati, OH  45246
                                  ABSTRACT

     Performance of the low temperature thermal desorption process for re-
moving volatile contaminants from soils was evaluated.  The data obtained
were necessary to assist EPA in the study of alternatives for treating Super-
fund soils.  Soils from two Superfund sites were selected for treatment.  The
effect of temperature and residence time was determined using a tray-type
furnace.  Temperatures of 350°F and 550°F, and residence time of 30 minutes
were tested.  The differences in concentration before and after treatment of
volatile and semivolatile organic compounds were used as a measure of treat-
ment effectiveness.  Metal concentrations before and after treatment were
also determined.  Results from these tests on actual Superfund soils (Phase
II) were also compared to earlier results of tests on synthetic soils (Phase
I).  The Phase II results showed that over 87 percent of volatile organic
compounds, and over 79 percent of semivolatile organic compounds were removed
at the 550°F temperature.


                                INTRODUCTION

     The thermal treatment of solids has been practiced for many years to
effect a chemical change in the solid or to separate components based o.n a
physical property such as vapor pressure.  The application of thermal treat-
ment to hazardous waste problems has utilized both physical and chemical
processes to decontaminate soils or other solids containing hazardous
constituents.

     Historically, thermal treatment has been most commonly practiced in
direct-fired incineration systems which heat contaminated solids to high
temperatures.  These systems are effectively used to decontaminate solids
that contain hazardous organic compounds.  The organic constituents on the
                                     137

-------
heated  solids  are  removed  and/or  rendered nonhazardous through a combination
of chemical  reactions  and  physical  transformations.  The heated solids are
discharged from  the  incineration  system after the organic contamination has
been removed.  The organic compounds removed from the solids are then de-
stroyed.  Treatment  of contaminated solids by this technique is currently
practiced on a commercial  scale on  a wide variety of solid waste problems.

     The thermal desorption  process takes advantage of thermal driving forces
to remove organic  contamination while avoiding typical incineration process-
ing conditions which are expensive  or have negative public perception.
Thermal desorption is  conducted at  lower operating temperatures, offering
significant  fuel savings over high  temperature incineration.  The heat re-
quired  for thermal desorption is  provided by indirect heating of the soils as
opposed to direct-fired heating of  solids in an incineration process.  This
greatly reduces  the  quantity of off-gases which must be cleaned prior to
discharge.   This design aspect not  only reduces the cost of subsequent air
pollution control  but  also facilitates the design of a closed system with no
visible plume.   Thus treatment of contaminated soils by thermal desorption is
potentially  more cost  effective on  low level organically contaminated soils.

     Thermal desorption has  been  successfully tested, at both the bench and
pilot scale, on  a  wide  range of solids contaminated with organics (EG&6 1988;
IT Corp., 1986).   The  organic compounds that have been successfully removed
from different soil  types  include:  polynuclear aromatic compounds from soils
contaminated by  coal gas manufacturing plants, polychlorinated biphenyls
(PCBs)  from  soils  contaminated by spills of oils, and priority pollutant list
compounds from surrogate soils spiked with these compounds (PEI and IT 1987,
Szabo,  et al.  1988). In these tests the technology was found to be effective
in removing  the organic contamination to the desired levels.  The specific
treatment conditions for these compounds varied according to the chemical and
physical properties  of each  contaminant and the matrix containing the con-
tamination.

     Because the nature of the process is physical and chemical, it lends
itself  well  to widely varying soil  types and soil characteristics as might be
found at Superfund sites.  Often  the nature of the soils found at Superfund
sites varies in contaminant  depth,  geographical site location, and existing
mineralogical  conditions.  The types of contaminants also vary widely but
they are usually compounds with moderate vapor pressures.  Compounds with
lower vapor  pressures can  also be removed using higher treatment tempera-
tures.  Figure 1 provides  a  general idea at what temperature several organic
compounds will vaporize; however, vapor pressures will vary depending on
whether the compound is the  sole  contaminant or there is a mixture of con-
taminants.   It is  interesting to  note that the final results of this study
did, for the most part, show the  highest percent removals for compounds with
the highest vapor pressures  (see  Figure 1).

     It is the ability to  treat a variety of soils with varying types and
levels  of contamination without other pretreatment that makes thermal desorp-
tion an attractive technology for treatment of Superfund soils.  The treated
soil often requires  no further treatment and can be immediately returned to
the site.
                                      138

-------
   en
   X

   E
   E,

   UJ
   rr
   ID
   CO
   CO
   UJ
   rr
   DL
   DC
   O
   CL
       600

       760
       700 -
       600 —
500  -
400
300  -
       200
       100
    = Trichloroethylene

    = Toluene

Q  = Tetrachloroethylene

^  = Pentachlorobenzene

|  |  = Hexachlorobenzene
                      TEMPERATURE, °C(°F)

                       Figure 1. Vapor pressure of selected volatile and
                                 semivolatile organic compounds.

     An additional  design aspect  of thermal  desorption which facilitates
application  to  the  wide range of  Superfund  site problems is the flexibility
to desian the off-gas treatment for specific site requirements.  The off-gas
from the thermal  desorber can be  incinerated in a high temperature secondary
combustion chamber  if it is desirable  to destroy the organic contamination.  .


                              EXPERIMENTAL DESIGN

     In the  Phase I study, conducted from May to September, 1987, the effec^
tiveness of  low temperature thermal desorption on spiked synthetic soils  was
evaluated.   Samples of soil were  thermally  treated in static trays in an
electric oven for specified periods of time.  Removal  of organic constituents
was measured after treatment at different test temperatures for a duration of
30 minutes.  The effectiveness of thermal treatment was measured by analysis
of the treated  residue for the known contaminants in the soil samples.

     The experimental approach for  Phase II testing was similar to Phase  I.
The test runs for Phase II consisted of each Superfund site soil being  treated
                                       139

-------
 at two different temperatures (350°F and 550°F).   Each  temperature was eval-
 uated by conducting four test runs.   A test run consisted of two  separate
 batch runs, with the residue from the two batches  being composited into a
 single test run sample.

      The experimental  procedures for Phase II  testing are described  in the
 following Experimental  Apparatus and Experimental  Procedure  sections.   The
 soils used were from two Superfund sites:   the Berlin-Farro  Site  and the Old
 Mill  Site.  The Berlin-Farro Site soils consisted  generally  of  glacial till,
 while the Old Mill  Site  soils were a glacial silty clay with sand, gravel,
 and boulders.  Table 1  presents  the  physical analysis of the two  site soils
 used  in Phase II and the synthetic matrix used in  Phase I.   Generally, soils
 lower in clay content  (and conversely higher in sand content) should be
 easier to treat with thermal  desorption because the organic  contaminants are
 not as tightly bound to  the sand as  to the clay.

                  TABLE  1.   PHYSICAL ANALYSIS  OF TEST SOILS
Phase  II

Berlin-Farro


Old Mill


Phase  I

Synthetic soil
 matrix
                 Coarse  sand
                  (>0.5  mm)
 4.2
 3.3

28.2
37.2
 7
 6
             Fine sand
           (0.05-0.5 mm)
                 Silt
            (0.002-0.05 mm)
                Clay
             (<0.002 mm)
35.2
36.1

40.3
35.6
48
48
33.4
34.1

22.6
19.6
33
33
27.2
26.5

 8.9
 7.6
12
13
EXPERIMENTAL APPARATUS

     The experimental apparatus used in Phase II was the same equipment used
in Phase I.  The primary piece of test equipment was a Lindberg furnace,
Model 51848, with an electronic temperature controller and 1600 watt heater
system.  The o^en has double-shell construction with interior surfaces made
of Moldathernr ', a molded aluminum-silicate insulation material.  This oven
is capable of operating up to 1100°C and has a relatively fast heat-up rate
due to its low mass.  The interior space is approximately 10 cm (3.9-in.)
wide x 11 cm (4.3-in.) high x 21 cm (8.3-in. deep).  A loose block (1/2 in.
thick) of Moldatherm is placed on the bottom of the oven to provide addi-
tional separation between an object placed in the oven and the hot interior
surface of the oven.
                                     140

-------
     The oven was continuously  purged  during  each  test by nitrogen  from an
Incoloy (3/8 in.) tube  inserted through  the back wall.  The purge gas  was
directed against the back wall  to  promote  preheating  and distribution.  The
purge gas flow rate, which was  measured  using a standard rotameter, was
maintained at approximately  90  cc  per  minute.  This flow rate resulted in a
complete turnover of the oven environment  every 20 minutes.

     Two thermocouples  were  used to measure temperature.   One of these was an
NBS traceable, type K,  sheathed thermocouple  placed approximately 3 cm above
the soil at the center  of the oven.  This  was the  thermocouple used to measure
the  "test temperature".  The other thermocouple was  used to measure tempera-
ture within a soil layer during selected test runs.   These thermocouples, the
oven temperature indicator,  and the purge  gas rotameter were calibrated
according to standard engineering  practices.   Temperatures were recorded
using a Cole Farmer 3-pen recorder, Model  595.  The oven, nitrogen  purge
line, and thermocouple  arrangement are shown  in Figure 2.
      INTERIOR OF
      OVEN CHAMBER
                                                   OVEN INDICATOR
                                                   THERMOCOUPLE
                                                           PURGE
                                                  TEST THERMOCOUPLE

                                                  SOIL THERMOCOUPLE
                            GAS EXIT AT DOOR SEAL
      Figure 2.  Schematic of interior of static tray test oven with the tray inserted.
                                    141

-------
     A specially made tray was used to contain the soil  within the oven.   The
tray which weighed approximately 430 grams, was 8.9 cm (3.5-in.) wide x 3.3
cm (1.3-in.) high x 19.3 cm (7.6-in) long and made of Incoloy to resist
oxidation.  A separate Incoloy lid was used to cover the tray while cooling.

TRAY TEST PROCEDURE

     A single procedure was established for all thermal  treatment tests; four
replicates were run for each sample.  The detailed tray test procedure is
delineated in the EPA draft report entitled "Alternative Treatment Technology
Evaluation of CERCLA Soil and Debris."  Briefly, the procedure entails trans-
ferring approximately an 80 gram sample of soil to a clean tray and spreading
the soil to achieve a uniform layer on the bottom of the tray, usually about
2.5 to 3 mm deep.  The tray is then inserted into the oven at ambient tem-
perature and heated to the target temperature.  When the prescribed residence
time at the target temperature is reached, the oven is shut off and the
sample is removed and allowed to cool for about an hour.  The sample is
weighed and an aliquot is transferred to a VGA vial.  The vial is properly
sealed with a Teflon-lined septum cap and the sample is sent for analysis.
The tray is cleaned and prepared for the next run.

                                   RESULTS

     Soil samples from the Berlin-Farro Site and the old Mill Site were
analyzed for all volatile organics listed on the Hazardous Substance List
(HSL).  Only the compounds that were detected are reported (see Tables 2 and
3).  Only those semivolatile organics detected in the untreated soils were
analyzed for in the treated soils.  Berlin-Farro soils were analyzed for the
semivolatile organics listed in Table 2.  Samples were also analyzed for
pyrene and bis-(2-ethylhexyl)phthalate; however, the results were inaccurate
because of analytical problems at the laboratory and therefore, were riot
reported in Table 2.  No semivolatile organics were detected in the untreated
Old Mill Site soils.  Aroclor 1260, however, was present and was analyzed for
in the treated soils (see Table 3).  TCLP extracts of both soils were ana-
lyzed for arsenic, barium, chromium, copper, lead, nickel, vanadium, and
zinc.

     The results of the thermal treatment test runs show that thermal desorp-
tion is effective in removing organic contaminants from Superfund soils.  The
treatment of Berlin-Farro soil is shown in Table 2.  The reduction of the
volatile organic compounds was typically greater than 90 percent at tempera-
tures of 350°F and 550°F.  The exception to this result was the behavior of
2-butanone at 350°F.  The concentration of 2-butanone appears to increase 18
percent from the initial concentration.  This compound, however, was detected
in all of the blanks corresponding to analysis of this sample; therefore,
this apparent increase probably resulted from laboratory contamination.

     The percent reduction of semivolatile organic compounds from BerlinFarro
soil was slightly lower than for the volatile compounds.  This is logical be-
cause of the higher vapor pressure of the volatiles.  The analytical results
are summarized in Table 2.  Data indicated an increased concentration of
pentachlorobenzene and hexachlorobenzene at the 350°F test temperature.  As
                                     142

-------
LU
CS
^C
CfC
UJ
^»
•a:
•»— '
_i
i—i
o
CO
0
a:
C£
^c
u_
l
«£!.
1— <
ce:
LU
CO
u_
o

^
o
1 — 1
1—
o.
a:
0
CO
LtJ
a

wJ
<^
s:
O**
UJ
^L.
H-

LL.
O

CO
J—
— J
,r J
oo
LU
a:

O
1— <
Zp
CO
o

LU
o

^"
Q£
<^
Sfe
it.
•_J
oo


*
C\l
LU
_J
CO

CO
a>
4-j
c
wj






















— <
c:
0

4-9
CJ
3
•a
CD
ct:
^


CO

*^^
co
^3.




-1
0
•r—
^ ^

•^
"^-^
co
=1






co
«^£
*^^
CO
3.
•t
^~
•^
o
CO


















i.
0)
+J
ro
CO
a.


CM CO CT> «3- •-•
f**^ co cr* cr» cr>
A A







O LO CO »>» LO
CO CM CM CM
V V








*• ^
CO "5j" CM ^O ^*
«-l CO CTi CT» tTi
*— ' A A A







CM
co co co cr» ro
«d- CM CM t-l CM
co v v v











CD f>- O CO f»»
cr> «a- oo co co
CM t— 1 CM «3- CO



















co a>
u c
E c .c ^
IO (I) 4J CO
Ol f 
S- 4-> O O
O O) O) i_ 4->
coo — •
OJ 0 S- •—
t— c o J= a, co
•r- CO f— (J C O)
4-> +J J= CO O) C
CO 3 O t. 3 O)
r— CQ «r- 4-» r— r—
o i &- a> o >>
> CM t— 1— 1— X
4
































f













c


















C/)
0
•r—
c:
CO
Ol
S-
o
olatile
(U
CO
5(-



CJ CO LO LO









o o
o o
CO CO
A •
co co
V V






P-. co








o

0 0
CO LO
•>!- O

CO







o

o o
O 0
en o
•V A
T— 1 ^O
•sf








0)
c
0!
•f—
T3
ro
 ro
C X
Ol (U
Q- 3C









OJ
3
CO

'o


>1

•a
0)

a>
S-
S-
o


1 *
o


«
c
o
•f—

CO

[ ^
c
(U
o
c
o
o

r—
to
•^
^_J
•r—
C
•r-

ps
O
i.
**~
d)
CO
c
CO
^Z
u

^5

CO
CO

"Q
O)
4->

O
a.
QJ
s_
o
•r-
4->
O
3
•o
OJ
C£
i— 1




































•
t/1
O)
CO
a>
^
4-»
C
CD

CO
a

^
•r—
C

O

CO
2!
ro

C
O
•^~
4^
ro
s_
4->
C
QJ
U
C
O
u
c
reases i
u
c
1— <















































•

t"
rO


^.
y^
0
CO
S-
o
**^
ro


QJ
g^
4^

C
•r-

•o
QJ

0
(U
(U
HO
c
o
a.
o
o
CM




































t
>.
s-
o
4->
CO
s_
o
/-I
CO
r—

Q.
_J
CJ
a>

4_>

^3

QJ

C
QJ
to
QJ
S-
o.

CO
CO

CO

3
CO
QJ
4J
CO
•r*
CO
LU
co
• A
t*
o

CO

o
-Q
ro
QJ
-t-5
^^
.a

•o
QJ
•a
0)
ai
o
X
QJ
CO
ro
3-

CO

CO

*^
ro
C
ro
QJ
J3


tj
O

4.)
•r-
c
•r—


C
O

1 <
o
QJ

(1)
HO

O •
J= -0
4-J Q,
QJ 4->
E CO

QJ 3
J= U
1— r-
CO
O

HO a,
(U J3
^ ^
CO 4->
i — O
3 C
U C
r— ro
ro O
0
' C
4-> O
O-r-
C 4J
U
II 3
O O)
«3-
143

-------
LU
CD
§
LU
^>
2.

_J
O
CO

7;
i— i
2:

o
_J
o
Lu
O

z:
0
»— i
i—
o_
ce:
o
CO
LU
0
^
ex.
LU
n:
|— .

Lu
O

00
f—
_l
=3
CO
LU
f^

U

•z.
1
0

LU
O

>—
(V^
•a:
§F
^^
co


•
co

LU
^J
CD
cC
i—










c
o
•r~*
O
3
•o

Di
&«



Lu cn
O *s
O *-s>
LO cn
LO a.



t— i
c
o
•r-
«^_>
U
3
"^
(U
C£



LU cn
O -*"*
O "-^
LO cn
CO 3.







cn
*o -^
 cn
re 3.
(U
t. «
C *f™
=> o
CO




















t-
CO
*4^
fli
(Ss
re
t-

Q-



CM CO CO LO
C7% C7l LO CM
A A







LO LO r^> oo
CM CM LO «5l-
V V








CO O CM O
CM CM r*«* r^^








CO LO CO LO
1*^ CO «* 00
i-H CM









O CM CM O
O LO LO LO
^- CO t-l CM

CM















co cu
u c
•r- CO O) • — .
C C .C t—
re cu 4-> re
cn.c cu ••->
t. -4-3 O O
O CO i. +>
o o - —
CJ J_ i—

•i- r— O C CO
+j J= re co c
re u s- 3 co
r- -i- +J i— r—
O J- CO O >>
s»t-^t-x



LO
CM
A




CO
o"
o
r-4
SM«* '

o
•z.




CM
ff-~^
o
LO
*a^^







o
o
o

CO







o
o
o

CM











CO
u
•r-
C
re
cn
i.
o

CO O
r— LO
••- CM
4"^ i~H
re
i— S-
o o
> r—
•r- 0
E 0
 i-
CO «C







CO
3
4J
to
o
£

>^
J2

-o
U
CO
s-
s_
o
o

4_)
o
c


.0
re

.}_>
c
(U
u
c
o
o

^—
re
•r-
^J
•t—
C
•r-

ps
£
**~
CO
cn
c
re
.£=
O

&«
CO
re

•o
CO

j_
o
0.

i.

c
o
•^
4_)
u
3
•o

on

t-H
































•
CO
o>
CO
CO

4_)
C
CO
i.
re
Q.

c
•r-

C

O
CO
CO
i.
re

c
o
•f—
re
^
4^
C
CO
o
c
o
u

c
•1—

CO
CO
CO
re
CO
i-
o
c


CM



































•
*~**
^^
•r—
E
•f—
|—

C
o

•1^
o
CO

CO
•o
-o
o
.c
^_>
CO

^VM^

CO
•I-*
u
CO
•40
CO
•o

^^
0


II

o
o
t— 1
*H^t

0


CO
144

-------
can be seen in Figure 1, the vapor pressure of these two compounds at 350°F
is low.  As the sample dries out in the heating process it looses mass, thus
concentrating any non-volatilized contaminants.  The apparent increase in
concentration of these two compounds at 350°F may be a result of this phenom-
enon.  It should be noted here that several problems were associated with the
semivolatile analyses, and these data should be used with caution.  Specifi-
cally, hold times for all semivolatile analyses were exceeded and method
detection limit requirements were not met.

     The test results for the Old Mill soils, presented in Table 3, are
similar to the Berlin-Farro soil data.  The volatile organic data trended
toward reduced levels of volatile organics at higher treatment temperatures.
The data for the Old Mill site soils showed an increase in concentration of
Aroclor (PCB) after treatment at 350°F, but greater than 95 percent reduction
at 550°F.  Again, the apparent increased concentration of Aroclor at 350°F is
likely a result of the concentrating of the compound as the sample mass
decreased during the heating process.  Ardor is successfully volatilized at
550°F.

     Tables 4a and 4b summarize the results of TCLP extract metals analyses
for the two site soils.  Total metals concentrations in both soils are also
given in the tables.  No significant reductions were noted, nor was reduction
of metals expected as a result of the low temperature desorption process.
Increases in particular metals concentrations are a result of sample weight
loss as the sample dries out during the heating process.  This behavior is
comparable to that described above for the semivolatiles and PCBs at 350°F.
The concentration of metals in the TCLP extract compared to the total con-
centrations of metals in the soil shows that very few metals are leaching out
of the soil and, therefore, further stabilization treatment to bind the
metals would probably not be required for these soils.

     The test results for the synthetic soils are presented in Table 5.  The
data show that a high percentage of volatiles were removed at lower tempera-
tures and that semivolatiles required the higher temperature of 550°F for
more complete removal.

     The experimental results from the tray tests involves the discussion of
three separate issues.  These issues are: 1) the Superfund site samples used
in the Phase II tests; 2) the experimental activities; and 3) the analyses of
the thermally treated samples.

     The Superfund site soil samples used in the Phase II study were received
at the Illinois Institute of Technology Research Institute (IITRI) in five-
gallon containers.  Inspection of the soil indicated the samples had not been
homogenized (presence of roots, stones, and discontinuity of color) prior to
shipment to IITRI.  A decision was made at that time to not homogenize the
samples in order to prevent the loss of volatile components.

     The experimental activities for the Phase II investigation used the same
experimental approach and similar experimental procedures as Phase I and
other EPA-sponsored thermal treatment evaluations.  The experimental program
                                     145

-------
                 TABLE 4a.  SUMMARY OF METALS ANALYSES FOR
                 THERMALLY TREATED BERLIN-FARRO SITE SOILS
Parameter
Arsenic
Barium
Chromium
Copper
Lead
Nickel
Vanadium
Zinc
Total analysis
untreated
(mg/kg)
13
83
36
'117
32
22
20
63
TCLP
Untreated
0.004
0.331
0.015
0.257
0.027
0.087
ND (0.014)3
0.157
Concentration
350DF
0.0092
0.359
0.013
0.317
0.03
0.104
ND (0.014)3
:7 0.223
(mg/1)
550" F
0.0042
0.318
0.023
0.2
0.025
0.097
ND (0.014)3
0.499
TABLE 4b. SUMMARY OF METALS ANALYSES FOR
THERMALLY TREATED OLD MILL SITE SOILS
Parameter
Arsenic
Barium
Chromium
Copper
Lead
Nickel
Vanadium
Zinc
Total analysis
untreated
(mg/kg)
7.0
45
8.5
71
65
9.0
5.0
169
TCLP
Untreated
. 0.007
0.337
ND (0.01)3
0.065
0.103
0.046
ND (0.014)3
1.26
Concentration
350°F
0.0092
0.258
ND (0.01)3
0.089
0.038
0.064
ND (0.014)3
0.921
(mg/1)
550°F
0.0052
0.343
ND (0.01)3
0.092
0.074
0.096
ND (0.014)3
1.80

       w

2 Detected in blank.
q
  ND (value) = Not detected (method detection limit).
                                    146

-------














LU
1—
LU
QC
_J
__J :

1
LU
t-
O tt>
U_ O)
(O
to s-
LU O>
to >
>- res
2: _J
o
o to
z: o
0 P
rv* LU
SI
CD >-
CO
a;
«=C
E: .,
to


LO
LU
CQ
^£
• '

















t
U
2 C
•o o

o ^^
o *>».
LO O)
LO a.

u
=3
T3 C
OJ O
^ '•£
*S
LL. O>
O ^^
CD "-^
LO O>
CO 3.

,
U
3 C
•a o
QJ *r™
Q£ <4~'
^
U_ CD
O -^
CD -•*.
LO C7)
r-l a

CD
•0 J^
+-> o>
«™<
»^
(I)
S- •>
=> o
I/)












S-
OJ
0)
E
(O
J-
o.






CO CTk CTk CTt O*t CT) CT>
CTt CT) CTk CX> O> CTt C71)
A A A A

o •«* LO o*-< «— i in
CD V V f*"*.^-"^ CM
T-H LO LO
•t ^ -"- -
i-H Q O
.......


co en 01 CD crt CTi en
en en en en en en en
A

o en co o o r-i o
O LO CO CO CM LO
f»» «3- f-i- en
* V
o
to




en en en en en en en
A A

o o o o o
O O-l O-« O O
o o *— » o -— - •-* o
•> «LO «LO « »
O CO ^— -*O *— **d" O
CM »-« Q CO Q CM
z 1-1 z co


O O O O O O 0
ooooooo
CD O O O O O O
o CM «a- vo i— i co r>.
CO CM O t— 1 LO CM f^
CO CO CO «—t VO >
-l-> .C
O) 0) <1) -f->
c o c a)
O) i- tt) O.
Nl O N S- O)
C i— C O r—
tt) ^ "5 ^ O) !c 4J
tt) C O -i- C (_) O) fO
^™ O i- CD r*~ tt)  O 1 >, S- S_ O) O
4-> tt>i— CM X: >,+J i— >
i— eC t_> <-H LU to (— x* E
O O)
> to


en en LO
en en co
A A

o o o
O CD O
i-H «— 1 O
A A «t
to to co
t-l r-i LO
V V


i-l O CO
CO «* CO


o o o
o o o
o o o

en co LO
LO CM CO
CO LO CM
I— 1 «— 1



•-H r>. co
CM «-»«—!


0 O 0
0 00
o o o
-'• ft fl ft
o o o
O O CO
00 rH CO
A A
LO CM
j

0 CD O
§o o
o o
i-T t-" 7
»-H CM CO
CM LO CO
r-^ CM








^•^ ^M.
r— O
>> c
X OJ
tt> JT
^z o.
i — O
tt) >> O) 1-
C JC •«-> O
tt) 4^ (0 r—
rd 1 ro O
1- CM .C «O
4-> to .C C
C ••- Q. tt)
«C CQ d.






























•
+J
c
0
4J
IJ
tt>
"O
•o
o
.c

" -E-
•a
4J
o
tt)
tt)
•a

1 %
o
2:

II

a>
3
1

CD.

frH •'"
147

-------
was carefully structured to obtain the desired level of precision to allow
accurate interpretation of the test results.  The Quality Assurance Project
Plan sought to control and measure experimental precision by requiring a
total of four replicate runs for each test condition on each Superfund soil.

     The results from Phase II show similar reduction trends for volatile
organics as the Phase I results as illustrated in Figure 3.  Treatment of the
actual Superfund site soils (Phase II) was not as efficient as the treatment
of the synthetic soils (Phase I).  The surrogate soils, however, were highly
contaminated; a high percentage removal is easier to obtain when the original
concentration is very high.  Furthermore, the difference in removal
efficiency is likely to be related to the natural aging and weathering of the
contaminated Superfund soils used in Phase II.  The synthetic soils used in
Phase I were spiked and then treated within a relatively short time frame.
Generally, however, volatile organic compounds were readily removed, even at
350°F.  The differences in percent reduction for some of the organic results
are due to differences in the detection limits reported by the analytical
laboratory.  Higher detection limits mathematically reduce the percent
reduction for specific analytes.

     Although the quality of the semivolatile data was lacking, the trend for
semivolatile removal was very similar for both Phase I (surrogate soil) and
Phase II (actual Superfund soil).  Semivolatiles required treatment at 550°F.
Figure 4 shows the percent reduction in semivolatile organics compared between
Phase I and Phase II.  The data points from the Phase II tests bracket the
line drawn to represent the results of the Phase I test and illustrate the
similarity in effectiveness of thermal desorption in both tests.

                                 CONCLUSIONS

     Low temperature thermal desorption was tested at the bench scale on two
different Superfund site soils.  Tests were run at 350°F and 550°F, residuals
were analyzed for volatile and semivolatile organic compounds, and TCLP ex-
tracts were analyzed for metals.  The results indicate that low temperature
thermal desorption removes over 86 percent of volatileorganics  at 350°F and
over 87 percent at 550°F.  An average of over 79 percent of semivolatile
organics were removed at 550°F; semivolatile organic reduction at 350°F was
inconsistent and highly related to the vapor pressure of the compounds.
Analyses indicate that this treatment process is not effective in reducing
the concentration of metals.

     The results of the Phase II testing (Superfund site soils) of low tem-
perature thermal desorption are similar and seem to support the Phase I
results (synthetic soil matrix).  That is, significant reduction of volatile
and semivolatile organic compounds in contaminated soils was accomplished by
the low temperature thermal desorption process.  As was expected, thermal
desorption appeared to be more effective in treating the synthetic soil
matrix (Phase I) than the actual Superfund soils (Phase II).  This behavior
is likely a result of the natural aging, weathering, and biological processes
to which the Superfund soils were exposed.
                                    148

-------
0>
•5   100.
_CO
5
CO
•5
o>
o>
CO
CO
D
111
oc
     95
      90
      85
      80
                       Synthetic Soil #1
                    (high organic/low metal)
                      Berlin-Farro Site Soils
                                     Old Mill Site Soils
                I     I    I     I     I    I     I     I    I     I     1    I     I
                   100       200       300      400       500       600      700

                                 TEMPERATURE, °F

             Figure 3.  Reduction efficiency of low temperature thermal desorption
                          as a function of temperature for volatile organics.
*-^   100-
 
-------
     Because sample hold times were exceeded and method detection limits for
semivolatile analyses were exceedingly high, the quality of the semivolatile
data is questionable.  The trends for reduction that were indicated, however,
are smiliar to those seen in Phase I of this study.  Volatile organic, PCB,
and TCLP extract analyses all met the quality assurance requirements for this
project.

     Bench scale tests of low temperature thermal desorption indicate that
this technology is promising for effective treatment of soils contaminated
with volatile and semivolatile orgam'cs and PCBs.  Further investigation of
this technology at the engineering or pilot scale may be warranted.

                                 REFERENCES

EG&6 Idaho.  "Thermal Desorption/UV Photolysis Process Research, Testing and
Evaluation Performed at Johnston Island for the U.S. Air Force Installation
Restoration Program".  Document No. AD-A195613.  1988.

IT Corporation.  Laboratory Investigation of Thermal Treatment of Soil Con-
taminated with 2,3,7,8-TCDD, U.S. Environmental Protection Agency, Edison,
New Jersey.  July 1984.

IT Corporation.  "Technology Demonstration of Thermal Desorption/UV Photo-
lysis Process for Decontaminating Soils Containing Herbicide Orange".
Presented at the Spring 1986 American Chemical Society Conference, New York,
New York.  1986.

PEI Associates, Inc. and IT Corporation.  "Low-Temperature Thermal Treatment
of Surrogate CERCLA Soils - Bench-Scale Tests."  Prepared for U.S. EPA Risk
Reduction Engineering Laboratory, Cincinnati, Ohio under Contract No.
68-03-3389, October 1987.

PEI Associates, Inc. and IT Corporation.  "Alternative Treatment Technology
Evaluation of CERCLA Soils and Debris (Phase II Results), Draft Report."
Prepared for U.S. EPA Risk Reduction Engineering Laboratory, Cincinnati,
Ohio, under Contract No. 68-03-3389.   February 1989.
Perry, R.H. and D. Green, (ed.).  Perry's Chemical Engineers' Handbook.
edition.  McGraw-Hill Book Company, New York, New York.  1984.
6th
Szabo, M. F., R. D. Fox, and R. C. Thurnau.  "Application of Low-Temperature
Thermal Treatment to CERCLA Soils", Proceedings 14th Annual Hazardous Waste
Research Symposium, Cincinnati, Ohio.  May, 1988.
                                     150

-------
   DETECTION OF MACRO DEFECTS IN SOIL-BENTONITE CUTOFF WALLS

                              by

     Andrew Bodocsi, Richard M. McCandless and Koon Wah Ling
                   University of Cincinnati
        Department of Civil & Environmental Engineering
                    Cincinnati, Ohio 45224
                           ABSTRACT
     Two methods for the detection of buried hydraulic defects
or "windows" in  cutoff  walls  are described.   The first method
can be  used with existing  cutoff walls,  and  is  based on  the
monitoring  of  groundwater  levels,  employing  a  system   of
standpipes spaced at intervals along the wall.  The method  was
tested with a model  cutoff  wall, where a series of standpipes
were  used   to   measure  the  drawdown   curves  for  various
artificial "windows" of known geometry and location.  The same
conditions  were  then examined  using a  numerical groundwater
transport model  in  order to validate  the numerical model  for
use  in  a  parameter study  of field-scale  "window" detection
cases.

     Preliminary results  include calculations  of  the maximum
field spacing of standpipes that can be  used  to detect point
source leaks or  "windows" of  specified size or discharge rate
for a) varying differential head conditions across the cutoff
wall,  and  b)   varying  ratios   of  horizontal  to  vertical
permeability of the in situ soils adjacent to the cutoff wall.

     The  second  method  is  applicable  to cutoff  walls under
construction.  The  detection  system in this  case included  an
impervious  liner   on   the  downstream   side   of .the  wall,
perforated vertical  standpipes  attached to the upstream side
of the liner at  intervals,  and  strips of geotextile that were
wrapped around the  standpipes and were  extended and attached
to the liner on  each side of  each standpipe.   This system  was
capable of detecting the approximate location of a "window"  in
the  model  cutoff  wall,  and  the  existence   and  approximate
location of a tear or a hole in the liner.
                              151

-------
                         INTRODUCTION
     Soil-bentonite  cutoff  walls  may  be  used  to   contain
leachate  generated at  hazardous waste  sites as  a temporary
containment measure alone,  or  as a part of permanent  remedial
work.   A preceding study by these  authors (1987)  showed  that
during  the placement  of the  soil-bentonite backfill  into  a
model cutoff  trench,  bentonite  slurry  may be  entrapped in  a
recessed corner  (at the base of a step) or in narrow  fissures
in the bottom of the trench.   The study  also demonstrated  that
sediment on the  bottom of the cutoff trench will most  likely
be  entrapped by  the   advancing  soil-bentonite backfill   and
result  in highly  pervious  sand  "windows" in a  cutoff wall.
According  to Evans  et.  al.,   (1985)   there  is  also  a  high
likelihood that  pockets of bentonite slurry  are entrapped at
many  locations  within  the   backfill  during  construction.
Regardless of the  mechanism, as-built or long-term chemically-
induced  hydraulic  defects, or  "windows",   in  cutoff walls
represent continued environmental risk.

     This paper describes the  detection  of buried "windows" by
two  distinct  techniques.   The first method can be  used  with
existing cutoff walls  and is based on monitoring the  drawdown
in  groundwater  levels  using  standpipes  spaced  at  intervals
along the wall.    In  this  research the piezometric  drawdown
curves for various artificial  "windows" of known location  and
size  were  measured  for  a model   cutoff wall.    The  same
conditions were  then  examined using the  MODFLOW Groundwater
Transport  Model   (McDonald  and  Harbaugh,  1988)  in  order to
validate the  numerical model.    Once validated, MODFLOW   was
used  for a  parameter  study of field-scale  window detection
cases.   Results  are  reported  as the maximum field standpipe
spacings  that can be  used to detect  in  situ point-source
"windows" of  specified size or  leakage discharge  for  various
differential head  conditions across the wall,  and for  varying
site soil permeability conditions.

     The  second  method  tested  could   be  installed   on   the
downstream side of new cutoff  walls during construction.   The
system  modeled  consists  of a  continuous  impervious  membrane
liner, perforated  standpipes attached to the upstream  side of
the  liner,  and  strips  of  filter  fabric  wrapped  around  the
standpipes and also attached to the liner.  The system  can be
used to locate "windows"  in the cutoff wall and holes  or tears
in the liner.
                             152

-------
           WINDOW DETECTION BY DRAWDOWN MEASUREMENTS
              ADJACENT TO AN EXISTING CUTOFF WALL

EQUIPMENT,  PROCEDURES AND RESULTS

     Figure 1  is an  end view of the  flume apparatus used  in
this phase of the testing.   In  this model,  a  solid plexiglass
partition-" was  used  to  simulate an  impervious  soil-bentonite
cutoff  wall.   The partition was provided with  orifice  plates
at two  locations to  create various  size "windows",  or  holes.
Each  orifice   plate  contained  sixteen  23.8  mm   (15/16  in.)
diameter  holes, and each hole could be  independently opened  or
closed  to give a  specified size "window"  opening.   A  medium
masonry sand with a horizontal permeability of 2 x 10~2  cm/sec
was  used to   simulate  the  in  situ  site  sand  outside the
hazardous waste containment area.  The water table in the sand
represented the unpolluted groundwater table.    Flow through
the  orifice plate  "windows"  was  caused  by  differences  in
hydraulic head  across the  partition  (model cutoff wall)  that
were controlled by overflow ports in the downstream reservoir.

     Several 12.7 mm  (1/2  in.)  o.d. perforated standpipes were
placed  within  the  sand adjacent to  the plexiglass partition
(simulated cutoff wall)  as also  shown in Figure  1.   The water
levels  in  the  pipes  were  measured by  a 2  mm  (1/8  in.)  o.d.
flexible  nylon  tube lowered into the pipes  and  connected to a
pressure  measuring transducer on the other  end.
            Downstream reservoir          Upstream reservoir
                           Site Zone (Sand)
                 152 mm
                              30S mm
                                         7fl mm
         Overflow
          port «
         (typical)
         a 10
                  o
~l\l: Stondplpe •  .' *  '•.,
    {Varied locations)  '.-
                                                Continuous water
                                               - supply

                                               . Constant head
                                                overflow port
                                               Perforated Inflow
                                                 panel
       Drain fitting
                                       Note:. 25.4 mm " \'
      Figure 1.  End View of Window Detection Model for the Case of
               Existing Cutoff Walls.
                               153

-------
      Numerous drawdown tests were conducted on this model with
a  variety of combinations  of window openings  and differential
head  conditions.  Typical of the  experimental  drawdown curves
is the curve shown  in Figure 2.   The curve drawn through the
triangle symbols  was  measured for  a composite  window opening
(one  orifice plate) of  51.5 sq. cm.  (8.0 sq.  in.)  and a head
difference of 190 mm  (7.5  in.).   The curve drawn through the
asterisk •:symbols  (*)  was  computed  for" the  same  conditions
using MODFLOW.    Figure 3  shows  similar  results  but  for  a
double window case (both orifice plates used).

      Having  close  agreement  between  the experimental  and
numerical  results as  shown in Figures  2  and  3,  the numerical
model was then  used  for a more extensive  parameter  study of
field-scale "window"  detection scenarios.   Specifically,  the
MODFLOW  model was  applied  to varying size  "window" openings,
various   differential  head   conditions,   varying  ratios  of
horizontal to vertical permeability of the site sand  (kh/kv),
and  different  ratios of the  permeability  of  the "window  -
forming" material to  that of  the  site sand  (kw/ks).   For each
set   of  chosen  parameters,  the   field  drawdown  curve  was
computed  and  then  plotted  to  scale.     Using  a  practical
accuracy of  +/-  2.5  cm  (+/-. 1  in.)  for  standpipe  readings
measured   under   field   conditions,    and   the   graphical
construction  technique  shown   in  Figure  4,   the   maximum
allowable  horizontal spacing of standpipes  along a cutoff wall
was  established  for  each window  configuration  studied.   The
results  are summarized both graphically and in tabular  form.

      Figures 5a  through  5d  present the   maximum  standpipe
spacings   that   could  be  used  and  still  detect  the  stated
      SOO-i
    §450
    o
    o
    Nl
    UJ
    D_
      400-
      350-
Notes:
1. window size — SI.5 aq cm
2. differenUol head - 190 mm

***** computed head
/=4AAA observed head
            i i i i i 11 | 11 i 111 11 i | 11 i i i i M i |'i i t iTTTi i fi'i'i i i i i i'T f
         0      500     1000     1500    2000     2500
               DISTANCE ALONG. CUTOFF WALL A30S  (mm)

    Figure 2.  Typ.ical Comparison of Measured and Predicted Drawdown
             Curves for a Single "Window".
                              154

-------
   50CH
 §450^
 o
   400-
   350
                                   Notes:
                                   1. window size «» 51.5 sq cm
                                   2. differentiol head = 190 mm
                                   3. two windows case

                                   ***** computed head
                                   A&&A& observed head
                        I I I I I I 1 I 1 1 I 1 I I I I 1 I 1 I 1 I I I I I I
                            1500      2000      2500
      DISTANCE ALONG CUTOFF WALL  AXIS  (mm)
Figure 3.   Typical Comparison of Measured and Predicted Pravdown
            Curves For Two Adjacent '^Windows11 .
    •o
    oi
    o
    o
    s
    CL
                   Maximum Standpipe Spacing, S
                                                    	Ji25 mm
Tangent Line

to Computed
Drawdown Curve
Lower Confidence Limit

 for Field Groundwater
 Level Measurements
                  Distance along Cutoff Wall Axis


Figure 4.  Graphical Construction Technique to Estimate Maximum Allow-
           able Spacing of Standpipes Assuming a Practical Accuracy of
           ± 2.5  cm (± 1 in.) For Water Level Measurements.
                                  155

-------
1U.U -
CO
0)
Q.
Q.
TJ
a 5.0 -
•*-»
CO
«4-
^J
i '-
s. :
CO
00-
W» \J
--**
,--"•" 	
«*•*"""'
•**
— — ^"*^
^ — — """"""^

^ 	 * Notes:
_, — ^^ 1. differential head
«* 1.525 m
2. Kw/Ks « 1
— *
^ _- — — -" AWAQ - 20 I/day
j.— — "" sW«***Q « 80 I/day
. _ — — »«"" Q - 160 I/day
*• »*«««Q - 310 I/day

. . A
23456789 10
                       Kh/Kv  of  Native  Soil

 Figure 5a.  Maximum Standpipe Spacing For a.Differential Head of
              1.525 m (5.0  ft.)  to Detect Various Size  'Windows' as a
              Function of kh/kv of the Site Soils.
    15.0  -
  CO
  CD
 •§10.0

 "i
 s.
 CO
 OJ
 s.
 o
 CO
     5.0 -
    0.0
Figure 5b.

  *-
                                     __ —	*
                                   Notes:
                                   1. differentia! head
                                      = 3.050 m
                                   2. Kw/Ks » 1
                                                •*••• Q
   40 I/day
  160 I/day
- 320 l/doy
-, 620 I/day
2
3
4
Kh/Kv
'4'
of
• 1 1 1 1 1 1
6
Native
1 1 1 1 1
7
Soil
1 1 1 1 1
8
,,,,,
1
0
Maximum Standpipe Spacing For a Differential Head of
3.050 m (10.0 ft.)  to Detect Various Size 'Windows' as a
Function of kh/kv of'the Site Soils.
                                      156

-------
   20.0  -i
 w.
 Q)
 Q.
 '5.
 TJ

 iio.o
 •*-»
 CO
 D)


 I
 Q.
 CO
    0.0
                 ___•---*•""
                                               Notes:
                                               1. differential head
                                           -*     - 4.575 m
                                               2. Kw/Ks - 1

                                               /VWAQ -  65 I/day
                                               *****Q = 250 l/doy
                                                    i Q - 490 I/day
                                                    | Q - 930 I/day
TT
2
1 1 1 1 1
3
4
Kh/Kv
5
of
6
Native
•y
Soil
-r-| ri
8
TT'ri
9
i i i 1
10
Figure  5c.   Maximum Standpipe Spacing For a Differential Head of
             4.575 m (15.0  ft) to Detect Various Size 'Windows'  as a
             Function of kh/kv of'the Site Soils..
   20.0 -
  CO
  CD
  Q.
 "5.
 •g
                                                 Notes:
                                                 1. differential head
                                                   = 6.100 m
                                                 2. Kw/Ks = 1
£10.0  -
  o>

 'o
  a.
 CO

     0.0
Figure 5d.  Maximum Standpipe Spacing For  a Differential Head of
             6.10 m  (20.0 ft.) to Detect Various Size  'Windows' as a
             Function of kh/kv of the Site  Soils.
*- -
A 	
1 2
. — - — A- —
Vi""*1"
Kh/Kv
	 	 	 •
5 6
of Native
— — 	 A
7 8
Soil
A/WWQ ~ 85 I/day
MHMM Q «= 650 I/day
»»••• Q « 1250 I/day
9 10
                                       157

-------
 CD


 I-
 w II
 a
 c  o
 BJ.H
•P 4J
CO.H
 ft
   CO
   S
 fij C
sl
Q)
rH
•S
DIFFERENTIAL HEAD H (m)
o
o
H
*
10
in
c^
in
•
«*
o
in
o
•
CO
in
oa
to
•
H
£~
•3 £
to
ft CO
CO
Discharge
Q (L/day)
2
Is
ta
aco
co
Discharge
Q (L/day)
BU
* A
(0
04 CO
CO
Discharge
Q (L/day)
o>
IS
(0
ftCO
CO
Discharge
Q (L/day)
EC ~
H < &
S w
                              in
             CO  CO


             CM  CO
             H  H
in   o   o  o  o
co   r»   <*  in  m
     H   n  vo  oj
                 H
                              o\
                                  eg   in   o
                                           H
                                               in
                                               H
                              in  o  o   o   o
                              vo  n  in   en   n
                                  H  N   **   o\
co  in

O  H
                                           n   vo
                 H
                 H
                              o  o  o  o   o
                              **  CO  VO  
§£
Sfi
£*
r; W
•H C
O BJ
p<
., o
 f-l
H
g»ll
c
»fl
:*
o IH
T|

|J|

2
g^
|

fe
[Zj
1*1
H
Q



















o
0
,_)

VO






in
in

••*







o
in
o
CO










in

in
H





id
ft CO
CO

a) ~^
P*&
h rt
RJ-O
tt
W ^
•H
Q oe
&
^
•H e
0^

ft CO
CO

a) ^
os
M (d
(OtJ
'•S>
to •— •
•H
Q ty

£_
C ^-»
•Si
ft CO
CO
Q) -^
pn>»
££
•s>
ra v'
•H
o a

&i
c ^
•H g
0-2-
ft CO
CO

<1)^
0>>i
M n)
id -o
^>
01 >-*
•H
Q ~»
8< e
W 0
H § O1
S W
                                                                                        vo   vo
                                                                                                M
                                                                                                H
O   O   O  O  O
o   o   o  in  o
O   H   H  H  O
H   n   ^f  co  in
                 H
                                                                                       o   vo
                                                                                                        eo
                                                         CO
                                                              «*   O  CO  O\
                                                                  H  H  H
                                                         o   o   o  o  o
                                                         in   vo   n  r»  vo
                                                         r-   in   H  H  H
                                                              H   n  vo  H
                                                                          H
**  p

n  r»
                                                                                                H

                                                                                                vo
             r«   in
             H   H
                                                         o  o   o   o  o
                                                         o  o   o   o-  **
                                                         in  o   H   o  •«*
                                                             H   CM   <*  t»
    rf  O  H   CM


    CM  CO    vo   CM
                                                         CM  «#  en  co   r«
                                                                     H   co
                                                    158

-------
leakage in  liters/day  as a function  of  kh/kv ratio (ratio  of
horizontal  to  vertical  permeability  of the  site  soil).    In
each  figure,  kw/ks  equals  unity  for the  general  case where
site sand has  been trapped to  form the  "window".   Each graph
was prepared for  the .case of a single  window of varying  size
and for differential head conditions  that varied from  1.525  m
to 6.1 m  (5ft.  to 20 ft.),  as measured between  the prevailing
groundwater  table  on   the  outside,  and  the  leachate level
inside  the  cutoff  wall.    Note  that   since  the   reported
standpipe spacings  are for a single  window,  the spacings  are
conservative and  would also be sufficient to detect  multiple
windows  i.e.,   the presence  of  more than one  window would
broaden the drawdown  curve while  the  spacing  of standpipes
remained the same  (compare Figures 2 and 3).

     Tables  1  and 2 give the  maximum  standpipe spacings  and
flow  rates  in  liters/day  for  four  different  assumed  head
conditions  and five different  window sizes.   For  the  results
shown  in  Table  1, values  of  kw/ks  =  1  and kh/kv = 2  were
assumed.  Table 2 presents results for an  assumed  kw/ks equal
to 10, and  kh/kv  of  1.   The kw/ks ratio  of 10 implies  that the
entrapped material  in  the "window" is a bentonite  slurry  with
very  high permeability.  When  the results in the two tables
are compared,  it can  be seen  that a  "slurry window"  gives  a
much  greater flow than a sand-filled  "window",  and a larger
allowable maximum spacing between standpipes.

     As an  example of  using  either  the graphs  or  the  tables,
assume that as a result of a field pump test, a "window"  that
allows  approximately  300 liters/day  leakage is   believed  to
exist  at a site  where the  ratio of  horizontal   to  vertical
permeability  of the natural  soils is 2.0.   Also  assume  that
there  is  only one window,  that it consists  of  entrapped  sand
 (thus  kw/ks =  1.0),  and that from field measurements  the  head
difference  is  known to be 3.0 meters.  In  order to detect the
location  of the "window", it is  necessary  to know  the maximum
allowable   spacing  of   standpipes   along  the  cutoff   wall
perimeter.   From  Figure  5b  the maximum spacing of standpipes
that  could  be  used  is 7.3  m  (24   ft.).   A  corresponding
"window"  size  of 1860  sq. cm.  is indicated in Table 1.  Figure
5b-shows  that if a  site  sand with a kh/kv equal 4 (a typical
ratio  for  natural  stratified  soils)  prevails at the  waste
site,  then  the maximum standpipe spacing could be increased to
9. m  (29  ft.).

      It should be noted that  the sensitivity  of  a standpipe
system can  be  improved by increasing  the difference between
the head  and tail water levels.  This could be achieved in
practice  on a  periodic monitoring basis  by temporarily pumping
down  the  leachate level in  the  containment zone.
                              159

-------
        WINDOW DETECTION USING A SYSTEM OF INTERMITTENT
                 STANDPIPES AND IMPERVIOUS LINER

     Figure  6   shows  the  front  and   end   elevations  of  a
different   window  detection  model  used  for  this  phase  of
testing.   The  end elevation  shows a plexiglass wall  used to
model  an"' impervious  liner  that  has  been   installed  on  the
downstream  side  of  a  soil-bentonite  cutoff  wall..    Vertical
standpipes  were  attached  to  the  upstream side  of  the  liner
(plexiglass wall)  at 114 mm (4.5 in.) spacing.   The perforated
standpipes  were 12.7 mm (1/2 in.) o.d.  and  extended the full
depth of  the model  cutoff wall, each having a drain valve at
the bottom.  Wrapped around  and attached to each pipe  was a
50.8 mm  (2 in.)  wide strip of  free-draining geotextile.   The
model cutoff wall  was  102 mm  (4  in.)   thick and contained a
50.8 mm  (2  in.) diameter by 102  mm (4  in.) long  horizontal
sand  "window"  which  extended  completely through  the  cutoff
wall.  The  upstream side of the  cutoff wall was  supported by a
perforated  plexiglass wall, which  separated the wall from the
upstream  "leachate"  reservoir,  but let the "leachate"  flow
through.   The water  in  the reservoir represented the leachate
contained by  the cutoff  wall.
                 Medium sand window
                  through backfill
                     Soll:bentonlte backfill
13mm SS
Vertical
Generalized
plezomeirlc
profile • — _
Indicating
location of
burled window
..
t/e tjrji.

* • • '
\'\:


T?
PP/
•'0 '
•
r "
' '•**7i

]Lf

i * * •
1 ':-]

13 Drain valves
I3i
Front Eievatlc
Fliter-fabr
standpl
I
1; n
\
9
f
— 102— j-7fl-|
.*- Fabric wrapped standplpe

• • *»
: .* •*•
fftSJtff?*
, , * *
• •"*."" * •
AiS&^l';
rf,
~^~



^.Overflow port
Constant head
'"water reservoir
^.Sombentonlte
-^backfill
Medium sand
-"window
ponfll
11 V 	 -"•
3 Drain valve
Note: All dimensions In
25.4mm "1*
End Elevation
 Figure 6.  Schematic of Standpipe/Liner Window Detection Model for
           New Cutoff Walls.
                              160

-------
     The typical window detection test started by draining  all
standpipes  through  their  individual drain valves.    After
closing the drain  valves,  the reservoir  was filled with water
(model  leachate).    This initiated  flow through  the window.
The leachate flowed through the window, but was stopped by  the
impervious liner (the plexiglass wall) on the downstream side.
However, very slowly, the  leachate would flow laterally along
the  interface  between  the   liner  and  the  soil-bentonite
backfill  toward   the  closest  geotextile  strips  and,  once
reaching the  fabric strips,  into  the two closest standpipes.
After a few hours  elapsed time, the two closest standpipes  had
filled  up  with   leachate  to  the  level  of  the   upstream
reservoir, while the other standpipes remained  empty, or  had
filled  to  a much  lower  level.   The dashed line  in  Figure  6
shows the leachate levels  in  the standpipes at the conclusion
of  the  test.    From  the  peak  in  leachate  levels  in   the
standpipes one  could estimate the most probable location of  a
buried window.

     The same system could also be used to detect leaks in  the
impervious  liner.    To  demonstrate this,  a  6.4 mm  (1/4 in.)
diameter hole,  representing a tear  in the liner,  was drilled
in the  downstream  plexiglass  wall.   To start a leak  detection
test, all  standpipes were  filled with water to  the  leachate
level in the upstream reservoir.  After a  few hours,  the water
levels  in the standpipes adjacent to  the tear were observed to
drop off, while those in the other standpipes remained high.
The  geotextile strips   adjacent  to  the  tear  facilitated
drainage  from  the standpipes  to which  they  were   attached.
Consequently, by finding no drop in the  water levels  in a  row
of standpipes,  one could conclude that there were no  tears in
the liner.   Conversely,  if there are drops in the water level
in two  adjacent pipes,  then one could strongly suspect a tear
or hole in the  liner in the zone between the two pipes.

     The two general detection  scenarios described above would
be  initiated either by  draining  (pumping)  the  standpipes in
the  case  of   a   window  in   the  backfill,  or  filling   the
standpipes  to  check for  a tear  in  the membrane  liner.    The
time required to observe the response of the standpipes could
vary  from  days to  weeks  depending upon  numerous   variables
including  standpipe spacing,  the permeabilities of  the site
and backfill soils and  the size and location of the  defect to
be  detected.   Future  work  will focus  on  the relationship
between  these   (and  other)   variables   under  field  scale
conditions.
                              161

-------
                          CONCLUSIONS
     Cost-effective  performance monitoring  of  soil-behtonite
cutoff walls could be  implemented by  one  of  the  two techniques
described  herein.   The spacing  of  conventional  standpipes
required to detect a  point-source  leak  can be determined  in
cases  where the subsurface  geology  and hydrology  are  well
characterized,  and  where  a  site-specific  design  study  has
defined a maximum tolerable leakage discharge.

     In  the case  of  existing  cutoff  walls where leaks  are
suspected,  a  series  of  standpipes  installed  downgradient
adjacent to a portion of  the  cutoff wall  could be  combined
with localized pumping tests to locate the lateral  position  of
a window within  a few meters.  The estimated location of  the
defect  could  then be  confirmed after  seasonal  variations  in
groundwater levels and other  site-specific  factors were  taken
into account.  Once located, the defect could be eliminated  by
in situ deep soil mixing  or some other suitable technique.

     In the  case of new  cutoff walls,  a permanent  standpipe/
liner leak detection system could be  designed and  installed  as
part  of the long-term remedial strategy.   The system .could
also  serve as  a post-construction  Quality Control  check  to
determine compliance with job specifications.
                        ACKNOWLEDGEMENT
     The research described herein was supported wholly by the
Waste Minimization, Destruction and Disposal Research  Division
(WMDDRD) of the  U.S.EPA Risk Reduction Engineering  Laboratory
(RREL).  The  authors wish to  thank  Project Officer Joseph K.
Burkart  and  Work  Assignment Managers  Herbert R.  Pahren and
Walter  E.  Grube   for  their  administrative  and  technical
support.   We  also thank Graduate  Assistants  Steve Liatti and
Carl Huntsburger and the UC contract management team at Center
Hill for their contributions to this study.
                             162

-------
                          REFERENCES
McCandless,  R.  M.,  Bodocsi,  A.  and  Ling,   K.   W.,   1987.
Evaluation of Slurry Cutoff Wall/Permeable  Barrier  System With
Oraanics.  Interim Report:   EPA Contract No.  68-03-3379;  Work
Assignment #0-1, 75 pp.

Evans, J.  C.,  Lennon, G. P.  and Witmer, K.  A.,   Analysis  of
Soil-Bentonite Backfill Placement in  Slurry Walls;  Proceedings
of  the   Sixth   National   Conference  on   the   Management  of
Uncontrolled  Hazardous   Waste  Sites,   Washington,  D.   C.,
November, 1985.

McDonald, M. G.,  and Harbaugh,  A.  W., 1988.   A  Modular  Three-
Dimensional   Finite-Difference   Ground-Water   Flow	Model.
(MODFLOW),   U.S.   Geological  Survey  Techniques   of   Water-
Resources Investigation, Book 6, Chapter Al.
                              163

-------
        A FIELD TEST OF HYDRAULIC FRACTURING IN GLACIAL TILL
                                 L.C. Murdoch

                          Center Hill Research Facility
                            University of Cincinnati
                             5995 Center Hill Rd.
                            Cincinnati, Ohio 45224
                                 ABSTRACT
       Hydraulic fracturing, a method of increasing fluid flow within the subsurface,
will improve the effectiveness of several remedial techniques, including pump and
treat, vapor extraction, bio-reclamation, or soil flushing. The basic method is
straightforward: the casing of a well is perforated and fluid is injected until pressures
exceed a critical value, fracturing the material enveloping the well. Sand pumped
into the fracture holds it open and provides a high-permeability channelway suitable
for either delivery or recovery.

       Although it has been used to improve the recovery of oil for more than half a
century, hydraulic fracturing has apparently never been used to improve the
remediation of contaminated sites. The goal of this research, therefore, is to
evaluate the feasibility of using hydraulic fractures for remediation. To do so, we
must first establish whether hydraulic fractures can be created under conditions
typical of waste sites. Then, if they can be created, we will assess their use in the
practice of remediation. Our initial investigations consist of theoretical analyses of
both the creation of fractures and the flow of groundwater to or from fractures, and
bench-scale experiments of the hydraulic fracturing process in soil. Recently, we
have conducted a field test of the method at shallow boreholes in an
uncontaminated site.

      _ Preliminary results of the bench-scale experiments, the theoretical
investigations, and the field test are all encouraging. In the experiments, clay-rich
colluvium was remolded and consolidated into rectangular blocks of various water
contents. The blocks were loaded in a triaxial cell and hydraulic fractures were
created by injecting dyed fluid through a tube resembling a borehole. Hydraulic
fractures have been successfully created in all tests conducted thus far. Remarkably,
even extremely soft samples of saturated, loosely consolidated clay were readily
fractured in the bench-scale apparatus.

      Particularly interesting results have come from the field experiment, in which
ten hydraulic fractures were created at shallow depths (roughly 2m) in a tight, silty-
clay till. Subsequent excavation and mapping have yielded three-dimensional images
of the fractures. In general, they were slightly elongate in plan and they dipped
gently (14° to 25°) toward the borehole.  All but two of the tests ended when the
fractures vented at the ground surface. The largest one covered 90 m2 in plan and
extended 13.5 m from the borehole when it vented. More typically, however, the
fractures covered roughly 20 m2 in plan and extended 5 to 8 m from the borehole. A
maximum thickness of 1 cm of sand was observed in the excavated fractures.
                                     164

-------
                               INTRODUCTION
      Removing contaminants from tight soil or rock is a particularly difficult
problem because most available techniques of remediation require fluid flow either
into, or out of, the contaminated region. Where low permeabilities inhibit flow,
remediation will at best be slow, at worst it will be impossible. Hydraulic fracturing
is a method of increasing fluid flow within the subsurface that should improve the
effectiveness of a variety of remedial techniques. Recovery rates in pump and treat,
vapor extraction, or soil flushing systems, for example, could  be increased by
hydraulic fracturing. Moreover, hydraulic fractures will facilitate the delivery of
treating materials, such as nutrients for microorganisms in bio-restoration systems.

       The basic technique of hydraulic fracturing is straightforward: fluid is
injected into a borehole until pressures exceed a critical value, fracturing the
enveloping material. Sand pumped into the fracture holds it open and provides a
high-permeability channelway for either delivery or recovery. Although it has been
used to improve the recovery of oil for more than half a century, hydraulic fracturing
has apparently never been used to improve the remediation of contaminated sites.

       At the Center Hill Research Facility (Department of  Civil and
Environmental Engineering, University of Cincinnati), we are evaluating the
feasibility of using hydraulic fractures to remediate contaminated sites. The
feasibility study was funded by the USEPA in May 1987, and  we will complete the
study in September 1989;

      Theoretical analyses, lab experiments, and  field testing comprise the scope of
the currently funded project. We have conducted theoretical  analyses of the
processes involved in creating hydraulic fractures and the effects that fractures will
have on the flow of groundwater at a site. According to the theoretical results,
shallow hydraulic fractures are expected to have a sheet-like  form, measuring
several tens of meters or more in length and up to several cm in thickness. The
orientation of the fractures will be governed primarily by the  state of stress in the
subsurface. In over-consolidated soil or bedrock, high lateral  stresses will result in
sub-horizontal fractures, whereas in normally consolidated soil or fill, weak lateral
stresses will result in sub-vertical fractures. These  findings are supported by field
observations and lab experiments.

       Results of theoretical  analyses of groundwater flow, based on standard
methods used by hydrologists, indicate that hydraulic fractures could significantly
increase the yields of recovery wells. Immediately after fracturing, the yields could
increase tenfold or more. The yield diminishes with time, but even after a long time
the yield from a fractured well is more than twice  that of an unfractured well,
according to the analyses. Observations at many oil wells indicate that hydraulic
fracturing increases yields by amounts similar to those in the  analyses (Howard and
Fast, 1970).

       Laboratory experiments have been conducted to determine the soil
conditions required to create a hydraulic fracture. Hydraulic  fractures have been
created in all the samples tested thus far. Remarkably, even soft clay that is poorly-
                                       165

-------
 consolidated and saturated with water can be readily fractured. Based on the lab
 tests, we expect that hydraulic fractures can be created in most naturally-occurring
 materials.

       The most exciting results have come from a field test, where 10 hydraulic
 fractures were created from shallow boreholes (2 m depth) in uncontaminated
 glacial till. The purpose of this paper is to describe our field test of hydraulic
 fracturing. The description includes the geologic setting, method used to create the
 fractures, and the forms of the fractures determined from excavation.

       Editorial restrictions permit only a summary of the field test in the following
 pages. An extensive description of the test, however, will be available in our final
 project report to be submitted in September 1989.
                              THE FIELD TEST
       A field test of hydraulic fracturing was conducted at a site 12 km north of
downtown Cincinnati on the western side of the valley of Mill Creek, a southerly-
flowing tributary of the Ohio River. The site is on the southeastern side of an area
owned by the ELDA Company, who currently uses it as a municipal landfill.

     _  Glacial till, which is probably of Illinoian age, underlies the test site. The till
is unlithified; that is, it is uncemented and readily softens or crumbles when
moistened. Hydraulic fractures were created in two stratigraphic units within the till.
The lower unit is massive, dense (bulk s.g.: 2.29) silty-clay containing 10 to 20
percent pebbles and  cobbles. The upper unit consists of beds of silty clay and
irregular graded beds of silt, sand and gravel. The beds are flat-lying and they are
typically several dm in thickness. Seven of the hydraulic fractures were created in
the lower, massive unit,  and three (2,12 and 13) were created in the bedded unit.

       The in-situ state of stress was measured'at the ELDA site using testing
equipment developed for this project. At the depth of initiation of the fractures
(roughly 2 m), the vertical stress is 35 KPa, whereas the horizontal stress is 340 KPa.
Roughly horizontal hydraulic fractures were expected due to the large ratio (roughly
10:1) or lateral to vertical stresses.

       The till was unsaturated, but contained small amounts of local perched water
in some gravel lenses. Water contents in the silty-clay were  11 to 13 percent by
weight. Saturated hydraulic conductivity of the silty clay is between 1.5 x lO"6 cm/sec
and 1.9 x 10'7 cm/sec, whereas in silty sands and gravel it is between 1.0 x 10"5
cm/sec and 3.5 x 10"5 cm/sec, according to in-situ measurements made using a
borehole permeameter.  Other details of the characteristics of the till will be in the
final report.
BOREHOLES
      Eleven boreholes were drilled along a narrow strip trending roughly NE.
                                       166

-------
Most borings are approximately 10 m from their neighbors, except at the southern
end of the site where four borings are clustered together, roughly 2.5 m from one
another (Fig. 1). Depths to the bottom of casing range between 1.64 m and 1.95 m
for most (nine of 11) boreholes. The other two boreholes were deeper; 2.72 m and
3.81 m. The boreholes were 7.5 cm in diameter.               ,

      The boreholes were designed for the purpose of creating hydraulic fractures
in over-consolidated material. In general, a borehole consisted of a steel tube
cemented into a boring, and open at both ends. A basket was fixed to the lower end
of the casing to prevent cement from plugging the bottom of the borehole. The open
boring extended several dm below the basket and was partly filled with fragments
(cuttings) of till. A narrow notch, oriented normal to the axis of the borehole, was
cut in the wall of the boring several cm below the bottom of the casing (Fig. 2). The
notches extended 4 cm into the till.

      Roughly horizontal hydraulic fractures were expected to develop in the till,
and the design of the boreholes was intended to nucleate a horizontal fracture at the
notch. Excavations of the fractures, however, indicated that the notches were
ineffective at nucleating hydraulic fractures (vertical fractures developed in the walls
of the open boring). We conclude that deeper notches will be required to nucleate
horizontal hydraulic fractures at the borehole.
METHOD OF FRACTURING


       Hydraulic fractures were created by Halliburton Services, a subcontractor,
using equipment designed to hydraulically fracture oil wells. The equipment consists
of a truck containing a blender, a centrifugal and a positive displacement pump; two
trucks containing sand; a truck containing water; and a van containing monitoring
and control equipment.

       During our hydraulic fracturing tests, water was pumped from a water truck
and mixed with sand and chemicals-dye and a gel-in the blender. A centrifugal
pump was used for most of the injection of the mixture of water, sand and chemicals
into the boreholes. Occasionally, the positive displacement pump was used when
pressures in excess of 480 KPa were required to initiate fracturing. A backhoe was
driven next to each borehole and the blade lowered onto the wellhead to prevent
the casing from lifting during injection.

       Pumping rate ranged from 0.075 to 0.227 m3/min (20 to 60 gpm), which is
approximately the lower limit that could be maintained by the equipment. The
duration of the tests ranged from 2 to 10 minutes, and the average volume of
injected water was 0.6 rrr (150 gals). Sand was mixed with water at ratios of 0.1 to
0.2 by volume. The volume of sand in fracture 13 was 0.12 m3 (4.3 ft3), based on
calculations using many measurements of the thickness of the fracture.
 MONITORING


       A variety of techniques of monitoring the process of hydraulic fracturing

                                      167

-------
FIGURE 1. The site of the hydraulic fracturing field test.
                              Portland cement
                            3—Basket
                                 Notch
                               	Boring partly filled
                                  with cuttings
  FIGURE 2. Schematic of a borehole used in the test.
                            168

-------
have been developed within the petroleum industry (a review of monitoring
techniques is contained in our first interim report). Such techniques are crucial
because they can provide information on the size, shape, location, and orientation of
a hydraulic fracture as it is formed. This information is required to design the proper
use of a fracture in delivery or recovery. Moreover, real-time monitoring can be
used to ensure that only the intended material is cut by the hydraulic fracture.

       Three different monitoring methods were evaluated during the field test. The
simplest method involves recording the injection pressure as a function of time.
Under ideal conditions, pressure increases during initial injection and then
decreases abruptly as the fracture begins to propagate (Haimson and Fairhurst,
1970). Accordingly, the peak injection pressure is commonly regarded as an
indication of the beginning of fracturing. We measured injection pressure as a
function of time in the field and found that the form of the function is similar to the
form that occurs when hydraulic fractures  are created in rock:  pressure increased to
between 240 and 760 KPa, and then decreased abruptly. We conclude, therefore,
that the onset of hydraulic fracturing in soil can be determined by monitoring
injection pressure (our lab experiments indicate that the fracture actually begins to
grow slightly before the pressure reaches a maximum).

       Another monitoring technique involves measuring  the deformation of the
ground surface over a hydraulic fracture. The measured deformations are compared
to the results of a theoretical model that determines surface deformations as a
function of the geometry of an idealized fracture at depth  (e.g. Pollard and
Holzhausen, 1979; Davis, 1983). We tested this technique  by measuring
deformations using highly accurate tiltmeters obtained from Applied Geomechanics
Inc., Santa Cruz, CA. The equipment yielded strong signals, indicating relatively
large tilts. We  are currently inverting the tilt data and comparing the results to the
actual geometry of the fractures as a detailed evaluation of this monitoring
technique.

       The electric geophysical methods Mise-a-la-masse, Dipole-Dipole, Wenner,
and Spontaneous potential were evaluated as techniques of monitoring hydraulic
fractures at shallow depths (Steirman, 1984). Mise-a-la-masse yielded anomalies
associated with the hydraulic fractures, and it shows promise as a monitoring
technique.
HYDRAULIC FRACTURES
      A backhoe was used after the fracturing operation to dig networks of
trenches in the vicinity of the boreholes, exposing the hydraulic fractures on the
trench walls. In general, the fractures are elongate in plan and the parent borehole
lies near one end of the fracture. Most of the fractures vented to the surface and the
long axis of each fracture lies on a line between the borehole and the vent. The
major axes, measured from borehole to vent, range from 1.8 to 13.5 m, and the
average is roughly 6 m (Table 1). Accordingly, the length of the fractures averages 3
times more than the depth of their initiation.

      The areas covered by the fractures range from 2.2 to 90 m2, and the typical
area is between 20 and 30 m2. The two deeper fractures (2 and 4) did not vent and
our excavations were insufficient to determine their sizes. We were unable to create
                                     169

-------
 a fracture at one well (8) because sand plugged the casing.

       All of the hydraulic fractures created beneath level ground dip toward their
 parent borehole, and the dip angles are remarkably similar, ranging from 12° to 25°
 (Table 1). One hydraulic fracture (13) was created beneath sloping ground, and it is
 nearly horizontal.

       The three-dimensional forms of the hydraulic fractures were determined
 from detailed cross-sections mapped on the walls of trenches cutting the fractures.
 The maps and cross-sections will be presented in our final report, but they will be
 summarized here in the form of an idealized hydraulic fracture (Fig. 3). The
 idealized form for fractures beneath level ground consists of the following zones:

    Zone lr Adjacent to Borehole: Vertical fracture containing the axis of the open
        part of the borehole (Fig. 3).  The strike of the vertical fracture is
        perpendicular to the major axis of the fracture. The vertical fracture
        changes orientation abruptly  to a sub-horizontal fracture within one to
        several dm of the borehole.
    Zone 2. Vicinity of Borehole: Sub-horizontal fracture extending as much as 2 m
        from borehole (Fig. 3). The maximum extent of this zone is roughly equal to
        the depth of initiation of the fracture. The sub-horizontal fracture either
        terminates or changes orientation abruptly to Zone 3.

    Zone 3. Majority of the Hydraulic Fracture: Planar to trough-like fracture
        dipping shallowly toward parent borehole (Fig. 3).

    Zone 4. Vent: Steeply-dipping fracture intersecting the ground surface. Strike of
        the fracture is parallel to the strike of the fracture in Zone 1 (Fig. 3). The
        fracture at the vent is several dm to 1 m in length and extends to a depth of
        several dm.

       The transition between zones is abrupt and marked by a sharp change in
orientation of the fracture. Zone 2 is absent from some of the fractures, but the
other three zones occur in all the fractures created beneath level ground.

       The idealized fracture is asymmetric in plan with respect to the borehole.
The long axis of the fracture is on the  side opposite from the backhoe that was used
to prevent movement of the casing during injection.
             HIGHLIGHTS AND SHORTCOMINGS OF THE TEST
      The successful creation of hydraulic fractures in unlithified material is the
main highlight of the test. This result, combined with the results of our lab
experiments, suggests that hydraulic fractures can be created in a variety of near-
surface settings.

      The fractures that were created are several times larger than anticipated.
Moreover, we expect that the maximum size attainable by a hydraulic fracture will
increase with increasing depth of initiation. This is important because the area
                                      170

-------
  Load due to
      backhoe
          Zone 2
     Horizontal frx
                               Zone  1
                           Vertical frx
FIGURE 3. Oblique view of an idealized hydraulic fracture created during the field test.
            Based on exposures of ten hydraulic fractures in the walls oftrenches.
TABLE 1.: SIZES AND DIPS OF HYDRAULIC FRACTURES
Fracture#
       2
       4
       5
       6
       7
       8
       9
       10
       11
       12
       13
Depth
(m)

2.77
3.84
1.64
1.85
1.83
1.89
1.75
1.83
1.67
1.98
1.83
Plan Area
[m

unknown
unknown
 13.
 28
 2.2

 20
 12
  9
 30
 90
Major axis
in Plan (m)

unknown
unknown
3.6
6.4
1.8

5.5
3.3
4.1
8.2
13.5
Average
  Dip  ,

shallow
shallow
 25°
 14°
variable

 17°
 22° and 25°
 24°   „
 12°
sub-horiz.
                                     171

-------
covered by a fracture will play a key role in how effective it is in improving delivery
or recovery.

       We should caution, however, that the orientation and size of a hydraulic
fracture will be strongly dependent on the properties (primarily the in-situ state of
stress) at a particular site. Site assessment in general, and measurements of the in-
situ state or stress in particular, will be crucial in evaluating where hydraulic
fracturing could be beneficial. We are developing a site assessment tool that
measures in-situ stress in soil by creating a small hydraulic fracture.

       The major shortcoming of the test was a paucity of sand propping some of
the hydraulic fractures. The fractures will be ineffective in improving recovery if
they lack sand to prop them open. We know that it is possible to inject sand because
one fracture (13) contained several cubic feet of sand, reaching thicknesses of one
cm. In other fractures, however, sand was nearly absent even though it was mixed
into the injection fluid.

       The lack of sand is probably due to methods and equipment used to create
the hydraulic fractures. Following fracturing, sand was found in pipes extending
from the blender truck to the well head. We suspect that  the sand settled out in the
pipes during pumping and never reached the well head, and this suspicion is
confirmed by mass balance calculations at well 13. The equipment used during the
test was designed to create fractures several orders of magnitude larger than the
ones we created. Different equipment, designed specifically to create small fractures
at shallow depths, should improve the placement of sand.
                                DISCUSSION
       The orientations of the hydraulic fracture in the four zones are consistent
with a conceptual model based on experiments and fracture mechanics. According
to the conceptual model, we expect a hydraulic fracture to propagate at an
orientation that requires the least expenditure of energy. Typically, hydraulic
fractures grow normal to the direction of least principle compression in the
enveloping material. In over-consolidated till, such as at the ELD A site, lateral
compression exceeds the vertical stress due to weight of the overburden.
Accordingly, under ideal conditions (uniform loading of an isotropic material of
infinite extent) we expect sub-horizontal fractures in the till. Sub-horizontal
fractures were observed in Zone 2, but elsewhere the fractures were dipping,
suggesting that conditions are different from the ideal.

       Several conditions at the test site cause the fractures we created to differ
from fractures created under those ideal conditions. Fluid pressures acting on the
wall of the open segment of the borehole affect the stresses local to the borehole. As
a result of the fluid pressures, the circumferential stress at the wall of the borehole
diminishes, potentially  to values less than the vertical stress. Apparently the fluid
pressure resulted in local stresses that favored the nucleation of a vertical fracture in
the wall of the borehole (the fractures  in Zone 1). Vertical fractures in the walls of
open boreholes were created in our lab experiments, as well as in the lab
experiments of others (Haimson and Fairhurst, 1970; Medlin and Masse, 1979).
                                      172

-------
      Shear stresses develop at the tip of a horizontal fracture when it grows to a
length roughly equal to its depth beneath the ground surface. This type of
mechanical interaction between the hydraulic fracture and the ground  surface will
cause the fracture to turn upward and propagate toward the ground surface (Pollard
and Holzhausen, 1979; Narendran and Cleary, 1983). Apparently, mechanical
interaction results in the inclination of hydraulic fractures in Zone 3 (e.g. Narendran
and Cleary, 1983; fig. 9), and perhaps in Zone 4.

      The asymmetry of the fractures with respect to the borehole seems to be due
to vertical loading caused by the backhoe-fractures propagated away from the.
backhoe. Accordingly, it could be possible to control the direction or the long axis of
a shallow fracture by artificially loading the overlying ground surface.

      We are currently developing a theoretical model, based on principles of
fracture mechanics and fluid mechanics, that will explain the development of the
forms of hydraulic fractures observed in the field experiment.
                               CONCLUSIONS
       The conclusions of the field test include results related to the fractures that
were produced, the methods used to monitor the fractures, and the equipment used
in the fracturing process.

    1. Fracturing: Hydraulic fractures can be created at shallow depths in glacial till.
        The fractures are elongate in plan and dip gently toward their parent
        borehole. The maximum dimensions of the fractures are several times
        greater than the initiation depths.

    2. Monitoring: Injection pressure can be used to determine the onset of
        hydraulic fracturing in till. Tiltmeters can be used to monitor the growth of
        hydraulic fractures at shallow depths. Electrical geophysical methods show
        promise as monitoring tools.

    3. Equipment: Equipment used to create hydraulic fractures at oil wells can be
        used to create hydraulic fractures at contaminated sites. New equipment,
        designed specifically for creating hydraulic fractures at shallow depths in
        soil or rock, should perform better than equipment used by the oil industry.

       Field tests planned for FY 1989 are designed to evaluate new equipment
used to create hydraulic fractures in soils. One of the systems that we are developing
is based on equipment used by construction and geotechnical contractors. A
pneumatic rig is used to drive a casing to a desired depth. Then, a specially-designed
point on the casing is withdrawn and a hydraulic fracture is created at the bottom of
the open casing. The point is then replaced and the casing driven to a greater depth
where  another fracture is created. In this manner, we expect to be able to create
multiple hydraulic fractures at various depths from a single boring.
                                    173

-------
                          ACKNOWLEDGEMENTS
      This project was funded by USEPA Project 68-03-3379-2-8, and I appreciate
the support and guidance of the current USEPA Work Assignment Managers, H.
Pahren and M. Roulier. Previous WAMs, D. Keller and D. Ammon, have also
played important roles in the success of the project.

      The field test would have been impossible without the cooperation of John
Stark,the manager of the ELD A Landfill. I also thank Mark Roberts, Halliburton
Services; Don Steirman, the University of Toledo; and Gary Holzhausen, Applied
Geomechanics, for then- efforts in the project.

      I appreciate the help of Joe Wilmhoff, who prepared the illustrations.
                               REFERENCES
Davis, P.M. Surface deformation associated with a dipping hydrofracture. J.
      Geophys. Res., v. 88, pp. 5826-5834,1983.

Haimson, B. and C. Fairhurst. In-situ stress determination at great depth by means
      of hydraulic fracturing, in Rock Mechanics—Theory and Practice,
      Proceedings llth Symposium on Rock Mechanics, pp. 559-584,1970.

Howard, G.C. and C.R. Fast. Hydraulic Fracturing. SPE AIME, New York, 1970.

Medlin, W.L. and L. Masse. Laboratory investigation of fracture initiation pressure
      and orientation. Soc. Pet. Eng. J. (April, 1979, pp. 124-144,1979.

Narendran, V.M. and M.P. Cleary. Analysis of growth and interaction of multiple
      hydraulic fractures. SPE Paper 12272, presented at 7th SPE Reservoir
      Simulation Symposium, pp. 389-398,1983.

Pollard, D.D. and Gary Holzhausen. On the mechanical interaction between a fluid-
      filled fracture and the Earth's surface. Tectonophysics, 53, pp. 27-57,1979.

Steirman, DJ. Electrical methods of detecting contaminated groundwater at the
      Stringfellow waste disposal site, Riverside County, California. Environ. Geol.
      Sci., 6, pp. 11-20,1984.
                                    174

-------
  COMPUTER-BASED METHODS OF ASSESSING CONTAMINATED SITES:
                             A CASE HISTORY
              W.G. Harrar, L.C. Murdoch, P.R. Cluxton, M.S. Beljin

                         Center Hill Research Facility
                           University of Cincinnati
                            5995 Center Hill Rd.
                           Cincinnati, Ohio 45224
                                ABSTRACT
      A Computer-Assisted Engineering (CAE) system, based on an AT-style
microcomputer, has been developed to evaluate data from hazardous waste sites.
Commercially available software packages  are used to store, manipulate, analyze,
and graphically represent site information.  Utility programs, written specifically for
the project, are used to transfer data from one software package to another.  System
capabilities include the generation of maps and cross-sections showing the geology,
hydrology, and distribution of contaminants; the calculation of volumes or masses of
contaminated material; and the modeling of ground water flow and contaminant
transport.

      The Queen City Farms Superfund site has been characterized and evaluated
using the CAE system. Maps and cross-sections of the geology, hydrology, and
distribution of contaminants at the site were created. A conceptual model of the
groundwater flow and advective transport of the contaminants was established
based upon the maps and cross-sections.  Preliminary numerical modeling of the
groundwater flow and advective transport of contaminants was conducted. The
results of the numerical models confirm our conceptual model and predict a
possible scenario for the migration of ground water contaminants. The volume of
contaminants in the soil was also calculated, and maps of the top and bottom
surfaces of the contaminated zone were generated to aid possible excavation.
                                    175

-------
                             INTRODUCTION
      Investigation and remediation of hazardous waste sites results in large
amounts of information that can be stored, analyzed, and displayed using a personal
computer. We have developed a computer system intended for that purpose under
the Computer Assisted Engineering (CAE) project funded by the USEPA.  The
system can be used throughput the remedial process. It is designed to characterize a
site from data obtained during a remedial investigation and feasibility study, to
estimate the volumes of contaminated material, to aid in the design of remedial
procedures, and to evaluate the effectiveness of the implemented remedial actions.
In the following paper, we describe the hardware and software components, and
some of the basic capabilities that have been developed. A summary of one of the
Technical Assistance projects serves as an example of the CAE system.
               DESCRIPTION OF THE COMPUTER SYSTEM
      The early stages of the. project involved selecting hardware and software, and
      iping basic capabilities of the system. Additional capabilities have been
developedin response to requests from EPA regional offices.
HARDWARE
      The components of the CAE system are based on an AT-style personal
computer containing 3.6 Mb of RAM and a 40 Mb hard disk. Peripheral
components include two digitizing tablets, a drafting plotter, a screen camera, and a
high resolution color graphics display.
SOFTWARE
      The CAE system consists primarily of readily available software that we have
tailored to the analysis of problems at contaminated sites. A Database
Management System (DBS) stores information from site investigations. A
Geographic Information System (GIS) performs spatial analyses.  Ground water
Modeling Programs (GMP) analyze me movement of water or chemical
contaminants in the subsurface. A Computer-Aided Design and Drafting (CADD)
package receives information from elsewhere in the system and renders it as elegant
drawings.

      Other software has been written during the project to perform tasks that
were impossible to accomplish using commercial software. The transfer of data
between various software packages, for example, typically requires a program that
translates the output format of one package into the input format of another.
                                     176

-------
USES OF THE SYSTEM
       The DBS is the heart of the CAE system. It contains any information
generated during site investigation or remediation that is required for analysis of the
site. Accordingly, the format of the database is flexible enough to permit recovery of
the types of data needed during the analyses, yet is compact enough to facilitate the
data entry. The formats that we use vary slightly from one site to another depending
on the detail of sampling, and on the type or data that are stored. An example
format for a site sampled at several different times consists of one record, or row,
for each contaminant analyzed at each well. The records are divided into fields, or
columns, containing information on (a) Well identification, (b) contaminant name,
(c)  contaminant concentration, and (d) sampling date. Maps of the distribution of
each contaminant during each sampling period can be constructed using this format.

       The CAE system is capable of generating maps and cross-sections that
characterize hazardous waste  sites. In many cases, the maps and cross-sections are
virtually identical to those created by hand. In general, however, they are intended
only as first approximations, the details of which will be modified by trained  	
investigators. The interpolation of continuous fields from data taken at points is
fundamental to site characterization, and it is a primary role of the GIS. Inferring
fields, or plumes, of contamination from analyses of samples taken at wells is one
example of interpolation that is used extensively.

       Another application of the GIS is creating spatial or three-dimensional
models of site geology or hydrology.  The models can be sliced to generate cross-
sections or maps.

       The GIS is powerful tool for inferring continuous fields from point data.  The
difference between the GIS and other contouring packages lies in the control that
the user has on the interpolation procedure. Inferring concentrations in regions
where neighboring wells lack contamination is one difference between the GIS
method and a contouring package that uses linear interpolation.  We use the GIS to
infer that points whose nearest neighboring well is uncontaminated are themselves
uncontaminated. In contrast, the linear interpolation package infers  that points
near uncontaminated wells are contaminated at concentrations equalling a weighted
average of concentrations at neighboring wells.  Several other methods, such as
kriging, are currently available for generating continuous fields from point data. The
other methods will yield slightly different results compared with those presented
here.

       The level of confidence of a line on a map or cross-section is commonly
indicated by demarking the line as solid, dashed, or dotted. We have developed a
method of assigning line style based on how close the line is to a data point
containing known values.  Lines that are relatively close to a point are solid, those at
intermediate distance are dashed and those at great distance are dotted. The
distance used to discriminate between line styles can be rigorously determined using
a value we term the proximity index. The proximity index, Pi, is defined as the
average minimum distance between points divided by the square root of two.
According to this definition, a Pi of 1 is the minimum distance required to
completely cover an area sampled on a square grid. Solid lines are used in areas of
Pi < 1, dashed lines for  1 < Pi < 2 and dotted lines for Pi > 2. The GIS  allows us to
incorporate the proximity index into the creation of maps and cross-section. The
reader should keep in mind that the proximity index is intended only to indicate the
                                      177

-------
relative distance to a point of known value; the level of confidence associated with
that distance must be determined from the geology, hydrology, and other factors.

      Contaminated soils must be excavated at many sites during remediation. The
GIS is currently being used to design strategies of excavation. For example, it is
used to estimate the volumes of soil of various concentrations that will be generated.
Moreover, the GIS produces maps of the depths of excavation required to remove
soil contaminated with concentrations greater than allowable by EPA guidelines.

      Flow of ground water and contaminant transport are analyzed using
numerical and analytical models. A typical flow analysis involves generating files of
the location of saturated intervals from well-logs in the DBS, and inferring the
locations of saturated zones using the GIS. These data are then used to determine
boundary and initial conditions in a flow model.

      Analyses of flow and transport are conducted first to characterize the site.
Subsequent analyses, based on those done during characterization, are used to
assess potential remedial action procedures.
    APPLICATION OF THE CAE SYSTEM: THE QUEEN CITY FARMS SITE
      The CAE system has currently been used to analyze data from a half dozen
sites, with each site requiring one or several system capabilities. The Queen City
Farms site is one example that illustrates many of the capabilities, ranging from the
creation of maps and cross-sections to the analysis of ground water hydrology and
the calculation of excavation parameters.

      The site is located near Seattle, in King County, Washington. The primary
source of contamination is thought to be from ponds where industrial and municipal
wastes were stored during the 1950s and 1960s. The ponds are on the southeastern
and southwestern edges of the Queen City lake (Figure la). Other potential sources
of contamination include sludge ponds and a leachate treatment systems associated
with the Cedar Hills Landfill, and the Cedar Hills landfill itself (Figure la). Figure
la was created by digitizing several base maps of the site.
 DISTRIBUTION OF CONTAMINANTS
       Contamination has been identified as a black, oily compound in soil, and
aromatic hydrocarbons, metals, sulfates, and nitrates in ground water. We received
data from the office of USEPA Region X on the concentration of contaminants in
wells sampled during three periods from fall 1986 to spring 1987.

       Maps were created showing the distribution of contaminants after each
sampling period. For example, Figure Ib is a map of the concentration isopleths for
methylene chloride in the shallow aquifer between 3 Dec. 1986 and 8 February
1987. It shows a maximum concentration of 260 ppb in the vicinity of the Queen
City ponds, and a secondary maximum of roughly 200 ppb south of the sludge
                                      178

-------
 Queen City Forms Site and Cedar Hill Landfill
          King County. Washington
                BASE MAP
 Leachate Treatment
     System       °
                           CITY FARMS
                        DISPOSAL SITE
    Absent
o Sample Location
Queen City Farms Site and Cedar Hill Landfill
         King County. Washington
M«thylen« Chloride Concentration (ppb) Winter 1086-87
                                                        V X,
                                                             \
                                                              1
                                                              I  •
                                                              i:
                                                             i :
                                                       100 - •

                                                      • 50	'"
    FIGURE 1. a. Base map of study area showing potential sources of
        contamination and end points of cross-section line B-B'.  b. Methylene
        Chloride concentrations (maximum = 260 ppb) in shallow aquifer for
        sampling period 3 December, 1986 to 8 February, 1987.
                                          179

-------
disposal ponds. A zone of lesser contamination of methylene chloride occurs north
of the Cedar Hills landfill.

      The map (Figure Ib) was created in the GIS by analyzing data from 16
monitoring wells.  The nearly straight isopleths near the sludge pond result from the
absence of contaminants in wells west of the sludge ponds and on the east-central
side of the study area (see previous section). The proximity index for this map is 725
feet, so the solid lines indicate areas that are within 725 feet of a monitoring well;
dashed lines are between 726 and 1450 feet; and dotted lines are greater than 1451
feet from a monitoring well. A final copy of the map was created by transferring
data from the GIS to the CADD.
GEOLOGY AND HYDROLOGY
      The site is underlain by Recent alluvium and deposits of Pleistocene glacial
      ••   .  , • A*  1  1 • Pi _ ._ J A* 11 /TT?i	_ — —. ^\  T— XWA-M A***i1 4-\-i r\ \*
-------
181

-------
Conceptual Model
      The till forming the base of the shallow aquifer is absent in an area south of
the Queen Qty Farms site (Figure 1). This is important because the upper surface
of the till dips toward that area. Thus, shallow ground water flows towards the area
of absent till, where it presumably flows from the shallow aquifer downward to
recharge the intermediate aquifer.

      The piezometric surface of the intermediate aquifer slopes generally
northeastward except beneath the recharge area where it is mounded. Flow
directions are inferred to be roughly radial from the recharge area and northeasterly
elsewhere.

      The piezometric surface of the deep aquifer slopes generally northeastward.
The elevation of the piezometric surface of the deep aquifer was 10 to 30 feet above
that of the intermediate aquifer during the times when measurements were made.
Accordingly, flow in the deep aquifer is inferred to be generally northeastward, and
there should be minor flow upward from the lower to the intermediate aquifer.

      Migration of contaminants due to advective transport in ground water is
inferred to be generally southward in the shallow aquifer, downward to the
intermediate aquifer through the zone of recharge, and generally northeastward in
the intermediate aquifer.  Contamination of the deep aquifer is possible if the
hydraulic heads in the intermediate aquifer locally exceed those in the deep aquifer.
Numerical Models
      Numerical modeling was conducted to test our conceptual model and to
evaluate the possibility of flow from the intermediate to the deep aquifer in the area
of recharge. Flow rates through the area of absent till were estimated using two
techniques: a flow rate of 1.2 ft3/sec was estimated by averaging the flow rate
calculated from several cross-sections of existing water levels in the shallow aquifer;
a flow rate of 5.6 ft3/sec was estimated by measuring the area draining to the
recharge zone and multiplying by the maximum rainfall, 6.0 in/month.

      Those recharge rates (and other parameters that will be described in our
final report) were used in a 3-D ground water model, MODFLOW, (McDonald and
Harbaugh, 1983) to calculate steady-state heads in the intermediate and deep
aquifers. Streamlines derived from the head calculations indicate that flow in the
intermediate is generally toward the northeast, but locally radiates from the zone of
recharge.  Flow in the  deep aquifer is also to the northeast with local effects at the
zone of recharge (Figure 3). These results support our conceptual model.

      Analyses indicate that recharge results in mounds in  the local distributions of
head in the lower aquifers, and the relative heights of the mounds are sensitive to
the rate of recharge. The lesser recharge rate results in a mound in the
intermediate aquifer that is lower than the piezometric surface of the deep aquifer,
implying that flow is upward through the silt bed separating the aquifers. In
contrast, the greater recharge rate results in a mound in the intermediate aquifer
that is higher than the  piezometric  surface of the deep aquifer. Accordingly,
                                     182

-------
      Shallow Aquifer
N 0 500
    ''////'Till absent
Intermediate Aquifer
                                        Deep  Aquifer
                                                        N o soo
Flow direction
          Potential source of
          contamination
Possible extent of
contamination
   FIGURE 3. Flow patterns in the shallow, intermediate, and deep aquifers; the
       latter two are generated from computed heads with maximum estimated
       recharge rate applied as a point source in the intermediate aquifer. Shaded
       area in deep aquifer depicts possible extent of contaminant migration.
                                       183

-------
 downward flow from the intermediate to the deep aquifer is implied for the greater
 recharge rate. Additional modeling suggests that recharge rates in excess of 1.6
 ft3/sec could result in flow from the intermediate to the deep aquifer.

       We conclude that it is possible for the mound in the intermediate aquifer to
 exceed the height of the piezometric surface of the deep aquifer. As a result, the
 migration of contaminants from the intermediate to the lower aquifer is possible,
 according to the preliminary modeling done using the CAE system. This possibility
 is based on preliminary modeling and its validation will require obtaining additional
 data at the site.

      The modeling described above is intended solely as a demonstration of the
 CAE system.  The capabilities of the CAE system involve the use of interpolation
 procedures and models that have been tested and are generally accepted by the
 scientific community. To date, however, the accuracy of the maps generated with
 the system have yet to be varified by field checking. For this reason, we are
 currently unable to estimate how well the interpolated fields will predict, for
 example, the distribution of contaminants in the field, and we advise the reader to
 keep this uncertainty in mind when interpreting the maps. We suggest that our
 maps and models, or any other equally preliminary maps and models, be thoroughly
 checked in the field before they are used with confidence. The calculations and
 conclusions presented here are not necessarily  those of the USEPA.
                                 SUMMARY
      An AT-style computer system has been tailored to fit the needs of storing,
analyzing and displaying data from the investigation and remediation of hazardous
waste sites. The system is designed to generate maps and cross-sections of the site
geology, hydrology, and distribution of contaminants. As such, the system is
designed to characterize a site from data obtained during a remedial investigation
and feasibility study, to estimate the volumes of contaminated material, to aid in the
design of remedial procedures, and to evaluate the effectiveness of the implemented
remedial actions.

      Models of ground water flow and transport are integrated into the system to
expedite the preliminary analysis of existing hydrology, or the evaluation of
contaminant recovery.

      As an example of the CAE system capabilities the Queen City Farms
Superfund site has been characterized and evaluated. Maps and cross-sections of
the geology, hydrology, and distribution of contaminants in ground water at the site
were created. A conceptual model of the site hydrogeology and advective transport
of contaminants in a three aquifer system was developed based on the maps and
cross-sections.

      The site in underlain by Recent alluvium and Pleistocene glacial outwash,
stratified drift, and till. The hydrology consists of a shallow, unconfined aquifer
overlying a till of low permeability; an intermediate, unconfined aquifer overlying a
silt layer; and a deep aquifer that is semi-confined by the silt layer. The till forming
the base of the shallow aquifer is locally absent in an area south of the Queen City
                                     184

-------
Farms site. Migration of contaminants due to advective transport in ground water
is inferred to be generally southward in the shallow aquifer, downward to the
intermediate aquifer through .the zone of recharge, and generally northeastward in
the intermediate aquifer. The elevation of the piezometric surface of the deep
aquifer, which slopes northeastward, was higher than that of the intermediate
aquifer when measurements were made. Contamination of the deep aquifer is
possible if the hydraulic heads in the intermediate aquifer locally exceed those in
the deep aquifer.

      Preliminary numerical modeling of the ground water flow and advective
transport of contaminants was conducted to test our conceptual model and
determine critical recharge values to the intermediate aquifer that induce ground
water flow downward into the deep aquifer. The results of the numerical modeling
confirm our conceptual model and predict a possible scenario for the migration of
contaminants in the ground water. A value of 1.6 fir/sec was determined as a
critical recharge rate that induces flow into the deep aquifer.

      The volume of contaminants in the soil was also calculated, and maps of the
top and bottom surfaces of the contaminated zone were produced to aid possible
excavation.

      The capabilities of the CAE system are available to regional offices of the
USEPA under the Technical Assistance Program. Direct inquiries to Gene Harris,
RRELSTDD.
                          ACKNOWLEDGEMENTS
       Appreciation goes out to Herbert Pahren, Eugene Harris, and former project
WAM's Douglas C. Ammon and Douglas Keller for their guidance and cooperation.
Thanks goes to personnel at the EPA Region X, Seattle, Washington for providing
the data. Joseph Wilmhoff contributed his drafting skills.
                               REFERENCE
 McDonald, M.G., and Harbaugh, AW. A modular three-dimensional groundwater
        flow model. Open-File Rep. 83-875. U.S. Geol. Surv., 1983. 528 pp.
                                     185

-------
                   RESULTS AND PRELIMINARY ECONOMIC ANALYSIS
                        OF AN APE6 TREATMENT SYSTEM FOR
                            DEGRADING PCB'S IN SOIL

                                      by:

                                John A.  Wentz,
                           Michael L. Taylor, Ph.D,
                             William E.  Gallagher
                             PEI Associates, Inc.
                               Cincinnati, Ohio

                            D.B. Chan, Ph.D, P.E.,
                      Naval  Civil  Engineering Laboratory
                           Port Hueneme,  California

                               Charles J. Rogers
                     Risk Reduction Engineering Laboratory
                     U.S. Environmental  Protection  Agency
                               Cincinnati, Ohio
                                   ABSTRACT

     This  paper  describes  the  system  and  operational  procedures utilized as
well as  results  obtained when  the  APE6  chemical  dechlorination process was
scaled up  to  field-scale and employed to  dechlorinate PCB-contaminated soil
on the Island of Guam,  U.S.A.  The APEG system consisted of a steam jacketed,
mixer, steam  generating plant, and condensate collection system.  Approxi-
mately 15  cubic  yards of soil  in batches  of  1.5  to 2  cubic yards each with
average  initial  PCB concentrations of 3430 ppm Aroclor 1260 were KPEG treat-
ed.  PCB concentrations of treated soil were reduced  by more than 99.999
percent with  no  individual PCB congener exceeding 2 ppm.  The demonstration
proved the efficacy of  the APEG process to chemically dechlorinate PCB con-
taminated  soil without  the use of  DMSO  or TMH.   Field-scale demonstrations
are being  planned for Fall 1989 with  modified reagents and optimized operat-
ing parameters where the APEG chemical  dechlorination process is estimated to
cost $200-300  per ton of contaminated soil.


                                INTRODUCTION

     Halogenated chemical  contaminants  such as chlorinated dibenzodioxins
(PCDD's),  chlorinated dibenzofurans (PCDF's), and polychlorinated biphenyls
(PCB's) have contaminated  soil, water, and other matrices in various loca-
tions throughout the United States and the world.  Because many of these
halocarbon contaminants have been  found to be highly toxic in laboratory
animal  studies, human exposure is  undesirable.  To date, only limited dis-
posal or treatment options are being  developed for ^these contaminants and the
matrices they contaminate—particularly soil.  The large quantities of con-
taminated soil have created a need for a safe, cost-effective, cleanup
                                      186

-------
process as an alternative to the current practices of secured on-site stor-
age, Class B landfill ing, or incineration;

     In 1978, at the Franklin Research Center in Philadelphia, Pennsylvania,
a reagent was identified and successfully utilized to dechlorinate PCB's in
oil.1  The reagent consisted of an alkali metal hydroxide (AOH) and poly-
ethylene glycol (PEG) mixture which became known generically as APEG (alkali
polyethylene glycolate).  The U.S. Environmental Protection Agency's Risk
Reduction Engineering Laboratory (U.S. EPA/RREL) initiated further develop-
ment of the APEG chemical dechlorination process for PCB oils to include
soils contaminated with PCB's, PCDD's, and other potentially toxic, halo-
genated aromatic compounds.  Initial laboratory findings indicated that
PCB-contaminated soils could be decontaminated and .that further investigation
of the process including assessment for full-scale service was warranted.

     A typical laboratory-scale procedure for dechlorinating PCB-contaminated
soil entails mixing potassium hydroxide (KOH) and PEG-400 (average molecular
weight 400) to formulate the reagent known as potassium polyethylene glyco-
late (KPEG).  The KPEG reagent is mixed with the contaminated soil, heated to
150°C, and held at that temperature for 1 to 4 hours to allow completion of
the reaction.  Excess reagent is decanted, the soil neutralized with sulfuric
acid and rinsed two or three times with water, and the decontaminated soil
discharged.

     The PCB's, PCDD's, and PCDF's are dechlorinated in a reaction with the
APEG mixture.  The reaction of AOH with PEG-400 produces an alkoxide (ROA)
(see Equation 1) that, in turn, reacts with a chlorine atom on the aryl ring
to produce an ether and chloride salt (AC1) (see Equation 2).  Replacement of
the chlorine atom on the aryl ring with an ether linked PEG detoxifies the
compound.2  The dechlorination process is described in general terms in
Equations 1 and 2:
                            ROH + AOH •> ROA + HOH
ROA + ArCl
                                  n
ArCln_1OR
                                                  AC1
(Eq.. 1)

(Eq. 2)
Early APEG reagent formulations included solvents such as dimethyl sulfoxide
(DMSO) and triethylene glycol methyl ether (TMH).  The DMSO and TMH were
believed to serve as cosolvents to the APEG formulation to enhance reaction
rate kinetics by improving rates of extraction of the aryl halide compound
into the alkoxide phase.3'4  Later findings, subsequent to the first pilot-
scale APEG chemical dechlorination demonstration on PCB-contaminated soils,
indicated that DMSO and TMH could be removed from the APEG formulation with-
out hindering the dechlorination process or extending the reaction time.

     In June 1987, a pilot-scale APEG chemical dechlorination demonstration
was performed on a PCB-contaminated site.  The pilot-scale demonstration was
one of the earliest attempts to dechlorinate PCB-contaminated soil at pilot
scale using a reactor.  The system consisted of a reaction .vessel, electrical
heating elements, and a condensate collection system to collect moisture
driven off of the soil and KOH solution at the elevated temperatures.  The
                                      187

-------
reactor vessel consisted of a  16-inch-diameter pipe, 40 inches long, that was
loaded with approximately 35 Ib of RGB-contaminated soil per batch.  A pre-
mixed reagent formulation of DMSO, TMH, and PE6-400 was added to the soil,
then 45 percent KOH solution added separately.  The treatment parameters
utilized mimicked those used at laboratory scale.  Initial PCB concentrations
of the batches ranged from 133 to 7013 ppm and averaged 1990 ppm.  PCB con-
centrations of the treated soil batches ranged from 1.09 to 12.4 ppm and
averaged 5.6 ppm, representing an overall PCB destruction rate of more than
99.7 percent.

     The satisfactory reduction of PCB's in the soil at the pilot scale lead
to the U.S. EPA/RREL's decision to design, construct, and demonstrate the
efficacy of a larger KPEG chemical dechlorination system.  The proposed
field-scale system would be capable of treating 1 to 2 cubic yards of con-
taminated soil per batch at a remote location.


                            EXPERIMENTAL METHODS

DESCRIPTION OF SELECTED SITE FOR FIELD-SCALE KPEG DEMONSTRATION

     The U.S. Navy Civil Engineering Laboratory (NCEL) and U.S. EPA/RREL
agreed upon a U.S. Navy site for the field-scale demonstration.  The U.S.
Navy Public Works Center (USN PWC) site on the Island of Guam, U.S.A, was
selected when analytical results of the collected soil samples indicated
average PCB concentrations of 2500 ppm with "hot spots" as high as 45,860 ppm
(4.58 percent).  Soil contamination found mainly in a nearby storm drainage
ditch resulted from leaks from a transformer rework building that had been
used as early as World War II.  The waste PCB oil that was stored outside
leaked and was carried by surface runoff into the ground.

     In preparation for the field-scale KPEG treatment demonstration, a 60-ft
by 40-ft metal building was constructed on a 100 ft by 100 ft concrete pad
and was used to stockpile the 20 cubic yards of excavated PCB-contaminated
soil.  The excavated soil was screened mechanically to separate particles
1/2-inch and smaller.  Of the 20 cubic yards, approximately 15 cubic yards
passed the 1/2-inch screen and were stockpiled for treatment.   The remaining
5 cubic yards consisted of coral and rock ranging from 1/2 to 12 inches in
diameter.  The oversized material  was stockpiled separately for subsequent
special processing.

BRIEF OVERVIEW OF MIXER SELECTION

     The type of reaction vessel used for the pilot-scale demonstration
alluded to above was not conducive to the order of magnitude of size for the
proposed field-scale demonstration.   A mixer system was required that would
provide sufficient capacity and mixing capabilities for the KPEG/soil mixture
as well as provide efficient heat transfer.   The mixer was selected based
upon the demonstrated mixing range and heat transfer efficiency as determined
by mixer manufacturer facility tests and scale-up potential.
                                      188

-------
     Various mixers were evaluated at several mixer manufacturer's facil-
ities.  At each facility, prototype mixers were charged with noncontaminated
soil, DMSO, TMH, PEG-400, and 45 percent KOH.  The mixers were turned on and
heated with either hot oil or steam through the external jackets.  Physical
and operational data were collected, and the mixer was selected that afforded
the widest range of mixing and greatest heat transfer efficiency.

DMSO REMOVAL FROM KPEG FORMULATION

     During the design phase of the field-scale KPEG system, the U.S. EPA/
RREL initiated laboratory treatability studies to determine the KPEG chemical
dechlorination effectiveness on PCB-contaminated soil without DMSO or TMH.
U.S. EPA/RREL laboratory results indicated that DMSO and TMH could be removed
from the KPEG formulation without hinderance to the chemical dechlorination
process or extension of reaction time.  The removal of DMSO from the formula-
tion was also appropriate from a health and safety point of view.  The ex-
cellent solvent characteristics of DMSO, coupled with the known rapid rate of
skin penetration by DMSO, posed serious concerns for workers in the presence
of compounds such as PCB's and PCDD's.

FIELD-SCALE KPEG TREATMENT SYSTEM

     A block flow diagram of the KPEG chemical dechlorination system designed
for the demonstration on the USN PWC site on the Island of Guam, U.S.A., is
provided in Figure 1.  The diagram illustrates that the mixer was the primary
component of the system where the chemical dechlorination process occurred
and was supported by ancillary equipment to make the system functional.  An
extensive pipe network was required for interconnecting the auxiliary systems
to each other and to the mixer.  To reduce the complexity of the piping
network, a centralized pipe rack was installed.  Figure 2 provides a layout
of the site plan of the KPEG demonstration in Guam.  The site plan illus-
trates the location of the equipment, pipe rack, and exclusion zone.  The
site plan indicates that the majority of the auxiliary systems were located
outside the exclusion zone to provide easy access.  Only equipment contacting
the soil in its contaminated state or required to be located near the mixer
because of physical limitations was located within the exclusion zone.

Mixer

     The selected mixer was designed with a total capacity of 793 gallons
(106 cubic feet) and a working capacity of 490 gallons  (65 cubic feet).  The
mixer was equipped with a 2-speed, 75-horsepower motor and gear box capable
of providing mixer shaft speeds of 30 and 60 rpm.  All potentially wetted
parts were comprised of 316 stainless steel to prevent corrosion from chemi-
cal attack.  The mixer was provided with an 8-inch-diameter shaft that ran
the length of  the mixing cylinder, which was supported at each end by posi-
tive-flow nitrogen purged seals.  Extending radially from the shaft were arms
with plows that maintained a 5/8-inch tolerance from the wall.  A maximum
tolerance of 5/8 inch between the plow and wall was recommended by the manu-
facturer without substantially sacrificing mixing and heating efficiency by
creating a dead zone where caking could occur.  This 5/8-inch tolerance also
established the maximum allowable particulate size of 1/2 inch to prevent
particulate jamming between the plow and wall.
                                      189

-------
                                                CO

                                                o
                                                .0

                                                Eo
                                             _. CD
                                             So
                                             O) —
                                          ELt-O
                                             COCO


                                                "55
                                                LE

                                                CD
                                                LU
                                                QL
                                                CO


                                                CD
190

-------
                                                      CO


                                                       O


                                                      1
                                                      •c
                                                       o
w  co

£51
=3  m
                                                       !
                                                       ,0)
                                                       CO
                                                       o
                                                       CO
                                                       LL

                                                       CD
                                                       LJLJ
                                                       d.
                                                       ^


                                                       CO
                                                       3
                                                       O
o
o
            191

-------
      Figure 3 is a diagram of the selected mixer and designed features.
 Three 2-inch and one 1-inch flanged liquid charge ports were located along
 the top of the mixer for PE6-400 and sulfuric acid addition.  One 1-inch
 charge port was provided at each headwall of the mixer for water addition.
 12-inch Teflon-coated rupture disc was provided at the top of the mixer as a
 precautionary measure against over-pressurization of the mixer.  A 2-inch
 vent was also provided that vented to the condensate collection system.

      Solids were loaded into the top of the mixer through a 20-inch by
 24-inch rectangular flange.  A 16-inch flanged screen assembly with a 2-inch
 drain was provided at the bottom of the mixer for reagent draining following
 treatment.  Treated soils were discharged from the mixer through an 8-inch
 air-operated ball  valve located on the bottom center of the mixer.
     0T\e mixer cyTinder was provided with a steam jacket rated at 80 psi
 (156°C).  A manifold was provided across the top and bottom of the steam
 jacket.   During heating, steam entered the top manifold, traversed downward
 through  the jacket, and exited the bottom manifold.   The manifolds were
 designed to serve as part of the cooling system as well.  Rearrangement of
 the valves allowed for upflow of cooling water through  the jacket.

 Platform

      The entire mixer/motor assembly was mounted on  a platform to elevate the
 discharge port sufficiently to allow for placement of a  soil-collection
 hopper underneath the mixer for discharge.   The platform was designed with
 catwalks around the mixer for access.   The  catwalks  folded down for transport
 in  order that  legal  road widths were not exceeded.   The  platform  was designed
 with  an  integrally mounted jib crane for lifting drums of soil and dry KOH
 for charging into the mixer.

 Liquid Reagent Loading

      The liquid reagent  loading system  consisted of  a pallet scale  and air-
 operated diaphragm pump.   PEG-400 was placed on  the  scale  and  tared.  The
 positive displacement pump was  used  to  charge  proper quantities of  PEG-400
 into  the mixer.

 Heating  System

      The heating  system  was a  leased package steam generating  plant.  Design
 calculations based upon  approximated soil and moisture content and reagents
 indicated that a  600-lb-per-hour, 80-psi unit was required to  heat the mixer
 contents  from  ambient  temperature to 150°C within a 4-hour timeframe.  Greater
 pressure  steam (higher temperature) could not be used because  the steam
jacket rating  on  the mixer was 80 psi.  The mixer steam manifold included  a
steam pressure  relief  valve specified at 80 psi.

Nitrogen System

     The nitrogen system was provided in the design as part of the safety
consideration  in the event DMSO and TMH were not removable from the KPEG
formulation.  The nitrogen system consisted of a pressure regulator and flow
                                      192

-------
                   m
                   2
                   til
                   CO
                   CO
                                                                  til
                                                               CO
CO
               193.

-------
 controller.   Nitrogen was purged into the mixer through  the  seals  to  displace
 ambient air.   The removal of DMSO meant the nitrogen was no  longer required
 and its use was discontinued when it was learned the seals would not  become
 damaged without the nitrogen flow.

 Condensate Collection System

      The condensate collection system was designed  to collect  and  condense
 moisture vapors vented from the mixer while at  elevated  temperatures.  The
 condensate collection system consisted of a 2-inch  vent  line connected to a
 knock-out tank that was used to remove any solids which  may  have been thrown
 into  the vent line from the rigorous mixing action.   The knock-out tank
 vented  to a fan-cooled condenser where 9600-cfm ambient  air  was blown over
 the condenser coils.   The fan condenser drained into a condensate  collection
 tank, which originally vented into  a secondary  condenser consisting  of a
 copper  coil submerged in an ice bath.   Restrictions  in the line size at the
 ice condenser created back-pressure on the system, which necessitated its
 removal.   The condensate collection tank was  then vented directly  to an
 activated-carbon column for collection of any remaining  volatilized organics.

 Process  Cooling Water System

     The hot  treated  soil  contained within the  mixer cylinder  required cool-
 ing prior to  further  processing.  The  process cooling-water  system consisted
 of  a 1250-gallon water tank and centrifugal pump assembly.   Piping on the
 mixer manifolds  was revalved to allow  cooling water  to upflow  through the
 mixer jacket  and return to the  water tank.  The process  cooling-water system
 also provided  the feedwater directly into  the mixer  cylinder when  the soil
 was rinsed in  an attempt to recover reagent.

 Reagent  Collection System

     The  reagent collection system  was  included in the design  to attempt to
 recover  and reuse a portion of  the  KPEG  dechlorination reagent.  The reagent
 collection system consisted of  two  1-inch  liquid charge  ports  on the mixer
 headwa-lls, a  16-inch  screened flange assembly with a 2-inch drain  on the
 bottom of the mixer,  and two 500-gallon  steel tanks  immediately adjacent to
 the mixer.  Ideally,  following  treatment,  the soil was washed  by pumping
 process cooling  water  directly  into  the mixer through the liquid charge ports
 and allowing mixing.   The  reagent would  be allowed to drain through the
 screen assembly  and 2-inch  drain line  into the 500-gallon tanks for potential
 reuse.  The first attempt  at the soil wash and reagent drain proved futile.
Subsequent batches were not washed  or drained of KPEG reagent  prior to neutra-
 lization.  Alternative methods for  KPEG  reagent recovery would be required
 should reagent recovery prove essential to the economics of the process.

Neutralization System

     The neutralization system consisted of a pallet scale,  sulfuric acid,
and drum pump.  A stoichiometric quantity of sulfuric acid was  pumped into
the mixer to neutralize the known quantity of KOH.  The high calcium carbonate
 (CaC03)  content of the soil required additional  acid to reduce  the  mixer
content pH to  a range within 6 to 9.
                                      194

-------
OPERATION OF THE FIELD-SCALE KPEG TREATMENT SYSTEM

     The mixer was charged with 3400 Ib of PCB-contaminated soil, 1555 Ib of
PEG-400, and 285 Ib of KOH.  The mixer was turned on to high speed (60 rpm)
to mix the soil and reagents.  The vent line from the mixer to the condensate
collection system was opened, the fan condenser turned on, and the steam
generating plant ignited.  Eighty psi (156°C) steam was circulated through
the mixer jacket until the mixer contents reached 150°C, as indicated by the
thermocouple readouts.  Steam pressure was reduced to 70 psi (150°C) and the
temperature and mixing maintained for a 4-hour period.  At the completion of
the 4-hour period, the steam generator and mixer were shut down.  The fan
condenser was turned off and the contents allowed to cool overnight.

     Following overnight cooling, the mixer contents had dropped from 150° to
90°C.  Additional cooling was performed by recirculating cooling water from
the process cooling water system in the upflow manner through the mixer
jacket until the mixer contents were cooled to 50°C.  The cooling water
remained on and a stoichiometrically calculated quantity of sulfuric acid was
pumped into the mixer in 20-1b increments.  Because of the known presence of
high CaC03 concentrations in the Guam soil, additional sulfuric acid was re-
quired to adjust the pH to within a range of 6 to 9.  Samples were collected
from the sample collection port on the mixer, and the slurry pH was measured.
Additional 20-1b increments of sulfuric acid were added, and the pH measure-
ment process was repeated until the pH was within the 6 to 9 pH range.  The
strong exothermic reaction during acid addition reelevated the temperature of
the mixer contents.  The cooling water continued to pass through the mixer
jacket until the mixer temperature was returned to 45°C.  During the entire
cooling process, cooling water initially at an ambient temperature of 25°C
was elevated to 40°C, which represented a significant transfer of heat away
from the mixer.

