EPA/600/9-87/015
July 1987
LAND DISPOSAL, REMEDIAL ACTION, INCINERATION
AND TREATMENT OF HAZARDOUS WASTE
Proceedings of the Thirteenth Annual Research Symposium
at Cincinnati, Ohio, May 6-8, 1987
Sponsored by the U.S*. EPA, Office of Research & Development
Hazardous Waste Engineering Research Laboratory
' x Cincinnati, OH 45268
Edison, NJ 08837
Coordinated by:
JACA Corp;
Fort Washington, PA 19034
Contract No. 68-03-3258
Project Officers:
Naomi P. Berkley
John F. Martin
Cincinnati, OH 45268
HAZARDOUS WASTE ENGINEERING RESEARCH 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.
-IT-
-------
FOREWORD
As hazardous waste continues to be one of the more prominent
environmental concerns to the people of the United States and other
countries throughout the world, there are continuing needs for research to
characterize problems, and develop and evaluate alternatives to addressing
those problems. The programs of the Hazardous Waste Engineering Research
Laboratory (HWERL) are designed to contribute to satisfying these research
needs.
These Proceedings from the 1987 Symposium provide the results of
projects recently completed by HWERL and current information on other
projects presently underway. Those wishing additional information on these
projects are urged to contact the author or the EPA Project Officer.
Thomas R. Hauser, Director
Hazardous Waste Engineering
Research Laboratory
-m-
-------
ABSTRACT
The Thirteenth Annual Research Symposium on Land Disposal, Remedial
Action, Incineration and Treatment of Hazardous Waste was held in
Cincinnati, Ohio, May 6 through May 8, 1987. The purpose of this Symposium
was to present the latest significant research findings of ongoing and
recently completed projects funded by the Hazardous Waste Engineering
Research Laboratory (HWERL) to persons concerned with hazardous waste mana-
gement.
These Proceedings are for Session A, Hazardous Waste Land Disposal;
Session B, Hazardous Waste Incineration and Treatment; and Session C, HWERL
Posters. Papers presented by Symposium speakers and poster presentation
abstracts are compiled. Land disposal subjects discussed include remedial
action treatment and control technologies for soil and water, landfill
liner and cover systems, geotechnical aspects of earthen barriers, leachate
composition and migration, underground storage tanks and emergency
response. Incineration and treatment subjects include thermal destruction
of hazardous wastes, field evaluations of treatment methods, control of
volatile emissions, waste minimization and emerging physical, chemical and
biological processes for hazardous waste destruction.
-iv-
-------
CONTENTS
SESSION A - HAZARDOUS WASTE LAND DISPOSAL PAPERS
Technical Resource Documents and Technical Handbooks for Hazardous
Wastes Management
Norbert B. Schomaker, U.S. Environmental Protection Agency 1
Implications of Current Soil Liner Permeability Research Results
Walter E. Grube, Jr., U.S. Environmental Protection Agency ..... 9
The Behavior and Assimilation of Organic and Inorganic Priority
Pollutants Codisposed with Municipal Refuse - A Progress Report
Frederick G. Pohland, Georgia Institute of Technology 26
Field Verification of FMLS — Assessment of an Uncovered
Unreinforced 60-mil EPDM Liner after 18 Years of Exposure
Henry E. Haxo, Jr., Matrecon, Inc 38
Geosynthic Design Considerations for Double Liner Systems
Gregory N. Richardson, Soil & Material Engineers, Inc 51
Inspection Procedures/Criteria for Installation of Flexible
Membrane Liners
William M. Hel-d^SCS Engineers ...... . . 59
An Assessment of Materials that Interfere with
Stabilisation/Solidification Processes
R. Mark Bricka, U.S. Army Engineer Waterways Experiment
Station. 64
Mine ,Waste/Overburden Analytical Techniques - Characterization
and Simulation of Mine Tailings Weathering Environments
Frank T. Caruccio, University of South Carolina . 72
The (Effects of Overburden Pressure and Hydraulic Gradient on the
Performance of Model Soil-Bentonite Slurry Cutoff Walls
Richard M. McCandless, University of Cincinnati. . 80
i • • - • ' ' ' . -
Expert Systems to Assist in Decisions Concerning Land Disposal
of ^azardous Wastes
Daniel G. Greathouse, U.S. Environmental Protection Agency 89
Modeling Soil Water Movement in Minimum Technology Waste
Management Facilities
Dfavid H. Gancarz, Radian Corporation 97
Remediation of an Industrial Dump Site - A Case History, Part II
David S. Kosson, Rutgers University 105
-v-
-------
age
SESSION A (Continued)
Capillarity and Anlsotropy Effects on Ground-Water Flow to
Excavation
Forest 0. Mlxon, Research Triangle Institute 113
Pathways for the Removal of Volatile Organics from Surface
Impoundments
Crowley Clark Allen, Research Triangle Institute 121
Composition of Leachates from Actual Hazardous Waste Sites
Glenn D. McNabb, Science Applications International Corp. 130
Decontamination Techniques for Mobile Response Equipment Used at
Waste Sites
Mary K. Stlnson, U.S. Environmental Protection Agency 139
Leak Prevention in Underground Storage Tanks: A State-of-the-art
Survey
A.C. Gangadharan, Enviresponse, Inc 149
A Preliminary Analysis of Underground Tanks Used for CERCLA
Chemical Storage
Ihor Lysyj, Environmental Monitoring and Services, Inc 156
U.S. EPA Evaluation of Volumetric UST Leak Detection Methods
James W. Starr, Enviresponse, Inc 164
NATO/CCMS Pilot Study on Demonstration of Remedial Action
Technologies for Contaminated Land and Groundwater
Donald E. Sanning, U.S. Environmental Protection Agency . . 172
Reactivity of Various Grouts to Hazardous Wastes and Leachates
Andrew Bodocsl, University of Cincinnati 184
i
Electro-Decontamination of Chrome-Contaminated Soils
Sunlrmal Banerjee, University of Washington. ..... 193
Current Status of the Designation' and Adjustment of CERCLA
Hazardous Substances and their Associated Reportable Quantities
K. Jack Kooyoomjian, U.S. Environmental Protection Agency 201
The EPA Personnel Protection Technology Research Program
Michael D. Royer, U.S. Environmental Protection Agency 210
Application Opportunities for Canine Olfaction: Equipment
Decontamination and Leaking Tanks •
Herbert S. Skovronek, New Jersey Institute of Technology 217
Nondestructive Testing for Location of Containers Buried in
Soil
Robert M. Koerner, Drexel University 224
-vi-
-------
SESSION B - HAZARDOUS WASTE INCINERATION AND TREATMENT PAPERS
Page
Thermodynamic Analysis of Post-Flame Reactions Applied to Waste
Combustion
Daniel P.Y. Chang, University of California, Davis ......... 235
Influence of Atomization Parameters on Droplet Stream Trajectory
and Incineration
James A. Mulholland, U.S. Environmental.Protection Agency. ..... 246
Distribution of Volatile Trace Elements in Emissions and Residuals
from Pilot-Scale Liquid Injection Incineration
Johannes W. Lee, Acurex Corporation 254
Assessment of Residues from Incineration of RCRA Wastes
Joan V. Boegel, Metcalf & Eddy* Inc. . 262
Waste Characterization and the Generation of Transient Puffs in a
Rotary Kiln Incinerator Simulator
William P. Linak, U.S. Environmental Protection Agency ....... 283
On-Line Monitoring of Organic Emissions with a Mobile Laboratory
Sharon L. Nolen, U.S. Environmental Protection Agency. . 297
Total Mass Emissions from a Hazardous Waste Incinerator
Andrew R. Trenholm, Midwest Research Institute ........... 304
Incineration of Cleanup Residues from the Bridgeport Rental and
Oil Services Superfund Site
Larry W. Waterland, Acurex Corporation . . ,. 318
Pilot-Scale Testing of Nonsteady Boiler Waste Cofiring
Howard B. Mason, Acurex Corporation. 326
Technical/Economic Assessment of Selected PCB Decontamination
Processes
Ben H. Carpenter, Research Triangle Institute. . ' 332
Mobile KPEG Destruction Unit for PCBs, Dioxins and Furans in
Contaminated Waste
Charles J. Rogers, U.S. Environmental Protection Agency 361
Supercritical Solvent Extraction
Charles A. Eckert, University of Illinois.
366
Supercritical Fluid Extraction from Catalytic Oxidation of Toxic
Organics from Soils
F. Carl Knopf, Louisiana State University 373
Microbial Degradation of Synthetic Chlorinated Compounds
Richard A. Haugland, University of Illinois at Chicago 388
-vri-
-------
SESSION B (Continued)
Bacterial Oxidation of Polychlorinated Biphenyls
Louis M. Nadim, The University of Texas at Austin 395
Engineering P450 Genes in Yeast
John C. Loper, University of Cincinnati College of Medicine 403
Blodegradation of Organopollutants by Phanerochaete Chrysbsporium:
Practical Considerations
John A. Bumpus, Michigan State University 411
Growth of the White-Rot Fungus Ptianerochaete Chrysosporlum in Soil
Richard T. Lamar, USDA Forest Products Laboratory 419
Biological Treatment of Selected Aqueous Organic Hazardous Wastes
Richard J. Lesiecki, University of Cincinnati. .... 425
Assessment of Alternative Technologies for Treating Spent
Electroplating Solutions and Sludges
Catherine Drlscoll, Metcalf & Eddy, Inc 431
Solvent Recovery Technologies
Robert A. Olexsey, U.S. Environmental Protection Agency 444
Evaluation of Hazardous Waste Recycling Processes in the Printed
Circuit Board Industry
Thomas J. Nunno, Alliance Technologies Corporation 452
The California Innovative Alternative Treatment and Recycling
Demonstration Projects Program
Robert Ludwig, California Department of Health Services 460
Field Assessment of Steam Stripping Volatile Organics from,
Aqueous Waste Streams
Marvin Branscome, Research Triangle Institute 468
Field Assessment of the Fate of Volatile Organics in Aerated
Waste Treatment Systems
David Green, Research Triangle Institute 478
Pilot-Scale Evaluation of a Thin-Film Evaporator for Volatile
Organic Removal from Land Treatment Sludges
Coleen M. Northeim, Research Triangle Institute 487
-viii-
-------
SESSION C - HWERL POSTER PRESENTATIONS
EPA/DOE Hazardous Waste Control Technology Data Base
Cathy S. Fore, DOE Hazardous Waste Remedial Actions Program 495
Analysis of Samples from the Gateway National Recreation Area at
Jamaica Bay, New York
Dave Olsen, NUS Corp./Enviresponse Inc. 496
Case Evaluations of RD&D Permit Applications
Wyman Clark, EER Corp 497
Update on Status of EPA Mobile Incineration System
A.C. Gangadharan, Enviresponse, Inc. . . . 498
Boiler Cofiring of Chlorinated Hydrocarbons
John W. Wasser, U.S. Environmental Protection Agency . >.' 499
Demonstration, Testing and Evaluation of Commercial
Technologies under SITE Program
Seymour Rosenthal, Enviresponse, Inc. 500
Conditions Which Enhance Biodegradation of Organic Compounds by
White Rot Fungi
Steven Aust, Michigan State University 501
Demonstration and Evaluation of the EPA Mobile Carbon Regenerator
Patricia M. Brown, Enviresponse, Inc 502
Pretreatment of Land-Treated Wastes
Thomas C. Ponder, Jr., PEI Associates, Inc. 503
Geotechnical Analysis for Review of Dike Stability
Mark S. Meyers, University of Cincinnati . . . . 504
Land Ban Data Needs
Ron Turner, U.S. Environmental Protection Agency 505
Demonstration of Computer Assisted Engineering Techniques for
Remedial Action Assessment
Phillip R. Cluxton, University of Cincinnati 506
Hazardous Waste Residuals Characterization
H. Paul Warner, U.S. Environmental Protection Agency . 507
Cost Engineering Models for Remedial Response Technologies
Wiliam Kemner, PEI Associates, Inc. ; . 508
Trial Burn Measurement Guidance
Roy Neulicht, Midwest Research Institute 509
-ix-
-------
SESSION C (Continued)
Page
Microscopic Microchemical Anajyses of Solidified Inorganic Wastes
Containing Interference Compounds
Harvlll C. Eaton, Louisiana State University 510
Vacuum-Assisted In-Situ Steam Stripping to Remove Pollutants from
Contaminated Soil
Arthur E. Lord, Jr., Drexel University ........... 511
Use of Modified Clays for Adsorption and Catalytic Destruction of
Contaminants
Steven A. Boyd, Michigan State Univeristy 512
Stringfellow Leachate Treatment with Rotating Biological
Contactor .. • ..
Edward Opatken, U.S. Environmental Protection Agency ... 513
Separation and Recovery of Hazardous Wastes
Paul R. Anderson, IIT Research Center 514
Treatment of Aqueous Metal and Cyanide Bearing Hazardous Wastes
Sardar Q. Hassan, University of Cincinnati 515
An Experimental Investigation of Single Droplet Combustion of
Chlorinated Hydrocarbons
Nelson Sorbo, University of California, Davis 516
Catalytic Destruction of Halogenated Hazardous Waste
Howard Greene, University of Akron . 517
Expert System Screening of Remedial Action Technologies for
CERCLA Sites
Lewis Rossman, U.S. Environmental Protection Agency 518
Activities at Louslana State University's Hazardous Waste
Research Center
Louis Thibodeaux, Louisiana State University 519
Partitioning of PCDDs and PCDFs in Soils Containing Wood
Preservative Fluid
Danny Jackson, Radian Corp 520
Technical Resource Documents
Norman Surprenant, Alliance Technologies Corp 521
-x-
-------
SESSION C (Continued)
Oxidation of Persistent Aromatic Pollutants by Lignin-Degrading
Enzymes
John Glaser, U.S. Environmental Protection Agency 522
Laboratory Study of the Thermal Decomposition of Sulfur
Hexafluoride
Philip H. Taylor, University of Dayton Research Institute 523
The U.S. EPA Combustion Research Facility
R. W. Ross, Acurex Corp. 524
Construction, Testing, and Shakedown of an Environmental Testing
Chamber of Soil Reagent Testing
Michael Black, U.S. Environmental Protection Agency 525
Earthen Liners: Prototype of a Field Study of Transit Time
Karen A. Albrecht, Illinois State Geological Survey 526
-xi-
-------
-------
TECHNICAL RESOURCE DOCUMENTS AND TECHNICAL HANDBOOKS
FOR HAZARDOUS WASTES MANAGEMENT
Norbert B. Schomaker
and
Daniel W. Farrell
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Environmental Protection Agency is preparing a series of Technical Resource
Documents (TRD's) and Technical Handbooks to provide best engineering control technology
to meet the needs of the Resource Conservation and Recovery Act (RCRA) and the Compre-
hensive Environmental Response Compensation and Liability Act (CERCLA) respectively.
These documents and handbooks are "basically compilation of research efforts of the Land
Pollution Control Division (LPCD) to date. The specific areas of research being cond-
ucted under the RCRA land disposal program relate to laboratory, pilot and field valid-
ation studies in cover systems, waste leaching and solidification, liner systems and
disposal facility evaluation. The specific areas of research being conducted under the
CERCLA uncontrolled waste sites (Superfund) program relate to pilot and field validation
studies in barriers, waste storage, waste treatment, modeling and postclosure evaluation.
.The technical resource documents are intended to assist both the regulated community and
the permitting authorities, as well as support the RCRA Technical Guidance Documents
prepared by EPA'"s Office of Sol-id Waste (OSW). The technical handbook's provide the EPA
Program Offices and Regions, as well as the states and other interested parties, with
the latest information relevant to remedial actions.
INTRODUCTION
Land disposal of hazardous waste is
subject to the requirements of Subtitle C
of the Resource Conservation and Recovery
Act (RCRA) of 1976 and to the 1984 Amend-
ments to this Act. This Act requires
that the treatment, storage, or disposal
of hazardous waste be carried out in
accordance with RCRA regulations. Owners
and operators of new facilities must
apply for and receive a RCRA permit
before beginning operation of such a
facility..
The clean-up or containment tech-
nology associated with remedial action at
an existing uncontrolled hazardous waste
is subject to the requirements of the
Comprehensive Environmental Response,
Compensation, and Liability Act of 1980
(CERCLA or Superfund). This Act requires
, evaluation of remedial action clean up
technologies.
To meet the Control Technology
aspects of RCRA and CERCLA as related
to Land Disposal Facilities, the research
program has relied heavily on the cont-
ainment aspects of the wastes at the
facility or site. The containment
aspects for waste disposal onto the land
needs to address the development of
performance and operational standards for
new waste disposal facilities (RCRA
sites) and the containment or destruction
of pollutants emanating from existing
waste disposal facilities (CERCLA sites).
The control technology research approach
being pursued by the USEPA is to develop
an, improved data base so that current
waste disposal practices can be upgraded
-1-
-------
by developing proper site selection
criteria and control technology for the
establishment of new waste disposal
facilities, and to develop improved
containment technology for existing waste
disposal sites by minimizing pollutant
generation and release to the environment.
The development of proper control
technology for new (RCRA) waste disposal
facilities will combine information from
laboratory, pilot and field validation
studies in the research areas of cover
systems, waste leaching and solid-
ification liner systems and disposal
facility evaluation. Cover systems
research is developing and evaluating the
effectiveness of various material comp-
onents (i.e., vegetation, soils,
membranes, and drainage blankets) in
relation to the cover function of minim-
izing moisture ingress and gas egress.
Waste leaching research is investigating
techniques for predicting the composition
of actual field leachates from samples of
wastes or mixtures of wastes. Waste
solidification research is evaluating the
effectiveness and performance with time
as liners to contain and minimize the
release of-leachate and gas pollutants to
the environment. Disposal facility
evaluation research is evaluating effect-
ive techniques to ensure that land dis-
posal facilities are built as designed
for either permanent disposal or short/
long term storage. Incorporated through-
out the research for development of
control technology for new waste disposal
facilities is a continuous technology
transfer and assistance program of activ-
ity for the program office and regional
offices and user communities. The Tech-
nical Resource Documents (TRD's) are
considered to be primary documents for
transferring current RCRA control tech-
nologies to the user. The TRD's are
being developed and published to assist
the permit applicants and permit review
officials to assure that the latest
containment facility technology is being
utilized. The TRD's are to be used in
conjunction with the Technical Guidance
Documents being prepared by OSW. The
documents contain guidance, not regula-
tions or requirements which the Agency
believes comply with Design and Operating
Requirements and the Closure and Post-
Closure Requirements contained in Part
264 of the regulations. The information
and guidance presented in these documents
constitute a suggested approach for
review and evaluation based on good
engineering practices. There may be
alternative and equivalent methods for
conducting the review and evaluation.
However, if the results of these methods
differ from those of the Environmental
Protection Agency's method, they may have
to be validated by the applicant.
The development of proper contain-
ment technology to upgrade existing
(CERCLA) waste disposal facilites will
combine information from the above
described program for new waste disposal
facilities along with pilot and field
validation studies in the research areas
of barriers, waste storage, waste treat-
ment, modeling, and post-closure evalua-
tion. Barrier research is developing and
evaluating insitu control technologies to
contain or minimize pollutant releases
from uncontrolled sites and to predict
performance with time. In-situ control
technologies such as slurry walls, grout
curtains, cutoff walls, and covers are
being evaluated. Waste storage research
is evaluating the cost-effectiveness of
placing wastes from the clean up of
uncontrolled sites into mines or above
ground storage facilities. Also, pack-
aging of hazardous wastes in various
containers is being investigated. Waste
treatment research is evaluating the
effectiveness of various techniques for
treating the wastes or collected leachates
in-place or on-site. Techniques such as
stabilization, encapsulation, permeable
treatment walls, microbial degradation,
and physical/ chemical treatment are
being investigated. Modeling technology
has been developed for evaluation of
remedial action alternatives- and will be
updated to reflect improved performance,
reliability, and cost information from
field scale studies, case studies and
other research areas. Post-closure
research is evaluating the criteria for
final site usage once the disposal facil-
ity has been remediated. Incorporated
-2-
-------
throughout the research for development
of control technology for existing waste
disposal facilities is a continuous
technology transfer and assistance prog-
ram of activity for the program office
and user communitiess. The Technical
Handbooks are considered to be the primary
documents for transferring current CERCLA
containment technology to the user.
TECHNICAL RESOURCE DOCUMENTS
Eight TRD's have been completed to
date. Six of the TRD's are related to
landfills, and one document each is
related to surface impoundments and land
treatment. A listing of these documents
is shown below, along with a brief
decryption, publication number, and the
project officer's name is parentheses.
Evaluating Cover Systems for Solid and
Hazardous Waste (SW-867)
A critical part of the sequence of
designing, constructing, and maintaining
an effective cover over solid and hazard-
ous waste sites is the evaluation of
engineering plans. This TRD presents a
procedure for evaluating closure covers
on solid and hazardous wastes sites. All
aspects of covers are addressed in detail
to allow for a complete evaluation of the
entire cover system. , There are eleven
sequential procedures identified for
evaluating engineering plans.
The document describes current
technology for landfill covers in three
broad areas: data examination, evaluation
steps and post-closure plan. The data
examination discusses test data review
procedure, topographical data review and
climatological data review procedures.
The evaluation steps.include cover com- .
position, thickness, placement, config-
uration, drainage and vegetation. The
post-closure aspects include maintenance
and contingency plan evaluation proced-
ures. There are 36 specific steps,
regarding the preceding factors, which
are recommended to be followed in eval-
uating a permit for a cover for hazardous
waste. *055-000-00228-2 (R.E. Landreth)
Landfill and Surface Impoundment Perform-
ance Evaluation (SW-869)
The evaluation of leachate collection
systems using compacted clay or synthetic
liners to determine how much leachate
will be collected and how much will seep
through the liner into underlying soils
is presented. The adequacy of sand and
gravel drain layers, slope, and pipe
spacing is also covered. The author has
allowed for the widely varied technical
backgrounds of his intended audience by
presenting, in full, the rigorous math-.
ematics involved in reaching his final
equations. Thus, any evaluator can take
full advantage of the manual up to the
level of his own mathematical profic-
iency. *055-000-00233-9 (M.H. Roulier)
Lining of Waste Impoundment and Disposal
Facilities (SW-870)
This document provides information
on performance, selection, and install-
ation of specific liners and cover mater-
ials for specific disposal situations,
based upon the current state-of-the-art
of liner technology and other pertinent
technologies. It contains descriptions
of wastes and their effects on linings, a
full description of various natural and
artificial liners, service life and
failure mechanisms; installation problems
and requirements of liner types, costs of
liners and installation, and tests that
are necessary for pre-installation and
monitoring surveys. A revised version
should be available in late 1986.
*055-000-00231-2 (R.E. Landreth)
Management of Hazardous Waste Leachate
(SW-871)
This document has been prepared to
provide guidance for permit officials and
disposal site, operators on available
management options for controlling,
treating, and disposing of hazardous
waste leachates. It discusses consider-
ations necessary to develop sound manage-
ment plans for leachate generated at ',
surface impoundments and landfills.
Management may take the form of leachate
-3-
-------
collection and treatment, or pretreatment
of the wastes.
*055-000-00224-0 (S.C. James)
Guide to the Disposal of Chemically
Stabilized and Solidified Wastes (SW-872)
The purpose of this TRD is to prov-
ide guidance in the use of chemical
stabilization/solidification techniques
for limiting hazards posed by toxic wastes
in the environment, and to assist in the
evaluation of permit applications related
to this disposal technology. The document
addresses the treatment of hazardous waste
for disposal or long term storage and
surveys the current state and effective-
ness of waste treatment technology. A
summary of the major physical and chemical
properties of treated wastes is presented.
A listing of major suppliers of stabiliz-
ation and solidification technology and a
summary of each process is included.
*055-000-00226-6 (R.E. Landreth)
Closure of Hazardous Waste Surface
Impoundments (SW-873)
The methods, tests, and procedures
involved in closing a surface impoundment
are discussed and referenced. Problems
related to abandoned methods such as
waste removal, consolidating the waste
on-site and securing the site as a land-
fill are also discussed. It is written
primarily for staff members in EPA
regional offices or state regulatory
offices, who are charged with evaluating
and approving closure plans for surface
impoundments under regulations of the
Resource Conservation and Recovery Act of
1976. Methods of assessing site closure
considerations are documented.
*055-000-00227-4 (M.H. Roulier)
Hazardous Waste Land Treatment (SW-874)
One objectives of "Hazardous Waste
Land Treatment" are to describe current
technology and to provide methods for
evaluating the performance of an applic-
ant's hazardous waste land treatment
facility design. Land treatment is
approached comprehensively from initial
site selection through final closure, and
additional information sources are refer-
enced liberally. Land treatment, which
involves using the surface soil as the
treatment medium, is already widely
practiced by some industries for handling
their hazardous wastes.
*055-000-00232-1 (C.C. Wiles)
FUTURE TECHNICAL RESOURCE DOCUMENTS
Additional TRD's now being developed
or in planning stages by the Office of
Research and Development are included
here for reference. A listing of these
documents is shown below, along with a
brief description and the project
officer's name in parentheses.
Soil Properties, Classification and
Hydraulic Conductivity Testing
This report is a compilation of
available laboratory and field testing
methods for the measurement of hydraulic
conductivity (permeability) of soils.
Background information on soil class-
ification, soil water, and soil compaction
are included along with descriptions of
sixteen methods for determination of
saturated or unsaturated hydraulic cond-
uctivity. This TRD (SW-925) was published
by OSW in March 1984 for public comment.
It is being revised to incorporate public
comments that were received. A draft
copy of this document has been sent to
ORD where it has been stalled, awaiting
clearance. (G.K. Dotson)
Solid Waste Leaching Procedures Manual
This is a report on laboratory batch
procedures for extracting or leaching a
sample of solid waste so that the comp-
osition of the lab leachate is similar to
the composition of ieachate from waste
under field conditions. This TRD "(SW-924)
was originally published by OSW in March
1984 for public comment and has subse-
quently been revised with their incorpor-
ation. At present, the draft document is
completed and awaits clearance by ORD.
(C.I. Mashni)
-4-
-------
Batch^Type Adsorption Procedures for
Estimating Soil Attenuation of Chemicals
This TRD summarizes laboratory batch
procedures for assessing the capacity of
soils to attenuate chemical constituents
from solutions such as leachates. It
explains the scientific basis and ration-
ale for these procedures and the use of
data in designing soil liners for pollu-
tant retention.
It will be issued for public comment
in May 1987. Copies will be available for
inspection at EPA Libraries in Cincinnati,
Washington D.C., Research Triangle Park,
and in all ten Regional Offices. It may
also be purchased, on paper or microfiche,
from the NTIS as PB 87-146155.
(M.H. Roulier)
Methods for the Prediction of Leachate
Plume Migration and Mixing
This project has developed a variety
of computer programs for hand-held calc-
ulators, microcomputers, and macrocomp-
uters. The programs predict leachate
plume migration from single and multiple
sources. The document also contains
discussions of sorption, case histories
and a field study. A draft for public
comment has been postponed until late
1988. (M.H. Roulier)
Hydrologic Evaluation of Landfill
Performance (HELP) Model
The HELP Model is a modification of
the original waste disposal site hydrol- '
ogic model entitled, "Hydrologic Simul-
ation on Solid Waste Disposal Sites."
This update has incorporated the two-
dimensional aspects of landfill cover
systems, as well as the addition of the
leachate collection system. OSW published
this TRD (SW-84-009 and SW-84-010) for
public comment in two volumes. These two
volumes are available from NTIS (tPB-85-
100-840 and tPB-85-100-832, respectively)
and include the user's guide for Version
1 and documentation and description of
the program. The HELP Model (Version 1)
is also available for the IBM PC/XT or
compatible computers. Version 2 of the
HELP Model is being developed to incorp-
orate public comments and results from
verification studies and will be published
in late 1987. (D.C. Ammon)
Design, Construction, Maintenance, and
Evaluation of Clay Liners for Hazardous
Waste Facilities
This 600-page TRD summarizes the
state-of-the-art for clay liners as of
August 1985. It was issued for public
comment in December 1986, is currently
being revised, and will be reissued for
inspection at EPA libraries in Cincinnati,
Washington D.C., Research Triangle Park,
and in all ten Regional Offices. The
draft TRD may be purchased, on paper or
microfiche, through NTIS as PB 86-184496/
AS. (M.H. Roulier)
TECHNICAL HANDBOOKS
Eleven technical handbooks have been
completed to date. These handbooks cover
a variety of techniques on general remed-
ial action guidance, in-place treatment,
barriers, decontamination and modeling.
A listing of those documents which
contain unique state-of-the-art inform-
ation is shown below, along with a brief
description, publication number, and the
project officer's name in parentheses.
Review of In-Place Treatment Techniques
for Contaminated Surface Soils
This two-volume report presents
information on in-place treatment tech-
nologies applicable to contaminated soils
less than 2 feet in depth. Volume 1
discusses the selection of the appropriate
in-place treatment technology for a
particular site and provides specific
information on each technology. Volume 2
provides background information and -
relevant chemical data.
Selection of in-place treatment
technologies follows the process outlined
in the National Contingency Plan. The
type of in-place treatment (extraction,
immobilization, degradation, attenuation,
or reduction of volatiles) is determined
on the basis of information available
-5-
-------
from the remedial investigation. Select-
ion of a specific technonlogy involves
assessment of waste, soil, and site-spec-
ific variables. The technology is
implemented if it is considered more
cost-effective in comparison with the
other alternatives.
tPB 85-124881 Vol. 1 (N.P. Barkley)
tPB 85-124889 Vol. 2 (N.P. Barkley)
Handbook for Remedial Action at Waste
Disposal Sites (Revised)
The objectives of -the Handbook are
twofold: (1) to provide the reader with a
generalized understanding of the pollutant
pathways involved in waste disposal sites,
the remedial actions as they apply to
each pathway, and the process of selecting
the appropriate remedial actions; and
(2) to provide detailed information on
specific remedial actions including
applications, state-of-the-art, design,
construction, and/or operating consider-
ations, advantages, disadvantages and
cost. EPA/625/6-85/006 (D.E. Banning)
Handbook for Evaluating Remedial Action
Technology Plans"
This Remedial Action Technical
Resource Document describes how the tech-
nologies and methods for evaluating prop-
oposed RCRA new hazardous waste disposal
sites can be applied to site-specific
remedial response activities for uncont-
rolled hazardous waste sites. The
, Remedial Action Document is based on the
state-ofthe^art technical and cost
information in eight TRD's for design and
c-valuation of new hazardous waste disposal
•jites under RCRA. That information was
reviewed for relevance to remedial resp-
onse at uncontrolled hazardous waste
disposal sites, and then edited to address
the needs of personnel involved in resp-
onse and remedial action planning-under
CERCLA. tPB 84-118-249 (H.R. Pahren)
Slurry Trench Construction for Pollutant
Migration Cbntrol
A guidance manual for slurry trench
cut-off wall design, construction, and
performance evaluation provides recommend
ations on a variety of scientific and
technical parameters relevant to using
this approach to isolate hazardous chem-
icals in near-surface groundwater regimes.
The accomplishment of this effort required
extensive information gathering and
integration of technical data gathered
from a diverse array of experience and
authorities. tPB 84-177-831 (W.E. Grube)
Guide for Decontaminating Buildings,
Structures and Equipment at Superfund
Sites
A decontamination manual was designed
for EPA Program Offices and Regional
Superfund Programs as part of the restor-
ation profile of Superfund sites. The
manual gives guidelines on 1) the extent
to which contamination of buildings,
structures and construction equipment can
be reduced or eliminated, 2) decontamin-
ation methods, 3) economics, 4) health
hazards, and 5) availability of equipment/
personnel for the detoxification proce-
dures. Specific waste types found in
contaminated buildings, structures and
equipment at Superfund sites are ident-
ified. Potential secondary impacts of
available and potential decontamination
treatment methods are addressed in this
study. Costs versus risk and projected
ultimate site usage are addressed.
Methods for monitoring the successfullness
of various procedures are defined.
tPB 85-201-234 (N.P. Barkley)
Modeling Remedial Actions at Uncontrolled
Hazardous Waste Sites
The objective of this document is to
provide technical guidance on the select-
ion and application of models for evalua-
ting remedial action alternatives at
uncontrolled hazardous wastes sites. The
volumes cover selection of models, simp-
lified methods for subsurface and waste
control actions, numerical modeling of
surface, subsurface and waste control
actions, and analytical and numerical
models for evaluation of surface water
remedial actions.
tPB 85-211-357 (D.C. Ammon)
-6-
-------
Leachate Plume Management
This Handbook summarizes information
of leachate plume dynamics and plume
management alternatives gained by a study
of leachate plume management techniques.
Factors that affect leachate plume move-
ment key considerations in delineating the
current and future extent of the leachate
plume, technologies for controlling the
migration of plumes, and criteria for
evaluating and selecting plume management
alternatives are discussed.
tPB 86-/22330 (N.P. Barkley)
Systems to Accelerate the Stabilization
of Waste Deposits
This document investigates in-situ
systems which accelerate the stabilization
of waste deposits. In-situ applications
involve three essential elements: select-
ion of a chemical or biological agent
(reactant) which can react with and
stabilize the waste, a method for delivery
of the reactant to the deposit and a
method for recovery of the reaction
products or mobilized waste.
Four reactant categories-have Been
examined: biodegradation, surfactant-
assistant flushing, hydrolysis, and
oxidation. Methods of delivery of react-
ants based upon gravity include surface
flooding, ponding, surface spraying,
ditching, and infiltration beds and
galleries. Forced injection (pumping)
may also be used. Recovery systems using
gravity include open ditching and buried
drains, and pumped methods include .well-
point and deep well systems.
tPB 87-112-306/AS (W.E. Grube)
Covers for Uncontrolled Hazardous Waste
Sites
A handbook has been developed which .
can be used as a guidance document for
the selection, design, installation, and
long-term maintenance of covers as remed-
ial actions. This handbook provides
technical information for regulatory
personnel as well as guidance for cover-
system designers and construction engin-
eers. tEPA/540/2-85/002 (J.M. Houthoofd)
Stabilization/Solidification of
Hazardous Waste
Another handbook has been developed
to provide guidance for the evaluation,
selection, and use of solidification/
stabilization technology as a remedial
action alternative at uncontrolled
hazardous waste sites. The planning for
the application of solidification/stabil-
ization is divided into two phases:
process selection and scenario selection.
Process selection is concerned with the
chemistry of the stabilization/solid-
ification processes in the identification
of the composition of the waste. Pres-
ented in the handbook are testing and
analysis techniques for characterizing
waste as a basis for selection of pre-
treatment and stabilization/solidification
processes. Also data are developed on
the compatibility of additives and spec-
ific classes of waste, and testing systems
for the evaluation of stabilizaed/solid-
ified wastes are reviewed. Scenario
selection is concerned with the develop-
ment of equipment requirements, construc-
tion sequencing, and exist estimating for
the chosen solidification/stabilization
process. The handbook presents four basic
field scenarios, based on field surveys,
that have been used successfully.
tEPA/540/2-86/001 (J.M. Houthoofd)
Fugitive Dust Control at Hazardous Waste
Sites
Field studies were performed to
determine the effectiveness of dust
control technologies at hazardous sites.
In the first field study, dust suppressants
were tested to determine the effectiveness
of fugitive dust control against wind
erosion from exposed areas. Based on a
tracer sampling.protocol, the suppressants
were 100 percent effective for- 1 to 4
weeks after application, with declining'
control efficiencies thereafter. The
second field study was an evaluation of
the effectiveness of windscreens and
windscreen/dust suppressant combinations
in controlling fugitive dust from storage
piles. tEPA/540/2-85/003 (S.C. James)
-7-
-------
CONCLUSION
The "technical Resource Documents and
the Technical Handbooks that are being
prepared and updated by the Land Pollu-
tion Control Division (LPCD) are a series
of documents which provide best engineer-
ing control technolgoy to meet the needs
of RCRA and CERCLA, respectively. The
TRD's provide design, operation, and
evaluation information related to new
RCRA hazardous waste disposal facilities
to assist the regulated community and the
permitting authorities. The Technical
Handbooks provide reliable and cost
effective remedial action technology
information related to Superfund facil-
ities to assist the user community and
on-scene coordinators. These documents
and handbooks present the sum total of
the body of information and experience
gained by the Agency over the years on a
given topic. As new information is
developed, the Agency intends to update
each of these documents and handbooks so
that they reflect the latest state-of-
the-art information.
More information about a specific
project of study can be obtained by
contacting the project officer referred
to in the text. Project Officers can be
contacted by writing or telephoning the
USEPA, Hazardous Waste Engineering
Research Laboratory, Land Pollution
Control Division, 26 West St. Clair
Street, Cincinnati, Ohio 45268.
Phone: (513) 569-7871
KEX TO SYMBOLS:
* These documents have been published and
the reports are available from GPO by
requesting the stock number. Copies can
be obtained for a price from the
Superintendent of Documents
U.S. Government Printing Office
Washington, D.C. 20402.
Phone: (202) 782-3238.
t These documents have been published and
reports are available from NTIS by
requesting the stock number. Copies can
be obtained for a price from the
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161.
Phone: (703) 487-4650.
-8-
-------
IMPLICATIONS OF CURRENT SOIL LINER PERMEABILITY RESEARCH RESULTS
W. E. Grube, Jr., M. H. Roulier, and J. G. Herrmann
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Since the 1970's hazardous waste 1egisi ation has specified soil
liner permeability (hydraulic conductivity) as a major criterion for
design of soil liners to contain hazardous wastes disposed on land.
HWERL studies since 1972 have included over 30 individual projects
investigating hydraulic features of compacted soils. Soi1/permeant
liquid compatibility, applicability of laboratory permeability tests,
field permeability testing, effective porosity, and solute transit time
have been investigated.
The type of test cell does not appreciably affect the value for
permeability measured in the laboratory for water alone. When the
permeant liquid contains high concentrations of solutes, rigid wall
permeameters show greater increases in permeability (relative to water
alone) than triaxial cells. Increasing the hydraulic gradient when
using a triaxial cell results in lower permeability values. Triaxial
cells and consolidation cells can simulate overburden loading stresses
on a soil liner but no data are available to determine how well these
laboratory values predict permeability of the soil liner after a
landfill cell is completely filled. The compaction mold modified for
permeability measurement is the most sensitive geotechnical laboratory
method to measure 1iner/1eachate compatibility.
Construction acceptance is being tested with infiltration test
devices. Optimum size of infiltration test area versus the number of
test sites is being studied. Collection lysimeters installed beneath
soil liners offer the most practical means to reliably quantify seepage
through the liner system. They also provide an opportunity to collect
tracer compounds or leachates solutes for calculation of transit times
through the 1i ner.
Examination of soil liners in the field consistently shows
heterogeneity in materials, fabric, moisture content, density, and
texture. Macrostructural units that provide preferential pathways for
rapid flow of liquids are common, but the amount of liquids moving in
these pathways is unknown.
Laboratory-measured values demonstrate the 1owest permeabi1ity
that a soil liner material can provide; these values should be regarded
as a goal that may be achieved in the field by skilled personnel only
under optimum conditions. The uncertainties about permeability/
-9-
-------
compatibility testing and the effect of construction practice on
permeability of the completed liner dictate that soil liners for
hazardous waste facilities must be designed, tested, and built with a
much higher degree of care than has been considered routine for civil
works.
RATIONALE
RESEARCH
FOR PERMEABILITY
The Resource Conservation
and Recovery Act (RCRA) of 1976
(PL 94-580); as amended in 1984,
(HSWA, PL 98-616; and.
regulations derived from these
laws provide that soils can be
used as a component of the liner
system of a waste disposal
facility. EPA regulations
specify that the soil liner shall
have a permeability of no more
than 1 x 10E-07 cm/sec and that
the liner shall prevent migration
through it during the life of the -
facility. Specific regulatory
requirements are found in the
Code of Federal Regulations, 40
CFR Parts 264 and 265, available
1n pub!ic 1ibraries.
Fifteen years ago, questions
based on principles of soil
chemistry arose suggesting that
solvents or other strong chemical
compounds in waste and landfill
leachates would be expected to
attack a soil liner and make it
more permeable with time.
Results of early research pointed
toward several unsolved, related
questi ons:
permeability methodology,
effective porosity,
field-scale measurement,
Since then, EPA's HWERL in
Cincinnati conducted over 30
studies evaluating clay soil
liner permeability and related
issues.
EFFECT OF SOLVENTS AND DILUTIONS
ON PERMEABILITY
Our early efforts addressed the
question of soil-liner
compatibility with leachates (27
30). Recent articles
(2,7,18,30,39) contain the-full
bibliographies that document work
we accomplished determining the
value of a permeability test to
measure 1iner/1eachate
'compatibi 1 i ty .
Although much data were
generated which demonstrated the
.adverse effect of pure solvents
and high concentrations of other
chemical compounds on soil liner
permeability, studies using
actual landfill leachates of
diluted solvents or salt
solutions showed an almost
undetectable effect in the
laboratory studies. A single
long-term laboratory study was
reported (30) that incorporated
actual industrial waste liquids.
These results over 10 years show
no clear change in soil liner
permeability. Recent data
(1,6,14) and others continue to
demonstrate that in laboratory
studies dilute solutions or
leachates of only several hundred
ppm solutes have no effect.
ASTM methods (mainly D-2434
adapted to low-flow rates) ha,ve
been applied to determine soil
liner permeability (1,2,15,23).
Since the early 1980's, other
geotechnical engineering testing
laboratories and academic civil
engineering departments have also
conducted numerous experiments
investigating various aspects of
liner permeability, and its
measurement. Citations listed
(28), document this activity.
Johnson and Richter (33)
provide an extensive compilation
of permeability studies up to
1967. A critical
evaluation of all of these
published reports is beyond the
scope of this paper, but many of
the data generated by these
studies are important in
understanding the current state
of soil permeability knowledge.
-10-
-------
Extensive data have been
reported demonstrating that the
permeability of a compacted soil
increases as the dielectric
constant of the permeant liquid
decreases (8,28,36). Although
many pure organic liquids have a
low dielectric constant, dilute
inorganic solutions also possess
a low dielectric constant.
Neither dilute inorganic
solutions nor dilutions of low-
dielectric-constant solvents have
been shown to significantly
affect soil liner permeability.
Thus, dielectric constant of a
permeant liquid is not per se a
meaningful measure of its
permeability performanceunless
it is a pure liquid or a mixture
of but a few pure liquids.
PERMEABILITY OF SOIL LINERS
CONSTRUCTED IN THE FIELD
Laboratory permeability test
results frequently demonstrate
that the tested soil presents a
hydraulic conductivity less than
that of the regulatory threshold,
1 x 10E-07 cm/sec, especially
where solute concentrations are
in the few hundred ppm range.
Recent reports (confirming data
first compiled by Olson and
Daniel(38), and still widely
cited) observe that a liner
constructed in the field with the
same soil tested in the
laboratory has a seepage rate
representing a much higher
permeability than laboratory test
results show. Since the early
1980's our Laboratory has been .
investigating the reasons for
this discrepancy. Thus recent
studies have emphasized field
performance of hydraulic
barriers. Our current research
efforts include attempts to (1)
quantify the areal variability of
soil liner hydraulic properties,
(2) obtain an accurate measure of
solute transit time through
several lifts of liner
constructed by field equipment,
and (3) determine the feasibility
of improving present construction
practices in order to obtain a
soil, ,liner with a lower overall
permeability. These studies are
in the early stages and will not
be discussed here.
It is clear from numerous
laboratory permeability
measurements that many soils
demonstrate a permeability of 1 x
10E-09 cm/sec or lower. These
values can be lowered in many
cases by additional measuress
such as additives, modified
compaction, or. other soil
handling. Additions of
bentonite, lime, Portland cement,
and other components to soils in
laboratory studies have resulted'
in substantially lower
permeability. Where the
hydraulic conductivity is lower
than about 1 x 10E-08 cm/sec,
diffusion processes move chemical
species at the same rate as
hydraulic flow. 'Thus by
constructing a relatively thic'k
soil liner, with chemical
diffusion governing mobility,
pollutants can be prevented from
leaving the lined facility
through the'soil liner for the
same or longer time as that
provided by a relatively thin
liner of-synthetic material.
Therefore, soil 1 i ner-thickness1
of several feet, with attendant
permeability of 10E-09 or lower
will provide liquid containment '
at least equal to that obtainable
with synthetic materials. The
field scale liner must not
contain faults at any greater
magnitude or frequency than those
now found in comparable areas of
synthetic materials piaced as
large sheets. The liner
construction industry must
demonstrate this at full field
scale. The demonstration must
include long term containment as •
measured by lack of seepage
through the constructed liner.
It will challenge present
environmental monitoring
techniques to accurately
determine that little or no'
-11-
-------
liquids seep through a buried
soil structure during a time of
many decades.
LABORATORY PERMEABILITY TESTS
Laboratory measurements are
made in order to determine
whether a candidate soil liner
material meets the permeability
specification of 1 x 10E-07
cm/sec, or to obtain an
evaluation of the compatibility
of a soil liner material with a
leachate or other solution. The
latter purpose is served by
noting whether a significant
change in permeability occurs
when the leachate is introduced
into the testing system.
Numerous laboratory permeability
tests are also being conducted
simply to better understand soil/
solution interactions.
Soils tested for
permeability in the laboratory
are first characterized by
standard engineering criteria.
These normally include
-grain size distribution,
including percent fines,
-Atterberg Limits
-water content
-classification by the
Unified Soil Classification
System.
ASTM procedures for these
tests specify that only field-
collected soil samples that pass
through, at largest, a sieve with
3/4-inch (19mm) openings are to
be tested. Indeed, ASTM allows
up to 30% of a field sample to
contain fragments larger than
3/4-inch before disallowing a
Standard Proctor compaction test.
It is at this stage--when a soil
is specified as being suitable
for liner service-- that
laboratory testing results
exhibit serious shortcomings as
predictors of field performance.
Three major apparatuses are
used to routinely determine soil
hydraulic conductivity in the
laboratory: rigid wall cell, or
modified compaction mold;
triaxial cell modified for
permeability testing; and
consolidation eel 1 (oedometer)
modified for directly measuring
permeability. In evaluating the
data generated by our studies and
those supported by other
institutions, it is clear that
the differences among
permeability values for each of
the cell types are not
significant when using water and
the same soil. Thi s is
particularly true if one
considers than many commercial
geotechnical testing laboratories
consider agreement within a
factor of two or three to be the
best obtainable on replicate
permeability analyses of fine-
grained soils. A point must be
made here that publication of
soil liner permeability data in
detail greater than two
significant figures is probably
inappropriate, unless accompanied
by detailed description of the
laboratory technique.
Permeability testing using a
solvent, mixture, or actual
landfill leachate presents
several problems in a soil
permeability test. These are not
insurmountable, however. The
apparatus degradation problem is
solved by using resistant
materials, such as stainless
steel, plastics, or other
substances resistant to chemical
attack. In the case of the
triaxial cell, several different
materials have been used to solve
the problem of chemical attack of
the confining membrane. Here the
user must be certain, however,
that some of the stiffer
materials used to resist chemical
attack, such as Teflon film,
aluminum foil, or multiple-layers
-12-
-------
of these materials, do not
interfere with the often-stated
advantage of the triaxial method-
-conforming the Latex membrane to
the soil sample surface to remove
the possibility of sidewall
leakage. If too rigid a membrane
is used, the lateral pressure
that must be applied so the
membrane will conform with the
surface of the soil sample may be
so high that lateral
consolidation occurs.
The potential effect of
lateral consolidation must be
accounted for in each application
of the triaxial apparatus to a
study of 1 iner/1 eachate
compatibility. So many data now
exist showing that soil
permeability decreases with
increasing hydraulic gradient--
and concurrent increases in
lateral confining stress — that
the triaxial cell must be
discouraged for use in soil-
1iner/1eachate compatibility
measurement (31,p.375; 9,pp.72-
82; 10,pp47-8; 35,p.364;
22,p.1654; 45,p.405; 17,Fig.7;
25). We recognize that triaxial
confinement is the only practical
apparatus to measure the
hydraulic conductivity of
undisturbed samples of a soil
liner, but caution must be urged
to apply a confining stress no
greater than is present in the
field. A few reports in widely
distributed technical literature
detail the application of the
triaxial cell to soil liner
permeability testing
(10,15,23,29). They commendably
present valuable details of the
method, but do not adequately
emphasize the the need to apply
only the lateral stress seen in
the field. This restriction
effectively prohibits the use of
high hydraulic gradients commonly
applied to reduce testing time.
The consolidation cell is
useful in measuring volume
changes of a test specimen during
permeation. Applied stress
should not exceed field values.
Arguments that this apparatus
simulates the overburden .loading '
of landfilled waste and that
these pressures will heal cracks
that may form from leachate
interaction with soil liner
materials fail to acknowledge the
liklihood of variable loading
pressures over a liner surface.
Placement of drummed waste, for
example, may create bridged voids
within the landfill, where the
liner underneath a drum may
experience much higher loading
pressure than areas between
drums. There are no data
available showing the degree of
consolidation present in liners
beneath land disposal facilities.
The modified compaction mold
permeameter described by EPA (43)
provides the most immediate
information regarding the
detrimental effect that a
leachate may have on a soil liner
material. We recognize that
laboratory technicians need to
exercise care to ensure that
sample preparation within rigid
wall cells does not provide
immediate sidewall leakage. Soil
shrinking because of
1iner/1eachate interaction may
rapidly increase measured
permeability and cause the
testing laboratory to conclude
that an incompatibility may
exist. The value of the rigid-
wall permeability cell in
providing a conservative estimate
of 1iner/1eachate compatibility,
by means of demonstrating a
change in permeabi1ity outweighs
disadvantages such as lack of
achieving complete sample
saturation. Daniel and
Broderick(16) used the compaction
mold cell because of its greater
sensitivity to potential
soi1/Ieachate interactions.
Bowders et al.(5) and Daniel et
al.(17) discussed the advantages
of the rigid and flexible wall
hydraulic conductivity cells for
studying compatibility. The
"double-ring compaction mold
permeameter" has been proposed as
-13-
-------
a device more sensitive to
detection of side-wall flow.
While this apparatus has
desirable features, no published
data describing its performance
on compacted clay have been
found.
Several recent reports
relate the results of studies
using constant flow rate pumps at
high pressure to obtain
laboratory permeability results
more rapidly (37,12,31).
Although data from these
investigations provide useful
characterization of various soil
materials, the sample size of the
soil being tested with this
technique has been very small.
We believe that the larger sample
sizes common in commercial
geotechnical laboratories reduce
data dependence on the careful
techniques necessary in these
research studies. Testing times
of several weeks or months result
in long 1iner/1eachate contact
times, which improves the
credibility of extrapolating
laboratory compatibility data to
long term field performance. The
long testing times necessary to
obtain reliable data from the
larger sample sizes and modified
classical methods have not
appeared to present any practical
problems to commercial
laboratories.
Deciding when stable
permeability has been reached and
a test can be ended is of greater
significance in obtaining a
realistic permeability value than
the type of laboratory
permeability cell. There is lack
of agreement within the
geotechnical testing community,
and we have found that
"experience" has dominated the
decision-making process. Some
laboratories rely upon equal
rates of inflow and outflow of
permeant liquid to define an
equilibrium. Peirce and
Witter(39) have recently
published a more objective
approach to determining whether
hydraulic equilibrium has been
reached. They point out that
frequency of data collection can
influence the test termination
decision. Cell inflow or
effluent volume data can be
collected using a variety of
buret sizes and collection ti.me
points. One should not use
calibrated columns of small
diameter to increase the accuracy
of small flow measurements, but
rather base the data collection
on porevolume increments that
have passed through the soil
sample. At this time we suggest
0.05 to 0.10 porevolumes as the
smallest increment. We do not
know the optimum smallest
increment of porevolume of
permeant liquid flow that will
provide reliable data from which
to calculate a permeability
value. But, a meaningful and
unbiased series of measurements
is essential in order to
calculate a series of successive
permeabilities that may be input
to statistical formulas to
determine equilibrium. Bryant
and Bodocsi (9) also present
methods to confirm that an
equilibrium permeability has been
reached.
FIELD PERMEABILITY TESTS
Demonstrating that a soil
liner constructed in the field
has the required low permeability
is the "proof of the pudding."
Field testing of soil liner
permeability can be considered
from three aspects:
1. Acceptance testing is a
measure of the quality of
construction in achieving the
specified permeability.
2. Performance monitoring
provides a measure of the seepage
through the soil liner over long
periods of use.
3. Transit time is a measure
of the time that a soil liner
will prevent a leachate solute
from migrating out of the bottom
-14-
-------
of the liner into the subsoils
below a waste disposal facility
Size of Soil Mass
Tested
Number of
Samples Tested
Size or Frequency
of Property
Figure 1. Interrelated
problems affecting accurately
characterizing the hydraulic
performance of a soil liner.
Figure 1 depicts the
problems facing us in accurately
characterizing the hydraulic
performance of a soil liner in-
place. This picture does not
address methodology problems
inherent in trying to quantify
very low flow rates characterized
by a hydraulic conductivity of
10E-07 cm/sec or lower. Factors
such as evaporation loss, liquid
flow volume quantification, leak-
proof apparatus, determination of
when to stop measurements because
steady-state fl ow-has been
achieved, and statistical
treatment of raw data are aspects
that must be dealt with.
Numbers of soil samples
tested(NSST) to determine the in-
the-field permeability is being
addressed by geostatistical
analyses of data obtained from
our studies (41). NSST should be
directly related to the size and
frequency of occurrence of soil
liner discontinutities that
provide higher flow rates than
allowed by a 10E-07 cm/sec
specification. The NSST are
likely to be severely affected by
economic considerations and by
the need to protect the
environment from landfill
seepage. The size of the soil
mass tested refers to the
diameter of a core or to the
surface area of the sample being
examined. The sample needs to be
sufficiently large to include
representative soil properties,
both well-mixed homogeneous mass
as well as occasional regions of
heterogeneities that permit
higher hydraulic flows.
The property of the soil
liner being tested is the
hydraulic flow rate. As
discussed later in this paper,
flow appears to be dominantly
controlled by porous regions in a
heterogeneous soil liner. The
spatial extent of these
discontinuties in the liner
matrix needs to be known so that
they can be accounted for in
sampling schemes. You need to
know the size and distribution of
•hydraulic discontinuities in,
order to reliably detect
hydraulic discontinuities. In
practice the trade-offs depicted
in Figure 1 need to be recognized
and accounted for in sampling
schemes.
EPA's Minimum Technology
6uidance(44) states a purpose of
the compacted soil liner is to
decrease leakage in the event of
leakage through the FML
component, and increase the
efficiency of the .secondary
leachate collectib'n system. In
order to be effective for this
purpose, the compacted soil liner
must be uniform in its property
of low permeability. Field
permeability tests must be aimed
toward evaluating the
permeability uniformity of as
large soil liner masses as
possible. Determination that a
soil liner meets a hydraulic
conductivity specification must
therefore take into account the
area of liner tested, as well as
the validity of the measurement
method. If one suspects that a
certain proportion of the soil
liner contains flaws that will
permit high seepage rates, then
these flaws should be included
within the area of. soil liner
evaluated to ascertain whether
the permeability specification
has been met. In this case
either a large area should be
evaluated for permeability, or
-15-
-------
adequate area! representation by
numerous small samples must be
provided and should be consistent
with accepted statistical
practice. The questions of how
many samples, and of what size,
must be addressed. We believe
that large area infiltration and
seepage data obtained from a test
section constructed with full
scale equipment represents the
current state-of-the-art.
Construction Acceptance Testing
In reports of soil liner
design and construction it is
often stated that the lowest
permeability is obtained by
compacting at a moisture content
wetter than the optimum found on
a moisture-density curve after
Proctor compaction. Recent
laboratory studies(30, 9, 4,)
reinforce the data from early
studies by Lambe and Mitchell
showing lower hydraulic
conductivity as the soil is
compacted at moisture contents
wet of optimum. It is unclear
exactly at what moisture content
wet of optimum the field
workability of a clayey soil
becomes impossible. Our
experience is that construction
contractors tend to be quite
conservative--!'n that they will
err on the dry side because they
know that soil that is extremely
wet is more difficult to compact,
Figure 2 s.hows the variation
1n compaction moisture content
across an i-ntensively studied
soil liner section of 1/20 acre.
In this case the optimum moisture
was 17.8%; 95% and 105% of
optimum are 16.9% and 18.7%
respectively. We have collected
both infiltration and seepage
data from this soil liner area
and have calculated comparable
permeabilities. We are currently
measuring the permeabilities of
intact cores taken from an
intensive sampling array across
this experimental area. These
data will comprise an extensive
set comparing laboratory and
field permeability under
carefully controlled conditions
Lift!
Lift 2
Lifts
II<95% OPT
>105% OPT
95-105% OPT
Figure 2. Distribution of
compaction moisture contents in
lifts of an experimental soil
liner based on nuclear surface
moisture meter (from 41).
Large infiltration rings,
particularly the recently
developed closed inner ring
types, are being rapidly accepted
by state and private
organizations. Using large rings
appears to be the best current
approach to more rapidly verify
the quality of soil liner
construction. Steady state flow
has been achieved after only a
few days (19). Objections have
been voiced that these devices
only evaluate the surface few
i nches of a total 1i ner
installation. Known applications
so far have been on soil liner
test sections, completed facility
liners, and clay cover soil
layers. They can also be used to
test different lifts in a liner
section several feet thick.
Infiltration ring approaches
simulate an event where a small
pond of leachate comes into
contact with a soil liner, such
as leakage from above. Large
-16-
-------
closed-top infiltration rings are
easy to install and monitor and
can represent a significantly
large area. Small double-ring
apparatuses, such as applied to
"perc-tests" for wastewater
application to agricultural
fields, should not be used for
soil liner testing unless
extensively modified.
Evaporation control and accurate
measurement of small flows are
essential in determining field
permeability. Several in-situ
permeability test methods have
been recently described (20); it
was concluded that an objective
of an in-situ test should be to
evaluate a representative volume
of soil. Clearly, this objective
can be better met by using larger
devices.
Our observations of dye
infiltration from water ponded on
test sections and our morphologic
examination of soil fabric lead
us to conclude that soil liner
test sections must be constructed
in at least four lifts. We have
observed large variability in
amounts of unbroken native soil
clods in the first lift above the
liner foundation and variable
thickness in the surficial lift
where the surface has been bladed
to reach a specified grade.
Reades (40) also appears to
negate the value of the surface
lift. If one constructs a test
section of only four lifts,
potential construction
compromises that are present in
the basal and surficial lifts
leave only the middle two lifts
as the effective hydraulic
barrier. Thus a fault, such as
inter-lift transport paths
between the two middle lifts,
brings about high permeability
results and/or short solute
transit times through the total
test section. The design,
construction, and operation of a
field-scale permeability test on
the test section of a soil liner
has evolved into a much greater
task than the usual compaction
test section practiced by all
earthwork engineers. Further,
present construction quality
assurance guidelines require
documentation that the entire
facility liner was constructed in
the same manner as was the test
section, so that test section
data are accurately
representative of the full-scale
soil liner performance.
The presence of
heterogeneity in the soil mass of
the test section requires that
the factors depicted in Figure 1
be applied to permeability and
other tests applied to the test
section. Cores taken for
laboratory analyses or
installation of small diameter
in-situ permeability sensors may
evaluate too small a volume of
soil to capture major hydraulic
pathways. Thorough soil pulver-
ization or mixing may reduce
these problems.
The nuclear surface
moisture-density meter, in
backscatter mode, is routinely
used to verify that the design
specifications are met on a
compacted soil liner. Figure 3
clearly shows that the detected
neutron backscatter is reflected
by the total soil mass in perhaps
a 9-inch cube beneath the
instrument. Our observations of
permeant liquid flow along
fracture planes, faces of
adjoining unbroken native soil
clods, and small-scale less dense
regions where compactor feet have
not directly compacted the soil
suggest that the soil volume
exposed to the area and volume
averaging effect of the neutron
probe is simply too large for the
neutron probe to detect
significant liquid flow pathways.
Where research data have
consistently shown that only a
small percentage of the liner
mass is active in permeant liquid
transmission, it follows that a
measurement of bulk properties of
a soil liner is unlikely to
detect small flaws that are
-17-
-------
hydraulically significant. This
conclusion simply means that the
Industry would benefit from a
construction quality measurement
technique that can reliably
detect small soil liner
discontinuities .
-SUB-BASE GRADE
Source
Surface
Detectors
Photon Paths
Figure 3. Geometry and scale
considerations of nuclear
moisture-density probe in
backscatter mode (from 42).
Long-Term Performance
The Hazardous and Solid
Waste Amendments of 1984, and
ensuing regulations, specify that
no constituent shall migrate
through a lower soil liner
"...during the period such
facility remains in operation
(including any post-closure
monitoring period)." This period
normally is several decades. A
recent article (11) referred to a
Canadian study that found that
numerous hazardous waste disposal
sites "leaked after an average of
14 years." Long-term containment
performance of a soil liner can
only be measured by detecting
leakage, or its absence.
Collection lysimeters appear to
be the most credible means to
measure the quantity and quality
of liquid that seeps through a
soil liner over an extended time
period(Fig. 4). Some type of
underdrain system is also needed
to collect seepage that can be
analyzed to evaluate the transit
time of solutes through a soil
liner. The State of Wisconsin
has for several years been
requiring the installation of
collection lysimeters beneath
lined disposal facilities (34).
Clearly, a device that intercepts
and quantifies leachate exiting
COMPACTED
SUB-BASE SOIL
COLLECTION DRAIN
GRANULAR-
BLANKET GRAVEL
Figure 4. Cross section of a
leachate collection lysimeter.
at the bottom of a liner is the
best means of determining whether
and when a constituent has passed
through the liner. The
collection lysimeter, however, is
not a simple answer to satisfying
the regulations. Sizes of the
leachate interception area in a
compilation (34) of Wisconsin
lysimeters ranged from 25 sq ft
to 7,400 sq ft. The installation
of these collection lysimeters
must be made very carefully to
ensure that any and all liquids
collected are not lost but are
conveyed to an outlet for
measurement (or treatment if in
significant quantities). Long-
term performance measurement
requires that a collection
lysimeter be placed beneath the
entire volume of the liner and
not arbitrarily located in some
convenient area of the surface or
other region of the liner.
Design, materials, and
construction must incorporate the
highest possible quality for two
important reasons: (1) the
collection structure must perform
perfectly for an unknown long
time period, and (2) after
construction, the lysimeter will
be unseen and unaccessible for
any future maintenance.
Alternative long-term performance
monitoring schemes, such as
monitoring well arrays are not
within the scope of this paper.
-18-
-------
Transit Time
To gai n approval , an
applicant for a land disposal
facility construction permit
normally provides hydraulic
conductivity data on the soil to
be used in a liner. "Knowledge
of the liner permeability is not
sufficient information to
accurately estimate the length of
time of liner effectiveness"
(32). To demonstrate that
leachate constituents will not
pass through a proposed liner
system, a test section of
compacted soil liner is an
accepted structure to demonstrate
hydraulic performance.
Installing a collection lysimeter
beneath this test section
provides an opportunity to
collect any seepage passing
through if water or other liquid
is ponded on the surface. Figure
5 depicts a possible approach to
such a test section. Adding
tracer compounds to surface
ponding allows ready observation
of whether and when a solute
exits the bottom of the liner.
We find very few published data
indicating liquid or solute
transit time for soil liners
built in recent years. One
company has installed detection
electrodes to try to monitor
passage of a wetting front
through a soil liner.(3)
Our experimental studies
show a rapid transit time of
tracer solutes through several
lifts of compacted soil; these
soil liners demonstrated low
permeabilities, less than 1 x
10E-07 cm/sec. Careful
examination of the compacted soil
lifts involved have clearly shown
that dyed liquids follow a few
preferential pathways through a
lift: commonly moving laterally
across a lift interface and
thence again downward through a
small zone in the next lift.
Footed compaction rollers are
widely used to achieve the shear
deformation of a kneading process
shown by laboratory studies to
result in lower permeability than
static or impact compaction.
Figure 5. Ponded or
underdrained test fill.
Figure 6. Compaction of a
multi-lift soil liner by a footed
roller.
Figure 6 illustrates the
variability in compaction
produced'by a footed roller where
the number of passes has been
inadequate to completely cover
the surface of the lift being
compacted. Although the number
of passes must depend on the foot
area and the roller drum's foot
density, at least 18 passes must
be made with commonly available
footed rollers to completely
-19-
-------
compact an area. Our
observations of vertical cross-
sections of soil liners several
feet thick show amazing
heterogeneity within a compacted
lift and adjacent lifts. We
conclude from liner profile
studies that regions of lesser
density, even as small as a
centimeter in dimension, comprise
a significant fluid flow pathway.
Our observations are
consistent with the critical view
of liner macrostructure by Folkes
(24). A percolating liquid seeks
the path of least resistance,
which need not be a clear
desiccation crack or fracture
within a liner of otherwise
plastic soil. These observations
support hypotheses that lack of
uniform soil compaction is a
major contributor to high seepage
rates through soil liners.
Figure 7 compares the constructed
fabric of a compacted soil where
it must support a load and where
it must contain a liquid. We
believe that structural
competence has not yet been shown
to provide the liquid containment
required of a landfill liner.
Structural Support
Containment
2
—
V
_—
V
~
"7~7 /T-/ /
Figure 7. Multi-lift soil
compacted by footed roller and
used for support or containment.
Where cracks and/or less-
dense soil regions are connected,
a ponded liquid will penetrate
the liner system much more
rapidly than transit time models
predict. We have only been able
to characterize these flow
pathways in field-constructed
liners using dyes in ponded
water, followed by careful soil
morphology studies of excavated
liners cross-sections. Liner
morphology studies have also
illustrated the dramatic effect
of unbroken native soil clods in
providing liquid flow short-
circuits. Daniel's laboratory
study of chunk si ze vs
permeability (13) has been widely
cited as illustrative of the
effect of soil clod size on
permeability. Folkes (24)
brought out the dependence of
compacted soil permeability on
soil structure, aggregate
orientation, and other factors.
He also cited several earlier-
studies that concluded that
permeability is controlled by
macropore distribution. Dunn
(21) has recognized the
importance of the structure of
the field-compacted soil liner in
restricting fluid flow. The
chunk-size in Daniel's study--
1/16 to 3/8 inch — represents
laboratory sieving, and field
liner soils are unlikely to be
limited to uniform clods this
small. Field-run borrow soils
commonly contain gravel-sized
stones or soil clods, unless
carefully screened. In examining
field-constructed soil liners, we
invariably see evidence of fluid
flow around the surfaces of
gravel or soil clods retaining
their native size. We also see
evidence that these native gravel
or clods are pushed intact
through the matrix of surrounding
soil during lift compaction.
Frequently a relatively smooth
plane results where a gravel or
clod has been so pushed. Close
examination has clearly shown
these planes to be paths of
liquid flow. Failure to break
down clods of native soil during
liner construction creates the
mechanism for rapid transit time
through the soil liner simply by
increasing the heterogeneity of
the liner.
-20-
-------
Laboratory studies of liquid
or solute transit time have tried
to identify a more accurate
measure of "effective porosity "
of soil liners. Effective
porosity has been described as
that portion of the total liner
porosity that contributes
significantly to fluid flow. We
have conducted several
investigations showing that
effective porosity determined on
laboratory prepared compacted
clay soils and in limited field
investigations is only in the
range of a few percent (32).
That is, effective porosity
comprises only about 10% of the
total liner porosity. Gordon
(26) presents a model for
effective porosity where fracture
characterization and continutiy
are known. It requires input of
field data for fractures, but
continuing porous zones could
probably be handled as well.
Collection lysimeters,
porous probes placed at different
depths, or relatively unverified
technologies of wetting front
detectors represent the range of
methods available to quantify
leachate transit time. Computer
model predictions of transit time
have thus far been unable to
accurately predict field events.
This is because soil liners
constructed in the field contain
properties and features which
have been excluded from models.
CONCLUSIONS
Laboratory-measured liner
soil permeabilities demonstrate
the lowest values that a soil
material can provide; these
values should be regarded only as
a goal that may be achieved in a
field structure by skilled
personnel under the best
conditions.
Use of the modi fi ed
compaction mold for
1iner/1eachate compatibility
tests removes an unaccounted-for
variable in consolidation cells
and triaxial cells. This
variable is the applied stress
that is likely to close actual
desiccation cracks or densify
small porous zones. Although
substantial stresses are present
on soil liners in place
underneath waste fills, it is
likely that this will not be
uniform, and unstressed or weakly
stressed regions may allow
permeation of leachates.
A substantial data base
exists demonstrating the adverse
effect of pure or relatively pure
organic solvents on compacted
soil liners. Long-term effects
of soil liner exposure to dilute
chemicals or leachates have not
yet been demonstrated. Available
data have resulted in regulations
that now prohibit land disposal
of solvents or free liquids.
Traditional soil moisture-
density measurement procedures
are unlikely to detect soil liner
features that are hydraulically
impo rtant.
Innocuous tracer compounds,
such as dyes and inorganic ions,
when added to water ponded on
liner test sections can be
detected in pore water or seepage
water. Collection lysimeters or
porous probes within the soil
liner can be monitored for the
presence of tracers. Time of
detection after addition can.be
used' to calculate realistic
solute transit times through the
soil liner.
Collection lysimeters must
be installed under liner areas
that are representative of the
total liner area. The
underdrained area must not be
constructed with special care
that is not applied to the rest
of the facility liner. Long-term
soil liner performance can only
be determined by long-term
monitoring.
-21-
-------
The references cited in this
paper should be examined, but
also the references in the cited
papers should be studied because
they provide essential
background.
REFERENCES
1. Acar. Y. B., A. Hamidon, S.
Field, and L. Scott. 1984.
The Effect of Organic Fluids
on Hydraulic Conductivity of
Compacted Kaolinite. in
Hydraulic Barriers for Soil
and Rock, ASTM STP 874,
avail. from ASTM,
Philadelphia, PA.
2. Acar, Y. B, and A. Ghosh.
1986. Role of Activity in
Hydraulic Conductivity of
Compacted Soils Permeated
with Acetone, in Proc. Int.
Symp. on Environmental
Geotechnology, ed. H. Y.
Fang, Envo Publ . Co., Inc.
pp. 403-412.
3. Bacopoulos, A. 1986.
Development and Operation of
the Keele Valley Landfill
Site, Maple, Ontario,
Canada, in Proc. Ninth
Annual Madison Waste
Conference, Univ. Wisconsin-
Madison, pp. 355-376.
4. Boutwell, G. P., and V. R.
Donald. 1982. Compacted Clay
Liners for Industrial Waste
Disposal. ASCE National
Meeting, Las Vegas, NV,
April , 23 pp.
5. Bowders, J. J., D. E.
Daniel, G. P. Broderick, and
H. M. Liljestrand. 1985.
Methods for Testing the
Compatibility of Clay Liners
with Landfill Leachate. in
Hazardous and Industrial
Solid Waste Testing, 4th
Symposium, ASTM STP 886, pp.
233-250.
6. Brown, K. W., and J. C.
Thomas. 1985. Influence of
Concentrations of Organic
9.
10,
11.
12.
Chemicals on the Colloidal
Structure and Hydraulic
Conductivity of Clay Soils.
in Proc. llth Rsch Symp. on
Land Disposal of Haz. Waste,
EPA/600/9-85/013, HWERL,
USEPA, Cincinnati,.Oh 45268
Brown, K. W., J. C. Thomas,
and J. W. Green. 1986. Field
Cell Verification of the
Effects of Concentrated
Organic Solvents on the
Conductivity of Compacted
Soils, in Haz. Waste and
Haz. Matls., 3(1):1-20.
Brown, K. W. 1986. Use of
Soils to Retain Waste in
Landfills and Surface
Impoundments, in
Utilization, Treatment, and
Disposal of Waste on Land,
E.C.A. Runge, ed.'publ. by
Soil Science Society of
America, Madison, WI, 23 pp.
Bryant, J., and A. Bodocsi.
1986. Precision and
Reliability of Laboratory
Permeability Measurements.
EPA-600/2-86-097, HWERL,
USEPA, Cincinnati, Ohio
45268. 177 pp. avail, from
NTIS as # PB87 113791/AS.
Carpenter, G. W., and R. W.
Stephenson. 1986.
Permeability Testing in the
Tri axi al Cell . in
Geotechnical Testing
Journal ,GTJODJ, 9(1) :3-9,
March. '
CE/ASCE, 1987
issue, p. 49
February
Coldewey, W. G. 1984.
Measurement of Low
Permeability Coefficients by
Means of Electronic
Instruments, in Proc. 5th
Nat!. Conf. on Mgmt. of
Uncontrolled Hazardous Waste
Sites, Washington, DC.
available from HMCRI, 9300
Columbia Blvd., Silver
Spring, MD, pp. 584-587.
-22-
-------
13. Daniel, D. E. 1981. Problems
i n Predi cti ng the
Permeability of Compacted
Clay Liners, in Proc. Symp.
on Uranium Mill Tailings
Management, Civil Eng,
Dept., Colorado State
University, Ft. Collins, pp.
665-675.
14. Daniel, D. E. and H. M.
Liljestrand. 1984a. Effects
of Landfill Leachates on
Natural Liner Systems.
Geotech. Engnrg, Rpt. GR83-
6, Civil Eng. Dept., Univ.
Texas, Austin, TX. 169 pp.
15. Dani el, D. E., S. J .
Trautwein, S. S. Boynton,
and D. E. Foreman. 1984b.
Permeability Testing with
Flexible-Wall Permeameters.
in Geotechnical Testing
Journal, GTJODJ, 7(3):113-
122, Sept.
16. Daniel, D. E. and G. P.
Broderick. 1985.
Stabilization of Compacted
Clay Against Attack by
Concentrated Organic
Chemicals. Geotechn Eng.
Report GR85-18, Civil Eng.
Dept., Univ. Texas-Austin,
105 pp.
17. Daniel , D. E., D. C.
Anderson, and S. S. Boynton.
1985a. Fixed-Wall vs.
Flexible-wall Permeameters.
in Hydraulic Barriers for
Soil and Rock, ASTM STP 874,
pp.107-123.
18. Daniel, D. E., D. E.
Foreman, and S. R. Day.
1985b. Effects of Hydraulic
Gradient and Field Testing
on Hydraulic Conductivity of
Soil. Draft Final Project
Report of Cooperative
Agreement CR810165,
submitted to HWERL, USEPA,
Cincinnati, Ohio 45268,.434
pp.
19. Dani el , D. E. and S. J.
Trautwein. 1986. Field
Permeability Test for
Earthen Liners. Proc. In-
Situ '86, ASCE Specialty
Conf. on Use of In-Situ
Tests in Geotech. Eng.,
Blacksburg, VA, Jun 1986, 15
pp.
20. Daniel, D. E. 1987.
Hydraulic Conductivity Tests
for Clay Liners. in Proc.
Ninth Annual Symposium on
Geotechnical and
Geohydrological Aspects of
Waste Mgmt. Colo. State
Univ., Ft. Collins, 20 pp.
21. Dunn, R. J. 1986. Clay
Liners and Barriers -
Considerations of Compacted
Clay Structure. in Proc.
Intl. Symp. on Environmental
Geotechnology, H. Y. Fang,
ed . , Lehi gh Univ.,
Bethlehem, PA, pp. 293-302.
22. Dunn, R. J., and J. K.
Mitchell. 1984. Fluid
Conductivity Testing of
Fine-Grained Soils. J.
Geotech. Eng. ASCE, Vol.
110, No. 11, pp.1648-1665.
23. Evans, J. C. and H. Y. Fang.
1986. Triaxial Equipment for
Permeability Testing with
Hazardous and Toxic
Permeants. in Geotechnical
Testing Journal, GTJODJ,
9(3) :126-132.
24. Folkes, D. J. 1982. Fifth
Canadian Geotechnical
Colloquium: Control of
Contaminant Migration by the
Use of Liners. Can Geotech.
J. 19:320-344.
25. Foreman, D. E. 1984. The
Effects of Hydraulic
Gradient and Concentrated
Organic Chemicals on the
Hydraulic Conductivity of
Compacted Clay. M.S.
Thesis, Univ. of Texas-
Austi n , 345 pp.
26. Gordon, M. J. 1986.
Dependence of Effective
Porosity on Fracture
Continuity in Fractured
-23-
-------
Media. Ground Water,
24(4):446-452.
27. Green, W. J., G. F. Lee, and 32,
R. A. Jones. 1979. Impact of
Organic Solvents on the
Integrity of Clay Liners for
Industrial Waste Disposal
Pits: Implications for
Groundwater Contamination.
Final Project Report, Grant
No. R-804549 from the USEPA.
avail, only from NTIS, 33,
Springfield, VA as # PB-81-
213423, 149 pp.
28. Grube, W. E. ,Jr. 1986.
Reference List on Hydraulic
Conductivity Measurements
for Containment of Landfill
Leachate. in Geotechnical
News, 4(2):16-18, June. 34.
BiTech Publishers Ltd.,
Suite 801-1030 W. Georgia
St., Vancouver, BC, Canada
V6E 2Y3.
29. Haji-Djafari , S. and J. C.
Wright,Jr. 1983. Determining
the Long-Term Effects of
Interactions between Waste
Permeants and Porous Media.
in Hazardous and Industrial 35.
Solid Waste Testing:Second
Volume, ASTM STP 805.
30. Haxo, H. E., Jr., R. S.
Haxo, N. A. Nelson, P. D.
Haxo, R. M. White, and S.
Dakessian. 1985. Liner
Materials Exposed to
Hazardous and Toxic Wastes.
EPA/600/2-84/169. HWERL,
USEPA, Cincinnati, Ohio 36.
45268. available from NTIS
as PB 85-121-333.
31. Henry, H. R., G. P. Whittle,
T. A. Carlton, and R. J.
Graves. 1985. Effects of
Hazardous Waste Chemicals on
the Permeability of Rock and
Clay Contaminant Containment
Media as Related to the
Transport of Pollutants to
Groundv/ater. in 37.
Hydrogeology of Rocks of Low
Permeability, Memoires Int.
Ass'n. of Hydrogeologists,
17th Internat. Congress,
Tucson, AZ, pp. 370-381.
Horton, R., M. L. Thompson,
and J. F. McBride. 1987.
Method of Estimating the
Travel Time of
Noninteracting Solutes
Through Compacted Soil
Material. Soil Sci . Soc.
Am. J. 51:48-53.
Johnson, A. I., and R. C.
Richter. 1967. Selected
Bibliography on Permeability
and Capillarity Testing of
Rock and Soil Materials, in
Permeability and Capillarity
of Soils, ASTM STP 417, pp.
176-210.
Kmet, P., and D. E.
Lindorff. 1983. Use of
Collection Lysimeters in
Monitoring Sanitary Landfill
Performance, presented at
Nat'l Water Well Assn. Conf.
on Characterization and
Monitoring of the Vadose
(Unsaturated) Zone, Las
Vegas, NV, Dec. 1983, 19 pp.
Korfi at is , G. P., A. C.
Demetracopoulos , and J. R.
Shuring. 1986. Laboratory
Testing for Permeability and
Dispersivity of Cohesive
Soils. in Proc. Int. Symp.
on Envi ronmental
Geotechnology , H. Y. Fang,
ed. , Lehi gh Univ.,
Bethlehem, PA, pp. 363-369.
Mashni , C. I., H. P. Warner,
and W. E. Grube, Jr. 1985.
Laboratory Determination of
Dielectric Constant and
Surface Tension as Measures
of Leachate/Liner
Compatibility, in Proc.
llth Annual Rsch. Sypm. on
Land Disposal of Haz. Waste,
EPA/600/9-85/013, avail from
NTIS as # PB 85 196376.
Olsen, H. W., R. W. Nichols,
and T. L. Rice. 1985. Low
Gradient Permeability
Measurements in a Triaxial
-24-
-------
38,
39,
40,
41,
42,
43,
System. Geotechnique, June
1985, pp.145-157.
Olson, R. E. and D. E.
Daniel. 1981. Measurement of
the Hydraulic Conductivity
of Fine-Grained Soils, in
Permeability and Groundwater
Contaminant Transport, ASTM
STP 746, T. F. Zimmie and C.
0. Riggs, eds. pp. 18-64.
Peirce, J. J.
Witter. 1986.
Criteria for
Permeabi1i ty
Geotech. Eng.
, and K. A.
Termi nation
Clay
Testing. J.
ASCE,
Double Liner Systems for
Landfills and Surface
Impoundments --Design,
Construction, and Operation,
EPA/530-SW-85-014. USEPA,
401 M. St. SW, Washington,
DC 20460, 71 pp.
45. Zimmie, T. F., J. S. Doynow,
and J. T. Wardell. 1981.
Permeability Testing of
Soils for Hazardous Waste
Disposal Sites, in Proc.
Intl Conf. on Soil Mechanics
and Foundation Eng.,
Stockholm, pp. 403-406.
44,
112(9):841-854, Sept.
Reades, D. W. 1986.
Laboratory and Field
Permeability Tests on Clay
Liners. in Proc. "Waste
Tech '86", Nat'l Solid Waste
Mgmt. Assn and Waste Age
magazine, Chicago, IL, Oct.
1986, 7 pp.
Rogowski, A. S. 1985.
Effectiveness of a Compacted
Clay Liner in Preventing
Ground Water Contamination.
in Proc. Fifth Nat'l Symp.
and Exp'n on Aquifer
Restoration and Ground Water
Monitoring, NWWA,
Worthington, OH 43085, pp.
412-429.
Rogows ki , A. S. , B. E.
Weinrich and D. E. Simons.
1985. Permeability
Assessment in a Compacted
Clay Liner. in Proc. 8th
Annual Madison Waste
Conference-Municipal and
Industrial Waste, Univ. of
Wisconsin-Madison, pp. 315-
336.
USEPA. 1983. Lining of Waste
Impoundment and Disposal
Facilities. SW-870. Office
of Solid Waste and Emergency
Response, Washington, DC.
avail from NTIS as PB 81-
166365. 448 pp.
USEPA. 1985. Draft Minimum
Technology Guidance on
-25-
-------
THE BEHAVIOR AND ASSIMILATION OF ORGANIC AND INORGANIC
PRIORITY POLLUTANTS CODISPOSED WITH MUNICIPAL REFUSE -
A PROGRESS REPORT
Frederick G. Pohland, Wendall H. Cross and Joseph P. Gould
School of Civil Engineering
Georgia Institute of Technology
Atlanta, GA 30332
ABSTRACT
The behavior and possible assimilation of organic and inorganic priority pollutants
oodisposed with refuse are being investigated in ten simulated landfill columns operated
under single pass leaching or leachate recycle. The priority pollutants include
selected organic compounds and three different"loading levels of heavy metals mixed with
municipal refuse. After being brought to indicated field capacity with water additions,
leachate and gas"from the ten columns were analyzed for routine indicator parameters as
well as the selected priority pollutants.
Preliminary results indicate that the presence of priority pollutants exhibited
little apparent influence on the progress of refuse conversion into the acid
fermentation phase of stabilization. Trends indicative of assimilative capacity,
particularly with respect to the inorganic priority pollutants, are beginning to be
established as the various biological and physical-chemical mechanisms of attenuation
take effect. Microbial mediation of the chemical environment has encouraged
precipitation and complexation or sorption of admixed species as gas and leachate
constituents are partitioned and released from the waste mass. Moreover, leachate
recycle tends to regulate this process, contain the various leached ingredients in a
more homogeneous medium, provide greater saturation and contact opportunity, and permit
better inspection and operational control of the overall mechanisms of in situ
assimilation.
INTRODUCTION
Landfill codisposal of municipal
refuse and organic and inorganic priority
pollutants emanating from households and
small quantity hazardous waste generators
has become common practice in many areas
throughout the country. Unfortunately,
this practice has received limited
scientific scrutiny and a general lack of
understanding prevails with respect to
loading limits and associated potentials
for adverse environmental impact.
Therefore, comprehensive evaluation of
these impacts and the possibility for
assimilation of hazardous wastes at such
landfill disposal sites is being addressed
by simulated codisposal investigations at
the Georgia Institute of Technology.
The information presented and
discussed herein is a summary of research
progress into the mechanisms controlling
and defining loading capacities and
impacts on the jLri situ processes of waste
stabilization when influenced by single
pass leaching or leachate recycle in
landfills. Therefore, comprehensive gas
and leachate analyses have been scheduled
over a three-year project period. From
these analyses, early results on"the
variations in leachate mercury and lead,
as well as measurable organic priority
pollutants, are used to typify stabiliza-
tion trends and reveal the complexity of
the landfill environment and its inherent
ability to accommodate and attenuate toxic
loadings.
-26-
-------
MATERIALS AND METHODS . .
Column Design and Operation
The ten simulated landfill columns
were designed in pairs with the operation-
al features illustrated in Figure 1. Each
column was constructed of two 0.9-m'diame-
ter steel sections, with a total height of
3.0 m. Five of the columns were designed
to operate with single pass leaching,
whereas the other five were designed to
facilitate leachate collection and recycle.
All columns were provided with appropriate
appurtenances and were sealed gas-tight
after loading.
QAS METES
TEMPEFUTtJSE INDICATOR
QAS SAMPLS.O VALVE
GAS TRAP
CHECK VALVE
PRESSURE GUAGE
DISTRIBUTOR A3M
RECVCLE P'JVP
FLANGE
THSRMOCCL'»LB
HOPE LINE*
JN-LINE F!!_T=3
STEEL
LEACHATE 5aAM .
LIQUID SAV='_E PCrIT
LIQUID '_=*=•_ CONTROL
GRAVEL. SA.-.3, AMD
13 GSO'rEXTH-S. SA,'iO.
GEOTEXTH.E. AND CRAVE!
LAYESS
13 110 V AC
20 MOV AC
TO PUf.tP
21 110 V AC P?3M
LIQUID LE'/ = _ CONTROL
22 VEflT TO ATVOSrHERE
23 SH3EODEO ==?USE
03ALL VALVE
SINGLE ?ASS UNIT
Figure 1. Simulated landfill columns.
The pair of single pass and recycle
control columns were loaded with shredded
municipal refuse only, whereas the remain-
ing four pairs of test columns were loaded
with selected organic priority pollutants
codisposed with shredded municipal refuse.
Three pairs of these latter test columns
also received incremental loadings of
toxic heavy metals'in the form of alkaline
metal finishing waste treatment sludge
spiked with additional heavy metals. The
loading levels for .the 'simulated landfill
columns are indicated in Table 1.
After loading and sealing had been
completed, tap water was added to bring
each column to indicated field capacity
and initiate the immediate production of
leachate for recycle and/or analysis.
Thereafter, water was added at a rate of
six liters per week for the single pass
columns and similarly to the recycle
columns until sufficient leachate had
accumulated to accommodate recycle and
analysis schedules.
Leachate and Gas Analysis
Once leachate was generated from each
of the ten columns, a comprehensive analy-
tical program was established for the
common indicator parameters as indicated
in Table 2. For the purposes of this
TABLE 2. LEACHATE AND GAS MONITORING
' PARAMETERS AND METHODS
Measurement
Conductivity
PH
Alkalinity
01", SOs"2,
P0,f3, s-2
HH3-N
OHP
BODg
COD
TOO
CHij/C02/H2
Cadmium
Calcium
Chromium
Iron
Lead
Magnesium
Manganese
Mercury
Nickel
potassium
Sodium
Zinc
Lithium
Solid Waste
Calorific Value
Solid Waste
Moisture
Volatile
Organic Acids
Reference
EPA 600/1-79-020
Method 120.1
EPA 600/1-79-020
Method 150.1
EPA 600/11-79-020
Method 310.1
Standard Methods
Method 129
EPA 600/1-79-020
Method 350.3
ASTM Method 1198-99
EPA 600/1-79-020
Method 105.1
EPA 600/1-79-020
Method 110.1
EPA 600/1-79-020
Method 115.1
Gas Chromatography
EPA 600/1-79-020
Methods 213.1 & 213.2
EPA 600/1-79-020
Method 215.1
EPA 600/1-79-020 :
Methods 218.1 & 218.2
EPA 600/1-79-020
Method 236.1
EPA 600/1-79-020
Methods 239.1 1 239.2
EPA 600/1-79-020
Method 212.1
EPA 600/1-79-020
Methods 213.1 4 213.2 ,
EPA 600/1-79-020
Method 215.1
EPA 600/1-79-020
Methods 219.1 & 219.2
EPA 600/1-79-020
Method 258.1
EPA 600/1-79-020
Method 273.1
EPA 600/1-79-020
Methods 289.1 & 289.2
Standard Methods
Method 317B
Parr Instruments
Tech. Manual J130
Ohaus Instruments
Tech. . Manual
Direct Aqueous
Injection capillary
Column G.C.
Std. Dev.
±6*
±0.1 SU
±5% '
±10*
±5*
-
±20*
±10*
±10*
±5*
not
±5*
±10*
±10*
±10*
±5*
±10*
±20*
±10*
±5*
±5*
±101
±5*
15*
±10*
Accuracy
95-1 05 J
±0.1 SU
95-105*
90-110*
90-110*
-
-
90-110*
90-110*
90-110*
90-110*
90-110*
90-110*
'90-110*
, 90-110*
90-110*.
90-110*
80-120*
90-110*
90-110*
90-110*
90-110*
95-105*
, - -
90-110*
90-110*
-27-
-------
TABLE 1. LOADING LEVELS FOR SIMULATED LANDFILL COLUMNS
Loading
Shredded Municipal
Refuse, kg (dry)
Priority Organic
Pollutants, g
Dioctyl phthalate
1 , E.'l-Triohlorobenzene
Dibrcmomethane
Llndane
Hexachlorobenzene
2,4-Dlchlorophenol
2-Nitrophenol
1,4-Dichlorobenzene
Naphthalene
Dieldrin
Nitrobenzene
Trlchloroethene
Priority Inorganic
Pollutants, g
Column Identity
1(CR) 2(C) 3(0) 4(OL) 5(OM) 6(OR) 7(OLR) 8(OH) 9(OMR) 10(OHR)
267 26? 267 267 267 267 267 267 267 267
120
120
120
120
120
120
120
120
120
•30
120
120
120
120
120
120
120
120
120
120
120
"30
120
120
120
120
120
120
120
120
120
120
120
-30
120
120
120
120
120
120
120
120
120
120
120
'30
120
120
120
120
120
120
120
120
120
120
120
"30
120
120
120
120
120
120
120
120
120
120
120
•30
120
120
120
120
120
120
120
120
120
120
120
•30
120
120
120
120
120
120
120
120
120
120
120
"30
120
120
Cadmium
Chromium
Mercury
Nickel
Lead
Zinc
- 26
- 45
20
- 39
- 101
- -45
.1
:3
:3
:7
:9
:7
52
90
40
79
209
91
.3
;6
;6
;4
;8
:4
26
45
20
39
104
- '45
.1
:3
.'3
.'7
:9
:7
104.6
181 ;2
-81 ;2
158:8
41 9. '6
182;8
52
90
40
79
209
91
.3
;6
:6
;4
;8
;4
104.6
181; 2
•81. -2
158. '8
419:6
182:8
Letters in parentheses indicate control (C) or recycle (R) columns; organic (0)
loadings; and, low (L), moderate (M) or high (H) inorganic loadings.
presentation, the protocols established
for the determination of leachate heavy
metals, selected anions, ORP, pH and
conductivity were of particular interest
with the latter analysis being used to
estimate the influence of ionic strength
and activity on solubility or complexation
equilibria.
In the case of the selected organic
priority pollutants, leachate samples were
spiked with surrogate compounds and
extracted for four hours with methylene
chloride using a continuous vapor phase
procedure. After drying over anhydrous
sodium sulfate and concentration to a
volume of 1.0 to 4.0 mL in a Kuderna-
Danish apparatus, the samples were
analyzed by capillary column gas
ohromatography-mass spectrometry (GC-MS)
using an internal standard. Results were
validated by a number of quality assurance
quality control (QA/QC) procedures,
including analysis of duplicate samples,
spiked samples and blanks. In addition,
calibration curves were developed for each
set of samples using serial dilutions of
standard solutions.
For those organic compounds capable
of being transferred by volatilization or
stripping to the gas phase, gas traps
containing 2 to 2.5 g of Tenax adsorbent
resin were inserted into the gas collec-
tion systems of each simulated landfill
column. These traps were removed periodi-
cally and the organics desorbed by heating
at 150°C for two to four minutes. The
desorbed organic compounds were trans-
ferred directly to a capillary column and
measured by GC-MS.
-28-
-------
RESULTS AND DISCUSSION
Changes in Selected Heavy Metals
As indicated previously, this presen-
tation will focus on the behavior and fate
of selected priority pollutants as the
landfill columns progressed through the
initial phases of landfill stabilization.
Accordingly, as has been described in
detail elsewhere (1), current operation of
the columns has been intentionally main-
tained in the acid formation phase, as
indicated by the low pH and high chemical
oxygen demand (COD) and total volatile
acids (TVA) concentrations illustrated in
Figure 2. This operational condition is
necessary to assess the effects and
implications of the test loadings during a
period when the leachate is most aggres-
sive and mobility generally most enhanced.
Based upon a fundamental understand-
ing of the reactivity of the various heavy
metals within the landfill environment
(2), mercury (Hg) and lead (Pb), were
selected as a focus for this presentation.
Accordingly, the two most significant
inorganic anions capable of affecting the
mobility of these metals were sulfate
(S0ij=) and chloride
-------
3000.
2000.
1000.
3000. :
2000.
1000.
0.
0.
500
100. 200. 300. 400.
Time since loading, days
Figure 3. Leachate sulfate concentra-
tions for recycle and single
pass landfill columns.
o.
-600,
100. 200. 300. 400.
Time since loading, days
Figure 5. Leachate ORP for recycle and
single pass landfill columns.
500
TOO. 200. 300. 400.
Time since loading, days
Figure 4. Leachate chloride concentra-
tions for recycle and single
pass landfill columns.
tracer, also decreased with time. In
searching for a cause for such a'decrease,
possible mechanisms influencing removal of
Pb and Hg included complexation and preci-
pitation and reduction of sulfate to
sulfide under the anaerobic conditions
prevailing within each landfill column.
As indicated in Figure 5, the likeli-
hood of reduction of sulfates to sulfides
was reinforced by the presence of reducing
conditions characterized by negative redox
potentials (ORP). Since sulfides serve as
potent precipitating agents for most heavy
metals, their presence in even very low
levels will control metal solubilities.
Moreover, ORP is significant in determin-
ing the chemical state of heavy metals,
their potential for reaction and their
ultimate mobility.
The perturbations in the figure indi-
cate that ORP is a procedure and condition
sensitive analysis which frequently
measures only trends rather than absolute
potentials. It is affected by the medium
within which the measurement is made and
by external influences during sampling and
500
-30-
-------
analysis. Exposure to air, contamination
of electrode surfaces, and non-homogeneous
mixing all tend to hinder absolute measure-
ments and often cause irregularities in
determination. Nevertheless, the consis-
tently negative ORP values were sufficient-
ly indicative of reducing conditions to
allow the presumptive diagnosis that
follows.
Similar difficulties were encountered
with the absolute measurement of sulfide
due to fouling of the electrode surface.
Moreover, during this period of investiga-
tion, sulfide levels were often below
detectable concentrations and were also
Influenced by reaction with the heavy
metals.
In terms of the two heavy metals
chosen for discussion here, mercury
solubility and speciation can be subject
to a complex array of factors including
reduction, precipitation and complexation.
Although the mercury was added in the form
of mercuric compounds, reduction to
mercurous ion or to metallic mercury could
occur under conditions of relatively low
ORP. Moreover, unlike the other heavy
metals loaded into the columns mercury
shares with lead the propensity to form
sparingly soluble chlorides, sulfldes and
sulfates. Sulfide forms very sparingly
soluble precipitants with both mercuric
and mercurous ions and can control mercury
solubility even if present in below
detectable levels. Likewise, mercuric ion
is subject to moderately strong complexa-
tion by chloride and, in the presence of
existing leachate chloride levels (Figure
4), the neutral complex, HgCl2, would be
the overwhelmingly dominant soluble
species of divalent mercury.
To inspect the behavior of mercury as
landfill stabilization progressed with
time, leachate mercury concentrations have
been plotted in Figures 6 and 7 on both
normal and expanded scales, respectively.
The expanded scale plot was necessary to'
2000. -
100. 200. 300. 400.
Time since loading, days
500
Figure 6. Leachate Hg concentrations
for recycle and single pass
landfill columns.
200. 300. 400.
Time since loading, days
Figure 7. Leachate Hg concentrations
for recycle and single pass
landfill columns (expanded
scale).
500
-31-
-------
compare the dissolved mercury levels when
leachate mercury concentrations had become
very low. Examination of these figures
indicates that, following the initial
elevated levels of leachate mercury for
the six spiked columns, the concentration
decreased rapidly to levels generally
below 50 pg/L.
To help explain this behavior, a
pC-pE diagram for the Hg+2/HgCl2/Hg°
system at pH 5.2 is presented in Figure 8.
The component equilibria suggest that, af
the negative redox potentials and acid pH
conditions prevailing in the landfill
columns, the mercuric compounds added will
be reduced to metallic mercury. This
finding is consistent with other reports
of metallic mercury in water at concentra-
tions of 20 to 1)0 ug/L (3,1)), and is of
major significance in determining eventual
fate, including the possibility of volati-
lization, particularly when methane
fermentation with gas stripping prevails.
pCI
14-
H-
P« u-
H.
u-
u-
II-
c
*
H«C1,
Hg"
Hg'
0
•1100
•1000
•BOO "'
r mV
Uaoo
•100
•94fi
•74B
•B45
-64S
-44S
•346
Figure 8. pC-pE diagram for
• Hg*2/HgCl2/Hg°
system at pH 5.2.
Lead, the other heavy metal selected
for analysis here, contrasts with mercury
in that it will be expected to remain in
the divalent state under current operating
conditions. This state subjects lead to
two significant reactions which control
its solubility. First, lead is suscepti-
ble to precipitation as sulfide, sulfate
or chloride. Secondly, while not as prone
to form very strong complexes as the
ionized forms of mercury, lead will form
such complexes with sulfate and chloride
and will thereby tend to increase lead
solubility beyond that otherwise predicted
by simple dissolution. Moreover, if there
is a preference for reduction to metallic
mercury, then in the absence of mercury
ions, lead will preferentially react with
sulfide over the other heavy metals loaded
into the columns.
To associate these concepts with an
interpretation of changes in leachate lead
concentrations, Figure 9 indicates that
initial high lead levels decreased rapidly
and then tended to increase again, particu-
larly in the most heavily loaded columns.
A possible explanation of these changes '
could be predicated on two considerations.
First, sulfide, even at very low concentra-
tions, would be the primary precipitant
for lead. Secondly, if sulfide is absent,
either sulfate or chloride will function
as the reacting anions.
500
100. 200. 300. 400.
Time since loading, 'days
Figure 9. Leachate Pb concentrations
for recycle and single pass
landfill columns.
Based on known solubility equilibria
(5,6), a'predominance area diagram has
been constructed in Figure 10 for the
important lead species under the
prevailing pH conditions. From this
diagram, it is apparent that the solid
-32-
-------
pSO,
Figure 10. Predominance area
diagram for Pb for
the recycle and
single pass landfill
columns.
species controlling lead solubility in the
absence of sulfide is PbSOi). Therefore,
in consideration of the complexity of
• equilibria for this system, the low
leachate lead concentrations are
reasonably consistent with sulfide as the
precipitant anion, while the more recent
higher concentrations are well within the
range expected in the presence of sulfate
control.
When coupled with the importance of
biological mediation, the results suggest
the possibility of very limited initial
sulfate reduction, but of sufficient
magnitude to control lead solubility.
Thereafter, possible heavy metal or acid
inhibition may have retarded further
generation of sulfide, thereby shifting
lead solubility control to the sulfate ion.
As the experiments progress and methane
fermentation is initiated with the further
production of sulfide, solubility control
for lead may again shift to sulfide. The
onset of methane fermentation with its
concomitment requirements for increasingly
low redox potentials will also continue to
sustain mercury in the metallic state.
Changes in Organic Priority Pollutants
Although the results of analyses on
the organic priority pollutants added to
the test columns have not yet revealed as
dramatic an impact on leachate quality,
the waste mass appears to be influencing
release consistent with solubility and
octanol/water partition coefficients (Kow)
as summarized in Table 3. Since Kow is a
measure of the tendency of a compound to
partition between water and a hydrophobic
phase, with increasing log Kow indicating
increasing hydrophobicity, each of the
listed compounds could be inspected
relative to their ability to dissolve in
the leachate and to be sorbed by the
refuse mass. Consequently, the highly
soluble compounds would be expected to
appear rapidly in the leachate at rela-
tively high concentrations, while the less
soluble would tend to be retained.
To test this hypothesis within the
landfill setting, it was necessary to
inspect the order and intensity of appear-
ance of the organic priority pollutants,
taking into account the position of
loading (near the bottom of each column),
the possibility of volatilization, and the
mode of operation (single pass or recycle).
As indicated in Table 4, all organic
priority pollutants with vapor pressures
greater than 1.9 kPa (Table 3) have been
observed in the gas phase of the columns
at least once. Gas phase transfer,
although minimal due to low gas evolution
under acid formation conditions, was most
frequently detected for trichloroethene
and dibromomethane. To assess the
magnitude of this transfer, gas phase
concentrations and masses have been
estimated for these two compounds in Table
5. Based upon this analysis, gas phase
transfer was not yet considered
significant, but would possibly be more so
when methane fermentation with gas
evolution and stripping of volatile
compounds Is established.
With respect to the leachate analyses
of Table 6, lindane and dioctyl phthalate
did not appear during the first year of
column operation, whereas the more soluble
trichloroethene and dibromomethane
appeared almost immediately. However, as
time passed, these latter compounds
eventually disappeared. In comparison,
the phenolic compounds'(2,4-dichlorophenol
and nitrophenol), although quite soluble,
have been observed only at relatively low
concentrations. Similarly, 1,2,4-tri-
chlorobenzene,'with its low solubility and
high log Kow, has been detected at low
concentrations, whereas naphthalene, being
only moderately soluble and fairly
hydrophobic, consistently appeared at low
to moderate concentrations.
-33-
-------
TABLE 3. PHYSICAL-CHEMICAL CHARACTERISTICS OF ORGANIC PRIORITY POLLUTANTS
Organic Compound
Molecular
Weight
Aqueous Solubility
mg/L
1,1, 2-Trichloroethene
Dibroraoraethane
1 , H-Dichlorobenzene
Nitrobenzene
2-Nitrophenol
Naphthalene
2,1-Dichlorophenol
1 ,2,1-Trichlorobenzene
Lindane
Dieldrin
Dioctyl phthalate
Hexaohlorobenzene
131.4
173:9
147:0
123:1
139:1
128:2
163:0
181 ;5
290:9
381 ;0
391:0
284:8
1,100
11,700
•• 49
1,900
2,100
•26.2
4,500-
19
17
•0.1
0:285
0:11
(20°C)
(15°C)
(22°C)
(20°C)
(20°C)
(20°C)
(20°C)
(22°C)
(24°C)
(24°C)
(24°C)
Log Octanol/
Water
Partition
Vapor Pressure Coefficient,
at 100°, kPa
150
101
• 8.1
3:3
2:54
2:5
1:9
3:3
0:17
2.39 x 10~5 (25°C)
0:0002
133'(114.4°C)
Knw
2.29
1:45
3:38
1:85
1:76
3:37
2:75
4:04
3^72
>5:6
5:2
5:61
Adapted from Verschueren (7).
TABLE 4. FREQUENCY OF OCCURRENCE OF ORGANIC PRIORITY POLLUTANTS IN GAS TRAPS
Organic Compound
Frequency of Detection
Samples
Percent
1 ,1 , 2-Trichloroethene
Dibromomethane
2, i|-Dichlorophenol
2-Nitrophenol
Nitrobenzene
1 , l|-Dichlorobenzene
1 ,2,4-Trichlorobenzene
Naphthalene
Hexachlorobenzene
Lindane
Dieldrin
Dioctyl phthalate
44
32
25
13
10
22
1
14
•3
ND
ND
ND
92
67
52
27
21
46
2
29
6
0
0
0
ND • Not detected.
*Total of 18 samples from gas traps obtained throughout experiments to Day 400.,
-34-
-------
TABLE 5. SUMMARY OF GAS PHASE TRANSFER OF TRICHLOROETHENE AND DIBROMOMETHANE
Trichloroethene
Landfill
Column*
KCR)
2(0
3(0)
MOD
5(OM)
6(OR)
• 7(OLR)
8(OH)
9(OMR)
10(OHR)
Average Con- Range of
centration, Concentration,
ppb ppb
6.48
0;49
135' '
764
385
137
247
267
378
757
BDL
BDL
' 6.25
585
10.3
13:8
58;7
55 :1
87; 6
12;3
- 39.1
- 2.53
- 820
- 958
- 1250
- 878
- 903
- 607
-1182
- 3500
Total Mass
Volatized,
Ug
300
6
3,800
2,600
4,000
1,500
- 720
2,160
2,020
8,900
Dibromomethane
Average Con- Range of
centration, Concentration,
ppb ppb
0
0
0.81
18:2-
•5:1
57"
71
2.0
0;88
0:33
BDL
BDL.
BDL
BDL
BDL
BDL
BDL
BDL
_
—
- 4.85
- 91.2
- 26;2
- 238
- 354
- 8.54
- 4;52
- .86
Total Mass
Volatized,
yg
0
0
22
82
71
830
280
21
6
5
As on Table 1.
TABLE 6. ELUTION ORDER FOR ORGANIC PRIORITY POLLUTANTS
First Day Observed
Compound in Leachate
Single Pass
1 , 1 ,2-Trichloroethene
Dibromomethane
Nitrobenzene ,
1 , 4-Dichlorobenzene
2-Nitrophenol
2 , 4-Dichlorophenol
Naphthalene
1 ,2,4-Trichlorobenzene
Dioctyl phthalate
Lindane
Hexachlorobenzene
Dieldrin
29
29
29
29
36
36
36
54
64
64
410
ND
Maximum Concentration
Final Day Observed Observed in Leachate,
in Leachate mg/L
Recycle Single Pass
29
29
29
36
36
36
. 36
43
64
111
389
ND
*
478
239a
85b
288a
*
*
»
400
410
410
ND
Recycle Single Pass Recycle
*
453
260d
85°
288e
*
x
*
473
*
453
ND
16
•3
26
5
11
15
'7
0.8
2'
0.5
. 0.-15
14
•5
28
6
2
2
1
0.9
1 '
0.3
0;2
ND = Not detected.
*Currently present.
aObserved in leachate from two columns subsequently.
^Reappeared in leachate on Day 260. ,
°Reappeared in leachate on Day 355:
dReappeared in leachate on Day 370'and after Day 480.
eReappeared in leachate on Day 488.
-35-
-------
Although it is difficult to establish
potential attenuation mechanisms for the
organic priority pollutants in terms of
physical-chemical and/or biologically
mediated interactions, all have now
appeared in the leachates from the test
columns and some notion of elution order
can be established. Combining the data of
Table 3 with those of Table 6 suggests
that Kow is a useful parameter for
establishing elution patterns. Hence, all
compounds with log Kow values of 3.4 or
lower were present in the leachate no
later than Day 36, between Days 40 and 120
for log Kow ranging from 3-7 and 5.2,' and
at Day ^00 or longer for log Kow in excess
of 5.6. Additional examination of these
elution patterns will be necessary to
provide a more comprehensive understanding
of these changes and the possible
influences of a transition from acid
formation to methane fermentation under
either single pass or recycle operations.
The opportunity for reinjection of leached
compounds for repartitioning or possible
extended attenuation in the case of the
recycle columns is expected to account for
some of the alternating elution patterns
detected.
CONCLUSIONS
Preliminary analyses of results from
investigations on the codisposal of
priority pollutants with municipal refuse
in simulated landfill columns, operated
under the influence of single pass
leaching or leachate recycle, have begun
to suggest possible attenuating mechanisms.
The removal of heavy metals is strongly
influenced by the loading intensity, the
physical and chemical characteristics of
the admixed waste and leachate transport
medium, the progress of microbially
mediated stabilization, and the presence
and availability of inhibitors and/or
reactive species.
The reducing environment helps to
determine metal speciation and reaction
potential in terms of possible precipita-
tion or complexation. Metallic mercury is
the suggested form of that metal, with a
potential for release through the leachate
or gas phases. Sulfur species are impli-
cated as controlling lead mobility either
through precipitation or complexation.
The distribution of oxidized or reduced
sulfur is determined by the opportunity
for biologically mediated reduction which
may be itself inhibited by the heavy metal
loadings, acid pH conditions and unfavora-
ble redox potentials. These contingencies
should become more evident as the results
of the heavy metal loadings are compared
under the succeeding methane fermentation
phase of landfill stabilization.
The impact of the simulated landfill
conditions on the release or possible
attenuation of the organic priority
pollutants remains uncertain, but several
logical patterns are beginning to emerge.
Solubility and/or reaction in the waste or
leachate and gas transport media in
accordance with physical-chemical proper-
ties is somewhat predictable, log octanol/
water partition coefficient being particu-
larly useful. Alternating periods of
appearance and disappearance of organic
priority pollutants in the leachate or gas
may reflect flow patterns established by
single pass or recycle operations as well
as the potential for physical-chemical
partitioning and possible bioattenuation.
As the microbially mediated processes of
landfill stabilization become more
actively established under the ensuing
methane fermentation phase, the effect of
the imposition of the loadings and the
most probable assimilative mechanisms will
become clearer.
ACKNOWLEDGEMENTS
The research described in this
article has been funded wholly or in part
by the United States Environmental
Protection Agency through Cooperative
Agreement No. CR 812158, to the Georgia
Institute of Technology. It has been
subject to the Agency's review and
approved for publication. Approval does
not signify that the contents necessarily
reflect the views and policy of the
Agency, nor does mention of commercial
products constitute endorsement or
recommendation for use. Technical
guidance from the Project Officer,
Jonathan G. Herrmann, is gratefully
acknowledged.
REFERENCES
1. Pohland, F. G. and S. R. Harper.
Critical Review and Summary of
Leachate and Gas Production from
Landfills. U.S. EPA Cooperative
Agreement CR809997.
-36-
-------
7.
Pohland, F. G., J. P. Gould, R. E.
Ramsey, B. J. Spiller and W. R.
Esteves. Containment of Heavy Metals
in Landfills with Leachate Recycle.
In: Proceedings of the Seventh Annual
Research Symposium, Land Disposal:
Municipal Solid Waste.
EPA-600/9-8l-002a. Philadelphia., PA,
March 1981 , p. 179.
Hughes, W. L. A Physicochemical
Rationale for the Biological Activity
of Mercury and Its Compounds. .Annals
of New York Acad. of Sci., 66, H5^,
1957. '
Mercury in the Environment, Chapter 3-
L. Frlberg and J. Vostal, eds. CRC
Press, Cleveland, OH, 1972.
Sillen, L. G. and A. E. Martell.
Stability Constants of Metal-Ion
Complexes. Special Publication No. 17
of the Chem. Soc., London, 1964, 1150
pp.
Sillen, L. G. and A. E. Martell.
Stability Constants of Metal-Ion
Complexes. Special Publication No. 25
of the Chem. Soc., London, 1971,
839 pp.
Verschueren, K. Handbook of Environ-
mental Data on Organic Chemicals, 2nd
Ed. Van Nostrand, 1983-
-37-
-------
FIELD VERIFICATION OF FMLS--ASSESSMENT OF AN UNCOVERED
UNREINFORCED 60-MIL EPDM LINER AFTER 18 YEARS OF EXPOSURE
Henry E. Haxo, Jr., Robert S. Haxo, Gary L. Walvatne
Matrecon, Inc.
Alameda, California 94501
ABSTRACT
Samples of a 60-mil vulcanized ethylene propylene rubber (EPDM) flexible membrane
liner (FML) were recovered for analysis and physical testing from different locations
within a basin that was being decommissioned after 18 years of service as an emergency
pond for "red-water." Observations on the in-place uncovered liner, description of the
sampling procedure and collection of the samples, and results of the laboratory testing
of the samples are presented.
The properties of the liner samples varied significantly depending on the location
in the basin from which they were taken. Samples taken from the liner on the dike
slope facing south had less extractables (i.e., oily plasticizers) and generally had
higher tensile strength and modulus than the reference FML. The samples taken from the
liner on the bottom of the basin had high extractables and properties comparable to
those of the reference. As no retained sample or analytical and physical test data on
the original unexposed sheeting were available for comparison, data on a 1972 EPDM
liner were used for reference.
No failures were observed in the factory seams of the in-place liner. Most of
the field seams on the slopes opened, resulting in sloughing of the liner. Those field
seams that were high on the slope showed an apparent increase in crosslinks of the
adhesive; however, the adhesive retained tack in most of the tested field seams,
indicating low cure.
Rodents had gnawed holes in the liner from the top surface in the upper slope
areas. There was no indication that the rodents had burrowed below the liner and
gnawed from beneath the liner. The design of the anchor trench using wooden planks was
unsatisfactory, as many of the panels of sheeting pulled out.
INTRODUCTION
Considerable laboratory test data
have been generated over the past several
years with respect to the properties of
FMLs and their chemical compatibility with
hazardous liquids and waste liquids. Such
data are continually being generated in
research and in EPA 9090 liner/waste
compatibility tests. On the other hand,
the data that are available on the effects
on FMLs in service in waste containment,
particularly data in the open literature,
are limited. Several studies have been
undertaken by the EPA to assess liners
that have been removed from service
(Roberts et al, 1983; Nelson et al,
1985; Emcon Associates, 1983; Matrecon,
1983); however, these studies are rel-
atively few in number and the exposures
are limited, both with respect to
length of service and severity of ex-
posure. Much more information on the
field performance of lining materials
in service needs to be obtained and
analyzed in order to develop correla-
tion between laboratory test data
and actual performance. Such data are
-38-
-------
also needed to set criteria for perfor-
mance or magnitude of the changes in
properties that will affect performance.
These criteria are needed for use in
assessing the results of EPA 9090 liner
compatibility tests.
An opportunity arose to assess the
effects on an EPDM liner of 18 years of
exposure in a surface impoundment when an
Emergency "Red-water" Basin located at the
Joliet Army Ammunition Plant in Joliet,
Illinois, was scheduled to be decommis-
sioned. Arrangements were made through
the Hazardous Waste Engineering Research
Laboratory of EPA in Cincinnati for
Matrecon to observe some of the cleanup of
the basin and the removal of the lining
material, and to collect samples of the
EPDM liner which had been in service under
relatively mild conditions that may not be
typical of most landfills or surface
impoundments.
The objective of the work was to
assess the overall condition of the
uncovered unreinforced FML after 18 years
of exposure to the waste liquid and
weather, and to assess the overall condi-
tion of the basin below the EPDM liner.
This paper presents a description
of the basin at the time of sampling, a
description of the sampling procedure,
observations made during the sampling
of the liner, and the results of the
laboratory testing of the exposed samples.
DESCRIPTION OF THE BASIN AND THE LINER
The basin was constructed in late
1967 and early 1968 to act as a surge
pond for "red water" produced as a waste
from TNT production. The pond covered an
area of 3.1 acres and contained "red
water" waste liquid, which was usually
concentrated by evaporation and then
disposed of by incineration. It must be
recognized that the composition of the
wastewater that was contained was highly
variable with time as the basin was used
intermittently. The constituent con-
centration for selected analytes from a
sample collected on May 28, 1981 is
presented in Table 1.
The basin was last used during TNT
production for the Viet Nam war and was
dismantled in May -1985, because it was no
TABLE 1. COMPOSITION OF SURFACE
WATER SAMPLED
Concen-
tration
Analyte (pg/L)
TNT-Related Organics Compounds
2,4,6-Trinitrotoluene (TNT) <0.29
2,6-Dinitrotoluene (DNT) 196.0
2,4-Dinitrotoluene (DNT) 1.00
2-Nitrotoluene 5.7
1,3,5-Trinitrobenzene <2.2
Anions
Nitrite
Nitrate
Sulfate
Phosphate
<250
433,000
6,690,000
390
Heavy Metals
Arsenic
Cadmium
Chromium (hexavalent)
Chromium (total)
Copper
Iron
Lead
Manganese
Mercury
<13.0
8.8
1.5
125.0
141.0
5,360.0
40.2
195.0
<0.35
aSample number: SW109; sample date:
May 28, 1981.
Source: Tom Erdman of the Joliet Army
Ammunition Plant.
longer needed. The capacity of the
basin was 4.06 million gallons. It had
an average depth of about 5 ft, and the
dike slope was 3:1. The liner was an
unreinforced EPDM FML of 60-mil nominal
thickness. It was uncovered, that is,
no soil cover was placed on the liner.
Figures 1 and 2 present a panaromic
view from the southeastern corner of
the basin at the time the liner was
being sampled. The water table in the
area appeared to have been about the
same as that of the basin bottom as a
partially-filled water drainage ditch
ran along the outside of the dike
on the north and the east sides.
Most of the field seams along the
basin slopes failed, but, from what could
-39-
-------
Figure 1. Part 1 of the panoramic view of the basin at
the time liner samples were collected. View
is from the southeastern corner of the basin.
Note the sloughing from the liner and the
hosing of the sludge to maintain its slurry
consistency to make it easy to pump.
Figure 2. Part 2 of the panoramic view of the basin at
the time the liner samples were collected.
View is from the southeastern corner of the
basin. Note the anchor trench failure and the
field seam failures.
-40-
-------
be observed of the basin bottom, the
field seams were mostly intact; we did
not observe a single failure of the fac-
tory seams. (Figure 3 shows a partially-
opened field seam on the slope.)
Figure 3. Partially-opened field se.ajn or,
the '.slope. Note the apparent
tension, and shrink.ag.e away, from
the seam. Nearby factory,seams
are intact.
The anchor' trench along the berm top
was completely 'inadequatel for anchoring
(Figure 4) the top of"the.liner. In com-
bination with the failure of the field
seams, this poor anchorage "resulted in
large sections of the basin slopes not
being covered by the liner. The extent of
failure of the FML is documented in the
site photographs (Figures 1, 2, and 5).
Gas generation below the liner re-
sulted in the formation of "whales" or
areas of the liner which lifted off the
floor of the basin. No means for bleeding
off the gas appeared to have been incor-
porated into the pond design. It was
reported that these "whales" were punc-
tured to release trapped gases shortly
before the liner sampling was started.
An attempt had been made earlier to
relieve the trapped gases by attaching
vent pipes to the liner, but these vents
appear to have been ineffective.
The bulk of the liquid in the basin
had been removed before the sampling was
started and the cleanup of the sludge
(which was drying) was underway. By
dragging a sled across the bottom, the
sludge could be kept wet (Figures 1 and
2) so that it could be handled, as a
slurry. It was also felt that the
residual TNT in the sludge would be less
active when wet.
SAMPLING OF LINERS FOR TEST
For assessing the overall effects
of exposure and aging on the properties
of an exposed liner from a lined surface
impoundment, it is desirable to obtain
samples that have been exposed to a range
of conditions from the mildest condition
that is believed to exist in a given
impoundment to that believed to be the
most severe exposure (Nelson et al,
1985). To meet these requirements, we
tried to:
• Obtain a 2-ft wide "strip"
sample that extended from the
anchor trench down the slope and
onto the bottom of the impound-
ment. The location should be on
; the north side, facing south.
This sample would yield suf-
ficient material for testing the
principal areas of exposure that
a liner would encounter in an
impoundment, from the top to the
bottom.
• Collect several samples of fac-
tory and field seams.
>' : • In addition to the "strip"
sample, collect samples of
the completely submerged liner
with seams, particularly at the
lowest point of the liner, e.g.,
the sump area.
• Collect any samples that showed
abnormal exposure or other ab-
normalities.
-41-
-------
Figure 4. Anchor trench construction. This photograph
shows the top of Sample A immediately prior to
cutting.
Figure 5. Failure of the anchor trench due to the
pullout of the membrane.
-42-
-------
Four major liner samples were ob-
tained and were designated Samples A, B,
C, and D. In addition, seven smaller
samples containing holes, apparently
caused by rodents, were also collected
(Figure 6).
The reason why each of the major
samples was taken are discussed briefly
below:
• Sample A was the major sample
from this liner. It was taken
from the northeastern dike
(southern exposure) and extended
from the anchor trench at the top
of the dike, down the dike, and
on to the basin bottom. A factory
seam extended most of the sample
length and a field seam extended
into the bottom section. Due to
its size, the strip was cut into
three sections. The schematic
drawing (Figure 7) shows the
different sections and areas from
which test specimens were drawn.
The drawing also shows the seams
and the locations that were
tested.
• Sample B included a field seam
which was partly intact and
partly failed. It was taken from
the northeastern dike between
Sample A and the northeast corner
of the basin.
• Sample C was from the basin
bottom. Before collection, it
was covered with very wet sludge
and exposed to very wet mud on
the underside. A small strip cut
from the liner adjacent to Sample
C was collected for volat.iles
determination.
• Sample D included a partly
intact/partly failed field seam
from a "whale." Sample D was
chosen because "whales" are
considered to be sites of stress
on the liner and because of the
presence of a field seam.
The locations where the samples were
collected are shown in Figure 8, a
simplified drawing based on the "as
built" drawing of the basin.
Figure 6. Holes gnawed by rodents in the liner at the
top of the slope. The 1 and 2 refer to
samples cut from the liner.
-43-
-------
Seam test
.area
p&l^JI
~*X*X*X*X*X*X*Xv
- Factory
seam
mmsmm
%iiA1-3^!;
- Factory
seam
r-~.'v.'v.;rr.~
*A2-5i:
tow
seam SeaJ,test
area
Field
seam
Section buried in
the anchor trench
Weather-exposed
section from the
top of the dike
Sections exposed
alternately to
weather and waste
Section at the
bottom of the
dike slope
Field seam connecting
panels on the dike
slope and the basin
bottom
Section from the
basin bottom
Figure 7. Schematic drawing of Sample A,
a strip sample cut from the
liner on the northeastern dike
of the basin. The strip ex-
tended from the anchor trench
down to the bottom of the
basin.
No retained sample of the original
liner was available for use as a baseline
reference. The top portion of Sample A,
which was buried in the shallow anchor
trench on top of the dike, probably had
only a modest exposure to either waste or
sun, but it appeared to have aged and
lost extractables and thus could not be
used as a baseline reference.
The temperature on the day the
samples were collected was in the
mid-70's. The weather varied from clear
in the morning, to cloudy in the early
afternoon, to clear again in the late
afternoon.
The samples cut from the liner were
kept as cool as possible, particular-
ly those that were removed from under
sludge or soil (i.e., from the anchor
trench) or had been below the liquid.
The samples were washed without scrub-
bing to remove loose material, then
rolled and placed in aluminum tubes. A
small amount of water was added to each
of the tubes that contained moist
liners and the tubes were sealed at the
ends with metal caps.
TESTING OF SAMPLES
Analytical and Physical'Properties
The liner samples were photographed
and measured at the laboratory; Sample A,
the "strip" sample, was tested in the
following areas, as shown in Figure 7.
o Section A1, the upper section of
the "strip", was tested in four
areas. Test areas were design-
ated Samples A1-1, A1r2, A1-3,
and A1-4. These samples had
been exposed in the anchor
trench, at the top of the slope,
at mid-slope, and at the toe of
the slope, respectively.
o Section A2 was tested for ana-
lyticaland physical properties
in an area that had been at the
bottom of the slope. The test
area was designated Sample A2-5.
o Section A3, which had been ex-
posedit" the bottom of the
basin, was tested for physical
properties. The; test area was
designated Sample A3-6.
-44-
-------
RODENT HOLES 1 AND 2>
RODENT HOLES 3,4 AND 5>
-80'
SAMPLE
\ A
SAMPLE
B
Toe of Dike
FLUME
Top of Dike
101 Typical
Figure 8. Schematic drawing of the basin showing the location
where, the liner samples were collected.
The other three samples were tested as
follows:
• Sample B, taken at a field seam,
was tested in two areas designated
Samples B1 and B2 which correspond
to two different layers of the
same sheeting. Sample B1 was ex-
posed to the weather and possibly
waste; Sample B2 was the underflap
part of the intact portion of the
field seam.
• Sample C--the strip taken ad-
jacent to Sample C from the
bottom of the basin was analyzed
immediately for volatiles upon
receipt at the laboratory.
Physical tests and additional
analyses were later performed on
the main portion of Sample C.
• Sample D, cut from a "whale", was
tested for analytical and physical
properties and the intact part of
the seam was tested in the peel
mode.
The properties of the exposed FMLs
that were measured and the specific
test methods that were used are presented
in Table 2. The results of the testing
are presented in Table 3.
TABLE 2. PROPERTIES AND TEST METHODS
Property
Test method3
Tensile properties
Tear strength
Puncture resistance
Hardness
Volatiles content
Extractables content
Ash content
Specific gravity
Thermogravimetric
analysis
ASTM D412
ASTM D624
FTMS 101C,
Method 2065
ASTM D2240
MTM 1
MTM 2 (with methyl
ethyl ketone)
ASTM D297
ASTM D792
Haxo (1983)
aASTM = American Society for Testing and
Materials; FTMS = Federal Test Method
Standard; MTM = Matrecon Test Method
(Matrecon, 1983).
Seam Strength
Many of the areas tested for physi-
cal properties included factory or
-45-
-------
z
i
Si
5.
UJ
X
s
>.
W
M
1
i
CC
Q
H-
0
UJ
d
U
UJ
in
0=
5
J
ul
_j
r-4
1
O
VJ
&
UJ
Ul
cc
UJ
.
K\
UJ
r
r—
.Q
CMS U
r* Q co
CNC_ C
r- UJ ••-*
O ^J
rH
O.BQ
S
cB
CO XI
a o
CD J3
8.1 gb
o a jj a cj
>-f iJ >^ca
Cd *O OB
-< m — i
C Q
o •**
U-
ei ir a. o «-
c xi o >>ca
.*--- r- CO
-J I rH
§VO
1
4J K>
x) > eo
»-t o CL
£ "* t5
c
•"4 $
U xi
3°
X
b
Q.
O
O.
8
t-i
CU
|
O.
rH
CO
u
X)
(-1
ts
5
CO t- CO t-V
t*"\ H
*> *i-f X)
T3 > r-4
a? 03 CO CO
QJ U d
-»H Dl O
0) CD U Q.
rH JJ -H
•H O &? U- rH
JJ CD •«-• CO
CD b - 0 U
rH X) JT CO -H
o x eg o. to
1
lAO OO OO OO VO T— O ON VO CO • P—
fA lA CMO IACM O-3" O T- vo • -^ • • lA
VOIA IAIA rA rA cor- CMCM vo « 3 • N
C C 4-> C ••-• J3
O O CJ -H ^H rH
•H .H C CO CO
•H COCO D.^* (-« oT C
CL J£ C C CD4JCDCCD'o
CO O O -H OCOtjCCXOlOl
- CO rH rH CX C XI **J* C C
^ t-t CO CO CL COrHCOCOCJ<-r-«.H
CO g? 5-5 *< COECDCOCDJZOCaCD
t-t JJ O O JZ -H [-. (-1 JJ (li 0 CD
XI CO O O 4-> CO «4J C 4-> 3 tj t-t
T— CM Ol COtOCOOCOrHQ
4J C CtjCO-H-H-O'O
CD O X> X) CO CO E JJ E E ^ C C
•H CO CO t-t COC3CODI COOO
t— C CO CO CO CO 3CJ-HC-HO CO CD CO
•H Ol CO CO X>-HXOXt— CCOCO
co cr co CD t4 cjx:corHco "oil
C O t-t {-, CO Cr-ELdS tj«-LA
CD rH 4J x> CD 3 CO
i— uj in co i— a. x
to
to
o
• f— t
£ 1
CO
CO t-t
CL CL
o
rH 0
to jj
4- CO
0 C
•H
CL 4->
0
4J rH
E B
t-t W
c»-
C -H
CD
j^ -a
CO CO
CO
rH (0
a
CO CO
C CO
O JJ
•H O
JJ (0
C rH
0) rH
t-t 0
•H O
c
CD CO
C CO
•H XI
111
CD jj ca
x: co c
4J x: -i-1
CD S
XI C_) CD
O -O 0
4-> C -O
CO
T3 CO
CO lA CD
> 1 1— 1
CD CM -H
.H <=C 4_)
rH CO
CD t*- rH
XI O O
CO W
•H CD (-J
rH O
4-1 Q. t*-
CO CD
C 4J
•H rH O
CO rH CD
"b E rH
CO CO O
CJ C
3 O lA
|
JJ n • CM
CO >-. CD <
CD iH C
JJ • CD O 0-
--^ > 4J O
d- lA -H CD
O CO 4-) -i! CO
Ov CJ i-H
C • r- CD rH Q.
O 4J D. >% E
•H C "CO XT CO
CJ rH CO (-t CD
t-> 4J - i-H rH
•H C CD "O X CO
Q O CD XI E
•HOC XI CO
JJ X -H CO
* -H CO CO E CD
IE 0 4->
lA E CO -H C
1 «5 CO SO
>> O CO "0*0
T) E 2: 2 CD Q)
CO (4 C C
> < tj S? -H -H
•H CD fA E CO
(D JJ C CM b JJ
CJ CD -H • CO XI
t-t rH 1 O
O 1 -O TD CO
co n CD c CD
i— I U CO CD S
CL co c f-4
§.c CD a? co a?
JJ f-« fA S CO
CO CD CO • CM
4- (4- • CO CO •
>•. O CD O CO CO CM
[_l f_, r-| rH CM
CO C d- -H XI
C CO CD O JJ CO U-
•H E C CO 4J O
E "O -H CO rH O
• H b i-H CD O CO CO
rH UJ CO 3 > (-( 3
CD CO rH JJ i-H
b • CO CO o LJ >
CO XI CJ 13 CD
-46-
-------
field seams that were tested for seam
strength in shear and peel modes. The
location and type of seam samples that
were tested for seam strength are indi-
cated in Table 4.
Seam strength in shear mode was
measured in accordance with ASTM D882 and
D3083, modified for testing exposed FMLs.
Testing in peel mode was performed in
accordance with ASTM 0413 in 90° peel.
In both modes of testing, 1-in. wide
strip specimens were tested at a jaw
separation rate of 2 inches per minute.
The results of the seam testing are
presented in Table 4.
SEAMS
The seams in the liner were prepared
both in the controlled environment of the
factory and in the uncontrolled outdoor
environment of the field. Roll stock
of EPDM sheeting was manufactured and
fabricated into large panels at the
factory that were then installed in
the basin. Vulcanized seams were made
to join the sheeting into panels at the
factory, while vulcanizable adhesives
were used to join the panels to form the
liner in the field.
The vulcanized factory seams ap-
peared, after exposure to "red water" and
weather, to be similar to new factory
seams. The factory seam was 3.2-in.
wide and the edges of both the top and
bottom sheets were beveled to a thickness
of about 30 mils up to 0.75 inches in
from the edge of each sheet. The field
seam was 6-.in. wide and was bonded with a
vulcanizable adhesive. A low-temperature
vulcanizable tape was placed along the
edge of the top sheet. In addition to
its function in bonding the sheets
together, the tape also served to round
the edge of the top sheet. However, many
of the field seams displayed an opening
in the tape along and adjacent to the
edge of the top sheet. This opening
TABLE 4. SEAM STRENGTH IN SHEAR AND PEEL MODES OF 60-MIL EPDM SEAM SAMPLES
COLLECTED FROM THE EMERGENCY "RED-WATER" BASIN
Strip Sample A (type
Anchor Top
trench slope
A1-1 ' A1-2
Seam test (Factory) (Factory)
Seam strength in shear3
Maximum strength^3, ppi
Strength at break, ppi
Locus of break0
FTBd
Non-FTBd
64.0
64.0
0
5 AD
Mid-
slope
A1-3
(Factory)
65.6
65.6
0
5 AD
of seam)
Lower I
slope
A1-4
(Factory)
68.1
68.1
1 SE
, 4 AD
3n basin
bottom
A2-5
(Field) (
69.2
69.2
0
5 AD
High on On basin Top of
slope bottom "Whale"
A3-6 B1/B2 C , D
.Factory) (Field) (Field) (Factory) (Field)
64.4
64.4
0 . ... ...
3 AD
Seam strength in peel6
Maximum strength, ppi ' 11.7 6.5 7.8 8.2
Locus of break0
FTBd 0000
Non-FTBd 5 AD 5 AD 5 AD 5 AD
9.2
0
5 AD
10.1
0
5 AD
0
5 AD
17.6
0
5 AD
8.7
0
5 AD
8.9
0
5 AD
aASTM D882/D3083, modified; five specimens tested per sample, except where otherwise noted. Two specimens of Sample B2
slipped in the clamps during testing; results declared void.
DMaximum value corresponds to tensile strength at break for all specimens tested.
GLocus-of-break determined from the following code:
Locus of Break
Description
Classification
FT8e
CL Break at clamp edge
SE Break at seam edge
BRK Break in sheeting FTBe
AD Break in adhesion non-FTBe
dFTB = film-tear bond.
eASTM D413, modified; five specimens tested per sample except where noted otherwise.
-47-
-------
may have been caused by differential
shrink and swell of the sheeting and
the adhesive.
SELECTION OF A BASELINE REFERENCE
No retained sample of the liner was
available that could be used as a baseline
reference. The samples that were re-
covered from the basin were 18 years of
age; consequently, there was a question as
to whether any of these samples was suit-
able for use as a baseline reference.
However, we had data on EPDM FMLs that had
been produced in 1972 and tested in
earlier work performed by Matrecon for the
EPA on a study of liners for municipal
solid waste landfills (Haxo et al, 1982;
Haxo et al, 1985). A review of the
analytical results for an FML produced in
1972 indicated that it was essentially the
same as that of the liner installed in the
basin in 1967. A comparison of analytical
properties of the 1972 EPDM liner with the
liner recovered from the basin is given in
Table 5. The data as a group constitute
a fingerprint of the liner, such as des-
cribed by Haxo (1983).
The physical properties for the
1972 membrane are presented in Table 3.
Tensile strength, stress at 100,°o elon-
gation, and puncture resistance are
comparable to the data obtained on the
liner taken from the basin, assuming
somewhat higher extractables content.
Using these data, one has a baseline
reference against which to compare the
effects of the exposure on the liner
samples recovered from the basin.
RESULTS AND DISCUSSION
Data on the samples taken from the
various sections of the "strip" sample
can be used to assess the effect of the
different exposures from the top to the
bottom of the basin. Some of the basic
conclusions that can be drawn are:
• The extractables decrease with
increasing distance from the
bottom of the basin which has
been under "red water" or sludge.
• The sheeting on the top part of
the slope, which was exposed to
the sun, contains about one-third
TABLE 5. COMPARSION OF ANALYTICAL
PROPERTIES OF EXPOSED SAMPLE AND
BASELINE REFERENCE
1972
Goliet Liner
Analysis Sample C No. 8
Extractables3, % 21.62 23.41
Thermograyimetric
analysis:
Polymer + oils, % 57.2 57.4
Polymer (cal-
culated)13, % 35.6 34.1
Oil (from extract-
ables), %
Carbon black, %
Ash, %
Total
Ash, K
Specific gravity
21.6
35.2
7.6
100.0
7.24
1.183
23.4
35.0
7.5
100.0
6.78
1.173
aExtractables consist of oils + extract-
able curatives and antidegradants
(determined with methyl ethyl ketone).
^Calculated by subtracting the extract-
ables from the thermogravimetric analy-
sis (TGA) determination for polymer +
oil.
less of oily plasticizer than was
in the sheeting on the floor of
the basin. The latter sheeting
has the highest extractables
values. These high values ap-
proximated that of the "baseline
reference" (i.e., the 1972 EPDM
liner). Also, the tensile
strength, modulus (stress at 100%
elongation), and puncture resis-
tance values tend to increase
with decreasing extractables,
that is, with distance up from
the bottom of the basin, as is
shown in Figures 9 through 11.
On the other hand, the elonga-
tion at break decreases as the
extractables decrease during the
exposure, as shown in Figure 12.
-48-
-------
1900
Extractables,%
Figure 9, Tensile at break of the samples
of exposed FMLs as a function
of their extractables. Tensile
data are the averages of the
values obtained in both machine
and transverse directions. The
samples with the low values for
extractables were cut from the
liner at the top of the slope.
Those with the high values were
cut from the liner on the
bottom. R = Baseline Reference
EPDM, produced in 1972.
100
- so
26 24 22 20 18 16 14 12
Extractables, %
Figure 11. Puncture resistance of the
samples of exposed FMLs as a
function of their extract-
ables. Puncture resistance
is the maximum stress value
calculated for 100-mil
thickness of the FML. The
samples with the low values
for extractables were cut
from the liner at the top
of the slope. Those with the
high values were cut from
the liner on the bottom. R =
Baseline Reference EPDM,
produced in 1972.
900
#_ 800
o a- 700
•Si 600
(» 8 500
£ o 400
300
55
200
'26 24 22 20 18 16
Extractables,%
14
12
600
500
400
o> 300
5
UJ
200
26 24 22 20 18 16
Extractables, %
14
12
Figure 10. Stress at 100?o elongation
(S-100) of the samples of
exposed FMLs as a function of
their extractables. S-100
rfata are the averages of the
values obtained in both
machine and transverse direc-
tions. The samples with the
low values for extractables
were cut from the liner at the
top of the slope. Those with
the high values were cut from
the liner on the bottom. R =
Baseline Reference EPDM,
produced in 1972.
Figure 12. Elongation at break of the
•samples of exposed FMLs as a
function of their e^tract-
ables. Elongation data are
the averages of the values
obtained in both machine and
transverse directions. The
samples with the low values
for extractables were cut from
the liner at the top of the
slope. Those with the high
values were cut from the liner
on the bottom. R = Baseline
Reference EPDM, produced in
1972.
-49-
-------
• In the case of sheeting protected
at the seams by an upper layer of
sheeting, the extractables are
greater and the tensile strength,
modulus, and puncture resistance
are less than those of the upper
layer. These differences again
appear to reflect the increased
plasticizer content of the pro-
tected sheeting.
• The factory seam results tend to
show increased strength values in
both shear and peel modes with
increased distance from top to
bottom down the slope. The peel
adhesion values increased from 6.5
to 10.1 ppi with distance down
from the anchor trench; all the
failures in the tests were ad-
hesion failures between the
sheetings.
• In the case of field seams, there
is a difference between the ex-
posure locations. A high value
of 17.6 was obtained on an un-
failed part of the seam that was
high on the slope. This seam had
partially failed during exposure.
Testing was performed on an un-
failed portion of the seam. In
this case, the adhesive appeared
to have crosslinked more than had
the seams in the other locations.
In all cases, the preponderant
failure was a cohesive failure in
the adhesive.
ACKNOWLEDGMENTS
The work reported in this paper was
performed under EPA Contract 68-03-3169
with the Hazardous Waste Engineering
Research Laboratory of the U.S. Environ-
mental Protection Agency, Cincinnati,
Ohio. The EPA Project Officer was Mr.
Robert E. Landreth, who located the site
for this study. The author wishes to
acknowledge the participation of Mr.
Robert Hartley of the EPA in the recovery
of the samples and Mr. Thomas Erdman of
the Doliet Army Ammunition Plant, with
whom arrangements for access to the basin
were made and who furnished a drawing of
the basin layout and the analytical data
on the wastewater.
REFERENCES
Emcon Associates. 1983. Field Verifi-
cation of Liners from Sanitary Land-
fills. EPA/9/-83-046. U.S. EPA,
Cincinnati, Ohio, p 32.
Haxo, H. E. 1983. Analysis and Finger-
printing of Unexposed and Exposed
Polymeric Membrane Liners. In:
Proceedings of the Ninth Annual
Research Symposium: Land Disposal,
Incineration, and Treatment of
Hazardous Waste. EPA-600/9-83-018.
U.S. EPA, Cincinnati, Ohio. pp.
157-171.
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. U.S. EPA, Cincinnati,
Ohio. NTIS No. PB 83-147-801.
Haxo, H. E., R. S. Haxo, N. A. Nelson, P.
D. Haxo, R. M. White, and S. Dakes-
sian. 1985. Final Report: Liner
Materials Exposed to Hazardous and
Toxic Wastes. EPA-600/2-84-169.
U.S. EPA, Cincinnati, Ohio. NTIS
No. PB 85-121-333.
Matrecon, Inc. 1983. Lining of Waste
Impoundment and Disposal Facil-
ities. ' SW-870 Revised. U.S. EPA,
Washington, D.C. 448 pp. GPO
#055-00000231-2.
Nelson, N. A., H. E. Haxo, and Peter
McGlew. 1985. Recovery and Testing
of a Synthetic Liner from a Waste
Lagoon After.Long-Term Exposure.
In: Proceedings of the Eleventh
Annual Research Symposium: Land
Disposal of Hazardous Waste.
EPA/600/9-85/013. U.S. EPA,
Cincinnati, Ohio. pp. 296-306.
Roberts, S., N. A. Nelson, and H. E.
Haxo. 1983. Evaluation of a Waste
Impoundment Liner System After Long-
term Exposure. In: Proceedings of
the Ninth Annual Research Symposium:
Land Disposal, Incineration, and
Treatment of Hazardous Waste. EPA-
600/9-83-018. U.S. EPA, Cincinnati,
Ohio. pp. 172-187.
-50-
-------
GEOSYNTHETIC DESIGN CONSIDERATIONS FOR
DOUBLE LINER SYSTEMS
Gregory N. Richardson
S&ME
Gary, North Carolina 27511
Robert M. Koerner
Drexel University
Philadelphia, Penn. 19104
ABSTRACT
The "minimum technological requirements" of the Hazardous and Solid Waste Amendments
of 1984 require a double liner system in most hazardous waste land disposal cells and
surface impoundments. Ensuing guidance recommended the use of two flexible membrane liners
(FML). Each FML in a landfill and the bottom FML in a surface impoundment is covered by a
leachate collection/removal (LCR) system to aid in preventing leachate from standing on
the FMLs. This paper reviews design considerations for the FML and LCR systems within the
.double liner system.
Potential failure modes for geosynthetic FMLs and LCRs are described in this paper
and design procedures are reviewed for each of the failure modes. Each design procedure
calculates the actual service stress or flow conditions and compares this required
performance to the limiting performance of the component itself. The limiting performance
is typically obtained from laboratory testing. A Design Ratio (DR) is defined'as the ratio
of the laboratory limiting performance'divided by the calculated service performance.