     A soil collection hopper was placed under the mixer discharge valve.
The air-operated valve was controlled from the mixer control panel mounted on
the mixer, which was  accessible from the platform catwalk.  The mixing action
internal to the mixer cylinder directed the contents to the discharge port.
After the treated soil was collected in the soil-collection hoppers, the
hopper  lids were  securely  fastened.  The soil collection hoppers were stored
in a secured area on  site, awaiting  analysis.


                                    RESULTS

VERIFICATION OF  PCB  DECHLORINATION

     On-site analyses of untreated  and  treated  soil were  performed  by using
extraction  and gas  chromatographic  mass  spectrometric  (GC/MS) methods adapted
from EPA Methods  (SW 846,  3rd  Edition).   Corroborative  analyses  on  duplicate
samples  of  the untreated and treated soil  and  on  collected  condensate and
recoverable reagent samples  were  performed by  an  independent  laboratory  in
 the  United  States by GC/MS as  well.  Table I  provides  the  analytical  results
                                       195

-------



CD
UJ
O_
vx

a
z:
*f-
uT
to
UJ
r- 1

O
0

•%
CO
H-4 CO
O _1
co »— t
o
E co
=3 C£
CD O
feo
*T-» tn«J
3E >
j— • ifj
25
QiCJ
Hc^
g§
Q. t— 1
1—
0 0
^^ *•— *
UJ
UJ
C£ UJ
O_ CD
i— « z:
UJ
a o
z: a:
rD UJ
O OL.
u_
Q
co z:
CQ
O f—
Q. z:
UJ
U_ CD
OeC
UJ
co o;
«££
o
>— 4
1—

cc
J— •
UJ
C-?
z:
CD
o

t-H
UJ

CQ
£
i—


C -P
•r- C
C CD
O "O CftjQ ^~s
•r- C tO tO to
•P 3  — - 0
O Q. O Z5
o o
0 I.


•a
c s= cu
o 3 -p .a--^
•r- O CO CO to
•P 4- e/l _l +J
co c  >->^-^
cn c co to
fO 0 _J -P
H-> •!- (O
c: -p cn-o
(DOC
O 3 -r- CM
s- -a to
cu cu 3 •
a. s- — 6
z:


,1 ^
•t-*
c c;
•r— CU
CQ -f->
O «tJ
P^ CL)
t(_ +j
o
cn
c c

'•P 0
fO •—
S- i—
•P 0
 i-
C C


+ CM
-|*
JD •
(O O
—i z:





—1 O
•^


x:
o •
•P 0
fCj p^
ca







r- 1
CD
•
0









t-H
CO
CO
CD





r--.
in
in
01



^•^
CM
r-l
CO
I—I





CM
in
i — i



CM

r— 1
*~^
vo
CM
CO


~-^
F-H
*^ff
0
Ol
o>




T-H








si-
t-H
•
O







^—^
co
r5
t-H
CD





CO
Ol
Ol
o>




0
t— 1





ai
Ol
r-H








CO
CM
oo
CO




o




CM








vo
LO

o









^
CO






0
vo
t-H
Ol
Ol
Ol




vo
co
CM





LO
t-H








vo
o
sl-
co




0
Ol




co










1 1 1 1










t-H O - — •
co in si-
co co in
t-H CM
t-H




oo co r»
Ol Ol t-H
t-H CD VO
01 01 r*~ i
Ol Ol Ol
Ol Ol Ol




CO t-H I — VO
CD co r-x in
co co t^ co





st- t-H OO CO
o> 01 co st-
t-H t-H St- CO





^-^* ^— ^
CM CM

CO r-H O *-*
r** in vo un
r-~ vo CM — '
CO CO CO




o o o o
«d" t-H t-H si"
LO O •* t-H




st- in vo p^










i i i i i








^__^
o
«-H O CD O
t-H t-H i— 1
J*^
— CD. O Q
o z: z: z:
z:




st- O Ol sl-
1 OO VO VO VO
VO O LO VO
Ol OO CD 00




vo
"•—* 1 1 1 1





st- co in in vo
Ol s|- 00 t— i VO
sf O O O 0








^_^
LO 1 1 1 I





oo oo si-
O CM CO si" Ol
vo • • • •
CO LO Si- CO St-


00
s 	 • r** r^ rv*
oo a: i i i
1 VO t-~ CO




fli
^*
.


CJ
o.
CO
ID
C
o

?^


CO
O
CO
r~"
UJ
i |
(JU> .
to
-
^>
CO
CO
^>

>,
i.
o
•p
CO
%-
o
J3
CO
r—

•P
to
1
C

UJ
a:
o;

a.
Ill
l-l-l
CO
«_
*



t— I
,
o

^~-
^^
o
•p
cb
O
^\
-If
CO










,
rtl
tu
to
(U
c
cu

A
0)
,__ ^
I""
"*Z'
X
c

^
o
i-
o
JD.
.
-.^J
1—4
C
o
« %
*r*
to
i.
o
Q.
S~
0
0
cn

o

ai
*-~
c-
o

•p
CO

c»
S-
O)
-p

*~H

s
CM
. • .
0

^~
*^.: ;
o
•p
CO
0

CO


+
                                               TJ
                                               ai
                                               3
                                               C
                                               •r-
                                               •P

                                               O
                                               O
196

-------
                    Q.
                    (/)


                    I/)
                     cu  w -o
 CO J=  CU O
 S- -4-> -C .C
4J    f_ +J
 c  •>     cu
 CU CU     E
 o s-   •
 c: o oj o
 O M-     •«-
 O CU  >> i-
    .S-  S- -M
CQ CU  O CU
O JE -t-> E
CX +•>  CO O

r— ••>  O 4->
 CO CO -Q O
•i-     CO CU
+J J= _J
•i- OJ
 C 3  >,
•i- o .a
    s_
 CO .C T3
.E -tJ  cu
I—     C
    t-H -I- C
        CO O
  • V) +•> T-
•o cu .a +->
 cu.f:  o 3
•1->.,0     i—
 O 4->  to O
 CU CO  CU 
  I  <*-
 >>   "O .
 s- >>  cu -^
    S- +J O
        O «r-
        CU f
        S- CX
        S- CO
        o s-
        o a>
         1

    o
    4-
OJ
    -o
 >v CU
 s- •«->
 o o
 4-> CU
 CO S-
          cu

           I

          o
   cu

   o
   U
   cu

co     T  6
s- s-  >,+->
CO O  S- CO
           _  fl
           E  O
           0-H
           s-  w
          Sd
        cu

        o
        o
        O)

        V  cyj™
 ._  _  cu  co -a
 s-  s- ^  en s-
 O  O +->     CO
ja  o    -o
 CO     .C  $=   •
_|  +J 4->  CO ^O
    O T-     «5f
 >,C  3  COO
ja         o
                                 co

                                 -o
                                  c
                                  (O

                                 OJ

                                  in
                                  0)

                                  o
                                 •)->
                                  CO
                                 CQ
                                  C
                                  O)
                                  CO
                                  (U
s
es
                c
                cu

                CO
                ex
•o
 cu
 c
 o
 o
UJ

s
•o  s-
 CU>cO

 S- f-t

 §L»
 cu  s-
 S-  O

 vt  ca
 c  s-
 o  o
•r- JD
4->  CO
 CO —1
       O) •!-> CO
       S-' O *^^
       CO. CO
       ex s- vt
       E 4-> -O
       O X O
       o cu .c:
 CO
 o -o
 c  cu
 O -M
 o  s-
     o
 cu  ex
 jr:  cu
 I—  s-
       E  CD O!
       cu  c: s:
       JZ •!-

           3 a.
       S    UJ
       o -a
       r-  cu E
           c o
       CU T- S-
       S-  CO 4-
       co -l->
          ja "O
       c/)  O CU
       cu    ••->
       3  CU CX
       i—  S- CO
       CO  CU T3
       >  S CO
                             (n
                            •I—
                             to
                    ca

                    CO

                    cu

                    ca
                    o
                    CX

                   T3
                                  s_

                                  co
                                  CO
                                  (/)
                                  o
                                  U
                                  o

                                  cu
                             en   i/i
                             CO   O
                             S-   CX
                             co   E
                             >   o
                            
                                                  
                                                  CO
                                                  cu
                                                  3
s
                                                    •o
                                                   .cu
                                              -p   -I-"
                                              •r-   CO

                                              •P   £
                                              i—   4-»
                                                    CO
                                       CO        t—
                                       ca
                                       
 CX

 r
4->   -O   >5
      CO  i—
S-    N   CU
 o    >>  +J
     i—   ca
 CU    CO   C
 CD   C  -r-
 ca    ca  -a
 S-         S-
 CU   +J   O
 >    o   c
                                              c
                                              o
                                               o
                                               CO
                                                       cu
•o
 cu
•*->
 o
 CO

 cu
                                               cu
                                               c
                                               o
                                               II
                                              a
      cu

      I
                                                             o
                                                    to
                                                    cu
                                                    o
                                                    CO
                                                   ca
                           CM   CO
                                                          CO
                                                                 197

-------
 of the pre- and post-KPEG-treated soils from both laboratories,  as  well  as
 the percent PCB reduction, PCB concentration in the collected condensate, and
 PCB concentration in the neutralized KPEG reagent.   The  PCB  concentrations  of
 the untreated soil  from the corroborative analyses  were  reported as re-
 covery-corrected.  On-site PCB concentrations of the untreated soil  were not
 recovery-corrected  when reported, thus  justifying the consistent discrepency
 between the two laboratory facilities.

     The PCB analyses  in Table I  indicate that the  lowest  PCB reduction  was
 99.58  percent,  while the majority of the reduction  rates exceeded 99.9
 percent.   Low concentrations of PCB  were identified in the collected conden-
 sate.   For each 1.5-cubic-yard batch of soil  that was treated, approximately
 50  gallons of condensate was collected.   Using Batch 4 as  an  example, initial
 PCB concentration of the 3400 Ib  of  soil  charged into the  mixer  was  3778 ppm;
 thus,  the total  quantity of PCB can  be  calculated by Equation 3:
          Total  Ib  PCB  in  soil  =  Ib  soil  x
                   (Eq.  3)
 Placing  Batch 4  values  into  Equation  3  indicates  that  12.8  Ib of PCB were
 contained within the  batch.  Analysis of  the  condensate from Batch 4 indi-
 cated a  PCB  concentration of 13.81  ppm.   The  total quantity of PCB in the
 condensate can be calculated by  Equation  4:
Total Ib PCB = Gal condensate x pCB
in condensate      1 x 10
>  (ppm)
                                                                         > 4)
Again using values obtained from Batch 4, the total quantity of PCB's trans-
ferred from the soil to the condensate was 0.006 Ib of PCB.  Therefore, the
total quantity of PCB's transferred from the soil in the mixer to the conden-
sate was less than 0.04 percent of the overall quantity, assuring that the
reduction of PCB's in the soil was not the result of PCB relocation into the
condensate by steam stripping.
     The collected condensate was passed through an activated carbon system
to remove the residual PCB's.  All analyses of carbon-treated condensate were
reported as nondetectable.  Treated condensate was collected and transported
to the sanitary sewer for discharge.

     The results presented in Table I indicate the efficacy of the KPEG
treatment for the dechlorination of PCB's.  The operation of the system,
which was capable of treating 1.5 cubic yards per batch utilizing equipment
that is readily available for scale-up, indicates that scale-up to full scale
is conceivable, assuming favorable economics.  The operation of the system
was performed without major mechanical or operational  problems.  Therefore,
full-scale design need only enlarge the system and incorporate minor changes
to improve operations, particularly the materials handling aspects of the
treatment process.


                       DISCUSSION OF PROCESS ECONOMICS

     The economics of the KPEG chemical  dechlorination process must indicate
a favorable advantage when compared with other treatment and disposal
                                       198

-------
practices for the development of the process toward a full-scale system.

     The demonstration performed on the RGB-contaminated soil in Guam was
performed solely to demonstrate the feasibility and potential of the KPEG
technology.  The system was purposefully designed to be labor-intensive for
two reasons:  (1) as a demonstration focusing on the efficacy of the KPEG
reagent to dechlorinate the PCB, the system was not automated in order that
capital expenditures be reduced, and (2) the use of hands-on operation would
allow for observations and overall intimate knowledge of the system to be
acquired.  To reduce capital costs even further, all system equipment that
could be leased was utilized in place of purchased equipment.  These lease
charges were initially incorporated into the operational costs per unit of
contaminated soil, which artificially elevated the operational costs.  Other
factors included in the initial operation costs were all labor, associated
per diem, and automobile charges; major equipment items leased from the U.S.
mainland; major equipment items leased while in Guam; diesel fuel; process
chemicals; electrical consumption; and contractor profit and overhead.
Taking into account a weekly expenditure (excluding unusually uncharacteris-
tically high costs associated with performing work in Guam) where six batches
of 1.5 cubic yards (3400 Ib) of PCB-contaminated soil were treated per batch,
the operational costs were calculated by PEI to be $1700-1800/ton of PCB-
contaminated soil.

     The system utilized in Guam only demonstrated the potential use of KPEG
reagent for PCB dechlorination.  During the demonstration, no attempt was
made to optimize the reagent formulation or operating parameters.  Since
returning from Guam, U.S. EPA/RREL and PEI have continued laboratory treat-
ability studies to optimize the reagent formulation and operating parameters.

     From the data concerning a modified KPEG reagent formulation and reduced
constraints upon operation, as determined by the U.S. EPA/RREL laboratory op-
timization studies, a full-scale, portable, self-supportive treatment system
has been preliminarily costed.  The estimated capital expenditures for a
full-scale system, including equipment and construction costs, are $3.5 to
$4.5 million.  The system would theoretically be capable of dechlorinating 72
tons of PCB-contaminated soil daily, with a more realistic throughput of 54
tons per day.  Assuming that all equipment is purchased outright and mate-
rials handling is automated to reduce excessive labor, operational costs are
estimated to be $200 to $300/ton.  This cost includes all direct labor and
indirect living costs for an out-of-town operational crew, diesel fuel con-
sumption, and chemical usage.  This cost, however, does not include exca-
vation of the contaminated soil or placement back onto the ground following
treatment; therefore, the overall cost to the client will be slightly higher.

     The purpose for presenting the operational costs without including the
excavation or replacement costs was so that a cost comparison can be made
with the current treatment practice of incineration for PCB-contaminated
soils.  A telephone poll of several independent PCB-permitted incinerators
throughout the country indicated an incineration charge averaging $1713 per
ton of PCB-contaminated soil.  This incineration charge does not include
excavation; loading into Department of Transportation (DOT) .approved drums,
since the majority of incinerator facilities will not accept contaminated
                                      199

-------
 soil  in  bulk;  transportation  of  soil  to  the  incinerator site; ash disposal;
 or  any cost  associated with replacement  of excavated soil with clean fill.


                                 CONCLUSIONS

      Based on  the projected cost estimates that have been performed, the APEG
 chemical dechlorination  process  appears  to be superior in economics, assuming
 the optimizations obtained in the  laboratory are suitable for full-scale
 scenarios.   Advantages of the APEG process are that transportation costs of
 contaminated soil and replacement  costs  of clean fill are eliminated, since
 treatment is performed on site and treated soil should be suitable for place-
 ment  back onto the ground.  Prior  to  initiating design of the full-scale
 system,  additional field-scale demonstrations will be performed utilizing the
 optimized parameters determined  by U.S.  EPA/RREL, NCEL, and PEL  These
 field-scale  demonstrations are currently being planned for Fall 1'989.


                              ACKNOWLEDGEMENTS

      PEI is  grateful to  the technical support and guidance provided through-
 out the  course of the KPEG. field-scale demonstration and on-site sample
 analysis on  the Island of Guam,  U.S.A., by Dr. Alfred Kernel and Mr. Harold
 Sparks of the U.S. EPA Risk Reduction Engineering Laboratory, Cincinnati,
 Ohio.  PEI is also grateful to Mr. Gorman Dorsey for use of the USN PWC FENA
 laboratory where on-site sample  analysis was performed as well as Mr. Jess
 Lizama, Supervisory Environmental  Engineer of the USN PWC site, Guam, for
 arrangements of construction equipment, supply of utilities, and USN PWC
 personnel for assistance of systems assembly and disassembly.


                                 REFERENCES

 1.    laconianni, F. J.  Destruction of PCBs—Environmental Applications of
     Alkali  Metal Polyethylene Glycolate Complexes.  Prepared for U.S.  En-
    vironmental Protection Agency, Hazardous Waste Engineering Research
     Laboratory, Cincinnati, Ohio.  Franklin Research Center, Philadelphia.
     May 31, 1985.

 2.   DeMarini, D. M., J.  E. Simmons.  Toxicological Evaluation of By-Products
     From Chemically Dechlorinated 2,3,7,8-TCDD.   Accepted for publication in
     Chemosphere, 1989.

3.   Peterson, R. L., E.  Milicic, and C.  J.  Rogers.  Chemical Destruction/
     Detoxification of Chlorinated Dioxins in Soils.   In:   Incineration and
     Treatment of Hazardous Waste, Proceedings of the Eleventh Annual Re-
     search Symposium.  EPA/600/9-85/028, September 1985.

4.   Peterson, R. L.  1986 Method for Decontaminating Soil,  U.S.  Patent
     Number 4,574,013, March 4,  1986.
                                      200

-------
         DESTRUCTION OF CYANIDES IN ELECTROPLATING WASTEWATERS USING
                              WET AIR OXIDATION
                  by:   H.  Paul Warner
                       USEPA
                       Cincinnati, OH  45268
                                   ABSTRACT

      Many of the technologies normally applied for the destruction of cyanide
in waste streams containing cyanides and metals result in the generation of
sludges which contain high concentrations of cyanide (200 to 1000 mg/kg), most
of which is strongly complexed with constituent metals.  In order to reduce
the total cyanide content of these sludges, we investigated the possibility of
treating the original wastestream by a technology which would destroy the
cyanide, both free and complexed, prior to precipitation of the metals.  This
paper presents the results of the application of wet air oxidation for cyanide
destruction prior to sludge generation.  Experience and engineering judgment
strongly suggest that this technology could also be applied to liquids and
sludges generated by other technologies which contain high concentrations of
cyanides.

INTRODUCTION

      Pursuant to Section 3004  (m) of the Resource Conservation and Recovery
Act  (RCRA), enacted as part of the Hazardous and Solid Wastes Amendments
(HSWA) on November 8, 1984, the Environmental Protection Agency (EPA) is
investigating alternatives for treating cyanide-containing electroplating and
heat treating wastes prior to placement in landfills.  The Agency has
previously established treatment standards for metals in the sludges from
these wastes with the First Third Listed Hazardous wastes  (53 FR 31137, August
17,  1988).  Treatment standards for the wastewater from these wastes were soft
hammered with the First Third rule.  The Agency is now developing standards
for  all of these wastes and will, with additional regulations, set standards
for  cyanide.  Cyanide standards were reserved by the Agency in the First
Thirds  rule.  This paper will discuss the application of wet air oxidation as
one  of  the technologies which can be applied for treatment of cyanide-
containing electroplating, specifically, spent plating bath  (F007).

      For the discussion that follows, it must be pointed out that data
related to the raw waste selected for treatment have been claimed as
Confidential Business Information  (CBI) by the generator and cannot be used  in
                                      201

-------
 this presentation.   However,  by using  constituent  concentrations  of  a
 "typical" F007 waste,  a  relatively  accurate  evaluation  of  this  treatment
 process  can be made.   The  constituent  concentrations  of  this  "typical"  F007
 waste will be the average  of  other  F007  wastes  as  found in the  Agency's
 proposed background  document  for  the regulation of  cyanides (1).

      It should also be  pointed out that prior  to  the pilot scale work, a
 bench scale run was  made which, due to the corrosivity  of  the raw waste,
 mandated the use of  titanium  as the material of construction  for  the treatment
 system.  Only the smallest scale  system  was  contructed  of  titanium and
 therfore was chosen  for  this  project.

 APPROACH

      Alkaline Chlorination is one  of  the most  commonly applied treatment
 technologies for F007  wastes, and is usually followed by precipitation,
 clarification, and finally dewatering  of the generated  sludge.  In a well
 operated system, the final liquid discharge  meeting effluent  discharge
 concentration regulations  may be  directly discharged  to a  surface stream or
 Publicly Owned Treatment Works (POTW).   However, the  dewatered  sludge will
 normally contain, along  with  varying concentrations of  regulated  metals, high
 concentrations of total  cyanide (200 to  1000 mg/kg).  These cyanide
 concentrations would more  than likely  restrict  continued land disposal  of the
 dewatered sludge without additional treatment.   Taking  into consideration the
 necessity of disposal, alternative  treatment techniques  were  evaluated.  Wet
 air oxidation of F007  wastes  has  been  successfully  demonstrated by
 Zimpro/Passavant in  Casmalia, California. Their technology was  selected for
 further  evaluation,  and  for the generation of data  in support of  the Agency's
 regulatory program.

 WET AIR  OXIDATION PROCESS  (2)

      Wet air oxidation  is the liquid  phase  oxidation of organics  or
 oxidizable inorganic components at  elevated  temperatures and  pressures.
 Oxidation is brought about by combining  the  wastewater with a gaseous source
 of oxygen (usually air)  at temperatures  and  pressures in the  range of about
 175° to  327°C (360° -  620°F)  and  2069  to 20,690 kPa (300 - 3,000  psig),
 respectively.  The solubility of  oxygen  in aqueous  solutions  is enhanced at
 elevated pressures, and  the elevated temperatures provide  a strong driving
 force for oxidation.

      Wet air oxidation  has been  demonstrated at bench-scale, pilot-scale, and
 full-scale as a technology capable  of  breaking  down hazardous compounds to
 carbon dioxide and other innocuous  end products.  Cyanide  in  electroplating
wastes is converted to carbonate  and ammonium ions  when oxidized  as  shown by
 the reactions:

            2 MTaCN + 02 + 4 H,0 =  Na2COa + (NH^COj

      The major processing steps  in the  wet  air oxidation -process  are
wastewater pressurization/air compression, preheat, reaction, cooling,
depressurization, and  liquid/gas  separation.  Figure  1 is  a flow diagram of
the wet  air oxidation process.
                                      202

-------
r
in
CO
0
o
o
                                                                                                                    RJ
                                                                                                                   13

                                                                                                                    X
                                                                                                                   o
                                                                                                                   <

                                                                                                                   *"*
                                                                                                                    ©
                                                                                                                    c
                                                                                                                    (0
                                                                                                                    >
                                                                                                                    C0
                                                                                                                    in
                                                                                                                    in
                                                                                                                    rt
                                                                                                                   CL
                                                                                                                    Q.
                                                                                                                    E

                                                                                                                   N
                                                                                                                    O

                                                                                                                    o

                                                                                                                    4^
                                                                                                                    co

                                                                                                                    E
                                                                                                                    
-------
       The wastewater or slurry is brought to system pressure by a high
 pressure pump.  Air from a compressor may be added directly to the waste or to
 dilution water and preheated to raise the temperature of the mixture at the
 reactor base such that the exothermic heat of reaction will increase the
 mixture temperature to the desired maximum.   Preheating can be accomplished by
 an external heat source as shown in Figure 1 or by the reactor effluent.
 Startup energy is provided by the external heat source to the preheater or an
 auxiliary heater.  Residence time for the oxidation reaction is provided by
 the reactor; the temperature of the wastewater-air mixture rises as the
 reaction occurs.  The reactor effluent is cooled with cooling water (as shown
 in Figure 1) or with the wastewater-air mixture.  Cooling is usually to about
 95° to 135°F (35-57°C).   A control valve reduces the pressure of the oxidized
 liquor-spent air mixture.   The gas phase is  disengaged from the liquid phase
 in the separator vessel.  Off-gas from wet oxidation systems is usually
 treated to reduce the concentration of hydrocarbons.  Wet scrubbing,  which is
 commonly used to cool the gas stream, results in some reduction of
 hydrocarbons.   Adsorption columns and afterburning provide additional organic
 emissions reduction.

       The overall F007  treatment system at Zimpro/Passavant may be described
 in three operations:   (1)  a blending step to control feed parameters;  (2 wet
 air oxidation process;  and (3)  treatment of  oxidized liquor.

 Feed Blending Operation

       Four fifty-five gallon drums of F007 waste were mixed in a water-heated
 stainless steel tank  to  ensure  a homogenous  feed composition.   The waste
 required heating to about  110°  to 130°F (43-54°C)  to maintain its liquid
 state.   Below that temperature  range,  the waste crystallized  because  of  the
 high concentration of sodium carbonate,  which made handling very difficult.
 After the waste was thoroughly  mixed,  the drums were refilled and placed in  a
 hot  water bath,  maintained at 110°F (43°C).   As waste was  needed in the
 performance of  the test  run,  a  drum of  waste  was removed  from the water  bath,
 thoroughly agitated with a portable mechanical  mixer,  and  then pumped  into the
 treatment system feed tank.   The feed tank is  equipped with a  heating  coil,
 mechanical mixer, and a  recycle line to  ensure  the feed is  maintained  at  the
 proper  temperature and is  homogenous.

 Wet  Air  Oxidation Treatment

      The steady-state operating conditions of  the Zimpro/Passavant wet  air
 oxidation process are shown  in  Table 1.   All  the parameters listed  in  Table 1
 are  key  operating parameters; however, the single-most important  parameter
 used  in  determine steady-state  was  maintaining  the reactor  outlet  temperature
 of about  450°F  (232°C).

      Operating  phases of  the wet air oxidation  process include:  warm-up
 {with tap water  followed by waste feed),  stabilization, steady-state,  and cool
down  (with  tap water).  The warm-up  period with  tap water typically requires
about four  hours  followed by an  additional two hour warm-up with  the
waste feed.  Waste feed warm-up  continues until  the operating  conditions were
                                     204

-------
within approximately 5 percent of the desired conditions, which signals the
beginning of the stabilization period.  During the stabilization period the
operating conditions were continually fine tuned.  Typically, 1 to 3 residence
times ( i.e., about 1 to 3 hours) were allowed before steady-state is begun.
The steady-state period continues as long as needed to collect all the
required samples or until the system falls outside steady-state conditions
specified in Table 1.  Following steady-state, the system is switched to tap
water and cooled down over a period of about 6 to 8 hours.

TABLE 1.  SELECTED OPERATING PARAMETERS OF THE WET AIR OXIDATION PROCESS
      Operating parameter
Steady-State conditions
Reactor inlet temperature

Reactor outlet temperature

Reactor pressure

Waste feed rate

Dilution water feed rate

Dilute nitric acid feed rate

High-pressure air injection rate

Residual oxygen content of the off-gass
430° - 470°F

440° - 480°F

1700 psig

2.5 - 3.0 gal/h

2.5 - 3.0 gal/h

0.5 - 0.6 gal/h

60 -80 scfh

16 - 20%
      In the wet air oxidation treatment process dilution water (i..e., tap
water) is pumped at a rate of 2.72 gal/h (1745 psig) and combined with 1.1
fta/min  of  compressed air prior to passing  through an oil preheater (Figure
1).  The preheater heats the tap water/air stream to about 520°F (271°C) as it
enters the base of the pilot-scale titanium reactor at 1710 psig.
Approximately 2 feet from the bottom of the 3-in. i.d., 15-ft-long reactor,
the waste feed is pumped into the reactor at a rate of 2.75 gal/h.  Total
influent flow rate to the reactor is 5.47 gal/h.  Heat tapes, spaced evenly
along the length of the reactor at 15-in intervals, along with the heat of
reaction maintain the temperature at about 450°F  (232°C).  The oxidized liquor
exits the reactor through one of two exit ports.  One port is used as the
reactor outlet for the oxidized liquor, and 18 percent nitric acid is pumped
at a rate of 0.55 gal/h  (1735 psig) through the other port to remove any
carbonate plugging at the exit port.  Every four hours through the test
run, the valves controlling the exit ports are reversed to ensure that the
reactor does not plug.  From the reactor, the oxidized liquor passes through a
tube-in-tube water cooler that brings the temperature down-to about 106°F.
After the cooler, the liquor passes through a pressure control valve that
returns the wastewater to atmospheric pressure.  The wastewater then enters a
gas/liquid separator.  The oxidized liquor is collected from the bottom of the
                                      205

-------
separator and the off-gas passes through a caustic scrubber containiing a 5
percent NaOH solution.  The oxygen content of the off-gas is monitored
continuously with an Oj  Meter,  and a dry gas meter measures  the off -gas flow
rate prior to release to the atmosphere.  Off -gas samples before and after the
scrubber are analyzed for methane, and total hydrocarbons by a gas
chromatograph .

Treatment of Oxidized Liquor

      In order to discharge the oxidized liquor to the POTW, the wastewater
required neutralization and metals precipitation.  This was accomplished by
the following process by use of nitric acid and sodium sulfide:
1.

2.
4.
5.
            The oxidized liquor was pumped into a 100-gallon holding tank.

            Nitric acid was added for pH adjustment.  Two types of acid were
            used:  a) spent acid from the reactor acid wash or b)
            concentrated, 38 degree Baume1 acid.

            Sufficient acid was added to lower the pH to 8.0 to 8.5.

            NOTE:  Care must be taken when the acid is added because a large
            quatitity of carbon dioxide is liberated and the reaction is
            somewhat violent.

            Sulfide in the form of sodium hydrosulfide (NaHS) or sodium
            sulfide  (NaS) -was added to the tank and mixed.
            The sulfide precipitates are very difficult to filter; therefore,
            diatomaceous earth was added as a filter aid to the slurry to
            improve filterability.

            A sample of slurry from the tank was filtered and a chip of
            sulfide was added to the filtrate.  If the filtrate remained
            clear, the metal ion precipitation was complete.  If a brown
            precipitate appeared, more sulfide was required.  Additional
            sulfide was added t'o the holding tank until the filtrate remained
            clear when a sulfide chip was added.

            A small plate and frame filter press was used for the filtering.
            The press cloths were precoated with diatomaceous earth prior to
            filtering the sulfide solids to improve filterability and prevent
            the sulfide solids from blinding the filter cloth.  Filter cake
            generated from the treatment of the oxidized liquor was disposed
            of at a hazardous waste landfill.
Treatment Operational Problems

      The most significant operational problems were  plugging  and/or  scaling
within the oil preheater and  reactor  .  These problems were anticipated prior
to the initial run due to high suspended  and dissolved solids  concentrations
                                      206

-------
in the raw waste and did, in fact, result in the termination of the initial
run.  There were other operation problems (valve and guage failure), however
they were more than likely due to the fact that the pilot systems had not been
operational for over a year.  After the initial run termination, modifications
to the reactor (removal of internal mixing baffles), the oil preheater
(repiping around the preheater), and further dilution of the raw waste,
allowed quite successful completion of the second run.  The sampling program
required 24 hours of steady state operation. An engineering representative of
Zimpro/Passavant estimated that the system could be operated for from 7 to 10
days before requiring a complete system acid purge.

      Following the completion of the sampling period, the treatment system
was allowed to cool and then the reactor was broken down to visually check the
plugging/scaling effects of the second run.  A brown, sand-like deposit was
found at the bottom of the reactor.  This material did not adhere to the
reactor walls, but was deposited loosely at the reactor bottom.  This would
indicate that it was held in suspension in the reactor during operation and,
given sufficient operating time, would probably discharge from the reactor
with the oxidized liquor (3).

      A second type of material found in the reactor was a scale deposit on
the reactor walls.  This material adhered to the reactor walls and would not
discharge from the reactor during operation.  Removal of this scale would be
required by the previously mentioned acid purge after a yet undetermined
operating period.  Neither the loose material in the bottom of the reactor nor
the reactor scale should prevent effective treatment of F007 waste by the wet
air oxidation technique.  However, it could lower the heat transfer
coefficient.

RESULTS AND CONCLUSIONS

      As previously mentioned, in order to evaluate the efficiency of wet air
oxidation in this presentation, a "typical" F007 waste constituent
characterization had to be generated for the feed stream (raw waste data
protected by CBI).  Incorporating this "typical" waste data in the calculation
of percent removal for cyanide, 99.9% removal is observed for total and
amenable cyanide as seen in Table 2.  The destruction of cyanide in the wet
oxidation process increased ammonia concentrations from minimal in the raw
waste to an average of 5400 mg/1 in the oxidized liquor.  Cyanide was not
detected in the off-gas scrubber water above its practical quantitation limit
of 0.25 mg/1.

      As seen in Table 2, a considerable amount of the copper and zinc is
apparently removed by the wet oxidation process.  A mass balance around the
process, including concentrations of the metals in the scale, bottom solids,
and the acid wash following completion of the test run, reveals that the
metals are concentrated in the scale and the bottom solids.  In the mass
balance calculations, metals concentrations from the actual raw waste were
used.  However, as previously pointed out, the raw data are not presented
because of a claim of CBI by the generator.
                                     207

-------
TABLE 2.  PERCENT REMOVALS ACROSS OXIDATION/PRECIPITATION PROCESS
F007*
Raw Feed
mg/1
CN, 33000
total
CN, 31000
amenable
Cu 5000
Zn 10000
Oxidized Percent Filter Press Percent
Liquor Removal Filtrate Removal
mg/1 mg/1
2.51 99.9 1.65 99.9
0.05 99.9 ND 100
802 84.0 2.45 99.9
6.47 99.9 2.45 99.9
*Rounded averages from other F007 wastes.
ND=Not detected.

      Conclusions from this study are as follows:
            1.
            2.
REFERENCES

Document    1.



Report      2.




Report      3.
Wet air oxidation is an effective treatment method for the
destruction of cyanides in F007 wastes, including complexed
cyanides.

Wet air oxidation, when followed by sulfide precipitation of
metals, is an effective treatment system for complete
treatment of F007 wastes.

Engineering judgment and years of experience predict that
other cyanide wastes containing metals, both liquids and
sludges, could effectively be treated, after appropriate
concentration or dilution, by the wet air
oxidation/precipitation technology.
USEPA, Office of Solid Waste, Proposed Best Demonstrated
Available Technology (BDAT) Background Document for Cyanide
Wastes, December 1988.

USEPA, 1988 of, Office of Solid Waste, On Site Engineering
Report of Treatment Technology Performance and Operation for
Wet Air Oxidation of F007 at Zimpro/Passavant, Incorporated
in Rothschild, Wisconsin, Washington, DC.

Zimpro/Passavant, Final Report for the Pilot Plant
Demonstration Study on Wet Air Oxidation of F007
Electroplating Cyanide Wastes, Rothschild, Wisconsin, June
1988.
                                      208

-------
         DETERMINING COST EFFECTIVE APPROACHES TO THE ENVIRONMENTAL
                    CONTROL OF ELECTROPLATING OPERATIONS

                by:    John 0. Burckle
                       U.S. Environmental Protection Agency
                       Risk Reduction Engineering Laboratory
                       Cincinnati, OH  45268
                                  ABSTRACT
     The U.S. EPA has sponsored a number of studies over the  past  few years
to determine improved approaches available to achieve more  cost-effective
control  of  electroplating  operations.    These  studies  indicate  that  a
multimedia process  systems approach is likely to  be  the most effective  in
attaining the environmental   pollution goals required by regulatory agencies
while minimizing costs to do so.  The approach involved requires  a  number of
steps.   These include:  (1)  the careful  analysis  of  the plating  processes
utilized and definition of the characteristics and the pathways of  pollutant
generation;  (2)  the  analysis  of the contribution  of pollution prevention
techniques to  minimize the  quantity  of  wastes formed;  (3)  selection of  a
combination of pollutant controls and ultimate  disposal technologies which are
allowable under the FWPCA, CAA and RCRA;  (4)  and  finally, the  development of
a route  to achieving the  required  environmental  protection goals  at  least
cost.   This  approach  is  based upon an iterative  process  which  takes  into
consideration  the  capital and  operating  costs of  a  number of  alternative
pathways consisting of various combinations of pollution prevention, control,
and  ultimate disposal  alternatives.  The target  sought is the  scenario  in
which the multimedia environmental goals  are  obtained  for all  aspects of the
operation at the minimum annualized cost  for capital and operating  expenses.
An acceptable target, from the perspective of cost, is that  scenario in which
the  annualized cost  for  installation  of pollution prevention  and  control
technologies,  including the  costs of residual management,  is  offset  by  the
savings achieved in  raw  materials wastage and waste treatment and disposal
costs for the existing waste management case.
                                    209

-------
                                INTRODUCTION
     This  paper attempts  to  summarize  the  results of  the  work we  have
conducted  over  the last several years to  address pollution control  in  the
electroplating sector on a multi-pollutant, multimedia basis.  Currently the
Agency promulgates standards on a media-specific basis, i.e., for air,  water,
and hazardous wastes.  The  current  standards-setting process does not provide
a "systems analysis" approach to minimizing residual  pollution discharged to
the environment for all media or all process pollutants.  The requirements for
each media are considered  separately  and in a stepwise fashion.   It was the
purpose of our efforts to  develop  an  integrated,  systematic approach  to the
selection  of various techniques for  the reduction and treatment  of  wastes
which would  optimize the  effectiveness  of reducing the  overall  multimedia
impacts.

     The discharge of wastewaters generated in electroplating processes into
receiving  waters  or sewers  are regulated  at the  Federal  level  under  the
Federal Water Pollution  Control  Act.    Regulations  to achieve the effluent
limitations established for electroplating  operations require the use of Best
Applicable Technology,  "BAT", for existing  sources  and New Source Performance
Standards, "NSPS",  for new sources.   The effect of these  regulations  on the
treatment  of electroplating wastewaters are  illustrated  in Figure 1.   The
National Pollutant  Discharge  Elimination System,  "NPDES",  requires a  permit
for  any discharge  into a receiving waterway (or  tributary) and  imposes
requirements for pretreatment of wastes discharged to sewers.
    SKIMMED OILS
                                                             'ASTE TOXIC ORGANICS
                                                                   HAUL OR
                                                                   RECLAIM
                                                                   ^ TREATED
                                                                   EFFLUENT
                                             CONTRACTOR
                                             REMOVAL
   Figure 1.  Treatment Required by the Federal Water Pollution Control  Act
              (BAT Level-NSPS is the same except for requirements for
               recovery of Cadmium)
                                    210

-------
     The disposal of residual  wastes to the land is regulated under provisions
 of the  Hazardous and Solid Waste Amendments of 1984.   These  amendments of the
 Resource Conservation and Recovery Act require that the Agency ban RCRA wastes
 from land  disposal  unless  it  is determined  that   land  disposal   is  more
 protective of the public health and the environment than available alternative
 technologies.   This act also  requires that  the  Agency assess and report on
 strategies to increase the use of recycling and waste minimization techniques.
 Under  the requirements  of this Act, the  Agency has  established  the "Best
 Demonstrated   Available  Technology"  for   electroplating   wastes   and  is
 investigating the contribution which can be made  through implementation waste
 minimization techniques.                             .

     Historically, we have advanced from a time when no treatment  was required
 to a time of zero discharge under the FWPCA,  with pollutants removed from the
 air and water being discarded onto the land to leach back into the ground and
 surface waters.  The enactment of the Hazardous & Solid Waste Amendments with
 its "Landfill  Ban"  and "waste minimization"  provisions has  brought us  full
 cycle.  There  is now no place left to go for the disposal  of the pollutants
 and other hazardous wastes.  There is now little  alternative to taking action
 to reduce the production of toxic  pollutants  through product substitution and
 process changes, to  utilize  recycle and  reuse opportunities,  and,  as a  last
 alternative, to detoxify as  best  we can  any  residuals which remain  prior to
 land disposal.

                           OVERALL CONSIDERATIONS


     The major source of  pollution generation in a  well-operated and maintained
 electroplating shop is the plating chemicals which are lost from the plating
 bath when the  plated parts are removed.   This is known as  "drag-out".    In
 plating operations which do not take measures to control drag-out,  plating
 chemical losses range from 50 to 90% -- that  is from 50  to 90% of the plating
 chemical is removed from the bath  in the  form of  drag-out and rinsed down the
 drain.  The  drag-out  is  removed from the plated workpiece by rinsing in water.
 Free, or once through rinsing had  been the usual  practice until recently.  It
 has  been recognized  by those "in the trade"  that this type of rinsing  produced
 by  far  the major  waste stream  requiring  extensive  treatment  to meet  the
 requirements of the FWPCA.

     A number of approaches are available to control the polluted  wastewaters
 generated in  electroplating.   The  most obvious  is  the treatment  of  the
 wastewater to remove pollutants to  levels acceptable  for discharge  into  the
 environment.   However,  wastewater treatment  can prove very expensive when
 applied to conventional  electroplating  processes owing to large wastewater
 volumes resulting from free rinse operations.

    The size and cost of the equipment required to remove pollutants to levels
 acceptable for discharge depend, in large  part, upon the wastewater volumetric
 flow rate in  addition to the inlet  and outlet concentrations  of the polluting
 substance.  Where chemical processes are used  for pollutant reduction,  greater
quantities  of  wastewater  treatment  chemicals   are   required  per  unit  of
pollutant removed as the wastewater concentration decreases.   This results in
                                    211

-------
proportionally greater sludge  formation.   Inefficient design and operation,
therefore,  significantly  affects  the  interrelated   factors  of  materials
consumption, pollution control, and costs.

    Because of increasing prices for electroplating chemicals, process water,
pollution control  systems and reagents, sewage  charges,  and residual waste
disposal, it is highly desirable to incorporate certain in-plant changes into
existing plating operations.   These in-plant changes  result in the dramatic
reduction of the production of wastewaters requiring treatment and disposal.
Reduction  of wastewater  flow  with recovery  and recycle  of  plating bath
chemicals has become  the  preferred technique,  not only for enhanced control
of environmental  pollution, but also  for the significant economic advantages
inherent  in  the application.    Wastewater  flow reduction  can  be achieved
through  two approaches  (Figure 2).   The  first approach  involves certain
in-plant process changes such  as application of techniques to reduce drag-out,
minimize rinse water  usage, and recycle rinse  water directly to the plating
bath.  The second involves application of controls at  the electroplating bath
to concentrate  plating chemicals from rinse waters for recycle to the bath
such as evaporation, ion exchange, or electrolyte recovery.  Those techniques,
when systematically implemented,  result in significant cost savings in bath
chemicals, process  water, wastewater  treatment,  sewage, and waste  disposal.

             Workpiece
• -»—
Chemical
recycle
Bath
purification
n' r
L__J
Plating
bath
•^—
•*~

i r
i • i.
L_>f

Recovery
unit



B— 4j
L.4


1 r
i 	 r
S 1
1 	 -J
Rinse tanks
«—
Rinse recycle
Makeup
^~ water
             (•) Closed Loop
             Workpiece
••*
Chemical
recycle
Bath
™ purification
1 1 	
1 1
1 1
l_ J p.
Platin
bath
i r
'1
Recovery
unit
•
1 r
H 	 h
I.J
Rinse tanks
•i

"~1 r
i
L_J
L
             (b) Open Loop
                                                        To waste
                                                        treatment
                                                        Makeup
                                                        water
               Figure 2.  Waste Minimization Approaches
                                     212

-------
     The resulting cost savings may  be so significant that a  plater  cannot
ignore their economic potential even  if pollution control was not recognized.
As shown in Table 1(1), the replacement cost for plating chemicals is a major
operating cost  item.   How many platers can  afford  to discard 50 to  90% of
their chemical raw materials? In  addition  to providing chemical  savings and
reducing water  use costs,  in-plant changes provide a  basis for  a  pollution
control system design.  Waste treatment equipment needs,  whether wastewater
concentrating techniques   such  as ion exchange or  conventional  end-of-pipe
treatment systems, often will be  reduced  significantly by in-plant  changes.
 And the potential impact on pollution control is also significant.
            TABLE 1.  ESTIMATED COSTS OF PLATING CHEMICAL LOSSES
                                                       Cost
Plating Chemical Rep!
Nickel
As NiS04
As NiCl2
Zinc cyanide as Zn(CN)2
Using C12 for cyanide oxidation
Using NaOCl for cyanide oxidation
Chromic acid as H2Cr04
Using S02 for chromium reduction
Using NaHS03 for chromium reduction
Copper cyanide as Cu(CN)2
Using C12 for cyanide oxidation
Using NaOCl for cyanide oxidation
Copper sulfate as CuS04
acement

1.19
1.05

2.00
2.00

1.18
1.18

2.62
2.62
0.88
Treatment

0.20
0.30

1.03
2.02

0.51
0.84

1.02
2.30
0.20
Disposal

0.35
0.41

0.57
0.57

0.52
0,52

0.50
0.50
0.36
Total

1.74
1.76

3.60
4.59

2.21
2.54

4.14
5.42
1.44
 Wastewater concentration of 100 mg/L assumed.
 Disposal of dewatered sludge (20 percent solids).

                             MINIMIZING  DRAG-OUT
    Drag-out losses can be minimized either by reducing the amount of plating
solution which leaves the plating bath, or by recycling plating chemicals in
the rinse  water  to the plating batlr  .  Many devices  and  procedures can be
used  successfully   to reduce   drag-out.   These  techniques    usually  are
employed to alter viscosity*  chemical concentration,  surface tension,  speed
                                    213

-------
of  withdrawal,  of the  workpiece  from  the  plating tank,  position  of  the
workpiece on the plating rack, and temperature.  Also drag-out tanks are used
for  capturing lost  plating  solution and  returning it  to  the bath.   Most
drag-out  reduction  methods   are  inexpensive to  implement  and  are  repaid
promptly through  savings in plating chemicals.   An additional  savings many
times the cost of the changes  will be realized once a pollution control system
is  installed.   The  reduced  drag-out will  decrease the need  for treatment
chemicals   and  subsequently,   the  volume  of   sludge  produced.     The
cost-effectiveness of each of these methods  is discussed below.

    There are also several simple methods of drag-out recovery that should be
considered.  Four simple drag-out recovery methods are:  drainboard, drip tank,
spray  rinse,  and  air knife.   The drain  board  is  the simplest method  of
drag-out recovery.    It  can capture drips of plating  solution  as racks and
barrels are transferred  between  tanks  (Figure 3).  Not only do drain boards
save  chemicals  and  reduce   rinse  water  requirements, they  also  prevent
unnecessary floor wetting.

               Workpiece
                                                     Drip bar
                                                          Concentrated
                                                          solution
                Figure 3.  Drain Board
 A drip  tank is an ordinary  rinse  tank that,  instead of  being  filled with
water, simply collects the drips from racked parts and barrels after plating
and before rinsing.  When a sizable volume of solution has been collected in
the drip  tank,  it can be returned  to  the plating bath.   Using  a  drip tank
tends to restrict the potential  use of a rinse  tank.   As will  be discussed,
an additional rinse tank used as  a  drag-out  tank or  in a series  arrangement
may be more  beneficial.  The  determining factors are  the volume  of drag-out
and the evaporation in the plating  bath.

    Spray  rinses are  ideal   for  reducing drag-out  from  plating  tanks  on
automated lines.  As the workpiece  is withdrawn from   the plating  solution,
                                     214

-------
the workplace, is rinsed by a spray returning as much  as 75% of the  drag-out
chemicals  back to the plating tank.  Spray rinsing is best suited for flat
parts, but will reduce drag-out effectively on  any plated  part.   The  volume
of spray  rinse may not  exceed  the volume of  surface evaporation  from  the
plating tank.

    An air knife reduces drag-out in much the same way as a spray rinse,  and
is particularly useful when the surface evaporation rate in the plating  bath
is low.  Air  knives reduce the volume  of drag-out  adhering to  the workpiece
by subjecting the workpiece to a high-velocity  stream  of air.   The  drag-out
is returned to the plating bath without changing its concentration.

                           RINSE WATER MANAGEMENT


    The rinsing operation can account  for as much  as  90  percent of  a  shop's
water usage.  Therefore, careful  management and conservation of water  in the
rinsing operation offers the greatest opportunity for  significant reductions
in water consumption and wastewater operation.  To achieve the desired  degree
of cleaning, the rinsing operation must include:   (1) turbulent motion between
the workpiece and rinse  water,  (2)  adequate  time of contact between  the
workpiece and rinse water, and (3)  contact with  rinse  water of sufficiently
low concentration to effect the dilution of the  plating  bath washed off the
surface  of  the workpiece.    These three  principles  apply  to all  rinsing
operations,  including those using   flow-through  or still rinse  tanks.   They
can be implemented  in  a number of ways, depending upon  the  feasibility  for
application to a specific plating line.

    The amount of make-up water required to dilute the rinse solution depends
on the quantity of chemical drag-in from the  upstream  rinse or  plating tank,
the   concentration  of chemicals  in  the  rinse water,  and the  contacting
efficiency   between  the  workpiece and  the  water.    It  is important  that
sufficient turbulence and adequate  contact time  be employed for all rinsing
operations  to minimize  the  introduction  of  excess  quantities of make-up
rinsewater.    The  maximum allowable concentration  becomes  a very important
parameter when  the other  two parameters  are  satisfied.   In   fact maximum
allowable concentration is the governing factor  with respect to water  use in
a well-operated shop.

    Reductions in rinse water volumes  can be achieved through use of  a  number
of conservation techniques  including the control  of  fresh  make-up water
introduction  based upon  dissolved  solids  content of the rinse  water, rinse
water recycling, multiple  rinse  stages,  use  of other  rinse techniques,  and
minimization of drag-in.

    Use of a simple method of water conservation is becoming more widespread.
It involves the reuse of rinse water  at two  or more  rinse tanks where  the
contaminants  in the rinse water after  a processing step  do not  detract  from
the rinse water quality at another station.  This method is applied most often
to the rinses  following acid dips and alkaline cleaners.  For example, instead
of using 19 liters per minute of rinse water in  each rinse tank for a total
of  38 liters  per minute, the rinse  water used following an acid dip  can be
                                    215

-------
 reused  as rinse water directly after an alkaline cleaner.  This practice will
 reduce  the water  use  for  these  two tanks  by  50 percent.   In most  cases,
 contamination does not appear to be a problem.  In fact, the rinsing following
 the   alkaline cleaner   appears   to improve.     The   mass  transfer  action
 attributable  to  diffusion  is  accelerated as the concentration of  alkaline
 material  at  the  interface  between  the alkaline  drag-out  film  and  the
 surrounding water is reduced by the chemical   reaction there.  Also,  alkaline
 solutions usually are more difficult to rinse off than acid solutions because
 of the  higher viscosities,  so  neutralization aids in  this respect(3).

    Other  reuse  arrangements  can  be  employed  where  the less  contaminated
 overflow from critical or final rinsing operations is  reused for intermediate
 rinse steps,  such as acid dips and alkaline  cleaning  steps.  The rinse water
 following a nickel plating bath can be routed to the rinse tank following the
 acid  dip.  This rinse water, in turn, could be routed to the alkaline cleaner.

    Choosing  the  optimal configuration  requires  analysis of the  particular
 rinse water needs.  Interconnecting rinsing systems might make operations more
 complicated,  but  the cost advantage justifies the extra  attention  required.
 The  benefits  from  reusing  rinse  water  are  limited,  however,  because  that
 method  of  conservation  cannot be applied to all rinse operations.   Methods
 exist,  however,  that can  be applied  more widely,  and  that result in  more
 dramatic  water savings.   The three most common methods  are  parallel  and
 countercurrent rinsing and  the use of  still  rinse tanks.
Workplace ••*•
                     Overflow pipes
                                      Air agitation
          Figure 4.  Three-Stage Countercurrent Rinse Arrangement
                                     216

-------
    Using a countercurrent rinse tank arrangement, the plater can achieve even
greater water  savings  than  with the  parallel  system.   In  this  arrangement
(Figure 4), make-up  water flows  into the rinse  tank  farthest  away from the
plating tank and moves  (countercurrently  to  the  work-flow)  toward the rinse
tank  closest  to the plating tank  either by  gravity or  by  pumping.   The
workpiece is dipped  in  the least pure water  first and in the  cleanest water
last.    The  quantity  of  chemicals  entering  the  final  rinse  will  be
significantly  smaller  than  that entering  a  single-tank rinse system.   The
amount of rinse water required for dilution will  be reduced the same degree.

     The relation  between volume of rinse water required  as  a  function of
initial concentration in the plating bath, required concentration  in the final
rinse tank, and number  of rinse tanks is  shown in Figure 5.  For example, a
typical Watts-type nickel plating  solution contains  270,000  milligrams per
liter of total  dissolved solids (Cp), and the final rinse  must contain no more
than 37 milligrams per liter  of dissolved solids  (Cn).  The  ratio of Cp/Cn is
7,300; hence,  7,300 liters of rinse  water would be required  for each liter of
process solution drag-in,  assuming a single-tank  rinse system.  By installing
a two-stage rinse system, the same  degree of dilution is achieved with only
86 liters of water per liter of process solution drag-in,  a reduction in rinse
water consumption  of almost  99%.   The  mass  flow  of  pollutants  leaving the
rinse system remains constant.
           100.000 E~
            10.000 =-
            1.000 —
              100
                    • iiitml   i i  11 mil   11 11 iiill	i  i i inn!  i  i 11 mil
                         10
100      1.000     10.000     100.000

 RINSE RATIO
        Figure  5.   Rinse Water  Dilution  vs.  Rinse  Ratio  for  Multitank
                   Countercurrent  Rinsing
                                     217

-------
     A three-stage countercurrent rinse arrangement would further reduce water
 consumption  to 76  liters/liter  of drag-in.  The  resulting cost savings  by
 going from a one-stage to a three-stage rinse system would  include  reducing
 water use  and sewer fees and reducing the size of the required waste treatment
 systems.   The investment  cost  to add two  additional  rinse tanks is  highly
 site-specific.  For manual plating operations, the major factor affecting cost
 is  the  availability  of  space  in the process area.   For automatic plating
 machines,  the cost  of modifying the unit to add additional  stations may be  as
 high as  $20,000  per  station.    Rubber-lined , steel  open-top  tanks  with
 appropriate  weir plates  and nozzles cost  anywhere  from  $1,000 to  $3,000
 depending  on the cross-sectional  area required  for the workpiece.

     A parallel rinse tank arrangement using three  rinse  tanks  is  illustrated
 in  Figure  6.  With the parallel  feed system,  each tank is  individually fed
 with fresh water.   The  rate of water flow  to each  tank should  be  the same  to
 obtain  the optimal  water savings.  In this case, each rinse tank receives  a
 fresh water  feed and  discharges  the overflow to waste treatment.  The rinse
 ratio required for a parallel rinse arrangement is defined by r = n (Cp/Cn)1/n.
 If the  rate  given in Figure 5 for a countercurrent  rinse  system with  the same
 number  of rinse tanks is multiplied by the number of rinse tanks, the  parallel
 rinse water  rate can be estimated.  Rinse water rates are  significantly  higher
 for  parallel rinsing than for countercurrent (series).
                 Drag-out
        Plating
        bath
                                                                         Water
                    To
                    waste
                    treatment
             Figure 6.  Three-Stage  Parallel  Rinse Arrangement
                          RINSING RECOVERY SYSTEMS
    The drag-out losses from the plating process can be significantly reduced
by  relatively low cost  process modifications which  lead to  an integrated
rinsing-recovery  system.   This system will offer  substantial  savings  in
plating chemicals where  it  can  be  employed.   These alternatives are usually
considered only  after  steps have been  taken to minimize  drag-out and rinse
water usage.
                                     218

-------
    The techniques  for  drag-out and rinse water  management can be  used  to
formulate a strategy for simple recovery systems  using multiple rinse tanks
and a minimum of additional equipment.  The  strategy  takes  advantage of the
need to resupply the plating bath losses, particularly the return of the more
concentrated solutions of dragged-out plating chemicals  to  the  plating bath
to make up  for  water lost  by  surface evaporation.  The  amount  of  chemicals
actually recovered depends on the amount of  chemicals lost  from the plating
tank, the number of rinse tanks used, the concentration of chemicals permitted
in the final rinse tank,  and the rate at which rinse water can be recycled to
the plating tank.

    Of these, the rate  at  which rinse water can  be recycled  to the plating
tank is usually  the  most critical;   it is  primarily dependent on the amount
of surface evaporation from the plating tank.  If  the evaporation rate can be
matched to  the  required  rinse water rate,  the entire volume  of rinse water
could  be  returned to the  plating  bath.   This  set-up is  referred  to  as a
closed-loop recovery system.   In  a closed-loop rinse water system the only
chemical  loss  is from the drag-out  after  the  last rinse tank,  which has a
dilute  concentration of plating  chemicals.   A  closed-loop  system  may  be
impractical when:

    o     a  very low final  rinse concentration is  required  and only
          achievable  through a larger number  of rinse  stages;

    o    excessive drag-out is unavoidable because of product configuration;

    o     plating tank surface evaporation  is minimal.

    The tank arrangement shown in  Figure  7, which  consists of  a drag-out tank
followed  by a flow-through rinse tank, is  the simplest recovery  system.  The
drag-out  tank is a  rinse tank that  initially is filled with pure water.

                  Evaporation
              Workpiece
                    Plating
                    bath
                  Figure  7.   Recovery with  a  Drag-Out Tank
                                     219

-------
     The drag-out rinse collects  a significant portion of the process solution
 carried on the parts,  rack or barrel.  Air agitation must be used to aid the
 rinsing process  because there  is  no water  flow within  the tank  to cause
 turbulence.   The presence of  a  wetting agent is helpful(3).   As the plating
 line is operated, the salt concentration" increases as more work passes through
 the  rinse tank.  Periodically,  the strong solution in the  drag-out tank is
 returned to the plating tank.  The volume returned  is  limited to the volume
 made available in the  process tank by evaporation.

     As  a rule, the use of a drag-out tank will  reduce  chemical  losses by 50
 percent or more.   The  efficiency  of the drag-out  -tank arrangement  can be
 increased significantly by adding a second drag-out  tank.  Use of a two-stage
 drag-out system usually reduces  drag-out losses by 70 percent or more.

     The use of drag-out  tanks usually results in less water savings than does
 parallel  or series rinsing.   The operational  procedure used with  drag-out
 tanks is responsible  for this effect.  The rinse water  in  the drag-out tank
 increases in  plating  salts  concentration until  a portion  is  returned to the
 plating bath to compensate for evaporative losses.  The concentration of salts
 in the  drag-out  tank  can  reach as  high  as  75  percent  of the  plating bath
 concentration.  Consequently, a significant water flow in the rinse following
 the  drag-out   tank  would  be   necessary to  meet  the   maximum  allowable
 concentration.

     The  low final concentration problem can  be overcome  in many  cases  by
 operating  the  final rinse  as  a free (running) rinse, and  using  the  upstream
 tanks as a countercurrent  rinse-and-recycle system.   Using this  approach,
 significant drag-out recovery can  be realized while providing a final  rinse
with  a  low  level    of contaminants.     Figure   8   shows  an   automatic
 rinse-and-recycle system with a  running rinse.   Level-control  devices in the
 plating and  rinse tanks control  the flow of rinse  water through the system.
                                             Makeup waten
                     ILC1	
               Surface
              evaporation j
           Workpiece
                  Plating bath
Two-stage recovery rinse
           LC •> Level Control

1
1_
1 1
1
1
	 1
Running rinse
                           To waste
                           treatment
              Figure  8.   Automated Rinse-and-Recycle System
                                     220

-------
                100
                80
              a 60
                40
                20
                                                          0.2
                                                          0.4
                                                          0.6
                                                          0.8
                                                             O

                                                          1.0
                 02       4       6       8
                              RECYCLE RINSE RATIO (r)

                  Notes:
                  n = number of counter-flow rinse tanks in recovery use.
                   r = recycle rinse (gal/h) + drag-out (gal/h).
                  Recycle rinse set equal to surface evaporation from batch.

                Figure 9.  Drag-Out  Recovery  Rate  for
                           Rinse-and-Recycle  Systems
                                                         10
the percent recovery of drag-out for such  a system as function  of the recycle
ratio,  defined as  the volume  of recycled  rinse divided by  the volume  of
drag-out in a  given time is shown in  Figure  9,  (the recycle rate  is assumed
to be equal to the evaporation rate).   As shown,  such systems can recover from
40 to 100  percent of the drag-out.

    A   relatively   new   application   of  multiple  rinse  tanks   is   the
drag-in/drag-out  configuration  (Figure 10).  With the drag-in/drag-out system,
the rinse  tank preceding the plating  bath (drag-in  tank)  is  connected to the
recovery rinse (drag-out tank)  following the bath;  the recovered  drag-out
solution is circulated by a  pump.  The concentrations of salts  in the drag-in
and drag-out  tanks remain about equal.   When a  rack or barrel  is  processed,
it drags in plating solution to the plating tank, thereby increasing recovery.

    The drag-in/drag-out system finds  application with plating baths that have
a  low  evaporation  rate.   The  recycle  ratio,  which  determines  recovery
efficiency, is calculated as the volume of recycled rinse plus the volume of
drag-out divided  by the  volume  of  drag-out.   The recycle ratio, therefore, is
greater with  a drag-in/drag-out system than  a common recovery tank.  If the
evaporation rate  is low,  the difference between  the  recycle  ratios for common
recovery and drag-in/drag-out systems  is  significant.   When evaporation rate
is high, the  difference is less.
                                      221

-------
                   Recirculate
                                                                       Rinse
Workplace
       Figure 10.  Drag-In/Drag-Out Recovery Arrangement
     Various other rinsing configurations could be developed by adding tanks.
 The choice  of a  best  arrangement  is  difficult  because  of the  trade-offs
 involved between further  reducing  chemical  losses and further reducing  the
 rinse flow rate.  Obviously, the value of the lost chemicals is a  significant
 cost.    Chemical  losses  also result  in additional  rinse  water and  waste
 treatment chemical requirements  and more sludge.(4)   Although complex,  the
 evaluation and selection  of a multiple  rinse tank system can be accomplished
 by analyzing each rinsing configuration and comparing cost  factors,  such as
 water,  sewer,  and waste treatment.  The results of the evaluation  will  enable
 the plater  to  determine  whether  a  multiple  rinse  tank  arrangement   is
 beneficial  and to identify the most appropriate configuration.

                             WASTEWATER TREATMENT


     When  the  techniques  for flow reduction have been implemented as  far as
 possible,  wastewater  treatment techniques must then be considered to  reduce
 the remaining pollution burdens to produce a wastewater suitable for discharge
 and a concentrated waste  for recycle or  for waste treatment.

     There are a number of processes used to concentrate plating chemicals in
 rinse  waters  for  recycle  to  the  plating  bath.   These  processes  include
 evaporation, reverse  osmosis,  ion exchange, electrodialysis, and electrolyte
 recovery.  Several new technologies  are being investigated.   These  are coupled
 transport,  reversible  gel absorption,   and  freeze  crystallization.   These
                                     222

-------
processes differ in selectivity for the concentration  of  the  chemicals  that
are dissolved  in  the  rinsewater.    Raw  wastewaters may  be concentrated  or
treated directly  by  one of several precipitation  techniques  to reduce  the
dissolved solids content of metals to a level which will  allow  discharge of
the cleaned wastewater.  There are  several concentration techniques available,
etc.   The concentrated  solutions may be  recycled to recover the  plating
chemicals of value or treated to remove the metals via chemical  precipitation,
electrolytic recovery of the metal, etc.

     Plating shop wastewaters  can  be classified into seven groups, each group
relating to a specific type of generic pollutant removal problem.  These are
treatment of acidic metal-bearing wastewaters,  recovery of precious  metals,
destruction of cyanide, control of hexavalent chromium,  treatment of complexed
metals, removal  of oil  and grease,  and  control of toxic organics.    It  is
considered best practice to segregate wastewaters containing oil and  grease,
toxic organics, precious metals,  cyanides,  hexavalent  chrome,  and  complexed
metals, as these require pretreatment  before precipitation  (Figure 1).  While
a broad range of alternative technologies exists for treating the wastewaters
from these processes,  treatments most  widely practiced  today are based on the
following technologies.

     Common acidic, metal-bearing  wastewaters are usually chemically treated,
usually with lime  or alkali-earth  hydroxide to  achieve adjustment  of pH and
precipitation of metals.  The precipitates are separated from the liquid phase
by  flocculation  and  clarification.   Further separation may be  achieved by
filtration,  where  required.    A  large  range  of  alternative  treatment
technologies, which provide wastes which are more environmentally stable than
hydroxide sludges or which provide for removal  and capability for recycling,
are also  available.   These techniques,  which  are addressed briefly  in  the
following section, are more costly for wastewater treatment; however, some may
prove useful in controlling overall  processing costs when the impact of waste
management costs incurred  to comply with land disposal  regulations are fully
incorporated into the overall system.

     The recovery of precious metals, the  reduction of hexavalent chromium,
the  removal  of  oily  wastes,  and  the  destruction   of  cyanide  must  be
accomplished prior to metals removal.  Oils and greases  are removed by gravity
separation  and  skimming  of  free oils  followed by  chemical   and  emulsion
breaking and subsequent skimming  for  the removal of emulsified  oils.

     Cyanide bearing wastes  are treated with  an oxidizing agent  (ozone or
chlorine) to destroy the cyanide in the wastewater.  Cyanide, as well as being
a highly toxic pollutant, will complex metals such as copper, cadmium and zinc
and  prevent  efficient  removal  of these  metals.   Wastewaters  containing
hexavalent chrome  are treated with  a  reducing  agent to reduce the chrome to
the trivalent form which can  then  be precipitated from  solution by hydroxide.

     Following separate stream treatment, the effluents are combined  and the
metals are removed by precipitation and subsequent clarification. Most metals
precipitate   as   hydroxides   although  some,  such  as  lead   and  silver,
preferentially  form other compounds  (e.g.,  carbonates  or chlorides).  The
                                     223

-------
 optimum pH for precipitation is generally in the range of 8.9-9.3, although
 it will vary somewhat depending on the specific waste composition.  The use
 of coagulants or flocculants to  enhance  the effectiveness of clarification
 is also required.

      Chelating agents are often  used in electroplating operations.   When metal
 chelating agents are  present,  the wastewater  streams  containing complexed
 metals  must be segregated and separately treated.  Chelating agents react with
 dissolved metal ions to form "chelate complex", which is  usually quite soluble
 in neutral or slightly alkaline solutions.  A waste stream containing metal
 complexes may then  be  separately  treated  by  adjusting the pH to 11.6 to 12.5
 with  lime to  break the complex.   Other  alternatives  may be used  such  as
 sulfide precipitation,  ion  exchange, or starch  xanthate precipitation.

                                 CONCLUSIONS


     Applications of waste minimization principles is the most important step
 in achieving  the  desired environmental  protection goals.   Those principles,
 briefly outlined in  this  paper,  are  now broadly  used because  they  offer
 significant  cost  savings  potential  through  the  conservation   of  process
 materials, reduction in waste treatment requirements, and, in many situations,
 the elimination of large quantities of  wastes requiring land disposal.  Waste
 minimization  techniques in  practice offer a six-month  to  four-year payback,
 usually a 15 to  20 percent return on investment, and a  50 to  90  percent
 reduction  in  chemical  wastage and in wastewater  operation.   The  techniques
 should  be evaluated in the sequence illustrated  in Table 2 and compared cost-
 wise to the present  plant performance based on the data derived from the plant
 audit.

     After the various  alternatives of waste minimization  are evaluated  and
 the generation of a waste requiring land disposal  is found to be  unavoidable,
 it is then necessary to  select an optimum wastewater treatment technology from
 among the many alternatives.  The  relative  attributes of several  alternatives
 are given in Table 3.   The selection process  is not necessarily straight
 forward as the technical and  economic considerations are affected by a number
 of site specific factors.  For example, in comparing hydroxide precipitation
 processes, the  costs  of using  caustic  soda and magnesium  hydroxide  appear
 similar.   However,  use  of  magnesium  hydroxide may  be  more desirable when
 disposal costs rise because  of  haul distance as  less sludge  is  produced per
 pound  of  metal  precipitated.    Also  in  comparing hydroxide  and  sulfide
 precipitation, the reagent costs are similar for caustic and soluble  sulfide
while most disposal cost for soluble sulfide is twice that  of caustic  soda.
However,  sulfide  precipitates  containing   high  levels of  copper,   lead, and
zinc,  especially  in association  with  precious  metals,  may  be  of value to
primary or secondary metal  smelters and refiners.  If so the disposal  cost is
eliminated, and there may even be a net positive cash flow generated.
                                    224

-------
     In  addition,  the wastewater  treatment  technique  employed  should be
selected to produce a waste compatible with  the  treatment  standards  obtained
from the application of the "Best Demonstrated Available  Technology  required
under the provisions of the land disposal restrictions of the Solid and
Hazardous  Waste Act  of  1984.    Standards   to  be  applied  to  F006 wastes
containing regulated  metals  have been promulgated  based  upon a  BOAT  using
solidification in a matrix of cement kiln dust.   However,  cyanide-containing
wastes are currently regulated under the  "soft hammer" provisions of ^he rule
(Federal Register Vol .53,  No.l59/Wednesday, August 17, 1988/pgs 31138-31222).
These   provisions   require   that  disposal   facilities  meet   the   minimum
requirements of RCRA 3004(o), i.e.rdouble liner, leachate collection system,
and ground water monitoring  -  or equivalent  performance as  provided in 3004
           TABLE 2.   SELECTING TECHNICAL OPTIONS  FOR ENVIRONMENTAL
_     CONTROL  OF  ELECTRQPI ATING  OPERATIONS _

1.   Process Audit
           a.    Chemicals Used
           b.    Processes used
           c.    Pathways of pollutant  generation - multimedia
           d.    Characterization of pollutant generation - multi pollutant
2.   Waste Minimization Techniques
           a.    Drag-out minimization
           b.    Rinse water minimization
           c.    Rinsing recovery systems
3.   Waste Water Treatment
           a.    Concentration techniques
           b.    Other recovery techniques
           c.    Precipitation techniques
4.   Hazardous  Waste Treatment
           a.    Land Disposal
           b.    "Soft Hammer" Provisions
5.   Environmental  Comparison Test
           Does the results  of controls selected comply  with  the       .
           requirements  of  the CAA,  FWPCA,  and RCRA?
 6.   Cost Comparison Test
           Does the pathway  selected yield  an affordable or optimal
           cost when compared with  the possible  pathways?


     This consideration will  have  an  effect  on  the cost figures for  sludge
 disposal given in  Table 2.   These cost values are for conventional  level
 disposal as practiced prior to the  "land ban  provisions".  As the provisions
 for disposal under the land ban,  including  interim soft  hammer requirements,
 are added, the disposal costs are  likely to  escalate.   This  escalation will
 make  non- and  low-sludge producing  waste  treatment alternatives even  more
 attractive.   Higher cost processes which  are capable  of producing a waste
 which can be delisted also may become more cost-competitive.
                                      225

-------
     TABLE 3.
Cost ($/lb of metal)
Precipitating
Aaent
Hydroxide
Sodium
Calcium
Magnesium
Sulfide
Soluble
Insoluble
Calcium
Carbonate
Sodium
Sodium bi-
Calcium
Sludge Chemical
Hb/lb metal) Usaae

150
300
150

300
900
300

300
200
200

0.32
0.06
1.20

0.36
0.34
0.27

0.30
0.40
0.30

- 0.37
- 0.07
- 1.30

- 0.48
- 0.45
- 0.34

- 0.35
- 0.48
- 0.35
Sludge*
Disoosal

2.00
4.00
1.10

4.00
12.00
4.00

2.60
2.60
2.60
Total

2.40
4.10
2.40

4.50
12.50
4.35

3.00
3.10
3.00
Other
 Sodium biohydrate      90
 Dithiocarbonate       490
 Starch xanthate     1,000

Adsorption
 Ferric chloride
 Alum

Electrodeposition
 Electrowinning           0

 Special  electrodes	0
                                         16.00
                                         25.00
                                         77.00
                                         0.46
                                         5.80
                                            0

                                            0
1.20
6.10
17.20
31.40
   0

   0
 2.00

 6.00
based upon a dewatering cost of one cent per pound of raw sludge (1 to 2%
solids  by  weight)  and a disposal cost of ten  cents  per pound of dewatered
siudge.
without credit for resale of recovered metals into the secondary metals
scrap market.
                                REFERENCES


    Environmental Pollution Control Alternatives:  Reducing Water Pollution
    Control Costs in the Electroplating Industry; EPA 625/5-85-016;
    September, 1985.

    Control and Treatment Technology for the Metal  Finishing Industry:
    In-Plant Changes;  EPA 625/0-82-008; January, 1982.
                                   226

-------
Kushner, Joseph B.;  Water and Waste Control for the Plating Shop.
Cincinnati, OH, Gardner Publications, 1976.

Roy, Clarence.; "Methods and  Technologies  for Reducing the Generation of
Electroplating Sludges."  In U.S. Environmental Protection Agency and
American Electroplaters' Society, Inc.  (cosponsors), Second Conference
on Advanced Pollution Control for the Metal Finishing Industry.
EPA 600/8-79-014.  NTIS No. PB 297-453. Feb. 1979.

This paper was adapted by John Burckle from references 1 and 2 above.
                                227

-------
                    PCB DEGRADATION: STATUS AND DIRECTIONS

                   by:    P. R.  Sferra
                         Risk Reduction Engineering Laboratory
                         U. S.  Environmental Protection Agency
                         Cincinnati, Ohio   45268
                                   ABSTRACT

     There is varied experimental activity to develop effective means of
destroying PCBs that are polluting the environment as hazardous wastes in
soils, sediments, water, and in storage.  The recognized potential of
achieving environmentally harmless clean-up of PCBs by the use of
biological processes has resulted in increasing research activity.  This
paper reviews the status of PCBs as pollutants and the current knowledge Of
microbial degradation of PCB compounds and the attempts to develop
technology for their control.  Current information on degradation by mixed
cultures, pure  cultures,  aerobic and anaerobic processes  and the degradative
relationships  between  various  microorganisms  and  substrate  structure  are
discussed. Emphasis  is on the  importance of knowledge of breakdown pathways
resulting from metabolic action of microorganisms  on PCBs.  This information is
vital to the successful development  through genetic  engineering techniques of
strains with improved capability for  degradation.  Pertinent metabolic activity
and related genetic  processes  are presented as a basis for  the rationale for
ongoing research utilizing applied genetics  to  develop  microbial strains that
will be super degraders of the PCBs polluting the environment.
                                      228

-------
INTRODUCTION

     Polychlorinated biphenyls  (PCBs) have been of worldwide concern as
polluting chemicals for over 20 years since Jensen (39) in 1966 first
reported their presence in wildlife  samples.  Because of  their abundance,
persistence, and apparent hazard to  wildlife and human health these
chlorinated hydrocarbons have been subjected to considerable study and as a
consequence by this time more than 1100 research articles on various
aspects of PCBs have been published.

      Commercial production of PCBs, which began in 1929, was accomplished
by the synthetic process of direct chlorination of biphenyl with anhydrous
chlorine and a catalyst of iron filings or ferric chloride.  Control of the
reaction conditions resulted in varied degrees of chlorination.  With
chlorines replacing hydrogens, a large number of possible congeners can be
produced.  Thus, a polychlorinated biphenyl is one of 209 compounds having
the formula:  C12H10_nCln  including  3 possible monochlorobiphenyls.  The
209 congeners  can be subdivided into  10  groups of homologs  according to the
degree of chlorination and the number of isomers per homolog can vary from
1 to 46.  PCBs were manufactured in  the United States and Great Britain
under the tradename Aroclor, in Japan under the names Kanechlor and
Santotherm, in Germany Clophen, in Italy Fenclor, and in France Phenoclor
and Pyralene.  The commercial products are complicated mixtures of chloro-
biphenyls.   The  different  Aroclors  are  given  a  4-digit  designation  that
represents the type of molecule and  the weight percent of chlorine.
Aroclor 1242 consists of 12 = chlorinated biphenyl and 42 = 42% chlorine
per weight.

      Two characteristics of the commercial products complicate attempts to
understand their behavior in the environment.  One is that any one of the
commercial products may consist of a great number of different PCB
congeners,  therefore the task of destroying PCBs is not as simple as
destroying  one  particular  type  of  molecule.    In  addition,  polychloro-
dibenzofurans (PCDFs) have been found in microgram per gram levels in
commercial PCB mixtures; these  impurities therefore may be responsible
for at least some of the toxicological effects  attributed to the
commercial PCB mixtures.

      The utility of PCBs as industrial materials is based mainly upon their
chemical and physical  stability and their electrical insulating  properties.
These compounds have been used as dielectric fluids in capacitors and
transformers, heat transfer fluids,  hydraulic fluids,  lubricating and
cutting oils, dedusting agents, additives to pesticides, printing inks,
paints,  pesticides,   copying paper,  carbonless copy paper, adhesives,
sealants, and plastics.  Because of their properties the PCBs do not
readily degrade in the environment (22).

      From 1929 through 1976 the world production of PCBs was about 1.3
billion pounds.  Monsanto,  in the U.  S., produced about 1.25  billion
pounds before they stopped production in 1977.   Production elsewhere
continued  through  at  least 1983  with total  world  production  through  1980
estimated to be about 2.4 billion pounds (22).   It is also estimated that
                                      229

-------
 in the U. S. about 404 million pounds of PCBs are still in use and
 accessible,  about 262  million  pounds  are  still  in  use  and generally  not
 accessible,  and  about  585  million  pounds  are  in storage,  landfills,  the
 environment, or have been destroyed.  The PCBs still in use and accessible
 are in transformers,
 capacitors, and other electrical equipment.   Within this  last category  are
 PCBs in sediments, soil, vegetation and animals, the atmosphere,  and fresh
 water totaling about 24 million pounds,  another 13 million pounds are in
 the oceans, about 385 million pounds are in landfills and other storage,
 and an estimated 162 million pounds has been degraded (46).  Several major
 contaminations involving PCBs resulting from  localized  discharges  have become
 notorious.   These occurred in the Hudson River, Waukegan, Illinois, New
 Bedford,  Massachusetts,  and Bloomington, Indiana.   Also, Yusho,  an incident of
 mass food poisoning western  Japan in 1968 affected more than  1600 people  who
 had consumed rice-bran oil contaminated  with PCBs.

       PCBs  have been found in marine mammals,  marine and freshwater fish,
 shellfish,  birds,  bird eggs,  adipose tissue in humans,  human milk,
 terrestrial animals,  wastewaters, drinking water,  household products,  and
 in foods.  PCBs are both hydrophobic and lipophilic and because they are
 soluble in  lipids  they tend to accumulate in organisms,  especially  those
 high in the food chain.   Studies  of the  toxic  effects of PCBs on  organisms
 have found  inhibition of cell division,  reduction  in RNA levels,  reduction
 in chlorophyll index,  reduced growth rate,  reduced cell  population  size,
 and inhibition of  carbon fixation (27).

       The amounts  of  PCBs  in  the  environment are considered to  be
 significant,  a situation perceived to be dangerous both  to  wildlife,and
 human health.   As  a consequence there has  been considerable effort  to
 determine how the  polluting PCBs  can be  destroyed  and those still in
 service prevented  from becoming new hazardous  wastes.  The  Risk Reduction
 Engineering Laboratory (RREL) of  the U.S. Environmental Protection Agency  has
 expended  considerable effort  in various  aspects of the PCS  pollution problem.
 Physical, chemical, and  biological means  of destroying PCBs  are being
 investigated  under RREL  sponsorship.   Basic knowledge of  the anabolic and
 catabolic capability of the living  cell by means  of its enzyme systems  leads
 to  the expectation that  cost-effective highly efficient biological control of
 hazardous pollutants  such as  PCBs with minimum adverse impact upon  the
 environment can be developed.  Mondello and Yates  (51) will  report  in these
 proceedings  on recent progress made in PCB biodegradation research under RREL
 sponsorship.

 DISPOSAL AND DESTRUCTION OF WASTE PCBs

       In  1982,  Fradkin and Barisas  (26) assessed the  available and  emerging
 technologies for the disposal of  PCBs and PCB-contaminated materials and  in
 1988,  Carpenter arid Wilson (15) published a technical and economic
 evaluation of processes  for the removal of PCBs from  sediments.  Exner
 (23) and Weitzman  (78) also prepared  information on the destruction or
 decontamination of PCBs.  Non-biological treatment technologies for PCB
wastes include wet air oxidation,  supercritical water oxidation, catalytic
                                      230

-------
dehydrochlorination, sodium-based nucleophilic substitution, alkali metal
potassium glycolate-based nucleophilic substitution, UV/photolysis,
microwave plasma, pyrolysis, removing and concentrating by heated air
stripping, extraction, adsorption, vitrification, stabilizing, and
bottom recovery by dredging (15).

      In 1973, publications by Gibson et al.(36), Wallnoffer et al. (76),
Catelani et al.  (17), Ohmori  et  al.  (55),  and Ahmed and Focht (3, 4) reported
on the degradation  of biphenyl and PCBs  by single strains of bacteria.  Since
that time it has become evident  that PCBs can be degraded by pure cultures and
by naturally occurring populations of microorganisms.  Data have also been
collected to determine the  conditions  that affect the degradation of PCBs and
to determine the metabolic pathways of PCB degradation by microorganisms.
Factors such as temperature, nutrients, oxygen, and pH all influence the
ability of the organism not only to function normally but function as a
degrader of xenobiotic compounds such  as  PCBs.   Other factors, related to the
PCBs, such as volatility, water  solubility, emulsification, adsorption, and
degree of chlorination  of the PCBs all influence  the degradative activity of
the microorganism (28).

      Furukawa et al. (35) determined relationships between chemical structure
and breakdown of biphenyl and PCBs.  General conclusions from  their results
were as  follows:  (1)  As  chlorine substitution increases, the  degradation rate
markedly decreases, (2) PCBs containing 2  chlorines in the ortho position of a
single ring (2,6-) and each ring (2,2'-)  show a high resistance to degradation,
(3) PCBs with all chlorines on a single ring are generally degraded faster than
those with the same number of  chlorines on  both rings, (4) PCBs with 2 chlorines
at the 2,3- position of one ring are more  susceptible to microbial degradation
at  least compared  with  other tetra-  and pentachlorobiphenyls,  and  (5) ring
cleavage occurs with a nonchlorinated or lesser chlorinated ring of the biphenyl
molecule.

      Tucker et  al. (74)  determined  that  the  rate  of  PCB biodegradation by
activated sludge decreased with  increasing chlorine content of commercial
mixtures.  Their studies  compared the  biodegradation  rates of  biphenyl and
the Aroclors 1221,  1016,  1242, and 1254 showing that  the average percent
degradation of the  commercial mixtures is  directly related to  their weight
percent  chlorine  content  (Aroclor  1016  contains 41%  chlorine).   Further,
metabolism of  PCBs  by bird and mammal species decreases in rate as the number
of chlorines in  the biphenyl molecule  increases.

      In 1973, Gibson et  al.  (36) isolated a species  of  Beijerinckia that
utilizes  biphenyl as  a sole source of  carbon for growth.  This organism
metabolized biphenyl  to cis-2,3-dihydroxy-l-phenylhexa-4,5-diene  (cis-2,3-
dihydro-2,3-dihydroxybiphenyl).   This  reaction is  catalyzed  by 2,3-
dioxygenase.   A  further intermediate,  catalyzed by dihydrodiol
dehydrogenase, was  identified as 2,3-dihydroxybiphehyl.   Continued
research (16,  17,  18, 47) verified  these metabolic steps  and  determined
that meta cleavage, catalyzed by 2,3-dihydroxybiphenyl  dioxygenase, between
Cl and  C2 of  the 2,3-dihydroxy  compounds  yields 2-hydroxy-6-oxo-6-
                                       231

-------
 phenylhexa-2,4-dienoate which can degrade to benzole acid.   Bacterial
 strains  are known to convert PCBs into the corresponding chlorobenzoates.
 Mondello and Yates (51) expand on this subject of metabolic pathway to degrade
 PCBs.

       Mlcrobial degradation of PCBs has been and is being  studied  by  a number
 of biochemists, microbiologists,  analytical chemists,  and molecular
 geneticists.   The widespread presence  of  PCBs in  the environment  is of concern
 to governments and the public worldwide thus there has been an abundance of
 research on various aspects of PCBs,  e.g.,  ecology,  toxicology,  biochemistry,
 and applied genetics.   There has  been  increasing  effort to  provide  a
 technology  utilizing some  process of biodegradation to destroy these polluting
 compounds leaving no harmful effect upon the environment.  The living  cell  is
 capable  of  highly efficient enzyme-catalyzed reactions and  man is on the verge
 of mastering the use of cells, most likely  specially  developed bacteria, to
 clean  up hazardous wastes  such as the  PCBs.   Table 1 lists  a few  references  to
 publications  on a variety  of aspects of the PCB problem  and Table  2.  consists
 of a list with references  of most of the  species  and strains of microorganisms
 that have  been used  in PCB biodegradation  research.   These publications can
 provide  excellent background material  on  the subject.
Table 1.  PCB Research References
SUBJECT: REFERENCE
SUBJECT: REFERENCE
Activated sludge: 40, 74, 75
Anaerobic strains: 59
Anaerobe strain B-206: 71, 72
Analog enrichment: 24
Analysis, all 209 congeners: 52
Analytical chemistry: 22
Animal toxicology: 42
Attenuation
  by earth materials: 37
Biodegradation
  of congeners: 7, 9
Biodegradation
  of hydroxybiphenyls: 43
Biphenyl metabolism: 47
Carcinogenic
  and other chronic effects: 41
Chemistry: 58
Early report: 57
Effect of glucose uptake: 64
Effect on nitrification: 65
Environmental
  dechlorination: 10, 12
Freshwater
  microbial populations: 64, 67
General: 21, 27, 61, 63
Human exposure: 20
Marine bacteria: 14
Metabolism
  and biochemical toxicity: 49
Molecular toxicology: 62
PCB intake in humans: 45, 78, 81
Persistence: 5
Photochemistry: 60
Progenitive manifestation:  44
Rapid assay: 8
Screening:  70
Structure:  69
Treatment technologies:  26
                                      232

-------
Table 2.   Microbial taxa  tested for PCS  degradative ability
TAXON
Achromobac ter spp.
Achromobacter sp. strain B-218
Achromobacter sp. strain BP
Achromobacter sp. strain pCB
Acidovirans group
Acinetobacter sp. strain P6
32, 34, 43
Acinetobacter sp. strain 4-CB1
Alkaligenes sp.
Alcaligenes sp.
Alcaligenes eutrophus strain H850
Alcaligenes faecalis strain Pi434
Alcaligenes odorans
Alcaligenes denitrif icans
Arthrobacter sp. strain BIB
Arthrobacter sp strain M5
Aspergillus flavus
Azotobacter sp. strain 4CB
Bacillus brevis sp. strain B-257
Beijerinckia sp. strain B8/36
Corynebacterium sp. strain MB1
Escherichia coli strain TB1
Escherichia coli strain FM4560
Nitrobacter agilis
Nitrosomonas europaea
Pseudomonas spp.
Pseudomonas sp. strain 1008
Pseudomonas sp. strain H1130
Pseudomonas sp. strain HBP1
Pseudomonas sp. strain JB1
Pseudomonas sp. strain LB400
Pseudomonas sp. strain MB86
Pseudomonas sp. strain Pi304
Pseudomonas sp. strain Pi939
Pseudomonas aeruginosa PA01161
Pseudomonas cepacia strain H201
Pseudomonas cepacia strain Pi704
Pseudomonas cepacia strain RJB
Pseudomonas paucimobilis Ql
Pseudomonas pseudoalcaligenes KF707
Pseudomonas putida
Pseudomonas putida
Pseudomonas putida strain LB400
Pseudomonas putida strain LB410
Pseudomonas testosteroni strain H128
Pseudomonas testosteroni strain H336
Pseudomonas testosteroni strain H430
Rhizoous -iaoonicus

80
48
3
3
76
1
1
31
32
7
76
19
19
43
30
53
38
48
36
11
51
50
65
65
80
66
76
43
56
50
6
76
76
29
76
76
76
73
29
17
16
54
76
76
76
76
77
REFERENCE
, 4
, 4
, 2, 13, 25, 30,
, 2
,34
, 9, 51, 54, 76, 82
, 37
, 37
, 7, 9, 76
, 51, 82
, 33
, 33, 73
, 18
, 68, 76
                                 233

-------
 1.
 2.
                              REFERENCES

Adriaens, P., and D.D. Focht.  Biodegradation of 4,4'-Dichlorobiphenyl
(4,4'-DCBP) by a Coculture of Two Acinetobacter Species.  Ann. Meeting
of the American Society for Microbiology, Miami Beach, FL. p.222, 1988.

Adriaens, P., and D.D. Focht.  4-Chlorobenzoate Metabolism by
Acinetobacter sp. 4-CB 1 and its Relevance to Biodegradation of 4,4'-
Dichlorobiphenyl.  In Environmental Biotechnology. Reducing Risks from
     Environmental Chemicals  through Biotechnology.  G.S.Omenn,  Ed.,
     Life Sciences,  Plenum Press,  New York.  Vol.  45,  p.442,  1988.
                                                                Basic
3.  Ahmed, M.  and D.D.  Focht.   Oxidation of  Polychlorinated  Biphenyls  by
    Achromobacter pCB.   Bull.  Environ.  Contam.  &  Tox.,  10(2):70-72,  1973.

4.  Ahmed, M.,  and D.D.  Focht.  Degradation  of  polychlorinated  biphenyls
    by  two species of Achromobacter.  Can. J^ Microbiol.  19:47-52, 1973.

5.  Ballschmiter,  K., M.  Zell,  and H.J.  Neu.  Persistence of  PCB's in  the
    Ecosphere:  Will Some PCB-Components  "Never" Degrade?  Chemosphere,
    2:173-176,  1978.

6.  Barton, M.R.,  and R.L. Crawford.  Novel  Biotransformation of
    4-Chloro-biphenyl by a Pseudomonas sp. Appl.  and Environ. Microb.,
    54(2):594-595,  1988.                          "

7.  Bedard, D.L.,  M.L. Haberl,  R.J. May,  and  M.J. Brennan.  Evidence for Novel
    Mechanisms  of  Polychlorinated Biphenyl Metabolism in Alcaligenes eutrophus
    H850.  Appl.  and Environ. Microb., 53(5):1103-1112, 1987T

8.  Bedard, D.L.,  R. Unterman,  L.H. Bopp, M.J.  Brennan, M.L. Haberl, and C.
    Johnson.   Rapid Assay for Screening and Characterizing Microorganisms for
    the Ability to  Degrade Polychlorinated Biphenyls. Appl. and Environ.
    Microb.. 51(4);761-768. 1986.                                     ~~

9.  Bedard, D.L.,  R.E.  Wagner,  M.J.  Brennan, M.L. Haberl, and J.F. Brown, Jr.
    Extensive Degradation of Aroclors and Environmentally Transformed Poly-
    chlorinated Biphenyls by Alcaligenes  eutrophus H850.  Appl. and Environ.
    Microb.. 53(5):.1094-1102, 1987.

10. Brown, J.F.,  Jr.,  D.L.  Bedard,  M.J.  Brennan, J.C.  Carnahan,  H.  Feng, and
    R.E. Wagner.  Polychlorinated Biphenyl Dechlorination in Aquatic
    Sediments.  Science, 236:709-712, 1987.

11. Brown, J.F.,  Jr.,  R.E.  Wagner,  D.L.  Bedard,  M.J.  Brennan,  J.C.  Carnahan,
    R.J. May, and T.J.  Tofflemire.  PCB Transformations in Upper Hudson
    Sediments. Northeast. Environ. Science, 3(3/4):167-179,  1984.

12. Brown, J.F., R.E. Wagner, H. Feng, D.L. Bedard,  M.J. Brennan, J.C.
    Carnahan, and R. J.  May.   Environmental  Dechlorination of PCBs.   Environ.
    Toxicol.  and Chem., 6:579-593, 1987.
                                      234

-------
13. Brunner, W., F.H.  Sutherland,  and D.D.  Focht.   Enhanced Biodegradation of
    Polychlorinated Biphenyls in Soil by Analog Enrichment and Bacterial
    Inoculation. J. Environ. Qual., 14(3);324-328, 1985.

14. Carey, A.E., and G.R.  Harvey.   Metabolism of Polychlorinated Biphenyls by
    Marine Bacteria.  Bull. Environ. Contain. Toxicol., 20:527-534, 1978.

15. Carpenter, B.H., and D.L. Wilson.  Technical/Economic Assessment of
    Selected PCB Decontamination Processes.  J._ Haz. Mat., 17:125-148, 1988.

16. Catelani, D., and A. Colombi.  Metabolism of Biphenyl. Structure and
    Physicochemical Properties of 2-Hydroxy-6-oxo-6-phenylhexa-2,4-dienoic
    Acid, the Meta-cleavage Product from 2,3-Dihydroxybiphenyl by
    Pseudomonas putida, Biochem. J., 143:431-434, 1974.

17. Catelani, D., A. Colombi, C. Sorlini, and V. Treccani.  Metabolism of
    Biphenyl. 2-Hydroxy-6-oxo-6-phenylhexa-2,4-dienoate: The Meta-Cleavage
    Product from 2,3-dihydroxybiphenyl by Pseudomonas putida.  Biochem. J.,
    134:1063-1066, 1973.        ,                            .

18. Catelani, D., C. Sorlini, and V. Treccani.  The Metabolism of Biphenyl
    by Pseudomonas putida.  Experientia, 27(10):1173-1174, 1971.

19. Clark, R.R., E.S.K. Chian, and R.A. Griffin.  Degradation of Poly-
    chlorinated Biphenyls by Mixed Microbial Cultures.  Appl. and Environ.
    Microb., 37(4):680-685, 1979.

20. Cordle, F., P. Corneliussen, C. Jelinek, B. Hackley, R. Lehman, J.
    Mclaughlin, R. Rhoden, and R. Shapiro.  Human Exposure to
    Polychlorinated Biphenyls and Polybrominated Biphenyls.  Envirbn.
    Health Persp., 24:157-172, 1978.   ...,.•

21. Dagley,  S.  Microbial  Degradation of Organic Compounds in the Biosphere.
    American Scientist, 63:682-689, 1975.

22. Erickson, M.D.  Analytical Chemistry of PCBs. Butterworth Publishers,
    Boston, MA, 508 pp.,1986.

23. Exner, J.H.,  Summary  of Polychlorinated Biphenyl Treatment Alternatives.
    In Detoxication of_ Hazardous Waste, J.H. Exner, Ed., Ann Arbor Science
    Publishers, Inc., Ann Arbor, MI, pp. 119-120, 1982.

24. Focht, D.D.  Analog Enrichment Decontamination  Process.  U.S.  Patent.
    Patent Number: 4,664,805. U.S. Patent and Trademark Office. 1987.

25. Focht, D.D., and W. Brunner.   Kinetics  of Biphenyl and Polychlorinated
    Biphenyl Metabolism in Soil.   Appl.  and Environ. Microb., 50(4): 1058-
    1063, 1985.
i
26. Fradkin, L., and S. Barisas. Technologies for Treatment, Reuse, and
    Disposal of_ Polychlorinated  Biphenyl Wastes.   Argonne National Laboratory
    ANL/EES-TM-168, 47 pp., 1982.
                                      235

-------
27. Furukawa, K.  Microbial Degradation of Polychlorinated Biphenyls
    (PCBs).  In Biodegradation and Detoxification of Environmental Pollutants,
    A.M. Chakrabarty, Ed., CRC Press, Inc., Boca Rafon, FL, pp. 33-57, 1982.

28. Furukawa, K.  Modification of PCBs by Bacteria and Other Microorganisms
    In PCBs and the Environment. J.S. Waid, Ed., CRC Press, Inc., Boca
    Raton, FL, Vol. II, pp. 89-100, 1986.

29. Furukawa, K., and N. Arimura.  Purification and Properties of 2,3-
    Dihydroxybiphenyl Dioxygenase from Polychlorinated Biphenyl-Degrading
    Pseudomonas pseudoalcaligenes and Pseudomonas aeruginosa Carrying the
    Cloned bphC Gene.  J. Bact., 169(2):924-927, 1987.

30. Furukawa, K., and A.M. Chakrabarty.  Involvement of Plasmids in Total
    Degradation of Chlorinated Biphenyls.  Appl. and Environ. Microb.,
    44(3):619-626, 1982.

31. Furukawa, K., and F. Matsumura.  Microbial Metabolism of
    Polychlorinated Biphenyls.  Studies on the Relative Degradability of
    Polychlorinated Biphenyl Components by Alkaligenes sp.  J_^ Agric. Food
    Chem.. 24(2):251-256, 1976.

32. Furukawa, K., F. Matsumura, and K. Tonomura.  Alcaligenes and
    Acinetobacter Strains Capable of Degrading Polychlorinated Biphenyls.
    Agric. Biol. Chem., 42(3):543-548, 1978.

33. Furukawa, K, and T. Miyazaki.  Cloning of a Gene Cluster Encoding
    Biphenyl and Chlorobiphenyl Degradation in Pseudomonas
    pseudoalcaligenes.  J. Bact., 166(2):392-398, 1986.

34. Furukawa, K., N. Tomizuka, and A. Kamibayashi.  Effect of Chlorine
    Substitution on the Bacterial Metabolism of Various Polychlorinated
    Biphenyls.  Appl. and Environ. Microb., 38(2):301-310, 1979.

35. Furukawa, K., K. Tonomura, and A. Kamibayashi.  Effect of Chlorine
    Substitution on the Biodegradability of Polychlorinated Biphenyls.
    Appl. and Environ. Microb., 35:223, 1978.

36. Gibson, D.T., R.L. Roberts, M.C. Wells, and V.M. Kobal.  Oxidation of
  .  Biphenyl by a Beijerinckia Species.  Biochem. and Biophys. Res. Comm.,
    50(2):211-219, 1973.

37. Griffin, R.A., and E.S.K. Chian.  Attenuation of Water-Soluble
    Polychlorinated Biphenyls by Earth Materials.  U. S. Environmental
    Protection Agency Research Report EPA-600/2-80-027, 93 pp., 1980.

38. Hernandez, B.S., and D.D. Focht.  Metabolism of Biphenyls, Chlorinated
    Biphenyls, and Ethylene by Azotobacter 4CB in N-free Media. Ann. Meet. Am.
    Society for Microbiology, Miami Beach, FL, p. 222, 1988.

39. Jensen, S.  Report of a new chemical hazard.  New Sci. , 32:612, 1966.

40. Kanekb, M., K. Morimoto, and S. Nambu.  The Respone of Activated Sludge
    to a Polychlorinated Biphenyl (KC-500).  Water Res., 10:157-163, 1976.
                                      236

-------
41. Kimbrough, R.D..  The Carcinogenic and Other Chronic Effects of
    Persistent Halogenated Organic Compounds.  In Health Effects of
    Halogenated Aromatic Hydrocarbons. Ann. N. Y. Academy of Sciences, The
    New York Academy of Sciences. New York. 320:415-418, 1979.

42. Kimbrough, R., J. Buckley, L. Fishbein, G. Flamm, L. Kasza, W. Marcus,
    S. Shibko, and R. Teske.  Animal Toxicology.  Environmental Health
    Perspectives, 24:173-185, 1978.

43. Kohler, H.-P.E., D. Kohler-Staub, and D.D. Focht.  Degradation of 2-
    Hydroxybiphenyl and 2,2'-Dihydroxybiphenyl by Pseudomonas sp. Strain
    HBP1. Appl. and Environ. Microb., 54(11):2683-2688, 1988.

44. Krockel, L., and D.D. Focht.  Construction of Chlorobenzene-Utilizing
    Recombinants by Progenitive Manifestation of a Rare Event.   Appl.  and
    Environ. Microb., 53(10):2470-2475, 1987.

45. Kuwabara, K., T. Yakushiji, I. Watanabe, S. Yoshida, K. Yoyama, and N.
    Kunita.  Increase in the Human Blood PCB Levels Promptly Following
    Ingestion of Fish Containing PCBs.  Bull. Environ. Contam. Toxicol.,  ,
    21:273-278, 1979.

46. Lauber, J.D..  Disposal and Destruction of Waste PCBs.  In PCBs and
    the Environment, J.S. Waid, Ed., CRC Press, Inc., Boca Raton, FL, Vol.
    Ill, pp. 83-151, 1987.
47. Lunt, D., and W.C. Evans.  The Microbial Metabolism of Biphenyl.
    Biochemical Journal, 118:54p-55p, 1970.
The
48. Masse, R., F. Messier, L. Peloquin, C. Ayotte, and M. Sylvestre.
    Microbial Biodegradation of 4-Chlorobiphenyl. A Model Compound of
    Chlorinated Biphenyls. Appl. Environ. Microbiol., 47(5):947-951, 1984.

49. Matthews, H., G. Fries, A. Gardner, L. Garthoff, J. Goldstein, Y. Ku,
    and J. Moore.  Metabolism and Biochemical Toxicity of PCBs and PBBs.
    Environ. Health Persp., 24:147-155, 1978.

50. Mondello, F.J.  Cloning and Expression in Escherichia coli of Genes
    Encoding Polychlorinated Biphenyl Degradation from Pseudomonas Strain
    LB400.  J_._ Bact., 171(3) :1725-1732, 1989.

51. Mondello, F.J.,  and J.R. Yates.   The  Development  of Recombinant Bacteria
    for Polychlorinated Biphenyl Degradation.  Proceedings Fifteenth Annual
    Hazardous Waste Symposium, Remedial Action, Treatment and Disposal of
    Hazardous Waste, In press.

52. Mullin, M.D., C.M. Pochini, S. McCrindle, M. Romkes, S.H. Safe, and
    L.M. Safe.  High-Resolution PCB Analysis: Synthesis and Chromatographic
    Properties of All 209 PCB Congeners.  Environ. Sci. Technol., 18(6):468-
    476, 1984.
                                      237

-------
53. Murado, M.A., MaC. Tejedor, and G. Baluja.  Interactions Between Poly-
    chlorinated Biphenyls (PCBs) and Soil Microfungi.  Effects of Aroclor-1254
    and Other PCBs on Aspergillus  flavus  Cultures.   Bull.  Environ. Contam. &_
   Tox., 15(6):768-774, 1976.

54. Nadim, L.M., M.J. Schocken, F.K. Higson, D.T. Gibson, D.L. Bedard, L.H.
    Bopp, and F. J. Mondello.  Bacterial Oxidation of Polychlorinated
    Biphenyls. Proceedings  of the Thirteenth Annual  Research Symposium,  Land
    Disposal, Remedial Action,  Incineration and Treatment  of Hazardous Waste,
    EPA/600/9-87/015, pp. 395-402, 1987.

55. Ohmori, T., T. Ikai, Y.  Minoda,  and K.  Yamada.   Utilization of polyphenol
    and polyphenol-related compounds by microorganisms. Agric. Biol. Chem.,
    37:1599, 1973

56. Parsons, J.R., and D.T.H.M. Sijm.  Biodegradation Kinetics of Polychlor-
    inated Biphenyls in Continuous Cultures of a Pseudomonas Strain.
    Chemosphere, 17(9):1755-1766, 1988.

57. Peakall, D.B., and J.L.  Lincer.   Polychlorinated Biphenyls. Another Long-
    Life Widespread  Chemical in the  Environment.  BioScience, 20(17):958-964,
    1970.

58. Pomerantz, I., J. Burke, D. Firestone, J. McKinney, J.  Roach, and W.
    Trotter. Chemistry of  PCBs and PBBs. Environ.  Health  Persp., 24:133-146,
    1978.

59. Quensen  III,  J.F.,  J.M. Tiedje, and  S.A.  Boyd.   Reductive Dechlorination
    of Polychlorinated Biphenyls by  Anaerobic Microorganisms from Sediments.
    Science, 242:752-754, 1988.

60. Ruzo, L.O., M.J. Zabik, and R.D. Schuetz.  Photochemistry of Bioactive
    Compounds: Photoproducts and Kinetics of Polychlorinated Biphenyls.  J_^
    Agr. Food Chem. , 22(2):199-202, 1974.

61. Safe, S.H.  Microbial Degradation of Polychlorinated Biphenyls.  In Micro-
    bial Degradation of Organic Compounds, D.T.Gibson, Ed., Marcel Dekker,
    Inc., New York, NY, pp.361-369, 1984.

62. Safe, S., L.W.  Robertson,  L. Safe, A. Parkinson,  S. Bandiera,  T. Sawyer, and
    M.A. Campbell.  Halogenated Biphenyls: Molecular Toxicology.  Can. J._
    Physiol. Pharmacol., 60:1057-1064, 1982.

63. Sahasrabudhe,  S.R.,  and V.V.  Modi.   Microbial  Degradation of Chlorinated
    Aromatic Compounds.  Microbiological Sciences, 4(10):300-303, 1987.

64. Sayler, G.S.,, L.C.. Lund, M.P. Shiaris, T.W. Sherrill, and R.E. Perkins.
    Comparative Effects of Aroclor 1254 (Polychlorinated Biphenyls) and
    Phenanthrene on Glucose Uptake by Freshwater Microbial Populations.
    Appl. and Environ. Microb., 37(5):878-885, 1979.

65. Sayler, G.S., M.P. Shiaris, W. Beck, and S. Held.  Effects of Polychlor-
    inated Biphenyls  and Environmental Biotransformation  Products  on Aquatic
    Nitrification.  Appl. and Environ. Microb., 43(4):949-952, 1982.
                                      238

-------
66. Sayler, G.S., M. Shon, and R.R. Colwell. Growth of an Estuarine
    Pseudomonas sp. on Polychlorinated Biphenyl. Microb. Ecol., 3:241-255,
    1977.                                              ;

67. Shiaris, M.P. , and G.S. Sayler.  Biotransformation of PCB  by Natural
    Assemblages of Freshwater Microorganisms.  Environ^ Sci. Technol.,
    16(6):367-369, 1982.                                           ;

68. Singh, S., F.K. Higson, L.M. Nadim, and D.T. Gibson.  Oxidation of Poly-
    chlorinated Biphenyls by Pseudomonas putida LB400.  Proceedings of the
    Fourteenth Annual Research Symposium, Land Disposal, Remedial Action,
    Incineration  and Treatment  of  Hazardous  Waste, EPA/600/9-88/021,  pp. 346-
    354, 1988.

69. Sissons,  D.,  and D. Welti.  Structural  Identification of Polychlorinated
    Biphenyls in Commercial Mixtures by Gas-Liquid Chromatography Nuclear
    Magnetic Resonance and Mass Spectrometry.  J._ Chromatogr., 60:15-32, 1971.

70. Sylvestre, M..  Screening Method for the Isolation of Polychlorinated
    Biphenyls Degrading Bacteria.  80th Annual Meeting American Society for
    Microbiology, Miami Beach, Fla., May  11-16,  1980.  Abstracts  of the Annual
    Meeting of the American Society for Microbiology,  Abstract 236,  Vol. 80,
    1980.

71. Sylvestre, M., and J. Fauteux.   A New Facultative Anaerobe Capable of
    Growth on Chlorobiphenyls.  J.  Gen. Appl. Microbiol., 28(l):61-72, 1982.

72. Sylvestre, M., R.  Masse,  F.  Messier,  J. Fauteux,  J.-G.  Bisaillon,  and R.
    Beaudet.  Bacterial Nitration of 4-Chlorobiphenyl.  Appl. and Environ.
    Micro., 44(4):871-877, 1982.                                  ::[~

73. Taira, K., N. Hayase, N.  Arimura,  S.  Yamashita,  T. Miyazaki,  K.  Furukawa.
    Cloning and  Nucleotide  Sequence of the  2,3-Dihydroxybiphenyl Dioxygenase
    Gene from the PCB-Degrading Strain of Pseudomonas paucimobilis Ql.
    Biochemistry, 27:3990-3996, 1988.

74. Tulp, M.T.M., R.  Schmitz, and  0. Hutzinger.   The  Bacterial Metabolism of
    4,4'-Dichlorobiphenyl and Its Suppression by Alternative Carbon Sources.
    Chemosphere, 1:103-108,  1978.

75. Unterman, R., D.L.  Bedard, M.J. Brennan,  L.H. Bopp, F.J. Mondello, R.E.
    Brooks, D.P. Mobley, J.B.  McDermott,  C.C. Schwartz, and O.K.  Dietrich.
    Biological Approaches for Polychlorinated Biphenyl Degradation.  In
    Environmental Biotechnology. Reducing Risks from Environmental
    Chemicals through Biotechnology. G.S. Omenn, Ed.,  Plenum Press,  New York,
    pp. 253-269, 1988.

76. Wallnofer, P.R.,  G. Engelhardt, S.  Safe,  and 0. Hutzinger.   Microbial
    Hydroxylation of  4-Chlorobiphenyl and  4,4'-Dichlorobiphenyl.   Chemosphere,
    2:69-72, 1973.                                                           '
                                      239

-------
77. Watanabe, I., T. Yakushiji, K. Kuwabara, S. Yoshida, K. Maeda, T.
    Kashimoto, K. Koyama, and N. Kunita.  Surveillance of the Daily PCB Intake
    from Diet of Japanese Women from 1972 Through 1976.  Arch. Environ.
    Contam. Toxicol., 8:67-75, 1979.
                             'k
78. Weitzman, L.   Treatment and Destruction of  Polychlorinated Biphenyls and
    Polychlorinated Biphenyl-Contaminated Materials.  In Detoxication of_
    Hazardous Waste, J.H. Exner, Ed.,  Ann Arbor  Science Publishers, Inc., pp.
    131-142, 1982.

79. Wong, P.T.S., and K.L.E. Kaiser.  Bacterial Degradation of Polychlorinated
    Biphenyls II.  Rate  Studies.   Bull. Environ. Contam. & Toxic., 13(2):249-
    256, 1975.

80. Yakushiji, T., I. Watanabe, K. Kuwabara, S. Yoshida, K. Koyama, and N.
    Kunita.  Levels of Polychlorinated Biphenyls (PCBs) and Organochlorine
    Pesticides in Human Milk and Blood Collected in Osaka Prefecture from 1972
    to 1977.  Int. Arch. Occup. Environ. Health,  43:1-15, 1979.

81. Yates, J.R., and F.J. Mondello.  Sequence Similarities in the Genes
    Encoding Polychlorinated Biphenyl Degradation from Pseudomonas  Strain LB400
    and Alcaligenes eutrophus H850.  J^ Bact., 171(3):1733-1735, 1989.
                                      240

-------
                 The Development of Recombinant Bacteria for
                    Polychlorinated Biphenyl Degradation

                    Frank J. Mondello and James R. Yates

                     GE Research and Development Center
                              S chenec tady, N.Y.
                                  ABSTRACT

    Pseudomonas strain LB400 and Alcaligenes eutrophus strain H850 have
previously been demonstrated to degrade an unusually wide variety of
polychlorinated biphenyls  (PCBs).  The genes encoding the PCS degradative
enzymes (the bph genes) from these organisms were isolated using the
broad-host-range cosmid vector pMMB34, and found to be expressed in
Escherichia coli.  A comparison of the PCB degradative capabilities of the
wild-type and recombinant  strains was conducted using resting-cell assays.
Significant improvements in the activity of the recombinant strains were
observed after plasmid modifications.  The degradation of a variety of PCB
mixtures including Aroclor 1242 was found to be comparable for LB400 and the
recombinant strain FM4560.
    DNA:DNA hybridization  analysis was used to determine that the genes
encoding PCB degradation in LB400 and H850 are genetically distinct from
those in a variety of other organisms.  This indicates the existence of at
least two different classes of genes for PCB metabolism.  The availability of
DNA probes for the LB400/H850 class of bph genes will make it possible to
determine the fate of recombinant strains or plasmids containing these genes
in bioremediation processes.


INTRODUCTION

    Polychlorinated biphenyls (PCBs) consist of a biphenyl molecule
containing from 1 to 10 chlorines, making it possible to produce 209
different PCB congeners (1).  Commercially, PCBs were used and discarded as
complex mixtures (known as Aroclors or Kanechlors),  which contain 60-80
different congeners (1).   Effective processes for the biodegradation of
environmental PCBs will therefore require organisms capable of attacking a
wide variety of these congeners. Few PCB degrading strains isolated thus far
are capable of degrading a broad range of highly chlorinated PCBs.  Two
strains with outstanding PCB degrading ability are Pseudomonas strain LB400
and Alcaligenes eutrophus H850,  which can degrade PCB molecules containing up
to six chlorines (2,3).

    The PCB degradative abilities of LB400 and H850 are similar.  In both
organisms this process involves two different types of oxidative attack (4).
At least one type of attack is performed by enzymes of the bph pathway,  which
are known to be involved in the metabolism of biphenyl (Figure 1).   The first
enzyme in this pathway is a 2,3-dioxygenase which inserts oxygen at carbon
                                     241

-------
positions 2 and 3 of biphenyl.  This type of activity has been found in all
known PCS degrading bacteria and results in the formation of 2,3-dihydrodiol
(5). Dihydrodiol dehydrogenase then converts this compound to 2,3-dihydroxy-
biphenyl which undergoes extradiol cleavage by 2,3-dihydroxybiphenyl
dioxygenase to form a yellow meta-cleavage product.  This compound is
converted by a hydrase to benzoic acid (1,4,6).
          bprtA
                                         ,OH

                                        COOH
                                       P     bphD
                                                                      COOH
 Clx
Clx
Clx
Clx
Clx
Figure 1.  Degradation of Biphenyl and PCBs by the 2,3- Dioxygenase Pathway.
Gene designations: bphA, biphenyl 2,3-dioxygenase; bphB, dihydrodiol
dehydrogenase; bphC, 2,3-dihydroxybiphenyl dioxygenase; bphD, meta-cleavage
product hydrase


    LB400 and H850 also contain a much less common 3,4-dioxygenase activity,
which adds oxygen atoms at carbon positions 3 and 4 (4).  This activity
may explain the unusually broad congener specificity of these organisms,
enabling them to degrade congeners recalcitrant to degradation by other
organisms.  For example, many PCB degrading bacteria are unable to degrade
2,5,2',5'- tetrachlorobiphenyl, presumably because all of the 2,3-ring
positions are blocked by chlorines. This congener is readily attacked by
LB400 and H850, resulting in the formation of a 3,4-dihydrodiol (4). It is
not yet known if 3,4-dioxygenase activity is encoded by bphA of the
2,3-dioxygenase pathway.

    The exceptional degradative capabilities of LB400 and H850 make the
isolation and analysis of the bph genes from these organisms an important
first step towards the development of new strains with enhanced abilities to
degrade PCBs.  The purpose of this report is to summarize recent progress
made toward this goal.  A more detailed description of some of this work will
appear in the March 1989 issue of the Journal of Bacteriology.

CLONING THE LB400 AND H850 GENES FOR PCB DEGRADATION

    Genomic libraries of LB400 and H850 were constructed using the
wide-host-range, mobilizable cloning vector pMMB34 (7), and introduced into
an Escherichia coli host via standard procedures.  Ampicillin resistant
colonies were tested for the presence of 2,3-dihydroxybiphenyl dioxygenase
activity.  This enzyme is encoded by the bphC gene of the biphenyl/PCB
degradative pathway, and catalyzes the conversion of 2,3-dihydroxybiphenyl to
a yellow compound, 2-hydroxy-6-oxo-6-phenyl hexa-2,4-dienoate (4).  Colonies
expressing bphC were identified by their accumulation of visible quantities
                                     242

-------
of this product.  Clones from both LB400 and H850 were found to express
2,3-dihydroxybiphenyl dioxygenase activity.  The bphG-containing plasmids
from LB400 were designated pGEM410, 420, 430.  Those from H850 were
designated pGEMSOO, 810, 830, and 850.  These strains were examined for the
presence of additional enzymes of the PCB/biphenyl pathway by using a series
of rapid screening procedures (4,6).  Each of the recombinant plasmids
contained several bph genes and many encoded all of the enzymes required to
convert PCBs to chlorobenzoic acids (Table 1). These data indicated that the
genes for PCS degradation in both LB400 and H850 were closely linked. A
similar result has been reported by Furukawa and Miyazaki for bphA, B, and C
of Pseudomonas pseudoalcaligenes KF707 (8).                         —      -


        TABLE 1.  BPH-PATHWAY GENES ENCODED BY RECOMBINANT PLASMIDS
                                 Genes Expressed
PLASMID
pGEM410
pGEM420
pGEM430
pGEMSOO
pGEMSlO
pGEM830
pGEM850
pMMB34
bphA bphB bphC bphD
+ + > • + +
        Enzymatic activity present (+) or absent (-)


PCB METABOLISM BY RECOMBINANT STRAINS                                 ;

     E. coli strains containing pGEM plasmids were tested for their ability
to degrade polychlorinated biphenyls in resting cell assays.  The strain with
the highest activity was FM4100 (containing plasmid pGEM410) and was chosen
for further study.  Incubation of this strain with either 2,3-, 2,5-, or
2,2'-di- chlorobiphenyl resulted in the accumulation of,2,3-, 2,5- and
2-chlorobenzoic acid, respectively.  This unequivocally demonstrated the
expression of at least the first four enzymes of the 2,3-dioxygenase pathway
of PCB metabolism.