OVERVIEW
On November 8, 1984, the Resource
Conservation and Recovery Act (RCRA) was
amended by the Hazardous and Solid Waste
Amendments (HSWA). Among the provisions that
went into effect were minimum technological
requirements for hazardous waste land
disposal facilities. HSWA requires new units
and lateral expansions of existing
facilities to use two liners with a leachate.
collection system above (in the case of a
landfill) and between such liners.
Draft EPA Minimun Technology Guidance
(MTG) (1,2) for liners and LCRs was
published on May 25, 1985. Proposed
codification of the MTG is outlined in the
Federal Register, Vol.51, No.60, March 28,
1986.
The MTG standard 'double-liner' system
for a landfill is shown on Figure 1 and
consists of a primary LCR and FML system
over a secondary LCR'and a composite
FML/clay secondary liner. The primary LCR
minimizes the amount of leachate that is
allowed to stand on the primary FML. The
primary FML must be designed to allow no
more than de minimis quantities of leachate
to pass through the liner. The•secondary LCR
collects leachate that has passed through
the primary FML and in this fashion bears
'witness' to the integrity of the primary
FML. The secondary FML/clay liner must be
designed to prevent greater than de minimis
quantities of leachate from leaving the
system during the minimum 30-year post- •
closure monitoring period.
MTG MINIMUM DESIGN PROPERTIES
The draft MTG provides minimum stand-
ards for each of the components within the
recommended double liner system. Shown on
^Figure 1, these are as follows:
-51-
-------
Filter Layer
Primary FML
SecondaryvFML\
Compacted Clay Liner
Native Soil Foundation
Figure 1 MTG Double-Liner System
o Primary LCR: This system is placed above
the primary FML and is separated from the
waste by a filter layer. It consists of a
minimum of 30 centimeters (cm) of drainage
stone having a minimum hydraulic
conductivity of 10~2 cm/second. A 15-cm
thick filter layer is placed between it and
the waste, and an internal pipe drain system
must be designed to keep less than 30 cm
maximum head of leachate acting on the
primary FML .
o Primary Liner: The primary liner must be
of synthetic material at least 0.75 milli-
meters (mm) thick. This liner must be
designed to allow no more than de minimls
quantities of leachate to pass through to
the secondary LCR. A composite soil/FML
primary liner is allowed under proposed
codification.
o Secondary LCR: This witness system has the
same minimum properties as the primary LCR
but lacks the overlying filter layer.
o Secondary Liner: This is a composite
FML/clay liner having an FML with the same
minimum thickness as the primary FML,
0.75mm, and an underlying clay having a
maximum permeability of 10~7 cm/sec and a
minimum thickness of 90 cm. The secondary
FML is typically of the same material and
thickness as the primary FML.
DESIGN CONCEPT - FML
A large number of failure scenarios can
be developed for an FML. These failures have
in common the generation of large tensile
stresses within the EML. Design for these
failure mechanisms is based on the stress-
strain curves for the particular polymeric
material used in forming the FML. For HOPE,
the short term stress vs strain curves yield
a well defined yield stress while for other
polymers such as PVC or CPE, the stress vs
strain curves show no distinct yield, Figure
2. Allowable stress limits for such curves
are based on a limiting strain level. In
both cases, the design philosophy is that a
significant amount of FML deformation
capacity remains even if the yield stress or
strain is reached.
Figure 3 presents four tensile failure
mechanisms for an FML. All mechanisms can be
quantified on the basis of free-body
diagrams that sum the forces or stresses
parallel to the surface of the FML. These
mechanisms include the following:
o Dead Weight - The self weight of the FML
places the FML in tension as it is draped
down the sideslope. A high value of DR,
greater than 10, insures that excessive
elongation of the FML does not occur during
placement of the FML. Fortunately, the
minimum MTG thickness is normally adequate
to generate this level of DR unless
extremely steep slopes are present.
-52-
-------
Thickness = 0.75 cm
To 1180%
To 500%
To 445%
40 80 120 160 180
Strain - percent
Figure 2 Stress vs Strain for FML
o Sliding Stability - Each synthetic layer
forming the liner systems forms a potential
shear failure plane. Shear is transferred to
and from each layer through the friction or
adhesion that exists between that layer and
the adjacent layers. Shear stresses in
excess of that which can be transferred by
these surface forces must be resisted in
tension by that layer. This reduces the
shear stresses transferred to underlying
layers.
o Waste Lift Stability - the placement of
waste against the liner can produce
significant downdrag forces on the liner
system. DR values for such problems can
easily fall below one (3) if lift heights .
exceed 3 meters.
o Settlement of Liner - Settlement of the,
subgrade beneath the liner system can lead
to excessive strains within an FML (4). Such
settlement could result from the collapse of
a drainage line or the uneven densification
of poorly compacted soils. The designer must
predict subgrade settlements and verify that
an adequate DR will exist under these
conditions.
With the exception of settlement, these
failure mechanisms are also appropriate for
evaluation of synthetic LCR systems.
With the exception of the settlement
mechanism, all of the above design concerns
predict a tension force T acting in the
plane of the FML at the shoulder of the
sideslope. This force must be resisted by
anchorage of the FML beyond the shoulder.
Such anchorage is commonly provided by
running the FML beyond the shoulder and
either burying the end of the FML in a
trench or by simply placing sand on top of
the FML. Typical FML anchor geometries are
shown on Figure 4. The maximum anchorage
tension, Tmax, for the horizontal anchor is
given by
qLtan 8
cos - sinS tan '%>
S
(1)
where S is the friction angle between the
subgrade soil and the FML. The maximum
anchorage of the trench system cannot be
rigorously calculated at present but can be
bounded as follows:
Ttch = Tho
(Ka+K' JtanS [0.5 If d? + qd]
cos ^ - sin^ tan &
(2)
where d is the depth of embedment, Ka is
the active earth pressure coefficient and
K' is the at-rest or the passive earth
pressure coefficient. The actual anchorage
-53-
-------
•S.eiu-
DEAD WEIGHT
\vi •FVAL
« V4
VI
FM.L-
SLIDING STABILITY
FML (T)=
vgetx
WASTE LIFT STABILITY
w
10
0-1 O'Z. . 0.1
SETTLEMENT for LINER
Figure 3 FML Tensile Failure Models
-54-
-------
Surcharge = q
Horizontal Anchor
Load from above
Point load with
hydrostatic
Protective layer
Shallow "V Trench
Support layer
' Point load without
hydrostatic
Figure 5 FML Compressive Stresses
X
X
X
Anchor Trench
Figure 4 FML Anchorage
capacity lies between that predicted hy the
two values of K'. The design is then based
on the smaller anchorage force to predict
DRs based on pullout and the higher anchor-
age force to calculate stresses in the FML
itself.
In service, the FML is also subjected
to large normal stresses due to the weight
of the overlying waste. These stresses tend
to push the FML into the void spaces within
.the underlying subgrade. Knipshield (4)
portrays the stresses acting on a localized
portion of the FML as shown on Figure 5.
Large subgrade particle size may create
void spaces large enough to generate
localized failure. While this situation
could be minimized by placing a geotextile
beneath the FML, this would destroy the
intimate contact bewteen the FML and the
underlying clay. The limiting normal stress
is determined from laboratory tests that
simulate the field system described above.
DESIGN CONCEPT - LCR
Geosynthetic LCR systems play an
important role in the facility by collecting
leachate at any location on the liner and
conveying it to a low point or sump where it
can be removed. The design of the LCR based
on in-plane flow is controlled by both the '
minimun MTG properties and the requirement
that no more than 30 cm head of leachate
may act on the underlying FML. The head ,
acting on the FML is controlled by the rate
at which leachate is being generated and
collected within the system, the hydraulic
properties of the LCR, and the spacing of
the collector pipes within the LCR. These
parameters are shown on Figure 6. Early work
(5,6) related these parameters to the rate
of leachate generation which is commonly not
known. The maximum head acting on the FML is
then given by
L/C"
Hmax= ~~~~ t
tano<
/•
/tan2«X
(3)
c ]
where c is defined as the inflow rate
divided by the hydraulic conductivity of the
LCR. This'method has been supplemented by an
alternate procedure (7) that is based on the
percolation velocity of the leachate. The
maximim leachate head using this method is
given by
(4)
-55-
-------
whore e Is the percolation velocity based on
conversion of the annual percipitation rat.e
into a uniform velocity (cm/sec).
A synthetic LCR may be composed of a
homogeneous material, e.g. a thick nonwoven
needled geotextile, or a composite formed of
a core that provides planar flow capacity
and a surface geotextile that acts as a
filter to prevent the adjacent soil from
intruding and blocking the core.- The planar
flow capacity of the LCR is defined by
Careys equation as
Kp i A
q - Kp [dh/L] W t
(5)
(6)
wnere Kp is the permeability in the plane of
the LCR, dh is head loss, L is flow length,
W is the width of the flow path, and t is
the thickness of the LCR. Equation (6) is
typically expressed as
q - 6 [dh/L] W
(7)
where 9 is defined as transmissivity and is
equal to the product of Kp and t (8).
The transmissivity of an LCR can be
reduced by compression of the core and
Intrusion of adjacent geotextiles or geo-
membranes due to soil pressure. These
effects can occur elastically as normal
loads are increased on the LCR or plastical-
ly over time in the completed cell.
Laboratory data are shown on Figure 7 for a
typical synthetic LCR. These data reflect
only the elastic reduction of transmis-
sivity. Long-term tests are rarely performed
to evaluate the future reduction in trans-
ralssivity due to plastic deformations.
Procedures (3) have been presented, however,
for estimating these losses based on an
analysis of core and filter material creep
properties.
q = inflow
.11 mm 111 m i m m m
LCR
Clay Liner
Figure 6 Calculation of Drain Pipe Spacing
.01
O
o
M
Tr
.
0
0
o
H*
.00001
Increasing
Confining "
Pressure
Figure
0.25 0.50 0.75 ' 1.0
Hydraulic Gradient
7 Transmissivity Data
Synthetic LCR
Leachate must first flow through the
filter layer of a synthetic LCR before it
can be drained away. The filter layer must
be selected so that it will filter out the
soil particles from the,adjacent soil and
yet-hot become clogged by these same soil
particles. No true design procedure is at
present available to evaluate filtration and
clogging potential. There are, however,
index tests that indicate the tendency of a
given fabric to filter or be clogged by a
given soil.
The filtering ability of a. fabric has
traditionally been related to the Apparent
Opening Size (AOS) of the fabric. The AOS is
defined as the diameter of glass beads with
5* retained in the fabric after shaking. AOS
is appropriate for woven and lightweight
fabrics but is questionable for heavy weight
nonwoven fabrics. In general, filter
criteria guidelines are of the form
°fabric
(8)
where Ofabric is usually the AOS of the
fabric, dsoii is a soil particle diameter
obtained from a grain size analysis, and A
is an emperical factor. One example of this
concept is the filtration criteria by Giroud
as presented in Table 1. This method (9)
incorporates grainsize data and relative
density of the soil. Assuming the fabric
manufacturer supplies the AOS for the
: fabric, the filtration design requires
;knowledge of fundamental properties of the
site-specific soil.
-56-
-------
Table 1 Geotextile Filter Criteria (9)
Relative Density,Dr
Loose(Dr<50j£)
Intermediate
(50£80£)
Dense(Dr>
13
095<(9d50)/CU
095<(13.5d50)/CU
095<(18d5o)/CU
Where Dr is relative density, dgo is the grain size corresponding to 5055
passing, 09g is still equal to the AOS of the geotextile, and CU is the
coefficient of uniformityCd^o/dio) of the soil.
Clogging of the fabric- by soil grains
occurs with time. The clogging potential of
a given fabric and soil is evaluated in the
laboratory using the gradient, ratio test or
longterm flow tests. The equipment for this
test consists of a soil column resting on a
geotextile as shown schematically in Figure
8. The test does not reproduce in-sltu
conditions and as such is only an index
test. As water is run through the column,
the. fabric becomes clogged and a hydrostatic
pressure gradient develops across the
fabric. If the gradient ratio, as defined on
Figure 8, exceeds 3, then there is a
potential for clogging of the fabric by the
soil. Recent studies (10) have shown that
•7
Leachate
\ ds&y/&y/Ad 'tf^t**
\ Soil \ jl"
^yfggC^jttl
^—Fabric
|
'i"+fabric
HI
H2
H9
• GRADIENT RATIO <= gj
flow
Figure 8 Gradient Ratio Test - Clogging
the gradient ratio test may not predict
clogging in many critical applications.
Long term flow tests use the same
apparatus as the gradient ratio test but •
measure the actual flow rate of leachate
passing through the sample over an extended
time period. Typical data from a long-term
test are shown on Figure 9. The slope of the
curve is the focus of attention. If the
slope continues to "be negative, then the
fabric will eventually clog. The combination
of an acceptable flow rate and a final slope
of zero indicates an acceptable soil/fabric
combination.
ADDITIONAL CONSIDERATIONS
Additional geosynthetic components are
used in building interior ramps, berms,
drainage standpipes, and the cap structure.
These ancillary components are not reviewed
in this paper.
The designer must also verify that the
synthetics used to build the system are
resistant to chemical attack from the
leachate. While beyond the scope of this
paper, guidelines for chemical evaluation
are presented elsewhere (11). The designer
must also be aware of the need for a better
assesment of. long-term creep performance of
the synthetics, hydraulic problems related
to biological growth within an LCR system,
and field construction quality control
problems. These unknown long-term
performance factors and the impact of a
failure in these facilities force today's '
designer to use Design Ratios significantly
higher that conventional designs require.
-57-
-------
['I
A
T <•
E
e
e 31
*.
SCO
1000 1500 2000 2500 3000
TIME (HOURS)
Figure 9 Long Tersi Flow Test - Clogging
SUMMARY
This paper elaborates on the basic
design considerations needed, to verify the
mechanical and hydraulic performance of the
FML and LCR components for double-liner
systems for both landfills and surface
impoundments. The designs are based on
estimates of actual field stress and
hydraulic conditions and on laboratory
measures of'the limit capacities of the
synthetic components. These procedures
are valid for both short and long-term
conditions if appropriate laboratory tests
are used.
ACKNOWLEDGEMENT
This work has been sponsored by the
U.S. Environmental Protection Agency, ' ;
Hazardous Waste Engineering Research
Laboratory, Cincinnati, Ohio.
REFERENCES
l.EPA, 1985, Minimum Technology Guidance on
Double Liner Systems for Landfills and
Surface Impoundments-Design, Construction,
and Operation, 2nd Draft Version, May 24.
(EPA/530-SW-85-014)
2.EPA, 1985, Minimum Technology Guidance on
Single Liner Systems for Landfills, Surface
Impoundments, and Waste Piles-Design,
Construction, and Operation, Draft, May 24.
(EPA/530-SW-85-013)
S.Koerner, R.M. and G.N. Richardson, 1987.
Design of Geosynthetic Systems for Waste
Disposal, Proc. ASCE-6T Conf. Geotechnlcal
Practice for Waste Disposal, Univ. Mich.,
Ann Arbor, Mich., June.
4.Knipshield, William K., 1985. Material
Selection and Dimensioning of Plastic Sheet
to Protect Groundwater, Waste and Refuse.
Schmidt Publisher, Vol. 22.
B.Wong, J», 1977. The Design of a System for
Collecting Leachate From a Lined Landfill
Site, Water Resources Research, Vol. B,
No.2.
6.Demetracopolous, A.C., et al, 1984.
Modeling for Design of Landfill Bottom
Liners, Journal Environmental Engineering.
ASCE, Vol. 110, No. 6, December.
7.EPA, 1983, Landfill and Surface Impound-
ment Performance Evaluation, SW-869, April.
S.koerner, R.M., 1986.Designing With
Geosynthetics. Prentice-Hall.
9.Giroud, J.P., 1982. Filter Criteria for
Geotextiles, Proc. Second Intl. Conf. on
Geotextiles. IFAI, Las Vegas, Vol.1, June.
lO.Halse, Y. ,et.al., 1987, Filtration
Properties of Geotextiles Under Long Term
Testing, Proc. ASCE/PennDot Tech. Seminar.
Hershey, Pa, April.
ll.Matrecon, Inc. , 1987, "Lining of Waste
Impoundment and Disposal Facilities,
(Draft), Third Edition, SW-870, U.S.
Environmental Protection Agency,
Cincinnati, Oh.
-58-
-------
INSPECTION PROCEDURES/CRITERIA
FOR -INSTALLATION OF FLEXIBLE MEMBRANE LINERS
Thomas D. Wright and J. Rodney Marsh
SCS Engineers
Long Beach, California 90807
William M. Held
SCS Engineers
Covington, Kentucky 41017
Louis R. Hoyater
Hovater Engineers
Laguna Hills, California 92653
. ABSTRACT
Inspection of flexible membrane liner (FML) installation is important in assuring
required hazardous waste containment. This paper outlines the procedures and criteria for
inspecting the installation of the four most commonly used FMLs for land containment of
hazardous wastes:
• Polyvinyl chloride (PVC).
• Chlorosulfonated polyethylene (CSPE).
• High-density polyethylene (HOPE).
• Chlorinated polyethylene (CPE).
The FML installation steps for which inspection procedures are provided include:
« Unloading and storage of FML. /
• Preparation and maintenance of the FML supporting surface (earth and other supporting
surfaces are discussed).
• Placement of FML on supporting surface.
• FML seaming operations.
• FML anchoring and sealing including anchoring in earth and anchoring to concrete and
other materials.
0 FML testing including seam testing and testing of the integrity of the entire FML
installation.
• FML cover operations.
The FML installer and inspector should be able to follow the guidelines developed by
this study in preparing construction quality assurance plans, and in assuring that design
plans and specifications are met. ,
-59-
-------
BACKGROUND
FMLs are thin prefabricated polymeric
layers of sheets of plastic or rubber com-
monly used for the containment of hazard-
ous and other wastes. Inspection of an
FML installation is important in assuring
containment of wastes.
This paper discusses a manual pre-
pared for the EPA which contains the pro-
cedures and criteria for inspecting the
Installation of the four most commonly
used FMLs (PVC, CSPE, HOPE, and CPE) for
land containment of wastes. The manual is
intended for use by regulatory officials,'
engineering firms, and facility owners/
operators. The manual can also be used as
a guide in preparing quality assurance
plans and as a guide for construction
quality assurance inspectors for assuring
that design plan and specification re-
quirements are met.
In order to develop a practical docu-
went questionnaires were sent to, and
interviews were conducted with, manufac-
turers (raw materials and finished mate-
rials), fabricators, lining contractors,
testing laboratories, and consulting engi-
neering firms to solicit their comments
regarding FML installation. In addition,
a thorough review of over 300 documents
regarding FML installation was conducted
as a further basis for manual preparation.
MANUAL CONTENTS
Discussions among in-house and EPA
personnel regarding the presentation for-
mat best suited for use by an inspector or
other user resulted in the manual being
divided into seven chapters that sequen-
tially cover each FML installation proce-
dure. These are:
• On-site unloading/storage of FML.
• Preparation and maintenance of the FML
supporting surface.
• Placement of FML on supporting surface.
• FML seaming operations.
• FML anchors/attachments.
• FML testing, including seam testing and
testing of the integrity of the entire
FML installation.
• FML cover operations.
This format allows the inspector the
capability of removing chapters indivi-
dually as he inspects the various instal-
lation procedures (assuming the manual is
kept in a three ring binder or each chap-
ter is bound separately). It is hoped
that the ease of carrying smaller, less
cumbersome inspection materials will en-
courage the inspector to keep the appli-
cable portion of the manual with him as
required to assure proper installation.
A brief discussion of each chapter is
given below. Copies of the completed
manual will be made available through
NTIS.
On-Site Unloading/Storage
The initial step of FML inspection is
to make sure the specified FML type and
accessories are delivered to the job site
undamaged. This is primarily the respon-
sibility of the installation supervisor.
However, the inspector should also make
spot checks.
Once the FML is accepted as undam-
aged, it is important that it be carefully
unloaded onto a surface that is relatively
level and smooth, free of rocks, holes and
debris to prevent damage to the FML.
FML materials and accessories are
best kept out of direct weather conditions
to prevent possible damage from sun, wind,
or moisture. Quite often the storage area
is in the same location as the unloading
area. If this is the case, the area must
be large enough so that unloading, stor-
age, and transport can operate smoothly.
If the storage area is in a building,
proper accessibility and maneuverability
are of primary importance. A loading dock
is also of great value. If the storage
area is open it should generally be fenced
for security.
Specific sections and subsections in
the manual discussing inspection criteria
and procedures for on-site unloading/
storage are listed below.
-60-
-------
• Material check.
- FML.
- Accessories.
0 Unloading and handling equipment.
• Unloading area:
- Accessibility.
- Type.
- Location.
- Unloading and handling.
• Storage area:
- Type.
- Location.
- Climatic conditions.
- Other factors.
Preparation and Maintenance of Supporting
Surface
For purposes of the manual, a sup-
porting surface is defined as the surface
on which the FML will be placed. For a
single FML system, the supporting surface
will probably consist of compacted earth,
concrete, asphalt, or other material. If
a second FML is specified (a double FML
system), the supporting surface for the
top FML will generally consist of a drain-
age net, a granular,soil layer, and/or a
geotextile fabric.
Whether a single or double FML sys-
tem, it is assumed that the supporting
surface has been brought to final grade,
and that the earth subgrade (if appli-
cable) has been compacted and is struc-
turally sound.
If the supporting surface is earth,
the inspector should be aware that the
soil type will effect the installation
procedure. For example, clayey soils form
surface cracks when dried, and sandy soils
form depressions under foot and vehicular
traffic. A table containing soil charac-
teristics pertinent to FML installations
is included in the manual.
An earth supporting surface should be
free of all vegetative growth prior to
final preparation. Burrowing animals
(e.g., gophers and ground squirrels)
should also be removed, and/or their tun-
nels destroyed. Final preparation con-
sists of removing all clods, pebbles,
etc., filling in voids, and rolling/com-
pacting the surface.
After the supporting surface has been
accepted for FML placement, it is impor-
tant to maintain its integrity by miti-
gating or correcting damage caused by wind
or storm water erosion, and saturation or
ponding by storm water.
Supporting surfaces other than earth
(e.g., concrete and asphalt) must also be
inspected to assure that they are smooth,
free of surface voids or depressions, and
that there are no abrupt changes in abut-
ting surface elevations.
The specific sections and subsections
as they appear in the manual are as fol- ,
lows:
« Earth supporting surfaces:
- Type of soil.
- Vegetation removal.
- Burrowing animals, ants, and other
pests.
- Preparation of finished earth sup-
porting surface.
- Maintenance of earth supporting sur-
face.
• Other supporting surfaces:
- Concrete.
- Asphaltic supporting surfaces.
- Geotextile fabric.
- Drainage layers.
Placement of FML on Supporting Surface
Placement of FML on the supporting
surface includes: transporting the FML to
the working area; removing the FML from
its packaging; and spreading the FML
sheets in their appropriate locations,
making sure each FML is not damaged (e.g.,
has no holes, etc.) and is ready for seam-
ing.
Prior to commencing placement of the
FML, the inspector should first ensure
that placement equipment is on site and in
working order; a sufficient number of
qualified placement personnel are on site;
and weather conditions are suitable.
Specific sections and subsections in
the manual discussing placement inspection
criteria and procedures are listed below.
-61-
-------
• Placement equipment.
• Personnel.
• Weather conditions.
* FML layout:
- Type of FML.
- Thickness of FML.
- Accessibility.
- Placement on slopes.
- Placement around penetrations.
Seaming Operations
FML seams can be prepared either in
the factory or in the field. Factory
seams are manufactured in a controlled
environment, and seam quality is generally
superior to field seams. Since the field
inspector is not responsible for observing
the making of factory seams, only a brief
summary of factory seaming techniques is
provided in the manual as background.
However, the overall quality assurance/
quality control should require this fac-
tory inspection.
The primary objective of the manual
is to discuss field seaming operations and
the inspection thereof. However, the dis-
cussion is necessarily general because
each FML manufacturer/fabricator has spe-
cific (and often proprietary) seaming
equipment and seaming procedures.
Prior to the start of FML seaming
operations, the manufacturer should pro-
vide the inspector with precise specifica-
tions on the equipment and procedures that
will be used to seam their material in the
factory and field. In addition, the
inspector should meet with the manufac-
turer's/fabricator's representative for an
explanation and demonstration of the seam-
ing equipment and procedures to be used on
the job.
Seaming methods described include
liquid applied solvent or adhesive
methods; thermal methods; vulcanizing
tapes and adhesives; and tape and mechani-
cal seaming methods.
Specific sections and subsections in
manual discussing seaming inspection cri-
teria and procedures are listed below.
• Seaming methods:
- Factory seams. '
- Field seams.
• Equipment.
• Personnel.
• Environmental conditions.
• Seaming.
Liner Anchors/Attachments
One of the most common sources of
failure in an FML installation is the
attachment of the FML to another surface.
In general, these attachments consist of
perimeter anchors or attachments to struc-
tures, such as pipes or columns, within
the facility.
The manual provides the inspector
with guidance on specific types of
anchors/attachments, and also delineates
several general practices and procedures
that should be followed. These include
ensuring that placement equipment is on
site, in working order, and that a suffi-
cient number of qualified placement per-
sonnel are on site.
Specific sections and subsections in
the manual discussing liner anchors/
attachments inspection criteria and proce-
dures are listed below.
• Placement equipment.
• Personnel.
• Anchors/attachments:
- Earth anchor trenches.
- Inspection procedures.
- Concrete and piping.
- Concrete.
- Battens and bolts.
- Reglets.
- Piping.
FML Testing
This chapter of the manual introduces
the inspector to methods that are used to
test factory and field seams and patches.
These test methods can be destructive
and/or nondestructive. Nondestructive
tests are performed in the field on in-
place FML. Testing is performed by the
-62-
-------
contractor and should be observed by the
inspector. This type of test retains the
integrity of the FML seam or sheet being
tested. Destructive tests are performed
in either the field or laboratory. The
intent is to determine the strength char-
acteristics of a seam sample by stressing
the sample until either the seam or the
FML sheeting fails. Only tests performed
in the field should be observed by the
inspector; however the inspector should
review laboratory test results to ensure
acceptable results were obtained.
Destructive tests can be performed on
samples taken from either in-place FML
seams, sheets, and patches (destructive
sampling), or on samples of representative
FML seams and patches fabricated by the
seaming/patching crew from the same mate-
rial, and using the same seaming methods
as those being used to make the in-place
seams and patches (nondestructive sam-
pling). The inspector should verify that
the type of samples used are those speci-
fied in the quality assurance/quality con-
trol plan.
Field test methods described in the
manual are probe, air lance, vacuum box,
ultrasonic pulse echo, ultrasonic imped-
ance plane, spark, pressurized dual seam,
electrical resistivity, hydrostatic, seam
strength peel, and,seam strength shear
tests. Laboratory test methods described,
are the bonded seam strength shear and
bonded seam strength peel tests.
Specific sections and subsections in
the manual discussing FML testing inspec-
tion criteria and procedures are listed
below.
• Field tests:
- Air lance.
- Probe.
- Vacuum box.
- Ultrasonic pulse echo.
- Ultrasonic impedance plane.
- Spark testing.
- Pressurized dual seam.
- Electrical resistivity.
- Hydrostatic.
- Seam strength peel.
- Seam strength shear.
• Laboratory tests:
- Bonded seam strength, shear.
- Bonded seam strength, peel.
FML Covers
The primary function of the inspector
is to ensure that the FML is not damaged
during cover placement operations. Covers
over an FML will generally consist of
earth (clay [natural and/or bentonite
added], sand, silt, gravel, or a combina-
tion of soils) or some other material such
as portlahd cement/gunite, geotextile or
drainage net.
The majority of the above covers are
placed over an FML as a protective layer
against mechanical, weather or other
potential damage. Drainage nets and sand
and gravel are used as a permeable layer
to convey leakage and/or leachate that may
accumulate on the upper and/or lower FML
after being put into use.
Specific sections and subsections in
the manual discussing cover type and
placement inspection criteria and proce-
dures are as follows:,
• Earth covers:
- Equipment.
- Personnel.
- Weather conditions.
- Earth cover placement.
• Portland cement concrete/gunite:
- Equipment.
- Personnel.
- Weather conditions.
• Drainage nets and geotextiles.
SUMMARY ,
An improperly installed FML is little
better than no liner at all, and may actu-
ally be worse, because it provides a false
sense of security. Consequently, the role
of the inspector in ensuring proper FML
installation can be critical. The inspec-
tor needs to be aware of all procedures
and criteria essential for proper instal-
lation, from unloading and storage at the
site through actual placement and final .
cover.
The use of the manual described herein
•should play a major role in standardizing'
inspection procedures and in ensuring .
proper installation of FMLs at hazardous
waste sites.
-63-
-------
AN ASSESSMENT OF MATERIALS THAT INTERFERE WITH
STABILIZATION/SOLIDIFICATION PROCESSES
M. John Cullinane, Jr., R. Mark Bricka, and
and Norman R. Francingues, Jr.
U. S. Army Engineer Waterways Experiment Station
Vicksburg, MS 39180-0631
ABSTRACT
Stabilization/solidification of hazardous waste involves mixing the waste with a
binder material to enhance the physical properties of the waste and to immobilize con-
taminants that may be detrimental to the environment. Many hazardous wastes contain
materials that are known to inhibit the setting and strength development properties of
commonly used stabilization/solidification binder materials. This paper describes the
initial results of an evaluation into the affect of ten interfering agents: oil, grease,
lead nitrate, copper nitrate, zinc nitrate, sodium hydroxide, sodium sulfite, phenol,
trichloroethylene, and hexachlorobenzene on the setting and strength development prop-
erties of a metal hydroxide sludge stabilized/solidified with three binder materials
(Portland cement, lime/flyash, and cement/flyash).
The setting and strength development properties of the waste-binder-interfering
agent mixture were evaluated using the unconfined compressive strength (UCS) test. UCS
data indicated that relatively low levels (less than 8 percent) of phenols, lead nitrate,
oil, and grease can result in an 80 percent reduction in the 28-day UCS developed by the
waste-binder-interfering agent mixture as compared to control specimens.
Xrichloroethylene and hexachlorobenzene where shown to have little effect on strength
development.
The results of this research confirm the need for waste-binder specific studies
prior to the selection of a chemical stabilization/solidification process for the treat-
ment of hazardous wastes.
INTRODUCTION
Background
The Environmental Protection
Agency (USEPA) is responsible for evalu-
ating the suitability of hazardous waste
and materials for land disposal. Chemical
stabilization/solidification (S/S) is one
technique that has been proposed as a
means of controlling the release of con-
taminants from landfilled wastes to sur-
face and ground waters. Indeed, S/S of
hazardous wastes is recognized in regula-
tions implementing both the Superfund
Amendments and Reauthorization Act of 1986
(SARA) and the Hazardous and Solid Waste
Act Amendments of 1984 (RCRA).
A variety of S/S technologies have
been proposed for treating hazardous
wastes. The most commonly applied tech-
nologies use cement, pozzolan, or cement-
pozzolan combinations as the primary means
of contaminant immobilization (U. S. EPA
1980, Cullinane et al. 1986). A potential
problem with using S/S technology involves
chemical interferences with the hydration
reactions typical of the cement and
pozzolan processes. Experience in the
cement and construction industry has
demonstrated that small amounts of some
chemicals can significantly affect the
setting and strength development charac-
teristics of concrete. Consequently, the
cement industry has developed fairly
stringent criteria for the quality of
-64-
-------
cement, aggregate, water, and additives
(accelerators or retarders) that are
allowed in concrete (Jones et al. 1986).
Of particular concern to S/S tech-
nology is the affect of organic compounds
on the strength and contaminant immobili-
zation characteristics of the final
product. It is well documented that small
concentrations of organic compounds (Young
1972, Young 1973), sugars (Ashworth 1965)
formaldehydes (Rosskopf 1975) and various
chemical contaminants typically found in
hazardous waste affect the setting mecha-
nisms of pozzolan cements and lime flyash
pozzolans. Roberts (1978) and Smith
(1979) reported on the affects of metha-
nol, xylene, benzene, adipic acid, and an
oil and grease mixture on the strength and
leaching characteristics of a typical
lime/flyash S/S formulation. Smith (1979)
concludes that there was a good correla-
tion between the effects of organic com-
pounds on lime/flyash pozzolanic systems
and the reported effects on the hydration
of Portland cement. More recently,
Chalasani, et al. (1986) and Walsh, et al.
(1986) , using x-ray diffraction and scan-
ning electron microscopy techniques,
reported on the affects of ethylene glycol
and p-bromophenol on the microstructure of
Portland cement hydration products. Eth—
ylene glycol was found to produce signifi-
cant changes in the microstructure up to
at least a year of curing time.
Purpose and Scope
The purpose of the research described by
this paper is to develop data on the com-
patibility of ten waste constituents with
three binding agents - Portland cement,
Portland cement/flyash, and lime/flyash
pozzolan cement. Only the results of the
unconfined compressive strength (UCS) test
are presented in this paper. The
remainder of the data will be presented in
a comprehensive report scheduled for'
publication in 1988, after completion of
the project.
MATERIALS AND METHODS
Jones et al. (1985) described the basic
research program to evaluate the effects
of interfering compounds on S/S processes
'at a previous conference. The study
reported on herein was conducted in three
phases: (1) preparation of a synthetic
wastewater and sludge; (2) addition of a
binder and interfering material to the
sludge; and (3) UCS testing of cured spec-
imens containing sludge, binder, and
interference chemicals.
Synthetic Wastewater and Sludge Production
Initial laboratory tests revealed that a
synthetic wastewater containing nitrate
salts of cadmium, chromium, nickel, and
mercury at 600 times the EPA extraction
procedure limit could be treated with
calcium hydroxide to produce a hydroxide
sludge with typical metal concentrations
of 86.2, 84.1, 18.8, and 0.137 mg/g (dry
weight basis) of nickel, chromium,
cadmium, and mercury respectively.
Typically, the raw sludge contained
8 percent solids and was very fluid. The
sludge was dewatered to approximately
30 percent solids using a rotary drum
vacuum filter. A constant moisture con-
tent between sludge batches was maintained
by adjusting the solids content of the
dewatered sludge to 25 percent using the
supernatant liquid from the sludge produc-
tion process as a dilution liquid.
Specimen Preparation
The 25 percent solids content sludge
was divided into three 150 gallon samples
and binder material was added to each at
the following ratios.
Binder
Portland Cement
(Type I)
Portland Cement
(Type I)/Flyash
(Type F)
Lime/Flyash
(Type C)
Binder/Sludge
Ratio
0.3:1 Cement:Sludge
0.2:1 Cement:Sludge
0.5:1 Flyash:Sludge
0.3:1 Lime:Sludge
0.5:1 Flyash:Sludge
After mixing the sludge with the binder,
each binder/sludge sample was subdivided
into four equal parts. One of the ten
interfering chemicals was added to each of
the subsamples at ratios of 0, 0.02, 0.05,
and 0.08 (by weight) interference chemical
to binder/sludge material. The subsample
to which no interference chemical was
added was used as a control specimen. A
control specimen Was prepared each time an
interference chemical was processed, thus
-65-
-------
r
accounting for variability between
batches, so that the UCS results between
batches could be compared.
The interference/binder/sludge mix-
ture (I/B/S) was then molded into two inch
cubes in accordance with ASTM Method
C-109-77/86 (ASTM 1986). Because the
I/B/S mixture was usually viscous and
could not be tamped into the molds, the
ASTM method was modified to include vibra-
tion of the I/B/S mixture to remove any
air pockets that developed during the
molding process.
The specimens were cured in the molds
at 23° C and 98 percent relative humidity
for a minimum of 24 hours and removed from
the molds whenever they developed suf-
ficient strength to be free standing.
After removal from the molds, the speci-
mens were cured under the same conditions
for periods of 4, 11, and 28 days. At the
end of each curing period, the UCS of the
specimens was determined in accordance
with ASTM C 109-77/86.
DISCUSSION OF RESULTS
Space limitations do not allow presenta-
tion of all the study results. Typical
results for selected interferences are
discussed below.
Table 1 presents the results,
reported as the percent increase or
decrease in 28-day UCS from the control
specimen. Figures 1 through 5 present a
graphical representation of the results of
the study. Figure 1 presents the UCS
versus curing time for the cement binder-
sodium hydroxide interference. Each curve
represents the strength development curve
for one interference concentration. Fig-
ure 2 presents the 28-day UCS versus
interference concentration for each of the
three binders for the sodium hydroxide.
Figures 3, 4, and 5 present a graphical
representation of the relative effects of
interference concentration on the 28-day
UCS of the three binders.
The data for the 28 day UCS are an indica-
tor of the UCS trends observed at the
earlier curing periods. This is clearly
illustrated in Figure 1 which is a plot of
cure time versus UCS for one interference
material. Although the slope of the
curves varies between binder and inter-
ference treatments, in most cases, the
lines of constant interference concentra-
tion do not cross.
Portland Cement Binder
The data presented in Table 1 clearly
indicate that the interference effects may
be positive or negative, depending on the
concentration of the interfering material.
The addition of oil or grease at a
0.08 ratio resulted in a 44 percent
decrease in the 28-day UCS. The addition
of copper nitrate, resulted in an increase
in UCS for all concentrations. The addi-
tion of copper nitrate at a 0.05 ratio
resulted in a 181 percent increase in the
28-day UCS. The addition of zinc nitrate
resulted in the an increase in UCS for the
0.02 ratio; however, when added at a
0.08 ratio, it resulted in the largest UCS
decrease, 85.5 percent, observed for any
of the interference compounds. For
cement, most specimens exhibited a
decrease in UCS with increasing inter-
ference concentration.
By comparing the UCS results for the
control specimens for different binders,
it is evident that the cement specimens
developed less strength than the lime/
flyash or the cement/flyash specimens.
Figure 2 however illustrates that
increasing the concentrations of the
interference chemical does not necessarily
affect the different binders to the same
degree.
Portland Cement/Flyash
The addition of a 0.08 ratio of lead
nitrate, copper nitrate, zinc nitrate, and
phenol resulted over a 90 percent decrease
in 28-day UCS. The effect of sodium
hydroxide, sodium sulfate, and copper
nitrate addition was interesting because
of the increase in UCS at the lower con-
centrations and decreases in UCS at the
higher concentrations. The addition of a
0.02 ratio of sodium hydroxide resulted in
a 22.8 percent increase in 28-day UCS
while the addition of a 0.08 ratio caused
a 19.5 percent decrease in 28-day UCS.
The effect of hexachlorobenzene was not
concentration dependent and resulted in a
consistent 9 percent decrease in 28-day
UCS, regardless of concentration.
-66-
-------
Lime/Flyash
The addition of a 0.08 ratio of lead
nitrate, copper nitrate, zinc nitrate,
sodium sulfate, and phenol resulted in an
80 percent decrease in 28-day UCS. The
addition of trichloroethlene, hexachloro-
benzene, sodium hydroxide, and sodium
sulfate resulted in a gain in 28-day UCS
(51.0, 6.2, 71.1, 65.2 percent respec-
tively) at the 0.02 ratio; however, these
same compounds resulted in a decrease in
strength (34.4, 1.2, 3.6, and 81.8 percent
respectively) at a 0.08 ratio. The addi-
tion of oil or grease at a 0.08 ratio
resulted in a 32 and 54 percent decrease
respectively, in 28- day UCS.
CONCLUSIONS
Several conclusions can be drawn that
characterize the effects of the inter-
ference materials investigated in this
project on the UCS of stabilized/ '
solidified waste materials.
1. The interference chemicals tested
had a measurable effect on the setting and
strength development properties of the
stabilized/solidified waste. The magni-
tude of the effect depended on the type ,of
binder, the curing time, 'and the type and
concentration of the interfering compound.
2. Stabilized/solidified waste will
show decreases in UCS development with
increasing oil or grease concentrations.
3. Although the waste stabilized
with Portland cement resulted in lower
28-day strength development, it appears
that the concentration of interference
material had less of an effect on the UCS
development 'properties for the Portland
cement binder than for the Portland
cement/flyash or lime/flyash binders.
4. Zinc and phenol concentrations
above 5 percent resulted in marked
decreases in 28-day UCS development for
all the binders tested.
5. Sodium hydroxide tends to
increase the UCS at lower concentrations
and decreases the UCS at higher concentra-
tions for all binders.
6. The chlorinated hydrocarbons
evaluated in this study had little effect
on the UCS development properties.
ACKNOWLEDGEMENTS
The tests described and the resulting
data presented herein, unless otherwise
noted, were obtained from research con-
ducted by the U. S. Army Engineer Water-
ways Experiment Station and sponsored by
the U. S. Environmental Protection Agency,
Hazardous Waste Engineering Research Lab-
oratory, Cincinnati, Ohio, under Inter-
agency Agreement DW96930146-01.
Mr. Carlton Wiles, Hazardous Waste Engi-
neering Research Laboratory was the EPA
project officer. Permission to publish
this information was granted by the Chief
of Engineers and the U. S. Environmental
Protection Agency.
REFERENCES
1. Ashworth R. 1965. "Some Investiga-
tions Into the Use of Sugar1 as an Admix-
ture to Concrete", Proceedings of the
Institute of Civil Engineering, London,
England.
2. ASTM 1986. Annual Book of ASTM Stan-
dards; Construction, Volume 04.01,
Cement; Lime; Gypsum, American Society
for Testing Materials, Philadelphia, PA.
3. Chalasani, D., Cartledge, F. K.,
Eaton, H. C., Tittlebaum, M. E. and
Walsh, M. B. 1986. "the Effects of Eth-
ylene Glycol on a Cement-Based Solidifica-
tion Process," Hazardous Wastes and
Hazardous Materials, vol 3, no. 2,
New York, NY.
4. Cullinane M. J., Jones L. W., and
Malone P. G. 1986. "Handbook for
Stabilization/Solidification of Hazardous
Waste", EPA/540/2-86/001, Hazardous Waste
Engineering Research Laboratory,
U. S. Environmental Protection Agency,
Cincinnati, OH.
-67-
-------
5. Jones, J. N., Bricka, M. R., Myers,
T. E., and Thompson, D. W. 1985.
"Factors Affecting Stabilization/
Solidification of Hazardous Wastes," Pro-
ceedings: International Conference on New
Frontiers for Hazardous Waste Management,
EPA/600/ 9-85-025, Hazardous Waste Engi-
neering Research Laboratory, U. S. Envi-
ronmental Protection Agency, Cincinnati,
OH.
6. Roberts, B. K. 1978. "The Effect of
Volatile Organics on Strength Development
in Lime Stabilized Fly Ash Compositions,"
M. S. Thesis, University of Pennsylvania,
Philadelphia, PA.
7. Rosskopg P. A., Linton F. J., and
Peppier R. B. 1975. "Effect of Various
Accelerating Chemical Admixtures on
Setting and Strength Development of Con-
crete", Journal of Testing and Evaluation,
Vol 3, No. 4.
8. Smith, R. L. 1979. "The Effect of
Organic Compounds on Pozzolanic
Reactions," I. U. Conversion Systems,
Report No. 57, Project No. 0145.
9. U. S. Environmental Protection Agency
1980. Guide to the Disposal of Chemically
Stabilized and Solidified Waste, SW-872,
Office of Research and Development,
Municipal Environmental Research Labora-
tory, Cincinnati, OH.
10. Walsh, M. B., Eaton, H. C.,
Tittlebaum, M. E., Cartledge, F. K., and
Chalasani, D. 1986. "The Effect of Two
Organic Compounds on a Portland Cement-
Based Stabilization Matrix," Hazardous
Waste & Hazardous Materials, Vol 3, No. 1,
New York, NY.
11. Young J. F. 1972. "A Review of the
Mechanisms of Set-Retardation of Cement
Pastes Containing Organic Admixtures",
Cement and Concrete Research, 2, No. 4
(July 1972).
12. Young J. F., Berger R. L., and
Lawrence F. V. 1973. "Studies on the
Hydration of Tricalcium Silicate Pastes.
Ill Influences of Admixtures on Hydration
and Strength Development", Cement and Con-
crete Research, 3, No. 6.
TABLE 1. 28-DAY UNCONFINED COMPRESSIVE STRENGTH AS A PERCENT OF CONTROL SPECIMEN
Interference
Chemical
Oil
Grease
Lead Nitrate
Copper Nitrate
Zinc Nitrate
Xrichloroethylene
Hexachlorobenzene
Sodium Hydroxide
Sodium Sulfate
Phenol
Portland Cement Binder
I/B/S Ratio
0.02 0.05 0.08
-20
-12
- 2
+91
+3
-28
-5
+14
-7
-22
-38
-25
+18
+181
-73
-36
-6
-33
-13
-26
-44
-45
+10
+4
-86
-27
+15
-52
-53
-54
Cement /Flyash
I/B/S Ratio
0.02 0.05 0.08
-8
-48
-51
+17
-47
-7
-10
+23
+23
-49
-28
-40
-75
-84
-93
-33
-9
+5
+16
-82
-42
-20
-97
-98
-95
-29
-10
-20
-64
-92
Lime/Flyash
I/B/S Ratio
0.02 0.05 0.08
—7
-7
-40
-48
-53
+51
+6
+1
+65
-65
-27
-27
-77
-48
-77
-20
+9
-7
+38
-88
-32
-54
-90
-98
-88
-34
-1
-4
-82
-96
1 All results reported as percent increase (+) or decrease from the control specimen
rounded to nearest whole percent.
2 Interference to binder/sludge ratio.
-68-
-------
BINDER: CEMENT
INTERFERENCE: NaOH
' 1 —
12 16 20
CURVE TIME, DAYS
24
28
Figure 1. Unconfined Compressive Strength as a Function of Curing Time
and Interference Concentration
800
0.02 0.05
PHENOL RATIO
0.08
Figure 2. Unconfined Compressive Strength for Three Binders as a
Function of Interference Concentration
-69-
-------
320
55 300
£280
£ 260
§240
gJ220
oi 200
w 180
tu 160
| 140
§120
0100
2 80
H 60
8 40
=> 20
BINDER: CEMENT
CURE TIME: 28 DAYS
0
OIL
CU CONTROL
E3 0.02
EZ3 0.05
• 0.08
I
GREASE Pb Cu Zn TCE HCB NaOH Na2SO4 PHENOL
INTERFERENCE CHEMICAL RATIO
Figure 3. Effect of Interference Concentration on 28-day UCS for
Type I Portland Cement Binder
I
fe
tu
900
800
700
500
I
400
300
200
100
0 CONTROL cSlMEf^DAYS^8"
1
•;
:•
;;
•;
';
,'f
'/
>',
,'
r/t
rt<
V
V
1
n
n
! y
: /
. /
; /
; /
' '
\ /
', t
'. j
7
/
E3 0.02 ic
1771 0.05 1C
H 0-08 1C
__
•
;*
;;
^;
^;
'•
'!
^;
;i
^
™
^
^
7m
7f
/
/
/
/
/
/
/
^ _ .
^/
x/
/
' /
/
I
1
t
1
1
t
1
/
T
/
^
/
',
/
j
.
f.
/
'
'
'
y
'•
J;
7]
/
/
//
/
/
/
i
A
A
/•
|
/
x
^
/
V
'y
)'
^
1
OIL GREASE Pb Cu Zn TCE HCB NaOH Na2SO4 PHENOL
INTERFERENCE CHEMICAL RATIO
Figure 4. Effect of Interference Concentration on 28-day UCS for
Cement/Flyash Binder.
-70-
-------
1200
S 1100
•if
{= 1000
o
i 900
EC
M 800
UJ
> 700
2 600
cc
a.
2 500
Q 400
2 300
u.
| 200
8
§ 100
BINDER: LIME FLY ASH
- CURE TIME: 28
•v
_
^
!fti
/I
/I
~
>
^
,>
^
.;
;'
TV
1
i
1
1
1
^
.>
;^
I;
DAYS
,
>
^
:J
:•
7
^
/
OIL GREASE Pb Cu
CU CONTROL
£3 0-02 1C
El 0.05 1C
H 0.08 1C
-
H
.^ .
S
i
rq
; p
;
' —
> /
;/
:/
;/
; /
\/
I',/
^71
1
1
^|
I
I
•*
• r
; ^
< ^
; •
J
7
/
X
/
/
' /
(ff
~v/
* /
$/
1
p^
J
1.
<1
'l
1
/
/
t
I
/
X
'/
/
/
^
:>
[;
1
Zn TCE HCB NaOH Na2SO4 PHENOL
INTERFERENCE CHEMICAL RATIO
Figure 5. Effect
of Interference Concentration on 28-day UCS for
Lime/Flyash Binder.
-71-
-------
MINE WASTE/OVERBURDEN ANALYTICAL TECHNIQUES -
CHARACTERIZATION AND SIMULATION OF MINE TAILINGS WEATHERING ENVIRONMENTS
Frank T. Caruccio, Gwendelyn Geidel
Department of Geology
University of South Carolina
Columbia, S.C. 29208
ABSTRACT
The prediction of mine drainage quality is based on the chemical weathering
attributes of the mine waste/overburden that is produced by the mining operation. These
are generally assessed through overburden analyses, which fall into two broad categories,
static or dynamic techniques. In the static tests, drainage quality projections are
based on whole rock analyses. Alternatively, dynamic tests subject the samples to
simulated weathering tests and monitor the quality of the effluent produced. Within
these two categories several mine waste/overburden analytical techniques are available,
each having advantages and disadvantages relative to the other. An extensive computer
literature search identified the most popular analytical techniques and includedcolumn
leaching tests, humidified cells, soxhlet reactors, BCR initial and confirmed (bacteria)
tests, and acid/base accounting. These analytical procedures were further structured to
evaluate the effect of alkaline pore water, air lock, and interstitial sulfide
reactivity.
In a preliminary study, several mine waste/overburden analytical techniques were
evaluated to determine which one most closely approximates observed field conditions.
Fractions of an acid producing pyrite rich ore found at a gold mine in South Carolina
were contained in plastic tubs and exposed to the atmosphere. The volume and quality of
the leachate produced after each rain event were related to the weight of the sample.
These field derived data provide the background against which the laboratory analyses of
splits of the samples were compared and evaluated.
In the preliminary test, we found the finer the particle size, as prescribed by the
particular test, the greater the amount of acidity produced. In essence, the acid
production potential is an artifact of the particular test used and, to a lesser degree,
the chemistry of the sample.
INTRODUCTION
Acid drainage is a problem common to
operations that expose metallic disulfides
(most commonly pyrite and marcasite) to
atmospheric conditions and oxidizing
environments. In the process a highly
acidic, sulfate-iron enriched solution is
generated that may severely impact the
environment.
The evaluation of the acid production
potential of a sample is based on the
accurate assessment of the sample's chemical
weathering attributes. These assessments
are generally performed through mine
waste/overburden analyses, which can be
grouped into two broad categories, static or
dynamic techniques. In the Static tests,
whole rock analyses are used to predict mine
drainage quality; the assumption being the
specific minerals comprising the mine
waste/overburden will react with water to
produce varying degrees of alkalinity and
'acidity (the projected concentrations of
which are balanced to determine the
character of the drainage). Alternatively,
dynamic tests empirically determine leachate
quality by subjecting the samples to
simulated weathering conditions while
monitoring the quality of the effluent
produced. Within these two categories a
variety of mine waste/overburden analytical
techniques is available, each having
advantages and disadvantages relative to the
other.
-72-
-------
Through a computer literature search,
examining 679 titles and 36 abstracts
dealing with various aspects of mine
drainage quality and predictive methods, we
identified several mine waste/overburden
analytical techniques widely used in the
United States and Canada, both in the coal
fields and sulfide mines. These methods
include simulated weathering chamber tests
(column and humidity cells), soxhlet
reactors (infrequently used), biological
confirmation tests (commonly referred to as
B.C. Bacterial test), and whole rock
analyses (acid/base accounting, extensively
used in coal fields), and B.C. Research
Initial test (extensively used for base
metal and gold mines in Canada and western
U.S.). The test involving the peroxide
oxidation of the sample has been recently
published and lacks widespread usage. The
various tests are summarized and described
in Table 1. .
Many questions exist as to the accuracy
of the testing procedures and the ability of
the tests to adequately project the long-
term (decades)'chemical weathering
attributes of mine waste/overburden material
based on laboratory results obtained in
relatively short time periods (days to
weeks). In addition, the variety ,of
environmental conditions relative to
climate, mining operation, enrichment or
recovering processing methods and manner of
disposal, have a. mitigating effect on the
oxidation of the sulfide and manner by which
the weathering products are mobilized from
the site. For example, the acid potential
of some tailings will be greatly affected by
the manner of disposal. Consider the fine
grained particle size of the tailings
material which will afford the sample with a
large water holding capacity (specific
retention) and create an "air lock" that
will effectively inhibit the transfer of
oxygen required for pyrite oxidation and
restrict the oxidation zone to the near
surface meter of the disposal site.
The fact that many of the tailings piles
are located in areas where the
evapotranspiration exceeds precipitation,
requires that the analytical results
obtained from dynamic tests be reexamined in
the context of non-leaching environments.
Assuming oxidation of the sulfide mineral to
take place, the dissolution and mobility of
the acid producing weathering products is
precluded by the lack of internal drainage.
Further, some of the processing plants
utilize slaked lime to enhance recovery of
the ore and in the procedure render the
discharge water/tailings mixture alkaline.
Other than the inhibitory effect that the
alkalinity has on pyrite oxidation, the
alkalinity of the pore water may have a
substantial neutralization potential and
effectively neutralize any acidity produced.
Thus, other than the intrinsic chemical
nature of the tailings material, the
environmental factors that may play a role
in the overall weathering process and affect
the sample's acid production potential, must
be included as part of the assessment.
The objectives of this study are 1) to
determine which of the variety of mine
waste/overburden analytical techniques
methods most closely approximate field
conditions and 2) to evaluate the effects of
porewater chemistry, "air lock" potential
and leaching interval (frequency) on acid
production potential.
METHODS
The study was divided into two parts,
one dealing with the intrinsic nature of the
analytical procedures, normalizing the
environmental effects and utilizing a pyrite
rich ore from a gold mine, in South Carolina.
The second part deals with an e'valuatioft of
the environmental effects using selected
analytical procedures and mine tailings from
several copper mines in the West.
We report on the results of the first
part of this study, evaluating the
analytical procedures, and describe the
strategy of the second part of the study
that wag initiated during the Spring of
1987, and for which preliminary results will
be presented during the Symposium.
First Part of the Study
Samples for this study were collected
from a gold mine in South Carolina. Idle
for about 40 years, this site was mined,
initially for gold and subsequently for
pyrite. Acidic waters drain from a large
acid lake and small, highly acidic seeps
emanate from the talus and gangue heaps.
For the first part of this study,
various size fractions of an acid producing
pyrite rich ore were contained in plastic
tubs (approximately 0.5 m x 0.5 m x 0.3 m
deep) and exposed to rainfall.
Representative splits from each were
collected for simulated weathering tests and
whole rock analysis. The leachate generated
-73-
-------
TABLE 1. SUMMARY OF MINE WASTE/OVERBURDEN ANALYTICAL TECHNIQUES
COMMONLY USED IN THE UNITED STATES AND CANADA
Static Tests
Method
Advantages
Disadvantages
Whole Rock Analyses
Acid/Base
Accounting
CD
British
Columbia
Research
(BCR) test
(2)
Dynamic Tests
• Soxhlet
Reactor (3)
Whole rock analyses Easy to perform,
completed on a quick turn-around
pulverized sample. time, useful in
Acid potential areas that are
related to sulfur acid or alkaline
content, neutral- prone.
ization potential
determined by hot
acid digestion with
HC1.
Whole rock analysis Same as above
completed on
pulverized sample.
Acid potential related
to sulfur content,
neutralization capacity
obtained by titration
with
Does not relate to
kinetic data.
Assumes parallel
release of acidity
and alkalinity which
provides inaccurate
results.
Same as above
Leachate generated Easy to perform
on pulverized sample quick turn-around
which is cycled in time, purported
a soxhlet reactor. kinetic data.
During interim,
sample is allowed
to dry at 105° C.
Expensive apparatus,
extremely aggressive
oxidation of sample
which is not related
to any natural
process.
Simulated
Weathering Tests
• Humidity
cells (4)
Crushed rock is
placed in humidified
atmosphere and
leached periodically.
Volume and character
of leachate related
to rock weight to
produce alkaline/
acid production
potential.
Produces kinetic
data, rates of
acidity per unit
weight of sample
obtained,
approximates field
conditions.
Long time required,
large data base
generated.
-74-
-------
• Column
Weathering
test (5)
Beaker
Leachate
test (1)
BCR
Bacteria
test (6)
Peroxide
test (7)
Field sample placed
in large columns and
and leached period-
ically. Leachate is
analyzed and related
to rock weight.
Pulverized sample is
placed in water and
chemistry monitored
through time.
Pulverized sample is
oxidized in presence
of bacteria. pH
monitored through .
time, sample
continuously shaken.
Pulverized sample
is oxidized in the
presence of hydrogen
peroxide and rate
of pH change
monitored through
time.
Best approximator
of field
conditions.
Simulates sub-
merged conditions
of mine waste,
some kinetic data.
Easy to use,
incorporates
bacteria in
reactions.
May be completed
in field, very
rapid results.
Same as above, in
addition large volume
of samples required,
channelization
problem encountered.
Does not allow for
easy transfer of
oxygen which may be
rate limiting.
Data identify sample
as acid producing;
results cannot be
correlated or related
to kinetic test data.
Sample pH response is
to be calibrated against
larger sample to translate
data to field conditions;
indirect testing of
alkaline samples.
is collected and analyzed for volume, pH,
specific conductance, acidity, and sulfate.
In turn, these data are converted to
milligrams of acid produced per kilogram of
sample and provides the basis against which
the results of the laboratory overburden
analyses are compared and evaluated.
Field Tub-Control Samples
The tests performed to date utilized
three different sized populations (2.5 cm or
small-S, 2.5-5 cm or medium-M, and 7-10 cm
or large-L) selected from the pyrite rich
ore found at the site.
The plastic tubs, containing known
sample weights, were placed in an open
field, one meter above the ground.
Following a rain event, the amount of
rainfall (cm) and the volume of effluent
(liters) produced by the tubs (which drain
into sealed plastic barrels placed beneath
the tubs) were recorded. An aliquot of
effluent was collected and analyzed within
24 hours for specific conductance, pH (hot
and cold), acidity, and sulfate.
The analyses, expressed as
concentrations 1 (mg/1), are converted to
loads (mg) of component produced by
multiplying the concentrations by volume of
the effluent. In turn, these loads may be
normalized for particular sample weights to
a kg base and percentage sulfur. Plotted
versus time, cumulative acid loads produced
per kg of sample, yield rates of acid
produced against which the laboratory
analyses are being compared and evaluated.
Column Leaching
Samples of known weight and size were
placed in plexiglass columns of 9 cm
diameter and 30 cm length. Column S
contained 3090 g of sample S and filled the
column to approximately 4/5 its length (23
cm). Column M contained 2181 g of sample M
and filled the column to a height of 18 cm.
Column L contained 2089 g of sample L,
filled to a height of 17 cm. Each column
drained into a 1000 ml sample bottle. The
top of each column was sealed with plastic-
wrap to prevent contamination and to
maintain humid conditions between leachings.
The columns were leached every seven
days by spraying the top of the column with
500 ml of deionized water and allowed to
drain for 12 hours. The leachate collected
was analyzed for specific conductance, pH
(hot and cold), acidity, and sulfate
concentrations.
Humidity Cell Weathering Tests
Samples of known weight (approximately
200 grams (g)) and crushed to pass 4
-75-
-------
millimeters (mm) were placed onto a filter
paper in separate chambers. Duplicate
humidity cells chambers were constructed for
each sample size (S, M, L), for a total of
six chambers. The chamber was a plastic
cylindrical pan, 7 cm deep and 15.5 cm in
diameter. A plastic lid prevented
contamination and evaporation. Humidified
air, generated by bubbling air through water
in a closed aquarium, is circulated into the
chamber to maintain saturated conditions.
The chamber is opened only to add water
during the leaching cycle.
Samples were leached every seven days by
slowly pouring 100 ml of deionized water
onto the sample and stirring for one minute.
Following, the lid was replaced and the
water - rock mixture was drained through a
basal outlet into a graduated cylinder for
12 hours. The leachate collected was
analyzed for specific conductance, pH (hot
and cold), acidity, and sulfate.
Soxhlet Extractors
Pulverized (to pass 125U) samples of
known weight were placed in separate
cellulose extraction thimbles and stored in
a drying oven at 100° C. Every seven days
the samples (in thimble) were placed in
separate soxhlet reactors and 250 ml of
deionized water was circulated through the
sample for 4 hours. The heat source was a
GLAS-COL heating mantle (run at 120 volts).
After 4 hours, the heating mantle was turned
off and the soxhlet reactor allowed to cool
to room temperature. The leached sample was
returned to the drying oven and stored at
100° C in preparation for the next leaching
cycle. The effluent collected in the
soxhlet was analyzed for specific
conductance, pH (hot and cold), acidity, and
sulfate.
B.C.R.
This test was run in triplicate, three
equal weight pulverized (to pass 125y)
portions of each sample (S, M, L) were
weighed out and each placed in separate 300
ml Erlenmeyer flasks to which were added 70
ml of 9K medium (8). Sufficient sulfuric
acid was added, if necessary, to bring the
initial pH to 2.5. The flasks were
continuously shaken for 4 hours to
homogenize the sample and medium, and
innoculated with 5 ml of viable Thiobacillus
ferrooxidans culture. The pH was monitored
every day for the first three days and every
other day thereafter. This was continued
until the pH readings maintained a constant
value (indicating cessation of
microbiological activity) or until the pH
dropped to 1.8. At this point, half of the
weight of the original sample was added and
the sample agitated for 24 hours. If the pH
was then greater than 3.5 the test indicates
the samples to be non-acid. Otherwise, an
additional sample portion (again half the
weight of the sample) is added and the
sample agitated for 24 hours. At this point
the pH was either greater than 4.0 or less
than 3.5 (the sample is an acid producer),
the test was to be terminated. Samples with
pH falling between 3.5 and 4;0 were to be
shaken for 48 hours and a final pH
recorded (9).
The shaker table and sample flasks were
housed within a plastic chamber that trapped
the heat generated by the operating shaker
table and maintained conducive to bacterial
activity (30-35° C).
RESULTS
The results of the tests are presented
for the first 70-80 days of the study. With
'the exception of the acid/base accounting
data and the B.C.R. Bacteria test results,
all leachate data have been adjusted and
normalized to cumulative mg of acid produced
per 1 kg of sample. These acid production
trends are shown in Figures 1-7 and
summarized in Table 2. The laboratory
analyses of the samples spanned an 82 day
time period; the data presented in Table 2
are for an 80 day interval.
For control tubs, humidified cells, and
column tests the time scale (x-axis) is in
days. The soxhlet samples were leached
weekly for four hours at a time, and
accordingly, the time scale for the soxhlet
experiment is in hours.
The tub-field data (Fig. 1) show a
nearly identical behavior between samples L
and S. Sample M exhibited more than double
the cumulative acidity of S and L throughout
' much of the 80 day period. All three curves
show a pronounced decline in acid production
for the last two weeks of this period,
The results of the column tests most
closely parallel those in the field, in
terms of both cumulative acidities and
intersample variation (Fig. 2). The
increases in cumulative acidity were more
-76-
_
-------
TABLE 2. NORMALIZED CUMULATIVE ACIDITIES
FOR VARIOUS ANALYTICAL TECHNIQUES
Sample Analyses
Cumulative Acidity
(mg/kg/80 days)
Control (Tub)
Column Test
Humidity Cell
Soxhlet
250 - 750
300 - 695
1,600 - 2,100
11,000 - 15,500 mg/kg
PER 50 HOURS
linear for the column tests than for those
in the field. This is an artifact of the
constant periodicity and leaching volumes
which were controlled in the laboratory
tests.
The simulated weathering chamber tests
demonstrated much less inter—sample
variation, though the two M samples had
higher cumulative acidities towards the end
of the 80 day leaching period (Fig. 3). The
humidity cell tests clearly over-predicted
the cumulative acidity found in the field.
The results of the soxhlet tests exhibit
a different inter-sample variation than
those of the other lab and field results.
While sample M greatly outproduces sample S,
both duplicate L samples attain greater
cumulative acidities than M towards the end
of the peripd. Cumulative acidities for the
soxhlet extractors are an order of magnitude
greater than those found in the field (Fig.
4)..
DISCUSSION
Of the simulated weathering tests it
appears that the column tests more closely
approximate the quality of the leachate
derived under natural conditions. The
humidity cell overestimates the amount of
acidity, while the soxhlet greatly
overestimates the acidity. The observed
variations in acid production occurring
between the control and humidity cell are
explained by differences in grain size (the
humidity cell sample utilizes sample sizes
of 2-4 mm). The soxhlet utilizes pulverized
samples (125y) and, in addition, subjects
the samples to 105° C drying temperatures at
weekly intervals, generating skewed and
exaggerated results.
The whole rock analyses tests (B.C.R.
initial and Acid/Base Accounting) indicated
potentially acidic samples. Because the
samples used in this study are pyrite-
enriched - calcareous material deficient, we
should not expect otherwise.
The B.C.R. bacteria tests, in essence,
also confirmed the samples to be acid. In
all sets of tests, the pH was reduced from
.2.6 to 1.7 in 4 to 17 days (Figures 5-7).
The biological confirmation test results are
difficult to interpret in either a
quantitative or comparative manner. The
data generated are pH (the test parameter)
versus time (days) and the results indicate
the variability between triplicates of a
particular sample to be as great as that
found between samples.
For the B.C.R. tests the parameter
monitored is pH and at.the levels measured
for these tests the system is buffered.
i Accordingly, there is no indication of the
amount of acid produced, and the B.C.R. test
data (which monitors pH) cannot be compared
to acid production rates (measured by
titration with a standard base). The test,
in its present form, is primarily
qualitative in nature.
In summary, for the tub, column, and
humidified cell tests, it appears that the
acid production potential derived from a
particular overburden analytical technique
is more a function of the sample size
required to perform the particular test than
the chemistry of the sample. In the soxhlet
tests further comminution of the sample,
coupled with an unnatural aggressive
oxidizing environment, produces orders of
magnitude more acid than the other leaching
•tests. All things considered, the column
tests more closely approximate the observed
field results than the other analytical
tests.
SECOND PHASE OF THE STUDY
The analytical techniques used in this
study can also be used to evaluate the
various environmental factors affecting
leachate quality. Using copper mine
tailings as the test sample, heavy liquid
separation methods are used to concentrate
the heavy mineral fraction (which includes
the sulfides and is the acid producing
component). In turn, the heavy fraction is
rinsed with acetone and deionized water and
analyzed with the soxhlet reactors.
-77-
-------
Inasmuch as the tailings will not be crushed
further, the variability in acid production
by sample size is eliminated. Thus, the
acid producing component of the sample can
be evaluated.
Using the weathering chambers, tailings
samples as received and rinsed with
deionized water, can be tested to evaluate
the effect that pore water chemistry and
matrix mineralogy have on the tailings'
potential to produce acid. This will
simulate the conditions expected to occur at
the near surface environment of tailings
ponds. In addition to a humidified air
environment, the test may also be structured
to simulate oxidation processes occurring
under semi-arid conditions and in the
presence of ultra-violet light.
To test the effect of "air lock" as
might occur at depth within the tailings
ponds, the column tests can be used with
rinsed and un-rinsed samples. The leachate
qualities derived from the columns and
weathering chambers can be compared to
evaluate the magnitude of the environmental
factors as thought to occur near surface and
at depth within tailings ponds.
Finally, the leaching interval is
controlled and the mass tranfer of water
documented (i.e. volume of water added and
leachate volume are recorded), which, in
turn, can be related to the mass transfer of
water through a tailings pond in response to
various climatic conditions. Given an
evaporation loss, precipitation rate and
effective porosity, pore volume
displacements can be calculated and related
to geochemical mobilities within the pond
body.
This phase of the study is currently
underway and preliminary results should be
available for presentation at the Symposium.
REFERENCES
1. Sobek, A. A., W. A. Schuller, J. R.
Freeman, and R. M. Smith, 1978. Field
and laboratory methods applicable to
overburdens and minesoils. EPA-600/2-78-
54, National Technical Information
Service, Springfield, Virginia.
2. Bruynestein, A. and D. Duncan, 1979.
Determination of acid production
potential of waste materials. AIME, A-
79-29.
3. Renton, J., R. Hildalgo, and D. Streib,
1973. Relative acid-production
potential of coal. W.V. Geological and
Ecological Survey, Bulletin No. 11.
4. Caruccio, Frank T., 1968. An evaluation
of factors affecting acid mine drainage
production and the ground water
interactions in selected areas of
western Pennsylvania. Proceedings of
the Second Symposium on Coal Mine
Drainage Research, Monroeville,
Pennsylvania, pp!07-151.
5. Hood, W. and 0. Oerter, 1984. A
leaching column method for predicting
effluent quality from surface mines.
Proceedings of the Symposium on Surface
Mining Hydrology, Sedimentology, and
Reclamation. pp271-277.
6. Bruneysteyn, A. and R. Hackle, 1984.
Evaluation of acid production potential
of mining waste material. Mining and
the Environment, Vol. 4.
7. Finkleman, R. and D. Giffin, 1986.
Hydrogen peroxide oxidation: an
improved method for rapidly assessing
acid-generating potential of sediments
and sedimentary rocks. Reclamation and
Revegetation Research, Vol. 5, pp521-
534.
8. Kleinmann, R. L. and P. Erickson, 1983.
Control of acid drainage from coal
refuse using anionic surfactants.
Bureau of Mines, Rep. Inv. 8847,
Department of the Interior.
9. Bruynesteyn, A. and R. Hackle, 1984.
Evaluation of acid production potential
of mining waste materials. Minerals and
the Environment, Vol. 4, pp5-8.
-78-
-------
HAILE MINE STUDY: TUB EXPERIMENT
CUMULATIVE ACJOflY vs. TIME
HAILE MINE STUDY: COLUMN EXPERIMENT
°
HUE Hoy.)
* MEDIUM
160
X LARGE
Figure 1
fc
CUMULATIVE AOOOY v». °T1UE
40
a SMALL
Figure 2
HAILE MINE STUDY: LEACHING EXPERIMENT
Figure 3
BUCTEWLCONTOLEIPBiiUpir
2.4
2.3
2i
il-
10
15-
1.7
•I 6 S 10 12 14 16 18 20
c S, ,T1«. S. . .
HAILE MINE STUDY: SOXHLET EXPERIMENT
cuuuunvEAcnxn- v>, HME
Figure 4
E*ClH!W.'cONIH)LEXFBi)»e(r (uffiuu)
V IARGE-2
TIE (C«S)
p III t 111 < HI
9 11 13 15 17 19 21
THE (MIS)
u t a i a
. Q U I- U3 9 UR
Figure 5- pH vs Time (S) Figure 6-pH vs Time (M) Figure 7- pH vs Time (L)
-79-
-------
THE EFFECTS OF OVERBURDEN PRESSURE AND HYDRAULIC GRADIENT ON
THE PERFORMANCE OF MODEL SOIL-BENTONITE SLURRY CUTOFF WALLS
Richard M. McCandless and Andrew Bodocsi
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Model soil-bentonite cutoff walls roughly 508 mm (20 inches) in diameter, 559 mm
(22 inches) in height and 102 mm (4 inches) thick were constructed and tested in an
instrumented tank. The effects of overburden pressure (vertical consolidation) and
hydraulic gradient (horizontal consolidation) were investigated followed by tests to
evaluate the potential for closure of artificial windows representing small pockets of
entrapped bentonite slurry in the backfill.
The average hydraulic conductivity of one model was measured for three hydraulic
gradients under each of three applied overburden pressures. Decreases in conductivity
were observed for incremental increases in both overburden pressure and hydraulic gradi-
ent as well as their combined effect. The tests were interrupted on two occasions by
hydrofracture near the base of the model. A reduction in effective stress with increased
depth in the model wall was evidenced by unit weight, water content and vane shear
strength data. By incrementally increasing overburden pressure it was possible to "heal"
two slot-like windows in a subsequent wall suggesting that in situ consolidation of the
backfill may serve to eliminate minor as-built or chemically-induced hydraulic defects in
real slurry walls.
INTRODUCTION
Slurry trench cutoff walls were first
used in the United States in the early
1940's. Since that time, their use has
become more widespread and now includes
application as hydraulic barriers to
control the movement of contaminated
groundwater from hazardous waste disposal
sites. Each application is unique and
requires site-specific engineering evalua-
tion. Nevertheless, the current state-of-
the-art involves fundamental concepts,
performance criteria, and methods common
to all applications.
One area of special interest deals
with questions of in situ consolidation
of soil-bentonite backfill after place-
ment. If significant consolidation does
occur after construction, the average
hydraulic conductivity of the barrier may
be significantly different from that meas-
ured in the laboratory during the design
and construction phases. Moreover, the
finished barrier may have an inherent
ability to eliminate minor construction
defects (via consolidation) given suffi-
cient time. Field observations suggest
that only limited in situ consolidation
may occur (evidenced by lack of subsidence
of the backfill surface) which, in turn,
suggests that friction between'the backfill
and the soils comprising the walls of the
cutoff trench may strongly affect the ver-
tical distribution of effective stress in
the backfill.
An experimental tank in which model
soil-bentonite walls could be constructed
and tested under various hydraulic and
surcharge loading conditions was construct-
-80-
-------
ed for this study. The tank accomodates
circular cutoff walls roughly 559 mm
(22 inches) in height, 102 to 152 mm (4 to
6 inches) thick, and up to 610 mm (24 inch-
es) in diameter. The tank is of stainless
steel construction and employs a pneumatic
bladder system to vertically confine and
consolidate the model wall during permea-
tion in the horizontal direction. A sche-
matic of the system is shown as Figure 1.
Although relatively small and constructed
using non-typical methods (described
later), the models were representative
of real soil-bentonite slurry walls in
most other respects.
CONSTRUCTION AND TEST METHODS
A total of three model walls were
constructed and tested during the course of
this study. The first model was used for
preliminary "shakedown" testing of the
system. The second model was built to
study the effects of surcharge pressure .
and hydraulic gradient on hydraulic con-
ductivity. Testing involved the sequen-
tial application of three hydraulic
gradients (i = 21, 42, 83) under effective
overburden pressures of 41.4, 82.7 and
165.5 kPa (6, 12, 24 psi) as measured at
the surface of the wall. The third wall
was used to investigate the surcharge pres-
sure necessary to close two slot-like win-
dows during permeation. The remainder of
this article describes the latter two
models and the methods and results.
Both models comprised a soil-bentonite
backfill designed for a target hydraulic
conductivity of 1.0 x 10~7 cm/sec under
nominal levels of consolidation and low
hydraulic gradients. The backfill consist-
ed of 1% ordinary (unaltered) bentonite
in a base soil of approximately two-thirds
fine to medium brick sand and one-third
clay of moderate plasticity (CL classifi-
cation per USCS) which contained minor
amounts of silt and fine sand.
The model walls were constructed
between two concentric PVC (polyyinyl
chloride) slip forms representing the walls
of a circular cutoff trench. The forms
were positioned in the tank and backfilled
with clean fine sand in 102 mm (4 inch)
lifts creating an empty 102 mm (4 inch)
wide annular space between the forms. This
space was, then filled with a 5% bentonite:
water slurry (weight-.volume basis) compris-
ing the same bentonite .used in the soil-
bentonite mix. The soil-bentonite backfil-
ling operation varied slightly for differ-
PERMEANT RESERVOIR j
(typ.)
f | PORE PRESSURE
1 PROBE
1
I V8B
e e—•
I C3t=3t=!
___ ; ___ _£ ----
I PVC MEMBRANE ] ___
BEARING T
DRAIN
AIR PRESSURE LINE
PERMEANT FLOW LINE
Figure 1. Schematic of the slurry wall tank system.
-81-
-------
ent models but generally involved raising
both forms about 102 mm (4 inches), allow-
ing the bentom'te:water slurry to penetrate
the sand and form a surface filtration
slurry seal, and then backfilling with
soil-bentonite using a pressurized tremie
pipe. This general procedure was repeated
until the surface of the model wall was
level with the surface of the center core
of sand (sand encircled by the model wall)
and outer ring of sand (sand encircling the
model wall).
After construction, the model was
readied for testing by installing a com-
bination membrane/hydraulic cutoff over its
surface and positioning concentric load-
bearing plates over each element of the
model (core sand, soil-bentonite wall,
outer ring of sand). This arrangement
allowed for differential loading and con-
solidation of the soil-bentonite wall
relative to the adjacent sand bodies. The
surface membrane/hydraulic cutoff over the
top of the soil-bentonite wall was designed
to promote horizontal flow through the wall
during permeation and eliminate leakage
over the top of the model. The cutoff
system comprised three concentric PVC rings
bonded to the PVC membrane and forced into
the surface of the soft soil-bentonite. A
similar configuration involving two cutoff
rings was used on the bottom of the model
as well. Schematics of both the top and
bottom cutoffs appear in Figure 2.
The typical testing procedure used in
evaluating the effects of overburden pres-
sure and gradient involved saturation of
the sand elements of the model, application
of a selected surcharge pressure, consoli-
dation of the soil-bentonite wall under the
applied surcharge (time estimated from con-
ventional consolidation tests performed on
the backfill material), application of the
design hydraulic head pressure at both the
top and bottom of the saturated center core
of sand (Figure 2.), and the measurement of
hydraulic head and volumetric inflow at
prescribed time intervals.
This test procedure was performed for
the three hydraulic gradients under each of
the three overburden pressures reported
earlier. In each case, equilibrium flow
conditions were established before applying
a new set of test conditions.
Similar procedures were used in the
construction and testing of the third model
b)
Figure 2. Schematic of a) surface membrane
and upper hydraulic cutoff,
b) lower hydraulic cuttoff.
wall to evaluate the closure of artificial
slot-like windows via surcharge pressure.
The slots were intended to model macro-
defects such as small pockets of entrapped
slurry remaining after construction of the
wall. Two slots approximately 7.9 mm (5/16
inch) wide by 1.6 mm (1/16 inch) high were
cut into the third wall after preconsolida-
tion under an effective overburden pressure
of 41.4 kPa (6.0 psi) as measured at the
surface of the wall. The windows were
positioned 180° apart at a depth of about
127 mm (5 inches) below the top of the
wall. Both ends of each slot were covered
with a fabric-wrapped wire mesh to prevent
washing the core sand into the slot during
permeation. The test procedure involved
incremental increase of overburden (sur-
charge) pressure until the slots were
effectively closed as evidenced by a return
to the predetermined baseline hydraulic
conductivity of the model.
The method used to compute hydraulic
conductivity was the same for the over-
burden pressure/hydraulic gradient series
of tests and the single window closing
test. Using measured values of total head
and the geometric constants of the models,
the average hydraulic conductivity was
computed using the falling head relation-
ship:
-82-
-------
a«L
,
k =
where: k = hydraulic conductivity (cm/sec)
a = cross-sectional area of permeant
reservoir (cm2)
L = thickness of the wall (cm)
A = cross-sectional area of the wall
perpendicular to the direction
of flow (cm2)
hg = initial total hydraulic head (cm)
h} = final total hydraulic head (cm)
t = time interval between two read-
ings (sec)
RESULTS AND DISCUSSION
Overburden Pressure and Hydraulic Gradient
The testing of the second soil-benton-
ite wall, termed Sequence 2 tests, involved
staged increases of overburden pressure and
hydraulic gradient over a period of six
months, followed by sampling and measure-
ment of unit weight, vane shear strength
and moisture content as a function of depth
in the model. Figure 3 represents a com-
posite of Sequence 2 hydraulic conductivity
results for the six sets.of test conditions
indicated (2(a) through 2(g)). All six
curves are of similar shape and for the
most part their relative positions present
a logical picture of decreasing equilibrium
hydraulic conductivity with increasing
hydraulic gradient for any given
overburden pressure. Note however, that
test 2(c) is missing, test 2(f) was termi-
nated prior to achieving equilibrium condi-
tions, and that data for tests 2(f) and
2(g) do not conform to the otherwise
logical trend described above. The expla-
nation appears in Figure 4 which presents
a chronological summary of the final equi-
librium conductivities measured for each
set of test conditions.
In the typical test, the wall model
: exhibited a high initial hydraulic conduc-
tivity represented by an open triangle,
then dropped off and gradually approached
the final equilibrium value designated by
an open circle. Two incidences of hydro-
fracture, one resulting from a ruptured
surcharge bladder (loss of applied over-
burder pressure), are indicated by solid
triangles.
The first incidence of hydrofracture
occurred near the end of test 2(b) strong-
ly suggesting that the effective stress at
some point in the model (presumably near
its base) was less than the effective over-
burden pressure. Because of the hydrofrac-
ture, test 2(c) was not attempted. In-
stead, test 2(d) involving a doubling of
applied overburden pressure and a reduction
in hydraulic pressure was initiated in an
attempt to heal the first hydrofracture.
Test 2(f) also suffered a rupture of the,
surcharge bladder resulting in a second
hydrofracture of the soil-bentonit.e wall.
To estimate the extent of damage,
test 2(g') was conducted at overburden and
hydraulic pressures matching those of test
2(d). As shown in Figure 4, the results of
test 2(g') did not match those of test 2(d),
having an equilibrium hydraulic conductiv-
ity almost an order of magnitude higher.
These data clearly indicate permanent dam-
age of the model as a result of the two
cases of hydrofracture. Results for test
2(g), however, illustrate that a large
increment in surcharge pressure can par-
tially offset the effects of hydrofracture.
In the field, both vertical and hori-
zontal consolidation of the backfill occur
simultaneously. The equilibrium conductiv-
ity.values reported in Figure 4 may there-
fore be regarded as representing the com-
bined effect of both vertical (surcharge)
and horizontal (hydraulic gradient) consol-
idation stresses for the test conditions
applied.
Except for test 2(g), the data sug-
gest a logical trend of decreasing equili-
brium hydraulic conductivity as a func-
tion of either increasing surcharge pres-
sure (compare results of tests 2(b) and
2(e)) or increasing hydraulic gradient
compare results of test 2(a),(b) and 2(d),
(e),(f)). Although the observed trend is
logical, the data fail to reflect the
correct magnitude of change in hydraulic
conductivity between successive tests in
several cases. For example, the actual
equilibrium hydraulic conductivity for test
2(d) (and therefore also for all subsequent
tests) is lower than reported. As describ-
ed earlier, the reason is that hydrofrac-
ture permanantly changed the properties of
the wall, thus artificially offsetting
groups of data measured after hydrofrac-
ture from other groups of data measured
before hydrofracture. In terms of inter-
nally consistent groups of data then, the
following two sets of tests may be recog-
nized: 2(a),(b) and 2(d),(e),(f) (as
-83-
-------
o -r
a
o
o
cc
Q
10 •
;
5
-
—
5
—
—
r
^
—
—
a
~~
)
\
tU
l\
\I\J
V\ v
'C\
TEST SEQUENCE
2
Effective
Overburden Hydraulic Gradient
TEST Pressure Pressure
(psi) (psi)
2(a) 6 3 20.8
2(b) o o 41.6
2(d) 12 3 20.8
2(e) 12 6 41.6
2(f) 12 12 83.2
2(g) 24 3 20?8
^r^^
\
.^^-2(1)
A>
f\-^
1 X--
V
\^.". —
—• ~2(d)
\x\
2(g>
/2<9)
0 200 400 600 800 1000 1200
TIME (HOURS)
Figure 3. Sequence 2 hydraulic conductivity
test results
projected).
The equilibrium values shown in Figure
4 represent the combined effect of vertical
and horizontal consolidation. Plotted in
terms of combined (total) vertical and hor-
izontal effective stress with the corre-
sponding permeability data on an arithmetic
scale, these same data would appear as
shown in Figure 5. The plot very closely
resembles a typical void ratio versus pres-
sure curve obtained from a conventional
consolidation test. The interpretation of
the data is also similar in that beyond a
certain point, the rate of change in perme-
ability (or void ratio) greatly diminishes
for a unit change in total stress, i.e.,
the material becomes more difficult to con-
solidate the more consolidated it is.
Although the two sets of data (2(a),(b) and
2(d),(e),(f)) are not strictly compatible
as explained earlier, they do reflect a
logical and expected trend, and permit the
following general observations:
• Test set 2(d) ,(e),(f), exhibits greater
changes in hydraulic conductivity for the
same change in hydraulic pressure reflect-
ing the reconsolidation or "healing" of a
soil-bentonite wall that had been damaged
(hydrofractured) to an undetermined extent.
• The change in equilibrium hydraulic con-
ductivity due to either a unit change in
hydraulic head pressure or a unit change
in surcharge pressure can be on the same
order of magnitude, i.e., the effect of
horizontal (gradient-induced) consolidation
can be as large as the effect of vertical
(surcharge) consolidation for a comparable
pressure change.
After the completion of Sequence 2
test 2(g) reported in Figure 4, the tank
-84-
-------
was opened to permit visual inspection of
conditions and allow for sampling and test-
ing of the consolidated backfill. The sam-
pling included undisturbed tube samples and
water content samples of the soil-bentonite
at various locations and depths. Testing
involved measurements of unit weight, vane
shear strength, and water content. Data
for these parameters appear as a function
of depth in Figure 6.
Despite the semi-qualitative nature
of these data (due to hydrofracturing of
the model) they clearly demonstrate that
the effective overburden stress applied
at the surface of the wall did not act
over the full depth of the wall. In other
words, overburden stress was dissipated
with depth, with the net result that the
final values of dry density and vane
shear strength are highest near the top
of the wall (highest degree of consolida-
tion) and final water content is highest
at the base of the wall (lowest degree of
eonsolidation).
Window Closure
The probability of "windows" existing
in a field-scale slurry wall, whether re-
presenting an as-built condition related to
materials or construction technique, or
resulting from long-term exposure to chemi-
cals, is not known. Presumably, undetected
subsurface windows might develop due to en-
trapment of slurry or sloughed trench wall
materials within the backfill during the
backfilling operation. According to Evans
et al.t1) the probability of such entrap-
ments along the surface of the advancing
soil-bentonite "mud wave" is high. More-
over, limited evidence of these types of
"windows" was observed during the post-test
inspection of the model.
After Sequence 2 testing a third wall
-s
i
..~
—
-7
-8
-9
<
\
, .
^D
r
J,
1
e/3 2(b)
8/6
3
0 8
1
\
}
\
\
\
I
2
12/3
\
\
\
) ** * * \ ?
1
•. 2 C 1 ) llj
\12/12 2(8)
! 0 1
iO 1
. EXPLANATION
surcharge pressure, psl
»• 6/3 -* — ^
hydraulic pressure, psi'
w— initial
©— equilibrium
V~ hydrof racture
9 — surcharge bladder
rupture
0 — projected to
equilibrium
10
a
z
o
o
TIME (Days)
Figure 4. Chronology and results of Sequence 2
hydraulic conductivity tests.
-85-
-------
t 90
P
u
2 40
20 -
2(a)
(d)
^
\
\
(a)
V
l
18
l
21
t2 IS
COMBINED VERTICAL AND HORIZONTAL
EFFECTIVE CONSOLIDATION PRESSURE. Pll
r
24
Figure 5. Equilibrium hydraulic conductivity as a function
of combined effective overburden (vertical) and
hydraulic (horizontal) pressure.
DEPTH
(In.)
|
4-
1 2-
DRY
1 1
UNIT WE
(Ib/ft3)
4 116 11
/
•
X-"
• — .
*-
->'
S
K.«
._-}
^m^.
X
X
r*"
/
t •
WATER CONTENT (%)
14 16 18 20
«• I
vl
«•
\
i
\
\
\
\>
\
note: symbols designate sample location along different imaginary vertical
lines through model
Figure 6. Results of tests on soil-bentonite backfill after
completion of Sequence 2 testing.
-86-
-------
was constructed for the window closing
test. Before the slot windows were formed
the new wall was preconsolidated under an
effective overburden pressure of 41.4 kPa
(6 psi) and a hydraulic pressure of 20.7
kPa (3 psi) representing a gradient of
20.8. Baseline (no window) results for
test 3(a) are shown in Figure 7. The
hydraulic conductivity versus time curve
has the typical shape of the previous
Sequence 2 data shown in Figure 3.
After establishing a baseline or ref-
erence value of hydraulic conductivity, the
two slot windows were formed at the loca-
tions and depths previously described.
Overburden pressure was then gradually
increased causing the apparent hydraulic
conductivity of the model to decrease
until the windows had been effectively
closed as evidenced by a return to the
measured baseline conductivity. These
data also appear in Figure 7 and show
that the average conductivity of the
model was lowered relative to the target
baseline value under the final set of
test conditions. For this reason it was
not possible to identify a specific
threshold value of surcharge pressure
necessary to heal the two slot windows.
The success of this window closing
test has important ramifications for real
-8
10 ~
-7
10 -
Baseline Curve
TEST 3(a)
window Cloilng Curve
Overburden / Hydraulic
j Initial : 1.0/1.0 pal
p : 1.5/2.0 pal
I
i Final : 2.0/3.0 pal
io-8-
Window Closing Test Initiated
I
800 1200 1800
TIME (HOURS)
2400
Figure 7. Baseline and window closing hydraulic
conductivity test results.
-87-
-------
slurry walls. It means that the effective
overburden pressure in the wall may serve
to close residual slurry windows and may
even close a multitude of micro shrinkage
cracks that may develop in the backfill
over the life of the barrier due to the
effects of chemical leachates.
ACKNOWLEDGEMENTS
The research described herein was sup-
ported wholly by the Land Pollution Control
Division of the U.S. EPA Hazardous Waste
Engineering Research Laboratory under
contract #68-03-3210 to the Department of
Civil and Environmental Engineering, Uni-
versity of Cincinnati. Appreciation is
expressed to project officer Joseph K.
Burkart and Work Assignment Manager Naomi
P. Barkley for their administrative and
technical support. The authors also wish
to acknowledge and thank graduate student
Jong Jen (Steve) Lin for his enthusiastic
dedication to the work and Frank E.
Weisgerber for his technical contributions.
Finally, we gratefully acknowledge base
operations manager Gerard Roberto for
his considerable efforts on behalf of
this project and others at the Center
Hill Solid and Hazardous Waste Research
Facility.
REFERENCE/SELECTED BIBLIOGRAPHY
1. Evans, J.C., 6.P. Lennon and K.A.
Witmer. Analysis of Soil Bentonjte
Backfill Placement in Slurry Walls,
in: Proceedings of the Sixth Annual
National Conference on Management of
Uncontrolled Hazardous Waste Sites,
November 4-6, 1985, Washington, D.C.,
pp. 357-361.
Ayers, J.E., D.C. Lager and M.J.
Barvenik. The First EPA Superfund
Cut-off Wall: Design and Specifica-
tions. Presented at the Third National
Symposium on Aquifer Restoration and
Groundwater Monitoring, 1983.
Barvenik, M.J., W.E. Hadge, and D.T.
Go!berg. Quality Control of Hydraulic
Conductivity and Bentonite Content
During Soil/Bentonite Cutoff Wall
Construction, in:Land Disposal of
Hazardous Waste. Proceedings of the
Eleventh Annual Research Symposium,
Cincinnati, Ohio, April 1984.
EPA/600/9-85/013, pp. 66-79.
D'Appolonia, D.J. Slurry Trench Cut-
off Walls for Hazardous Waste Isolation.
Technical Paper.Engineered Construc-
tion International, Inc., Pittsburgh,
Pennsylvania, April 1980.
Evans, J.C., H.Y. Fang and I.J.
Kugelman. Containment of Hazardous
Materials with Soil-Bentonite Slurry
Walls, in": Proceedings of the Sixth
National Conference on the Management
of Uncontrolled Hazardous Waste Sites,
November 4-5, 1985; Washington, D.C.,
pj>. 369-373.
Evans, J.C., H.Y. Fang. Geotechnical
Aspects of the Design and Construction
of Waste Containment Systems. Proceed-
ings of the National Conference on the
Management of Uncontrolled Hazardous
Waste Sites, November 1982.
JRB Associates. Slurr.
for Polution Control
Trench Construction
-y Tr
EPA-
•540/2-84-001.
U.S. Environmental Protection Agency,
Municipal Environmental Research Labora-
tory, Cincinnati, Ohio, February 1984.
-88-
-------
EXPERT SYSTEMS TO ASSIST IN DECISIONS CONCERNING LAND DISPOSAL OF
HAZARDOUS WASTES
Daniel G. Greathouse
United States Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
In FY'84 the Hazardous Waste Engineering Research Laboratory successfully
developed a small proof-of-concept expert system to assist in interpretation of
chemical immersion test (EPA Method 9090) data for PVC liner materials. This was the
beginning of an orderly progression of efforts to assess the feasibility of using expert
systems to assist in permit reviews for hazardous waste land disposal sites. Permit
review decision areas amenable to expert system applications have been identified and
several systems are in various stages of development and testing. The rationale for this
approach to provide decision support aids for permit review include the complexity of the
required engineering evaluations; availability of extensive relevant research results and
known subject-specific specialists (experts); concern that permit reviewers do not have
all of the required expertise and that they have little, if any access to subject specific
specialists; concern that the reviewers do not have sufficient time to assimilate all
regulatory policy and research information; and concern that decisions may not be
consistent among reviewers or with EPA regulations and policies. The decision areas
selected for expert system development and the progress on the ongoing development efforts
will be presented.
INTRODUCTION
Review of permits for land disposal
of hazardous wastes requires numerous
decisions concerning technical and policy
issues. Some require interpretation and
application of information in research
reports, others involve interpretation and
evaluation of specialized test data, and
others involve assessment of compliance
with latest regulatory policies. Special-
ized knowledge concerning a number of
technical areas and a broad base of
environmental regulatory experience are
necessary to adequately perform these
reviews. This need for current knowledge
and background in addition to the concern
that reviews be consistent (i.e., permits
are judged the same by all reviewers)
prompted our interest in expert systems.
Expert systems are computer programs
(software) designed to provide advice
concerning specialized areas. The design
objective of the programs is to emulate
the advice of subject specialists by
incorporating the decision rules or
criteria that they use in terms of IF —-
THEN statements. For example, IF the
carbon black content of an HOPE liner is
less than 2 percent THEN the liner is
inappropriate. Two characteristics of
these programs differentiate them from
traditional programs, namely (1) they are
essentially large pattern matching rou-
tines that seek a solution (advice) that
corresponds to the pattern of input data
and (2) the program logic is separated
from the rules (IF -— THEN statements) to
facilitate ease of rule modification or
refinement. Like traditional programs
these systems can perform engineering and
statistical calculations, but these are
not the primary feature of these systems.
This paper presents the history,
current status, and future direction of
the expert systems development program
supported by the Land Pollution Control
Division of the Hazardous Waste Engineering
Research Laboratory. The development
methods being used and some of our
experiences are also presented.
-89-
-------
HISTORY OF EXPERT SYSTEMS DEVELOPMENTS
In 1984 a small proof-of-concept
expert system was developed to assist in
assessment of chemical compatibility of
PVC liner materials based on chemical
Immersion test (EPA Method 9090) data.
This system demonstrated that at least
for this specific application it is
feasible to extract sufficient information
from an expert to formulate advisory
rules for some decisions that must be made
by permit reviewers. This system contained
20 rules and was written in BASIC for the
IBM PC microcomputer. To determine what
other permit review decision areas would
be amenable to expert system applications,
a Requirements Study was initiated.
Other issues addressed were development
strategy, software and hardware require-
ments, and associated resource requirements.
Table 1 presents the prioritized list of
decision areas identified by this project,
which was completed in the fall of 1985.
In order to facilitate compatibility among
the systems designed to provide assistance
1n the different decision areas, and to
facilitate information sharing among
them, it was recommended that a supervisory
system with decision-specific submodules be
developed to address permit review. It
was also recommended that development be
performed on AI dedicated hardware with
specialized software due to the limitations
of the PC-based expert systems software
at the time. The former recommendation
was rejected due to the concern that the
deadlines for performing most permit
reviews would be past before a comprehensive
permit review system could be built, hence-
the strategy of developing independent
modules to address individual, narrowly
defined decision areas. The latter
recommendation was rejected due to the
concern over potential high equipment
costs (for development and possibly for
running the systems) and the uncertainty
that it would be feasible to develop
software on the specialized hardware for
use on the available micro-computers. Our
objective was to make the systems available
for use by everyone in the EPA or the
states responsible for permit review. Due
to these concerns and the rapidly advancing
capabilities of micro-computer based
expert systems languages and shells the
decision was made to develop the systems
on the PC for use on the EPA standard IBM
PC/AT microcomputers. Also the goal of
wide distribution of the systems has been an
important factor in selection of particular
microcomputer "software. Strong preference
for minimal licensing costs for run time
versions in addition to flexibility for
tailoring to particular applications,
have been the driving considerations for
software selection. The recommendations
of the Requirements Study concerning
decision areas provided direction for
initial development efforts and raised
the issues of development strategy and
software/hardware selection for consider-
ation and resolution.
The initial proof-of-concept PVC
liner resistance system was redone by an
EPA contractor. It was expanded to
include advisory rules for HOPE and CSPE
liner materials and refined by incorpo-
rating the advice of additional experts.
This system is ready for limited use by
permit reviewers. A Waste Analysis
advisory system has also been developed
and is ready for limited use. Current
development efforts include systems to
assist in evaluation of Closure Plans,
Surface Impoundments, and Site Selection
and to assist in screening of control
technologies for CERCLA sites. Each of
these systems is described in further
detail later in this paper.
SYSTEM DESCRIPTIONS
The FLEXIBLE MEMBRANE LINER RESISTANCE
system includes rules to assist in inter-
pretation and evaluation of chemical
immersion (EPA Method 9090) data for PVC,
HOPE, and CSPE liner materials. The
primary input data for the system are
physical property measurements obtained
during a four-month time period; the 9090
test requires immersion of coupons of the
liner material in the waste to be disposed.
Information is also included concerning
the carbon black content (for .HOPE mate-
rials), intended service life, etc.
Based on this input the system determines
if there is evidence that the proposed
liner material may not be resistant to
the wastes that will be deposited at the
site. An example of input data is pre-
sented in Table 2 and the resulting
conclusions are shown in Table 3. The
data are entered on Lotus 1-2-3 type
screens and can be edited, saved, plotted,
and/printed out on the screen or hard
copy. This is different from the typical
interactive, or question-driven, input of
many expert systems. Advice corresponding
-90-
-------
to. any set of input data is derived by
transferring the data and some simple
statistics (means, percent changes, etc.)
to an'expert system.written in Arity
Prolog. Conclusions that a liner material
is not chemically resistant to the subject
waste/1eachate are presented separately
for each physical property examined.
Nothing is stated concerning those proper-
ties that do not indicate a failure to be
resistant. For example, in the sample
data in Table 2, since weight does not
change over time, nothing is said concern-
ing weight change in the conlusions in
Table 3. Excessive scatter in the data,
which makes trend analysis uncertain, is
also noted as illustrated by tensile
strength. All of the systems have been
developed for application on an IBM PC/AT
microcomputer. This system is currently
being reviewed by the Office of Solid
Wastes in preparation for release and
introduction to the EPA Regional Offices
and the States in performing permit
reviews of land disposal facilities.
This system was developed by an.EFA
contractor and is an expansion and refine-
ment of the original prototype system
developed by the Laboratory. Since the
expert used as the resource for the initial
system is a well known recognized expert
in the field of flexible membrane testing,
he was also used as one of the subject
specialists (experts) and consultant for
this phase of the development effort. The
fact that he gave different responses to
questions by the contractor personnel
than to the same questions by the EPA
person, and did not recall the advice «
recorded by the EPA person, illustrates
the importance .and difficulty of acquiring
expert opinions. This problem has been
alleviated in this follow-up system by
acquiring the input of multiple experts •
and by comparisons of their advice with,
that of the finalized system on sample
test data. Further testing of the rule
bases and software is also being performed
by the Office of Solid Waste to insure
compliance with regulatory policies and
guidance documents, to identify any errors
in the software, and insure that the
users' manual i.s correct and adequate.
Thorough, documentation, including identi-
fication of the source for each rule
in the rule base, is a requirement for all
of the systems.
The, WASTE ANALYSIS PLAN EVALUATION
SYSTEM assists in the identification of
chemical incompatibilities,that may occur
when different waste chemicals are handled4
stored, treated, or buried together. ,It
also recommends the appropriate sampling
equipment and analytical methods for
monitoring the waste stream. The knowledge
base (rules and information) for this
system was developed by the expert who
prepared the manual entitled "Waste
Analysis Plans: A Guidance Manual" for
the Office of Solid Waste. Other resource
materials included regulations in 10 CFR
Part ,21 "Identification and Listing of
Hazardous Wastes" and 40 CFR 264 Waste
Management Facility Regulations; and
guidance from regulatory personnel in the
Office of Solid Waste and permit reviewers
from two EPA Regional Offices. This
system was also developed in Arity Prolog
for.application on the IBM PC AT micro-
computer. It has been tested in two of
the EPA Regional Offices and one state.
.The waste analysis evaluation system
will be finalized by the EPA contractor
by early spring of 1987 and submitted to
the Office of Solid Waste for Review and
testing. This system is effectively an
intelligent data base system that facili-
tates implementation of the evaluation
procedures specified in the Waste Analysis
Plan Evaluation Document prepared by OSW..
As such it helps to illustrate one of the
uses of expert systems methods. One of
the primary features of the system is the
identification of chemical incompatibilities
among the waste compounds. Typically
this would be done by a professional
chemist and/or by consulting available
tables. Many of the permit reviewers are
not chemists, however, and looking up
every pair of chemicals in an appropriate
table is time-consuming. The waste
analysis system reduces the time for this
determination, which is another reasonable
purpose of expert systems.
The SUBSIDENCE EVALUATION SYSTEM was
developed to identify waste characteristics,
disposal procedures, climatic conditions,
etc. that increase the risk of differential
subsidence of a landfill after closure.
This system incorporates the observations
and insights of coprincipal investigators
on an EPA-supported project to identify
and evaluate the causes of, post-closure
-91-
-------
differential subsidence. As are the
earlier described systems, this one is
also written in Arity Prolog for applica-
tion on the IBM PC/AT microcomputers.
The rules for this system are ready for
inclusion as part of a system to evaluate
the adequacy of proposed cover systems.
The subsidence system was a learning
experience by the EPA staff to assess the
feasiblity of summarizing the results of
a recently completed research project in
terms of an expert system. As a result
of this effort, one can learn to understand
the factors that contribute to post-closure
differential subsidence by either reading
the final project report or using the
parallel expert system. This further
illustrates potential applications for
expert systems techniques. A diskette
could be provided with each major research
report or as a stand-alone product. The
expert system would facilitate use of the
basic material and the final report would
provide further discussion and/or resource
materials.
The CONTROL TECHNOLOGY SCREENING
SYSTEM was designed to assist in screening
of technologies to contain or remove
hazardous wastes from CERCLA (Superfund)
sites. The technologies included in the
system database are arranged in a four-
level hierarchy. These levels consist of
1) generic action (containment or removal),
2) action class (e.g., capping, subsurface
barriers, etc.), 3) technology class (e.g.,
single layer caps, horizontal barriers,
vertical barriers, etc.), and 4) distinct
technologies (e.g., soil liner cap,
synthetic liner cap, slurry wall, etc.).
The rules for this system were developed
from information contained in available
reports and documents published and/or
supported by the EPA (see Table 3). The
system is not complete and ready to use
since it contains a number of information
gaps and needs refinement. It is an
excellent prototype system that demon-
strates the potential for this type of
application. It has also been written
in Arity Prolog for application on the PC
AT microcomputer.
The technology screening system is
being directed towards the need of the
EPA Site Coordinators responsible for
initial feasibility investigations of
Superfund sites. In addition to its basic
screening role, it has features that
could be used'as a training tool and
information resource. Thorough Help
screens are included to define and explain
terms and selection criteria, and the
rules for assessing suitability of a
particular technology are presented in
self-explanatory terms. The input data
for each run can be saved and selectively
modified as part of a sensitivity analysis.
This system, which is based on the infor-
mation abstracted from mutiple resource
materials, illustrates the feasibility of
using expert systems concepts to integrate
information from multiple sources into a
unified information resource.
The SURFACE IMPOUNDMENT SAFETY
ANALYSIS SYSTEM is being designed to
provide user-friendly access to available
analytical routines for estimating safety
of selected components of surface impound-
ments. The system will also assist in
interpretation of the safety factors that
results from the calculations. The
development language or expert system
shell has not yet been decided but it
will likely be written in Insight 2 (a PC
expert system shell) with the analytical
routines written as Fortran subroutines
that are accessed thru links written in
C. The delivery system will be written
for application on the IBM PC AT micro-
computer. Scheduled completion is late
summer or fall of 1987.
The CLOSURE PLAN EVALUATION SYSTEM
is being developed to assist in review of
closure plans (for land disposal sites
that will close due to failure to apply
for a part B permit or failure to receive
approval of their part B permit) and
permits for closure (for land disposal
sites that submit a part B and receive
approval). The initial system will not
be applicable to "clean" closures for
surface impoundments. The system has
been divided into the following 6
components:
1. Cover systems
2. Runon/runoff control
3. Final vegetative cover
4. Ground-water monitoring program
5. Leachate collection and treatment
systems
6. Scheduled maintenance and monitoring
of al1 systems
-92-
-------
Development is being phased and staggered.
Initial efforts will be devoted to initia-
tion and development of one system.
Midway through the time for completion of
the first project, the second one will be
initiated. When the first system is
finalized the third one will be initiated
while efforts on the second continue. It
is anticipated that the first system will
be completed by the first of August of
1987 with each successive system due for
completion 2 months after completion of
the former, i.e., the second is scheduled
for completion by the first of October,
1987, the third by the first of December,
1987, etc. The total time for the 6
modules is scheduled at 14 months. The
delivery machine will be the IBM'PC/AT
microcomputer with minimal required
enhancements such as a graphics card.
The decision concerning software has not
yet been made.
DISCUSSION " . ,
All of the expert systems supported
by the Hazardous Waste Engineering Research
Laboratory address issues related to
containment of hazardous wastes in a land
disposal site or removal to another site.
The decisions addressed by each are
sufficiently limited in scope that the
systems are small enough for implemen-
tation on the standard IBM PC/AT micro-
computer. The primary focus of the
systems are technical issues that have
been researched by the Laboratory during
the past years. Hence EPA-sponsored
research reports have been a major source
of information for knowledge-base formation
for a number of the systems. That is one
reason we believe that expert systems (or
knowledge-based systems) can play a
significant role in terms of automated
technology transfer systems for the
Laboratory.
Based on our experience, prototype
systems can be developed to address
relevant questions without a major commit-
ment of resources and time. Some of the
smaller inhouse systems were developed
within 2 or 3 months by one person assisted
by a microcomputer specialist. Development
of prototype systems, however, is not the
major task. The major effort is to
prepare a finalized system for general use
by decision officials both within and
outside the Agency. Some of the time-
consuming and difficult tasks include
knowledge-base refinement to address latest
research1 results, policy considerations,
and user needs and desires; testing the
knowledge bases to insure they reflect the
opinions of the experts or advice contained
in the source materials; testing of the
source code to insure freedom from coding
and programming errors; documentation of
the source code and preparation of an
users' manual; maintenance and refinement
of the systems; and provision of readily
available and accessible user support.
In order to insure that the systems
satisfy the needs of the user community,
we are working interactively with the
users as the systems are initially con-
ceived, developed, and refined. One of
the systems (Flexible Membrane Liner
Resistance) is being tested by comparing
the conclusions of the system on sample
test data with those reached by the
experts without assistance from the
system. All rules are being documented
by source (expert or specific report
citation). The other issues are being
discussed with representative from the
Office of Solid Waste but have not yet
been resolved.
Based on the success of our develop-
ment efforts, responses from the user
community, and the interests expressed by
management and regulatory personnel , the
next major expansion area will be treat-
ment technologies. This will include
screening of treatment methods or techno-
logies for the RI/FS process for Superfund
sites and/or remediation of RCRA sites,
evaluation of treatment test data (for
example trial burn data), and evaluation
of proposals for particular treatment
regimens to handle particular waste
streams.
CONCLUSIONS
It is feasible to develop usable and
useful expert systems. Expert systems
appear to offer significant opportunities
for insuring expedient implementation of
latest technologies, compliance with
regulatory policies, and decision consist-
encies. Hence we expect to continue
efforts to complete systems under develop-
ment, maintain ongoing systems and develop
new ones to assist in other amenable decision
areas.
-93-
-------
TABLE 1
PRIORITIZED LIST OF PERMIT REVIEW DECISION AREAS AMENABLE TO EXPERT
SYSTEM APPLICATIONS (As Of December 1985)
Priority 1
Priority 2
Priority 3
Priority 4
Priority 5
- Facility Closure Plans
- Facility Location Information
- Synthetic Liner Design
- Dike Design and Construction
Lower Soil Liner Design
Secondary Leachate Collection
Primary Leachate Collection
Facility Post-Closure Plan
Landfill Run-on Control System Design
Lower Soil Liner Construction
Landfill Run-off Control System Design
Waste Analysis Plan
Plan for Management of Units Associated with Landfill Run-
on and Run-off Control Plan
Chemical and Physical Analysis of Hazardous Waste to Be
Handled
Prevention of Overtopping
Landfill Wind Dispersal Flow Control Plan
Construction Quality Assurance Plan
Synthetic Liner Construction (upper and lower)
Inspection Plans
Plans for Preparedness and Prevention
Contingency Plans
Precautions Taken to Prevent Ignition or Reaction of Wastes
Planned Traffic Pattern
Closure Cost Estimates and Financial Assurances
Post Closure Cost Estimates and Financial Assurances
Lists of Wastes to be Placed in each Surface
Impoundment/Landfil1
Certification by Qualified Engineer of Dike's Integrity
General Facility Description
Procedures and Equipment to Prevent or Mitigate Hazards
Training Program
Facility Deed Documentation
Insurance Policies
Coverage by State Financial Mechanisms
Topographic Map for the Facility and Environment
-94-
-------
TABLE 2
SAMPLE DATA FROM 9090 TESTING OF HOPE LINER MATERIALS IMMERSED IN A
HYPOTHETICAL WASTE
WEIGHT (grams)
Time
(months)
Time
(months)
Time
(months)
Time
(months)
Sample Coupon
1
2
3
4
5
Average
% Change
0 1
1.83 1.83
1.87 1.87
1.85 1.86
1.85 1.86
1.87 1.88
1.86 1.86
0.19
0 2
1.87 1.87
1.87 1.87
1.87 1.88
1.84 1.84
1.87 1.88
1.86 1.87
0.28
0 3
1.85 1.86
1.85 1.86
1.86 1.87
1.87 1.87
1.87 1.87
1.86 1.86
0.41
0 4
1.84 1.85
1.83 1.84
1.87 1.87
1.87 1.87
1.86 1.87
1.85 1.86
0.25
TENSILE AT BREAK (Ib/in)
Time (months)
Sample Coupon
1
2
3
4
5
Average
% Change
0
3800
3800
3800
3800
3800
3800
0.00
1
2750
3650
4490 .
4190
4020
3820
0.53
2
2450
2520
3810
2570
3490
2958
-22.16
3
2470
2640
2570
2670
2870
2644
-30.42
4
3740
3720
2470
4560
3550
3610
-5.00 :
MODULUS (Ib/in)
Time (months)
Sample Coupon
1
2
3
4
5
Average
% Change
0
82900
84100
83000
84900
85200
84020
0.00
1
96000
95500
93000
93800
94500
94560
12.54
2
87500
86200
87500
89000
85500
87140
3.71
3
71500
72000
79500
74200
76500
74740
-11.04
4
60400
64200
61300
62300
61800
62000
-26.21
-95-
-------
TABLE 3
CONCLUSIONS OF FLEXIBLE MEMBRANE LINER EXPERT SYSTEM FOR SAMPLE
PRESENTED IN TABLE 2
DATA
The liner is fundamentally inappropriate for the application.
The liner is fundamentally inappropriate for the proposed application because either
initially or at some point in the immersion test one or more of the liner's key
physical properties was below the minimum standards established by the National
Sanitation Foundation (NSF). NSF Standard Number 54 was developed in consultation with
several liner industry experts. Standard Number 54 specifies minimum values for several
physical properties for HOPE liners of two thicknesses 0.08 and 0.1 inches (i.e., 80 and
100 mils). There is no NSF standard for HOPE liners of less than 80 mils.
The NSF standard requires modulus of elasticity be more than 80,000 Ibs/sq in. For
this applicant the modulus of elasticity is 62,000.
The immersion test results indicate that the liner is not chemically resistant to the
waste/1eachate.
The immersion test results indicate that the liner material is excessively softened by
the waste/1eachate. Excessive softening is a sign of poor chemical resistance. A
liner material that softens excessively may not be able to withstand the stresses of
the waste site and may fail due to tearing or puncture.
Excessive softening of the liner material is evidenced by an excessive loss in its
modulus of elasticity. At a lower modulus a given force will produce a greater elonga-
tion of the material. Excessive losses in modulus suggest the potential for easier
tear, puncture, and break of the liner.
The loss in molulus of elasticity was -25.21%. This is beyond the threshold limit of
25% used by many liner material specialists.
No conclusion can be made as to the chemical resistance of the liner material to the
waste/1eachate.
The results are inconclusive because there is excessive scatter (i.e., lack of preci-
sion) in the immersion test results.
The average standard deviation of the data for the four-month immersion is greater than
102 for the measurements of changes in:
The tensile strength at break of the liner material.
-96-
-------
MODELING SOIL WATER MOVEMENT IN MINIMUM TECHNOLOGY
WASTE MANAGEMENT FACILITIES
David H. Gancarz1 and Timothy J. Durbin2
1 Radian Corporation
Austin, TX 78766
2S.S. Papadopulos and Associates, Inc.
Davis, CA 95616
ABSTRACT
The hydrology of landfills, surface impoundments, and waste piles is dominated by unsaturated flow of soil water.
Unsaturated conditions normally persist in the cover, solid waste, liner system, leachate collection and removal system,
and in the soil below these systems. A finite element model, UNSAT2D, developed as a generalized computer program
based on the two-dimensional equation of saturated/unsaturated flow, is described. This model can be used to simulate
moisture movement through a two-dimensional vertical section of a facility.
Results from a series of simulations of alternative "minimum technology" designs are discussed, with emphasis on
bottom liner design and leak detection. These data show that leak rates into leachate collection and removal (LCR)
systems in excess of 100 gallons/acre-day are necessary for LCR system drains to flow when constructed over three-foot
thick compacted soil bottom liners with a hydraulic conductivity of 10-7 cm/s. LCR systems built in conjunction with com-
posite (flexible membrane over low-permeability soil) bottom liners are significantly more effective. Time to drain flow is
reduced, minimum leak rate resulting in LCR system drain flow is reduced, and leachate collection efficiency increases.
INTRODUCTION
The Hazardous and Solid Waste Amendments of 1984
require that certain landfills and surface impoundments
have two or more liners designed to contain hazardous
constituents, and a leachate collection system designed to
promptly detect any leaks, should they occur. Similar rules
apply to waste piles.
These requirements constitute the "minimum tech-
nology" for waste management facilities constructed to
treat, store, or dispose of hazardous waste. To predict the
effectiveness of different leachate collection and liner
system designs, it is necessary to have a prior understan-
ding of saturated and unsaturated moisture movement
within such systems. A useful approach to gaining insight
into the hydraulic behavior of these systems is to apply
computer simulation techniques. In this way a broad spec-
trum of design, siting, operational, and failure scenarios
may be modeled and the results used to evaluate signifi-
cant factors that influence leachate collection and liner
system performance.
This paper describes the computer program
UNSAT2D that was developed to model saturated and un-
saturated flow in landfills, surface impoundments, and
waste piles. Following the description, results from a
number of computer simulations are discussed in terms of
leachate collection and liner system performance.
MODEL DESCRIPTION
Soil moisture movement within a facility is described
by the two-dimensional equation of saturated/unsaturated
soil moisture flow within a specified flow domain. A
generalized representation of the flow domain Q is shown
in Figure 1. That domain is defined by the union of boun-
dary surfaces S.| and S2, On the boundary surface S^
head is specified. On the boundary surface S2, flux is
specified. The specified head and specified flux are each
functions of location and time.
-97-
-------
r
X-Axis
Figure 1. Diagram representative of the flow domain.
The initial and boundary value problem associated
with this two-dimensional flow domain is described by the
following set of equations, written in repeating index
notation:
dp
ft
for i « x,y
h(x,,t) - hB(X|,t) on S,
and
on
(1)
(2)
(3)
Where h is hydraulic head, hB is hydraulic head on the
specified head boundary, Ky is the hydraulic conductivity
tensor, nj is the outward normal vector on the specified
flux boundary, Ss is specific storage, qB is the flux across
the specified flux boundary into the flow domain, N is a
source or sink, X| are coordinates x and y, t is time, 6 is
soil moisture content, p is soil moisture potential, and
d9/dp is the soil moisture capacity. Additionally, total head
and soil moisture potential are related by the expression:
p(x,y,t) = h(x,y,t) - y
(4)
Equation 1 describes the conservation of mass at a
point. To solve the partial differential equation, additional
specification of the problem is needed. In particular, the
solution depends on the solution domain, soil character-
istics, boundary conditions, and initial conditions. The
solution domain consists of the flow and time domains.
The flow domain is the region O to be modeled that is sur-
rounded by the boundary S. The time domain is the period
over which a solution is to be obtained. Soil characteristics
are the hydraulic properties of the soils, waste, and mem-
branes. Boundary conditions are a mathematical state-
ment of soil moisture conditions at the boundary of the
flow domain. Initial conditions are the specification of
pressures at the start of the time domain.
Equation 1 is nonlinear because the hydraulic con-
ductivity (K) and soil moisture content (9) depend on the
soil moisture potential (p). For a particular soil, the K-6-p
relationship defines a pair of characteristic curves. Figure
2 shows two such pairs of characteristic curves. If the ef-
fects of hysteresis are neglected, hydraulic conductivity is
a monotonic function of soil moisture potential. For all non-
negative soil moisture potentials, hydraulic conductivity is
10 V
0.
iov
0.0 0.1 0.2 0.3 0.4 0.5
6 (v/v)
10 ~11 10"" 10 ~7 10
K (cm/s)
10
Figure 2. Volumetric moisture content (a) and hydraulic conductivity (b) versus soil moisture potential for soils used in
the simulations. For sand, a = 0.3658 ft ~1 and n = 3.00. For clay, a = 0.0616 ft~1 and n = 1.59
(Equations 5 and 6).
-98-
-------
the saturated hydraulic conductivity..For negative poten-
tials (positive soil moisture tensions), however, the
hydraulic, conductivity decreases with decreasing soil
moisture potential. Likewise, the soil moisture content has
a similar functional relationship to soil moisture potential..
The soil moisture capacity, however, has a different func-
tional relationship. Soil moisture capacity is zero near.
saturation, it is zero for large soil moisture tensions, and it
has positive values between those points.
The van Genuchten equations (2) were selected to
describe parametrically the unsaturated hydraulic proper-
ties of soils. These equations are:
= 6r
[.1 + (a\p\nm
K(p) = KsHl/2[1-(1-H1/m)m]2
-- - (0, - 6r) a"mn(1
(5)
, (6)
(7)
(8)
and
H =
6 -
,es - er
(9)
where 8S is the saturated soil moisture content, 8r is the
residual soil moisture content, Ks is the saturated hy-
draulic cpnductivity, a and n are parameters that
characterize the soil, and H represents the dimensionless
soil moisture content.
The van Genuchten equations have a number of
desirable attributes. First, the equations describe
characteristic curves with plausible shapes throughout the
range from zero to high moisture tensions, which include a
zero value of soil moisture capacity at zero pressure. Se-
cond, the soil moisture content and hydraulic conductivity
relations form a consistent pair of relations, which means
that the relations as a pair are physically plausible. Third,
the relations are adapted to particular soils by fitting only
five parameter values (6S, 0r, Ks, a, and n). Two of the
parameters are empirical (a and n), but their values can
be derived for a particular soil from data on moisture con-
tent and soil moisture tension for the soil. Values of a and
n used in the simulations are reported in Figure 2.
Moisture flux across a membrane is modeled as be-
ing proportional to the difference in head on either sur-
face. The constant of proportionality is termed "leakance."
Flux is given by:
where qT is the flux per unit length, Cj is the leakance of
the membrane, and AhT is the head differential across the
membrane. The leakance is the sole hydraulic property of
membranes, because it is assumed that they do not store
moisture. The leakance represents the ability of the mem-
brane to allow soil moisture movement under a unit head
differential across the membrane. An impermeable mem-
brane has a leakance value of zero. A permeable mem-
brane has a non-zero leakance value. That non-zero
leakance can be the result of the normal installed proper-
ties of the membrane or the result of distributed
perforations.
Numerical Solution
Equation 1 is solved by the Galerkin finite element
method (3). The fundamental idea of the finite element
method is to replace the exact continuous solution of the
original partial differential equation by an approximate
piecewise continuous solution. The piecewise continuous
function is described by soil moisture potentials specified
at a finite number of .discrete points called nodes. Poten-
tials between these points are calculated by using inter-
polating functions defined over a finite number of subdo-
mains called elements. The interpolating functions will pro-
vide an exact representation as the element size ap-
proaches zero in the limit. For a finite number of elements
and nodes, the approximation will not exactly satisfy Equa-
tion 1, and a residual will result. The Galerkin method
forces this residual to zero, in an average sense, through
selection of coefficients for the interpolating functions.
An interpolating function of the form:
h(x,y,t)«h(x,y,t)
(11)
is used, where h is a series approximation to h.^ are
linearly independent interpolating functions defined over
the flow domain Q, H, are undetermined coefficients, and
n is th,e number of nodal points.
The Galerkin finite element method yields a system of
n ordinary differential equations that can be solved for the
n values of Hj(t). The matrix form of the system of equa-
tions is:
[A]{H> + [B]
dt
+ {F} = 0
(12)
where the typical elements of [A], [B], and {F} are:
'
qT
(10)
(13)
-99-
-------
r
and
(14)
(15)
The matrices [A] and [B] have the dimensions [nxn] and
the vector {F} has the dimensions [nx1].
The matrices [A] and [B] and the vector {F} are com-
monly referred to by names taken from structural
, engineering, where the finite element method was first ap-
plied. The matrix [A] is referred to as the stiffness matrix,
because in structural problems the matrix [A] expresses
the stiffness of a structure. From a similar background, the
matrix [B] is referred to as the mass matrix, because it ex-
presses the mass of a vibrating structure. Finally, the vec-
tor {F} is referred to as the load vector, because in struc-
tural problems it expresses the external forces on a
structure.
To facilitate the integrations in equations 13 -15, the
interpolating functions are defined piecewise in each ele-
ment and their union produces global interpolating func-
tions within the flow domain. The elemental interpolating
functions are linear and are defined over two-dimensional
triangular elements within zones occupied by soils or
waste and over one-dimensional linear elements where
membranes occur. Triangular elements, which are used to
represent soils and soil-like porous media, have three
nodes (Rgure 3), and yield [3x3] elemental matrices.
Linear elements, which are used to represent membranes,
have four nodes (Figure 4), and yield [4x4] matrices.
«
3
Node-
X-Axis
Figure 3. Triangle element representing
soils and soil-like porous media.
Element
X-Axis
Figure 4. Linear element representing a membrane.
Equations 13 -15 are performed on an element by
element basis, but the results of those integrations are
assembled into global matrices. Elemental stiffness and
mass matrices are generated, and the results are then
transferred to the global stiffness and mass matrices. The
global matrices are obtained by summing, for a given
node, the contribution to that node from each elemental
matrix. This procedure is described by Wang and Ander-
son (3) for each node in an element and for all elements in
the domain Q.
Equation 12 is discretized in time by an implicit finite
difference representation of the time derivative, which
leads to a system of algebraic equations for the n coeffi-
cients Hj(t) at time t. However, the coefficients of the
matrices [A] and [B] depend in part on hydraulic head,
and the system of algebraic equations is nonlinear
because of the dependence of hydraulic conductivity and
soil moisture on head. Fortunately, the nonlinearity in-
troduced by the soil characteristics is not too severe, and
a solution of Equation 12 can be obtained by a simple
iterative procedure. The iterative procedure is a two step
predictor-corrector scheme. In the first iteration, heads at
the one-half time step are predicted using soil properties
based on heads at the beginning of the time step. Using
heads at the one-half time step, soil properties are
reevaluated. Then, using these corrected soil properties,
heads at the end of the time step are computed.
The mathematical basis described above has been in-
corporated into the computer program UNSAT2D. The pro-
gram includes components for data input and for the
march of the numerical solution through time. The data in-
put component reads information on the problem
geometry, material hydraulic properties, initial conditions,
and boundary conditions. The time march component in-
cludes an outer time step loop and an inner predictor-
corrector loop. The time step loop is repeated for the
specified number of time steps in the simulation, and the
-100-
-------
predictor-corrector loop is repeated twice for each time
step.
APPLICATION
To investigate moisture movement at facilities equip-
ped with leachate collection and removal (LCR) and liner
systems, a series of computer simulations have been per-
formed.
The focus of analysis in this study is the performance
of the LCR and liner systems. Insofar as the function of
these systems is to contain hazardous constituents and
promptly detect any leaks that should occur, measures of
performance include:
• the minimum top liner leak rate that will cause LCR
system drains to flow;
• the time required for a given leak to cause drain flow;
• the pattern of drain flow a given leak evokes;
• the leachate collection efficiency expressed as the ratio
of LCR system drain flow to top liner flux; and
• the total volume of moisture lost from the facility to the
surrounding environment.
The simulated facility is a double-lined surface im-
poundment with various degrees of top liner failure and
different bottom liner designs. The physical system, il-
lustrated in Figure 5, consists of a 300-foot wide by 10-foot
deep surface impoundment with 4:1 sideslopes and 2 per-
cent lower slopes.
In the series of simulations discussed in this paper,
the LCR system consists of a one-foot thick layer of sand
(saturated hydraulic conductivity of 10'3 cm/s) and a
system of five drains resting on the bottom liner. Figure 2
describes the hydraulic properties of the sand. The
drainage layer is bounded above-by a flexible membrane
liner (FML), which constitutes the top liner. A two-foot thick
sludge layer blankets the top liner and is assigned a
saturated hydraulic conductivity of 10~4 cm/s.
Two types of bottom liners have been simulated. Com-
posite bottom liners consist of an upper FML component
underlain by a low-permeability compacted soil compo-
nent. Non-composite bottom liners lack the upper FML
component. The compacted soil layer in the model has a
clay texture with a saturated hydraulic conductivity of 10'7
cm/s. Hydraulic properties of the clay are described by
Figure 2. A static water table exists at a depth of 20 feet,
and the surrounding loam-textured soil has a saturated
hydraulic conductivity of 10'4 cm/s.
The sand drainage layer is assigned an initial soil
moisture tension of 0.1 bar (24 percent moisture v/v), and
the compacted soil layer is assigned an initial soil
moisture tension of 0.33 bar (27 percent moisture v/v). In-
itial moisture tension in native soil and sludge is
hydrostatic with respect to the water table and impounded
liquid surface, respectively. Before leakage into the LCR
system occurs, the facility is allowed to equilibrate for ap-
proximately three years. At the end of this period, leakage
across the top liner begins.
The model is bounded by the sludge layer at the top,
a ten-foot native soil fringe to the sides, and by three feet
of soil below. Boundary conditions are prescribed poten-
tial. The boundary condition on the top sludge surface is
calculated as the pressure potential with respect to the
surface of the impounded liquid. Boundary conditions at
the bottom and sides are calculated as the pressure
potential with respect to the water table. Because the
facility is symmetrical about a vertical centerline, only half
the problem is modeled.
Oj
"$ --s
d) *J
H—
T~10
I -15
1-20
-150 -120 -90 -60 -30 0 30 60 90 120 150
Horizontal Distance (feet)
Figure 5. Schematic of the surface impoundment used in the simulations.
-101-
-------
Non-composite Bottom Liners
In a set of three simulations, moisture passes into the
drainage layer across a uniformly leaky top liner at rates of
14 gallons/acre-day, 95 gallons/acre-day, and 795
gallons/acre-day. These non-composite (NC) bottom liner
simulations are identified as NC-14, NC-95, and NC-795,
respectively. In each case, the bottom liner is a three-foot
thick layer of compacted clay-textured soil. Figure 6 shows
the soil moisture potentials that surround each of the
drains as functions of time. Negative potentials exist in un-
saturated soils and potentials rise towards zero as soils
approach saturation. At nonnegative potentials, soils are
saturated and drain flow occurs. Summarized results from
these simulations are reported in Table 1.
The simulation data suggest that a top liner flux in ex-
cess of 95 gallons/acre-day is necessary to achieve
saturation and drainage in the LCR system. In simulation
NC-95, soil moisture potential at the drains increases over
time but stabilizes after six to eight years (Figure 6). Soil
moisture potential at the centerline drain (Drain 1) reaches
-9 cm at steady state. Drains 2 and 3 do not approach zero
potential as closely because their vertical position realtive
to Drain 1 causes the equilibrium soil moisture potential
around them to be more negative because of their position
with respect to the water table. A top liner flux of 95
gallons/acre-day causes soil potentials around drains to
approach closely to zero which suggests that this flux is
near the threshold leak rate required to cause drain flow in
a facility of this design.
In simulation NC-795, drains begin to flow 3.1 months
after the top liner leak begins. When steady state condi-
tions are realized after roughly five months, flux across the
-50
-100
-150
-200
-250
NC-14
E
o
-50
:p -100
c
-2 -150
o
-200
3-250
OT
1 -
'I -50
t/)
-100
-150
-200
NC-95
Drain 1
— — — Drain 2
Drain 3
NC-795
234
Time
567
(years)
8 9 10
Figure 6. Soil moisture potentials at LCR system drains for
non-composite bottom liner simulations.
TABLE 1. LCR/LINER SYSTEM PERFORMANCE SUMMARY
Simulation
NC-14 NC-95 NC-795 NC-795-6 NC-64-6 NC-76-6 C-777 C-60
14
95
795
795
64
76
777
60
Top Liner Rux
(gallons/acre-day)
Tbp Liner Leak Location Uniform Uniform Uniform Uniform Uniform Uniform Uniform Sidewall
Bottom Liner Thickness 3 3 3 6 6 6 33
(feet)
Time to Drain Flow
(months)
Steady State Flux into
Bottom Liner
(gallons/acre-day)
Steady State Leachate
Collection Efficiency
Cumulative Loss from Unit 213,000 1,320,000 2,530,000 2,070,000 875,000 943,000 300
after 40 years
(gallons/acre)
no flow no flow
14 95
3.1
174
0.78
2.8
144
0.82
no flow
64
71
68
0.11
1.0 16.5
0.002 0.002
>0.999 > 0.999
300
-102-
-------
bottom liner is 174 gallons/acre-day. Leachate collection
efficiency, defined as the ratio of steady state drain flux to
steady state top liner flux, is 0.78.
In simulation NC-14, soil moisture potential at the
drains declines over time as a result of net moisture move-
ment out of the drainage layer and through the underlying
liner (Figure 6). Net drying of the sand layer occurs
because it was initially assigned a greater soil water
potential than the underlying compacted soil layer. After a
period of roughly five years, moisture conditions at the
drains stabilize in response to the combined effects of
moisture influx from above, unsaturated hydraulic proper-
ties of the drainage laye.r, bottom liner, and native soil, and
distance to the water table beneath the facility.
Increasing the thickness of a compacted soil bottom
liner from three to six feet does not significantly improve
its ability to act as an impermeable barrier to moisture
movement. A fourth simulation was identical to NC-795 ex-
cept that:the bottom compacted soil liner was six feet thick
rather than three feet thick. Table 1 summarizes results
from this simulation, identified as NC-795-6. Increasing
liner thickness from three to six feet results in a slight
decrease in time to drain flow (3.1 to 2.8 months) and a
slight increase in leachate collection efficiency (0.78 to
0.82).
The minimum top liner leak rate necessary to cause
drain flow in an LCR system above a six-foot thick com-
pacted soil bottom liner was found to be between 64 and
76 gallons/acre-day. In simulation NC-76-6 (Table 1) a flux
of 76 gallons/acre-day across a uniformly leaky top liner
causes the centerline drain (Drain 1) to flow after 71
months. Soils at Drains 2 and 3 never reach saturation
and never flow. In simulation NC-64-6 (Table 1), a top liner
flux of 64 gallons/acre-day is insufficient to cause drain
flow.
Composite Bottom Liners
When a composite bottom liner is substituted for a
non-composite bottom liner, LCR system effectiveness is
greatly improved. Time to drain flow is reduced, minimum
top liner flux that will cause drains to flow is reduced, and
leachate collection efficiency increases. This improvement
is demonstrated by 'composite (C) bottom liner simulation
C-777, in which moisture enters the drainage layer across
a uniformly leaky top liner at 777 gallons/acre-day. The
bottom liner consists of an FML component above a three-
foot thick compacted soil component. The FML compo-
nent of the bottom liner is allowed to pass moisture at a
rate of approximately 0.02 gallons/acre-day, which is a
typical rate for vapor phase transport across an intact FML
(1). Tabular performance data for C-777 are reported in
Table 1, and soil moisture potentials at the drains are
shown in Figure 7.
In simulation C-777, a top liner flux of 777 gallons/
acre-day causes drain flow 1.0 month after leakage begins.
This delay is a consequence of the absorptive capacity of
I- 0
a>
-*-*
o
5 -25
E
vH» ~50
~ -75
Drain 1
Drain 2
Drain 3
i ii i i i
jO 1000 12345 6 7 8 9 10
Time (weeks)
Figure 7. Soil moisture potentials at LCR system drains for
simulation C-777.
the granular drainage material. Only when a sufficient
volume of moisture is added to the drainage layer will soils
around the drains become fully saturated and drain flow
begin. Since the FML component of the composite bottom
liner allows very little moisture to pass into the bottom
liner, virtually all the moisture that leaks into the drainage
layer during this period is used to satisfy this absorptive
capacity. Because the absorptive capacity of a soil is a
function of initial moisture content and moisture
characteristic curve (e.g., Figure 2a), the wetting period
may be reduced with an initially wetter drainage layer
and/or with a coarser drainage material such as coarse
sand or pea gravel, for example.
Since a net increase in the moisture content of a
drainage layer is possible only when top liner flux exceeds
flux into the bottom liner, the bottom liner leak rate defines
the minimum detectable top liner leak rate. An LCR/liner
system using a composite bottom liner in good condition
can be expected to have a bottom liner leak rate of less
than 1 gallon/acre-day. Constant wetting of the drainage
layer is possible for top liner fluxes greater than this and
drain flow will eventually occur. This top liner leak rate is
much lower than the approximately 100 gallons/acre-day
required to cause drain flow in LCR systems above three-
foot thick non-composite bottom liners.
Leachate collection efficiency is greatly improved
when composite bottom liners are used. A collection effi-
ciency approaching 1.0 is achieved by simulation C-777,
compared with an efficiency of 0.78 achieved by NC-795.
The preceding simulations use uniformly leaky top
liners that allow moisture to enter the drainage layer along
its entire length. In simulation C-60, the top liner is im-
permeable except for ten-foot wide sections centered on
either sideslope through which moisture enters the
drainage layer at a rate of 60 gallons/acre-day. Otherwise,
the facility design of C-60 is identical to that of C-777. ,
Table 1 summarizes data from this simulation. Soil
moisture potentials at the drains are presented in Figure 8.
Simulation C-60 illustrates the response of an LCR
system to a localized top liner leak. As evidenced by the
change in soil moisture potential at the drains (Figure 8),
moisture that enters the drainage layer on the facility
-103-
-------
SUMMARY
o
E
u
.2 -7
s
u
-25
-50
-75
_^Xx ^--
" ' _•'•-•*
---"" -"
,~~ •*
S X* r\rn\n 1
17""" Drain 2
. / 1 1 1
] 6 12 18 2
Time (months)
Figure 8. Soil moisture potentials at LCR system drains for
simulation C-60.
sldeslope travels downslope past Drains 2 and 3 by un-
saturated flow and collects in the vicinity of Drain 1 on the
facility centerline. Drain 1 eventually flows after 16.5
months. Drains 2 and 3 never flow.
Soil around Drain 3, at the toe of the sideslope,
begins to wet almost immediately (Figure 8) in response to
the sidewall leak approximately ten feet away. After
roughly six months, the influence of the sidewall leak
reaches Drain 2 and soil there begins to wet. In eight
months, soil around Drain 1 begins to wet and that drain
flows 16.5 months after the start of the sidewall leak.
Since drains operated at atmospheric pressure cannot
remove unsaturated soil moisture and since they are
assumed not to contribute moisture, they exert no in-
fluence when soil moisture conditions around them are
unsaturated. Only when Drain 1 flows does it affect sur-
rounding soil moisture potentials. Drain flow limits adja-
cent soil moisture potential to a near-zero value and, as a
result, the gradual wetting at the other drains is arrested.
Figure 8 shows this effect clearly for Drain 2. Eventually, a
steady state condition develops in which all moisture
passes Drains 2 and 3 in the unsaturated state and only
Drain 1 flows. Under these conditions, increasing the
number of drains in the LCR system will have no effect on
LCR system performance. At large enough leak rates,
however, soil moisture potential in the vicinity of the leak
can become positive (fully saturated) and the presence of
a drain in this zone of saturation will improve the LCR
system performance in two ways: the leak can be detected
sooner than if a drain is not located nearby, and informa-
tion about the location of the leak is provided.
The computer program UNSAT2D is a useful tool for
hydrologic analysis of land storage and disposal facility
designs. The model simulates both saturated and un-
saturated moisture movement in facilities that may contain
flexible membrane liners and leachate detection drains.
Application of this model to a range of landfill, surface im-
poundment, and waste pile designs and operating condi-
tions has provided considerable insight into their
hydrologic behavior.
ACKNOWLEDGMENT
This work was accomplished in part under EPA Con-
tract 68-01-7310 to provide technical assistance to the EPA
in developing regulations for certain land disposal
facilities. Technical monitors were Mr. Doug Ammon of the
Hazardous Waste Engineering Research Laboratory and
Messrs. Alessi Otte and Walter DeRieux of the Office of
Solid Waste.
REFERENCES
1. Haxo, Henry E. Jr., Jelmer A. Miedema, and Nancy A.
• Nelson, 1984. Permeability of polymeric membrane lin-
ing materials. Technical paper, Matercon, Inc.,
Oakland, California, 10 pp.
2. van Genuchten, Martinus Th., 1978. Calculating the
unsaturated hydraulic conductivity with a new closed-
form analytical model. Research Report 78-WR-08,
Water Resources Program, Department of Civil
Engineering, Princeton University, Princeton, New
Jersey, 63 pp.
3. Wang, Herbert F. and Mary P. Anderson, 1982.
Introduction to Groundwater Modeling. Freeman and
Co., San Francisco, California, 237 pp.
-------
REMEDIATION OF AN INDUSTRIAL DUMP SITE - A CASE HISTORY, PART II
David S. Kosson, Irene A. Legiec and Robert C. Ahlert
Department of Chemical and Biochemical Engineering
Rutgers, The State University of New Jersey
College of Engineering
P.O. Box 909
Piscataway, NJ 08855
ABSTRACT
The case history of the design and implementation of a remediation strategy for a
hazardous waste disposal site is described. Sludges resulting from treatment of diverse
chemical manufacturing effluents had been deposited in an unlined surface impoundment over
several decades. "Remediation of an Industrial Dump Site - A Case History," presented at
the 12th Annual Research Symposium, described laboratory and pilot-scale investigations of
a proposed remedial strategy. The proposed process consists of in-situ alkaline sludge
extraction coupled with on-site, sequential aerobic-anaerobic, soil-based microbial
destruction of recovered organic contaminants. First year pilot plant results, presented
previously, indicated rapid organic species extraction from sludges and greater than 95%
destruction of recovered extract TOC. This paper will focus on results from second year
pilot plant operation and analyses, which have been carried out on residuals present after
extraction and treatment processes.
INTRODUCTION
Site Description
Over a period of several decades,
industrial waste sludges were landfilled
in an unlined surface impoundment. During
the period of operation, the compositions
and rates of depositions of sludges varied
greatly. Primary and secondary sludges
were deposited. The primary sludges were
lime-neutralized inorganic matter,
including neutralization wastes, spent
catalysts, and solid residues from diverse
chemical manufacturing operations. The
secondary sludges were biomass from
aerobic treatment of aqueous effluent from
the same manufacturing activities.
The resulting 4.1 acre site contains
approximately 30,000 cubic yards (yd3) of
sludge. Two principal layers are found
within the fill" material. The first layer
of approximately 15,000 yd3 of
secondary sludges 'deposited over a ten-
year period, prior to 1967. In response
to complaints from local residents about
obnoxious odors, lime was applied to the
secondary sludge. Subsequently, the limed
sludge was covered with clean fill and
plastic sheeting. The second layer
consists of approximately 5,000 yd3 of
primary sludge that was transferred to the
site from another lagoon. This sludge was
deposited over the layers of fill and
plastic sheeting covering the secondary
sludges. Mounds of shale fill were placed
over the primary sludge to minimize odor
problems. This gave rise to approximately
10,000 yd3 of contaminated fill material.
At present, the physical state of the
sludges ranges from solid to gelatinous.
Leachate from the sludges can impact local
groundwater resources. Detailed
descriptions of the leachate and sludge
characteristics have been presented
previously (1).
-105-
-------
Pilot Plant Description
Lagoon clean-up is viewed as two
interrelated problems. The first problem
is the extractive removal of contaminants
from the lagoon without major excavation.
The second problem is treatment of the
extract stream containing the stripped
contaminants, which include organic and
inorganic species. A research program was
designed to evaluate several treatment
options for the lagoon. Elements of this
program included forced extraction of
representative sludge samples and
laboratory evaluation of aerobic
(secondary) mixed microbial treatment and
sequential aerobic/anaerobic soil-based
microbial treatment of naturally occurring
leachate and forced extracts (2,3,4).
The laboratory program led to the
design of a pilot plant incorporating the
most promising renovation strategy. The
pilot plant employed in-situ extraction of
sludge deposits with aqueous sodium
hydroxide solution (pH between 9 and 13).
(Sodium hydroxide (0.05N) was shown
(1,2,5) to be a very effective extractant
for this lagoon sludge; significantly
different sludges may require other
extractants). Recovered extract was
treated on-site using a soil-based, mixed
microbial treatment system. The pilot
system was designed to consist of several
sequential process steps. A simplified
process flow diagram is presented in
Figure 1; additional detail is available
(1,5). The first process step was
extraction of sludges present in a
representative section of the lagoon.