     The ability of LB400 and H850 to degrade 2,5,2',5'-tetrachlorobiphenyl
(2,5,2',5'-CB) sets them apart from other PCB degrading strains.  This
congener has no free adjacent 2,3 ring positions, and is not readily oxidized
by the 2,3-dioxygenase enzymes of most PCB degrading bacteria (2).  LB400 and
                                     243

-------
H850 catabolize 2,5,2',5'-CB via a 3,4-dioxygenase to a compound identified
as 3,4-dihydroxy-3,4-dihydro-2,5,2',5'-tetrachloro- biphenyl (3,4-dihydro-
diol) (4).  Degradation of 2,5,2',5'-CB by FM4100 resulted in the production
of a metabolite with a GG profile identical to that of purified
3,4-dihydrodiol (provided by D.T. Gibson) demonstrating the existence of
3,4-dioxygenase activity in the recombinant strain.

MAPPING STUDIES

    Digestion of pGEM410 with the endonuclease EcoRI results in the
formation of nine DNA fragments whose sizes and relative order are 15.5, 6.1,
2.9, 0.8, 2.0, 6.7, 2.9, 2.3, and 0.5 kilobase pairs (kb).   The relative
positions of these fragments were determined by analyzing subclones produced
from an incomplete EcoRI digest of pGEM410 DNA.  The partially-digested DNA
was ligated into pUC18 and this mixture was transformed into E. coli.
Plasmids were isolated from several transformants and subjected to
restriction analysis with the enzyme EcoRI. These subclones contained
overlapping multi-fragment segments of the original recombinant plasmid.
Alignment of these fragments resulted in the map shown in Figure 2.
DC DC CC DC
0 O Ol O
O O 0 O
in uj nil uj
ft^^^MMMMM^^^^^^MMMBMMB^^^^H^^^^^^^^^W^tHM^B^H
6.1 2.9 0.8 2.0



SC37
SC46
SC46A
SC56
O/^CT
DC CC DC DC DC
O O O O O
0 O O O O
UJ UJ UJ UJ 111
2.9 6.7 2.3 0.5
II
J
II I
II I
II II
I III
I I I
II I
III
Figure 2. Mapping the EcoRI sites of the pGEM410 insert. The positions of
these sites are  indicated at the top. The inserts of several subclones were
mapped and aligned as shown. Subclone designations are shown at the left.
                                     244

-------
    Recombinant strains containing these subclones were tested for various
enzymatic activities associated with biphenyl metabolism. An examination of
the activities associated with these fragments revealed that the bph genes
were grouped together within a 12.4 kb region of DNA.  A partial restriction
map of this region is shown in Figure 3.
           2.9kb
                                                  2.3kb
   0.5kb
                                                                     VECTOR
                               bph B,C
bph D
Figure 3.  Partial restriction endonuclease map of the region encoding PCB
degradation in pGEM410.  Gene designations are as described in Figure 1.


COMPARING PCB DEGRADATION BY RECOMBINANT AND WILD-TYPE STRAINS

    The PCB degrading abilities of two recombinant strains, FM4110 and
FM4560, were compared to that of LB400.  FM4110 is E. coli strain TB1
containing the original recombinant plasmid pGEM410.  FM4560 is E. coli
strain TB1 containing a derivative of pGEM410. This derivative was produced
by inserting the 2.9 and 6.7 kb EcoRI fragments (encoding the initial three
enzymes of the Bph pathway) into the vector pUClS.

    Resting cell assays were used to test the ability of FM4110 and FM4560
to degrade PCB mixes IB and 2B (9).  The recombinant strains were grown using
succinate as the carbon and energy source since they will not grow on
biphenyl.  PCB degradation by FM4110 was significantly lower than that by
LB400 for many tetra-, penta- and hexachlorinated congeners (Table 2).
Furthermore, PCBs with chlorines at both para ring positions were not
attacked by FM4110.

    FM4560 showed a substantial increase in PCB degrading ability over
FM4110, and demonstrated activity similar to that of LB400 for a wide variety
of tetra-, penta-, and double-para substituted PCBs (Table 2) .  FM4560 was
also similar to LB400 in its ability to degrade Aroclor 1242, a complex PCB
mixture containing 60-80 different congeners. In resting cell assays with 10
ppm of Aroclor 1242, FM4560 and LB400 degraded 85 and 91% of the PCBs,
respectively.  As shown in Figure 4, the patterns of depletion demonstrated
by the two strains appear nearly identical.  These results demonstrate the
ability of this recombinant strain to degrade a PCB mixture that is. an actual
environmental contaminant.
                                     245

-------
    TABLE 2.   DEGRADATION OF PCB  CONGENERS BY LB400 AND RECOMBINANT STRAINS


PCB
Congener
2,3
2,2'
2,4'
2,5,2'
2,5,4'
2, 5, 2', 5'
2, 3, 2', 3'
2,3, 2', 5'
9 5 3' 4'
^» J » J »**•
2452' 5'
i. ,*-t-, -i , ^ , _/
234?' 5'
^ , .J ,*+, Z. , J
9 4 5 9 ' 3'
^»^» J ,^ » °
4,4'
2,4,4'
2, 4,3', 4'
2, 4,2', 4'
3,4, 3', 4'
2, 4, 5, 2', 4', 5'


LB400
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
*****
****
***
*****
***
*****
**
***
Percent Depletiona
FM
4110
*****
*****
*****
*****
*****
****
***
***
***
*
-
- -
_
-
-
-
-
—

FM
4560
*****
*****
***** '•'
*****
*****
*****
*****
*****
*****
*****
***
***
**
****
*
**
-
—
    testing Cell Assays- Mix  IB,  2B. Percent depletion: ***** - 80-100%;
    **** - 60-79%;  *** = 40-59%; ** = 20-39%; * = 10-19%; - = 0-9%.
B
      JLL
.ju-M-jJ	WU~
Figure 4.  Degradation of Aroclor
1242 by E. coli FM4560 and
Pseudomonas LB400.  (A) Aroclor 1242
(10 ppm) incubated with mercury
killed cells; (B and C) Aroclor 1242
incubated at 30 C for 24 h with cells
(optical density at 615 nm of 1.0)
of FM4560 and LB400 respectively.
                                      246

-------
COMPARING THE GENES FOR PCB DEGRADATION FROM DIFFERENT ORGANISMS

    The PCB-degrading abilities of Pseudomonas strain LB400 and A. eutrophus
H850 are significantly different from those of other bacteria.  This is
reflected both in the wide range of congeners they can degrade and in the
presence of 3,4-dioxygenase activity.  This might indicate that the bph genes
of LB400 and H850 are similar to each other, and substantially different from
those in other PCB degrading strains.  This possibility was examined using
DNA:DNA hybridization experiments.

    Genomic DNAs from nine different bacterial strains were examined for
sequences similar to the bph genes of LB400.  The selected strains varied
widely in PCB degrading ability. This group included representatives from
four different genera, and six species (see figure legend for a list of the
organisms tested).   Figure 5 shows the results of an experiment in which a
plasmid containing the bphAB and C genes of LB400 was used as a probe.  This
probe hybridized to fragments of the A. eutrophus H850 genome. Similar
results were obtained using probes containing bphD (data not shown). Probes
with other fragments of LB400 DNA (i.e., non-bph gene fragments) did not
hybridize to H850 DNA. Genomic DNAs from the other PCB degrading strains did
not have sequences that hybridized to these probes. Therefore, the bph genes
of these strains cannot be very closely related to those of LB400 and H850.
These data indicate that there are at least two distinct varieties of genes
for PCB degradation.  Also, similarities in the LB400 and H850 bph genes
explain the very similar PCB-degrading capabilities of the two organisms.

FUTURE PROSPECTS

    Recombinant strains of E. coli that express the LB400 and H850 bph genes
have been constructed.  While significant improvement in the PCB degrading
ability of these strains has already been achieved through genetic
manipulations, further studies designed to increase and regulate the
expression of the LB400 bph genes are currently in progress.

    The discovery that the entire bph pathways of LB400 and H850 were
genetically distinct from those of many other PCB degrading strains was
surprising.  All of the organisms examined are capable of converting PCBs to
chlorobenzoic acids and it might be expected that some of the genes in these
pathways would be similar to those of LB400. As bph genes from other
organisms become available we will attempt to examine their relationship to
those of LB400, and attempt to correlate specific biochemical activities with
specific types of bph genes.

     The recombinant strains have thus far been examined only for their
ability to degrade PCBs in resting cell assays.  Since it is important to
evaluate the degradation of PCBs as they are actually found in the
environment, the abilities of these, and future, recombinant organisms to
degrade PCBs on soil will be examined in the laboratory.

    Previous studies using LB400 to degrade PCBs on contaminated soils have
been promising, but it is likely that degradation could be increased if cell
survivability could be improved (10).  This could be accomplished by using a
wide-host-range mobilizable plasmid to introduce the cloned LB400 bph genes
into an organism indigenous to a contaminated soil. The resulting strain may
                                     247

-------
Figure 5. Autoradiogram of Southern blot where several PCB-degrading strains
were examined for bph genes similar to those of LB400. All DNAs were digested
with EcoRI and probed with pGEM415 (containing bphAB and G).  pGEM415,
positive control; C600, E. coli (negative control); LB400, Pseudomonas sp.;
H850, Alcaligenes eutrophus; MB1, Corynebacterium sp.; H336 Pseudomonas
testosteroni; H430, P.  testosteroni; Pi434, Alcaligenes fecalis; H1130,
Pseudomonas sp.  (acidovarians group); H201, Pseudomonas cepacia.
                                     248

-------
combine the superior PCB degrading ability of LB400 with the survivability of
the indigenous organism.

    The fate of recombinant strains and plasmids in the environment is a
question which must be addressed for any bioremediation process utilizing
genetically engineered organisms.  Studies to assess the environmental
transfer, mobility, and persistence of recombinant molecules often employ DNA
hybridization methods because they are highly sensitive and specific.  The
availablity of DNA probes for the LB400/H850 bph genes makes it possible to
use these techniques to determine the stability of the cloned genes in the
environment and to detect them even in organisms where they are not
expressed.

                              ACKNOWLEDGEMENTS

    This work was supported by Grant CR812727 from the U.S.  Environmental
Protection Agency, Office of Research and Development, Hazardous Waste
Engineering Research Laboratory, Cincinnati, OH.  The authors thank Dr. P.R.
Sferra, EPA Project Officer for his interest, support and suggestions.


                                 REFERENCES

1.  Rochkind, M.L., Blackburn, J.W., andG.S. Sayler.  Chlorinated biphenyls.
    In, Microbial decomposition of chlorinated aromatic compounds.  United
    States Environmental Protection Agency, Cincinnati OH, 1986.  129 pp.

2.  Bedard, D.L., R. Unterman, L.H. Bopp, M.J. Brennan, M.L. Haberl, and C.
    Johnson.  Rapid assay for screening and characterizing microorganisms
    for the ability to  degrade polychlorinated biphenyls.  Appl. Environ.
    Microbiol. 51: 761, 1986.

3.  Bopp, L.H.  Degradation of highly chlorinated PCBs by Pseudomonas strain
    LB400.  J. Ind. Microbiol.  1: 23, 1986.

4.  Nadim, L., M.J. Schocken, F.J. Higson, D.T. Gibson, D.L.  Bedard, L.H.
    Bopp, and F.J. Mondello.  Bacterial oxidation of polychlorinated
    biphenyls.   Proceedings of the 13th Annual Research Symposium on Land
              Remedial  Action, Incineration, and Treatment of Hazardous
               395, 1987.
Disposal,
Waste,  p
                                                                       A.M.
Furukawa, K.  Microbial degradation of polychlorinated biphenyls.
Chakrabarty, (ed.), In, Biodegradation and detoxification of
environmental pollutants, CRC Press, Inc., Boca Raton, Florida, p.  33,
1982.

Mondello, F.J. and L.H. Bopp.  Genetic and cell-free studies of PCB
biodegradation in Pseudomonas putida LB400.  Proceedings: Biotech USA
1987.  Online International Inc.  p. 171, 1987.

Frey, J., M. Bagdasarian, D. Feiss, F.C.H. Franklin, andJ.  Deshusses.
Stable cosmid vectors that enable the introduction of cloned fragments
into a wide variety of Gram-negative bacteria. Gene 24: 299, 1983.
                                     249

-------
8.  Furukawa, K.,  and T. Miyazaki.  Cloning of a gene cluster encoding
    biphenyl and chlorobiphenyl degradation in Pseudomonas
    pseudoalcallgenes.   J. Bacteriol. 166: 392, 1986.

9.  Bedard, D.L.,  M.L.  Haberl, R.J. May, and M.J. Brennan.  Evidence for
    novel mechanisms of polychlorinated biphenyl metabolism in Alcaligenes
    eutrophus H850.  Appl. Environ.  Microbiol. 53:  1103,  1987.

10. Unterman, R.,  D.L.  Bedard, M.J. Brennan, L.H. Bopp,  F.J.   Mondello,
    R.E. Brooks, D.P. Mobley, J.B. McDermott, C.C.  Schwartz and D.K.
    Dietrich.  Biological approaches for PCB degradation.   In: Omenn
    et.al., (eds.), Reducing Risks From Environmental Chemicals Through
    Biotechnology.  Plenum Press, London.  1988.  p. 253.
                                    250

-------
                  Treatment of Wood Preserving
                        Soil Contaminants
                       by White Rot Fungus
                         John A.  Glaser
          United States Environmental Protection Agency
              Risk Reduction Engineering Laboratory
                   26 Martin Luther King Drive
                     Cincinnati,  Ohio 45268
          Rich Lamar,  Diane Dietrich, and T. Kent Kirk
             United States Department of Agriculture
                   Forest Products Laboratory
                     1 Gifford Pinchot Drive
                    Madison, Wisconsin 53706

                            ABSTRACT
      A wood degrading fungus, Phanerochaete chrysosporium has
been the object of considerable attention for its potential
application to hazardous waste degradation. The development of a
field soil treatment technology based on this fungus has been the
focus of an intense research program. Early stages of this work
sought to determine ways to assist the growth of the fungus in
soil, an environment not known to be sought by this fungus. Once
general methodology was established to promote its sustained
growth then studies were pursued to quantitatively determine the
extend of biodegradation attributable to the fungus using
surrogate soils(artificially contaminated with wood preserving
waste constituents). Work with pentachlorophenol and related
polyaromatic hydrocarbons has paved the way to undertake limited
field trials during the North American growing season of 1989.
The related bench scale experimental work and some information
concerning the field trials will be presented.

                          Introduction

     The use of specialized or selected microorganisms  to degrade
soil bound contamination has received considerable interest. This
situation is attributable in part to the interest of those
responsible for cleanup actions to use cost effective and
environmentally compatible means to treat wastes.  In spite of
these very promising aspects, biological detoxification, as a
site cleanup technology, must be recognized as a developing
technology. Biological  treatment has excellent credentials in the
areas of municipal and  industrial wastes but awaits development
for  the  treatment of mixtures of more toxic and  persistent
chemicals found as components of hazardous waste sites.
                                251

-------
                        Wood Treating Waste [1]

      A major  hazardous  waste problem confronting regulatory
 authorities  in the  United .States is  the waste associated with the
 wood  treatment industry. Depending on the  age of a facility,  at
 least three  technologies have contributed  to  the accumulated
 waste.  Creosote  treatment was followed by  pentachlorophenol which
 was replaced  with copper chromated arsenite.  Each of these
 technologies  present  its special conditions  for  cleanup.  Creosote
 is derived from  coal  tar production  and usually  contained a host
 of compounds  ranging  from straight aromatic  compounds  to
 polyaromatic  species  including smaller quantities of aromatic
 nitrogen bases and  an array of phenolic compounds.  The lower
 vapor pressure components of this mixture  contribute to the
 residuals found  at  such sites.  Pentachlorophenol,  a potent
 fungicide, is  a  major contaminant at wood  treating site that
 exhibits significant  toxicity towards microflora.  The  analysis of
 wastes  derived from this technology  have identified other
 potential toxic  components.  For our  current development efforts,
 we have narrowly focussed on a significant portion of  the waste
 including the  major contributors that are  polycyclic aromatic
 compounds and  phenols.  It is necessary to  limit  the scope of
 contamination  treatable by  this technology to permit the
 development activity  to be  achieveable in  a reasonable time
 frame.

                        Bacteria vs  Fungi  [2]

      Microorganisms (both bacteria and fungi) are  known to
 possess a variety of  detoxification  skills [3].  Xenobiotic
 chemical pollutants generally  do not  provide  sufficient energy to
 sustain many microorganisms.  The biological degradation of  such
 substrates occurs as  part of a  cometabolic activity, where  the
 organism's growth is  maintained by specific substrates  and  the
 detoxification activity of  other materials non-growth  substrates
 ensues. Many bacteria and fungi  can  accomplish simple
 transformations  on  organic  substrates  but  often  fail to complete
 the conversion of toxicant  substrate  to  carbon dioxide  or
 generate a toxic intermediate  that can impair the  growth  of the
microorganism. The  use of bacterial  communities  recognizes  these
deficiencies through  the combined  use  of many species  where the
abilities of one species supplants the  inadequacies  of  another.
Since the collective  action  of  these  communities  is  important to
 treatment success,   it is important to  protect them  from
environmental  effects that may  adversely affect  the  communities
 [4].

      Fungi have not been investigated  to any extent for use  as
degraders of waste materials until recently [5].  Sewage treatment
                               252

-------
operations steered clear of filamentous fungi due to processing
problems and the possibility that such fungi may be pathogenic.
Exceptions to these generalization do exist  [6]. A wood rotting
basidiomycetes,  Trametes versicolor, was studied, twenty-five
years ago, in an attempt to quantify its ability to degrade
chlorinated phenols [7]. Wood preservative chemicals have been
found to "be degraded by fungi [8,9].

     A wood degrading fungus, Phanerochaete  chrysosporium
characterized by fast growth and easy reproductive cycles
degrades an extensive list of hazardous waste consistuents under
laboratory conditions. This ability to degrade hazardous
pollutants appears to corellate well with the fungus1 ability to
degrade lignin, a complex natural polymer composed of
phenylpropane units that is resistant to decay by many
microorganisms. This lignin degrading ability is attributable to
a complex mixture of enzymes secreted by the fungus to the
extracellular medium. The enzymes are peroxidases that utilize
hydrogen peroxide from complementary enzyme  systems to perform
the initial oxidative conversion of pollutant substrates.

     Some of the more common substructures of lignin resemble the
chemical structure of many persistent organic compounds
contaminating the environment.  This structural similarity gave
sufficient reason to pursue application of a white rot fungus,
Phanerochaete chrysosporium to the biodegradation of hazardous
waste constituents [10].

     Phanerochaete chrysosporium is a filamentous, white, wood
rotting fungus and has been classified as a  member of the
Hymenomycetes subclass of Basidiomycetes  [11]- To distinguish
them from bacteria, fungi are eukaryotic, ie. they possess a
nuclear membrane and as microorganisms are considered to be
plantlike without chlorophyll having no photosynthetic abilities
[12].

       Carbon Substrate Degradation by Wood  Rotting Fungi

     White rot fungi are primary wood degraders in nature[10].
They excel in their ability to recycle carbon of wood origin when
compared with brown rot fungi. The naturally occuring polymers  of
cellulose and lignin are degraded by these fungi forming the
major sources of carbon to assist fungal growth. Lignin is by far
the more difficult to degrade due to its  composition as a
heteropolymer formed from the cross linking  of three precursor
cinnamyl alcohols and cannot serve as the sole carbon source for
growth of the fungus  [13]. The fungus must be able to switch its
ability degrade  these various polymers as the concentration  of
polymer varies with the composition of the wood. This ability for
non mutant strains of P. chrysosporium is controlled by the
                                253

-------
 absence of certain nutrients. Nitrogen deficiency is generally
 used to induce this secondary metabolic cycle of lignin
 utilization [14] .

      The enzyme systems responsible for the initial attack on
 lignin  require unusual abilities due to the complexity and
 resistance of  the  lignin structure.  The 600-1000 k-dalton size
 range for lignin  is far too large to enter the cells of
 microorganisms by  known transport systems. An enzyme system
 permitting the microorganism to  overcome this limitation would
 most likely be extracellular, non-specific (due to the
 heterogenity and  large molecular weight -of the substrate),  and
 resistant to protease  destruction.   It is important to realize
 that the ability  of P.  chrysosporium to degrade lignin by these
 extracellular  enzymes  occurs in  a secondary metabolic cycle.  The
 fungus  uses cellulose  as  its primary growth substrate but when
 large quantities  of lignin  are encountered or certain nutrients
 are  not present the secondary metabolic cycle is entered [15].
     The  extracellular  lignin  degrading  enzymes  serve  to  fragment
lignin  into  pieces  that  can  be assimilated  by  the  fungus.  This
conceptualization of  degradation  activity stresses the  importance
of the  individual enzyme's wide range  activity and function.  The
intracellular  enzyme  components complete the conversion of  the
lignin  fragments into carbon dioxide.

      Lignin cleavage reactions are  catalyzed  by a hemoprotein
ligninase  [16-18].  Hydrogen peroxide  is consumed  in this
reaction  that  degrades  lignin  indicating a  peroxidative
mechanism. The generation of hydrogen  peroxide has  been
attributed to  three different  enzymes: glucose oxidase  [19],
pyranose-2-oxidase  [20], and methanol  oxidase  [21]. The glucose
oxidase enzyme, considered the major contributor to hydrogen
peroxide production, may be  located  in unique periplasmic
microbodies. Stoichiometries of product formation  as well as
hydrogen peroxide and oxygen uptake  are consistent with a radical
pathway [16].  These results established the one-electron
oxidative mechanism as the primary extracellular oxidative
pathway for P.  chrysosporium.

            Degradation Studies of Waste Constituents
     Radiorespirometric studies of the degradation of [U
pentachlorophenol in aequous media indicated .that  the substrate
was rapidly converted to carbon dioxide.  Enzyme studies showed
that pentachlorophenol is converted  to the  1,4-
tetrachlorobenzoquinone by the fungus  [22.].  The quinone was
difficult to quantify due to its propensity to form, charge
transfer complexes with cellular materials.  Further elucidation
of the metabolic pathway is  in progress.  Several aromatic
                               254

-------
hydrocarbons, benzo[a]anthracene, pyrene, anthracene,
benzo[a]pyrene and perylene, (constituents of creosote), were
converted to carbon dioxide by the fungus in liquid
culture[17,18]. This latter finding serves to differentiate the
fungus from bacterial species since few bacteria have the ability
to utilize the higher molecular weight aromatic polycyclics.

           Life Cycle of Phanerochaete    chrysosporium

     To adequately harness the striking abilities of the wood
rotting fungi, it is necessary to understand their life cycle in
order to optimize the treatment process. The life cycle of
Hymenomycetes fungi is characterized by many structures formed
during vegetative, sexual, and asexual reproductive phases  [12].

      The fungal mycelium, a mass of interwoven filamentous
hyphae, is usually submerged in growth medium when cultured in
liquid. The mycellium passes through three distinct stages of
development. The primary mycellium growth phase is not vigorous.
Once secondary mycellium is formed subsequent growth is
frequently different from the primary mycellium. As the mycellium
tissues organize and specialize the tertiary phase is initiated.
Secondary and tertiary mycellia comprise the vegetative segment
of the life cycle. The vegetative phase is the longest and
dominant growth phase. The highest concentration of extracellular
enzymes are secreted during the vegetative phase. Eventually the
tissues of the tertiary mycellium differentiate into fruiting
bodies that are shed depending on environmental conditions.
Asexual reproduction, continued maintenance of current degrading
abilities,  can occur anytime during the vegetative growth phase.
P. chrysosporium produces asexual spores prolifically and at all
stages of the life cycle [23].

     It has been shown that P. Chrysosporium produces at least
ten extracellular hemoproteins and roughly half have ligninase
activity [24]. The heterogeneity among the various extracellular
proteins produced by P. chrysosporium points to possible
functional differences among them important to pollutant
degradation.

           Soil Detoxification Technology Development

     The general success of liquid phase biodegradation studies
with the fungus stimulated speculation that this microorganism
may be an appropriate candidate for the treatment of contaminated
soils. Attempts to innoculate environmental matrices with non-
native microorganisms have met with varying degrees of success
[25]. The elucidation of optimal practices leading to successful
innoculation  of contaminated environmental materials remains to
be discovered  [26]. At the outset of this research, P.
                                255

-------
 chrysosporium was not known to inhabit the soil. Due to this
 general lack of knowledge of the habitat,  a rather cautious
 research effort was engaged to determine the ability of the
 fungus  to inhabit and thrive in the soil.  Early work,  indicated
 that  P. chrysosporium did not grow well in non-sterile soils;
 this  may be  attributable in part due to ineffective competition
 with  the indigenous microflora. These results were anticipated
 since the soil is not the normal habitat of P^ chrysosporium.
 Lately, it has been found that growth within the soil  can be
 accomplished through the use of larger quantities of inoculum
 [27,28].

      Recent research has assessed the effects of selected soil
 types,  temperatures,  pH,  and water potentials on the growth of
 the fungus in sterile and non-sterile soils.   Three well
 characterized soils(topsoil and subsoil) were used in  this work
 (Table  I). The effect of soil type,  temperature, and water
 potential on the  growth  of  P.  chrysosporium in three sterile
 soils was evaluated in a factorial experiment. Soils were
 sterilized by fumigating with methyl bromide  to avoid  confounding
 effects  from native microflora.  Growth of  the fungus was
 evaluated at five  soil temperatures  ranging from 25 to  39°C  and
 four  water potentials ranging from -0.03 to -1.5 MPa.  The extent
 of growth was  determined  by measuring the  amount of ergosterol
 that  could be  extracted  from two  sub-samples  of the soil  from
 each  test  at  the  end  of  a two  week incubation period and  reported
 as ug.  of  ergosterol/ g.  of soil  [29,30].

     Growth  of  the  fungus was  the  greatest  in the  Marsham,
 intermediate  in the  Xurich  and  least  in  the Batavia soil.  The
 same trend was  observed when a  visual assessment  system of  growth
was used.  Soil  water  potential  had a  significant  affect on  the
growth of  the  fungus  in  soil.  As the  .soil water  potential  was
increased  (corresponds to decrease in soil  water  content),  fungal
growth decreased.  Water  potential  is  another  easily controlled
soil factor  [31]. Growth  of  the  fungus between 25  and 35°C was
unaffected but  significantly  decreased at 39°. These results do
not agree with  earlier work  using  the  visual  growth estimation
technique. The difference may  be attributable  to an increase of
sporulation  of  the fungus at  the soil  surface  with  increased
temperature leading to a  biased measurement of  growth.   Soil
temperatures  under field  conditions  can be  controlled by
selecting the normal warm months for  operation  and  by soil
solarization. Biomass  accumulations as well as  growth habit  of P.
chrysosporium were greatly influenced by soil  type.


                       Growth Measurement

     The measurement of growth has presented major  problems for
                               256

-------
this study. Assessment of growth in the early stages of these
studies was done by means of visual estimation on a ranking
basis. Due to the three dimensional growth patterns of the fungus
in a solid substrate such as soil, it was necessary to find a
more reliable means to determine the extent of fungal growth. The
normal means of assessing growth in bacterial systems have no
application to the present study. Based on work  investigating the
infestation of cereal grains by fungi, a mycosteroid, ergosterol
(ergosta-5:6,7:8,22:23-trien-3-ol) has been employed as a
quantitative means to determine growth of the fungus [32]. It has
been shown to correlate well with the visual estimation
technique.

     The application of the fungus to soil treatment is the
remediation of wood treating sites. Target pollutants identified
for treatment at these sites are pentachlorophenol(PCP) and the
major aromatic hydrocarbon contaminants found in creosote
(napthalene, anthracene, and phenanthrene). Creosote has been
extensively characterized and new substrates will be added to the
mixture when deemed necessary  [33]. The degradative ability of
the fungus in the soil has been evaluated  through the measurement
of evolved labelled carbon dioxide. Disappearance of the parent
compound was monitored by GC or HPLC techniques.  Separation of
the soil into solvent extractable, humic  acid,  fulvic acid and
humin fractions permitted material balance evaluations  [34].

      The degradation of 14C [UL]-pentachlorophenol was  studied
over  an eight week period in the  three  soils. Mineralization,
volatile losses, extractable PGP  in  the  soil, and soil  residuals
containing bound PGP and transformed products were  measured to
develop a  tight material balance  [35].  A  very small percentage  of
the total  14C was  accounted  for. by mineralization and
volatilization. Both mineralization  and  volatilization  were
significantly greater  in innoculated  than in  non-innoculated
cultures of  the three  soils. The  extractable  quantities  of  PGP
were  greatly  reduced by  innoculation with P.  chryspsporium. The
greatest rate of PGP removal due  to  fungal  activiy  was  found  with
the Marsham  soil.  Extractable  PGP  after 14 days  was about  2 ppm
of  the  original 50  ppm  spiked  amount  and  that was reduced
roughly 1  ppm after  an  additional  14  days of  treatment.  Decreases
of  PGP  concentration  in  the  control  tests are attributable  to
several potential  causes: abiotic  avenues of  degradation or  loss,
irreversible  binding  to  the  soil  or  degradation due to  the
regrowth  of  native  organisms.

    The closure  of  material  balances  for these  experiments  is
made  possible  by  the  careful  and  detailed analysis  of  the  soil
fractions  for  labelled  carbon  content.  The fate of  ^C PGP in the
soil  was  determined  by analysis of the recoverable carbon label
from:  an  organic  extractable  fraction,  the soil organic matter
                                257

-------
 (humic  and  fulvic  fractions),  and  the  non-extractable  humin
 fraction. Combustion  analysis  was  used to  assay  the  amount of  ^C
 associated  with  humic and  fulvic acid  fractions,  and the  non-
 extractable humin  fraction.

     The percent of total  14C recovery  ranged from 55 to 84%  over
 a 56 day period  in the Marsham soil  tests.  Volatilization losses
 were less than 3%  for the  three soils.  There  was  a significant
 amount  of labelled carbon  activity associated with nonextractable
 fractions,  indicating that there is  possibly  incorporation in  the
 soil material of the  pollutant substrates  during  its metabolism.
 Bollag  has  shown the  possible  polymerization  reactions  between
 pentachlorophenol and  soil chemicals such  as  syringic  acid [36].
 Polymeric forms  of a  series  of pollutants  were constructed by
 Haider  and Martin, who showed  that P.  chrysosporium  would degrade
 these higher molecular weight  materials but at a  slower rate than
 the parent pollutant  substrate [37].

     Future research  in the  soil application will include small
 scale treatment of selected  pollutants at  environmentally
 significant concentrations,  the evaluation  of admendments on
 primary and secondary metabolism,  and  the  delivery of oxygen
within  the soil  to the growing fungus.
                               258

-------
Table I. Physical and Chemical Characteristics  of  Test  Soils
                       Macro Features
Soil Type
Texture
Horizon
Cation Exchange
Capacity (meq/lOOg)
Base Saturation (%)
PH
Organic Matter(%)
Nitrogen(%)
Batavia
silty clay/loam
Bt2
17
29.5
5.4
0.5
0.05
Marsham
sandy loam
A
38
66.4
6.8
12.0
0.46
Xurich
sandy loam
A
14
24.0
,7.1
39.0
0.18
Trace Constituents(ppm)
Calcium
Magnesium
Potassium
Phosphorous
Boron
Mnaganese
Zinc
Sulfur
1950
850
145
75
0.6
12.5
1.9
9.7
4900
1650
90
17
1.3
9.0
12.2
11.3
1675
640
80
17
0.8
77.5
6.4
3.9
                              259

-------
                             References

 1.  U.S.  Department of Agriculture. 1980. The Biological and
 Economic Assessment of Pentachlorophenol, Inorganic Arsenicals,
 Creosote vol 1 and 2. Technical Bulletin 1658-1 and 1658-11.

 2.  Griffin,  D.M. 1985. A Comparison of the Roles of Bacteria and
 fungi.   In Bacteria in Nature Vol 1.  (ends) E.  R. Leadbetter and
 J.S.  Perlmutter, pp.  221-255.  New York:  Plenum.

 3.  Wainwright,  M.  1988.   Metabolic Diversity of Fungi in
 Relation to  Growth and Mineral Cycling in Soil  - A Review.
 Trans.   Br.  Mycol.  Soc.  90 150-170.

 4.  Slater,  J.H.  and D. Lovatt. 1984.  Biodegradation and the
 Significance of  Microbial Communities.  I_n Microbial Degradation
 of  Organic  Compounds  (ed) D.T. Gibson,  pp.  439-4

 5.  Eaton, D.C., Mineralization of polychlorinated biphenyls
 by  Phanerochaete chrysosporium; a ligninolytic  fungus. Enzyme
 Microb.  Technol..  1985,  7,  194-196.

 6.  Bollag, J.-M. 1972. Biochemical Transformation by Soil Fungi.
 Grit. Rev. Microbiol.  2,   35-58.

 7.  Lyr,  H.,  Enzymatische  detoxifikation chlorieter phenole,
 Phytopathol.  Z.,  1963, 47,  73-83.

 8.  Duncan, C.G.,  and  Deverall,  F.J.,  Degradation of wood
 preservatives  by fungi, Appl.  Microbiol..  1964,  12,  57-62.

 9.  H.H.  Unligli,  Depletion  of  pentachlorophenol  by fungi,  Forest
 Prod. J.  18  45-50  (1968).                                  	

 10. Bumpus,  J.A., and  Aust,  S.D.,  Biodegradation of
 Environmental  Pollutants  by  the White Rot Fungus     Phanerochaete
 chrysosporium; Involvement  of  the  Lignin    Degrading  System.
 BioEssays, 1987,  6, 166-170.

 11. Burdsall,  H.H.  and Eslyn, W.E., A new Phanerochaete with  a
 chrysosporium  imperfect state. Mycotaxon. 1974,  1,  123-33;

 12. Deacon, J.W., Introduction  to  Modern Mycology,  Blackwell
 Scientific Publications, Oxford, 1984,  pp.1-24.

 13. Kirk, T.K. and  Shimada, M., In Biosynthesis  and
Biodegradation of Wood Components, ed.  T. Higuchi, Academic
Press, N.Y., 1985,  pp  579-605.
                               260

-------
14. Kirk, T.K.
of Organic Compounds,
1984, pp. 399-438.
                Degradation of
                      ed. D.T.
                  Lignih. In Microbial Degradation
                  Gibson, Marcel Dekker, New York,
15. Kirk, T.K. and R,L. Farrell.  1987.  Enzymatic  "Combustion":
The Microbial Degradation of Lignin. Ann.  Rev.  Microbiol.  41,
465-505.

16. Hammel, K.E., Tien, M., Kalyanaraman,  B.,  and  Kirk,  T.K.,
Mechanism of oxidative C-C cleavage  of  a  lignin model  dimer
by Phanerochaete chrysosporium  ligninase:  Stoichiometry  and
involvement of free radicals. J.  Biol.  Chem.,  1985,  260,
8348-53.

17. Hammel, K.E., Kalyanaraman,  B.,  and Kirk,  T.K.,  Oxidation  of
polycyclic aromatic hydrocarbons  and dibenzo[p]dioxins by
Phanerochaete chrysosporium ligninase.  J.  Biol. Chem., 1986,
261, 16948-52.

18. Haemmerli, S.D, Liesola, M.S.A., Sanglard,  D.,  and Feichter,
A., Oxidation of benzo(a)pyrene  by extracellular  ligninases,
from Phanerochaete chrysosporium, J. Biol.  Chem.,  1986,  261,
6900.
19. R.L. Kelley, and C.A. Reddy
Phanerochaete chrysosporium.  In
BIOMASS Part
S.T. Kellogg
             B, Lignin,
              PP
           Pectin,
     307-315.
 1988. Glucose Oxidase of
Methods in Enzymology, Vol 161,
and Chitin. (eds) W.A- Wood, and
20. Vole. J. and K.-E. Eriksson.  1988.  Pyranose-2-Oxidase from
Phanerochaete chrysosporium  In  Methods  in  Enzymology,  Vol 161,
                                 and  Chitin.  (eds)  W.A.  Wood,  and
BIOMASS Part
S.T. Kellogg.

21. Eriksson,
Phanerochaete
B, Lignin, Pectin,
 pp. 316-321.
              K.-E.,  and  A.  Nishida.  1988.  Methanol Oxidase of
              chrysosporium  In  Methods  in  Enzymology,  Vol 161,
BIOMASS Part
S.T. Kellogg
             B, Lignin, Pectin, and Chitin.  (eds) W.A. Wood,
              pp. 322-326.
                                                              and
22. Hammel, K.E.  and  P.J.  Tardone.  1988.  The  Oxidative 4-
Dechlorination of  Polychlorinated  Phenols is  Catalyzed by The
Extracellular Fungal  Lignin  Peroxidases.  Biochemistry 27, 6563-
6568.

23. Gold, M.H, and  Cheng,  T.M.,  Conditions for  fruit body
formation in the  white-rot basidiomycete  Phanerochaete	
chrysosporium. Arch.  Microbiol.,  1979,  121,  37-41.
                                261

-------
 24. Kirk, T.K., Groan, S., Tien, M., Murtaugh, K.E.,  and  Farrell,
 R.L., Production of multiple ligninases by Phanerochaete
 chrysosporium; Effect of selected growth conditions and use of
 mutant strain. Enz. Microb. Tech.. 1986, 8, 27-32.

 25. Zaidi, B.R., Stucki, G., and Alexander, M., Low Chemical
 Concentration and pH as Factors Limiting the Success  of
 Innoculation to Enhance Biodegradation, Environ. Toxicol.
 Chem., 1988, 7, 143.                                   	
 26. Goldstein, R.M., L.M. Mallory, and M. Alexander. 1985.
 Reasons for Possible Failure of Innoculation to Enhance
 Biodegradation. Appl. Environ. Microbiol. 50, 977-983.

 27.  Lamar, R.T., Larsen M.J., Kirk, T.K., and Glaser, J.A.,
 Growth of the white-rot fungus Phanerochaete chrysosporium in
 soil.  In Land Disposal, Remedial Action, Incineration and
 Treatment of Hazardous Waste Proceedings of the 13th Annual
 Hazardous Waste Symposium, EPA/600/9-87/015, \T. S~T
 Environmental Protection Agency, Cincinnati, Ohio, 1987, pp. 419-
 429.

 28. Lamar, R.T.,  Larsen, M.J., Kirk, T.K., and Glaser, J.A.,
 Effect of Biotic  and Abiotic Soil Factors on Growth and
 Degradative Activity of the White-Rot Fungus Phanerochaete
 chrysosporium Burds. In Chemical and Biochemical
 Detoxification of Hazardous Waste,(ed.) J.A. Glaser, Lewis
 Publishers, Ann Arbor, MI., 1988. (In Press)'.

 29. Matcham, S.E.,  B.R. Jordan, and D.A. Wood.  1985. Estimation
 of Fungal Biomass in Solid Substrate by Three Independent
 Methods.  Appl. Microbiol.  Biotechnol. 21, 108-112.

•30. Seitz, L.M.,  D.B. Sauer, R. Burroughs, H.E. Mohr, and J.D.
 Hubbard.  1979. Ergosterol  as a Measure of Fungal Growth.
 Phytopathology 69,  1202-1203.

 31.  Sommers,  L.E.,  Gilmour, C.E., Wildung,  R.E.,  and Beck, S.M.,
 The effect of  water  potential  on decomposition processes in
 soils.  In  Water  potential relations in soil microbiology,
 SSA Special Publication Number 9,Soil Sci.Soc.of Amer.,
 Madison,  WI.,  1981,  pp.97-117.

 32. Weete, J.D.,  and D.J.  Weber.  1980. Lipid Biochemistry of
 Fungi  and Other Organisms.  Plenum Press, New York.

 33. Nestler,  F.H.M., U.S.  Dept.  Agri. For.  Ser.  Res. Pap.  FPL
 195. U.S.  Department of Agriculture,Forestry Service,Forest
 Products  Laboratory, Madison,  WI.
                               262

-------
34. Stevenson, F.J. 1982. Humus  Chemistry,  Chapt  2.  pp
Wiley, New York.
26-54.
35. Mayaudon, J. 1971. Use of Radiorespirometry  in Soil
Microbiology and Biochemistry,  In  Soil  Biochemistry Vol 2  (eds)
A.D. Mclaren and J. Skujins. pp. 202-256.

36. Bollag, J.-M. and S.-Y. Liu, Coploymerization of halogenated
phenols and syringic acid, Pest. Biochem.  and  Physiol.  23,
261 (1985).

37. Haider, K.M. and J.P. Martin.  1988.  Mineralization  of  14C-
Labelled Humic Acids and Humic-Acid  Bound  1^C-Xenobiotics by
Phanerochaete chrysosporium. Soil  Biol.  Biochem  20, 425-429.
                               263

-------
                BIOLOGICAL TREATMENT OF PETROCHEMICAL SLUDGES

         Stephen D. Field, Associate Professor of Civil Engineering
     Andrew K. Wojtanowicz, Assistant Professor of Petroleum Engineering
                         Louisiana State University
                        Baton Rouge, Louisiana 70803
                                  ABSTRACT

     The research results of using a microbial population acclimated to
petrochemical waste sludges to degrade the complex hazardous organic sub-
strates found in these sludges are presented.  The results obtained include
the determination of microbial population stability, rates of growth and
substrate utilization for selected single compound substrates1and two API
separator sludges.  Specific growth rates as high as 0.38 hr~  have been
obtained for an API separator sludge treated in a batch reactor loaded with
five percent oil (v/v).  Cultural stability appears to have been established
after three months of continuous loading in complete-mixed stirred tank
reactors for two different API separator sludges; microbial composition
differs for each of the API sludges.  These results are being included in a
growth rate model to relate specific growth rate to substrate concentration
and desorption rates of organic compounds from the inorganic solid consti-
tuents as a function of microbial culture activity. The results of this
research are the initial step towards a detailed description and evaluation
of biological treatment of complex organic hazardous waste sludges.


                                INTRODUCTION

     Hazardous waste sludges from the petroleum industry (e.g., API  separator
sludges, DAF floats, storage tank still bottoms) and contaminated soils from
hazardous waste spills and at 'Superfund' sites lack technical and cost
effective treatment methods.  Petroleum industry wastes primarily rely on
landfarming for degradation of the petroleum compounds and sludges.   This
process suffers from low loading rates (less than 10% oil in the cell), long
treatment periods (1-3 years), high odor emission, and potential for ground-
water pollution.  In addition, landfarming is considered a land disposal
method by the USEPA and falls under the land ban requirement.  Therefore, an
efficient alternate technology needs to be developed.  In a similar  manner,
effective soil cleansing of petrochemical components at Superfund and other
                                     264

-------
spill sites is lacking; incineration is not an optimal technology for soils
and fixation/stabilization still requires removal and disposal in a secure
landfill.

     The expected promulgation of a stricter test for defining hazardous
wastes by EPA, and the existing "land disposal ban" will bring substantially
greater amounts of wastes under the hazardous waste regulatory authority.
Increased public opposition to off-site transport and incineration of
hazardous wastes will require the development of alternative, economical
treatment methodologies such as bioengineered systems, in order to meet the
goals of a clean environment and economic growth.  The support of the
development of bioengineered systems to treat wastes on-site offers potential
economic benefits through cost savings to the petroleum and petrochemical
industry, as well as a methodology to treat hazardous wastes.

     The isolation and development of biological cultures in the laboratory
to degrade  toxic and bioinhibitory materials has been previously estab-
lished.  However, there is a dearth of concurrent development of engineering
data and information necessary to translate the laboratory performance of the
microbial populations to successful full-scale applications.

     A microbial culture has been developed which has the ability to survive
in and consume natural crude-oil long-chain hydrocarbons and processed hydro-
carbons, including diesel oils.  This culture, which relies on the cometabo-
lism of several species of bacteria, fungi and protozoa, has has degraded 80
to 90% of the oil present in slurries containing as much as 24% oil and 24%
solids.  This success in degrading high strength petroleum and petrochemical
wastes has been tested and found to be reproducible for selected USEPA
classified hazardous sludges (e.g., API separator sludges, silt pond sludges)
generated in the petroleum processing industry.

     Previous research demonstrated the effectiveness of the microbial
populations for the biodegradation of oilbased drilling muds and oilfield
production wastes (1,2,3).  Specifics of these studies included investigation
of aerobic biodegradation of drill cuttings containing 58% oil v/v and 48%
solids w/w and an oilfield production sludge containing 13% oil v/v and 30.4%
solids w/w.  Results indicated that the hydrocarbons were tightly bound to
the soil and clay structure.  The process of aerobic biodegradation, however,
quickly destroyed the bonding mechanisms and released the oil for rapid
biodegradation.

     The microbial culture under investigation has demonstrated excellent
potential for degrading Principal Organic Hazardous Constituents (POHC's) in
waste oil and API separator sludges. The microbes have been found to
effectively consume long-chain hydrocarbons (up to C-36), aromatics and
chlorinated hydrocarbons.  Two API separator wastes containing 26-28% oil
w/w, 30% solids w/w and significant amounts of up to 19 different POHC's,
including:  benzene, phenol, chloroform, 1,2-dichloroethane, ethylbenzene,
chlorobenzene and toluene, have also been significantly biodegraded.  In
similar tests of a 1000 mg/L chlorinated pesticide substrate and a 14000 ug/L
benzene substrate, the population exhibited microbial viability and proli-
feration under the toxic environments.
                                     265

-------
      The overall objective of this research program is  to  examine  the feasi-
 bility of using a heterogeneous  microbial population acclimated  to petroleum
 and petrochemical waste sludges  to degrade the  complex  hazardous organic
 substrates found in these wastes.   This  paper presents  the preliminary
 results on the determination of:   microbial population  stability for two API
 separator sludges, and specific  growth rates and  substrate utilization
 measured for an API separator sludge.


                             MATERIALS AND METHODS

      The measurement of specific growth  rates and substrate utilization rates
 were performed in batch reactors using sealed 250 ml erlenmeyer  flasks
 incubated in a constant temperature shaker bath.  Procedures are outlined in
 previous publications (4,5).   From the specific growth  rates and substrate
 utilization rates, Monod growth kinetic  coefficients (maxium specific growth
 rate and half-velocity constant) and yield coefficients were able  to be
 calculated (6).

      Two API separator hazardous wastes  were used for continuous loading
 experiments.   These materials were designated as BRE API Separator and PLS
 API Separator sludges,  and were obtained from different sources.   Samples of
 these sludges contained oil-rich and water-rich layers  above the bottom
 sludge.   Removal of the surface oil and  water left oil  saturated settled     s
 sludge for use  in the continuous reactor program.  This pretreatment was
 instituted since a commercial processing route in a  continuous reactor
 processing system would typically  use a  preseparator to remove the free oil
 and surplus water.   The two  continuous culture aerobic  systems have been
 operated for  over a six month period and periodically monitored for substrate
 removal  and microbial population assay.  Microbial population assays included
 microorganism identification  and enumeration, and metabolic capability by
 testing  for percentage  of  organisms  in the  total population that can metabo-
 lize various  substrates.   Details  of the  reactors' specifications operation
 and testing protocol have  been published  elsewhere (4).  A summary of key
 operational data is  presented in Table 1.
                           RESULTS AND DISCUSSION

     The results of microbial growth and respiration measurements for the BRE
separator oil are presented in Figure 1.  The microbial growth and respira-
tion rates increase with increasing substrate concentration; the maximum
growth rate measured was 0.38 hr~  for a substrate concentration of five
percent oil (v/v).  This growth rate is well above the required growth for
effective treatment to reduce the oil content of typical waste API separator
sludges^ For comparison, Gaudy et al. (7) was able to obtain growth rates of
0.35 hr   on sucrose using municpal sewage organisms.  The upper range of
growth in activated sludge systems treating municpal sewage is approximately
0.42 hr   (8).  An average true yield coefficient (gm TOG cell mass pro-
duced/gm TOG substrate consumed during log growth) of 0.28 was determined
(5).  This value is similar to those reported for heterogeneous populations
grown on mixed priority pollutant substrates (9).
                                     266

-------
        TABLE 1.   OPERATING DATA FOR TWO CONTINUOUS  CULTURE REACTORS
                      USING BRE AND PLS HAZARDOUS  WASTE  SLUDGES
             Parameter
BRE Sludge
PLS Sludge
Settled Sludge Feed, g/d
Feed Sludge Density, gm/ml
Feed Sludge pH
Oil Content, v/v %
Reactor Volume, ml
Mean Cell Residence Time, d
Redox Potential, mv
Reactor pH range
10
1.87
6.3
3.7
1000
5
+295
6.5 - 7.2
10
1.11
7.1
6.7
1000
5
+355
5.7 - 7.2
     The growth curves for the API separator sludge substrates demonstrate a
typical lag phase prior to log growth.  During this period,  the free phase
hydrocarbons floating on top of the water phase can be seen to be emulsified
and made readily available to the microbial population.  This period of time
seems to be somewhat variable, typically ranging from 10 to 20 hours.  This
lag phase can be shortened by using stock cultures which are presently in
log growth to initiate the experiment.

     The growth rate results presented must be viewed as encouraging but
preliminary; they are being  repeated for verification.  In addition, other
substrates are being tested to examine the effectiveness of biological
treatment for these types of wastes.  In particular, the attendant principal
organic hazardous constituents in the separator sludges currently are being
monitored for degradation in the API separator oil experiments.

     The microbial cultures grown in continuous culture loading experiments
for the two different API separator sludges showed adaptation to the specific
sludge source.  The BRE API separator sludge generally was the more difficult
to achieve growth and substrate removal, even though the kinetic experiments
demonstrated excellent growth rates.  The initial microbial culture screening
techniques revealed no actinomycetes growth, two fungal populations, four
varieties of free swimming protozoa, and bacteria tentatively identified as
Pseudomonas fluorescens, and strains of either Azotobacter, Azomonas, or
Beijerinckia.  The PLS separator sludge culture differed by demonstrating the
presence of actinomycetes, and only one fungal population.  The ciliates and
bacteria were similar to the BRE separator culture.
                                     267

-------
                  cn
                  *-•
                  u
6000 -i

5000 -

4000-

3000 -

2000-

1000 -
                                                  a 0.195
                                                  « 0.595
                                                  n 1.055
                                                  * 5.095
                           ' • I " " I1' " I	|n i u ii i n 1111111111
                        20 25  30  35  40  45  50 55  60
                                 Time (hrs)
                      10 n
                   31
                   *->
                   «r-
                   W
                   c
                   CO
                   o

                   15
                   o
                   *->
                   a
                   O
   1 -
                      .01
                        Q 0.1 S5 (1=0.04
                        + 0.595 (1=0.11
                        » 1.0% (1=0.12
                        <• 5.0S5 |l=0.3S
                        20  25  30  35 40  45  50  55  60

                                  Time (hrs)


                 Figure 1.  Respiration and growth measurement
                            for BRE-sludge oil.
     Examination of the common fungal  population to both separator  sludge
sources has  been tentatively identified as Phialophora j eanselmei,  a dark
green  fungus common to polluted waters.   The other fungal growth present in
the BRE separator sludge culture has yet to be identified.

     Observations after one month of loading reveal population changes  in
both separator sludge systems, but no  reduced efficiency of microbial
consumption.   The BRE separator fed culture shows slower growths on solid
media, no  growth on oil droplets in solution, one dominate ciliate  type and
different  fluorescent pseuomonads than initially identified, and two to
three  fungi  types.   The BRE separator  culture shows no actinomycetes present.
                                      268

-------
 The  PLS  culture  grows  faster on  solid media, demonstrates the growth on oil
 droplet  surfaces, has  three different kinds of ciliates all of which are
 different  from the  BRE population, different fluorescent pseudomonads, and
 the  same single  fungal population.  These populations are being freeze dried
 and  stored for further investigation at a later date, when the evolution of
 the  culture to a steady  state population can be completely investigated from
 an ecological  standpoint.

      From  these  observations, the cultures are exhibiting a dynamic state of
 population diversity.  The overall reactor performance in fterms of waste
 treatment  efficiency has not been significantly afffected, however.  The
 implications are that  the continuous loading reactors need to be continued
 until a  stable population develops in terms of both numbers and species
 diversity,  and reactor preformance monitored to detect any effects on process
 performance associated with these changes.  It also is evident that waste
 type will  influence the  species  composition.  This logically would be antici-
 pated.   The extent  of  such changes need to be considered and evaluated,
 especially for those wastes exhibiting variable composition.

      The results of the  endeavor to explore some of the possible substrates
 available  to the microbial populations using solid media are presented in
 Table 2.   These  results  reflect  only a portion of the microbial populations'
 capabilities since  not all organisms identified in the cultures would be
 expected to survive on pour plate culture techniques.  The results indicate
 that metabolically, both cultures possess the ability to metabolize similar
 substrates.  These  substrates include the standard nutrient agar, tryptic
'soy  agar and dextrose  for comparison.  As is evident in Table 2, a wide
 variety  of  substrates  are available as a sole carbon source including the
 cyclic compounds benzene, toluene and xylenes, the long chain hydrocarbon
 dotriacontane  (C-32),  and methylene chloride.  No growth was attainable for
 pentachlorophenol or para-cresol using solid media at concentrations of 800
 and  1000 mg/L, respectively.  Growth was attained at 50 mg/L pentachloro-
 phenol on  solid  media.  It is evident that the higher concentrations are
 toxic to the populations on solid media.  Separate liquid cultures fed the
 pentachlorophenol and  para-cresol at 800 mg/L and 1000 mg/L, respectively,
 were capable of  supporting growth.  Monitoring of the pentachlorophenol
 liquid culture chloride concentration with time showed 60 percent of the
 pentachlorophenol was  dechlorinated within five days, verifying substrate
 consumption.   Chloride production was used to verify consumption of methylene
 chloride in liquid  culture also.

      Reactor performance in terms of polycyclic aromatic hydrocarbon (PAH)
 compounds destroyed for the BRE  sludge reactor are presented in Table 3.
 Average  PAH reduction was approximately 90 percent with the reactor operating
 on a 5 day  hydraulic and mean cell residence time.  The more readily degrad-
 able PAH's  (napthalene, fluorene, phenanthrene and anthracene)  were 85  to
 99 percent  removed  in  the reactor.  The more difficult substrates (pyrene,
 benzo (a) anthracene, chrysene)  were removed with 75 to 86 percent effective-
 ness.  These removal efficiencies were achieved with light loading rates
 (approximately 130 mg/kg PAH in  feed sludge)  but are significant in that  this
was  achieved with an abundance of more readily available substrates present.
                                     269

-------
TABLE 2.  MICROORGANISM GROWTH ON SELECTED ORGANIC CARBON SOURCES
                  FOR BRE AND PLS CONTINUINS CULTURES
Carb'on
Source
                             Concentration
                                 mg/1
              Agar
              Media*
Liquid
Media*
 Nutrient Agar

 Tryptic Soy Agar

 Dextrose

 2-2-4 Trimethyl pentane

 Benzene

 Toluene

 Methylene Chloride

 Mixed Xylenes

 Pentachlorophenol

 Pentachlorophenol

 P-Cresol

 Succinic Acid

 Dotriacontane

 BRE Separator Oil

 PLS Separator Oil
 500

 690

 700

 870

 660

 860

 800

  50

1000

1250

1250

 525

 525
  (+) Growth attained, (-) no growth
                               270

-------
              TABLE 3.  PAH REMOVAL EFFICIENCIES IN CONTINUOUS
                               LOADED REACTOR
                  PAH
Percent Removed
                  Napthalene

                  Fluorene

                  Phenanthrene

                  Anthracene

                  Fluoranthrene

                  Pyrene

                  Benzo(a)anthracene

                  Chrysene

                  Benzo(a)fluoranthene

                  Benzo(a)pyrene

                  Benzo(g,h,i)perylene
      99

      96

      86

      85

      71

      75

      86

      82

      85

      87

      97
Total oil content of the feed was 100 fold higher (approximately 16,000 mg/
kg).  Increased loading rates and biokinetic parameter measurements are being
made to examine for maximum process efficiencies for PAH destruction.


                                 CONCLUSIONS

     The results obtained thus far on the biological treatability of petro-
leum and petrochemical wastes are very encouraging.  The growth rates
obtained on API separator sludge oil are well within the requirements for
efficient process design for reducing oil content of the sludges.  Substrate
utilization at low concentrations indicates that process design can achieve
quality effluent and waste solid streams.  Further work needs to be expended
on the fate of the principal hazardous constituents present in the sludges.
Additional work on complex substrate availability, substrate destruction,
kinetics of reactions for process design and scale up, and environmental
growth limitations need to be addressed.  Microbial stability appears not to
be a concern for a specific waste stream.  Continuous loading experiments
demonstrate an ability to maintain an effective population over time.
                                     271

-------
Population composition has been observed to change with different waste
sources; this factor needs to be addressed further to determine the potential
implications for those waste streams that have temporally variable composi-
tion.
                              ACKNOWLEDGEMENTS

     This work was partially supported by the LSU Hazardous Waste Research
Center.  Lee Forbes and Robert Marks, graduate assistants in civil engi-
neering, are gratefully acknowledged for their data collection efforts on
which most of this paper is based.


                                 REFERENCES

1.  Marks, R. E., Field, S. D., Wojtanowicz, A. K., "Biodegradation of
    Oilfield Production Pit Sludges," Proceedings of the 42nd Industrial
    Waste Conference, Lewis Pub. Inc., Chelsea, MI, (1988).

2.  Marks, R. E., Field, S. D., Wojtanowicz, A. K. "Oil Reduction in Spent
    Drilling Muds by Biotreatment," paper presented at the Third National
    Conference on Drilling Muds, Norman, OK, May, (1987).

3.  Marks, R. E., Field, S. D., Wojtanowicz, A. K., "Biodegradation of
    Oil-Based Drilling Muds and Production Pit Sludges," Journal Energy
    Resources Technology, 110, 183-188, (1988).

4.  Field, S. D., Forbes, L., Marks, R. E., Wojtanowicz, A. K. "Biological
    Treatment Oilfield and Petrochemical Hazarodus Wastes," Proceedings of
    the Engineering Foundation Conference on Biotechnology Applications in
    Hazardous Waste Treatment, Longboat Key, FL, Oct. 1988 (in press).

5.  Forbes, L.  "Biodegradation of Petroleum and Petrochemical Wastes," MS
    Thesis, Dept. Civil Engineering, Louisiana State University, Baton
    Rouge, LA, 88 p., Dec. 1988.

6.  Braha, A., Hafner, F., "Use of Lab Batch Reactors to Model Kinetics,"
    Water Research, 73-81, (1987).

7.  Gaudy, A. F. Jr., Yang, P,. Y., Bustamante, R., Gaudy, E. T.,
    "Exponential Growth in Systems  Limited by Substrate Concentration,"
    Biotechnology Bioengineering, 15, 589-596, (1973).

8.  Medcalf and Eddy, Inc., "Wastewater Engineering: Treatment, Disposal,
    Reuse," 2nd Edition, McGraw Hill, NY, (1979).

9.  Kincannon, D. F., Stover, E. L., "Determination of Activated Sludge
    Biokinetic Constants for Chemical and Plastic Industrial Wastewater,"
    EPA-600/2-83-073, USEPA, Ada, OK, (1984).
                                     272

-------
    THE DETERMINATION OF BIODEGRADABILITY AND BIODEGRADATION KINETICS OF
    ORGANIC POLLUTANT COMPOUNDS WITH THE USE OF  ELECTROLYTIC RESPIROMETRY

                               Henry H. Tabak
                    U.S. Environmental Protection Agency
                     Office of Research and  Development
                    Risk Reduction Engineering  Laboratory
                           Cincinnati,  Ohio   45268

                                     and

                       Sanjay Desai and Rakesh Govind
               Department of Chemical and Nuclear Engineering
                          University of Cincinnati
                           Cincinnati,  Ohio   45221
                                  ABSTRACT
     Electrolytic respirometry involving natural sewage, sludge and soil
microbiota is being applied to the fate studies of priority pollutant and
RCRA toxic organics to generate data on their biodegradability and on
biodegradation/inhibition kinetics.  This paper discusses the experimental
design and procedural steps for the respirometry biodegradation and toxicity
testing approach for individual organics or specific industrial wastes.  The
discussion also includes a review of the electrolysis BOD measuring system
inherent in electrolytic respirometry and the factors affecting
respirometric determination and measurement of respiration rate.

     A developed multi-level protocol is presented for determination of the
biodegradability, microbial acclimation to toxic substrates and first order
kinetic parameters of biodegradation (n and n') and for estimation of the
Monod kinetic parameter (JL, Ks and Y ) of toxic organic compounds, in order
to correlate the extent and rate of biodegradation of these organics with a
predictive model based on chemical properties and structure of these
compounds.

     Respirometric biodegradability/inhibition and biodegradation kinetic
data are provided for representative RCRA alky! benzenes, phenolic
compounds, phthalate esters, and ketones.
                                     273

-------
                                 INTRODUCTION
      Electrolytic  respirometry  is  attaining prominence in biodegradation
studies  and  is  becoming  one  of  the more  suitable experimental methods for
measuring the biodegradability  and the kinetics of biodegradation of toxic
organic  compounds  by  the sewage, sludge  and soil microbiota and for
determining  substrate inhibitory effects to microorganisms in wastewater
treatment systems.

      Biodegradation of toxic and hazardous organic compounds holds a great
promise  as an important  fate mechanism in wastewater treatment and in soil
detoxification.   Information about the extent  and rate of biodegradation is
a prerequisite  for informed  decision making on the applicability of the
biodegradation  approach.  Unfortunately, relatively little quantitative data
are  available from which engineering judgement can be made, because of the
large effort required to assess biodegradation kinetics.

      Current research in our laboratories has  shown that it is possible to
assess biodegradation kinetic parameters from  oxygen uptake data, obtained
through  the  use of electrolytic respirometry.  This methodology greatly
reduces  the  work and  expense involved  in evaluation of biodegradation
kinetics.  The  ongoing biodegradation  studies  are concerned with the
generation of biokinetic database  so that it can be ultimately used to
establish a  possible  correlation between molecular substrate configuration
(chemical/physical  characteristics) and  biomass activity (kinetic
parameters)  as  an  index  of biodegradation.  The experimental respirometry
testing  is also providing data  on  the  concentration levels of toxic organics
inhibitory to microbial  activity.

      Initially,  the  inter-laboratory,  ring test, Organization of Economic
'Cooperation  and Development (OECD) studies at  the EPA laboratory,
Cincinnati,  Ohio,  were undertaken  to develop confirmatory respirometric
biodegradability testing procedure.  Respirometric biodegradability and
biokinetic data are  provided for the selected  non-inhibitory and non-
adsorbing compounds,  tetrahydrofuran,  hexamine, pentaerythritol, 1-napthol,
sodium benzene  sulphinate, thioglycolic  acid and the biodegradable reference
compound, aniline.

      Subsequently, similar electrolytic  respirometry studies were  initiated
to  determine biodegradation kinetic parameters for selected representative
toxic compounds of varied classes  of organics  included  in the Priority
Pollutant, RCRA list, and to demonstrate presence of any inhibitory effects
of  these organics of specified  concentration levels on  the  sludge  biomass
and on the metabolism of biogenic  compounds.

      The objectives of the present study were  to  utilize the electrolytic
respirometry oxygen uptake data to:   (1) determine the  biodegradability of
selected RCRA alkyl  benzenes, phenols,  phthalates, and  ketones,  (2) generate
 information  on  their acclimation  time  (lag)  values  (t0) and the initiation
and termination time values for the declining  growth phase  (t^ and t2); (3)
determine their first order kinetic parameters of  biodegradation  (specific
growth rate  constants for the exponential  growth  phase  (/*)  and  for the
declining growth phase  (#'); (4)  estimate the  Monod  kinetic  parameters  (Mm,
                                      274

-------
Ks and Yg) of  these  compounds  without  initial  growth  or growth yield
assumptions;  (5) demonstrate presence of any inhibitory effects of these
compounds on the metabolism of biodegradable reference compound, aniline;
and (6) to correlate the extent and rate of biodegradation of these
compounds with a predictive model based on chemical properties and structure
of these compounds.

     The purpose of this study was to obtain information on biological
treatability of the benzene, phenol, phthalate and ketone organics in
wastewater treatment systems which will support development of an EPA
technical guidance document on the discharge of the above organics to POTWs.
Respirometric biodegradability, biokinetic and inhibition data are provided
for the selected RCRA benzene, phenol, phthalate and ketone compounds.

                                 BACKGROUND
MEASUREMENT OF OXYGEN CONSUMPTION

     Measurement of oxygen consumption is one of the oldest means of
assessing biodegradability.  Time consuming manual measurement of oxygen
uptake (dilution BOD measurements) was replaced gradually by a more direct
and continuous respirometric method for measurement of oxygen consumption in
biochemical reactions, for use in routine examination of sewage and in
control of sewage treatment process.

     A rather comprehensive review of the use of respirometers for the study
of sewage and industrial waste and their application to water pollution
problems was published by Jenkins in 1960 (21).  Montgomery's (29) review on
respirometric methods summarized the design and application of respirometers
for determination of BOD.

     The application of respirometry was gradually directed to research
studies to assess the toxicity and bipdegradation of specific wastes or
compounds, to evaluate factors affecting biological growth and to provide an
insight into nitrification reaction.  Of the commercial respirometers which
have been developed for respirometric studies, the electrolytic
respirometers were shown to be most applicable for measurement and
quantitation of biodegradation activity because they automatically produce
oxygen as needed, thereby eliminating some of the limitations of other
techniques and allowing output data to be collected automatically for direct
recording and processing (3, 7, 26, 30, 53, 56, 57).  A recent detailed
review of respirometric techniques and their application to assess
biodegradability and toxicity of organic pollutants was published by King
and Dutka (23).

RESPIROMETRIC BIODEGRADABILITY TESTING

     Most uses of electrolytic respirometry in biodegradability testing have
been for screening purposes to measure the extent of biodegradation as a
percentage of the theoretical oxygen demand exerted in some time period (3,
6, 7, 17, 27, 50, 52, 54, 55).  A more recent study by Painter and King (38)
concluded that a procedure based on electrolytic respirometry was reliable
                                      275

-------
for assessing biodegradability, and could serve as an adequate Level I
screening test for biodegradability (33).

     A considerable amount of studies using electrolytic respirometry to
determine the biodegradability of wastes and specific organics is available
in published literature and significant data on biodegradation of pollutants
based on oxygen uptake have been generated (1, 15, 16, 20, 24-26, 28-32, 37,
40, 44, 48, 49, 51, 58).

     There are many techniques that have been used to evaluate
biodegradation kinetics and these were reviewed in detail by Howard et al.
(18, 19) and Grady (11)-  These techniques utilize continuous, fed-batch and
batch type reactors for providing data from which kinetic parameters can be
evaluated.  The use of batch systems in biotechnology and biological
wastewater treatment represents a less labor intensive, less expensive and
much faster way to model biokinetics.

     The kinetic parameters obtained by the above techniques should be
intrinsic, that is, dependent only on the nature of the compound and the
degrading microbial community and not on reactor system used for data
collection.  If this condition is satisfied then the parameters obtained can
be used for any reactor configuration and can be used.in mathematical models
to estimate the fate of toxic organics.

     Batch techniques are successful in obtaining intrinsic kinetic
parameters by applying non-linear curve fitting techniques to single batch
substrate removal curves, provided initial conditions are selected with
proper care (Simkins and Alexander (42, 43); Robinson and Tiedje (41); Cech
et al. (5); Braha and Hafner (2)].  Batch systems can be used with either
acclimated or unacclimated biomass for providing kinetic data and require
that samples be taken at discrete time intervals during the course of
biodegradation [Tabak et al. (45); Larson and Perry (25); Paris and Rogers
(39)].

     Measurement of oxygen consumption through electrolytic respirometry is
a batch type technique which has been shown to be very promising for
automating data collection associated with biodegradation and intrinsic
kinetic parameters.

USE OF ELECTROLYTIC RESPIROMETRY TO GENERATE BIOKINETIC DATA

     Whereas the electrolytic respirometry is becoming the most commonly
employed method for automatically collected data associated with
biodegradation, with the exception of the studies performed by Dojlido  (6),
Larson and Perry (25), Tabak et al. (46), Oshima et al.  (35, 36), Gaudy et
al. (9, 10), Grady et al. (12-14) there have been very few investigations
into the use of respirometry to generate biodegradation kinetic data.
Larson and Perry (25) showed that the electrolytic respirometer can be used
to measure biodegradation of complex organics in natural waters when
specific analytical methods or radio!abeled materials are unavailable.
However, they used empirical kinetic expressions which were system specific.

     Dojlido (6) divided the oxygen uptake curve into seven different phases
and then proposed an empirical model for each phase and evaluated the
                                     276

-------
faiodegradability and toxicity of a test compound by measuring empirical rate
constants and time interval associated with each phase.  Various phases were
distinguished by identifying inflection points in the curve through the
plots of the logarithm of the slope versus time.

     Tabak et al. (46) and Oshima et al. (36) sought to capitalize on
Dojlido's method of identifying inflection points in order to quantify more
fundamental rate coefficients.  In order to identify more fundamental (and
therefore intrinsic) kinetic coefficients they used the generalized concept
of oxygen uptake by Gaudy and Gaudy (8).  Substrate removal was divided into
exponential and declining phases separated by an inflection point, and the
endogenous phase.  They coupled substrate removal and cell growth to oxygen
consumption by imposition of an electron balance and consequently were able
to evaluate um from the oxygen consumption data up to the inflection point.
They proposed the use of the lag time as an indicator of how difficult it is
to achieve acclimation to a test compound and their respirometric studies
were carried out with an unacclimated biomass.

     The justification for using respirometry to obtain intrinsic kinetic
lies in the concept of oxygen consumption as an energy balance [Busch et al.
(4); Gaudy and Gaudy (8)].  This concept states that all of the electrons
available in a substrate undergoing biodegradation must either be
transferred to the terminal acceptor or be incorporated into new biomass or
soluble microbial products.  If the concentrations of the substrate, the
products and the biomass are all expressed in units of chemical demand
(COD), then the oxygen uptake can be calculated in a batch reactor from an
equation relating oxygen uptake to substrate, biomass and soluble products
[Grady et al. (12)].  Furthermore, since biomass growth and product
formation are proportional to substrate removal, this suggests that an
oxygen uptake curve can provide the same information as either a substrate
removal curve or a biomass growth curve.  This latter concept has recently
been used by Gaudy et al.  (9, 10) to calculate biodegradation kinetics.
Specific growth rates obtained from growth studies as slopes of plots of
Ln(X) versus time at different substrate concentrations, compared favorably
with those obtained from exponential phase of respirometric oxygen uptake
curves as slopes of plots of Ln (doydt) versus time.

     Studies of Grady et al. (12, 13) have demonstrated that it is possible
to determine intrinsic kinetics of single organic compounds by using only
measurements of oxygen consumption in respirometric batch reactors.  With
the use of computer simulation techniques and non-linear curve fitting
methods, intrinsic kinetic parameters were obtained from oxygen consumption
data and were shown to be  in agreement with those obtained from traditional
measurement of  substrate removal  (DOC, SCOD, 14C)  or cell  growth.
                            MATERIALS AND METHODS
 EXPERIMENTAL  SETUP
      The  electrolytic  respirometry  studies were  conducted  using  an  automated
 continuous  oxygen  uptake  and  BOD  measuring Voith Sapromat  B-12  (12  unit
 system) electrolytic respirometer-analyzer.  The instrument  consists of  a
 temperature controlled waterbath, containing measuring  units* a  recorder for
                                      277

-------
 digital indication and direct plotting of the decomposition velocity curves
 of organic compounds; and a cooling unit for the conditioning and continuous
 recirculation of waterbath volume.  The recorder shows the digital
 indication of oxygen uptake and constructs a graph for these values of each
 measuring unit.  The cooling unit constantly recirculates water to maintain
 constant temperature in waterbath.  Each measuring unit as shown in Figure 1
 is comprised of a reaction vessel with a carbon dioxide absorber mounted in
 a glass joint flask stopper, an oxygen generator and a pressure indicator.
 This measuring unit is interconnected by hoses, forming an air sealed
 system, so that the atmospheric pressure fluctuations do not adversely
 affect the results.

      The activity of the microorganisms in the sample creates a vacuum which
 is recorded by the pressure indicator, which triggers the oxygen generator.
 The pressure conditions are balanced by electrolytic oxygen generation.   The
 quantity of the sample, the amperage for the electrolysis and the speed of
 the synchronous motor are so adjusted that,  with a sample of 250 ml,  the
 digital counter indicates the oxygen uptake  directly in mg/L.   The C02
 generated is absorbed by soda lime.   The nitrogen/oxygen ratio in the gas
 phase above the sample is maintained throughout the experiment and there is
 no depletion of oxygen.  The recorder-plotter concomitantly constructs  an
 oxygen uptake graph for the selected values.   The oxygen generators of the
 individual  measuring units are electrolytic  cells which supply the required
 amount of oxygen by electrolytic dissociation of a copper sulfate solution
 combined with sulfuric acid.

      The nutrient solution used in these studies was an OECD synthetic
 medium (33,  34)  consisting of measured amounts per liter of deionized
 distilled water of (1)  mineral  salts solution;  (2)  trace salts solution;  and
 (3)  a solution (150 mg/L)  of yeast extract as a substitute  for vitamin
 solution.

      The microbial  inoculum was an activated  sludge from The Little Miami
 wastewater treatment plant in Cincinnati, Ohio,  receiving municipal
 wastewater.   The activated sludge  sample was  aerated  for 24  hours  before  use
 to bring it  to an endogenous  phase.   The sludge biomass  was  added  to  the
 medium at  a  concentration  of 30 mg/L total solids.   Total volumes  of  the
 synthetic medium in the 500  mL  capacity  reactor vessels  were brought  up to  a
 final  volume  of 250 mL.

      The test  and control  compound concentrations  in the media were 100
 mg/L.   Aniline was  used  as the  biodegradable  reference compound, at a
 concentration  of 100 mg/L.

      The typical  experimental system consisted  of duplicate  flasks  for the
 reference substance  aniline and  the  test compounds, a single flask  for the
 physical-chemical  test  (compound control), a  single flask for toxicity
 control  (test  compound  plus aniline  at 100 mg/L each) and an inoculum
 control.  The  contents  of the reaction vessels were preliminarily stirred
 for an  hour to ensure endogenous respiration  state at the initiation of
 oxygen  uptake measurements.  Then  the test compounds and aniline were added
to it.   The reaction vessels were  then incubated at 25°C in the dark
 enclosed in the temperature controlled waterbath and stirred continuously
throughout the run.  The microbiota of the activated sludge used as an
                                      278

-------
Inoculum were not preacclimated to the substrates. The incubation period of
the experimental run was between 28 to 50 days.  A more comprehensive
description of the procedural steps of the respirometric tests is presented
elsewhere (33, 46).

     For fully automatic data acquisition, frequent recording and storage of
large numbers of oxygen uptake data, the Sapromat B-12 recorders are
interfaced to an IBM-AT computer via Metrabyte interface system.  The use of
Laboratory Handbook software package allows the collection of data at 15
minute intervals.

DETERMINATION OF SUBSTRATE BIODEGRADABILITY FROM OXYGEN UPTAKE DATA

     In this study, biodegradation was measured by two approaches:  the
first, as the ratio of the measured BOD values in mg/L (oxygen uptake values
of test compound minus inoculum control - endogenous oxygen uptake values)
to the theoretical oxygen demand (ThOD) of substrate as a percent; the
second as a percentage of the test compound as measured by dissolved organic
carbon (DOC) changes  (OECD Guidelines for Testing of Chemicals)  (34).

     Graphical representation of percent biodegradation based on the
BOD/ThOD ratio were developed against time for each test compound.  The
experimental DOC data for the initial samples and samples for reaction
flasks collected at the end  of experimental run were used to calculate the
percent biodegradation based on the percent of DOC removal in the culture
system.

DETERMINATION OF KINETIC  PARAMETERS OF BIODEGRADATION

     The first order  kinetic rate constants were determined by the
linearization of the  BOD  curves (oxygen uptake of test compound  minus oxygen
uptake of inoculum control), which gives straight lines expressing the
exponential and declining endogenous phases of the BOD curve as  shown in
Figure 2.  The slope  of the  Ln(6 oxygen uptake/St) versus t as described by
Dojlido (6), Tabak et al. (46), Oshima et al.  (35, 36), and Tabak et al.
(47). give specific rate constants of the exponential growth phase (/* values)
and  the declining growth  phase  (#' values) of the BOD curve.

     Acclimation time values (t0) and the time values for the initiation and
termination of the declining growth phase (t1 and t2)  for each test  compound
were determined from  linearized expressions of BOD curves.

     The estimations  of the  Monod Kinetic parameters, maximum specific
growth rate constant, 0m, half saturation constant, Ks and growth yield
constant, Y  were determined directly from experimental oxygen uptake curves
without the consideration of initial growth  and growth yield assumption
[Jobbagy et al.  (22); Tabak  et  al.  (47)].

A. Determination of Yg Constant

     Y  - the true yield  parameter or the ratio of growth of biomass to
substrate utilization, can be obtained for the experimental oxygen uptake
curve at the  initiation  of the  plateau of the  curve as  shown  in  Figure  3.  A
vertical line  is drawn at the point of intersection of  the tagents of the
                                     279

-------
exponential and plateau phases of the curve.  The oxygen uptake value
obtained at the point of  intersection of the vertical line and oxygen uptake
curve is the Oupt  value  -  oxygen  uptake  value at the  initiation of plateau.

     The Oypt value  is then  plugged  into the equation Y  = (S_-0 t/S0)-Y
(where S0 is initial concentration of substrate and Y  is  soluble  product
concentration formed divided by  initial substrate concentration) for Y_
determination.                                                        9

B.  Determination of um Constant

     l^ - the maximum specific growth rate can  be determined from
experimental oxygen uptake  curve plot in the following manner:

(1)  Values of the change of Ou with time (60u/6t)  or slopes  are determined
     along the entire experimental oxygen uptake curve as shown in Figure 4.
(2)  These 60u/6t (slope)  values are then plotted against  the cumulative 0
     values for each time interval, as  shown in Figure 5.
(3)  The slope of the developed linearized form of oxygen uptake curve is
     the estimated /tm value.
C.
    Determination of K. Constant
     Ks - the half saturation constant or the substrate concentration at
which the specific growth rate is 1/2 the maximum specific growth rate can
be obtained from the experimental oxygen uptake curve in the following
manner:
     Value of 0 t  can  be  calculated  from  the  plot  of  (50u/5t) versus Ou
     provided the Ks value is 1 or less (insignificant in comparison to S0
     value) and the plot contains a linear section with the slope p,  as
     shown in Figure 5.
     Other (SOJSt)  versus 0  plot in which the  slope  deviates  from  nm
     because of larger Ks values (more  significant in comparison to  S  )  is
     illustrated in Figure 6.
     The value of 60 /6t  is determined  at the intercept of the  straight line
     developed from the plot of 60u/5t  versus Ou (Figure  5) which contains a
     linear section with slope 0T.
     Beginning with the value of 1/2 the intercept value, another straight
     line (b) is constructed with the slope 1/2 that of the slope of
     original line (a) whose slope is /im.
     At the point where line (b) intercepts the declining experimental curve
     of the plot,  a vertical line from that point  of interception can
     provide the value of Out on the x  axis.
(6)  This Out value  is  then  used in the determination  of  K, with the use of
(1)
(2)
(3)
(4)
(5)
     the equation
                                -  (Out/(l-Y-YJ) =
                                          p  gy
     Where S0 = initial  substrate  concentration
     St  =  substrate  concentration  at  time  t
     Yp  -  soluble product  concentration  formed  (i.e.  intermediate
        metabolites) divided by the initial  substrate concentration
                                    280

-------
(7)   When the Out, Ya, YD,  and S0 values  are  plugged  into  the  equation,  value
     of St can be calculated - which is  the  value of Ks (in systems where Ks
     value is 1 or less).

                           RESULTS AND DISCUSSION


     Respirometric biodegradability, biokinetic and Monod kinetic data for
selected RCRA alkyl benzenes, phenols, phthalates and ketones are reported
in this paper.  The electrolytic respirometry oxygen uptake data for the
test compounds, the control reference compound aniline, the inhibition and
endogenous control systems were generated revealing the lag phase
(acclimation phase), the biodegradation (exponential) phase, the different
bio-reaction rate slopes (characteristic of the test compound) as well as
the plateau region at which the biooxidation rate reaches that of the
endogenous rate of microbial  activity.  Figure 7 illustrates a
representative oxygen uptake  curve  for aniline and the endogenous controls.
Figure 8 shows the replicate  pentaerythritol oxygen uptake curves and the
toxicity control  (pentaerythritol plus aniline) curve. Figure 9 illustrates
a representative graphical treatment of the percent biodegradation of
pentaerythritol with time, which was developed for each test compound  (OECD
studies).

     Based on the biokinetic  equations relating growth rate of microbiota in
presence of above compounds,  the substrate  utilization rate, and rate of
oxygen uptake  (BOD) curves,  specific growth rate kinetic parameters
(biodegradation  rate constants) were derived as  slope values of the
linearized plots  (plots of the  log  of DO/dt) of exponential and declining
growth phases of the BOD curve.  The acclimation time values  (t0), and time
values for the  initiation  and the termination of the declining growth  phases
(tj and t2) for the test compounds and aniline were also  generated.

     The  estimations of the  Monod kinetic parameters for benzene, phenol,
phthalate, and  ketone compounds reported here, were determined directly from
experimental  oxygen  uptake curves without the consideration  of initial
growth and growth yield assumption.

RESPIROMETRIC STUDIES WITH SELECTED RCRA ALKYL BENZENE COMPOUNDS

     The  biodegradation of benzene,  toluene, ethyl  benzene,  m- and
p-xylenes, tert-butyl benzene,  sec-butyl benzene,  butyl  benzene, cumene,  1-
phenyl benzene  and  the  reference compound,  aniline  at  100 mg/L concentration
by  30 mg/L sludge biomass  (as measured  by oxygen consumption by  sludge
microbiota in mg 02/L) was followed over a  period of 20 days.  The
electrolytic respirometry  oxygen uptake and BOD  curves were  generated  and
graphical  treatment of  the percent  biodegradation  was  established  for  each
compound.  Figure 10 demonstrates typical oxygen uptake  and  BOD  curves for
p-xylene and p-xylene + aniline and Figure  11  illustrates graphically  the %
biodegradation of p-xylene with time.

     The percent biodegradation data based  on  the BOD/ThOD  ratios  for
benzene,  toluene, ethyl  benzene,  m- and p-xylene and the reference  compound,
aniline,  are summarized in Table 1.  All  of the  above alkyl  benzene
compounds were shown to be biodegradable substrates at concentration levels
                                     281

-------
 of 100 mg/L when exposed to 30 mg/L of activated sludge biomass under the
 environmental conditions of the respirometric testing procedure, and within
 the period of 20 days of incubation.

      The toxicity test control flask respirometric data revealed no
 inhibitory effects by these test compounds at the 100 mg/L concentration
 levels on the bio-oxidation of aniline by sludge microbiota.

      Table 2 summarizes the bio-kinetic data for the benzenes studied,
 showing the specific growth rate constants for the exponential  growth phase
 0* values) and for the declining growth phase (/{' values)  of  the linearized
 form of the BOD curves of these compounds, as well  as the  t0, t, and t?
 kinetic parameters.  Figure 12 shows a typical  plot of Ln(60u/5t) vs. time
 for toluene,  from which the kinetic parameters were determined.
     Table 3 summarizes the Monod kinetic parameter
these benzene compounds.
                                                         Ks, Y )  data for
                                                              9
 RESPIROMETRIC STUDIES WITH SELECTED RCRA PHENOLIC  COMPOUNDS

      The  biodegradation  of phenol,  resorcinol,  o-, m-  and  p-cresols,
 catechol,  2,4-dimethyl phenol  and the  reference compound aniline  at  100  mg/L
 concentration levels  and exposed to 30 mg/L  biomass was followed  over  a
 period  of 20  days.

      All  of the  phenols  were  shown  to  be biodegradable substrates under  the
 conditions of the respirometric testing procedure.  The toxicity  test
 control flask respirometric data revealed no inhibitory effects by these
 compounds  at  the 100  mg/L levels on the biodegradation of  aniline by the
 sludge  biomass.

      Table 4  summarizes  the bio-kinetic data for the phenols studied,
 showing the specific  growth rate constants as well as the  t0, t,,  and t2
 kinetic parameters.   Table  5 provides  the Monod  kinetic parameter data for
 these phenolic compounds.

 RESPIROMETRIC STUDIES WITH  SELECTED RCRA PHTHALATE ESTER COMPOUNDS

      Evaluation  of the biodegradability and  determining of bio-kinetics  of
 degradation of phthalate  compounds,  dimethyl phthalate, diethyl phthalate,
 dipropyl phthalate and butyl benzyl  phthalate was achieved with use of
 respirometric oxygen  uptake data.

     All of the  above phthalates were  shown  to biodegradable under the
 conditions of the respirometric tests  and were shown not to exhibit any
 inhibitory effects at the 100 mg/L  levels on aniline biodegradation by the
 sludge microbiota.

     Tables 6 and 7 summarize respectively the biokinetic (first order) and
Monod kinetic parameter data for the selected phthalate esters under study.
                                     282

-------
RESPIROMETRIC STUDIES WITH SELECTED RCRA KETONE COMPOUNDS

     Respirometric oxygen uptake data from the studies with the selected
ketone compounds, acetone, 2-butanone, 4-methyl-3-pentanone and a cyclic
ketone, isophorone were utilized to determine their biodegradability and
biodegradation kinetic parameters.

     All the ketones were shown to be biodegradable at 100 mg/L
concentration levels in media containing 30 mg/L biomass and did not exhibit
any toxicity to aniline biodegradation at these concentrations.

     Tables 8 and 9 summarize respectively the first order and Monod kinetic
parameter data for these ketones.

                                 CONCLUSIONS


     The experimental data of respirometric studies with several classes of
organic compounds definitely demonstrate that it is possible to measure the
biodegradability  (percent biodegradation - as a ratio of BOD to ThOD) and to
determine the kinetics of degradation of single organic compounds by using
only measurements of oxygen consumption in respirometric batch reactors.
The values of the kinetic parameters determined from oxygen consumption data
were demonstrated to be similar to those based on the measurements of
substrate removal and those made with cell growth data.

     The generated data on biodegradation, biodegradation rates and
substrate inhibition kinetics through the use of electrolytic respirometry,
will enable the  classification of biodegradability of toxic priority
pollutant and RCRA toxic  organic compounds and ultimate projection of the
fate of organic  compounds of similar molecular structure to those
experimentally studied by way of the established predictive treatability
models based on  structure-activity relationships.

     With the electrolytic respirometry approach, data base of the removal
of the above compounds by biodegradation fate mechanism can be adequately
generated to support the  development  of predictive models on fate and
removal of toxics in industrial  and municipal waste treatment  systems.  A
possible relationship  between the  kinetic parameters  and the effect of
different factors on these parameters, as determined  through electrolytic
respirometry and the structural  properties  of the organic pollutant, can
eventually facilitate  prediction  of the extent and the rate of
biodegradation of organic chemicals  in the  field of wastewater treatment
systems from the knowledge of the  structural properties of the pollutant
organics.

      A preliminary predictive biodegradation  - structure/activity model
based on  the group contribution  approach was developed from the generated
biodegradation  kinetic data  (first order  kinetic parameters) with the  use  of
electrolytic  respirometry.   It  is  expected  that  the model will closely
 predict the  results  found experimentally.   In  this way,  the fate  of  other
 organic compounds may  be anticipated  without the time and  expense of
 experimental work.
                                     283

-------
      The electrolytic respirometry biodegradation studies will  provide basic
 pilot scale treatability information and data which will  be used to confirm
 methods  to predict treatability and the need for pretreatment of
 structurally related pollutants (e.g.,  by structure,  anticipated
 treatability properties,  etc.).  This study will  thus provide a more
 extensive list of pollutants than was covered by experimental  data, for
 consideration in  guiding the Agency to  predict the fate of such compounds
 without  costly experimental  testing.
  1.



  2.


  3.


  4.



  5.
                                  REFERENCES
Arthur, R.M.  Twenty years of respirometry.  In:
Thirty-Ninth Annual Industrial Waste Conference.
West Lafayette, IN, 1984.
Proceedings of the
Purdue University,
Braha, A. and Hafner, F.  Use of lab batch reactors to model
biokinetics.  Water Res.  21(1): 73, 1987.

Bridie, A.L.A.M.  Determination of Biochemical oxygen demand with
continuous recording of oxygen uptake.  Water Res.  3: 157, 1969.

Busch, A.W., Grady, L., Jr., Rao, T.S., and Swilley, E.L.  Short-term
total oxygen demand test.  J. Water Pollut. Control Fed.  34: 354,
1962.                      ~
Cech, J.S., Chudoba, J., and Grau, P.  Determination of kinetic
constants of activated sludge microorganisms.  Water Sci. Technol.
259, Amsterdam, 1984.
                                                                          17:
 6.  Dojlido, J.R.   Investigation of biodegradability and toxicity of
     organic compounds.  EPA-600-2-79-163.  U.S. Environmental Protection
     Agency, Cincinnati, Ohio,  1979.

 7.  Fuhs, G.W.  Some factors affecting biochemical oxygen demand as
     determined in manometric or manostatic devices.  Wasser Abwasser-
     Forschung.  5:  161, 1968.

 8.  Gaudy, A.F., Jr. and Gaudy, E.T.  Biological concepts for design and
     operation of the activated sludge process.  EPA-17090 FQJ.  U.S.
     Environmental Protection Agency, Cincinnati, Ohio, 1971.

 9.  Gaudy, A.F., Jr., Rozich, A.F., Garniewski, S., Moran, N.R., and
     Ekambaram, A.  Methodology for utilizing respirometric data to assess
     biodegradation kinetics.  Paper presented at the 42nd Annual Industrial
     Waste Conference, Purdue University, West Lafayette, Indiana, 1987.

10.  Gaudy, A.F., Jr., Ekambaram, A., and Rozich, A.F.  A respirometric  .
     method for biokinetic characterization of toxic wastes.  Paper
     presented at the 43rd Annual Industrial Waste Conference, Purdue
     University, West Lafayette, Indiana, 1988.

11.  Grady, C.P.L., Jr.   Biodegradation:  Its Measurement and
     Microbiological  Basis.  Biotechnol. Bioenaineer.   27: 660, 1985.
                                     284

-------
12.
13.
14.
15.
16.
 17.
 18.
 19.
 20.
 21,
 22.
Grady, C.P.L., Jr., Dang, J.S., Harvey, D.M., Jobbagy, A., and Wang,
X.-L.  Determination of biodegradation kinetics through use of
electrolytic respirometry.  Water Sci. Techno!.  21: 957, Brighton,
1989.

Grady, C.P.L., Jr., Dang, J.S., Harvey, D.M., and Jobbagy, A.
Evaluation of biodegradation kinetics with respirometric data.  Water
Pollut. Control Fed.  1989: (submitted).

Grady, C.P.L., Jr., Aichinger, G., Cooper, S.F., and Naziruddin, M.
Biodegradation kinetics for selected toxic/hazardous organic compounds.
Hazardous Waste Treatment:  Biosystems for Pollution Control, AWMA,
1989:  (in press).

Halbartschlarger,  J., Kohler,  H., Szwerinski, H., and Bardtke, D.
Investigations on  the biological degradation  of chlorinated
hydrocarbons  using dichloromethane  (methylene chloride)  as an example.
Gwf-Wasser/Abwasser.  125(H.8): 380,  1984.

Hickey, C.W.  and Nagels,  J.W.  Modifications  to electrolytic
respirometer  systems for  precise determination of BOD exertion kinetics
in  receiving  waters.  Water Res. 19:  463, 1985.

Howard, P.H.  and Banerjee, S.   Interpreting  results from
biodegradability tests  of chemicals  in water and soil.   Environment.
Toxicol.  Chem.  3: 551,  1984.

Howard, P.H., Banerjee,  S., and  Rosenberg, A. A review and  evaluation
of  available  techniques for determining  persistence and routes of
degradation of chemical  substances  in the environment:   An update  of
the 1975  Report.   EPA-560/5-81-011.   U.S. Environmental  Protection
Agency, Cincinnati, Ohio, National  Technical  Information Services  No.
PB84-168731,  1981.

Howard,  P.H., Saxena,  J., Durkin,  P.R.,  and  Ou,  L.-T.   Review and
.evaluation of available techniques  for determining  persistence and
routes of degradation  of chemical  substances in  the environment.   EPA-
 560/5-75-006.  U.S. Environmental  Protection Agency, Cincinnati,  Ohio,
National  Technical Information Service No.  PB243825, 1975.

 Huang, J.Y.C., Cheng,  M.D.,  and Mueller, J.T.  Oxygen uptake rates for
determining microbial  activity and  application.   Water Res.   19(3):
 373, 1985.

 Jenkins,  D.   The use of manometric methods  in the study of sewage and
 trade wastes.  In:  P.C.G. Isaac (ed.),  Waste Treatment.  Pergamon
 Press, Oxford, 1960.  p. 99.

 Jobbagy,  A.,  Grady, C.P.L.,  and Tabak, H.H.   Characterization of
 biodegradation through respirometry:  Graphical  analysis and
 theoretical   considerations.    Clemson University Research Report,
 Clemson,  South Carolina.  Water Res. 1989:  (to be submitted).
                                     285

-------
 23.  King, E.F. and Dutka, B.I.  Respirometric techniques,  in:  G. Britton
      and B.J. Dutka (eds.), Toxicity Testing using Microorganisms.  Vol. 1.
      CRC Press, Inc.  Boca Ruton, Florida, 1986.  p. 76.

 24.  Klecka,  G.M.  Fate and effects of methylene chloride in activated
      sludge.   Appl. Environ. Microbiol.  44: 701, 1982.

 25.  Larson,  R.J. and Perry, R.L.  Use of the electrolytic respirometer to
      measure  biodegradation in natural waters.  Water Res.  15: 697,  1981.

 26.  Liebman, H.  and Offhaus,  F.   Volumetric BOD measurements with the help
      of 'Sapromat" a new apparatus for determining 5-day BOD and toxicity.
      Abwassertechnik.   17: 4,  1966.

 27.  Madden,  M. and Tittlebaum, M.  Oxygen uptake rates associated with
      biological treatment of pentachlorophenol wastewater.  J.  Environ. Sci.
      Health.   A19(3):  321, 1984.                             ~  	  	

 28.  Manios,  V. and Balis, C.   Respirometry to determine optimum conditions
      for the  biodegradation of extracted olive press-cake.  Soil  Biol.
      Biochem.  15(1):  75,  1983.

 29.   Montgomery,  H.A.C.   The determination of biochemical  oxygen  demand by
      respirometric methods. Water Res.   1:  631,  1967.
30.
31.
32.
33.
34.
35.
Montgomery,  H.A.C.,  Oaten,  A.B.,  and  Gardiner,  O.K.  An  automatic
electrolytic respirometer--Its  construction  and use.   Effluent Water
Treatment J.   11:  23,  1971.                            ~         ~~~

Nochi, K.  Oxygen  consumption due to  decomposition of  chemical
substances.   J. Water  Waste.  22(11):  1285,  1980.

Nochi, K.  Oxygen  consumption due to  decomposition of  chemical
substances.   J. Water  Waste.  26(7):  751, 1984.

OECD, "OECD  Guidelines for  Testing of Chemicals", Section 3,
Degradation  and Accumulation, Method  301C, Ready Biodegradability:
Modified MITI Test (I) adopted  May 12, 1981  and Method 302C Inherent
Biodegradability:  Modified MITI  Test  (II),  adopted May  12, 1981,
Director of  Information, OECD,  Paris,  France, 1981.

OECD, "OECD Guidelines for Testing of  Chemicals", EEC Directive 79/831,
Annex V, Part C:  Methods for Determination  of  Ecotoxicity.  5.2
Degradation.  Biotic Degradation.  Manometric Respirometry.  Method
DGX1, Revision 5, 1983.  p. 1-22.

Oshima,  A.,  Tabak, H.H., and Lewis,  R.F.   Electrolytic respirometry
biodegradation study - OECD ring test  biodegradation method evaluation
and development.  EPA Draft Report,  MERL.  U.S. Environmental
Protection Agency, Cincinnati, Ohio,  1985.
                                     286

-------
36.  Oshima, A., Tabak, H.H., and Lewis, R.F.  The evaluation of biological
     treatability and removability of toxic organic chemicals by
     respirometry.  EPA Draft Manuscript, MERL.  U.S. Environmental
     Protection Agency, Cincinnati, Ohio, 1985.

37.  Pagga, U. and Gunthner, W.  Biodegradation and toxicity studies with
     microorganisms:  A comparison of laboratory tests with a treatment
     plant model system.  In:  Proceedings of the International Symposium
     on Principles for Interpretation of the Results of Testing Procedures
     in Ecotoxicology.  Valbonne, France, 1980.

38.  Painter, H.A. and King, E.F.  Environment and quality of life-ring test
     programme 1983-84 - Assessment of biodegradability of chemicals in
     water  by manometric respirometry.  Final Report, Contract No.
     XI/W/83/238; Directorate-General Environment, Consumer Protection and
     Nuclear Safety, Commission of the European Communities, Report No.
     EUR9962EN, 1985.

39.  Paris, D.F. and Rogers, J.E.  Kinetic concepts for measuring microbial
     rate constants:  Effects of nutrients on constants.  ADD!. Environ.
     Microbiol.  51: 221, 1986.

40.  Rigin, V.I., Golovin, Y.G., and Tyuneva, G.S.  Determination of the
     biochemical oxygen demand of natural waters and effluents with the
     electrolytic generation of oxygen.  Khimiva 1 Teknologiva Vod.v.  4(2):
     180, 1982.

41.  Robinson, J.A. and Tiedje, J.M.  Nonlinear estimation of Monod growth
     kinetic parameters from a single substrate depletion curve.  Appl.
     Environ. Microbiol. 45: 1453, 1983.

42.  Simkins, S. and Alexander, M.  Models for mineralization kinetics with
     the variables of  substrate concentration and population density.  Appl.
     Environ. Microbiol. 47: 1299, 1984.

43.  Simkins, S. and Alexander, M.  Nonlinear estimation of the parameters
     of Monod kinetics that  best describe mineralization of several
     substrate  concentrations by dissimilar  bacterial densities.  ADD!.
     Environ. Microbiol.  50: 816, 1985.

44.  Simpson, J.R. and Nellist, G.R.  Development and use of a large-volume
     automatic  respirometer.  Water Pollut.  Contr.   69: 596, 1970.

45.  Tabak,  H.H., Quave, S.A., Mashni,  C.I., and Barth, E.F.
     Biodegradability  studies with organic priority  pollutant compounds.
     JWPCF.   53(2):  1503,  1981.

46.  Tabak,  H.H., Lewis, R.F., and Oshima, A.   Electrolytic respirometry
     biodegradation  studies, CEC/OECD ring test of respiration method  of
     determination  of  biodegradability,  Ring Test Program 1984.   EPA Draft
     Final  Report, MERL.  U.S. Environmental  Protection Agency, Cincinnati,
     Ohio,  1984.
                                     287

-------
47.  Tabak, H.H., Desai, S., Govind, R., and Grady, C.P.L.  Evaluation of
     biodegradability and biodegradation kinetics of organic pollutant
     compounds with the use of electrolytic respirometry.  Presented at the
     61st Annual Conference of Water Pollution Control Federation, Dallas,
     Texas.  October 2-6, 1988.

48.  Therien, N., and Ilhan, F.  Relating BOD5 with on-line oxygen uptake
     rate measurements using automatic respirometers in view of process
     monitoring and control.  In:  Procedures and Practices in Activated
     Sludge Process Control.  Vol. 3.  1982.  p. 113.

49.  Urano, K. and Kato, Z.  Evaluation of biodegradation ranks of priority
     organic compounds.  J. Hazard. Materials.  13: 147, 1986.

50.  Verstraete, W., Voets, J.P., and Vanlocke, R.  Three-step measurement
     by the Sapromat to evaluate the BOD5,  the mineral  imbalance and the
     toxicity of water samples.  Water Res. 8: 1077, 1974.

51.  Wojnowska-Baryla, I. and Young, J.C.  Measuring the effect of
     biocatalytic additives on treatment process performance.  J. WPCF.
     55(11): 1373, 1983.

52.  Yoshimura, K. and Masuda, F.  Biodegradation of Sodium Alkyl
     Poly(oxyalkylene)-sulfates.  Am. Oil Chem. Soc. J.  59: 328, 1982.

53.  Young, J.C., Garner, W., and Clark, J.W.  An improved apparatus for
     biochemical oxygen demand.  Anal. Chem.  37(5): 784, 1965.

54.  Young, J.C. and Baumann, E.R.  Demonstration of the electrolysis method
     for measuring BOD.  Draft Final Report to the U.S. Environmental
     Protection Agency from Iowa State University for Grant WP16020 DUN and
     2-80036, 1972.

55.  Young, J.C. and Affleck, S.B.  Long-term biodegradability tests of
     organic industrial wastes.  In:  Proceedings of the 29th Industrial
     Waste Conference.  Purdue University, Extension Series 145, West
     Lafayette, Indiana, 1974.  p. 154.

56.  Young, J.C. and Baumann, E.R.  The electrolytic respirometer-I.
     Factors affecting oxygen uptake measurements.  Water Res.  10: 1031,
     1976.

57.  Young, J.C. and Baumann, E.R.  The electrolytic respirometer-II.  Use
     in water pollution control plant laboratories.  Water Res.  10: 1141,
     1976.

58.  Young, J.C.  Biochemical oxygen demand:  Measurement and application.
     Iowa State University, Ames, Iowa, Final Report, 1977.
                                    288

-------
3 *-t a m
W  V *i U
«  H *• -rt

e  Sirs
                              «  g-s
                              O  -rt C 3
                                 •MUM
                              H  U DIM
                              «. « >«4>
                              n  a HM
                              tK*> U O Pi
                              (••rt
                              •* a
                              o a«au
                                        N
                                        «.-»
                                            U
                                            o
                                        1
                                        4JO tl C «l
                                        M m J]*H 41 H
                                         MM  >• V
                                        u »•» o « f<-a u
                                        'ft  • « O *« *1 r^ 4)
                                        6 MU 0.41 41 U



                                        •-i« n 
-------
          FIGURE 4   PLOT OF OXYGEN UPTAKE RATE
                              ( Ou VERSUS  TIME )
              10     20     30     40     50   .60    70
                      UM-0.1. KS-1. YC-0.4. XO-5. SO-100
FIGURE 5  Ptot showing (AQu/At) versus Out curve containing a
             ifnoAP Oa/*tu*vn tAjHh a *tts**>ui.»  4* WA UMA^ L _._&*	A.'	
     6.0
     4.5
     4.0
     as
   £3.0
   g2£
   02.0
     15
     to
     0£
     0.0
             linear section with a slopes mto be used In estimating
             the values of K Jn cases whore K_values are nealioible
             In comparison t
-------
    500
en
             10      20      30      40



                 ELAPSED TIME (Days)




FIGURE 7    OXYGEN UPTAKE DATA ON ANILINE



  500





  400
«   300
z
a
u


U)
         — — — PenlMqrthrilol A .

         —— Ptntxrirthfitol B .

              PentMrythrilol It Aniline
    200
     100 •
               "TO      20      30      40



                   ELAPSED TIME <0ayo>
 FIGURE 8   BIOLOGICAL OXYGEN UPTAKE CURVE

 r'°U              (Run 1 Sample No. 10-12)
     100
 a
 a
 m
               Paitaajthritol A


          — — — PenlMnrthritol 0
                    ELAPSED TIME  (days)


FIGURE 9  BIODEGRATION (% BOD REMOVAL) CURVE

                     ( PENTAERYTHRITOL )
                       291

-------
             1. Inoculum Control

             2. p-Xjtcn.

             3. p-Xylm BOD

             4. p-X;l
-------
    TABLE 1.   SUMMARY OF RESPIROMETRIC  BIODEGRADATION DATA FOR SELECTED BENZENES
                  PERCENT BIODEGRADATION (BASED ON % BOD REMOVAL)
Time
(days) Aniline
0.0 0.0
1.0 2.58
2.0 3.43
3.0 3.94
4.0 14.95
5.0 102.0
6.0 102.0
7.0 102.0
8.0 102.0
9.0 102.0
10.0 102.0
11.0 102.0
12.0 102.0
13.0 102.0
TABLE Z.
COMPOUNDS
Aniline
(Experiment 1)
(Experiment 2)
Benzene
Ethyl benzene
Toluene
p-Xylene
m-Xylene
tert-Butyl benzene
sec-Butyl benzene
Cumene
Butyl benzene
1-Phenyl hexane
Benzene
0.0
2.11
2.11
4.15
4.97
71.5
74.22
81.85
93.37
93.89
95.45
95.45
96.13
97.46
SUMMARY OF
ThOD
for 100
310
310
308
317
313
317
317
322
322
320
322
326
To! uene
0.0
2.78
8.81
85.11
89.04
94.34
97.92
100.0
101.18
103.48
103.48
103.48
103.48
103.48
BIO-KINETIC DATA
mg (days)
4.00
4.00
4.50
4.00
2.00
3.90
2.00
4.40
3.50
2.40
3.30
4.00
Ethyl benzene
0.0
2.27
2.27
2.99
3.78
4.29
40.25
73.91
73.78
83.94
. 88.86
88.86
89.65
91.00
FOR SELECTED














BENZENE
t; t2
(days) (days)
4.70
4.65
4.87
4.21
2.20
4.22
2.35
5.12
4.00
2.79
3.92
4.55
4.83
4.79
5.00
4.83
2.42
4.83
2.50
5.70
5.70
3.00
4.56
5.15
m-Xylene
0.0
1.35
2.14
69.68
69.68
76.68
82.58
84.67
87.03
87.79
90.12
90.15
90.97
91.80
COMPOUNDS
/I
(day-1)
2.78
3.80
8.57
8.33
8.75
9.94
6.60
1.21
0.78
2.31
2.42
1.85
p-xylene
0.0
1.58
1.58
1.58
9.4
70.38
82.68
95.58
99.49
101.76
104.7
105.78
106.7
108.6

*'
(day-1)
3.29
8.60
25.30
44.80
14.93
4.50
29.60
1.69
0.73
2.24
2.42
1.90
It  - specific growth rate constant for exponential  growth  phase of BOD curve.
p' » specific growth rate constant for declining  growth  phase of BOD curve.
                                        293

-------
  TABLE 3.  SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED BENZENE COMPOUNDS
COMPOUNDS
Aniline
Benzene
Ethyl benzene
Tolouene
p-Xylene
•-Xylene
tert-Butyl benzene
sec-Butyl benzene
Cunene
Butyl benzene
1-Phenyl hexane
Lag Time

-------
  TABLE  5.   SUMMARY  OF  MONOD KINETIC  PARAMETER DATA FOR SELECTED  PHENOLIC COMPOUNDS
COMPOUNDS
Aniline
Phenol e
Resorcinol
p-Cresol
o-Cresol
m-Cresol
Catechol
2,4-dimethyl phenol
Lag Time
(t0)
days
4.00
1.00
1.50
1.00
1.20
1.44
0.85
2.00
Y9
mg biomass
mg substrate
0.38
0.58
0.48
0.33
0.41
0.46
0.49
0.39
Mm
(day-1)
6.15
9.82
12.22
6.11
4.10
7.97
12.80
5.62
mgs/l
6.10
9.43
6.31
27.78
16.41
17.62
43.87
14.07
   H - maximum specific growth rate.
   K, - half  saturation constant; concentration of substrate at
   Yg - growth  yield, mg biomass formed/mg substrate coriiumed.
    TABLE 6.  SUMMARY OF BIO-KINETIC DATA FOR SELECTED PHTHALATE ESTER COMPOUNDS
COMPOUNDS
                       ThOD          t0         t,         t2          it           v'
                     for 100 mg     (days)     (days)     (days)     (day-1)     (day-1)
Aniline
Dimethyl phthalate
Diethyl phthalate
Dipropyl phthalate
Butyl benzyl phthalate
310
168
195
211
226
4.00
3.46
2.00
2.40
2.00
4.70
3.98
2.97
2.87
2.28
4.83
4.25
3.30
3.40
2.80
2.78
2.76
2.16
2.04
4.12
3.29
4.71
2.92
2.00
2.33
       specific growth rate constant for exponential  growth phase of BOD curve.
       specific growth rate constant for declining growth phase of BOD curve.
                                           295

-------
TABLE 7.  SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED PHTHALATE ESTER COMPOUNDS
COMPOUNDS Lag Time
(t0)
days
Aniline
Dimethyl phthalate
01 ethyl phthalate
Dlpropyl phthalate
Butyl benzyl phthalate
H * maximum specific
4.00 .
3.46
2.00
2.40
2.00
Y j^
ma biomass (day-1)
mg substrate
0.38 6.15
0.43 7.07
0.46 3.00
0.48 5.78
0.61 7.80
«g'/i
6.10
41.68
11.67
15.81
36.25
growth rate.
K, « half saturation constant;
Y, - growth yield, ng
biomass
concentration of substrate at 0j/2.
formed/ing substrate consumed.


        TABLE 8.  SUMMARY OF BIO-KINETIC DATA FOR SELECTED KETONE COMPOUNDS

COMPOUNDS              ThOD           t0          t,         t,          n           n'
                     for 100 ng     (days)      (days)      (days)      (day-1)     (day-1)
Aniline
Acetone
2-Butanone
4-Methyl -2-pentanone
Isophorone
310
221
244
272
278
4.00
3.70
2.00
1.85
22.30
4.70
3.99
2.20
2.24
23.70
4.83
4.18
2.35
2.35
25.40
2.78
2.45
2.41
2.31
0.73
3.29
3.98
4.98
4.80
0.38
  p  « specific growth rate constant for exponential growth phase of BOD curve.
  It' • specific growth rate constant for declining growth phase of BOD curve.
   TABLE 9.  SUMMARY OF MONOD KINETIC PARAMETER DATA FOR SELECTED KETONE COMPOUNDS
COMPOUNDS
Aniline
Acetone
2-Butanone
4-Hethyl -2-pentanone
Isophorone
Lag Tine
(to)
days
4.00
3.70
2.00
1.85
22.30
ma biomass
ng substrate
0.38
0.36
0.39
0.45
0.43
(day-1)
6.15
4.86
5.11
6.40
1.57
*s
«>9/l
6.10
9.76
10.79
24.70
27.42
  /i - naxlnua specific growth  rate.
  K, - half saturation constant; concentration of substrate at
  Y, - growth yield, *g bloaass formed/eg substrate consulted.
                                           296

-------
       PREDICTION AND MODELING OF BIODEGRADATiON KINETICS OF
                     HAZARDOUS WASTE CONSTITUENTS

                          Rakesh Govind and Sanjay Desai
                      Dept. of Chemical & Nuclear Engineering
                             University of Cincinnati
                              Cincinnati, OH 45221
                                      and
                                 Henry H. Tabak
                    RREL, U S Environmental Protection Agency
                              Cincinnati, OH 45268

                                  ABSTRACT

     Biodegradation is the  most important mechanism in  controlling the concentration
of chemicals in an  aquatic  system  because it  can mineralize  toxic pollutants to
innocuous forms.  So the  fate  of organic  chemicals  in  an aquatic environment,  is
dependant on  their  susceptibility  to  biodegradation.  Because  of  a large  number of
chemicals, it will  be expensive and labor intensive  to gather this information in a
reasonable amount of time. Hence, there is  need for a  prediction method to obtain
these data.

     Experiments  using  electrolytic respirometer,  were conducted to collect oxygen
consumption data  of some RCRA compounds and  first  order kinetic rate constants
were  obtained.   Using  these  data  along with those from literature, a  structure-
activity relationship was developed.
                                       297

-------
                                  INTRODUCTION

      It has been  estimated that 50,000 organic chemicals are commercially produced in
 the United States  and a  large number  of new  organic chemicals  are added  to the
 production each year  (1).  The presence of many of these chemicals in the environment
 is  a  serious public health  --•-•---   -••  •               •• •     .......
 disposal techniques.
ilth problem. Their presence could be  attributed to inadequate
      Since  many of these 'hazardous  chemicals can be  detected  in wastewater,  their
fate in wastewater treatment system  is  of great  interest.  Of the many factors that
affect the  fate  of these  compounds,  microbial  degradation  is  probably  the  most
important  (2).  The   high  diversity  of  species   and  the  metabolic  efficiency  of
microorganisms suggest  that  they play a  major role  in  the ultimate degradation  of
these   chemicals   (3).   Biodegradation   can   eliminate  hazardous   compounds   by
biotransforming  them  into innocuous  forms,  degrading  them by  mineralization  to
carbon dioxide  and water. So the information regarding the extent and  the rate  of
biodegradation  of  these  chemicals  is  very  important   for  regulation   of   their
manufacture and  use.  Due to a  large number of  these  chemicals, gathering of this
information in a reasonable amount of time will be  both expensive and  labor intensive.
Thus, the only practical way out is to  develop correlations  and predictive techniques
to assess biodegradability (4). Lack of an adequate database on  biodegradation kinetics
prevents the development of such techniques.

                           EXPERIMENTAL TECHNIQUES

     There are many techniques  for collection of biodegradation  data and  these  are
reviewed  in great detail by Howard,  et  al. (5)  and Grady  (6). Experimental techniques
for biodegradation  data collection fall  into tnree  broad categories :  continuous,  fed-
batch and batch reactor systems.

     Continuous culture reactors require an acclimated biomass.  They also require long
transitional  time intervals for  reaching quasi steady state condition (7). So  it is time
consuming, tedious and expensive. However, the analysis of data  obtained is simpler
because  the equations for continuous  reactors, operating at  steady  state,  reduce  to
algebraic  equations that are easily solved.  This technique is  used  more for evaluating
parameters  for the design of  treatment systems rather  than  for biodegradation kinetics
(8).   Because of the microbial competition, each continuous  reactor will have  a unique
microbial  community associated with it  (9); hence the evaluated  kinetic  parameters are
system specific and are not intrinsic.

     Fed—batch  reactors have also been used  to  estimate  the kinetic parameters  of
biodegradation. In these reactor  configurations, a quantity of biomass  is added into a
reactor and a substrate stream is  continuously added in negligibly  small amounts with
respect to  the  reactor  volume.  Because  of small amounts  of both  substrate  and
microorganisms, a  pseudo steady state is achieved. So this technique reduces  the time
required  and also alleviates the  problems associated with  changes in the composition
of the microbial community. This configuration requires an acclimated biomass because
the microbes must be capable of responding instantaneously  to the input of substrate.
However,  this technique cannot be used to determine the  Monod biokinetic parameters,
but it is  an excellent  procedure to determine the parameters for  treatment  plant
design or operation.