Sodium hydroxide solution was mixed (Tank
1) and injected into the sludges or
applied to the surface of the extraction
bed. Extract was recovered from the
extraction bed using two wells and
collected in a process tank (Tank 2).
The second process step was the
adjustment of pH, dilution if necessary,
and addition of nutrients to extract
stored in Tank 2. This occurred
continously in a baffled process tank
(Tank 3). The third process step was
treatment of the modified extract
(effluent from Tank 3) in an
aerobic/anaerobic bioreactor. Treatment
occurred in a lined soil bed in which an
aerobic microbial population was
maintained in the upper region and an
anaerobic microbial population was
maintained in the lower region. Extract
applied to the surface of the treatment
H,O
TO -*•
DISCHARGE
ITRIENTS
MIXING TANK
(Tank 3)
TREATMENT BED
Figure 1. Simplfied process flow diagram.
-106-
-------
bed percolated through the soil column
where organic contaminants were
biodegraded. Effluent from the treatment
bed was recovered through a screened well
and pumped into a storage tank (Tank 4)
prior to recycle or discharge.
Operation of the pilot plant commenced
in July 1985 and continued until freezing
conditions necessitated a winter shutdown.
During 1985, extraction bed effluent pH,
total organic carbon (TOC), and total
dissolved solids (IDS) were initially
approximately 10, 11,000 mg/1, and 28,000
mg/1, respectively. Recovered extract pH
initially declined as previously applied
lime was extracted, and then increased to
9.5, in response to additipn of sodium
hydroxide extractant. Extract TOC and IDS
both declined during 130 days of operation
to approximately 4,000 mg/1 and 8,000
mg/1, respectively. Influent TOC to the
treatment bed was controlled during
prolonged intervals to 1000 mg/1 and 2000
mg/1. Integrated TOC reduction for the
entire 1985 period of operation was
greater than 95% on a mass basis. Steady
state TOC reduction, after microbial
population development, was greater than
PILOT PLANT OPERATION DURING 1986
Extraction Bed Results
Completion of pilot plant repairs to
correct cold weather damage resulted in
full operation start up for the second
operating season in June 1986. During
1986, extraction bed effluent pH, TOC and
TDS were intially approximately 9.7, 6,000
mg/1, and 13,000 mg/1, respectively.
Extract pH remained generally between 9.4
and 9.8 during operation, while extract
TOC and TDS declined rapidly after 40 days
of operation. Extract TOC and TDS were
approximately 2,200 mg/1 and 7,500 mg/1,
respectively. The TOC response of
recovered extract is presented in Figure
2. The initially elevated extract
EXTRACTION BED
1986 EFFLUENT TOC
S -
'4 -
3 -
40
—i—
60
BO
10O
120
TIUS (days)
O WELL f + ' WELL
Figure 2.
-107-
-------
concentrations, relative to 1985 results,
are attributed to the extended extract-
sludge contact period during the winter.
The apparent response delay during 1986
was between 40 and 50 days. Average
hydraulic flux through the extraction bed
during 1986 was 8.6 1/nrday, as compared
to 3.7 l/ra2day for 1985.
Several physical changes were
apparent in the extraction bed during
operation. Extended periods of extraction
during 1985 and 1986 resulted in
settlement of the sludges within the
extraction bed of between 0.5 and 1.0 ft.
In addition, as extraction progressed, the
surficial sludges appeared to become more
friable. This observation was further
supported by increasing sludge
permeability during operation.
Treatment Bed Results
Influent and effluent volumetric
fluxes for the treatment bed varied
between 3.6 and 9.4 1/nrday, depending on
mode of operation* as compared to between
15.1 and 17.9 l/mzday during 1985.
Throughout operation, influent pH was
targeted to be maintained between 7.5 and
8.0 although occasional power supply
interruptions resulted in brief periods of
influent pH between 8.0 and 8.5. Effluent
pH remained between 6.3 and 6.8 throughout
operation.
Influent and effluent TOC data for
the treatment bed are presented in Figure
3. Influent TOC was maintained at
approximately 1,200 mg/1 for 30 days
following a brief start up with a
glucose/nutrient medium. Note, no
microbial inoculum was used for startup in
1986. After day 40 of operation, the
influent TOC to the treatment bed was
adjusted to between 2,000 and 2,500 mg/1.
In contrast to the microbial acclimation
period observed during 1985, no
substantial microbial acclimation period
was apparent from the effluent TOC
responses. Increased effluent TOC after
day 70 was a delayed response to
increased influent TOC starting day 40.
Integrated TOC reduction over
the entire period of operation was greater
than 99%, on a mass basis. Typical TOC
treatment capacity ranged from 7 to 22
g/nrday, which was in close agreement with
laboratory soil column data (6).
TREATMENT BED
1988 INFLUENT AND EFFLUENT TOC
4000
3000 -
2000 -
1000 -
0-
*•
100 -
a a
o a a
a Da DQ
a D
ZO 40 BO 80
TIUB (days)
0 INFLUENT + EFFLUENT
Figure 3.
100
120
-108-
-------
EXTRACTION BED RESIDUALS ANALYSIS
Extraction of Split-Spoon Core Samples
At the conclusion of operation of the
extraction bed, split-spoon sampling was
carried out to retrieve sludge samples for
laboratory analysis. Four cores were
obtained at the conclusion of operation in
1985 and two cores were obtained at the
conclusion of operation in 1986. Core
locations within the extraction bed are
indicated in Figure 4. The samples
retrieved were segregated by core number
and sampling depth. Each split spoon
sample retrieved represented a two foot
deep core section. A description of
sludge composition as a function of depth
is presented in Table 1.
Sixteen of the core samples obtained
in 1986 were extracted in the laboratory
with 0.05N sodium hydroxide solution to
evaluate the uniformity and effectiveness
of the extraction bed operation. Each
core sample tested was homogenized
TABLE 1. EXTRACTION BED CORE SAMPLE
DESCRIPTIONS
Depth
(ft)
Sludge Description
0-2 Primary
2-4 Primary
4-6 Primary
6-8 Primary
8-10 Primary with some secondary
10-12 Even mixture of primary and
secondary
12-14 Secondary
14-16 Secondary with some clay
16-18 Secondary with some clay
18-20 Even mixture of secondary
and clay
20-22 Clay with some secondary
EXTRACTION BED
3('8S) !
2('86)
TREATMENT
BED
• 1986 CORE SAMPLES
« 1985 CORE SAMPLES
• INJECTION POINTS (6)
Figure 4. Extraction bed core locations.
-109-
-------
Individually and two 10-g aliquots were
shaken with 100 ml of the sodium hydroxide
extractant for 40 hours on a rotary
shaker. The resulting mixture was
separated into solid and liquid phases
through centrifugation and filtration.
Recovered extract was analyzed for TOC;
results are presented in Figure 5. The
extract TOC data indicates a pocket of
inefficiently extracted sludges between
depths of 6 and 12 ft, while sludges
present outside of that pocket were
extracted fairly thoroughly. Review of
previously obtained information indicated
that plastic sheeting was present at
depths of "approximately 4 and 12 ft, while
injection points were screened at depths
between 11 and 13 ft. Although plastic
sheeting was pierced using a rod at
regular intervals, pockets of irregular
hydraulic flow persisted. This indicates
the necessity of careful placement of
injection points during scale-up.
In addition to extractable TOC
analysis, volatile solids content (weight
loss upon ignition at 550 C from sludge
previously dried at 105 C for 24 hr) was
determined for each homogenized sludge
sample (7). Results of analysis carried
out in 1984 on untreated primary and
secondary sludge samples was compared to
results obtained from Core 1 (1986) at
depths of 4-6 ft and 12-14 ft. These
particular samples were considered
representative of primary and secondary .
sludges before and after effective in-situ
extraction. The comparison is presented
in Table 2.
Neutralization and teachability of Final
Residuals
Laboratory experiments were carried
out to determine the quantity of acid
required to neutralize the solid residuals
present after in-situ extraction was
completed and assess the Teachability of
neutralized residuals. Thus, a possible
process end-state was investigated. The
quantity of acid required was determined
by titration of 10-g aliquots of the
homogenized core samples, shaken in 100 ml
of distilled water and titrated with 1.0 N
sulfuric acid. Samples were shaken on a
NAOH EXTRACTION OF CORE SAMPLES
1.2
19BB EQUILIBRIUM TOC
0-S'
4-B'
DEPTH (feet)
Q CORE 1 + CORE ,
Figure 5.
-110-
-------
rotary shaker until a constant pH was
obtained after each addition of titrant
(approximately 40 hrs). Samples were
considered neutralized after an
equilibrium pH of 5.0 was achieved. Solid
and liquid phases of the resultant
neutralized mixture were separated by
centrifugation and filtration. The
recovered aqueous phase subsequently was
assayed for TOC. The number of
equivalents of acid required to neutralize
each titration mixture and the resultant
equilibrium TOCs are reported in Figure 6.
The number of equivilents of acid required
to neutralize sludge samples correlated
closely with the amount of Teachable
residual TOC. In addition, TOC
Teachability from core samples was
approximately ten times less for
neutralized samples than for samples
similarly extracted with 0.05 N sodium
hydroxide solution (Figure 5).
TABLE 2. SLUDGE CHARACTERISTICS
BEFORE & AFTER IN-SITU
EXTRACTION
Before After Change
(1984) (1986)
fmg/q) _ fmg/g)^f%)_
TOC Extracted with 0.05 N NaOH:
Primary 3.2 1.1 65
Secondary 8.7 1.4 84
Volatile Solids:
Primary
Secondary
200
320
45
22
78
93
NEUTRALIZED CORE SAMPLE RESULTS
I
0.60-
0.40 -
0.30 -
0.20 -
0.10 -
i
o.oo -
«_ OPEN SYMBOLS A FILLED SYMBOLS _,„
(# of Equivalents (TOC)
of Acid) *
A
•»
A
* : * :
o A i
I i i i i i i i i i i r •'
-110
-too
- 90
- 80
-"8
- 60 ^
- SO ^
-40
- 30
- 20
- 10
- n
0-2
4-e
e-a
8-10 1O-12 12-14 16-18 20-22
DEPTH (Jaot)
CORK 1 A CORE 2
Figure 6.
-Ill-
-------
CONCLUSIONS
Operation of a large-scale pilot
plant on-site at an industrial sludge
impoundment successfully demonstrated the
feasibility of in-situ alkaline extraction
followed by on-site soil-based microbial
treatment as a renovation process, for
this case. Extraction uniformity was
influenced heavily by extractant injection
locations and local hydraulics. Solid
residuals present after in-situ sludge
extraction can be neutralized through
addition of limited quantities of mineral
acid. Neutralized residuals leach greatly
reduced amounts of TOC. In addition,
soil-based sequential aerobic/anaerobic
microbial treatment removed greater than
99% of the TOC, on a mass basis, present
in the recovered extract.
ACKNOWLEGEMENT AND DISCLAIMER
The work described in this paper was
funded in part by U.S.EPA under
Cooperative Agreement CR807805. Dr. John
Brugger was the project officer. The use
of trade names does not constitute or
imply endorsement of any kind.
4. Kosson, D.S., R.C. Ahlert, J.D. Boyer,
E.A. Dienemann, and J.F. Magee II,
1985. "Development and Application of
On-site Technologies for Sludge Filled
Lagoons", Proceedings International
Conference on New Frontiers for
Hazardous Waste Management, EPA-600/9-
85/025, 118-127.
5. Kosson, D.S., E.A. Dienemann, and R.C.
Ahlert, 1986a. "Field Studies of In-
Situ Extraction and Soil-Based
Microbial Treatment of an Industrial
Sludge Lagoon", Proceedings of
Hazardous Hastes and Hazardous
Materials (HMCRI), March 4-6, Atlanta,
6A.
6. Kosson, D.S., I.A. Legiec, E.A.
Dienemann, and R.C. Ahlert, 1986b.
"Operation and Control of a Soil-Based
Microbial Treatment System", AICHE
Summer National Meeting, August,
Boston, MA.
7. Standard Methods for the Examination of
Mater and Uastewater, 1985, 16th Ed.,
APHA, AWWA, WPCF, Washington, D.C.,
96-98.
REFERENCES
1. Ahlert, R.C., and D.S. Kosson, 1986.
"Remediation of an Industrial Dump Site
- A Case History", Proceedings of 12th
Annual Research Symposium, EPA-600/9-
86/022, 73-89.
2. Boyer, J.D., M.B. King, D.S. Kosson,
and R.C. Ahlert, 1985. "Aerobic
Biodegradation of Leachate and Forced
Extract from a Sludge Disposal Lagoon",
Toxic and Hazardous Wastes, Technomic
Publishing Co., Lancaster, Pa., 497-
508.
3. Dienemann, E.A., J.F. Magee II, D.S.
Kosson, and R.C. Ahlert, in press.
"Rapid Renovation of a Sludge Lagoon",
Environmental Progress.
-112-
-------
CAPILLARITY AND ANISOTROPY EFFECTS ON GROUND-WATER FLOW
TO EXCAVATION
Forest 0. Mixon
Research Triangle Institute
Research Triangle Park, NC 27709
ABSTRACT
In a hazardous waste disposal facility located in a saturated soil, the local water
table and the capillary fringe zone are modified by the presence of the facility. In
this paper, a conformal mapping solution to the accompanying ground-water flow is
discussed. The flow net around an excavation is calculated and displayed in terms of
the capillarity, local geometry and flow properties.
Capillarity can significantly influence net flow into an excavation; typical values
can cause net flow to be 10 to 20 percent higher than predicted without capillarity.
Anisotropic behavior can also be important. It is shown that anisotropy favoring
horizontal flow can greatly alter the flow net and the total flow.
INTRODUCTION
There is increasing interest in the
possibilities of siting hazardous waste
disposal facilities in locations at which
the water table (or the tension-saturated
zone immediately above) is near the
ground surface (1, 2). This situation is
normal in saturated low-permeability
soils which exist at various locations of
the country but which predominate in the
eastern United States because of higher
rainfalls and local geologic conditions.
A disposal facility so located would
be beneath the water table, and thus
would influence the local flow dynamics,
including the water table and the
capillary fringe (3). The prediction of
these effects is the problem of concern
in this paper. Also considered are the
effects of anisotropy on the local ground
water flow.
The problem to be addressed is
presented schematically in Figure 1,
which shows a disposal facility,
idealized as a two-dimensional
rectangular excavation, sited such that
its lower surface is beneath the normal
water table. As in a well, there is a
tendency for ground water to flow into
the excavation. A complicating factor,
however, is capillary attraction, which
draws fluid into the soil interstices and
creates a zone of negative pressure, the
capillary fringe or tension-saturated
zone.
•Figure 1 shows the excavation, the
flow profile into the excavation
including lines of constant potential and
stream function, the free surface
(immediately beneath which P = -Hc), and
the water table (along which P = 0). (A
-113-
-------
-202
Horizontal Distance — halfwidths
Figure 1. Schematic of disposal facility slightly below
normal water table.
List of Symbols is provided at the end of
the paper).
In the capillary fringe, the region
between the water table and free surface,
the physical picture is that of individual
granules whose interstices contain no
significant vapor bubbles (4, 5). This
zone 1s normally referred to as the
tension-saturated zone or the negative
pressure zone.
Above the top of the capillary fringe,
the physical appearance is that of
Individual granules whose interstices are
filled primarily with vapor or gas bubbles
(4, 5). This region is referred to as the
unsaturated zone and its moisture content
normally decreases with elevation or
Increases with depth.
APPROACH
Our purpose is to assess the effects
of the capillary fringe on the fluid flow
patterns 1n the neighborhood of the
excavation. This is done by formulating
and solving the equations of motion for
the flow field with the equilibrium
capillary rise, Hc, as an adjustable
parameter. Comparative analysis of the
solutions for various values of the
equilibrium rise should reveal and
quantify its effects.
Classical potential theory (6), which
applies here, and in which the flow is
proportional to the gradient of some
potential, provides a relatively easily
visualizable and familiar representation
of the flow field in the form of the flow
net. This net is comprised of lines of
constant potential, say 0, and orthogonal,
or perpendicular, lines of constant stream
function, say ^. The lines of constant
stream function, or stream lines,
represent particle tracks in steady flow;
hence no fluid flows across these lines.
Any two such lines form a stream tube,
through which the total flow is constant.
A particularly simple flow net is
shown in the rectangle in the center of
Figure 2. In this flow net, the
horizontal lines represent lines of
constant potential, 0, and the vertical
lines represent stream lines or lines of
constant stream function, $. Thus the
flow field represented by this rectangle
is uniform flow from high potential to low
potential, thus uniformly upward.
Imagine now that the lines of constant
potential and flow lines in this rectangle
-114-
-------
Ground Surface
Figure 2. Schematic of overview mapping.
(Figure 2) are drawn on an infinitely
stretchable thin sheet of rubber. Imagine
moreover that this rectangular piece of
rubber is positioned and deformed as
indicated in the figure so that the
outside edges correspond as indicated to
the centerline of the excavation, the
bottom of the pit, and the free surface.
If this stretching is done in such a way
that angles are preserved, e.g.,
conformally, then the resulting
curvilinear network comprising the (now
deformed) lines of constant potential and
constant stream function, represents a new
flow field, the one sought in this study.
Conformal mapping is a mathematical
formalism for accomplishing this
stretching. In a typical application, a
simple flow configuration (the rectangle)
is mapped onto a more complicated geometry
(the disposal facility) in such a way that
the important mathematical features of the
flow net are preserved and the resulting
curvilinear flow net is a theoretically
accurate representation of the flow field
in the more complex geometry. The
mathematical formalisms are developed
elsewhere (7) and solutions are presented
from which the behavior at various values
of the capillarity can be compared.
SCALING . , . .
In this analysis, all distances have
been scaled to the excavation half-width,
including the equilibrium capillary rise,
Hc. Moreover, all velocities have been
scaled to the free-drain velocity, which
is superficial velocity resulting from a
head gradient of unity (1 ft/ft). Thus, a
value of A = 1 would correspond to a flow
velocity given by
operating over the area of the excavation
bottom.
For present purposes, it is adequate
to consider the quantity A as simply a
measure of relative flow into the
excavation.
ANALYSIS OF CAPILLARITY
Computed results illustrating the
effects of capillarity are shown in
Figures 1 and 3. Figure 1, in addition to
serving as a schematic of the problem, is
also a flow map included to establish
axes, scales, and formats for Figure 3.
The latter figure represents a compilation
-115-
-------
0.4
OS
0.2
0.4
Flow Into Excavation
Fraction of Free-Drain Flow
OS
Figure 3. The effects of total flow into the excavation and the
equilibrium capillary rise on the resulting net rise.
of cases, plotted to the same scale as in
Figure 1, but omitting, for convenience,
the location of the water table, and
reduced so as to fit on the same sheet for
comparative purposes.
Referring to Figure 3, moving to the
right increases A, the relative flow into
the excavation. The second and third
columns of figures represent progressively
lower positions of the excavation base
with respect to the water table, hence,
higher flows into the facility.
Moving downward in Figure 3
corresponds to increasing the equilibrium
capillary rise, Hc. An interesting
feature is the extent to which the
elevation of the free surface at the
excavation wall is dependent upon A, the
relative flow. The point of intersection
is bounded from below by the equilibrium
capillary rise and increases with flow, A.
The effects of capillarity can be
further quantified as shown in Figures 4
and 5. To generate these figures, the
hydraulic gradient in the vicinity of the
excavation was first estimated for each
case in Figure 3. This was done by first
estimating the gradient as the slope of
the line segment connecting the corner of
the excavation to the location of the
water table 4.5 half-widths away.
Crossplotting the resulting information as
flow vs. gradient with capillarity as a
parameter gives Figure 4. Expressing flow
as a percent increase over the no
capillarity case gives Figure 5.
The analysis shows that capillarity
does, indeed, increase the net flow into
the excavation by small, but significant
amounts, with the increase being roughly
proportional to the equilibrium capillary
ri se.
Figures 4 and 5 can be utilized to
estimate the flow into the excavation.
First, one estimates the local gradient of
the water table and then, with Figure 4,
determines the no-capillarity flow. Next,
one estimates the capillarity from grain
size or by some other method. Figure 5 is
then used to predict the net increase in
the flow.
-116-
-------
0.04
0.08 0.12 0.16
Approximate Hydraulic Gradient
Figure 4. Effect of hydraulic gradient on relative flow.
0.2 0.4 0.6
Equilibrium Capillary Rise — halfwidths
Figure 5. Effect of capillarity on relative flow.
Physically, one can reason that flow
increases with capillarity because the
capillary rise provides more lateral area
for flow to cross for a fixed hydraulic
gradient.
ANALYSIS OF ANISOTROPY
Soils can be quite anisotropic, with
horizontal permeabilities exceeding values
by several orders of magnitude (3).
Omitting mathematical details,
anisotropic flow profiles cases have been
computed with values of Kx/Ky from 1 to
64. Graphical results from these
-117-
-------
calculations are summarized in Figures I
and 6.
Figure 1, as before, establishes axes,
scales and formats for Figure 6. The
latter figure represents a compilation of
cases, plotted to the same scale as that
used in Figure 1, reduced so as to fit on
the same sheet for comparative purposes.
Referring to Figure 6, moving downward
Increases the capillarity, Hc. Moving to
the right corresponds to increasing the
anisotropy, specifically the x-direction
permeability relative to that in the y-
direction. The results are as one would
expect, though the magnitude is
surprisingly large. Further analysis
follows in Figure 7 which was generated
from the computed results for the special
case of intermediate capillarity (Hc =
0.4) by the following procedure.
Approximate hydraulic gradients were
estimated as the slopes of the line
segments connecting the corner of the
excavation to the water table 4.5 half-
widths way. Relative flow from (into) the
excavation, the A values, were then
plotted against these hydraulic gradients
with the permeability ratio as a
parameter.
Intuitively, it is expected that
increasing the horizontal permeability
while holding the vertical permeability
constant would have the effect of
increasing the flow from (into) the
excavation for a given hydraulic gradient.
The figure shows that this is indeed the
case, quite dramatically. Consider a
hydraulic gradient of 0.1. Increasing the
permeability ratio fourfold, from 1 to 4,
results in an approximate threefold
increase in flow into the excavation, from
A « 0.5 to A « 1.5. Another fourfold
increase in the permeability ratio (to 16)
results in an approximate fourfold
increase in flow, to A «6. Still another
fourfold increase in the permeability
ratio (to 64) gives quite a dramatic
increase in the flow, from about 6 to much
greater than 32.
The following generalized rule-of-
thumb seems to apply: In the two-
dimensional approximation, the ground-
water flow from (into) an excavation
0.05
0.1
0.2
1.0
4 16
Horizontal to Vertical Conductivity Ratio
64
Figure 6. Effect of anisot-ropy on flow map.
-118-
-------
32
30
28
26
1-24
I 22
| 20
u5 18
I «
1 14
C 12
l»
IT 8
6
4
2
0
0.1 0.2 0.3
Approximate Hydraulic Gradient
Figure 7. Effect of anisotropy on flow into excavation.
approximately doubles for every doubling
of Kx with constant Ky. The effect is
stronger at large values of the
permeability ratio, e.g., for Kx/Ky > 16.
PRACTICAL IMPLICATIONS
Fine-textured soils, e.g., clays have
capillary radii in the range of 10~4 to
10~5 centimeters. Corresponding values
for the free-drain velocity, v0, are in.
the range of 5 x 10~6 to 5 x 10~8 cm/sec
(1 to 0.001 gal/day-ft2). Moreover,
equilibrium capillary rise values for such
soils can exceed 50 ft.
For illustrative purposes, assume a
soil with free-drain velocity of 1 x 10~6
cm/sec (0.2 gal/day-ft2). Assume moreover
that one is considering the ground-water
flow into a long trench-like disposal
facility whose width is 100 ft which is
located in a soil exhibiting an
equilibrium capillary rise of 25 ft. Then
the equilibrium capillary rise is 0.5
excavation half-widths. If, moreover, the
local hydraulic gradient in the vicinity
of the excavation is 0.05, then, from
Figure 4, the relative flow into the
excavation in the absence of capillarity
is about 0.14. Hence the estimated total
flow is 0.14 x 0.2 gal/day-ft2, or 0.028
gal/day-ft2.
What effect would one anticipate from
including capillarity? From Figure 5, at
an equilibrium capillarity rise of 0.5
half-widths, one would expect the flow
into the excavation to be almost 20
percent higher than estimated for no
capillarity.
CONCLUSIONS
A method has been demonstrated for the
formulation and solution of ground-water
flow problems involving a free surface,
e.g., a conceptual interface between
liquid-filled interstices and vapor-filled
interstices. This method has been applied
to predict the effects of capillarity and
anisotropy on flow patterns in the
vicinity of a disposal facility located
beneath the local water table.
The solution to the problem for zero
capillarity provides a baseline estimate
of the flow into the excavation in terms
of the local hydraulic gradient.
The effect of capillarity is to draw
fluid vertically into the soil interstices
and thence to open the area through which
-119-
-------
fluid can flow into the excavation.
Capillarity thus causes the flow into an
excavation to increase over the no
capillarity case; the magnitude of the
increase can be on the order of 10 to 20
percent of the total flow and is higher
for higher hydraulic gradients.
Soils frequently exhibit anisotropic
flow behavior in that the horizontal
hydraulic conductivity is greater than the
vertical. This phenomenon can increase
the net flow substantially at a given
hydraulic gradient as a result of the more
open horizontal flow path for liquid
motion.
LIST OF SYMBOLS
P
x, y
a
v
P
Value of stream function
at lower right corner of
excavation, hence a
measure of flow volume to
(from) excavation, L.
Gravitational acceleration
L/T-2
Equilibrium height of
capillary rise above water
table.
Heterogeneity half-length,
L.
Permeability, permeability
components, L2
Manometer pressure, L.
Horizontal and vertical
position coordinates, L.
(Ky /K
Kinematic viscosity, L /T.
Bulk mass density of
fluid, M/L3 .
Piezometric potential, P +
y, L.
Stream function, L.
ACKNOWLEDGMENTS
The research described in this article
has been funded wholly or in part by the
United States Environmental Protection
Agency through Contract No. 68-03-3149,
Task No. 26-1, to the Research Triangle
Institute. It has been subject to and
approved for publication. Approval does
not signify that the content necessarily
reflect the views and policies of the
Agency, nor does mention of tradenames or
commercial products constitute endorsement
or recommendation for use. Technical
guidance from the Project Officer,
Jonathan G. Herrmann, is greatly
acknowledged.
REFERENCES
1. Proceedings, U.S. EPA Workshop.
October 4-5, 1985. Monitoring
considerations in the siting and
operation of hazardous waste disposal
facilities in temperate zone wet
environments, Tallahassee, FL.
2. Smith, E. D. 1974. Hydrogeologic
assessment of zone-of-saturation
landfill design. CONF-8405143,
Environmental Sciences Division, Oak
Ridge National Laboratory, Oak Ridge,
TN.
3. Martin, J. P. and R. M. Koerner. 1984.
The influence of vadose zone
conditions in groundwater pollution.
Part II: Fluid movement, J. Hazardous
Materials, 9, 181-207.
4. Freeze, R. A. and J. A. Cherry. 1979.
Groundwater, Prentice-Hall, Englewood
Cliffs, NJ.
5. Sowers, G. B. and G. F. Sowers. 1961.
Introductory soil mechanics and
foundations, 3rd Edn., MacMillan, New
York. .
6. Lamb, H. 1945. Hydrodynamics, Dover
Publications, New York.
7. Mixon, F. 0. 1987. Capillarity and
anisotropy effects on ground-water
flow to excavation, prepared for
Ground-Water.
-120-
-------
PATHWAYS FOR THE REMOVAL OF VOLATILE ORGANICS FROM SURFACE IMPOUNDMENTS
Crowley Clark Allen and Jeffrey Bryan Coburn
Research Triangle Institute
Research Triangle Park, North Carolina
ABSTRACT
A series of surface impoundments have been investigated to determine the significance
of biological oxidation as a major pathway for volatile organic removal. Measurements of
the volatile organic concentrations, pH, and dissolved oxygen have been taken from active
surface impoundments. Wastewater was removed from selected impoundments, and the rate of
oxygen uptake and the compound specific fate of the volatiles was evaluated in the labora-
tory under both anaerobic and aerobic conditions. A biocide was used to evaluate the
significance of biological activity relative to chemical reactions. The results'indicated
that biological activity is common in surface impoundments. The biological removal of
specific components at one impoundment was low relative to the anticipated volatilization
rates.
INTRODUCTION
The United States Environmental
Protection Agency is developing regula-
tions to control the emissions of volatile
organics (VO) from hazardous waste treat-
ment, storage, and disposal facilities.
Facultative lagoon systems present a
potential source of emissions of this type
as they provide an opportunity for mass
transfer of VO from the aqueous phase to
the atmosphere. Properly designed and
operated biological wastewater treatment
systems protect water quality by removing
organic chemicals from wastewater prior to
discharge; the actual fate (biodegraded,
transferred to the air, or removed in a
separate sludge stream) of this material
is uncertain. Visits to three non-aerated
wastewater management facilities were made
to determine the composition of wastewater
in the lagoon and the potential for bio-
degradation and air emissions. (1, 2, 3)
SITE DESCRIPTIONS
Description of Site A
Site A produces aldehydes, glycols,
glycol ethers, nitriles, esters, and other
products. Wastewater and runoff are col-
lected at different points within the
manufacturing area of the plant. The
wastewater flows to a series of seven
oxidation basins. Some of the effluent
from the basins is pumped to a series of
four large unlined facultative basins
prior.to discharge. Wastewater samples
were collected from the first facultative
basin (60 acres in area). The discharge
permit application for Site A includes the
following information about the long-term
average, final effluent: BOD, 21 mg/L;
COD, 384 mg/L; TOC, 127 mg/L; TSS, 43
mg/L; methylene chloride, 18 /»g/L;
acenaphthylene, 10 /tg/L; bis (2-ethyl
hexyl) phthalate, 24 ug/L; naphthalene, 4
(1)
-121-
-------
Description of Site B
DO Profile Results
Site B has two adjacent facilities (a
refinery and a lube oil plant), and each
facility has its own wastewater treatment
system. Each wastewater treatment system
contains a RCRA-regulated polishing pond
near the end of the system. Wastewater
samples were collected from each of these
polishing ponds for chemical analysis and
biological activity testing. In June
1986, upstream from the polishing pond,
the wastewater contained 82.7 mg/L COD,
33.7 mg/L TSS, and 0.9 mg/L phenols. (2)
Description of Site C
Site C 1s a commercial waste facility
which treats dilute aqueous waste. The
following four lagoons were sampled:
Lagoon B, an evaporation pond; Lagoon C,
an evaporation pond which receives the
overflow from B; Lagoon D, a .holding pond;
and Lagoon E, a solids settling pond.
Samples were taken from the southeast
corner and the west end of Lagoon B; only
one sampling location was employed in
Lagoons C, D, and E. The wastewater at
Site C contains a variety of dilute «20
ppm) organic compounds. (3)
TEST PROCEDURES AND RESULTS
Samples for chemical analysis were
collected in one liter amber glass bottles
with Teflon-Hned screw caps and in 40 ml
zero headspace septum bottles. The sam-
ples were transported on ice and refriger-
ated until analyzed. Aside from refriger-
ation, the samples were not preserved.
At the chemical analysis sampling
points, additional samples were obtained
for biodegradation rate studies. Waste-
water was pumped into five gallon Nalgene
containers and shipped to RTI without
refrigeration or preservation.
Dissolved oxygen (DO) measurements
were made at different depths at the sam-
pling points using a YSI Model 54A dis-
solved oxygen meter calibrated against
saturated air. The pH of the wastewater
at each sampling point was determined
using short range pH paper.
The results of the DO profile measure-
ments are presented in Figure 1. At Site
A the DO was greater than saturated except
for the bottom of the lagoon. A high
concentration of algae in the upper level
of the lagoon in combination with the
bright sunlight produced super-saturated
conditions throughout much of the water.
This also is the likely cause for the high
pH (approximately 9) as carbon dioxide was
removed from the water photosynthetically.
The DO concentration in the refinery
polishing pond at Site B was about 8 mg/L
independent of liquid depth; the pH was
approximately 7. The DO concentration in
the lube oil polishing pond at Site B was
about 3 mg/L independent of liquid depth;
the pH was approximately 7.
All lagoons at Site C had DO concen-
trations of less than 1.0 mg/L indicating
that the. lagoons were primarily anaerobic.
A1J of the wastewater collected except
from Lagoon E, had a pH of 8-8.5; Lagoon E
wastewater had a pH of approximately 13.
Analytical Results
The wastewater samples at Sites A and
B were analyzed for volatile organics
using EPA Method 624 for purgeables and
EPA Method 625 for base/neutral and acid
extractables. None of the Method 624 or
625 compounds were present above detection
limits in the lagoons.
Samples from Site C were analyzed for
specific volatile organics using EPA
Method 624 for purgeables and EPA Method
625 for base/neutral and acid extract-
ables. Acetone was present at concentra-
tions of 1.6, 0.05, 2.8, and 16 ppm for
Lagoons B, C, D, and E, respectively.
Lagoons D and E also contained over 10 ppm
methylene chloride, and approximately 1
ppm 1,1,1-trichloroethane, Freon 113,
toluene, and total xylenes. No other VO
were detected in Lagoons B and C.
Microorganisms
The presence of microorganisms in the
wastewater at Site A was Initially con-
firmed by microscopy studies. Several
different microorganisms were observed
using wet drop slides. These include:
-122-
-------
coccoid blue-green algal (Phylum
Cyanophyta, anacystis sp.); protozoans
(both flagelates and cilliates, numerous
species); and green algae (particularly
arthrospira).
Site B also contained microorganisms.
Several different microorganisms were
observed using wet drop slides. These
include: protozoans (both flagelates and
cilliates, numerous species); and green
algae (particularly arthrospira).
The wastewater from Site C contained
no motile microorganisms that were ob-
served using wet drop slides. Wastewater
samples taken from two locations in Lagoon
B appeared to have agglomerations of
coccoid blue-green algae. The abundance
of inorganic solids, however, especially
in the Lagoon D sample, hindered the wet
drop slide studies. Both filamentous and
nonfilamentous bacteria were observed
using Gram-stained slides of Lagoon B, Q,
and D samples. Both gram-positive bac-
teria (stained purple) and gram-negative
bacteria (stained red) were observed. No
cell cultures were grown to further char-
acterize the bacteria.
Aerobic Biological Testing
The first experiment to measure the
oxygen consumption rate of the microorgan-
isms measured the dissolved oxygen (DO)
depletion rate. The procedure employed
was as follows. A wide mouth, amber
glass, pint bottle was filled with the
wastewater sample and allowed to come to
thermal equilibrium. Air was then bubbled
through the sample for approximately 5
minutes to raise the initial DO concentra-
tion. A magnetic stir bar was added to '
the sample bottle. The lid, fitted with a
DO probe, was secured allowing the waste-
water to overflow in order to insure zero
headspace within the bottle. The sample
was stirred using a magnetic stirrer and
the DO concentration was recorded with
time. ,
In the second method for measuring the
oxygen uptake rate (the BOD-type experi-
ment), 250 mL of sample was added to a
pint amber glass respirometry bottle. The
respirometry bottle lid has a tube fitting
to allow the bottle to be connected to a
mercury manometer. A T-connector was
inserted in the manometer tubing; lithium
hydroxide was poured in the side tube to'
absorb produced carbon dioxide and the
side tube was sealed. The bottle was then
clamped in a wrist-action shaker and suf-
ficiently agitated to ensure that oxygen
transfer was n6t rate limiting. The pres-
sure drop resulting from aerobic (oxygen
consuming) biological activity was mea-
sured with the mercury manometer as a
function of time.
- ^
The initial oxygen consumption rate in
the wastewater from Site A was 2.4 * 0.2
mg/L-hr (95%) and 137 mg/L BOD was con-
sumed after 90 hours. The data are sum-
marized in Table 1.
The data from the oxygen depletion
from Site B are presented graphically in
Figure 2. A least squares linear regres-
sion was performed on the data and the
calculated slopes provide the observed
zero order oxygen utilization rates. Note
that the data for the refinery wastewater
sample do not fit a linear curve very well
(R2 = 0.9813).
Presuming that the reaction is first
order in substrate concentration (i.e.,
BOD), the reaction rate equation becomes':
rate
d(DO)
~~dt
k(BOD).
(D
Since the DO uptake (DOUn t) Is simply the
difference in the original BOD (BOD0) and
the BOD at time t, Equation 1 can be
rewritten as:
- DOuprt/BOD0) = -kt.
(2)
The BOD0 value which best fit Equation 2
(i.e., y-intercept of zero) was 8.89 mg/L;
the value of k at this BOD0 was 0.0185 *•
0.011 hrs-1 (R2 = 0.9949). Equation 2
yields a calculated BODs value of 7.9 mg/L
for the refinery wastewater sample. This
is commensurate with the expected effluent
BODs values of 2.4 to 7.7 mg/L since some
reduction in BOD will occur between the
sampling point and the effluent.
The DO uptake rate for the lube oil
wastewater sample remained fairly constant
even though the absolute DO concentration
fell below one. For the range evaluated,
both the zero-order and the first order in
BOD rate models were acceptable models in
-123-
-------
describing the rate of oxygen consumption.
For the BOD rate model, the best BOD0
value was 18.53 mg/L, and the calculated
first order rate constant was k = 0.0111 ±
0.0005 hours'1 (R2 = 0.9969). The BOD5 of
the lube oil wastewater sample calculated
using the above equation is 13.6 mg/L.
Again, this 1s commensurate with the
expected effluent BODs values of 8 to 12
mg/L since some reduction 1n BOD will
occur between the sampling point and the
effluent. Since the BOD rate model is the
best model for both sets of data as seen
by the correlation coefficients, the low
concentrations of substrates are thought
to be responsible for the slow oxygen
uptake rates observed.
Some of the results of the oxygen
uptake experiments for the Lagoon B sam-
ples at Site C are shown in Figure 3. A
linear regression analysis was performed
on the data and the calculated slopes
provide the observed oxygen utilization
rates. The DO uptake of the poisoned
Lagoon B samples were minimal; when a BOD-
type experimental sample was poisoned,
oxygen consumption ceased. The oxygen
uptake rates observed in the DO depletion
experiments were slower than those ob-
served in the longer term BOD-type experi-
ments. Also, there was a noticeable lag
phase in the BOD-type experimental oxygen
utilization curves. These results are
predictable 1n light of the in situ DO
measurements. The quick oxygen uptake
rate observed experimentally and the low
DO concentrations measured iji situ suggest
that the bacteria present in the pond
could utilize oxygen faster than it was
provided. Consequently, the low DO con-
centration in the collected sample re-
quired much of the bacteria to acclimate
to the oxygen-rich conditions of the
experiments causing the lag phase and slow
Initial oxygen utilization rates. Appar-
ently, the potential for aerobic biologi-
cal activity was significant in Lagoon B,
but its in situ importance was limited by
the slow aeration rate.
The long-term average oxygen uptake
rates of the Lagoon B samples are very
similar. The total amount of oxygen con-
sumed by the different samples was also
nearly Identical. The total observed BOD
of the Lagoon B wastewater was 2,050 mg/L;
however, oxygen was still being consumed
when the experiment was terminated.
The results of the oxygen uptake
experiments for the Lagoon C wastewater
sample were similar to those for Lagoon B.
Specifically, these results were: the
poisoned sample showed minimal oxygen
utilization; the DO depletion experiment
revealed a slower initial oxygen utiliza-
tion rate than was observed with the
longer BOD-type study; and there was a
noticeable lag phase in the BOD-type
experimental oxygen utilization curve.
The experimental results of Lagoon D
wastewater suggested that chemical oxida-
tion, not biooxidation, was the most sig-
nificant oxygen utilization mechanism.
The oxygen depletion rate of the poisoned
and non-poisoned samples were identical.
The high pH of Lagoon E wastewater is
prohibitive to biological growth. As this
pond is upstream of Lagoon D, chemical
oxidation is again believed to be respon-
sible for its low iji situ DO concentra-
tion.
For each bulk wastewater sample at
Site C, 700 mL of wastewater was placed in
a 1-L amber glass bottles with teflon-
lined screw caps. Identical samples with
700 mg of mercuric acetate was also
examined. The 1-L bottles were continu-
ously agitated and vented twice daily by
blowing air into the headspace for 30
seconds. Relative component concentra-
tions were determined by GC using head-
space injections. Analytical samples were
collected on Day 1, Day 3, and Day 8 of
the study. (3)
The wind was flowing toward the east
at approximately 200 ft/min (8 mph) during
the sample collection at Site C. The
sample taken from the west end of Lagoon B
was near the impoundment inlet and was
somewhat lower in concentration than the
southeast corner sample initially.
Three major VO peaks were observed by
GC headspace analyses. There was no sig-
nificant trend of the first peak (34 sec.)
area changes during the eight days of the
test for the sample taken from the west
end of the lagoon, but there were average
increases of 20 percent per day for the
first peak area from the sample taken from
the southeast corner. These increases
were only observed in one sample and could
be due to experimental error.
-124-
-------
For the second peak (55 sec.) there
was no effect of biocide on the rate of
change of the peak area. The sample taken
from the southeast corner of the lagoon
did not demonstrate a significant reduc-
tion in the 55 sec. peak area. The 55
sec. peak of the sample from the west end
of the impoundment, however, declined 6
percent per day for both the sample of
wastewater with biocide and without the
biocide.
For the sample taken from the west end
of the lagoon, the third peak (60 sec.)
increased in area at roughly the same rate
as the rate of decrease of the second peak
(in the same sample). For the sample of
wastewater taken from the southeast
corner, there was no significant change in
concentration, with biocide or without
biocide, for either the second or third
peaks.
Thus, no clear trends were observed
for the aerobic destruction of any of the
three major VO components in the headspace
of the Lagoon B wastewater samples. The
average percent change of total headspace
VO (total peak area) over the eight day
period with biocide present was zero, with
a standard deviation of approximately 10
percent of the concentration. The average
change for the total headspace VO without
biocide was 18.3 percent, with a standard
deviation of approximately 10 percent.
These results suggest the possibility of a
small change in VO due to biological
activity over an eight day period that is
somewhat greater than the experimental
error. Lagoon B would be expected to have
lower jji situ aerobic rates than measured
in the laboratory, since it was primarily
anaerobic at the time of the sample col-
lection.
For a typical mass transfer rate for
volatilization of 2 • 10~5 g mol/cm^-sec,
the time constant for a two meter deep
impoundment is approximately 6 days (63%
volatilization loss). This predicted
volatilization rate is significantly
greater than the aerobic VO removal rates
measured in the Lagoon B samples (the
samples with the highest biological oxygen
consumption rates). Lagoon B at Site C
did not have an aerobic zone more than a
few centimeters, additionally suggesting
that aerobic biodegradation does not com-
pete well with volatilization for VO re-
moval in at typical surface impoundments.
Anaerobic Biological Activity Testing
When, the iji situ DO measurements indi-
cated the presence of an anaerobic zone
(DO < 0.5 mg/L), a wastewater sample was
withdrawn from the bottom of the lagoon
and tested for anaerobic biological
activity as follows. Nitrogen was first
bubbled through the five gallon sample to
purge any oxygen from the system that may
have been introduced at the time of sample
collection. The lid, modified to accommo-
date a small tube, was secured, and the
tube was run to a water filled inverted
graduated cylinder. Gas produced by the
system was thus collected by water dis-
placement, and the volume of gas produced
was measured.
Figure 4 shows the cumulative volume
of gas produced by the anaerobic system
from Site A. During the first month, only
35 mL of gas was collected. After this
lag phase, the system exhibited an expo-
nential growth phase as gas was produced
quickly. The growth phase was followed by
a saturated or steady state phase in which
gas production (assumed to be related to
metabolic activity) proceeded at a con-
stant rate. In single population systems,
the saturated phase is typically followed
by a decline phase caused by substrate
limitation. However, in this system,
another exponential growth phase followed
the saturated phase. This is believed to
be caused by the old population shifting
to a new substrate or a new population
exhibiting growth on the decaying matter
of the old population.
Since no anaerobic zones were found at
Site B with the jji situ DO measurements,
no anaerobic biological activity tests
were performed.
Although the wastewater impoundments
at Site C are oxygen deficient, there was
little anaerobic gas generation in the
four wastewater samples tested from
Lagoons B, C, and D. Some carbon dioxide
absorption was possible due to the high
pH. The rates are reported in Table 1.
Samples were withdrawn from the Site C
anaerobic test samples periodically for
chemical analysis. These samples were
collected in duplicate in 40 mL VOA vials
and killed at the time of collection by
the addition of 1 mL of a 40 mg/mL mer-
-125-
-------
curie acetate solution. Analytical sam-
ples were collected at t=0, Day 8, Day 14,
and Day 55 of the study; these samples
were refrigerated until analyzed. One set
of samples was analyzed by GC using head-
space injections (3). The duplicate set
of samples was analyzed by GC using purge
and trap (3). The results of the head-
space analyses on lagoon samples in the
anaerobic study show a dramatic increase
in total area counts for each lagoon sam-
ple after 55 days. Sample D contained
more compounds and in greater quantities
than the other samples.
The limited anaerobic biological acti-
vity of Site C wastewater samples does not
demonstrate removal of volatile organics
from the lagoon; however, in the waste-
water sample from the west end of Lagoon
B, a VO peak which corresponded to toluene
and one other VO peak showed major reduc-
tions with time. There was a large
Increase in a VO peak which corresponded
to methanol with time.
CONCLUSIONS
1. A sampling and analysis method is
presented to evaluate the rate of
biological activity in surface
Impoundments. The measured rates of
biological and chemical VO removal can
be compared to predicted air removal
rates to determine the relative
importance of competing VO removal
pathways.
2. Surface impoundment wastewater samples
from al1 three wastewater treatment
facilities demonstrated biological
activity. Microorganisms were ob-
servable and oxygen utilization was
measurable in each lagoon wastewater.
3. The rate of oxygen consumption varied
greatly from site to site, ranging
from 0.1 to 35 mg/L-hr. An unusually
high rate of oxygen consumption, 38 to
48 mg/L-hr, in one lagoon was attrib-
uted to chemical processes rather than
biological processes.
4. Compound specific headspace of aerated
samples from Site C indicated a half-
life of approximately 8 days for one
of the compounds. Other compounds
either did not change in concentration
or exhibited an increase in concentra-
tion. Thus, the measured biological
oxidation rates are less than the
expected volatilization rates of VO
from surface impoundments.
5. The rate of anaerobic gas generation
was slow in the lagoons which were
anaerobic in the field. Several weeks
were required before the anaerobic gas
generation began. The compound
specific analysis of VO indicated
either no change in VO concentration
or an increase in VO concentration due
to anaerobic processes. Two VO peaks
showed major reductions in one waste-
water sample. These results suggest
that anaerobic processes are not a
significant pathway for VO removal in
impoundments.
REFERENCES
1. Site A Visit Report, EPA No. 68-03-
3253, Research Triangle Institute,
August 26, 1986.
2. Site B Visit Report, EPA No. 68-03-
3253, Research Triangle Institute,
August 27, 1986.
3. Site C Visit Report, EPA No. 68-03-
3253, Research Triangle Institute,
September 23, 1986.
-126-
-------
TABLE 1. , MEASURED BIOLOGICAL RATES
Aerobic oxygen uptake rates (mg/L-hr)a
Sample
Site A
Site B
refinery
lube oil
Site C
Lagoon B, west
Lagoon B, SE
Lagoon C
Lagoon D
DO depletion
2.4 *
0.079 *
0.171 *
7.19 ±
..- 12.1 *
2.85 ±
38C *
0.2
0.009
0.014
0.37
0.45
0.07
5
BOD-type
1.57 * 0.10
NA
NA , .
34.9 ± 1.0
33.8 * 0.6
5.75 ± 0.34
47. 8C ± 1.1
Anaerobic
gas generation
(mL/L-hr)
0.022b
NA '
NA
9 • ID'5
2.3 • 10-3
2.6 • 10-3
1.2 • 10-3
aOxygen uptake rates were determined by using a least squares linear regres-
sion on the data; the 95% confidence interval is also reported.
"Initial rate was lower.
cNot biological uptake, see text.
8
§
FIGURE 1. DISSOLVED OXYGEN PROFILES
COMPARISON OF IN SITU 00 MEASUREMENTS
10 14 18 22 26 30 34
DEPTH FROM SURFACE (Inches)
-127-
-------
ui
I
I
FIGURE 2.
8.0
7.0 -
6.0 -
5.0-
4.0 -
3.0 -
2.0 -
1.0-
0.0 Hi1
REGRESSION OUTPUT:
y-Int — 0.204 mg/L
•lop* — 0.171 mg/L— hr
R2 - O.S882
0 20
D REFINERY SAMPLE
DO UPTAKE CURVE
SITES
REGRESSION OUTPUT;
y-fcrt. — 0.433 mg/L
slope — 0.079 mg/L— hr
R2 - 0.9813
40
TIME (hrs)
60 80
100
<• LUBE OIL SAMPLE
FIGURE 3. DO UPTAKE CURVE
STIE d LAGOON B (SE CORNER)
REGRESSION OUTPUT:
y—hit •• —0.082 mg/L
a!op« — 12.1 mg/L-hr
R2 - 0.9946
REGRESSION OUTPUT:
y—Int. •• 0.036 mg/L
•lop* —
R2 - 0.9955
RUN 1
TIME (mln)
+ RUN 2
A KILLED
-128-
-------
I
FIGURE 4. ANAEROBIC GAS PRODUCTION
SITE A: BOTTOM SAMPLE
80 100 120 140
160
TIME (doya) ,
-129-
-------
COMPOSITION OF LEACHATES FROM ACTUAL
HAZARDOUS WASTE SITES*
Glenn D. McNabb, James R. Payne, Paul C. Harkins
Science Applications International Corporation
San Diego, CA 92038
William D. Ellis, Jennifer A. Bramlett
Science Applications International Corporation
McLean, VA 22102
ABSTRACT
This presentation addresses the analytical methodology used in a follow-on effort of a
U.S. Environmental Protection Agency sponsored project. The project was initially
undertaken to gather data on the composition of hazardous waste leachates and to support
the development of multi-component synthetic leachate. These synthetic leachates will be
used to evaluate the effectiveness of various liner materials used in landfills and other
hazardous waste storage, treatment, and disposal facilities. As such, the formulated
leachates should be representative of typical compositions of actual hazardous waste site
leachates. During the initial study, the routine organic analyses of thirteen leachates
accounted for only approximately four percent of the overall Total Organic Carbon (TOG).
As a result, a more rigorous and complex analytical method was developed and is presently
being employed in the follow-on study to obtain a more comprehensive characterization.
During the first phase of this study, a hazardous waste leachate sample was characterized
by the new analytical procedure with the intent of maximizing the percent of TOC
accountable by specific compounds or by functional groups. Overall, approximately 48% of
the TOC was accounted for by the new method. This included approximately 20% attributed
to individual components and 28% accounted for by functional groups. In order to obtain
more information on the actual composition of hazardous waste leachate, a second phase of
this study is employing the same new analytical method to characterize two additional
leachate samples. Based on the results of these more thorough characterizations,
recommendations will be made regarding the composition of representative synthetic
leachates for linear compatibility testing.
INTRODUCTION
A data base containing information on
hazardous wastes and their associated
leachate compositions is being developed by
the Hazardous Waste Engineering Research
Laboratory (HWERL) of the U.S. EPA Land
Pollution Control Division. The data base
will be used to develop a multi-component,
synthetic hazardous waste leachate. The
formulated leachate will be used to test
the integrity of containment liners under
consideration for use in landfills and
other hazardous waste storage, treatment,
and disposal facilities.
*The work reported herein was performed by Science Applications International Corporation
under U.S. Environmental Protection Agency Contract No. 68-01-7043, Work Assignment No.
P-28. The content of this publication does not necessarily reflect the views or policies
of the U.S. Environmental Protection Agency nor does mention of trade names, commercial
products, or organizations imply endorsement by the U.S. Government.
-130-
-------
In a previous study (1), thirteen
hazardous waste disposal site leachates
were analyzed for priority pollutant
metals, organic compounds, total cyanide,
total organic carbon, chemical oxygen
demand, and a variety of other general
parameters. The authors completely charac-
terized the leachates for metal concentra-
tions; however, the overall sum of identi-
fied organic constituents averaged to only
about four percent of the sample TOG.
Because liner compatibility tests have
shown that heavy metals in saturated
solutions are generally compatible with
flexible membrane liners, the focus of this
follow- on study has been to more
thoroughly characterize the unidentified
portion of the TOG.
Therefore, in order to maximize the
accountable organic carbon, the first phase
of this study was devoted to the develop-
ment of a more rigorous analytical proce-
dure, which would identify organic contri-
bution by functional group classes or by
individual compounds when possible. The
method was evaluated by using it to charac-
terize a leachate sample, and the results
presented here indicate that the procedure
is capable of a more complete organic
characterization of hazardous waste site
leachates. Additional insight into the
actual organic composition of waste site
leachates is the objective of the second
phase of study, which will also allow for
minor modifications to the method through
the analysis of two different leachate
samples. This paper describes the analy-
tical methodology, results of the first
phase analysis, recommendations for
modifications to the procedure, and
suggestions for areas of further study.
METHODS AND MATERIALS
The analytical scheme that was deve-
loped is outlined in Figure 1. The
procedure was designed to not only charac-
terize organic priority pollutants, but
also to identify polar compounds not usual-
ly analyzed in standard, procedures. In
addition, the method allowed for the char-
acterization of higher molecular weight
compounds through the identification of
functional groups. Total organic carbon
(TOG) measurements were made at various
critical steps in order to track the orga-
nic carbon mass which thereby allowed for
an assignment of the TOG by functional
group.
, As seen in the figure, Initial charac-
terization included total organic halides
(TOX) , total nitrogen, sulfate, sulfide,
pH, conductivity, and methylene blue active
substances (MBAS) as well as TOG. In addi-
tion to the more routine analyses, MBAS
analysis allowed for an estimation of the
percent TOG due to anionic surfactants or
detergents which may not be chromatograph-
able.
The next step (B) identified and
quantified individual volatile compounds
through gas chromatography/mass spectro-
scopy (GC/MS) purge and trap techniques.
The purged sample was then subjected to a
molecular weight fractionation using gel
filtration chromatography (GFC). The >500
molecular, weight -(MW) fraction was analyzed
for TOG, TOX, and MBAS for mass balance of
those parameters and by UV^visible, and
infrared spectroscopy and C-NMR for func-
tional group characterizations. After TOG
and MBAS analysis, the <500 MW was extrac-
ted (neutral and basic) followed by silica
gel chromatography of the extract into
aliphatic, aromatic, and polar fractions,
with subsequent GC/MS analysis. The
aqueous remainder was re-analyzed for TOG,
then extracted after acidification. After
GC/MS analysis the extract was subjected to
derivitization with diazomethane followed
by reanalysis in o.rder to verify compound
identities. The aqueous remainder after
acidic extraction was analyzed by high
pressure liquid chromatography (HPLC) and
aqueous compatible capillary column GC/MS
and, finally, TOG.
A brief description of each step shown
in Figure 1 follows. The figure also lists
the QA/QC measures taken at each of the
steps including replicates, field blanks,
spike and recovery, and method blanks.
Step A Raw leachate was obtained from a
hazardous waste site in New Jersey and sub-
mitted to SAIC with a field blank of
distilled water. The collection of the
-131-
-------
LEACHATE(500MLS)
TOC, TOX, N (TOTAL),
S, pH, EC, MBAS
El
SiOa
CHROMATOGRAPHY
F11 F21 F3 I
ORGANIC
GC/MS
GC/MS
GC/MS
PURGE FOR GC/MS
VOA
GFC
> 500 MW
< 500 wiw
I
BASE/NEUTRAL
EXTRACTION
AQUEOUS
TOC
TOC, TOX, I Ci
UV-VIS, MBAS I
LYOPHILIZE
C2
ORGANIC
ACID EXTRACTION
AQUEOUS
TOC, HPLC,
AQUEOUS GC/MS
GC/MS I G1
DERIVATIZE
W|TH CH2 N2
GC/MS
G2
QA ANALYSES:
REPLICATES AT STEPS A, B, C1, D
FIELD/METHOD BLANKS AT A, B, C1, D, El, E2, F, G2, H
SPIKED BLANKS AT STEPS E, E1, E2, G, G1
Figure 1. Diagram of phase II analytical steps.
-132-
-------
field blank consisted of opening the appro-
priate sample container (filled with
distilled- water) for the duration of the
leachate sampling. Aliquots of the leach-
ate and field blank were taken for TOG,
TOX, total nitrogen, sulfate, MBAS, sul-
fide, pH and conductivity (EC).
TOG measurements were made using an
O.I. Corporation Model 700 TOC analyzer,.
and followed the US EPA 600/4-79-020 Method
160.4(4). TOX was measured on a Xertex/
Dohrmann DX-20 analyzer, following the US
EPA Method 9020. All the other analyses
were done following procedures outlined in
Standard Methods for the Examination of
Water and Wastewater (5) and the particular
methods for each analyte are as follows:
nitrogen by method 420A; sulfate by method
426D; MBAS by method 512B; sulfide by
method 427; EC by method 205; and pH by
method 423, using an Altex 4500 pH meter
with an Orion 8104 probe.
Step B A 5 ml aliquot of the raw leachate
was analyzed by GC/MS for volatile organic
compound characterization according to EPA
Test Method 624.
The remaining leachate was purged with
Ultra pure nitrogen for 24 hours at a rate
of approximately 100 mis N./minute. An
aliquot was analyzed by GC/MS to assure
total removal of volatile organic
compounds. A 500 ml aliquot of the purged
leachate was used in Step C.
Step C The molecular weight fractionation
step presented here was completed using gel
filtration chromatography techniques. In
order to calibrate the GFC separations, one
column (20 mm x 300 mm with a 250 ml reser-
voir) was static-packed to the 300 mm mark
with Sephadex®G-10-120, which has an
average pore exclusion size of 700 daltons.
The water exclusion volume was calculated
(3) to be 31 ml and, after a 10 ml applica-
tion of the leachate, this quantity was
collected and discarded. Next, 10 ml frac-
tions were eluted and collected for three
bed volumes. Three additional bed volumes
were taken in 15 ml fractions, for a total
of fifteen fractions. These fractions were
extracted and analyzed by GC/MS for deter-
mination of molecular weight.
The results obtained were plotted as
molecular weight versus volume eluted in
mis. The relation was linear, and indica-
ted that all compounds of molecular weight
greater than 500 were eluted in the first
34 ml (very close to the calculated vol-
ume) . All compounds were eluted in 142 ml.
Steps Cl, C2, & D Aliquots of the MW >500
fraction were analyzed for TOC and TOX, in
duplicate, with accompanying field blanks.
Aliquots of the <500 MW fraction were
analyzed for TOC only. The analytical
scheme calls for MBAS analysis of Steps Cl
and D; however, because no MBAS were detec-
ted during Step A, further MBAS analyses
were "eliminated.
The >500 MW fraction was also scanned
on a Hitachi 100-80 UV/VIS spectrophoto-
meter (with recorder) from 800 nm to 190
nm. The scan was conducted with the field
blank in the reference cell using matching
quartz cuvettes.
Lyophilization of this fraction (to
remove water) was accomplished by placing
approximately 75 mis of sample at a time in
a 500 ml round bottom flask fitted with a
ground glass adapter. Vacuum tubing con-
nected the flask and adapter to a trap
which was placed in a dewar containing
liquid nitrogen. A vacuum was applied to
the system using a Welch Duo-Seal®vacuum
pump. Ice was removed from the trap
periodically and discarded. The complete
contents of the fraction were lyophilized
in this fashion and the resulting solid was
quantitatively recovered.
Upon completion of the lyophilization,
a small quantity of the solid was pressed
into a potassium bromide (KBr) pellet, and
subjected to infrared spectrophotometry
using a Perkin-Elmer Fourier Transform
Infrared Spectrophotometer connected to a
Perkin-Elmer 7300 Series computer. Scans
were from 4000 cm to 600 cm and refer-
enced against..air. The solid was' also
subjected to C NMR, using a GE 300 series
model. The sample was dissolved in D-0,
spiked with dioxane (as an internal stan-
dard) and run at a 45 degree pulse to allow
for relaxation.
-133-
-------
Steps E. El. E2 and G The <500 MW fraction
was spiked with Base/Neutral/Acid surrogate
standards to a final concentration of 200
ng//Jl. The fraction was then extracted
With methylene chloride at pH 7, pH 12 and
pH 1. A small aliquot was taken for TOC
analysis before acid extraction (Step F).
The base and neutral extracts of the spiked
leachate were combined. The combined
extract was volume reduced in a
concentrator tube to -1 ml then exchanged
Into hexane. The reduced extract was then
subjected to silica gel chromatography,
using a column of 10 mm i.d. x 240 mm with
A 200 ml reservoir. Three fractions were
collected (aliphatic, aromatic, and polar)
corresponding to elution by 1) 50 mis of
pentane, 2) 200 mis of 1:1 pentane; methy-
lene chloride and 3) 50 mis of methylene
chloride. Each fraction's volume was
reduced and then analyzed by GC/MS.
Steps Gl and G2 The acid extract was also
volume reduced and screened by GC/MS for
compound identification and quantification.
After initial screening and compound
Identification, the extract was derivatized
with dlazomethane.
This derivitization converts carboxylic
acids to esters and alcohols to ethers.
Diazomethane for the reaction was produced
according to the method of H.M. Fales, T.M.
Jaouni and J.F. Babashak (2). This method
is of considerably lower hazard than most
conventional methods, while producing CH2N2
In good yield. Once derivatized the sample
extract was reanalyzed by GC/MS for
compound-specific identification confirma-
tion.
Step H After acid extraction the aqueous
remainder was aliquoted for TOC, HPLC and
aqueous GC/MS techniques. Aqueous GC/MS
analysis was conducted using a J&W Scienti-
fic Carbowax Fused Silica capillary column.
HPLC was done using a Waters Associates
Model 244 LC with a R401 Refractive Index
Detector. This particular model does not
run gradients in the mobile phase; there-
fore, runs were done with 100% methanol;
75% methanol and 25% H-0; 50% methanol and
50% H00; 25% methanol and 75% H20; and 100%
H.O.- All runs utilized a C-18 reverse
pnase column.
RESULTS AND DISCUSSIONS
The analytical approach described above
was applied to a waste site leachate sample
with the initial (Step A) characteristics
shown In Table 1. As seen, this sample has
elevated TOG and conductivity and a
depressed pH, moderate to high levels of
nitrogen, sulfate and TOX, and undetected
levels of sulfide and MBAS.
GC/MS purge and trap analysis for vola-
tile organic constituents identified 18
compounds mostly present in the low part-
per-billion (ppb) range. The most abundant
volatile compounds were toluene at 143 ppb
and vinyl acetate at 114 ppb. The remai-
ning compounds included ketone solvents,
chlorinated solvents, aromatic and ali-
phatic hydrocarbons and alcohols in the low
Table 1. INITIAL LEACHATE CHARACTERIZATION
(STEP A)
Analyte
Units
Field Sample
Blank Average
Nitrogen (total)
Sulfate
Sulfide
MBAS
pH
Conductivity
TOC
TOX
mg/1
mg/1
mg/1
mg/1
umhos/cm
mg/1
mg Cl/1
<10
<3
<10
<0.1
6.9
10
<3
0.07
635
210
<10
<0.1
4.3
19,500
16,000
166
(<50) ppb range. The total level of vola-
tile organic carbon (374 /tg/1) represents
<0.01% of the initial sample TOC.
Figure 2 tracks the TOC distribution by
weight percent according to the analytical
steps employed. As seen, the molecular
weight fractionation step by gel filtration
chromatography resulted in the loss of 29%
of the TOC due to adsorption or reactions
with the Sephadex®gel. The remaining TOC
was almost evenly distributed between the
less than and greater than 51j>g molecular
weight fractions. Based on C-NMR and
proton NMR, the 39% TOC observed in the
greater than 500 MW fraction was determined
to be comprised primarily of aromatic
hydrocarbons. This data was in good agree-
ment with fourier transfer infra-red (FTIR)
-134-
-------
100% I INITIAL
VOLATILES
100% I NON-VOLATILES
29%
(ADSORBED TO
SEPHADEX®
J
32% I < 500 MW
B/N
EXTRACTABLES
AQUEOUS
REMAINDER
ACID
EXTRACTABLES
3.0%
30%
AQUEOUS
REMAINDER
Figure 2. TOC distribution flow scheme by percentage.
-135-
-------
spectroscopic analysis which showed only an
OH absorbance and aromatic overtone reso-
nances. In addition, this higher molecular
weight fraction contained nearly 20% of the
original TOX.
The results of base/neutral extraction
of the <500 MW fraction with subsequent
silica gel chromatography indicate that the
aliphatic fraction (Fl) consisted solely of
low level n-alkane concentrations. The
aromatic fraction (F2) had only three com-
pounds present, two of which were not
identifiable by GC/MS. The highest levels
of base/neutral extractable carbon were
present in the polar fraction (F3) with
phenol exhibiting the highest concen-
trations. The total contribution to the
TOG resulting from base/neutral extractable
compound level is approximately 1.0%.
Another 3.0% of the initial TOG was
accounted for during acid extraction. The
largest contribution is due to benzoic acid
and phenol, with pentanoic and hexanoic
acids also present in substantial concen-
trations. Subsequent diaxomethane derivi-
tization was used primarily for compound
identification confirmation.
The remaining aqueous phase after both
GFC and base/neutral and acid extraction
was found to retain 30 percent of the
original TOG. The composition of this
"remainder" fraction was determined by
aqueous GC/MS, and was found to be composed
primarily of benzoic acid with some phenol
as well. The benzoic acid concentration
was exceedingly high in this fraction, and
this individual compound alone accounts for
17% of the entire initial TOG value. Table
2 shows .the other major individual com-
pounds contributing to the TOG. As seen,
over 20% of the leachate TOG can be
attributed to benzoic acid and phenol.
TABLE 2. TOG CLASSIFICATION BY MAJOR
INDIVIDUAL COMPOUNDS
Compounds
% of Original TOG
Benzoic Acid
Phenol
Pentanoic Acid
Hexanoic Acid
Trimethyl-1, 3-pentanediol
17.
3.1
.07
.06
.08
Table 3 presents the leachate composi-
tion as a function of organic function-
ality, with representative compounds and
their contribution. These six general
organic functional classes are often used
to assess the compatibility of organic
compounds with various contaminant liners.
The largest contribution was from aromatic
hydrocarbons of molecular weights greater
than 500. The only other class of com-
pounds to contribute substantially to the
TOC was organic acids, with benzoic acid
responsible for the vast majority of the
TOC. All the other classes were respon-
sible for less than 1% of the overall TOG,
with organic bases not contributing at all.
CONCLUSIONS AND RECOMMENDATIONS
The results of the waste site leachate
characterization by the analytical
procedure developed and utilized during
this study provide the basis for the
following conclusions:
• The method accounted for approximately
48% of the organic carbon present.
• The major constituents are organic
acids (20%) and aromatic hydrocarbons
(26%). The organic acid group consists
primarily of benzoic acid. The aromatic
hydrocarbon contribution is attributed
to high (>500) molecular weight
aromatic compounds.
• The procedure allows for a more tho-
rough characterization of the sample by
providing steps to ensure the recovery
of highly water soluble, polar
compounds. Standard separatory funnel
solvent extraction methods were capable
of recovering only a small fraction
(13%) of the total mass of benzoic acid
present. Therefore the procedure calls
for aqueous GC/MS, and HPLC steps,
which were used successfully to
quantify the organic acid contribution.
• Reactions with and/or adsorption to the
Sephadex®gel during GFC resulted in
the loss of 29% of the sample TOC.
Based on the conclusions above, recom-
mendations for further studies would
include 1) the optimization of the molecu-
lar weight separation step and 2) a more
detailed investigation into the composition
of the >500 molecular weight fraction.
-136-
-------
TABLE 3. ORGANIC CONTENT BY CHEMICAL CLASSIFICATION
Chemical Classification
Percent (WT)
Representative Compounds
Organic Acids
Oxygenated Hydrocarbons
Halogenated Hydrocarbons
Organic Bases
Aromatic Hydrocarbons
Aliphatic Hydrocarbons
20,3
0.08
0.86
0.0
26.8
0.002
Benzoic Acid (17.1%)
Phenol ,(3.1%)
Alkanoic Acids (0.13%)
Substituted Benzoic Acids (0.01%)
Substituted Phenols (0.002%)
Ketone Solvents (0.0003%)
Alcohols (0.0002%)
Trimethylpentanediol (0.08%)
TOX (0.86%)
Chlorinated Solvents (0.001%)
None detected
Aromatic Compounds >500 MW (26.8%)
Benzene and Alkyl-substituted
benzenes (0.001%)
n-alkanes (0.002%)
Because the molecular weight separation
step was directly responsible for the loss
of 29% of the TOG, this procedure should be
modified to affect a clean separation while
minimizing the loss of organic carbon. In
addition to testing other available gel
chromatography materials, efforts should be
directed towards resources (such as frac-
tional vacuum distillation) that might
circumvent organic carbon losses. With
that additional organic carbon available
for characterization, accountable TOG may
exceed 80% (based on the loss experienced
using Sephadex®gel for molecular weight
separation).
The molecular weight separation step is
important because it can effectively
isolate a large portion of the organic
carbon that normally would not be accounted
for. Of the six chemical classifications,
the >500 molecular weight fraction of
aromatic hydrocarbons exhibited the highest
percentage of organic carbon (26.8%).
Because this fraction represents a
substantial portion of the overall organic
carbon and because synthetic leachate
formulations should include representative
constituents from this fraction, additional
studies should focus on the specific
chemical composition of the >500 MW
fraction.
Once a more efficient molecular weight
separation step has been devised we recom-
mend that a statistically sufficient number
of representative leachate samples be
subjected to the procedure. A comprehen-
sive data base, containing the fully
characterized organic portion of actual
waste-site leachates, will greatly facili-
tate the formulation of a synthetic
leachate. This approach, in combination
with the newly developed analytical method,
will ensure that containment liner compati-
bility tests are conducted with a realistic
and accurate representation of actual waste
site leachates.
-137-
-------
REFERENCES
1. Bramlett, J.A., Repa, E.W., and Mashni, C.I., 1986. Leachate Characterization and
Synthetic Leachate Formulation for Liner Testing. Proceedings to the Seventh National
Conference on the Management of Uncontrolled Hazardous Waste Sites.
2. Fales, H.M., Jaouni, T.M., Babashak, J.F., 1973. Simple device for preparing ethereal
diazomethane without resorting to codistillation. Analytical Chemistry, Vol. 45, No.
14.
3. Kremmer and Boros, 1979. Gel Chromatography. J. Wiley and Sons, New York, N.Y.
4. United States Environmental Protection Agency, 1986. Test Methods 415.1, 624, and
9020. Federal Register Vol. 51. No. 125.
5. Standard Methods for the Examination of Water and Wastewater, 1985. Test Methods
420A, 426D, 512B, 427, 205 and 423. 16th Edition.