     The use of batch cultures  in biotechnology and  biological  wastewater treatment
                                          298

-------
represents  a less  expensive  and much  faster way  to model  biokinetics  in  fermenters
and  in  activated  sludge  tanks  (7). The batch  method  commonly  used for a  large
number of  compounds (10)  is  one in  which the substrate of  interest, at  different
concentrations, are inoculated with the small amount of biomass.  Then the  increase in
biomass  concentration  in  each  reactor is measured as  a  function  of  time.  Another
technique focuses  on the substrate removal rather  than the microbial growth. This is
commonly  used in engineering  studies. The batch reactors are inoculated  with  large
quantities of biomass and the  substrate  removal is  measured  as a function  of  time.
Both  of  these  techniques  have  been widely  used using general measures as  five day
BOD  and  COD.  Batch  reactor can  be used  with either acclimated or unacclimated
biomass. It requires that samples be taken at discrete time intervals during  the course
of biodegradation.  So if unacclimated biomass is  used, the number of samples required
may  be  large  depending  on  the acclimation  time. Tabak, et  al. (11)  have  collected
degradability and   acclimation  data on 96  compounds  by  static  culture  screening
procedure and culture enrichment process. Larson and Perry (12) and  Paris,  et al. (13)
have  done  biodegradation studies with  unacclimated biomass in  the batch reactors and
have evaluated the kinetic parameters.

      In the above  procedure, the number of data points  collected are less  because of
manual sampling. This  can be avoided by monitoring oxygen consumption as  an indirect
measure  of  biodegradation  using  an  electrolytic  respirometer.  The automatic   data
collection and  recording allows  sufficient accumulation of data,  so  the reliability  of
the kinetic parameters evaluated is maximized.

      In the respirometer methods of BOD measurement,  wastewater samples  are kept
in contact  with the gas phase source of oxygen. Oxygen uptake by the microogranisms
over  a period of time  is measured by  the  changes in  volume  or pressure  of the  gas
phase. An  alkali is included in the apparatus to absorb  carbon dioxide produced during
biodegradation.  Samples are  usually incubated at  constant temperature  and  are  kept
away from  light.  The latest development  in  respirometric techniques  has been  the
advent of  an electrolytic respirometer.  It supplies  oxygen, produced  from  electrolysis
of water, to the  air space above the sample in a completely  sealed reaction vessel.
The  production of oxygen is triggered  due to the  pressure changes in the reaction
vessel. Studies by Larson and Perry (12), Young  and  Baumann  (14), Tabak,  et al. (15)
and  Dosanjh and  Wase (16) have  shown that the electrolytic  respirometer eliminates
most of  the technical difficulties associated  with other  methods for determining  BOD.
It is  particularly  useful for  the rate studies because  it  provides both,  a  continuous
record of oxygen uptake and it  maintains an unchanging atmosphere over the sample
regardless of the length of the test.

                             EXPERIMENTAL SET-UP

      The electrolytic respirometer Sapromat B—12, consists of a temperature controlled
waterbath,  which  contains the  measuring units, a  recorder for digital  indication and
direct plotting  of the oxygen uptake curves  and  a  cooling unit. The waterbath has  12
reaction  flasks, each connected to the  recorder. Each  unit  as shown in  figure  1
consists  of a  reaction vessel  C,  with a  carbon dioxide absorber (sodalime) 3,  mounted
in  a   stopper,  an oxygen generator  B, and a pressure indicator  A;  The vessels
interconnected  by  hoses,  form  a  sealed measuring system so  that the  barometric
pressure  fluctuations do not  adversely  affect the result. The magnetic stirrer 1,  in the
sample 2,  to  be  analyzed,  provides  vigorous  agitation, thus ensuring an  effective
exchange of  gases.  The  activity  of  the  microorganisms in  the sample  creates  a
reduction in the  pressure  which  is  recorded by  the  pressure  indicator.   It controls
both,  the  electrolytic oxygen generation  and  the  indication  and the plotting of the
                                         299

-------
measured values.

     The consumption of oxygen by microorganisms  creates a reduction in  pressure in
the  reaction  vessel. As a  result, the  level of 0.5%  sulphuric  acid  in  the pressure
indicator  rises and  comes  in  contact with the platinum electrode. This  completes the
circuit and triggers  the  generation  of oxygen  by  electrolytic cell. The oxygen  gas is
provided  to the  reaction  vessel, alleviating the  negative  pressure.  So, the level of
electrolyte in the pressure indicator  drops down,  breaking  contact  with the electrode.
This switches  off the electrolytic cell. The amount of oxygen supplied to the sample is
recorded directly in  milligrams per litre by  the recorder. The recorder is  connected to
an IBM AT which records data from the measuring units every 15 minutes.
                         Figure 1. Diagram of a Measuring Unit

                      A.  Pressure indicator
                      B.  Oxygen generator
                      C.  Reaction vessel
                      1.  Magnetic stirrer
                      2.  Sample (250ml)
                      3.  CC>2 absorber
                      4.  Pressure indicator
                      5.  Electrolyte
                      6.  Electrodes
                      7.  Recorder
                            BIODEGRADATION KINETICS

     A  considerable  amount  of  information  concerning  biodegradation   kinetics  is
available  in  the  published  literature.  Early  literature  shows  widely  differing  kinetic
rates  in different  studies. The evaluation and prediction of  the  extent and  rate  of
biooxidation  is affected by methodological and  experimental  factors.  Regardless of the
different  assumptions  involved in  the  measurement  of  biodegradation rates,  it  is
                                          300

-------
 generally considered associated with microbial cell growth  and so most of the models
 for it are the same as those used to model growth and substrate removal.

      Although  many models  have  been proposed for microbial  growth, the  Monod
 relation  is the most popular kinetic expression (17). Monod model, in combination with
 the  linear law for substrate removal can provide an adequate description of microbial
 growth  behavior. It states that the  cell growth is  first order with  respect  to the
 biomass concentration  (X) and mixed  order with respect to the substrate concentration
 (S)

                            dX/dt = (S u™ X) / (Ks + S)  .  .  .  .  .  (1)

 Cell growth is related to the substrate removal by the linear law

                                dX/dt = - Yg (dS/dt)	(2)
      The  kinetic  parameters of interest  are  maximum  specific growth  rate |xm, half
 saturation constant Ks, (it is the  concentration of substrate when u,=0.5jxm),  and the
 yield coefficient Yg. The Monod equation has two limiting cases.  When the substrate
 concentration  is  much  greater than  the  saturation  constant  the term  (S/KS+S)
 approaches 1.0 and cell  growth and  substrate removal  are zero order with respect  to
 substrate concentration.   When the substrate concentration is much smaller than the
 saturation constant,  the term  (S/KS+S)  approaches  (S/KS)  and  cell growth and
 substrate removal  are first order  with respect to the  substrate concentration.  Many
 researchers have  used  either of  the above  two  approaches. The characteristics  of
 these kinetic expressions have been discussed by Sim kins and Alexander (18).

      The  electrolytic  respirometer  has been  mostly used to  measure the  extent  of
 biodegradation as a percentage of  the theoretical oxygen demand over a certain period
 of  time.  Several  researchers  have tried  to  extract kinetic  parameters from  oxygen
 uptake  data.  Larson  and  Perry  (12) used  empirical kinetic  expressions which  were
 system  specific.  Dojlido  (19)  divided  the  oxygen  uptake curve into  seven  different
 phases  and  then  proposed  an empirical model  for each  phase  and evaluated the
 biodegradability and toxicity of a test compound by measuring empirical rate  constants
 and time  intervals  associated with  each phase. Tabak, et al. (15) divided the substrate
 removal  region  of  oxygen  uptake  curve  into two  regions, separated by  an  inflection
 point. The  first  period  was  called  exponential phase where it was  assumed that the
 substrate  was  not  limiting and cell growth was occuring,  while the second period was
 called declining phase  and here it  was assumed  that the substrate  was limiting. So
 work done to obtain kinetic parameters from oxygen  uptake curve has been empirical.

                      STRUCTURE-ACTIVITY RELATIONSHIP

      The  structure—activity relationships (SAR) have  been widely used in pharmocology
 and medicinal chemistry. The different  methods  and  models  used are free energy
 models, Free-Wilson mathematical model, discriminant analysis, cluster analysis, pattern
 recognition, topological methods, and quantum mechanical  methods.

     The  free energy  model of Hansch, et al. (20)  is widely  used. They incorporated
octanol—water coefficient,  log P, in the linear free energy relationship as a measure of
lipophilicity. This provided a  general SAR model for biological  activity. The success of
this model has led many  researchers  to include additional physicochemical parameters
and properties,  structural,  topological  and molecular indices.  Using similar  principles
other researchers have  proposed models to include more  complex relationships between
                                         301

-------
the bioactivity and the  chemical  structure or the properties. Martin  (21) has dicussed
these models.

     Free and Wilson proposed a mathematical model to assess  the  additive effects of
substituents  and quantitatively estimate their magnitude  (22). According to their model
the structure of  a  compound is composed of different groups or  a  core which is
substituted in various positions, resulting in a  series of linear equations of the form
where  BA is  biological activity, X: is the j**1 group, aj is  the  contribution of the jtn
group  and (3  is the overall average activity.  These linear equations are solved by least
square method for a;  and (3. This method requires a large number of compounds for a
meaningful  analysis and  it  will breakdown  if there  are interactions between different
groups. Fujita and Ban (23) suggested that BA should  be expressed as log(l/C), where
C  is the  concentration of the compound  that produces a constant biological response.
This modified Free— Wilson model is in common use in medicinal chemistry.

     Discriminant analysis is used where  only semiquantitative or qualitative  data have
to  be  evaluated. In  this method,  a linear combination  of parameters  called  linear
discriminant function  is  formed,  which  classifies the  observations.  Martin  (21)  has
discussed the background of this method  and  has given examples.  Principal component
method is  used  as a  preliminary step in multiple regression  analysis  of  the  Hansch
type   (24).   Factor   analysis   is  used   to  gain  insight  into  the  structure   of
multidimensional data set and involves manipulation of the eigenvectors of  variance—
covariance  matrix of the dataset  (25).  Cluster analysis  is  used  to group similar
substituents when various combination of parameters are considered (26).
     Pattern  recognition   is   used  to  examine   structural   features  and   chemical
properties  for the  patterns  associated  with different biological activities  (27). In this
method, a set  of descriptors  is  generated  for  each  compound  and  then  suitable
algorithm  is applied to find some combinations  and weight  of the descriptors which
give a perfect classification  for a set of compounds. This classification is then applied
to another set of compounds of known classification and performance is judged by the
percentage of correct prediction.

     Various methods  have been proposed to relate topology of the molecule  with its
biological activity.  Verloop  (28)  has proposed  a method to treat directionality of  steric
effects. A  computer representation of  a  compound  is created  and then  measured by
tangential  planes which results  in five STERIMOL parameters. Kier and  Hall  (29)  have
used   molecular  connectivity   index,  a  number  calculated   from  graph  theoretical
principles for SAR correlation. Blankley (30) has reviewed all  these and other methods
used in SAR.

     In the field of biodegradation there are several studies which have attempted to
correlate  some  physical,  chemical or  structural  property  of  a  chemical   with its
biodegradation.  Based  on the type  and the location of the substituent groups, Geating
(31)  developed  an algorithm  to  predict  biodegradation. Qualitative relationships for
different compounds have  been investigated  by others,  but  quantification is  required
for regulatory purposes.

     In the literature,  first  order rate of biodegradation or five day BOD of  chemicals
have  been  correlated  with  physical  or  chemical  properties.   Paris, et  al. (32,33,34)
established a correlation between  second order  biodegradation rate constant  and the
                                          302

-------
van der Waal's  radius of substituent group, for substituted anilines and for a series of
para-substituted phenols. Wolfe, et al. (35)  correlated second order alkaline hydrolysis
rate  constant and  biodegradation rate constants for selected  pesticides  and phthalate
esters.  Several  workers  have  observed  a  correlation  between  biodegradability  and
liphophilicity,  specifically  octanol/water partition coefficients (log  P).  Paris, et  al. (36)
found a good correlation  between biodegradation rate  constant and log P, for  a series
of esters  of 2,4—dichlorophenoxy acetic  acid. Banerjee, et al.  (37) obtained a similar
relationship for  chlorophenols.  Vaishnav, et al.  (38)  correlated  biodegradation  of  17
alcohols  and  11  ketones with  octanol—water  coefficients  using 5—day  BOD data.
Pitter(39)  has found  a  dependence of biodegradation rate on electronic  factors,  like
Hammett  substituent constant, for a  series  of  anilines  and phenols.  Dearden  and
Nicholson  (40) have  correlated  5—day  BOD  with  modulus difference  of  atomic charge
across a selected  bond in a molecule for amines,  phenols, aldehydes, carboxylic acids,
halogenated   hydrocarbons   and  amino  acids.  A  direct  correlation  between  the
biodegradability  rate  constant and  the  molecular  structure  of the  chemical has  been
used  by Govind  (41) to relate  the first order  biodegradation  rate constant with the
first   order molecular  connectivity  index  and  by  Boethling  (42)   to  correlate  the
biodegradation rate  constants  with  the  molecular  connectivities  for  dialkyl  esters,
carbamates, dialkyl ethers, dialkyl phthalate esters and aliphatic acids.

                       GROUP CONTRIBUTION APPROACH

     Using a  group  contribution  approach,  a very   large number  of chemicals  of
interest can  be  constituted  from  perhaps  a  few  hundered  functional  groups. The
prediction  of  the  property is based on  the  structure  of the compound. According to
this  method the molecules of a  compound are structurally decomposed into functional
groups  or their  fragments,  each having unique contribution  towards  the compound
property.

     The  biodegradability  rate  constant  k, is expressed as  a  series  function  of
contribution a;, of each group of the compound. The  first order approximation of this
series function representing biodegradation rate constant can be expressed as
                         Ln (k) = 2
(3)
where  Nj is  the number of groups of type j in  the  compound, oj is  the  contribution
of group of type j and L is the total number of groups in the compound.

CALCULATION OF GROUP CONTRIBUTION PARAMETER, a

     Using equation 3  for each compound,  a  linear equation in ot's is  constructed.  For
a  given  dataset this generates a  series of linear equations which are solved for  ex's
using regression analysis.

     If there are n a's  and m  compounds (n
-------
                             m
       S = 2 { Ln(ki)
          1=1
                                               j }
(4)
The estimates of a's should  be such that it produces the least value of S. These
estimates are obtained by differentiating equation 4 with respect to a  and setting it
equal to zero
                       m
= -22{[Ln(ki)-2Njjaj]Nki} =
                                                                   (5)
This will generate a series  of n  linear  equations which  are solved for a's. If N is the
matrix of coefficients  of a's  and Y is the vector  of Ln(k)  values then the solution
vector a is given by

                                a=(N'N)~1 (N'Y)

where ' denotes the transpose of matrix.

                              MODEL VERIFICATION

     The  study of Urano and Kato (43) was  used to determine  the  contribution of
several different groups.  This  study  was selected because  it involved testing  of a large
number of chemicals using a consistent method. The requirement of consistent method
is  necessary  to ensure  that  the  difference  in  rate  of  biodegradation of different
compounds  is due  to the difference  in  their chemical  structure and not  the  test
conditions.  The  range  and average of Lnfk) values, for compounds used for this model,
are given in  Table  1.  The average values  were used for  calculation of  the group
contribution parameters given in Table 2.

     The biodegradation experiments were carried  out for cresols, phenol, 2,4— dimethyl
phenol,  2— butanone, acetone, 1— phenyl hexane and butyl  benzene using an electrolytic
respirometer Sapromat B— 12 (Voith— Morden, Milwaukee, Wl). The chemicals  were from
Aldrich  chemical company with 99+%  purity. These compounds were used to validate
the model, but  were not used in the calculation of the  group contribution  parameters.
The experimental conditions were : temperature 25°C, biomass concentration  30 mg/L
and  compound concentration lOOmg/L. The biomass was  obtained from Little  Miami
wastewater  treatment  plant in  Cincinnati, which receives 95% domestic waste.  The
biomass  was   aerated  for  24  hours  and  then   used  for   the  experiment.   The
biodegradation  rate   constants   for  these  compounds  were  determined   using   the
following empirical first order rate  equation given by Urano and Kato (44).
                             dBOD/dt = k BOD.
Integrating this equation
                        log (BOD) = kt + constant
                                                 (6)
                                                 (7)
is  obtained. The above equation was used  for only the rising part of the BOD  versus
time curve, and initial data and data for endogenous phase were neglected. Note that
the selection of the  rising  part of the BOD  curve  was  arbitrary and was  based on
visual  inspection of  the BOD  curve.  The biodegradation  rate  constants  were also
predicted using the group contribution parameters of  Table 1. The comparison of these
values  is given in Table 3  and  Figure  2. The % error in BOD values were calculated
                                         304

-------

                                                   
-------
by
                      % Error = (1/n) 2 {(BODe - BODp)/BODe}*100
where subscripts e and p  denote  experimental and  predicted values  respectively and n
is  the total number of data points used.  Figure 3  shows the best and  worst case of
prediction  obtained.  The  experimental  conditions  in  both  the  studies  were  similar
except the nature and the source of biomass.

                            RESULTS AND DISCUSSION

     The average error in  the prediction of BOD values varies from  13% to 85%. It is
important to emphasize that there are several reasons for the rather large discrepancy
between the predicted and experimental values. These reasons are as follows :

1.    The data used for calculating the group contribution parameters were  developed
by  Urano  and  Kato (refer Table 1).  The  experimental  BOD curves, obtained from
electrolytic  respirometry,  were   approximated  by  an  exponential   relation   (refer
equations 6  and  7}.  The  range  of curve  to  which this equation was applied was
selected arbitrarily. This resulted in some  error between the experimental data and the
calculated  value  of  the  kinetic constant,  k. Furthermore, a  range  of k values were
calculated  by Urano and  Kato  due  to  differences  in  the replicate samples.  In  our
calculation  of the group  contribution parameters,  an  average  value  for the  k  values
was used.  This introduced another error in  the input data  for the group contribution
estimates.

2.    The group contribution method described here is the first order approximation. It
assumes that the contributions of the groups  are  independent of each  other. So  the
rates predicted are within an order of magnitude.

3.    Biomass used in both the studies were different. Urano and Kato had  acclimated
their biomass to the compound  before using it in  the  experiment, while our biomass
was from  domestic wastewater  treatment  plant  and  was not  acclimated  to  the
compounds used in the study.

     Inspite  of  these errors, the  prediction is  within  an order of  magnitude. With
availability  of more  data  and using a  detailed  kinetic model rather  than a  first order
exponential equation, the prediction error may  be reduced considerably.

ACKNOWLEDGMENT  :This  research  was  supported by  co-operative agreement  CR
812939—01 from U. S. Environmental Protection Agency.

                                   REFERENCES

1.    Blackburn, J.W., Troxler, W.L. Prediction of the fates  of organic chemicals in a
     biological treatment process — An overview. Environ. Prog. 3: 163,1984.
2.    Howard, P.H.,  Saxena,  J., Sikka, H. Determining the fate of chemicals. Environ.
     Set. Tech. 12: 398,1978.
3.    Alexander,  M. Biodegradation of chemicals of environmental concern. Science 211:
     132,1980.
4.    Strier, M.P. Pollutant  treatability  :  A molecular  engineering  approach. Environ.
     Sci. Tech. 14: 28,1980.
5.    Howard, P.H.,  Banerjee, S., Rosenberg, A.  Review and evaluation of available
     techniques for  determining  persistance and  routes  of degradation of  chemical
     substances in environment. 560/5—81-011, U.S.  Environmental Protection Agency,
                                         306

-------
 o
 Q
 O
 m

 o
 o
 m
 s^»
 0>
o
Q
O
m

Q
o
m
•s^X
o»

3
0.9




0.8 -




0.7 -




0.6 -




0.5 -




0.4 -




0.3 -




0.2 -




0.1 -




  0
         0.05
      0.8
      0.7 -
      0.6 -
      0.5 -
0.4 -
      0.3 -
      0.2 -
      0.1 -
                 2,4-Dimethylphenol
                  0.15
0.25
0.35
                                                                    0.45
                    1-Phenylhexane
                            0.2
                                        —1—

                                         0.4
                                                                  0.6
                                     Time, days
         d   Experimental Data                      +   Predicted Data



               Figure 3.  Best and worst case  of prediction
                                    307

-------
     1981.
6.   Grady,  C.P.L.,  Jr. Biodegradation  : Its measurement  and microbiological basis.
     Biotech. Bioengg. 27: 660, 1985.
7.   Braha,  A., Hafner, F.  Use of lab batch reactors  to  model kinetics.  Wat Res. 21:
     73,1987.
8.   Chudoba,   J.,  Cech,   J.S.,  Farkac, J., Grau,   P.  Contro  of  activated  sludge
     filamentous bulking ;  Experimental  verification of a  kinetic selection  theory. Wat.
     Res, 19: 191, 1985.
9.   Bull, AT., Slater,  J.H. In: AT. Bull, J.H.  Slater feds.), Microbial Interaction and
     Communities, Vol. 1., Academic Press, New York, NY, 1982.
10.  D'Adamo,  P.C., Rozich, A.F., Gaudy, A.F., Jr. Analysis of growth data  with
     inhibitory carbon sources. Biotech Bioengg. 26: 397,1984.
11.  Tabak,  H.H., Quave,  S.A., Mashi,  C.I., Barth, E.F. Biodegradability studies  with
     organic  priority pollutant compounds. JWPCF 53: 1503,1983.
12.  Larson,   R.J.,  Perry,  R.L.   Use  of  electrolytic  respirometer   to  measure
     biodegradation in natural waters. Wat. Res. 15: 697,1981.
13.  Paris, D.F.,  Rogers, J.E. Kinetic  concepts for measuring  microbial rate constants
     : Effects of nutrients on rate constants. Appl. Environ. Microbiol. 51: 221,1986.
14.  Young,  J.C.,  Baumann, E.R. The electrolytic respirometer —  II —  Use in water
     pollution control plant laboratories. Wat. Res. 10: 1141,1976.
15.  Tabak,  H.H.,  Lewis,  R.F.,  Oshima,  A. Electrolytic respirometry  biodegradation
     studies,   OECD   Ring  test   of  respiration   method   for   determination   of
     biodegradability. Draft  Manuscript,  MERL,  U.S. Environmental  Protection Agency,
     Cincinnati, OH, 1984.
16.  Dosanjh, M.K.,  Wase, D.A.J. Oxygen uptake  studies on various sludges adapted to
     a waste containing chloro—,  nitro—, and amino—substituted xenobiotics. Wat.  Res.
     21: 205,1987.
17.  Howard, P.H.,  Banerjee,  S. Interpreting results from  biodegradability  tests  of
     chemicals in water and soil. Environ.  Tox. Chem. 3: 551, 1984.
18.  Simkins, S., Alexander, M.  Models for mineralization  kinetics with  variables  of
     substrate  concentration and population density. Appl.  Environ.  Microbiol.  47:
     1299,1984.
19.  Dojlido, T.R. Investigations of biodegradability and  toxicity of organic compounds.
     EPA-600/2-79-163, U.S. Environmental Protection Agency, Cincinnati, OH, 1979.
20.  Hansch, C. On the structure of medicinal chemistry. J. Med. Chem. 19:1, 1976.
21.  Martin,  Y.  C.  Quantitative  Drug Design  :  A  Critical  Introduction,  Dekker,
     NewYork, 1978.
22.  Free, S. M., Wilson,  J. W. A  mathematical  contribution to  structure-activity
     studies. J. Med. Chem. 7:395,1964.
23.  Fujita,  T.,  Ban, T. Structure-activity studies of phenylamines as  substrates  of
     biosynthetic enzymes of sympathetic transmitters. J. Med. Chem. 14:148, 1971.
24.  Mayer,  P. P. Drug Design, Ed.,  E. J.  Ariens, Medicinal Chemistry, Vol.  IX,
     p.187, Academic Press, NY, 1980.
25.  Weiner, M.  L., Weiner, P.  H. A  study of  structure-activity  realtionships  of  a
     series of diphenylaminopropanols by factor analysis. J. Med. Chem. 16:655, 1973.
26.  Hansch, C., Leo, A. J. Substituent  Constants for Correlation  Analysis in
     Chemistry and Biology, Wiley (Interscience), NY, 1979.
27.  Dunn,  W.  J.,  Ill,   Wold, S.  Structure—activity analysed by  pattern  recognition :
     The asymmetric case. J. Med. Chem. 23:595,1980.
28.  Verloop, A.,  Hoogenstranten, W.,  Tipker, J.  Drug  Design, Ed., E. J. Ariens, Vol.
     VII, p.165, Academic Press, NY,  1977.
29.  Kier, L. B.,  Hall,  L.  H. Molecular Connectivity in Chemistry and Drug Research,
     Academic Press, NY, 1976.
30.  Blankley,  J.  C.  Quantitative  Structure-Activity  Realtionships  of  Drugs,   Ed.
                                          308

-------
     Topliss, J. G., p.l, Academic Press, NY, 1983.
31.  Geating,  J. Literature study of the biodegradability of chemicals in  water.  EPA—
     600/2-81-175, U.S. Environmental Protection Agency, 1981.
32.  Paris,  D.  F., Wolfe,  N.  L.  Realtionships between  properties of a series of anilines
     and their transformation  by bacteria. Appl. Environ. Microbiol. 53:911,1987.
33.  Paris,  D.  F.,  Wolfe, N. L., Steen,  W.  C.  Structure-activity relationships  in
     microbial transformation  of bacteria. Appl. Environ. Microbiol. 53:971,1987.
34.  Paris,  D.  F.,  Wolfe, N.  L.,  Steen,  W. C.,  Baugham, G. L.   Effect of  phenol
     molecular  structure  on  bacterial transformation rate constants in pond and river
     samples. Appl.  Environ. Microbiol. 45:1153,1983.
35.  Wolfe, N.L., Paris,  D.F., Steen,  W.C., Baugham, G.L. Correlation  of microbial
     degradation rates with chemical structure. Environ. Sci. Tech.  14: 1143, 1980.
36.  Paris,  D.  F.,  Wolfe, N. L., Steen, W. C.  Microbial  transformation  of ester  of
     chlorinated carboxylic acids. Appl. Environ. Microbiol. 47:7,1984.
37.  Banerjee,  S.,  Howard, P. H., Rosenburg, A. M.,  Dombrowski,  A.  E., Sikka,  H.,
     Tullis, D.  L.  Development of a general kinetic model for  biodegradation and its
     application  to  chlorophenols and related compounds.  Environ. Sci.  Tech. 18:416,
     1984.
38.  Vaishnav,   D.   D.,   Boethling,   R.   S.,   Babeu,  L.   Quantitative   structure—
     biodegradability  relationships  for  alcohols,   ketones   and  alicyclic   componds.
     Chemosohere. 16:695, 1987.
39.  Pitter, P.  Correlation between the structure of aromatic compounds and the rate
     of their biological degradation. Collection Czecoslovak Chemical  Comm.  49:2891,
     1984.
40.  Dearden,  J.C.,  Nicholson, R.M.  The prediction of biodegradabilities by the  use of
     quantitative  structure—activity  relationships  : Correlation  of  biological  oxygen
     demand with atomic charge difference. Pestici. Sci. 17: 305, 1986.
41.  Govind, R., Treatability of toxics in wastewater systems. Hazardous Substances 2:
     16,1987.
42.  Boethling,  R.S.  Application of molecular  topology  to quantitative  structure—
     biodegradability relationships. Environ. Tox. Chem.  5: 797, 1986.
43.  Urano,  K.,  Kato,  Z.  Evaluation  of  biodegradation  ranks  of priority organic
     compounds. J>  Hazardous Mtl. 13: 147, 1986.
44.  Urano, K., Kato, Z.  A method to classify biodegradabilities of organic  compounds.
     J. Hazardous Mtl. 13:135, 1986.
                                            309

-------
TABLE 1. DATA USED TO CALCULATE GROUP CONTRIBUTION PARAMETERS

Compound
Ethyl alcohol
Butyl alcohol
Ethylene glycol
Acetic acid
Propionic acid
n— Butyric acid
n— Valeric acid
Adipic acid
Methyl ethyl ketone
Hexamethylenediamine
n— Hexylamine
Mono ethanol amine
Acetamide
Benzene
Benzyl alcohol
Toluene
Acetophenone
Hydroxybenzoic acid
Aminobenzoic acid
Aminophenol
Ln(k)
-2.9004 ~
-3.0791 "
-3.3524 "
-2.6037 "
-2.6736 "
-2.7031 ~
-2.6310 -
-2.8134 "
-3.4738 ~
-4.3428 "
-2.8647 "
-3.3242 "
-2.9957 ~
-2.8647 "
-2.7806 ~
-2.6037 "
-3.1701 "
-2.3538 "
-2.6592 "
-3.2442 "

-3.1465
-3.3242
-3.6496
-2.7181
-2.9759
-3.0576
-2.6593
-3.1235
-3.6889
-4.5099
-3.0576
-3.3814
-3.0576
-2.9759
-3.1701
-2.8647
-3.5404
-2.9188
-2.9374
-3.2968
Average Ln(k)
-3.0159
-3.1942
-3.4900
-2.6593
-2.8134
-2.8674
-2.6451
-2.9565
-3.5755
-4.4228
-2.9565
-3.3524
-3.0262
-2.9188
-2.9565
-2.7257
-3.3382
-2.6363
-2.7984
-3.2702
                             310

-------
        TABLE 2. GROUP CONTRIBUTION PARAMETERS
  Group
Methyl
Methylene
Ketone
Amine
Acid
Hydroxy
Aromatic CH
Aromatic Carbon
CH3
CH2
CO
NH2
COOH
OH
ACH
AC
-1.3667
-0.0438
-0.5073
-1.4654
-1.3133
-1.7088
-0.5016
1.0659
TABLE 3. COMPARISON OF ACTUAL AND PREDICTED Ln(k) VALUES

Compound
o— Cresol
m— Cresol
p— Cresol
Phenol
2,4— Dimethylphenol
2— Butanone
Acetone
Butylbenzene
1— Phenylhexane
Actual
Ln(k)
-2.6878
-2.3694
-2.4647
-3.0006
-2.8460
-3.1326
-3.1161
-3.1285
-3.3971
Predicted
Ln(k)
-2.9501
-2.9501
-2.9501
-3.1509
-2.7493
-3.2845
-3.2407
-2.9402
-3.0278
Error %
41.71
27.94
57.10
42.28
13.13
50.57
13.86
28.78
85.31
                         311

-------
           PRELIMINARY RESULTS ON THE ANAEROBIC/AEROBIC
      BIOCHEMICAL REACTOR FOR THE MINERALIZATION OF ORGANIC
                 CONTAMINANTS BOUND ON SOIL FINES
               Robert C. Ahlert, PhD, PE, Dist. Prof.
               David S. Kosson, PhD, Asst. Prof.
               William V. Black, Graduate Student

               Chemical & Biochemical Engineering
                       Rutgers University
               P.O. Box 909, Piscataway, NJ 08855
                         ABSTRACT

The goal of the overall research program, a part of which is
discussed in this paper, is to demonstrate a sequence of
aerobic/anaerobic microbial process steps for degradation of
contaminated soil fines and slurries of soil fines.  Toward this end,
it must be possible i) to assay individual organic species and total
contaminant organic carbon in soils of varying properties, ii) to
separate whole soils into fractions according to particle size, and
iii) to assay [as in i)] reactor slurries containing suspended soil
particles, microbial culture and dissolved, dispersed and sorbed
organic contaminants and metabolites.  These techniques are required
to define the nature of the contamination, devise operating
conditions to facilitate microbial contact, and assure complete
mineralization of target contaminations and "clean" residuals.

The first major section of this paper describes the development of
analytical methodology for whole soil and soil fractions; in
parallel, techniques for mixing/homogenizing, fractionation and
extraction used in sample preparation are discussed.  It has been
possible to separate soil fines and some "bulk" organic matter.  A
large part of the total organic chemical contamination is due to
sorption and physical pore interactions with the fine particle
fractions [clay minerals and humic substances] of whole soil.

A second section describes microbial degradation experiments.
Systems and procedures for microbial reactions were designed and
implemented to accomodate the properties and behavior of target
substrates.  Both shake flask and fermentation reactions are being
carried out on slurries of soil fines.  Low molecular weight
polynuclear and chlorinated aromatic hydrocarbons are readily
biodegraded.
                                  312

-------
INTRODUCTION

Distillation bottoms and sludges from benzene-toluene-xylenes [BTX]
production were impounded for several decades.  The production process
consisted of the catalytic cracking of naphtha, in the presence of fuming
sulfuric acid, and distillation.  Therefore, lagoon contents include
naphtha-related compounds, distillate residues and compounds resulting from
reactions of these species and sulfuric acid.  Possible contaminants
include, but are not limited to, simple aromatic species, polynuclear
aromatic hydrocarbons [PAHs], phthalates, as well as sulfonated derivatives
of these compounds.  Many of these species have slight solubility in water
and/or an affinity for some of the constituents of soil and have migrated
into and through the soil immediately adjacent to the lagoon.

During this study, soil samples were obtained from the containment area
surrounding the lagoon. The impoundment has been designated a CERCLA-NPL
site; it exceeds ten acres in extent and contains an estimated 100,000 cubic
yards of residues.  The contents of the lagoon have separated into several
distinct layers that include, in bottom-to-top sequence, a solid mixture of
organic and inorganic substances, a tar-like layer, a layer of viscous
organic matter, and a floating aqueous layer.

ANALYTICAL APPROACH

Initially, soil samples are mixed and homogenized.  The resulting material
is air-dried and sieved through a 3-cm brass screen to remove debris, rocks
and gravel; this procedure also breaks up macro-agglomerates.  A second
sieving, with a 5-mm screen, improves homogeneity, enhances mixing and helps
toward analytical reproducability.  Direct solvent extractions of
homogenized, contaminated soil utilize methanol, cyclohexane, or methylene
chloride.  Methanol has relatively high polarity, cyclohexane is a model
cyclic compound, and methylene chloride is a moderately polar, volatile
compound with broad solvent capabilities.

Gas chromatography [GC] is used to identify and quantify compounds in soil
extracts.  USEPA Test Methods 602 and 610 are employed in these analyses.
Method 602 is used to assay aromatic species in GC column effluent with a
photoionization detector [PID] in series with an electrolytic conductivity
detector [ELCD].  Method 610 is employed to detect PAHs and phthalates,
utilizing a flame ionization detector [FID].  Standard solutions are assayed
in sequence with solvent solutions to match retention times for compound
identifications.  Peak areas are used to construct standard curves,and
provide a basis for determination of contaminant concentrations.  Compound
concentrations are calculated from both PID and FID output to check
analytical consistancy.

Experimental Methods

Soil contaminant levels were initially estimated to fall between 2 and 5 %
on a dry weight basis.  Direct solvent extractions are carried on varying
masses of soil with the goal of limiting contaminant concentrations to
about 100 mg/L in extract solutions.  This target concentration was adopted
                                    313

-------
to  avoid  overloading  the  detectors.  Soil masses varying from 0.15 to 0.38 g
are extracted with  a  fixed  volume of solvent.

Duplicate amber  serum bottles, each with label, septum and aluminum cap, are
weighed with a Mettler PE3600 balance.  Bottles are 100 ml in volume  Soil
is  added  until target weights are attained; approximately 50 ml of methanol,
cyclohexane or methylene  chloride is added to both bottles.  Bottles are
sealed with the  septa [Teflon-coated neoprene] and reweighed.  Experimental
errors include the  small  discrepancies in obtaining target soil masses and
measuring solvent volumes.   In general, these are accounted for in
concentration calculations.  A second form of the experiment is carried out
to  facilitate compound identification.  It is the same in all respects,
except that 20 g of soil  are added; since quantification is not desired,
duplicates are not  performed.

Extraction vessels  are shaken for approximately one hour.  This time was
found to  be adequate  in earlier studies; however, it assumes that only
readily reversible, high-rate sorption processes are involved.  Higher
energy binding processes  and sorbate trapped by capillary forces would not
participate in such short-term partitioning.  Extract solutions are filtered
through 0.2-um MSI  Cameo  II 25-mm disposable syringe filters, into duplicate
5-mL vials.  Samples  are  stored at 4°C, to minimize volatilization losses,
and enclosed to exclude light and avoid photolytic chemical reactions.

Method 602 utilizes a 1.8-m long by 2-mm ID stainless steel GC column packed
with 100/200 mesh Supelcoport, coated with 5 % SP-1200 and 1.75 % Bentonite-
34.  Oven temperature is  held at 50°C for 2 min; a 6°C ramp takes the oven
to  90° for a final  period of 23 min.  The PID has detection limits of 1 to
10  pg for unsaturated carbon bonds found in aromatic compounds.  The ELCD
has detection limits  of 0.1 to 1 pg when used in the detection of
chlorinated compounds or  to verify PID results.  Sample size is 2 uL.
Method 610 is applicable  to PAHs and phthalates.  It utilizes a 1.8-m long
by  2-mni ID glass GC column packed with 100/200 mesh Chromosorb W-AW-DCMS,
coated with 3 % OV-17.  Oven temperature is held at 100°C for 4 min; a 8°C
ramp takes the oven to a  final temperature of 280°C.  The FID has detection
limits of 10 to  100 pg.   Sample size is 5 uL.

Chemical  Oxygen Demand [COD] is determined for some contaminated soil
samples.  This procedure  is identical to that described in Standard
Methods".  The COD  has some value for comparison with carbon in identified
species,  to estimate  extraction and identification efficiencies.

Analytical Results

Extracts  generated  in  Experiment 12888 were distinctly different in color.
After filtration, the  cyclohexane extract was translucent orange, methylene
chloride  gave an opaque brown liquid, and the methanol solution was clear
and tan.  This appeared to be evidence for susbtantial variation in
extraction efficiency.  In addition, the high polarity of methanol leads to
destruction of soil aggregates; thus, methanol is capable of extracting
contaminants held in  micro- and macro-pores by capillarity and interfacial
tension.  Solvents can be compared on the basis of the mass of naphthalene
                                     314

-------
extracted.  No naphthalene was extracted by cyclohexane; methanol and
methylene chloride extracted 1,089 and 1,403 mg/kg dry soil, respectively.
Thus, for naphthalene, solvent power varied considerably.

The PID sees cyclohexane and impurities in methylene chloride.  The ELCD
detects chlorinated compounds and is overloaded by methylene chloride
solutions.  The consequence is chromatogram baseline fluctuations and large
peak area integration inaccuracies.  Methanol was the only solvent suited to
Method 602.  Methylene chloride was used in conjunction with Method 610.
Eight major organic compounds were identified and quantified in methanol
solution.  In order of decreasing concentration [mg/kg], they are:
naphthalene - 1,090; 1,2-dichlorobenzene - 360; toluene - 150; xylene
isomers - 145; benzene - 113; and, ethyl benzene - 28.  The eight species
account for 64 % of the total peak area of chromatographic responses.

Benzene, toluene and the xylenes are primary products of naphtha
distillation.  GC residence times for standard solutions and extract
solutions, with PID detection, varied less than 0.004 sec for this group.
Naphthalene is also a major component of naphtha; residence times differed
by 0.004 sec.  Ethylbenzene [EB] is formed by catalytic reaction of benzene
with ethylene, an olefin found in industrial naphtha. In BTX production,
sulfuric acid is the catalyst.  EB residence times varied by 0.008 sec.

The appearance of 1,2-dichlorobenzene was signaled by the PID and verified
by the ELCD, a halogen-specific detector.  Chlorinated compounds are not
normally found in naphtha nor are they produced by sulfuric acid catalysis.
The presence of this compound may indicate disposal to the impoundment from
another manufacturing source or a spill clean-up activity.  Two substantial,
unidentifiable peaks were encountered with Method 610; neither was observed
with Method 602.  These peaks correspond to compounds that are believed to
be sulfonated aromatic hydrocarbons; operating temperatures for Method 602
preclude elution of such higher boiling species.  Naphthalene sulfonic acids
represent compounds of higher molecular weight and boiling point, requiring
increased GC oven temperatures  [Method 610] and extended residence times.
Standard  solutions
[see Table  la]  and
solvent  [see Table
Method 610  for  the
Similarly,  Figures
purgable  aromatic
is  the naphthalene
 included seven purgable aromatic compounds in methanol
 fifteen PAHs in a 50:50 methanol:methylene chloride mixed
 lb].  Figures 1 and 2 are gas chromatograms obtained by
 PAH standard solution and methanol  extract, respectively.
 3 and 4 are GC outputs obtained by Method 602 for the
standard and the methanol extract, respectively.  Figure  5
 case run with a modified version of Method 602.
 COD  has  been  measured  for  several  soil  samples.  This  assay  is  used  to
 determine  the total  oxygen required  to  fully  oxidize all  reduced  species;  it
 does not distinguish between  contaminant  organic carbon  and  hydrogen, soil
 organic  matter,  and  metals in reduced or  partially  oxidized  states.  The COD
 of sieved,  air-dried,  unextracted  soil  corresponds  to  about  50  mg C/kg  soil.
 In comparison,  the total organic carbon [TOC]  associated with the eight
 quantified contaminant species cited above  is  approximately  1,6 mg C/kg soil
 or 3.2 % of the COD  carbon equivalent.  This  is not a  reflection  of
 extraction efficiency. However, CODs run  on soil samples taken  at points
                                    315

-------
remote to the disposal lagoon average close to 50 mg C/kg soil.  Thus,
extraction efficiency with methanol is probably relatively good.
                              Table la

              Purgable Aromatic Standard Mixture [602-M]
                            (in Methanol)
              Compound

              Benzene
              Toluene
              Ethyl benzene
              Chlorobenzene
              1,2-Di chlorobenzene
              1,3-Di chlorobenzene
              1,4-Di chlorobenzene
Concentration (mg/L)

      2000
      2000
      2000
      2000
      2000
      2000
      2000
                              Table Ib

      Polynuclear Aromatic Hydrocarbon Standard Mixture  [610-M]
                fin 50:50 Methanol:Methv1ene Chloride)
              Compound

              Acenaphthene
              Fluoranthene
              Naphthalene
              Benzo(a)anthracene
              Benzo(a)pyrene
              Benzo(b)f1uoranthene
              Benzo(k)f1uoranthene
              Chrysene (93 %)
              Acenaphthylene
              Anthracene
              Benzo(g,h,i)pyrene
              Fluorene
              Phenanthrene
              Dibenzo(a,h)anthracene
              Indeno(l,2,3-cd)pyrene
              Pyrene
Concentration (mq/L)

      1000
       200
      1000
       100
       100
       200
       100
       100
      2000
       100
       200
       200
       100
       200
       100
       100
                                    316

-------
                         Method 610
figure i
            6.2S
                                                                   13  7.62
                                     ftuorene.
                        12.72. Phenanthrene and
          !
                                16.86
                      16.58
20-
             18.48




             19.4S
                            . 17 8en20 (a)anthracene and
            22.42
                                        ffTamb)fiuoranth&ne4i\d Senze
-------
 5--
10-
15-
20-
                                      Method 610

                                             2
               ,  16. 69
               )  17.03
               P  17,53
                  J&.20
                    .^*
                  >Z1*ft»

                      •8«2^
                   23. 3Z
                                          318

-------
   seoo-
 O
 -P
 U
   «<">-
1
   2JOO-
    JOO
   9200
                   Chromatograph  for Standard
                 Solution of Purgable Aromatics

                          Method 602

                           Figure 3
                                                    Ch lorobenzene
                                                    1 , 4-dichlorobenzene
                                                    1 ,3 -diehlorobenzene
                                                    1 ,2-dichlorobenzene
   6ZOO-
 S-i
 O

 O
 JJ

Pi

Q
M
PH
   zeno-
    zon
                                Benzene
                                Toluene
                                Ethyl benzene
                                Chlorobenzene
                                              -1,4-dichlorobenzene
                                              -1,3-dichlorobenzene
                "T"
                 T
                        'V _______
                           18
                                                       1 , 2-dichlorobenzene
                                    'T'
T"
"I"
 3£
T"
                                             319

-------
C37S-
                                    Hethanol  Extract

                                      Method 602

                                       Figure  4
 tn
 O
 4J
 O
 \.
                          ~r
                           10
                                          .-'~\--.
                                                      Naplithalene —-'"
                                                      1 , 2-dichlorobenzene
                                                          25

                                                                           I
                                                                           55
                                            320

-------
  3SBII -i
o rsn-
4J
o
0)

0)
o  .,
                      Naphthalene by Modified

                            Method 602


                             Figure 5
                                                             Naphthalene-
                                                              	j
   B'iWI-l
 U
'QJ
 4J
 (1)
                           .f


                           1C
"1	"
15
                                             	"1	
                                                  20


                                             Standard
r
T"
 4.0
                                                           Naphthalene
       :i
                                                                A.
                 5
                            14
                                      ._,	r—=_n.
                                       IS          20    •   •   ZS
                                         Methanol Extract
                                                                       T
                                                                       30
T-
                                              321

-------
 MICROBIAL  APPROACH

 The  solubilities  of PAHs  in  aqueous  media  decreases with  increasing
 molecular  weight;  for  four rings  or  more,  saturation  concentrations  are very
 low  and difficult  to measure.   Within  this context, rates of  aerobic
 metabolism of  PAHs are insignificant.   In  addition, this class of aromatic
 compounds  has  a high affinity  for several  constituents of natural soils,
 i.e., soil  organic matter [humic  substances] and clay minerals.  PAHs are
 lipophilic,  i.e.,  prefer  sorptive association and/or  dissolution in  organic
 rather than  hydrophilic phases.   In  addition, many multi-ring molecules can
 migrate into stable positions  inside clay  mineral structures, either
 diffusing  between  laminae in mica-like  structures or  entering the crystal
 lattice directly.   The latter  process  is a form of clathration.  These
 preferred  soil components are,  in general,  the smaller particle fractions of
 the  soil system.   Thus, it is  possible  to  introduce PAHs into a microbial
 systems in slurry  form, with substrate(s)  bound to particulate matter.

 A slurry form  of bioreactor has a number of possible  thermodynamic phases
 present, including:  aqueous medium with dissolved substrate and nutrients,
 substrate  bound to soil particles, suspended single and clustered cells,
 substrate  emulsions  and colloids,  and cells attached  to the exterior and
 macro-pore surfaces  of dispersed  particles.  Sampling and analytical
 methodology  are difficult throughout the design and implementation of
 biodegradation experiments.

 An earlier Section dealt  with  the problems  of accurate chemical contaminant
 identification and quantification for soils prior to  chemical or biochemical
 reaction.  Often tightly-bound  PAHs must be extracted from a soil mass, with
 uncertain  efficiency or recovery.  Method  verification is critical.  Extract
 solution is  assayed  by GC or High Performance Liquid  Chromatography  [HPLC].
 Multiple phases must be sampled and extracted, during and after biochemical
 transformations.   The  possible  appearance  of intermediate or final
 metabolites, i.e.,  incomplete mineralization, adds to the complexity of mass
 balances for substrate species.   COD is a  useful tool for quasi-continuous
 monitoring of  the  progress of  slurry-type  bioreactor  systems.  In well-
 aerated aqueous systems,  metals are not a  serious factor. However, variable
 cell mass and  natural  soil organic matter  severely limit this approach.

 Before slurry  reaction can be undertaken, whole soil must be reduced to
 several particle size  fractions.   Contaminant PAHs favor finer particles,
thus, fractionation  is a  useful way of concentrating these compounds prior
to degradation experiments.   A reproducible method for soil  classification
has been developed and has been demonstrated to lead to the desired
concentration of target substrates, as follows.   Whole soil  is separated
 into three phases: a tar-like organic phase, a coarse sandy phase and an
aqueous suspension of  soil fines.  The latter is suited to slurry reactor
experiments.  Experiments consist of small-scale studies in shake flasks,
performed in matrix  formats,  and  fermentation studies in reaction vessels  of
larger volume.
                                    322

-------
Background

The rates of microbial assimilation of PAHs have been demonstrated to be
functions of solubility, molecular weight, number of six-member rings,  ,
degree and type of substitution, as well as environmental conditions, such
as temperature, pH and oxygen concentration.  The solubility of
unsubstituted PAHs,in water, drops sharply as the number of rings increases.
It rapidly diminishes to levels that are too low to support significant
biological activity; see Table 2 for data.  Compounds of six or more rings
have vanishing solubility in water.

The number and type of substituents on or in a PAH molecule have a marked
influence on solubility.  The solubilities of phenols, nitrogen
heterocyclics, polynuclear polyols, sulfonates and other mono- and poly-
substituted PAHs are often significantly higher than the basic hydrocarbons.
Therefore, substituted compounds are more likely to be observed as solutes
in contaminated groundwater.  Also, surfactants increase PAH solubility.
However, these compounds complex or "react" with the high molecular weight
polynuclear species to create a composite hydrophilic exterior.  The result
is either a stable emulsion, colloidal suspension or micro-dispersion; it
cannot properly be classed as dissolution in the thermodynamic sense of a
homogeneous liquid phase.  Sodium  laurylsulfate increases the solubility of
2- to 7-ring PAHs by  2 orders-of-magnitude or more.

Biodegradation of 2-  to 3-ring  PAHs by pure microbial cultures has been
demonstrated;  naphthalene, phenanthrene and anthracene have been shown to be
assimilated quantitatively.  Higher molecular weight compounds, i.e.,
benzo(a)anthracene  and  benzo(a)pyrene, can  be degraded to simpler
intermediates  in the  presence of  supplementary carbon sources or
cometabolites,  i.e.,  biphenyl and  succinate.

Bacteria concentrate, grow,  and form  bioslimes  in  aqueous boundary layers at
liquid/liquid  and  liquid/solid  interfaces.  Organic  cosolvents can transport
PAHs  to  such  interfaces and  increase  the  rate of biodegradation.  There  is
no information to  support  microbial metabolism  of  solid  PAHs;  similarly,
there is little  data  on the  bioreaction of PAHs  sorbed on nonreacting
 surfaces.
       Compound
                    Table 2. Solubilities of PAHs in Hater
Mol. Wt.
Naphthalene
Acenaphthene
Fluorene
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo(a)pyrene
Dibenzanthracene
128
154
166
178
178
202
202
252
278
Solubility (ug/L)

     31,700
      3,200
      2,000
      1,300
         73
        260
        140
          4
# of Rings

    2
    3
    3
    3
    3
    4
    4
    5
    5
                                     323

-------
 Experimental

 i) Apparatus

 Aerobic biodegradation  experiments are carried out in 60-mL Ehrlenmeyer
 flasks, on a laboratory shake device, or in 3-L  [working volume]
 fermentation vessels.   The shaker studies are arranged in matrix format with
 the following common composition: 30 ml soil slurry, 20 ml inoculum and 10
 ml nutrient medium  [see Table 3].  The composition of the inoculum is varied
 to provide a basis  for  the evaluation of substrate volatility losses,
 reactor surface wetting and sorption onto biomass as mechanisms of substrate
 disappearance.  The 20  ml "inoculum" is live culture; in sterile controls,
 the 20 ml of seed is replaced by autoclaved culture or deionized water.
 Controls are intended to illustrate the extent to which volatilization and
 inorganic surface sorption influence sub-strate fate.  Given that biomass
 rendered unviable by autoclaving retains substantial sorption capacity,
 autoclaved culture  is designed to investigate this loss pathway.  Also, this
 control provides a  zero-time or baseline measurement for carbon or oxygen'
 demand.  The pH of  the  composite solution is adjusted to 7.15 by addition of
 a mixture of solid  potassium monobasic and dibasic phosphates.

 Fermentation studies are carried out with a working volume of 3 L prepared
 in the same ratio of 3:2:1 for soil  slurry:inoculum:nutrient medium as in
 the shake flask studies.  The reactor is sparged with air at the rate of 5
 to 6 L/min; it is stirred at 300 rpm.  The pH is maintained at 7.15 by
 periodic additions  of 0.25M sodium hydroxide regulated by a pH controller.
 Samples are taken at 24-hr intervals, to monitor the course of reaction.
Separate studies utilizing deionized water are used to evaluate
volatilization losses.
                     Table 3 - Medium Composition

                     Constituent    Concentration [ma/Li
                     (NH4)2S04

                     MgS04.7H20

                     FeCl3.6H20

                     MnS04.H20

                     CaClo
1,500

  100

  0.5

   10

  7.5
                                    324

-------
  ii) Soil Slurry

Soil slurry is prepared by a sequence of homogenization, extraction, and
fractionation steps.  This procedure is designed to create a suspension of
soil fines that displays minimum variation from batch-to-batch.   Whole soil
samples are homogenized by passing the air-dried material through a 5-mm
screen, quartering the resulting solids cone through the apex, segregating
the quarters, and sieving each quarter to form a new cone. Soil is sieved
three times.  A prescreening with a 3-cm sieve removes rocks and
miscellaneous debris and, also, serves to break-up larger clumps of packed
soil.

Homogenized soil [84 g on a dry basis] is extracted with 350 ml of water at
pH 7.  Extraction separates the contaminated soil into three phases: a tar-
like [smell, sticky, viscous, etc.] organic phase corresponding to 0.65 -
0,75 % of the initial dry mass; a mixture of larger, heavier particles
[sand]; and, an aqueous supension [slurry] of fine soil particles.  The
aqueous suspension  is separated from the settleable solids by screening
through a 10-micron sieve; filtered solids are washed with 650 ml of water
to dilute the filtrate slurry to 1 liter and the final concentration.  The
fractionation procedure has been found to retain approximately 65 % of the
initial dry mass of soil on the sieve.  The final slurry of fine particles
is stable and does  not show any evidence of settling under experimental
conditions.

iii) Inoculum

The  aerobic  inoculum is obtained as waste sludge from the Somerset/Raritan
Valley Sewage Authority.

iv)  Nutrient Medium

The  nutrient medium is prepared from a conventional recipe for aerobic
cultures  and does not employ a primary or supplementary carbon source; see
Table 3.  The soil  slurry supports microbial activity without the use of a
co-substrate; none  is added.

v) Analytical

The  progress of substrate conversion is monitored by COD and GC analyses.
Samples are  prepared for the latter by extraction with methylene chloride at
8 ml of solvent for 20 ml of slurry.  The remainder of the assay is carried
out  in accordance with USEPA Method 610.  A 1.8-m long by 2-mm ID glass
column is packed with a  stationary phase consisting of  100/200 mesh
Chromosorb  W-AW-DCMS coated with 3 % OV-17.  Oven temperature  is held at
100°C  for 4 min; a  8°C/min ramp  increases the temperature to  280°C.  An FID
is  used to  determine residence times and peak areas.  Gas pressures are 70,
40  and 65 psi for nitrogen,  hydrogen and dry air, respectively.  A  5-uL
sample is injected  into  the  GC.
                                     325

-------
   Results

   Table 4  summarizes  moisture contents  and  COD  analyses  for  whole  soil
   and  the  fractions generated by repeated sieving,  fractionation and
   extraction.   COD determinations are referred  to  1  kg of  air-dried whole
   soil. The COD balances  vary about 15 % for a typical  set;  this  is  a
   consequence  of an error  of at least ± 5 % in  this  measurement.   Slurry,
   diluted  with wash water,  has a COD of 7.1 g 02/L.
                 Table  4  -  Moisture  Content  and  COD

        Sample  Type            Composition  F%1      COD  Fa Oo/kol
        Whole  Soil
           Water
           Large Particles
           Soil  Fines
        Tar-like Residue
        SIurry
16
53
31
 89.6
  0
trace

 19.4
 84.5
When an active microbial  inoculum  is combined with a slurry of soil fines, a
lag period of approximately 6 hr is observed.  Acid  production in shake
flasks and fermentations, and carbon dioxide generation by fermentations,
are not observed until after the lag phase.  There is no loss of COD, as
might accompany volatilization or  sorption.  It is assumed that COD
attributable to biomass remains unchanged during the experiment, i.e.,
growth is negligible.  The reduction of COD in the flasks inoculated with
live cultures is indirect evidence for substrate mineralization, as opposed
to physical uptake  (sorption) by the biomass.  The results of a shake flask
matrix study are summarized in Table 5.  Flasks contain a working volume of
60 ml; total reaction time is 80 hr.  The slurry has an initial soil fines
concentration of 30 g/L and a COD of 7.0 g 02/L; inoculum COD is 4.75 g
02/L.

Table 6 sumarizes results of a larger scale fermentation study.  Reactor
working volume is 3 L and reaction time is 68 hr.   The slurry has an initial
soil fines concentration of 45 g/L and a COD of 7.5  g 02/L;  as in the shake
flask illustration, inoculum COD is 4.75  g 02/L.   Initial  reactor COD was
calculated to be 15.9 g 02; an experimental determination gave 18.9 g 02.
The difference is probably measurement error, due to the several  phases
present,  i.e., cells', contaminated fines, suspended tarry material  and
dissolved PAHs.   A volatilization loss study was carried out  with the
fermentor.  The COD of air sparged slurry did not change in 4 days; it
remained at 3.1  g 02/L.
                                    326

-------
                 Table 5 - Typical  Shake Flask Matrix Study
   Initial  COD [mg 02]
   Final  COD   [mg 02]

   COD Change [%]
   COD Change [%]
   {corrected for  inoculum)

   Final  Dissolved TOC
           [mg C]
           [mg C/L]
   COD Equivalent  [mg 023
      Live
    Inoculum

      310
      204

      -34
      -50
        8
      131
       18
   Autoclaved
    Inoculum

      310
      318

      nil
      nil
        8
      131
       18
      No
   Inoculum

      210
      240

      +14
      +14
        8
      133
       18
                   Table 6 - Fermentation Study
   COD [mg 02]
21 hrs

 13200
   COD Change [%]          30
   COD Change [%]          39
   (corrected for inoculum)

   Final Dissolved TOC
           [mg C]         318
           [mg C/L]       106

   COD Equivalent [mg 02] 848
46 hrs

 10200

    46
    61
                 315
                 105

                 840
68 hrs

 6900

   63
   84
                288
                 96

                768
  Figures 6 and 7 are gas chromatographs for the fermentation liquor
  after 21 hrs and at the end of  the experiment.  The characteristics of
the soil slurry without medium or culture added are described in Figure 8.
The contents of the fermentor and the original soil slurry were extracted
with methylene chloride; the volume ratio was 5:2 for aqueous
suspensionrsolvent.  Figures 6 and 7 show definite declines in the number
and size of peaks, especially those corresponding to low molecular weight
PAHs.
                                    327

-------
Figure 6:   21  Hours
                                                                (nt.n
                                                                til?.91
                                                                i •••M. 31
                                                               !«I*.*f   MM
                                                                             7.211
                                                                             t.flZ
                                                                             j.irr
Figure  7:   68 Hours
                                                     II.Jt
                                                     »•.**
                                                     '*•"
               *n« •-.

                1.972

               "tt'.t J
                •'.I »
               I?.9 9
                J.r 1
                :.i i
                •I.5-.I
                i.in
                3..-JI
 ^Figure  8:   Original  Slurry
                                                    If.11

                                                     1.1?
                                                     ]!'«•
 a»ii.ti  v*    /.jjj
ICIH.^f  vv    11.59:

ll-f.-|l«  vv    Il'll9

 n*i,:i  w    i!*,
-------
CONCLUSIONS

1) Bench-scale shake flask studies are performed with slurries of soil
   fines and mixed microbial seed.  COD, corrected for the presence of
   the inoculum growing at a trivial rate, is reduced by 50 % in
   approximately 80 hours.  Similarly, TOC is measured on settled [filtered]
   aqueous phase and remains low throughout.  The latter assays are a
   reflection of limited hydrocarbon solubilities.

2) Larger-scale fermentations are carried out in 3-liter, stirred, air-
   sparged reactors.  Inoculum and nutrient medium are mixed with slurry.
   Biodegradation is monitored by assays on samples of aqueous dispersions
   and measurement of carbon dioxide generation rates.  COD reductions
   exceed 84 % in 68 hours.

3) Whole soil is separated into size fractions to characterize contaminant
   distribution by soil constituent type and particle size.  It is possible
   to separate whole soil into larger, settleable particles [primarily sand
   and silt], slurried fines, a clarifiable aqueous phase and a bulk organic
   phase.  Recovery of initial whole soil COD in the slurry and a tar-like
   organic phase is nearly quantitative.

4) Analytical techniques have been developed and demonstrated with whole
   soil, soil fractions, slurries of fines and filtered liquids.  These
   techniques are essential to the identification of contaminant species and
   quantification of individual and total contaminant concentrations.
   Assays are necessary to define initial and intermediate conditions and to
   demonstrate that contaminant destruction [mineralization] by microbial
   reaction  is effective and approaches completeness.
                                   329

-------
                  FATE AND EFFECTS OF RGRA AND CERCLA TOXICS
            IN ANAEROBIC DIGESTION OF PRIMARY AND SECONDARY SLUDGE

                 by: Richard A. Dobbs,  2Rakesh Govind,  2Peter A. Flaherty,
                     3Thomas  L.  Crawford,  2Kaniz Siddiqui, ^arry M. Austern

                      xRisk Reduction Engineering Laboratory, United States
                      Environmental Protection Agency, Cincinnati, OH  45268

                     Department of Chemical and Nuclear Engineering,
                      University of Cincinnati, Cincinnati, OH  45221

                     3Department of Civil and Environmental Engineering,
                      University of Cincinnati, Cincinnati, OH  45221
                                     ABSTRACT
    Municipal wastewaters have been shown to contain toxic organics,  many of
which are anthropogenic.  Sorption onto solids is one of the primary means of
removal of these chemicals.  This study investigates the steady-state fate and
effects of organic priority pollutants sorbed on primary and secondary sludge
during subsequent treatment in anaerobic digesters.  The investigation
consisted of two separate studies of pollutants:  volatile organics chosen from
the RCRA list, and semi-volatile organics from the CERCLA list.  Simulating
typical digestion processes, three bench-scale, semi-continuous, complete-mix
units with a total volume of 40 liters each were operated with a solids
retention time of thirty days at 35°C.  Primary and secondary sludge were
combined in equal weight ratios prior to feeding to the digesters.
Conventional operating parameters were monitored for the one control and two
organic-spiked digesters to assess differences in performance.
    The fates of seven of the volatile compounds which showed consistent
behavior, including several chlorinated aliphatics and ethylbenzene,  were
determined by purge and trap gas chromatographic analyses.   Steady-state fates
of twelve of the semi-volatile organics, which included dichlorobenzenes,
phthalates, p-cresol and lindane, were determined by gas chromatographic/mass
spectroscopic analyses.
                                     330

-------
                                 INTRODUCTION
    Municipal wastewaters have been found to contain a number of toxic
organics (1).  In response, effluent regulations have also restricted the
particular pollutant discharge levels, thus increasing the interest in the
fate of specific toxic organic compounds in wastewater treatment processes.
One area which presents potential problems is the handling and disposal of
sludge.  Sorption of toxics onto solids is one of the fundamental removal
mechanisms of pollutants from wastewater.  Previous investigations have shown
that organic compounds tend to accumulate in sludge at concentrations several
orders of magnitude higher than the influent concentration (2,3).  One study
reported some organic component concentrations in sludge greater than 10,000
times that found in the aqueous phase (4).  A dilute toxics-laden wastewater
could potentially generate hazardous primary and/or secondary sludge, creating
more difficult and costly final disposal options such as landfilling or
incineration as a hazardous waste.
    Anaerobic digestion is often utilized in the stabilization and reduction
of wastewater treatment sludge.  A survey of ninety-eight municipal wastewater
treatment plants in the United States found that seventy-three incorporated
anaerobic digestion as a means of sludge stabilization and volume reduction
(5).  Anaerobic treatment systems offer several advantages over aerobic
treatment in terms of lower energy requirements, higher process loading, and
the potential energy recovery in the form of methane gas production.
Anaerobic processes may also control the discharge of volatile compounds to
the atmosphere via biotransformation or capture of the volatilized material
with the off gas.
    Research on the fate of toxic organic material in aerobic systems has been
extensive,  yet relatively little work has been conducted to determine the
fates of organic toxics in typical anaerobic processes.   Investigations
generally have been conducted with bench or pilot scale processes,  serum
bottles with digester sludge, and serum bottle studies with specific or
enriched anaerobic cultures.  Most studies have been conducted by spiking the
compound or compounds directly into the influent process stream or serum
bottle.  This investigation, however, attempts to better typify actual
processes by determining the fate and effects of compounds already sorbed onto
the digester feed sludge.  The goal of this study is to determine the steady-
state fates of several volatile and semi-volatile organic priority pollutants
in the anaerobic digestion of primary and secondary aerobic sludge, and to
determine the effects of these toxic sludge on the digestion process.

                              EXPERIMENTAL METHOD
AEROBIC PILOT PLANT

    The United States Environmental Protection Agency (USEPA) has evaluated
the removal and fate of selected toxic organic compounds during primary-
activated sludge treatment of municipal wastewater (6).   Volatile and semi-
volatile organics were investigated under separate studies with the same
aerobic treatment system. Two parallel pilot plants were studied: one
continuously spiked to allow for biomass acclimation, and the other
                                      331

-------
Intermittently spiked to simulate unacclimated operation.  The treatment
systems chosen as a representative municipal wastewater treatment plant design
consisted of primary clarification followed by conventional plug-flow
activated sludge treatment.  Each system was operated with a sludge retention
time of 4.0 ± 0.3 days over a period of six months.  Screened and degritted
raw wastewater from the Cincinnati Mill Creek Treatment Plant was used as feed
for the aerobic systems.  Mill Creek is a combined residential/industrial
treatment facility.
    Volatile and semi-volatile toxics studied were selected from the RCRA and
CERCLA lists of organic priority pollutants, respectively.  A miscible mixture
of the compounds listed in Table 1 (volatiles) and Table 2 (semi-volatiles)
was used to spike the raw wastewater influent to the pilot plant at 0.25 mg/L
of each compound for both studies.  The primary and secondary sludge from the
continuously spiked aerobic system were used as feed to the bench-scale      T
anaerobic digesters.

ANAEROBIC BENCH-SCALE DIGESTERS

    Three bench-scale digesters (see Figure 1) were constructed from
Plexiglass cylinders with quarter-inch thick walls.  Influent and effluent
ports were constructed from polyvinyl chloride pipe.  One-inch valves were
used at each port location to provide for sludge feeding and digester mixed-
liquor withdrawal. *A Plexiglass cup with approximate capacity of one liter
was threaded at the top of the inlet pipe to act as a funnel during digester
feeding.  A long thermistor probe was inserted through the top endplate to

        TABLE  1.   VOLATILE ORGANIC  COMPOUNDS  IN RAW WASTEWATER  SPIKE
      Acetone
      Atrazine
      Carbon tetrachloride
      Chloroform
      Cyclohexanone
      1,2-dichloroethane
      1,2-Dichloropropane
      2,4-Dimethylphenol
      2,4-Dinitrophenol  •
      Ethylbenzene
      Furfural
Methylene Chloride
Methyl ethyl ketone
Methyl isobutyl ketone
4-Ni tropheno1
Phenol
Tetrachloroethylene
Tetrahydrofuran
Toluene
1,1,1-Trichloroethane
1,1,2-Trichloroethane
Trichloroethylene
       TABLE  2.   SEMI-VOLATILE ORGANIC  COMPOUNDS IN RAW WASTEWATER SPIKE
      1,2-Dichlorobenzene
      1,3-Dichlorobenzene
      1,4-Dichlorobenzene
      1,2,4-Trichlorobenzene
      Nitrobenzene
      1,3-Dinitrobenzene
      2,6-Dinitrotoluene
      E-Cresol
      Hexachloroethane
      Hexachloro-1,3-butadiene
N-nitrosodiphenyl amine
Dimethyl phthalate
Diethyl phthalate
Di-n-butyl phthalate
Butyl benzyl phthalate
(Bis) 2-ethylhexyl phthalate
Napthalene
4-Chloroaniline
Lindane
Dieldrin
                                      332

-------
measure digester temperature.   All fittings were sealed with silicone caulking
and each unit was pressure  tested to insure air tightness.  The gas
delivery/mixing system was  constructed from one-eighth inch stainless steel
tubing and stainless  steel  fittings.   The H-shaped sparger was placed one-half
inch above the endplate  and levelled.
    On the outside of the digester,  a two-inch long,  stainless steel tube was
inserted through the  upper  endplate to act as the digester gas outlet port.
Butyl rubber tubing was  inserted over the inlet and outlet stainless steel
tubes with hoseclamps.   The tubing ran from the gas recycle outlet port to a
gas pump and from the gas pump  to the gas recycle inlet port.  A glass tee was
inserted between the  gas recycle outlet port and the pump to allow for the
escape of digester off-gases to a Tedlar gas collection bag.  A stainless
steel tee with a septum  and cap was placed in line with the gas collection bag
for withdrawal of gas samples for analysis.
                BUTYL RUBBER
                STOPPER
                TEMPERATURE
                PROBE

                STAINLESS
                STEEL TUBING
                 SPARGER
                 BUTYL
                 RUBBER SEAL
                 (TYPICAL)

                 r BRASS
                 SVAGELDK
                 VALVE
                 (TYPICAL)
                                                        GAS RECYCLE
                                                        PUMP
                                                        TO GAS
                                                        COLLECTION
                                                        BAG
                                          61 cm
SIDE
VIEW
                                        EFFLUENT
                                          PORT
                          30.5 cm —
Figure 1.  Schematic of anaerobic digester for volatile
            and semi-volatile studies.
DIGESTER OPERATION

    During startup  for both studies,  the three 40-liter digesters were seeded
with 30 liters of anaerobic biomass from the Mill Creek plant.  The digesters
were cleaned out and reseeded between volatile and semi-volatile experiments.
Initially, all three digesters were fed one liter of unspiked primary and
secondary sludge for three  or four days to achieve start-up.  After start-up,
digester 1 continued to  receive unspiked sludge from the Mill Creek plant,
while digesters 2 and 3  were fed sludge from the continuously-spiked aerobic
                                     333

-------
 pilot plant.  The feed rate of one liter per day resulted in a 30 day solids
 retention time.  The digesters were maintained at 35 + 1°C in an environmental
 room.  Off-gases were recirculated through the sparger for 15 minutes every
 six hours to keep the contents of the digesters well mixed.   Conventional
 parameter evaluations were performed biweekly to monitor digester biomass
 acclimation and the approach of steady state.  Steady state operation was
 achieved after seven weeks for the volatiles study and five weeks for the
 semi-volatiles study, after which analysis for the spiked organics began.
     The daily draw and feed proceeded after digester biomass was mixed
 sufficiently uniformly distribute the cpntents.  One liter of mixed liquor was
 purged from the effluent port and then, reintroduced through the feed port to
 insure representative sampling.  A second one liter sample was withdrawn and
 used for the analysis of conventional parameters and organic compounds.   One
 liter of feed sludge was then added to the digester through the influent port.

 FEED SLUDGE

     For the volatile organics study,  unspiked feed sludge for digester 1 were
 gathered weekly from the Mill Creek treatment plant to act as control.
 Primary sludge was drawn from the gravity thickeners and diluted to 4% total
 solids (if necessary) with clarified primary influent.   Secondary sludge was
 collected from the return activated sludge line, thickened by means of a
 perforated bowl centrifuge and diluted with the resulting centrate to 4% total
 solids.  Both unspiked sludge were then combined in equal volumes and the
 resultant feed sludge was stored at 4°C in a large carboy until needed for
 digester feeding.
     Volatile spiked feed sludge for digesters 2 and 3 were gathered daily to
 minimize component loss.   The primary sludge was collected from the clarifier
 of the continuously-spiked aerobic pilot plant.  Thickening was achieved by
 overnight settling at 4°C in completely-filled glass containers to avoid loss
 of volatiles in the headspace.  Secondary sludge was drawn from the waste
 activated sludge line of the pilot plant and thickened by overnight settling
 in glass jars at 4°C followed by centrifugation.  After achieving 4% total
'solids, the spiked primary and secondary sludge were combined in equal volumes
 and stored in glass containers at 4°C (with no headspace)  until needed for
 digester feeding.   Slight losses of volatiles occurred during processing,
 however all feed sludge were analyzed as fed to the digesters.
     For the semi-volatile organics study,  unspiked sludge were taken weekly
 from the Mill Creek facility and spiked sludge from the T&E  pilot plant.
 Unspiked primary sludge was gathered  from the gravity thickeners and diluted
 to 4% total solids,  when necessary, with primary influent.   Secondary sludge
 was gathered from the thickened secondary sludge line (after polymer addition
 but prior to gravity-belt thickening).   Return activated sludge was used for
 dilution when necessary.   Both primary and secondary sludge  were combined in a
 1:1 volume ratio to yield the final feed sludge for digester 1.
     Semi-volatile spiked sludge for digesters 2 and 3 were gathered biweekly
 from the pilot plant, as  the danger of component loss by volatilization  was
 not significant as that for the volatile organics.   Primary  sludge was
 thickened by overnight settling at 4°C in large Nalgene carboys.   Secondary
 sludge was thickened by the perforated bowl centrifuge and diluted with  the
 centrate to 4%.   Both spiked sludge were combined in equal volume ratios prior
 to storage at 4°C in a large carboy.
                                     334

-------
DIGESTER PERFORMANCE

    Conventional parameters were measured for the sludge feed, mixed-liquor,
and gas samples from each of the three digesters in order to assess digester
performance.  Digester temperature, pH, and gas production were measured
daily.  Chemical oxygen demand, alkalinity, volatile fatty acids, total and
volatile suspended solids, total Kjeldahl nitrogen, ammonia nitrogen, and
total and organic phosphorus were measured biweekly throughout the study.
Analyses of off-gases for methane and carbon dioxide content were performed
weekly.
    For organics analysis, the primary/secondary sludge mixtures used as
digester feed were analyzed daily for the volatile organics or composited
daily and analyzed weekly for the semi-volatile organics.  Mixed liquor was
collected for analysis once a week.  A portion of the mixed liquor was
centrifuged, and the centrate was collected for analysis.  For the volatiles
study, gas samples were collected and analyzed weekly.  Gas samples for the
semi-volatiles study were analyzed near the end of the testing period.

ANALYTICAL METHODS

    Conventional parameters were evaluated by EPA Methods (7).  All analyses
were performed in duplicate, when possible, for each sample.  Volatile organic
compounds were analyzed using EPA Method 601 (8) and EPA Method 602 (9).  The
analytical system was composed of a Tracer LC2 Sample Concentrator, a Tekmar
Model ALS Automatic Laboratory Sampler, and a Tracer 585 Gas Chromatograph
with a Nelson Analytical 900 Series Interface to an IBM Personal Computer AT.
Digester off-gas samples were analyzed for specific volatile organics by
injecting a 5-cc sample directly into the purging chamber with a gas-tight
syringe.  From this point, the methods used for gas analysis were identical to
those described above for the aqueous samples.
    Semi-volatile organics analysis was performed by EPA Method 1625 (10).
Analytical equipment included a Varian gas chromatograph and an Incos 50 mass
spectrometer.  Off gas was analyzed as described above.  For those compounds
not quantifiable in the spectra, gas-phase concentrations of the semi-volatile
compounds were estimated by pure component Henry's Law constants evaluated at
35°C.  The chemical component in the gas was assumed to be in equilibrium with
the mixed-liquor centrate.
    Mass balance calculations were performed for each of the sampling events
from the sludge feed, mixed liquor, mixed-liquor centrate, and gas
concentrations, multiplying by the appropriate volumetric flows and gas
productions for each digester yielded the mass flows.  The amount of component
sorbed on solids was determined by mass difference between the mixed liquor
and the mixed-liquor centrate.  Fates of the specific organics were then
determined.

                            RESULTS AND DISCUSSION
CONVENTIONAL DIGESTER PARAMETERS

    Conventional parameters are reported in Table 3 to document operation and
performance of all three digesters during the 7-week test period for the
                                      335

-------
volatile organics study.  Parameters for digesters 2 and 3 were very similar,
and are thus presented as an average for the spiked system.  Results in Table
3 indicate that all digesters functioned within the limits of normal operation
over the course of experimentation.  Of particular interest is the methane
content of off-gases from the digesters.  As shown, the difference in methane
production (per gram VSS destroyed) between the spiked and control digesters
is only 13%.  Further, for each digester the methane produced per gram COD
reduced was above the theoretical value (11) of 0.35 liters(at STP), again
signalling normal operation.  The total gas production from each digester was
also similar, with a difference of 6% between the spiked and unspiked units.
    Conventional operating parameters for the control and spiked digesters
during the 9-week test period of the semi-volatile organics study are
presented in Table 4.  As with the volatiles study, the operational variations

  TABLE 3.  SUMMARY OF ACCLIMATED DIGESTER CONVENTIONAL OPERATING PARAMETERS
                           FOR RCRA VOLATILES STUDY
  Parameter:
                         Digester
    CONTROL
              SPIKED
                    CONTROL
                                    SPIKED
Temp. (C) 35.1 ± 0.8
pH 7.34 ± 0.07
TSS (mg/L) 29,200 ± 1,000
VSS (mg/L) 13,600 ± 1,500
COD (mg/L) 22,000 ± 3,600
TEA (mg/L) 3,560 ± 100
VFA (mg/L) <50
TKN (mg/L) 895.0 ± 40.0
TP (mg/L) 51.0 ± 4.0
GAS (Lgjp/d) 14.0 ± 3.6
CH4 (L/gVSS) 0.62 ± 0.02
CH4 (L/gCOD) 0.41 ± 0.07
35.2 ± 0.8
7.32 ± 0.12
33,200 ± 2,500
15,300 ± 2,200
24,000 ± 4,000
3,610 ± 320
<50
860.0 ± 25.0
28.0 ± 9.0
13.1 ± 3.8
0.71 ± 0.11
0.39 ± 0.17
...
	
39,400 ± 3,500
24,500 ± 2,100
42,300 ± 5,700
675 ± 80
360 ± 90
200.0 ± 45.0
42.0 ± 4.0
	
	
...
	
	
42,700 ± 4,700
24,700 ± 6,400
44,000 ± 9,500
1,360 ± 515
470 ± 195
175.0 ± 25.0
22.0 ± 5.0
	
	
— — —
  TABLE 4.  SUMMARY OF ACCLIMATED DIGESTER CONVENTIONAL OPERATING PARAMETERS
                        FOR CERCLA SEMI-VOLATILES  STUDY
                         Digester
  Parameter:     CONTROL           SPIKED
                                               Feed
                                     CONTROL
                                               SPIKED
  Temp. (C)
  PH
 TSS (mg/L)
 VSS (mg/L)
 COD (mg/L)
 TBA (mg/L)
 VFA (mg/L)
 TKN (mg/L)
  TP (mg/L)
 GAS (Lsxp/d)
 CH< (L/gVSS)
 CHA (L/gCOD)
  35.1 ±
  7.08 ±
31,000 ±
17,400 ±
30,700 ±
 4,150 ±
     <50
 709.1 ±
  87.9 ±
 12.1 ±
  0.70 ±
  0.48 ±
 0.8
 0.03
 1,400
 1,100
 9,400
 150

 123.1
 16.9
2.1
 0.12
 0.25
32
17
35.2 ± 0.8
7.07 ± 0.03
       1,600
       800
       4,300
       170
800
700
27,400 ±
 3,670 ±
     <50
 569.2 ± 93.9
  23.9 ± 5.8
 10.7 ± 1.4
  0.65 ± 0.12
  0.34 ± 0.09
43,600
30,200
55,900
   560
   890
4,800
3,000
12,600
120
220
               181.0 ± 43.1
                88.2 ± 20.6
43,500 ± 5,400
28,700 ± 3,000
49,700 ± 8,900
   450 ± 140
   570 ± 170
  90.8 ± 22.3
  19.1 ± 5.7
                                      336

-------
between the two spiked digesters were slight, therefore parameters are
tabulated as an average of both.  No major operational difference between the
spiked and control digesters is seen.  Methane production was near the
theoretical value of 0.35 l(STP)/g COD destroyed for all digesters.  Average
daily gas production from the control and test units differ by 12%.  An
increase in gas production from digester 1 near the end of the test period is
attributed to an increase in COD in the control feed, while the COD of the
spiked feed and spiked units' gas production remained relatively constant.  As
with the volatiles study the semi-volatile spike does not appear to have
hindered digester operation.

FATE OF CHEMICAL COMPONENTS

    The specific fate of each compound attaining steady-state conditions is
summarized in Table 5 according to treatment mechanism.  Volatile and semi-
volatile concentrations in samples from the two spiked digesters were averaged
together in the mass balance calculations.  For both studies,  it was found
that weekly pollutant concentrations, especially in the sludge feed, varied
considerably in some instances.  Fluctuation in the feed concentration was due
to variability in the background chemical concentrations in the raw wastewater
and the method of the feed preparation.  An additional (and unquantifiable)
complication in assessing the fate of individual components is that some of
the degradation products are also components of the spike mixture (i.e.,
tetrachloroethylene to trichloroethylene).  Finally, it should also be
emphasized that accurate analytical data for specific components in an
anaerobic digester matrix are difficult to obtain.   In spite of these
difficulties,  the experimental results provide useful information on the fate
of sorbed organic toxics in anaerobic digesters.
    For the volatile organics study, eleven of the compounds originally spiked
into the raw wastewater were consistently present in the digester feed sludge.
Other compounds in the original spike were either not present in the feed or
were present sporadically or in trace quantities.  Steady-state conditions
were achieved during the test period for seven of these volatile compounds,
shown in Table 5.  Biodegradation was the most significant removal mechanism
for all of these compounds except 1,2-dichloropropane.  As would be expected,
volatilization of the components into the digester off-gas was also a primary
removal mechanism.  Four other compounds, including chlorobenzene,  chloroform,
1,2-dichloroethane, and toluene, initially accumulated in the digesters, but
all showed evidence of degradation by the end of the testing period.
    Analytical results of the semi-volatile organic analysis showed twelve
compounds exhibiting steady-state behavior.  Degradation was the predominant
removal mechanism for lindane,  dibutyl phthalate, butyl benzyl phthalate, and
2,-cresol were almost completely degraded.  The other compounds,  while
generally showing significant biotransformation,  tended to sorb or remain
sorbed on the solid material.  Results of off gas analysis showed only the
dichlorobenzenes, trichlorobenzene and hexachloro-1,3-butadiene to be
quantifiably present.  As expected from the high Kows of the components,
solubilization was not significant except for 4-chloroaniline.  As mentioned
previously, some components may biotransform into some of the other spiked
components, thus affecting the apparent fates.  Of note here is 1,2,4-
trichlorobenzene, which has been shown to degrade to 1,2-dichlorobenzene (12).
                                     337

-------
                                    SUMMARY
    The performance of the anaerobic digesters were not hindered by either the
volatile or semi-volatile organic spiked sludge.  Gas production and methane
content for spiked and unspiked systems were nearly identical.  All parameters
were within typical operating limits.  Aliphatic volatile compounds were
significantly degraded or otherwise volatilized.  Removal of volatile organics
by sorption was generally not a significant mechanism.  The semi-volatile
compounds tended to sorb onto the mixed-liquor solids or be degraded.
Solubilization and volatilization generally did not play key roles.  For both
studies, all compounds showed some evidence of degradation.  Anaerobic
digestion is a viable method of treating toxic-laden primary and secondary
sludge, however, more research is necessary to determine specific fates of
individual compounds in actual digestion processes.

                               ACKNOWLEDGEMENT
    This work was performed jointly by the USEPA and the University of
Cincinnati under cooperative agreement no. CR812939-01.

            TABLE 5:  STEADY-STATE FATES  OF SPIKED ORGANIC  COMPOUNDS
I. VOLATILES:
Fate Mechanism Average
SOL VOL SORB DEC FEED LOAD
(%) (%) (%) (%) (mg/kg)
1,1,2-Trichloroethane
1,1,1-Trichloroethane
Trichloroethylene
Tetrachloroethylene
Methylene Chloride
Ethylbenzene
1,2-Dichloropropane

II. SEMI-VOLATILES:
0
0
0
0
0
0
5
 0
 3
 8
 8
14
26
46
 0
 2
 0
 1
 3
14
10
100
 95
 92
 91
 83
 72
 39
 39
 64
 84
 37
 33
10
14
144 ± 45
61
20
25
 61 ± 14
Lindane
Butly benzyl phthalate
Di-n-butyl phthalate
2-Cresol
4- Chloroaniline
Napthalene
1 , 3-Dichlorobenzene
Bis (2-ethylhexyl) phthalate
1,2, 4-Trichlorobenzene
Hexachloro -1,3 -butadiene
1 , 2-Dichlorobenzene
1 , 4-Dichlorobenzene
0
1
1
6
31
4
3
5
3
1
4
4
0
0
0
0
0
4
10
0
5
8
14
16
2
2
3
20
34
65
61
69
66
73
66
68
98
97
96
74
34
27
26
26
26
18
16
13
490
290
270
13
12
230
320
1100
750
1100
280
275
± 140
± 190
± 140
± 6
± 3
± 40
± 110
± 700
± 320
± 560
± 70
± 60
SOL—solubilized(amount in mixed-liquor centrate)   VOL=volatilized
SORB-sorbed on solids   DEG=degraded   FEED LOAD=mg organic/kg dry wt.  solids
                                      338

-------
                                  REFERENCES
 3.
 4.
 5.
 6.
 7.
 8.
 9.
10.
11.
12.
Hannah, S. A., and Rossman, L.  Monitoring and analysis of hazardous
organics in municipal wastewater--a study of twenty-five treatment
plants.  Paper presented at the Seminar on Hazardous Substances in
Wastewater, Toronto, Ontario.  November 3, 1982.

Hannah, S. A., Austern, B. M: , Eralp, A. E.,  and Dobbs, R. A. J. WPCF.
60(7): 1281, 1988.

Dobbs, R. A., Jelus, M., and Cheng, K.  Partitioning of toxic organic ''
compounds on municipal wastewater treatment plant solids.  EPA/600/D-
86/137, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1986.

Malz, F. Heavy metals and chlorinated hydrocarbons in sewage sludge.  In;
Proceedings of the Workshop on Removal of Chlorinated Hydrocarbons and
Heavy Metals from Wastewater by Advanced Treatment Systems.  IUPAC
Commission on Water Chemistry (VI.6), Frankfurt, FRG, 1986.

Sludge handling and disposal practices at selected municipal wastewater
treatment plants.  MCD36.  U.S.Environmental Protection Agency, Office of
Water Program Operations, Washington, DC, 1977.

Bhattacharya, S. K., et al.  Fate and effects of selected RCRA and CERCLA
compounds in activated sludge systems.  Paper presented at the 15th
Annual EPA Research Symposium, Cincinnati, OH.  April 11, 1989.
EPA methods for chemical analysis of water and wastes.
U.S. Environmental Protection Agency, 1979.
EPA-600/4-79-020.
Test methods - Methods for organic chemical analysis of municipal and
industrial wastewater:  Purgeable halocarbons - Method 601.  EPA-600/4-
82-057.  U.S. Environmental Protection Agency, 1982.

Test methods - Methods for organic chemical analysis of municipal and
industrial wastewater:  Purgeable aromatics - Method 602.  EPA-600/4-
82-057.  U.S. Environmental Protection Agency, 1982.

EPA Method 1625 - Semivolatile organic compounds by isotope dilution
GC-MS.  Federal Register. 49(209): 1984.

Parkin, G. F., and Owen, W. F. Fundamentals of anaerobic digestion of
wastewater sludge.  J. Environ. Eng.  112(5): 879, 1986.

Tsuchiaya, T. and Yamaha, T.  Reductive dechlorination of 1,2,4-
trichlorobenzene by Staphylococcus epidermidis isolated from intestinal
contents of rats.  Agric. Biol. Chem.. 48: 1080, 1984.
                                      339

-------
                  FATE AND EFFECTS OF SELECTED RCRA AND CERCLA
                     COMPOUNDS IN ACTIVATED SLUDGE SYSTEMS

                                       by

                    Sanjoy K. Bhattacharya, Rao V.R. Angara

               Department of Civil and Environmental Engineering
                            University of Cincinnati
                            Cincinnati, Ohio  45221

                    Sidney A. Hannah, Dolloff F. Bishop, Jr.
                     Richard A. Dobbs, and Barry M. Austern

                      U.S. Environmental Protection Agency
                            Cincinnati, Ohio  45268
                                    ABSTRACT
     Two separate studies were conducted to investigate the removal and fate
of 28 selected RCRA compounds (0.25'tng/l of each compound) and 19 selected
CERCLA compounds (0.5 mg/1 of each compound) in conventional activated sludge
treatment.  In each study, two pilot-scale (35 gpm) activated sludge systems
(SRT: 4 days for RCRA study and 8 days for CERCLA study) were operated in
parallel at the USEPA Test & Evaluation Facility in Cincinnati, Ohio.  One
system was spiked continuously with either RCRA or CERCLA toxics to produce
an acclimated biomass; the other was spiked intermittently with the same
toxics and sampled to determine performance under unacclimated conditions.
The selected RCRA or CERCLA compounds did not cause any adverse effects on
COD and SS removals.  The concentrations of organics (RCRA study) in the air
emissions indicated that the chlorinated aliphatic solvents were essentially
volatilized into the plant air emission stream, whereas the aromatic volatile
benzenes were substantially degraded.  Additional work is planned to attempt
to reduce the analytical variability encountered in these studies.
                                    340

-------
                                  INTRODUCTION
     A survey of publicly owned treatment works (POTW) showed that concentra-
tion of priority pollutants in the influent wastewater to many of these
plants exceeded the allowable concentrations for these chemicals (1).
Petrasek, et al. (2) studied the fate of 22 toxic organics in wastewater
treatment plants.  They reported that a typical POTW significantly (up to
90%) reduced the concentrations of most of these compounds.  However certain
compounds were present in the activated sludge effluent in relatively high
(20-30 jug/L) concentration.  Hannah, et al. (3) investigated the comparative
removal of priority pollutants by six biological and physical-chemical treat-
ment processes.  They reported that activated sludge process provided the
best results.  A further review of the literature indicated that only limited
data are available for many priority pollutants.

     In this study, the removal and fate of selected RCRA and CERCLA toxic
organic pollutants were evaluated with two pilot-scale activated sludge
systems fed municipal wastewater at the USEPA's Test and Evaluation Facility
in Cincinnati, Ohio.  The 28 RCRA (semivolatile and volatile) and 19 CERCLA
(semivolatile only) chemicals spiked into the systems are shown in Table 1.
The selected RCRA and CERCLA toxics were spiked into the raw wastewater in
two separate test periods.
                   EXPERIMENTAL SYSTEMS AND TESTING APPROACH
     The twq activated sludge systems were operated at a flow rate of 35 gpm
and a hydraulic retention time (HRT) of 7.5 hours.  An operational sludge
retention time (SRT) of 4 days was used in the RCRA study period. In the
CERCLA study period, the SRT was 8 days.  Each compound was spiked at 0.25
mg/L for the RCRA study and at 0.5 mg/L for the CERCLA study.  The operating
conditions and design characteristics for the two systems used in the study
are given in Table 2.  Both RCRA and CERCLA studies were performed with an
acclimated (continuous addition of toxicants) and an unacclimated (intermit-
tent spiking of toxicants) system.  These systems were operated in parallel.

     To sample from the air space above the primary clarifier, the units were
covered and vented through a duct to the roof.  An air sweep equivalent to a
5 kilometer per hour wind was maintained over the surface of the primary
clarifier by exhausting air at 14,000 liters/min.  The aeration basin was
fitted with an air tight cover and the off-gas was also vented to the roof.
Air flow in the aeration basins averaged 5,600 liters/min.
                                     341

-------
               TABLE  1.   RCRA AND  CERCLA TOXIC ORGANIC  POLLUTANTS
         RCRA Study  Period
     CERCLA Study Period
    acetone
    cyclohexanone
    furfural
    2-butanone
    4-methyl-2-pentanone
    tetrahydrofuran
    carbon tetrachloride
    chlorobenzene
    chloroform
    1,2-di chloroethane
    1,2-dichloropropane
    methylene chloride
    tetrachloroethylene
    trichloroethylene
    1,1,1-tri chloroethane
    1,1,2-tri chloroethane
    ethyl benzene
    toluene
    total xylenes
    bis(2-ethylhexyl) phthalate
    butyl benzyl phthalate
    1,4-dichlorobenzene
    2,4-dimethylphenol
    2,4-dinitrophenol
    naphthalene
    nitrobenzene
    4-nitrophenol
    phenol
1,2-dichlorobenzene
1,3-dichlorobenzene
1,4-dichlorobenzene
1,2,4-tri chlorobenzene
nitrobenzene
1,3-di ni trobenzene
2,6-dinitrotoluene
p-cresol
4-chloroaniline
hexachloroethane
hexachlorobutadi ene
dimethyl phthalate
diethyl phthalate
dibutyl phthalate
butyl benzyl phthalate
bis(2-ethylhexyl) phthalate
naphthalene
lindane
dieldrin
     Automated analytical procedures were used for the conventional pollu-
tants (COD, BOD, NH4-N, N03-N and TKN) and 6C/MS procedures were used for
the toxic organic compounds.  RCRA samples were analyzed by a contract
laboratory (PEI Associates Inc., Cincinnati, OH).  Air samples were collected
in stainless steel canisters, and were analyzed by GC/MS.  From these data,
masses of each RCRA compound stripped during the sampling event were calcu-
lated.  Sludge and liquid samples were also analyzed by GC/MS according to
approved USEPA methods (4).  Semi-volatile RCRA compounds were extracted from
the samples using continuous liquid-liquid extraction.  Prepared portions of
the extracts were injected into the GC/MS for analysis.  The semi-volatile
CERCLA compounds were analyzed following Method 1625 (5). Analyses were per-
formed at the EPA, RREL.  Details of the analytical procedures were reported
elsewhere (6).
                                     342

-------
          TABLE 2.   OPERATING CONDITIONS AND DESIGN  CHARACTERISTICS  OF
                    THE PILOT SYSTEMS
         I.


        II.




       III.




        IV.
Design Flow
= 2.2 Us
= 191 m3/d
Primary Clarifiers - Diameter              = 2.9 m
                     Weir Diameter         = 2.8 m
                     Surface Area          =6.8 m2
                     Surface Overflow Rate =28.0 m3/m-d

Aeration Basins - L:W:D                    = 5.4 m:3.0 m:3.6 m
                  Surface Area             = 16.3 n£
                  Volume                   = 59.7 m3
                  Hydraulic Residence Time = 7.5 hrs.

Secondary Clarifiers - Diameter              = 3.6 m
                       Surface Area          = 10.4 n£
                       Surface Overflow Rate =18.4 m3/mz.d
     Three tests  (sample collection events) were performed during the RCRA
study period.   For the CERCLA study, eleven tests on the continuously spiked
(acclimated) system and 4 tests on the intermittently spiked  (unacclimated)
system were performed.


                                    RESULTS


     The presence of the spiked toxic organics in the wastewater produced no
major adverse effects on the treatment of conventional pollutants.  Average
removals of conventional pollutants in the pilot systems during the two
studies were between 94 and 97 percent for SS and between 81 and 88 percent
for COD (Table 3).  In the RCRA study period nitrification in the activated
sludge processes produced average NH4-N reductions between 76 and 81
percent.  In the CERCLA study, the NH4-N removal was between 88 and 98
percent.  The CERCLA toxics (0.5 mg/L) did interfere with nitrification in
the acclimated system.

     Substantial variability occurred in the reported results with some toxic
compounds, especially in the RCRA study period.  Table 4 lists average
measured concentrations of the selected RCRA organics in wastewater and
sludges for the acclimated system.  The difference between the concentration
of most toxics in the spiked wastewater feed and primary effluent was very
low.  The primary sludge showed enhanced concentrations of the two phthalates
and naphthalene along with reduced concentrations in primary effluent indica-
ting adsorption onto sludge solids.  Four compounds (tetrahydrofuran, 1,2-
dichloroethane, methylene chloride and 1,1,2-trichloroethane) were present in
                                      343

-------
      TABLE  3.   AVERAGE PERCENT REMOVALS  OF  CONVENTIONAL  POLLUTANTS
                 DURING THE  RCRA AND  CERCLA STUDIES
                       Acclimated  System
                     % Removal    Standard
                     	Deviation
                         Unacclimated System
                         % Removal   Standard
                        	Deviation
   RCRA Study

      SS

      COD

      NH4-N

   CERCLA Study
97

82

76
 4

 7

19
97

81

81
 3

 8

18
SS
COD
NH4-N
95
88
88
3
4
14
94
87
98
6
10
3
the secondary effluent stream  in high concentrations  (between 95 and 140
ug/L) indicating poor removals of these organics.  Five other compounds
(cyclohexanone, furfural, 2,4-dimethyl phenol, 2,4-dinitrophenol and 4-
nitrophenol) were not evaluated due to inconsistent results.  The average
removals of the toxic compounds in the RCRA study are summarized in Table 5.
The removal of RCRA organics with primary treatment was between 3 and 44
percent.  The total removal was between 36.6 and 99.0 percent.  The calcu-
lated percent stripped for the individual volatile compounds varied from 1 to
139 percent.  No air analyses were performed for the semivolatiles.  Bio-
degradation of a compound was estimated by subtracting the measured removals
by adsorption and stripping of the compound from the total removal.  The
estimated biodegradation was between -42 and 97 percent.  A negative bio-
degradation indicated inconsistent mass balance and the problems of
estimating biodegradation using this approach.  Biodegradation appeared to be
the predominant removal mechanism for the polar solvents, and the aromatic
volatiles (e.g., toluene: 72%, xylenes: 66%, chlorobenzenes: 60%).  The
unacclimated system (e.g., biodegradation of toluene: 56%) showed similar
results and no significant advantage of acclimation was observed (6).

     Table 6 lists the average concentrations of the CERCLA organics.  Five
compounds (1,2,4-trichlorobenzene:  89 ug/L, 2,6-dinitrotoluene: 125 ug/L,
p-cresol: 156 ug/L, lindane: 198 ug/L, and dieldrin: 99 ug/L) showed high
concentration in the secondary effluent (Table 7).  The average removals of
the toxic compounds in the CERCLA study are summarized in Table 7.  The
                                      344

-------

s:
Oi
co co
0 -Zi
i — i ZD
K
CH. CO
O LU
o: co
o
o: LJ_
ll_ °
CD LU
CD
CO ==C

O LU
i — i ^*


£ «5"
•y* LU
LU 1 —
O CO
Z >-
O CO
0
Q
LU LU
CD 1 —

LU >— l
•=C 0

CO UJ
UJ
tD
a:
UJ
<;

CkC O
Q H-
3: <
I— >
to uj
o
UJ


1


g 0

C> 1—
z ^

00 UJ
o

UJ
C3
C£.
%












UJ
5
z







II
II
II
II
II
II
II
II
II
II
II
II
—1 II
-^ 11
O> II
D II
11
11
II
II
II
II
II
II
II
II
II
II
—1 II
^ II
a> u
n n
n
n
n
n
n
n
u
u
u
n
u
ii
n
n
—i u
^, n
CD II
D II
II
11
II
II
II
II
II
II
II
II
II
II
||
II
-J II
^« II
a> n
=i II
n
u
ii
n
n
n
ti
u
it
n
u
It

n
-i u
•-- II
O) II
3 II
II
II
II

II
II
II
jj
II
II
It
II
It
II
II
II
II
II
II
II
II
II
It
II
11
II
II
II
II
tl
II
II
II

OJ
«— OJ CO
•<»• %o oj
OJ «-
«— ro in
JO 0 ,0

o in ro
CM



0s *"~ *O
ro o CM
CO




CM CO CM
SCO NO
«-




ro co N-
*- CO CO

*—




N- --* co
*»j" ~ ro ro
v± «— CM













c
OJ
4-J
c
C 0-
C Q) CM
OJ C «
O Q) OJ C >-
c JT t- m j=
O O D *J -M
4-> — ' M- D 0)
. 3 « i
<£ CJ u. CM ^f


CM
1
st

co
OJ



tn
OJ





in
OJ
OJ




o-






1













C
m
L.
D
M-
o
|
0)
t_
4-*
Q)


in
1
«-

ro



o
*~




JJ





j:






1













C.
O
^1
2
4J
0)
1—
c


OJ
o


ro
s.
o^
in
OJ
»-

ro



o

O
O
.r:
0
Q
i
OJ



sf
o
CO
o
OJ
o
OJ

Nf
in



R





OJ
OJ




o
in






OJ












c
1

o
t.
o
u
Q

OJ



' R ~* -
co -o ro
ro o o
in o in
t- OJ OJ
ro sj- ro
>o in r^
«— OJ OJ
oj ro oj in

r^ o in in
o »-



r^ -o m co
*~




co sf -o ro
o- o- co ro
ro «— «— oj




i? 5 in 5






sf «- «- O
ro *— o ^
-d- oj oj oj









 • ai o
O JS — ' t-
-• 4J X 0
jr 4) .c — •
O O 4J JZ
c. v o
QJ O O "^
C — • c- t-
CJ JZ O H-
— ' U — ' <
>* (0 J= «-
JZ L. U *

01 O tl
5E 1— 1— «-


O OJ
ro
fe M
in ^~
•o o
r- OJ
r- OJ
 1C





OJ . OJ
0 O
OJ «—




Sf !J






^ t>










c
ro

o co
S. x± <— c\j . , t\j
' ^_ ]
1
«— O CD O CD *—
1
. .
1
1
>t in in in in in



»•$• o« N- ' 'CM co *o
CM ro . in ro o
T-
.



(
f
in ro in to ro CD
o ro o- ro oj o
i— «— CM




ro ^o ro *— r^- CD
ro fM ro NQ >d- co

-•

•
.
•
o CM CM *4- ON . in
OJ vO O -
ro
_.
03 .
JZ
-M
.C
Q. . OJ N C O
x ^: c .— ' t- >. O 0) C C
f-v>ni^c_c~aj-UOQ4J O-M O
1 >- 1 - 1 i .C L- ••- C
yj-+j^j--J--st O. *J Z 
-------


CO
o- —
l— l CO
2: 2:
— >
5:
OH-
— °°
^C L fJ
S1-
Q£ CO
1 1 l ^t^
o o
CO LU
—I CD
11

z: LU
it
.11. 1
-
II 0
II I—
II
II
II
!!
n
n
u
u
H
IS »
ii s
ii »_
n a:
It Q.
n
ii
n
ii

ii
11
n
n
n
n
n
it
n
it
n
it
ii
ii
n
ti
u
u
u
11
u
II
tl
n
H
"

1 1
| REMOVAL STANDARD |
(%) DEVIATION |
1 1
[REMOVAL STANDARD
(%) DEVIATION

1
| REMOVAL STANOARt
| (%) DEVIATIOI
& 1
STANDAF
DEVI ATI

_,
^
o *!

UJ

g I
«C t—
0 <

^c ^>
1— UJ
CO O

— I
o S
UJ
C£













Ul

z







II
ii in t- NO
it
it
it
ii
n
n
n
ti
n
n
ll oj -si- in
II ON ON ON
II
II
II
II
II
II
II
II *~ "~ *~
II
It
II
II
II
It
II
II
II
ll ro <— oj
it
ii
ii
ii
u
It •
tl
n *~
it
ll
it
it
u
ll
ll
It OJ T- «-
II
II
II
It
ll NO co in
u oj oj ro
u
u
u
ii
u
HO CD in
UN. in S-
jl ON ON ON

II
II
II CO T— CO
II NO «— 0
ll oj ro NO

ii
ll
it
ll
it
u
IIO 00
II NO -.
II C J= t-  JC
ii w — < H- D a!
Ii 8 S. D ? ?
It 
OJ
O
t_


o
CO

1 —

NO
i

-
in


ro
£



sf
^_
OJ





ro
ON



















O
ON






0
•J-













OJ
•M
CO
CD
JCZ
4->
•>.
C
OJ
J.3
D
CO

ro c\j vQ c\J
.


(\J CM *— T—
«— in in o


ro ro N. CD -
O O O o



*O OJ f\J *—
S- >d" N" r—






ro N- ro o
co NQ -j- ro













• O  O
i . . Ji t_ .~ C
"^ *^«. "^ (D 't- ^ .C
T— OJ OJ Z Z -J- CL
I
I
1
1
1
I
1
1
1
1
1
1
1
I
J
1
)
1
I
1
1
1
t
1
I
1
1
1
1
t
1
1

1
:
|
I
I
1
1
I
1
I
1 CD
f
l Crt
l
1 (U
1 4-»
(D
f >
I "-
t O
I CU
1 *-*
t M
1 03
I 13
t +
I
t I-

1 E
t ••-
1 f-
1 Q_
1
t
t
( 0)
! f
) — •
1 C/)
1
1 *


1
t
I
1
1
r
t
i
l
i
i
l
i
i
l
i
i
l
i
i
i
l
i
l








t
f
o
o

-------

1
u_
co
i— t CO
 CD
O &
ii ^^
i— >
A? ^j

UJ UJ
CJ 1—

o >-
o co
LU Q
CD UJ
fff |— •
Dig
UJ S
^> i — i
0

UJ
ca
r


















ii
u
n
u
n
n
n
u
ii
u
it
it
n
n
u
it
u
n
n
ii
u
n
u
n
n
u
n
n
n
n
n
ii
u
n
u
n
n
n
n
n
n
u
u
n
u
u
n
n
n
it
n
n
u
n
u
n
n
n
u
n
n
n
ii
n
n
n
n
n
n
n
n
u
u
u
n
!!
jj

j!
II

Jj

U

II

II

It
11

It
'j
II
II
ii
SECONDARY SLUDGE
PRIMARY SLUDGE |
1
EFFLUENT |
1

>-
a
 Q 11
II
II
UJ It
(D —1 II
K ~O> II
UJ 3 II
> II
 II
t— UJ II
CO O II
II
UJ II
0 _J II
ce. 01 n
UJ 3 II
> tl
< it
n
Z II
Q O II
 II
h- LU II
CO O II
II
UJ II
C3 —I II
< — II
UI 3 II
> II
< II
II '
II
II
II
II
II
II
It
II
II
II
II
LU II


CM ro
ro in
in co
o ro
ro in
- <*^ *^ ^ *~~ N- ^O
fo «— *sj~ in «— in in
^- r— N-
CM«-inNOCQ COS-O
NO-J-O>OCO OfMO
O «— in ^a- «— N- -o
«- ^- ro
in QS ON r**- ^f ^o *o ^~
«— «— rocoo mco«—
o O* o NO in co
sjO^*oOi>fs- N-O>O-
coco^ro>o ru <— **
tn^fON-NO fMN-in
T— «— m 0* o in NO
-r -j-ro-st
rONOcor^-cM CO^NQ
cviin-4-OJOj r^irorvj



rONOororo rotoo
m-ON-coco <— in ON
ONN-ONC\JON rOf\Jf>
roin*4-"J'ro -vf-^d-in



UJ

-~i
UJ -
< < _I _J Q. X
*— 31 ^ «4 UJ

CO 3= 1— H- >- -J UJ
CZ Q. D_ 2 3= UJ
O -J UJ h- -J 2
CJi— rct-_jfMj-
to

c
i
S
*




















347

-------
       II
       II
       II
       II
       II

       !!     ~
       n
       n
       n

       n
       n
       n
       n
  .
CO

0
   co
-
uj co  N

°"Q  |i
LU LU  M


^S£  "
LU I
   0
CO

e£
       It




1—

_j


z
*~*

UJ
>
g

o:




•K


<
O
1—














«I
S

CL















UJ
z:

o
E
LU
o:
g
«c
o
z

UJ
a

&?



g
i—
<

>
MJ
*£



s^



g
§

>
UJ
0


s«





g

1—



UJ
o



































II
II
11
11
11
II
II
II
II
It
It
II
II
II
II
II
11
11
II

II



11
|j

11

ii

II
1!
ii

n
n


n
n



it
ii
n
n
n
n
n
n
n
n
n

n
n
n

n
"
H
||
II


11


II

II
II
II

II


II
II


II



II
II

O







to




0.






o
CM





ro
•*





-3-
ro
o





f-
tr-







*—
in






CM





ca

m
0
a_


UJ
z
LU

Z
LU


g

3:
0

a









e




o






vO






,_
in





O
CO
CO





o
0







o
in






ro





CO

ca
0
a:
o_


UJ
z
UJ
IM
Z
UI


g


0

a

o







&




0






N.






CM
in





ro
in
o





*o
,_







ro
,„.






^3-





CO

ca
0
ee.
o_


UJ
LU
IM

UJ


g


0

Q

^







CO
•J-




CM






N.
ro





CO
1^





«-
in
CO





o
in







CM
N.
CM












LU
Z
Si
z
LU
CO
O
g


O

en
t-
1


CM

^—

ro







c>




r-






^_






<0
CM





CO
CM
O





CO
CM







O.
>o




















UI

UI
IM
Z
UJ
m
o

h-

z




























r^
ro





N.
§





co
o







o
o>
















LJJ

UI
IM
Z
UI
CO
o

1—

z
a
i
ro

^

CO
ro






r-
in




in






CM






ro
in





CM
S





in
co







ro
ro
CM















LU
z
LU
3
O
t—
O

1—

z

1


CM


CM






O
•a




,-






CM






O
CM





O
R








































_i
o
LU
C£
U

Q.

co







IS




,-






CM






,_
5Q





o
So





CM
in
ro






si-
,_
CM

















LU
_J

Z

O


o

(J

^^


*
*





o
o
^~"



JUt
*
4c





O






in
CM





«-
fe





CM
^O







ro
^.
ro

















UI
|
i—
LJJ
O

o
_J
3:
0
1


CO

o

CM CO CO ^t
v- ro CM






f^~ "^ CO *O
 co ** i
•*3" T— in i

i
i

	 i
i
i
i
i
UJ 1
h- 1

_J I
UJ - I
CL X 1
UJ 1
_J 3T 1
>- — 1 LLJ 1
rsl >- z I
Z 3= HI 1
UI 1— — J -SS. \
03 LU < LU — 1
—1 CM 1— - N-^ 3= Q _J |
t-  i-, < ^-. •-• i
CO CO Z -J O 1
1
1
1
1
1
1
1
1
1
1
i
i
i
i
i
i
i
i
i
i
t
i
i
i
i
i
i
i
i
i "
i
i
i
i
i
i o
1 O)
1 TJ
1 3
1 (/)
\
\ TJ
1  — •

1 *-• >. (D
1 - O t_ — •
i 
1 4-« O 

.^ •*- TJ
U I_
a. a. .
TI *J "c
3 O O
-^ 1—
a> *
* *
* * *
_

,
,



.



.

•

-

' ,
-

.
'

,



                                            348

-------
removal of CERCLA organics in primary treatment with the acclimated system
was between 3.5 and 79.0 percent.  The total removal varied between 55.9 and
98.1 percent.  Biodegradation was estimated by subtracting the removal by
adsorption from the total removal of a particular compound.  The extent of
biodegradation varied between 28 and 100 percent.  The unacclimated system
also exhibited similar removals.  Biodegradation values were similar for both
acclimated and unacclimated systems (e.g., naphthalene: 79% and dimethyl
phthalate: 85%).  Like the RCRA study, no significant advantage of acclima-
tion was observed for the CERCLA compounds (6).  In the CERCLA study period,
the amounts of organics found in the complex primary sludge samples were
substantially lower than the measured removals across the primary process.
Due to the analytical variability encountered in these studies, additional
work has been planned.
                                  CONCLUSIONS
     The following conclusions were drawn from this study:

 1. The polar solvents and aromatic volatiles were biodegraded to a great
   extent.  For example, toluene exhibited 72 percent and total xylene  showed
   66 percent biodegradation.

 2. A significant  amount of chlorinated aliphatic solvents may be volatilized
   from an  activated sludge  system.  The percent stripped varied between  1
   and 139.

 3. Pesticides (lindane and dieldrin) were removed by both adsorption onto
   primary  and secondary sludge and biodegradation  in the secondary tank.
                                    REFERENCES
 4.
 5.
Fate of priority pollutants in publicly owned treatment works, 1,
EPA440/1-82/303, USEPA Effluent guidelines div.. WH-552, Wash. D.C,
Sept. 1982.
Petrasek, A. C, et a!., "Fate of toxic organic compounds in wastewater
treatment plants", Journal of Hater Pollution Control Federation 1983,
1286-1296.
Hannah, S.A, et al., "Comparative removal of toxic pollutants by six
wastewater treatment processes", Journal of Water Pollution Control
Federation 1986, 27-34.
USEPA Methods for Evaluating Solid Waste, SW846, 3rd edition, Nov. 1986.
Federal Register. Guidelines establishing test procedures for the analysis
of pollutants under the clean water act; 40 CFR, 136, 49, No. 209, 1984.
Bhattacharya, S.K., et al., "Removal and fate of RCRA and CERCLA toxic
pollutants in wastewater treatment", Final Report, Contract No. 68-03-4038.
                                      349

-------
                COMPATIBILITY OF FLEXIBLE MEMBRANE LINERS AND
                       MUNICIPAL SOLID WASTE LEACHATES
                   by:  Henry E. Haxo, Jr.
                        Matrecon, Inc.
                        Alameda, CA  94501
                                 ABSTRACT

      This paper describes the results of a survey of the  open technical
 literature relating to the composition of currently produced  municipal solid
 waste (MOT)  leachate,  the compatibility of flexible membrane  liners  (FMLs)
 with  such leachate, and the results  of limited experimental work  on  the
 absorption of  organics from dilute aqueous solutions that  simulate MSW
 leachates.

      The  results of the survey revealed little information on the compati-
 bility of FMLs  currently being used  in the construction of disposal  facili-
 ties  with MSW  leachates.   Some information was available from studies per-
 formed in the  1970s.   The little information  that  was available on the
 composition  of  MSW  leachates  currently being  produced indicates (1)  that
 concentrations  of potentially polluting species in current leachates appear
 to be low and  (2) that  current leachates  may  contain more  potentially pol-
 luting organics which are less biodegradable  and which may be more aggressive
 to FMLs than leachates  generated during the 1970's.

      Because of uncertainties  regarding the compatibility  of FMLs and MSW
 leachates and the absorption  of organics  from dilute solutions, such as
 leachates, limited laboratory  experiments were performed to measure the
 partitioning of selected  organics from  dilute  aqueous solutions to various
 types of FMLs.  The results indicate that, depending on the similarity of
 the solubility parameters of the organics and  the FMLs, some of the organics
even at low concentrations can partition  from  the water in which they are
dissolved and significantly swell an FML.
                                    350

-------
               COMPATIBILITY OF FLEXIBLE MEMBRANE LINERS AND
                      MUNICIPAL SOLID WASTE LEACHATES
                               INTRODUCTION

    Under Subtitle "C" of the Resource Conservation and Recovery Act (RCRA),
the EPA requires all materials that are used in constructing hazardous waste
disposal facilities and may come in contact with leachate to be tested for
chemical resistance to the specific leachate.  At the present time, the EPA
is considering extending this requirement for chemical resistance testing to
all materials which are used in constructing municipal solid waste (MSW)
landfills and which may come in contact with leachate.

     Previous studies indicate that commercially-available flexible membrane
liners (FMLs) are chemically resistant to typical MSW leachate (1, 2).  These
studies were conducted with leachate generated in simulators containing
freshly collected and shredded MSW.  Analyses of that leachate showed in-
organic salts, volatile organic acids, and suspended solids as the only
contaminating species.  However, specific analyses of additional organics do
not appear to have been perfromed.  In addition, adding small amounts of
organics from small quantity generators and nonhazardous industrial wastes
to an MSW stream could generate a leachate containing measurable quantities
of organic compounds that could chemically attack an FML and reduce its
service life.

     Even though commercially available FMLs are generally chemically
resistant, some could be adversely affected by specific constituents of
the leachate.  The effects can depend on the type of constituent, its con-
centration in the leachate, and the specific type of FML.  Furthermore, the
effects of the constituents can be synergistic and can vary with time as the
concentrations change with the aging  of the waste.   Organic molecules  [as
indicated by such analyses as volatile acids, volatile solids, total organic
carbon  (TOG), and chemical oxygen demand (COD)] may be damaging to some FMLs.
For example, specific organics can cause some types of FMLs to swell, to
become  softer and more permeable, and to lose in mechanical properties, such
as tensile strength and tear resistance; thus, exposure to some organics can
allow FMLs to be more easily torn and damaged.  Other constituents of a
leachate, e.g. water, can also cause  some FMLs to swell.