-138-
-------
DECONTAMINATION TECHNIQUES FOR MOBILE P^PONSE EQUIPMENT USED AT WASTE SITES
Mary K. Stinson
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Releases Control Branch
Edison, NJ 08837
Gary Kepko
U.S. Environmental Protection Agency
Region VII
25 Funston Road
Kansas City, KS 66115
ABSTRACT
Any cleanup equipment used at waste sites must be decdntaminated after
use. This paper highlights a published EPA report on the state-of-the-art
review of decontamination techniques for cleanup equipment and discusses
field experience with decontaminating equipment presently in use.
For those who prepare decontamination plans for cleanup equipment at
hazardous sites, the EPA report provides background material on decontamin-
ation methods, contamination assessment, and contamination avoidance. The
EPA report particularly stresses the importance of contamination avoidance.
Such measures as use of enclosures for equipment, safety features on equip-
ment to prevent spills and leaks, and protective coatings on equipment sur-
faces reduce hazard, time, and cost of the final decontamination task.
Though chemical methods are being developed to degrade contaminants on
equipment surfaces, use of physical removal methods prevails in the field.
This will be shown in discussing decontamination procedures of equipment
presently in use, such as the EPA Mobile Incineration System operating at
the Denney Farm Site, Missouri, on dioxin-contaminated oils, sludges, and
soils.
-139-
-------
INTRODUCTION
Any equipment used at a waste site
must be decontaminated prior to its
removal from the site. This paper
highlights a published EPA state-of-
the-art review entitled "Decontamina-
tion Techniques for Mobile Response
Equipment Used at Waste Sites," and
discusses field practices with decon-
taminating equipment presently in use
at dioxin-contaminated sites in EPA
Region VII. Personnel protective
clothing and equipment and decontam-
ination of personnel is not a subject
of this paper, though mention of it
is made where appropriate. Decontam-
ination is a separate field activity
with full-time personnel assigned to
it. Like any other field operation
at a site, decontamination must be
planned for, executed in the field,
and tested for its effectiveness. A
decontamination plan constitutes a
separate section of any operating per-
mit or any other workplan pertaining
to the site activities. The decontam-
ination plan for the EPA Mobile Incin-
eration System operating at the Denney
Farm Site in Missouri is an example of
such a plan.
The EPA report is particularly useful
to those who are new to the responsi-
bility for equipment decontamination,
either in the planning stage or in
the field. For example, the report
should be useful to those who develop
and design new treatment systems for
use at waste sites and to those who
introduce existing systems to process
hazardous waste. The report gives
background material on decontamina-
tion methods, contamination assess-
ment, and contamination avoidance.
Decontamination methods for large and
expensive components of mobile re-
sponse equipment should be non-destruc-
tive to the equipment itself, but ef-
fective in removing contaminants from
interior and exterior surfaces of the
equipment. Small items may be sacri-
ficed and/or disposed of. Determina-
tion of an acceptable level of contam-
ination is essential. Development and
testing of decontamination methods may
be necessary prior to taking equipment
to the field. Contamination assessment
ment is done with chemical and physi-
cal test(s) that are able to test down
to the determined acceptable level of
contaminants on equipment. Some test
methods may need to be developed.
The report stresses the importance of
contamination avoidance because, if
appropriate measures are taken, both
the decontamination and the contamin-
ation assessment can be much reduced
and simplified.
Field decontamination practices pre-
sented here are those of EPA Region
VII. EPA Region VII has gained con-
siderable experience with decontamin-
ation of various equipment used at
many dioxin-contaminated sites in
Missouri. Decontamination has been
performed at all levels of site activ-
ities such as field investigation,
excavation and removal of contaminated
materials, and ultimate disposal and
deli sting of residues with the use of
the EPA Mobile Incineration System at
the Denney Farm Site, Missouri. Reg-
ion VII practices contamination avoid-
ance. The amount of equipment that
would require decontamination is kept
to a minimum. They accomplish it by
using disposable equipment and by
wrapping reusable equipment with
disposable coverings. For equipment
that cannot be protected, Region VII
has developed a rigorous decontamina-
tion procedure. Equipment decontamin-
-140-
-------
inated with this procedure is then
wipe-sampled and released only when
analysis of wipe samples show non-
detectable levels of dioxin.
IMPORTANCE OF CONTAMINATION AVOIDANCE
Contamination avoidance is an important
consideration of any planned or ongoing
activity at a waste site because it
reduces hazard, time, and cost of the
final decontamination task. The report
discusses such measures as use of en-
closed structures and secondary con-
tainment for the mobile response units,
equipment safety features, and protec-
tive coatings for mobile response equip-
ment. Region VII has developed a num-
ber of practical ways to minimize con-
tamination in all of their activities
at dioxin-contaminated sites. Contam-
ination avoidance has been carefully
considered for the EPA incinerator at
Denney Farm.
Enclosures for Equipment
A combined use of an overhead struc-
ture with a secondary containment is
an effective measure to minimize
transport of hazardous materials dur-
ing operation of the mobile response
equipment and to facilitate the decon-
tamination task. An example would be
a "Butler" steel building attached to
a concrete slab.
Housing of equipment by an overhead
structure is not always required, but
it minimizes air transport of contam-
inants during its operation. A var-
iety of commercial temporary or semi-
permanent structures are available at
a cost of $30-$400 per square meter
as shown in Table 1.
Table 1.Overhead Temporary
Structures
Cost Per
Structure Type Square Meter
Air Inflatable $ 30 - $100
Steel Arch $100 - $200
Fabric Supported
by Arch $200 - $400
Prefabricated Steel
("Butler") $ 80 - $100
Secondary containment to catch spills,
leaks, or wash solution from the de-
contamination operation is always
required. It can be applied either
under the entire operating unit or at
points of the unit where leaks or
spills are most likely. A simple and
effective secondary containment can
be a sloped, combed concrete slab,
coated with a polymeric material and
with a collection sump. It can also
provide structural support for the
operating system. Another simple and
commonly used secondary containment
is a polymeric fabric cover, such as
polyurethane or polyethylene.
Safety Features
Safety features on equipment to pre-
vent spills and leaks is another means
to prevent contamination of equipment
surfaces. Most important here is the
design of the equipment itself to have
few potential sources of fugitive
emissions. The next step is to exam-
ine system components, identify those
that may break during operation, and
either change or reinforce them.
Failures typically occur in seals and
fittings used in such components as
pumps, valves, and piping systems.
See Table 2.
-141-
-------
Table 2. Equipment Components Using Seals and Fittings
Component
Pumps
In-line Valve
Open-ended Valve
Piping System
Point of Potential Failure
Drive shaft seals
Stem and bonnet
Stem, bonnet, and flow seal
Flanges, elbows, and tees
Protective Coatings
Protective coatings of paint or poly-
meric materials, either permanent or
temporary, can provide a barrier be-
tween equipment surfaces and contamin-
ants. For mobile equipment, those that
Comment
Use seal less pumps
Use bellows-sealed valves
and diaphragm valves
Minimize sharp elbows. Use
high-temperature/high-pres-
sure seals. Use stainless
steel or Inconel pipes.
are promising are temporary coatings
that can be peeled off and disposed of
by incineration, for example. For
small areas protective tapes can be
used. Protective coatings on response
equipment, except for tapes, are not
used at waste sites. See Table 3.
Table 3. Protective Coatings for External Equipment Surfaces
Coatings
Permanent
(Epoxy resins)
Temporary
(Poly vinyl
Chloride)
Quantity of
Waste Produced
Moderate
Moderate
to Large
Cost
Moderate
Moderate
to Large
Danger to
Operator
Low
Low to
Moderate
Comment
Not suitable for
mobile equipment
Two or More
Coatings Needed
Contamination Avoidance at Region VII
Region VII's approach to contamination
avoidance on dioxin-sites is to use
disposable equipment and wrapping
equipment with disposable coverings.
Among disposable equipment is most of
the personnel protective gear and
equipment such as Tyvek® suits, spent
cartridges from respirators, occasion-
ally boot covers, and inner and outer
gloves. These items are discarded
whenever they are removed at lunch
breaks or at the end of the day.
-142-
-------
Other disposable equipment are sampl-
ing spoons, sample trays, and survey
flags. Disposing of small sampling
equipment reduces chances for cross-
contamination of samples* which is
another benefit. Disposable plastic
covers are used to protect surfaces
of reusable items. The exterior of
sample jars is kept free of contamin-
ation with this method. Prior to
sample collection, the jar is placed
into a plastic bag with the opening
sealed around the top of the jar with
a rubber band. After filling the jar
with the sample, the bag is pulled
off and discarded. Disposable cover-
ings are also used to protect large
containers, so called roll-off boxes,
for transporting bulk excavated soil
from other sites to the site where
the incinerator operates. The boxes
are first triple lined with polyethy-
lene sheeting. After filling with
soil, the sheeting is sealed and
boxes are topped for transport. At
the feed storage building, the back
of the boxes are opened, the boxes
are tipped, and the plastic wrapped
soil package slides out of the box
into the storage building. The in-
terior and exterior of the roll-off
boxes remains free from contamination
and does not require decontaminating.
Also, the exterior of the wrapped
soil package is free from contamina-
tion. Of course, discarded items and
coverings add to the bulk of hazard-
ous waste. Availability of the incin-
erator as at the Denney Farm makes it
possible to dispose of this waste by
incineration. The contamination
avoidance for the EPA incinerator at
Denney Farm is achieved by situating
main components of the system in the
clean zone. Only the material storage
building for loose soil and the build-
ing housing the shredder and waste
oil tank are situated in the contamin-
ated zone. A completely enclosed con-
veyor brings the shredded feed to the
completely enclosed loading area of
the incinerator, which is located in
the clean zone. The incinerator is
also housed in a building, though the
primary reason for constructing this
building was weather protection of
the system and comfort of operating
personnel.
Pre-cleaning of Equipment Before and
After Travel
Another helpful technique to prevent
buildup of contaminants on equipment
surfaces is to clean the equipment
prior to its transport from the home
base to the site and upon arrival at
the waste site. The entire incinera-
tion system was steam-cleaned within
three days before leaving its home
base in Edison, New Jersey, to remove
oil and dirt, and this operation was
repeated upon arrival at Denney Farm
to remove accumulated road dirt.
DECONTAMINATION PROCEDURES
Decontamination procedure by a single
method or a sequence of methods re-
moves or detoxifies the residual con-
taminants from the internal and exter-
nal surfaces of the equipment. The
decontamination procedure has to be
designed for the particular hazardous
substance(s) to be treated, and also
tested on equipment prior to its use
in the field. Any decontamination is
usually time consuming and labor in-
tensive. Region VII designed and
tested a decontamination sequence for
the dioxin-contaminated equipment.
The report divides decontamination pro-
cedures into three categories: solu-
bilization methods, which use solvents
to solubilize and wash off contamin-
ants, chemical degradation methods
which chemically degrade contaminants
at the surface, and mechanical methods,
which mechanically detach contaminants
from the surface with abrasive and non-
abrasive methods.
-143-
-------
Solubillzation Methods
For the mobile field equipment, solu-
bilization methods are primarily used
because they are available, effective,
non-destructive, and suitable for both
internal and external surfaces. These
methods use water or steam, with or
without detergents, organic solvents,
and foams, gels, and pastes. Use of
foams, gels, and pastes may give better
coverage of. equipment surfaces than a
thin organic solvent. Application of
wash solutions is usually by immersion
or spray. The disadvantage of these
methods is that they generate waste,
usually liquid, that contains the
removed contaminant. See Table 4.
Table 4. Decontamination of Surfaces by Solubilization Methods
Method
Water
Steam
Organic
Solvents
Foams
Gels and
Pastes
Quantity of
Waste Produced
Large
Small
Moderate
Small
Small
Cost
Low
Moderate
Moderate
to High
Moderate
to High
Moderate
to High
Danger to
Operator
Low
Moderate
Moderate
Low
Moderate
Comment
Detergents and/or surfactants
increase effectiveness.
Detergents and/or surfactants
increase effectiveness. Can
spread contaminants.
May be toxic or flammable.
Not for internal surfaces.
Chemical Degradation Methods
Chemical degradation of surface con-
taminants can be a one-step treatment,
and generated waste, if any, could
contain no contaminant. These methods
are mainly in the developmental stage.
Methods presented in Table 5 are com-
mercially available. Flash lamp clean-
ing with pulse ultraviolet radiation
that delivers localized, short-lasting,
very high temperatures is a promising
emerging method for destruction of
surface contaminants. The removal
mechanism depends on the contaminant
and treatment conditions. The contam-
inant can be photochemically reduced,
vaporized, incinerated, pyrolyzed,
or, in the presence of water, "steam
cleaned." Some such lamps are used
for specialized cleaning operations.
Lamps can be operated safely.
-144-
-------
Table 5. Decontamination of Surfaces by Chemical Degradation Methods
Method
High Inten-
sity Light
(Xenon Lamp)
UV Light
Cleaning
Electro-
polishing
of Metal
Surfaces
Quantity of
Waste Produced
Small
Smal 1
Moderate
Cost
Low to
Moderate
Low to
Moderate
Moderate
to High
Danger to
Operator
Low
Low
Moderate
Comment
Most effective on flat
surfaces. For external
surfaces only.
Most effective on flat
surfaces. For external
surfaces only.
Developmental stage.
For small parts that
can be immersed.
External and internal
surfaces.
Electropolishing systems can be used
on a variety of alloy systems if
parts can be disasembled and immersed
in the liquid bath. After treatment,
the metal must be rinsed off, dried,
and painted.
Mechanical Methods
Mechanical decontamination methods can
be abrasive or non-abrasive to the
equipment surfaces. Abrasive methods
are destructive to the equipment and
generate large amounts of waste.
Table 6. Decontamination of Surfaces by Abrasive Methods
Method
Pigs
Brushes
Wet Abrasives
Dry Abrasives
Dry Ice
Blasting
Quantity of
Waste Produced
Moderate
Moderate
Large
Large
Smal 1
Cost
Low
Low
Moderate
Moderate
Moderate
Danger to
Operator
Low
Moderate
Moderate
High
Moderate
Comment
For internal surfaces
like pipes
Destructive to equipment
Destructive to equipment
-145-
-------
Table 7. Decontamination of Surfaces by Non-Abrasive Methods
Method
High-Pressure
Water
(200-700 atm)
Ultra-high
Pressure Water
(1000-4000 atm)
High-Pressure
Freon 113™
Ultrasonic
Vacuum with
Filters
Quantity of
Waste Produced
Moderate
to Large
Moderate
to Large
Small
Small
Small
Cost
Moderate
Moderate
to High
High
Low to
Moderate
Low
Danger to
Operator
Moderate
Moderate
Low
Low
Low
Comment
For loosely adhering
contaminants.
For tightly adhering
contaminants.
Cleans cloth, rubber,
and plastic.
Effective for cleaning
small parts that can
be immersed.
For loosely adhering
contaminants.
Of mechanical methods presented in
Table 6 and Table 7 in field use are
vacuum systems with a HEPA filter or
an activated carbon filter, brushes,
ultra high-pressure water at 1,000
atm, and high-pressure Freon 113™.
All these systems are commercially
available.
Decontamination of Equipment at Reg-
Ion VII
The EPA Region VII decontamination pro-
cedure for dioxin-contaminated equip-
ment consists of a five step sequence.
This procedure is:
- Tap water and detergent (Alconox)
wash/scrub (1 part "Alconox" and
3 parts water);
- Tap water rinse;
- Deionized water rinse;
- Methanol or Isopropyl alcohol rinse;
- Air dry.
The above procedure was developed
early in the Missouri dioxin investi-
gations and was tested with quality
assurance/quality control (QA/QC) of
wipe and rinsate samples prior to its
present routine use. The procedure
is used on all field equipment such
as backhoes, forklifts, front-end
loaders, drill rigs, trucks, and
trailers. The cleaned equipment is
wipe-sampled and released only when
the analysis shows acceptable levels
of dioxin, not higher than 100 ng/rn^.
This procedure will be used on some
components of the EPA incineration
system before it leaves the Denney
Farm site. The procedure is labor
intensive, particularly for large
equipment. It takes three people a
day or more to decontaminate a large
piece of equipment.
The operators performing decontamina-
tion wear protective clothing and
-146-
-------
equipment that is required by the
level of hazard in the operating zone;
usually it is Level C of personnel
protection.
Decontamination equipment in use at
Region VII consists of the following:
steam jenny; vacuum with a HEPA fil-
ter; tanks or containers for clean
solutions such as tap water, deionized
water, and alcohol; garden hoses and
spray nozzles; brushes and rags; and
wipe test sampling sets and jars for
liquid or solid samples. If possible,
spent wash solutions are discarded
into the ground in the contaminated
zone; some of it is used to moisten
the feea going to the shredder; some-
times alcohol is allowed to evaporate.
However, at the end of the cleanup
operation, and after the incinerator
leaves, the remaining wash solutions
must be treated.
Most decontamination operations at
dioxin sites in Region VII are per-
formed in the open. The location of
the operation is always at the border
between contaminated and clean zones
so that decontaminated equipment is
moved into the clean zone. However,
use of an enclosed structure with
catchments for fumes, wash solutions,
or solid waste, may be desirable when
many items of equipment are to be
decontaminated with a variety of
methods.
Also, some enclosed decontamination
systems for equipment are commercially
available.
CONTAMINATION ASSESSMENT OR TESTING
The need for decontamination and
effectiveness of decontamination are
determined by analyzing samples taken
from interior and exterior surfaces
of the equipment, and samples of wash
solutions or other wastes generated
by the decontamination operation are
also analyzed. The starting point is
to determine the acceptable level of
decontamination of equipment surfaces.
This enables the selection of both
suitable decontamination and, analy-
tical procedures. In the case of
dioxins, the decontamination level
is the same as the detection level,
which is the the worst case.
Sampling and analytical procedures
for decontamination assessment and
the number and location of samples to
be taken are specified in the decon-
tamination plan.
Most convenient for analysis of sur-
face contaminants in the field would
be an instant instrumental reading.
The report describes a portable fluoro-
metric monitor for polynuclear aromatic
hydrocarbons on a variety of surfaces,
such as metals, plastics, and fabrics.
The monitor was in the developmental
stage at the time the report was
written.
For analysis of dioxins, wipe samples
are taken from surfaces and sent to
the laboratory. In Region VII, wipe
samples are analyzed within 24 hours
so the decontaminated equipment can
be released the next day.
-147-
-------
SUMMARY
Decontamination of mobile response
equipment operating at a hazardous
waste site is a requirement. Any
measure taken to prevent contamination
will reduce hazard, time, and cost of
the future decontamination operation.
Both decontamination and contamina-
tion assessment procedures should be
selected and tested before taking
equipment to the field. In some
cases, development of either decon-
tamination procedure or testing pro-
cedure, or both, may be necessary or
desirable.
BIBLIOGRAPHY
1. Meade, J.P. and W.D. Ellis.
Decontamination Techniques for
Mobile Response Equipment Used
at Waste Sites (State-of-the-Art
Survey), EPA/600/2-85/105, U.S.
Environmental Protection Agency,
Cincinnati, Ohio, 1985, 74 pgs.
2. RCRA Permit, Part B. Applica-
tion for the Operation of the
U.S. EPA Mobile Incineration
System at Penney Farm. McDowell,
Missouri, 10/19/84.~
3. Personal communications with
Gary Kepko, On-Scene Coordinator,
U.S. Environmental Protection
Agency, Region VII, 25 Funston
Road, Kansas City, Kansas 66115.
-148-
-------
LEAK PREVENTION IN UNDERGROUND STORAGE TANKS:
A STATE-OF-THE-ART SURVEY
Anthony N. Tafuri
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, NJ 08837
A.C. Gangadharan, T.V. Narayanan,
R. Raghavan, G. Amoruso
Enviresponsei Incorporated
Foster Wheeler Development Corporation
Livingston, NJ 07039
ABSTRACT
The objectives of this state-of-the-art survey were to examine the design and opera-
tional practices associated with underground storage tank (UST) systems and to identify
areas for further research and development that would advance leak prevention technology.
Many standards, guidelines, and recommended practices for the design and operation
of UST systems are currently promulgated by several professional and industrial organiza-
tions. However, many of these procedures have overlapping requirements and there is no
way of confirming how widely they are understood or followed in the field. Consequently,
there is a need for a cohesive and coordinated set of rules and standards that apply to
various types of UST systems, including those that store chemicals, and for further work
to assess and improve operating practices, including spill prevention and leak detection
methods and devices.
BACKGROUND
Protecting the nation's groundwater
resources from contamination by regulated
substances* that leakt from underground
storage tank (UST) systems has emerged as
a major environmental issue. More than
50 percent of the nation's population draw
drinking water from underground sources.
There are between 2 and 3.5 million under-
ground tanks buried across the nation, of
which some 100,000 tanks are estimated to
be presently leaking, and some 350,000
are expected to leak within-the next 5
years. Accordingly, the need to improve
leak prevention technology is evident.
*Regulated substances are those substances defined in Section 101 (14)
of the Comprehensive Environmental Response, Compensation, and Liability Act
of 1980, and petroleum, including crude oil or any fraction thereof which is
liquid at 60°F, and 14.7 pounds per square inch absolute pressure.
tin this paper, the word "leak" denotes all unauthorized releases.
-149-
-------
FACTORS AFFECTING LEAK PREVENTION
Preventing leaks in UST systems
requires consideration of several factors,
including the system characteristics
(e.g., age, ownership, product stored,
size, etc.) and the elements of the
solution scheme (e.g., design and engi-
neering, fabrication and installation,
operation, corrective actions, etc.).
System Characteristics
The age of an UST system influences
the solution option. Leak prevention
technology is designed and engineered
into new systems. In old installations,
it is a reactive step requiring system
monitoring, retrofitting, remediation and
restoration. Identifying the problem is
the significant step in an abandoned
installation.
UST system ownership affects leak
prevention. Large industrial owners have
the technical, managerial, and financial
resources and the economic and legal
impetus to employ effective prevention
strategies. Small owners lack money and
organization to develop their own methods
and procedures to prevent and remediate
leaks. They require training in under-
standing the dimensions of the problem,
reliable methods and procedures for
solution, and incentives that compel them
to apply leak prevention programs.
The largest class of regulated sub-
stances stored in UST systems is gasoline
and other petroleum products. Accordingly,
these have been the primary focus of leak
prevention investigations thus far. The
problem, however, extends far beyond
this. The list of regulated substances
includes 698 chemicals which are stored
in USTs. The different physical and
chemical properties, toxicity, transport,
and fate characteristics of such regulated
substances require different approaches
to leak prevention.
The size of USTs has a significant
influence on leak prevention strategy.
Leak rates from large installations are
likely to be higher than from smaller
installations. Size differences also pre-
sent problems as regards to construction
materials, design, inspection procedures,
leak monitoring, repair, maintenance, and
replacement schedules.
Elements of the Solution Scheme
A leak prevention strategy should
ensure the integrity of the containment
boundaries for the life of the system;
avoid or minimize accidental spills and
overflows; provide warning of impending
leaks; and prevent the spread of leaking
products. These tasks require: proper
design, engineering, fabrication and
installation, correct operation, and
appropriate corrective actions through
inspection, repair, and maintenance.
These requirements can be satisfied by
applying valid principles of mechanics
and other engineering sciences; ensuring
quality of materials and workmanship;
providing appropriate tools and equipment;
establishing appropriate schedules for
inspection, repair and maintenance; and
enforcing appropriate standards and
regulations.
DESCRIPTION OF UNDERGROUND STORAGE TANK
SYSTEMS
A basic UST system includes tanks,
piping, and accessories.
Tanks
Eighty-nine percent of existing USTs
are carbon steel. Carbon steel is compat-
ible with petroleum products, inexpensive,
available, easy to fabricate and repair,
strong, and relatively resistant to damage.
However, most carbon steel USTs are not
protected against corrosion.
Tanks that have corrosion protection
include steel tanks with internal and
external coatings; cathodically protected
steel tanks; fiberglass-reinforced plastic
(FRP) tanks, and steel-FRP-bonded composite
tanks.
FRP tanks are a composite consisting
of a plastic resin matrix with a fiber-
glass reinforcement. The resins must be
compatible with stored product(s) and the
surrounding environment. An estimated 24
percent of USTs storing petroleum products
-150-
-------
use FRP tanks. Steel-FRP-bonded tanks
have a steel inner shell and an FRP outer
layer fused by a polyester resin bond.
These combine the strength and stiffness
of steel with the corrosion resistance of
FRP. Tanks made of stainless steel,
aluminum, and various plastics are used
in limited applications.
Pipes used in USTs are made of a
variety of materials, including: carbon
steel; cast iron; stainless steel; gal-
vanized steel; rubber, plastic, or epoxy-
lined steel; plastic; and fiberglass-
reinforced plastic. The same considera-
tions that enter into the selection of
the tank material apply to the piping
selection. For example, carbon steel
pipes are compatible with petroleum;
however, they are susceptible to corro-
sion when they are kept in contact with
corrosive chemicals. On the other hand,
cast iron pipes resist corrosion well,
and can be used to carry concentrated
acids. They are brittle, however, and
can break on impact or shock. Both
carbon steel and cast iron are relatively
inexpensive.
Accessories
Accessories in an LIST system include
valves, pumps, joints, fittings, vapor
recovery systems, overfill prevention
systems, and leak monitoring ports.
Secondary Containment
Secondary containment retains leaks
from a basic UST system, aids their
detection, and facilitates their cleanup.
Secondary containment can be accomplished
in two ways: (1) by building a barrier
between the basic system and the surround-
ing ground with flexible membrane liners,
a concrete vault, clay liners, or soil
sealants; or (2) by using a double-wall
structural configuration for tanks and
pipes. Both methods can be used in
combination.
DESIGN AND ENGINEERING PRACTICES
Properties of Products
Physical, chemical, and hazard
characteristics of stored products are
important UST design considerations.
Critical characteristics include the
product's physical state at ambient
temperature, melting point, boiling
point, specific gravity, vapor pressure,
explosivity, flammability, combustibility
and corrosivity.
When products stored in an UST system
comprise a mixture, the consequences of
combining the constituent chemicals must
also be evaluated. One method for deter-
mining these consequences is a chemical
class compatibility matrix (1). The
method is based on grouping chemicals
into 38 classes based on similar molecular
structure and similar reactivity charac-
teristics. Possible consequences (e.g.,
heat generation, explosion, violent poly-
merization, etc.) of mixing one class of
chemicals with another can be indicated
for ready design reference.
Mechanical Forces
Mechanical forces imposed on an UST
system and its components include: dead
loads due to product weight, self weight,
weight of soil overlay, reaction forces,
etc; live loads due to internal pressure,
thermal expansion forces, vehicular
traffic; and environmental loads due to
wind, buoyancy pressure due to ground-
water table, seismic load in earthquake-
prone zones. While most current designs
are based on manufacturer specifications
and industry standards, there are sophis-
ticated design methods and analytical
tools available to determine optimal
configurations, dimensions and layouts
of UST systems.
Corrosion
Corrosion is a major cause of deteri-
oration and failure of metallic UST
-151-
-------
systems. It can occur internally and
externally. Internal corrosion is largely
due to the incompatibility of the stored
product with materials of construction.
External corrosion is usually influenced
by soil resistivity; moisture content;
type and concentration of salts in the
soil; presence of certain types of bac-
teria; temperature; permeability of
surface film; presence of adjacent under-
ground metallic structures; and stray
underground electrical current.
Corrosion can be prevented by pro-
viding cathodic protection that forces an
electric current toward, rather than away
from, UST components, or by selecting
materials of construction that inhibit
electric flow altogether. Two design
concepts used for cathodic protection
include sacrificial anodes and impressed-
current systems.
Materials of Construction
Materials used in UST systems include
various types of metals and polymeric
materials. Structural strength and com-
patibility with products and soil environ-
ment are two key factors that determine
the choice of materials. The American
Society for Testing Materials (ASTM) has
the most universally accepted standards
for construction materials. However,
these standards do not include exposure
to all hazardous chemicals listed as
regulated substances. Consequently, UST
manufacturer guidelines and specifications
should be consulted for selection of
appropriate materials.
Codes and Standards
For the purpose of design classifi-
caton, UST systems may be grouped into:
systems that operate essentially at atmos-
pheric pressure; low-pressure systems
that operate at pressures up to 15 psig;
or high-pressure systems that operate at
pressures higher than 15 psig.
Many technical standards, guidelines,
and recommended practices generated by
professional, trade, and industrial
organizations exist for the design of
these systems. Most of the documents that
apply to atmospheric systems are developed
by the American Petroleum Institute, Amer-
ican Water Works Association, and Under-
writers Laboratories. Most standards and
design guidelines that apply to lower-pres-
sure systems are from the American Petro-
leum Institute and the American Society of
Mechanical Engineers. The ASME Boiler and
Pressure Vessel Code, by adoption, is a
legally binding standard in most states
and local jurisdictions for design, con-
struction, and operation of high-pressure
systems. The American Petroleum Insti-
tute and the National Fire Protection
Association also have some codes perti-
nent to high-pressure systems.
Several states have already estab-
lished, or are in the process of estab-
lishing, UST regulations incorporating
many of the previously mentioned guide-
lines. These include California,
Connecticut, Florida, Maine, Maryland,
New Jersey, New Hampshire, New York, and
Rhode Island.
INSTALLATION TECHNIQUES
Improper installation of UST systems
is a common cause of leakage. Proper
installation and testing procedures that
are based on sound engineering principles
can reduce such leakages. These proce^
dures include: investigation of soil
conditions and characteristics; selection
of materials of construction appropriate
for design conditions; selection of
proper bedding and backfill material;
handling and care of equipment during
construction; tightness testing require-
ments; and supervision requirements.
Although state regulations vary,
most of them require the installation of
double-walled tanks or secondary contain-
ment systems such as flexible membrane
liners, concrete vaults, clay liners,
and soil sealants; overfill protection;
leak monitoring wells; and leakage alarm
systems. Each system has very specific
guidelines pertaining to installation
procedures, cost differentials, and
environmental conditions.
-152-
-------
Piping and Accessories Installation
Vapor Recovery Systems
Selection, installation, and testing
of UST system piping must be based upon
appropriate standards and guidelines.
Field connections must be made properly
and protected against corrosion. This
involves the cleaning and preparation of
the surfaces to be connected, proper use
of thermowelding or mechanical clamps,
and application of effective corrosion
protection to the bare areas before back-
filling.
OPERATING PROCEDURES AND GUIDELINES
Professional, trade, and industrial
organizations have developed operating
procedures and guidelines concerning over-
fill prevention, transfer spill prevention,
vapor recovery, and leak detection and
monitoring. There procedures are designed
to prevent the release of products during
filling and transfer operations, and to
enable prompt recognition of underground
leaks that result from impairment of tanks,
pipes, or accessories.
Overfill Prevention
An ideal overfill prevention system
includes a level-sensing device equipped
with an alarm to alert the operator of an
impending overfill and an automatic
product shutoff when the tank is full.
Many states have specific regulations
pertaining to overfill protection. These
range from automated overfill systems to ,
delivery operator determination to manhole
containment.
Transfer Spill Prevention
Proper operating practices that
should be followed to prevent transfer
spills are well documented. These prac-
tices require: tight connections between
the hose and fill pipe; periodic inspec-
tion of all transfer hoses; inspection of
tank ullage before product delivery to
ensure sufficient capacity; proper identi-
fication of stored products and container
capacities; and proper training of all
operators who perform loading or
unloading operations.
Gasoline vapors and volatile organic
compound emissions from UST systems may
violate ambient air quality standards.
These releases occur during UST filling
and vehicle refueling. Vapors are also
emitted from truck tanks as gasoline
displaces the gasoline-enriched air in
the.tank. These vapors can be controlled
by venting through charcoal filters in the
truck tank itself or back into the UST.
Leak Detection
Leak detection is an integral part
of the regulatory requirements associated
with preventing leaks. Several states
already have such requirements. Methods
and strategies for leak detection include:
inventory control, continuous in-tank leak
monitoring, nonvolumetric methods, leak
effects monitoring, and tank integrity
testing.
Inspection
Proper inspection of tanks and other
UST components is carried out before,
during, and after the system is installed
and operated to determine the structural
integrity of the system and to evaluate
possible corrective actions should problems
be found. Inspection of existing UST
systems is difficult, if not impossible,
unless provisions have been made for
inspection ports, manways, and other means
of access.
A quality inspection program should
identify excessive corrosion, erosion of
interior parts due to abrasion by particles
suspended in moving fluids, structural
fatigue or cracking, deterioration of
liners and accessories, and weakened or
cracked welds and joints. Formal check-
lists, records of inspection, and frequent
inspections ensure a quality program.
Physical, nondestructive inspections
of UST systems are not always possible
nor do they always give a reliable assess-
ment of the UST system's integrity. Thus,
predictive methods, based on theoretical
or empirical models, supplement physical
inspections and aid in scheduling tests,
-153-
-------
maintenance, and repair. Two predictive
models proposed and used by the Petroleum
Association for Conservation of the Cana-
dian Environment (PACE) are the Soil
Aggressiveness Value (SAV) method (2) and
Roger's Regression Analysis (3).
The SAV method is based on the pre-
mise that the age at which an underground
storage tank leaks is directly related to
the surrounding soil condition. The soil
condition is characterized by a SAV which
is an aggregate number determined on the
basis of the following soil properties:
average values of soil resistivity, soil
pH, and soil moisture; differential values
(i.e., the ranges of resistivity and pH);
and presence of sulfides (which promote
bacterial action on the surface of the
tank).
The data obtained by the PACE on tank
age at failure, and the corresponding
SAV, have been developed into a tank
evaluation graph (Figure 1). The region
below the curve S, given by the equation
SAV x Age = 69, represents a 95 percent
confidence bound for occurrence of no
leaks. The higher the SAV x Age value,
the higher the probability that the tank
will leak. The curve SAV x Age = 180
envelopes 40 percent of all leaks. Based
on these analyses, PACE recommended the
following corrective actions:
Region
Recommended Action
1. I _> 180
2. 69 <_ I < 180
3. I < 69 and SAV _> 4
4. I < 69 and SAV < 4
Replace tank
Test, and replace or retrofit
Retrofit
Benign, no corrective action warranted
where I = SAV x Age
Roger's regression analysis method
is based on a statistical analysis of the
age-to-leak data correlated to measureable
characteristics of the tank environment.
The correlation equation for mean age-to-
leak is given by:
L = 5.75 R-05 T--107 exp (.12? - .42 M - .265)
where L is the mean-age-to-leak in years;
R is the soil resistivity in ohm-cm; T is
the tank size in Imperial gallons; P is
the soil pH; M is a factor related to
moisture content in the soil (1 for sat-
urated, 0.5 for damp, and 0 for dry); and
S is a factor related to sulfides content
in the soil (1 for strongly present, 0.5
for trace, and 0 for no sulfides).
It is claimed that approximately 75
percent of the total variability in the
dependent variable L is explained, with
a high degree of statistical significance,
by the full set of independent variables
included in the model. Roger's equation,
exercised with four sets of values for
independent variables, results in a mean
age-to-leak range from 13.5-16 years,
with an average of 14.9 years. This pre-
diction is close to variously reported
mean ages-to-leak of 17-19 years.
It should be noted that this method
does not allow for changes in the values
of the independent variables that would,
most likely, occur, e.g., in soil resis-
tivity and moisture content, during the
life of the tank system. Also, the
method assumes that tank failure occurs
as a consequence of only the external
soil properties without any influence of
internal condition.
MAINTENANCE AND REPAIR
There are presently no standard in-
spection, maintenance, or repair practices
available for tank owners. However, many
states are introducing regulations for
such practices. This will require quali-
fied inspectors, testers, and maintenance
personnel trained to recognize impending
failures, and to respond with appropriate
corrective actions. Designers, manufac-
turers, installers, and suppliers must
provide input to develop procedures and
practices that can be easily implemented
by the owners and operators of UST
systems.
-154-
-------
13
12
11
ut 10
I 9
3 8
ui
LU >"
55 2, 8
UI -
EC 5
C3
< A
o
-------
A PRELIMINARY ANALYSIS OF
UNDERGROUND TANKS USED FOR CERCLA CHEMICAL STORAGE
Ihor Lysyj
Environmental Monitoring and Services, Inc.
Camarillo, California
Robert Hillger, John S. Farlow and Richard Field
U.S. Environmental Protection Agency
Edison, New Jersey
ABSTRACT
The scope and severity of leaking underground storage tanks (USTc) containing chemi-
cals have not been well defined. A study was undertaken for the United States Environ-
mental Protection Agency (EPA) to collect and analyze data on USTs with the goals of (1)
obtaining better information on the chemical UST population and (2) developing a strategy
to rank underground tanks according to the hazard potential of their stored chemicals.
The study addresed only Comprehensive Environmental Response, Compensation, and Liability
Act (CERCLA) chemicals. Information sources included State surveys in California and New
York and data from the Chemical Manufacturers Association (CMA). The analysis considered
the nature (physical-chemical and toxicological properties) of the stored chemicals, tank
population, size and age of tanks, materials of tank construction, and means of tank
corrosion protection.
Solvents constitute the bulk (70-90%) of the organic CERCLA substances stored in
USTs. As reported by CMA, the most prevalent organic solvents (acetone, methanol,
toluene, methylene chloride, and xylene) constituted over 50% (both by number of tanks and
volume stored) of all CERCLA substances in USTs. The average tank size reported by
California and New York was 6,000 gallons, while those reported by the CMA was 15,000
gallons. The average tank age reported by CMA was 18 years. The majority are single
walled, steel tanks that are protected against corrosion only by paint.
INTRODUCTION
The recent and rather sudden realiza-
tion of the real and potential dangers to
human life, environment, and the economy
from leaking underground storage tanks
(UST) has resulted in major regulatory
efforts by federal, state, and local
governments. On the federal level the
Resource Conservation and Recovery Act
(RCRA) was reauthorized and signed by the
President in November of 1984. Subtitle I
of this Act directs the United States
Environmental Protection Agency (EPA) to
promulgate regulations to control leakage
from USTs. Deadlines mandated are:
registration by owners of all tanks with
the states by May 1986, estabishment of new
standards for petroleum USTs by February
1987, and establishment of new standards
for chemical USTs by August 1988. Some
states have undertaken parallel action by
promulgating their own regulations for UST
control [1]. The major thrust of ongoing
research is currently directed toward pet-
roleum products, while relatively little
-156-
-------
attention is being paid to chemicals stored
in underground tanks. This paper is based
on a project that defined the physical uni-
verse of CERCLA chemical USTs. Included
are the nature (physical-chemical and toxi-
cological properties) and the quantity of
stored chemicals, as well as the size,
design, and construction characteristics of
USTs. The objective is to provide a fac-
tual basis for the prioritization of UST
leak control research (as a function of the
stored chemicals).
The EPA divided UST leak control
research into three main categories, i.e.,
leak detection/monitoring, leak prevention,
and corrective action. As indicated,
research has begun for petroleum tanks.
The main emphasis is on volumetric leak
detection methods (e.g., those based on
changes in liquid level). State-of-the-art
information has also been compiled for
petroleum tank leak prevention and correc-
tive action; however, chemical tanks have
not yet been addressed. As a first cut,
this project used a matricized ranking
method for CERCLA chemical hazard potential
to enable EPA to optimize a research
program strategy for chemical tanks. Along
with volume and number of tanks, the matrix
included toxicity, ignitibility/reactivity,
and viscosity/density of stored chemicals.
The results of this analysis will pri-
marily be used for an intelligent estimate
of the national hazard potential of stored
chemicals. From this, chemicals can be
grouped and prioritized for UST research
and regulatory purposes. Chemicals with
similar physical-chemical properties will
be grouped together. The properties
(including toxicity) will be used for hard-
ware compatability requirements, industrial
hygiene standards (human handling
requirements), and signal-sensing require-
ments for volumetric (liquid level) leak
detection. These groupings will further be
used in developing a logical research
progression in the prevention and correc-
tive action categories, and will provide
an important reference/database for on-
scene decision-making in the case of
corrective action needs.
SCOPE
Since EPA has mandated deadline for
new standards for CERCLA chemical USTs,
this project was designed to provide a
logical foundation for chemical UST leakage
control research. Aside from petroleum,
Subtitle I of the reauthorized RCRA, defi-
nes regulated substances as those chemicals
that are listed under Section 101 (14) of
CERCLA [2], excluding hazardous waste.
All 611 CERCLA chemicals (excluding
those substances that fall under the CERCLA
categories of waste streams) defined by
Section 101 (14) were considered in this
study. Specifically, an attempt was made
to determine:
a) composition of regulated substances
stored underground
b) number and volume of tanks storing
each substance
c) UST design characteristics, including
materials of construction and corro-
sion protection
d) age of tanks.
SOURCES OF INFORMATION
Chemicals that require underground
storage because of either flammability or
other safety hazards are produced and
handled by the commercial community that
can be classified as:
0 Primary Producers
0 Formulators - Distributors
0 Ultimate Users
The relative number of underground
storage tanks increases from top to bottom
in this classification scheme.
Commerce generally dictates the nature
and quantity of chemicals developed by the
primary producers. The primary producers
provide large quantities of these chemical:
in bulk storage for use by the chemical
formulators who combine these chemicals
into useful products and sell them to
industrial, agricultural, and household
users. Chemical distributors perform
similar function by repackaging. The
majority of tanks are owned by chemicalj
users, such as manufacturing, agricultjj
and service operations. Principal sou
of UST information used in this study]
State surveys in California [3,4] anc
York [5]. Additional information
vided by CMA, a trade organization
represents primary chemical produce
-157-
-------
nationally [63.
REPRESENTATIVENESS AND COMPLETENESS OF DATA
Available state sources of UST infor-
mation included: summary reports from
California representing 100% of its UST
users [3,43 and a computer data base file
of hazardous substances stored in USTs in
New York [53 representing approximately 90%
of the New York UST population. A CMA sur-
vey of its members (178 companies repre-
senting 90% of primary chemical producers)
[63 resulted in a 34% response of its mem-
bership. The product of this survey is a
list of tank capacities, contents, status,
(in use or abandoned), construction, corro-
sion protection, and age.
Each information source provided a
different perspective on the universe of
USTs. The CMA data are national in scope
and, while somewhat incomplete, neverthe-
less indicate general patterns prevailing
in the primary producing industry.
California is a large southwestern coastal
state with an integrated industrial/agri-
cultural base and a significant contribu-
tion of new technologies. New York is a
large northeastern state, with a broad
industrial/agricultural base and a prepon-
-------
UST Design Characteristics
The data (Table 3) indicates that the
great majority of New York and CMA USTs
that store chemicals are made of steel
(94%) and are single walled. The remainder
are made of concrete, copper, fiberglass,
aluminum, or polymeric materials. More
than one-half (56%) of the tanks are pro-
tected only by paint and only 18% have
cathodic protection; the remainder use tar
and fiberglass coatings. The age of USTs
ranges between brand new and 60 years. The
mean and .median ages are each 18 years 16].
CONCLUSIONS
Both organic and inorganic substances
are stored in underground tanks. The orga-
nics constitute the major part (92%) of
CERCLA chemicals stored underground by
major chemical producers. Significant
amounts of inorganic substances, however,
are stored underground by users of chemi-
cals in hardware manufacturing and in
electroplating. Seventy to 90% of the
organic substances stored underground are
solvents, including:, alcohols; ketones;
alicyclic, aromatic, and chlorinated hydro-
carbons; esters; and ethers. Monomers and
miscellaneous chemicals constitute the
balance. Based on the California and New
York data, it appears that only a small
fraction (1 to 2 percent) of the total UST
population is devoted to storing CERCLA
chemicals. Petroleum products comprise
98-99% of the total liquids stored in USTs.
The top five CERCLA chemicals stored
underground are acetone, methanol, toluene,
xylene, and methylene chloride. These five
solvents represent over 50% of the CERCLA
organics stored underground.
The majority of underground tanks used
for storing chemicals are made of steel and
are single walled. More than half are
corrosion protected only by paint. The
average tank size reported by the two sta-
tes is 6,000 gallons, while CMA data show
that primary chemical producers use tanks
of a larger size (15,000 gallons).
The conclusions from this study are
based on an organized and logical ranking
of the stored substances and will aid the
EPA in developing a research and rulemaking
program for leak detection, prevention, and
corrective action of CERCLA hazardous
substances stored in USTs.
REFERENCES
1. "California Underground Storage Tank
Regulation." California State/Water
i. Resources Control Board, Sacramento,
CA, August 1985.
2. Code of Federal Regulations, Title 40,
Part 302, US GPO, Washington, DC,
1986.
3. "Underground Container Program - Con-
tainer Summary by County". State
Water Resources Control Board, Sacra-
mento, CA, October 1985.
4. Jack Kooyoomjian, "CERCLA Hazardous
Substances in California USTs". Priv-
ate communication, EPA/OERR.
5. Richard Coriale and Russell Brauk-
sieck, "Revelation Computer File and
Data Summary". New York Dept. of En-
vironmental Conservation, Division of
Water, Bureau of Spill Prevention and
Response, Albany, NY, January 1987.
6. "Tank Notification". Jeff Reamy,
Chemical Manufacturers Associationj
Washington, DC, August 1986.
-159-
-------
TABLE 1
PREDOMINANT TYPES OF ORGANIC CHEMICALS STORED UNDERGROUND
(by percent*)
SOLVENTS
Ketones/al dehydes
Aromatic hydrocarbons
Alcohols
Chlorinated hydrocarbons
Esters
Ali cyclic hydrocarbons
Total solvents
MONOMERS
MISCELLANEOUS CHEMICALS
PESTICIDES
California
1 2
35.6 32.9
22.2 21.1
10.2 8.8
12.5 14.0
6.0 4.4
0.6 0.7
87.1 81.9
3.6 6.2
7.4 7.0
1.4 4.2
New York
1 2
25.2 31.5
37.8 32.9
16.5 17.2
5.7 4.0
6.2 4.4
91.4 90.0
2.8 1.6
6.0 8.0
CMA
1 2
23.5 21.7
21.8 22.3
18.8 16.8
12.6 10.3
1.2 0.8
0.4 0.4
78.3 72.3
13.3 22.2
8.8 5.0
Note: l=number of USTs; 2=volume of USTs
* Totals may not sum to 100% because of rounding
-160-
-------
TABLE 2
MAJOR ORGANIC CERCLA SUBSTANCES STORED IN USTs
Chemical Groups
Ketones/Aldehydes
Acetone
Methyl ethyl
ketone
Methyl isobutyl
ketone
Cyclohexanone
Formal dehyde
Acet aldehyde
Total
Aromatic Hydro-
carbons
Toluene
Xylene
Benzene
Ethyl Benzene
TOTAL
Alcohols
Methanol
n-Butanol
iso-Butanol
TOTAL
Ali cyclic Hydro-
carbons
Cyclohexane
California
1
441 22.8
205 10.3
38 1.9
7 0.4
14 0.7
2 0.1
707 35.6
265 13.3
162 8.1
9 0.5
5 0.3
441 22.3
131 6.6
46 2.3
25 1.3
202 10.2
12 0.6
2
2334 18.0
1254 9.6
235 1.8
45 0.4
401 3.1
10 0.1
4282 32.9
1847 14.2
818 6.3
28 0.2
47 0.4
2740 21.1
714 5.5
277 2.1
158 1.2
1149 8.8
67 0.7
New York
1
43 12.0
32 9.0
12 3.8
3 0.8
90 25.2
80 22.4
54 15.5
1 0.3
135 37.8
41 11.5
14 3.9
4 1.1
59 16.5
2
319 18.3
122 7.0
78 4.5
30 1.7
549 31.5
369 21.1
205 11.7
1 0.1
575 32.9
149 8.5
146 8.4
5 0.3
300 17.2
CMA
1
98 17.8
21 3.8
8 1.5
1 0.2
1 0.2
129 23.5
72 13.1
31 5.6
10 1.8
7, 1.3
120 21.6
87 15.8
11 2.0
4 1.0
102 18.8
2 0.4
2
1682 19.1
145 1.6
66 0.8
2 <0.1
20 0.2
1915 21.7
1481 16.9
300 3.4
120 1.4
56 0.6
1957 22.3
1309 14.9
130 1.5
37 0.4
1476 16.8
63 0.4
Note: l=number of USTs; 2=volume of USTs in thousands of gallons
-161-
-------
TABLE 2 (cont.)
MAJOR ORGANIC CERCLA SUBSTANCES STORED IN USTs
Chemical Groups
Chlorinated Hydro-
carbons
Methylene chloride
Methyl chloride
1,1,1-Trichloro-
ethane
Trichloroethylene
Dichloropropane
Chlorobenzene
1,2-Dichloro-
propane
Chloroform
Ethyl ene di-
chloride
Tetrachloro-
ethylene
Pentachlorophenol
Carbon tetra-
chloride
Trichlorobenzene
1,2 Dichloro
ethyl ene
Methyl bromide
Ethyl ene di bromide
1,1,2-Trichloro-
ethane
TOTALS
Esthers
Ethyl acetate
n-Butyl acetate
iso-Butyl acetate
Amyl acetate
TOTALS
California
1
%
56 2.8
6 0.3
71 3.6
18 0.9
1 <0.1
3 0.2
—
12 0.6
13 0.7
36 1.8
12 0.6
3 0.2
3 0.2
1 <0.1
1 <0.1
1 <0.1
3 0.2
247 12.5
22 1.1
94 4.7
—
3 0.2
119 6.0
2
%
269 2.1
21 0.2
250 1.9
71 0.5
10 0.1
3 <0.1
—
465 3.6
367 2.8
146 1.1
144 1.1
28 0.2
11 0.1
10 0.1
10 0.1
9 0.1
3 <0.1
1823 14.0
111 0.9
431 3.3
—
22 0.2
564 4.4
New York
1
%
5 1.4
1 0.3
9 2.5
2 0.6
—
1 0.3
—
—
1 0.3
—
—
1 0.3
—
—
—
—
—
20 5.7
7 2.0
2 0.6
13 3.6
—
22 6.2
2
%
13 0.7
4 0.2
12 0.7
11 0.6
—
30 1.7
—
—
1 <0.1
—
—
1 0.1
—
—
—
—
—
71 4.0
23 1.3
8 0.5
46 2.6
—
57 4.4
C
1
%
47 8.5
10 1.8
6 1.1
2 0.4
3 0.4
1 0.2
1 0.2
—
—
—
—
__
__
—
—
—
--
69 12.6
2 0.4
2 0.4
2 0.4
—
6 1.2
MA
2
%
666 7.6
31 0.4
99 1.1
80 0.9
24 0.3
3 <0.1
4 <0.1
—
" —
—
—
—
—
—
—
—
907 10.3
44 0.5
12 0.2
15 0.2
. —
71 0.8
Note: l=number of USTs; 2=volume of USTs in thousands of gallons
-162-
-------
TABLE 2 (cont.)
MAJOR ORGANIC CERCLA SUBSTANCES STORED .IN USTs
Chemical Groups
Monomers
Aero lain
Ethyl ene oxide
Styrene
Acryl on i tr i 1 e
Vinyl acetate
Vinyl chloride
Methyl metha-
cryl ate
Vinyl i dene
chloride
Propylene oxide
l-Chloro-2,3-
epoxypropane
Ally! chloride
Ethyl acrylate
TOTALS
Miscellaneous
Chemicals
Pesticides
TOTALS
California
1
%
—
3 0.2
33 1.7
1 <0.1
11 0.6
1 <0.1
6 0.3
4 0.2
7 0.4
—
--
3 0.2
69 3.6
145 7.4
37 1.4
1979
2
%
—
40 0.3
340 2.6
20 0.2
165 1.3
10 0.1
61 0.5
58 0.4
58 0.4
—
—
51 0.4
803 6.2
951 7.0
598 4.2
12997
New York
1
%
—
1 0.3
--
4 1.1
—
1 0.3
1 0.3
—
—
3 0.8
10 2.8
21 6.0
— —
357
2
%
—
—
6 0.3
—
6 0.3
--
1 0.1
—
10 0.6
—
—
6 0.3
29 1.6
144 8.0
— —
1725
C
1
%
15 2.7
14 2.5
13 .1.4
9 1.6
7 1.3
5 0.9
5 0.9
1 0.2
1 0.2
1 0.2
2 0.4
--
73 13.3
48 8.8
_-
549
MA
2
%
787 9.0
345 3.9
305 3.5
52 0.6
151 1.7
163 1.9
48 0.5
7 0.1
20 0.2
50 0.6
20 0. 2
--
1948 2.2
452 5.0
• __.
8725
Note: l=number of USTs; 2=volume of USTs in thousands of gallons
TABLE 3
UST DESIGN CHARACTERISTICS
by percent
STEEL
Carbon
Stainless
Double Wall
TOTAL
Concrete
Fibreglass
Copper
Aluminum
OTHERS
New York
86.4
3.4
0.5
90.3
3.4
3.4
2.9
CMA
87.7
6.0
ND*
93.7
3.5
0.7
1.1
6.2
0.8
*ND - Not determined
-163-
-------
U.S. EPA EVALUATION OF VOLUMETRIC UST LEAK DETECTION METHODS
Joseph W. Maresca, Jr. and Robert D. Roach
Vista Research, Inc.
Palo Alto, CA 94303
James W. Starr *
Enviresponse, Inc.
Livingston, New Jersey 07039
John S. Farlow
U.S. Environmental Protection Agency
Edison, NJ 08837
ABSTRACT
This report summarizes the quantitative results through January 12, 1987 of the
ongoing U.S. Environmental Protection Agency's (EPA) Hazardous Waste Engineering Research
Laboratory program to evaluate the performance of commercially available, volumetric test
methods for detecting leaks in underground petroleum storage tank systems. Volumetric
methods (i.e., those operating in or on the tank that yield a quantitative estimate of
the leak rate) can be influenced by a wide variety of environmental factors, all of
which can significantly reduce the accuracy of the measurement.
The first set of full-scale product temperature experiments on a 30,285 L (8,000 gal)
tank were conducted to assess the impact of thermally-induced volume fluctuations on the
testing of overfilled tanks, the most common test condition. The initial data indicate
first that thermal effects are large when the temperature of the added product is differ-
ent from that of the in situ ground and stored product temperature, even 24 h after pro-
duct delivery, and second that these effects can significantly impair a method's ability
to detect small leaks unless the thermally-induced volume changes are compensated. When
the vertical and horizontal distribution of temperature was investigated, the results
indicated that volume-weighted temperature changes measured by a single vertical ther-
mistor array located at the fill hole of the tank would be adequate for compensation of
thermally-induced volume changes of the product throughout the tank. The results also
indicated that a 20 cm (8 in.) vertical separation of thermistors on the array was ade-
quate to characterize the temperature fluctuations that caused these volume changes.
An estimate of the technological limits of detecting leaks with volumetric test
methods is also being made. These results can be used to assess the performance of
existing test methods, as well as new ones that might be developed in the future. The
analysis suggests that, with proper instrumentation and procedures, a leak rate of
0.19 L/h (0.05 gal/h) can be detected with a probability of detection (PD) of 0.95 and
a probability of false alarm (PFA) of 0.001, providing no other sources of ambient
noise are present (e.g., tank deformation, vapor pocket).
* Now with Vista Research, Inc., Palo Alto, California 94303
-164-
-------
INTRODUCTION
Leaking petroleum underground storage
tanks (USTs) represent a serious environ-
mental threat. The United States Environ-
mental Protection Agency (EPA) recently
estimated that approximately 25% of the
UST throughout the United States are
leaking at a rate of 0.23 L/h (0.06 gal/h)
or greater (1). Records from past release
incidents indicate that, without the use
of release detection, a release can
become substantial before it is detected.
Most releases are first detected by
people seeing, test tasting, or smelling
the released material in the environment.
Only about 20% of the releases have been
detected by inventory reconciliation or
tank tightness testing procedures. The
voluntary practice of release detection
by UST owners and operators gives rise
to the low percentage of incidents
reported by release detection methods (2).
The 1984 Hazardous and Solid Waste
Amendments to the Resource Conservation
and Recovery Act of 1976 have charged the
EPA with developing regulations for the
detection of releases from UST. Develop-
ment of technically sound and defensible
regulations requires that 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 transport,
fate, and amount of petroleum considered
hazardous in the environment are not
sufficiently known. A performance stan-
dard that is based on the current techno-
logy will minimize the uncontrolled
release of petroleum product. Unfortu-
nately, the data required to formulate a
realistic regulatory policy are incomplete
or nonexistent. Many commercially avail-
able leak detection methods can be used
to detect small releases. Claims of
excellent performance are made, but little
evidence, theoretical or experimental, is
provided to support these claims. Most
manufacturers of test methods claim to
detect leak rates of 0.19 L/h (0.05
gal/h) or smaller, the practice recom-
mended by the National Fire Protection
Association (NFPA) Pamphlet 329 (3).
Limited evidence suggests that the most
common methods are not reliably meeting
this claim (4).
The environmental hazards posed by
uncontrolled releases from UST are so
formidable that they require the system-
atic application of current and evolving
release detection technology. The
Hazardous Waste Engineering Research
Laboratory (HWERL) is currently conducting
a research program to characterize the
performance of the commercially available
volumetric tightness test methods and to
determine the technological limits of
detection using this approach. A detailed
description of the project is presented
in "Protocol For Evaluating Volumetric
Leak Detection Methods for Underground
Storage Tanks" (5). An overview of the
approach, including a discussion of the
major variables affecting volumetric leak
detection methods and a description of
EPA's full-scale, environmentally safe
test apparatus, has also been published
(6). Over 40 volumetric detection
methods have been identified so far.
The purpose of this engineering
research program is to evaluate commercial
volumetric methods for detecting leaks in
underground storage tanks containing
petroleum motor fuels such as gasoline,
kerosene, and diesel oil. The results
will be used to develop regulations to
reduce the pollution of groundwater by
leaks from these tanks. This progress
report briefly summarizes the important
results to date derived from the experi-
ments being conducted in the specially
designed test apparatus located at the
EPA's Hazardous Waste Engineering Research
Laboratory (HWERL) facility in Edison,
New Jersey. A complete research report
will be published at the conclusion of
this research project.
OVERFILLED TANK TEST EXPERIMENTS
The most common method of release
detection is a volumetric tank tightness
test. Product level changes in the tank
are measured and converted to volume
changes through the knowledge of the
tank geometry. Product level measurements
alone are not sufficient to detect a small
leak, because the height changes that
normally occur in a tank that is not leak-
ing are large enough to mask a small leak.
These volume changes can arise from expan-
sion or contraction of the product in the
tank by temperature changes, expansion or
contraction of the volume of any trapped
-165-
-------
vapor by temperature and pressure changes,
structural deformation of the tank ends
and walls, evaporation or condensation of
the product, and surface and internal
waves. A method of compensating for each
of these sources of ambient noise is
required to detect small leaks.
Three classes of volumetric tank test
methods are being evaluated: overfilled
tank test methods, partially-filled
tank test methods, and methods that
circulate product in the tank. The most
common method of tank testing is the
overfilled tank test. This type of
method tests the entire system (tank,
piping, and associated underground equip-
ment) for leaks, but requires the purchase
of sufficient additional product to fill
the tank system completely. For a low-
budget operator, this requirement can
impose a substantial financial hardship.
It should also be remarked that many UST
systems are never filled this full at any
time during routine operations.
In this method, product is added
to the existing product in the tank to
obtain a level in the fill tube itself
or in an above-ground extension of the
fill tube. Accurate measurement of
product level in the fill tube is easy;
for example, a 0.04 L (0.01 gal) volume
causes a 0.37 cm (0.15 in.) product
level change in a system with a 10 cm
(4 in.) fill tube with a 5 cm (2 in.)
vent tube.
Ambient Noise
A comprehensive set of experiments
is being conducted in the test apparatus
to characterize and model the magnitude
of each source of ambient fluctuations
that limit the performance of volumetric
leak detection test methods. The first
tests focused on the volumetric fluctua-
tions that limit the performance of test
methods that overfill the tank. The
dominant ambient volume fluctuations for
these test methods are:
o expansion and contraction of the
product in the tank by temperature
changes,
o expansion and contraction of trapped
vapor in the tank and piping by temper-
ature and pressure changes, and
o structural deformation of the tank by
changing product levels before or
during the test.
Product level changes in overfilled
tank test methods will be generated by
summing the volume changes effected by
these three sources of noise. Three types
of experimental runs are being conducted
to collect the data required to develop
these three models and to validate them.
This is accomplished by designing a set
of experiments to isolate each noise
source and show that the residual volume
fluctuations, after subtracting the
volume fluctuations predicted by the noise
model from the measured product fluctua-
tions, are small enough to be negligible
(i.e., less than 0.04 L/h).
Product level changes due to evapor-
ation or condensation are small because
of the small surface area open to the
atmosphere. Surface waves are essentially
nonexistent, although a long period stand-
ing wave (seich) in the tank is possible
due to multiple openings in the tank
which may be vented to the atmosphere
during a test. Internal waves in a tank
will cause periodic temperature changes
and sometimes periodic product level
fluctuations that are unrelated to the
temperature changes required to estimate
thermal expansion or contraction of the
product. These effects, which are charac-
teristic of the temperature field, can be
minimized by properly sampling and filtering
the data.
Thirteen 24- to 48-h tests have
already been conducted to characterize
the temperature field and volume changes
generated by temperature fluctuations.
Each run has been conducted by adding
15,142 L of product to the half-filled
30,245 L capacity steel tank at tempera-
tures that are 0 to 10°C cooler or warmer
than the temperature of the ground, and
of the in situ stored product. After
adding product, all tests were conducted
with the fluid level between 234 cm
(92 in.) and 236 cm (94 in.) This level
permits temperature changes in a full tank
to be approximated accurately without
trapping vapor pockets.
-166-
-------
Product Temperature Analysis
Three analyses of data from these
thirteen tests have been performed. The
first analysis estimates and tabulates
the magnitude of the thermal volume
changes as a function of time after deli-
very. The results indicate that ther-
mally-induced volume changes are large,
even 24 h after adding product to the
tank, and that temperature compensation
is necessary to conduct an accurate tank
test. In the course of a 24-h test,
uncompehsated volume rates of 1 to 3 L/h
were observed.
The second and third analyses charac-
terize the vertical and horizontal spatial
inhomogeneities of the temperature field,
respectively. The coefficient of thermal
expansion, the volume of the product in
the tank, and the temperature change of
the product is required to estimate
thermally-induced volume changes of the
product. The second temperature study
was performed to compare the ability of
each of the three thermistor arrays to
measure thermally-induced volume changes
in the tank. In particular, this analysis
examined whether one thermistor array is
sufficient to characterize the temperature
field of the whole tank. The data were
analyzed as follows:
o A thermally-induced volume change time
series was generated for each ther-
mistor array as well as for the average
of the three arrays.
o The average thermally-induced volume
change computed for all three arrays
was subtracted from each array's
volume, and the residuals were differ-
entiated and smoothed with a 1-h box
car window (running average).
o The root mean square (rms) error was
calculated for successive 2-h segments
to give an estimate of the error in a
thermally-compensated leak rate deter-
mined from an individual array.
Results suggest that a single verti-
cal array of thermistors having a spacing
of 20 cm is sufficient to characterize
the temperature field of the whole (2.4 m
diameter by 6.1 m long) tank when testing
is begun at least 4 to 6 h after product
delivery. In the first 4 to 6 h, large
differences in temperature between the
three horizontally spaced arrays are
observed; during this interval even three
arrays are not sufficient to characterize
the temperature field. However, after 6 h
the measurements indicate that the
difference in the temperature fluctuations
between arrays is small.
The average rms errors for each
array are summarized in Table 1. This
analysis suggests that a single array can
measure temperature changes to within
0.04 L/h, which is required to reliably
detect leak rates, of 0.19 L/h. Because
of the turbulence caused by the physical
disturbance of mixing and by the thermal
settling of the fluid in the tank, the
rms errors calculated for the first 6 h
of each experimental run were excluded
from the analysis. Due to the presence
of a bad thermistor in the array, the rms
of Array 3 appears to be somewhat higher
than those of Array 1 or Array 2.
TABLE 1. AVERAGE RMS* ERROR FOR THREE
VERTICAL THERMISTOR ARRAYS
Array 1 ' 0.041 L/h
Array 2 , 0.039 L/h
Array 3 0..046 L/h
Mean 0.042 L/h
* Root Mean Square
The cumulative rms errors were plot-
ted as a function of .the time of day to
observe any diurnal variations of the rms
error. An increase of approximately
0.01 L/h in the rms error is observed .
during the 6 h period from 10 a.m. to
4 p.m., coinciding with the period of .
most rapid change of air temperature.
For the next 18 h, the rms error settles
to a level of approximately 0.038 L/h..
The third analysis was performed to
investigate whether the 20 cm vertical
separation of thermistors on each array
was adequate to characterize the vertical
distribution of temperature in the tank.
The mathematical coherence was formed
between adjacent pairs of thermistors and
was found to be near unity for frequencies
less than 1 cycle per 2 h.
-167-
-------
These results have a significant
implication for tank testing. First,
temperature compensation is essential
to the conduct of an accurate tank test.
Second, independent of structural defor-
mation effects, a waiting period of at
least 4 to 6 h after topping off a par-
tially filled tank with product is re-
quired before testing begins. Third,
a single array of thermistors, at the
fill hole, is sufficient to obtain the
necessary data for effective temperature
compensation. The effect of the number
of thermistors, test time, and sample
interval is being quantified by simula-
ting the performance of six canonical
test methods.
PERFORMANCE OF CANONICAL TEST METHODS
The performance of six canonical
(representative) methods for testing
overfilled tanks is being evaluated
while detailed descriptions of commer-
cially available methods are being
collected and modelled. These "theo-
retical test methods" invoke sound
operational, data collection, and data
analysis practice. The implied prac-
tice can readily be incorporated into
commercially available methods. This
preliminary evaluation will support the
development of performance standards
sooner than would otherwise be possible.
The canonical methods are described as
follows:
o product level measurements only
(no temperature measurements);
o product level and one temperature
measurement at the center of the ..tank;
o product level and three temperature
measurements, volumetrically weighted
by tank volume, and located at three
equally spaced vertical intervals;
o product level and five temperature
measurements, volumetrically weighted
by tank volume, and located at five
equally spaced vertical intervals;
o product level and the average tempera-
ture computed from a vertical array of
twelve equally spaced temperature
sensors; and
o product level and a volume-weighted
average temperature computed from a
vertical array of twelve temperature
sensors.
Performance will be expressed in
terms of the probability of detection
(PD) and probability of false alarm
(PFA). PD is the probability that a
given method will detect an actual leak
at or above its threshold rate in an
UST. PFA is the probability that a
given method will declare that a tank
is leaking at or above its threshold
rate when, in fact, the tank is tight.
The data for each test method are
collected and processed at a 1 sample/
min rate. The precision of the height
measurement sensor is assumed to be
0.25 mm (0.01 in.), corresponding to a
volume change of 0.002 L (0.0005 gal).
The temperature sensor precision is
taken to be 0.001°C. All tank tests are
conducted between 12 and 24 h after
product delivery to a half-filled tank.
The estimated volume rate for each tank
test is calculated by subtracting the
temperature time series from the product
level time series after converting each
to equivalent volume, and fitting a
least squares line to the residual volume.
Preliminary performance curves that
display PD versus PFA as a function of
leak rate have been generated for each
canonical method and for test periods of
1, 2, 3, and 4 h. The curves are gener-
ated from a histogram of the temperature
compensated volume fluctuations compiled
using the data from the EPA test appara-
tus. Typical results of these analyses
are shown in Figure 1 for a canonical
method employing five equally spaced
thermistors in a single array. The
results presented in this figure include
only the volume fluctuations from one
source of ambient noise (thermal expan-
sion or contraction of the product).
The performance results indicate that
temperature compensation is essential
and that performance is improved with
increased test time and number of ther-
mistors used for temperature compensa-
tion. Without temperature compensation,
only leak rates of at least 4.75 L/h
(1.19 gal/h) are detectable with a PD
= 0.95 and a PFA = 0.001.
-168-
-------
8
CO
' o
NOI133130
in
o
•e
o ^
il
W
K
D
o
o
33N3H8nD30 JO
-169-
-------
The improvement in test method per-
formance when the volume fluctuations are
temperature compensated is reflected in
Table 2. From this it can be seen that
a 1-h test with a single thermistor
compensation scheme will detect a leak
rate of 0.42 L/h (0.11 gal/h) with a PD
» 0.95 and a PFA = 0.001. A 2-h test,
however, with temperature compensation
provided by a complete vertical array
that has been volumetrically weighted,
will detect a leak rate of 0.17 L/h
(0.04 gal/h) at this performance level.
Tests are currently under way to
quantify the magnitude of the volume
changes caused by structural deformation
of the tank and by trapped vapor in the
tank. When these are complete and more
temperature runs are added to the data
base, a final performance estimate can
be made. These results should be indi-
cative of the performance that the
technology is capable of achieving.
As part of the experiments completed to
date, a method of measuring the volume
of trapped vapor pockets in any UST has
been developed to afford estimates of
volume fluctuations produced by a vapor
pocket. This is an important measurement
because all overfilled tanks trap vapor,
and heretofore there has been no means to
measure the volume of these pockets.
With all proturberances removed, the
tanks of the test apparatus trap approx-
imately 40 L of vapor when overfilled.
These tanks were installed with end to
end elevation changes of less than
0.6 cm. Field installations have a
typical elevation difference of 5 cm
and will, therefore, trap larger volumes
of vapor. As a consequence, the 40 L of
trapped vapor represents a very much
smaller volume of trapped vapor than is
typically found in the field.
TABLE 2. DETECTABLE LEAK RATE IN LITERS PER HOUR FOR THREE OF THE CANONICAL
TEST METHODS FOR A PROBABILITY OF DETECTION OF 0.95 AND A PROBABILITY
OF FALSE ALARM OF 0.001. (This performance estimate includes only the
volumetric fluctuations caused by product temperature fluctuations.
A 0.05 gal/h leak rate is equal to a 0.19 L/h leak rate.)
Canonical Test Method
One Thermistor
Five Thermistors
Vertical Array, Volumetrically
Weighted (12 Thermistors) ...
0.32
Leak Rate (L/h)
Test Duration
1-h
0.42
0.38
2-h
0.37
0.31
3-h
0.34
0.28
4-h
0.33
0.27
0.17
0.13
0.12
CONCLUSIONS
Preliminary results from the first
set of ambient noise experiments in the
test apparatus are available. These early
tests have been conducted to assess the
impact of temperature fluctuations on
testing in overfilled tanks (the most
common condition for volumetric testing).
The initial results indicate that product
thermal effects are large when the temper-
ature of the added product is different
from that of the in situ ground and the
stored product. During the time period
immediately after product addition (i.e.,
the first 4 to 6 h), the temperature
fluctuations are not horizontally coher-
ent, so that a single thermistor array
cannot adequately compensate for thermal
fluctuations. After this initial period,
the horizontal coherence of the long term
trends is of order one, and a single
vertical array of thermistors at the fill-
hole of the tank can provide adequate
-170-
-------
thermal compensation. Coherence analysis
between pairs of thermistors separated
vertically by 20 cm indicates that this
separation is generally adequate to define
the vertical temperature distribution of
the temperature field in the tank.
At the same time the experimental
work was being carried out, a theoretical
basis was developed that EPA's Office of
Underground Storage Tanks (OUST) could
use to rapidly assess the performance of
existing test methods, as well as the
performance of any newly developed tests
that might be brought to EPA's attention
in the future. When presently available
commercial volumetric leak detection
tests that overfill the tank were exa-
mined, the majority were found to fall
into six classes. For each of these
ACKNOWLEDGMENT
This work was performed for the EPA
under Contract No. 68-03-3255 with
Enviresponse, Inc., Livingston, New
Jersey.