     Determining  the  chemical resistance of  FMLs and other materials of
construction is necessary when the waste liquid or leachate is. known to
contain constituents  that are aggressive to  these materials and when the
                                     351

-------
           concentrations of these constituents are sufficiently high.  The current
           method for determining chemical resistance of FMLs to waste liquids and
           leachates for permitting purposes under RCRA is EPA Method 9090.  However,
           this method as presently performed may not be adequate for assessing FML
           resistance to MSW leachate due to the instability of the leachate during the
           required 120-day exposure at the temperatures (23° and 50°C).  To assess FML
           resistance, the FMLs under test should be exposed to leachates that are in an
           in situ condition.

                This paper describes the results of a survey of the open technical
           literature relating to information that would be appropriate to a discussion
           on extending the requirement for chemical resistance testing of materials
           used in constructing MSW landfills, particularly FMLs.  The information
           sought in the survey included data on the composition of currently generated
           MSW leachates, the compatibility of FMLs with such leachates, and basic
           information regarding dilute aqueous solutions.   Results of limited experi-
           mental work on the partitioning of selected organics from dilute aqueous
           solutions to various types of FMLs are also presented in this paper.   Because
           MSW leachates are essentially complex dilute solutions,  these results  can
           supply information on the tendency of FMLs  to absorb organics from leachates
           and, ultimately, on the compatibility between FMLs and MSW leachates.


           SURVEY OF THE SCIENTIFIC AND OPEN TECHNICAL LITERATURE

                The  information sought in the survey included the following:

                o Data on MSW leachate composition and characteristics,  particularly
                  currently generated leachate.

                o Data from the compatibility testing of FMLs  with MSW  leachates.

                o Information on the  preservation  and stabilization of MSW  leachates,
                  particularly at  temperatures  somewhat higher than 23°C  so  that  these
                  leachates can be  used  in  conventional compatibility tests,  e.g.  EPA
                  Method  9090.

                o Data on the  resistance  of  FMLs and  polymeric compositions  to dilute
                  aqueous  solutions of organics and inorganics.  Such solutions simu-
                  late MSW  leachates  in many respects.

                o  Basic information from  physical  chemistry  regarding such factors as
                  the  chemistry of dilute aqueous  solutions  of  organics,  solubility
                  parameters,  constituent activity, partitioning of dissolved organics
                  between phases, and  transport of organic species.

          The information resulting from  this literature search is  summarized in the
          following subsections.
                                               352
_

-------
Composition of MSW Leachate

     The literature search revealed that most of the available information
on the composition and characteristics of MSW leachate had been generated in
the 1970s when the disposal of MSW was a principal concern of the EPA (3, 4,
5, 6, 7, 8, 9, 10, 11, 12).  The data reflect the characteristics of leach-
ates generated in both laboratory and pilot-scale projects and in actual
full-scale MSW landfills; these data include information such as chemical
oxygen demand (COD), biological oxygen demand (BOD), total organic carbon
(TOG), hardness, pH, electrical conductivity (EC), total dissolved solids
(TDS), trace metal analysis, volatile organic acids, and dissolved inor-
ganics.  Almost no specific information on organic species was found, except
for information on the organic acids, even though many are potentially pol-
luting or aggressive to FMLs and other polymeric .construction materials.
Much of the information on leachate composition was considered in developing
the chapter on wastes for the Technical Resource Document, SW-870, "Lining of
Waste Impoundment and Disposal Facilities" and its subsequent revision (2,
13).  Overall, the information indicates the following:

     o  MSW leachate is a complex mixture of inorganics, organics, and
        bacteriological constituents usually generated in anaerobic environ-
        ments in MSW landfills.  MSW leachate is generally mildly acidic,
        and many of the constituents are biodegradable.

     o  Leachates from different MSW landfills vary widely; the precise compo-
        sition is waste- and site-specific and depends on such variables as
        type of waste, amount of infiltrating water, age of the landfill, and
        pH.  Table 1 presents typical results of analyses of MSW leachates
        that were obtained and reported in the 1970s.

     o  MSW leachate is highly oxidizable and unstable and subject to rapid
        changes in composition on removal from the in situ anaerobic environ-
        ment in which it is usually  generated.  Even when cooled and sealed
        in bottles, the composition  of leachate will change for a limited
        time (4).

     Data in the open literature on  the composition of recently generated
MSW  leachates are limited; data that are available indicate the presence of
priority pollutants, aromatic hydrocarbons, and other constituents which
may  be  absorbed by FMLs  (14, 15, 16).  Table 2 presents a statistical
analysis of data on the concentrations of organic constituents of leachates
generated between  1980 and 1985.  These data include data from 15 case
studies performed by the EPA (15); data from landfill leachate sampling by
Minnesota and Wisconsin, and data from the National Pollutant Discharge
Elimination System  (NPDES) discharge permits for leachates from landfills
in New Jersey.  Even though only a relatively small number of facilities
were surveyed,  the data  resulting from those studies were considered reli-
able by the EPA (16).  The concentrations of most of the organics are low
and, even  if  they partition to  the FMLs, the amounts may not  be sufficiently
large  to cause  significant changes in properties of the FMLs, even after long
exposures.
                                     353

-------
       TABLE 1.  COMPOSITION OF MSW LANDFILL LEACHATES GENERATED BEFORE 1980
    Concentration of Constituents. (mg/L), Except pH and Electrical Conductivity

Constituent
BOD5
COD
TOC
Total solids
Total dissolved solids
Total suspended solids
Total volatile acids as acetic acid
Acetic acid
Fropionic acid
Butyric acid
Valeric acid
Organic nitrogen as N
Ammonia nitrogen as N
Kjeldahl nitrogen as N
Total phosphorus
PH
Electrical conductivity (ftmho/cm)
Total alkalinity as CaC03
Total acidity as CaCC>3
Total hardness as CaC03
Metals and anions:
Arsenic
Boron
Cadmium
Calcium
Chloride
Chromium
Copper
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
Phosphate
Potassium
Silica
Sodium
Sulfate
Zinc

Reference
No. 17
• • *
42,000
• • •
36,250
• • •
• • •
• • •
• • •
• • •
• • •
• • •
• • •
950
1,240
• • •
6.2
16,000
8,965
5,060
6,700

• • •
* • •
• • •
2,300
2,260
• • •
• • •
1,185
» • •
410
58
* • •
• • •
82
1,890
• • •
1,375
1,280
67
Source
Reference
No. 18
13,400
18,100
5,000
12,500
• • •
85
9,300
5,160
2,840
1,830
1,000
107
117
• • •
• • •
5.1
• • •
2,480
3,460
5,555

• • •
• • •
• • •
1,250
180
• • •
• • •
185
• • •
260
18
• • •
• • •
1.3
500
• • •
160
...
...
of Data
Reference
No. 19
• • •
1,340
• • •
• • *
• • •
• • •
333
• • •
• • •
• • •
• • •
* • •
862
• • •
• • •
6.9
• • •
* • •
• • •
• • •

0.11
29.9
1.95
354.1
1.95
<0.1
<0.1
4.2
4.46
233
0.04
0.008
0.3
• • •
• • *
14.9
748
<0.01
18.8

Reference
No. 4
81-33,360
40-89,520
256-28,000
0-59,200
584-44,900
10-700
• •
• •
• •
• •
• •
• • •
0-1,106
• • •
0-130
3.5-8.5
2,810-16,800
0-20,850
i • • «
0-22,800

• • •
• • *
0.03-17
60-7,000
4.7-2,467
• • •
0-99
0
0-2,820
17-15,600
0.09-125
• • *
• • •
• • •
28-3,770
• • •
0-7,700
1-1,558
0-370
Based on:  Reference No. 2, p A-5.
                                        354

-------
             TABLE 2.  CONCENTRATIONS OF ORGANIC CONSTITUENTS/
                    MSW LEACHATES GENERATED AFTER 1980a
(Units in ppb)
Constituent
Acetone
Benzene
Bromome thane
1-Butanol
Carbon tetrachloride
Chlorobenzene
Chloroethane
Bi s ( 2-chloroethoxy )me thane
Chloroform
Chloromethane
Delta BHC
Dibromome thane
1 ,4-Di chlorobenzene
Dichlorodifluoromethane
1 , 1-Di chloroethane
1 ,2-Dichloroethane
Cis 1 , 2-dichloroethene
Trans 1 , 2-dichloroethene
Di chloromethane
1 ,2-Dichloropropane
Diethyl phthalate
Dimethyl phthalate
Di-n-butyl phthalate
End r in
Ethyl acetate
Ethyl benzene
Bis(2-ethyl hexyl) phthalate
Isophorene
Methyl ethyl ketone
Methyl isobutyl ketone
Naphthalene
Nitrobenzene
4-Nitrophenol
Pentachlorophenol
Phenol
2-Propanol
1,1,2, 2-Te trachloroe thane
Tetrachloroethene
Tetrahydrofuran
Toluene
Toxaphene
1,1, 1-Trichloroethane
1,1, 2-Trichloroethane
Trichloroethene
Trichlorof luoromethane
Vinyl chloride
m-Xylene
p-Xylene and o-Xylene
Minimum
140
2
10
50
2
2
5
2
2
10
0
5
2
10
2
0
4
4
2
2
2
4
4
0
5
5
6
10
110
10
4
2
17
3
10
94
7
2
5
2
0
0
2
1
4
0
21
12
Maximum
11,000
410
170
360
398
237
170
14
1,300
170
5
25
20
369
6,300
11,000
190
1,300
3,300
100
45
55
12
1
50
580
110
85
28,000
660
19
40
40
25
28,800
10,000
210
100
260
1,600
5
2,400
500
43
100
100
79
50
Median
7,500
17.
55
220
10
10
7.5
10
10
55
0
10
7.7
95
65.5
7.5
97
10
230
10
31.5
15
10
0.1
42
38
22
10
8,300
270
•8
15
25
3
257
6,900
20
40
18
166
1
10
10
3.5
12.5
10
26
18
aThis table was developed by the U.S. EPA, Office of Solid Waste,  Economic
 Analysis Branch.

Source:  Reference No. 16.
                                  355

-------
 Compatibility of FMLs with MSW Leachate

      The only data on the compatibility of FMLs with MSW leachate found in
 the literature survey had been developed by Haxo et al (1, 2, 13).  In this
 work the major types of liner materials that were available and used for
 containment in the early 1970s were evaluated on exposure to MSW leachate for
 up to 56 months.  The leachate used in these studies was generated in pilot-
 scale MSW landfill simulators filled with a shredded residential solid waste.
 As the work was part of an exploratory research program performed before
 Method 9090 was developed, no immersion testing that approximates EPA Method
 9090 or other short-term exposures were conducted.  All of the exposure
 testing was performed at 23°C or less with leachate that was continually being
 generated in the landfill simulators.  Complete analyses of the leachate were
 not performed, though several parameters, i.e. solids, pH, COD, and total
 volatile acids (TVA), were followed.  Overall, the changes in the physical
 properties of the FMLs resulting from 56 months of exposure were relatively
 minor.   It should be noted,  however, that the leachate generated for this
 study probably did not contain aromatics, chlorinated hydrocarbons,  and other
 volatile organics that are known to affect hydrocarbon FMLs.   Consequently,
 these results may not be completely applicable to leachates being generated
 currently in active landfills.


 Preservation of  MSW Leachates for Use in Compatibility Tests

      An important factor in  conducting a compatibility test is  that  the
 composition of the test  leachate  reflects  the  composition  of  in  situ
 leachate.   Thus,  the  test  leachate should  maintain a constant~~concentration
 of those constituents  that could  affect an FML during  extended  service.
 Because MSW  leachate  is  highly oxidizable  and  unstable and  is subject to
 change  almost  immediately after removal from a service environment,  which
 generally is anaerobic,  methods of preserving,  stabilizing, or sterilizing
 leachates need to  be  developed  for possible use in  performing exposure  tests
 to assess the  compatibility of  FMLs  with MSW leachate.

     For analytical purposes, refrigeration has been used to protect leach-
 ates from bacteriological changes  before they  can be analyzed.   However, EPA
 Method  9090, the standard test  for determining the  compatibility of an  FML
 with a waste liquid, is  performed  at higher temperatures (23° and 50°C);
 therefore, refrigeration can only  be used to store  the  leachate  before use
 in exposure testing.

     No  information was  located in the open literature  indicating any methods
 of long-term stabilization of MSW leachates that could  be used in the com-
 patibility testing of FMLs at 23° and 50°C.

     Analysis of a single sample of stored MSW leachate indicated that stabi-
 lization of the leachate may be possible with sterilization.  A  1-gal bottle
of MSW leachate, which had been generated in October 1976 in a pilot-scale
simulator (1) and had been stored in a sealed brown glass bottle, was found
in September 1988 to be only slightly changed in composition since 1976.
                                      356

-------
There was no indication of volatile chlorinated hydrocarbon organics in
the leachate.

     If the EPA should require compatibility testing of FMLs with MSW leach-
ates in accordance with EPA Method 9090, additional research is needed to
develop adequate means of preserving or sterilizing MSW leachates to prevent
changes in leachate composition during the test, except for those changes
resulting from absorption by the FML under test.


Applicable Information on the Chemical Resistance of FMLs

     Solubility parameters are used in polymer science and technology to
assess the solubilities of polymers in different organic solvents (20).
Data generated in a recent study (21, 22) on the solubility parameters of FMLs
are applicable to assessing the compatibility of FMLs and MSW leachates if the
composition of the leachate is well established.  In that study, the solubility
parameters of polymers used in manufacturing FMLs were determined using their
equilibrium swelling in a variety of different-organics, the solubility
parameters of which are well documented in the literature (23, 20, 24, 25).
However, the solubility parameters of an FML are only one property of an FML
that can affect the magnitude of its swelling when it is in contact with a
leachate or waste liquid; for example, crosslinking, crystallinity, and filler
content of the FML compound can significantly affect the amount of swelling.


Applicable Information from Physical Chemistry on Dilute Solutions

     When a substance such as an organic solvent is added to a two-phase
system such as a polymer and water It will, in general, distribute at equi-
librium with different concentrations in the two phases (26).  The equi-
librium concentrations are related to relationship between the solubility
parameters of the phase and those of the solvent.  The ratio of the con-
centrations at equilibrium of the solute in the two phases, also known as the
distribution coefficient, remains essentially constant over a wide range of
concentrations.  Therefore, in the case of an organic distributed between
water and a polymer, a decrease in the concentration of the organic in the
aqueous phase would eventually result in a decrease in its concentration in
the polymer phase.  This characteristic has been found to be applicable to
dilute aqueous solutions in contact with a polyethylene FML.  Limited data
have been reported on the distribution coefficients of selected organics in
dilute aqueous solutions between the water and selected polyethylene FMLs
(22, 27, 28).


EXPERIMENTAL WORK ON DISTRIBUTION COEFFICIENTS OF FMLS

     On reviewing the data obtained in the literature survey, it was obvious
that there were areas of uncertainty which should be resolved before valid
recommendations could be made as to the type of testing needed to assess the
long-term compatibility of a specific FML with a specific MSW leachate.
                                   357

-------
      Inasmuch  as  only a limited  amount  of  experimental  work could  be  per-
 formed  in  this study, it was  decided  to measure  the  absorption  of  three
 organics by  different FMLs  from  dilute  aqueous solutions.   Samples of FMLs
 based on four  different polymers,  including  a linear low-density polyethyl-
 ene  (LLDPE), polyvinyl chloride  (PVC),  chlorinated polyethylene (CPE), and
 chlorosulfonated  polyethylene (CSPE), were placed in vapor-tight cells con-
 taining a  dilute, unsaturated aqueous solution containing  three different
 organics.  The concentrations of the  organics in the aqueous solutions were
 monitored  until they  had become  relatively constant  at  which time  the cells
 were  opened  and the FML samples  analyzed to  determine the  concentration of
 the organics in the FML samples.  Analyses of both the  solutions and  the FMLs
 for the organics were performed  using gas  chromatographic  (GC)  procedures.
 Experimental details  and results are  presented in this  section.


 Gas Chromatography Procedures

      A  Perkin-Elmer Sigma Three  Series  gas chromatograph with a flame ioni-
 zation  detector was used for  the GC analyses.  The instrument was  fitted with
 an open-capillary column coated  with polyethylene glycol.   Details  of the
 GC analyses  are presented in  Table 3.

      The concentrations  of  the organics in the aqueous  solutions were
 determined by  injecting  samples  removed from the test cells  directly  into
 the GC.  These cells  featured septums through which  the aqueous  solutions
 could be sampled during  the test to determine whether equilibrium had been
 reached.

      The concentrations  of  the organics in the FML samples were  determined
 by headspace GC.  In  this procedure, an FML  specimen containing  absorbed
 organics is placed in a  small vapor-tight  can provided with  a septum
 through which  vapors  from the specimen  can be sampled.  The  can  is  placed
 in an oven at  105°C and  heated for approximately an  hour.  A sample of the
vapors  is drawn from  the  can  and injected  into the GC.  The  FML  specimen is
 removed from the sampled  can  and placed in a new can which is then heated
 in a  105°C oven for approximately  an hour.   Once again, the  vapors inside
 the can are sampled and  injected into the  GC.  The process of heating,
 sampling, and  injecting  is repeated until  no organics are detected  in the
 sampled vapors  by the  GC.  The amount for  each organic were  summed  to
represent a total for  the organic  in the sample.

      The concentrations  of the organics in the injected samples  were
calculated by  comparing  peak  height values resulting from the GC analyses
with  calibration curves.  The calibration  curves for the aqueous solutions
were  determined by injecting  1 /*L  of various solutions of known  concentra-
tions of the different organics  into the GC  column.  Injections  of each
standard were  performed  five  times to ensure reproducibility  of  injection
techniques.  The calibration  curves for the  headspace GC analyses were
prepared by analyzing  a  specific 100 /xL volume of vapor from headspace cans
filled with different  amounts of the organics.
                                      358

-------
          TABLE 3.  GAS CHROMATOGRAPHY CONDITIONS FOR LEACHATE
                           AND VAPOR ANALYSIS
            Condition
            Setting
Detector range

Injector and detector temperature

Initial temperature

Initial holding time

Final temperature

Final holding time

Temperature program rate

Detector



Column^
Chart speed

Carrier

Specimen size:
  Liquid
  Vapor from headspace

Attenuation
x 10

250°C

60°C

1 min.

120°Ca

1 min.

20°C/min.

Flame ionization:
  H2 30 cc/min.
  02 40 cc/min.

Polyethylene glycol coated
fused silica open tube
(FSOT) capillary: 0.53 mm
in diameter and 15 m in
length

30 cm/hour

Helium, 10 cc/min.
1 ML
100 /uL

32 (for 500 ppm concentration)
down to 4 (for low concentration)
aCan vary with the solvent mixture from 120°C to 260°C, the maximum
 temperature for the column.                        ,
^Trade name Superox (Altec), Megabore DB WAX (J and W).
                                   359

-------
 Selection of Volatile Organics

      For this study, three types of volatile organics representing a range
 of chemical characteristics were desired, including a volatile containing
 oxygen,  a volatile that was a chlorinated solvent, and a volatile that was
 an aromatic.  The three selected were methyl ethyl ketone (MEK), trichloro-
 ethylene (TCE),  and toluene.  All of these organics have been observed in
 MSW leachates.  Properties of these solvents are presented in Table 4.
            TABLE 4.   ORGANICS USED IN ABSORPTION EXPERIMENTS
                      WITH DILUTE AQUEOUS SOLUTIONS
Organic
Toluene
Trichloroethylene
Methyl ethyl ketone
Mole-
cular
weight
92.13
31.40
72.10
Density
at 20°C,
K cm""
0.866
1.476
0.805
Boiling
point ,
°C
110.6
87.2
79.6
Vapor
pressure
at 25°C,
mm Hg
31.96
80.30
100.0
«o
8.9
9.2
9.3
Solubility
parameters3
<5d See Reference No. 29.
 cChemical Abstract Services' number.
Immersion Test Cells

     The immersions were  performed in vapor-tight cells consisting of 8-oz
glass jars with ground  polished top edges, and Teflon-lined phenolic resin
tops (Figure  1).  Each  of these tops was fitted with a Swagelock sampling
port and a Teflon-lined silicone rubber septum for withdrawing samples for
GC analysis.  Special arrangements were made in the cells to suspend the FML
samples in the solutions.
Experimental Procedure

     The work performed  in  this  study developed out of previously reported
work (22, 27, 28).  In the  earlier  study,  samples of a polyethylene FML were
placed in the same type  of  test  cells as used in the present work.  These
cells were then filled with a  series  of dilute but saturated aqueous solu-
tions, each of which contained a single organic.  These solutions were
prepared with an excess  of  the organics to maintain saturation.   After
analyzing the aqueous solutions  and FML samples at the end of the immersions,
the distribution coefficients  (i.e. the ratio of the concentration of the
organic in the FML to its concentration in the aqueous solution)  were"cal-
culated.  It was noted,  however,  that the  amounts of organics absorbed by the
FML approached the amounts  the FML  absorbed when immersed in neat organics.
It appears that the water had  acted as a permeable medium between the FML and
                                      360

-------
the reservoir of excess organics which allowed the organics to be absorbed by
the FML until saturation of the FML was reached.  These conditions are not
representative of those in a landfill due to the improbability of an excess
of the organic being maintained until saturation was reached in the installed
FML.
              Teflon
              Septum
                  \
   Teflon-lined
   Screw cap
                            Swagelock
                            Assembly
         Washer
                          Nut
             TOP ASSEMBLY
Jar  with
ground and
polished
edge
                                                 8 OZ JAR
           Figure 1.  Schematic of the immersion test cell.
     In the experimental work conducted in the present study, the aqueous
solutions were prepared at concentrations less than saturation and the
number of FMLs in test were increased to include the four basic polymer
types.  Four immersion test cells were each filled with an aqueous solution
containing a mixture of 500 ppm (on a weight basis) each of MEK, TCE., and
toluene.  Samples of CPE, fabric-reinforced chlorosulfonated polyethylene
(CSPE-R), LLDPE, and PVC FMLs were placed in separate test cells, and the
concentrations of the volatile organics in each cell were followed by GC.
In all cases, the initial concentrations of the MEK in solution showed
little change and, thus, little distribution to the FMLs, whereas the
toluene and TCE showed drops in concentration and thus a transfer from the
water to the FMLs.  The cells were dismantled and the respective FMLs
were analyzed for the volatile organics by headspace GC.


Experimental Results

     The results of analyzing the aqueous solutions and the FMLs at the end
of the immersion are presented in Table 5.  The results show that there was
a significant increase in the weight of the FML samples; however, the dif-
ference between the weight gains of the FML samples and the total amount
of organics detected in each FML sample indicates that not all of the weight
gains could be attributed to the absorption of the organics.  Thus, consider-
able water had also been absorbed by the specimens.  In addition, it is also
                                     361

-------
TABLE 5.  ABSORPTION BY IMMERSED FMLS OF A MIXTURE OF ORGANICS3
                  FROM DILUTEb AQUEOUS SOLUTIONS
Parameter
Contents of cell at beginning
of experiment
Amount of water in cell, g
Amount of organics in cell:
MEK, mg
TCE, mg
Toluene, mg
Total
Original weight of FML
specimen, g
FML specimen at end of
immersion
Swollen weight , g
Weight gain, g
Weight gain, %
Headspace analysis of
swollen FML specimen
Amount in swollen FML
specimen:
MEK, mg
TCE, mg
Toluene , mg
Total
% of original weight
of FML
Concentration of organics
in swollen FML specimen
MEK, ppm
TCE, ppm
Toluene , ppm
PVCC
579


225.5

112.8
112.8
112.8
338.4

3.4580


3.668
0.210
6.1




1.35
54.0
55.6
110.95

3.21


370
14,720
15,200
LLDPEC
580


226.6

113.3
113.3
113.3
339.9

2.5337


2.626
0.093
3.7




0.08
36.3
44.5 .
80.88

3.19


30
13,800
16,900
CPEC
581


228.4

114.2
114.2
114.2
342.6

5.4670


6.246
0.779
14.2




1.16
63.7
65.5
130.36

2.38


190
10,200
10,490
CSPE-RC
582


225.2

112.6
112.6
112.6
337.8

6.1601


6.702
0.542
8.8




0.55
34.3
38.8
73.65

1.20


80
5,120
5,790
                                                      continued  .
                              362

-------
                            TABLE 5.   CONTINUED
Parameter
FML specimen after head-
space analysis
PVCC
579

LLDPEC
580

CPEC
581

CSPE-RC
582

 Weight.g                      3.4501     2.5310      5.4551       6.1253

 Loss in weight  (based on
    original weight),  %           0.23       0.11        0.22         0.56
at end of experiment
Amount in aqueous solution:
MEK, mg
TCE, mg
Toluene , mg
Total
Concentration of organics
in aqueous solution (CH_Q):
MEK, ppm
TCE, ppm
Toluene , ppm
Total amount of organics
in celld:
MEK, mg
TCE, mg
Toluene , mg
Total
Organic accounted for at
end of experiment , %
Distribution coefficient
(CFML/CH2o):
MEK
TCE
Toluene


112.5
48.3
39.2
200.0


500
214
174


114
102
95
311

92


0.74
68.8
87.4


110.1
87.5
80.4
278.0


486
386
355


110
124
125
359

106


0.06
35i.8
47.6


91.4
41.6
27.9
160.9


400
182
122


93
105
93
291

85


0.48
56.0
86.0


111
46.6
38.1
195.7


493
207
169


111
81
77
269

80


0.16
24.7
34.3
aMethyl ethyl ketone (MEK), trichloroethylene (TCE), and toluene.
CPVC » polyvinyl chloride; LLDPE = linear low-density polyethylene;
 CPE = chlorinated polyethylene; CSPE-R = chlorosulfonated polyethylene
 (fabric-reinforced).
^Initial concentration of each organic was 500 ppm.
     of organics in water and in the FML.
                                    363

-------
possible that some of the organics not accounted for at the end of the experi-
ment had been absorbed by the FML but were not recovered during the headspace
GC procedure.  Regardless, the results do show a large partitioning of the
organics from the water to the FMLs and, furthermore, they show that the FMLs
varied considerably in their absorption of the different organics.  The total
amounts of organics and water that were absorbed ranged from 3.7% for the
polyethylene to 14.2% for the CPE.  This experiment was repeated with similar
overall results.  The results of very limited physical testing, which was
performed on percut specimens immersed in the second experiment, consistently
indicated moderate losses in tensile property values.


CONCLUSIONS

     The results of the literature survey and the limited experimental work
that was conducted in this study indicate that:

     o  There is little information in the literature regarding the compo-
        sition and characteristics of current generation leachates and the
        compatibility of FMLs with these leachates.

     o  The limited data available indicate that MSW leachates currently
        being generated may contain low concentrations of organics having low
        biodegradability; in addition, depending on their concentration, some
        of these organics are aggressive towards FMLs.

     o  It is questionable whether EPA Method 9090, as presently designed,
        can result in a consistent and realistic assessment of the compati-
        bility of an FML with a MSW leachate without modification of the
        exposure cells to maintain anaerobic conditions and a standard means
        of assuring stabilization of the leachates so that little change in
        concentrations of the constituents during the required four months is
        ensured.

     o  The use of distribution coefficients that represent the distribution
        of dissolved organics between aquous and polymeric phases appears to
        be applicable to assessing the compatibility of FMLs  and other poly-
        meric construction materials with MSW leachates,  which are basically
        dilute aqueous solutions of organic and nonorganic solutes.  Thus, the
        amount of organics absorbed by an FML from a dilute aqueous solution
        with which it is in contact may be estimated.  For example, at low
        concentrations of organics in solution, the amount of the organics
        that an FML will absorb will diminish as the concentration in the
        leachate is lowered.

     o  The solubility parameters of an FML and the individual organic are
        important tools in assessing the level of absorption  of the organics
        by the FML and in determining the distribution coefficient of an
        organic between an aqueous solution and an FML.
                                    364

-------
RECOMMENDATIONS

     It is recommended that further experimental research be performed:

     o  To develop quantitative data regarding the distribution of dissolved
        organics between various FMLs and aqueous solutions or actual MSW
        leachates in which the FMLs are immersed.

     o  To determine threshold concentrations of various organics found
        in MSW leachates, below which levels absorption will not signif-
        icantly affect the properties of FMLs and other materials of con-
        struction used in liner systems.

     o  To assess possible interactions between organics in dilute aqueous
        solutions and the effects of the absorbed organics on permeability
        and properties of FMLs.

     o  To perform a round-robin series of exposure tests performed in
        accordance with EPA Method 9090 to establish error and bias in
        the test method.

     o  To perform more in-depth analyses of MSW leachates than have
        normally been made to determine the presence of specific organics
        that swell various FMLs.  These organics include chlorinated
        solvents, aromatic solvents, and some aliphatic solvents, all of
        which which are known to deteriorate the properties of polymeric
        compositions.
ACKNOWLEDGMENT

     The work on this project was performed by Matrecon under Work Assignment
No. 0/01 of Subcontract 771-87 of Environmental Protection Agency Contract
68-03-3413 to PEI Associates, Inc.
REFERENCES

 1.  Haxo, H. E., R. M. White, P. D. Haxo, and M. A. Fong.  1982.  Final
     Report: Evaluation of Liner Materials Exposed to Municipal Solid Waste
     Leachate.  NTIS No. PB 83-147-801.  U.S. Environmental Protection
     Agency, Cincinnati, OH.

 2.  Matrecon, Inc.  1988.  Lining of Waste Containment and Other Impoundment
     Facilities.  EPA-600/2-88-052. [SW-870, second revised edition.]  U.S.
     Environmental Protection Agency, Washington, B.C. 991 pp.

 3.  Chian, E. S. K., and F. B. DeWalle.  1976.  Analytical Methodologies for
     Leachate and Gas Analysis.  In: Proceedings of a Research Symposium on
     Gas and Leachate from Landfills: Formation, Collection, and Treatment.
     EPA-600/9-76-004 (NTIS No. PB-251-161).  U.S. Environmental Protection
     Agency, Cincinnati, OH.  pp 44-53.
                                    365

-------
 4.  Chian, E. S. K., and F. B. DeWalle.  1977.  Evaluation of Leachate
     Treatment.  2 volumes.  EPA-600/2-77-186 a,b.  U.S. Environmental
     Protection Agency, Cincinnati, OH.

 5.  Dunlap, W. J., D. C. Shew, J. M. Robertson, and C. R. Toussaint.  1976.
     Organic Pollutants Contributed to Groundwater by a Landfill.  In:
     Proceedings of a Research Symposium on Gas and Leachate from Land-
     fills: Formation, Collection, and Treatment.  EPA-600/9-76-004 (NTIS
     No. PB-251-161).  U.S. Environmental Protection Agency, Cincinnati,
     OH.  pp 96-110.

 6.  Ham, R. K.  1975.  Milled Refuse Landfill Studies at Pompano Beach, FL.
     Approx. Range, Three Cells Aged One Year.  21 pp.

 7.  Ham, R. K.  1976.  Solid Waste Degradation Due to Shredding and Sludge
     Addition.  In: Proceedings of a Research Symposium on Gas and Leachate
     from Landfills: Formation, Collection, and Treatment.  EPA-600/9-76-
     004 (NTIS No. PB-251-161).  U.S. Environmental Protection Agency,
     Cincinnati, OH.  pp 168-176.

 8.  Ham, R. K., K. Hekimian, S. Katten, W. J. Lockman, R. J. Lofty, D. E.
     McFaddin, and E. J. Daley.  1979.  Recovery, Processing, and Utili-
     zation of Gas from Sanitary Landfills.  EPA-600/2-79-001.  U.S.
     Environmental Protection Agency, Cincinnati, OH.  133 pp.

 9.  Pohland, F. G.  1975.  Sanitary Landfill Stabilization and Leachate
     Recycle and Residual Treatment.  EPA-600/2-75-043.  U.S. Environmental
     Protection Agency, Cincinnati, OH.  105 pp.

10.  Pohland, F. G.  1976.  Landfill Management with Leachate Recycle and
     Treatment: An Overview.  In: Proceedings of a Research Symposium on Gas
     and Leachate from Landfills Formation, Collection, and Treatment.
     EPA-600/9-76-004 (NTIS No. PB-251-161).  U.S. Environmental Protection
     Agency, Cincinnati, OH.  pp 159-167.

11.  Pohland, F. G., D. E. Shank, R. E. Benson, and H, H. Timmerman.  1979.
     Pilot Scale Investigations of Accelerated Landfill Stabilization with
     Leachate Recycle.  In: Municipal Solid Waste: Land Disposal.  Proc.
     5th Annual Res. Sympos.  EPA-600/9-79-023a.  U.S. Environmental Pro-
     tection Agency, Cincinnati, OH.  pp 283-295.

12.  Pohland, F. G., W. H. Cross, and J. P. Gould.  1987.  The Behavior and
     Assimilation of Organic and Inorganic Priority Pollutants Codisposed
     with Municipal Refuse - A Progress Report.  In: Proceedings of the
     Thirteenth Annual Research Symposium: Land Disposal of Hazardous Waste.
     EPA-600/9-87-015.  U.S. Environmental Protection Agency, Cincinnati, OH.
     pp 26-37.

13.  Matrecon, Inc.  1983.  Lining of Waste Impoundment and Disposal Facili-
     ties.  SW-870 Revised.  U.S. Environmental Protection Agency, Washington,
     D.C.  448 pp.
                                   366

-------
14.  Kmet, P., and P. M. McGinley.   1982.  Chemical Characteristics of Leach-
     ate from Municipal Solid Waste  Landfills in Wisconsin.  In: Proceedings
     of the 5th Annual Madison Conference of Applied Research and Practice on
     Municipal and Industrial Wastes, September 22-24, 1982.  Dept. of Eng.
     and Applied Science, University of Wisconsin Extension, Madison, WI.
     pp 225-254.

15.  EPA.  1986.  Muncipal Landfill  Case Studies (unpublished).  These
     studies were prepared by PEI, SRW, and ICF.  U.S. Environmental Pro-
     tection Agency, Office of Solid Waste, Washington, D.C.  Cited in: U.S.
     EPA.  1986.  Subtitle D Study - Phase I Report.  EPA/530-SW-86-054.
     U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
     D.C.

16.  EPA.  1986.  Subtitle D Study - Phase I Report.  EPA/530-SW-86-054.
     U.S. Environmental Protection Agency, Office of Solid Waste, Washington,
     D.C.
17.  Wigh, R. J.  1979.  Boone County Field Site.  Interim Report, Test Cells
     2A, 2B, 2C, and 2D.  EPA-600/2-79-058.  U.S. Environmental Protection
     Agency, Cincinnati, OH.  202 pp.  (NTIS PB-299-689).
18.
19.
20.
21.
22.
23.
Breland, C. G.   1972.  Landfill Stabilization with Leachate Recir-
culation, Neutralization, and Sludge Seeding.  CE-756A6.   School of
Civil Engineering, Georgia Institute of Technology, Atlanta, GA.  80
pp.

Griffin, R. A.,  and N. F. Shimp.   1978.  Attenuation of Pollutants in
Municipal Landfill Leachate by Clay Minerals.  EPA 600/2-78-157 (NTIS
PB 287-140).  U.S. Environmental Protection Agency, Cincinnati, OH.
146 pp.

Barton, A. F. M.  1983.  Solubility Parameters and Other Cohesion
Parameters Handbook.  CRC Press, Boca Raton, FL.

Haxo, H. E., and P. J. Pick.  1986.  Determination of the  Solubility
Parameters of FMLs for Use in Assessing Resistance to Organics.  In:
Proceedings of the Twelfth Annual  Solid Waste Research Symposium: Land
Disposal, Remedial Action, Incineration and Treatment of Hazardous
Waste.  EPA/600/9-86/022.  U.S. Environmental Protection Agency,
Cincinnati, OH.  pp 132-145.

Haxo, H. E., T.  P. Lahey, and M. L. Rosenberg.  1988.  Factors in Asses-
sing the Compatibility of FMLs and Waste Liquids.  EPA/600/2-88/017
(NTIS No. PB 88-173-372/AS).  U.S. Environmental Protection Agency,
Cincinnati, OH.  143 pp.
Barton, A. F. M.
75(6):731-753.
1975.   Solubility Parameters.  Chemical Reviews
                                    367

-------
24.  Leo, A. and C. Hansch.  1970.  Linear Free-Energy Relationships Between
     Partitioning Solvents Systems.  Journal of Organic Chemistry 36(11):
     1539-1544.

25.  Leo, A., C. Hansch, and D. Elkins.  1971.  Partition Coefficients and
     Their Uses.  Chemical Reviews, Vol. 71, No. 6.  pp. 525-554.

26.  Daniels, F., and R. A. Alberty.  1961.  Physical Chemistry.  2nd
     Edition.  John Wiley and Sons, NY.

27.  Haxo, H. E.  1988.  Transport of Dissolved Organics from Dilute Aqueous
     Solutions Through Flexible Membrane Liners.  In: Proceedings of the
     Fourteenth Annual Solid Waste Research Symposium: Land Disposal,
     Remedial Action, Incineration and Treatment of Hazardous Waste, May
     9-11, 1988.  U.S. Environmental Protection Agency, Cincinnati, OH.
     21 pp.

28.  Haxo, H. E., and T. P. Lahey.  1988.  Transport of Dissolved Organics
     from Dilute Aqueous Solutions Through Flexible Membrane Liners.
     Hazardous Waste and Hazardous Materials.  5(4):275-294.

29.  Riddick, J., and W. Bunger.   1970.  Techniques of Chemistry, Volume
     II - Organic Solvents, Physical Properties and Methods of Purification.
     Wiley-Interscience, NY.
                                    368

-------
        GEOSYNTHETIC CONCERNS  IN LANDFILL LINER AND COLLECTION SYSTEMS

        by: Robert M. Koerner, Arthur E. Lord,  Jr.,  and Yick H.  Halse
           Geosynthetic Research Institute
           Drexel University
           Philadelphia, PA  19104
                                  ABSTRACT

     The  use of geosynthetic materials in landfill liner and closure  systems
is commonplace. This use includes the containment  of  all types of solid waste
materials,   e.g.,  hazardous,  municipal,   industrial,   ash,   low  level
radioactive, and heap leach, ores. The geosynthetic materials involved  include
the following;
     •  geomembranes  or flexible membrane liners (FMLs),
     •  geotextiles as filters and cushions,
     •  geonets  as drains,
     •  geogrid  reforcing elements, and
     •  geocomposites for various  uses.

While it is felt that current systems can indeed be designed and constructed
with confidence, there are  certain  aspects  in need of further investigation
and/or clarification. This paper highlights several of them. Included are an
investigation of FML behavior in anchor  trenches,  stress cracking behavior of
HOPE seams,  flow  rate  reduction in  geonets,  and  particulate  and biological
clogging of geotextile filters.  Some amount  of  data is presented  on each
topic but,   clearly,  the paper represents research-in-progress from the point
of view  of  definitive guidelines.  The  goal  of each  individual  topic is to
generate  a data  base  so  as  to   avoid long-term  problems  when  using
geosynthetics in association with waste  containment facilities.

                          INTRODUCTION AND OVERVIEW

     Current waste  containment  facilities  represent  a bevy of geosynthetic
materials.  Included in the group  are  FMLs  which  form  the primary and/or
secondary liners against escaping liquids, geonets which are commonly used as
drainage materials  in  primary and/or secondary leachate  collection systems
and  geotextiles  used as  filters to allow leachate  to enter  into  drainage
collectors.  When the waste containment facility is doubly lined,  as are many
hazardous,  municipal and industrial landfills, the cross  section of Figure 1
                                    369

-------
is common. Here one can itemize the following geosynthetics and their primary
functions (progressing from the waste downward);
     •  geotextile filter beneath the waste,
     •  geonet as a  primary  leachate collector  on side slopes,
     •  geotextile filter around perforated pipe drains,
     •  geotextile cushion above primary FML,
     •  the primary  FML (i.e.,  a geomembrane),
     •  geotextile separator between primary clay and geonet drain,
     •  geonet as secondary  leachate collector,
     •  the secondary FML  (i.e., a geomembrane), and
     •  geotextile cushion against soil subgrade

Thus as many as nine (9)  geosynthetics are  used  in  the liner system shown.
When adding the various geosynthetics placed  in the cap, or closure, of such
facilities,   it  is  easily seen that  these materials  play a key  role in the
proper functioning of the  system.  This  is tantamount  to  stating that they
must be properly evaluated, selected,  designed, and installed.

     Toward  this end we have selected to investigate  a number of areas which
are  felt  to  be of  concern.  They are = identified on  Figure  1  and will be
described in the sections to follow.

1.   FML Behavior  in Anchor Trenches - As shown in Figure 1 at Location "1",
FMLs will typically come out  of  a lined facility and terminate around the
perimeter of  the site. There  is  a short horizontal runout length and then a
vertical drop into an anchor trench.  The lengths involved can be  estimated by
available design models (1,2)  but their evaluation awaits field verification.
As an  intermediate step  we have constructed  a large  scale test facility to
model  the situation in  the  laboratory.  A  steel reinforced wooden  box of
internal size 3 ft. height by 3ft.  width by 6 ft.  length has been constructed
which  can contain various  anchor  trench configurations,  see  Figure 2. The
jacking system to which the outside of the FML is  attached  can be oriented so
as to  exert  a downward force thereby simulating an actual situation. Normal
Stress is exerted by an air-bag system to mobilize overburden pressures up to
1500  lb/ft2.  While not necessary  for the perimeter  anchor  trench shown in
Figure 1, these high normal stresses are mobilized on the interior berms of
discrete cells when a zoned landfill expands laterally and  then vertically.

     Strain  gages  are attached to  the  FML test specimens at four locations.
The initial data indicate that the largest FML stresses are in the portion of
the FML closest to the pullout force. These  stresses  rapidly dissipate with
increasing  distance  into  the test box  as seen  in  the curves  of  Figure 2
moving from  locations Gl  through G4.  Obviously,   soil friction  plays an
important role in  this  stress dissipation.  Also note that the FML in the
vertical anchor trench is essentially non-stressed, i.e. gage number G4. This
has been  typical of tests  conducted to  date.  The  planned experiments  include
30 and 60 mil HDPE,  40 mil LLDPE, 30 mil PVC and 36 mil CSPE-R.  Normal stress
will  vary from 100  lb/ft2  to 1500 lb/ft2. Results will be  compared to the
various  design models  so as to  verify,  modify  or refute  the  existing
literature.
                                     370

-------
2.   Stress Cracking Behavior  of HOPE Seams - ASTM  defines  stress cracking
as,  "an internal or external rupture in  a  plastic  caused by tensile stress
less than  its short-term mechanical strength."  Thus  any mechanism which cause
premature  failure  of  the FML  can fall  into  this  general  category,  e.g.,
environmental stress cracking,  residual stress  cracking,  scratches  and cracks
resulting  from grinding  marks,  geometrical irregularities,  and/or fatigue
failure. Many  (if not  all) of these situations are encountered at HDPE liner
seam locations.

     An evaluation  of  HDPE seams  was  initially addressed using  a modified
version of  ASTM D-2552  test  method  which immerses dumbbell-shaped  test
specimens  in a surface wetting agent at an  elevated temperature. A number of
HDPE  seam types were evaluated  under  different   constant stress  levels
resulting  in a large  number that  cracked  (3), see Tables l(a) and (b)  for
test results of 168 and 1000 hour test durations,  respectively.
     TABLE l(a)  - LABORATORY STRESS CRACKING RESULTS FOR 168 HR.  DURATION
Type of Seam
fillet extrusion
flat extrusion
hot wedge
hot air
ultrasonic
Number of
Tests
179
40
60
80
40
Elastic
74
15
25
27
28
Results
Plastic
13
19
19
22
5
Cracked
92
6
15
31
7
Percent
Cracked
51
15
27
39
25
    TABLE l(b) - LABORATORY STRESS CRACKING RESULTS FOR 1000 HR. DURATION
Type of Seam

fillet extrusion
flat extrusion
hot wedge
Number of
Tests
20
20
20

Elastic
14
11
8
Results
Plastic
1
0
0

Cracked
5
9
12
Percent
Cracked
25
45
60
To be  emphasized is  that these  laboratory tests are  index tests  wherein
stress relaxation of  the polymer cannot  occur  and that they are  conducted
under very harsh conditions.  The  extent of the problem in the  field is  not
known,  although it is known to exist at a few surface impoundments  where  the
FML was exposed to the atmosphere <4>, see Figure  1  at  Location "2".

    Work is  currently ongoing  in  evaluating  various  seaming procedures
(e.g.,  temperature,  pressure and  time),  determining  residual  stresses,
evaluating cracking initiation mechanisms, measuring crack growth  rate in
various  HDPE  formulations, and  attempting to  optimize  seam geometry.  The
ultimate goal is to  assess the magnitude of the  stress cracking  problem,  its
optimization and/or  its remediation.
                                     371

-------
3.   Flow Rate Reduction of Geonets -  The  intrusion  of a geotextile  (or any
other material) into the flow apertures of a drainage net will surely reduce
its flow rate capability. While a number of cross  sections can be envisioned,
the  situation of  a clay over  a geonet with a  geotextile  separator between
them is of great concern, see Figure 1  at Location "3".

     Flow rate tests using  a  cross  section  consisting of  a geonet between two
60 mil HDPE liners versus the same type of  geonet  between a  60 mil HOPE liner
beneath  it  and a  geotextile/clay layer above  it have  been  conducted.  The
geotextile used as the  separator was  a needle punched, continuous filament,
nonwoven  fabric of 8.0  oz/yd2 mass per unit  area.  The test  procedure was
performed in  accordance with ASTM D-4716. Results are given in Table 2 where
one  can  see  flow  rate decreases of 19%  to  41%  depending upon the hydraulic
gradient and applied normal stress. Clearly,  intrusion  of the geotextile/clay
is occurring as it spans  the open apertures  of  the geonet. Note  that the
geotextile  acts in a  membrane  reinforcement  mode and  must  do so  for the
design lifetime of the  system.  If it  fails,  the clay will immediate extrude
into the  geonet thereby blocking  all  flow.  Thus long term creep  tests are
warranted for many situations.

  TABLE 2 - FLOW RATES (IN GAL/MIN-FT)  AND  REDUCTIONS  (IN %) FOR DIFFERENT
                       GEONET DRAINAGE CROSS SECTIONS
Normal
e>4-
(Ib/ft2)
5000



10,000



15,000



Cross


HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
HDPE (both sides)
GT-Clay (one side)
Difference
Reduction
Hydraulic Gradient "i"

i = 0.25
1.6
JUJi
0.3
19%
1.4
•J-l
0.3
21%
1.2
ILu2
0.3
25%

i = 0.50
3.7
2^1
1.0
27%
3.4
2^3.
1.1
32%
3.2
JU&
1.4
44%

i = 1.00
7.1
±*&
2.3
32%
6.4
AJ.
2.3
36%
5.6
3.3
2.3
41%
 4.   Parfciculate  Clogging  of  Geotextile  Filters  -  Shown in  Figure 1  at
 Locations  "4"  is  a geotextile covering a primary leachate  collection  geonet
 on  the  sideslope and  drainage  gravel on  the bottom.  In both  cases  the
 geotextile must act as a filter. Hence, it must provide adequate  permeability
 (thus  open voids) and  soil retention  (thus tight  voids),  i.e., a  balanced
 void structure must be achieved. While these considerations  can be  adequately
 handled by proper design,  there  is  a third necessary condition, that  being
 prevention of long-term clogging.
                                     372

-------
    The  customary approach toward the assessment of geotextile  clogging is
by the  gradient  ratio  test,  CW-02115.  This  U.S. Army  Corps of  Engineers
developed  test,  however,  was  directed  toward  sandy  soils  and  woven,
monofilament geotextiles of relatively high open areas (5). For the very fine
sediment carried by leachate,  the authors prefer the  use  of long-term column
tests  (6,7).  Here the  actual  cross section  being  considered is  simulated
using  site-specific cover  soil,  the  candidate geotextile  and actual  (or
simulated)  leachate.  The  long  term  flow through  this  cross-section  is
measured over a period of time. As seen in Figure 3,  the  flow rate initially
decreases until a transition time is reached.  This is due to soil compaction
and an  initial  "tuning" of the  geotextile  to  the  upstream soil and  the
permeating liquid. Beyond this point, the  flow  rate  either becomes constant
(thus  equilibrium is  established),  becomes  uncertain  (which  requires
continued testing), or decreases to  zero  (hence the system is clogged).

    A  broad base study is ongoing in this  regard  using four  geotextiles
(woven  monofilament,   nonwoven  melt-bonded,  lightweight  needle-punched
nonwoven, and heavyweight needle-punched  nonwoven); four conditions above the
geotextile  (no soil, Ottawa sand, clayey silt and a  combination  of sand and
clayey silt);  and two  permeating liquids  (water  and sediment  laden water).
Upon establishing one  of the trends shown  in Figure 3,  the  system will be
epoxy-set,  properly sectioned and microscopically viewed  so as to understand
the mechanisms involved. The goal of this  study  is a laboratory test  matrix
illustrating the behavior and proper design against fine particulate clogging
of geotextile filters.

5.  Biological  Clogging  of  Geotextile Filters  -  To  be sure,  municipal
landfill leachates  have many viable (active)  bacteria present in  them,  see
Figure 4. The geotextile filters  shown  in Figure 1 at Locations "5" interface
with these bacteria. Columns of the type shown in Figure  3 have  been  set up
at six  municipal landfills under aerobic  conditions resulting;  in the flow
rate trends shown in Table 3(a). Relatively large  reductions over the elapsed
times are indicated, e.g.,  50%  reduction  (or higher)  occurred in one-third of
the test materials. A separate  phase of the study was also aimed at anaerobic
bacteria clogging of  geotextiles  which produced the results  of  Table 3 (b) .
Here the flow  rate reductions  are much less  than noted previously, but are
nonetheless quantifiable in all cases of 7 months or  longer.  See reference 9
for additional details.

    This  study is also ongoing into a second phase  where  backflushing with
water and  with biocide-treated water will be performed.  The  resulting flow
rate response will be evaluated.  If  flow  is reconstituted to its original, or
near to original,  flow  rate  the  recommendation will  be  forthcoming  to
incorporate such  procedures  into  the  design of  the  collection  system.  Such
designs are certainly possible  and,  if  justified, should be implemented.

                           SUMMARY AND CONCLUSIONS

    Numerous types  of geosynthetic materials are currently seeing widespread
use in  waste containment facilities.  Use  of these  materials  has led to  a
certain  degree of  confidence which  is augmented by the installation  of
redundant  liner  and  leachate  collection  systems.  Due  to their   rapid
                                      373

-------
 co

 in
 m
 d>
-H  r-i
 S   W
    •P
 w   to
\p   a)
 (0  4J
 0)
4J  T)
     (0
 0)   0)

^•^
 M   tn

 S  -H
 O  H
H  H

           Cu  Cn
           Cu  o
           I  rH
                                    id H
                                    *  3
                                       <4H
                                    o
                                   -H -P
                                    1-1  10
                                   X>  <1)
                                    (0 JJ
                                   <4H *
                                                                             H
                                                      fa

                                                      O
                                                      HI
                                                         co
                                                                            CO
                                                          «.
                                                                                                       I   rH
                                                                                                             df>  dP O  O
                                                                                                      IT) O  O  O

                                                                                                       I  rH  t-l  H

                                                                                                          I   I    I
                                                      dp dp dp  dp O  U
                                                      m o o  o --^  ~~.
                                                       I  rH rH  rH S  2
                                                                                   O  O


                                                                                   S3 ^
                                                                                                      dP dp  6IP o\»
                                                                                                      o in  o o
                                                                                                      I   I    I   I
                                                                                       co  to  w   to
                                                                    rH  O
                                                                    rH  rH CT>
                                                                                                                    in
                                                                                                      r- r-
                                                                                                             00 CO  00  00
                                                                                                      rH rH  in
                                                                                                                00  OO
                                                                                                                        OO
                                                                                                             eg in  vo  CN
                                                                                                     rH  H  rH
                                                                                                                    VO
                                                      rH eg  ro i1  in  vo
                                                      I   I    I   I   I    I.
                                                                                                                                a<
                                                                                                         CO
                                                                                                                                Q.
                                                                                                                               •o
                                                                                                                               o
                                                                                                                               II   M
                                                                                                                               _  <1>
                                                                                                                               B  4J
                                                                                                                                   a.
                                                                                                                                  H
                                                                                                                               IS CU
                                                               374

-------
development,  however,  there are areas where additional inquiry is warranted.
Some of these areas have been the  focus of this paper.

     •  FML  anchor trench behavior  and mobilization mechanisms will soon have
       a  meaningful  data  base via  large  scale laboratory evaluation.  A
       rigorous critique  of available design methods will result.

     •  Stress cracking of HDPE field  seams  appears to be strongly related to
       the  workmanship involved  in  making  the  seam.  Strict  construction
       quality control (10)  should alleviate the concern for most situations
       except possibly the exposed berms  of surface impoundments.  In  these
       areas   proper  design  with  sufficient  expansion  and contraction
       capability must  be  considered.(11)

     •  Flow rate  reduction of  geonets  via  intrusion of materials  above  and
       below  the  net's apertures  or openings  is very  quantifiable.  Proper
       design of  the  geotextile  or geomembrane  above  and  below the net is
       necessary. The  laboratory simulation test  is  available  to  assess  the
       adequacy of  the  design.

     •  Fine  precipitate  clogging of  geotextile  filters  associated with
       leachate  collection systems  is of concern for those  leachates with
       high particulate content. A laboratory simulation test is proposed  and
       a  broad based series of  experiments is ongoing. If such clogging is of
       serious concern, backflushing might  be required for  remediation  and
       reconstitution  of  flow.

     •  Biological  clogging  of geotextile filters associated  with  leachate
       collection  systems is  a  likely  concern for  municipal  landfills.
       Clearly,  biocide  treatment either within  the geotextiles or  in  the
       backflushing liquid as proposed above are possible remedial measures.
       Work is ongoing to evaluate these various remediation schemes.

                              ACKNOWLEDGEMENT S

     This project is funded by  the U.S. Environmental Protection Agency under
Project No. CR-813953-01. Our sincere appreciation is extended to the Agency
and in particular to the Project Officer, Robert  E. Landreth.

                                 REFERENCES

1. Koerner,  R. M.,  Designing with Geosynthetics, Prentice  Hall  Publ. Co.,
   Englewood Cliffs, NJ, 1st  Ed. 1986,  2nd Ed.  1989  (to appear).

2. Richardson, G.  N.  and Koerner, R. M., "Geosynthetic Design Guidance  for
   Hazardous Waste Landfill Cells  and Surface Impoundments," EPA Contract  No.
   68-03-3338, 1987, GRI, Drexel University, Philadelphia, PA.

3. Halse, Y.  H.,  Koerner, R. M. and Lord, A. E., Jr., "Laboratory Evaluation
   of Stress Cracking  in HDPE Geomembrane Seams," Proc. Aging and Durability
   of Geosynthetics, Dec. 1988, GRI,  Drexel University, Philadelphia, PA.
                                     375

-------
4. Peggs,  I.  and  Carlson,  D.  S.,  "Stress  Cracking  of  Polyethylene
   Geomembranes:   Field   Experience,"  Proc.  Aging  and  Durability  of
   Geosynthetics,  Dec.  1988, GRI, Drexel University,  Philadelphia,  PA.

5. Haliburton,  T.  A.  and  Wood,  P.  D.,  "Evaluation of U.S.  Army Corps of
   Engineers Gradient  Ratio Test for Geotextile  Performance,"  Proc. 2nd Int.
   Conf. on Geotextiles, Las Vegas, NV, Aug. 1-6, 1982,  IFAI,  pp.  97-101.

6. Koerner, R. M. and  Ko,  F.  K.,  "Laboratory Studies on Long-Term Drainage
   Capability of Geotextiles," Proc.  2nd.  Int.  Conf. Geotextiles,  Las Vegas,
   NV, Aug. 1-6, 1982,  IFAI, pp. 91-95.

7. Halse, Y. H., Koerner,  R. M. and Lord, A. E. Jr., "Filtration  Properties
   of Geotextiles  Under Long  Term Testing," Proc. ASCE/PennDOT Conf.  on
   Advances in Geotechnical Engineering, Hershey, PA, Apr.  1987, pp. 1-13.

8. Rios,  N.  and  Gealt,  M.  A.,  "Biological Clogging  Growth in Landfill
   Leachate  Collection   Systems,"  Proc.  Aging  and  Durability  and
   Geosynthetics,   Dec., 1988, GRI, Drexel University, Philadelphia, PA.

9. Koerner,  G.  R.  and Koerner,  R.  M.,  "Biological Clogging  in Leachate
   Collection Systems," Proc.  Aging and  Durability of Geosynthetics, Dec.
   1988, GRI, Drexel University, Philadelphia, PA.

10. Rollin,  A.,  "Factors  Influencing Geomembrane Seam  Quality,"  Proc.
   Geosynthetics '89,  San Diego, CA, IFAI.

11. Peggs,  I.  D.,  "Failure  and  Regain  of Geomembrane  Lining System,"
   Geotech. Fabrics Report, Vol. 6, No. 6, 1988,  pp.  13-16.
                                         GEOTEXTILE FILTERS
                                                           P-FML
                                                           S-FML
 Figure 1 - Cross Section of Double Lined Landfill Facility Often Used for
             Municipal/Industrial Solid Waste Disposal
                                     376

-------
                                                                  2'-0"
      D
01
n
* ^
10 •
400 -jw —
__ - J» H
200-1
nJ
'fa
!y


*• 	 ••"




> 	 o
a
a


+
o
*
p
D
J








O
a

«
!l>"J'™"

Q
D
3 	









H
a


* 	 	














u 	



*

c;
"~ET~















»

Q


































*

E
Q

















Fl


















c
















































...








                             1              2
                            TRUE STRflIN (7o)
                                                               El  G1
                                                               *  G2
                                                               a  G3

                                                               «  G4
             - Pullout of 60mil HOPE with 350 Ib/ft2 Surcharge

Figure 2  - Details of Test Facility and FML Test Results for Anchor Trench
          Pullout Experiments
                                 377

-------
                                                  TIME
  Figure 3 - Long-Term Flow Column and Usual Trend in Flow Rate  Data versus
             Time
Figure 4 - Viable  Bacterial  Titer  of Each  of  the  Leachate  Samples as
           Determined by Multiplying Viability by the Total Direct Count  TDC,
           after Rios and Gealt<8)
                                   378

-------
                ATTENUATION OF PRIORITY POLLUTANTS CODISPOSED
                       WITH MSW IN SIMULATED LANDFILLS

                        by:  Dr.  Frederick G.  Pohland
                             Department of Civil Engineering
                             University of Pittsburgh
                             Pittsburgh,  PA  15261

                                     and

                             Dr.  Wendall H. Cross, Dr.  Joseph P.
                             Gould and Ms. Debra R.  Reinhart
                             School of Civil Engineering
                             Georgia Institute of Technology
                             Atlanta, GA 30332
                                   ABSTRACT

     Organic  and  inorganic  priority  pollutants  codisposed  with  shredded
municipal solid  waste (MSW)  in  ten pilot-scale simulated landfill columns,
operated under single pass leaching or leachate  recycle, were capable of being
attenuated  by microbially-mediated  landfill stabilization  processes.   The
results of detailed investigations have indicated that inorganic heavy metals
(Cd, Cr, Hg,  Ni, Pb and  Zn)  were subject to a complex  array of attenuation
mechanisms  within the  MSW matrix,  including  precipitation,  encapsulation,
complexation, reduction,  adsorption and ion exchange.   Similarly,  the major
classes  of  organic priority  pollutants  (aromatic  hydrocarbons,  halogenated
hydrocarbons, pesticides,  phenols and phthalate  esters) were attenuated mainly
by  sorption,  volatilization and bioassimilation and  release  of identifiable
by-products within the leachate and gas transport phases.

     Collectively these in situ processes  constitute the assimilative capacity
of a landfill for priority pollutants,  the magnitude of which is dependent on
microbial viability during the sequential phases of landfill stabilization as
affected by loading intensity and contact opportunity.  Data are presented to
demonstrate  some of  the  principal  assimilative mechanisms  as  well  as  the
efficacy of in situ process control through leachate and gas management.  Based
upon these findings, a landfill management strategy is proposed for landfills
receiving inputs of organic and inorganic priority pollutants.
                                      379

-------
                                  INTRODUCTION
      There exist few more complex and challenging problems than those associated
 with the management of municipal and  industrial  solid wastes.   Because these
 wastes often contain toxic and hazardous substances,  they may become a threat
 to human  health and  the  environment unless proper  treatment and  disposal
 techniques are employed.   Of the array of options available for such treatment
 and disposal,  codisposal  in sanitary landfills is probably the most prevalent,
 considering the magnitude  of household  and  small quantity generator sources.
 Therefore,  it  is imperative that the efficacy of codisposal of hazardous wastes
 with MSW be properly assessed,  and that decisions on codisposal be approached
 in as scientifically and  technically sound a manner as possible.

      One approach to such  an assessment is  to examine the progress  of waste
 stabilization  with time and under various operational conditions, either in the
 laboratory or  in the field.  Laboratory or pilot-scale  investigations are often
 more cost effective and permit evaluations with greater operational control and
 less parametric variability.  Hence, pilot-scale simulations were chosen for a
 3-year investigation on the behavior and fate of selected inorganic and organic
 priority pollutants codisposed with  shredded MSW under the influence of single
 pass leaching  or leachate recycle.   This paper presents  a progress  report on
 some of the research results  to  date.

                            EXPERIMENTAL PROCEDURE


 Column Construction and Loading

      Since  the behavior and fate of toxic inorganic and  organic compounds within
 a  landfill  setting are controlled by the  related effects of various  mass trans-
 fer and removal mechanisms,  including solubilization,  sorption, volatilization,
 chemical transformation  and bioassimilation, it  was  necessary to develop  an
 experimental protocol sufficient to  achieve an adequate description and control
 of the waste mass and the gas and  leachate transport phases.   Accordingly,
 shredded MSW was selected  as the primary waste  matrix, and was augmented  by
heavy metal sludge  and selected classes of  organic  priority pollutants.   In
addition, to provide  a range of loadings under both single pass leaching  and
leachate recycle, 10 simulated landfill  columns were constructed and prepared
for operation  as indicated  in Figure 1 and Table 1.

      The  simulated  landfill  columns   were   constructed  from  0.3-cm  steel
cylinders,  0.9 m in diameter and 3 m in  height. The columns were lined with  30-
mil HOPE which  was supported by an underdrain  system composed of 20 cm of gravel
and 2.5 cm  of  sand.  The underdrain  system allowed collection of leachate  for
sampling,  discard  (single  pass  columns) or  recirculation (recycle columns)
through a perforated pipe distribution system  installed  after loading.  Moisture
addition (or recycle) was facilitated through this pipe distribution system.

     A leachate sight glass was also  connected to the underdrain  system  as well
as the headspace under the cover.  Gas  generated during landfill stabilization
                                       380

-------
                                 1.22m
                                  1.83m
                                  0.61m
  LEGEND
1 GAS METER
2 TEMPERATURE INDICATOR
3 GAS SAMPLING VALVE
4 GAS TRAP
5 CHECK VALVE
6 PRESSURE GUAGE
7 DISTRIBUTOR ARM
8 RECYCLE PUMP
9 FLANGE
10 THERMOCOUPLE
11 HOPE LINER
12 IN-LINE FILTER
13 STEEL
14 LEACHATE DRAIN
15 LIQUID SAMPLE PORT
16 LIQUID LEVEL CONTROL
17 GRAVEL, SAND, AND
 GEOTEXTILE LAYERS
18 GEOTEXTILE, SAND,
 GEOTEXTILE, AND GRAVEL
 LAYERS
19 110 V AC
20 110 V AC
 TO PUMP
21 110 V AC FROM
 LIQUID LEVEL CONTROL
22 VENT TO ATMOSPHERE
23 SHREDDED REFUSE
® BALL VALVE
                  RECYCLE UNIT
                                                              SINGLE PASS UNIT
Figure  1.   Construction and Operational Features of Simulated Landfill Columns


was collected from this system and measured with an automatic gas-displacement
meter.

     Each of the 10 columns  received 42 individual  9-kg batches of shredded MSW
placed  and compacted over an 8-hour period.   Columns 1 and 2,  designated control
columns (Table 1) , received only shredded MSW and an equivalent amount of  saw-
dust  also  added to  the other  test  columns because  of  its  use  to facilitate
homogeneity in the sludge-loaded columns.   This loading  strategy provided  five
pairs,  each pair  identically  loaded,  but  operated  with either  single  pass
leaching or leachate recycle.

      Batches of metal sludge were prepared by mixing with sawdust and augmenting
with metal oxides as necessary to deliver low, medium  and high loadings  (Table
1) .  The  organic priority pollutants (Table  1) were mixed together and added at
the  surface of  the  first 30 cm of compacted MSW to increase the potential for
detection during operations.  After column  loading, an 8-cm layer of washed pea
gravel  was added to the surface of the MSW,  the columns were sealed, and initial
                                       381

-------
3
I

53
M

£
             •§
             o-g

              •. a


             il
     II
                                                    *   =8   *
                                       ro   M
                     OJ   M   (Nl   c3
                                          >«.
                                                   fl)   —*   «4

                                                                       O) CD O) O)
                                                                          lit
                                                                       '1!
                                                                  -   2.
                                                                         I-S
                                                                       1'
                                                                        a. ro
                                                                          §
                                                                        8.-I
                                                                                         ITS
                                                 382

-------
moisture  (tap  water)  was introduced to  satisfy  field capacity and  encourage
leaching.   Once  leaching  commenced,  the  five single  pass  columns  received
periodic  moisture additions  equivalent  to  local  annual  rainfall  (123  cm),
whereas  the five  recycle  columns  were  subjected  to  leachate  recycle  with
moisture  additions only of  quantity necessary  to  replace  that  removed  for
sampling and analysis.

Analytical Methods

     Leachate and gas samples were analyzed routinely for both gross parameters
and  priority pollutants once sufficient quantities were  produced to  permit
analysis.   Leachate  samples were  subjected  to  immediate  analyses  for  pH,
conductivity,  chemical  oxygen demand (COD),  and  oxidation-reduction  potential
(ORP), supplemented by  biochemical  oxygen  demand (BOD5) , total organic  carbon
(TOG), volatile organic acids  (TVA),  alkalinity,  nutrients (N,P),  selected
anions (Cl~, SOA~2, S"2,  Br") and cations (Ca+2, Mg+2,  Na+, K+, Li+) ,  heavy metals
(Cd, Cr,  Fe, Pb, Hg, Ni, Zn) and organic priority pollutants (Table 1)  according
to standard techniques.  Gas  samples were analyzed for C02, 02, N2, H2, CHA and
volatile  organic priority pollutants using GC and GC-MS techniques.

                      EXPERIMENTAL RESULTS AND  DISCUSSION


     Since  the primary  purpose  of the research investigations was to evaluate
the  fate  of selected organic  and inorganic priority pollutants cpdisposed with
MSW  under the  influence of single pass leaching and leachate recycle, emphasis
here is placed on these pollutants as the  simulated landfills progressed through
acid formation and into the  methane fermentation phase of  landfill  stabiliza-
tion.    Accordingly,   experimental  data  are  presented to  reveal  potential
attenuating mechanisms  operative during these phases  and  contributing  to the
overall  assimilative  capacity of the landfill  environment.

Progress  of Landfill  Stabilization

     The  progress  of  landfill stabilization  can  be described by  certain
indicator parameters  (1).   For  purposes  here,  gas production and leachate COD
and  TVA  concentrations and pH were selected as  indicated  for the single pass
and  recycle columns in Figure 2. As illustrated  in this figure, acid formation
(acid phase) became prominent and was intentionally sustained for the majority
of   the  report period.   Reductions  in  leachate  strength  (COD  and TVA)  or
increases in gas production,  as  well as some increase in pH, occurred only after
the  onset  of  methane  fermentation (methane phase),  which was  initiated by
digested sludge seeding with neutralization and  took effect on about Day 720.
Thereafter,  leachate and gas quality changed dramatically,  particularly for the
control  columns (1 CR,  2 C) without heavy metal and organic priority pollutant
codisposal.

     The differences  in results between  the  individual  single pass and recycle
test columns  also  are  indicative  of loading  implications,  particularly with
respect  to  the heavy metals,  and the washout by  single pass  operations  as
contrasted  to  leachate  recycle.  The apparent release of leachate constituents
through washout contributed to a decrease in concentrations, but did not greatly
                                      383

-------
W  S0.n
+j
o

J  40.
3

15  30.-4
 n  20.-
   10.-
ti

I
  0.

100.
       »-o 1 C
       o-o 5 0
       *-* 70L
       *-* 9 OM
       «—f 10 OH
                                                o-o 2 C
                                                o-a 3 0
                                                A-A 4 OL
                                                «—» 5 OM
                                                *-* 8 OH
                                                                                  smr.LC
                                                                                o I C
                                                                                aao
                                                                                • 4 01
                                                                                « 5 OW
                                                                                • 8 OH
        200.    400.   600.    800.   1000.
            Time since loading, days
                                            1200.  0.
                                                      200.    400.    600.    800.   1000.
                                                          Time since loading,  days
                                                                                       1200.
 Figure
       2.   Changes  in Gas Production and Leachate  COD,  TVA  and pH during
            Simulated Landfill Investigations
                                              384

-------
enhance  the  extent of  stabilization in the  single pass  columns  where heavy
metal and  organic priority pollutants  were admixed.   With recycle, leachate
concentrations  and gas production  tended  to stabilize  initially  and then
decreased or increased, respectively, when methane fermentation ensued. There-
fore, the containment for conversion of these  leachate  constituents  to gas, by
better contact and controlled in situ treatment,  emerged as a major advantage
of recycle over single  pass operations.

Mechanisms of Change in Inorganic Priority  Pollutants (Heavy Metals)

     A principal focus of these investigations was the behavior and fate  of  the
toxic heavy metals in  terms  of their mobility in the leachate (or gas) phases
as landfill stabilization progressed.   Therefore, special emphasis was placed
on analyses of these metals and potential attenuating mechanisms.

Hydroxide Precipitation--

     Of  the heavy metals added to the sludge loaded columns (Table 1), only
Cr+3  was  subjected to significant solubility control by hydroxide.   Based upon
its extremely low  solubility  (pKso = 30.5),  precipitation of the chromic  ion as
the hydroxide, Cr(OH)3, at equilibrium concentrations of  chromium  on the order
of 1 mg/L, would be attained  even at pH levels as low as  5.0.  Therefore, even
during the acid phase  of landfill stabilization  (pH 5.5  to  6.0),  the mobility
of trivalent chromium would be limited by hydroxide precipitation.   This control
was demonstrated by leachate chromium analyses  (Figure 3), although some, initial
fluctuation  occurred  for  both  the  single  pass  and recycle  columns.   These
fluctuations  coincided  with  early  fluctuations in pH  and,  no  doubt, were
affected somewhat by the respective  column operations  and  the possibility of
short-circuiting.   Moreover,  since leachate chromium concentrations  decreased
sharply  to very low levels during  the  acid phase and  into  the methane  phase,
coupled with the absence of other important precipitants for chromium in typical
landfill  leachates,   hydroxide  precipitation  was  considered   the  primary
attenuating mechanism  for this heavy metal.
  30.

-------
Redox Processes --

      The ORP is particularly important  in defining  the chemical character of
the landfill environment.  As illustrated in Figure 4, negative  leachate redox
values provide reducing conditions  that mediate  the behavior  of many  of the
codisposed inorganic species.   The  impact  of this  condition  can be  either
directly through a modification of the nature  of  the pollutants with  a change
in  their mobility,  or  indirectly  by  reductive  generation   of a  potent
precipitant.
        •400 L
                                                                         SINGLE PASS
                                                                        ate
                                                                        030
                                                                        • 4 01
                                                                        + S OU
                                                                        • tl Oil
               800.    900.   1000.   1100.
                 Tlmo since loading, days
                                        1200.
                                             0.0
                                              700.
800.    900.    1000.   1100.
  Time since loading, days
1200.
           Figure 4.   Changes in Leachate ORP,  Sulfate and Sulfide  during
                       Simulated Landfill Investigations
                                       386

-------
     An  important  example of  the latter  indirect  case is  the  reduction of
sulfate to sulfide  and subsequent precipitation of very sparingly  soluble metal
sulfides.  This is  an  extremely important attenuating mechanism in the landfill
environment  since  many heavy  metals are  precipitated and  removed from  the
leachate, particularly when leachate recycle provides  an additional mechanism
by filtration through the  waste matrix.  However, the efficiency of attenuation
is also  a  function of the availability of sulfides, primarily provided by in
situ microbially-mediated reduction of sulfates.

     As  also  illustrated  in  Figure 4,  sulfate reduction to  sulfide tended to
coincide with  the  onset of methane fermentation, most  noticeably  in the least
loaded  recycle columns.   The  associated impact of available sulfide on  the
leachate lead  and zinc concentrations is dramatically illustrated in Figure 5.
Both  metals rapidly  decreased in concentration in  correspondence with  the
reduction of sulfate to sulf ide.  Therefore, the important attenuating mechanism
for these  heavy metals is precipitation  as PbS  (pKso  = 28.0)  and ZnS (pKso  =
24.8).

      Similarly,  a  direct  impact of reducing conditions on attenuation  of toxic
heavy metals is also  apparent with mercury (Figure 5).  Following an initial
peak, leachate mercury decreased rapidly to concentrations below 100  /zg/L and
eventually stabilized at about 10 /ig/L.   However, in spite of the  exceptionally
low solubility of  mercuric  sulfide (pKso  = 52.5),  leachate mercury concentra-
tions  tended   to  exhibit  no  additional decrease with  sulfate  reduction.
Collectively,  these observations  are consistent with reduction of mercuric ion
to metallic mercury, which has been shown to have a  water solubility of 5 to 30
jzg/L  (2).   This behavior has  also been documented for bottom sediments  (3),
where similar reducing conditions prevail.  Hence, reduction to volatile mercury
and  subsequent sorptive containment within the  waste  matrix would constitute
important attenuating mechanisms  for this heavy  metal.

Other Precipitants --

      During the acid phase,  when  significant  sulfate reduction  and sulfide
generation were not in evidence,  control  of metal solubility may have involved
such  generally  abundant  anions  as chloride,  sulfate,  carbonate  and  possibly
phosphate.  Within the landfill environment,  these potential precipitants may
have  only transient significance,  particularly where  leachate  recycle  and
promotion  of  sulfate reduction  and reducing  conditions  are provided under
controlled operating conditions.

Heterogeneous Physical-Chemical Processes --

      Codisposal of alkaline metal sludges with MSW presents the potential for
 generation  of chemical  microenvironments within  the waste matrix,   and the
 opportunity for a complex array of heterogeneous  (liquid/solid) reactions  of
major  importance  to  the overall  attenuation  of heavy metals.   These micro-
 environments  provide a  source  of alkalinity and  acid neutralizing  capacity
 which will lessen  leachate metal mobility, while  the hydroxide sludge will react
 with leachate anions such  as  sulfide,  phosphate  and sulfate  to develop  an
 encapsulating  layer capable of impeding dissolution of the  sludge  metals  into
 the  leachate.  Therefore,   these alkaline microenvironments  will contribute
                                       387

-------
                                 BtCYCtC.
                                  1 CD
                                  6 OS
                                  7 OtR
                                  9 OUR
                                 10 OHR
    0.
         200. •  400.   600.    800.   1000.
             Time  since loading, days
1200.
                                             0.
                                           2000.'
        n    700.   400.    600.   800.   1000.  1200.
                                            1500. -
                                            1000,-
                                            500.
                                              0.
             200.   400.    600.    800.   1000.
                 Time since loading, days
1200.
         Figure 5.  Changes  in Leachate Lead,  Zinc and Mercury  during
                    Simulated Landfill Investigations


effectively to in situ attenuation during the acid phase of landfill stabiliza-
tion  when  it  is most needed, whereas  this contribution  will  diminish  as
encapsulation  proceeds  and  other  attenuating  mechanisms,   such  as  sulfide
precipitation, gain prominence.

     The  MSW  itself also  provides abundant surface  area for  sorptive  inter-
actions with leachate constituents.  These processes include physical adsorption
on  the solid  surface,  ion exchange  (particularly  on soil),  chemisorption  by
                                       388

-------
complexation with insoluble  ligands  associated with  the MSW,  and physical
containment  in  transiently  stagnant  void  volume liquid.     In addition,
complexation of metals by soluble ligands, particularly with moderate  to high
molecular weight humic-like substances, may enhance sorption by  incorporating
the metal into  a relatively hydrophobic molecule.  Current analyses on leachate
aromatic hydroxyl concentrations and molecular weight distributions are being
provided  to help  determine   the  significance  of  this  latter  attenuating
mechanism.

Mechanisms of Change in Organic Priority Pollutants

     Since  the  landfill  environment  provides  numerous possibilities   for
contaminant transport and transformation, an  equally important focus of these
investigations  was  the  behavior and fate of  the organic priority pollutants
during both acid and methane  phases of  landfill  stabilization.   In this case,
mobility and possible attenuation  were  not  as clear as for the  heavy  metals.
However,  scrutiny of the  data  (Figures  6  through  8)  suggests  that  the  12
compounds added to the test columns (Table 1)  can be divided into four  general
groups in terms  of their relative mobility and reactivity.

     Four  of  the compounds,   dieldrin,  hexachlorobenzene, bis-2-ethylhexyl-
phthalate  and  7-1,2,3,4,5,6-hexachlorocyclohexane (lindane) essentially have
not  emerged in  either  the leachate  or gas  phases  although in the   case  of
lindane, three samples from over 600 analyzed  revealed its presence at about 10
to  20  /ig/L.     Three  of  the  compounds,  dibromome thane,  2-nitrophenol   and
nitrobenzene,  appeared  early  in appreciable concentrations and then decreased
to below detection  limits  (Figure  6), particularly  after methane fermentation
had  been established.    In  contrast,1,2,4-trichlorobenzene has  been detected
regularly in the leachate (Figure 7) , but  at relatively low levels  (< 1 mg/L) .
The  remaining  four  compounds, naphthalene,  1,4-dichlorobenzene,  2,4-dichloro-
phenol  and 1,1,2-trichloroethene,  appeared  in  the   leachate  in   varying
concentrations  throughout the project period  (Figure 8).

     Headspace  gas analysis has indicated the presence of only three of the test
compounds;  trichloroethene  fairly  regularly at 1 to 20  jug/m3, and naphthalene
and  dichlorobenzene sporadically  at  equally  low concentrations.  Moreover,
evidence  of biodegradation  of  at  least two of the added organic priority
pollutants has  been obtained.  Bromide ion (Figure 9) has been detected in the
leachates  from all  test columns;  slowly leaching from the single pass  columns
due  to  washout, while  remaining relatively constant in the leachates of  the
recycle  columns (-150  mg/L).   Since no leachate bromide  was detected  for  the
control columns, and appeared from the test columns coincident with  the decrease
in  dibromomethane  (Figure 6),  microbially-mediated debromination  would  be
suggested.  Likewise, vinyl chloride (as high  as  300 /*g/L) has been detected in
the  headspace  gas  of  the  test  columns   after  reductions   in   leachate
trichloroethene were observed, suggesting partial bioconversion  to this  known
intermediate  (4).

Primary Attenuating Mechanisms --

     In  addition to the suggestion of bioconversion of some of  the  organic
priority pollutants into detectable intermediates and reaction products, several
                                      389

-------
                                                                        SINGLE PASS.
                                                                       o 2 C
                                                                       0 3 O
                                                                       » 4 OL
                                                                       + 5 OU
                                                                       * B OH
     0.
  200.   400.   600.   800.   1000.
      Time  since loading, days
                                       1200. 0.
                                               200.    400.    600.    800.   1000.
                                                   Time since loading, days
1200.
Figure
6.  Changes in Leachate, Dibromomethane, 2-Nitrophenol and Nitrobenzene
    during Simulated Landfill Investigations
other potential attenuating mechanisms could be envisioned.   Inspection of the
concentration data,  as well as the mass of  each  added compound released from
the test columns (Table 2) , provides some basis for deducing possible mechanisms
controlling relative mobility in the leachate or gas.  Although £here does not
appear to be  a variation in perceptible patterns  of release of the organic
priority pollutants between the single pass and recycle columns,  the operational
procedure in  either  case would determine whether  the leached compounds would
                                      390

-------



UJ
1
c_>
LJJ
O
i
CO
o
O£.
Q

O
Ul
0
&
111
OK
01 CO
UJ 5
GO —1
CO O
O O
(O — 1
rf II
rz* s^
r— O
^ Z
	 | ^
_I _j
O
a. a
UJ
^ <
|i
a.: co
UJ
u =

3 o
0£ 0£
O "-

CO
UJ
(__
_J
13
Q
*>
o

LL
O
CO





UJ
3
i—








ujapug;
a^BjBtpqd


3T19 Z U9 q
OJOTUOBX9H



suBpufT;






euazuaq
— o aoxqoTax— -17 * 3 ' j


•[ouaqdoj^Tji-^


auazuaq

••OJt OTUO"E(T™^7 T



auazuaaoaaxM






3Ua dBN
+



d OTUOTfT-17'7






aUBUaaOJOTUOTJT






*



A^j^uapi uuinjog



II II II II II



II It II II II



ii ii ii *4~ r- ii
" O

V
^/


it t i O '"* O *•* O '•^
in *^ CM CM N. %o
o o o


t i ii ^D ^^ C3 ^^ O ^^
v^ vx v^

^D ^^ CZ3 *"* C3 *"*
^D ^— CD O* ^5 r1*
ii ii C\l *^ O • O •
» » S.X >^



O ^N 0 ^ O '^
ii ii «— ^" r^in in'O
»— • in » N- •





C5 ^% ^D ^^ O ^^ C3 ^^ CD ^^
CDO OO CD -J- O O OCM
^T ,-T ^^ ^^



O • *"* O ^N O *^
SN- O sf O *-
»- . ro •

v— vx C? ^^ iN- ^^
. . .


§x*x ^3 ^N CO ^^
CO ON- O CM

*~ ^ O. ^f O> ^



^D ^^ ^D ^^ CD ^^
ON- O CD O CO
ii ii OK) CD jTO CD CM
•S* %O -*
** to ro


0 CJ O O S
T- CM ro st in


it 11 ii ii ii



11 ii it 11 ii


^•n' " XX
II II 11 *Q «— ""d" «—
• O • O
«- O O

N.X V
^^


^D ^^ ^D ^* CD ^"* CD *"* CD ^^
%oro I'-NO -OCM oro ro«—
ro • o CM • sa- • «- •


§**\ O ^^ O *"* O '•^ O *•*»
^o co CM o «~ in ro r*- ^3
-* ~

O«— OTO OCM OCM 0*4-

»t«— »*— •>•:— -*v- ».«—



CM-4- ON- O-« OCO Oo «- in • to • •* • IO,'~

N- O^-* v— v-' v-:^^ in



o *** o ^^ o ^^ o ,^* o ^^
ot- OTO OTO o*r oin
o« in*— ^is" ^o*" r^«—

t*» T- in co ^ *O t*-
•r- v^ T- a- CON- ^CM or- OCM
^^ ^^ TO t^> ^O
in co TO «- • CM

a OL ai
CZ —1 31 £ 3=
o o o o o
-o N. co o o

























•
c
3
"o
o
~o
L.
c
o
0
c
"^
T3

UJ
CA
o
(/)
w
l_
o
• • M-
't-> "D
CO  ?S 3
01 t- >
£ ° ~
O O> 55
•M
c -c c
c 0) "ra
- O l_ •!->
fi. 0 J=
S Q. Q-
O 
-------
  3.0
 .2,0
s>

I
B9
J3

oW

o
  0.0
                          RECYCLE.
                         Q  1 CR
                         O  6 OR
                         »  7 OLR
                         +  9 OUR
                         « 10 OHR
 SINGLE PASS.
D 2 C
030
* 4 OL
+ 5 OM
« 8 OH
    0.
200.   400.    600.    800.
    Time since loading, days
                                       1200.
                                                 200.   400.    600.    800.   1000.
                                                     Time since loading,  days
      1200.
               Figure 7.   Changes in Leachate 1,2,4-Trichlorobenzene during
                         Simulated Landfill Investigations
      escape due to washout (single pass) or be more  likely  subjected to additional
      attenuation opportunity with recirculation (recycle).

           Evidence to date indicates  that  the organic priority pollutants  will be
      influenced by the equilibrium  conditions established at the point of  contact
      between the gas, liquid and solid phases.  Supplemental  studies with the test
      compounds have demonstrated that other leachate constituents  have enhanced the
      solubility of the more hydrophobic molecules, but that physical  sorption was
      a primary attenuating mechanism.  Moreover, a strong correlation of sorbability
      existed with the octanol/water partition coefficient (Figure 10).   Therefore,
      sorption of  the test compounds  on the  waste matrix  could  be  attributed to
      hydrophobic interactions and removal  from the aqueous phase onto  nonspecific
      surface sites, particularly such MSW constituents as fats, oils,  waxes, biomass,
      humic-like substances,  lignin,  plastics and leather.

           In the final analysis,  only soluble, less hydrophobic compounds would be
      expected to elute from the wastes in the  test columns.  Hence,  compounds would
      be  expected to emerge in the approximate order of increasing affinity  for the
      waste.    However,  the  degree   of MSW   stabilization  and  the  corresponding
      microbially-mediated phase  (acid or  methane) would influence  this pattern,
      i.e. ,  pH  changes  could impair  sorption  of ionizable  compounds and the more
      stable the waste,  the  greater the mobility.  Hence, upon  completion   of  the
      stabilization process through methane fermentation and into  final  maturation,
      the liquid (leachate) transport phase should be  removed  to preclude continued
      fractionation of many of the residual organic priority pollutants from the waste
      matrix.  With leachate containment and recycle, such operational control can be
      easily accommodated.

                                 SUMMARY AND CONCLUSIONS
           Simulated landfill  columns,  containing  shredded MSW and  any array  of
      inorganic  and organic  priority  pollutants  have been  shown  to  possess  a
      significant  assimilative capacity due to a variety of attenuating mechanisms.
                                           392

-------
200.
           400.   600.'   800.   1000
         Time  since loading, days
                                   1200. 0.
240.     480.     720.     960.
  Time since loading,  days
                                                                         1200.
Figure  8.   Changes  in Leachate Naphthalene, 1,4-Dichlorobenzene,
            Trichloroethene  and 2,4-Dichlorophenol during
            Simulated Landfill Investigations
                                     393

-------
   400.
   300.-

              o
              D  Soil/Sediment Data
                   (Karlckhoff, 1981)
                                                                      -b-
                 R»lu«« Beat Fit Curv* Plu* 86 Percent
                   Confidence Interval
                                     D
           1234

                     Log  Octanol/Water  Partition Coefficient

           Figure 10.   Relationship Between Log Octanol Water Partition
                        Coefficient and Log  Sorption  Partition Coefficient
                        Normalized to Carbon Content.
                                               394

-------
     Metal mobility  was minimized by microbially-mediated  physical and both
homogeneous   and   heterogeneous  chemical  processes,   including  hydroxide
precipitation, direct and indirect reductive removal, sorption, and retention
in  stagnant  intersticial  void  liquid.    Associated  specific  attenuative
mechanisms, such as encapsulation, tended to function during  operational phases
when other mechanisms, such as sulfide precipitation, were unavailable.

     Organic priority pollutants appeared to be removed primarily by sorption
and/or biodegradation, with identifiable metabolic intermediates and reaction
products being detected. The sorptive capacity of MSW results in extremely Ipng
contact  times within  the  landfill  setting,  and  enhanced  opportunity  for
acclimation and further degradation of more recalcitrant compounds.

     Leachate  recycle  offers  the advantage  of  not only  retaining leached
materials within  the landfill system, but also  providing better contact and
redistribution of components contained in the leachate, thus  facilitating and
beneficiating the overall efficiency of stabilization.

     Landfills containing MSW and operated with leachate recycle, can provide
opportunities for codisposal of organic and inorganic priority pollutants with
reasonable assurance that mobility and potential release can be controlled to
safeguard human health and the environment.
                                  REFERENCES
1.  Pohland, F. G. Dertien, J. T., and Ghosh, S. B.  Leachate and Gas Quality
    Changes During Landfill Stabilization of Municipal Refuse.  In: Proc. 3rd
    Intl. Symp. on Anaerobic Digestion, Boston,  MA, 1983.

2.  Hughes, W. L.  A Physicochemical Rationale for the Biological Activity of
    Mercury and Its Compounds.  Ann. N.Y. Acad.  Sci.  65:454, 1957.

3.  Mercury in the Environment. An Epidemiological and Toxicological Appraisal.
    L. Fribery and J. Vostal, [Eds].  CRC Press, Cleveland, OH,  1972.

4.  Wilson, B. H., Smith, G.  B.  and Reas, J.  F.  Biotransformation of Selected
    Alkylbenzenes and Halogenated Aliphatic Hydrocarbons in Methanogenic Aquifer
    Material:  A Microcosm Study,  Environmental Science and Technology 20, 997-
    1002,,1986.

5.  Karickhoff,  S. W.   Semi-Empirical Estimation  of Sorption of Hydrophobic
    Pollutants on Natural Sediments and Soils, Chemosphere, 10(8),  833,  1981.
                                     395

-------
               SITE DEMONSTRATION OF HAZCON PROCESS
                        Paul R. de Percin
             SITE Demonstration and Evaluation Branch
              Risk Reduction Engineering Laboratory
               U.S.  Environmental  Protection Agency
                         Cincinnati,  Ohio
ABSTRACT

     In October 1987 the HAZCON stabilization process was tested
and evaluated at the Douglassville, Pa. superfund site.  There
was extensive physical and chemical testing of the untreated and
treated waste samples.  Stabilized wastes were also stored
underground on site as a field durability test.  After 270 days
(July 1988) these wastes were dug up and the samples obtained
were tested for any changes in chemical and physical properties.
The wastes were reburied and will be sampled again in July 1989.

     The durability of the treated samples was determined by
comparing-the 28-day and the 270-day test results.  Test results
of the HAZCON 270-day samples indicated a slight loss in long-
term durability, but because of the data scatter this may not be
statistically important.  There was some loss in UCS strength in
the low oil and grease samples (<5%) , but the high oil and grease
samples (16-25%) the UCS strength increased.  TCLP leachate
concentrations for lead remained low (ppb) ; and the VOC, BNA and
PCB leachate concentrations appeared lower than the 28-day sample
results.  A noticeable decrease in the porosity of the samples
was observed.  Overall, there was not a major change in sample
characteristics and some durability was demonstrated.
                               396

-------
INTRODUCTION

     The HAZCON solidification/stabilization treatment process
was demonstrated at the Douglassville, Pennsylvania superfund
site under the Superfund Innovative Technology Evaluation (SITE)
program in October 1987.  Reliable field performance and cost
information were the major objectives of this demonstration.  A
secondary objective was to study the long-term durability of the
treated waste under field conditions.  The treated wastes
resulting from the HAZCON demonstration were stored on-site,
underground and isolated from the untreated waste.

     One aspect of solidification/stabilization processes is that
the waste is not destroyed, but macro- and micro-encapsulated
(trapped)  or  fixated  (reacted) in/to the stabilizer matrix.
Thus, in future years there is potential for the contaminants to
leach from the treated waste as the stabilizer degenerates.
Long-term durability or lifetime treatment effectiveness is a key
questions both regulatory agencies and responsible parties have
about stabilization.

     After being stored on-site for nine months (July 1988),
sample cores were bored from the treated wastes and the samples
were tested using physical and chemical procedures.  The sampling
and the sample testing procedures were the same as the procedures
used to obtain and test the one month samples.  One and nine-
month sample test data were compared to determine if any
degradation of the treated waste was evident.

BACKGROUND

     The Douglassville, Pa superfund site is a former oil
recovery facility covering about 50 acres near Pottstown, Pa.
There are six contaminated areas: two large lagoons referred to
as lagoon north (LAN) and lagoon south (LAS), an oily filter cake
storage area (FSA) , an oil drum storage area (DSA) , an area where
generated sludge was landfarmed (LFA), and the plant facility
area (PFA).  The major contaminants at this site are oil and
grease, semivolatile organics (BNAs) and lead; minor contaminants
were PCBs and volatile organics (VOCs).

     The HAZCON proprietary solidification process involves the
mixing of hazardous waste material and cement with a patented
nontoxic chemical called Chloranan.  Chloranan is claimed to
neutralize the inhibiting effect that organic contaminants
normally have on the crystallization of cement-based materials.
For this treatment, the wastes were immobilized and bound into a
hardened, leach-resistant concrete-like mass.  Waste from the
Douglassville superfund site was selected because of the high
levels of oil and grease (up to 25%) and lead (up to 2.2%)  in the
FSA waste.   This combination was considered very difficult to
                             397

-------
process successfully.

     During the HAZCON SITE demonstration at Douglassville, five
cubic yards of treated waste was produce for each of five wastes.
The sixth waste (LAS) generated, twenty-five cubic yards of waste
as a test for equipment reliability.  Waste was mixed with the
Chloranan at a ten to one  (10:1) ratio and with the portland
cement at a one to one (1:1) ratio.  Water was added as needed.
A 1:1 ratio of waste to cement  is high, but it was believed
necessary because of the high level of organics and heavy metals
in the FSA waste.  These mixing ratios were not changed for the
other wastes.  The treated waste was poured into 1-cu-yd molds
(as a slurry) and allowed to harden.  The molds were stripped
from the hardened waste after one to two days.

     During the waste processing, the holes created excavating
the feed waste were enlarged to accommodate burial of the treated
waste blocks.  Before the blocks for a particular waste area were
buried, the hole was lined with plastic (to prevent seepage of
contaminated water into the hole) and a one foot layer of clean
soil was deposited.  After the  blocks were placed into the
excavation hole, additional soil was added to cover the blocks.
Stakes were planted to identify the location of each block.
SAMPLING AND ANALYSIS PROCEDURES

     One month and nine months after the blocks had been buried,
samples were collected from two blocks in each area for physical
property tests and from one block for chemical tests (soil
analyses and leaching tests) and weathering tests.

     Samples of solidified waste were collected using a rotary
rig with diamond-tipped core barrel.  The outside diameters for
the cores were 7 cm for unconfined compressive strength,
permeability, and leaching tests; and 4.5 cm for wet/dry and
freeze/thaw tests.  These multiple samples of different diameters
required the boring of at least two holes in each of twelve
blocks, two blocks in each area.  Cores were removed from the
core barrel and sealed in aluminum foil, placed in glass jars
closed with a custody seal, placed in zip-loc bags, and stored
with ice packs during storage and shipment to the laboratory.


     Both standard methods from SW-846 and experimental test
methods were used to evaluate the core samples.  The test
procedures are listed in Table 1 below:
                                398

-------
                             Table 1
     Physical Tests
     Chemical Tests
     Micro Scale Tests
Moisture
Bulk Density
Unconfined Compressive Strength (UCS)
Wet/Dry Weathering
Freeze/Thaw Weathering
Permeability - Falling Head

Leaching   -   TCLP - VOCs, BNAs,  PCBs
Priority Pollutant Metals - Pb, Cr, Ni,
               Zn, Cu, Cd
Total Oil and Grease

X-Ray Diffraction
Scanning Electron Microscopy
Unconfined compressive strengths were performed on the Wet/Dry
test specimens and controls after the twelve-cycle weathering
tests, except for LAS where a permeability test was performed on
the test specimen.  Unconfined compressive strengths were
performed on the Freeze/Thaw test specimens and controls after
the twelve-cycle weathering tests, except for DSA and FSA where
permeabilities were performed on the test specimens.  The listed
metals were the only metals detected in significant quantities.


TEST RESULTS

     The results are presented in three parts; physical, chemical
and microstructural properties.  The physical properties include
moisture content, bulk density, permeability, unconfined
compressive strength, and weathering effects.  The chemical
analyses were for treated soil and TCLP leachate analyte
concentrations for volatile organic compounds (VOC), base
neutral/acid extractables (BNA), six priority pollutant metals,
and polychlorinated biphenyls (PCB).  Microstructural analysis
included X-ray diffraction and.scanning electron microscopy.
It should be noted that the 28-day (i.e., 1 month) sample data is
from the report "Technology Evaluation Report, SITE Program
Demonstration Test, HAZCON Solidification, Douglassville,
Pennsylvania - Volume I (EPA/540/5-89-001a) .
Physical Properties

     The results of the physical tests are summarized in Tables
2, 3 and 4; and are discussed below.

                                399

-------
     The moisture content ranged from 8.8% by wt for PFA-1 to
20.4% by wt for FSA-3.  All the values are approximately 1% to 2%
by wt less than for the 28-day samples, with the largest
reduction for FSA.  This may indicate that the blocks continued
to cure thus reducing the free moisture content.

     The bulk densities were essentially equivalent to those
obtained in the 28-day cores.  They ranged from 1.51 g/ml for FSA
to 2.05g/ml for PFA.  The values at LAS and LAN increased by 5%
by wt and 3% by wt, respectively.  This increase is most likely
due to sample variations.

     The permeabilities ranged from 2.2xlO"10 cm/sec at DSA-3 to
2.3X10"7 cm/sec at  DSA-1.  These values are in the same general
range as the 28-day cores but appear to be slightly larger.  The
apparent increases in permeability are larger than the decreases,
and eight of the .twelve samples increased.  This trend is not
conclusive as data on permeability variation within individual
blocks is needed.  This apparent increase could be explained by a
small breakdown in the porous structure of the blocks, which
would allow interconnection of the pores and a path for water
flow.

     The unconfined compressive strength  (UCS) ranged from 230
psi at FSA-3 to 1170 psi at PFA-3.  Comparing the results to
those of the 28-day cores, nine of twelve samples had a lower
UCS.  The largest  losses occurred for the areas with low oil and
grease content.  Averaging all of the values across the six plant
areas shows an overall decrease of about 20%.  This apparent
•reduction in UCS is consistent with an increase in permeability,
if the solid structure is deteriorating.

     The oil and grease content in the treated soil ranged from
0.48% by wt at DSA-3 to 12.1% by wt for FSA-1.  All values are
consistent with 28-day cores.  The samples are also consistent on
a material balance basis to the untreated soil oil and grease
concentrations.

     The wet/dry weathering tests weight  losses of the test
specimens were slightly greater than  for the 28-day samples, with
losses ranging from 0.7% to 1.7% by wt.  The unconfined
corapressive strength  for the  long-term and 28-day samples were
the same order of  magnitude.  A permeability was performed on the
weathered test specimen from  LAS-LM1  and was 5.5x10"  cm/sec
which agreed with  the unweathered  long-term sample but is greater
by a factor of 30  compared to the unweathered 28-day  cores.

     The freeze/thaw weathering test  weight losses of the test
specimens was nearly  double that  of the 28-day  cores, with values
ranging from 0.99% by wt at PFA-3  to  3.19% by wt at LAS-LM1.  The
loss for the controls was only  slightly greater than  for the  28-

                                400

-------
day core controls.  Therefore, unlike the wet/dry samples, the
corrected cumulative weight  loss  (wt. loss of test specimens
minus wt loss of control each divided by its dry starting weight)
for the long-term samples were larger than for the 28-day cores.
This indicates, along with UCS _and permeability results, that the
freeze/thaw cycles at Douglassville during the winter of 1987-
1988 may have damaged the internal structure of the test blocks.
The unconfined compressive strengths of the 12-cycle weathered
test specimens ranged from 430 psi at LFA-5 to 860 psi at LAN-1.
These values on average are  larger than the UCS values from the
28-day weathered cores.  However, individual UCS values are so
scattered that a comparison  of UCS values between long-term and
28-day cores shows them to be approximately equal.


Chemical Properties

     The results of the chemical  tests are summarized in Tables 5
and 6; and are discussed below.

     The total PCBs ranged from <0.64 mg/kg at DSA-3 to 5.9 mg/kg
at FSA-1.  These values are  considerably lower than anticipated
based upon untreated soil sample  analysis performed during the
Demonstration Test.  Analytical difficulties were encountered.
Analysis on 28-day core samples were not performed.  PCBs were
not detected in the TCLP leachates.

     The VOC content ranged  from  210 ug/kg at DSA-3 to 70,800
ug/kg at FSA-1.  These values are considerably lower than i the
28-day cores where the values ranged from 3,490 ug/kg at LAN-1 to
108,700 ug/kg at FSA-1.  The VOC  content in the TCLP leachate
ranged from 13 ug/1 for LAN-1 to  493 ug/1 for LAS-LM1.  A
calculation of the migration potential, which is defined as
analyte weight in the leachate divided by analyte weight in the
solid, provided erratic results,  but appear to be equivalent to
the 28-day cores.  The Demonstration Test results showed
equivalent migration potentials for the treated and untreated
soils.  Therefore, no immobilization of VOCs could be seen.

     The total metals content in  the long-term treated soil
samples was equivalent to those obtained for the 28-day cores
with lead values ranging from 980 mg/kg at DSA-3 to 8,600 mg/kg
at FSA-1.  As previously, the primary metal contaminant is lead.
Leachate results from the TCLP leach test showed total metals
content ranging from 3 ug/1  at DSA-3 to 120 ug/1 at LFA-5.   The
FSA-1 results were quite low, 39  ug/1, wit only lead detected in
the leachate.  A calculation of migration potential produced
values equivalent to those for the 28-day samples,  which showed
high immobilization of the metals.  The migration potential for
lead ranged from 2.1x10  ug  leachate/ug soil for LAN-1 to
6.3x10  ug leachate/ug soil  for LFA-5.
                              401

-------
     The quantity of BNAs for the long-term soil samples was
considerably less than reported for the 28-day cores.  However,
the detection limits used by the laboratory contractor were a
factor of 50 to 100 higher due to laboratory difficulties.  For
example, at FSA-1, no BNAs were_ detected in the long-term samples
while the 28-day cores showed a value of 210 mg/kg.  However, the
long-term and 28-day samples had TCLP leachate concentrations
that were equivalent.  The long-term leachate values ranged from
35 ug/1 for PFA-3 to 3,109 ug/1 for FSA-1, the latter being about
the same as that of the 28-day cores.  Therefore, it appears that
no further immobilization of these organics occurred.

     The oil and grease levels in the long-term sample leachates
were below 1.0 mg/1, which is less than for the 28-day samples,
where the values were primarily between 2.0 mg/1 and 4.0 mg/1.
Oil and grease was detected in the TCLP leachates only for LAN-1
and FSA-1.  The other four samples showed values below the
detection limits for oil and grease of 0.4 mg/1.  It appears that
the mobility for oil and grease is equivalent for the treated and
untreated soils.  This is an apparent improvement over the
earlier results.

     Analysis of clean soil samples taken from the backfill used
around the buried blocks in each plant area showed no
distinguishable change in its measured physical properties or
organic content.  No VOC, BNA or PCBs were detected in the soil.
Microstructural Analysis

     Analyses  of thirteen  samples were performed on a
microstructural scale.  The  samples were  studied by scanning
electron microscopy  (SEM), optical microscopy  (OM), and X-ray
diffraction  (XRD).   The type of information to be  obtained from
the tests were:

     Microscopy      -    crystal appearances, porosity,
          fractures, and the presence of  unaltered soil
                                    crystalline  structure of the


The  results  can be summarized as follows:
X-ray Diffractometry
     hydration products.
      The samples were morphologically poorly  defined material
      compared to hardened portland cement/soil mixtures which do
      not contain waste material.

      The porosity of the long-term samples were  low, having
      decreased noticeably from examinations of the  28-day
      samples.  Decreased porosity suggests that  either
      microstructures vary significantly or that  dehydration

                               402

-------
     reactions continued after the previous samples were
     analyzed.

     Mixing appears to be poor as observed previously.  Brownish
     colored aggregates and opajgue particles, likely to be
     untreated waste material, could be observed in both long-
     term samples and those of the 28-day cores.

     X-ray diffraction analyses continue to support that
     encapsulation is the major process contributing to
     stabilization of these contaminated soils.

     As in the 28-day samples, previously examined, unhydrated
     clinker was detected in the long-term samples, suggesting
     poor mixing and/or incomplete hydration.


CONCLUSIONS

     Based on the comparison of the long-term and 28-day sample
test results, the following conclusions were drawn:

     *    The priority pollutant metals were immobilized,  with
          the long-term results equivalent to the 28-day results.

     *    The VOCs and BNAs leaching did not change.

     *    The clean soil backfill appears not to have been
          contaminated by the buried blocks.

     *    A small deterioration in physical properties appears to
          have occurred.  This was concluded from a series of
          small changes, which are as follows:

               a decrease in UCS for samples of low oil and
                    grease content

               an increase in permeability of the samples

               increased weight losses of test specimens and
               controls during the wet/dry and freeze/thaw
               weathering tests

               large increase (double)  in weight loss of
               freeze/thaw test specimens.

     *    The microstructural results are similar to earlier
          results except that the porosity was low having
          increased noticeably from the earlier samples.


     The general  conclusion is that the 9-month test results were
                               403

-------
similar to the earlier result, but where they deviated the
results appear to show a small reduction in the physical
properties.

     Further long-term samples .are scheduled to be collected in
June 1989 and analyzed as described in this study.  Comparison of
three sets of data will allow a stronger conclusion to be drawn
about the long term durability of solidification/stabilization
processes.
                                404

-------











trt


S
i
u_
0

IU
ot
IU
&
£
PHYSICAL
(M
IU
•J

















1
I


§
« X

i


•—
"2 .i **•
ill
S h. ?
a s I
5 o *-
s
p^ V
15

1*6
£ -
II

?£
-. i
O ci
L

i

o. <

!>.»«« COM CMK> MN «-O M^
•<) ^T ep O> eo •- u>K. NCM ino
ineo «7in •- ro o>o KM o- S
M N» 40 *• ^" co ** N. •o in
cMo  (O O> 03 •- -O
•*in -or-- M<- •* m coo inin
•-•- ^-»- CMCM »-»- «- •-»-





MK inp RS inia SS «l8
»- -* in h- CMCM coco (O'- eo I--


OO Kl CM OO -J- O -* 1M OO
JQO •*•"* '"t KO OO OM


£ S 3 S E 5 S S S S 35

57* o •» 0 in
l^>* v*i tin to <
-------
                              TABLE 3.  WEATHERING TEST RESULTS  - WET/DRY
'
Plant
Area
DSA-3
LAH-1
FSA-1
LFA-5
PFA-3
LAS-LH-1
Average
Long-term
Ut. Loss, X
Specimen Control
1.50 1.33
0.80 0.66
0.86 0.80
1.71 1.40
0.91 0.73
1.42 1.00
1.20 0.99

UCS, psi
Specimen Control
830 470
890 500
340 430
250 230
2290 1270
(a) 830
920 622
a) Permeability performed on weathered sample resulted


Plant
Area
DSA-3
LAH-1
FSA-1
LFA-5
PFA-3
LAS-LH-1
Average
TABLE
Long-term
Wt. Lose, X
Specimen Control
1.85 1.73
1.55 0.62
1.48 1.04
2.89 1.28
0.99 0.88
3.19 1.00
1.99 1.09
28-day
Ut. Loss, X UCS,
Specimen Control Specimen
0.90 0.88 1150
0.74 0.77 180
0.93 0.75 340
1.53 1.15 1230
0.73 0.66 1170
1.05 0.84 400
0.98 0.84 739
in 5.5x10"8 cm/sec.

psi
Control
750
610
330
500
1190
450
638

4. WEATHERING TEST RESULTS - FREEZE/THAW

UCS, psi
Specimen Control
(a) 1530
860 420
(b) 520
430 870
750 620
470 770
628 788
28-day
Ut. Loss, X UCS,
Specimen Control Specimen
1.29 1.24 660
1.07 0.60 370
0.53 0.53 400
2.17 1.20 520
0.58 0.49 980
0.95 0.73 210
1.10 0.80 523

psi
Control
1020
250
210
320
350
400
425
a)  Panaeability performed on weathered sample resulted in 4.0x10'  on/sec.
b)  Permeability performed on weathered sample resulted in 7.1x10   cm/sec.

                                                       406

-------





















S
1
CD
U
OL
i
UJ
X-
_J
_J

u
X
uu
ac
o


in
ui
i

































i
00
CM




























E
4)
i














+ S ci
f — • O
£55
* 8 s
8.

a.
u

c

X
*



c

5


8
S - S
— ex:
*rf **- O
« *-. ro
&• C 4)
01 C —
— *••
x 8. S




c








c
g
>




0.
CJ
»—
c


§

te

***
Cfl
(J
OL



J
Q.

•O
O
»
5

««
3

41
**
a
u
a
±
a,

3
_^
to
|

i.

T3
O
3



^
5?
«T
a
£
S


O*


o>
^
(0
"8
*J
0
«>
u

ff

V
a
f
a
a
.

"o
(A
U
1


I

in to o o in
•". ^ S S R S
iX o o o d o





§T- &t |W
O Q O
*^ X> •"
— *





0 O 0 0 0 0
in O- o K) in ry
Kl »» h- O> to «»
in" m to K? in" ^
3 «M *-



Kl 00 tO
AJ ro o *o -^ h*
• •••••
>» 0 0 U\ 0  to o> o o to
^r •- o o^ o o








o o o o o »-
«- Q Q O O O.
CM oo oo in in in
g" S






333333






>O Kl t> •- K IO
?.* in «- to •*




Ki »- -- in to 5
• 111*1
5 3 35 2 2 S
o -J u. -J a. -J



0)
0
«5
i.
^
u
0
C
o
i

2
0
1
S 0
•v. w
! oo a
*•» u
s -
§•
DC -O
S "
— M
I I
~a » ^
_o S
1 0
f. ,C
84* •
C
i- .•«- 5
0 ° 8
0 X i
M U
0 <->
« » 2
g S —

a 01.
> JC 0
1- *1 i
ii M U
•*" CB
•0 I ^


a u ^
£ a «
•a 5 - cT
CM O0
O ** C
SO. C 0
S St.
— — u a
o.
1 £ i1 i

f «D 0 0
1 si M
s - * Is
« 33 * ~*

£ 00 •«• CD
5 s 5 ^ «^
? C C ° 0
•^ O o 01*
0 — ~- .CO
oa u u n •*.
" " • £
an 80
3 xx t i£
. * :
407

-------
                                      TABLE 6.  LONG-TERM CHEMICAL ANALYSES - 8HA AND LEAD
Plant
Are*
DSA-3
LAN-1
FSA-1
IFA-5
PFA-3
LAS-LH-1
BHAs in
treated soil, mg/kg
DO* (12.15)
BD (17.90)
BD (534.2)
1.1 (36.70)
0.91 (18.45)
BO (39.60)
BKA in TCLP
leachate, ug/l
43 (14)*
1,260 (1.409)
3,109 (4,083)
60 (169)
35 (55)
216 (689)
Migration
potential - BNA,
ug teachate
ug solid
- (•)
- (.62)
- (.15)
1.09 (-)
.077 (.53)
• (.31)
Lead in
treated
soil, mg/kg
980 (570)*
3,800 (3,010)
8,600 (10,200)
3,800 (2,700)
3,400 (4,900)
4,300 (3,850)
Lead in TCLP
leachate, ug/l
3 (7)
4 (5)
39 (400)
120 (50)
38 (11)
23 (51)
Migration potential
lead, (104)
ug leachate
ug solid
0.61 (1.69)*
0.21 (0.36)
0.91 (8.0)
.6.32 (5.40)
2.24 (0.61)
1.30 (3.13)
*  Values In parentheses are from 28 day samples as reported in the Technology Evaluation Report.




+  Below detection limits.  The detection limits were quite high due to interferences.
                                                         408

-------
   SITE DEMONSTRATION OF THE TERRA  VAC  IN SITU VACUUM  EXTRACTION TECHNOLOGY
                by:  Peter A. Michaels
                     Foster Wheeler Enviresponse, Inc.
                     Edison, NJ  08837
Mary K. Stihson
USEPA, RGB, RREL
Edison, NJ  08837
                                    ABSTRACT


    Terra Vac Inc's vacuum extraction system was the subject of a SITE
program demonstration test in Grovel and, Massachusetts.  The site chosen was
contaminated with volatile organic compounds, mainly trichloroethylene, which
were used as degreasing solvents in an operating machine shop on the site.

    The eight-week test run produced the following results:

o   extraction of 1,300 Ib of VOCs                                    •  ..

o   a steady decline in the VOC recovery rate with time

o   a marked reduction in soil VOC concentration in the test area

o   an indication that the process can remove VOCs from clay strata

    The system operation proved to be very reliable.  Upon achievement of a
steady operation, the only stoppages occurred in order to replace spent
activated carbon canisters with fresh canisters.
                                     409

-------
                                 INTRODUCTION

    This SITE program demonstration test was planned to determine the
effectiveness of Terra Vac Inc's vacuum extraction technology in the
removal of volatile organic compounds from the vadose zone.  The location
of the test was on the property of an operating machine shop.  The property
is part of a Superfund site and is contaminated by degreasing solvents,
mainly trichloroethylene.

                                  OBJECTIVES

    The main objectives of this project were:

    The quantification of the contaminants removed by the process.

    The correlation of the recovery rate of contaminants with time.
o

o

o
    The prediction of operating time required before obtaining site
    remediation.

o   The effectiveness of the process in removing contamination from
    different soil strata.

                                   APPROACH

    The objectives of the project were achieved by following a
demonstration test plan which included a sampling and analytical plan.  The
sampling and analytical plan contained a quality assurance project plan.
This QAPP assured that the data collected during the course of this project
would be of adequate quality to support the objectives.

    The sampling and analytical program for the test was split up into a
pretest period, which has been called a pretreatment period; an active
period; midtreatment; and a posttreatment period.

    The pretreatment period sampling program consisted of:

o   soil boring samples taken with split spoons

o   soil boring samples taken with Shelby tubes

o   soil gas samples taken with punch bar probes

    Soil borings taken by split spoon sampling were analyzed for volatile
organic compounds (VOCs) using headspace screening techniques, purge and
trap, GC/MS procedures, and the EPA-TCLP procedure.  Additional properties
of the soil were determined by sampling using a Shelby tube, which was
pressed hydraulically into the soil by a drill rig to a total depth of 24
feet.  These Shelby tube samples were analyzed to determine physical
characteristics of the subsurface stratigraphy such as bulk density,
particle density, porosity, pH, grain size,  and moisture.  These parameters
were used to define the basic soil characteristics.
                                     410

-------
    Shallow soil gas concentrations were collected during pre-, mid-, and
posttreatment activities.  Four shallow vacuum monitoring wells and twelve
shallow punch bar tubes were used at sample locations.  The punch bar
samples were collected from hollow stainless steel probes that had been
driven to a depth of 3 to 5 feet.  Soil gas was drawn up the punch bar,.
probes with a low-volume personal pump and tygon tubing.  Gas-tight .50-ml
syringes were used to collect the sample out of the tygon tubing.

    The active treatment period consisted of collecting samples of:

o   wellhead gas

o   separator outlet gas

o   primary carbon outlet gas

o   secondary carbon outlet gas

o   separator drain water

    All samples with the exception of the separator drain water were  r
analyzed on site.  On-site gas analysis consisted of gas chromatography
with a flame ionization detector  (FID) or an electron capture detector
(ECD).  The FID was used generally to quantify the trichloroethylene (TCE)
and trans 1,2-dichloroethylene (DCE) values, while the ECD was used to
quantify the 1,1,1-trichloroethane (TRI) and the tetrachloroethylene (PCE)
values.  The use of two detectors, FID and ECD, was necessitated by high
concentrations of TCE in the extracted well head gas.  Owing to the high
TCE concentrations, most of the  samples injected on the ECD had to ,be
diluted.  Even with dilution factors of 333 to 1, the TCE concentration on
the ECD would exceed the linear  range of the detector, thus necessitating
the use of two detectors.