REFERENCES
1. Westat, Inc., Midwest Research Insti-
tute, Battelle Columbus Division, and
Washington Consulting Group, 1986.
Underground Motor Fuel Storage Tanks;
A National Survey, Vol. 1 Technical
Report, EPA 560/5-86-013, Office of
Pesticides and Toxic Substances,
USEPA, Washington D.C.
2.
-171-
-------
NATO/CCMS PILOT STUDY ON DEMONSTRATION OF REMEDIAL ACTION TECHNOLOGIES
FOR CONTAMINATED LAND AND GROUNDWATER
by
Donald E. Sanning
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
and
Robert Olfenbuttel
U. S. Air Force
Tyndall Airforce Base, Florida 32403
ABSTRACT
Groundwater and soil contamination by hazardous waste is a pervasive problem in
industrialized countries. As scarce resources, water and land must be returned to
productive use. Current cleanup efforts are hampered by limited technology options
and high costs. It's desirable to build up the knowledge base so that more efficient,
cost effective remedial technologies can be developed. However, the urgent needs of
society require that near-term solutions be found and applied to the most significant
pollution problems. Consequently, promising new technology must be tested and
demonstrated to determine their applicability and effectiveness for today's problems.
The U. S. Environmental Protection Agency has established a formal program to
enhance the development and use of new or innovative technologies for mitigating the
problems caused by releases of hazardous substances at uncontrolled hazardous waste
sites. In the United States the program is called the Superfund Innovative Technology
Evaluation or SITE Program.
In November 1986 the NATO-CCMS formally adopted a U. S. proposal for a new pilot
study entitled "Demonstration of Remedial Action Technologies for Contaminated Land and
Groundwater." The following NATO countries opted to participate:
0 Canada
0 Denmark
0 Federal Republic of Germany
0 Greece
0 Italy
0 The Netherlands
-172-
-------
0 Norway
0 Spain
0 United States
Two non-NATO countries, Australia and Japan have also, expressed an interest in
participating.
The purpose of this new study will be to field demonstrate and evaluate new
technology and/or existing systems for remedial action at uncontrolled hazardous waste
sites and is a logical international extension of the U. S. EPA, SITE program. This
study will offer the potential to obtain a multiple data base on various remedial
action unit processes, that is microbial degradation, on site treatment/destruction
fixation, without any single country having to commit a disproportional amount of its
internal resources to a specific research activity.
Simultaneously, along with the primary demonstration portion of the study, the
opportunity for long term technology transfer of environmental restoration technology
development will be provided to participating countries.
INTRODUCTION
The first working group meeting was
hosted by the Federal Republic of Germany
on March 16-20, 1987. At this meeting
experts from six countries (Canada,
Denmark, Federal Republic of Germany, the
Netherlands, Norway, and U.S.A.) agreed
on an initial program of 12 demonstra-
tions of technologies to clean-up con-
taminated soils and groundwater.
The sites were selected by a
majority vote of all participating
countries (no country was permitted to
vote for its own sites) after agreement
that each interested country would have
at least one site automatically accepted.
With this in mind, every attempt was made
to maximize diversity within the limits
of the 21 sites from which selections were
made. Special emphasis was placed on the
diversity and innovative aspects
of the technologies. Each project is
expected to have a substantial data base
on its application within the 1987
calendar year.
The International Study Group will
meet twice yearly to review progress on
each demonstration project and to consider
the inclusion of further projects in the
program. The results of the various pro-
jects will be published in a final "state-
of-the-art" report on treatment technolo-
gies for contaminated land and groundwater
by Plenum Press as an official NATO
report. Interim reports will be published
in international journals from time to
time as available data warrants.
Discussion:
Table 1 is a summary of the technolo-
gies accepted for the first year of the
study classified by country.
Further details on the sites can be
obtained from the individual referenced
sources. It is important to note that
substantial results are not yet available
on any of the sites as,the overall study
(NATO-CCMS) was only approved in Novem-
ber of 1986 and the sites selected in
March, 1987. Twelve site remediation pro-
jects from six countries were selected for
the first phase of this multi-year infor-
mation exchange project. The projects can
-173-
-------
TABLE 1. NATO-CCMS Pilot Study on Demonstration of
Remedial Action Technologies for Contaminated
Land and Groundwater
Country
Canada
Denmark
Germany
Japan
Netherlands
United States
Summary of Technologies
0 Groundwater treatment
0 Groundwater treatment and aerobic/
anaerobic biological treatment
0 Groundwater treatment, cleaning, and thermal treatment
0 Thermal treatment followed by calcination
and mercury recovery
0 Thermal treatment in consort with
biological treatment
0 In-situ biorestoration
0 Landfarming
0 Electric infrared incineration
0 In-situ biodegradation
0 In-situ soil vapor extraction
be grouped by the type of technology into
the following:
0 4 in-situ biological treatment
(Denmark, Germany, Netherlands
and U.S.A.)
0 3 thermal treatment (Germany,
Netherlands and U.S.A.)
0 2 groundwater treatment (Canada,
Denmark and Germany)
0 1 high pressure soil washing
(Germany)
0 1 land farming (Netherlands)
0 1 soil vapor extraction (U.S.A.)
0 1 chemical treatment (Japan)
Some of the sites have more than one
technology to be demonstrated.
Canada(1)
Canada has one site included in the
study. The Ville Mercier site is in a
small community located on the south shore
of the St. Lawrence River, about 20 km
southwest of the city of Montreal. Ground-
water is a significant source of supply for
the potable water needs of the local popu-
lation.
From 1968 to 1972, some 40,000 m3
of waste oils and liquid industrial wastes
from chemical and petrochemical industries
in the Montreal area were dumped into a
lagoon in an old gravel pit located several
kilometres southeast of Ville Mercier. In
October 1971, it was discovered that
-174-
-------
several wells in the vicinity were
contaminated. The dump site was closed
in 1972, and it was estimated that
20,000 m3 of wastes remained in the
lagoons. Sampling conducted by the
Quebec Ministry of the Environment
(MENVIQ) revealed the presence of phenols
and chlorinated organic compounds in the
groundwater in the area. Subsequent
studies indicated that approximately
30 km2 of the aquifer had become contam-
i nated.
Four specific zones of contamination
were delineated. The first two, where
the groundwater was highly contaminated,
occurred within 2 km of the source. More
than 80 organic compounds were identified
in zone 1. The principal contaminants
were phenols, trichloroethylene, dichlor-
ethane, trichloromethane, trichloroethane,
chlorobenzene, dichloroethylene and
Arochlor 1254. Two other zones of lesser
contamination were delineated with a
total areal extent of contamination
estimated at about 30 km2.
In addition to the inconvenience
caused to the population and the loss of
a water resource to thousands of people,
the direct cost in this case of ground-
water contamination may reach ten million
dollars (Cdn) or eight million U.S.
A program to rehabilitate the aqui-
fer was started by the Government of
Quebec in 1981. The liquid material
which was stored in the lagoon was first
removed (incineration, landfill) in order
to prevent additional contamination of
the groundwater. Following this, MENVIQ
awarded contracts to develop a purge-well
system and to design and construct a
groundwater treatment facility.
The pumping system consists of three
(3) extraction wells located a few hun-
dred meters downstream of the hazardous
waste dump site. These wells create a cone
of depression into which contaminated
groundwater is drawn, i.e., it forms a
hydraulic trap;
Denmark(2)
Denmark has one site included in the
study. The Skrydstrup special deposit site
was formerly a gravel pit which was later
used for the disposal of chemical/indus-
trial waste. Some of this waste was stored
in several hundred drums at the site.
There has been extensive chemical, geologi-
cal and hydrogeological investigations
conducted at Skrydstrup since 1986. Tri-
chloroethane, trichloroethylene, paint and
acid wastes have resulted in groundwater
pollution. The remedial action will con-
sist of digging up, removal and off-
site disposal of several hundred drums,
followed by groundwater pumping and treat-
ing and on-site aerobic/anaerobic
biological treatment of the polluted soil.
This remedial action is active at the
current time.
The remedial action program
of the Danish National Agency of Environ-
mental Protection will involve 4 actual
sites where remedial actions are enforced
by the environmental authorities and is
budgeted at 3,890,000 DKK or about $220,000 U.S.
Federal Republic of Germany(3)
The FRG has 3 sites included in the
study, they are as follows:
0 The Pintsch-uil Site(3) - is located
in an industrial estate of West Berlin.
The recovery of used oils has been carried
out since 1924. Residual products were
left in several pits and subsequently
seeped into the soil arid groundwater.
After initial inspection of the company in
1976 the water authorities in charge
directed that groundwater samples
-175-
-------
be taken. Based on the considerable
pollution detected in the groundwater, it
was directed that a supplemental project
be initiated with the objective of ground-
water sanitation.
In the autumn of 1984 a pilot plant
to purify the contaminated groundwater
was planned. The objective of this
plant was to compile techniques of
groundwater purification adapted to the
special contamination detected here.
The construction of this plant was
entrusted to Messr. Harbauer, Engineer-
ing Office for Environmental Technology,
Berlin.
At first the site was thoroughly
investigated and then 3 clearing wells,
having a maximum capacity of 40 m3/h at
25 m WC, were first designed.
A preliminary pollutant analyses of
groundwater samples indicate extensive
contamination with a wide variety of
organic compounds. Results showed
concentration for:
hydrocarbons
oils
phenols
up to
up to
up to
16,000 mg/1
1,000 mg/1
225 mg/1
In addition to the dissolved and undis-
solved oils and their compounds,
groundwater contamination is primarily
caused by volatile organic compounds
(VOCs), and chlorinated hydrocarbons.
The analytic measuring values were the
basis for the construction of a pilot
plant to purify ground and seeping water
on site.
This pilot plant has a capacity of
40 m^/h and consists of oil separator,
flotation, counter-current desorption
with exhaust purification and active
carbon purification.
Another major part of this remedial
activity will involve full-scale
cleaning of the contaminated soils
scheduled to begin in March 1987. No
data are currently available on this part
of the clean-up effort.
0 The scrap metal site in the Chariot-
tenberg area of Berlin^3/ is contami-
nated with cyanide, arsenic, cadmium,
mercury, PCB, lead, volatile chlorinated
hydrocarbons, and oil. Remedial operations
have been active since December, 1986 and
consist of high pressure soils washing.
A Netherlands company, Klockner Oecotec
GMHB, has used this high pressure soils wash-
ing process to clean 100,000 tonnes of
contaminated soils and removed over 95% of
the pollutants. High pressure water jets
are used to physically strip off hydrocarbons
and heavy metal residues from the soils.
This process is particularly promising for
sandy/gravelly soils but is less effective for
clays.
0 The coke oven site at Unna-Boenen,
Notherhine, Westphalia is contaminated with
aromatic hydrocarbons , tars and acid
resins. Volatile organic compounds comprise
5% of the waste, by weight. The site is
estimated to be 230,000 square meters in
area. It represents a number of similar
sites throughout the FRG.
A variety of technologies will be
applied to the site, individually focused
on the soil and aqueous phases of the
contamination. A 50 ton per hour,
transportable thermal unit will be used to
process the soil and redeposit it on-site.
The unit has been successfully pilot tested
on similarly contaminated soils at a 7 ton
per hour rate with a destruction effectiveness
of approximately 98%. In addition, in situ
enhanced indigenous microbial degradation
is underway, using nitrate as an electron
acceptor. For contaminated water, a
-176-
-------
combination of air stripping, floccula-
tion, sedimentation and filtration are
under consideration.
The Netherlands(4)
The Netherlands has 3 sites included
in the study, which are as follows:
0 The contaminated soil at the former
gasworks site in the Province of Zurd,
Rotterdam, will be excavated and trans-
ported to the thermal destruction in-
stallation of Ecotechniek. The soil is
contaminated primarily with polynuclear
aromatics (PNA's) and complexed cyanides
(ferri-ferro cyanides) caused by spillages
and dumping of waste materials. The in-
stallation of Ecotechniek B.V. has been
selected for evaluation because it has the
highest 'production1 of cleaned soil in the
Netherlands (over 300,000 m3). Although
the installation has proven to be appli-
cable in that it removes and destroys
many types of contaminants (oil, PNA's,
aromatics, cyanides) from soil, there is
a need to assess more accurately the
relation between type of contaminated
soil, process conditions and treatment
results (including air emissions).
The installation consists primarily
of an internally heated rotating kiln
(direct heat transfer) and an after-
burner for the off gases; these operate
at a maximum temperature level of 550°C
and 1100°C respectively.
Main purposes of the demonstration
project are:
a. Evaluation of the treatment re-
sults of the above described
thermal installation for cleaning
of soil from a former gasworks
site.
b. To obtain insight into the start-
up and reliability.problems at
or about $115,000 U.S. soil
treatment plants in a general
sense.
The evaluation which is planned
to start in May 1987. The evaluation
follows the approach developed by ('A Standard
Method for Evaluating Soil
Decontamination Techniques - a First
Outline'). It is planned to evaluate the
installation during periods of two days each,
in which fluctuations in process conditions
and type of contaminated soil should be
kept at a minimum.
Three periods will be evaluated:
1. Normal (or standard) process
conditions and moderately contam-
inated soil
2. Normal process conditions and
highly contaminated soil
3. Deviating process conditions (high
temperature) and moderately contam-
inated soil.
In all cases use is made of soil with a
rather high peat (humus) and clay
content.
During the evaluation attention
will be given to:
0 Gaseous emissions (CxHy, HCN, S02, NOX,
dust)
0 Characteristics of the contaminated soil
(feed) and the cleaned soil (product),
such as particle size distribution,
free-CN, total-CN, PNA's, aromatics.
Depending on the homogeneity of feed and
product, 8 to 20 soil samples will be
analyzed. The soil samples will also be
subjected to leaching tests and bio-
assays.
° The Asten Site in the province of Noord-
Brabant is a petro (gasoline) station,
where the soil is contaminated with petro
containing small amounts of lead, and a
small quantity of diesel oil. The contami-
-177-
-------
nation is caused by a leaking tank of the
petrol station. At least 30,000 liters of
normal gasoline have leaked and about 400
cubic meters of soil have been contami-
nated.
This project is aimed at optimizing
the treatment of deeper layers of con-
taminated soil by enhancing in situ
microbial activity. The development of
in situ techniques is important for the
reclamation of other sites similar to
gasoline stations. It is expected that
the use of this technique will result in
cost reduction since it will avoid costs
for excavation and allow the petrol
station to continue its operation.
The project is being carried with
the cooperation of the RIVM (National
Institute of Public Health and Environ-
mental Hygiene) and the TNO (Netherlands
Organization for Appled Scientific
Research). Funding support is provided
by the Ministry of VROM (Housing, Physical
Planning and Environmental Hygiene).
0 The third site submitted by the
Netherlands is the Wijster site in the
province of Drenthe. This site is the
former location of a waste disposal
company. The primary emphasis will be
on improving landfarming methods for the
bio-destruction of gasoil, crude oil and
halogenated hydrocarbons (for example,
hexachlorocyclohexane).
Contaminated soil will be excavated
and spread over a drained sand bed. The
sand bed will be isolated from the sub-
soils by a plastic membrane. Nutrients
Will be added and the contaminated soils
will be covered by an oxygen permeable
plastic membrane or plants to prevent
erosion.
Japan has one site included in the
study. Because no representative from
Japan attended the March site selection
meeting, this activity must still be
considered tentative. Until confirmation
is received from the Japanese Environ-
mental Agency - National Institute of
Environmental Studies - discussion on
their site will be limited to reporting
that it's a former electro-chemical
industry site where the remedial action
will involve thermal/ chemical treatment and
recovery of mercury in contaminated soils.
United States(6)
The U.S.A. has three sites included in
the study. They are as follows:
0 Peak Oil, Tampa, Florida. The Peak
Oil site is a former facility used by an
oil re-refiner. A former waste lagoon
contaminated with oil re-refining waste
containing PCB's and lead is contaminating
the local groundwater supply. The facility
has been inactive. In 1984, investigation
began to characterize the type and extent
of cleanup.
Remedial investigations conducted at
the Peak Oil site determined that the waste
lagoon was a source of PCB and lead con-
tamination. Because of contamination to
the local groundwater supply, it was
necessary to remove the source of contami-
nation. The soil in the area is sandy and
the groundwater table is very near the
surface. The lagoon has been drained of
water and sand, soil and lime have been
mixed with the lagoon contents to further
adsorb any water and neutralize the acidic
waste. The lagoon was contaminated with
acetic acid that was used in the re-
refining process. These materials were
mixed in place by a bulldozer to obtain a
homogeneous, dry material. The level of
PCB's measured was less than 50 ppm. Lead
has been measured at 10-15 ppm.
Based upon the waste and the oil-soil
matrix present at the site, thermal de-
destruction technology was selected for the
-178-
-------
cleanup of the waste lagoon. In May of
1986 the Shirco Infrared Systems pilot
incinerator was brought to Peak Oil for
a test burn of the PCB contaminated
soil. This test indicated that the
Shirco infrared technology was an
acceptable thermal destruction techno-
logy for the waste present at this site.
The infrared incineration system
consists of a continuous conveyor belt
furnace with associated material feed
and discharge systems, process control
and instrumentation equipment, emission
control systems, and heating element
power center. This system is mounted on
transportable trailers. Hazardous
material is conveyed to the furnace as
sludges or solid wastes. The feed
material drops onto a metering conveyor
located at the feed end of the furnace.
The metering belt is synchronized with
the furnace conveyor screw which dis-
tributes the material across the width
of the metering belt. Material enters
the furnace through a rotary airlock.
The material then moves through
the furnace where it is exposed to
infrared radiation in multiple
temperature-controlled zones. Zone
temperatures are controlled by vary-
ing the input power (electric) to
maintain preset zone setpoint tempera-
tures. As the material moves through
the furnace on the belt, it is exposed
to the thermal environment necessary for
oxidation of volatiles and solid
organics. This technology is primarily
applicable to solids, sludges and con-
taminated soils. The technology is also
applicable to both organic and inorganic
waste streams.
0 Eglin Air Force Base. This jet fuel
spill site is near the city of Fort
Walton Beach in Florida. In 1984 a
large area of dead grass appeared in the
base petroleum storage area. The smell
of fuel indicated that an underground fuel
leak had occurred. Pressure testing of
underground fuel lines revealed the source
of the spill, a one liter per minute leak
in a 15 cm diameter pipeline. The duration
of the leak is unknown; however, soil and
groundwater sampling indicates that as much
as 100,000 liters of jet fuel have contami-
nated 6000 - 8000 cubic meters of soil and
shallow groundwater.
After the spill was discovered, local
authorites characterized the extent of
contamination and established a series of
shallow trenches to recover the free
product. During 1985 over 30,000 liters,
of jet fuel were recovered using skimmer
pumps. In October of 1986, the Air Force
Engineering and Services Laboratory initi-
ated a full-scale research project at the
site to study enhanced biodegradation
methods for removing fuel residuals from
soil and groundwater.
Surface and underground spills of jet
fuel are the most common source of soil and
groundwater contamination at U. S. Air
Force bases. As a result, the Air Force
Engineering and Services Laboratory is
investigating cost effective methods for
decontaminating fuel saturated soils and
impacted groundwaters. In-situ biodegra-
dation provides a method of eliminating
fuel residuals from the unsaturated zone
and removing a long-term threat to the
groundwater. ,
The degradation of petroleum hydro-
carbons has been extensively studied over
the last 25 years" and is well understood
in the laboratory. However, extensive data
on full-scale field studies is seldom
available or published and much of the
experience rests with experts who have a finan-
cial interest in the technology. The goals
of this site demonstration will be to
generate new data on the in-situ degrada-
tion of specific jet fuel components, to
optimize nutrient/oxygen additions, and
-179-
-------
to maximize contact between nutrients,
soil microbes and fuel soaked soils.
The study will be conducted over a two-
year period to provide information on the
lower limits of biological removal of
fuel compounds. A control area has been
established in a fuel contaminated area
on the upgradient end of the site. The
purpose of this control area will be to
compare natural levels of degradation to
the enhancement provided by nutrient/
oxygen additions.
The total cost of this site demon-
stration is $780,000. Of this total,
approximately 50% is for design and oper-
ation of the system, 25% is for sampling
and analysis, 15% is for nutrients and
hydrogen peroxide and 10% is for well
installation and delivery equipment.
More precise cost information will be
available when the project is fully
operational.
0 Vernona Well Field. The Verona
Well Field site consists of several
distinct contaminated areas within
approximately 100 acres. The well field
itself contains 30 production wells that
supply the entire city of Battle Creek,
Michigan, including several major busi-
nesses. The site also includes a rail-
road marshalling yard and two solvent
facilities. The Thomas Solvent Raymond
Road (TSRR) facility is a former solvent
repackaging and distribution facility.
Solvents were stored in 21 underground
storage tanks which were later discovered
to be leaking. The TSRR facility is
located about one mile upgradient of the
well field. It is primarily a residen-
tial area surrounded by a few businesses.
The contamination problem at the
Verona Field site was first discovered
in August 1981, during routine testing
of the city's water supply. Volatile
organic compounds (VOCs) were discovered in
10 of the production wells. An area-wide
survey by the U. S. EPA Technical Assis-
tance Team (TAT) in the Spring of 1982
determined that the TSRR facility was a
potential major source of well field con-
tamination. This was confirmed during
remedial investigation activities. Chlor-
inated hydrocarbons are the most signifi-
cant environmental contaminants.
Groundwater and soil contamination within
the TSRR facility was found with VOCs as
high as 100,000 and 1,000,000 ppb respec-
tively. The total estimated mass of
organics in groundwater and soil at TSRR
was 440 Ib. and 1,700 lb., respectively.
A two-stage approach to remedial
action at the TSRR facility was adopted.
Each assembled remedial alternative evalu-
ated included separate but related alter-
natives for groundwater and soil. The
selected alternative for the site includes
a groundwater extraction (GWE) system in
conjunction with a soil vapor extraction
(SVE) system. Due to the significant
mass of contaminants in the soil, alter-
natives that employed both groundwater and
soil remediation were developed. Several
alternatives for soil cleanup were evalu-
ated, including SVE, excavation with
on/off-site disposal, site capping, soil
washing (saturating the unsaturated zone
with water extraction), and no action.
The key components of the SVE system
are extraction wells screened primarily in
the unsaturated zone and a vapor phase
activated carbon treatment vessel. The
entire system will be closed (piping will
lead directly from the extraction wells to
the carbon system). A performance goal of
10 ppm total VOCs in the unsaturated zone
was established. At present, eight ex-
traction wells are envisioned. Due to the
innovative nature of the SVE system, the
U.S. EPA has adopted a formal two-step
procurement approach for both SVE design
and construction. Technical proposals
for the design will be solicited, then
cost proposals will be evaluated only
-130-
-------
for those bidders submitting acceptable
technical proposals. However, at a
minimum, five extraction wells followed
by vapor phase carbon treatment will be
part of the system.
Summary and Conclusions
The NATO-CCMS pilot study program
provides an excellent forum for an
international exchange of current,
state-of-the-art, scientific information
pertinent to environmental protection.
One of the challenges in modern society
is dealing with the byproducts resulting
from manufacturing. Throughout the
world, improper handling and disposal
of hazardous materials and hazardous
waste has caused both groundwater and
soil contamination. This is especially
true in industrialized countries. Water
and land are scarce resources, and must
be returned to productive use. Current
cleanup efforts are hampered by limited
disposal sites, and high cost. It has
not been proved that land disposal,
particularly landfill ing of hazardous
waste, is effective over the long term in
containing hazardous wastes. Immediate
solutions must be found and applied to
the most serious pollution problems.
Promising new treatment technologies must
be tested and then demonstrated to see if
they apply to, and will be effective for
today's pollution problems. We must
build a knowledge base so that more
efficient, cost effective remedial
technologies can be developed. This
knowledge base must include "lessons "•>
learned" from the past. A knowledge of
technical and economic limitations, or
failures, of the various technologies is
as beneficial as the details of
successes. Through the exchange of
information on emerging remedial
technologies, nations may pool their
knowledge and experience and make the best
use of their own limited resources.
Acknowledgements
1. The plenary meeting group of the NATO-
CCMS for approving the study. They
were as follows:
0 Canada
0 Denmark
0 Federal Republic of Germany
0 Greece
0 Italy
0 Netherlands
0 Norway
0 Spain
0 United States
2. Special recognition to Mr. Allen
Si el en, NATO-CCMS Coordinator for
the Office of International Activ-
ities, U. S. EPA, Washington, DC
for his consultation and assistance
in getting the study adopted.
3. Mr. A. James Barnes, Deputy Adminis-
trator, U. S. EPA for chairing the
U. S. Delegation at the time of
adoption of the Study.
4. All the members of the first working
group-site selection meeting for their
presentations and technical input.
They are in alphabetical order as
follows:
Jan Willem Assink, TNO, P.O. Box
342, 7300 AH Apeldoorn, The
Netherlands.
Danna Borg, Ministry of Environ-
ment, Strandgade 29, Denmark.
Dr. Heinz-Jurgen Brauch, Engler-
Bunte-Institute of the Univer-
sity of Karlsruhe, D-7500
Karlsruhe 1, Richard-Willstatter-
Allee 5 Federal Republic of
Germany.
Winfried Brull, Klockner Oecotec,
4100 Duisburg, Nendorfer Str
3-5, Federal Republic of
-181-
-------
Germany.
Douglas C. Downey, USAF Engi-
neering and Service Center,
AFESC/RDV Tyndall AFB, FL
32403, USA.
Schmid Ernst, LFU, Benzstrasse
5, Germany.
Dr. Jurgen Fortmann, Ruhrkohle
Umwelttechnik, 4300 Essen
15, Federal Republic of
Germany.
Dr. Volker Franzius, Umwelt-
bundesamt, Bismarckplatz 1,
D-1000 Berlin 33, Federal
Republic of Germany.
Peter Fuhrmann, Landesanstalt
fur Umweltschutz, Gries-
bachstr 3, Federal Republic
of Germany.
Dr. Heimhard, Hans-Jurgen,
Klockner Oecotec Gmah, D 41
Duisburg, Neudorfer Str. 3-5,
Federal Republic of Germany.
Morten Helle, State Pollution
Control Authority, Box 8100
DEP, N-0032 OSL01, Norway.
Morten Hinseveld, TNO-Apeldoorn,
Box 342, Apeldoorn, Holland.
Stephen C. James, U. S. Environ-
mental Protection Agency,
26 West St. Clair Street,
Cincinnati, OH 45268, USA.
Walter W. Kovalick, Jr., U. S.
Environmental Protection
Agency, 401 M Street, S. W.,
Washington, DC 20460, USA.
Manfred Nussbaumer, Ed. Zublin
AG, albstadt Weg 3, 7000
Stattgart 80, Federal Republic
of Germany.
Joachim Ronge, Ruhrkohle Umwel-
ttechnik, Rellinghauser Str.
1, D-4300 Essen, Federal
Republic of Germany.
Jim Schmidt, Environment-Canada,
Wastewater Technology Center,
867 Lakeshore Road, burlington,
ONT, L7R4A6, Canada.
Dr. Seng, Hansjorg, Land Baden-
Wurttemberg, 75 Karlsruhe,
Griesbach Str. 3, Federal
Reppublic of Germany.
Geroges Simard, Environment -
Quebec, 3900 rue Marly, Quebec,
Glx4E4, Canada.
Michael Alan Smith, (HMRC), Bostock
Hill & Rigby, 288 Windsor Street
Birmingham, B74DW, UK, (o)
021-359-5951 (Birmingham).
Esther'Soczo, RIVM, P.O. Box 1,
3720 BA, Bilthoven, The
Netherlands.
Klaus Stief, Umweltbundesamt,
Bismarckplatz 1, D-1000 Berlin
33, Federal Republic of Germany.
Becker Thomas, Ministry of Environ-
ment, Denmark, Hogstorgdell 25,
2100 0, Denmark.
Peter Walter Werner, DVGW-forschung-
sstelle am Engler-Bunte-Institute
der Universitat Karlsruhe,
D 7500 Karlsruhe 1, Federal
Republic of Germany.
References
1. Schmidt, J., 1987 "Aquifer Decontam-
ination for Toxic Organics - The
Case Study of Ville Mercier, Quebec,
Canada," Wastewater Technology Centre,
Environment Canada, P.O. Box 5050,
Burlington, Ontario, Canada L7R 4A6.
2. Borg, D., 1987 "Skrydstrup Special
Deposit, Province of Ribe," Ministry
of Environment, National Agency of
Environmental Protection, Strandgade
29, 1401 Copenhagen K, Denmark.
3. Werner, W., 1987 "The Sanitation of
the Pintsch Site," Harbauer & Co.,
Ingenieurboro fur Unwelttechnik,
Bismarckstrasse 10-12, 1000 Berlin
12, Federal Republic of Germany.
Site 1 - Pintsch - nil
-182-
-------
Site 2 - Scrap metal site at
Charlottenberg Heimhard,
Klockner-Decotec GmDH,
Nuudorfer Str. 3-5
D-4100 Duisburg 1, Federal
Republic of Germany
Site 3 - Coke oven site at
Unna •? Boenen, Northrhine,
Westphalia
Krauss, Ruhrkohle Aktiengesell-
schaft, Abteilung P7, Postfach
10 32 62, D-4300 Essen 1,
Federal Republic of Germany
4. Site 1 - Gasworks Site, Province
of Zuid, Nedtherlands Assink J.W.,
TNO, P.O. Box 342, 7300 AH
Ape!doom, The Netherlands
Site 2 - Asten Site, Province of
Noord-Brabant, Netherlands Soczo
E. RIVM, P.O. Box 1, 3720 BA,
Bilthoven, The Netherlands
Site 3 - Wejster Site, Province
of Drenthe Soczo E. RIVM, P.O.
Box 1, 3720 BA, Bilthoven,
The Netherlands
5. Gotoh S. and Ikeguchi T., National
Institute for Environmental Studies,
Japan Environmental Agency Yatabe-
machi, Tsukuba, Ibaraki 305 Japan.
6. Site 1 - Peak Oil, Florida. James
S., U. S. EPA, 26 W. St. Clair St.
Cincinnati, Ohio 45268 USA.
Site 2 - Eglin Air Force Base, Florida
Downey D. C., USAF Engineering and
Service Center, AFESC/RDV, Tyndall
AFB, Florida 32403 USA.
Site 3 - Verona Well Field, Battle
Creek, Michigan. James S. U. S. EPA,
26 W. St. Clair Street, Cincinnati,
Ohio 45268 USA.
-183-
-------
REACTIVITY OF VARIOUS GROUTS TO HAZARDOUS WASTES AND LEACHATES
Andrew Bodocsi and Mark T. Bowers
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
A laboratory study was conducted to evaluate the potential of selected grouts for
controlling the percolation of leachates from hazardous solid waste landfills or hazardous
waste ponds. In the course of the study, seven different grouts were subjected to perme-
ability tests and three of the grouts were tested for their reactivity by an immersion
type test. Eight different chemicals, some with two concentrations, and two real-site
wastes were used as permeants in the permeability tests, and as liquids for the immersion
baths.
Of the seven grouts, the acrylate, cement-bentonite (mix 2), and urethane grouts had
the lowest baseline permeabilities with water, ranging from 2.3 x 10~10 to 3.6 x 10"9
era/sec.
During permeability testing with chemicals, the acrylate grout exhibited excellent
resistance to the paint and refinery wastes, 25% acetone, 25% methanol, and sodium hydrox-
ide, performed satisfactorily With cupric sulfate; ethylerie glycol, and xylene, and was
seriously damaged by aniline, 100% acetone, hydrocholoric acid, and 100% methanol.
The permeability of the cement-bentonite (mix 2) grout was tested with acetone,
aniline, cupric sulfate, hydrochloric acid, inethanol, arid sodium hydroxide. With every
one of these chemicals the permeability of the grout improved, ultimately reaching a
practically impervious state.
The urethane grout maintained its low permeability with acetone, aniline, ethylene
glycol, methanol, paint waste, refinery waste, and hydrochloric acid and it performed
marginally well with cupric sulfate. However, the urethane lost its low permeability
with sodium hydroxide and xylene.
Based on the comparision of permeability and reactivity test results, a scheme was
proposed to correlate the permeability changes of grouts to the weight and consistency
changes that may occur during their reactivity testing.
INTRODUCTION
One of the major environmental prob-
lems facing the nation is the threat of
contamination of groundwater from leaking
hazardous waste landfills and leachate
ponds. If a waste site is underlain by an
impervious stratum, the most cost-effective
remedy may be using a cutoff slurry wall
constructed around the site and keyed into
the aquielude. However, if there is no
impervious stratum below the waste, the
remedy may be the construction of a bottom
seal created by injection grouting, in
conjunction with a vertical slurry wall.
Alternately, both the bottom seal and side
wall may be made by injection grouting.
Injection grouting has been used for
many years for stabilizing soils, to pro-
-184-
-------
vide cutoff curtains under dams, for
stabilizing tunnels, and more recently it
was proposed for containment of hazardous
waste (2,3).
When constructing an impervious bar-
rier by injection grouting under a waste
site, the grout must thoroughly penetrate
the soil. After .expelling the waste from
the voids the grout must set and harden in
the soil. In addition, the hardened grout
must provide a durable impervious seal;even
when permeated by hazardous leachate.
The purpose of this study was to test
the permeability and reactivity of selected
grouts with ten chemicals to determine!if
the grouts had the durability to withstand
the typical hazardous waste site environ-
ment, arid thus could be considered for
horizontal seal construction.
The permeability test results indicate
that certain grout-chemical combinations
caused the deterioration of the permeabil-
ity of the grout, while others resulted in
; little or no detrimental changes. The
reactivity test results show what effect
chemical baths had on grout samples; some
combinations caused weight gains, others
weight losses, and still others caused no
changes. In addition, changes in sample
consistency were observed. The combined
analysis of the two tests resulted in
correlations that allow prediction of
permeability changes from reactivity test
results. ;
MATERIALS AND METHODS
All permeability tests and most
reactivity tests were conducted on grouted
soil samples. The permeability testing
was conducted in specially constructed,
permeameters and safe environmental boxes,
using selected chemicals as the permeants.
The reactivity samples were tested by
their immersion in selected chemical biaths.
Grouts
In the typical batch of the cement-
bentonite (mix 1) grout 3000 g Type III
cement, 120 g bentonite, and 6000 ml water-
was used, resulting in a water-cement ratio
of 2.0. The batches of cement-bentonite
(mix 2) grout were made up of 3000 g MC-500
microfine cement, 120 g bentonite, 30 ml
dispersant, and 2250 g water, yielding a
water-cement ratio of 0.86. For ease of
injection, pea gravel was used as the soil
with both grouts.
The sodium silicate grout selected was
SIROC 132, distributed by Raymond Inter-
national, Inc. It consisted of 60% modi-
fied silicate, 25% water, 10% formamide,
and 5% calcium chloride. A fine Mason's
sand was used as the soil to avoid synere-
sis. Later extensions to the work included
a glyoxal-modified sodium silicate grout
and a sodium aluminate-modified sodium
silicate grout in order to reduce the
permeability of these grouts.
The urethane grout selected was CR360,
a product of the 3M Company. A mixture of
89.2% water, 5.7% CR361 (gel inhibitor),
and 5.1% CR360 (urethane polymer solution)
was chosen. A silica sand was used with
this grout.
AC-400, distributed by Avanti Inter-
national, was chosen to represent the
acrylate grouts. A mixture of 73.44%
water, 24.99% AC-400, 0.74% triethanolamine
(catalyst), 0.74% ammonium persulfate (ini-
tiator), and 0.074% potassium ferricyanide
(inhibitor) was used. Silica sand was used
as the soil.
Grouting Procedure
The soil samples for the permeability
tests were pressure grouted in a 75 mm
inside diameter plastic mold. Sample
lengths could be varied as required. At
least two pore volumes of grout were pass-
ed through the samples before completing
the grouting process.
The 25.4 mm diameter by 25.4 mm high
cylindrical reactivity samples were prepar-
ed in acrylic molds by first placing the
sand in the molds and tamping the surface
to densify it. The grout was then injected
continuously from the bottom of each sample
using a syringe. The cement-bentonite
grout was poured into the molds, since
these samples were prepared without a soil.
Permeameters
The permeameters used were flexible-
wall type, specially built in-house to re-
sist the chemicals and grouts. The ele-
ments in contact with the chemicals were
made of one of the following: nylon,
stainless steel, PVC or teflon.
-185-
-------
Permeability Measuring Apparatuses
Research Facility (1).
The percolation of the leachates
through the grouted soil samples was
performed in two permeability apparatuses
that were housed in vented boxes utilizing
a one-pass air system to remove any
hazardous vapors. They had a total
capacity of thirty testing stations.
Each station had regulated chamber and
driving pressures, and means to measure
either leachate outflow only, or both
inflow and outflow.
Permeability Testing
Since the main objective of the
study was to investigate the effect of
the various chemical leachates on the
permeability of the grouts and compare
them with permeabilities established with
water, the first step was to permeate the
samples with deionized water until
equilibrium baseline permeabilities were
reached. After that, the samples were
permeated with their respective chemical
leachates until an equilibrium permeability
was established and a minimum of 2.0 pore
volumes of chemical had passed through
each sample.
Reactivity Testing
ASTH Standard C267-82 "Standard Test
Method for Chemical Resistance of Mortars,
Grouts and Monolithic Surfacings" was
followed as guidance for the preparation,
weighing, measuring and immersion of the
specimens in the selected chemicals.
Observations and weighings were made
after 1, 7, 14, 28, 56 and 84 days of
immersion. Plots were made illustrating
the percent change in weight versus time
for each sample and replicate series.
Chemicals and Hazardous Leachates Used
For this study, ten representative
chemicals were selected. Eight of these
organic and inorganic compounds represent
a particular functional group as set
forth in (3), and consisted of acetone,
aniline, cupric sulfate, ethylene glycol,
hydrochloric acid, methanol, sodium
hydroxide and xylene. In addition, a
real-site refinery waste and a real-site
paint waste were included in the testing
program, which were obtained as leachates
from lysimeter studies conducted at the
Center Hill Solid and Hazardous Waste
The chemical concentrations used
were: acetone (100% and 25%), aniline
(100%), cupric sulfate (20% and '10%),
ethylene glycol (100% and 25%), hydro-
chloric acid (IN and 4N), methanol (100%
and 25%), sodium hydroxide (25% and
10%), and xylene (100%). The real-site
wastes used were unaltered.
RESULTS AND DISCUSSION
Permeability Test Results With Mater
Three of the grouts tested had very
low baseline permeabilities with water:
acrylate (k = 5.1 x 10"10 cm/sec),
cement-bentonite (mix 2) (k = 2.3 x 10"10
cm/sec), and urethane (k = 3.6 x 10~9
cm/sec). The other four grouts tested had
permeabilities with water that exceeded
1 x 10~7 cm/sec.
Permeability Test Results For Acrylate
Grout With Chemicals
Figure 1 illustrates in a bar chart
form the overall permeability changes of
acrylate grout with ten selected chemi-
cals. On the left vertical axis of the
chart permeability is plotted on a log
scale. Along the horizontal axis the
chart is subdivided into ten boxes, each
representing one of the ten chemicals.
On the right vertical axis the equilib-
rium permeabilities of the acrylate
grout samples with water are indicated.
Each shaded bar represents the average
changes in the permeability of the grout
with one of the chemicals. Each bar
starts at the equilibrium permeability
of the grout with water, and may go up
or down, or remain unchanged, depending
on the reaction of the grout with the
specific chemical. In most cases the
bar first rises to a dashed line that
represents the permeability of the grout
at the first peak of its permeability
with flow. The bar may rise further and
terminate at a permeability level that
corresponds to the final equilibrium
permeability of the grout with the
specific chemical. The number on each
line gives the number of pore volumes of
chanical which flowed through the sample
before it reached the indicated equilibrium
permeability.
As shown in the bar chart, the acry-
-136-
-------
10'
1-4-
42pv
u
QJ
cn
u
10
r6..
JQ
ra
ID
£L
OJ
Q.
, I , . 1
I
1
34DV
I
I
1
^
1
I
p. Spy
//
i
\
>
I
1
\
i
m
Q.2DV
10.8
\
V •
k
j-i
ro
r)
•|H
C.
•rH
- >H
.^^
CT
UJ
Figure 1. Permeability of acrylate grout with various chemicals.
late grout exhibited excellent resistance
to real-site paint and refinery wastes,
and to 25% sodium hydroxide. Its perfor-
mance was satisfactory with 20% cupric
sulfate, 100% ethylene glycol and xylene.
The aniline, 100% acetone, 4N hydrochlo-
ric acid, and 100% methanol were very
detrimental to the permeability of this
grout. The permeability of the grout
decreased with the introduction of 25%
sodium hydroxide, however, since the flow,
volume was only 0.2 pore volume, this
trend may reverse itself with continued
flow, but only after a very long testing
time.
The effects of chemical concentration
on acrylate grout can be observed in
Figure 2. The figure shows pairs of bars
side by side, allowing the comparison of
the effects of concentration of selected
chemicals on the permeability :of the
grout. In the cases of acetone and
methanol t.he effects of the higher chemi-
cal concentrations are very pronounced. ,
On the other, hand, the ultimate perme-
ability of the grout was the same with
both concentrations of cupric sulfate,
and again with both concentrations of
hydrochloric acid. • • •
Permeability Test Results For Cement-
Bentonite (Mix 2) Grout With Chemicals
The way the cement-bentonite (mix 2)
grout behaved with the various chemicals
is summarized in the bar chart shown in
Figure 3. With both water and chemicals,
this grout was the most impervious of
all tested. The introduction of every
chemical caused the permeability of the
grout samples to decrease from their base-
line permeabilities with water. With
very small amounts of flow of between
0.1 and 0.6 pore volumes, the permeabil-
ities of the samples dropped to between
3 x lir11 and 6 x 10"11 cm/sec, or to -a
practically impervious state. Because
of their very low permeability, these '
samples allowed very low flow volumes,
even though they were tested on the
average for more than 120 days. It is
possible that given enough time and flow
these chemicals could increase the perme-
ability of this grout, but it would take
very long testing times.
-187-
-------
sec
c
>*
-M
•iH
t-H
•rH
J3
(O
m
s
c_
CD
Qu
!-»•
°
>
O)
CD
^
V/
-1 1 1 1 1
33pv
\
W/1
r/x
\
20pv
lOpv
1
25p
1
V
t \
o
B.3pv
o >•
30pv
!
i
O
&
&
rj
•rH
c.
cr
UJ
Figure 2. Effect of leachate concentration on
the permeability of acrylate grout.
u
(U
en
10
~6 '
JD
01
CL
O.EpV
O.lpv
0.4pv
O.lpv
O.HpV
O.ipv
c_
(U
JJ
ID
c.
JO
•r-l
Figure 3. Permeability of cement-bentonite
(mix 2) grout with various chemicals
-188-
-------
Permeability Test Results For Urethane
Grout With Chemicals
Permeability Test Results For Other
Grouts With Chemicals~
Figure 4 illustrates in a bar chart
form the overall changes in permeability
of urethane grout with ten selected
chemicals. As shown, the urethane grout
remained quite impervious with the majority
of the chemicals. With 20% cupric sulfate
and 4N hydrochloric acid the grout perform-
ed marginally, as its final permeability
slightly exceeded the 1 x 10~' cm/sec
level. The most detrimental to the
urethane grout were the 25% sodium
hydroxide and xylene. The sodium hydroxide
caused a 4.5 orders of magnitude increase,
raising the equilibrium permeability of
the grout to much above the 1 x 10~7 cm/sec
1evel.
The bar chart in Figure 5 shows the
effects of concentration of selected
chemicals on the final permeability of
the urethane grout. As shown, varying
the concentrations can have varying
effects. For example, with 25% ethylene
glycol the urethane grout samples did not
exhibit increases in their permeabilities,
while the samples with a 100% solution of
ethylene glycol did.
In addition to the above discussed
grouts, cement-bentonite (mix 1), sodium
silicate, glyoxal-modified sodium sili-
cate, and sodium aluminate-modified
sodium silicate grouts were also tested.
However, the baseline permeabilities of
these grouts with water ranged between
1.7 x 10"5 and 3.0 x 10"6 cm/sec, too high
for cutoff construction. Nevertheless,
the grouts were subjected to limited
testing with chemicals and showed rela-
tively small changes in their permeabil-
ities.
Reactivity Tests
The objectives of the reactivity
tests were to observe the weight and
volume changes that small samples of the
various grouts underwent during their
immersion in selected chemical baths,
and to explore if these observations
could be correlated with the corresponding
permeability test results.
In this research the three grouts
tested for their reactivity were acrylate,
u
cu
in
10-
10-
.5 10
CD
CU
e
C-
03
Q.
-8
10
rlO
7.9pv
r*
0.6pv
.0.8
8.7pv
13.2pv
IBpV
20pv
B.5pv..
c.
m
-M
to
t.
. . J3
-. CT
UJ
\«-
Figure 4. Permeability of urethane grout with various chemicals.
-189-
-------
£ 10
tn
-4..
10
-6--
to
03
e
c_
QJ
Q.
10
-S--
8.3pv
DpV
lO.Bpv
17pv
20pv
40pv
c.
QJ
-U
ra
C-
J3
13
cr
LU
Figure 5. Effect of leachate concentration on
the permeability of urethane grout.
ceraent-bentonite, and urethane. The
testing consisted of immersing the grout
samples in selected chemicals'and weighing
them after elapsed times of 1, 7, 14, 28,
56, and 84 days. Plots of percent weight
change versus time were prepared for each
sample and replicate series. In addition,
subjective observations were made and
recorded on the shrinkage, swelling,
spall ing, hardening, softening and
stickiness of the samples. The results
ranged from total disintegration to as
much as a 40X weight increase.
COMPARATIVE ANALYSIS OF PERMEABILITY AND
REACTIVITY TESTS
The results from the two types of
tests were analyzed, compared, and a
scheme was proposed to allow the prediction
of the permeability behavior of grouts
with various chemicals from their behavior
with the same chemicals in the reactivity
tests. This scheme is summarized in
Table 1. In this table, weight and volume
changes in the vertical columns, and
consistency changes in the horizontal rows
are correlated with permeability changes
indicated in the boxes. For example the
urethane samples in a xylene bath under-
went medium (10%, authors' classification)
weight losses, and they also became hard.
This case corresponds to the matrix loca-
tion column 2 and row 2, which reads:
"Permeability increases significantly."
Indeed, going back to Figure 4, it is seen
that the permeability of the urethane
grout with xylene increased almost three
orders of magnitude before it came to an
equilibrium. It is proposed that with
the presented correlations the choice of
the most suitable grout for a site could
possibly be made based on reactivity tests
only, and these could be conducted at the
site of the hazardous waste.
CONCLUSIONS
1. When tested for permeability, the
acrylate grout exhibited excellent to
satisfactory resistance to all chemicals,
except reagent grade aniline, 100% acetone,
IN and 4N concentrations of hydrochloric
acid, and 100% methanol. These caused
increases of several orders of magnitude
in its permeability.
-190-
-------
00
LU
a;
oo
LU
OO
=C
CO
Si
o
o;
UJ
a.
UJ
as
a-
33
>=c
a: UJ h-
il LU 00 -C 1— 1—
E h- =£ C3 :n oo • — •
13 et LU *— ' C5 LU OO
—I _J aS LU i— 1 1— Q
^ C_5 ^£ LU UJ
•=• •> 2: 3 00 LU
LU i-< _l <0
_i oo a: « o
=c < LU 1-1 oo 2: os
-« LU s I— oo i— > a.
5gg~
LU CD 3T t— H-
s: «a: as 3 oo LU
LU I — 1 «=C O
_i as oo <: •> o
=c o —i z oo as
-I -Z. OS I— i— 00 CL.
— »-i LU i— I <£ O
-i i— s: CD _j
-4 — I •
0
>- z
— CO )— 1
LD 00 _1
U O D.
3 _1 00
UJ
jj OO t—
E =C 3: 00
Z) LU CJ3 OO
—I OS >-> O
O O LU — 1
>• LU 3
o
UJ
JJ 00 H-
5 LU 55 —•
O C_> LU C3
~-~ oo /
h- UJ LU /
•J}=>*Z /
Jj 0 3: / 0
3 > e_> / 2= oo
/ LU LU
/ t — CI
/ oo ^
y' oo a:
C +J
ai eu -i-
P W J3
t- a>
t/J •»—
ra i—
O) -i-
0 «3
c: j
^~ C 4-9
(O r—
i. ,. £ ,
•i- C
r— W IO
•i- O) U
J3 tA *i—
IO IO M-
O) Ol -i-
= «- c
t- o en
0) C -i—
O- -r- t/>
to £- t/)
COO)
IO T3 O "U
E «- o «-
OS -C J3 J=
^J
l_>
^" ^)
t- T3 i—
^5 O! i —
IO 4-> IO
£ 0) -i-
£- >*- C
Q) 1- -i-
o. 10 E
£ s"
i|- O) -P O O> r—
O to fl3 5 t/> i —
IO IO J^ (O IO
CJ 5 -P O) E
E t- d> IO £_ -r-
£- o E o c:
}
C -P
0) O) -r-
IO IO IO
t_ O! t-
O = 0) E
t_ E to
O ^
C Q) (rt
i— 1 O-T-
10 >,
O) -^
4J 0
If- •!-
0 =3
oo cr
C -P
r- T-
0) •»- OJ
IO 03 &_
O) 0) «3
i— E ^~
C flj W)
-1 Q.II-
Ul
c >>
01 i—
P f
O r—
^0 U5 :
i — tSi
•r- O)
J3 01 >,
IO IO r—
0). 0) +J
E t- 10
t- O OJ
D c: t-
Q_ *r— Ol
to
O)
1 4->
C IO
.1- t-
to en
•i- O)
a -p ,
-191-
-------
2. With all the chemicals tested,
the permeability of the cement-bentonite
(mix 2) grout decreased from its baseline
permeability with water to a practically
Impervious permeability of approximately
3 x 10"11 cm/sec.
3. The urethane grout remained
quite impervious with the majority of the
chemicals tested, except it performed
only marginally with 20% solution of
cupric sulfate, and IN and 4N concentrations
of hydrochloric acid, and poorly with 25%
solution of sodium hydroxide and reagent
grade xylene.
4. The effects of the concentration
of chemicals on the permeability of grouts
varied. It was found from the limited
data that some chemicals with reduced
concentrations caused smaller increases
in the permeability of a grout than with
higher concentrations, while with other
chemicals, varying the concentration had
no significant effect.
5. From the analysis of reactivity
and permeability test results, a scheme
was proposed that correlates the weight
and consistency changes of the reactivity
samples of a grout immersed in a chemical,
to expected changes in its permeability
when permeated by the same chemical.
This may allow an engineer to make at
least the preliminary selection of suitable
grouts for a site by using reactivity
tests in place of the more costly
permeability tests.
ACKNOWLEDGMENT
The research described herein was
supported by the U.S. EPA Hazardous Waste
Engineering Research Laboratory, Cincin-
nati, Ohio. Appreciation is expressed to
work assignment managers Herbert R. Pahren
and Ronald F. Lewis, and project officers
John Martin and Joseph K. Burkart for their
technical and administrative support, and
to operations manager Gerard Roberto for
his support throughout this project. The
authors also appreciate the work of grad-
uate students Roddy Sherer, Brian Randolph,
Jim Cipoll one and Bruce Hick.
REFERENCES
1. Kinman, R.N., J. Rickabaugh,
J. Donnelly, D. Nutini and M. Lambert.
"Evaluation and Disposal of Waste
Materials Within 19 Test Lysimeters
at Center Hill," EPA-600/2-86-035,
U.S. Environmental Protection Agency,
Cincinnati, Ohio, March 1986.
2. May, J.H., R.J. Larson, P.G. Malone,
J.A. Boa, Jr., and D.L. Bean. "Grout-
ing Techniques in Bottom Sealing of
Hazardous Waste Sites," EPA-600/286-
020, U.S. Environmental Protection
Agency, Cincinnati, Ohio, 1985.
3. Spooner, P.A., G.E. Hunt, V.E. Hodge,
and P.M. Wagner. "Compatibility of
Grouts with Hazardous Wastes," EPA-
600/2-84-015, U.S. Environmental
Protection Agency, Cincinnati, Ohio,
1984.
-192-
-------
ELECTRO-DECONTAMINATION OF CHROME-CONTAMINATED SOILS
Sunirmal Banerjee
University of Washington
Seattle, WA 98195
ABSTRACT
A technique of in-situ treatment of inorganic waste-contamined soils is
being explored at a Superfund site in a current study. Transport of inorganic
ions under an imposed electric field is essentially the basis of this technique.
In this paper, the results of the laboratory experiments conducted on undisturbed
soil samples obtained from the site and the initial results of preliminary field
experiments are reported.
Generally, the laboratory results have shown that with appropriate combina-
tion of applied hydraulic and electric fields, it is possible to remove chromium
at a faster rate by this approach than by hydraulic leaching alone. The pre-
liminary field experiments also show that chromium concentrations can be altered
by electro-kinetic treatment alone.
INTRODUCTION
In the past, electro-kinetics have
been applied in a variety of engineering
projects. Mostly, these applications made
use of the electro-osmotic transport of
the liquid phase. The past applications
include dewatering and consolidation of
soils (3,5), recovery ,of embedded objects
from the ocean floor (4), alteration of
penetration or pull-out resistance of
piles (2), dewatering of sewage sludge
(8), and mine tailings ponds (9) and
dredge spoils (11), and increasing
recovery rates in oil fields (1) among
others. Some previous studies have also
indicated that, from a technical point of
view, electro-kinetic treatment of the
ground may be a feasible approach for
decontaminating waste-contaminated soils
(6,7,10).
However, the practical effectiveness
and feasibility of such a technique have
not been explored. Definitive field-scale
testing is required prior to widespread
application of the technique. This pro-
ject was, therefore, undertaken as a
field-scale study, the scope of which
included a) characterization of the soil-
water electrolyte system at a Superfund
site in Corvallis, Oregon, b) bench-scale
studies and design of in-situ electro-
kinetic experiments, c) performance of a
series of field tests and d) establishment
of some possible relationships between the
extent (and rate) of decontamination and
the important experimental variables based
on the field test results. The project is
only partially complete at present. The
underlying concepts of the treatment and
some of the up to date findings are
reported in this paper.
APPROACH
The processes associated with the
flow of direct-current electricity through.
soil-water electrolyte systems are collec-
tively termed electro-kinetics. These
-193-
-------
include electro-osmosis, electrophoresis,
and other electro-chemical processes.
These processes mainly cause relative
movement of electricity (current flow),
ions (ionic drift), charged particles
(electrophoresis) and of the liquid phase
(electro-osmosis). The major effects of
electro-kinetic processes are movement of
water, ionic drift, ion exchange, develop-
ment of osmotic and pH gradients, electro-
lysis, hydrolysis, oxidation, reduction,
heat generation, gas evolution, formation
of secondary minerals, physical and chemi-
cal adsorption, and many others. The
overall effect is quite complex and the
present-day understanding is not up to the
task of quantitatively accounting for all
of these effects of electro-kinetic pro-
cesses. Nevertheless, the empirical evi-
dence from the past applications leads one
to believe that electro-kinetics may have
some possible applications in waste-man-
agement. A partial list of the poten-
tial applications of electro-kinetics in
waste management is shown in Table 1. The
present project attempts to evaluate the
practical effectiveness of electro-kine-
tics as the primary method of treating
waste-contaminated soils.
TABLE 1
APPLICATIONS OF ELECTRO-KINETICS IN WASTE HftNAGEHENT
-- AS PRIMARY HETHOD OF TREATMENT
Valor-borne and adsorbed offensive eleaents can be made
to flow toward wells (electrodes) and can be extracted
froa the ground.
-- AS AID TO OTHER ALTERNATIVE TECHNIQUES
a) With Excavation of Contaminated Soil
Soils which are difficult to dewater by
conventional techniques can be dewatered; need to
excavate Bay be eliminated by lowering the water
table.
b) With Installation of Impervious Barriers
For grout curtains, sheet piles and floor seals,
etc., electro-kinetics can be used to enhance
placement at desired locations and thereby
reaove uncertainties in forming an impervious
seal.
c)
ijlth Injection of Chemical or Biological
Detoxifying Agents"
Electro-kinetics can be used to assure uniform
distribution and improve effectiveness.
d) With Collection of Contaminant Plumes by Pumping
Electro-kinetics can be used to reduce dilution by
surrounding fresh water. Contaminant plumes may
be directed back to waste site for containment or
treatment.
BACKGROUND
Site History
The field evaluation of electro-
kinetic treatment is being conducted in
Corvallis, Oregon within the confines of
an abandoned industrial hard chrome
plating facility, previously called United
Chrome Products, Inc. (UCPI). The site is
one of the nation's Superfund sites and is
situated next to the Corvallis Airport,
south of the Airport Road. The UCPI
facility and its immediate vicinity is
shown in the site location plan (Fig. 1).
UCPI conducted a chromium plating opera-
tion at the site from 1956 to 1985 and is
reportedly responsible for discharging
liquid plating wastes into the subsurface
soils at the site. The principal loca-
tions of discharge were the dry well dis-
posal pit on the west end of the building
and some leaky plating tanks inside the
building.
Subsurface Conditions
The project site is located on the
alluvial plains of the Willamette River
Valley and occupies about 1.5 acres of
level ground. Detailed investigation of
the conditions at the site was carried out
by the CHoM Hill Company under a contract
with theTJSEPA Hazardous Waste Site Con-
trol Division of the Region 10 office and
is available in the report for Contract
No. 68-01-6692. This report reveals that
the subsoils are characterized generally
by unconsolidated deposits of clay, silt
and gravel which can be broken into three
distinct units, an upper aquifer, a lower
aquifer and an aquitard separating the two
aquifers. The upper soils in the profile
consist of about 2 to 3 ft. of
miscellaneous fill or top soil and 15 to
20 ft. of clayey silt to silt ranging from
mottled grayish brown to bright yellow in
color. This layer of soil extends to a
depth of 17 to 21 ft. below the ground
surface. The less pervious soil layer
which separates the two saturated previous
zones, consists of light blue to dark gray
clay to silty/sandy clay ranging in thick-
ness from 2.5 to 12 ft. The bottom of
this layer extends from 22 to 29 ft. in
depth below the ground surface over the
site. The lower aquifer consists of wet,
fine to coarse sands and gravels and is at
least 15 ft. thick at the site. The
-194-
-------
AIRPORT ROAD
'BE2
N
Leaky
Storage Tanks
BE1
UNITED CHROME
Dry Well Area
- Soil Boring Locations.
Fig. 1: Location plan for soil borings.
groundwater table fluctuates seasonally
between 0 to 10 ft. below the ground sur-
face and has a slope of approximately 5
ft. per mile in the north-northeast direc-
tion.
Site Characteristics
The near-surface soils are heavily
contaminated with chromium and the distri-
bution of the contaminant concentration is
greatly variable (100 - 10000 mg/1) with
depth and distance from the source points.
Fortunately, the affected soils have low
permeability (ranging from 5 x 10~4
cm/sec, to 5 x 10 ° cm/sec.) and there is
very little flow of groundwater at the
site, the average hydraulic gradient at
the site being only 0.008 ft./ft. In
fact, the risk of proliferation is
increased by the presence of the open
ditch which runs along the northern edge
of the site. During the winter the
standing water in the ditch comes in
direct contact with the groundwater and
carries contaminants up to the Willamette
River.
The general characteristics of the
site were assessed in view of the poten-
tial application of electro-kinetic treat-
ment. These characteristics, e.g.
moderately large area! extent of the site,
nearly static groundwater regime and
saturated moderately permeable soils at
shallow depth and relatively simple
chemistry of the wastes, are ideally
suitable for in-situ electro-kinetic
treatment. The size of the site is
neither so large that an untested tech-
nique might be considered too risky for
the benefits expected, nor it is so small
that excavation and disposal at a secure
+BE3 facility can be considered desirable.
Again, the near-surface contaminated soils
which range from clayey silts to silty
clays are not permeable enough to be
easily treated by iri-situ leaching. The
major ionic constitutents of the contami-
nants at the site are trivalent and hexa-
valent chromium (Cr III cationic form, and
Cr VI anionic form). The anionic chromium,
being more stable in the pH range of the
groundwater at this site, is more predomi-
nant in the soil-water system. Hexavalent
chromium is relatively unreactive with
soil and is not expected to be
significantly retarded in its migration by
adsorption or precipitation. Furthermore,
the high enough concentration and ionic
mobility of chromium are also favorable
factors. Hence, it seems quite likely
that it can be transported with the
induced flow with relatively high
efficiency and with low electrical power
consumption. x
PRELIMINARY RESULTS AND DISCUSSION
Laboratory Tests
To .aid in the design and development
of a field test program, a field
exploration effort was first devoted to
obtaining representative samples of
contaminated on-site soils. These
samples were brought back to the labor-
atory to evaluate the pertinent geotech-
nical and chemical characteristics of the
soil and pore water and to develop some
bench-scale data on the effects of
electro-kinetic treatment.
The laboratory tests were carried out
in a specially designed electro-kinetic
cell. It was decided that combining of
hydraulic leaching with electro-kinetic
treatment may be appropriate, since main-
taining continuous flow of liquid across
the sample would eliminate desaturation
-195-
-------
due to trapping of evolved gases in the
sample. The effectiveness of electro-
kinetic treatment was observed by
comparing the results of hydraulic
leaching experiments with those of com-
bined hydraulic and electro-kinetic
experiments.
Chromium removal by leaching of the
soil column appeared to depend only on the
total water flow through the sample and
was apparently independent of the
hydraulic gradient (or pressure
difference) across the sample. Results
from the tests are presented as normalized
plots of cumulative fraction of chromium
removed against cumulative volume of water
leached per unit volume of soil sample.
Fig. 2 shows that no significant
difference in the total fraction of
chromium removed was noticeable even
though Experiments 1 and 2 were carried
out at variable pressure differences
across the sample. These results indicate
that in the case of hydraulic leaching
alone the principal mechanism of chromium
removal is miscible displacement. The
results of Experiments 1 and 2 were used
as a comparison basis for the electro-
kinetic effect on chromium removal in the
subsequent experiments.
In the later experiments, DC electric
fields were applied across the sample in
addition to the hydraulic leaching. The
first few experiments which combined
hydraulic leaching and electro-kinetic
treatment were carried out with frequent
reversal of the direction of fluid flow or
electric field. The superposition of
electric field was seen to either reduce
or enhance removal rate depending upon the
point of collection of net hydraulic flow
or effluent. While anodic effluent showed
an increase in the rate of removal,
cathodic effluents showed a decrease. In
Experiments 5, 7, and 8, the effluents
were collected at the anode. Fig. 3 shows
a comparison of the results of Experiment
1 (and 2 combined) with those of Experi-
ments 5 and 7 carried out at different
field strengths. It can be seen from Fig.
3 that in order to achieve 95% removal of
chromium, Experiments 1 (and 2 combined),
7 and 5 would require passing of net
fluid flow of 1.1., 0.8, and 0.5 times the
volume of soil treated. In other words,
electro-kinetics may significantly de-
crease not only the time of in-situ treat-
ment, but also the amount of leachate
water to be treated elsewhere.
1.0
0.9-
0.8-
0.7-
0.6
0.5
0.4
0.3
0.2-
0.1
0.0.
p=OJ>7
p=pressure difference
across sanpte
+ Exp1, p=0.14 Kg/cm2
O Exp2, variable p
OO
0.3 0.6 0.9 1.2 1.5 1.8 2.1
Cumulative Volume Ratio, I V/Vo
24
Fig. 2: Effect of hydraulic gradient
On Cr removal by leaching.
o D OB o
X Exp 1 & 2
D Exp 5 - 1.0 V/cm
O Exp 7 - 0.1 V/cm
0.0 03 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 30
Cnmolatire Volume Ratio, EV/Vo
Fig. 3: Comparison of effects of
combined treatment and
hydraulic leaching.
-196-
-------
Field Test Program
Although the laboratory test results
Indicate that simultaneous application of
hydraulic and electrical gradients in
opposite directions is necessary for re-
moval of (anionic) chromium, this may be
an artifact of the laboratory electro-
kinetic cell. In order to prevent de-
saturation of the laboratory sample near
the anode, it was necessary to maintain
flow through the sample by hydraulic
leaching. For field soils, this may not
be warranted and it may be possible to
subject the field soils to electro-
kinetics alone for treatment. Hence, it
was considered appropriate to adopt a two-
stage approach for field activities in
order that such uncertainties can be
resolved. The field-scale effort was
split into: a) feasibility-level experi-
ments and b) final experiments. The pur-
pose of the feasibility level experiments
was evaluation (if possible) of the effec-
tiveness of the process of ionic migra-
tion, identification and elimination of
possible experimental difficulties, cali-
bration of instrumentation and estimation
of hydraulic and electro-kinetic conduc-
tivity parameters in the field. The major
emphasis was to check in the field whether
anionic chromium can be mobilized towards
the anodes simply due to the process of
ionic migration under electrical field or
if migration of chromium ions takes place
in the direction of movement of the bulk
solution.
The field-scale operations were
designed in as much detail as possible by
giving proper consideration to: a) elec-
trode configuration and material, b)
hydraulic leaching system, c) monitoring
system and d) power source and con-
sumption. The shallow well BE1 shown in
Fig. 1 was chosen to serve as the cathode.
The anodes were installed at the vertices
of a regular hexagon (5 ft. spacing) with
the cathode being at the center of the
hexagon. The anodes were reinforcing
steel rods (0.25 in. diameter, 20 ft.long)
inserted in 15 ft. deep slotted PVC tube
wells. An alternate set of four such
anodes was also placed in a square pattern
(8 ft. spacing) around the cathode.
During the experiments, the response of
the soil-water electrolyte to the treat-
ment was designed to be monitored and
controlled by keeping measurement records
of the important experimental variables,
e.g. applied potential, total current flow
across the pairs of electrodes, pH, con-
O
SQ1
DIRECTION OF
GROUHBWATER flOW
_. J
••"F^-H
SYMBOLS:
-- KbdricUr UM lor 1ta*o&u«n.
(•) Vill «ad OtclKril.
• For* Fr*Mor« Tnuudtmr.
Fig. 4: Schematic plan for field
experimental setup.
I iMilYSIS 1
Fig. 5: Schematic plan for monitoring
and sampling system.
-197-
-------
ductivity, chromium concentration, temper-
ature, redox potential of the effluent,
time of sampling and length of treatment
time, hydraulic flow volume injected or
withdrawn and levels of standing water in
the wells. A schematic plan of the elec-
trodes and the electrical connections for
the field set up-is shown in Fig. 4. Fig.
5 shows the schematic plan of the moni-
toring and sampling system.
Three preliminary sets of experiments
with hexagonal and square anode geometry
have been completed. The early indi-
cations are very promising. The typical
results obtained from the first two con-
secutive experiments are shown in Figs. 6
and 7.
The first test was carried out for
26.5 hours with 5-volt potential
difference across the anodes arranged in
the hexagonal pattern and the cathode.
The second test was performed by applying
30 volts across the same electrodes for 5
hours. Figs. 6a) and 7a) show temporal
variations of pH values of the water
samples extracted from the cathode (SE1),
one of the anodes (H5) and the inter-
mediate monitoring well (Ml). The tem-
poral variations of concentration of
chromium in those samples are shown in
Figs. 6b) and 7b). It can be noted that
pH values are increasing at SE1 and HI and
decreasing at H5. This is expectedly due
to electrolysis of water at and near the
cathode and dissolution of ion at the
anode. Unfortunately, the trend in the
variation of chromium concentration is
apparently confusing. The decrease in
concentration of chromium at the anode is
unexpected. This anomolous change at the
anode may have been caused by copre-
cipitation of chromium with iron oxide
produced by anodic dissolution. The
decrease of chromium concentration at the
TEST 1:HBXACONAL PATTERN CSV FOR Z«:30HR
a) pH changes In Cathode, Anode, and
Monitoring Will.
SE1 (CATHODE)
D H5 (ANODE) •
O Ml (MONITORING)
0.0
00
TSiJT 22.5 30.0
TIME (HOURS)
37.5
4 .0
b) Cnong.ll of Total Cr Concentration In
Cathode, Anode, and Monitoring Well.
TEST 2:HEXAGONAL PATTERN O30V FOR 5 HR
a) pH changes In Cathode, Anode, and
Monitoring Well.
12.0
10.5
9.O
7.5
6.01
0.0
450!
18
21
"T 12 15
TUB (HOURS)
SE1 (CATHODE) a H5 (ANODE) O Ml (MONITORING)
b> Changes of Total Cr Concentration In
Cathode, Anode, and Monitoring Well.
24
+ SE1 (CATHODE)
a H5 (ANODE)
O Ml (MONITORING WELL)
00
7.5
15.0 22J536.0
TIUE (HOURS)
9 12 IS
TIME (HOURS)
Fig. 6: Temporal variations of pH and Cr
concentration during Test 1.
Fig. 7: Temporal variations of pH and
Cr concentration during Test 2.
-198-
-------
cathode with simultaneous Increase at the
monitoring well may be indicative of ionic
transport due to electrical treatment
during the second test. Further assess-
ment of the preliminary test data is con-
tinuing and a more coherent understanding
of the process may emerge in the near
future.
CONCLUSIONS
The following conclusions may be
drawn from the results of the experiments
conducted so far:
(1) Laboratory tests have shown that a
combination of electro-kinetic and
hydraulic treatment can accelerate
the process of chromium removal sub-
stantially compared to hydraulic
treatment by itself. This acceler-
ation should be much more pronounced
in the field operations where the
hydraulic gradient which can be prac-
tically applied (and consequently the
flow) will be very small.
ACKNOWLEDGEMENTS
The research described in this paper
has been funded wholly or in part by the
United States Environmental Protection
Agency through Cooperative Agreement No.
CR811762, to the University of Washington.
It has been subject to the Agency's peer
review and approved for publication.
Approval does not signify that the content
necessarily reflects the view and policies
of the Agency, nor does mention of trade-
names or commercial products constitute
endorsement or recommendation for use.
Technical guidance from the Project
Officer, Mr. Jonathan G. Herrmann, is
gratefully acknolwedged. Contributions by
Prof. J. F. Ferguson, co-investigator, and
Mr. J. J. Horng, graduate research
assistant, are also greatly appreciated.
(2) The possible mechanisms involved in
the treatment are dispersion due to
hydraulic flow, ion migration, elec-
trolysis of water, adsorption/de-
sorption and chromium reduction due
to imposed electric field. There are
significant experimental and theo-
retical problems in quantitatively
assessing these factors and their
interactions.
(3) Considerable work is still needed to
reach general guidelines for use of
this technology. Continued labor-
atory experimentation is needed to
accompany field studies and labor-
atory studies should evolve to extend
the range of physical and chemical
conditions studied.
-199-
-------
REFERENCES
1. Anbah, S. A., Chilingar, G. V. and 6.
Beeson, C. H., "Application of Elec-
trical Current for Increasing the
Flow Rate of Oil and Hater in a
Porous Medium," J. of Can. Petro.
Tech.. April-June 1965. 7.
2. Butterfield, R. and Johnson, I. H.,
"The Influence of Electro-Osmosis on
Metallic Piles in Clay," Geotech-
nlque, Vol. 30, No. 1, 1980, pp. 17-
38. 8.
3. Chappel, B. A. and Burton, P. L.,
"Electro-Osmosis Applied to Unstable
Embankment," Journal of the Geotech-
nical Division, ASCE, Vol. 101, GTS, 9.