    The separator drain water-was analyzed for VOC content using SW846
8010.  Moisture content of the separator inlet gas from the wells was
analyzed using  EPA Modified Method 4.  This method is good for the
two-phase flow  regime that existed in the gas emanating from the wellhead.
Table  1 lists analytical methods  used for this project.             !

    The posttreatment sampling essentially consisted  of repeating
pretreatment sampling procedures  at locations as  close as possible to the
pretreatment sampling locations.                          . ... .

    The activated carbon canisters were sampled,  as close to the center  of
the canister- as possible, and these samples were  analyzed for  VOC content
as a check on the material balance for the process.   The method  used was
P&CAM  12.7, which consisted of desorption of the carbon with CS2  and
subsequent gas  chromatographic analysis.

                             PROCESS DESCRIPTION

    The vacuum  extraction process is a technique  for  the removal and.
                                    411

-------
                         TABLE  1.   ANALYTICAL METHODS
   Parameter
Analytical method
Sample Source
Grain size
PH
Moisture (110°C)
Particle density
Oil and grease
EPA-TCLP

TOC
Headspace VOC
VOC
VOC
VOC
VOC
VOC
VOC
ASTM D422-63
SW846* 9040
ASTM D2216-80
ASTM D698-78
SW846* 9071
F.R. 11/7/86,
Vol. 51, No. 216,
SW846* 8240
SW846* 9060
SW846* 3810
GC/FID or ECD
GC/FID or ECD
SW846* 8010
SW846* 8010
Modified P&CAM 127
SW846* 8240
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings
Soil borings

Soil borings
Soil borings
Soil gas
Process gas
Separator liquid
Groundwater
Activated carbon
Soil borings
*Third Edition, November 1986.
                                     412

-------
venting of volatile organic constituents (VOCs) from the vadose or
unsaturated zone of soils.  Once a contaminated area is completely defined,
an extraction well or wells, depending upon the extent of contamination,
will be installed.  A vacuum pump or blower induces air flow through the
soil, stripping and volatilizing the VOCs from the soil matrix into the air
stream.  Liquid water is generally extracted along with the contamination.
The two-phase flow of contaminated air and water flows to a vapor liquid
separator where contaminated water is removed.  The contaminated air stream
then flows through activated carbon canisters arranged in a parallel-series
fashion.  Primary or main adsorbing canisters are followed by a secondary
or backup adsorber in order to insure that no contamination reaches the
atmosphere.  Figure 1 illustrates the layout of wells and equipment.

                     EQUIPMENT LAYOUT AND SPECIFICATIONS

    Specifications are given in Table 2 for the equipment used in the
initial phase of the demonstration.  This equipment was later modified when
unforeseen circumstances required a shutdown of the system.  The
vapor-liquid separator, activated carbon canisters, and vacuum pump skid
were inside the building, with the stack discharge outside the building.
The equipment was in an area of the machine shop where used cutting oils
and metal shavings had been stored.

    Four extraction wells (EW1-EW4) and four monitoring wells (MW1  - MW4)
were drilled south of the shop.  Each well was installed in two sections,
one section to just above the clay lens and one section to just below the
clay lens.  The extraction wells were screened above the clay and below the
clay.  As shown in Figure 2, the well section below the clay lens was
isolated from the section above by a bentonite portland cement grout seal.
Each section operated independently of the other.  The wells were arranged
in  a triangular configuration, with three wells on the base of the  triangle
(EW2,  EW3, EW4) and one well at the apex  (EW1).  The three wells on the
base were called  barrier wells.  Their purpose was to  intercept
contamination, from underneath the building and to the side of the
demonstration area, before this contamination reached  the main extraction
well (EW1).  The  area enclosed by the four extraction  wells defined the
area to be cleaned.

INSTALLATION OF EQUIPMENT

    Well drilling and equipment setup were begun on December 1, 1987.  A
mobile drill rig  was brought in, equipped with hollow-stem augers,  split
spoons, and Shelby tubes.  The locations  of the extraction wells and
monitoring wells  had been staked out previously based  on contaminant
concentration profiles  from a previously  conducted remedial investigation
and from bar punch probe  soil gas monitoring.

     Each well drilled was sampled at 2-foot intervals  with a split  spoon
pounded  into the  subsurface by the drill  rig  in advance of the hollow  stem
auger. The hollow stem auger would then  clear out the soil down to the
depth  of the split spoon, and the cycle would  continue in that manner  to  a
depth  of 24 feet.  The  drilling tailings  were  shoveled into 55-gallon  drums
                                     413

-------
                                   MW1
1   Schematic diagram of equipment layout.
                 414

-------
                 2//PVCPIPE
                             — BENTONITE
                           3'
                                SAND
                                SCREENING
                           12.67'
                          «	GROUT
                               " BENTONITE
                                SCREENING
                            24'
Fig.   2   Schematic diagram of an extraction well
                    415

-------
                            TABLE 2.  EQUIPMENT LIST
      Equipment
   Number required
           Description
Extraction wells

Monitoring wells

Vapor-liquid
  separator

Activated carbon
  canisters


Vacuum pump skid


Holding tank

Pump
4 (2 sections each)

4 (2 sections each)

         1
Primary:  2 units in
  parallel
Secondary:  1 unit
2" SCH 40 PVC 24' total depth

2" SCH 40 PVC 24' total depth

1000-gal capacity, steel


Canisters with 1200 Ib of carbon
  in each canister - 304 SS
4" inlet and outlet nozzles

25 HP motor - positive displace-
  ment lobe type blower - 3250 rpm

2000-gal capacity - steel

1 HP motor - centrifugal
for eventual disposal.  After the holes were sampled, the wells were
installed using 2-inch PVC pipes screened at various depths depending upon
the characteristics of the soil in the particular hole.  The deep well  was
installed first, screened from the bottom to various depths.  A layer of
sand followed by a layer of bentonite and finally a thick layer of grout
were required to seal off the section below the clay lens from the section
above the clay lens.  The grout was allowed to set overnight before the
shallow well pipe was installed at the top of the grout.  A layer of sand
bentonite and grout finished the installation.

                      VOC  REMOVAL  FROM THE VADOSE ZONE

    The permeable vadose zone at the Groveland site is divided into two
layers by a horizontal clay lens,  which is relatively impermeable. As
explained previously, each extraction well had a separate shallow and deep
section to enable VOCs to be extracted from that section of the vadose  zone
above and below the clay lens.  The quantification of VOCs removed was
achieved by measuring

    o    gas volumetric flow rate  by rotameter and wellhead gas VOC
         concentration by gas chromatography
                                     416

-------
    o    the amount of VOCs adsorbed by the activated carbon canisters by
         desorption into C$2 followed by gas chromatography.

    VOC flow rates were measured and tabulated for each well section separ-
ately.  The results of gas sampling by syringe and gas chromatographic
analysis indicate a total of 1,297 Ib of VOCs were extracted over a 56-day
period, 95% of which was trichloroethylene.  A very good check on this
total was made by the activated carbon VOC analysis, the results of which
indicated a VOC recovery of 1353 Ib; virtually the same result was obtained
by two very different methods.

    One view of the reduction in VOC concentrations in the vadose zone can
be seen from examining the three-dimensional shallow soil gas plots.  Soil
gas was collected during pretreatment, midtreatment, and posttreatment from
punch bar probes and shallow vacuum monitoring wells.  The collection
points were located on a coordinate system with extraction well 1 as the
origin (0,0).  Each collection point has an x and y coordinate, and TCE
concentrations are plotted on a "Z" scale, which gives a three-dimensional
plot across the grid.  Values of "Z" between data points and around the
grid are generated by the Kriging technique, which uses given data points
and a regional variable theory to generate values between and around sample
locations.  Kriging is the name given to the least squares prediction of
spatial processes and is used in surface fitting, trend surface analysis,
and contouring of sparse spatial data.

    A total of twelve shallow punch bar tubes were utilized along with the
four shallow vacuum monitoring wells.  The punch bars were driven to a
depth of 3 to 5 feet, and as with the vacuum wells, soil gas was drawn up
the punch bar probes with a low-volume personal pump and tygon tubing.
50-ml gas-tight syringes were used to collect the sample out of the tygon
tubing.  The gas samples were analyzed in the field trailer using gas
chromatographs with flame ionization detectors and electron capture
detectors.

    The soil gas results show a considerable reduction in concentration
over the course of the 56-day demonstration period as can be seen from
Figures 3 and 4.  This is to be expected since soil gas is the vapor halo
existing around the contamination and should be relatively easy to remove
by vacuum methods.

    A more modest reduction can be seen in the results obtained for soil
VOC concentrations by 6C/MS purge-and-trap analytical techniques.  Soil
concentrations include not only the vapor halo but also interstitial liquid
contamination that is either dissolved in the moisture in the soil or
existing as a two-phase liquid with the moisture.

    Table 3 shows the reduction of the weighted average TCE levels in the
soil during the course of the 56-day demonstration test.  The weighted
average TCE level was obtained by averaging soil concentrations obtained
every two feet by split spoon sampling methods over the entire 24-foot
depth of the wells.  The.largest reduction in soil TCE concentration.
occurred in EW4, which had the highest initial level of contamination.
                                    417

-------
                 MU2
a
a
      0
      Fig.   3    Pretreatment  shallow  soil-gas concentration
                                  418

-------
                                              MAP VIEW
    MW2
                             MU3
                                               MUJ4
Fig.  4    Posttreatment shallow soil-gas concentration
                            419

-------
           TABLE 3.  REDUCTION OF WEIGHTED AVERAGE TCE LEVELS IN SOIL
                     (TCE  Cone,  in mg/kg)
Extraction  Well
Monitoring Well

       1
       2
       3
       4
Pretreatment
Posttreatment
    1.10
   14.75
  227.31
    0.87
     0.34
     8.98
    84.50
     1.05
% Reduction
1
2
3
4
33.98
3.38
6.89
96.10
29.31
2.36
6.30
4.19
13.74
30.18 . .
8.56
95.64
   69.09
   39.12
   62.83
EW1, which was expected to achieve the greatest concentration reduction,
exhibited only a minor decrease over the course of the test.  Undoubtedly
this was because of the greater-than-expected level of contamination that
existed in the area around MW3 that was drawn into the soil .around EW1.
The decrease in the TCE level around MW3 tends to bear this out.

            EFFECTIVENESS OF THE TECHNOLOGY IN VARIOUS SOIL TYPES

    The soil strata at the Groveland site can be characterized generally as
consisting of the following types in order of increasing depth:

    o    medium to very fine silty sands

    o    stiff and wet clays

    o    sand and gravel

    Soil porosity, which is the percentage of total soil volume occupied by
pores, was relatively the same for both the clays and the sands.
Typically, porosity over the 24-foot depth of the wells would range between
40% and 50%.  Permeabilities, or more accurately hydraulic conductivities,
ranged from 10"* cm/sec for the sands to 10"° cm/sec for the clays,
with corresponding grain sizes equal to 10"1 mm to 10"3 mm.

    Pretest soil boring analyses indicated in general  that most of the
contamination was in the strata above the clay lens with a considerable
quantity perched on top of the clay lens.  This was the case for EW4, which
                                     420

-------
showed an excellent reduction of TCE concentration in the medium to fine
sandy soils existing above the clay layer,  with no TCE detected in the clay
in either the pretest or posttest borings (see Table 4).   One of the wells,
however, was an exception.  This was MW3, which contained the highest
contamination levels of any of the wells, and was exceptional in that most
of the contamination was in a wet clay stratum.  The levels of
contamination were in the 200-1600 ppm range before the test.

    After the test, analyses of the soil boring adjacent to MW3 showed
levels in the range of ND-60 ppm in the same clay stratum.  The data, as
shown in Table 5, suggest that the technology can desorb or otherwise
mobilize VOCs out of certain clays.

    From the results of this demonstration it appears that the permeability
of a soil need not be a consideration in applying the vacuum extraction
technology.  This may be explained by the fact that the porosities were
approximately the same for all soil strata, so that the total flow area for
stripping air was the same in all soil strata.  It will take a long time
for a liquid contaminant to percolate through clay with its small pore size
and consequent low permeability.  However, the much smaller air molecules
have a lower resistance in passing through the same pores.  This may
explain why contamination was generally not present in the clay strata, but
when it was, it was not difficult to remove.  Further testing should be
done in order to confirm this finding.

                 CORRELATION OF DECLINING VOC RECOVERY RATES

    The vacuum extraction of volatile organic constituents from the soil
may be viewed as an unsteady state process taking place in a nonhomogeneous
environment acted upon by the combined convective forces of induced
stripping air and by the diffusion of volatiles from a dissolved or sorbed
state.  As such  it is a very complicated process, even though the equipment
required to operate the process is very simple.

    Unsteady state diffusion processes  in general correlate well by
plotting the logarithm of the rate of diffusion versus time.  Although the
representation of the vacuum extraction process presented here might be
somewhat simplistic, the correlation obtained by plotting the logarithm of
the concentration of contaminant  in the wellhead gas versus time and
Obtaining a least squares best fit line was reasonably good.  This type of
plot, shown in Figure 5, represents the data very well and is more valid
than both a linear graph or one plotting concentration versus log time, in
which a  best fit curve would actually predict gas concentrations of zero or
less.

    Looking at the plots  for EW1,  shallow  and deep, equations are given for
the least  squares  best fit line for the data points.   If  the vacuum
extraction process is run long enough so that the detection  limit for TCE
on the  ECD, which  is 1 ppbv,  is reached, the length of time  required to
reach that concentration would be approximately 250 days  on  the  shallow
well  and approximately 300 days on the  deep well.
                                     421

-------
 TABLE  4.   EXTRACTION  WELL  4:
          TCE REDUCTION IN SOIL STRATA
Depth
ft
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
Description Permeability TCE Cone, com
of strata
Med. sand w/gravel
Lt. brown fine sand
Med. stiff It. brown fine sand
Soft dk. brown fine sand
Med. stiff brown sand
V. stiff It. brown med. sand
V. stiff brown fine sand w/silt
M. stiff grn-brn clay w/silt
Soft wet clay
Soft wet clay
V. stiff brn med-coarse sand
V. stiff brn med-coarse w/gravel
cm/sec
1°1
10-J
10-5
10";
lo-J
lo-J
lO-J
IQ-f
10"°
10"!
lo-J
lO'3
pre
2.94
29.90
260.0
303.0
351.0
195.0
3.14
ND
ND
ND
ND
6.71
post
ND
ND
39
9
ND
ND
2.3
ND
ND
ND
ND
ND
TABLE 5.  MONITORING WELL 3:
          TCE  REDUCTION  IN SOIL STRATA
Depth
ft
0-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
16-18
18-20
20-22
22-24
Description
of strata
M. stiff brn. fine sand
M. stiff grey fine sand
Soft It. brn. fine sand
Lt. brn. fine sand
Stiff V. fine brn. silty sand
Silty sand
Soft brown silt
Wet green -brown silty clay
Wet green -brown silty clay
Wet green-brown silty clay
Silt, gravel, and rock frag.
M. stiff It. brn. med. sand
Permeability
cm/sec
1Q-5
10"?
10"!
10-J
lo-J
10-J
10-J
10 Q
1° "I
10"?
10-J
ID'4
TCE Cone.
pre
10.30
8.33
80.0
160.0
ND
NR
316.0
195.0
218.0
1570.0
106.0
64.1
ppm
post
ND
800
84
ND
63
2.3
ND
ND
62
2.4
ND
ND
          422

-------
CROVELAND/TERRA-VAC DEMONSTRATION


 _ 1000
 £
 CL
O
cr
 LU
 CJ

 O
 C>

 Ld
 O

    °'1
   0.0.1
                              EXTRACTION WELL #1
                              SHALLOW
                                                     CURVE COEFFICIENT- R2=0 62.
        0       20       40       60       80       100
             DAY  OF  ACTIVE  TREATMENT
          Fig.  5    Wellhead TCE concentration vs. time
                             423

-------
               PREDICTION OF TIME REQUIRED FOR SITE REMEDIATION

    The soil concentration  that would be  calculated from the wellhead gas
concentration using Henry's Law is  included  in the last column of Table 6.
Calculations for the predicted soil concentrations were made assuming a
bulk density of the soil of 1761  kg/nr3, a total porosity of 50%, and a
moisture content of 20%.  The calculated  air filled porosity of the soil is
approximately 15%.  Henry's constant was  taken to be 0.492 KPa/m3-gmol at
40°F.

    Given the nonhomogeneous nature of the subsurface contamination and
interactions of TCE with organic  matter in the soil, it was not possible to
obtain a good correlation between VOC concentrations in wellhead gas and
soil in order to predict site remediation times.  Henry's Law constants
were used to calculate soil  concentrations from wellhead gas concentrations
and the calculated values obtained, correcting for air filled porosity,
were lower than actual soil  concentrations by at least an order of
magnitude (see Table 6).

                   TABLE 6.   COMPARISON  OF WELLHEAD GAS VOC
                              CONCENTRATION AND SOIL VOC CONCENTRATION
Extraction Well
TCE concentration
 in wellhead gas
      ppmv
TCE concentration
  in soil ppmw
Predicted by
Henry's Law
    ppmw
IS
ID
2S
2D
3S
3D
4S
9.7
5.6
16.4
14.4
125.0
58.7
1095.6
54.5
7.2
ND
20.4
20.9
18.0
9.1
0.11
0.07
0.20
0.17
1.5.3
0.74
12.49
    Before one can attempt to make a rough estimation of the remediation
time, a target value for the particular contaminant in the remediated soil
must be calculated.  This target concentration is calculated by using two
mathematical models, the Vertical and Horizontal Spread ModelW and the
Organic Leachate Model^a'.  The mathematical models allow the use of a
regulatory standard for drinking water in order to arrive at a target soil
concentration.

    The VHS model is expressed as the following equation:

    Cy = C0 erf (Z/(2(azY)°-5)) erf (X/(atY)°-5)
                                     424

-------
    where:

    Cy   =    concentration of VOC at compliance point (mg/1)
    CQ   =    concentration of VOC in leachate (mg/1)
    erf  =    error function (dimensionless)
    Z    =    penetration depth of leachate into the aquifer
    Y    =    distance from site to compliance point (m)
    X    =    length of site measured perpendicular to the direction of
              ground water flow (m)
    at   =    lateral transverse dispersivity (m)
    az   =    vertical dispersivity (m)

    A simplified version of the VHS model is most often used,  which reduces
the above equation to:

    cy = cocf

    where:

    Cf   =    erf (Z/(2(azY)°-5)) erf (X/(atY)°-5), which is
              reduced to a conversion factor corresponding to the amount of
              contaminated soil

    The Organic Leachate Model (OLM) is written as:

    C0 = 0.00211 cs°'678S°-373
    where:
              concentration of VOC in leachate (mg/1)
              concentration of VOC in soil (mg/1)
              solubility of VOC in water  (mg/1)
                                                                   This
    The regulatory standard for TCE in drinking water is 3.2 ppb.
regulatory limit is used in the VHS model as the compliance point
concentration in order to solve for a value of the leachate concentration.
This value of leachate concentration is then used in the OLM model to solve
for the target soil concentration.

    Once the target soil concentration is determined, a rough estimation of
the remediation time can be made by taking the ratio of soil concentration
to wellhead gas concentration and extrapolating in order to arrive at a
wellhead gas concentration at the target soil concentration.  The
calculated target soil concentration for this site is 500 ppbw.  This
corresponds to an approximate wellhead gas concentration of 89 ppb for
EW1S.  The equation correlating wellhead gas concentration with time (see
Figure 5) is then solved to give 150 days running time.

    After 150 days the vacuum extraction system can be run intermittently
to see if significant increases in gas concentrations occur upon
restarting, after at least a two day stoppage.  If there are no appreciable
increases in gas concentration, the soil has reached its residual
                                    425

-------
equilibrium contaminant concentration and the system may be stopped and
soil borings taken and analyzed.

(a) EPA Draft Guidelines for Petitioning Waste Generated by the Petroleum
    Refinery Industry, June 12, 1987.

                               ACKNOWLEDGMENTS

    The authors wish to thank Mr. James S. Ciriello, formerly of the U.S.
Environmental Protection Agency, Region I, Boston, Massachusetts at the
time of the project, for his efforts during the course of this project.  A
special note of gratitude is to be given to Mr. Thomas Quinlan of the
Valley Manufactured Products Company, Inc. for his special  support and
cooperation that helped make this project a successful one.
                                    426

-------
              THE OFFICE OF RESEARCH & DEVELOPMENT
                          WRITE PROGRAM
                          Ivars J.  Licis
                     Research Program Manager
              Risk Reduction Engineering Laboratory

                               and

                           M.  Lynn  Apel
                Chief,  Process Engineering  Section
              Risk Reduction Engineering  Laboratory,
              U.S.  Environmental Protection Agency
                      Cincinnati,  OH  45268
     Passage of the 1984 Hazardous and Solid Waste Amendments
(HSWA) to the Resource Conservation and Recovery Act (RCRA) of
1976 marked a strong change in the policies of the United States
concerning the generation of hazardous and nonhazardous wastes.
In addition to authorizing very stringent treatment and disposal
regulations, the Amendments also indicated, as the Nation's top
waste management priority a redirection toward "waste
minimization" as a preferential strategy for encouraging
improvement in environmental quality.

     To carry out the intent of the Amendments to reduce the
generation of waste, the U.S. Environmental Protection Agency
(EPA) has developed a comprehensive pollution prevention program
addressing the release and transport of hazardous, toxic, and
nonhazardous materials through air, water, and solid media.  The
EPA pollution prevention program includes information gathering,
research and development, demonstration, support of state and
local government waste minimization and pollution prevention
programs, training and education, technology transfer activities,
waste minimization assessments, and extensive communication with
universities, state and local governments, and the general
public.

     This paper describes one of the major pollution prevention
research programs being undertaken by EPA, the Waste Reduction
Innovative Technology Evaluation [WRITE] Program.  The WRITE
Program has been designed to identify, evaluate, and demonstrate
new ideas and technologies that lead to waste reduction.  The
WRITE Program involves the cooperative efforts of the EPA, state
and local governments, private industry, universities,  and other
organizations to encourage the research and development of
                               427

-------
effective techniques  and technologies  for multimedia pollution
prevention.
BACKGROUND

     During the last decade, waste management and disposal
regulations have changed drastically.  During the 1970's, the
seriousness of the hazardous waste problem prompted the Congress' *
to enact the Resource Conservation and Recovery Act in 1976.  By
the early 1980's, it had become apparent that even well-regulated
land disposal could cause significant environmental damage.  The
Amendments to RCRA resulted in the banning of the land disposal
of over 400 chemicals and hazardous wastes.  During the 1980's,
significant effort was devoted to research, development, and
commercialization of hazardous waste treatment technologies.
Although some efforts were successful in substantially mitigating
the problem of existing waste, it became apparent that treatment
alone would not solve all of the problems associated with the
increasing amounts of waste being generated.

     Congress recognized this fact while drafting the national
policy on hazardous wastes and directed EPA to report on the
feasibility and desirability of developing mandatory requirements
to compel the adoption of pollution prevention techniques.  The
1984 HSWAs to RCRA define the national policy on hazardous waste
management as follows:

          "The Congress hereby declares it to be the
          national policy of the United States that,
          wherever feasible, the generation of
          hazardous waste is to be reduced or
          eliminated as expeditiously as possible.
          Waste that is nevertheless generated should
          be treated, stored, or disposed of so as to
          minimize the present and future threat to
          human health and the environment." [1]

In 1986, EPA responded to the Congressional request  with a
Report to Congress on the Minimization of Hazardous Waste [2].
In this report, the Agency defined waste minimization as:

          "The reduction, to the extent feasible, of
          hazardous waste that is generated or
          subsequently treated, stored or disposed of.
          It includes any source reduction or
          recycling activity undertaken by a generator
          that results in either (1) the reduction of
          total volume or quantity of hazardous waste
          or (2) the reduction of toxicity of
          hazardous waste, or both, so long as the
          reduction is consistent with the goal of
                                428

-------
          minimization of present and  future threats
          to human health and environment."
     In January 1989, the Agency issued a Pollution Prevention
Policy Statement [3].  This policy encourages organizations,
facilities and individuals to fully utilize source reduction and
recycling practices and procedures to reduce risk to public
health, safety and the environment.  Source reduction is
emphasized as the preferred approach since it eliminates or
significantly reduces the quantity of pollutants generated,
thereby significantly reducing the potential risk to health and
the environment.  As a second choice, recycling reduces the
amount of waste to be treated or sent for disposal.
Additionally, many of the waste reduction efforts have proven to
be both more economical, in the long run, as well as beneficial
to the environment producing a win-win situation.  On this basis,
source reduction and recycling are the primary two elements in
the Waste Management Hierarchy  preceding treatment and disposal.
In this context, source reduction has been defined as the
reduction or elimination of waste at the source, usually within a
process.  Source reduction measures include process
modifications, feedstock substitutions, improvements in feedstock
purity, housekeeping and management procedural changes, increases
in the efficiency of equipment, and recycling within a process.
Likewise, recycling has been defined as the use or reuse of a
waste material as an effective substitute for a commercial
product or as an ingredient or feedstock in an industrial
process.  It includes the reclamation of useful constituent
fractions within a waste material or the removal of contaminants
from a waste to allow it to be reused [4].

EPA'S POLLUTION PREVENTION PROGRAM

     In its Report to Congress, the Agency explored various
technical, economic, and policy issues relevant to the reduction
and recycling of hazardous and nonhazardous wastes, and concluded
that it would be counterproductive for EPA to establish a
mandatory program for waste minimization at this time.  The
Report to Congress stressed that the most constructive role
government could assume is to promote voluntary waste
minimization by providing information, technology transfer, and
assistance to waste generators.  The Agency proposed a waste
minimization program to encourage industry to accelerate efforts
to reduce the generation of wastes through implementation of
process changes and/or the incorporation of recycling methods.

     In its efforts to pursue the objectives set forth by
Congress, EPA has established a national program to effect waste
minimization, or what has since become included in the term
"pollution prevention".  Pollution prevention is a term that has
been used more frequently within the last year by several
organizations including EPA to describe techniques, practices, or
procedures implemented by the private and public sectors to
                               429

-------
prevent the generation of pollutants.  As it is used today,
"pollution prevention" has replaced the term "waste minimization"
which was generally applied to reducing the generation of
hazardous wastes.  Through elimination of this term, that may be
perceived as closely tied to RCRA, EPA is emphasizing that it's
pollution prevention policy has applicability beyond the RCRA
hazardous waste context.

     EPA's pollution prevention program is based on several
elements including: (1) promoting waste reduction by transferring
technical information to firms and state and local government
technical assistance officers; (2) encouraging the adoption of
waste reduction by identifying, evaluating, and demonstrating
appropriate technologies and by promoting the use of waste audits
or assessments; (3) keeping Congress advised on national progress
of waste reduction techniques and practices; (4) fostering
development of state and local government pollution prevention
programs; (5) developing outreach and communication programs with
the goal of raising awareness of the benefits of waste reduction
practices within government, industry, and the public.

     To accomplish some of these objectives and encourage the
identification, development, and demonstration of processes and
technologies that result in less waste being generated, EPA's
Office of Research and Development (ORD) has initiated several
multimedia pollution prevention research programs.  These
programs are designed to be the cornerstone of the Agency's
pollution prevention research and demonstration effort and will
provide much needed data on new technologies to the generating
sectors. Addressed under these programs are hazardous,
nonhazardous, industrial and municipal wastes.  Design and
implementation of these programs is being undertaken by the
Pollution Prevention Research Branch  (formerly the Waste
Minimization Branch) of the Risk Reduction Engineering Laboratory
of ORD in Cincinnati, Ohio.

     The Agency has also established a Pollution Prevention
Office (PPO) within the Office of Policy, Planning and Evaluation
in Washington, D.C.  The Pollution Prevention Office is
responsible for award and oversight of the Source Reduction and
Recycling Technical Assistance grants, establishing a National
Pollution Prevention Information Clearinghouse, and publishing a
quarterly newsletter on pollution prevention topics.  PPO
together with ORD and the Agency's media-specific offices  (i.e.,
Office of Solid Waste, Office of Toxic Substances, etc.) will
develop and implement EPA's multi-media pollution prevention
program.


THE WRITE PROGRAM

     Reducing the generation of industrial and other wastes can
be achieved in many ways.  Process chemistry can be changed;
potential waste streams can be recycled within a manufacturing
                               430

-------
process or back into the process; process technology and/or
equipment can be modified to produce products more efficiently,
resulting in less waste; plant operations, i.e., "housekeeping"
methods, can be changed or controlled to produce fewer and
smaller waste streams or less waste in general; changes in raw
materials (feedstocks) can lead to fewer waste streams or
less-hazardous waste streams; finally, changes in the end
products from manufacturing operations can, in some instances, be
made so as to affect the types and quantities of wastes emitted.
The EPA WRITE Program outlined below is designed to identify,
evaluate and enhance the application of technologies and
practices containing these attributes.

     The WRITE Program is a research program designed to
identify, evaluate, and/or demonstrate the use of innovative
engineering and scientific technologies to reduce the volume
and/or toxicity of wastes produced from the manufacture,
processing,  and use of materials.  The WRITE Program is broad in
technical scope and addresses the reduction of pollutants across
all environmental media: air, land, surface water, and
groundwater.  Attention is directed toward methodologies with the
potential for reducing the quantity and/or toxicity of waste
produced at the source of generation, or to achieve practicable
onsite reuse or recycling of waste materials.  Strong
consideration is given to the applicability of a technique on an
industry-wide basis and across industries.  Industries of primary
interest under the WRITE Program include chemical, fabricated
metal, electronic, printing and publishing, lumber, petroleum,
transportation, food, and textile [5].

     The objectives of the WRITE Program are:

        0  To establish reliable performance and cost
           information on pollution prevention techniques by
           conducting  evaluations or demonstrations of the more
           promising innovative technologies.

        0  To accomplish an early introduction of waste
           reduction techniques into broad commercial practice.

        0  To encourage active participation of small- and
           medium-sized companies in evaluating and adopting
           pollution prevention  concepts by providing support to
           these companies through State and local government
           agencies.

           To encourage the transfer of knowledge and technology
           concerning pollution prevention practices between
           large, medium-sized, and small industries.

        0  To provide solutions to important chemical-,
           wastestream-, and industry-specific pollution
           prevention  research needs.
                              431

-------
     For the purposes of this program, innovative waste
minimization technologies are defined as those technologies that
represent an alternative approach to conventional waste
management methods  (i.e., alternatives to incineration, land
disposal, treatment, etc.).  This approach involves substantially
reducing the volume or toxicity of waste generated at the source
and/or recycling or reusing the waste materials.  In some cases,
the technologies may incorporate novel modifications to a
component of an existing process or operation, or the adaptation
of an existing technology to a new process.

        To date the WRITE Program encompasses approximately 12
research and demonstration projects totalling over $4.0 million.
Under these projects, approximately 30 waste reduction
technologies will be evaluated throughout the next three years,
and several long-term waste- and industry-specific research
studies will be undertaken.  These activities are conducted under
three subprograms of the WRITE Program; the WRITE Pilot Program
with State and Local Governments, the WRITE Program With
Industry, and the WRITE Research Program [see WRITE Program
Chart).


WRITE PILOT PROGRAM WITH STATE AND LOCAL GOVERNMENTS

        At the moment, the "WRITE Pilot Program with state and
Local Governments" is the largest subprogram under WRITE and
addresses immediate information transfer needs between government
and industry.  Through the joint efforts of EPA and various state
and local governments, technical and economic evaluations of
source reduction and recycling technologies are being conducted
of manufacturing and  processing operations across approximately
twenty industries.  This joint approach was chosen because State
and local government officials are often more familiar with local
industrial practices and regional manufacturing and economic
interests that can affect the potential success and widespread
applicability of proposed pollution prevention technologies.
States currently participating in this program include
California, Connecticut, Illinois, Minnesota, New Jersey, and
Washington.

     Under this program, $100,000 per year is provided by EPA to
each participating state/local government.  The state/local
government also contributes additional matching funds ranging
from 33% to 50% of the cost of the research.  An average of five
waste reduction technologies will be evaluated by each
state/local government during the 3-year period of this pilot
program.  Waste reduction technologies evaluated under this
program are based on several selection criteria.  These include:
(1) type of waste minimization technology, (2) status of
development, (3) unique nature of the technology, (4)
application, (5) source reduction performance capability, (6)
extent of process modification,  (7) cost effectiveness of the
technology, (8) process safety and health considerations, (9)
                               432

-------
cost to the EPA and state/local government, and  (10)
legal/contractual issues.  A worth assessment model is then used
as one of the decision tools to evaluate and rank potential waste
minimization technologies.

     The technical and economic evaluations conducted for each
technology include an in-depth study of the process, a literature
review of comparable processes, material and energy balance
computations, a field demonstration, and determination of cost
estimation parameters including itemization of capital and
operational costs, calculation of the payback period and return
on investment.  A summary of the types of information collected
during a technical and economic evaluation of a waste
minimization technology under this program is shown in Table 1.
The example concerns the modification of a cold solvent cleaning
process.  In the cold cleaning of ball bearings with solvents,
using a two-step countercurrent cleaning sequence can increase
the cleaning efficiency.  It can also substantially reduce the
solvent requirement and, hence, the waste generation.  This
process does not involve substantial equipment modification.
Material balance calculations indicate a waste reduction of 50
percent and a 33 percent reduction in fresh solvent requirements
[6].


WRITE PROGRAM WITH INDUSTRY

     The "WRITE Program With Industry" focuses on  evaluations of
waste reduction technologies currently in use or under
development by large industries.  One of the  objectives of this
program is to encourage the transfer of knowledge and technology
concerning pollution prevention practices between large,
mid-size, and small industries.  Under the "WRITE Program With
Industry", evaluations of waste reduction technologies are
performed directly with industrial firms or through industrial
trade associations and/or technical societies.


WRITE RESEARCH PROGRAM

     In addition to evaluation programs, the WRITE Program  has a
research subprogram.  This subprogram focuses on pollution
prevention research needs, i.e., the generation of data to allow
the future demonstration of emerging new pollution prevention
techniques.  Projects under this component of the WRITE  Program
address various technical obstacles to waste  reduction and to
chemical-, wastestream-, and industry- specific waste
minimization issues.  These research efforts are conducted with
industrial firms,  universities, other government agencies,
technical societies, and industrial trade organizations.

        While emphasis to date has been on industrial processes
the area of focus is being enlarged to include  techniques for
pollution prevention for any source and especially those that are
identified with posing significant risk in terms of high toxicity
                               433

-------
or high threat or those producing large amounts of pollution.


CONCLUSIONS

        The WRITE Program will continue to expand and address
pollution prevention research needs within government and
industry.  A number of challenging areas have been identified and
are receiving attention.  An effort is being launched to provide
a way to prioritize needed work.  This is especially significant
with the change of focus from the industrial and hazardous waste
area to all sources of pollution.  Another area deals with the
means by which technologies can be best evaluated in terms waste
reduction and cost when the differences between the old and the
new technology involve different pollutants at differing
toxicities, concentrations and exposure pathways.
  Additional information concerning this and other EPA pollution
prevention research programs can be obtained from the  Pollution
Prevention Research Branch  of the Risk Reduction Engineering
Laboratory, U.S. EPA, Cincinnati, Ohio.
REFERENCES
1.
4.
6.
U.S. Congress.  Hazardous and Solid Waste Amendments.
Washington, D.C. 1984.

U.S. Environmental Protection Agency.  Report to
Congress, Minimization of Hazardous Wastes.  U.S.   EPA,
Office of      Solid Waste, Washington, DC., EPA/530-SW-
86-033. 1986.

U.S. Environmental Protection Agency.  Pollution
Prevention Policy Statement.  Federal Register, Vol,. 54,
No. 16, January 26, 1989, p. 3845.

Freeman, H.M., and J. Lounsbury.  The U.S. EPA Hazardous
Waste Minimization Program.  Proceedings of American
Institute of  Chemical Engineers Annual Meeting,
Washington, D.C.  1988.

Apel, M.L., H.M. Freeman, M.F. Szabo, S.H. Ambekar.
Guidance Document for the WRITE Pilot Program With State
and Local Governments.  U.S. EPA, Risk Reduction
Engineering Laboratory,  Cincinnati, OH. 1988.

Jacobs Engineering.  Waste Minimization Audit Report:
Case Studies  of Minimization of Solvent Waste from
Parts Cleaning and From Electronic Capacitor
Manufacturing Operations.  U.S. EPA, Risk Reduction
Engineering Laboratory, Cincinnati, OH,
EPA/600/52-87/057.  1987.
                                434

-------
  TABLE 1.  SUMMARY OF ENGINEERING EVALUATION FOR COLD SOLVENT
                    CLEANING PROCESS MODIFICATION [4]
Type of
Application

Stage of
Development

Unique Nature of
Technology

Applications of
cold solvent
Performance
Need for
Modification
Process modification
Demonstration
First-of-a-kind demonstration
Reduces hazardous waste generation in the
cleaning operations of the parts cleaning
industry, which  is a medium/small-scale
operation.

Achieves 50% waste reduction by reducing the
fresh  solvent requirement by 33%

Requires essentially only minor equipment
modification.
Cost Effectiveness  Added capital costs = $ 600
of Technology       Net operating savings = $ 380 per year
                    Payback period = 1.6 years
                    Note: Net operating savings include savings
                    resulting  from  reduced waste  disposal,
                    reduced  solvent requirement,  and operation
                    and  maintenance expense.
Safety & Health
 Properly designed  system  is considered  safe.
 Metal  cleaning  systems are routinely used  in
 industry without any  safety or health
 problems.
                               435

-------
CC
O
O
cc
CL
111

5E
UJ


UJ
DC
cn
CL
CQ

CO
                 CO
   CO


o<^



ZDOO

CQ
UJZ>
   CO
                                          O
                                          rr
                                          "3 «3
                                          < ^
                                          LU<
                                          CODC
                                          cc
                                                 «
                                                       i=Ol-
                                       w
DCt±
CD CO
OZ>
DCQ


[r
DC
                                  CM
                                         ; Q ^ LJLJ uj
                                         • ^ {/J O ft"
                                       3 pc N o S
                                                          : LU
                                                   ZZQi
                                                                           CO

                                                                           O)
                                                                           LLJ


                                                                           DC
                                                                           CD
                                                                         CO
                                                                        'a
                                                                                       Q.
                                          into
                              DC
                              CD

                              S:
                                            cc
                                            UJ

                                           ,
                                            Q
                                                     cc
                                                    2 Q
                                                    > y
                                                    o!
                                        Q O
                                            436

-------
                      Solidification/Stabilization as a
                 Best Demonstrated Available Technology for
                Resource Conservation and Recovery Act Wastes

                                     by

                  R. Mark Bricka and M. John Cullinane, Jr.

                          Environmental Laboratory
                           Department of the Army
              Waterways Experiment Station, Corps of Engineers
                                P.O. Box 631
                      Vicksburg, Mississippi 39180-0631

                                  ABSTRACT

     Early in 1987, the U.S. Army Engineers Waterways Experiment Station
(WES) was tasked by the U.S. Environmental Protection Agency (USEPA) to
investigate solidification/stabilization (S/S) as a Best Demonstrated Avail-
able Technology (BOAT).  Under this program EPA supplied WES with a number
of Resource Conservation and Recovery Act (RCRA) listed hazardous wastes for
evaluation.

     Nine listed wastes were processed using three generic S/S processes:
Portland cement, kiln dust, and lime/flyash.  The physical and leaching
characteristics of the treated materials were evaluated.  Physical charac-
teristics were evaluated using the unconfined compressive-strength (UCS)
test.  The leaching characteristics of the wastes were evaluated using the
Toxicity Characteristic Leaching Procedure (TCLP).  The TCLP extracts were
analyzed for a total of 22 metals.  No attempt was made to analyze for semi-
volatile and volatile compounds in the TCLP extracts, due to the absence of
these compounds in the raw wastes.  The effectiveness of the S/S processes
in immobilizing the metals was evaluated by comparing the TCLP extract
result of the raw (unsolidified/unstabilized) waste to the TCLP extract
result of the solidified/stabilized waste.

     Preliminary results of the physical tests indicate that the solidified/
stabilized wastes developed strengths ranging from not measurable to
4400 psi.  The results of the TCLP extractions indicate that mercury and
lead concentrations as high as 26.8 and 1.16 mg/1 respectively, were leached
from the solidified/stabilized wastes.  The effectiveness of the S/S pro-
cesses appears to be dependent on the specific waste and metal of interest.
                                     437

-------
            INTRODUCTION
Background
     Amendments to the Resource Con-
servation and Recovery Act  (RCRA)
enacted through the Hazardous and
Solid Waste Amendments of 1984
(HSWA), impose substantial new
responsibilities on handlers of haz-
ardous waste.  In particular, these
amendments prohibit the continued
land disposal of untreated hazardous
waste beyond specified dates,"
...unless the Administrator deter-
mines that the prohibition...is not
required in order to protect human
health and the environment for as
long as the waste remain haz-
ardous..." (RCRA sections 3004 (d)
(1), (e) (1), (g) (5), 42USC 6924 (d)
(1), (e) (1), and (g) (5) ).

     To expedite the development of
treatment standards, various dead-
lines have been established for
Agency action.  Land disposal of a
particular group of hazardous wastes
is prohibited after certain deadlines
if the USEPA has not set treatment
standards under RCRA section 3004 (m)
for such wastes or determined, based
on a case by case specific petition,
that there will be no migration of
hazardous constituents from the
wastes for as long as the wastes
remain hazardous.  Additional dead-
lines result in conditional restric-
tions on land disposal taking effect
if treatment standards have not been
promulgated or if a petition has not
been granted.

     Treatment standards will be
established based on Best Demon-
strated Available Technology (BOAT)
and developed in accordance with RCRA
section 3004 (m).  USEPA (1986a)
defined a technology as best, demon-
strated and available as follows:
     a.  Best—if several technologies
are available for treating the same
(or similar) waste(s), the waste
treatment method which reduces the
concentration and/or the migration of
contaminants most effectively is con-
sidered best.

     b.  Demonstrated—for a waste-
treatment technology to be considered
demonstrated, a full scale facility
must be known to be in operation for
treating the waste.

     £.  Available—for a waste-
treatment technology to be considered
available, it must:  (a) be able to be
purchased or licensed from the propri-
etor if a technology is a proprietary
or patented process; and (b) provide
substantial treatment.

Solidification/stabilization (S/S) is
one technology that meets the demon-
strated and available criteria (USEPA
1986b).  This process has been pro-
posed as a hazardous waste treatment
method that has the potential to sub-
stantially reduce the likelihood of
contaminant migration at land disposal
sites.

Purpose and Scope

     The purpose of this study was to
determine if S/S techniques could be
applied to specific wastes and to
determine the effect of S/S on the
wastes.  The physical and chemical
properties of the treated wastes were
evaluated to determine if S/S tech-
niques substantially reduced the
amount of hazardous contaminants in
the leachate and improved the physical
handling properties of the wastes.
The study described in this paper was
initiated by the USEPA to evaluate S/S
technology as a BDAT and to develop
data to support the establishment of
treatment standards.
                                     43G

-------
     This paper is not intended to,
nor does it, make any attempt to
determine whether S/S is a BOAT for
any specific waste.  This determina-
tion will be made by the USEPA in
accordance with their regulatory
procedures.

        MATERIALS AND METHODS

Wastes of Interest

     Nine waste types listed under the
first third scheduled wastes (USEPA
1988) were evaluated in this study.
Table 1 lists a description of these
wastes, their appropriate USEPA waste
identification numbers, and a single
contaminant of interest.  While many
contaminants for each waste were mea-
sured, only a single contaminant is
discussed to simplify this paper.

General Description
of the S/S processes

     Three S/S processes were used to
treat the wastes and are differen-
tiated by the type of binder material
used in the process.  The three pro-
cesses included:  Portland cement,
kiln dust, and lime/flyash.  The S/S
processes involved the addition of
water and binder material to a waste
followed by a mixing and curing
period.  A schematic diagram of S/S
processing performed during this
study is shown in Figure 1.

Initial Screening Test

     The moisture content of the
wastes varied from almost zero per-
cent to 95 percent water (water/
weight basis).  The first step was to
determine if enough water was avail-
able in the waste for binder hydra-
tion.  An initial screening test was
used to determine the appropriate
water-to-waste/binder ratio for each
S/S process and also to narrow the
range of binder-to-waste ratios used
in the detailed evaluations.  For the
purpose of initial screening, the Cone
Index Test (CI) (Bricka et al. 1989)
was used to measure the strengths of
various water—to-waste/binder mixtures
after they had cured for 48 hours.  By
preparing several water-to-waste/
binder mixtures and subjecting each
mixture to the CI test, the water-to-
waste/binder mixture that developed
substantial strength could be
selected.  An example of the screening
matrix for the Portland cement binder
is given in Table 2.

Preparation of Specimens
for Detailed Evaluation

     Based on the results of the ini-
tial screening test, one water ratio
for each binder was selected for the
evaluation of each waste.  A minimum
of 4 and a maximum of 9 binder-to-
waste ratios were selected for addi-
tional study.  Batches of the water-
to-waste/binder mixtures were prepared
and either poured or compacted into
2 by 2 in. brass molds and cured at a
temperature of 23° C and 98-percent
relative humidity for a minimum of
24 hr (Bricka et al. 1989).  Specimens
were removed from the molds when they
developed sufficient strength to be
free standing and were cured at the
same conditions until used for further
testing.

Physical and Con-
taminant Release Testing

     Unconfined Compressive Strength.
Unconfined Compressive Strength (UCS)
was used to define and characterize
the effects of S/S on the physical
characteristics of the wastes.  The
UCS of each waste specimen was deter-
mined using ASTM method C 109-86 (ASTM
1986).  The only deviation from this
method was the vibration or compaction
of the specimens.  UCS testing was
performed on the cubes after 7, 14,
21, and 28 days of curing.
                                     439

-------
      Contaminant Mobility Testing.
Only  one binder-to-waste ratio was
selected for TCLP  analysis  for each
binder.  The binder-to-waste  ratio
selected for evaluation was the ratio
that  exhibited an  UCS value closest
to but greater that 50 psi.   A UCS of
50 psi was chosen  based on  informa-
tion  found in the  USEPA Office of
Solid Waste and Emergency Response
Policy Directive 9487.00-2A (USEPA
1986c).  TCLP extractions were per-
formed in triplicate for the  selected
binder-to-waste ratio.  A total of
nine  TCLP extractions were  performed
for each waste, i.e. triplicate
extractions for the three binders.

        DISCUSSION OF RESULTS

      The following discussions are
presented on the UCS and TCLP tests.

Unconfined Compressive Strength

      Generally the UCS data,  for each
waste, is presented as UCS versus
curing time at the various binder-to-
waste ratios.  A typical example of
the relationship between UCS and cure
time  is illustrated by the K048/K051
ash waste for the  cement binder as
presented in Figure 2.  The shape of
the UCS binder ratio curves gives an
indication as to the degree of ulti-
mate  strength development that will
be achieved by a S/S waste.  Figure 2
illustrates that the 0.8 cement-to-
waste ratio (binder to raw waste
basis) has achieved the majority of
its strength by a  cure time of
28 days.

      The binder ratio selected for
TCLP  evaluation was based on the 50
psi criteria.  The ratios selected
for each binder and each waste are
presented in Table 3.  The 28 day UCS
for these ratios are presented in
Figure 3.  This figure illustrates
that  for each waste, except for K061,
the cement treated wastes developed
 substantially  higher  strengths than
 the  wastes  treated with  other binders.
 Figure  3  also  illustrates  that many of
 the  binder  ratios selected for TCLP
 evaluation  are in excess of the 50 psi
 criteria.   This can be attributed to
 the  fact  that  it is difficult to
 predict the 28 day UCS and that for
 some wastes it was difficult to add
 binder  and  not exceed the  50 psi
 criteria.

 TCLP Results

     Results for the  raw waste com-
 posite  analysis, the  TCLP  for the raw
 waste,  and  the TCLP for  the treated
 waste are presented in Table 4.  Data
 are  presented  for the contaminant of
 interest for each waste.   All TCLP
 data for the F006 waste  are below the
 detection limit.  In  addition, the
 data for the K031 waste  are not pre-
 sented  because the data  are not avail-
 able to date.   The data  in Table 4
 indicate generally, that there is a
 reduction in the concentration of con-
 taminant measured in  the TCLP extract
 of treated wastes.  Only the cement
 and  lime/flyash binders  for  the K106
waste have a higher concentration of
 contaminant in the TCLP extract of the
 treated waste  than that measured in
 the  TCLP extract of the raw waste.

     Due to dilution by the binder,
 the  TCLP data  for the raw  and treated
wastes  in Table 4 cannot be directly
 compared.  The leaching data can be
more accurately evaluated  if the data
are normalized to the dry raw waste
concentration and presented as the
percent of treatment effectiveness.
Treatment effectiveness was calculated
using the following equation:
            Cdr
                    Cr
                  Wr * Mr
(1)
                                     440

-------
where
     Cdr
      Cr
      Wr
      Mr
     TCLP contaminant  concen-
     tration mass/dry  weight
     untreated waste,  mg/g
     untreated waste TCLP mass
     for the contaminant of
     interest, mg.  (Calculated
     as:  TCLP contaminant con-
     centration mg/1 * TCLP
     extraction solution
     volume, 1)
     wet weight of  the waste
     extracted, g
     Solids content of the
     untreated waste used in
     the extraction
     expressed as a decimal
         Cdt
                    Ct
               Wt * Mt * Bt
                             (2)
where
     Cdt = TCLP contaminant
           concentration/dry weight
           treated waste after S/S,
           mg/g
     Ct  = S/S waste TCLP mass, for
           the contaminant of inter-
           est, mg.  (Calculated as:
           TCLP contaminant concen-
           tration, mg/1 x TCLP
           extraction solution vol-
           ume, 1)
     Wt  = weight of wet S/S waste, g
     Mt  = solids content of the S/S
           waste used in the extrac-
           tion, expressed as a
           decimal
     Bt  = weight fraction of waste
           in stabilized/solidified
           waste calculated as
           follows :
n  - weight of raw waste
   ~ weight of raw waste +
       weight of binder
PT
                    100
                                   *_,.
                                   (4)
                                   Figure 4 presents the normalized
                              data as the percent treatment.  As
                              illustrated by Figure 4 most of the
                              contaminants were effectively immobi-
                              lized by S/S.  The mobility of chrome
                              in the K048/K051-Ash waste and the
                              mobility of mercury for the K106 waste
                              were increased after S/S for several
                              of the binders.

                                           CONCLUSIONS

                                   Several conclusions can be drawn
                              that characterize the effects of the
                              S/S processes on the various wastes
                              investigated in this project.

                                   1.  Small quantities of the Port-
                              land cement binding agent produced
                              materials with an UCS well above the
                              50 psi criterion.

                                   2.  The binders were easily mixed
                              with the wastes, improving the
                              physical handling properties of the
                              wastes.

                                   3.  Typically, S/S resulted in
                              reduced mobility of most of the con-
                              taminants studied.

                                           REFERENCES

                              1.  American Society for Testing and
                              Materials.  1986.  "Construction;
                              Cement; Lime; Gypsum," Vol 0401,
                              Annual Bank of ASTM Standards, Phila-
                              delphia, PA.

                              2.  Bricka, R. M., Holmes, T. T.,
                              Cullinane, M. J., Jr.  1989.  "An
                              Evaluation of Stabilization/
                              Solidification of Fluidized Bed
                              Incineration Ash (K048 AND K051)",  (in
                              preparation), U.S. Army Engineer
                              Waterways Experiment Station,
                              Vicksburg, MS.
 where PT = percent  of  contaminant  not
 leached due to S/S.
                                     441

-------
3.  U.S. Environmental Protection
Agency.  1986a (Nov).  "Best Demon-
strated Available Technology (BDAT)
Background Document for F001-F005
Spent Solvents,"  EPA 1530-SW-86-056,
Vol I, Office of Solid Waste,
Washington, DC.

4.  U.S. Environmental Protection
Agency.  1986b (Jun).  Handbook for
Stabilization/Solidification of
Hazardous Wastes, Hazardous Waste
Engineering Research Laboratory,
Cincinnati, OH.

5.  U.S. Environmental Protection
Agency.  1986c.  "Prohibition on the
Placement of Bulk Liquid Hazardous
Waste in Landfills; Statutory Inter-
pretive Guidance," EPA 530 SW-86-016,
OSWER Policy Directive 9487.00-24,
Office of Solid Waste and Emergency
Response, Washington, DC.
6.  U.S. Environmental Protection
Agency.  1988 (Aug. 7).  Federal
Register, "Land Disposal Restrictions
for First Third Scheduled Wastes;
Final Rule', Vol. 53, No. 159, Office
of Solid Waste, Washington, DC.

ACKNOWLEDGEMENT:  The test described
and the resulting data herein, unless
otherwise noted, were obtained from
research conducted by the U.S. Army
Engineer Waterways Experiment Station
and sponsored by the U.S. Environ-
mental Protection Agency, Risk Reduc-
tion Engineering Laboratory (RREL),
Cincinnati, Ohio, under Interagency
Agreement DW96930146-01-5.
Mr. Carlton Wiles of RREL was the EPA
project officer.  Permission to pub-
lished this information was granted by
the Chief of Engineers and the U.S.
Environmental Protection Agency.
                                     442

-------
  ITER BINDER     WATER BINDER
WASTE
TO BE 	 *-
STABILIZED
V
WATER TO WASTE
AND
BINDER TO WASTE
RATIO
SELECTION


J V
BATCH
PREPARATION


CURING



DETERMINATION OF
UNCONFINED COMPRESSIVE
STRENGTH AT
7. 14. 21. AND 28 DAYS


TOXICITY
CHARACTERISTIC
LEACHING
PROCEDURE
(AFTER 28
DAY CURE)


ANALYSIS
OF
LEACHATE
J V
INITIAL UCS TESTING TCLP
SCREEN TESTING
TESTING ' •-- • •
Figure 1.  Schematic  Flow Chart for Solidification/Stabilization
           Processing
  4.500
                               CURE TIME. DAYS
Figure 2.  An example of UCS  verses Cure Time  at Different Binder
           Ratios
                                443

-------
2.500
2.000
1.500
IOflft

Q. t-/v*\
. 500
CO
o
400
300
zoo
100
••
-i— t
J*£
\i
•wp*
±
^r
=
=
	
~""

—
:•:•:
LEGEND
^^ CEMENT
ty-yj KILNDUST
^?3 LIME/FLYASH
I

0 = '•'•'-• ^ — *:-£S3
0 0
DE: K048/K051 K106

— — — —
liL 1^ |


^q
CvlTtCM CMTfCM IOWU)
ooo odd °. °. °.
CM CM ° ° ?
8
o
F006 K004 K061
E
I




CM




-H
JJJ
R

b





r~
= D
lit II

•»N loom ^ *-. in
" O Or^O WW1^ OO°
S S ° °' 2 §
o o u> o
o o
o'
K044 K046 K031 K031
Figure 3.  Twenty-eight day UCS for Each Waste Studied at the
Binder-to-Waste Ratios Utilized in the TCLP Evaluations
                             444

-------
1 1U
100
90
80
70
60
50
40
t- 30
z.
HI 20
OC 10
Ul
0. 0
en -10
Ul -20
LU
>
H- -7.0000
ill
£ 10,0000
ill
j_ -13,0000
Ul -16,0000
< -19,0000
H -22,0000
-25,0000
-28,0000
-31,0000
-34,0000
-37,0000
-40,0000
CONTAMINEN1
WASTE CODE:

« ^— ^—



_
-

A)"
rn




::•:•:•
X;X

1






^^
LEGEND
t=4 CEMENT
Ixwl KILNDUST
^^ UIME/FLYASH
i i i i t ii i
' i i 1 1
1": Cr Hg Pb Pb Zn
K048/KO51 K106 K004 K061 K044
IHiiiMiiHiiiimiij
m

i
Pb
K046
Figure 4.  The Percent Treatment Effectiveness for the Waste and
Binder Ratios Studied.
                              445

-------
                 TABLE 1.  DESCRIPTION OF RCRA WASTE EVALUATED

No.
1.

2.
3.
4.
5.

6.


7.

8.
9.
EPA Waste
Code
K048/K051-Ash

K106
F006
K004
K061

K044


K046

K031-Liquid
K031-Solid

Waste Description
Ash from the Incineration
of API Sludge
Mercury Cell Sludge
Electroplating Sludge
Chrome Pigment Sludge
Emission Control Dust from
a Steel Electric Furnace
Wastewater Treatment Sludge
from an Explosive Manu-
facturing Process
Lead Based Initiating
Compounds
Arsenic Salts
Arsenic Salts
Contaminant
of Interest

Cr(III)
Hg
Cr(III)
Pb

Pb


Zn

Pb
As
As
       TABLE 2.   INITIAL SCREENING MATRIX FOR THE PORTLAND CEMENT BINDER
 Binder-to-Waste
	Ratios
      0.1
      0.7
           Number of Specimens at
      Indicated Water-to-Waste Ratios
0.2                 07J077
 1                   1                   1
 1                   1                   1
                                    446

-------
        TABLE  3.   BINDER-TO-WASTE  RATIOS  SELECTED  FOR TCLP EVALUATIONS

Waste Code
K048/K051
K106
F006
K004
K061
K044
K046
K031-Liquid
K031-Solid

Cement
0.2
0.1
0.2
0.2
0.05
1.2
1.0
0.065
0.1
Binder-to-Waste Ratios
Kiln dust
0.2
0.5
0.4
0.4
0.05
1.4
1.4
0.09
0.1

Lime/Flyash
0.2/0.2
0.1/0.1
0.2/0.2
0.2/0.2
0.05/0.05
0.7/0.7
0.7/0.7
0.045/0.045
0.05/0.05
           TABLE 4.   WASTE ANALYSES FOR THE CONTAMINANT OF INTEREST


Waste
ID
K048/
K051-Ash
K106
F006
K004
K061
K044
K046


Contaminant
of Interest
Cr(III)

Hg
Cr(III)
Pb
Pb
Zn
Pb
Raw
Composite
Concen-
tration
(mg/1)
1520

25,900
DL
666,820
20,300
2.61
967
Raw TCLP
Concen-
tration
(mg/D
2.2

0.1
DL
1410
45.1
0.236
103
Average TCLP for


Cement
2.13

28.6
DL
0.54
1.16
0.02
0.078
the Treated
(mg/1)
Kiln dust
1.86

0.0139
DL
7.2
0.787
0.02
1.0
Waste

Lime/Flyash
1.14

6.65
DL
31.2
0.095
0.28
0.4
DL = Compound measured below the detection limit.
                                     447

-------
                     VOLATILE EMISSIONS FROM STABILIZED WASTE

                                           By

                             Leo Weitzman, Lawrence E. Hamel
                                    Acurex Corporation
                                Research Triangle Park, NO

                                 Paul dePercin, Ben Blaney
                           Risk Reduction Engineering Laboratory
                                      Cincinnati, OH
        The 1984 Resource Conservation and Recovery Act (RCRA) amendments, which limited
 the free liquid content of wastes placed in hazardous waste landfills, resulted in significant
 quantities of hazardous waste being solidified prior to disposal.  Recent "Land Ban" restrictions
 requiring that waste be pretreated to the standards of Best Demonstrated Available Technology
 (BOAT) prior to land disposal are also resulting an increased level of solidification/stabilization
 (SS). While the mechanical strength and leaching characteristics of stabilized wastes have been
 the subjects of investigation, there is very little data available on the emissions of organics from
 the S/S process and from the treated waste. The present program was initiated to address this
 data gap. This program measured the organic emissions from synthetic hazardous wastes during
 the mixing and curing stages of the S/S process.

 APPROACH

        The tests were conducted on two synthetic wastes (1) an inorganic "waste" made of soil
 and water and (2) an organic "waste" made of soil, water and latex paint sludge.  Each of these
 wastes was doped with either an equal weight mixture of trichloroethylene  (TCE) and acetone,
 representing a volatile organic, or N-pentanol (amyl alcohol) representing a semivolatile organic.
 Two stabilization agents were used, (1) a 50%, by weight, mixture  of portland cement (Type A)
 and flyash or (2) a 50%, by weight, mixture of lime kiln dust and flyash

        The program evaluated the emissions from 64 samples of  material. First an initial
 program was conducted on 14 samples. In this program,    both  the organic and inorganic
 "wastes" were doped with two levels (2 and 4 percent each) of acetone and TCE. The emissions
 from these mixes were measured during mixing, and at 7 and 14 days. On the basis of this
 preliminary program, a more detailed program was developed involving 50 samples. Of these 50
 tests, 30 were conducted on synthetic "wastes" contaminated with  TCE and acetone and
 20 additional tests on the same type of "wastes" contaminated with n-pentanol. Three levels of
 contaminant (1,2, and 4  percent) in the waste were evaluated as part of this program. In
 addition, the sampling of the emissions was continued for 30 days  rather than the 14 days of the
 preliminary work.

       The organic dopants used in this study were acetone and trichloroethylene, representing,
 respectively, a water soluble and a relatively insoluble organic compound.  Phenol was also
 examined as a potential candidate, but it posed major analytical difficulties  as it elution time
through the GC column is much longer than for the other compounds.  More importantly, its vapor
pressure is so low that it would be difficult to detect. Furthermore, phenol will react with highly
alkaline materials found in stabilization agents to form nonvolatile phenoxide. This would have
further reduced its vapor pressure to the point of nondetectability.
                                         448

-------
       Acetone was chosen because it is a polar, volatile organic material with a very high vapor
pressure.  It is soluble in water and the presence of water reduces its vapor pressure sharply.
Acetone reacts with alkaline materials relatively slowly, forming a salt. Over a period of time, this
reaction could actually tie up the acetone. Conversely, it is conceivable that the S/S process
could actually increase the acetone emissions over leaving the waste as a liquid.  The increase
would be due to the binder removing the free water but leaving the acetone unchanged. The
removal of the water could, conceivably increase the evaporation rate of the remaining acetone,
thereby increasing its emissions. It was hoped that testing over a 30 day period could identify
such a phenomena and help identify the important parameters for predicting organic emissions.

       The TCE was selected to represent a water insoluble, volatile organic material. Based on
discussions with WES researchers, it appeared that the heat associated with the fixation
processes using lime could result in rapid release of TCE during the mixing of the sludge with the
fixating agents.

       The n-pentanol was selected to represent a semivolatile compound.  It is a relatively
nontoxic material, whose vapor pressure is high enough to be measured with the same system as
used for the TCE and acetone, but sufficiently different from those to identify whether and how
the vapor pressure of the compound impacts its emissions.

       The following pieces of equipment were developed and used for the  emission
measurements.

             1.     "Wind Tunnel" System

             2.     "Modified Headspace" Sampling System

             3.     "Sample Venting" System

        The "Wind Tunnel" system (shown in Figure 1) consisted of a Lundberg mixer mounted in
an 8" x 18" (height x width) cross section rectangular duct. The top plate of the duct was tightly
fitted around the mixer head to minimize air infiltration.  A hole was cut directly below the mixer
head, and a clamping mechanism was constructed to hold a 5 gallon pail tightly against the duct
and mixer. A hinged door was installed next to the mixer to allow addition of materials while the
system was in operation. The end of the rectangular duct (at the far right on Figure 1) was open
to the air. The other end transitioned into a long 6" circular duct containing an orifice plate,
sampling ports leading to a GC, and  pitot tubes. The duct was heat taped and insulated to
maintain the gas temperature 80°-90 °F (26°-32 °C). The organic emission rates were measured
by measuring the airflow through the wind tunnel with the orifice plate and the concentration of
the analyte with the GC sampling system.

        The second system, the "Modified Headspace" sampling system, shown in Figure 2, was
used to measure the low organic releases from the  solidified "wastes".  It consisted of the same
pail with the lid on it, attached to a pump that pulled air at a measured rate over the surface of the
solidified wastes. The rate of flow of the air over the surface of the pail was  measured by a
calibrated rotometer and the concentration of the compound in question was measured with the
 GC. This allowed determination of the emission rate of the organic constituents over a period of
time.  The GC used here was equipped with a 10-port gas sampling valve which was controlled
 by the GC computer system.  This system switched the GC sampling loop from pail to pail so that
 emissions from up to 10 pails of material could be measured automatically (one at a time) over a
 period of several hours) without moving sampling lines or shifting pails.

        The third system, the "Sample Venting System", was like the modified headspace system
 of Figure 2, but without the GC and the associated gas sampling system.  The "Sample Venting
                                           449

-------
 System"was used to maintain a constant air flow over the curing "wastes" while they were being
 stored in between measurements. In this way, airflow was maintained at a measured rate over
 the surface of the solidified/stabilized material throughout the curing period of 30 days.

        The tests were conducted by putting a weighed quantity of sludge into a five gallon pail
 which could be (but was not at this point) tightly sealed with a lid. The pail was placed in the
 "wind tunnel apparatus and the air flow was turned on. The dopant, and binder were then added
 and the mixer was turned on.  Mixing was continued for 60 to 90 minutes while the gas flow rate
 and concentration were measured at 3-10 minute intervals. The time between measurements
 was fixed by the GC elation time for the constituent being measured, approximately 3 minutes for
 the acetone and TCE, approximately 10 minutes for the amyl alcohol.

        At the end of the mixing period, the mixer was stopped and the mixing arm removed.
 The pail was then covered and placed on a pallet of ten pails, each containing a sample which
 had been mixed in a similar manner. Each pail in the pallet was then  hooked up to the "Sample
 Venting" system for the duration of the testing.  Every three to four days, the emissions from that
 sample were measured by removing that pallet from the sample venting system and placing it
 onto the "modified headspace" system. The automated sampling and analytical system
 measured the emissions from each pail on a rotating basis (one every approximately 3 minutes)
 over a period of three to four hours. In this way, each sample was cured for 30 days under a
 constant airflow, with the organic emission rate measured over an extended period every 3 to 4
 days.

        Early in this program, a concern arose that the results from the "Headspace" and
 "Modified Headspace" measurements might be a function of the system, rather than an indication
 of the actual rate of release of the organic compound from the solidified "wastes." The concern
 was that the air flowing over the surface of the "waste" in the pail might pump or purge the
 volatiles out of the solid.

        To determine whether this was the case, a test was set up whereby a sample of the
 "waste," contaminated with TCE and acetone, solidified and placed in  the "Modified Headspace"
 system.  The flow of air in the headspace of the pail was varied in a random manner between 0.2
 to 5.0 liters per minute. After each change of air flow the system was allowed to reach steady
 state and the emission rate of the acetone and TCE were measured. The results are given in
 Rgure 3.

       As can be seen, the organics emission rate  is a function of the flow rate up to
 approximately 1 liter per minute.  Above that flow rate the emissions are limited by the diffusion
through the solid phase. Functionally, this means that above 1 l/m, the loss of the organics from
the solidified mass is limited by diffusion through the solid rather than through the boundary layer
between the solid and the bulk gas phase. On the basis of this test sequence, the air flow rate
through the pail was selected to be in the three liters per minute range. This is well above one
liter per minute, but still low enough so that the air does not dilute the organic compound being
measured to below the level of detectability.

RESULTS

       Including blanks and replicates, this program resulted in approximately 70 sets of
emission curves. Figure 4 shows a typical emission curve generated for a sample. The  left hand
curve shows the emissions (in g/min) during the mixing, and the right-hand curve shows the
emissions from the solidified material during curing.  As can be seen, the time-scales for  the two
parts of these curves are different, the left portion gives the time in  minutes, the right portion in
days.  Figure 5 presents the results for the same sample in a cumulative format. These curves
                                           450

-------
show the cumulative percentage of organic constituent released during the mixing and during the
curing.

       As discussed in the "Conclusions", below, a statistical analysis of the data indicated that
only the type of binder used and the type of organic constituent influenced the level of emissions
during the mixing step to a statistically significant extent. Six average emission curves one for
each of the three organics tested and for each of the two binders were, therefore calculated by
averaging the emissions for all tests at each of these conditions. The results are given in
Figures 6 through 8 which show the cumulative percent emissions from the "wastes" for each of
the three organic compounds. Each of these figures shows the cumulative emissions for the
wastes during mixing, when the LF and PF solidification agents were used, and for the blanks—
without solidification agent. The curves were obtained by averaging the results of all the runs for
the respective conditions over concentration. These curves can be interpreted as being a first
approximation to an emission factor—the percent of the given component released during the
mixing as a function of time.

       The tests showed that only the type of organic compound being emitted was significant
during the curing portion of the tests. Figure 9 was obtained (for the curing materials) in a similar
manner to Figures 6 through 8 (for the mixing) by averaging the emission measurements at the
specific times.  Note that for all of these cases, "blank" refers to samples of synthetic waste
containing appropriate levels of organic which were mixed and stored in the same way as the
other samples but which were not mixed with any binder. By comparing the  results of each of the
samples to those of the blanks, it is possible to determine what portion of the emissions are due
to the mechanical mixing and handling and what is due to the S/S process.

       The percentages shown in Figure 9 are based on the amount of organic constituent
remaining in the waste after mixing. For example, if 90 percent of the organic constituent was
emitted during mixing, the percentage given in Figure  9 is of the remaining 10 percent. The
curves of Figures 6 through 8 can be interpreted as showing the cumulative percent (the amount
emitted from the waste during mixing) of each organic constituent released from the
solidified/stabilized waste and from the quiescent blank, which had not been mixed with binder
but otherwise handled in the same way.

CONCLUSIONS

       The experimental program described herein was designed to determine the following:

       1.   Can a laboratory system be fabricated to readily measure organic emissions from all
            phases of the stabilization process and from the stabilized wastes?

       2.   Does  stabilization reduce the emissions of organic compounds  from sludges? If not,
            how does it affect them?

       3.  Do nonvolatile organic constituents affect the emissions of organic compounds?

       4.  Does  the type of solidification agents used affect the emissions or organic
            compounds?

       5.  Are the emissions of the different organic compounds proportional to their
            concentration or does another relationship apply?

        6.  How significant is the vapor pressure of the component on its emission?

        7.  How significant is the solubility of the component on its emission?
                                          451

-------
         The first question was addressed by evaluating the overall performance of the systems
 developed to measure the organic emissions from the waste solidification process and from the
 solidified wastes themselves. The systems developed performed very well.  Measurements were
 performed continuously for a period of more than two months without any loss of data. The
 systems proved to be flexible, and were able to measure organic emissions ranging from less
 than one mg/minute to tens of grams per minute—three orders of magnitude.

         Questions 2 through 7 were answered by performing an analysis of variance (ANOVA) on
 the results of the tests.  The tests were conducted by grouping the experiments into categories
 with common parameters of interest, and then comparing the variation between the categories to
 that within each category. This analysis allowed estimation of the probability that the two
 categories are, or are not, of the same population. This is a standard method of testing
 hypotheses.

        The act of solidifying wastes had its greatest impact on the organic emissions during
 stabilization, when all of the components were being mixed. At that time, stabilization with lime
 kiln dust often raised the temperature of the "waste," resulting in emissions that were significantly
 higher than those from wastes processed in a similar manner, but without the addition of
 solidifying agents—blanks. The tests indicated that there was a 99.99% probability that the
 emissions during mixing are higher when lime kiln dust and flyash solidifying agents are used
 than when no solidifying agent is used.

        Emissions during mixing of the components did not differ significantly between the blanks
 and those samples which used portland cement as the solidifying agent. The statistical means of
 the emissions during mixing of the blanks and of the "wastes" solidified with portland cement
 were equivalent respectively. The emissions from the blanks (i.e., unsolidified wastes) were a
 affected by the agitation during the tests, which was expected, although this affect was relatively
 small. The organic compounds were released from the liquid at a very low rate until the liquid
 was disturbed by processing.

        The organic emissions after mixing showed no significant difference between any of the
 samples for each of the three compounds (acetone, TCE, and n-pentanol). The results indicated
 a greater than 90% probability that the method used for solidification does not affect these
 emissions from the solidified wastes.  In fact, they indicated a greater than 90% probability that
 the organic emissions from a solidified waste are the same as those from the quiescent liquid
 waste.  It should be noted, however, that solidification does prevent the emissions from being
 Increased by environmental factors, such as wind or activities which would disturb a pool of
 standing liquid and increase emissions from it.

        The effect of nonvolatile organics in the wastes on the emissions was tested by
 comparing the emissions from the "organics wastes" (those prepared  with latex paint) to those   k
 from the corresponding "inorganic wastes."  The analysis showed that there was a small
 difference between the two populations during mixing, but no significant difference in the
 emissions from the stabilized wastes.

        The proportionality of emissions was tested to determine whether the emissions of each
 compound are proportional to the amount in the waste, or whether another dependency occurs.
 If the emissions are proportional to the concentration then they are governed mainly by
 equilibrium relationships.  If, on the other hand, the relationship is more complex, then diffusion
 and possibly other relationships would have to be included in subsequent modeling. This
 analysis can assist further research programs.

        The results showed no clear trends  in the emissions. Various runs tended to show
differences which  could not be correlated over groups of tests. The results seem to indicate that
                                          452

-------
the organic emissions are proportional to the amount in the "waste." That is, the emissions of
each component from the wastes containing 2% at the start are approximately twice those
containing 1%, and half of those from the wastes containing 4%.

       This analysis addressed whether the vapor pressure of each of the components was a
significant influence on the emissions. It was performed by comparing the TCE emissions
against the n-pentanol emissions for the same parameters. Both TCE and ri-pentanol are
insoluble in water and have widely different vapor pressures.

       The intent of this analysis was actually to see whether their emissions were proportional
to the vapor pressure, and hence, if the governing mechanisms are equilibrium related. The
results indicated that the emissions of each of the three components tested were different during
the mixing step, although the difference was by no means as great as the differences in vapor
pressure would indicate, even considering the impact of dilution on the vapor pressure. The
emissions from the solidified waste (during curing) were also similar The n-pentanol emissions,
while lower than those of the volatile compounds, were of the same order of magnitude. This
reinforces the conclusion that mechanisms other than straight volatilization are involved in all
cases.

       The solubility of the compound in water was tested by comparing the acetone emissions
to the corresponding TCE emissions. This must be classified as a very tentative comparison, as
TCE and acetone have different vapor pressures in the pure state and, clearly, the vapor
pressure of the acetone dissolved in the water is different from the TCE which remains in the pure
state.

       The emissions were significantly different for these two compounds throughout the tests;
however, it was not possible to  ascertain whether the difference was due to solubility or to vapor
pressure. The partial pressure  of acetone drops rapidly when it is mixed with water.  This
dependency on the water content of the system was not noted during these tests.  The
significance of the emission difference was, therefore, tentatively attributed to the solubility
variation.

        In summary, a system has been developed which can rapidly and with reasonable
accuracy measure the emissions of organic compounds from mixing processes  such as S/S.
This equipment was used to measure the emissions of organic constituents from synthetic wastes
and it showed that under the conditions of these tests, significant percentages of the organic
constituents were released to the atmosphere. The experiments resulted in the determination of
emission factors which could be used to estimate the likely levels of emissions from S/S
processes.

        The equipment developed here can also be used as an adjunct to a remediation  effort.
S/S is often used to as part of a cleanup of a contaminated site, to solidify/stabilize a siudge
pond, in place, for example. The air emissions from such an operation have to be considered in
evaluating the applicability of the technology.  The equipment developed under this program can
be used to determine the emissions from this application. Samples of the waste and binder can
be tested in the laboratory and the emissions measured, thus giving a reliable and defensible
 estimate of the environmental impact associated with the remediation.
                                           453

-------
                               ACKNOWLEDGEMENT

       The work described herein was performed under contract to the Environmental Protection
Agency Contract Number 68-02-3993, WD 32 and 37 to Acurex Corporation. This paper has not
been subject to the Agency's Peer Review Process butdoes not necessarily represent the views
of the Agency.  Mention of trade names or commercial products does not constitute endorsement
or recommendation for use. The final report, of the same title, is available from the Risk
Reduction Engineering Laboratory, Cincinnati, Ohio.
               ROOF LINE

        FAN
r
                                        if=TOGC
                                        JlSAMPLING LOOP
                                                                        HINGED PLATE
                  \/ PILOT PORTS SAMPLING PORT ORIFICE PLATE TRANSITION
               DAMPERS
                   Figure 1. Schematic of "Wind Tunnel" Treatment System.
                                        454

-------
          TO VENT
                                S SAMPLING
                               .OOP AND GC/
                  10 PORT STREAM
                  SELECTOR VALV
               PALLET OF 10 PAILS OF SOLIDIFIED-WASTE
                                                      SOLIDIFIED
                                                        WASTE
                                                          TO VENT
  Figure 2. "Modified headspace" sampling system used in WP 37.
0.9  -


O.8  -


O.7  -


O.6  -


O.5  -


O.4  -






O.2  -
                            O      O
                                                     a     a
02                       +

                           FLOW (lUen/mlnuto)
      O    ACETONE            +   TWGHLOROETHYUME


                         Figure 3. Results of air flow tests.
                                 455

-------
                     DOPANT  EMISSION  RATE
1
                        UNUTB
                        O   ACfTONC          •  TOC

                         Figure 4.  Dopant emission rate.
                           IL1B % EMISSION
                           (CUKUXIKC: ACC7QNC M 1CD
    i
                 20
                        UNUttS
                         o   Accrotc
                        Figure 5. IL1B percent emission.

                                  456

-------
in
3?
                  AVERAGE ACETONE EMISSIONS
                  DURING MIXING (SOLIDIFICATION AGENTS)
                              MINUTES
                 Figure 6. Cumulative acetone emissions.
1
in
                     AVERAGE TCE EMISSIONS
                  DURING MIXING (SOLIDIFICATION AGENTS)
                                  30
                              MINUTES
                   Figure 7. TCE cumulative emissions.

                               457

-------
114
5?
                 AVERAGE 1-PENTANOL EMISSIONS
                  DURING MIXING (SOLIDIFICATION AGENTS)
                                 30
                             MINUTES
                 Figure 8. 1-pentanol cumulative emissions.
60
                    AVERAGE VOC EMISSIONS
                            AFTER MIXING
                              DAYS
                  Figure 9. Total VOC emission by VOC.

                              458

-------
  TECHNOLOGIES APPLICABLE  FOR  THE  REMEDIATION  OF  SUPERFUND  RADIATION  SITES

              by:     Ramjee Raghavan
                      Foster Wheeler Enviresponse, Inc.
                      GSA Raritan Depot
                      Woodbridge Ave.
                      Edison, NJ  08837

                      Darlene Williams
                      Releases Control Branch
                      USEPA RREL
                      GSA Raritan Depot
                      Woodbridge Ave.
                      Edison, NJ  08837
                                   ABSTRACT
    This paper identifies technologies that may be useful in removing or
stabilizing radioactive contamination at uncontrolled hazardous waste
(Superfund) sites containing radioactive material.  The radioactive
materials at some Superfund sites consist primarily of waste from radium,
uranium, and thorium processing.  Twenty-five existing Superfund sites are
known to contain radionuclides.

    Sites contaminated with radioactive material pose a unique problem
because, unlike organic wastes,' it cannot be destroyed by physical or
chemical means; it can only decay at its natural rate.  Alteration of the
radioactive decay process thereby changing the fundamental hazard is not
possible.  Several technologies have potential for removing or stabilizing
radioactive material at Superfund sites.  These fall into the categories of
disposal, on-site treatment, chemical extraction, physical separation, and
soil washing.  Applicability of these technologies is controlled by
site-specific factors, and their feasibility must be determined on a
site-by-site basis.
                                     459

-------
                                 INTRODUCTION
    The United States Environmental Protection Agency (EPA) has identified
twenty-five Superfund sites in the country that are radioactively
contaminated.  These Superfund sites are briefly described in the Appendix.

    The radioactive materials at many Superfund sites are by-products of
uranium, thorium, and radium processing in the form of tailings,
contaminated buildings and equipment, and stream sediments.  These sites,
located across the United States, vary greatly in size and may involve
radiation exposure to people who reside on and around them.

    The primary public health threat is from contact with radioactive
materials through a) external whole body exposure to gamma radiation, b)
ingestion of radionuclides in food and water, and c) inhalation of radon
and radon progeny.  Radon and radon progeny are continuously produced
through the decay and decomposition of uranium, thorium, and radium.
Radioactive material cannot be destroyed by physical or chemical means;
they can only decay at their natural rate.  The health risks will persist
throughout the entire decay period if no remedial action is taken.
Possible effects on human health include the increased risk of cancers and
increased risk of genetic damage that may cause inheritable defects in
future generations.

    Technologies that have potential to remediate structures (buildings)
and groundwater are of interest at some Superfund sites, but these are
beyond the scope of this paper.  The soil remediation technologies
discussed in this paper fall into the categories of disposal, on-site
treatment, chemical extraction, physical separation, and soil washing.
Applicability of these technologies to Superfund sites is controlled by
site-specific factors; therefore, their usefulness must be determined on a
site-by-site basis.

                                   DISPOSAL
    The disposal technologies can be in one of two categories:  on-site
disposal and off-site disposal.

ON-SITE DISPOSAL

    Several technologies are available for on-site disposal.  These may be
applied in situ.  These technologies include capping and vertical barrier
wal 1 s.

Capping

    This concept involves covering the contaminated site with a barrier
sufficiently thick and impermeable to minimize the diffusion of radon gas .
(1).  Barrier materials can be either natural low-permeability soils (e.g.,
                                     460

-------
clay) or  synthetic membrane  liners,  or both.

Application:  Appropriate  for  large  discrete contaminated  areas, or  several
  smaller areas that  are close together.
Advantages:   Low  cost,  easily  applied, well-known, and a proven technology.
Disadvantages:  Limits  further use of the site.  The cap must be maintained
as
  long as the contaminant  exists  at  the site.  Also, horizontal migration
  of the  radioactive  material  in  groundwater could still occur.
Experiences:  Exists  for radioactive contaminated soils and tailings  (1,

Cost:  $13-200 per m3 (3).   Low cost for cap only, high cost for
excavation,
  transportation, legal assistance,  and cap.

Vertical  barrier  walls

    Vertical  barrier  walls may be installed around the contaminated zone to
help confine  the  material  and  any contaminated groundwater that might
otherwise flow from the site.  The barrier walls, which might be in the
form of slurry walls  or grout  curtains (4), would have to reach down to an
impermeable natural horizontal barrier, such as a clay zone, in order to be
effective in  impeding groundwater flow.

Application:  Could be considered for large discrete waste material or
  around  several  smaller areas that  are close together.
Advantages:  Simple to install, and  applicable to a variety of soil
  conditions.
Disadvantages:  Restricts  further use of the site, possible deterioration
  of the  barrier  walls resulting from the chemicals contained in the waste,
  would not stop  vertical  contamination to groundwater below.
Experiences:  Exists  for hazardous wastes and not for radioactive wastes
  (4).
Cost:  $33-377 per m2 of vertical face (4).

OFF-SITE  DISPOSAL

    Four  off-site disposal methods are briefly described in this paper:
land encapsulation, land spreading,   underground mine disposal,  and ocean
disposal

Land encapsulation

    Land  encapsulation has been the disposal method most used to this point
in time for low-level  radioactive waste materials.  Land encapsulation can
be as simple as excavating the contaminated material  and,  without further
treatment, hauling it to a secure site.

Application:  Appropriate for wastes that have not been treated,  as well as
  for radionuclides extracted from a soil  or other type of matrix.
Advantages:  Low cost, proven,  workable technology for the disposal of
                                    461

-------
  low-level radioactive wastes.
Disadvantages:  Finding a site is politically and socially difficult.
  Transportation of large volumes also carries certain costs and risks.
  Longevity is a consideration not fully addressed by this disposal method.
Experiences:  Exists for radioactively contaminated soils (3).
Cost:  $276-895 per m3 of contaminated soil (3)

Land spreading

    This technology involves excavation of the contaminated material,
transporting it to a suitable site, and spreading it on unused land,
assuring that radioactivity levels approach the natural background level
for these materials when the operation is completed.

Application:  Appropriate for dry, granular tailings and soils with very
  low level radioactivity.
Advantages:  Simple and relatively inexpensive.
Disadvantages:  Selecting a site is both politically and socially
  difficult.  Also, it could contribute to a non-point source pollution
  problem.
Experiences:  Very limited (1).
Cost:  Not available.

Underground mine disposal

    Underground mine disposal could provide secure and remote containment.
The radioactively contaminated wastes could be excavated and transported
without treatment to the mine site, pretreated for volume reduction, or
solidified to facilitate transport and placement, thus reducing associated
costs.  Movement of radionuclides into groundwater must be investigated and
prevented.

Application:  Appropriate for variety of radionuclides and matrix types.
Advantages:  Would provide a very secure and remote containment.
Disadvantages:  Expensive.  Transportation costs and associated risks need
  to be researched further.
Experiences:  Very limited (1).
Cost:  $399-942 m3 of contaminated soil (1)

Ocean disposal

    A possible alternative to land-based disposal options is ocean
disposal.  This alternative should only be evaluated for naturally
occurring mill tailing wastes and not considered for enhanced radioactive
materials.

Application:  Appropriate for low level radioactive waste.
Advantages:  Offers extreme isolation of low-level radioactive waste.
Disadvantages:  Stringent permit requirements.
Experiences:  Exists for radioactive wastes (1).
Cost:  $332-400 per nr of contaminated soil (2).
                                     462

-------
                              ON-SITE TREATMENT
    Two on-site treatment technologies are briefly described;
stabilization/solidification and vitrification.

Stabilization/solidification

    This method immobilizes radionuclides by trapping them in an impervious
matrix (4).  The solidification agent (i.e., Portland cement, silica grout,
or chemical grout) can either be injected in situ, or the waste can be
excavated, mixed, and returned.

Application:  Can be applied to buried and/or capped material.
Advantages:  Solidification may be able to reduce the release of radon and
  associated radioactivity to acceptable levels.  Solidification also may
  make it easier to transport and dispose of the waste material off site.
Disadvantages:  Long-term effects are not known.  There can be undesired
  reaction between the additives and other types of hazardous waste.
Experiences:  Exists for hazardous wastes and not for radioactive wastes
  (4).
Cost:  $33-248 per ton of contaminated soil (4).

Vitrification

    This technology immobilizes radioactive contaminants by trapping them
in an impervious matrix.  The in situ process melts the waste materials
between two or more electrodes, using a large amount of electricity.  The
melted material then cools to a glassy mass in which the radionuclides are
trapped.

Application:  Applicable for low-level radioactive wastes.
Advantages:  Minimal site preparation required.
Disadvantages:  Many substances volatilize, requiring gas collection
  system.  Radium may volatilize, therefore extra precautions are required.
Experiences:  Limited (1).
Cost:  $161-600 per m3 of contaminated soil (4)

                             CHEMICAL EXTRACTION


    Chemical extraction generates two soil  fractions.  One fraction
contains the concentrated radioactive contaminants and may require
disposal; the remaining material is analyzed for residual contamination and
evaluated for replacement at the point of origin or at suitable alternative
sites.  The various applicable chemical  extraction techniques include
extraction with water, inorganic salts,  mineral acids, and complexing
agents.
                                     463

-------
EXTRACTION WITH INORGANIC SALTS

    Radioactive contaminants can be extracted by thoroughly mixing soil and
mill tailings in an inorganic salt solution.  The slurry is filtered,
separating the extractant from the solids.  The radioactive contaminant is
separated from the extractant by ion exchange, coprecipitation, or membrane
filtration.

Application:  Should not be used as a pretreatment to an acid extraction
 •process.  The presence of sulfate and hydroxide in soils and tailings
  will negatively impact the efficiency of radium and thorium removals.
Advantages:  A high percentage of radium and thorium removal is possible.
  Simple extraction vessels are needed.  Recycling of salts is possible.
Disadvantages:  Large amounts of salts may be required.  Some salts such as
  chloride may be environmentally undesirable.
Experiences:  Limited laboratory and bench-scale testing (5,6).
Cost:  $50-150 per ton of soil.

EXTRACTION WITH MINERAL ACID

    In these processes, the ore is ground to 28 mesh and mixed with water
to form a slurry.  The slurry is pumped into an acid leach circuit,
maintaining a pulp consistency of 50 percent solids.  The solids are
separated from the leach liquid by physical methods.  The radioactive
material is removed from the leach solution by ion exchange, solvent
extraction, or precipitation. (1,7)

Application:  Removes most of the metals, both radioactive and
  nonradioactive.
Advantages:  High percentage of radium removal is possible.  Uranium and
  other metals would also be removed.
Disadvantages:  Increased operating and capital costs due to expensive
  reagents, higher operating temperatures, and the corosion resistant
  material required.  The resulting chemically leached material may create
  a harmful waste stream.
Experiences:  Extraction of radionuclides from ores (1,7,8).
Cost:  $50-150 per ton of soil.

EXTRACTION WITH COMPLEXING AGENTS

    This process differs from acid extraction in that complexing agents
like EDTA (ethylenediaminetetraacetic acid) are used instead of mineral
acids.

Application:  Radium from soils with low concentrations of thorium.
Advantages:  High percentage of radium removal.  Low reagent concentrations
  required, reagent can be recycled, reducing operating costs.  The process
  works at ambient temperatures, and many of the reagents are innocuous.
Disadvantages:  Reagents are expensive, process would not remove thorium.
Experiences:  Used in extraction of uranium from ores (1,9,10).
Cost:  $50-150 per ton of soil.
                                     464

-------
                             PHYSICAL SEPARATION
    The radioactive contaminants in soils and tailings in many cases are
associated with the finer fractions (9,11,12).  Thus, size separation may
be used to reduce the volume of concentrated material for disposal, leaving
a cleaner fraction.  Physical separation may be used with extraction to
further reduce contaminant volume.  Four physical separation technologies
are screening—both wet and dry, classification, flotation, and gravity
concentration.

SCREENING

    Screening separates soil (or soil-like material) on the basis of size
(13).  It normally is applied to particles greater than 250 microns.  The
process can be done dry or by washing water through the screen.  Screening
is not efficient with damp materials,  which quickly blind the screen.  It
may be particularly effective as a first operation to remove the largest
particles, followed by other methods.

Application:  Appropriate for all soils, can separate fractions as low as
  50 micron in size.
Advantages:  Simple and inexpensive method.
Disadvantages:  Noisy.  Dry screening requires dust control.  Wet screening
  will require separation of contaminants from water.
Experiences:  Mature technology (1,13,14).
Cost:  Equipment cost ranges from $2-5 per Ib/hr capacity.

CLASSIFICATION

    Classification separates particles according to their settling rate in
a fluid.  Several hydraulic, mechanical, and nonmechanical configurations
are available.  Generally, heavier and coarser particles go to the bottom,
and lighter, smaller particles' (slimes) are removed from the top.
Theoretically, classifiers could be used to separate the smaller particle
fractions, which may contain much of the radioactive contamination in waste
sites.  Classifiers could be used with chemical extraction in a volume
reduction process.

Application:  Effective for sandy soil with low clay and humus content.
Advantages:  Low cost, reliable, high continuous processing capabilities.
Disadvantages:  Humus and clay soil are hard to separate by classification.
Experiences:  Extensive use in industry (1,13,14).
Cost:  Equipment cost ranges from $0.10 to $0.50 per Ib/hr capacity.

FLOTATION

    Flotation is a liquid-froth separation process often applied to
separate specific minerals (particularly sulfides) from ores.  The process
depends more on physical and chemical  attraction phenomena between the ore
and the frothing agents, and on particle size, than on material density.
                                    465

-------
Ordinarily, flotation is applied to fine materials; the process often is
preceded by grinding to reduce particle size.

Application:  Effective for extraction of radium from uranium mill
  tailings (12).
Advantages:  If the particle fraction containing the contaminants can be
  collected by the froth, then flotation is a very effective tool.
Disadvantages:  New additives may have to be developed to permit successful
  flotation separation for radioactively contaminated materials.
Experiences:  Extraction of metals from complex ores (1,12,14).
Cost:  Equipment cost varies from $160 to $830 per gallons/minute capacity.

GRAVITY CONCENTRATION

    Gravity concentration is an old and proven technology that takes
advantage of the difference in material densities to separate the materials
into layers of dense and light minerals (14,15).  Separation is influenced
by particle size, density, shape, and weight.  Shaking and other motions
are employed to keep the particles apart and in motion.  Gravity separation
can be used in conjunction with chemical extraction.

Application:  Limited to those soils in which the contaminants are
  relatively coarse and capable of resisting breakage and sliming.
Advantages:  Highly efficient and proven process for a wide range of
  applications.  It gives a high-grade concentrate over a wide range of
  particle sizes and functions well with most soil types.
Disadvantages:  Low capacity.  High throughput requires multiple decks,
  clean water.  Must ensure there is no slime buildup in recycle water.
Experiences:  Extensive use in industry and ore processing (13,14,15).
Cost:  Equipment cost varies from $65 to $75 per gallons/minute handling
  capacity.

                                 SOIL WASHING
    Soil washing uses a combination of physical separation and chemical
extraction technologies.  Contaminated soil or tailings are mixed with
water and/or extraction reagents.  The clean coarse particle sizes are
separated from the liquid containing the fines and radioactive material by
a combination of physical separation methods.  The radioactive material
would then be extracted from the liquid by standard water treatment
processes such as filtration, carbon treatment, ion exchange, chemical
treatment, and membrane separation.

Application:  Depends on chemical reagents used.
Advantages:  High percentage of contaminant removal is possible.  Recycling
  of reagents is possible.
Disadvantages:  Expensive reagents. Chemically leached material may create
  a harmful waste stream.
Experiences:  Only laboratory and bench-scale testing (1).
                                    466

-------
Cost:  $50-100 per ton of soil.
                                 CONCLUSIONS
      The 25 Superfund Sites have radiologically contaminated soil spread
      over a total of 9500 acres and have several contaminated groundwater
      wells.

      Alteration of the radioactive decay process thus changing the
      fundamental hazard is not possible.

      Any choice of remediation technologies for radioactive waste at
      Superfund sites would have to be site specific.  Extensive site soil
      characterization studies would be required prior to development and
      application of most of the technologies.

      Since none of the chemical extraction and physical separation
      technologies has been used in a site remediation situation, their
      application must be approached cautiously.  The same holds true for
      solidification and stabilization processes.  Essentially, only land
      encapsulation has been used to remediate similar sites; ocean disposal
      has been used for low-level radioactive wastes.

      Various remediation technologies have potential to reduce the volume
      of the contaminated waste with an associated increase in concentration
      of the radioactive material.

      Every site remediation involving radioactive materials must involve a
      final environmentally safe disposal site for the radioactive
      materials.

      Even if it proves feasible at a particular site to lower the
      concentration of radionuclides in the soil by physical separation
      and/or chemical extraction to some acceptable level, the "clean"
      fraction is likely to contain traces of radionuclides.  Therefore,
      adequate attention must be given to whether the "clean" fraction may
      be returned to the original site or an unrestricted location or must
      be sent to a disposal site.

                               ACKNOWLEDGMENTS


  The authors wish to express their gratitude for the contribution of F.
Freestone, P. Shapiro, R. Hartley, W. Gunter, and G.  Snodgrass of USEPA; and
G. Gupta of Foster Wheeler Enviresponse, Inc.

                                  REFERENCES
1.    U. S. Environmental Protection Agency.  Technological  Approaches to
                                    467

-------
      the Cleanup of Radiologically Contaminated Superfund Sites, USEPA.
      Report EPA/540/2-88/002, August 1988.
2.    Camp, Dresser & McKee et al.  Draft. Final Feasibility Study for the
      Montclair/West Orange and Glen Ridge, New Jersey Radium Sites, Volume
      1.  USEPA Contract 68-01-6939, 1985.
3.    U. S. Department of Energy.  Long Term Management of the Existing
      Radioactive Wastes and Residues at the Niagara Falls Storage Site,
      DOE/EIS-0109D, Washington, DC, 1984.
4.    U. S. Environmental Protection Agency.  Handbook -- Remedial Action at
      Waste Disposal Sites (Revised).  EPA-625/6-806, Hazardous Waste
      Engineering Research Laboratory, Cincinnati, OH, 1985.
5.    Landa, E. R. Leaching of Radionuclides from Uranium Ore and Mill
      Tailings.  Uranium, 1:53-64, 1982.
6.    Organization for Economic Cooperation and Development (OECD).  Uranium
      Extraction Technology - Current Practice and New Development in Use
      Processing.  OECD, Paris, 1983.
7.    Clark, D. A.  State of the Art:  Uranium Mining, Milling and Refining
      Industry.  USEPA/60/2-74-038m 1974.
8.    Ryon, A. D., F. J. Hurst, and F. 6. Seeley.  Nitric Acid Leaching of
      Radium and Other Significant Radioncuclides from Uranium Ores and
      Tailings.  ORNL/TM-5944, Oak Ridge National Laboratory, Oak Ridge,
      Tennessee, 1977.
9.    Borrowman, S. R., and P. T. Brooks.  Radium Removal from Uranium Ores
      and Mill Tailings.  RI-8099, U.S. Bureau of Mines, Salt Lake City
      Research Center, Salt Lake City, Utah, 1975.
10.   Taskayev, A. I., V. Ya. Ovchenkov, R. M. Altkaskhin, and I. I.
      Shuktomova.  Effect of pH and Liquid Phase Cation Composition on the
      Extraction of 226 Ra from Soils.  Pochvovedeniye, 12:46-50, 1976.
11.   Garnett, John, et al.  Initial Testing of Pilot Plant Scale Equipment
      for Soil Decontamination.  U.S. Dept. of Energy, RFP 3022, 1980.
12.   Raicevic, D. Decontamination of Elliot Lake Uranium Tailing.  CIM
      Bulletin, 1970.
13.   Kelly, E. G., and D. J. Spottiswood.  Introduction to Mineral
      Processing.  John Wiley, New York, 1982.
14.   Wills, B. A. Mineral Processing Technology.
      1985.
15.   O'Burt, Richard.  Gravity Concentration Technology.  Elsevier, New
      York, 1984.
Pergamon Press, New York,
                                    468

-------
APPENDIX:  RADIOACTIVELY CONTAMINATED SUPERFUND SITES
Site Name
1.
2.

3.
4.
5.
6.
7.
8.
9.
-••'
10:

t1.
12.

13.


14.
15.
16.
17.

18.

19.
20.

21.
22.
23-
24.
25.
Shpack/ALI
Haywood Chemical Co./
Sears Property
U.S. Radium Corp.
W. R. Grace & Co.
Montclair, West Orange,
Glen Ridge Radium Site
Lodi Municipal Well
Lansdowne Property
Maxey Flats Nuclear
Disposal Site
West Chicago Sewage
Treatment Plant
Reed-Keppler Park
. Kerr-McGee Off -Site
Properties
Kerr-McGee Kress Creek/
West Branch of Dupage
River
The Homestake Mining Co.
United Nuclear Corp.
We I don Spring Quarry
Monticello Radioactivity-
Contaminated Properties
Denver Radium Superfund
Sites
Lincoln Park
U.S. DOE Rocky Flats
Plant
Uravan Uranium Project
Teledyne Wah Chang
Hanford 200-area (USDOE)
Hanford 300-area (USDOE)
Hanford 100-area (USDOE)
City/County State/EPA Region
Norton/Attleboro
Maywood/Bergen Co.

Orange, Essex Co.
Wayne/Passaic Co.
Essex Co.
Essex Co.
Lodi, Bergen Co.
Lansdowne
Fleming City/Hi I Isboro
- . . .-..••••
West Chicago :
, '•„'
West Chicago
West Chicago

West Chicago


Cibola Co.
Church Rock
St. Charles City
Monticello
San Juan, Co.
Denver

Canon City
Golden

Montrose City/Uravan
Albany
Bentorij cb
Bentoh, CO
Benton, CO
MA/ 1
NJ/II

NJ/II
NJ/II
NJ/II
NJ/II
NJ/II
PA/III
KY/IV

IL/V

IL/V
IL/V

IL/V


NM/VI
NM/VI
MO/VII
UT/VI 1 1

CO/VI 1 1

CO/VI 1 1
CO/VI I I

CO/VI 1 1
OR/X
WA/X
WA/X
WA/X
Acres
31
42

1
6.5

127.0
wells
1.9
25.0-
,: 	
25.0

0.25
--

--


245
170
220
;

40

900
6,550

900
--
--
--
--
Cu yds

270,000

10,000
120,000

350,000
_ _
2,000
178,000
.
40,000

20,000
61,000

..


16,500,000
4,700,000
780,000
182,000

106,000

1,900,000
--

10,000,000
..
1,000,000,000
27,000,000
4,300,000,000
                          469

-------
             RRELevant OA DIAGNOSTICS: MAJOR HNDINGS IN FY 88

                 by:  GuyF. Simes and Ronald D. Hill
                      Risk Reduction Engineering Laboratory
                      U.S. Environmental Protection Agency, Cincinnati, Ohio 45268
                      John R. Wallace, S-CUBED, San Diego, California 92121
                                   ABSTRACT

    This paper evaluates the RREL QA Program by considering its impact on specific projects
undertaken during FY 88. During this time, RREL reviewed 108 QA Project Plans (the major
document describing QA requirements for data generation). Typical problems identified during
these reviews included poorly defined objectives, incorrect sampling and analytical methods,
incomplete or inappropriate methods for measuring precision or accuracy, and sampling strat-
egies which did not match the project objectives. Before experimental efforts began, these con-
cerns were resolved, thereby avoiding potentially serious experimental flaws. A total of 61 on-
site technical audits have been carried out during sampling and analytical activities. At times,
these Technical Systems Audits have resulted in major changes, such as discontinuation of work
at a laboratory. More commonly, problems were corrected promptly by the analytical laboratory
or field team before the integrity of the project was compromised. Audits of Data Quality
(ADQs) were carried out less frequently during FY 88, but nevertheless resulted in substantial
impacts. For example, an ADQ of one project revealed that much of the required QA data had
not been collected, and consequently much of the analytical data were of limited usefulness.
This finding was valuable for EPA policy makers as well as for the management of follow-on
projects. The last component of the RREL extramural QA Program is review of final reports,
which provide constructive feedback to RREL regarding the success (or failure) of the project.
Typical  problem areas identified at this stage were insufficient presentation of QA data or overly
generalized conclusions.
                                        470

-------
                                   INTRODUCTION
     The formal Quality Assurance Program at the Risk Reduction Engineering Laboratory
 (RREL) originated in the late 1970s in response to the need to coordinate the data requirements
 in complex, mterdisciplinary projects. For the purposes of achieving project objectives, it was
 not sufficient that engineers, chemists, economists, regulators, and statisticians each performed
 well in their own areas; rather, it was also necessary that the data requirements and data gener-
 ation activities of all experts be coordinated and communicated in the project planning stage.

     In response to these needs, RREL (then the IERL laboratory) appointed the first Quality
 Assurance Officer in 1978.  Quality Assurance Programs originally emphasized the development
 of improved analytical and sampling methods. In the early 1980s, RREL began requiring that
 data generating projects include a formal quality assurance plan.  Also at this time, most of the
 formal quality assurance functions were initiated.  The period of 1981 to present has been char-
 acterized by a gradual evolution in intensity and approach.

     The primary purpose of the Quality Assurance Program at RREL is to assure that all data
 are adequate to  meet project objectives.  A corollary to this goal is that data be of known quality
 in terms of accuracy, precision, and detectability, and that samples be obtained in a meaningful
 and representative manner.  These goals apply to all measurements (including process measure-
 ments) that may be needed to complete technical, economic, and regulatory analyses.

     Quality assurance involvement in a particular project begins in the planning stage with the
 development of data quality objectives (DQOs).  These are not, as the name might suggest,
 objectives for precision and accuracy, but rather are statements of overall project objectives. For
 example, a data quality objective for an incineration process might be "to demonstrate a destruc-
 tion efficiency of 99.99 percent for a given pesticide at a confidence level of 95 percent." The
 development of DQOs particularly requires the interaction of various managers and disparate
 experts in order to assure that the end product meets the customer's needs.  The second stage of
 quality assurance involvement in RREL projects is the development and review of a Quality
 Assurance (QA) Project Plan which describes all procedures for data generation. This plan must
 be approved by  RREL before data generation can begin. The third stage of quality assurance
 consists of various on-site technical reviews (technical audits) that take place during sample col-
 lection, chemical analysis, or data reduction.  Such on-site inspections evaluate sampling and
 data generation  vis-a-vis the QA Project Plan. Fourth, final reports are reviewed with respect to
 the credibility of data and the validity of conclusions.

    Once data quality objectives are defined, the other QA activities (described above)  can and
 do affect the outcome of RREL projects.  The remainder of this paper presents case studies from
 RREL's experience with QA Project Plans, various types of technical audits, and final report re-
 views, and discusses their impacts on overall project performance. The final section of this
paper attempts to summarize what RREL has learned from these activities and the impact this
experience may have on future quality assurance activities.
                                         471

-------
                 QUALITY ASSURANCE PROJECT PLAN REVIEWS
    The purpose of the QA Project Plan is to relate the overall project objectives to the specific
measurements that are required. Emphasis is placed on sampling strategy and methods, analyt-
ical procedures, measurements of process conditions and requirements, and data treatment, since
these elements are crucial in evaluating the candidate treatment technologies. QA Project Plans
are reviewed to determine whether the proposed data generation procedures are adequate to
achieve project goals. Common deficiencies exhibited by the first drafts of QA Project Plans
include incorrect sampling and analytical methods; poorly defined objectives; incomplete or
inappropriate procedures for measuring precision, accuracy, or detection limits; or a sampling
strategy that does not match project objectives.  Examples of some of these problems are con-
tained in the following case studies.

Case No. 1: A review of a QA Project Plan for the chemical treatment ofvolatiles in water re-
vealed several problems that would have made much of the data  useless. These problems
included sample preservation methods that were inconsistent with the intended analytical pro-
cedures and the use of untested analytical methods that were of questionable validity. These
concerns were resolved in a face-to-face meeting between EPA and the contractors before field
operations began. In the process, approximately $80,000 in costs for unnecessary analytical
measurements were eliminated from the project budget.

Case No. 2: The QA Project Plan for this project, which pertained to the combustion of a herbi-
cide, originally showed several important deficiencies with respect to sample preparation and
analysis:

    (a)  The intended analytical method for the target herbicide was inappropriate and
         likely would have given erroneous results.  This error was especially important
         because this measurement was intended for the calculation of the Destruction
         Removal Efficiency (DRE).

    (b)  It provided for a single stack sample for several types of analytes, each requiring
         mutually exclusive sample preparation procedures.  Had sampling proceeded as
         planned, the required analyses would have been impossible to perform.

    (c)  The GCIMS target analyte list included labile compounds which are easily lost
         during sample preparation and analysis; in contrast, all of the intended matrix
         spike compounds were stable and easily analyzed, and were thus not represen-
         tative of the target analytes. Had these matrix spike compounds been used, the
         accuracy and precision of the key analyses would have been unknown, and the
         validity of any resulting conclusions would have been questionable.

    These three concerns were addressed to a  large extent in the next two drafts of the QA
Project Plan. While normally the second draft of a QA Project Plan is approved, in this case a
third draft was required, and a fourth was prepared that addressed several recommendations
that had been made.

Case No. 3: Another QA Project Plan for an incineration project exhibited serious problems in
the areas of analytical methods and quality control procedures:
                                         472

-------
     (a)  The analytical method for the most important compound, the Principal Organic
         Hazardous Constituent (POHC), had not been specified, although various alter-
         native, but unvalidated, methods were suggested. Under these conditions the
         probability of successfully achieving the necessary analyses was very small.

     (b)  Methods for assessing the accuracy of measurements for target compounds were
         inadequate. The only method for measuring the accuracy of some samples was
         surrogate spiking ofnontarget compounds at different points in sample recovery,
         but matrix spikes with the target compounds were needed. Data obtained under
         these conditions would not have been defensible.

     (c)  The QC procedures were not adequate, and for some analyses, including the
         most critical one, there was no discussion of QC procedures at all.  If analyses
         had proceeded as indicated, data would have been of questionable utility.

     These and other deficiencies were addressed in the second draft of the QA Project Plan,
providing a document that was complete and comprehensive enough that the objectives for the
project could be met.

Case No. 4: Another QA Project Plan which was reviewed during the last year pertained to the
chemical treatment process for PCBs in soil. Problems exhibited in this plan included an inap-
propriate sampling strategy, unclear QA objectives, and the use ofunproven methods. In addi-
tion, this plan indicated an excessive number ofGCIMS analyses that were not needed to
achieve project objectives. These concerns were corrected before the beginning of the project.
Regarding the  large number ofGCIMS analyses, it was recommended that analysis proceed to a
given level ofconfidence rather than to a fixed number of'samples.

     It is safe to say that the value of these and other projects would have been severely compro-
mised, had they proceeded as originally planned. Reviews of QA Project Plans has thus allowed
substantive concerns to be corrected before data generation begins.

           TECHNICAL SYSTEMS AUDITS DURING FIELD OPERATIONS
     During the field sampling and analysis stage of a project, a representative of the EPA fre-
quently visits the site to carry out a Technical Systems Audit (TS A). The purpose of this TS A is
to assure that sampling is being carried out as defined in the QA Project Plan and according to
scientifically sound procedures. Every effort is made to perform this TS A early enough in the
project to permit corrective action, if required, before data quality is seriously compromised.
The results of two field TS As are summarized in the case studies below.

Case No. 5: The first example of afield Technical Systems Audit (TSA) stems from an investi-
gation of the biodegradation rate of organic compounds in an aerated wastewater treatment
system.  This investigation sought to compare biodegradation rates in a full-scale treatment
plant with rates obtained from a bench-scale, in-vitro experiment. During the conduct of the
TSA, it was observed that the in-vitro test apparatus was housed in an air conditioned labora-
tory with an air temperature of 20 ° C, while the treatment plant itself was operating at 38 ° C.
Since biological reaction rates are temperature dependent, it was important that both tests be
conducted at the same temperature.

                   .                     473

-------
     Under the conditions described above, the in-vitro reaction rate could have been low by a
factor of four in comparison to the actual treatment plant results. (Biological reaction rates
decrease 50% for each 10° C decrease in temperature.) Based on the TSA observation, it was
decided to move the test apparatus from the air conditioned laboratory into an environment of
the correct temperature. Again, the problem was detected early enough in the program to pre-
vent significant loss of data. However, had this error in the basic experimental design not been
detected promptly, it would have seriously impacted the validity of the conclusions, and may
have required that the work be repeated.

    The findings of this TSA illustrate a common problem in which the field team is not
entirely informed regarding sampling and analytical requirements. It is unlikely that the field
team would have made the error described above had they been aware of the overall project
objectives.

    While some of the problems revealed during field TSAs are significant enough that they
might compromise the overall validity of the investigation, other less significant though impor-
tant problems are also found. One common problem encountered during field TSAs is that pri-
mary process wastes such as stack gases are sampled in a complete and thorough manner, while
secondary wastes or process feeds are sampled in a less rigorous manner,  as is illustrated by the
following case study.

Case No. 6: During an incinerator test burn it was observed that composite samples of the
feeds and wastes, such as scrubber water blowdown and bottom ash, were being collected by
combining aliquots of unequal volumes.  This procedure would have resulted in biased samples.
Since a primary objective of this project was to measure mass closure, representative samples
were essential, and the sampling procedure was promptly corrected as a result of this TSA.

                   LABORATORY TECHNICAL SYSTEM AUDITS
     Much like the field TSA, the laboratory TSA consists of an intensive on-site review of ana-
lytical and quality assurance procedures at a participating laboratory. Typically, such reviews
are carried out early during the laboratory phase of a project in order to make corrective action
on a timely basis. Most typically, a laboratory TSA reveals concerns or problems which the
laboratory corrects promptly. Perhaps the most common types of errors observed are associated
with a failure in communication between the various project participants, such as laboratory and
sampling teams.  The following case study provides an example of such an occurrence.

Case No. 7: During this TSA, the laboratory itself was seen to be performing satisfactorily. In
fact, the GCIMS analysis had been carefully optimized for the main compound of interest, a pes-
ticide, permitting reliable determination Of the target compound at concentrations below its
normal detection limit.  However, examination of the shipping records at the laboratory for the
project in question brought to light related problems with sample collection and shipment. Two
important concerns were noted:

     (a) Some samples which were specified in the QA Project Plan were absent from the
         shipping log because they were being discarded by the sampling personnel. The
         importance of this circumstance was somewhat mitigated by the fact that the
                                         474

-------
         samples in question were condensate waters from the volatile organic sampling
         train, and volatile organics were not critical compounds to the project.

     (b)  Some samples for VOA analysis were shipped in the same ice chest with samples
         containing high levels of organic compounds. No trip blank was included, unfor-
         tunately, and consequently it could not be determined whether any contamination
         tookplace. Again, the volatiles were not critical compounds.

     In this case, sampling requirements had not been entirely communicated to the sampling
group by those team members responsible for planning and documenting the test. Although the
major objectives of the project were not affected because VOAs were not a critical measurement,
the attainment of secondary objectives was questionable.

     In contrast to the latter case study, laboratory TSAs occasionally reveal much more severe
problems, as is illustrated by the following case study.

Case No. 8: A Technical Systems Audit of a general analytical service laboratory was per-
formed in which a number of concerns were noted, most notably in the areas of GCIMS and
trace metals analysis.

     In the GCIMS area, some of the findings were as follows:

     (a)  Tuning and calibration requirements for the GCIMS had not been met, and blank
         levels had not been checked, resulting in data of unknown quality.

     (b)  Documentation was inadequate in the GCIMS area, making it nearly impossible
         to reconstruct the analytical run sequences. It was thus nearly impossible to
         identify and isolate problems such as sample contamination that may have occur-
         red during a run, and data users could not realistically recheck their results or
         diagnose problems.

     (c)  In one case, a sample run was repeated on two consecutive days and gave results
         for some compounds differing by more than a factor of two. Although these data
         failed precision objectives, and were therefore of known poor quality, they were
         reported anyway.

     The area of trace metals analysis also exhibited some severe problems.

     (a)  An interference check sample had never been run for the ICP analysis; thus, the
         reported data was of suspect quality.                           .

     (b)  In one ICP run sequence, the calibration check failed, yet no recalibration was
         performed, and the suspect data were reported.

     (c)  A continuing calibration standard, which was run in the middle of a sequence
         during an  atomic absorption analysis, resulted in a measurement that was in
         error by a factor often. However, data were processed with no indication of the
         failed standard, meaning that the data from the run sequence were of unknown
         quality.

                                        475

-------
    As a result of this TSA, the GCIMS and trace metals analysis areas were deemed inade-
quate, andRREL suspended activities at this laboratory in these areas. Substantial corrective
actions have been made subsequently and the laboratory can now analyze RREL project
samples.

    This audit took place before the laboratory began running critical RREL samples. Had this
TSA not been carried out, it is safe to say that data quality would have been severely compro-
mised, unbeknownst to the principal engineers requiring these analytical services.

                            AUDITS OF DATA QUALITY
    Audits of Data Quality (ADQs) are performed during the latter stages of data interpretation
or early stages of report preparation. ADQs consist of a retrospective review of supporting data
beginning with raw data and proceeding to values included in the Final Report. ADQs have
been employed less frequently than other types of audits, but have nevertheless led to the iden-
tification of serious problems as well as recommendations for improving data quality.

    While it is impossible to affect the experimental part of a project at this stage, ADQs are
helpful in avoiding misinterpretation of results, as is illustrated by the following case study.

Case No. 9:  During the last year, anADQ was performed in support of a preliminary study of
an extraction technology for treating soils contaminated with volatile organic compounds. This
study required the long-term measurements of volatile compounds in gases, soil, and ground-
water.

    Review of the data from this study turned up several serious problems:

    (a)  Replicate analyses were not performed as required for obtaining an assessment
         of the precision of the methods.  Since it was important to the project to deter-
         mine differences in analyte concentration from the beginning to the end of the
         study, the lack of an adequate estimate of precision was a concern.

    (b)  An independent QC gas standard was not used as a check of the analytical
         instrumentation.  As a result, the data could not be compared from one cali-
         bration period to the next over the entire sampling period.

    (c)  Several sample concentrations fell outside the calibration range of the instru-
         ment.  For this reason, these data had to be regarded as estimates only, limiting
         their usefulness.

    (d)  Method blanks were not run on a daily basis. This means that it could not be
         determined whether interferences and contaminants were below the required
         levels.

    (e)  The holding times for VOST samples were exceeded, which could have affected
         the accuracy of these data.
                                        476

-------
     Because there were several problems found during the audit that could not be corrected,
some of the data from this preliminary study had to be considered of unknown accuracy and of
limited usefulness. While it was not possible to re-analyze the questionable samples, recommen-
dations resulting from this audit should be of benefit to similar projects in the future. It was
noted that many of the problems discovered during the audit were the result of lack of adherence
to the QA project plan.

     This example of an ADQ is perhaps extreme in its consequence, but nevertheless illustrates
that such ADQs are helpful in avoiding misinterpretation of data and potentially costly policy
decisions that might result from unjustified faith in poor data. Other ADQs completed during
the previous year have tended to support report conclusions and have been useful clarifying
ambiguous data sets.

                              FINAL REPORT REVIEW
     The primary objective of the Final Report Review is to determine how well the data support
the conclusions of the investigation. Toward this purpose, this review addresses the credibility
of the approach, sampling strategy and procedures, analytical methods, the implications of qual-
ity control results (e.g., spikes and replicates),  and the editorial quality of the reports.

     Perhaps the most common problem encountered in these documents is the presentation and
discussion of analytical results without discussions of the quality control results. These two
components are directly related, since it is impossible to judge the validity of the conclusions
without having a measure of data quality.  Final Report reviews consistently point out that the
presentation and discussion of QA data is crucial to assessing the validity of project conclusions.
This situation was illustrated by Case Study No.  10:

Case No. 10: This project dealt with the performance of a wastewater treatment system for
cyanide-bearing waste. While the report provided an excellent discussion  of the process itself, it
did not discuss the analytical results in sufficient detail. Crucial analytical data (TCLP results)
were not presented, nor was adequate discussion of the QA/QC analyses provided. From the
data presented it appeared that there was a serious problem with the cyanide analyses, yet no
explanation was provided. Based on the limited information provided in this report, it was not
possible to assess the credibility of the data presented, and the overall credibility of the report
was thus compromised. The review recommended that the complete analytical data set be pre-
sented, including the results of the QA/QC analyses. Given this information, it would be possible
to judge the overall effectiveness of the treatment process described.

     Another problem revealed less commonly by the Final Report review is conclusions that
are overly generalized and not entirely supported by the data. This situation is illustrated by the
following case study pertaining to a solidification process.

Case No 11: This report on a solidification project discussed the location and immobilization
of the hazardous trace metals within the solid matrix based on electron microscopy.  While this
technique can visualize some major constituents, it is not able to detect these metals at the con-
centrations of interest to this project. By identifying which conclusions were valid and by
recommending the removal of those conclusions which were not, the review enhanced the
reliability of this report for EPA's decision-making process.
                                          477

-------
     Another important long-term benefit of the final report review is that it provides "feedback"
 to the QA staff regarding any continuing problems, especially those that may affect future and
 ongoing projects. A case in point is the review of a Final Report for a PCB incineration project
 which brought to light several fundamental problems with implications for future and current
 projects:

 Case No. 12:  The first problem made apparent by this report was that the analytical methods
 were not sensitive enough to allow calculation of the required DRE. Referring back to the
 original QA project plan, it became apparent that project objectives had never been stated in
 numerical terms, and as a consequence, inadequately sensitive methods were selected for key
 analytes. These findings have sensitized the RREL QA staff to the importance of clearly stated,
 numerical project objectives.

     Another problem of broad interest that became apparent during this review was the dif-
 ficulty in interpreting PCB measurements made by GC/MS according to Method 680. These
 difficulties include, among others, the inability to detect PCBs at sufficiently low levels or to
 even define meaningful detection limits. This is a severe limitation for incineration processes,
 which must demonstrate a DRE of 99.9999%. In subsequent projects, RREL QA staff generally
 have recommended that pre-test samples be prepared to simulate the treatment process, and that
 these samples be analyzed and interpreted just as real treatment samples prior to the start of the
 test proper. This step helps to elucidate potential analytical and data interpretation problems
 prior to incurring the cost of a complete field test.

                                LESSONS LEARNED
     Perhaps one method of evaluating the vigor of the RREL Quality Assurance Program is to
ask whether it has had any beneficial effect on RREL projects. As is illustrated by the above
discussions and case studies, it is obvious that the in-depth objective reviews provided by the
QA Program have indeed been instrumental in avoiding numerous errors ranging from the use of
•improper analytical procedures to unsupported conclusions. The review process has also been
useful in facilitating communication between diverse interested parties, such as regulators, engi-
neers, chemists, statisticians, and developers. For this reason, the independent, objective review
has been the mainstay of the RREL Quality Assurance Program and will continue to be so in the
near future.

     Experience during FY 88 has also illustrated areas that will require additional attention in
the near future. One such problem area is that project objectives are often not as well defined as
is preferred. It is thus expected that additional effort will be directed towards the development
of (preferably numerical) data quality objectives. Another area of uncertainty is the selection of
best methods for measuring PCBs for incineration or other destructive processes, as discussed
above. Finally, experience during FY 88 has indicated that unneeded delays sometimes occur
simply because contractors are unaware of what is expected in the QA Project Plan.  For this
reason, it is hoped that additional instructional material will be prepared and disseminated during
the following fiscal year.

     Finally, it must be emphasized that the external review process afforded by the RREL
Quality Assurance Program is no substitute for the commitment of knowledgeable project staff,
                                         478

-------
who are directly responsible for the quality of work performed. However, unlike a project in-
volving a single principal investigator at a given location, RREL projects tend to require the
coordinated efforts of regulators, managers, engineers, chemists, statisticians, and economists,
and under these circumstances it is not surprising that many of the problems revealed by the
review process  arise not from lack of dedication on the part of project members, but rather from
communication problems among the diverse parties involved. Chemical analysts may be un-
aware of regulatory or engineering objectives, engineers may be misinformed regarding analyt-
ical limitations  and possibilities, the sampling team may lack instructions regarding sampling
requirements, and managers may have unrealistic expectations regarding the readiness of the
project team. It is precisely because of the complicated nature of RREL projects that such
potential problems arise, and it is for this reason that the quality assurance process has become
an integral part of RREL projects.
                                          479

-------
               EFFECT OF FEED CHARACTERISTICS ON THE PERFORMANCE
                      OF ERA'S MOBILE  INCINERATION SYSTEM
       by:  James P. Stumbar
            Robert H. Sawyer
            Gopal D. Gupta
            Foster Wheeler Enviresponse, Inc.
            Edison, NJ  08837
Joyce M. Perdek
Frank J. Freestone
Releases Control Branch, USEPA
Edison, NJ  08837
                                   ABSTRACT

    During the past four years, the EPA Mobile Incineration System (MIS),
has processed a wide variety of feeds.  Besides  ->  incinerating the
hazardous materials for which the MIS was designed, the unit has also
incinerated contaminated debris including wood pallets, steel and fiber
drums, and plastics.  This paper identifies significant physical and
chemical characteristics of various feed materials and their relationship
to MIS performance.  The paper also correlates the effect of these feed
characteristics on specific MIS components.  Corrective actions taken to
mitigate several problem characteristics are presented.  The operating
experience with the MIS has provided valuable data on the limits of
incineration capacity as well as reliability of the unit in relation to
various feed stocks.  This information is also discussed.  The information
contained in this paper is directly applicable to field use of mobile and
transportable incinerators at Superfund and other industrial cleanup sites,
                                    480

-------
INTRODUCTION

    Under the sponsorship of the Office of Research and Development of the
U.S. Environmental Protection Agency (EPA), the Mobile Incineration System
(MIS) was designed and constructed to demonstrate high-temperature
incineration of hazardous wastes (1).  The system essentially consisted of
a refractory-lined rotary kiln, a secondary combustion chamber (SCC), and
an air pollution control system.  These three components are mounted on
three separate heavy-duty semi-trailers.  Monitoring equipment is carried
by a fourth trailer.

    The MIS was rigorously tested in Edison, New Jersey, during 1982 and
1983 with PCB-contaminated and other chlorinated organic liquids (2).  The
system was transported to the Denney Farm site in McDowell, Missouri, in
December 1984 for a dioxin trial burn and field demonstration (3,4).  A
total of 900,000 kg of solid and 81,600 kg of liquid dioxin-contaminated
materials was incinerated between July 1985 and February 1986.  During
1987, the MIS was modified to double its capacity and to improve its
reliability.  A second trial burn was conducted on both solids and liquids
contaminated with chlorinated organic compounds and PCBs during August and
September of 1987 (5).  Since 1987, an additional 3,200,000 kg of solids
and 31,000 kg of liquids have been decontaminated or destroyed.

    Over the lifetime of the MIS, a wide variety of feed materials have
been processed.  These materials exhibited differences in characteristics
that affected the MIS in various ways.  Often a particular characteristic
or a combination of characteristics would affect the MIS performance
adversely.  The experiences gained from field operations of the MIS during
the past four years have increased the understanding of the interplay of
feed characteristics with hardware.

    This paper describes the effects of feed characteristics on the MIS
performance; correlates various feed characteristics with affected parts of
the system; describes actions taken to mitigate the resulting problems; and
discusses the limits imposed on*capacity and reliability by the various
feed characteristics.

FEED CHARACTERISTICS

    Both the physical and the chemical properties of the feed determine
incineration system performance (6).  Important physical properties
include: heating value, morphology, density, rheology, ash particle-size
distribution and fusion characteristics.  Important chemical properties
include the composition of the feed as shown by:  organic content, organic
hazardous constituents, acid forming elements such as sulfur and the
halogens, moisture content, and inorganic ash components.  These properties
can affect the operating parameters, the capacity, and the reliability of
the incineration system.  Many of these properties are interdependent as
far as their effect on the incinerator performance.  The manner in which
these properties affect the performance of incineration systems, based on
the experience with the MIS, is summarized in Table 1.
                                    481

-------

















s
zc
cf
OC
a
X
111

c/>
>-

g
i
o
a:
u
=s
§
u.
u.
M
UJ
i

-
UJ
_J
s
i—





















01

f~

u
o
CA
"8
0)
u.


01
i
t-
o
L.
a.
s
4->
U
>s
15





•§

o
"co
fl>
g
g.
g
ca
8.
C>


I
1
**-
ca
o
c.
cn
3:
(0
as
















1
CL
cu tu ••
O 4J 0)L.(U CU (U O
» _n> C/J •• D) DO) _O) •* P) _O)
,f •. iff CO •» "O 4-*^ T» — ' i? JCT3
co to 3 * o co D X 3 3—3 COD
C04-'— 'C0"~ 4-> — J , .E c -c •" '£ •£

M "O E QJ D) "O E E co E 32 E lr— 6 E co E
— f O C- 4-> -r- O 0) t- tic. 5 L. L.L.Q) — ' £-
a. IE ca crt t- n > ca H- CD s: co LU co > o. co
1 . . • o> S
(U CO (U CU* E* C (1)
O) 3 O) C  o> 4->
O) O CO — ' CO U> W
A) tO 3 (U t- "D
3 C co D - cu -
UU O O U. CO 4J • C 08
• p. 4-» (0 — '
*. 4_» * -O3; "-CO
>* Q. >*>s>s >s3C0^4J
• P. t_ IP- «r- IP- >p- CO C H- CU
CO C COCO CO CO O) U C. C-MO
T3 *n T3 T3 "O "O ^^ * 1— O) •'~ "~
m O (Del) CU u> DO*-H COD
flj 
S >: « S b
o> C4-» co *O4-* «
_/ .P- .P- O •— CO CO
% H>o CE&UJ
*. CO 4-* CD 1— •*
cu "n^c-^ CJCDCU cu
t. — i O D £ «- 4-» » t-
DO CU«- OO «p- 3 CO — ' D
+^CJ JSO> •*- 4-» 4-» — • O 4-*
CO CO C*O=C4-> CODE- CO
t, — » 'p- CO Q. (- CUC-O4-' C-
CUC » CO ••- -P- S- X H- O
^ O O O 4-» > C 01 ^i CO O COL.
c c o co 3: co ^ t- — « t_4-j
>. 5 4-fCUC-OO O)>-73COO TI C
co»2 "SlrcQ.* 4-' co
4-»C/l .P- CU — ' E O D4J— 'OC — •
OO CDL..P-3— » — • O •— X 0 -r-CNJ
fv» r T ~y ii y^ Q^ m LU Ctf *sjf f t f O ^ O
CA
S - ^ « - c" 3
CT ^ .i:-<0 Socx ° S
s: i- 3C>-0)C
<: u «4-'>i- c/) ca^^ocjo
ECCO .4J.r- ^.^
•. >. *. CUOCU •• CJCAC_4-> »O4-*
C E C EC4JOT3 S CJOOCO CCCO
^ S •'^4J''^>»C4JC4J 4JCOC"^3C

tT W [^ "^"SSS-lo " Suj-S'S tTuTS
CO "D COT3CO C — ^ "- t- "U — ' O4J I04J4-*
.M CD 4^ CU4-> 35 — ' 3 CU CU O4^CUCA 4-*CJCO
O CU O 0) O 3 O V Q- CU >s fl) t. C O 13 C
EV* ii rw n r^ Qf Q. 0) C3 t^ O 2E CL *~* O£ O "^
CO
CO
4-i C  CO •
c o o o at
* ir-
n u> o .a 4-<
. H- ii- CD W
"5 CO CU C
§M 01 .C C 0)
N O •£ 4-»

CO >* C W) C t_ CO

O >« CUOI ii-4-»CO
n> — • >.6)C "- — uia>j=
C 0 iVocu 30 DQO
•^ ^ ••- — • O) 4-> — f- »^
4J ^. « p 0 0) 4->
CO I- CCU — ' IP-C- JZ
cu o cujcca oco to
33 s: ac^sc z:a. «t
482

-------
EFFECT OF HEATING VALUE

    Heating value of the feed material affects both feed capacity and fuel
usage of the incinerator.  As the heating value of feed material increases,
kiln temperature can increase and sometimes become uncontrollable.  The
kiln also requires greater amounts of oxidant to complete combustion and
greater quantities of inert material to control kiln temperature.  This
temperature increase can limit feed capacity.  The MIS reached its capacity
limit at 1.33 to 1.61 megawatts (MW) heat input to the kiln.  Feed
materials, such as plastics, trash, wooden pallets, and brominated sludge,
had capacity constraints caused by high calorific values.  Maximum feed
capacities for these materials ranged from 90 kg/hr for pure plastics to
859 kg/hr for brominated sludge as shown in Table 2.

    Solid materials with high calorific values cause transient behaviors
that sometimes further limit feed capacity.  Plastics, trash, and wood
ignite almost immediately after they are fed to the kiln.  Gases evolved
from these materials burn rapidly producing a sharp increase in kiln
temperature and a sharp decrease in excess oxygen.  Prior to the 1987
modifications, the MIS was extremely sensitive to these transients, which
caused many feed stoppages.

    After the addition of the LINDER Oxygen Combustion System (OCS), the
MIS response to the above transient behavior was improved, and. feed
stoppages due to low oxygen were virtually eliminated (7).  However, there
were still many feed cut-offs caused by excessive kiln temperatures.  These
were minimized by operating the kiln at the lower end (790°C) of the
temperature range, allowed by the RCRA permit, and by using water injection
to control kiln temperatures.

    For brominated sludge, the behavior of the MIS was somewhat different.
Large oscillations of the kiln temperature and excess oxygen level occurred
even when the kiln was operated at 790°C.  The resulting
over-temperatures (greater than 1040 °C) caused feed cut-offs and loss of
the kiln burners.  Loss of the burners increased the length of the feed
cut-off period.  The operating changes required to alleviate this
phenomenon are described below.

    To reduce the amplitude of the temperature excursions, an automatic
feed cut-off was introduced into the kiln control system.  This stopped
feed whenever the kiln temperature exceeded 945°C or the oxygen level
dropped below 4% (wet).  This action minimized overtemperature  incidents,
but the oscillation frequency was still large.  As shown in Figure 1, about
four oscillations occurred per hour.  Feed was cut off for approximately
eight minutes during each oscillation.  Feed rate was limited to 450 kg/hr
under these operating conditions.

    Observations of the  kiln during the oscillations showed that the sludge
was not igniting rapidly.  Several batches of  sludge would be fed by the
ram before ignition occurred.  After  ignition, flame would fill the kiln
and oxygen flow and temperature would increase rapidly.  After the sludge
                                     483

-------











tu
z


g
u.
s

CO
X
g
CO
o
UJ
Ul
u.
CO
s
a:
u.


0
UJ
u.
u.
UJ


1
CO
u
CO


Ul

1


u
s
UJ
u.


CM
UJ

CD

fr—






















0.
•5 c
4-» (_
2 g
d§


o
H-
g

g
4->
S

]i
J3

o
g 4->
1 2 C
CO 09 CD

•- C
c o> ^

t» c O
c5 8

0)
3

U

1




_e 

co ca co
a> > o
3= **r

CA CO
G> C

U c.
••— O
4-> .=
u E
CO ^^
a. o jc
*o V S







"to

fe
4-»


1
U.


1
O£
«4-
O
i
—3
E
CO
a:

.1
4-*
CO
cu
u

u
u
c/>



o
o
o

f*.
CM
o
1
CO
ro
*";
o





in

CM





o
CM






f".
si-








co





X
CO
1


o
CO
E
c.
u.


0)

3 X
O (D

01 O
0 ~
0 t
% §2
*:
7 IT
ro t-






4-f
'o

a.
CO
o
















































ca
T3
o
0) X
L. 4-»
O "-
H- CO
Q 0)
T3
CO U
0. 0

I CD
8 j*
*-* v*





S

£^ g
— ' CO
0)
•t- T3
O CO
z !a
— • c.
1- O
C. 4->
CU U
> 0
o a
x -g

"° ~0 E
in XI o
N- C. sr
CM O «-
CM O «-

in
co

1

_ _














o
= h-
A





















ca

i
C-

4^
CO

18 ^

«- c/>
1
8 9
^^ £
1
CO
cu
o
c
X
•8
1
ro
c

• r-
***

c 'i x
S g. 1 -§
o -S ^ ro
g. e

4-> o:
c -
s .i
4J
ro D • X
CO. CO 4-*
•^ C 0) .ii
CD — t- 0
O) CO
ro D cj a.
— • I— U CO
co ca » u



o
o
o

N-
S!
1
i
















o
CM
















^j.

m
I-O

0)

4-» O
'« CU
t L.
— • O
• p. H-
O O
CO ^3

E MD
L. CO
CO O
U. T—
1
II
Ul NJ*


g.
CO
13
in
1


C

H-
o
8
H-
L.

°
>s
t.

1C
RJ

in
CO
i
CO
•R

j.








S





























C/l

i
L.
0)

H-
CD

§
O

1
8
\_x





M


._,
-s
CO
^ 2
"cu o

t. u
o o
u a
T.,

g
o






=














o

A



























'5
CO




s:

i
(G
O
0 g.
CO
_c -a
•— o
7 >
ro cu
c &
_j (D

^ ^

(tJ
^ g g.
O CD "-
4-* CD
8 CO 13
— • H-
a co ca



o
CO






_








5












O
o
in
















CO
u



J2
Q.




E

O
C
CO
_fl)
u
c
X
•8
1
ro
m
c

e
CO
0)
CO
JD
c_
o

u
o
• a


o
CM
T
o
o

tn
CO
I
g
r—
in
i
o




o
0


o
in



o


o

0
0
in
o
i
o
o
 r>
CD t~
co m
CO »


^ e
^ 4J























































u
o
•v,
O>

CM

CM
1
O
o
484

-------



























^
T3
V
c
4J
C
O
o

<\J
UJ

CO







































g
to
1

13
tu
4-
c
o
c
o
to
4-»
I

i

'x


c

s
c
Q)
0
"to
3:



o
'c
to
CO
t_
O
01
3
CO

o
ss

O)
c
4-f
to
tu
=c
CO
tu
"o
4J
t.
to
a.
H-
o
a*



to


tu

i

1

u-








co
concern








a>
eg

"g £
tU D)

*-*
Q. ^N

4-> --^
4-f
C

0 *"*
0




g ~
£ S
O
o








O D)
"to 'to
> 0
sx
CO
c
E
o
'i
O J=
V v^
- - •


















«^


CO
o
£ CO
•*_ 
£•
•*-
!5
8. x
ID 4->
O •*-
O
c. to
tu a.
•a to
"8
t- B
w ex

in
CJ

& s

in tn
CO CO
i i




o o






0 0




o !






CD O





0 0




to

tu tu

c. »-

"° 0

11 e
tu t-.
4J tU
CO >

c - c
CD CO 1 C O
tu >. — • at •—
— » to 4J ••- C 4-»
o -a u !M! ~^ (D _,

ftM T3 C 3 C O
«— s C. 4J •. t- O
CO tU O) tU CJ 4-* O
> c J5 cj co
>. tU -r- CO C 4->
T3 4J 3 t- -n "~ to
I 3 O O C •«- L.
ro o H- 3 to o -a
tA
TO
!S •
o tu
Co E

, i g i
~ -§ .£
.- 8 2
;*: t- ca
0
S3"
i
o
in

^

o
§'



«-





ro
i
in



t\j
i
in
ro
o
o

ro
i
*~




s.


o £
f3
4->
_, .X

CO E

•D C
tu t
4^ rj
to .a

o •
1
o
§


•o
CM





in
i



CO
t>






i





i










o
CO



ca
tu
0
o
c.
a.


4-*
instabili
water IDS

o
i
4J 0)
CD <-
O
O t-
tO CD
CO «-
O »-
o
o
i
o


=

=


-o
<>






o
o




o


o
co

^^
i
CD
CO
t>
ro




£






at


3
_cr
















v
B.
C
*
'il
o
1





















































485

-------
had burned several minutes, the flames would extinguish and the oxygen flow
and temperature would decrease rapidly to complete the cycle of
oscillation.  It became apparent that steady ignition of the material was
required to prevent the oscillations.

    The kiln temperature was increased from 790°C to 900°C to provide
the necessary energy to evaporate the water and volatile organics so that
ignition could be sustained.  This operating change was successful in
reducing the oscillations to a minimum as shown in Figure 2.  Maximum feed
rate was increased to 900 kg/hr by the above changes.

    For feeds with high heat content such as brominated sludge, the
capacity of the MIS is increased by water injection.  Due to its high heat
capacity, water provides a very effective heat sink.  Consequently, when it
is used to control kiln temperature, moisture increases the SCC residence
time as compared to the use of excess air.  Figure 3 shows that the use of
water injection can increase capacity by about 20% over the use of excess
air at a given SCC residence time.  The use of oxygen in the kiln enhances
the effect of water injection by allowing further capacity increases.  At
an enrichment to 40% 0? in the combustion air, capacity can increase by
60%.

EFFECT OF MORPHOLOGY

    The morphology of the feed material affects the feed system by causing
periodic jams.  Most of these problems are caused by materials that are
poorly prepared by the shredder due to their morphological
characteristics.  Problem materials consist of wooden pallets, metal drum
closure rings, plastics, trash, clothing, and mud (8,9).

    The feeding of these materials restricts the MIS capacity.  As shown in
Table 2, relatively dry soils can be fed at rates up to 2275 kg/hr but the
presence of plastics and mud reduces the feed rate to 680 kg/hr.

    The shredder is used to prepare the solid feed materials for
incineration.  For most materials, the shredder works extremely well.
However, wooden pallets and metal drum closure rings often cause feed
blockages when shredded in the present equipment.  While the shredder
breaks most of the wooden pallet into 5-cm wood chips, an occasional board
will position itself to go through the shredder as a 5-cm wide by 1.3-m
long sliver.  The same is true for the drum lid rings.  The shredder
sometimes drags a ring through, straightening it but not cutting it.  In
each case, plugging of the conveyor, weigh scale, or ram follows.  The best
solution has been to manually separate and prepare these feed materials by
cutting them into small pieces (about 24 cm in length) prior to shredding.

    The shredder is unable to handle pipe or thick metal pieces.  These
must be manually classified, cut to proper size, and placed on the main
feed conveyor downstream of the shredder.

    The shredder also performs poorly on materials such as plastic,
clothing, trash, and mud.  These poorly shredded materials often jammed at
                                    436

-------
                                                                        Excess
                                                                        Oxygen
                                                                        % Wet
                   -• [-l__ Excess Oxygen -±~ j
          NOTES:
          - Brominated Sludge
          - September 2,1988
          - Before Operating Changes
           Figure 1.   Oscillations In  temperature and
                     excess oxygen In rotary kiln.
Temp.
  °C

 1095
  955
  815
  675
Excess
Oxygen
% Wet

  12.5


  10.0


   7.5


   5.0


   2.5

    0
         NOTES:
         - Brominated Sludge
         - September 8,1988
         - After Operating Changes
         Figure 2.  K11n temperature and excess oxygen
                   after operating change.
                                      487

-------
the doctor blade or restricting dam that was originally used to level the
granular material on the conveyor belt as the belt exits the shredder
hopper.  The doctor blade worked quite well for granular material but it
created large blockages when materials such as shredded plastic, clothing,
trash, metal, or mud were fed.  A roller has been installed to replace the
doctor blade, and this has reduced the number of jamming incidents.

    In one instance, the combination of metal and mud damaged the main
conveyor belt.  The belt was slit for 21 of its 26-m length by a metal
shard which was embedded in mud.  The clump of mud containing the shard had
stuck to the belt, passed under the top-end belt wiper, and lodged in the
underside rollers.  The top-end wiper was repositioned to provide a more
positive wiping action.

    Finely granulated material affected the operation of the ram feeder.
It would bypass the ram head and collect on the backside.  The material on
the backside would periodically prevent the ram from fully retracting.  A
small chain-plug conveyor was installed and timed to convey the bypassed
material to the front side of the ram.  This solution has worked quite
well.

EFFECT OF DENSITY

    The density of the feed determines capacity for many feed materials.
For feeds of typical densities (1.5 g/cc), such as soils, the maximum feed
rates were 2275 kg/hr.  The maximum feed rate was obtained at a kiln
revolution rate of 1.6 rpm, which gives a typical solids residence time of
30 minutes.  For low density materials such as vermiculite (0.096 g/cc),
feed rates up to 364 kg/hr were feasible.

    The density of the feed determines capacity for many feed materials,
because the density of a material is inversely proportional to the volume
it occupies, and the ultimate feed rate for a given material  is limited by
the volume capacity of the kiln.  The volume capacity of the kiln is
determined by the amount of material that can be held by the kiln without
overflowing the corbel.  The corbel is an annular lip on the front end of
the kiln, which rotates with the kiln.  There is a space between the corbel
and the front plate of the kiln.  When material  gets into this space, the
seals of the kiln are damaged.  The maximum volume that the kiln can hold
without spilling over the corbel is about 15% of its total  volume.

EFFECT OF RHEOLOGY

    The rheology of the material affects either the feed system or the
decontamination behavior.  Muddy soils, fed to the MIS, formed clumps of
material, which were caught by the doctor blade and also would stick to the
conveyor belt, weigh scale and ram trough.  The resulting buildup would
periodically plug various parts of the feed system thus reducing overall
feed rates to about 1140 kg/hr.  This is approximately 50% of the maximum
rates achievable with dry soils.  Addition of vermiculite has eliminated
the sticking of the muddy material  while adding only a small  amount to the
throughput weight.
                                     488

-------
    The brominated sludge that was fed had a tendency to form balls up to 8
cm in diameter.  Since the time required for burnout of a sphere is
proportional to the square of its diameter, the large balls require a much
longer residence time in the kiln for decontamination.  At the normal 1.6
rpm revolution speed of the kiln for soil, small smoking particles would
exit the kiln with the kiln ash.  This required limiting feed rate to about
450 kg/hr.  However, when the residence time of the sludge was increased by
reducing the revolution speed to 0.8 rpm, feed rates up to 900 kg/hr were
achievable without smoking.

EFFECT OF HALOGEN AND SULFUR CONTENT

    Incineration of brominated and chlorinated wastes generate the acid
gases hydrogen bromide (HBr) and hydrogen chloride (HC1).  These acid gases
affect the capacity, the blowdown rates that control total dissolved solids
(TDS) in the process water system and the particulate emissions.

    A capacity limit of 115 kg/hr of acid-forming organic chlorides was
encountered during the 1987 trial burn (5).  The capacity limit was caused
by pump cavitation in the quench system, which cools the gases exiting from
the SCC to about 95°C.  The cavitation reduced the quench water flow rate
which activates the protective instrumentation resulting in cut-off of the
feed and shutdown of the burners.  This cavitation was produced by
excessive chlorinated waste feed rate as follows:  The quench water is
treated with caustic solution to neutralize acid gases.  Reaction between
HC1 in the combustion gases and sodium hydroxide (NaOH) in the quench sump
produces effervescence.  The amount of effervescence increases to violent
levels as HC1 flow rate increases.  The violent effervescence reduces the
available net positive suction head (NPSH) of the pump, which causes
cavitation at very high HC1 loads.

    High organic chloride loads also affect particulate emissions through
the phenomenon of mist carry-over into the stack.  The amount of carry-over
is determined by both the HC1 .loading of the flue gas and the TDS of the
process water (5,10).

    In tests performed prior to the trial burn, particulate emissions were
found to exceed the allowable emissions (180 mg/dscm) by as much as a
factor of three.  The relationship between particulate emissions and
organic chloride loading is shown in Figure 4.  Table 3 presents the
analysis of the Method 5 particulate filter cakes, which shows that the
major portion of the particulate was sodium chloride (NaCl) and sodium
hydroxide (NaOH).  The emissions were brought into compliance after a mist
eliminator was installed.

    However, data taken during the 1987 trial burn showed that TDS of the
process water also affected particulate emissions.  As shown in Figure 5,
particulate emissions were proportional to TDS during the trial burn
tests.  The data shows that operation with TDS at 20,000 ppm adequately
controls particulate levels at high chlorine loadings.

    The TDS is controlled by adjusting the blowdown rates as follows:  For
                                    489

-------
                     TABLE 3.  RESULTS OF ANALYSES OF METHOD 5
                         PARTICULATE FILTER CAKES FROM  HIGH
                       ORGANIC CHLORIDE LOADING TESTS OF MIS
   Test Description
   Date  (1987)
High Chloride
    without
mist eliminator

   6/18-19
High Chloride
     with
mist eliminator

     7/20
   Total particulates

        Iron  (Fe)
        Chromium  (Cr)
        Sodium (Na)
        Chloride  (Cl)
        Aluminum  (Al)

   As Sodium Chloride  (Nad)
   Remaining Na  as Sodium Hydroxide
   (NaOH)  (Excess caustic)
(weight in grams)

   0.3412

   0.0202
   0.0022
   0.1560
   0.1367
(weight in grams)

   0.0524

   0.0057
   0.2253
   0.1174
   0.0175
   0.0200
   0.0022

   0.0329
   0.0079
      Organic chloride loading » 71.5 kg/hr
      Organic chloride loading = 83.4 kg/hr
                 TABLE  4.   RESULTS OF ASH DEPOSIT ANALYSIS
                         FROM MIS CYCLONE RISER DUCT
Analyte
Silicon
Aluminum
Titanium
Iron
Calcium
Magnesium
Sodium
Potassium
Sulfur
Phosphorous

(Si)
(Al)
(Ti)
(Fe)
(Ca)
(Hg)
(Na)
W
(S)
(P)
Reported as
Silicon Dioxide
Aluminum Dioxide
Titanium Dioxide
Ferric Oxide
Calcium Oxide
Magnesium Oxide
Sodium Oxide
Potassium Oxide
Sulfur Trioxide
Phos. Pentoxide
Amount
(wt. %)
25.6
8.5
0.3
5.0
36.7
1.5
0.4
0.4
21.0
0.5
Analytical
technique
X-ray
X-ray
X-ray
X-ray
X-ray
AA
AA
X-ray
X-ray
X-ray
AA » Atomic Absorption Spectroscopy
                                       490

-------
      Ill
      H
      Ul
      IU
      a
             5.
            4 .
            3.
            2 . .
            1 -
                        \
                           Permit Limit
                    0.2
                         0.4
                               —i—
                               0.6
—1—
 0.8
T—r—
   1.2
                                                       1.4
                                                             1.6
—I—
 1.0
                                  FEED RATE (kg/hr)
               NOTES:
               Solid Feed Heating Value    1.50 Kcal/g
               Kiln Temperature           925°C
               SCC Temperature           1200°C
               A - Water injection using 40% oxygen-enriched air for
                   combustion
               B - Water injection using air for combustion
               C - Air cooled using either air or 40% oxygen-enriched
                   air for combustion

               Figure  3.  Effect of cooling media on SCC residence time.
VI
en
CO
Ul
I
     200 .
     100 .
                    20          40          60
                    ORGANIC CHLORIDE LOADING (kg/hr)

         NOTES:
         + Before Addition of Mist Eliminator
         0 After Addition of Mist Eliminator

      Figure 4.  Effect of organic chloride loading
                on particulate  emissions.

                                  491
                              100

-------
a-chlorinated waste 1.65 kg NaCl is formed per kg Cl in the waste.  To
maintain the IDS at 20,000 ppm, 82.4 kg of process water must be drawn from
the system for every kilogram of Cl processed.  This illustrates that the
acid content determines the required process water blowdown rate.

EFFECT OF MOISTURE

    Moisture content affects incinerator performance and can adversely
affect rheological behavior as described above.  Depending upon the heat
content of the waste, moisture can either improve or impede incinerator
performance.

    When using feeds with high heat content moisture acts as a heat sink to
control kiln temperatures.

    Conversely, when using feeds with low heat content, moisture increases
auxiliary fuel requirements and decreases SCC residence time causing a
capacity debit.  The effect of moisture on SCC residence time is shown in
Figure 6.

EFFECT OF PARTICLE SIZE DISTRIBUTION

    The particle size distribution of the ash generated from the waste
determines the amount of particulate carry-over from the rotary kiln to the
rest of the system.  The importance of this characteristic can be
demonstrated by the MIS experience with Denney Farm soil and Erwin Farm
soil.  As shown in Table 2, Denney Farm soil is much coarser than the Erwin
Farm soil.  Up to 25% of the Erwin Farm soil would carry over from the kiln
to the SCC,  This caused a rapid buildup of solids in the SCC.  The solids
buildup necessitated a 70-hr shutdown for removal of the slag after each 96
to 120 hours of operation (45,000 kg of soil processed).  The behavior of
the silt caused the unit to be unavailable for operation an average of 40%
of the time due to the need to clean out the SCC.  On the other hand, the
unit could process the coarser Denney Farm soil for about 600 hours
(270,000 kg of soil processed) before a shutdown for slag removal from the
SCC was required.  The unit was unavailable for 10% of the time due to SCC
clean-outs with the coarser Denney Farm soil.  In both cases, the buildup
of solids in the SCC significantly reduced the availability of the
incinerator.

    The problem was mitigated in 1987 when a cyclone was added, between the
kiln and the SCC, to remove the fines carried over from the kiln.  The
system operated over a three-month period and processed over 500,000 kg of
solid material without requiring a shutdown for slag removal from the SCC.

    Although the cyclone has alleviated the solids buildup in the SCC, fine
particulates still have caused problems with the operating instruments.
The large number of fine particulates associated with brominated sludge
fouled the kiln oxygen meter and the SCC thermocouple about once every
eight hours.  This increased the number of over-temperature incidents in
the rotary kiln, caused the incinerator feed cutoff to actuate due to a
false SCC low temperature measurement, and increased the fuel flow to SCC
                                     492

-------
     1
     55
     0)
     UJ
     3
     I
                                 20
                          TOTAL DISSOLVED SOLIDS - TDS (ppm)
               NOTES:
               Solids Feed Rate           1800 kg/hr
               Liquids Feed Rate          60-160 kg/hr
               Organic Chloride Feed Rate  50-74 kg/hr

         Figure 5.  Effect of TDS on particulate emissions.
ui
P
in
o
UI
Q
DC
8
CO
                               30
                             MOISTURE (%)

         NOTES:
         Solid Feed Rate         1820 kg/hr
         Kiln Temperature        925°C
         SCC Temperature       1200°C
         30% Oxygen-Enriched Air
       Figure 6.
Effect  of feed moisture content
on SCC  residence time.
                                  493

-------
       Sample 1
   QUENCH ELBOW
1.5x
   Sample 2
CYCLONE RISER
DARK SURFACE
Figure 7.  Macrophotograph of samples from Quench Elbow (1)
         and Cyclone Riser (2).
                            494

-------
Figure 8.   SEM photomicrographs of the cross section
           of the deposit from the Cyclone Riser.
                                   495

-------
burners.  The thermocouple problems have been eased by changing the
thermocouple location and using a thermowell rather than an aspirating
thermocouple.  No satisfactory solution has been found for the kiln oxygen
measurement.

EFFECT OF ASH FUSION CHARACTERISTICS

    The ash fusion characteristics of some feed materials caused the
formation of hard slag deposits in the kiln; consolidated deposits in the
ductwork between the kiln and cyclone, between the cyclone and SCC, and in
the cyclone exit tube; and consolidated deposits in the SCC exit venturi.
The ash fusion characteristics are determined by the chemical composition
of the ash.

    Ashes containing elements such as sodium (Na) and potassium (K) have
low slagging temperatures.  Ashes from plastics, glasses, wood, and other
components of trash are  rich in these compounds.  The increased slagging
tendency of trash was experienced in the rotary kiln, which required a
system shutdown about every ten days to remove the slag build up caused by
incineration of trash.

    Ashes containing significant quantities of calcium  (Ca), iron  (Fe),
sulfur  (S), or phosphorous  (P) have moderate slagging temperatures.
Although brominated sludge, containing both Ca and S did not slag  the kiln,
the ash produced a consolidated deposit, shown in Figure 7, which  fouled
the ductwork between the kiln and the SCC.  A consolidated deposit also
occasionally formed in the  quench elbow upstream of the quench nozzles.
This  fouling necessitated a system shutdown about every twelve days to
remove the deposits.

    Samples of the deposits were analyzed  to determine  the mechanism of
-deposition.  More details on this topic are provided in reference  11.  The
deposit mechanism was found to be similar  to those operative in boilers
fired with subbituminous coal.  The deposits were formed by  sintering of
calcium sulfate  (CaS04)  in  the temperature range of 870 and  980°C  in an
ash containing 14% CaS04, 23% calcium oxide  (CaO), and  about 2.5%  sodium
oxide (Na20).  The formation of fused calcium silicates as a result of
the decomposition of CaS04  in the presence of quartz and aluminosilicates
between 870-980°C was also  an important factor in the mechanism of
deposition.  A photomicrograph showing the sintering is presented  in Figure
8.  Table  4 gives the bulk  composition of  the deposit from the cyclone
riser.

    Most  ashes consist mainly of  aluminum  (Al) and silicon  (Si), which
generally  have good fusion  characteristics (fusion temperatures above
1650°C).   Both the Denney Farm and  Erwin soils were composed mainly of
SiOo  and AloOo.  The lack of  slagging  and  of troublesome deposits
experienced with the MIS when processing these materials demonstrates these
good  fusion  characteristics of Al  and  Si.
                                     496

-------
CONCLUSIONS

    The paper has  shown how various feed characteristics affected the MIS
performance.  Concerns stemming from these effects have been discussed and
are summarized below:

    o     Increased heating value of the feed often reduces the fuel
          requirements of the unit.  However, as heating value increases
          control of kiln temperature becomes important and feed capacity
          can be restricted.  Water injection can increase kiln capacity
          under these conditions.  The experience with the MIS also shows
          that operating conditions must be adjusted to insure rapid
          ignition of solids that have high heating value.

    o     Proper feed preparation is very important for maximizing
          throughput.  Problem materials should be sorted out and specially
          prepared prior to feeding.

    o     Rheology can affect either the feed system or the decontamination
          behavior.

    o     Increased halogen content increases mist formation and can
          increase particulate emissions.  TDS of process water can also be
          important in controlling particulate emissions.

    o    Moisture content increases fuel requirements and can create
         feeding problems for muddy feeds.

    o    Materials containing large quantities of micron-sized particles
         can foul critical instruments and decrease reliability of the
         system.

    o    Ash fusion characteristics can cause formation of slag or other
         deposits in various parts of the system.   This also decreases
         reliability.   The processing of pure trash produced slagging
         problems in the rotary kiln.

Most of this experience is directly applicable to  other mobile or
transportable incinerator systems.
                                    497

-------
                                 REFERENCES
1. Yezzi, J.J., Jr. et  al.   Results  of the  Initial Trial  Burn  of the
   EPA-ORD Mobile  Incineration Systems.   In:  Proceedings  of the 1984
   National Waste  Processing Conference,  ASME, pp. 514-534.

2. Yezzi, J.J., Jr. et  al.   The  EPA  Mobile  Incineration Systems  In:
   Proceedings of  the 1982  National  Waste Processing  Conference, ASME,  pp.
   199-212.

3. Lovell, R.J., et al.  Trial  Burn Testing  of the  EPA-ORD Mobile
   Incineration System.   EPA-600/D-84-054,  Municipal  Environmental
   Research Laboratory,  Cincinnati,  Ohio, 1984.

4  Mortensen, H. et al.   Destruction of  Dioxin-Contaminated Solids  and
   Liquids by Mobile  Incineration.   EPA  Contract 68-03-3255,  Hazardous
   Waste  Engineering  Research Laboratory, Cincinnati,  Ohio, 1987.

5. King,  G., and Stumbar, J.  Demonstration Test Report for Rotary  Kiln
   Mobile Incinerator System at  the  James Denney Farm Site, McDowell,
   Missouri.   EPA  Hazardous Waste Engineering Research Laboratory,
   Cincinnati, Ohio,  1988.

6. Brunner, C.   Incineration Systems Selection and Design. Van  Nostrand
   Reinhold Company,  New York,  1984.

7. Ho,  M.,  and Ding,  M. G.  Field Testing and Computer Modeling of  an
   Oxygen Combustion  System at the EPA Mobile Incinerator. JAPCA,  Vol.
   38,  No.  9,  September 1988.

8. Gupta, G.D.,  et al.   Operating Experiences with EPA's  Mobile
    Incineration  System,   In:  Proceedings  of the  International  Symposium
   on Incineration of Hazardous, Municipal, and Other Wastes.  American
   Flame Research  Committee, Palm Springs,  CA, 1987.

9.  Freestone,  F.J.,  et al.  Evaluation of On-site Incineration for  Cleanup
   of Dioxin-contaminated Materials.  Nuclear and  Chemical Waste
   Management,  Vol 7, pp 3-20, 1987.

10. Gupta, G.D.,  et al.   MIS Modifications Trial  Burn and  Operations
    February 1986 to September 1987  Draft Report EPA Contract 69-03-3255,
    Risk Reduction  Engineering Laboratory, Edison,  New Jersey, 1988.

11.  Bryers,  R.W.   Deposit Analysis:  Cyclone Riser/Quench  Elbow - Denney
    Farm Site,  Foster Wheeler Development Corp., Livingston,  New Jersey,
    1988.
                                    498

-------
           LONG-TERM FIELD  DEMONSTRATION OF  THE  LINDE®OXYGEN
       COMBUSTION SYSTEM INSTALLED  ON  THE EPA MOBILE  INCINERATOR

     Min-Da Ho                            Robert H. Sawyer, James P. Stumbar
     Linde Division                      F.  W.  Enviresponse,  Inc.
     Union Carbide  Industrial Gases Inc.  Edison, NJ  08837
     Tarrytown,  New York 10591

                           Joyce M. Perdek
                           U.S. EPA, Risk Reduction Engineering Laboratory
                           Edison,  NJ  08837

                                   ABSTRACT

     This  paper  summarizes the various  system performance tests and  the
long term  operating experience of  the LINDE®Oxygen Combustion System (OCS)
installed  on the EPA Mobile Incineration  System (MIS).  The LINDE OCS was
installed  on the MIS as part of a major modification program in 1987.  The
modified system  was demonstrated for  three months in 1987 (June-September).
During this period  various system performance tests were conducted.  Later,
the system resumed  operation in February  1988 to continue the incineration
of dioxin-contaminated materials from sites  in  southwestern Missouri.  Since
the implementation  of the  modifications,  over seven million pounds of
dioxin-contaminated material including  soil, lagoon sludge, plastics, trash,
grass, protective clothing, wood, etc.  have been processed.

     The Mobile  Incinerator's capacity  for treating contaminated soil more
than doubled because of the LINDE OCS and modifications to the feed system.
Test burns of the unit show destruction and removal efficiencies (DREs)
surpassing both  RCRA and TSCA standards.  Recently, the Mobile Incinerator
has been used successfully to burn brominated sludge materials with high
heating value.  Water was used to absorb excess heat release.  The LINDE OCS
and other modifications have been instrumental for maximizing the throughput
of  this brominated  sludge material.

     The LINDE OCS has shown excellent dynamic response to varying oxygen
demand.  This ability has  significantly reduced the frequency of waste feed
shutdowns due to  low oxygen and high carbon monoxide conditions resulting
from variations  in the BTU content of the waste feed.  It was also shown
that nitrogen oxides emissions from the oxygen enriched operation compare
favorably with the previous air-based operation due to the unique design of
the LINDE burner.  In addition to the oxygen burner,  a microprocessor-based
control system has been used in the OCS.  This control system enhanced the
MIS performance in processing high BTU-content waste materials.   Some common
considerations about oxygen combustion and experiences learned during this
operation will also be discussed.
                                    499

-------
                                INTRODUCTION

     The EPA Mobile Incineration System (MIS) was designed and constructed
to demonstrate high-temperature incineration of hazardous wastes[l].  The
original system consisted of a refractory-lined rotary kiln, a secondary
combustion chamber (SCO), and an air pollution control system.  These three
components are mounted on three separate heavy-duty semi-trailers.
Monitoring equipment is carried by a fourth trailer.  During 1987, the MIS
was modified to increase its capacity and reliability.  The addition of the
LINDE Oxygen Combustion System (OCS) and modifications to the feed system
were needed to increase the capacity of the MIS.  The cyclone and wet
electrostatic precipitator (WEP) were installed to increase the reliability
of the MIS.                                                          ,

     This paper summarizes the various system performance tests and  the
long term operating experience of the LINDE Oxygen Combustion System (OCS)
with the MIS.  The LINDE System was installed on the MIS as part of a major
modification program in 1987.  The modified system was demonstrated for
three months in 1987 (June-September).  During this period various system
performance tests were conducted in addition to the normal incineration
operation.  The capacity of the MIS for treating contaminated soil more than
doubled because of the addition of the oxygen system and modifications to
the feed system.  Test burns of the unit showed DREs surpassing EPA
standards.  All incinerator effluents such as kiln ash and scrubber water
passed the stringent delisting criteria established by the EPA.

     The system resumed operation in February 1988 to continue the
incineration of dioxin-contaminated materials from sites in southwestern
Missouri.  Since the implementation of the 1987 modifications, over seven
million pounds of dioxin-contaminated material including soil, lagoon
sludge, plastics, trash, grass, protective clothing, wood, etc. have been
processed.

        TECHNICAL CONSIDERATIONS IN THE USE OF OXYGEN IN INCINERATION

Oxygen Enrichment

     The use of pure oxygen (O_) or of oxygen-enriched air reduces the
volume of both the oxidant and the flue gas.  As the oxygen replaces part or
all of the combustion air, the nitrogen portion (which is 79% by volume of
air) is reduced or eliminated which results in decreased volumes of the
oxidant and flue gases per unit of soil processed.  This has the direct
benefit of reducing particulate carryover from the kiln, reducing the
pressure differentials through the system, increasing the gas residence time
for destruction, and reducing fuel consumption.   While various levels of O^
enrichment were considered, it was decided that the highest level of oxygen
enrichment be used in the kiln to attain the lowest possible gas flow, which
would minimize particulate carryover and, as a result, cause less downtime
of the system.                                             ,

     The use of the pure oxygen system produced a 60% reduction in gas
velocity through the kiln which helped to reduce the particulate carryover
                                    500

-------
problems.  When the MIS processed 4000 Ib/hr of solids at 20 wt% water with
the pure oxygen system, the gas velocity at the kiln exit was approximately
3.3 feet per second (fps) compared with 8.0 fps when it processed only 2000
Ib/hr of dry soil using the air system.(2)

Oxygen Burner Criteria

     The major considerations in the selection of an oxygen burner were the
burner's flame temperature and mixing characteristics.  The use of oxygen
normally increases flame temperatures significantly.  Previous studies have
shown that even an additional 300-400°F in flame temperature would have an
unacceptable effect on the NO  levels and would also increase the kiln shell
temperature.  This would increase the potential for kiln distortion, and
could also increase the maintenance on the seals, refractory, etc.  It could
also increase the tendency for slagging in the kiln with materials that have
low ash fusion temperatures.

     To minimize local hot spots and ensure good combustion, good mixing of
fuel, oxygen, and recirculated combustion products is needed.  High momentum
is needed to produce good mixing.  This may be obtained through either high
mass or high velocity or a combination of both.  Due to the lack of nitrogen
mass, an oxygen burner must rely primarily on high velocity jets to achieve
good mixing.

LINDE Oxygen Combustion System

     The LINDE Oxygen Combustion System, with its patented "A" Burner, was
selected based on Union Carbide's data which had substantiated low NO
levels and flame temperatures lower than other oxygen burners.  This Burner,
which uses 100% oxygen, replaced one of the conventional air burners in the
rotary kiln.  The remaining air burner has been used as a backup.  A
schematic of the "A" Burner installation on the rotary kiln is shown in
Figure 1.  A unique feature of this burner is that it produces a flame
having a temperature comparable to the conventional air burner flame
temperature.  This is accomplished by aspirating the furnace gases into the
oxidant jets prior to mixing with the fuel.  Referring to Figure 2, air
atomized fuel oil (or fuel gas) is supplied at the burner axis as a
relatively low velocity jet.  Most of the oxygen is supplied as a ring of
high velocity jets surrounding the center fuel stream.  The purpose of the
oxygen annulus around the fuel stream is to enhance flame stability.  The
aspiration of furnace gases into the oxygen jets prior to mixing with the
fuel is indicated in Figure 2 by the small arrows.  By maintaining
sufficient distance between the oxygen jets and fuel supply, enough of the
furnace gases are aspirated into the oxygen jets prior to their mixing with
the fuel so that the resulting flame temperature is reduced to values
substantially below the theoretical flame temperature.  At the point of
combustion, both conventional air burners and the "A" Burner using 100%
oxygen have an oxidant with a low concentration of oxygen.  For the air
burner, the dilutent is nitrogen while for the "A" Burner using pure oxygen,
the furnace gases make up the dilutent.  These high velocity oxygen jets
also cause vigorous recirculation of gases within the furnace which produces
a more uniform temperature distribution within the furnace.
                                    501

-------
     Flame  shape measurements were made with a suction pyrometer to confirm
 the  low  flame  temperature.   In Figure  3 is the result obtained when straight
 oxygen nozzles were used on  a natural  gas "A" Burner.  The peak temperature
 is about 2800°F which is only 800°F higher than the average furnace
 temperature.
 [2-5].
More details about the "A" Burner can be found in references
      In  addition to  the  "A" Burner,  the LINDE Oxygen Combustion System
 included flow control piping  skids and a microprocessor-based control
 system.   This control system  was  designed with  safety as a prime
 consideration.   This system is  able  to control  transient emissions  ("puffs")
 better than the  previous MIS  combustion control system because of its fast
 reaction time.

 NO Emissions
 —x	
      In  the field demonstration of the MIS  at the Denney Farm, the  NO^ level
 was continuously monitored and  recorded.  Due to the air infiltration, the
 effective O  enrichment  levels  were  40 to 50% O in the rotary kiln.  The
 mean  NO   emission levels reported in the 1987 trial burns[6] were between
 54.6  toX138.3 ppm at 15% C02  (or  0.07 to 0.18 Ib NOx/MMBTU)  (lb/MMBTU=pound
 per million BTU). This  compares  favorably  with the previous air system
 levels of 126 to 166 PPM at 11% CO2  (or 0.19 to 0.235 Ib NOx/MMBTU) which
 were  obtained in the 1985  trial burns.[1]   It should be noted, however, that
 the NO   emission level from the LINDE system was more sensitive to  the
 operating conditions than  the air system.   It is anticipated that additional
 NO reduction can be achieved by  (1) reducing air leakage into the  kiln, (2)
 using steam instead  of air as an  oil atomization flui'd for the oxygen burner
 and (3)  reducing the excess oxygen level in the rotary kiln.

                   SYSTEM  PERMITTING AT INCREASED CAPACITY

      Due to the  increase in solids feed capacity to 4000 Ib/hr and  the
•installation of  the  cyclone and WEP, operating  permit modifications for both
 the RCRA and air permits were filed  jointly with the Missouri Department of
 Natural  Resources (MDNR) and  EPA  Region VII.[7]  Since MDNR had not
 received final authorization  from EPA to regulate dioxins, EPA and  MDNR
 jointly  issued the final RCRA permit[8], and MDNR issued the revised air
 permit.[9]

      The modified permits  required compliance testing of the MIS for
 particulate emissions at the  new  feed rates.  Verification tests to show
 that  the byproduct residues passed the delisting criteria also were
 required.  However,  further Destruction and Removal Efficiency (ORE) testing
 was not  mandated by  the  regulatory agencies because the critical operating
 parameters remained  unchanged.

      All requirements for  compliance testing were met.  The  stack
 particulates during  these  tests were well below the regulatory limit of 0.08
 gr/dscf.  All effluent streams  (ash  solids  and  scrubber liquid) also met the
 delisting requirements,  and hence could be  discharged as non-hazardous
 wastes.
                                     502

-------
                         CONTAMINATED SOIL OPERATION

     The maximum contaminated soil throughput of the Mobile Incineration
System during its operation with air burners had been 2000 Ib/hr.  However,
this maximum rate was not sustainable.  For example, the "average throughput
rate of four test runs in the Spring of 1985 had been only 1478 Ib/hr of dry
soil (<10% moisture).  With the addition of LINDE Oxygen Combustion System
and other modifications, the MIS achieved a sustainable soil throughput rate
of 4000 Ib/hr even with moisture content up to 20%, as confirmed by
certified verification tests[11].

     As detailed in a complete report on subsequent TSCA/RCRA trial burn
tests conducted in the summer of 1987[6], the MIS was able to achieve the
required destruction and removal efficiencies (DREs) for all the principal
organic hazardous constituents (POHCs).  DRE for PCBs exceeded 99.9999%.
The DRE for carbon tetrachloride (CCl ), hexachloroethane (C Cl ), and
trichlorobenzene (TCB) exceeded 99.99%.

     With normal and relatively dry (about 20% moisture) soil, it was quite
easy to feed up to 4,000 Ib/hr.  Although the feed system had trouble
processing muddy soil, the kiln and SCC operated well with it.  Due to the
addition of a cyclone downstream of the kiln and the significant reduction
in the kiln combustion gas velocity, a dust carryover problem in the SCC,
experienced before the modifications!7], was greatly reduced.  The
particulate carryover was less than 15% of the feed rate for most feeds and
was easily removed by the cyclone.

     It should be noted that the current capacity limitation of the MIS is
about 4000 Ib/hr of soil, mainly due to the mechanical limitations of the
rotary kiln (which could be upgraded).  The capacity limitation is not due
to heat transfer or flue gas volume limitations.  Therefore, oxygen
enrichment of the SCC to obtain additional throughput is not currently
warranted.  The kiln limitation can be solved by using one or more larger
rotary kilns to feed common downstream equipment.  It is estimated that up
to 10 tons per hour (TPH) of soil could be handled by such an upgraded
system fired with LINDE oxygen burners in both the kiln and SCC.  Such a
throughput increase would make mobile incineration technology much more
economically attractive.

                         BROMINATED SLUDGE OPERATION

     In 1988 and 1989, the MIS processed over 2,150,000 Ibs. of
dioxin-contaminated brominated sludge from a site in southwest Missouri.
The sludge had characteristics which caused processing difficulties in the
MIS.  The sludge was a very heterogeneous material.  Analyses showed that
measured bromine content varied between 1 and 12%; moisture varied between
35% and 42%; and heating value ranged from 3200 to 4600 BTU/lb.  The sludge
required considerable testing before it could be processed as described
below.

     Previous testing!1] in 1985 had shown that this material was difficult
to process in the MIS partially due to the high moisture content and the
                                   503

-------
non-uniform heat content.  The two runs yielded throughput of less than 1000
Ib/hr.  The throughput was mainly restricted by the required gas residence
time in the SCC.  Also, 1000 ppm of organic bromide remained in the ash and
300 ppm was detected in the scrubber water.

     During the summer of 1937, a number of feasibility tests for
incinerating this sludge were run.  The program examined the effect of
blending contaminated soil, sludge and sodium sulfate (Na SO ) in different
ratios!. 11].  The three tests produced mixed results.  Mecnanzcal performance
of the unit was generally good, but chemical performance, ascertained by
analyzing the residues, ranged from poor to good.  Samples of purge water,
kiln ash, cyclone ash and separator sludges were analyzed for selected
constituents.  The analyses showed that the byproduct streams slightly
exceeded the stringent delisting criteria proposed for these residues
(Federal Register, Vol. 52, No. 171, Thursday, September 3, 1987).
Therefore, further testing was required to establish proper processing
conditions.

     Since it appeared that the unsatisfactory effluent quality was due
mainly to insufficient heating of the kiln ash, the residence time for
processing the solids was increased by decreasing the kiln rotation speed
and by adding tests using pure sludge.  The Na SO  was eliminated to permit
higher processing temperatures.  The new tests showed that straight sludge
could be processed successfully and provided the operating basis for the
commercial run which occurred between September 1988 and January 1989.

     At the beginning of the commercial run, a series of delisting tests
were conducted to demonstrate that the effluents of the incinerator were
delistable.  It was determined that oxygen enrichment, with the LINDE OCS,
should be continued for increased throughput and better controllability.
Since the sludge had substantial heating value, water was used to absorb the
excess heat.  Such a practice is more advantageous than using air injection
to absorb excess heat.  This is because water vapor contains approximately
twice the enthalpy per unit volume as nitrogen at 1800°F, so that for a
given heat load, the flue gas volume is smaller when using water spray than
when using excess air.

     Sludge throughput rates between 1600 to 1800 Ib/hr were achieved during
the delisting tests.  The kiln ash and other effluents met the delisting
criteria.

     The oxygen injection rate responded dynamically to control the kiln
oxygen level.  Instantaneous flow rates in the delisting test runs ranged
between 3000 and 12000 standard cubic feet per hour (scfh).  The average
oxygen flow rate was about 5000 scfh.  All of the excess pure oxygen was
also injected through the LINDE burner, and a small amount of diesel fuel
was also fired through the burner to stabilize the combustion.  An estimated
25000 scfh of air enters the kiln mainly due to infiltration.
                                   504

-------
          OTHER OPERATING EXPERIENCES AND BENEFITS OF THE LINDE PCS

Fuel and Cost Savings

     When treating contaminated soil, supplemental fuel is reguired to
provide the heat required to operate the rotary kiln at 1500-1600 F and the
SCC at about 2100 F since the contaminated soil does not have a sufficient
heating value to sustain self-combustion.  Even with high BTU waste
material, auxiliary fuel is still reguired in the SCC.

     Specific fuel savings of over 60% (7.2 MMBTU/ton soil) were achieved
during operation of the EPA/MIS with the LINDE system processing
contaminated soil.  This result can also be expressed as 50 MMBTU saved per
ton of oxygen used.[2]

     The economics of using oxygen to save fuel, of course, depend on the
relative cost of fuel and oxygen.  With No. 2 fuel oil costing $0.70 per
gallon (or $5.50 per million BTU) and a fuel savings of about 50 million BTU
for every ton of oxygen consumed, the breakeven oxygen cost is about $275
per ton of oxygen.  The cost of oxygen depends on methods of oxygen
generation, size of plant, and location.  For example, it ranges from about
$50 per ton of oxygen produced by a large on-site facility to about $120 per
ton for delivered liquid.

     The principal economic benefit from oxygen combustion is derived from
the very significant throughput improvement.  The large fixed portion of
daily incinerator operating costs is spread over a much larger quantity of
waste processed.  For example, for mobile/transportable incinerators a
doubled throughput can reduce the allocated incineration cost of
contaminated soil by typically $100 to $500 per ton of waste, while the cost
of oxygen reguired is typically between $10 to $50 per ton of waste
incinerated.

Flame Stability and Operational Flexibility

     During operation of the EPA/MIS, good flame stability and operational
flexibility were achieved with the LINDE OCS.  A number of system stability
tests were conducted with various feed materials at different feed rates.
By cycling liquid waste feed and water spray on and off, disturbances were
generated to test the dynamic response of the burner system.  Satisfactory
system response and flame stability were demonstrated in all the tests
conducted.[2]  It has also been demonstrated that very light vermiculite can
be processed through the system with modest particulate carryover (about
17%) from the kiln.  During the entire operation period, no significant MIS
downtime was attributed to the LINDE System.

Transient Emissions Control

     When high-BTU wastes are fed into rotary kiln incinerators in an
intermittent mode (typical of ram feed systems), the transient combustion
behavior of these materials creates unsteady releases of combustible gases
which may momentarily deplete the oxygen content of kiln gases.  These
                                    505

-------
temporary oxygen-deficient conditions could cause the release of products of
incomplete combustion (PICs).  These phenomena have raised public concern
recently and have been the subject of research projects sponsored by the
EPA.[12-14]

     In the field operation of the EPA/MIS, large quantities of high BTU
materials were burned periodically.  These materials were rain-fed into the
rotary kiln every 30 to 45 seconds.  To respond to the transient oxygen
demand resulting from the burning of these materials, a unique oxygen
feedforward-feedback control logic was designed into the LINDE System.  ,

     With the normal functioning of the oxygen control feature, the
transient upset conditions associated with the release of the combustible
gases were virtually eliminated in the operation of the MIS.  As, an example,
data taken during a delisting test are shown in Figure 4.  The oxygen level
of the gas entering the SCC (kiln O %) was controlled to be above 5% (wet),
while the O_ level at the SCC is maintained at about 7.5% (dry) except for a
short perioa when the kiln oxygen analyzer was plugged, as indicated in
Figure 4.  Note that due to safety consideration, the oxygen control system
does not allow the stoichiometric ratio at the burner to drop below a preset
minimum, even when the oxygen level may be higher than set point due to,
e.g., air leakage.  Carbon monoxide levels were consistently below 50 ppm.
Note also in Figure 4, the oxygen flow rate responded promptly to the
transient oxygen demand.  This can be attributed to the fast response of the
LINDE Oxygen System and the Thermox®WDG in-situ 0  analyzer.  In addition,
the high-momentum oxygen jets in the LINDE burner enhanced mixing in the
kiln to reduce any pockets of unburned combustibles.
     The transient puffs of the incinerator are minimized with this control
system for two reasons.  First, the oxygen feed dynamically follows the
trend of the demand, so that the oxygen level is sufficient at all times.
Second, a consistent oxygen level helps to maintain a stable and smooth
combustion, so that the flame fluctuation and rapid expansions due to
unstable combustion are minimized.  Consequently the release rate of
combustible gases is more consistent.  It is also important to maintain a
high turbulence in the combustion chamber and to maintain the entire chamber
above the auto-ignition temperature of the combustible gases.

Corrosion Resistance

     The presence of hydrogen chlorides and hydrogen bromides in the
combustion gases had caused some concern about potential corrosion problems
of the burner equipment.  As a precaution, the front end of the water-cooled
burner was coated with a layer of protective ceramic material (LC-5) by
Linde Coatings Service Co.  Hastelloy C-22 material was used for the oxygen
nozzles and for the patented oil atomizer.  Experience in the MIS was that
no noticeable corrosion of the burner occurred during more than a year of
operation.  The atomization nozzle showed signs of minor erosion and
significant coke buildup after more than one year of operation.  The coke
buildup was readily removed', and the nozzle then became fully functional
again.  In comparison, the separate waste oil nozzle made of stainless steel
                                   506

-------
had severe metal loss and had to be replaced within three months of
operation.

Slagging Problems

     Due to the varying characteristics of waste feeds in hazardous waste
incinerators, the control of temperature distribution is both challenging
and important.  Local hot spots can cause refractory damage, and more
frequently, clinker formation or slag buildup (slagging), which would
eventually require shutdown of the furnace for slag removal.

     Since the use of oxygen-enrichment was historically associated with
very high flame temperature, there was considerable concern that the use of
the LINDE OCS in a rotary kiln incinerator would create serious slagging
problems.  During the field operation at Denney Farm, there was no evidence
that the slagging tendency of the rotary kiln was aggravated by the use of
the LINDE OCS.  Experience has shown that when the system is operated with a
good understanding of the process and waste feed characteristics, the
occurrence of slagging is manageable.

     During the incineration of contaminated soil, slagging was normally not
a problem.  The heat release of the "A" burner was well controlled, and the
thermal mass of the soil absorbed the heat and dampened temperature
fluctuations.  However, the combustion of plastic materials tended to be
more challenging. [ 15]                                             .,

The principal causes of slagging formations noted were:

     (1)  Temperature excursions due to fluctuations in the feed rate of
          high BTU waste would sometimes cause rapid slagging because the
          kiln temperature rose above the fusion temperature of the ash.

     (2)  Occasional false readings of the control thermocouple due to
          abnormal air leakage through the ash gates caused slagging.  In
          this situation, the kiln was running at a temperature at least
          150 F higher than the temperature reading of the control
          thermocouple.

     (3)  The presence of a small fraction of low-melting point ash in the
          feed materials may cause the formation of eutectic solid solutions
          with low melting point.  The presence of glass or sodium sulfate,
          for example, is conducive to slag formation.

     (4)  The flame characteristics of the burner could also have a
          significant impact on the slagging problem.  It was found earlier
          that the mid-section of the kiln was somewhat hotter than the
          other areas when there was no heat load in the kiln.  This was
          possibly due to the flame shape created by the divergent angles of
          the original nozzles.  A set of nozzles generating higher momentum
          and a narrower flame, which were installed later, proved to be
          effective in producing a more uniform temperature distribution.
                                   507

-------
     Occasionally, there was a gradual slag buildup on the burner wall which
would distort the flame pattern.  In the future, retractable burner mounts
should be considered to allow easy slag removal from the burner area.   ;

                                 CONCLUSION

     The long-term reliable operation of the LINDE Oxygen Combustion System
has demonstrated the significant advantages of this technology on the MIS.
Much valuable experience has also been gained in the field application of
this technology to handle a large variety of waste streams.  It is believed
that similar advantages can be realized in the application of this
technology to other thermal treatment systems.

                               ACKNOWLEDGEMENT

     The authors would like to thank Mr. Frank J. Freestone of U.S. EPA and
Dr. G.D. Gupta of Foster Wheeler Enviresponse, Inc. for their support and
guidance during this project.
                                   503

-------
                                 REFERENCES

(1]   Mortensen,  H.,  et al,  "Destruction of Dloxin-Contaminated Solids and
     Liquids by Mobile Incineration," USEPA report,  EPA Contract
     #68-03-3255, Hazardous Waste Engineering Research Laboratory,
     Cincinnati, Ohio, April, 1987.

[2]   Ho, Min-Da and Ding, M. G., "Field Testing and Computer Modeling of an
     Oxygen Combustion System at the EPA Mobile Incinerator," JAPCA, Vol.
     38, No. 9,  September 1988.

[3]   Anderson, J. E., U. S. Patent Nos. 4,378,205 and 4,541,796, "Oxygen
     Aspirator Burner and Process for Firing a Furnace," March 29,  1983,
     September 17, 1985.

[4]   Anderson, J.E., "A Low NO , Low Temperature Oxygen-Fuel Burner,"
     Proceedings of the American Society of Metals, 1986 Symposium on
     Industrial Combustion Technologies, Chicago, Illinois, April 29, 1986.

[5]   Ho, Min~Da and Ding, M. G., "Proposed Innovative Oxygen Combustion
     System for the Incineration of Hazardous Waste," Hazardous Materials
     Management Conference & Exhibition/West, December 3-5, 1986, Long
     Beach, CA.

[6]   King, G. and J. Stumbar, "Demonstration Test Report for Rotary Kiln
     Mobile Incinerator System at the James Denney Farm Site, McDowell,
     Missouri," Enviresponse, Inc., Edison, New Jersey, January 1988.

[7]   Sherman, A., et al., "Final Permit Application USEPA Mobile
     Incineration System at the James Denney Farm Site, McDowell, Missouri,"
     USEPA, Hazardous Waste Engineering Research Laboratory, Releases
     Control Branch, Edison, New Jersey, April 3, 1987.

[8]   Brurmer, F.A., and Wagoner, D.A., "Hazardous Waste Facility Permit"
     State of Missouri, Department of Natural Resources, Jefferson City, MO,
     May 6, 1987.

[9]  Letter from F. A. Brunner and M. Nikkila to F. J. Freestone, State of
     Missouri, Department of Natural Resources, Jefferson City, MO, April
     12, 1987.

[10] Gupta, G.D., et  al, "Operating Experiences with EPA's Mobile
     Incineration System,"  Int'l Symposium on Hazardous and Municipal Waste
     Incineration, AFRC, Nov.  2-4, Palm Springs, CA.

[11] G. D. Gupta, et  al, "MIS  Modifications Trial Burns and Operations Feb.
     1986  to  Sept.  1987," EPA  Contract 68-03-3255, US EPA HWERL, Cincinnati,
     Ohio  45268.

[12] Linak, W.  P.,  J. D. Kilgroe,  J. A. McSorley, J.O.L. Wendt  and  J. E.
     Durn,  "On  the  Occurrence  of Transient Puffs in a Rotary Kiln
     Incinerator  Simulator,  Part I," JAPCA, VOL. 37, No. 1, Jan. 1987.
                                    509

-------
13]  Linak, W. P., J. D. Kilgroe, J. A. McSorley, J.O.L. Wendt and J.  E.
     Durn, "On the Occurrence of Transient Puffs in a Rotary Kiln
     Incinerator Simulator, Part II," JAPCA, VOL. 37, No. 8, Aug. 1987.

[14] Linak, W. P., et_al, "Rotary Kiln Incineration:  The Effect of Oxygen
     Enrichment on Formation of Transient Puffs During Batch Introduction of
     Hazardous Wastes," Proceedings of the 14th Annual Research Symposium on
     Land Disposal, Remedial Action, Incineration and Treatment Hazardous
     Waste, USEPA, Cincinnati, Ohio, May, 1988.

[15] Stumbar,  J.P., Sawyer, R., Gupta, G.D., Perdek, J. and Freestone, F.,
     "Effect of Feed Characteristics on the Performance of EPA's Mobile
     Incineration System," 15th Annual EPA Research Symposium,  USEPA,
     Cincinnati, Ohio, April, 1989.
                                  510

-------
  _l
  5

  cc
u.z
00

°Z
E2
   
   flC
   111
   z
   cc

   CD
                                511

-------
                        Figure 2
         SCHEMATIC SKETCH OF THE A - BURNER
                                FUEL TUBE
                                   ANNULUS OXYGEN
                                       OXYGEN
                                       NOZZLES
        OXYGEN
           JET
OXYGEN TO
  NOZZLES
                              OXYGEN
                              ENVELOPE
  FUEL
                            ^iW//* FURNACE
                            sife GASES
                                r^l
OXYGEN TO
ANNULUS
                        512

-------
    s  .
U_
o
 »\

 £



I
    CM


    O
    CM
I

8
CO
    10


    CO
                                    CO

                                    *

                                    N
                                Q.L.
                                               O
                                               N.
i  if;
LL  2 U. 3
                     O>

                     I

                     8
                     CO
                     c
                         513

-------
            20OO

  KILN TEMP. 1500
             5OO
  KILN O2%
20
15
10

 5
    KILN
 O2BURNER
FLOW(SCFM)  200
             1OO
 SCC TEMP
   SCC 02%
            2500
            15OO
              20
              1O
               0

             1OO

SCC CO (PPM)  50
    WEP
 GAS FLOW  5OOO
   (ACFM)
                   14
                                         Figure 4
                             MOBILE INCINERATOR OPERATION
                               DELISTING TEST No.4, 9/12/88
                                     XV	»v.
                                                   KILN
                                                O2 ANALYZER
                                                 PLUGGED
                                CALIBRATION
                                              RESULTED FROM
                                              KILN O2 ANALYZER
                                                 PLUGGAGE
            15
16
17
18
19
20
21
                                    HOURS OF THE DAY
                                    514

-------
   IN PLACE TREATMENT OF CONTAMINATED SOIL AT SUPERFUND SITES: A REVIEW

         M. Roulier, J. Ryan, J. Houthoofd, H. Pahren, and F. Custer
                    Risk Reduction Engineering Laboratory
                     U.S.  Environmental  Protection Agency
                           Cincinnati, Ohio  45268
                                   ABSTRACT

      The USEPA's Superfund research program is developing methods for in
place (in situ) removal of contaminants from soils and for in place treatment
of contaminated soils.  This work is motivated by the high cost for managing
large volumes of soil with low levels of contamination and because of the need
to comply with provisions of the Superfund Amendments and Reauthorization Act
(SARA) and the Resource Conservation and Recovery Act (RCRA).  This paper
summarizes available information sources and improvements in technology since
1984 when USEPA issued a summary report on in place soil treatment.  There
have been only a few instances of in place treatment based on aqueous solution
chemistry and these have involved primarily organic contaminants.
Biodegradation has been successful for some organic compounds and
stabilization/solidification is increasingly successful  for inorganics and
some organics.  In place vapor-phase removal processes such as vacuum
extraction, steam stripping, and microwave heating appear most promising for
low solubility, low boiling point organic compounds.  Improvements are needed
in methods for delivering and mixing treatment materials in soil and in
methods for recovering unreacted materials and reaction products.

                                 INTRODUCTION

      Work on in situ  (in place) treatment of contaminated soils at Superfund
sites is motivated by the need to comply with current regulations and by the
high costs of treating large volumes of contaminated soil often encountered at
sites.  The provisions of the Comprehensive Environmental Response
Compensation and Liability Act (CERCLA) as amended by SARA require the maximum
possible treatment of wastes and contaminated soils at Superfund sites and
restrict on-site containment and off-site disposal.  The provisions of RCRA as
amended by the Hazardous and Solid Waste Amendments (HSWA) establish treatment
standards and disposal limitations that further restrict the disposition and
encourage treatment of contaminated materials removed from Superfund sites.
      A number of accepted technologies (e.g. incineration,
stabilization/solidification, chemical treatment) are available or are being
tested for the broad range of Superfund wastes and soils (1,2,3,4).  These
                                     515 '

-------
technologies  are  cost-effective when  applied to concentrated wastes and highly
contaminated  soils  that  have  been  excavated.  They have not been shown to be
cost effective  for  the large  amounts  of  slightly contaminated soil that are
often encountered after  removal of wastes or near-surface soils.  Volumes of
soil remaining  are  usually  large;  the cost of moving such soils for treatment
1s excessive.   The  alternatives, for  reducing these costs, are to remove
contaminants  without moving the  soils (in situ removal) or to treat the
contaminants  in place (in situ treatment).
      The USEPA's first  report (5) in 1984 on in place treatment described a
large number  of chemical and  physical processes (e.g. oxidation, reduction,
precipitation)  that could potentially be used in situ to immobilize or
detoxify contaminants in soils.  The  authors noted that the majority of these
were conceptual or  had been tested only  in the laboratory.  Another USEPA
report  (6) published two years later  evaluated in situ (in place) methods for
stabilizing (treating) waste  deposits.   This report considered treatment
materials (reactants), methods for delivering the treatment material, and
methods for recovering the  products of the reaction.  The report noted that ".
. . the combination of injection,  reaction and recovery as a system for in
situ remediation  has scarcely been practiced and is in its infancy as an
integrated technology."
      Most of these proposed  treatments  involved aqueous solution chemistry
and it was hoped  that workable in  place  treatment technologies would be
developed along these lines.  Instead, the major in place developments since
that time have  been in biodegradation, stabilization/solidification, and
removal of contaminants.  This trend  is  likely to continue with contaminants
being increasingly  recovered  for treatment above ground rather than being
treated and left  in place.

                                 RECENT  WORK

BIOLOGICAL AND  PHYSICAL  TREATMENT

      Biodegradation (1,2,3,7,8,9,10,11,12,13,14) and stabilization/
solidification  (1,3,4,10,15,16,17,18,19,20,21) are rapidly developing
technologies that are distinguished, within the Superfund research program,  as
work areas separate from in place treatment; only a few of the many recent
references are  listed.   These technologies are being applied both in place and
in above-ground reactors and  batch plants.  Biodegradation is effective only
for organic contaminants; stabilization/solidification (S/S) processes are
most effective  for  inorganic materials but are being developed to stabilize
wastes containing organic contaminants.   Experience with biodegradation and
S/S suggests that these  technologies will achieve expanded coverage of waste
types and improved  performance.
      Both these technologies are limited by the problem of delivering
materials to subsurface  soils and achieving uniform mixing.   Two commercial  in
place mixing technologies are available  (15,18,22,23);  the cost and
performance of these are being evaluated (3,4) in the Superfund Innovative
Technology Evaluation (SITE) program.   In place mixing technologies will  also
have application for in  place vapor-phase removal  of organic compounds.
                                     516

-------
CHEMICAL TREATMENT

      There have been only a few examples of in place treatment of
contaminated soil by chemical processes.  Successful projects include
treatment of polymer waste in an impoundment (24), a spill of acrylonitrile
(25), various pesticides  (26,27,28) and arsenic compounds in groundwater  (29).
The limited amount of in  place treatment appears to be due to difficulty  in
applying treatment materials uniformly and recovering unreacted materials and
reaction products.  There is also a reluctance to apply treatment materials
that are dangerous substances (e.g. hydroxides and hypochlorites).  The
contaminants treated have generally been low-solubility organic chemicals.
Organic chemicals with high aqueous solubility or high vapor pressure are
transported out of soil by natural processes or are easily removed in water or
in the vapor phase for treatment above ground.  Most inorganic contaminants
are not treatable (cannot be truly degraded).  Stabilization/solidification
has been used for most of these but there is controversy whether S/S processes
are chemical treatments that form new low-solubility compounds or merely
occlude the contaminants  through physical processes and retard their release.

DELIVERY AND RECOVERY

      The limited success of in place treatment has led to an increase in work
on processes for adding treatment materials and distributing them in soil, and
on processes for removing contaminants for treatment above ground without
disturbing the soil (in place removal).  The Superfund research program
describes these collectively as innovative delivery and recovery processes.  A
recently completed review (30) identified 17 processes, proposed or being used
in other industries, that could be used for delivery or recovery in
remediation of Superfund  sites.   The authors noted that several  of the
technologies were commercially available in industries such as petroleum
production but none were  fully proven for use in waste site remediation.
Vapor (vacuum) extraction is being used to decontaminate soils affected by
leaking underground storage tanks and has successfully removed a variety of
volatile organic compounds from soil.

LIQUID PHASE REMOVAL

Withdrawal of contaminated groundwater for treatment (pump and treat) was the
first in place removal process used at Superfund sites.  The Superfund
research program is attempting to improve the pump and treat process through
the use of intermittent (pulse)  pumping for minimizing the amount of water
removed per unit of contaminant and for avoiding hydrogeologic "dead zones"
observed in many pump and treat operations.   The engineering component of this
research is being conducted by the USEPA's Risk Reduction Engineering
Laboratory (RREL) in Cincinnati, OH (31) and the evaluation/interpretation
component by the USEPA's  Robert S.  Kerr Environmental  Research Laboratory in
Ada, OK (32).   These efforts are likely to improve removal of some inorganics
and soluble organic compounds but not most inorganic cations or low-solubility
organic compounds that are highly partitioned onto the solid phase in soil-
water systems.
                                     517

-------
      The  aqueous  liquid  phase  with  chemicals  added  has  also  been  used  as a
displacing solution  to  extract  contaminants  from  soils;  this,  is termed  soil
flushing.   A  pilot test for removal  of  organics (33) was less successful than
full  scale operations for removal  of cadmium from contaminated soils  (34) and
mixed metals  and volatile fatty acids from a sludge  lagoon  (35).   Except in
confined or well-controlled hydrogeologic settings,  soil  flushing  will  be
limited for the same reasons as chemical treatment - difficulty in applying
and recovering materials  and reluctance to bear responsibility for adding
dangerous  chemicals  to  soil.  Several commercial  processes  are available and
are being  tested under  the SITE program (3,4).

LIQUID CONTROL

      Two  control  procedures are being  tested  for use with  liquid  phase
removal.   The movement  of cations  and soil water  in  response  to an applied
direct current gradient,  called electro-kinetics  or  electro-osmosis,  has been
adapted from  applications in the mining industry  and is  being considered as a
means for  directing  the movement of  water and  contaminants  during
removal/treatment  in low-permeability soils  (2,36).  The RREL sponsored one
field test (37) and  the SITE Emerging Technologies Program  is testing the
process when  used  in conjunction with an acoustic field  (4).   In place
freezing techniques  used  in the construction industry are being considered for
temporarily making soil impermeable  or  for concentrating  contaminants in front
of a  slowly moving freezing front  (38,39).   Both  of these control  procedures
are in the early stages of testing to determine their usefulness at Superfund
sites.

GAS PHASE  REMOVAL

      For  many low-solubility organic compounds,  equilibrium  thermodynamics
indicates  that the aqueous  liquid phase will contain greater  amounts of
organic contaminants than  an equal volume of the  gas phase.   However, the
rates of transfer  from  the  solid phase  to the  liquid phase  are often slow
relative to rates  of transfer into the  gas phase.   Additionally,  it is  easier,
because of relative  densities and viscosities, to move large  volumes of gas
through the subsurface  than  it  is to  move equal volumes of  liquid.  This
combination of greater  convective flow  and more rapid phase transfer makes
gas-phase  removal  an efficient  process  for many organic compounds,
particularly  in vadose  (partially saturated) zones where  it is difficult to
remove water.
      Gas  phase removal,  called  "in  situ vacuum extraction, air stripping,
volatilization, vapor extraction", or "forced air venting," is a commercially
available  process  that  has  been  used most extensively in decontaminating soils
affected by leaking  underground  storage tanks  (21,40,41,42,43,44); these often
contain fuels and  other volatile substances.  A combination of air injection
and vacuum  venting wells  are used to  induce air flow through the contaminated
soil.   Performance data for this technology have been summarized recently (45)
and detailed cost  analyses were  conducted for several typical  systems (46).
Several  gas phase  removal  processes are being tested under the SITE program
(3,4,22,23)  Although gas phase  removal  is used almost exclusively in vadose
zone soils, there  is at least one site  in Europe  (10) and one in  the U.S.  (47)
                                     518

-------
where it has been used to simultaneously remove organics from groundwater and
soil; air was injected in the saturated zone and the organics stripped from
the water were captured by the vacuum system in the vadose zone.

SOIL HEATING

      Simple gas phase removal is most effective for organic compounds with
relatively low boiling points.  Several methods are being tested for heating
contaminated soil to increase vapor pressure and allow gas phase removal of
higher boiling point compounds.  Radio frequency (microwave) heating has been
used successfully in a U.S. Air Force pilot-scale field test for removal of
fuel oil (48,49); a full scale test is being planned.  The RREL is sponsoring
a laboratory test of radio frequency heating for removing the higher boiling
compounds creosote and pentachlorophenol from soils contaminated with wood
treating fluids.  Steam injection, developed for petroleum recovery (30), is
being adapted for use with vacuum extraction at hazardous waste sites (50).
It has been tested at a small scale in the field for solvent removal (51) and
will be examined at a larger scale in a RREL field study that is being
initiated.  A combination of steam and hot air are used to heat soil for
several of the gas phase removal technologies being tested under the SITE
program (3,4,22,23,).  Another soil heating technology melts (vitrifies) the
soil (52,53) to destroy most of the organic compounds and capture the
remainder with a vacuum hood on the surface.  This technology had been
proposed for use at one Superfund site and is being tested under the SITE
program (3,4) for treatment efficiency and for ability to control the loss of
volatiles to surrounding soils.

                                   SUMMARY

      Since EPA's first report on in place treatment, the technologies for in
place removal, particularly in the gas phase, have developed much more rapidly
than those for in place chemical treatment.  Contaminants will be increasingly
recovered for treatment above ground rather than being left in place after
treatment.  Biodegradation and stabilization/solidification can successfully
treat some organic and inorganic contaminants in place and these two
technologies will continue to improve their ability in this regard.  Further
work on development of methods for delivering and mixing treatment materials
in soil and for recovering unreacted materials and reaction products would
expedite improvements in all in place treatment and removal processes.

                                  REFERENCES

1.    U.S. Environmental Protection Agency.  1985.  Handbook:  Remedial  Action
      at Waste Disposal Sites (Revised)  EPA/625/6-85/006.  USEPA Hazardous
      Waste Engineering Research Laboratory, Cincinnati, OH and USEPA Office
      of Emergency and Remedial Response, Washington, DC.

2.    Sanning, D. E., and R. F. Lewis.  January 1987.  Technologies for In
      Situ Treatment of Hazardous Wastes.  USEPA  Hazardous Waste Engineering
      Research Laboratory, Cincinnati OH.  EPA/600/D-87/014, NTIS-PB87-146007.
      24p.
                                     519

-------
3.    U.S Environmental Protection Agency.  1988.  Technology Screening Guide
      for Treatment of CERCLA Soils and Sludges, EPA/540/2-88/004.  USEPA
      Office of Emergency and Remedial Response, Washington. D.C.  127p.

4.    U.S. Environmental Protection Agency.  1988.  The Superfund  Innovative
      Technology Evaluation Program: Technology Profiles EPA/540/5-88/003.
      USEPA Risk Reduction Engineering Laboratory, Cincinnati, OH.  83p.

5.    U.S. Environmental Protection Agency.  1984.  Review of In place
      Treatment Techniques for Contaminated Surface Soils (Volume  1:
      Technical Evaluation)  EPA/540/2-84/003a.  USEPA Municipal Environmental
      Research Laboratory, Cincinnati, OH and USEPA Office of Solid Waste and
      Emergency Response, Washington, D.C.   163p.

6.    U.S. Environmental Protection Agency.  1986.  Systems to Accelerate In
      Situ Stabilization Of Waste Deposits, EPA/540/2-86/002.  USEPA Hazardous
      Waste Engineering Research Laboratory, Cincinnati, OH.  286p.

7.    Lee, M. D., J. M. Thomas, R. C. Borden, P. B. Bedient, J. T. Wilson, and
      C. H. Ward.  1988.  Biorestoration of Aquifers Contaminated  with Organic
      Compounds.  CRC Critical Reviews in Environmental Control, Vol. 18,
      Issue 1, pp. 29-89.

8.    Soczo, E. R., and J. J .M. Staps.  1988.  Review of Biological Soil
      Treatment Techniques In The Netherlands.  IN: . Contaminated  Soil '88
      Volume 1, Proceedings of the Second International TNO/BMFT Conference on
      Contaminated Soil, 11-15 April 1988, Hamburg, Federal Republic of
      Germany, K. Wolf, J. van den Brink, and F. J. Colon, eds., Kluwer
      Academic Publishers, The Netherlands, pp. 663-670.

9.    Mischgofsky, F. H. and R. Kabos.  1988  General Survey of Site Clearance
      Techniques: Trend Towards In Situ Treatment.  IN:  Contaminated Soil '88
      Volume 1, Proceedings of the Second International TNO/BMFT Conference on
      Contaminated Soil, April 11-15, 1988, Hamburg, Federal Republic of
      Germany, K. Wolf, J. van den Brink, and F. J. Colon, eds., Kluwer
      Academic Publishers, The Netherlands, pp. 523-533.

10.   U.S. Environmental Protection Agency.  1988.  Assessment of
      International Technologies for Superfund Applications:  Technology
      Review and Trip Report Results, EPA/540/2-88/003.  USEPA Office of Solid
      Waste and Emergency Response, Washington, D.C.  50p.

11.   Wetzel, R. S., D. J. Sarno, C. M. Durst, B. C. Vickers, P. A. Spooner,
      J. R. Payne, W. D. Ellis, M. S. Floyd, and Z. A. Saleem.  1986.  In Situ
      Biological Degradation Test at Kelly Air Force Base, Volume  I:  Site
      Characterization, Laboratory Studies and Treatment System Design and
      Installation.  (ESL-TR-85-52-I)  U.S. Air Force Engineering  & Services
      Center, Tyndall Air Force Base, FL  32403  291p.
                                     520

-------
12.
13.
14.
15.
16.
17.
18.
19.
20.
Wetzel, R. S., C. M. Durst, D. H. Davidson, and D. J. Sarno.  1987.  In
Situ Biological Treatment Test at Kelly Air Force Base, Volume II:
Field Test Results and Cost Model.  (ESL-TR-85-52-II)  U.S. Air Force
Engineering & Services Center, Tyndall Air Force Base, FL  32403  170p.

Wetzel, R. S., C. M. Durst, D. H. Davidson, and D. J. Sarno.  1987.  In
Situ Biological Treatment Test at Kelly Air Force Base, Volume III:
Appendices.  (ESL-TR-85-52-III)  U.S. Air Force Engineering & Services
Center, Tyndall Air Force Base, FL  32403  215p.

Olfenbuttel, R. F., E. Heyse, D. C. Downey, and T. L. Stoddart.  1988.
Lessons Learned: A Basis For Future Success.  IN: Contaminated Soil '88
Volume 1, Proceedings of the Second International TNO/BMFT Conference on
Contaminated Soil, April 11-15, 1988, Hamburg, Federal Republic of
Germany, K. Wolf, J. van den Brink, and F. J. Colon, eds., Kluwer
Academic Publishers, The Netherlands, pp. 471-479.
Ghassemi, M.  1988. Innovative In Situ Treatment Technologies for
Cleanup of Contaminated Sites.  Journal of Hazardous Materials 17:
206.
                                                                         189-
Puglionesi, P. S., J. Kesari, M. H. Corbin, and E. B. Mangel and.  1987.
Heavy Metals-Contaminated Soils Treatment.  IN:  Superfund '87,
Proceedings of the 8th National Conference.  The Hazardous Materials
Control Research Institute, Silver Spring, MD.  pp. 380-384.

Smith, R. L., D. T. Musser and T. J. DeGrood.  1985.  In Situ
Solidification/Fixation of Industrial Wastes.  IN:  Management of
Uncontrolled Hazardous Wastes Sites.  The Hazardous Materials Control
Research Institute, Silver Spring, MD. pp. 231-233.

Stinson, M. K., and S. Sawyer.  1988.  In Situ Treatment of PCB-
Contaminated Soil.  IN:  Superfund '88, Proceedings of the 9th National
Conference.  The Hazardous Materials Control Research Institute, Silver
Spring, MD.  pp. 504-507.

Wiles, C. C. and H. K Howard.  1988.  U.S. EPA Research in
Solidification/Stabilization of Waste Materials.  IN: Land Disposal,
Remedial Action, Incineration and Treatment of Hazardous Waste -
Proceedings of the Fourteenth Annual Research Symposium.  EPA/600/9-
88/021.  USEPA Risk Reduction Engineering Laboratory, Cincinnati, OH.
pp. 126-135.

U.S. Environmental Protection Agency.  1986.  Handbook for
Stabilization/Solidification of Hazardous Wastes.  EPA/540/2-86/001
USEPA Hazardous Waste Engineering Research Laboratory, Cincinnati, OH.
156 p.
                                       521

-------
21.   White, D. C., J. R. Dunckel and T. D. Van Epp.  1986.  Applying
      Alternative Technologies at Superfund Sites.  IN:  Management of
      Uncontrolled Hazardous Wastes Sites.  The Hazardous Materials Control
      Research Institute, Silver Spring, MD.  pp. 361-364.

22.   La Mori, P. N.  1987.  In Situ Treatment Process for Removal of Volatile
      Hydrocarbons From Soils:  Results of Prototype Test.  IN:  Proceedings,
      Second International Conference on New Frontiers for Hazardous Waste
      Management.  USEPA Hazardous Waste Engineering Research Laboratory,
      Cincinnati, OH  pp. 503-509.

23.   Van Tassel, R. and P. N. LaMori.  1987.  Innovative In Situ Soil
      Decontamination System.  IN:  Superfund '87, Proceedings of the 8th
      National Conference.  The Hazardous Materials Control Research
      Institute, Silver Spring, MD.  pp. 396-402.

24.   Bort, M. R.  1985.  In Situ Stabilization of Viscoelastic Polymer Waste.
      IN:  Management of Uncontrolled Hazardous Wastes Sites.  The Hazardous
      Materials Control Research Institute, Silver Spring, MD.  pp. 152-156.

25.   Harsh, K. M.  1978.  In Situ Neutralization of an Acrylonitrile Spill.
      IN:  Proceeding of 1978 Conf. on Control of Haz. Materials Spills, April
      11-13, Miami, FL, Information Transfer Inc., Rockville, Maryland pp.
      187-189.

26.   King, J., T. Tinto, and M. Ridosh.  1985.  In Situ Treatment of
      Pesticide Contaminated Soils.  IN:  Management of Uncontrolled Hazardous
      Wastes Sites.  The Hazardous Materials Control Institute, Silver Spring,
      MD.  pp. 243-248.

27.   Paulson, D. L., R. Honeycutt, Jr., J. Lebaron, and V. Seim.  1984.
      Degradation of High Concentration of Diazinon in Soil by Parathion
      Hydrolase, in 1984 Hazardous Material Spills Conference Proc.-
      Prevention, Behavior, Control, and Cleanup of Spills and Waste Sites,
      Ludwigson, J., ED., Government Institutes, Inc., Rockville, MD.  pp. 92-
      97.

28.   Ryckman, M. D.  1984.  Detoxification of Soils,  Water and Burn Residues
      from a Major Agricultural Chemical Warehouse Fire.  IN: Proceedings of
      the 5th National Conference on Management of Uncontrolled Hazardous
      Waste Sites.  Hazardous Materials Control Research Institute, Silver
      Spring, MD.  pp. 420-426.

29.   Stief, K. 1984.  Remedial Action for Groundwater Protection Case Studies
      Within the Federal Republic of Germany.  IN: Proceedings of the 5th
      National Conference on Management of Uncontrolled Hazardous Waste Sites.
      Hazardous Materials Control Research Institute,  Silver Spring, MD.  pp.
      565-568.   '
                                     522

-------
30.   Murdoch, L., B. Patterson, G. Losonsky, and W. Harrar.  1989.  A Review
      of Innovative Technologies of Delivery or Recovery for the Remediation
      of Hazardous Waste Sites.  Final report to the USEPA Risk Reduction
      Engineering Laboratory for work assignment 1-11, Contract 68-03-3379.
      Center Hill Research Facility, Cincinnati, OH  45221.  96p.

31.   Beljin, M. S. and J. F. Keeley.  1989.  In Situ Delivery and Recovery:
      Pulse (Cyclic) Pumping, Phase I.  Contract 68-03-4038 with the USEPA
      Risk Reduction Engineering Laboratory, Cincinnati, OH., Work Assignment
      2-WM32.0, University of Cincinnati.

32.   Keeley, J. F.  1988.  Remedial Performance Evaluation for Pump and
      Treat.  Cooperative Agreement CR812808 between the USEPA Robert S. Kerr
      Environmental Research Laboratory, Ada, OK. and the National Center for
      Groundwater Research, Rice University.

33.   Nash, J. H.  1988.  Project Summary:  Field Studies of In Situ Soil
      Washing.  EPA/600/S2-87/110, February 1988.  USEPA Hazardous Waste
      Engineering Research Laboratory, Cincinnati, OH.

34.   Urlings, L. G. C. M., V. P. Ackerman, J. C. V. Woundenberg, P.P. v.d.
      Pijl, and J.J. Gaastra.  1988  In Situ Cadmium Removal - Full-Scale
      Remedial Action of Contaminated Soil.  IN:  Contaminated Soil '88 Volume
      1, Proceedings of the Second International TNO/BMFT Conference on
      Contaminated Soil, 11-15 April 1988, Hamburg, Federal Republic of
      Germany, K. Wolf, J. van den Brink, F. J. Colon, eds., Kluser Academic
      Publishers, The Netherlands, pp. 911-920.

35.   Legiec, I. A., and D. S. Kosson.  1988.  In Situ Extraction of
      Industrial Sludges.  Environmental Progress 7(4):270-278.

36.   Herrmann, J. G.  1989.  Status Report on Electro-Kinetics Technology.
      January  23, 1989 memorandum from J. G. Herrmann, USEPA Risk Reduction
      Engineering Laboratory, Cincinnati, Ohio to J. Kingscott, USEPA Office
      of Solid Waste and Emergency Response, Washington, D.C.  3p.

37.   Horng, J. and S. Banerjee.  1987.  Evaluating Electro-Kinetics as a
      Remedial Action Technique.  IN: Proceedings, Second International
      Conference on New Frontiers for Hazardous Waste Management  EPA/600/9-
      87/018F.  USEPA Hazardous Waste Engineering Research Laboratory,
      Cincinnati, OH.  pp. 65-77.

38.   Iskandar, I. K. and J. M. Houthoofd.  1985.  Effect of Freezing on the
      Level of Contaminants in Uncontrolled Hazardous Waste Sites.  Part 1.
      Literature Review and Concepts.  IN:  Proceedings, Eleventh Annual
      Research Symposium (EPA/600/9-85/013) Cincinnati, OH, April 29 - May 1,
      1985.  pp. 122-129.
                                     523

-------
39.   Iskandar, I. K., and T.  F. Jenkins   1985.  Potential Use of Artificial
      Ground Freezing For Contaminant  Immobilization.   IN:   Proceedings,
      International Conference on New  Frontiers for Hazardous Wastes
      Management, EPA/600/9-85/025.  USEPA Hazardous Wastes  Engineering
      Research Laboratory, Cincinnati, OH.  pp. 128-137.

40.   Anastos, G. J. and P. J. Marks,  M. H. Corbin and  M. F. Coia.  1985   In
      Situ Air Stripping of Soils Pilot Study, U.S. Army Toxic and Hazardous
      Materials Agency, Report AMXTH-TE-TR-85026, Aberdeen Proving Ground
      (Edgewood Area), MD.  92p.

41.   Electric Power Research  Institute.   1987.  Remedial Technologies for
      Leaking Underground Storage Tanks (EPRI CS-5261).  Electric Power
      Research Institute, Palo Alto, CA.   288p.

42.   U.S. Environmental Protection Agency.  1987.  Underground Storage Tank
      Corrective Action Technologies   EPA/625/6-87/015  USEPA Office of Solid
      Waste and Emergency Response, Washington, D.C.  203p.

43.   Glynn, W. and C. Fan.  1988.  Underground Storage Tank Corrective
      Action: Application and  Field Evaluation of Vacuum Extraction
      Technology.  IN: Land Disposal,  Remedial Action,  Incineration and
      Treatment of Hazardous Waste - Proceedings of the Fourteenth Annual
      Research Symposium.  EPA/600/9-88/021.  USEPA Risk Reduction Engineering
      Laboratory, Cincinnati,  OH.  pp. 12-24.

44.   U.S. Environmental Protection Agency.  1988.  Cleanup of Releases from
      Petroleum USTs: Selected Technologies (EPA/530/UST-88/001)  USEPA Office
      of Underground Storage Tanks, Washington, D.C.

45.   Hutzler, N. J., B. E. Murphy and J.  S. Gierke.  1989.  State of
      Technology Review: Soil  Vapor Extraction Systems.  Interim Final Report
      to the USEPA Risk Reduction Engineering Laboratory for Cooperative
      Agreement CR-814319, Michigan Technological University, Houghton,
      Michigan  49931.  94p.

46.   Metzer, N., M. Corbin, and S. Cullinan.  1987.  In Situ Volatilization
      (ISV) Remedial System Cost Analysis.  (AMXTH-TE-CR-87123) U.S. Army
      Toxic and Hazardous Materials Agency, Edgewood Area, Aberdeen Proving
      Ground, MD 21010.  36p.

47.   Webster, K. T.  1988.  Aquadetox/VES Integrated System.  AWD
      Technologies, Fairfield, NJ.  (personal communication)

48.   Dev, H. P.  Condorelli, J. Bridges, C. Rogers, and D. Downey.  1986.  In
      Situ Radio Frequency Heating Process for Decontamination of Soil.  IN:
      Solving Hazardous Waste  Problems Learning from Dioxins.  Division of
      Environmental Chemistry at the 191 Meeting of the American Chemical
      Society, New York, NY. April 13-18.   pp. 329-339.
                                     524

-------
49.   Dev, H., G. C. Sresty, J. E. Bridges, and D. Downey.   1988.  Field Test
      of the Radio Frequency In Situ Soil Decontamination Process.   IN:
      Superfund '88, Proceedings of the 9th National Conference.  The
      Hazardous Materials Control Research Institute, Silver Spring, MD.  pp.
      498-502.

50.   Lord, A. E., Jr., R. M. Koerner and V. P. Murphy.  1988.  Laboratory
      Studies of Vacuum-Assisted Steam Stripping of Organic Contaminants from
      Soil.  IN:  Land Disposal, Remedial Action, Incineration and Treatment
      of Hazardous Waste, Proceedings of the Fourteenth Annual Research
      Symposium, EPA/600/9-88/021.  USEPA Risk Reduction Engineering
      Laboratory, Cincinnati, OH.  pp.65-92.

51.   Udell, K. S.  1988.  Results of a Pilot Study of Vacuum Extraction and
      Steam Injection at Solvent Service, Inc., San Jose, California.
      University of California, Berkeley, CA. (personal communication).

52.   FitzPatrick, V. F., C. L. Timmerman and J. L. Buelt.  1987.  In Situ
      Vitrification - An Innovative Thermal Treatment Technology.  IN:
      Proceedings, Second International Conference on New Frontiers for
      Hazardous Waste Engineering Research Laboratory, Cincinnati, OH  pp.
      305-322.

53.   FitzPatrick, V. F.  1988  In Situ Vitrification an Innovative Melting
      Technology for the Remediation of Contaminated Soil.  IN: Contaminated
      Soil '88 Volume 1, Proceedings of the Second International  TNO/BMFT
      Conference on Contaminated Soil, April  11-15, 1988, Hamburg, Federal
      Republic of Germany, K. Wolf,  J. van den Brink, and F. J. Colon, eds.,
      Kluwer Academic Publishers,  The Netherlands,  pp. 857-859.
                                    525

-------
                         RREL EXPERT SYSTEMS PROJECT;
               DEVELOPING TOOLS FOR HAZARDOUS HASTE  MANAGEMENT
                by:   Jay E. Clements and Daniel G. Greathouse
                                   ABSTRACT

This paper describes  the ongoing work of the  RREL  Expert Systems Development
Project.  Due  to  recent increases in the number  of environmental regulations
and the level  of  compliance  required,  EPA regulatory personnel  must monitor a
greater number of sites and  monitor these sites with  more  scrutiny than ever
before.   In  an attempt to reduce this workload and  improve the response time
of EPA Regional Offices  and  authorized  States, the  Risk Reduction Engineering
Laboratory  of  EPA's   Office   of  Research   and   Development  is  developing
microcomputer-based  decision   support   tools   (expert  systems)   to  address
commonly encountered waste management issues.

The  authors  explore  the  basic  concepts  of  expert  system technology,  the
potential  advantages  of  applying  this  technology,  and  several  specific
applications in the area of hazardous waste management.
                                   526

-------
 INTRODUCTION

 Protection  of human health and the environment  from  the risks of exposure to
 hazardous waste  involves consideration of numerous complex issues.   Decision
 makers  of  the U.S. Environmental  Protection  Agency,  bound by Congressionally
 enacted  regulations,   must oversee  the  handling  and  disposal   of  hazardous
 substances  on a  daily  basis.    In  addition  to  the  legal  implications,  EPA
 project  managers   must  be  aware  of  a  myriad  of  scientific,  technical,
 statistical,  and  economic factors  which ultimately  influence  the  decision
 making  process.   With the  preservation  of our communities  and  the  health of
 their people  at  stake,  these  decisions  are important  public issues.   Yet with
 many EPA regional  offices  characterized by high workloads and tight deadlines,
 industrial  hazardous waste management programs are in need of tools to improve
 productivity  and the quality of decision making.

 Recognizing this  problem,  the Risk Reduction Engineering  Laboratory  of EPA's
 Office of Research and  Development embarked several years ago on an effort to
 develop computer  based utilities  to  assist  regional  decision makers.   These
 utilities   will    allow   those   charged   with   mission   critical   decision
 responsibilities   to   become   more  productive,   better  informed,   and  more
 consistent  in the  execution of their duties.  In some cases, these systems can
 also be used  as training devices  for  new  or inexperienced employees.   In this
 way the Agency can use  computers,  which  are becoming  increasingly  common in
 the workplace, to  combat huge workloads.  The technology used to develop these
 decision aides is  commonly known  as expert systems.

 DESCRIPTION OF EXPERT SYSTEMS

 Expert systems are computer programs  (software)  designed to provide  advice in
 a narrowly  defined domain.  The  expert  system is an automated  process  which
 incorporates  the judgement, experience, rules of thumb,  and intuition  used by
 a human  specialist to  emulate that  specialist's  problem  solving  abilities.
 The  objective  of  these  systems  is  to  furnish  the  same  recommendations
 concerning particular decision scenarios as would an authority in the field.

 Expert  systems  are one  product  of artificial intelligence  research  programs
which have  been  ongoing  for thirty years in  university  and computer  industry
 labs.   Shortly after  the  invention  of  the modern digital  computer,  it  was
 realized that such  a machine was capable not only of mathematical calculations
 but  also  of  executing  a  series  of  logical  calculations  which simulates  a
 subset of the human thought process.  Although sound in  theory, development of
 such systems  proved to  be an  overwhelming  task.   Early  advisory  programs,
 completed  in  the  mid  to  late 1970's,  required scores  of skilled  software
 professionals  to  develop  and  expensive  dedicated  computer hardware to  run.
These restrictions  placed  expert  systems  technology out of  the  reach  of most
commercial   enterprises.  Therefore computers  have traditionally  been  used in
 "number crunching"  type applications; accounting, statistics, and mathematical
modelling for example.
                                   527

-------
Two recent  advances  in the computer field  have  paved the way  for  the expert
system's transition from research to application.  First, knowledge processing
software which enables  complex heuristic models to be coded  quickly  and with
minimal  low  level   symbolic   programming   requirements  has  become  widely
available.   Second,  breakthroughs in microelectronics design and fabrication
have increased the computing power to  price ratio of digital  machinery beyond
even the most optimistic  predictions of a decade  ago.   Consequently, serious
expert system applications can now  be  fielded economically  enough to render
the potential of swift return of the development costs.  Additionally, many of
these systems can run on computer hardware that is currently  in place for more
common engineering and data processing tasks.   Finally,  the  proliferation of
personal  computer  systems  has provided more workers than   ever before with
access to computational resources.

Knowledge based systems have several  characteristics which differentiate them
from conventional  software.   Whereas  conventional computer  programs rely on
numeric  algorithms  as the  basis  for  their operation,  knowledge  systems are
characterized by an emphasis on symbolic processing,  logical  inferencing, and
pattern matching.  Most knowledge based systems contain facilities which allow
the user  to obtain  an explanation concerning why  a  question  was asked and/or
to obtain  clarification for a  particular  question if the user requires more
information.   In  addition, the  user  is  often  supplied  with  the   reasoning
employed in reaching  a conclusion.  This allows the user to follow the logical
path from input data to  conclusion  when a justification for  the  results is
desired.    In  short,   expert  systems  are   interactive;  they  do  not   run  and
calculate so much as  they aid thought and offer advice.

Expert systems  have  their roots  in  traditional  data base management systems.
The standard data base however,  simply presents  relevant information to the
user and often only at the  expense of learning a cryptic query language.  The
expert system automates the process  of, not merely retrieving the appropriate
information,  but  also  deducing the  implications  of this information  in the
current  context.    It is  this application  of  information that  is  generally
referred to  as intelligence.

The substance of an expert  system is found  in its  knowledge base.  This  is the
collection  of facts and suppositions which when logically evaluated attempt to
reproduce the results of human  expertise.  This evaluation process is known as
inference  and software which performs  this  function  is  known as an  inference
engine.   Not long ago development of the  inference  engine was the first  step
in expert system development.   Today programs known  as expert system  shells
provide  completed  inferencing mechanisms  ready  to  accept  problem   specific
knowledge.   Programming  skills  are  still   necessary  for the  development of
expert systems  however these  skills  revolve around collection  and structuring
of  knowledge concerning  the  problem  domain.   These are  for  the most  part
unrelated   to   the   type   of  programming  expertise  necessary  to  build   a
generalized inference engine.
                                   528

-------
The  knowledge  base  is  really  just  a  sophisticated database  of  conditional
statements  or  rules.    For  example,  a  soil  mechanics  expert  system might
contain the following  rule:

IF the universal  soil  classification  is OL
THEN the soil type is  organic clay.

Statements following the IF are  known as the premise.  The premise  may  contain
a  combination of conjunctions  (two or more  statements connected  by the AND
operator)  or  disjunctions   (two  or  more  statements  connected  by  the  OR
operator).   In this way,  complex,  multi-conditional rules  can be developed.
The premise as  a whole evaluates to either true  or  false.   If the premise is
found to be true  then  the  statements following the THEN (the conclusion of the
rule) are executed.  When  this happens the rule is said to have fired.  In the
simple  example  above,  if it   is  known  that  the   variable   containing  the
universal soil  classification contains  the value OL  then  a  variable known as
soil type  is  given the  value "organic  clay".   If another rule is added, the
system can make a logical  deduction.

IF the soil type  is organic clay
THEN the soil is  not suitable for use in a liner system
BECAUSE organic  clays  may form  pores  and  thus increase  in  permeability with
time.

This rule makes a deduction about the  suitability of a  specific soil  type for
use in a  soil liner system.   If the goal  of  the system were  to make  such a
recommendation, this  rule would be  targeted  since the inference engine  sees
that  the conclusion   of  this  rule  can  provide  us with  a  recommendation.
However,  before  the  recommendation  is presented,  it must be proven  that the
premise  of  this  rule  is  true;  therefore a  determination of  the soil  type
becomes the new temporary  goal of the  system.  The knowledge  base is searched
for any fact  or rule which gives a  value to  the  soil  type variable.  Since no
facts concerning  soil  type have been recorded,  our first rule  is targeted.
This rule in  turn requires information on the universal  soil  classification.
Again the knowledge base is searched.  This time, neither a fact nor a rule is
located.   Usually when this occurs,  the user  is  asked to  supply the necessary
information although other options may be available.   If the user is asked and
indicates that  the  universal soil   code  is  OL  then  both rules fire and  a
recommendation  is formed.    The text  following  the  BECAUSE can be used  to
justify the recommendation if the user so requests.  If a soil  type other than
OL  is  specified  the  first  rule fails  and  the  knowledge  base   is  searched
further for rules which might lead  to  a value  for  soil  type.    This process
continues until   a  recommendation   is  found  or  until  all  possibilities  are
exhausted.

This  example   is  admittedly   trivial  however  it  illustrates  an  inference
strategy known as backward chaining.  Backward chaining  is only one of several
inference strategies  which  may be employed  in  an  expert  system.    It  is
represented here  because it is  the  most commonly used  technique  in knowledge
                                   529

-------
base construction.  It should also be noted that the rules here are written in
English  and therefore  would not  currently  be meaningful  to any  inference
engine other  than the human  mind.   Each  expert  system shell has  a specific
syntax.  This  syntax  defines the format that rules  and variables must assume
if the computer is to interpret them properly.

The process of collecting the rules  and  placing them in the shell is known as
knowledge  engineering.    In  general,  knowledge  engineering   is  the  most
important part of any expert system development endeavor.   Regardless of the
attractiveness and  features  in  a system,  if the program  does  not accurately
emulate the decision  process of the experts,  it  will  not  be of  value to the
end users.  The  knowledge engineering component  differentiates  expert system
development from conventional software engineering in several ways.  The first
difference is problem selection.  Not all problems are amenable to solution by
an expert system.   In order  for  an  expert  system  to be feasible there must be
recognized experts  in  the field that the  knowledge  engineer can  access.   The
problem  should not be  too  simple  and  it's solution  must represent  a  high
enough payoff  to  justify  the time and money  spent  developing  the system.   On
the other  hand,  the problem must not  be too complex or  the development  task
will quickly  become overwhelming.   In addition, a  successful  solution to the
problem  under consideration  must  be based upon  domain  specific  intellect
rather than common  sense or physical ability.

Once a problem is selected there is the matter  of shell selection.  Each shell
has a  unique  user  interface  and a  unique  knowledge representation structure.
A shell  should be selected which simplifies the knowledge engineering task as
much as  possible since this  is  a  very expensive aspect  of development.   The
user interface provided by the  shell  is  also of concern.   If users are unable
to efficiently communicate with the system they will be reluctant to use it at
all.

Finally,   there   are  the   concepts  of  rapid  prototyping   and  iterative
development.   Since outside  parties (the domain  experts)  are  involved in the
development  process  every  bit  as  much   as  the   system  designers,  it  is
imperative  that  close contact  be  maintained between  experts  and developers.
Rapid  prototyping  dictates   that  a  small,   naive  system  based  upon common
knowledge  about  the problem  be  developed  before any  experts  are approached.
This   information   can  be   collected  from  trade   magazines  and  reference
documents.   By demonstrating the  rapid  prototype,  the experts  are shown the
form and objectives  of the  system  at  the first interview.  The  experts  will
undoubtedly find many flaws  in  the prototype.   These  insights  are the basis
for  further development  of  the system.   After the  interview,  the knowledge
engineer implements additions and changes  suggested  by  the  experts and returns
to  them  for another demonstration  and  interview.   This  iterative development
process  insures  that  the  knowledge engineer does not misinterpret information
gained  through the interviews.   After  several iterations  the   system should
evolve to  the point where the  problem is  solved to  the  satisfaction of both
the domain  experts  and the system developers.
                                    530

-------
BENEFITS OF  EXPERT  SYSTEMS

The  benefits  of applying  expert  systems  technology  depend on  the  problem
domain to be addressed.   In most  cases, expert  systems  are  implemented  in  hope
of decreasing  operational  costs.   An expert's time can be a  costly  commodity.
Obviously,  if  an automated  process can successfully  distribute expertise to
many problem instances, this  system can yield considerable  savings.

Note that the  expert system  is not an  attempt  to replace domain specialists.
Currently, knowledge representation technology  is far from the stage where the
human thought  process can be  cloned.  Rather, a good expert system can  imitate
a human expert's responses in a  limited problem  scope  so as  to reach  similar
conclusions  in routine or monotonous cases.  This liberates the  expert  so that
time can be  spent on more complex or research oriented  problems.

Conservation of  expert  resources is not  the sole benefit  of expert systems.
By using an  automated system, every solution alternative  can  be  systematically
considered.  Expert systems do not  reach conclusions without  first considering
all important  factors and outcomes.  In addition, these factors  can  usually be
examined by  a  machine much  faster than  they could manually.  An expert system
may solve  a problem  in  minutes  that could require hours  or even  days  of a
human expert's time.  Expert  systems are also potentially more  available than
human  experts.  Domain  specialists are generally  in  high  demand.    In some
cases  there may be  a  significant time  interval  before  an expert   becomes
available to begin work on a  problem.

Another advantage of the  automated  system  is that machines are not  subject to
time,  pressure,  or  personal  constraints.    A  computer  will always  deliver
consistent responses  given consistent  input information.  In  an environment
where there  are  many decision makers and a need for uniformity among them, an
expert  system  can  simultaneously  increase the  quality,   consistency,  and
response time of decisions.   Furthermore, an expert system can often be geared
to act as a  training tool  for inexperienced employees.   Lastly,  creation of a
prototype knowledge  base can be  used as  a storage media  and  organizational
mechanism for  information in  a very dynamic field  of study.   A knowledge base
can depict   more  complex informational  relationships   than  a database while
maintaining  greater  accessibility to the  information  than  standard  technical
documentation.

RREL EXPERT  SYSTEM PROJECTS

For several  years  the EPA's  Risk  Reduction Engineering  Laboratory  (formerly
Hazardous  Waste  Engineering  Research   Laboratory)   has   been  active   in
development  of expert systems to  aid hazardous waste managers.  As a result of
this project, two tools have been developed, more are scheduled for completion
this year,   and  others  are in  the  beginning  stages of  development.   These
systems  will  run  on  standard  IBM  compatible  microcomputers.    Although
dedicated workstations   would  increase the performance of many  of  these
systems,  the laboratory has decided to develop for the PC platform in order to
                                   531

-------
minimize hardware  requirements  thus  making these systems available to as many
users as possible.  Not  all of  these programs are expert systems in the strict
sense  of the  term.   Each however  has the  potential  to  make environmental
decision  makers  either  better informed,  faster,   or  more  consistent  in  a
specific  area  of  concern.   The following  paragraphs  will  briefly  describe
these systems.

FLEX

FLEX is an  advisory system to assist in the analysis of chemical compatibility
data for flexible  membrane liners.   EPA specifies that geosynthetics which are
used  in hazardous waste  containment  applications must  first be  tested for
compatibility  with leachate from the contained  substances.  The test specified
for  this  determination  is  EPA Method   9090.    FLEX  assists  the   user   in
interpretation of  the  test results of  Method  9090.  This program was completed
in November of 1988 and  is currently  undergoing beta  testing in selected EPA
offices prior  to formal  release.

GM

The  GM system is   a  geosynthetic  modelling  advisor.   It  is not  so  much  an
expert  system  as a technology transfer tool.   It contains  knowledge concerning
the  appropriate engineering  calculations  used  in  geotechnical  designs  which
incorporate geosynthetic components.  It  is  used as a quick reference to the
equations  used in determining   safety  factors  for  a  variety of  geotechnical
structures.  This  system is available.

WAPRA

The  WAPRA system  assists  in the review  of the Waste Analysis Plan Section  of
RCRA  Part  B  facility  permit  applications.     Specifically,   this   program
automates  two  functions of waste  analysis  plan  review.    First,  the  system
"checks all wastes  which will  be processed  in  the same treatment  or  disposal
unit to  identify   possible chemical  interactions  which might  pose  a  health
threat to  unit operators  or demean  the  performance of the waste  processing
unit.   Second,  sampling methods are  analyzed  and judged  for  appropriateness
based  on guidelines in EPA Reference SW-846.   WAPRA is  better classified,as  an
 intelligent database than as an expert system.   It differs from a conventional
database in that  it  is  mostly  menu  driven and thus does not  require  the user
to learn a query  language.  In  addition,  WAPRA has  the  capability to  classify
wastes according  to  chemical  characteristics  and  decide  when two classes  of
 chemicals are  incompatible.  This  system is scheduled for completion  in June
 1989.

 FINAL COVER ADVISOR

 The final  cover  advisor makes recommendations  concerning standard  RCRA  cap
 design for hazardous  waste land disposal  facilities.   Increasing restrictions
 on  the land  disposal  of hazardous  waste  may force  closure  of many such
                                     532

-------
 facilities.   This  system assesses the  long term integrity  of a final  cover
 system based on  information provided in  the facility's  closure  plan    This
 system is  scheduled for completion in  June 1989.

 LEACHATE COLLECTION SYSTEM ADVISOR

 The  leachate collection and removal  (LCR)  system  analyzes  the suitability of  a
 leachate collection system based upon various site characteristics and design
 parameters.    It  makes  judgements on  operational  specifics  such  as   flow
 capacities,  pipe  strengths, and  clogging  potential.    Development  of   this
 system is  ongoing with  a tentative completion date of June  1989.

 VEGETATIVE COVER ADVISOR

 The  VEGCOVER system  contains knowledge concerning  the suitability  of  plant
 species proposed for  use in  the  vegetative cover  layer of landfill and surface
 impoundment   caps.     This  program   contains   information  on  the  growing
 characteristics  of  over  60 types  of  vegetation  commonly  used in  capping
 applications.   This  is compared  to  soil   type  and climate  designations for
 areas  within  the 50 United  States.  Development of this system  is ongoing with
 a tentative  completion  date  of June 1989.

 TECHSCRN

 TECHSCRN advises the  user of the technical feasibility of various remediation
 alternatives  for   Superfund  sites.     The  system   prompts  the  user  for
 characteristics  of the site  and   the  contaminants.    It  then  searches its
 internal  database  of  35  remediation  technologies   and   screens  out  those
 technologies  which  are  not appropriate for the cleanup task.   This  system is
 currently  a  rapid prototype.   Plans have been made  to  continue development of
 this system  but  a  completion date has not been  identified.   Future versions
 may  incorporate cost  analysis algorithms or a treatment train optimizer.

 CONSTRUCTION DESIGN REVIEWER

 This system will conduct a prefinal  review of design plans and specifications
 for  remedial  action at  a Superfund site.   The  objective of  the  system  is to
 assure  that  these  designs  are  biddable,  constructive,   operable,   and
 consistent with applicable environmental  regulations,   this system is still in
 the planning stage.

 CONSTRUCTION CLAIMS ADVISOR

This system will assist  the  site engineer  when a  contractor proposes  a change
 in  work   activity  at   a   Superfund   site.    The  system  will   assess   the
appropriateness  of  the  proposal,  suggest  responses   to  a   variety  of
unpredictable site  events  (eg.,  delays,  additional  work,  contract disputes),
and help determine whether there is a  legal  basis for  recovering  the  costs of
a change order.   This system is still  in  the planning stage.
                                  533

-------
RISK ASSESSMENT

This  system  helps  assess  health  risks  associated  with  hazardous  waste
exposure.   It contains information  on  maximum contaminant  levels  (MCLs)  and
alternate  contaminant levels  (ACLs)  and selects  the  appropriate model  or
methodology for predicting risk  levels.   A rapid  prototype has been developed
and work on this system is continuing.

QA/QC SAMPLING PLAN ADVISOR

This  system assists  the  user  in  the  analysis  of  QA/QC  sampling  plans  for
Superfund sites.  The objective of the system is to check for conformance with
Agency guidelines and procedures.   In  addition, data is examined to assure it
is sufficiently  accurate  to support  regulatory  and/or cleanup  decisions.   A
rapid prototype has been developed and work on this system is continuing.

LOOKING TO THE FUTURE

In general,  response to this project  has been enthusiastic.   It seems that,
given the  volume of applications that  must be processed  by EPA's regulatory
arm, reviewers are  eager  to  employ  methods which  might increase productivity.
Of  course,  as  with  any  new  technology,  quality  assurance  is  a  concern.
However, by  involving the prospective users as well  as domain experts in the
iterative  development process,  those  responsible  for  decision  quality  are
contributing  to  the shape of the product.   This  does  a lot to alleviate the
skepticism with which expert systems are  often met.

If  this  project is  successful,  individual  expert  systems  will  soon  make
employees  more  informed  and  more efficient  by  placing volumes of reference
materials  and expertise  at their  fingertips.    In  the  future  all  computer
software will be expected to incorporate  the cognitive  features  distinctive of
expert  systems.  When this  happens,  a network of advanced technology transfer
systems  will  allow each environmental  decision maker to  perform  as does the
best.   In this respect, RREL is working toward the future  today.

For  current availability information  regarding  any  of the  systems described
herein  contact:  USEPA, Risk Reduction  Engineering Laboratory, Municipal Solid
Waste and Residuals Management Branch  (MSWRMB), Cincinnati,  Ohio 45268.
                                   534

-------
                ASSESSMENT  OF  CHEMICAL AND  PHYSICAL METHODS  FOR
                      DECONTAMINATING BUILDINGS AND DEBRIS~
                               AT SUPERFUND SITES

                by:  Michael L. Taylor, Ph.D.,
                     Majid  A.  Dosani,
                     John A. Wentz,
                     Roxanne B. Sukol
                     Timothy L. Kling,
                     Jack S. Greber,
                     PEI Associates, Inc.
                     Cincinnati, Ohio

                     Naomi  P.  Barkley
                     Risk Reduction Engineering Laboratory
                     U.S. Environmental Protection Agency
                     Cincinnati, Ohio
                                  ABSTRACT

     Many Superfund sites contain buildings, building materials, and debris
that are contaminated with one or more toxic organic and/or inorganic chemi-
cals.  To date, no generally applicable decontamination technique has been
developed for the removal of organic contaminants such as polychlorinated
biphenyls (PCB's) from the various materials included in a modern-day struc-
ture.  The objective of this study was to evaluate chemical and physical
methods for decontaminating buildings and debris at Superfund sites.

     For evaluation of techniques designed to remove PCB's from concrete
floors in buildings, concentrations of PCB's in the top 1/2 inch of a con-
crete floor in a building located at a Superfund site were determined before
and after treatment by analyzing cores obtained from selected locations in
the floor.

     An innovative system for decontaminating debris was also designed, as-
sembled, and evaluated.  After bench-scale experiments were performed to de-
termine an optimal solution for cleaning PCB-contaminated, corroded, metallic
components, a 300-gallon, pilot-scale module was designed and field-tested at
a Superfund site.

     The results obtained during this study were very promising, and the
techniques evaluated showed a great deal of potential for removing PCB's from
concrete flooring and from the surface of the contaminated debris.  This
paper discusses the procedures developed and the results obtained by imple-
menting the chemical and physical decontamination techniques evaluated in
this study.
                                      535

-------
                              DISCLAIMER NOTICE
     This paper was prepared as an account of work sponsored by an agency of
the United States Government.  Neither the United States nor any of their em-
ployees, nor any of the contractors, subcontractors, or their employees make
any warranty, expressed or implied, or assume any legal liability or respon-
sibility for any third party's use or the results of such of any information,
apparatus, product, or process disclosed in this paper or represents that its
use by such third party would not infringe on privately owned rights.  The
views and conclusions contained in this document are those of the author and
should not be interpreted as necessarily representing the official policies
or recommendations of the U.S. Environmental Protection Agency or of the U.S.
Government.
                                      536

-------
KEY-WORD INDEX

Remediation
Decontamination
Super-fund
PCB
Concrete
Debris
Chemical Reagent
Shotblasting
                                    537

-------
                ASSESSMENT OF CHEMICAL AND PHYSICAL METHODS FOR
                      DECONTAMINATING BUILDINGS AND DEBRIS
                               AT SUPERFUND SITES

                                INTRODUCTION

     A large number of sites (more than 20,000) in the United States are con-
taminated with hazardous waste, and cleanup of these waste sites is the top
environmental priority of the decade.  As of this writing, about 1200 sites
are listed on the National Priorities List (NPL), and many more sites have
been proposed for inclusion on the list.  In the Superfund Amendments and Re-
authorization Act (SARA), a mandatory cleanup schedule is being considered
that would require EPA to clean up at least 375 of these sites over the next
5 years.

     Many Superfund sites contain buildings (e.g., office buildings, manufac-
turing facilities), building materials (e.g., glass, concrete, mortar, brick,
stone), and debris (e.g., scrap metal, pieces of wood, equipment or furni-
ture) that are contaminated with one or more toxic organic and/or inorganic
chemicals.  Decontamination of these items is important in preventing the
spread of contamination off site and in reducing exposure levels to future
users of the buildings or equipment.  To date, no generally applicable decon-
tamination technique has been developed for the removal of organic contami-
nants, such as polychlorinated biphenyls (PCB's), from the various materials
included in a modern-day structure.  At present, equipment is usually steam-
cleaned, and buildings and structures are frequently torn down instead of
being decontaminated.

     A recent U.S. Environmental Protection Agency publication (1) discusses
various methods including scarification, hydroblasting, and a variety of
chemical treatments for cleaning the surfaces of concrete and similar mate-
rials.  Many of these methods produce large amounts of liquid residues that
have to be collected and treated.  Moreover, the overall effectiveness of
these methods has not been carefully verified.  Reliable data are lacking
that provide an indication of the efficacy of either established or emerging
methods of decontaminating intact structures and structural components.

     Therefore, the major objective of this study was to evaluate the effi-
cacy of currently available chemical and physical decontamination methods for
removing contamination from intact buildings and restoring these structures
to a usable condition.  Another goal was the investigation of methods that
show promise for removing toxic contaminants from debris.  A successful
decontamination method can offset the high costs of dismantling and disposing
of contaminated structures, while at the same time salvaging or increasing
the value of the reconditioned buildings, equipment, or property.
                                     538

-------
      Phase  I  of  this  study was  directed  toward  locating an actual contami-
 nated site  or sites that would  be  suitable for  demonstrating various cleanup
 or  site  remediation technologies.   Several sites  (some NPL-listed and some
 non-NPL-listed)  were  visited and evaluated.  Ultimately, two PCB-contaminated
 Superfund sites  were  selected for  Phase  II of this study:  the Pioneer Equip-
 ment  site and the Carter Industrial  site, which are less than 1 mile apart in
 Detroit, Michigan.

      In  Phase II of this building  and debris decontamination study, the
 efficacy of selected  decontamination technologies for removing PCB's has been
 evaluated during field studies  performed at the two contaminated sites in
 Detroit.  Methods for removing  PCB's on the surface and in the top i inch of
 a concrete  floor at the Pioneer site were assessed in a comparative fashion.

      Also during Phase II, a system for decontaminating debris was designed
 and assembled.   Bench-scale studies were conducted to determine the optimal
 solution for  hydromechanically  cleaning contaminated debris.  Based upon the
 outcome of  bench-scale studies, a  pilot-scale version of the debris cleaning
 system was  designed and demonstrated at the Carter site.


      BUILDING DECONTAMINATION  DEMONSTRATIONS AT PIONEER EQUIPMENT SITE

      At the Pioneer Equipment Site, two decontamination techniques for re-
moving PCB's were demonstrated.  These two techniques were 1) a method for in
 situ  degradation of PCB's that entails application of an alkali metal/poly-~
 ethylene glycolate mixture directly to the concrete surface, and 2) a shot-
 blasting technique that entails use of steel  shot to cut away concrete sur-
 faces.  The decontamination tests were conducted in a statistically valid
manner in an attempt  to minimize both the effects of point-to-point varia-
 tions in PCB concentrations in or on the floor and the subtle effects of
variations  in concrete composition itself on the comparison of decontamina-
tion methods.

SAMPLING PROCEDURE

     Prior to implementation of the two decontamination technologies,  the
concrete floor (which is located in an abandoned building)  was divided into
sections, and each section subdivided to form test plots.   Within these test
plots, sampling locations were identified for baseline (pretreatment)  and
posttreatment sampling.   More specifically,  the floor was  sectioned off into
four 20-foot x 10-foot rectangular plots and  each was then  bisected to create
a total of eight 10-foot x 10-foot plots.  These plots were all  in the north-
west quadrant of the north room of the building and had a  north/south  longi-
tudinal axis.  Five sampling locations were  identified in  each of the  eight
test plots.   Two adjacent points were marked  off at each  of the five sampling
locations, and the first of the two adjacent  points was sampled as part of
the pretreatment sampling program.   The second  point was  sampled  subsequent
to treatment of the test plot.
                                      539

-------
     Concrete core samples were obtained by use of a 2-inch coring tool at
each of the five sampling locations in each test plot.  Each concrete core
had a diameter of approximately 1.75 inches.  From each of these eight plots,
five core samples were obtained for a total of 40 samples.  The top i inch of
24 of the 40 samples (i.e., 3 per plot) was analyzed for PCB content.  The
remaining 16 core samples (2 from each plot) were each divided into two sec-
tions (the top i inch and the next i inch, thus creating 32 samples) and each
section was subsequently analyzed for PCB's.  Thus, a total of 56 sample
analyses were completed before the two decontamination procedures were imple-
mented.

     After completion of the two decontamination technologies, another 40,
core samples were obtained from locations on the floor that were as close as
possible to the original sampling locations.  The samples were sectioned and
analyzed as described for the pretreatment samples, resulting in another set
of 56 analyses.  The locations of the top i inch and the next i-inch core
samples are shown in Figures 1 and 2, respectively.

DEMONSTRATION OF THE IT/SEA MARCONI REAGENT:  METHODOLOGY AND RESULTS

     The chemical reagent, also called the IT/SEA Marconi reagent, is a poly-
ethylene glycol-based mixture.  During the technology demonstration at the
Pioneer Equipment Site, cans of IT Marconi reagent were heated in a waterbath
to a temperature of 180° to 190°F.  When the reagent reached the correct
temperature, a small amount was poured onto the concrete floor within desig-
nated test plots and was spread evenly by a roller.  This process of applying
the reagent was repeated two additional times at 24 and 48 hours following
the initial application of reagent.  The reagent was then allowed to dry for
2 weeks.

     After 2 weeks, 20 posttreatment core samples were collected from the
areas that had been treated with the Marconi reagent and were analyzed for
PCB's.

     The concentrations of PCB's in the top i inch of concrete before and
after the application of IT/Sea Marconi reagent are summarized in Table 1.
The reduction in PCB concentration ranges from 11 to 97 percent (average
reduction of 73 percent), and these results seem to indicate that the reagent
was able to penetrate into the concrete floor and react with PCB's.  The
penetration of the reagent into concrete and reaction with PCB's is supported
by the results obtained for the next i inch of concrete, as shown in Table 2.
In this latter table, the percentage reduction in PCB concentration ranges
from 66 to 99 percent (average reduction of 91 percent).  In previous studies,
it has been found that PCB-containing oils can penetrate deeply into concrete.
For example, at the Dayton Tire Site (an NPL site in Dayton, Ohio), a poured
concrete slab that supported large transformers was sampled for PCB content.
PCB's in the concrete ranged in concentration from 47,400 yg/g at a depth of
0.25 to 0.5 in. to 33.700 yg/g at 3.25 to 3.50 in.  In an assessment of
another PCB-contaminated site for a confidential client, a similar finding
was made—PCB's spilled on concrete penetrated into and even through an 8- to
                                      540

-------
                           O
                               r-
                               co
                                       CO
                                       CO
                               CO
                               CO
o>
ffi
CO
SHOTBLASTING
SB11 SB12 I
•«•••




• •
m
CO
eo
m —
CO
5
CO
•^••M



MARCONI REAGENT
MR11 MR12





m
«
!
t
10
E
•t
••"
"^l
L— ^



    co
                  CO
CO
                  CO
                  CO
                          ID
                                  ffi
                                                                                                                        CO
                                                                                                                       .CD
                                                                                                                        Q.


                                                                                                                        05
                                                                                                                        CO
                                                                                                                        o
                                                                                                                        o

                                                                                                                        o
                                                                                                                        a.
                                                                                                                        o
                                                                                                                       *-
                                                                                                                        CD
                                                                                                                       O
                                                                                                                       CO

                                                                                                                       g


                                                                                                                       I

                                                                                                                       O)

                                                                                                                       "5.

                                                                                                                       03
                                                                                                                       CO
                                                                                                                       §
                                                                                                                       D)
                                                                                                                       U_
                                                          541

-------
                            S3
                            en
                            co
                       O

                       CO
       111

       i

       I
              to
              CM
              m
              co
                            CD
                            CO
                       I
                       g
                       5
(O
   o
       1
              to
a
                                                                                                                        (A
                                                                                                                       .0)

                                                                                                                        Q.

                                                                                                                        E
                                                                                                                        OS
                                                                                                                        CO

                                                                                                                        £

                                                                                                                        8
                                                                                                                        c=

                                                                                                                        0)
                                                                                                        CO

                                                                                                       JO

                                                                                                       "ca
                                                                                                        O)


                                                                                                       "o.
                                                                                                        E
                                                                                                        (0
                                                                                                       CO
                                                                                                                       CM
                                                                                                                        C5)
                                                       542

-------
   TABLE 1.  CONCENTRATION OF PCB's IN A SURFACE (TOP i INCH)  OF
CONCRETE BEFORE AND AFTER APPLICATION OF SEA MARCONI  REAGENT (ppm)
Sample No.
MR1
MR2
MR3
MR4
MR5
MR6
MR7
MRS
MR9
MR10
MR11
MR12
MR13
MR14
MR15
MR16
MR17
MR18
MR19
MR20
Pretreatment
52.0
6.5
22.0
5.7
8.3
6.5
9.2
42.4
22.1
23.3
33.2
21.7
50.0
19.8
39.8
4.6
8.1
60.0
10.2
12.3
Posttreatment
3.67
2.46
0.64
1.58
2.79
2.35
1.96
1.43
1.94
1.58
5.03
3.25
5.45
5.63
6.22
3.04
4.12
5.82
9.10
5.48
% Reduction
93
62
97
72
66
64
79
97
91
93
85
85
89
72
84
34
49
90
11
55
 TABLE 2.  CONCENTRATION OF PCB's IN A SUBSURFACE (NEXT i INCH)  OF
CONCRETE BEFORE AND AFTER APPLICATION OF SEA MARCONI REAGENT (ppm)
Sample No.
MR21
MR22
MR23
MR24
MR25
MR26
MR27
MR28
Pretreatment
3.5
5.0
6.3
13.0
4.1
1.8
2.7
3.0
Posttreatment
0.23
0.23
0.21
0.10
1.41
0.35
0.10
0.01
% Reduction
93
95
97
99
66
80
96
99
                               543

-------
12-in.-thick poured concrete slab.  Of course, the permeability of the concrete
is influenced by many factors, such as the density, moisture content, and
cracking.                           ,                      j

     On the basis of this rather limited study, it appears that the IT/SEA
Marconi reagent is effective in reacting with PCB's that a're on the surface
of or perhaps contained within the upper i inch of a concrete floor.  Addi-
tional studies involving analyses of concrete cores prior to and subsequent
to application of the reagent should be performed to corroborate these
initial findings.  Furthermore, additional work is needed 'to determine the
optimum reaction conditions for the IT/SEA Marconi reagent.

DEMONSTRATION OF THE SHOTBLASTIN6 PROCESS:  METHODOLOGY AN'D RESULTS

     Shotblasting is a destructive procedure that may be used to remove
surface layers of contaminated concrete.  By selecting the shot size and the
rate at which the machine traverses ;the concrete surface, one can control the
amount of concrete removed from the .surface of a concrete floor.  Typically,
a 1/16- to 1/8-inch layer of concrete can be removed by the shotblaster.
(Note:  The shot blasting machine used in this study was Blastrac Model 1-10D
with a Model 554-DC dust collector.)  The machine is equipped with a HEPA-
filtered vacuum system that captures nearly all of the paryticulate generated
during shotblasting.  The captured dust is periodically removed from the
vacuum system of the shotblaster and transferred to barrets for subsequent
disposal..  Despite the HEPA-filtered vacuum system, use of the shotblasting
process may generate airborne debris, which has the potential for cross-
contamination of test plots.  Therefore, the shotblasting evaluation at the
Pioneer Equipment Site was commenced after all activities related to the
evaluation of the IT/Marconi reagent were completed.      [

     Designated test plots on the co'ncrete floor (see Figures 1 and 2) were
shotblasted and a minimum of 0.125 to 0.250 inch of concrete was removed.
Fugitive dust generated by the process that was not captured by the vacuum
system was gathered and vacuumed separately, and disposed of as contaminated
concrete material.  Following shotblasting, 20 posttreatment core samples
were collected.
                                                          i
     The concentrations of PCB's in the top i inch surface of concrete before
and after the shotblasting process are summarized in Table 3.  The percentage
reduction in PCB concentration after1 shotblasting ranges fjrom 19 to 96 per-
cent (average reduction of 68 percen't), which indicates that the technique is
fairly effective in removing PCB's from the surface of the concrete floor.
However, two samples (SB2 and SB11) Ishowed higher PCB concentration after the
shotblasting treatment.  These latter results can be explained on the basis
that the distribution of PCB within ,the concrete is undoubtedly nonuniform,
and therefore it is probable that thje two posttreatment samples were obtained
from locations in the concrete floor where the initial concentrations of PCB
were much higher than the concentrations of PCB found in the SB2 and SB11
pretreatment samples.               <
                                      544

-------
TABLE 3. CONCENTRATION OF PCB's IN A CONCRETE SURFACE (TOP i INCH)
BEFORE AND AFTER SHOTBLASTING (ppm)
Sample ID
SB1
SB2
SB3
SB4
SB5
SB6
SB7
SB8
SB9
SB10
SB11
SB12
SB13
SB14
SB15
SB16
SB17
SB18
SB19
SB20
Pretreatment
4.4
0.13
65.0
7.7
26.0
17.2
11.8
35.7
13.4
4.6
1.6
4.4
4.7
23.5
3.2
25.0
22.8
21.4
40.9
42.5
Post-treatment
3.0
1.5
2.52
3.17
2.11
6.92
5.44
6.3
3.4
3.91
1.86
2.75
1.97
1.56
1.41
4.84
1.28
5.38
8.15
7.78
% Reduction
32
-1054
96
59
92
60
54
82
75
15
-16
38
58
93
56
81
94
75
80
82
545

-------
     These results indicate that in the case of the PCB-contaminated floor at
the Pioneer site, PCB's in the top i to i inch of the concrete floor are ef-
fectively removed by shotblasting.  However, further studies should be per-
formed to determine the depth of penetration of PCB's into the concrete.  As
discussed previously in this paper, PCB-laden oil can, at least in some
cases, penetrate deeply into concrete.  It may be necessary, therefore, to
repeatedly shotblast the concrete surface to remove the contaminated layers
of concrete.

     Only a small amount of fugitive dusts were generated by the shotblasting
process.  An estimated 95 percent of the concrete dust was captured by the
vacuum system, which is an integral part of the machine.  In total, a 1000-
square-foot area of concrete was shotblasted, removing approximately a 1/8-
to 1/4-inch layer of concrete.  The total quantity of dust collected amounted
to two barrels.

COST ESTIMATION OF MARCONI AND SHOTBLASTING TECHNOLOGIES

     On the basis of the experience at the Pioneer site, a cost analysis was
performed to determine the costs of large-scale implementation of each of the
two building decontamination techniques.  The cost of implementation of any
cleanup technology at a hazardous waste site can vary substantially depending
on the site's location, nature, types of contaminants, availability of facil-
ities, and many other variables that can unexpectedly increase or decrease
the estimated cost.  Therefore, to establish a base in the cost estimation,
several assumptions have to be made.

     A 10,000-square-foot plot was taken as a baseline area on which the cost
estimation was performed.  These costs are based on the following assump-
tions:  1) the site is within 50 miles of the contractor's facility, 2) no
pretreatment cleaning of the plot is required, 3) the plot is free of
obstructions such as debris, equipment, or machinery, and 4) the site has
electrical power and outlets.  The costs for the large-scale implementation
of the IT/SEA Marconi reagent and shotblasting techniques that were
calculated on the basis of these assumptions are summarized in Tables 4 and
5, respectively.

     As shown in these tables, the cost of Marconi reagent is 85<£ per square
foot, compared with $2.19 per square foot for shotblasting.  The shotblasting
technique is labor intensive and generates a significant quantity of contami-
nated waste, whereas the Marconi reagent technique requires minimal labor and
the reagent does not generate wastes.  On the basis of limited experience in
these studies, it appears that the cost of shotblasting is almost 3 times
higher than that of the Marconi technique.  These costs do not include  sam-
pling and analytical costs.

      DEMONSTRATION OF DEBRIS DECONTAMINATION:   BENCH-SCALE EXPERIMENTS

      In designing the debris decontamination system,  the goal was to produce
a portable, self-contained module  in which debris could be washed by a  non-
toxic cleaning solution.  The debris decontamination  system was designed to
                                      546

-------
        TABLE 4.  ESTIMATED COST OF IMPLEMENTATION OF MARCONI REAGENT
    TECHNIQUE BASED ON THREE APPLICATIONS TO A 10,000-FT2 CONCRETE PLOT
                       Description
  Cost
1.  Fixed rate labor (includes cleanup technicians),
     70 hours at $50/h
2.  Fixed rate equipment (includes reagent heating system,
     applicator)
3.  Other direct costs
      Expendables (includes Marconi reagent, protective
       clothing, empty drums, miscellaneous)
      Nonexpendables
      Travel (@ $50 per diem)
4.  Disposal (approximately 2 drums @ $125/drum)
5.  General  and administrative (calculated on Item 3 @ 16.1%)
6.  Fee (on  Items 3 and 5 @ 10.0%)
        Total estimated cost for 10,000 ft2
        Cost per square foot
$ 3,500

    500


  2,900

      0
    450
    250
    539
    389
$ 8,528
$  0.85
                                     547

-------
    TABLE 5.  ESTIMATED COST OF IMPLEMENTATION OF SHOTBLASTIN6 TECHNIQUE
    BASED ON THE REMOVAL OF TOP 1/4 INCH OF A 10,000-FT2 PLOT OF CONCRETE
                         Description
Cost
1.  Fixed rate labor (includes foreman, cleanup technician, and     $  9,000
     equipment operator), 180 hours at $50/h
2.  Fixed rate equipment (includes rental of shotblaster,              4,612
     HEPA vacuum, air compressor)
3.  Other direct cost
      Expendables (includes vacuum bags and filters, micro-            2,090
       traps, protective clothing, miscellaneous)
      Nonexpendables                                                       0
      Travel (@ $50 per diem)                                          1,500
4.  Disposal (approximately 30 drums of contaminated concrete          3,750
     @ $125/drum)
5.  General and administrative (calculated on Item 3 @ 16.1%)          ,  578
6.  Fee (on Items 3 and 5 @ 10%)                                         417
        Total estimated cost for 10,000 ft2                         $ 21,947
        Cost per square foot                                        $   2.19
                                     548

-------
include a solvent reclamation system that would permit the cleaning solution
to be reclaimed and reused, thereby minimizing the volume of contaminated
liquid produced during the debris washing process.  In order to test some of
these concepts, the bench-scale experiments described below were performed.

     A bench-scale version of a Turbo-Washer (Bowden Industries) served as
the debris washer for the initial studies.  This unit includes an axial flow
pump, a propeller shaft, a propeller, and a pressure chamber, all housed
within a heated tank also equipped with a rotating disc for removing oil that
rises to the surface of a protected (minimal turbulence) segment of the
cleaning tank.  During operation the Turbo-Washer's pump vigorously mixes the
cleaning solution by continuously recirculating the cleaning fluid in and out
of the pump.  The items to be cleaned are placed into a wire-mesh basket that
is lowered into the heated tank where the items are exposed to the turbulent
cleaning solution.

     The bench-scale version of the Turbo-Washer was used to evaluate four
cleaning solutions including tap water, 10 percent sulfuric acid, and aqueous
dilutions of two proprietary cleaning solutions:  BB-100 (Bowden Industries)
and Power Clean (Penetone Corporation).  The experimental procedure involved
the application to rusted iron parts of measured quantities of used motor
oil, grease, and soil to simulate the kind of grime likely to be encountered
on oily, PCB-contaminated metal parts and debris in the field.  Three tests
were performed with each cleaning solution, and a fresh set of oil/grease
contaminated metal parts was employed for each test.  Each set of contam-
inated parts was matched closely with regard to size, shape, and type of
metal.  The parts were arranged in the same order in the parts washer basket
during washing.

     At the completion of each of the bench-scale tests, two aliquots of
cleaning solution were collected from the heated tank, one aliquot was sub-
mitted for oil and grease analysis, and another for total suspended solids
analysis.  In addition, two surface wipe samples were collected from the
surfaces of the cleaned metal parts, and these were also analyzed for oil and
grease to determine the level of oil/grease remaining on the metal surfaces
after treatment in the debris washer.  The skimmer oil collected during each
of the three runs was combined for oil and grease analysis.  The amount of
oil and grease in each of the samples was determined by standard EPA methods
(2).

     The results of the oil/grease and total suspended solids analyses are
summarized in Table 6.  The analytical results of the wipe samples indicate
that> after cleaning, the amount of oil and grease on the metal surfaces was
significantly higher in the case of water or sulfuric acid and comparatively
lower for BB-100 and Power-Clean.  This indicates poor cleaning performance
of water and sulfuric acid.  Moreover, the handling of 10 percent sulfuric
acid was difficult, and it also had a severe corroding effect on the debris
washing equipment.  Hence, it was concluded that neither water nor sulfuric
acid should be considered as a potential cleaning solution for. oily PCB-con-
taminated debris.

     On the basis of the results of the surface wipe testing listed in Table
6, it was concluded that BB-100 solution is a more effective cleaning solution
                                      549

-------










CO
CO
1
^
CO
o
1— <
1
o
CO

0
Ul
0
i»^|
UJ
Ou
CO

co

^
0
I*"" *
o
§
UJ
CO
UJ
CO
^**»
P"j
o

C£
o
Li-

CO
1
1

co
UJ


o

5^»
cc

3IZ
=>
CO


•
(O

UJ
_J
CO
J—
























c
o
•r"
3
"o
CO

cn
c
c
to
at
o



























































o
C 4J
re re
CU S~
010
CU
(X C
o
u
>
="
1 rH
CO
CO •
§
u





"o
•o >

u •
re t>
3r
u
•r- ••
$- fc^
3O
CO U
c
0
u

1_
CU
.[ %
re





CO
co

r—
re
c







re
cu
Q_

**
CU
'S.
E
re
CO







re
+•>

CU O
E: 2E
'£ C
CU 3
OL S-
X
UJ





O 0 0
rH rH CM





"^ CO rH
rH CO














rH CO CO
ID «J- CO
rH rH rH





CM rH rH
rH CM



CU V CU
CO CO CO
to re re
 T3 4J
K-r- C •!- C •!-
re r-~ re r^ re r-~

r— COr— COr— CD



C C C
O O O
•r- 'r- •!-•
^ ^3 ^
r— r— r—
000
CO CO CO

C C C

C C C
co re re
01 CU 01

CJ O t_>









rH CM CO




1




CO CO "G}*
o r** co






C3 «4- O
o o o
rH













co in co
CM If) «3-
rH CM rH





cn r-.. co


i~ i. i-
01 01 01
01 'r- CU "r- CU *r-
•O i— -O r- X> i—
c •»-. c >^. c >^
O) en oj co o> cn
Q. E D» E o. E
CO CO CO
3 • 3 «3 »
CO CO CO CO CO CO
•o -o -o
r— • *r- r— «r- r~ »r"
re ^^ re r^ re r*~
•P O 4-> O 4J O
o to o to o to



c c c
o o o

4J +J 4J
r— r— r—
O O O
CO CO CO

C C C
•r~ «r- *r-
C C C

  • <_> rH CM CO O CTl CO CO «* rH CO *tf- ^}- «3* CO CO O O O O O O CM ID CO CM CO CO CO CM rH «3- CM CO • ••••• O O 0 O O O CO IT) rH r~- rH rH O ^ CO CO r-» CM •* o co f» oo cn «i- **• i rH rH O «* CM CU Ol Ol fll 01 CU CO CO CO CO tO CO re re re re re • re 01 Ol Ol Ol O) Ol cn cn cn cn cn cn T3N "ON "ON T3N "OM T3W C £E C £ C E C E c F" ci E reureureureureureo r— cnr— cnr— cnr— cnr— cnr— cn rH CM rH CM rH CM O O O O O O 01 01 01 Ol 01 0) O. CL O- O. Q. O. 3: !I !2E ^ 3! 3 rH rH CM CM CO CO O § VI CO § CO CO o rH 1 Ol CO re cn s- "O 4J C -r- fO r- if s_ 01 •r— CO § 4- r— O CO <*- o oa r— •> re CM o • r- rH CO 01 Q. re CO +•> C 01 B re £ -i i to o a. £ re to 01 "o. re to r_ cC irj 550

  • -------
    than a Power Clean solution.  This is shown graphically in Figure 3, which
    plots the results obtained for the wipe samples (in milligrams of oil and
    grease/square centimeter) from each test.  Table 6 results also show that
    cleaning with BB-100 removed solids from metal surfaces more effectively than
    Power-Clean.  The data also show that at the completion of the third run, the
    BB-100 solution still had more cleaning capacity to remove dirt from metal
    parts than did Power-Clean.  Hence, of the four cleaning solutions tried,
    BB-100 was selected as the cleaning solution best suited for cleaning oily
    PCB-contaminated metal parts and debris in the field.
    
         Throughout these cleaning solution evaluations, the Bowden Turbo-Washer
    performed well.  Good agitation of the cleaning solution was attained and the
    unit performed reliably.  Therefore, it was concluded that a large-scale
    version of the Turbo-Washer would be used in subsequent field tests.
    DEMONSTRATION OF DEBRIS DECONTAMINATION:
    CARTER INDUSTRIAL
    RESULTS OF PILOT-SCALE TESTS AT
         An Experimental Debris Decontamination Module (EDDM) was designed and
    assembled on the basis of the bench-scale results.  A 300-gallon-capacity
    Turbo-Washer was installed on a 48-foot semitrailer and the Turbo-Washer was
    modified by incorporating a particulate removal system and oil/water separa-
    tor.  Also mounted on the trailer was a carbon sorption system for removing
    PCB's from the cleaning solution.  Figure 4 represents a flow-diagram of the
    pilot-scale module.  The trailer-mounted module was transported to the Carter
    site for field-testing.
    
         At the Carter site, two 200-1b batches of metallic debris were cleaned
    by using the EDDM.  A solution of BB-100 surfactant was used as the cleaning
    solution.  Prior to the cleaning process, five individual pieces of metal
    from each batch were sampled for PCB's by a surface wipe technique.  The
    debris items were placed into a basket and transferred into the EDDM and the
    cleaning process was instituted.  Each batch of debris was cleaned for a
    total of 2 hours.  During the cleaning process, a portion of the cleaning
    solution contained in the Turbo-Washer was pumped through a closed-loop
    system in which the oil/PCB-contaminated wash solution was passed through the
    particulate filter and into the oil/water separator.  The effluent from the
    oil/water separator was then recycled into the module.  At the completion of
    the cleaning process, five additional wipe samples were obtained from the
    same pieces of metallic debris to assess the postdecontamination level of
    PCB's.  The surface wiping procedure was carried out as described in the
    "Field Manual for Grid Sampling of PCB Spill Sites to Verify Cleanup" (EPA
    560/5-86/017, May 1986, pg. 33), which entails use of the hexane-soaked
    cotton gauze pad to wipe a 100-cm2 area on the surface of the object being
    sampled.  In the case of the metallic debris sampled in this study, the
    posttreatment wipe sample was obtained from a location adjacent to that of
    the pretreatment sample.  The surface wipe samples were analyzed for PCB's at
    Hayden Environmental, Inc. (formerly PCS, Inc.) in Dayton, Ohio.
                                         551
    

    -------
        0.6 -
    u
    
    en
    UJ
    C_3
    CO
    
    
    1
    2=
    o
    UJ
    
    I
    CD
        0.5 •-
                                                      POWER CLEAN
        0.4 --
         0.3 - -
                                                        BB-100
         0.2
         0.1
           (0,0)
                                             RUN NO.
    
    
                Figure  3.   Amount of  oil  and grease  on surfaces  of ;metal
                            parts after completion  of cleaning cycle.
                                              552
    

    -------
                                    s
                                    9
                                    ^>
                                    5
                                    o
                                    <;=
                                    OS
                                    TO
                                    8
                                    e
                                    Q.
                                    J>
                                    8
                                    •93
                                   EL
    553
    

    -------
         The quantity of PCB's on  the  surface of each piece of metal before and
    after  cleaning  is summarized in Table 7.  The percentage reduction of PCB's
    achieved during cleaning ranges from 33 to 87 percent  (average reduction of
    58  percent) for Batch  1 and from 66 to 99 percent (average reduction of 81
    percent) for Batch 2.  In the  case of Sample 1  in Batch 2, however, the PCB
    analysis gave a higher concentration after the  cleaning process probably
    because the posttreatment wipe sample was obtained from a location on the
    debris surface  that was initially  more heavily  contaminated with PCB's.
         TABLE 7.  CONCENTRATION OF PCB's FOUND IN SURFACE WIPES AND BLANKS
    Sample No.
        Pretreatment,
         pg/100 cm2
    Posttreatment,
     ug/100 cm2
    % Reduction
    Batch 1
    
          1
          2
          3
          4
          5
    Batch 2
    
          1
          2
          3
          4
          5
             134
             490
            1280
              73
             203
    
    Blank:  <1.0 yg/100 cm2
             8.0
          6090
           374
            96
          1690
    
    Blank:  1.0 ug/100 cm2
         50
        178
        856
         43
         23
         13.0
       1800
        128
         10
         18
          63
          64
          33
          41
          87
                                                     Average % reduction:  58
         -63
          70
          66
          90
          99
                                                     Average % reduction:  81
         The results also indicate that the quantity of PCB's removed during
    cleaning of Batch 2 was greater than that of Batch 1.  The reason for better
    cleaning results for Batch 2 could be due to the following:  In the case of
    Batch 2, after 1 hour of cleaning the basket containing the debris was re-
    moved from the washer and the parts were manually rearranged so that all
    sides of the debris were exposed to the cleaning solution with the same force
    of the Turbo-Washer.  The basket was then lowered back into the washer and
    cleaning was continued for 1 more hour.  In the case of Batch 1, however, the
    cleaning process was continued for 2 hours without the debris in the basket
    being rearranged.
                                          554
    

    -------
         The surfactant solution in the Turbo-Washer was sampled twice during the
    actual cleaning process; the concentrations of PCB's found were 928 and 420
    yg/liter.  Following completion of the debris washing experiment, the clean-
    ing solution was pumped through a series of particulate filters and finally
    through activated carbon.  The PCB concentration was reduced to 5.4 yg/liter
    during this treatment.  Most municipalities allow water containing _
    -------
                                REFERENCES
    
    Esposito, M. P., McArdle, J. L., Crone,  A. H., Greber, J.  S.,  Clark,
    R., Brown, S., Hallowell, 0. B., Langham, A., and McCandlish,  C.  D.
    Guide for Decontaminating Buildings, Structures, and Equipment at Super-
    fund Sites.  EPA Report No. 600/1-85/028, U.S. Environmental  Protection
    Agency, Cincinnati, Ohio, January 1985.
    
    Test Methods for Evaluating Solid Waste.  Volume 1C, SW 846,  3rd  ed.,
    November 1986.Office of Solid Waste and Emergency Response,  Washing-
    ton, D.C.
                                     556
    

    -------
          IMMOBILIZATION  MECHANISMS IN  SOLIDIFICATION/STABILIZATION
                       USING CEMENT/SILICATE FIXING AGENTS
    
                   by:  L.  G. Butler, F.  K.  Cartledge,  D.  Chalasani
                           H.  C.  Eaton,  F.  Frey,  M.  E.  Tittlebaum,
                           S.  L.  Yang
                       Hazardous  Waste Research Center
                       Louisiana  State University
                       Baton Rouge, LA 70803
                                  ABSTRACT
    
          Lime treatment of aqueous solutions of heavy metal  ions prior
    to solidification/stabilization produces a sludge which  is or-
    dinarily assumed to consist of a mixture of the corresponding metal
    hydroxides.  The behaviors of Cd- and Pb-containing sludges toward
    cement solidification  is quite different.  We have investigated
    this behavior for both sludges and soluble Cd and Pb salts using
    TCLP leaching tests, conduction calorimetry and solid-state nuclear
    magnetic resonance spectroscopy as a function of time.   Concentra-
    tions of Cd in leachates are very low, while Pb concentrations are
    considerably higher and would represent a serious threat to ground-
    water.  We believe that the Teachability differences arise for the
    following reasons.  The Cd/cement system involves Cd(OH)2, which
    provides nucleation sites for precipitation of CSH and calcium
    hydroxide, resulting in Cd being in the form of the insoluble
    hydroxide with a very impervious coating.  On the other  hand, the
    Pb/cement system is much more complicated in terms of the precipit-
    ation reactions.  In the presence of hydroxide, sulfate and (in the
    case of the soluble Pb salt) nitrate, mixed salts are precipitated
    which contain all three anions.   These salts retard cement hydra-
    tion reactions by forming an impervious coating around cement
    clinker grains.   However,  as pH in the cement pore waters undergoes
    fluctuations during the progress of hydration,  the Pb salts undergo
    solubilization and reprecipitation on surfaces of the cement
    matrix.
                                 557
    

    -------
                   WASTE REDUCTION EVALUATIONS AT FEDERAL SITES
    
                    by:   James S. Bridges
                          Risk Reduction Engineering Laboratory
                          U.S. Environmental Protection Agency
                          Cincinnati, Ohio 45268
                                     ABSTRACT
    
        The Waste Reduction Evaluations at  Federal Sites  (WREAFS) Program
    focuses on waste minimization research opportunities and technical
    information transfer within the Federal community.  Many Federal
    activities are implementing waste reduction plans and programs to reduce
    the generation of wastes.  For example, an overall goal within the
    Department of Defense is to reduce hazardous waste generation a total of
    50% by 1992 compared to 1985.  There are several operation and process
    opportunities for pollution prevention practices at Federal sites which
    will reduce wastes and result in providing cost avoidance or cost
    reduction as well as increased productivity, and  environmental and human
    health benefits.  Waste minimization technologies can be applied
    throughout the Federal community in fabrication, production and/or
    maintenance processes depending on the waste generating function.
    Industrial activities which support the Federal community through in-house
    or contract efforts, whether on-site or off-site, are generally the same
    activities practiced within the private sector.  The application of waste
    minimization research may vary depending on the production or consumption
    function; however, waste reduction techniques and technologies will work
    in both the Federal or private sectors.  WREAFS Program objectives
    include:  conducting waste minimization workshops, developing technical
    information transfer opportunities between industry and government;
    providing EPA research information to the Federal community; performing
    waste minimization opportunity assessments at selected Federal sites;
    demonstrating waste minimization techniques or technologies at Federal
    facilities; and promoting waste minimization within all portions of the
    Federal community.
                                        558
    

    -------
            DEMONSTRATE COMPUTER ASSISTED ENGINEERING (CAE)
               TECHNIQUES FOR REMEDIAL ACTION ASSESSMENT
              by:  P.R. Cluxton, W.G. Harrar, L.C. Murdoch,
                   and M. Beljin
                   University of Cincinnati
                   Department of Civil & Environmental Engineering
                   Cincinnati, Ohio
                                ABSTRACT
         A  computer  workstation   dedicated   to  remedial  action
    assessment  is being  developed.   The  system  is  composed  of
    several  off-the-shelf  software  and  hardware  modules,  with
    software  development  limited  to   the   creation   of  utility
    programs  used to  transfer data  from one  software module  to
    another.   The component  modules  include a Geographic  Infor-
    mation System, a  Database Management System,  a Computer Aided
    Design and Drafting System, and a Groundwater  Modeling System.
    The completed system will  be an example of a Computer Assisted
    Engineering (CAE)  type system.
    
         The  purpose  of  the  project is  to  demonstrate how  the
    remedial  action   evaluation   process  can  be  improved  and
    expedited  thru  use  of  the  CAE  system.    The   CAE  system
    capabilities include the production  of maps and cross-sections
    showing the  geology,  hydrology,  and  distribution of  contam-
    inants; the  calculation of volumes  or masses  of  contaminated
    material;   the modeling  of  groundwater   flow  and  contaminant
    transport; and the modeling of remedial actions such as cut-off
    walls or pumping.   Several site studies have been completed for
    the regional offices,  including the  Queen City Farms  Superfund
    site and  adjoining Cedar Hill Landfill in Seattle,  Washington,
    a  soil gas  survey in Croydon,  Pennsylvania  and  groundwater
    contamination  mapping  in  the  vicinity  of  the  Taylor  Road
    Landfill in Hillsborough County, Florida.
                                   559
    

    -------
                 Vi.VilAULIC MECHANISMS OF A MULTIPLE SOIL LAYER COVER
        by: Ati hot (s)
           Ric ard C. Warner
           University of Kentucky
           Lex ngton, KY 40546-0075
    James A. Harned
    Greiner Engineering Services
    Tampa, FL 33609-3416
    Nathaniel Peters, III
    Law Environmental
    Louisville, KY 40220
                                           ABSTRACT
    
        Invest  ations of the hydraulic  mechanisms of a field scale cover, with  a  clay barrier,
    were collided under both stable and subsidence conditions.  Temporal water movement as
    influenced by precipitation and  evapotranspiration were documented.  These investigations
    encompassed:  11) runoff, drainage,  leachate and moisture migration throughout  the complete
    cover,  (2j  a  variety  of  field   scale  hydraulic   conductivity  testing  procedures,  (3)
    determination of tracer breakthrough times, (4) modeling of preferential flow paths, and (5)
    determination of clay cap subsidence as a function of a cavity formed beneath the cap.
        The research was conducted  on  a 27.4 m long by 6.1 m wide cover consisting of a topsoil
    layer,  sand  drainage  layer,  and  compacted  clay  barrier layer.    The  clay  barrier  was
    constructed in 5 lifts at  a moisture  content 2 to 3% above optimum with regard to Standard
    Proctor.  The hydraulic  conductivity of the clay cap was examined using  sealed double  ring
    infiltrator (SDRI) and borehole methods.  The SDRI tests yield consistent values  of 7.9 x
    10"   and  8.9 x 10"  cm/s with an average  standard error  p^  9.0  x   10"   cm/|.    The
    borehole   lethod  yield  hydraulic   conductivities  of  2  x  10   and  8.0  x  10    cm/s,
    respective1:    A  ;00 day  breakthrough experiment was conducted using a conservative anionic
    tracer ami  >ui;i were compared to the SOILINER model.  After 90 days  the tracer initially
    appeared   ,-ve  Background levels.  This time frame is analogous to the tracer replacing the
    pore voiuir s of the largest macropores prior to being discharged.  Over  the next 110 days
    the effluent increased to within 80% of the initial tracer concentration.
        Througi a continuous stirred tank reactors (CSTR) modeling analysis it  was concluded
    that prefer ntial flow paths, representing only 6.2% of the total pore volume, conveyed 95%
    of the totai  volume of tracer through the barrier layer.   CSTR modeling determined that
    complete  tr cer breakthrough would take 620 days whereas the SOILINER Model predicted 5.76
    years.
        Research  on the influence of a  known cavity size created  directly beneath  the multiple
    soil  layer eover is  continuing  this  spring.   Preliminary results  indicate that  a clay cap,
    constructed 1 to  3% above optimum Proctor moisture content and subjected to an overburden of
    water and saturated sand, collapsed above a 1.57 m diameter cavity.
        Previoii, centrifuge experiments concluded that potential subsidence would decrease  due
    to increased clay cap thickness, construction at moisture contents near Proctor optimum,  and
    increased dry density.   Thus the selection of the  water content, used in  compaction of the
    clay  layer,  appears  to  present   a tradeoff   between  decreasing  clay cap's  hydraulic
    conductivity  and the structural  capability  of  the clay layer  to  span a  given size cavity
    created beneath the cap.
                                              560
    

    -------
                      Evaluation of Solidification/Stabilization
                             Treatability Studies at the
                    Uni'ted States  Environmental  Protection Agency
                                 Center Hill  Facility
    
                                          by
    
                                 Edwin  F.  Barth,  P.E.
                    United States  Environmental  Protection Agency
                        Risk Reduction Engineering Laboratory
                                    Cincinnati,  OH
    
                                       Abstract
     Solidification/Stabilization  is  being  considered  as  either  a  treatment or
    
     post treatment  option  for  several  uncontrolled  hazardous waste  sites
    
     containing contaminated  soil.  The up  front  prediction  of the feasibility of
    
     this treatment  process is  complicated  because of  the many possible chemical
    
     interactions  between the soil, site contaminant,  and binder.  RREL has been
    
     active  in performing solidification/stabilization treatability  studies at
    
     the Center Hill  Facility as well as reviewing treatability  studies from
    
     other sources.   The following general  technical issue areas have been common
    
     in the  studies  evaluated:
    
                     * feasibility of stabilizing organic waste
                     * long term durability testing
                     * placability of material
                     * volatile emissions during mixing and curing
                     * ability  to mix in field
                     * construction quality assurance
    
    A preliminary testing evaluation process has been developed for determining
    
    the feasibility of a solidification/stabilization process for contaminated
    
    soil.   Results of the treatability studies reviewed have been favorable for
    
    certain contaminated matrices and less favorable for others.
                                       561
    

    -------
      THE EPA MANUAL FOR WASTE MINIMIZATION OPPORTUNITY ASSESSMENTS
                 by:  Mary Ann Curran, Harry Freeman,
                      Kenneth R. Stone
                      U.S. EPA, RREL
                      Cincinnati, Ohio 45268
                                 ABSTRACT
    
         A waste minimization assessment (WMA) is a systematic
    planned procedure with the objective of identifying ways to
    reduce or eliminate waste.  The assessment consists of a careful
    review of a plant's operations and waste streams, and the
    selection of specific areas to assess.  After a specific waste
    stream or area is established as the WMA focus, a number of
    options with the potential to minimize waste are developed and
    screened.  The technical and economic feasibility of the selected
    options is then evaluated.  Finally, the most promising options
    are selected for implementation.
    
         Waste minimization is fast gaining recognition as a means of
    contending with the nation's hazardous waste problem and other
    forms of environmental pollution.  Opportunities exist for waste
    minimization throughout industry and government.  The waste
    minimization assessment procedure described^in the manual offers
    a means of determining a facility's waste situation and
    identifying and evaluating potential viable options for reducing
    waste.
    
         The waste minimization  assessment offers opportunities to
    reduce operating costs, reduce potential liability, and improve
    the environment, while improving regulatory compliance.  The WMA
    results in a careful review  of a plant's operations toward
    reducing wastes.
                                    562
    

    -------
                       THE U.S. EPA COMBUSTION RESEARCH FACILITY
                       by:   Johannes W. Lee, Larry R. Waterland
                             Acurex Corporation
                             Jefferson, Arkansas  72079
    
                             Robert C. Thurnau
                             U.S. Environmental Protection Agency
                             Cincinnati, Ohio  45268
                                       ABSTRACT
    
         During FY'88, the EPA Combustion Research Facility  (CRF)  in Jefferson,
    Arkansas conducted 58 incinerator tests over 26 weeks. Completed test programs
    include: demonstration of the American Combustion,  Inc.  Pyretron Thermal
    Destruction System under the Superfund Innovative Technology Evaluation  (SITE)
    program; tests to evaluate the use of SFg as a surrogate for destruction and
    removal efficiency in liquid injection incineration; evaluation of the fate of
    hazardous trace metals fed to the rotary kiln incineration system; and tests
    to evaluate the valence state of chromium discharges from the  rotary kiln
    incinerator. The presentation focuses on the SFg and trace metal test results.
    
         During the surrogate-tests, SFg DREs ranged from 99.9 to 99.99999 percent
    and POHC DREs ranged from 99.993 to 99.999997 percent. SFg ORE were uniformly
    lower than the corresponding POHC DREs and increased with increased combustor
    temperatures. SFg and POHC DREs showed no clear dependence on flue-gas 02.
    
         The trace metals tests studied the fate of arsenic, lead, cadmium,
    chromium and barium (appendix VIII) and copper, magnesium, bismuth, and
    strontium when these are subjected to incineration. Except for arsenic, metals
    discharge distributions were consistent with their volatilities. Cadmium,
    lead, and bismuth appeared to be volatile and were found mainly in the flue-
    gas.  Barium, copper, strontium, arsenic, chromium, and magnesium appeared to
    be refractory and were found mainly in the kiln-ash. Presence of chlorine in
    the waste shifted cadmium, lead, copper and bismuth from the kiln-ash to the
    flue-gas but did not affect arsenic,  lead,  chromium, barium, magnesium and
    strontium.  Neither kiln nor afterburner temperature within the range tested
    affected metal discharge distributions significantly. The venturi/packed
    column scrubber-system metals-removal efficiency ranged from 31 to 88 percent
    and averaged at. 57 percent and were lower for the volatile metals.
    
         In addition,  reports for several FY'87 test programs were issued.  Several
    facility upgrade efforts were completed.  A major facility reconstruction was
    initiated to expand the CRF's physical plant and capabilities.
                                         563
    

    -------
                EVALUATING THE COST EFFECTIVENESS OF SITE TECHNOLOGIES
    
                 By:  Gordon M. Evans
                      USEPA, RREL
                      Cincinnati, Ohio
          The goal of the Agency's Superfund Innovative Technology Evaluation
    (SITE) program is to develop reliable performance and cost data of unique and
    commercially available hazardous waste destruction and treatment technologies.
    One major challenge which faced the SITE program was how best to insure that
    the cost evaluation process produced estimates which were useful to the
    Superfund decision making process.  Toward this end, broad objectives were
    established to guide project managers in the construction of cost evaluations
    for innovative technologies.  As they have evolved, the goals of SITE
    program's cost analysis are:
    
          1)  To provide the Agency decision makers with the developer's cost
          analysis as well as an independent cost analysis, prepared by the
          Agency, and based on well defined and appropriate operating assumptions.
    
          2}  To identify and highlight all critical operating and economic
          variables which the engineering analysis suggests are likely to have
          significant impacts on costs.
    
          3)  To insure that Agency decision makers are provided with the
          opportunity to reconstruct any cost analysis, substituting assumptions
          which are more consistent with their own situation. This in turn means
          that all assumptions must be clearly stated and all equations clearly
          presented.
    
          This poster will first review the specific legislative and program goals
    guiding the SITE program cost evaluations.  Then it will examine the twin
    problems imposed on these cost evaluations by: 1) the need to balance the
    goals of the public and private sector in the marketplace;  2) the R&D aspects
    of demonstrating innovative technology.  The poster will describe how the SITE
    program has organized itself to facilitate the attainment of the goals,
    mentioned above.  An economic framework will be presented which segregates
    estimated costs into one of 12 categories.  Lastly, the poster will examine
    the results of completed SITE cost analyses.
                                        564
    

    -------
                             SOLIDITECH  SITE  DEMONSTRATION
    
                    by:  Walter E. Grube, Jr.
                         U.S. Environmental  Protection Agency
                         Cincinnati, Ohio  45268
    
                         Kenneth 6. Partymiller
                         PRC Environmental Management, Inc.
                         The Woodlands,  Texas  77380
    
                         Danny R. Jackson
                         Radian Corporation
                         Austin, Texas  78720
                                       ABSTRACT
    
          The Soliditech SITE technology demonstration was carried out to  obtain
    reliable process performance and cost information.  The demonstration  took
    place at the Imperial  Oil Company/Champion Chemicals Superfund Site in
    Monmouth County, New Jersey.  Contamination
    various metals, and petroleum hydrocarbons.
    waste material with Urrichem, a proprietary
    materials (portland cement was used in this
    batch-type concrete mixer.
    at this site includes PCB's,
     The Soliditech process mixes the
    reagent; additives; pozzolanic
    demonstration); and water; in a
          Technical criteria applied to evaluate the effectiveness of the
    Soliditech process include contaminant mobility, based upon solute extraction
    and permeability tests; and the retention of structural  integrity of the
    solidified material, based upon physical  and morphological  tests.
    
          Extensive sampling and analyses of three different waste types from the
    New Jersey site showed (1) a reduction of heavy metals in the TCLP extract of
    solidified waste samples, (2) no volatile organic compounds in the TCLP
    extract of solidified wastes, (3)  detectable levels of phenols and cresols in
                                       565
    

    -------
    the TCLP extract of treated waste, (4) an increase in oil  and grease in
    extracts of solidified waste, and (5) good structural stability of solidified
    waste under freeze/thaw, wet/dry, and static testing.
    
          Long-term testing for extractable pollutant solutes, Teachable solutes
    mobile through diffusion, morphological  changes in the concrete-like matrix,
    and structural properties are planned to continue for an additional  4-1/2
    years.
                                        566
    

    -------
                     BIOTROL SOIL WASHING SYSTEM
    
                        by:  Steven B. Valine
                             BioTrol, Inc.
                             Chaska, Minnesota 55318
                               ABSTRACT
    
         BioTrol, Inc.  (Chaska, Minnesota) has developed an on-site
    treatment system for the remediation of contaminated soils from
    wood preserving  sites.   The  system  is based on  a series of
    intensive scrubbing and physical separation steps using water
    as  a carrier  for  the  soil  and  contaminants.    Contaminant
    removal efficiencies of 90  to 95  percent have been achieved.
    Process water from the system is treated in a fixed film bio-
    reactor using an amended bacterial consortium prior to recycle
    back to the  soil scrubbing system.    This  paper presents the
    results of two phases of  pilot testing recently completed at
    a wood preserving site in Minnesota using a mobile, 500 Ib/hr
    pilot-scale unit.
                                 567
    

    -------
                       SEPARATION OF HAZARDOUS ORGANICS BY
                    LOW PRESSURE REVERSE OSMOSIS MEMBRANES
    
                              M.E. Williams and D. Bhattacharyya *
                              Department of Chemical Engineering
                                    University of Kentucky
                                     Lexington, KY 40506
    
                                             and
    
                                Richard P. Lauch, Project Officer
                              U.S. Environmental Protection Agency
                             Risk Reduction Engineering Laboratory
                                     Cincinnati, OH 45268
                                         ABSTRACT
    
          Membrane processes are gaining considerable attention for the separation of organics and
    inorganics from hazardous wastes. Membrane processes can be divided into four categories:
    high pressure RO, low pressure RO, nanofiltration, and ultrafiltration. Low pressure processes
    with thin-film composite membranes have definite advantages in terms of energy savings and
    capital cost. Extensive experiments were conducted with two types of low pressure aromatic
    polyamide membranes.  Membrane experiments were carried out with various priority organic
    pollutants with and without partial ozonation. Partial ozonation converts many chlorinated
    organics to intermediate organic acids. These ionizable intermediates were effectively removed
    by charged and uncharged membranes without significant drop in water flux.  Charged reverse
    osmosis membranes provide additional advantages because high water flux (11 x 10"4 cm/s) can
    be obtained even at low pressures (1.4 MPa).
    
    
     INTRODUCTION
    
          Chemical manufacturers generate millions of tons of wastes containing various hazardous
    priority pollutants each year. The development of alternative technologies for the treatment of
    various hazardous wastes is becoming increasingly important as concerns grow over its disposal.
    These and other wastes, such as leachate from unsecured disposal sites, contain a wide variety of
    priority pollutants such as pesticides, herbicides, PCBs, chlorinated hydrocarbons, and heavy
    metals.  Much of this waste is relatively dilute and so must be concentrated before further
    treatment (1,2).
    
          Several methods have been used for the treatment of dilute wastewater.  These include
    biological treatment, stripping, and carbon adsorption  (3, 4).  Ozonation has also been found to
    be effective in oxidizing some hazardous organics (5,6) to less toxic compounds. Membrane
    processes can be used to purify wastewater and produce a 20 to 50 fold decrease in waste
    volumes that must be treated with other processes such as incineration or wet air oxidation,
    greatly reducing energy cost.
    * corresponding author
                                            568
    

    -------
     MEMBRANE CONCEPTS AND APPLICATIONS
    
          Membrane processes are gaining considerable attention for the purification and volume
    reduction of dilute hazardous wastes.  The development of low pressure reverse osmosis (RO)
    membranes such  as aromatic polyamide and sulfonated polysulfone has resulted in membrane
    processes which give high water flux at low pressures (< 2 MPa).  These thin-film composite
    membranes can provide high solute separations and have definite advantages in terms of energy
    savings, capital cost, and broad pH (2 to 12) operating range. The low pressure membranes can
    provide simultaneous separation of hazardous organics and inorganics.  Selective separations of
    ionizable compounds are also possible with membranes containing charged groups (7, 8).
    
          Membrane performance is measured in terms of membrane rejection (R), permeate water
    flux (Jw), and extent of water recovery (r).  The rejection is a measure of solute separation by the
    membrane and is defined as
                                                 C
    where C and Cs are the solute concentrations in the permeate and feed streams.
    
          Reviews of the development of membrane technology can be found in Sirkar and Lloyd
    (9), Sourirajan and Matsuura (10), Lee (11), and Belfort (12). Although many of the applications
    referred to in these involve the use of cellulose acetate membranes, works with thin-film
    composite membranes were more favorable for the separation of hazardous wastes. Chian (13)
    used reverse osmosis systems to remove more than 99 % of fifteen major pesticides.
    Carcinogenic substances were removed 92.5 % with both spiral wound and hollow fiber
    polyamide membranes in studies by Light (14). Shrem and Lawson (15) used reverse osmosis
    units to treat wastewater from an organic chemical manufacturing plant and found organic matter
    rejection over 90 %. Bhattacharyya et al. (16, 17) have done extensive work with various
    priority organic pollutants using thin-film aromatic polyamide (FT30) spiral wound modules.
    Experiments with priority pollutants showed that rejections of >98 % for PAH compounds
    (napthalene, anthracene, phenanthrene) were possible with little drop in permeate water flux.
    For ionizable organics such as phenol, chlorophenols, and nitrophenols, they found that
    rejections and flux drops were highly dependent on operating pH values. Membrane rejections
    (atpH 11) were 99.5-99.8 % for phenol, 2-chlorophenol, 2,4-dichlorophenol,
    2,4,6-trichlorophenol, etc.
    
          Studies have shown that high separations of hazardous organics such as chlorinated
    phenolics can be achieved by low pressure polyamide membranes. However, these studies have
    indicated substantial water flux drop for non-ionized chlpro and nitrophenols due to
    membrane-solute interactions (17). It is known that partial ozonation will convert many
    chlorinated organics to intermediate organics acids (oxalic, formic, etc.) (5, 6, 18). These
    organic acids do not interact as strongly with the membrane and so cause less flux drop. Also,
    since these intermediates are ionizable, charged membranes can be utilized to achieve high water
    flux and separation of partially ozonated hazardous organics at low pressures.
    
    
     OBJECTIVES
    
          This work deals with the use of low pressure, thin-film, composite membranes for the
    concentration and separation of selected chlorophenols and chloroethanes with and without feed
    pre-ozonation. Membrane feed pre-ozonation will result in the formation of organic acid
    intermediates which should result in less flux drop, and, since these compounds are ionizable,
    high separations by charged as well as uncharged membranes. Separation and flux
    characteristics of both charged and uncharged membranes were studied with both ozonated and
                                              569
    

    -------
    non-ozonated solutions to evaluate the effectiveness of the ozonation - membrane process for the
    enhanced removal of priority organic pollutants.
    
    
    EXPERIMENTAL
    
          Membrane experiments were conducted with a system (Figure 1) containing both a batch
    and a continuous flow cell so that the performance of two different membranes could be studied
    at the same time. The thin-film, composite membranes used in the experiments were the Film
    Tec NF40 and the FT30-BW; the NF40 was placed in the batch cell and the FT30 in the flow
    cell. The pump shown in the system served to provide flow of solution through the continuous
    cell and mixing in the batch cell; compressed nitrogen was used to supply the pressure driving
    force for the system. The operating conditions were: system pressure (P>VB) of 0.3 -1.4 MPa,
    feed pH*s of 3.0 - 9.4, and system temperature of 24° C.
    avg •
         The procedure used for experiments involving ozonation - membrane process is outlined'in
    Figure 2. Membrane feed solutions were ozonated in a 1.8 liter stirred reactor with a flow of
    0.20 standard liters per minute (SLPM) O2 containing 2 % ozone. Pre-ozonation times ranged
    from 0 to 60 minutes. After ozonation, solutions were mixed for several hours to allow
    decomposition of residual ozone. Membrane experiments were then carried out with the
    ozonated solutions.
    
         Membrane performance was measured in terms of flux drop from that of distilled water
    flux (DWF) and solute rejection. Membrane feed, concentrate, and permeate samples were
    analyzed by Total Organic Carbon and HPLC direct injection.  Total Organic Carbon (TOC) was
    measured using a Beckman Model 915-B Carbon Analyzer.  HPLC analysis (phenolics) were
    performed with a Varian 5000 liquid chromatograph using a MCH-5 column (reverse phase
    octyldecylsilane on silica) and a UV-50 variable detector at 220 and 280 nm.
    
    
    RESULTS AND DISCUSSION
    
         Membrane separation of selected hazardous organics with and without feed pre-ozonation
    were investigated with the NF40 and FT30-BW membranes. Single component studies were
    conducted with trichlorophenol while multicomponent systems examined consisted of
    trichlorophenol/humic acid mixtures and mixtures of chlorophenol, dichlorophenol,
    trichlorophenol, trichloroethane, and tetrachloroethane.  A wide range of ozonation times and
    pressures were studied, and, since many of the compounds studied were ionizable, several
    membrane feed pH values were investigated.  Membrane performance results are presented in
    terms of solute rejection and flux drop. The % flux drop (at a particular pressure) is defined as:
    „   ,      Distilled Water Flux- Flux with Wastewater   inn
    Flux drop =	^.  .„  ,„,	—	x 100.
                          Distilled Water Flux
    
         Membrane stability was checked by standard sodium chloride and sodium sulfate
    rejections and distilled water fluxes. For the NF40, a negatively charged membrane made of
    carboxylated aromatic polyamide, the DWF at 1.38 MPa was found to be 11 x 10"4 cmVcm2 s
    (23.3 gal/ft2 day).  Sodium chloride and sodium sulfate rejections were 30 % and 97 %,
    respectively. The DWF of the FT30, an aromatic polyamide membrane, was also 11 x 10"4 cm/s
    at 1.38 MPa; sodium chloride and sodium sulfate rejections for this membrane were 96 % and 97
    %.  The low sodium chloride rejection of the NF40 illustrates the principal advantage of the
    charged membrane over reverse osmosis membranes such  as the FT30. The FT30 membrane
    gives high rejections of most solutes, whereas the NF40 membrane can be used to selectively
                                             570
    

    -------
    separate compounds with different charges.
     Studies with Trichlorophenol
    
         Membrane feed solutions of 50 mg/L trichlorophenol (TCP) were ozonated from 0 to 60
    minutes. The pH of the solutions decreased after ozonation; this drop in pH indicated the
    degradation of the TCP to intermediate organic acid compounds.  HPLC analysis indicated that
    after 5 minutes of ozonation the TCP concentration was reduced to 24.7 mg/L, and after 15
    minutes, was <0.2 mg/L.  Formation of carbon dioxide during the ozonation was indicated by
    reduction in TOC and the presence of carbon dioxide in the high pH trap placed after the
    ozonation reactor.
    
         Figure 3 shows flux drop as a function of ozonation time for the NF40 membrane for low
    (3.3-3.6) and high (8.8-9.4) pH membrane feeds. At low feed pH, TCP is not ionized and so
    interacts strongly with the membrane; the adsorbed TCP displaces water in the membrane pores,
    causing significant water flux drop. Ozonation of TCP results in the formation of organic acids
    which do not interact as strongly with the membrane and so cause less flux drop. At the high
    feed pH the TCP is ionized; the negatively charged solute does not interact as strongly with the
    charged membrane and the flux drop is less than that at the lower feed pH.  Since the ozonation
    products are also ionizable, the flux drop for ozonated solutions are also smaller under the high
    pH conditions.
    
         TOC rejections for the NF40 membrane are shown in Figure 4.  It can be seen that both an
    increase in feed pH and ozonation time increased rejections. An increase in feed pH results in
    the ionization of the TCP or the organic acid intermediates formed during ozonation.  These
    negatively charged species are more highly rejected by the negatively charged NF40 membrane.
    Ozonation improves rejection since the organic acids that are formed ionize at lower pH's than
    does TCP and so these are rejected better than TCP at lower feed pH's.
    
         Figures 5 and 6 show the flux drop and TOC rejection for the FT30 membrane. Under
    non-ionized conditions (feed pH 3.3-3.6), the TCP interacts strongly with the membrane, causing
    a large flux drop. The drop in water flux is greater than that for the NF40; the charge on the
    NF40 weakens the interactions between the TCP and the membrane. As with the NF40,
    ozonation reduces flux drop for the FT30 membrane by reducing solute interactions with the
    membrane. Under ionized conditions (feed pH 8.8-9.4), the TCP and organic acids formed
    during ozonation do not interact with the FT30 membrane as strongly as under non-ionized
    conditions and so the flux drop is smaller. Figure 6 shows that feed pH and ozonation did not
    greatly affect TOC rejection for the FT30 membrane; this membrane does not depend on charge
    for separation as does the NF40 membrane.
    
    
     Studies with Trichlorophenol/Humic Acids
    
         Experiments to determine the effect of water recovery on water flux and solute rejection
    were conducted with mixtures of 50 mg/L trichlorophenol and 10 mg/L humic acids (TCP/HA).
    Humic acids are high molecular weight compounds that are present in  soils and so can be found
    in ground water containing hazardous organic leachates.  These compounds are highly rejected
    by the membrane but can cause large water flux drops due to adsorption.
    
         The fluxes for distilled water, TCP/HA, and TCP/HA ozonated for 30 minutes are given in
    Figure 7 as a function of water recovery for the NF40 membrane. For both non-ozonated and
    ozonated TCP/HA, no drop in flux was observed even at high recoveries; the TCP/HA and the
                                             571
    

    -------
     organic acids formed during ozonation were ionized at the high feed pH (9.1-9.3) and so little
     change in water flux occurred. The TOC rejections are shown in Figure 8. Rejection of the
     non-ozonated TCP/HA decreased from 80 % at 10 % recovery to 64 % at a recovery of 80 %;
     however, for the ozonated solution 84 % TOC rejection was maintained even at 75 % water
     recovery. The decrease in TOC rejection for the non-ozonated TCP/HA could have been the
     result of enhanced concentration of the HA at the membrane surface. Since degradation of the
     HA is expected during ozonation, HA concentration on the membrane surface did not increase
     for the ozonated solution, and TOC rejection remained high.
    
    
      Studies with Mixtures
    
          Studies were performed with mixtures of 50 mg/L chlorophenol (CP), dichlorophenol
     (DCP), and trichlorophenol (TCP) with 100 mg/L trichloroethane (TCE) and tetrachloroethane
     (TTCE) to determine the effect of multicomponent systems on flux and TOC rejections for
     ozonated and non-ozonated solutions. The flux behavior of the mixture with the NF40 was
     found to be linear over the pressure range 0.34 -1.38 MPa (50 -200 psi), indicating the absence
     of surface polarization phenomena. Flux drops as a function of pH for the NF40 membrane at
     1.38 MPa are shown in Figure 9. Flux improves substantially for increase in feed pH for both
     the ozonated and non-ozonated mixtures; ozonation did not greatly improve flux drops found for
     the mixture at a fixed feed pH. However, Figure 10 shows that TOC rejection is enhanced
     significantly for the mixture after ozonation for 60 minutes. While the non-ozonated mixture
     rejection is almost constant for the different feed pH's, it is increased to as high  as 80 % for the
     ozonated mixture. The increase is due to the formation of organic acids which ionize and are
     rejected by the membrane. Although the phenolics in the mixture are ionizable, the pKa of these
     compounds are much higher than those of the intermediates, and also, the chloroethanes present
     are not ionizable.  As a result, the rejection did not increase over the pH range studied. However,
     the organic acids formed after ioiu'zation have much lower pKa's and so are rejected by the
     charged membrane.
    
          The flux behavior of the mixture with the FT30 membrane was also found to be linear over
     the pressure range studied. FT30 membrane flux drops and rejections are shown in Figures 11
     and 12 for an operating pressure of 1.43 MPa. Flux drop is also a strong function of feed pH for
     this membrane, and feed pre-ozonation does improve flux drop over the range of feed pH's
     studied.  As  with the single component TCP system, ozonation produces the intermediates that
     do not interact with the membrane to same extent as the mixture and so flux is enhanced. TOC
     rejection slightly increases with ozonation and feed pH.
    
    
      Overall Removal of Trichlorophenol
    
          Figure 13 illustrates an example calculation of an ozonation - membrane process for the
     separation of hazardous organics; trichlorophenol is used as the model compound. After
     ozonation for 30 minutes the TCP concentration would be reduced to <0.2 mg/L; the TOC of the
     feed solution would be 66 mg/L due to the formation of carbon dioxide during the ozonation.
     For a membrane that rejects 90 % (either the NF40 or FT30), the permeate contains only 6.6
     mg/L TOC which would be due primarily to organic acids; the TCP concentration in the
    permeate would be less than 0.2 mg/L.  For 90 % water recovery, the concentrate TOC,  also
    mostly due to organic acids formed during ozonation, would be 1194 mg/L. This greatly
    reduced volume could be disposed of by incineration. The overall ozonation - membrane
    process would produce permeate water of high quality and greatly reduce the volume and risk of
    waste that must be further treated.  Also, since the process has been shown to be effective when
    utilizing charged membranes, selective separation of hazardous organics from feeds containing
                                             572
    

    -------
    high salt concentrations would be possible, allowing high fluxes at low operating pressures even
    for feeds with high osmotic pressures.
    
    
    CONCLUSIONS
    
          This study has shown that thin film, composite membranes can be used effectively for the
    separation of selected hazardous organic compounds. This waste treatment technique offers
    definite advantages in terms of high solute separations at low pressures (<1.4 MPa) and broad
    pH operating range, and the use of charged membrane would allow the selective separation of
    some organics from feeds containing high salt concentrations. Li addition, feed pre-ozonation of
    selected organics has been shown to improve flux and rejection characteristics for both charged
    and uncharged membranes due to the formation of ionizable organic acid intermediates during
    the ozonation that do not interact as strongly with the membrane. It has been shown that the
    ovejall ozonation - membrane process could be greatly effective in producing permeate water of
    high quality while minimizing the volume of waste that must be further treated.
    
    
    REFERENCES
    
     1.  "Industrial Hazardous Waste Management", Industry and Environment Special Issue - No.
          4(1983).
    
     2.  "Land Disposal of Hazardous Wastes", EPA Report, EPA-600/9-84-007 (1984).
    
     3.  Shuckrow, A.J., Pajak, A.P., and Osheka, J.W., "Concentration Technologies for
          Hazardous Aqueous Waste Treatment", EPA Report, EPA-600/2-81-019 (1981).
    
     4.  S.mith, J.K., "Laboratory Studies of Priority Pollutant Treatability", EPA Report,
          EPA-600/2-81-129 (1981).
    
     5.  Rice, R., and Browning, M., "Ozone for Industrial Water and Wastewater Treatment", EPA
          Report, EPA-600/2-80-060 (1980).
    
     6.  Baillod, C.R., Faith, B.M., and Masi, O., "Fate of Specific Pollutants During Wet
          Oxidation and Ozonation", Environmental Progress, 3, No. 1, pp. 217-227 (1982).
    
     7.  Bhattacharyya, D., Adams, R., and Williams, M., "Separation of Selected Organic and
          Inorganic Solutes by Low Pressure Reverse Osmosis Membranes" in "Synthetic and
          Biological Membranes", Alan R. Liss (1989).
    
     8.  Eriksson, Peter, "Nanofiltration Extends the Range of Membrane Filtration",
          Environmental Progress, 7, No. 1, pp. 58-62 (1988).
    
     9.  Sirkar, K.K., and Lloyd, D.R., "Membrane Materials and Processes for Separation", AIChE
          Symposium Series, 84. No. 261 (1988).
    
    10.  Sourirajan, S., and Matsuura, T., "Reverse Osmosis/Ultrafiltration Process Principles",
          National Research  Council Canada, Ottawa (1985).
    
    11.  Lee, E.K., "Membranes, Synthetic, Applications", Encyclopedia of Physical Science and
          Technology. 8, pp. 20-55 (1987).
                                              573
    

    -------
    12.  Belfort, G., "Synthetic Membrane Processes: Fundamentals and Water Application",
         Academic Press Inc., New York, New York (1984).
    
    13.  Chian, E.S.K., Bruce, W.N., and Fang, H.H.P., "Removal of Pesticides by Reverse
         Osmosis", Environmental Science and Technology. 9 (1975).
    
    14.  Light, W.G., "Chemistry and Water Reuse", 1, Ed. WJ. Cooper, Ann Arbor Science, Ann
         Arbor, Michigan (1981).
    
    15.  Shrem, M., and Lawson, T.,  Industrial Water Engineering, 14 (1977).
    
    16.  Bhattacharyya, D., and Madadi, R., "Separation of Phenolic Compounds by Low Pressure
         Composite Membranes: Mathematical Model and Experimental Results" in AIChE
         Symposium Series, 84> No.  261 (1988)
    
    17.  Bhattacharyya, D., Barranger, T., Jevtitch, M., and Greenleaf, S., "Separation of Dilute
         Hazardous Organics by Low Pressure Composite Membranes", EPA Report.
         EPA/600/S2-87/053 (1987)
    
    18.  Wang, Y., Pai, P., and Latchaw, J., "Effects of Preozonation on the Methanogenic Toxicity
         of 2.5-Dichlorophenol". Journal WPCF, 61, No. 3, pp. 320-326 (1989).
                                             574
    

    -------
                                                            8
                                                    CD
    
                                                    E
                                                   •c
                                                    CD
                                                    CL
                                                    X
                                                   UJ
    
                                                    CD
    
    
                                                    I
                                                   JD
    
                                                    CD
                                                   CD
                                                   CO
    
                                                   "o
                                                   CB
                                                   E
                                                   CD
    
                                                   O
                                                   CO
                                                   CD
    
                                                   3
                                                   O)
    575
    

    -------
    UL
    O
    X
    o
    H-
    O
    CO
    
    u.
    o
    
                                                                                          c
                                                                                          o
                                                                                          E
                                                                                          CL
                                                                                          X
                                                                                          LJJ
    
                                                                                          0>
    
    
                                                                                          I
                                                                                          J3
    
                                                                                          e
                                                                                          o
    
                                                                                          s
                                                                                           I
                                                                                          c
                                                                                          _o
    
                                                                                          "cO
                                                                                          c.
                                                                                          o
                                                                                          N
                                                                                          O
                                                                                          01
    
                                                                                          O
    
                                                                                          O
    
                                                                                          O
                                                                                          c
                                                                                          g
    
    
                                                                                          'c
                                                                                          o
                                                                                          N
                                                                                          O
                                                                            _o
                                                                            V-*
                                                                             CO
    
                                                                             E
                                                                             
    -------
      20
    
      IS
    
      IS
    
      14
    
    a.12
    2
    
    I10
    5  81
                                                                         10
                             System: NF40 Membrane
                                     50 me/L TCP
                                     DW = 10.8 x 10~4cm/!
                Membrane
                 FeedpH
                °3.3 - 3.6
                0 8.8-9.4
                  M""l""l""l"
     Figure
         0  5  10  15  20 25  30 35  40 45  50 55  60 65
                       Ozonation Time (min)
    
         3.   Effect of Ozonation on Flux Drop for the NF40
             Membrane with Trichlorophenol.
       100
                                System: NF40 Membrane
                                        50 mg/L TCP
        20 i
    
           0   5  10 15  20 25  30  35 40  45 50  55 80  65
                        Ozonation Time (min)
    
    Figure 4.   Effect of Ozonation on TOC Rejection for the NF40
               Membrane with Trichlorophenol.
                               577
    

    -------
                                                                       11
      30-
      25-
      20-
    o.
    2
    x
      15-
      10-
    System:  FT30 Membrane
            50 mg/L TCP
            DWF =10.5 x 10~4 cm/a
            PaTg=I.43MP.
            Temp. = 24°C
    
            Membrane
             Feed pH
            n 3.3-3.6
            0 a.a - 9.4
         inn [muni HI in|i mi" in ii "i"" i" "i ""i"" i1 "M"
         0  5  10  15 20  25 30  35 40  45  50  55  60 65
    
                       Ozonation Time (min)
    
    Figure 5.  Effect of Ozonation on Flux Drop for the FT30
               Membrane with TrichlorophenoL
    
    
    
    
    
    
    s
    75
    u
    «
    1
    t?
    
    
    
    
    100-
    95'
    
    on •
    ou
    85"
    
    80-
    
    75'
    
    70"
    65
    60
    55
    50
    
    3
    a °
    3 n n 	
    O
    a
    3
    
    
    
    Membrane
    Feed pH
    °3.3 - 3.8
    o 8.8 - 9.4
    System: FT30 Membrane
    SO mg/L TCP
    «vg~ _ *
    Temp. = 24°C
    111111
    
          0  5  10  15  20 25  30 35  40 45  50 55  60 65
    
                        Ozonation Time (min)
    
    Figure 6.  Effect of Ozonation on TOG Rejection for the FT30
               Membrane with Trichlorophenol.
                                573
    

    -------
                                                                       12
        13
    
        12-
    
        11-
     •yio-l
         8-
    
         7"
    
         8-
    
         5-
                                               o  a
              ° TCP/HA
              °0zon. TCP/HA (30 min.)
              System: NF40 Membrane
                      50 mg/L TCP - 10 mg/L HA
                      Feed pH = a. 1-9.3
                      Pavg = 1-3a MPa
                      Temp. = 24° C
             —i—'—i—'—i—-—i—•—i—•—i—•—i—.—i—•—r
          0   10   20   30   40   50   80   70   80   90
                            % Recovery
    
    Figure 7.  Flux as Function of Water Recovery for the NF40
              Membrane with Trichlorophenol/Humic Acid
              Mixtures.
      100
      90-
       80-
    
    s
    =  70-j
    o
    V
    as
    o  80-1
      50-
      40"
      30 i
             °No Ozon.
             A30 Min. Ozon
              System: NF40 Membrane
                      SO mg/L TCP - 10 mg/L HA
                      Feed pH 9.1-9.3
                      P   = 1.38 MPa
                      Temp. = 24°C
              10
                   —r—
                   20
    —r—
     30
    —r
     40
                                 50
                             Recovery
    —1—
     so
    —I—
     70
    —r-
     80
    —r
     90
    Figure 8.  TOG Rejection as Function of Water Recovery for the
               NF40 Membrane with Trichlorophenol/Humic Acid
               Mixtures.
                             579
    

    -------
                                                                       13
    40-
    
    
    35"
    
    
    30-
    
    
    25"
    2*20-1
    a
    x
    !15i
       ID-
        o-
      -5
                           System: NF40 Membrane
                                   50 m«/L CP, DCP, TCP
                                   100 mg/L TCE. TTCE
                                   DTTF = 11.2 x 10 ~4 cm/a
                                   P   =1.38MPa
                                   Temp. = 24° C
              n No Ozon.
              A60 Min. Ozon.
          234567
    
                             Feed pH
                                         S     9     10
    Figure 9.  Effect of Feed pH and Ozonation on Flux Drop for the
               NF40 Membrane with Hazardous Organic Mixtures.
    ioo-|
    
    
    
    
    _o
    "a
    0^
    IT
    a:
    1
    M
    
    
    
    
    
    
    
    .
    90-
    80-
    70'
    80-
    
    50-
    
    40-
    30-
    
    20-
    10-
    n-
    
    a No Ozon.
    A60 Min. Ozon..
    *
    S^
    /S
    .-^
    /
    ./ System: NF40 Membrane
    / SO m«A CP. DCP. TCP
    / 100 mg/L TCE, TTCE
    jf P^ = 1.38 MPa
    A Temp. = 24° C
    Q— 	 . 	 	 	
    	 a 	 a
    
    23456789 1C
    Feed pH
     Figure 10.  Effect of Feed pH and Ozonation on TOC Rejection
               for the NF40 Membrane with Hazardous Organic
               Mixtures.
                               580
    

    -------
                                                                         14
       50
    
       45
    
       40
    
       35
     2
     X
     3
     b
       20
    
       IB
    
       ID
    
        S
    
        o
                             System: FT30 Membrane
                                    50 mg/L CP. DCP, TCP
                                    100 mg/L TCE, TTCE
                                    DWF = 10.9 x 10 "*4 cm/a
                                          1-43 MPa
                                          = 84°C
               °No Ozon.
               A60 Min. Ozon.
                           587
    
                              Feed pH
                                                       10
    Figure 11.  Effect of Feed pH and Ozonation on Flux Drop for the
               FT30 Membrane with Hazardous Organic Mixtures.
       100
    
    
       95
    
    
       90-
    
    
    c  85-
    _o
    2
    o
    •S-  80-1
    a:
    u
    °f  75"
    
    
       70-
    
    
       65-
      60 i
              ° No Ozon.
              A60 Min. Ozon.
              T"
               3
                    nr
                     4
                            System: FT30 Membrane
                                   50 mg/L CP. DCP, TCP
                                   100 mg/L TCE, TTCE
                                   Pavg=sl-43MPa
                                   Temp. = 24° C
                          587
    
                             Peed pH
    T—i—r
     9     10
    Figure 12. Effect of Feed pH and Ozonation on TOC Rejection
               for the FT30 Membrane with Hazardous Organic
               Mixtures.
                               581
    

    -------
                                                                                        15
                                          CO
                                          ;g
                                          "o
                                          co
    
                                      •2  o  O
                                      co  7=  u-
    
                                      $  §  J
                                      cr  oj =1
                                      CD  ^r   OJ
                                      o  O   c
                                      cr       *-
                                      o  >» •*
                                      O  ~  en
                                          co  ^1
                          oo
    
                           it
                                                          o-S
    
                                                          O 'O        .
                                                          CD  CO      —j
    
                                                          "oT „,  o ^  °^
                                                          S .2  X OJ  CO
                                                           „  C  P d   II
                                                                     co
    
    
                                                                     Q.
                                                                     CD
    -3
    cr
    i
    c
    E
    C
    
    C
    
    A
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    c
    > £
    p-
    jj "?
    • t.
    > cc
    ) C£
    •
    
    CM
    )
    5
    
    k
    
    
    
    c
    0
    OS
    cr
    C3
    N
    O
    
    A
    
    
    
    
    ) "J
    ) ^
    | /1^
    * s
    •S f^
    it f~
    • ^"»
    : 3
    IS
    
    
    
    
    
    
    
    ^^^
    E
    
    o
    CO-
    
    k
    
    
    
    
    _J
    OJ -^ CO
    E ^"^ 3
    v '•a'
    v^ Q_ CO
    f- S i.
    ^- ^-
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    a
    c
    0
    i.
    j:
    £
    0
    •s
    ^
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    3
    Z
    3
    ^
    2
    "•
    j
    >
    «
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    >.
    CD
    O
    0
    CD
    DC
    rtl
    "cO
    ^o
    °
    in
    C33
    
    
    
    
    
    
    
    
    
    
    
    
    
    O OJ _j 11 OJ CO
    Q 0 "OJ CD = ^
    E E O E
    ^fc^ •• jjj* ^C 3«*
    S ~ CO — P ^
    tS co" "2 ^~ ^^.
    <5 Q. "E1
    D_ — o
    *^ CO
    ^> _ >—
    _ CO o5 ^
    5j "*S CO o^
    *- CJJ x-T4 Is-;
    **" S OT o ^
    o ^ A ^ C3J
    04 *- II II "
    CO 	
    II > >
    -• 1 1
    
    ^"
    ^ B
    •^ 1-
    co =
    *- CO
    CD >-
    > CD
    0 5
                                                                                        CO
                                                                                        CO
                                                                                        CD
                                                                                        o
                                                                                        o
    
                                                                                       al
    
                                                                                        CD
                                                                                        CD
                                                                                        i
                                                                                        cr
                                                                                        o
                                                                                        N
                                                                                       O
                                                                                        a
                                                                                        o
                                                                                        co
                                                                                       O
    
                                                                                       _CD
                                                                                        a.
    
                                                                                        co
                                                                                        x
                                                                                       LU
                                                                                        CO
    
    
                                                                                        CD
    
                                                                                        3
                                                                                        O)
    
                                                                                        ii.
    T3
     CD
     CD
    LL.
    0
    o
     0)
     £
    •r-   a  °
    
    
     CD    "   O)
     E   £   E
    •i   -i  O
    o   >   co
                                        582
    

    -------
                     Testing of a Leachate  Treatment System
                        Based on a Wood Degrading Fungus
    
                    by:  John A. Glaser, Henry Tabak, and Edward Opatken
                         United States Environmental Protection Agency
                         Risk Reduction Engineering Laboratory
                         26 Martin Luther King Drive
                         Cincinnati, Ohio 45268
    
                         Susan Strohofer, Carol Hummel, Marge Kupferle,
                         P.V. Scarpino, and M. Wilson Tabor
                         University of Cincinnati
                         Cincinnati, Ohio 45265
                                    Abstract
    
         The fungus, Phanerochaete chrysosporium, has been employed in the
    design of a rotating disk reactor for liquid treatment.  Early studies
    decolorizing Kraft waste liquor were very promising, and subsequent
    applications to individual hazardous waste constituents such as
    pentachlorophenol and trinitrotoluene were equally impressive.  The latter
    application to hazardous waste has become the subject of an intensive
    study to scale-up the early bench-scale reactors.  The research program
    and status will be presented with details of projected field activities.
                                        583
    

    -------
                       ASSESSMENT OF KPEG TREATMENT FOR PCB
                                CONTAMINATED SOILS
                    by:  Alfred Kernel
                         U.S. Environmental Protection Agency
                         Risk Reduction Engineering Laboratory
                         Cincinnati, OH
                                     ABSTRACT
    
          The novel reagent generically named APEG (alkaline polyethylene gly-
    colate) can be very effective for dehalogenation of a variety of halo aro-
    matic pollutants.  The PCBs, PCDDs, PCDFs even PCTs (polychlorinated
    terphenyls) can be effectively dehalogenated by APEG reagents to yield non-
    toxic products.  The reagent has been shown to be effective on these
    classes of pollutants in a variety of matrices from sediment to soils and
    waste oils.
    
          Early work in 1986 with KPEG performed on PCDD and PCDF contaminated
    oil (9000 gal.) in Butte, Montana demonstrated the effectiveness of the
    KPEG system on these types of hazardous wastes.  Early tests of the in situ
    application of KPEG on dioxin contaminated soil in Missouri demonstrated
    some limitations of the reagents applicability.  The most severe drawback
    for in situ KPEG application is the reagents extreme hygroscopisity.  These
    early experiences led to the design of the pilot-scale chemical reactor
    system for the effective use of the APEG reagent.  This reactor system was
    first demonstrated on a heavily PCB-contaminated site in Guam, USA during
    April-May 1988.  This pilot-scale unit was capable of treating from one to
    two tons of contaminated soil per batch.  Further refinements were made to
    the reactor system after careful examination and analysis of the first
    eight reactor batches.  During September-October of 1988 these refinements
    were employed in a second series of KPEG treatments in Guam.  Contaminated
    soil containing from 500 to 2,600 ppm of PCB closely resembling Aroclor
    1260 was treated with the KPEG reagent.  The only residual PCB peak
    detected in the soil treatments was a tetrachlorobiphenyl, ranging from
    approximately 1 ppm to nondetectable levels.  The KPEG reagent was also
    applied to treating all of the contaminted tyvek clothing, gloves and boots
    from the combined runs during the 1988 year.  Sampling fo the reactor con-
    tents revealed the presence of tetra-, penta-, and hexa- chlorobiphenyls
    but at below the one part per million range.  These results demonstrate the
    capability of the KPEG reagent to perfrom chemical dehalogenation on
    haloaromatics in a variety of matrices, ranging from oils to soils and
    contaminated clothing.
                                           584
    

    -------
                        STATE-OF-THE-ART FIELD HYDRAULIC
                     CONDUCTIVITY TESTING OF COMPACTED  SOILS
    
                  by:  Joseph 0. Sal and David C. Anderson
                       K. W. Brown & Associates, Inc.
                       College Station, TX  77840
                                    ABSTRACT
        The Congressionally mandated performance standard for soil liners of
    hazardous waste management facilities is a hydraulic conductivity of 1 x
    10.  m/s or less.  In response to this statutory requirement, the U.S.
    Environmental Protection Agency has issued guidance .requiring that facil-
    ities demonstrate this hydraulic conductivity in field tests.
    
        Hydraulic conductivity test methods currently used on soil liners
    were evaluated for their ability to meet the minimum requirements for
    field tests; i.e., that the test be capable of measuring hydraulic conduc-
    tivities as low as 1 x 10   m/s and that the values obtained be represen-
    tative of the overall soil liner.  Of the few methods capable of meeting
    the minimum requirements, even fewer are both practical to use and rarely
    give false low values*  Considering all the advantages of the methods
    evaluated, the best, most practical, and currently available technologies
    for -evaluating hydraulic conductivity are represented by large single-
    ring infiltrometers and sealed double-ring infiltrometers.  If correction
    factors are ngeded to bring the values obtained with single-ring devices
    below 1 x 10   m/s, confirmatory tests should be conducted with sealed
    double-ring infiltrometers.
    
        A long-term study is needed to allow a comparative evaluation of
    candidate hydraulic conductivity testing devices.  A large underdrain
    should be incorporated into the study to give the true overall hydraulic
    conductivity value with which other values should be compared.
                                       585
    

    -------
       REVIEW OF SOIL HASHING TECHNOLOGIES FOR SOILS  CONTAMINATED  WITH  HEAVY  METALS
    by:  Carl Gutterman
         Ramjee Raghavan
         Foster Wheeler Enviresponse, Inc.
         Edison, NJ  08837
                                                     Darlene Williams
                                                     Releases Control Branch, USEPA
                                                     Edison, NJ  08837
                                        ABSTRACT
    
        Heavy metals are estimated to contaminate forty percent of the Superfund
    sites on the National Priorities List (NPL).  Studies have shown that eight
    metals (lead, chromium, arsenic, cadmium, copper, zinc, mercury, and nickel)
    predominate at these sites.  These facts have led to a review of the current
    status of extraction technologies employed on excavated soils.
    
        The most advanced extraction technologies are identified and their basic
    processing steps discussed.  A review is presented of the established
    industrial practices incorporated within the technologies for cleaning soils,
    sludges, and sediments.  The results reported by the developers and operators
    of facilities employing these technologies are discussed, and questions to be
    explored with respect to application are addressed.  Recommendations are
    given as to the further development of extraction technologies for excavated
    soils.
                                         586
    

    -------
                              ROTARY KILN INCINERATION
    
                        by:    V.A. Cundy, T.W. Lester, S. Acharya, C. Leger,
                                     A.Montestruc,G. Miller
                              Department of Mechanical Engineering
                              Louisiana State University
                              Baton Rouge, Louisiana 70803
    
                              A.M. Sterling
                              Department of Chemical Engineering
                              Louisiana State University
                              Baton Rouge, Louisiana 70803
    
                              J.S. Lighty, D.W. Pershing, G.D. Silcox
                              Department of Chemical Engineering
                              University of Utah
                              Salt Lake City, Utah 84112
    
                              W.D. Owens
                              Department of Mechanical Engineering
                              University of Utah
                              Salt Lake City, Utah 84112
                                           ABSTRACT
    
    
           A multifaceted experimental and theoretical program aimed at understanding rotary kiln
    incinerator performance is underway.  The program involves university, industry, and
    government participation. This poster session summarizes work accomplished to date in the
    areas of full-scale insitu sampling from an operating rotary kiln facility, kiln-simulator
    experimentation, and numerical modeling.
    
    
           Full-scale insitu measurements are obtained from the Louisiana Division rotary kiln
    facility of Dow Chemical USA, located in Plaquemine, Louisiana. Summary results obtained
    from experiments that were performed during continuous processing of carbon tetrachloride
    and preliminary results  obtained during the batch mode processing of toluene-laden sorbent
    packs are presented.  In situ measurements from a rotary kiln incinerator during steady and
    transient loading are helping to clarify the local behavior inside an operating kiln. The infor-
    mation is needed to develop reasonable mathematical models of the process for use in inciner-
                                          587
    

    -------
    ator evaluation and incinerator design. Ultimately, the model may be developed to the degree
    that knowledge of the flame-mode chemistry is combined with knowledge of the transport rates
    throughout the kiln to tailor kiln design and operation achieving optimum destruction per-
    formance.  Results to date demonstrate the existence of significant spatial and temporal
    variations in temperature and species concentrations in the kiln even under quasi-steady oper-
    ating conditions.  The addition of turbulence air (to promote mixing) reduces, but does not
    eliminate, these variations. A consequence of the enhanced mixing, however, appears to be a
    reduction in processing rates in the upper regions of the kiln and an increase in the demand on
    the afterburner. Data obtained from the kiln/afterburner/stack train are presented to demonstrate
    these trends.
    
           Kiln simulator work conducted during this study has aided in providing an under-
    standing of the results observed at the full-scale. An objective of the laboratory kiln simulation
    work has been to determine bed evolution rates. Results clearly demonstrate the utility of the
    kiln-simulator in helping to understand processes occurring at the full-scale. The importance of
    kiln rotation rate (hence, bed motion) on the evolution characteristics of the studied sorbent-
    contaminant  matrix has been demonstrated.  The intermittence in contaminant evolution
    observed at the full-scale has been, at least in part, explained through careful interpretation at
    the kiln-simulator scale. The results also demonstrate that careful experimental design, ex-
    ecution of experiments and interpretation of results must be employed to yield meaningful
    information from the kiln-simulator. Various species and temperature data obtained from the
    kiln-simulator are presented to demonstrate these trends as well.
    
           The poster session also involves a presentation of the modeling efforts that have been
    undertaken to predict rotary kiln performance.  Work initiated with two-dimensional modeling
    efforts. Tliree-dimensional models that, although considerably more complicated, are required
    to obtain even qualitatively correct predictions. Realistic bed evolution models and
    fundamental chemical kinetic parameters that will be used in the three-dimensional models are
    also under development  Various results from these components of the modeling effort are
    shown.
                                           588
    

    -------
    BIOLOGICAL TREATMENT OF CHLOROPHENOL-CONTAMINATED 6ROUNDWATER
    
                       by:  Thomas J. Chresand
                            BioTrol, Inc.
                            Chaska, Minnesota  55318
                               ABSTRACT
    
         BioTrol, Inc. has developed a system for the treatment of
    chlorinated phenols  in wastewater streams.  The  phenols are
    degraded  by  microorganisms  which   are  immobilized  in  a
    submerged, fixed-film bioreactor.  An  indigenous consortium of
    microorganisms  is augmented  by  inoculation  of  a  specific
    bacterium with capability to degrade penta chlorophenol  (PCP)
    as  well  as  tetra-  and  tri-chlorophenol.     BioTrol  has
    demonstrated treatment  of wastewater  containing up  to 90 ppm
    PCP.  Removals from 95 to 99 percent were achieved with a one
    hour  hydraulic  residence  time.    The  system is  primarily
    applicable to treatment of groundwater; however,  treatment of
    process and lagoon waters has also been demonstrated.
                                 589
    

    -------
             COMPUTERIZED MANAGEMENT AND DISSEMINATION
            OF INFORMATION FOR RESEARCH AND DEVELOPMENT
              OPERATIONS AT THE TECHNICAL INFORMATION
                     EXCHANGE (TIX) EDISON. NJ
            May Smith,  Francine Everson,  and Pacita Tibay
                         Enviresponse, Inc.
                 6SA Raritan Depot, Woodbridge Ave.
                          Edison,  NJ  08837
                            Hugh Masters
                       Releases  Control  Branch
                Risk Reduction Engineering Laboratory
                U.S. Environmental Protection Agency
                      Edison, New Jersey 08837
                              ABSTRACT
        The Technical Information Exchange (TIX) at EPA's Edison,
    New Jersey Facility was established in 1980 to  support the
    research personnel at this location.  It  is operated  by a
    Research and Development  (R&D) on-site contractor, currently
    Enviresponse, Inc., for EPA's Releases Control Branch (RCB) of
    the Risk Reduction Engineering Laboratory  (RREL).
    
        TIX provides a wide range of support services for RREL
    research activities being conducted at the Edison Facility,
    including the preparation of  slides; selection  of slides,
    movies and videotapes from the extensive  library developed by
    
                                    1
                                 590
    

    -------
    RGB over the years for presentations; and dissemination of
    descriptive literature.  TIX provides immediate  access to its
    collection and assists  in locating or procuring materials from
    other sources (e.g., EPA's Center for Environmental Research
    Information, other Agency  information sources,  the National
    Technical Information Service, etc.).  A library  of commercial
    software is maintained  so users  may  evaluate various packages
    before purchasing additional copies.
    
        TIX also pursues research studies of its own into the
    computerized retrieval  and dissemination of this  information.
    It developed and maintains the computerized On-Line
    Information System (COLIS) designed to meet the needs of
    several specific R&D efforts. COLIS searches  by key words to
    retrieve files for display.  While COLIS is expandable to
    other needs as they arise, the intent is not to compete with
    broad-based, commercial types of information databases.
    Currently, while not all tasks are equally advanced in their
    development, COLIS is designed to service program efforts.
    
        These include a Case History File of hazardous materials
    spills, remedial  and removal  actions for Superfund sites, and
    corrective actions for underground storage tank problems; a
    SITE Applications Analysis Reports program containing
    performance and cost data from Superfund  Innovative Technology
    Evaluation  (SITE) Program  projects;  a Library Search System
    that provides descriptive records and abstracts  for documents
    on RREL research  topics in  the TIX files, including a special
    collection of Stormwater Research publications containing EPA
    reports, technical papers, etc., developed by the Storm and
    Combined Sewer Overflow Program during 1967-1982.
    
        In addition to a modest operating budget,  TIX is funded
    from other R&D projects to support their special needs.  To
    meet growing demands for technology  transfer of  RREL research
    at Edison, TIX recently expanded its workshop/conference
    capabilities with the addition of three modular units.
                                591
    

    -------
                        The EPA Treatability Database
                  by:  Stepahanie A. Hansen,
                       Radian Corporation
                       Milwaukee, ¥1   53214
    
                       Kenneth A. Dostal, and
                       Glenn Shaul
                       Environmental Protection Agency
                       Cincinnati, OH   45268
                                   ABSTRACT
    
         The purpose of this paper will be to introduce to the public the
    EPA Treatability Database.  Since the mid-1970's, EPA has accumulated
    a wealth of data on the treatability of compounds found in industrial
    and domestic wastewaters, groundwater, leachates, ponds, lagoons and
    other surface waters.  To date, various attempts have been made to
    organize selected segments of information, but a comprehensive
    evaluation of peer reviewed information encompassing compounds
    regulated by all pertinent environmental laws has never previously
    been undertaken.  This major research activity was initiated with the
    overall objective of providing a database on the treatability of
    priority pollutants and other hazardous compounds in water and/or
    wastewater.  The database summarizes years of studies on the treat-
    ability of compounds regulated under the Clean Water Act, Safe
    Drinking Water Act, Resource Conservation and Recovery Act, Toxic
    Substances Control Act, Superfund Amendments and Reauthorization Act
    and the National Priorities List.
    
         The database contains such compound specific information as
    physical-chemical properties (formula, melting point, boiling point,
    molecular weight, etc.), reference to environmental data (health
    advisories, water quality criteria, etc.), carbon isotherm
    (adsorption) data, as well as actual treatability data.  The treat-
    ability data summary tables present the percent contaminant removal
    achieved, effluent concentration for a compound, treatment technology
    (9 biological and 8 physical-chemical technologies were evaluated),
    treatment scale (bench, pilot or full-scale) and water matrix (12
    matrices were studied).
    
         The database has been computerized and can be accessed using a
    personal computer.  The program for the database was developed using
    dBase III PLUS programming and compiled with Clipper.  Thus,  no
    special language or software is needed to run the program.   The
                                         592
    

    -------
    .program is menu driven so  that persons with minimal computer experience would
    be able to access any information provided in  the database.
    
         This paper will highlight important aspects of the database and provide
    an overview of program availability and use by the public.
                                         593
    

    -------
    CHEMICAL TREATMENT OF METALS IN WASTEWATERS AND
    SLUDGES AT THE T&E FACILITY
    
    Douglas W. Grosse
    U.S. EPA, RREL
    Cincinnati, OH
    
    Sardar Q. Hassan and Mark J. Briggs
    University of Cincinnati
    Cincinnati, OH
    Hazardous waste treatment  technology  studies  are currently being
    conducted at  the  U.S.  EPA'S Test and  Evaluation (T&E)  facility.
    As  a part of this program,  alternative technologies  are being
    evaluated for the  treatment of hazardous wastes containing heavy
    metals and cyanide.  Both  bench-and pilot-scale research studies
    are being performed to evaluate the removal of cyanide from spent
    electroplating solutions and residues.  Findings to date have shown
    that it  is  important to destroy all  forms  of cyanide (amenable,
    non-amenable, and  complexed)  in cyanide/metal bearing wastewater
    prior to metals precipitation.  Otherwise, significant quantities
    of cyanide (complexed)  can be precipitated with heavy metals into
    a hydroxide sludge.  Although conventional alkaline chlorination
    is  a widely  employed technology used  to  destroy amenable (free)
    cyanide  in  solution,   often  strongly complexed  (non-amenable)
    cyanides  remain in  solution  impervious  to  chemical  oxidation.
    Several processes have shown potential for breaking these complexes
    (e.g.,ferricyanide)     and    destroying     the    cyanide.
    Ozonation/photolysis (UV light)  is one of them.  Bench studies will
    be  conducted  evaluating this technology.  In the area  of metals
    treatment, work will be performed which will evaluate the treatment
    of  an  EDTA-lead complexed  wash generated from  the treatment  of
    lead-contaminated  soils.   Additional  studies will  focus  on the
    treatment of chromate wastes resulting from tanning, wood-treating,
    and  ink  manufacturing  operations.   Results generated  from these
    studies will be presented.
                                   594
    

    -------
             BIOLOGICAL DEGRADATION OF CHLORINATED PHENOXY ACIDS
    
    by:  R. A. Haugland, D. Schlemm, R. Lyons and A. M. Chakrabarty
         Department of Microbiology & Immunology
         University of Illinois at Chicago
         Chicago, Illinois
    
         P. R. Sferra
         Risk Reduction Engineering Laboratory
         U.S. Environmental Protection Agency
         Cincinnati, OH 45268
                                   ABSTRACT
    
         Previous studies have shown that Pseudomonas cepacia strain
    AC1100 and Alcaligenes eutrophus strain JMP134 metabolize the chloro-
    phenoxyacetate herbicides 2,4,5-T and 2,4-D. respectively.  Our
    results indicate, however, that both of these activities are inhibited
    when co-cultures of the two organisms are confronted with mixtures of
    2,4,5-T and 2,4-D.  The basis for this inhibition may be related to
    the physical separation of the 2,4,5-T degradative activity in AC1100
    cells from the 2,4-D degradative activity in JMP134 cells since we
    have shown that the effect can be largely eliminated by transfer of
    the 2,4-D degradative genes via the plasmid PJP4 from JMP134 to AC1100.
    Results of studies on the biochemical and physiological factors respon-
    sible for this inhibition will be presented.
                                      595
    

    -------
               USE  OF  FOAM TECHNOLOGY  FOR  CONTROL OF TOXIC  FUMES DURING
                            EXCAVATION AT SUPERFUND SITES
         by: Ramjee Raghavan
             Patricia Brown
             Foster Wheeler Enviresponse, Inc.
             Edison, NJ  08837
    Mark Torpey
    Foster Wheeler Development Corp.
    Livingston, NJ  07039
                                John E. Brugger
                                Releases Control Branch, USEPA
                                Edison, NJ  08837
    
                                       ABSTRACT
        This poster depicts the development of "foam scrubbing" as a technique
    for minimizing hazardous gas and/or airborne particulate releases during
    cleanup operations at Superfund sites.
    
        In this application, foam is used to 1) incorporate and neutralize
    vapor phase emissions and 2) capture particulates.  A foam concentrate that
    contains neutralizing agents is converted to a foam by using the vapor and
    particulate-contaminated air to actually generate the foam; the foam is not
    preformed for use as a blanket.  Once the foam bubbles are formed, a large
    interior liquid surface area is available for gas, vapor, or particulate
    interaction.  A neutralization agent present in bubble walls reacts with
    the absorbed gas or vapor to render it innocuous.  The foam can then be
    collapsed and recycled with makeup reagents.
                                        596
    

    -------
    SOIL WASHING-REMOVAL OF SEMIVOLATILE ORGANICS USING AQUEOUS SURFACTANT
    SOLUTIONS
        by:    Edward Coles
               Ramjee Raghavan
               Foster Wheeler Enviresponse, Inc.
               Edison, NJ  08837
    Darlene Williams
    Releases Control Branch, USEPA
    Edison, NJ  08837
                                        ABSTRACT
    
        Semivolatile organic chemicals from soil (size range -20+200 mesh) may be
    removed using aqueous surfactant solutions as the extracting medium and high
    mass transfer rate ultrasonic cavitational excitation for liquid-solid
    contacting.  Tests were performed to confirm performance of the critical
    processing steps: 1) soil deagglomeration and clay separation; 2) ultrasonic
    cavitational extraction; 3) philic to phobic conversion of spent wash
    solution; and 4) hydrophobic organic removal.
                                         597
    

    -------
                      PREPARATION FOR SITE DEMONSTRATION OF A
                  POWDERED ACTIVATED CARBON TREATMENT (PACT)  UNIT
                 by:  John F. Martin
                      Risk Reduction Engineering Laboratory
                      U.S. Environmental Protection Agency
                      Cincinnati, OH  45268
                                     ABSTRACT
    
         As part of the Superfund Innovative Technology Evaluation (SITE)'
    program, Zimpro/Passavant Inc. will work with EPA to conduct a field
    demonstration of its mobile Powdered Activated Carbon Treatment (PACT)
    unit.  The project is designed to produce data on the feasibility and
    cost-effectiveness of using the process for Superfund site remediation.
    
        The PACT process is a biological treatment system which incorporates
    physical adsorption using powdered carbon.  It is particularly well suited
    to the treatment of combined industrial/municipal wastes.  In the PACT
    process,  powdered activated carbon is added to the aeration basin of a
    standard activated sludge system at a dose rate that varies considerably
    depending on the biodegradability and adsorptivity of the waste
    components.  Carbon dosages may range from a low of 10 mg/1 to more than
    1000 mg/1.
    
        During operation of the PACT unit, excess solids (biomass and carbon)
    are removed from the system by wasting a portion of the clarifier
    underflow.  The SITE demonstration of the PACT system will employ wet air
    oxidation (WAO) to  regenerate carbon and destroy organic contaminants
    carried with the  solids.
    
      .  The Zimpro demonstration will be performed at a Superfund site near
    Newark, New Jersey.  The 15-acre site had been used for production of
    alkyl resin carriers for paint and varnish products as well as for
    reprocessing off-specification resins from other suppliers.  Contamination
    at the site includes volatile and semi volatile organics, pesticides,
    polychlorinated biphenyls, and metals.  The demonstration will concentrate
    on treatment of the shallow aquifer (2-6 feet below the surface) which
    contains mostly organic solvents.
    
        Field work is scheduled for start-up in the summer of 1989.  During
    the spring of 1989, personnel from Zimpro/Passavant will be conducting
    treatability studies on wastewater collected from the demonstration site.
    These data will enable design of the treatment protocol for the full-scale
    field demonstration.
                                        598
    ftU.S.GOVERNMENT PRINTING OFF I CE t t 990-748-1 5 9/00404
    

    -------
    

    -------
    

    -------
    

    -------
    

    -------
    

    -------
    

    -------
    

    -------
    03 O
    £ S
    
    S1™
    ^|
    
    =•3
    
    18
    00
    0>
    CO
    o
    o
    rn
    PA/600/9-
    <£>
    0
    
    o
    o
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    
    If you do not wis
    detach, or copy 1
    upper left-hand
    o 3- 3"
    5 w o
    m ° ~*
    r 1 i
    
    QJ CD
    D —
    
    
    21 O
    3 5
    |E
    
    3- w
    CD O
    Q, X
    r Q. m
    a. o
    3 ^
    w -f-
    w m
    3 33
    — m
    
    Please make all
    detach or copy,
    left-hand corner
    3
    Q_ <0
    -i CD
    as
    C Q)
    3 •<
    o 3"
    <-* Q)
    IT 3
    CD (Q
    Q. M
    CL O
    3 3
    « s:
    ^ffl
    
    
    
    T3 S"
    m ^*
    "~
                    > m
                    
    -------