August, 1974, pp. 733-740.
4. Esrig, M. I. and Henkel, D. J., "The
Use of Electro-Kinetics in Raising
Submerged, Partially Buried Metallic 10.
Objects," Report to Department of
Navy, ONR, Research Project No. RR-
004-01-01, March 25, 1966.
5. Fetzer, C. A., "Electro-Osmotic 11.
Stabilization of West Branch Dam,"
Journal of SH and Fdn. Division,
ASCE, Vol. 93, SM4, July 1967, pp.
85-106.
Gibbs, H. J., "Research on Electro-
Reclamation of Saline-Alkali Soils,"
Trans. ASAE, Vol. 9, 1966, pp. 164-
169.
Hamnet, R., "A Study of the Processes
Involved in the Electro-Reclamation
of Contaminated Soils," M.S.C.
Thesis, Univ. of Manchester, U.K.,
1980.
Liang, L., "Electro-Osmotic De-
watering of Hastewater Sludges,"
Ph.D. Thesis, Dept. of Mechanical
Eng'g., M.I.T., Cambridge, MA, 1977.
Lockhart, N. C., "Sedimentation and
Electro-Osmotic Dewatering of Coal-
Washery Slimes," Fuel 60, 1981, 919-
923.
Puri, A. N., and Anand, B., "Recla-
mation of Alkali Soils by Electro-
dialysis, Soil Science, Vol. 42,
1936, 23-27.
Segal 1, B. A., O'Bannon, C. E. and
Matthias, J. A., "Electro-Osmosis
Chemistry of Hater Quality," Journal
of Geotechnical Division, ASCE, Vol.
106, GT10, October 1980, pp. 1148-
1152.
-200-
-------
CURRENT STATUS OF THE DESIGNATION AND ADJUSTMENT OF CERCLA
HAZARDOUS SUBSTANCES AND THEIR ASSOCIATED REPORTABLE QUANTITIES
K. Jack Kooyoomjian
John E. Riley
Office of Solid Waste and Emergency Response
Emergency Response Division
U.S. Environmental Protection Agency
Washington, DC 20460
Richard Field
Office of Research and Development
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, NO 08837
ABSTRACT
In this paper the U.S. Environmental Protection Agency (EPA) describes the-technical
methodology it has used to adjust reportable quantities (RQs) of CERCLA hazardous
substances, which when released into the environment must be reported to the National
Response Center (NRC), and the methodologies the Agency is considering for designation of
additional CERCLA hazardous substances. In accordance With CERCLA Section 102 the EPA
Administrator may promulgate regulations to establish the level of release of a hazardous
substance which must be reported to the NRC. The methodology considers the intrinsic
physical/chemical, toxicologic, and degradative properties of the hazardous substance
The Administrator issued Final Rules on April 4, 1985 and on September 29, 1986 which
a^J^n^nthe statutory RQS of 442 of the 717 CERCLA hazardous substances. Section 102(a)
of CERCLA provides the Administrator with the authority to designate additional hazardous
substances and adjust their RQs. The options available.to the Administrator for choosing
those substances most appropriate for designation are also described.
INTRODUCTION
This paper is a status report of the
U.S. Environmental Protection Agency's
(EPA's/Agency's) progress to date on the
adjustment of the statutory Reportable
Quantities (RQs) of the hazardous substances
(HS) as defined in the Comprehensive
Environmental Response, Compensation, and
Liability Act (CERCLA or "Superfund"), and
on the Agency's strategy for designating
additional CERCLA HS as provided for in
CERCLA Section 102(a). It also discusses
the status of other EPA regulations related
to the reporting of releases of HS and other
potentially harmful chemicals.
This paper is relevant to this
conference on hazardous waste research
because the reporting provisions of CERCLA
and its implementing regulations are not
limited to episodic releases such as those
which sometimes occur during transportation
incidents, for example, but apply to any
release which reaches the environment such
as those which occur at hazardous waste
sites.
BACKGROUND
The CERCLA Hazardous Substances
The term "hazardous substance" has a
very specific meaning according to CERCLA.
-201-
-------
Section 101(14) of CERCLA defines a
hazardous substance as:
1. Any substance designated pursuant to
Section 311(b) of the Clean Water Act
(CWA);
2. Any hazardous waste having
characteristics identified under or
listed pursuant to Section 3001 of the
Solid Waste Disposal Act, otherwise known
as the Resource Conservation and Recovery
Act (RCRA);
3. Any toxic pollutant listed under Section
307(a) of the CWA;
4. Any hazardous air pollutant listed under
Section 112 of the Clean Air Act (CAA);
5. Any imminently hazardous chemical
substance or mixture with respect to
which the Administrator of EPA has taken
action pursuant to Section 7 of the Toxic
Substances Control Act (TSCA); and
6. Any element, compound, mixture, solution
or substance the Administrator determines
to be hazardous pursuant to Section 102
of CERCLA.
There are currently 717 HS composed of
611 unique chemical compounds and 106 waste
streams. This total does not include
chemicals or mixtures that exhibit
characteristics of ignitibility,
corrosivity, reactivity, or extraction
procedure toxicity according to 40 CFR
261.20. Such chemicals and mixtures are
also considered HS, but because it would be
impossible to identify all such chemicals
and mixtures, they are not specifically
listed under RCRA, and therefore, they are
not specifically listed under CERCLA. As of
this date the Administrator has not
exercised his authority to designate any new
substances as CERCLA HS pursuant to Section
102.
Notification Requirements
Section 103(a) of CERCLA requires any
person in charge of an offshore or onshore
facility or a vessel to report to the
National Response Center (NRC) as soon as
that person has knowledge of any release of
a HS that is equal to or greater than the RQ
for that HS. The term "facility" includes
active and inactive hazardous waste sites.
Once the NRC is notified, it informs the
predesignated On-Scene Coordinator (OSC)
pursuant to the National Contingency Plan
(40 CFR 300) of the release. The OSC
evaluates the circumstances of the release,
gives the pertinent information to
appropriate state and local officials and
decides whether, and in what manner the
Federal government should respond to the
release.
Adjustment of Reportable Quantities
Congress assigned statutory RQs to the
HS in Section 102(b) of CERCLA. All of the
HS received statutory RQs of 1-pound except
those HS that were designated pursuant to
Section 311(b) of the CWA which received
statutory RQs equal to the RQs assigned
pursuant to Section 311(b)(4) of the CWA
and codified in 40 CFR 117. Section 102 of
CERCLA also authorizes the Administrator to
adjust the statutory RQs of the HS.
Rulemaking History
Adjustments to Statutory RQs., The Agency
has promulgated final adjusted RQs for 442
HS in two separate rulemaking actions. The
first of these final rules was published in
the Federal Register on April 4, 1985 (50
FR 13456 - 13513).It finalized adjusted
RQs .for 340 HS. The other final rule was
published on September 29, 1986 (51 £R
34534 - 34549) and it finalized adjusted
RQs for 102 HS. Most of these 102 HS had
been identified as chronic toxicants. On
March 16, 1987 the Agency published a
Notice of Proposed Rulemaking (NPRM) (52 FR
8140 - 8171) that proposed adjustments to
an additional 273 of the remaining 275 HS.
This proposed rule presents the Agency's
methodology for adjusting statutory RQs on
the basis of potential carcinogenicity.
The two remaining HS, lead and methyl
isocyanate, are still being studied by the
Agency, and may have their statutory
1-pound RQs adjusted in the future.
Also, in the Federal Register on March
16, 1986 the Agency published another NPRM
that proposed adjustments to the 1-pound
statutory RQ for Radionuclides (52 _FR 8172
- 8186). This broad generic category
includes thousands of separate substances,
and therefore neither the category nor the
individual radionuclides are counted among
the 717 CERCLA HS.
Designation of Additional Hazardous
Substances. The Agency has solicited
-202-
-------
comments on various methods that could be
used to designate additional HS in an
Advance Notice of Proposed Rulemaking
(ANPRM) published in the Federal Register on
May 25, 1983 (48 £R 23602 - 23605jT^ -
The Reauthorized Superfund
On October 17, 1986 the President
signed the Superfund Amendments and
Reauthorization Act of 1986 (SARA). The
SARA is a law that strengthens the EPA's
role in state environmental activities,
increases the liability of private
companies, and imposes strict requirements
on settlement provisions, judicial review,
and cleanup standards. In the reauthorized
CERCLA, Congress included Title III (The
Emergency Planning artd Community Right-to-
Know Act of 1986) which was designed to
prevent a catastrophe similar to the
incident which occurred in Bhopal, India in
1984 with the release of methyl isocyanate.
Title III provides for the
establishment of emergency plans at the
local level based on data from facilities in
the area. Chemicals stored at these
facilities are reported to a Local Emergency
Planning Committee (LEPC), which then will
consider how to handle a situation in the
event of a release of that chemical. In
order to implement the system, Congress has
defined the chemicals and their quantities
that shall be subject to this reporting
requirement. The Title also provides for
civil and criminal penalties for failure to
comply.
RQ ADJUSTMENT
A previous paper (1) provided a
detailed description of the methodology the
Agency uses to adjust the statutory RQs of
the CERCLA HS. In addition, the reader is
referred to the. Technical Background
Documents that provide the technical basis
for the RQ adjustment methodology (2)(3)(4).
The following is a brief description of the
RQ adjustment methodology.
The RQ adjustment methodology consists
of two major steps. The first major step
involves evaluation of the physical,
chemical and toxicological characteristics
of each HS. The primary criteria examined
are: aquatic toxicity, acute mammalian
toxicity (oral, dermal, and inhalation),
ignitibility, reactivity, chronic toxicity,
and potential carcinogenicity. For each
primary criterion a five-tier rating scale
is used, corresponding with RQ values of 1,
10, 100, 1000, and 5000 pounds. Since this
five-tier system was successfully used in
the CWA and both the regulated community
and response personnel are familiar with
it, the Agency has decided to use this
methodology for adjusting the statutory RQs
of the CERCLA HS.
The Agency has proposed to use only
three of the five RQ levels (1, 10, and 100
pounds) for assigning adjusted RQs to the
CERCLA HS that are potential carcinogens.
This decision is based on the special
properties associated with potential
carcinogens and an analysis of the risks
posed by their release. A complete
discussion is provided in (4).
Each HS is evaluated according to the
primary criteria and an RQ value is
determined for each applicable criterion.
The primary criteria RQ for each HS is the
lowest value of all the applicable
criteria. For example, if a particular
HS's ignitibility corresponds to a RQ of
1000 pounds, chronic toxicity to a 5000
pound RQ, aquatic toxicity to a 10 pound
RQ, mammalian toxicity to a 100 pound RQ,
and reactivity to a 1000 pound RQ, the
primary criteria RQ is 10 pounds, based on
the RQ for aquatic toxicity. The Agency •
then evaluates each HS according to a set
of secondary criteria consisting of.
biodegradation, hydrolysis, or photolysis-
(BHP). The primary criteria RQ is then
raised one level if an analysis indicates
that the HS naturally degrades by BHP into
less hazardous products when released into
the environment.
Research Needs
The Agency recognizes that technical
updates to the final rules on RQ
adjustments will be an ongoing process
because; 1) the RQ adjustment methodology
relies on existing data in the areas of
health effects, aquatic toxicity,
ignitibility, reactivity, and degradation
.in the environment, many of which are areas
of active research, and 2) the Agency is
active in the areas of generating and
reviewing data on the toxicology of
chemicals. Technical updates of
promulgated adjusted RQs are appropriate
because such updates will result in RQs
that are more accurate and which will
better protect the public health, welfare,
-203-
-------
and the environment, and in those cases
where additional data result in raising an
RQ, will lessen compliance burdens on the
regulated community.
The Agency will periodically review the
published and unpublished literature for the
purpose of using the highest quality data
for RQ adjustment purposes. When new data
are found that affect the adjusted RQ of a
HS the Agency will propose a new adjusted RQ
in the Federal Register in a technical
update rulemaking. In addition, the Agency
will accept for review data on chemicals
supplied by the public from the literature
or from laboratory studies, and if
warranted, propose adjusting the RQ of the
HS in question.
An analysis of the current data set on
the CERCLA HS reveals the following
statistics :
1. There are 215 HS (35%) without aquatic
toxicity data, although some of these are
gases or are insoluble in water.
2. There are 193 HS (32%) without acute
mammalian oral toxicity data, although
some of these are gases.
3. There are 497 HS (81%) without acute
mammalian inhalation toxicity data,
although some of these are solids not
normally available as dusts or low vapor-
pressure liquids.
4. There are 542 HS (89%) without acute
mammalian dermal toxicity data, although
many of these are gases.
5. There are six HS (1%) that have adjusted
RQs (either final or soon to be proposed)
based on either LDlo or LClo data (the
lowest dosage or concentration known to
have been lethal to an individual of the
test species). Such data are generally
from limited studies that were unable to
produce suitable LD50 or LC50 data, or
from accidental poisonings (see Table 1
for definitions of LD data).
6. It is believed that the data collected on
ignitibility and reactivity represent a
complete data set since ignitibility is
based on the fundamental physical
properties of boiling point and flash
point, and reactivity is based on well
known polymerization and water reaction
tendencies. The methods for measuring
these properties were established many
years ago and massive amounts of such data
have been generated.
DESIGNATION OF NEU CERCLA HAZARDOUS
SUBSTANCES
Section 102(a) of CERCLA provides the
Administrator of EPA the authority to
designate as hazardous any element,
mixture, solution, or substance, which,
when released into the environment, may
present substantial danger to the public
health or welfare or the environment. As
mentioned previously, the Agency published
an ANPRM on designation on May 25, 1983
(48 £R 23602 - 23605). The Agency is
considering several sources of candidate
substances for designation, including the
lists mandated by SARA and from other
sources.
Reasons for Designating Additional
Hazardous Substances
The reasons for designating additional
substances are:
1. The lists pursuant to CERCLA Section
101(14) are not all inclusive of
potentially harmful substances.
2 Designation of substances as HS
encourages greater care in handling.
3. Designation of substances as HS
encourages recycling.
4. Designation of substances as HS provides
the Administrator the mechanism of
recovering costs for cleanup.
Each of these reasons is discussed below.
CERCLA Section 101(14) Lists Are Not All
Inclusive. Many substances which are
hazardous to the public do not necessarily
qualify to be on the lists Congress
identified in CERCLA Section 101(14).
Those lists were developed to protect a
single medium, or, in the case of hazardous
wastes, to protect the public from improper
disposal and other mishandling of hazardous
wastes. Each of the lists is responsive to
one or more sections of environmental
statutes, i.e., the Clean Air Act, the
Clean Water Act, the Solid Waste Disposal
Act, and the Toxic Substances Control Act.
To meet the Congressional requirements of
those acts, the Agency developed technical
-204-
-------
criteria to be applied to chemicals and/or
wastes. These criteria were designed to
specifically meet certain requirements of
the acts.
Congress anticipated that by providing a
multiplicity of lists, taken from several
regulations developed pursuant to the acts
discussed above, most substances which
provide increased risks to the public,
welfare and environment would be covered.
However, Congress recognized that the
umbrella of these four major environmental
acts might not cover all the substances from
which the public should be protected (e.g.,
Congress did not include the Federal
Insecticide, Fungicide, and Rodenticide Act
[FIFRA] active ingredients). Accordingly,
Congress gave the Administrator authority to
designate additional substances as CERCLA
HS, thus removing the media constraints from
the Administrator in designating HS.
Designation Encourages Safer Handling
Procedures. Experience has shown that
liability for cleanup, pursuant to Section
107 of CERCLA, has made handlers of HS more
cognizant of the risks HS pose to the
public. Accordingly, more attention is paid
to handling, packaging, shipping, storage,
and general treatment of HS. Thus, there is
a tendency for industry to treat designated
HS in a safer manner than pollutants and
contaminants.
Designation Encourages Recycling. Many
chemicals produced, used, or processed by
manufacturers often are discarded when they
become contaminated or otherwise out of
specification. Frequently, it is more
economical to discard the chemical than to
recycle it. Discarding, of course, does not
eliminate the chemical; usually it is sent
to a landfill. If the chemical is a
pollutant or contaminant it may go to a
sanitary landfill where it can eventually
migrate to ground waters or perhaps become
airborne. If that chemical does not pose a
threat to the health, welfare or
environment, then disposal in a landfill is
both economical and safe. However, if the
chemical does pose a threat, in the absence
of regulation, it can still be discarded in
a sanitary landfill at a relatively low
cost. On the other hand, if that substance
is a designated HS, disposal in a sanitary
landfill is not acceptable, and the cost of
disposal in a suitable landfill is high. In
most cases the cost of disposal can exceed
the cost of recycling. Hence, in order to
encourage industry to both conserve
resources and to protect the public health,
welfare and the environment, substances
which do pose threats should be designated
as HS to reduce the total quantity of those
undesignated but harmful substances in
waste disposal sites.
Designation Provides the Administrator the
Mechanism for Recovering Costs of Cleanup.
Section 106 of CERCLA provides the EPA with
a mechanism to recover government costs for
cleanup of spills of HS from owners or
operators, or from other responsible
parties. CERCLA does not provide a
mechanism for cost recovery if the
substance for which remedial action was
taken is not a CERCLA HS. Furthermore,
owners or operators can be liable for up to
$50 million for damage to the environment
caused by releases of HS. No such
liability exists for damage due to releases
of non-CERCLA HS. Therefore, in order to
protect the SUPERFUND and to correct
environmental damages, it is incumbent on
the Administrator to designate as HS those
substances which when released into the
environment are likely to trigger a
government response.
Candidate Substances for Designation from
SARA ; :
Both Title III of SARA (Emergency
Planning and Community Right-to-Know Act)
and Title I of SARA (Provisions Relating
Primarily to Response and Liability)
provide source lists the Agency is
considering for designation purposes.
The Extremely Hazardous Substances.
Section 302 defines the specific list of
Extremely Hazardous Substances (EHS) and
requires the Agency to publish the list
within 30 days after the enactment of SARA.
The list of EHS is defined as "the list of
substances published in November, 1985 by
the Administrator in Appendix A of the
Chemical Emergency Preparedness Program
Interim Guidance." This list was
established by the Agency to identify
chemical substances which could cause
irreversible health effects from accidental
releases. Any facility making, using or
storing EHS above a promulgated threshold
planning quantity must report the quantity
to the Local Emergency Planning Committee
(LEPC). Furthermore, a release of an EHS
must be reported to the LEPC if the release
is in a quantity at or above the
-205-
-------
promulgated release notification quantity.
If the EHS is also a HS the release quantity
is the RQ of the HS. In the event that the
EHS is not a HS the release notification
value is that value promulgated in the April
22, 1987 Final Rule (52 £R 13378 - 13410).
However, there is a subtle difference
between notification requirements of Section
304 and Section 103. In Section 304,
notification must be made only if the
release leaves or threatens to leave the
perimeter of the facility, whereas in
Section 103, notification to the NRC must be
made if the release is into the environment.
Thus, the perimeter of the facility is not a
constraint for notification requirements
under Section 103. EPA published the list
Of 402 EHS on November 17, 1986 (51 £R 41570
- 41592) in an Interim Final Rule, and on
the same date, published an NPRM which
proposed the addition of five substances to
the list of EHS, and the deletion of 40
substances from the list of EHS (51 _FR 41593
- 41594). The criteria that are the basis
of the list of EHS are provided in Table 1.
On April 22, 1987, EPA published a
Final Rule which revised the list of EHS to
include four additional substances. The
Agency had decided to retain the 40
substances proposed for deletion pending the
development of criteria for determining
additional health effects resulting from
short-term exposure at specified levels.
EPA intends to reassess these 40 chemicals
when such criteria are available.
The Agency applied the criteria shown
in Table 1 to the data base of chemicals
contained in the Registry of Toxic Effects
of Chemical Substances (RTECS) and applied
the additional criterion that the EHS had to
be in current production. In addition, the
Agency included substances on the EHS list
that did not meet all of the formal criteria
discussed above, but are in high production
and have caused death and injury in
accidents.
There are currently 406 substances on
the list of EHS of which 145 are also CERCLA
HS. Therefore, 261 EHS are candidates for
designation as HS.
The Agency is considering designating
these 261 EHS as CERCLA HS to ensure that
releases of these substances are reported to
the NRC for possible response by the Federal
government, and to protect the SUPERFUND
against response costs associated with
releases of non-designated HS.
The Toxic Chemicals. Another list of
chemicals defined in SARA that the Agency
is considering for designation purposes is
the list of toxic chemicals defined in
Section 313 of SARA (Toxic Chemical Release
Forms) as those chemicals listed in
Committee Print Number 99-169 of the Senate
Committee on Environment and Public Works,
titled "Toxic Chemicals Subject to Section
313 of the Emergency Planning and Community
Right-to-Know Act of 1986," including any
revised version of the list as may be made
pursuant to subsection (d) or (e) of
Section 313 of Title III of SARA. The
Administrator may by rule add a chemical
if there is sufficient evidence to
establish that: 1) the chemical is known to
cause or can reasonably be anticipated to
cause significant adverse acute human
health effects at concentration levels that
are reasonably likely to exist beyond
facility site boundaries as a result of
continuous, or frequently recurring
releases; 2) the chemical is known to cause
or can reasonably be anticipated to cause
cancer in humans or teratogenic effects, or
serious irreversible reproduction
dysfunctions, neurological disorders,
heritable genetic mutations, or other
chronic health effects; or 3) the chemical
is known to cause or can reasonably be
expected to cause, because of its toxicity,
its toxicity and persistence in the
environment, or its toxicity and tendency
to bioaccumulate in the environment,
significant adverse effects on the
environment of sufficient seriousness to
cause reporting under Section 313 of Title
III of SARA. In addition, any person may
petition the Administrator to add a
chemical to the list of toxic chemicals.
The owner or operator of a facility
subject to the requirements of this section
of Title III of SARA must complete a toxic
chemical release form for each toxic
chemical listed that was manufactured,
processed, or otherwise used in quantities
exceeding the toxic chemical threshold
quantity defined in Section 313(f). Draft
release forms have been distributed to the
public for comment.
There are currently 329 substances on
the SARA Section 313 toxic chemicals list.
The Agency may designate some or all of the
substances on this list as CERCLA HS.
-206-
-------
The Hazardous Waste Site Substances.
Section 110 of SARA (Health-Related
Authorities) requires that the
Administrators of both the EPA and the
Agency for Toxic Substances and Disease
Registry (ATSDR) prepare a list of at least
100 HS which are most commonly found at
facilities on the National Priorities List
(NPL) and which they determine are posing
the most significant potential threat to
human health due to their known or suspected
toxicity to humans and the potential for
human exposure to such substances at
facilities on the NPL or at facilities to
which a response to a release or a
threatened release is under consideration.
This section also provides for regular
additions to this list. The Administrator
of the ATSDR must prepare toxicological
profiles of each of these substances.
The Agency may designate as hazardous
those Section 110 substances that are not
already HS to protect the Superfund against
response costs associated with releases on
non-designated chemicals.
Other Candidate Substances for Designation
In addition to the lists identified
above that contain chemicals that are
candidates for designation, the Agency has
identified a number of other lists that may
contain chemicals that should be designated
as HS. These list include but are not
limited to :
* The Federal Insecticide, Fungicide, and
Rodenticide Act (FIFRA) active
ingredients;
* The Occupational Safety and Health
Administration (OSHA) substances, i.e.,
those substances requiring a Material
Safety Data Sheet.
* Carcinogens identified by the Carcinogen
Assessment Group, the Office of Toxic
Substances and other Agency offices, as
well as those identified by the National
Toxicology Program (NTP) and the
International Agency for Research on
Cancer (IARC); and
The Agency may develop a priority-
setting process for selecting substances
from the lists that are of greatest concern
to the CERCLA program and focus its
resources on those highest priority
substances that are most likely to present a
threat to public health or welfare or the
environment. Criteria which could be
applied to the lists of chemicals for
purposes of prioritization includes degree
of replication of a .substances on various
lists, the level of production of a
substance, the RQ which would be assigned
to a substance using the RQ Adjustment
Methodology, the substances' history of
release and the environmental mobility and
persistency of a substance.
Options for Designation
The Agency is considering a number of
options for implementing its designation
strategy. It is cognizant of the
regulatory burden imposed by designation,
and at the same time is concerned that a
number of chemicals are released or have
the potential of being released which are
not currently CERCLA HS, and therefore, are
not reported at the federal level.
The Agency is currently considering
several strategies for designation,
including a limited designation strategy
whereby only those chemicals on the lists
identified above from SARA would be
considered for designation, with emphasis
on the Extremely Hazardous Substances. The
Agency may further prioritize these lists
to designate only the most dangerous of the
chemicals. The Agency is also considering
a broader designation methodology that
would include other lists in addition to
those from SARA as sources of candidate
substances. These options will be
presented and public comments solicited in
an NPRM to be published in the latter part
of 1987.
REFERENCES
1. Kooyoomjian, J.K., R. Field, S. Gibson,
M. Kirsch, and G. Ricci, 1986.
Reportable Quantity Guidelines for
CERCLA - Designated Chemicals.
Proceedings of the Twelfth Annual
Research Symposium on Land Disposal,
Remedial Action, Incineration an?
Treatment of Hazardous Waste. US
Environmental Protection Agency,
Cincinnati, Ohio, pp90-98.
2. Technical Background Document to Support
Rulemaking Pursuant to CERCLA Section
102, Volume 1. March 1985. Prepared
under contract No. 68-03-3182 to the US
Environmental Protection Agency.
-207-
-------
Technical Background Document to Support
Rulemaking Pursuant to CERCLA Section
102. Volume"^August 1986.Prepared
under contract No. 68-03-3182 to the US
Environmental Protection Agency.
Technical Background Document to Support
Rulemaking Pursuant to CERCLA Section
'102. Volume"!!December 1986.Prepared
under contract No. 68-03-3182 to the US
Environmental Protection Agency.
-208-
-------
Route of
Exposure *
Acute Toxicity Measure **
Value
Inhalation
Median Lethal
Concentration in Air
(LC50).
Less than or equal to
0.5 milligrams per liter
of air.
Dermal
Oral
Median Lethal Dose
(LD50).
Median Lethal Dose
(LD50).
Less than or equal to 50
milligrams per kilogram
of body weight.
Less than or equal to 25
milligrams per kilograms
of body weight.
Table 1.
Criteria to Identify Acutely Toxic Chemicals that may Present Severe Health
Hazards to Humans During a Chemical Accident or Other Emergency.
The route by which the test animals absorbed, the chemical, i.e. by breathing in air
(inhalation), by absorbing it through the skin (dermal), or by ingestion (oral).
LC50: The concentration of the chemical in air.at which 50 percent of the test animals
died. LD50: The dose which killed 50 percent of the test animals. In absence of LC50
or LD50 data, LClo or LDlo data should be used. LClo: Lethal Concentration Low, the
lowest concentration in air at which any test animals died. LDlo: Lethal Dose Low, the
lowest dose at which any test animals died.
-209-
-------
THE EPA PERSONNEL PROTECTION TECHNOLOGY RESEARCH PROGRAM
Michael D. Royer
Releases Control Branch
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Edison, NJ 08837
ABSTRACT
The Environmental Protection Agency's Personnel Protection Technology Research
Program provides data, information and technology to enhance the Agency's capability to
perform its mandated roles that require: 1) regulation of pesticides and toxic
substance handling and use; and 2) operation of EPA and contractor personnel at chemical
spills and uncontrolled hazardous waste sites. To meet this objective, the Program is
developing, evaluating and improving chemical protective clothing and equipment;
procedures to enhance the safety and cost-efficiency of working conditions; methods to
predict the effectiveness of chemical protective clothing; and detection methods and
devices that warn of imminent hazards to life and health.
INTRODUCTION
The Environmental Protection
Agency's (EPA) Personnel
Protection Technology Research
Program supports the activities of
the: 1) Office of Pesticide
Programs (OPP) mandated by the
Federal Insecticide, Fungicide and
Rodenticide Act (FIFRA), 2) Office
of Toxic Substances (OTS) mandated
by the Toxic Substances Control
Act (TSCA) and 3) Office of
Emergency and Remedial
Response(OERR) mandated by the
Superfund Amendments and
Reauthorization Act (SARA or
Superfund). The Program is
principally funded by the Water
Engineering Research Laboratory
and the Hazardous Waste
Engineering Research Laboratory
(HWERL); HWERL provides technical
management support.
Pursuant to FIFRA, OPP is
responsible for developing
agricultural protective clothing
regulations, implementing the
pesticide registration process,
reviewing and approving pesticide
labels, administering the
pesticide-related Farm Safety
Program, and supporting training
and education programs for
pesticide users through state
extension services. These efforts
are aimed at improving levels of
protection for the estimated 4
million farmers, farm workers and
farm families who may be directly
and chronically exposed to
pesticide sprays, dusts and
ambient residues. Program
research in this area currently
focuses on evaluating the
effectiveness of protective
clothing for pesticide mixers,
loaders, and applicators.
TSCA and subsequent
regulations require that producers
or importers of new chemicals
submit a Premanufacture
Notification (PMN) to OTS at least
90 days prior to beginning
manufacture or distribution. OTS
must review the PMN, which may
include a description of the
protective clothing and
respirators to be used, within 90
days and determine whether
-210-
-------
production of the new chemical
will result in unacceptable health
or environmental risks. Program
efforts in this area currently
focus on developing improved
methodologies (i.e., test methods,
predictive methods, and integrated
systems) to estimate the
effectiveness of respirators and
protective garments.
Under the provisions of SARA,
the EPA performs a variety of
roles that affect (or are affected
by) personnel protection
technology. These include
procuring and using (both directly
and through contractors)
protective clothing and equipment
for laboratory and field
operations, developing and
reviewing site safety plans, and
preparing and presenting safety
training courses for EPA employees
and client organizations. Program
efforts in this area focus on
improving personnel protection
technology in order to enhance
the safety, range, and cost
effectiveness of operations at
Superfund sites.
All of the above research
efforts are performed through
contracts, cooperative agreements
with state universities,
interagency agreements, in-house
projects, technology transfer
workshops and active coordination
with researchers, manufacturers
and users in the personnel
protection area. Through these
mechanisms, the Program: 1)
conducts desk-top, laboratory and
field evaluations; 2) prepares
guidance documents on the
selection and appropriate uses of
protective clothing and equipment;
3) evaluates, develops and
verifies methods for predicting
and testing the performance of
these products; 4) analyzes the
costs and benefits of alternative
technologies; and 5) identifies
research results that will
significantly enhance EPA
operations involving hazardous
materials. Table 1 summarizes the
outputs of these activities
through June 1987.
RESEARCH IN SUPPORT OF
FIFRA-MANDATED OPP ACTIVITIES
To support OPP's
FIFRA-mandated responsibilities,
the Personnel Protection Research
Program has initiated research
projects in three areas:
laboratory evaluations and field
tests of protective clothing
effectiveness, and production of
guidance manuals on the selection
of protective clothing for
agricultural pesticide operations.
Previous efforts by the OPP
and others have confirmed that the
hands are an important site of
pesticide exposure, and research
has demonstrated that chemicals
can permeate "impermeable" gloves
without degrading their physical
characteristics. To aid OPP and
pesticide users to identify gloves
that will provide adequate
protection, EPA contracted Arthur
D. Little, Inc., to: 1) produce,
assemble and evaluate data on the
effectiveness of polymer gloves
against pesticide exposures, and
2) identify correlations in the
collected data that will
facilitate glove selection. Three
types of testing are being
employed in this project: two
"rapid screening" procedures to
eliminate glove-chemical
combinations that result in
observable, rapid degradation of
glove polymers and ASTM method
F-739-85 to test for permeation.
Degradation tests on 92
pesticide-glove combinations have
been completed and at least 60
combinations are being selected
for subsequent permeation testing.
Permeation tests on two pesticides
are in progress. As the project
proceeds, the two rapid screening
tests will be compared and the
more satisfactory will be selected
for future use. Also, pesticides
which have low water solubility
and volatility do not lend
themselves to the ASTM permeation
test due to the difficulty of
collecting the permeant after it
has passed through the glove
sample. An effort is underway to
-211-
-------
identify alternative methods for
collecting the permeant.
Thermal comfort is a key
factor in the selection of
chemical protective garments for
agricultural operations. Existing
protective clothing provides a
trade-off between thermal comfort
and pesticide penetration
resistance. To obtain a better
understanding of this trade-off,
EPA entered a cooperative
agreement with the University of
Tennessee under which 40 fabrics
were screened for thermal comfort
properties and 32 were tested for
penetration resistance against
four pesticide formulations:
Dicofol, Ethion, Chiorobenzilate
and Terrazole. The results of
this work will be used to identify
fabrics with superior combinations
of thermal comfort and penetration
resistance. These fabrics will be
fabricated into coveralls and
field tested by citrus and
greenhouse workers as one part of
a cooperative agreement between
EPA and the University of Florida.
Under this agreement, the
University of Florida is
investigating the dermal pesticide
exposure received by greenhouse
workers and the mitigation of that
exposure through the use of
protective clothing. In phase I
of the field test, raw data was
collected by having pesticide
sprayers wear collection patches
on the inside and outside of their
clothes. The patches were then
removed and subjected to
extraction and chemical analysis.
In addition, pesticides were
removed from the hands with
ethanol rinses and breathing zone
air was sampled. Approximately
3000 pesticide samples were
generated in this phase, which
included various application
methods (e.g., tractor-drawn boom
and span, handgun, pulse fog and
drench spraying) in various
settings (e.g., open, enclosed and
partially enclosed greenhouses).
In the second phase, begun in
September 1986 and to be completed
soon, the effectiveness of work
pants, long-sleeved shirts,
T-shirts and two commercial brands
of coveralls will be evaluated.
Field evaluations of the fabrics
identified in the University of
Tennessee study have also begun
and will continue through 1988.
In addition to these field
and laboratory evaluations, the
Personnel Protection Research
Program has also initiated
development of a Guidance Manual
for Selecting Protective Clothing
for Agricultural Pesticide
Operations, with Arthur D. Little,
Inc., serving as the contractor on
the project. In the first phase
of this effort, an Interim
Guidance Manual (IGM) was prepared
as an internal resource document
for the OPP. The IGM focuses on
assessments of the effectiveness
of polymer gloves against
pesticide exposure and of the
limitations, imposed by thermal
comfort needs, on the use of
protective garments in
agricultural pesticide operations.
The IGM has been circulated among
the research and user community,
and the scope of the final
Guidance Manual will be based on
their comments, and on those of
OPP. Topics in the final version
may include: work task functional
requirements; manufacturing
methods and materials; performance
data; human factors data; product
lists; decontamination agents and
methods; and foot, eye, head,
forearm/arm and respiratory
protection.
RESEARCH IN SUPPORT OF
TSCA-MANDATED OTS ACTIVITIES
Because OTS is limited to 90
days to complete a PMN review, and
because they want to minimize
additional testing by
manufacturers, the Office requires
reliable model(s) or system(s) for
predicting the effectiveness of
protective clothing and
respirators in the workplace.
Therefore, to support these needs,
the Personnel Protection Research
Program is evaluating and
improving methods, equipment and
-212-
-------
procedures for estimating the
exposure protection provided by
these products. As methods are
identified or developed, they are
being incorporated into the PMN
review process, giving OTS the
opportunity to provide prompt
feedback.
In the first phase of these
model development efforts several
theoretical permeation models and
test methods for estimating
permeation related properties were
identified and compared to the
results of permeation tests. The
models and test methods chosen
were based on theories of the
solution thermodynamics of
polymer/solvent systems and the
diffusion of solvents in polymers.
The contractor further developed
these models and test methods to
estimate the solubility (S) and
the diffusion coefficient (D) for
a solvent in a glove polymer.
Given S and D, the permeation of a
glove by a solvent can be
predicted for various exposure
conditions using analytical or
numerical solutions to Pick's
1 aws.
Based on this work, a
computer model (written in Fortran
and operable on an IBM PC) was
developed and delivered to OTS for
hands-on evaluation. Using simple
inputs, this program can calculate
S and D, and generate curves used
in estimating the barrier
effectiveness of protective
garment materials. Also, the
capabilities and reporting
requirements of existing
permeation test methods were
analyzed and a preliminary
hierarchy for test method
specification was proposed.
Analyses of immersion and vapor
sorption tests are underway and
development of a splash test
method, in cooperation with
standards development
organizations, will be pursued in
1987. The results of these
efforts, which will continue
through 1989, will be integrated
into a system that will enhance
the performance of PMN reviews.
In addition to evaluating the
effectiveness of protective
clothing, OTS must also assess the
likely effectiveness of any
respirators included in a PMN.
Research to enhance OTS
capabilities in this area is being
conducted through an interagency
agreement with the National
Institute for Occupational Safety
and Health (NIOSH). This research
will determine the workplace
protection factors of respirators,
develop laboratory test methods to
predict these factors and
establish a decision logic for OTS
evaluation of PMN submittals
involving respirators.
The initial phase of this
work involved an assessment of
quantitative fit test methods,
which measure the effectiveness of
a respirator by comparing the
concentration of a test
contaminant inside the facepiece
to that outside. The reliability
of this test is critical to
meaningful measurements of
workplace protection factors, and
to individuals who use fit test
data to select respirators.
Laboratory evaluations
demonstrated that the measured
protection factor can be strongly
biased by the position of the
sampling probe and the location of
leaks into the facepiece. Work in
1987 will focus on eliminating
this bias.
RESEARCH IN SUPPORT OF
SARA-MANDATED EPA ACTIVITIES
A number of Personnel
Protection Research Program
projects support SARA-related
hazardous waste site cleanup and
emergency spill response
activities. These projects are
identifying, evaluating and
improving prototypical and
commercially available chemical
protective materials, clothing and
equipment, and related procedures,
that could potentially improve the
safety, range, and
cost-effectiveness of EPA and
EPA-contractor operations at
chemical spill and uncontrolled
-213-
-------
hazardous waste sites. Products
of particular interest include
disposable protective clothing,
vital signs monitors, personal
cooling devices, respiratory
protection devices, new garment
materials, totally encapsulating
ensembles and personal
communication devices.
Recent Program efforts to
evaluate and improve protective
clothing materials include an
assessment of the stiffness and
strength of Teflon /Nomex
laminate (a newly available
material resistant to a wide range
of chemical compounds) at cold
temperatures and an investigation
of the permeation and degradation
resistance of a 20 mil chlorinated
polyethylene (CPE) to liquid
chemicals. In the latter project,
CPE swatches were exposed to ten
chemicals (acetic acid, acetic
anhydride, acetone, bis
(2-chloroethyl) ether, carbon
tetrachloride, ethylene diamine,
isopropyl alcohol,
nitrosodimethylamine, phenol and
xylene) in order to determine
permeation rates, breakthrough
times and swelling. Mean
permeation breakthrough times ,
ranged from 0 to 170 mg'm" 's~ ,
and swelling and solubility data
also showed wide variation,
depending upon the chemical.
Soaking CPE test swatches in one
chemical resulted in a loss of
weight, while contact with four
other compounds resulted in weight
gains of over 100% by CPE. In
eight out of ten cases chemical
contact also reduced the capacity
of CPE to resist tearing.
In an interagency effort with
the U.S. Army's Chemical Research
and Development Center, advanced
development of a prototype,
long-term (2.5 hour),
self-contained, chemical
protective ensemble (based on an
existing Army/Coast Guard
prototype) was pursued. The
ensemble development was performed
by USD Corporation. Some
desirable modifications (e.g.,
lighter weight, simplified
maintenance and 2.5-hr service
life) were attained, but the
breathing apparatus failed NIOSH
certification testing twice. No
additional attempts to obtain
certification are planned. In
.related projects, the high
pressure oxygen compatibility of
the apparatus' materials of
construction and the
state-of-know!edge of the
physiological effects of routinely
breathing high concentrations of
oxygen were evaluated.
Other recently completed
Program efforts to improve
equipment include a preliminary
investigation of critical design
features of back-mounted equipment
(to minimize risk of back
injuries) and a field evaluation
of three recently developed vital
signs monitors, which are used in
heat stress management programs at
site cleanup operations. This
evaluation was conducted at a
dioxin cleanup site in Missouri to
assess the monitors' ease of use,
suitability for field operations
and accuracy. Further evaluations
of vital signs monitors and
evaluations of personal cooling
systems are planned for early 1987
in order to develop modified heat
stress management work practices.
Also in 1987, a project will
begin to identify and evaluate
commercially available or
prototypical personal hazard
detectors. This effort will
involve desk-top, laboratory and
field evaluations of personal
toxic gas, combustible gas and
oxygen deficiency detectors.
Evaluation criteria will include
reliability, ease of use,
sensitivity, portability, cost and
safety. Initial results indicate
that, under field conditions,
currently available detectors
exhibit some generic failure
modes: failure below 0 C and in
precipitation, lack of portability
and sensitivity, and selectivity
(or lack thereof).
With regard to evaluating and
improving personnel protective
-214-
-------
procedures for Superfund
operations, two Program efforts
are underway: development of a
field test kit for protective
clothing and evaluation of
decontamination agents and
methods. The field test kit has
been laboratory tested and the
results compared to those obtained
using ASTM Method F-739-85
(permeation test method). Three
kits are undergoing field testing
by EPA response groups. The kit
is relatively inexpensive and
readily field usable: Liquid
waste is placed in a shallow cup,
the cup is covered by a swatch of
clothing material and inverted,
and the rate of permeation is
monitored gravimetrically using a
two-place, battery-powered
balance.
COORDINATION ACTIVITIES
In a critically important set
of activities, the Personnel
Protection Research Program
continually promotes the transfer
-- both inward and outward --of
information on personnel
protection research needs,
activities, plans and outputs;
analyzes EPA needs in this area;
and sponsors investigations of the
state-of-the-art in this field.
Specific coordination and
technology transfer activities in
1987 will include: continuing
active participation in ASTM
Committee F-23 on Protective
Clothing, making presentations on
research progress at the Second
International Symposium on the
Performance of Protective Clothing
and the American Industrial
Hygiene Conference, participating
in an Interagency Memorandum of
Understanding Work Group with
representatives of the U.S. Coast
Guard, NIOSH, Occupational Safety
and Health Administration and
Federal Emergency Management
Agency; conducting a second
intra-EPA Workshop on Personnel
Protection Research and Research
Needs; and participating on the
American Conference of
Governmental Industrial Hygienists
Committee on Agricultural Health
and Safety.
Similar activities are
planned at least through 1991, and
are expected to continue to
provide ideas and information of
value to the planning and
performance of personnel
protection research in support of
EPA's needs.
-215-
-------
TABLE 1. SUMMARY OF EPA PERSONNEL PROTECTION PROGRAM OUTPUTS
Title
Long-term, Self-contained Chemical Protective Ensemble
el f- i ned
Decontamination Techniques for Mobile Response Equipment Used at
Waste Sites
Evaluation of Chlorinated Polyethylene Protective Garment Material
Test Methods to Predict the Permeation of Polymers by Organic Solvents
Interim Protocol for Diving Operations in Contaminated Water
Predicting the Effectiveness of Chemical Protective Clothing-- Model
and Test Method Development
Biases Associated with In-facepiece Sampling of Respirators
Interim Guidance Manual for Selecting Protective Clothing for
Agricultural Pesticide Operations
Review of Models for Predicting the Effectiveness of Protective Clothing
Effect of Fabric Weight and Thickness on Pesticide Penetration
Evaluation of Protective Clothing for Agricultural Pesticide Operations
Predicting Breakthrough of Chemicals through Protective Clothing
Field Evaluation of Protective Clothing: Experimental Design
Field Test Method to Evaluate Protective Clothing
Interpretation of the Results of Permeation Testing
Evaluation of Protective Gloves Used in Agricultural Pesticide Operations
Product
Type*
CP
PR,S
PR,S
CP
PR,S
PR,S
PR,S
IR
OP
IR
CP
CP
CP
CP ,
CP
OP
OP
Date
5/84
8/85
12/85
1/86
8/86
9/86
9/86
9/86
6/87
12/86
1/87
1/87
1/87
1/87
2/87
6/87
6/87
*OP - oral presentation, CP = article in conference proceedings,
PR,S » project report and summary, and IR = internal report
-216-
-------
APPLICATION OPPORTUNITIES FOR CANINE OLFACTION:
EQUIPMENT DECONTAMINATION AND LEAKING TANKS
Herbert S. Skovronek, Barbara Kebbekus, and Stewart Messur,
New Oersey Institute of Technology, Newark, N3 07102
Lorenz D. Arner, Biosensors, Inc., Skippach, PA 19474
Hugh Masters, US Environmental Protection Agency, Edison, NO 08837
ABSTRACT
Rapid screening of heavy equipment used in site cleanup for residual contamination
and scanning of underground storage tanks for leaks were identified as two promising
environmental applications for canine olfaction.
In equipment decontamination, the objective was to demonstrate that a trained dog
could detect and indicate extremely small residues of hazardous chemicals remaining on
heavy equipment such as bulldozers, baekhoes and front end loaders after washup. Using
xylene and 1,1,1-trichloroethane as models of common hazardous chemicals, a trained dog
reliably indicated hidden samples emitting as little as 0.5 ug/min. Gaussian dispersion
models Indicated that the dog is detecting 5-10 ppt or less at these emission rates.
Field tests on equipment indicated detection at emission rates as low as 1 ng/min.
From tests to evaluate the dog's ability to differentiate similar compounds, it was
concluded that, at least in some compound families, the dog does respond to both the
compound used for training and its congeners. This capability may be useful in finding
any members of such families at a site.
Gasoline was selected as the material of greatest importance when searching for
underground leaks. Water-washed gasoline, used to simulate underground leaks, did exhibit
minor changes in composition. However, possibly due to the training approach used, the
dog was unable to differentiate the washed gas from unwashed gasoline. Alternate
approaches are delineated for future study.
INTRODUCTION
Passage of the Resource Conservation
and Recovery Act (RCRA) and the Compre-
hensive Environmental Response Compensation
and Liability Act of 1980 (CERCLA), common-
ly known as Superfund, and most recently,
the Superfund Amendments and Reauthoriza-
tion Act of 1986 (SARA), has placed great
emphasis on the cleanup of hazardous waste
sites and spills. Careful and complete
delineation and monitoring of such sites,
both before and after cleanup, have become
increasingly important. The cost and
complexity of such monitoring programs have
been far from inconsequential and have led
to the development, testing, and use of
highly sophisticated analytical equipment
to provide more reliable results more
quickly and at lower ambient levels. In
some cases, however, highly sophisticated
technology is used solely because it is the
only or the fastest approach accepted by
the technical and regulatory community,
even when precise data are not needed.
In 1983, an alternate, innovative
approach to monitoring was brought to the
attention of the Edison, NO Releases
Control Branch of EPA's Hazardous Waste
Environmental Research Laboratory (Cinc-
innati). The technique, canine olfaction,
-217-
-------
has attained wide acceptance as an aid to
the military, to police, to rescue workers,
to custom agents, and to certain industries
to detect explosives, weapons, personnel,
narcotics, leaking gas, underground pockets
of leaking cable oils, and even termites,
often under conditions WHERE NO OTHER TECH-
NOLOGY IS EQUALLY CAPABLE. In some of
these areas, the dog remains the "state-of-
the-art".
The results of a preliminary study
funded by US EPA's Releases Control Branch
in 1983 to study the feasibility of the
technique and the breadth of its potential
applications essentially confirmed
expectations (1). The dog was, if any-
thing, more sensitive to chemical vapors
than expected, detecting "plumes" from
nuniscule samples at considerable distances
and tracking them back to their sources;
potentially providing environmental workers
with a tool to zero in on contamination
very quickly so that quantitation could
then be carried out in the most cost-
effective manner and without numerous "not
detected" results. A great deal also was
learned about the techniques suitable for
training dogs for such work, environ-
mental situations where a dog might be
useful and the reluctance of regulators,
contractors, analysts, and others
accustomed to dealing with instruments and
numerical results.
The current Cooperative Agreement (2)
between the USEPA and the New Gersey
Institute of Technology was undertaken to
explore the dog's ability to work usefully
in two different areas of considerable
interest to USEPA, cleanup contractors, and
state agencies. These two problem areas
were:
1. Assessing the effectiveness of heavy
equipment decontamination after use at a
site; and
2. Non-invasive screening for leaks from
underground storage tanks (UST),
particularly those containing gasoline.
While there have been no applications
of the dog in the environmental field per
se, there have been a number of reports
that really fit this category. Dogs were
used to detect 150 leaks, some too small to
confirm with conventional instrumentation,
over the 94 mile length of a new natural
gas pipeline across Canada, all within a
two-week period (3). They also were used
to locate leaks of electrical cable oils
below the paved streets of New York City
(4). Experiments have even demonstrated
that dogs were able to detect nitrogen,
Freon, and helium gas emanating from a
manifold system (5).
For the two phases of this project,
"decontamination" and "UST", two chemicals
were selected to represent hazardous
materials commonly encountered at sites.
These were m-xylene and 1,1,1-trichloro-
ethane (TCE). Xylene was chosen as a less
volatile, less toxic representative of the
aromatic solvents benzene and toluene
frequently encountered at spills and
cleanup sites. Of course, xylene is also a
component of gasoline and fuel oil.
Similarly, the non-carcinogenic
trichloroethane was chosen to represent the
chlorinated hydrocarbons frequently found
in contaminated groundwater and attributed
to industrial degreasing operations.
The project was carried out with one
primary dog, a 5-year old German Shepherd
Dog, Ramos, who had already demonstrated
his ability and desire to do scent work in
other areas. Ramos was trained and handled
by Mr. L. Don Arner. A second dog, a
female German Shepherd Dog named Anja, and
her handler, Mr. Don Bowman, were brought
into the project later for specific
segments.
The training protocol relied on
positive reinforcement (reward of some
type) for all successes, no matter how
slight and the withholding of reward for
failure (negative reinforcement). Rarely
if ever was discipline (aversion training)
necessary. The training further used a
technique called "chaining" in which each
small segment of the training is taught
separately and only combined as the dog
succeeds with the first effort. For
example, the dog,learns to search, he
learns to detect a specific odor, he learns
to respond by pawing or retrieving. Each
dog was taught to recognize a particular
scent, that of the target chemical, by
introducing the dog to a somewhat higher
level of the chemical (up to 0.5 gm)
protected in a hollow wooden dowel or a 35
mm film canister and rewarding the dog for
the very first alert or recognition. The
amount of chemical was then rapidly reduced
and/or the difficulty of finding the source
of the chemical was increased. In fact,
-218-
-------
the dog quickly learned to recognize and
detect both target chemicals (TCE, xylene),
even when the source was reduced to one
drop (about 0.05 gm). Because of the
project's goals, the objective was to
develop the dog's ability to detect low
concentrations of the vapor rather than his
ability to recognize aged samples or track
an aged scent great distances, as had been
demonstrated in the earlier study.
Calculations using a simple Gaussian
air dispersion model (6) were carried out
to provide some insight into the airborne
level of chemicals the dog might be
perceiving at different distances from a
source. These calculations suggested that
with a source emitting at 0.5 ug/min in a 5
mph breeze the dog could be responding to
as little as 1-5 parts per trillion in the
air, depending on how far from the source
he gave the first alert. Figure 1 is a
plot for such a model calculation.
Laboratory research by other investigators
has indicated that the dog was capable of
detecting chemical odors at the parts per
quadrillion level (7).
XYLENE CONCENTRATION
DOWNWIND
CENTERLINE
GROUND LEVEL
.0.5 ug/min, 5 mph
.0.5 ug/min," 1 mph
Figure 1.
i 2 3 4 5 T
DOWNWIND DISTANCE (meters)
Calculated plume concentrations
for xylene.
To work at even lower emission levels
and also to provide a constant emission
rate, as might be expected from a contam-
inated piece of heavy equipment or a tank
leak, permeation tubes, consisting of
sealed lengths of plastic that allow vapors
to permeate through the walls at uniform
rates as long as liquid is present (Figure
2). These efforts did not deter the dogs.
Ramos readily and happily found permeation
tubes emitting at only 0.4 ug/min, the
practical limit for calibration by weight
loss over a reasonable length of time.
Even a commercial tube calibrated as
emitting xylene at 10 ug/min at 90 C (the
calculated emission rate at 25 C is only
0.062 ug/min) was readily found by both
dogs. Of course, the tubes were never
handled without gloves and were never
touched by the dogs to avoid imparting
other scents as clues. From later work, it
appears that the dog is probably still one
or more orders of magnitude away from the
CANINE DETECTION LIMIT!
Figure 2. Permeation tubes used in
training.
DECONTAMINATION
To familiarize the dog with the type
of searches that could be encountered,
samples, either a drop on a surface or in a
dowel or film canister, or permeation
tubes, were hidden on, under, or in
"available" heavy equipment vehicles. The
dog was then brought to the site and
encouraged to search, usually first
"scanning" the entire vehicle on his own
and then proceeding to a more detailed
search under the guidance of the handler.
With rare exception, the article was
located quickly.
-219-
-------
To simulate a real situation, permis-
sion was obtained to use a front end loader
at a construction site in Morris Plains,
New Oersey. A mud consisting of 5 g of
alumina powder, water, and xylene was
applied to a spot on the vehicle and the
dog brought in to search. Blanks consist-
ing only of alumina plus water were also
applied. This sequence was repeated three
times, placing the xylene-tainted mud at
different points. In the first test, the
dog localized the odor but did not actually
locate the sample, which was on the
underside of the beam supporting the
shovel. In the second test, the dog found
the sample (in the opening of one of the
lower frame members) after considerable
searching and with some guidance. In the
third test, the dog quickly found the
sample on a horizontal surface at the rear
of the left track. (Figure 3)
Figure 3. Ramos searching vehicle.
Throughout the tests, it was observed
that the dog was very anxious to get under-
neath the vehicle. This behavior plus a
strong diesel fuel odor led the investi-
gators to discover a very significant fuel
oil leak from a ruptured line. Diesel fuel
had saturated the ground under the vehicle
and it was this odor that the dog was
concentrating on. Since diesel fuel is
rich in xylene, the dog's success in
locating added sources becomes even more
amazing. Even more intriguing was the •
dog's considerable interest in the front
bumper of one investigator's car during a
rest period. A sample alumina/xylene mud
had been placed on the bumper the day
before as practice.
To obtain a semi-quantitative idea
of the level of xylene that the dog was
detecting in these tests, air samples were
collected directly from each mud
immediately after the dog had completed his
search, as well as from the general area
(background). The sample from test 2 was
lost when the dog pawed it. These samples
were collected by holding an inverted 2.5"
funnel 0.5-1.0 inches over the mud and
drawing air into a Tenax adsorbent trap for
10 minutes at 3-4 cc/min (Figure 4).
Figure 4. Equipment used to collect air
sample over alumina mud.
The traps were then returned to the labora-
tory, desorbed thermally, and analyzed by
gas chromatography (GC), giving the levels
of xylene noted in Table 1.
TABLE 1. DETECTED XYLENE CONCENTRATIONS
Sample Cone. Emission Rate
(ppb) (ng/min)
background
sample. #1
sample. #2
sample. #3
car bumper
10
68*
LOST
24
517
0.3
1.0
-
0.77
16.0
*sample funnel contaminated with
mud, result high.
-220-
-------
It must be concluded that the dog is able
to detect and locate a source of xylene
(simulated residual contamination in this
case) — even when a significant amount of
the same vapor is present in the background
air.
In a second set of experiments at
another location, xylene-contaminated
alumina muds were applied to two of five
hand-shovels placed on snow-covered ground
and a blank water/alumina mud was applied
to a third one. The dog found one xylene
sample (A) rapidly, but he showed little
interest in the second xylene sample (B)
until directed to search that shovel. The
dog also did show considerable interest in
the blank shovel (C), but not at the mud
itself. Instead, his interest seemed to be
concentrated on an area to which some of
the melting snow and mud may have run. Air
samples were again taken and analyzed by
GC. While the GC of the air trapped over
the blank shovel showed no xylene, there
was another, as yet unidentified, peak.
TABLE 2. XYLENE DETECTION ON SHOVELS
Sample
shovel A
shovel B
Blank C
background
Cone.
(ppb)
23
11
0
0
Emission Rate
(ng/min)
3.6
0.2
0
On the basis of these experiments, it
appears that the dog can be a valuable
asset to EPA and cleanup contractors trying
to establish whether equipment can be
removed from a site. Since the decision to
use the dog and the selection of the target
compound can be made while the cleanup is
proceeding, there would be more than
adequate time to train or condition the dog
to the compound(s) of interest. And, as
noted earlier, the dog would be used only
to point out residual contamination,
whether on heavy equipment, hand tools, or
even personnel protective gear. It remains
the responsibility of the cleanup crew to
clean the indicated area further or
"measure" the dog's find by wipe tests and
quantification.
CHEMICAL DISCRIMINATION
One question kept surfacing as work
progressed. How selective was the dog?
Would 129 dogs be needed to find the 129
priority pollutants (and more for each
chemical 'on the RCRA lists) or could one
dog find them all? And, from the opposite
point of view, would a dog trained to find,
for example, xylene, give false positives
by also finding Parathion, trichloroethane,
acetic acid, diesel fuel, and anything else
in the area? These were interesting
questions and the team set out to get some
insight into what would happen.
Field-testing the question of
selectivity presented a unique problem. It
was realized that while the goal was to
observe the dog's natural response
(retrieve, curious, ignore, etc.) to any
new chemical that had not previously been
used in training, this natural response
would soon be replaced by a trained or
learned response. For example, rewarding
the dog for finding substances other than
the original target (trained) chemical
would be training him; discipline also
would be training him (aversion training)
not. to indicate those. Even no reaction by
the handler to "an improper find" could
cause the dog to work harder to gain
approval — by seeking more of the same
odor! Consequently, it was anticipated
that only limited testing could be done
before a learned response would be
produced.
The initial experiments consisted of
choices between structurally-similar
chemicals, xylene and ethyl benzene, or
trichloroethane and trichloroethylene,
placed on undisturbed twigs several feet
apart. Human odor thresholds for the
chlorinated hydrocarbons are higher (less
sensitive) than for the aromatics (Table
3). With xylene and ethyl benzene
TABLE 3. CHEMICAL ODOR THRESHOLDS
Chemical
trichloroethane
tr i ch 1 or oethy 1 ene
toluene
xylene
ethyl benzene
n-amyl alcohol
Threshold
(ppm)
20.0
20.0
2.0
0.5
2.0.
5.0
-221-
-------
the first reaction to the ethyl benzene
appeared to be recognition that the odor
was "interesting (i.e., similar) but not
quite right". However, in repeat tests the
dog then did retrieve articles tainted with
ethyl benzene, probably the learned
behavior expected from the noted uninten-
tional training. In experiments with
chlorinated compounds, the dog initially
ignored the trichloroethylene, but began to
show awareness in subsequent experiments.
The tendency to retrieve trichloroethylene
articles never was as strong as that for
ethyl benzene.
In a more complex experiment, undis-
turbed twigs 2-3 feet apart were impreg-
nated with 1 drop of water, amyl alcohol
(dissimilar odor), xylene, and ethyl
benzene. When allowed to explore the area
freely, the dog noted but ignored the amyl
alcohol, showed considerable interest (and
uncertainty?) about the ethyl benzene, and
ultimately retrieved the xylene.
Because of concern about "incidental
training", this question of selectivity was
not pursued further. It was the research
team's opinion that the find of any contam-
ination close in character to that of the
target material would be useful and worth
investigation. And, while it was unlikely
that a completely different chemical odor
would be present at a site, if it was,
monitoring personnel should be interested
in knowing of its presence.
Thus, after several months of
developing and modifying field techniques
the program had produced a dog that could
be reliably used to locate small residues
of chemicals such as might persist on heavy
equipment even after decontamination.
Where wipe tests only have a random proba-
bility of "hitting" a contaminated spot,
the dog could assist the contractor by
quickly indicating where such residues were
— even when inaccessible or invisible to
human observation ~ and could quickly
evaluate the effectiveness of subsequent
cleaning attempts. Of course, the results
of wipe or other tests at residues detected
by the dog could, at some point, be inter-
preted as "de minimus" because of level or
size of the area.
LEAKING UNDERGROUND TANKS
The area of leaking underground
tanks was, as anticipated, a much more
challenging problem. The goal was to use
the intelligence and olfactory acuity of
the dog to detect any vapors rising to the
surface from a leak, as reported with leaks
from buried natural gas lines and
electrical cable oils. The concern was,
however, that the dog would also find any
other sources of the target compound,
including every surface spill, drips at
loading valves, etc.
In seeking means to avoid this
problem, it was realized that while pure
chemicals stored in tanks would not change
when leaked, a complex mixture such as
gasoline might undergo significant extrac-
tion and/or fractionation as the vapors
from a leak permeated through the soil to
the surface. Such changes could, it was
hypothesized, provide a vapor profile
sufficiently different from that of surface
spills so that the dog could differentiate
the two. Attempts to simulate this
permeation phenomenon by analyzing gasoline
vapors passing up through soil columns were
not successful. Very little vapor was pro-
duced and no significant differences in
profile were observed.
In an alternate approach, regular
unleaded gasoline was extracted with
water. Gas chromatographic analysis of the
washed gasoline did indicate a reduction in
the lower aromatics, benzene, toluene, and
xylene. Efforts were then directed to
training the dog Ramos to detect this dif-
ference, primarily the ABSENCE of xylene,
with which he was familiar. In spite of
several approaches and many attempts, how-
ever, a reliable, preferential response by
the dog to the washed gasoline was never
achieved. To this time it has not been
possible to devise a training scheme or a
search protocol that would provide the
desired selective detection for vapors
leaking from an underground tank.
It is unlikely that we will be able
to solve this problem within the
constraints of this project, and three
challenging questions remain. First, the
dog used in the gasoline work had,
unfortunately, already been trained to
detect xylene. Perhaps a dog without that
training would be more sensitive to the
different profiles of washed and unwashed
-222-
-------
gasoline (the xylene is not totally removed
by extraction). The second dog, now being
trained only for chlorinated hydrocarbons,
may offer the opportunity to test this
point.
Second, to date the dog has been
asked to identify —- and differentiate —
the vapors emanating from small (1-2 drops)
liquid samples of washed and unwashed
gasoline. However, gas chromatography
suggests that there may be significant and
rapid changes in the vapor profiles over
these samples that would not be comparable
to the "steady state" that would exist over
an underground tank leak. Diffusion tubes,
vials equipped with small-bore vents,
should provide more uniform vapor profiles
for washed and unwashed gasoline that may
be more readily differentiated by the dog.
Unfortunately, preliminary gas chromato-
graphic analyses exhibit no differences in
the vapor profiles for the two samples.
Third, the investigators' experience,
and conversations with other trainers, ,
animal psychologists, and veterinarians
have failed to uncover any prior evidence
that an animal can be trained to respond to
the ABSENCE of a stimulus. If that is
true, a completely different training
approach may be needed. One such approach
would be the introduction of small amounts
of marker chemicals. While it is expected
that a dog would find a properly selected
marker at the right concentration, just as
dogs found the butyl mercaptan marker in
natural gas, supplier resistance to
contamination of their products at any
concentration is anticipated. In addition,
all surface leaks, spills and other sources
at a site would soon become contaminated
and would, presumably, be indicated by the
dog.
CONCLUSIONS
1. The dog/handler team is a viable tool
for EPA and their cleanup contractors to
use in evaluating the effectiveness of
decontamination efforts on equipment.
Presumably, the effectiveness of cleanup of
building surfaces as well as land sites,
hand tools, and protective gear could also
be evaluated by the dog.
2. The dog shows a degree of selectivity
for similar compounds that is useful but
not absolute. Greater selectivity could,
presumably, be achieved by training, but it
is doubtful if that would have any
advantage.
3. A viable technique for using the dog to
detect gasoline vapors from leaking
underground tanks only has not yet been
developed.
REFERENCES
1. Arner, Lorenz D., Glen, R. Johnson,
and Herbert S. Skovronek, Dec. 1985.
Delineating Toxic Areas by Canine
Olfaction, US Environmental Protection
Agency Project Summary,
EPA/600/S2-85/089.
2. New Jersey Institute of Technology,
Canine Olfaction: Evaluation of Canine
Olfaction Technology for Detection of
Hazardous Substances. USEPA
Cooperative Agreement No. CR
812180-01-3, 1985.
3. Johnson, Glen R., Tracking Dog Theory
and Methods, 1977, Arner Publications,
Westmoreland, New York, pp 14-21.
4. Johnson, Glen R., The New York
Experiment, June 1981, Off-Lead, pp
10-13.
5. Johnson, Glen R. Odorless Gas
Detection by Domestic Canines, December
1977, Off-Lead, pp 18,19.
6. Turner, D. Bruce, Workbook of
Atmospheric Dispersion Estimates, 1970,
US. Dept. of Health, Education and
Welfare Publication No. 999-AP26.
7. Moulton, D. G. and D. A. Marshall, The
Performance of Dogs in Detecting
alpha-Ionone in the Vapor Phase,
1976, Journal of Comparative
Physiology, 110, pp 287-306.
-223-
-------
NONDESTRUCTIVE TESTING (NOT) FOR LOCATION OF CONTAINERS BURIED IN SOIL
Arthur E. Lord, Jr. and Robert M. Koerner
Drexel University
Philadelphia, PA 19104
ABSTRACT
At the 12th Annual Hazardous Waste Research Symposium held at Cincinnati in 1986 the
authors reported on their work concerning the nondestructive testing (NOT) for location of
containers buried in soil. An overall view was presented at that time. In this paper
more detail is given about certain aspects of the testing, which could not be included
previously due to space limitations.
Experimental work is described where seven techniques were reduced to four for the
majority of the testing. (Originally 17 techniques were considered - 10 were eliminated
during the literature search.) The four techniques:
. metal detector (MD)
. electromagnetic induction (EMI)
. magnetometer (MAG)
. ground probing radar (GPR)
were looked at in considerable detail. In particular, results concerning the ability of
each method to detect the container(s) when not travelling directly over the container(s)
(the lateral scan sensitivity) are given. Also detailed results, in the form of response
contour diagrams, are given in the case of a "metal trash dump."
The effect of steel container burial orientation on the GPR is presented. Water
table depth determination (to 15 feet) with GPR is also demonstrated. Under near perfect
conditions of very little interference (i.e., low electrical conductivity, highly homo-
geneous, dry soil and absence of power lines, metal objects etc.), it is possible to
detect empty plastic drums to a depth of.three feet with EMI.
INTRODUCTION
An overview of the results of our
work concerning the location of buried con-
tainers using various nondestructive (re-
mote sensing) techniques were presented at
the 12th Annual Hazardous Waste Research
Symposium (Lord and Koerner, 1986a). This
paper briefly reviews that work and pre-
sents additional detail concerning portions
of the work which could not be presented
earlier due to space limitations.
The investigation of subsurface ob-
jects can be approached in two very dif-
ferent ways. The first type is by use of
a suitable destructive test method. This
category includes: test pits, excavation
trenches, auger holes, core borings, and
observation wells. While one does indeed
"see" the subsurface materials as they are
excavated for ease of examination of sub-
sequent testing, such methods have certain
drawbacks in identifying and locating
buried containers (Lord and Koerner, 1986a).
The second type of approach to identi-
fy and locate buried containers is by the
use of a suitable non-destructive testing
(NOT) method; These have also been called
remote sensing or short range geophysical
techniques. Within this category are the
following methods which have either been
used or appear to have applicability:
-224-
-------
seismic reflection, seismic refraction,
electrical resistivity, electromagnetic in-
duction, induced polarization, metal detec-
tor, magnetometer, continuous microwave,
pulsed radio frequency (ground penetrating
radar), infrared radiation, sonar (pulse
echo acoustic), electrical self potential,
optical (below water surfaces), and pene-
trating radiation. All of the above
methods are not equally suited for identi-
fying and locating buried containers.
Seventeen techniques were identified; as a
result of the literature search ten were
eliminated from further consideration;
seven were evaluated further in the experi-
mental phase of the project and are de-
scribed in the following paragraphs.
DESCRIPTION OF METHODS
The metal detector (MD) and electro-
magnetic induction (EMI) methods, are both
inductive methods. A transmitting coil
sends a continuous electromagnetic signal
to a receiving coil. As a simplified de-
scription, the signal arrives at the re-
ceiver essentially through two major paths.
One path is through the air and the induced
signal does not change with the search
position. The other path is through the
subsurface material and is affected mainly
by the local electrical conductivity of
material involved. If an anomaly in the
subsurface conductivity is encountered,
e.g., a buried metal drum, the induced
signal received through the earth path is
changed significantly and the instrument
indicates this change accordingly. Refer
to Lord, et at. (1982a) and McNeil! (1982),
where the commercially-available instru-
ments used in these studies are described
i n detai1.
The ground penetrating radar (GPR)
method operates on exactly the same princi-
ple as ordinary aircraft radar. A short
pulse of electromagnetic radiation is
beamed in the ground by a special highly-
damped antenna and reflections occur from
any subsurface discontinuity in dielectric
constant. The reflected pulse arrives
back at the receiving antenna and a dis-
play of reflected intensity versus depth
is presented on an oscilloscope and on a
recorder. This commercially-available
technique is described more fully in
Bowders, et al. (1982).
The,magnetometer (MAG) method mea-
sures the local magnetic field strength
(essentially the earth's field) and, with
it,'any changes in this magnetic field.
The type used in this study was a proton
precession model. The local magnetic field
is determined by measuring the precession
frequency of the proton magnetic moment..
This rate is linear in the magnetic field
and, as the frequency can be measured very
precisely, the magnetic field can also be
measured very accurately. (The source of
precessing protons is an organic liquid;
the liquid, generator and associated fre-
quency measuring apparatus are contained
•in the MAG unit). A steel drum, being
ferromagnetic, changes the local value from
the earth's magnetic field and hence can
be detected. The MAG technique (commer-
cially-available) is described in more de-
tail in Tyagi, et al. (1983).
The electrical resistivity (ER) method
applies current to the ground through
electrodes and depends, for its operation,
on the fact that any subsurface variation
in conductivity alters the form of the
current flow within the earth. Therefore,
the distribution of electrical potential
is affected. The degree to which the po-
tential measured at the surface is affected
depends on the size, shape, location and
electrical resistivity of the sub-surface
mass. It is therefore possible to obtain
information about the sub-surface distri-
bution of various bodies at the surface.
This method is used extensively in the oil
and mineral prospecting area, but has not
been widely used in shallow monitoring for
small buried objects. It has seen some
use for tracing sub-surface liquids.
There are many sources of commercial equip-
ment to choose from.
Seismic refraction (SR) is a well
established geophysical method. In this
method a seismic impulse (a hammer blow or
explosive charge) is applied to the ground
and the time to reach a transducer is mea-
sured for varying distances between the
impulse and the transducer. The time is
plotted as a function of distance and, if
there is a well-defined stratigraphic
layer beneath the surface, a characteristic
break in the curve is found from which the
depth to the layer can be determined. This
method has not been used to detect small
objects but is widely used in oil and
mineral prospecting and exploration, and
can be used to determine generalized topo-
graphy at a dump site. Details concerning
ER and SR can be found in the.book by
-225-
-------
Dobrin (1976).
The continuous microwave (CWM) method
Is similar to ground probing radar except
that a continuous wave (CW) is used. The
CW is swept in frequency and the wave re-
flected from the ground surface and the
wave from a subsurface reflection inter-
fere with each other. The spacing (in fre-
quency) between the interference maxima
(or minima) as the frequency is swept gives
the depth of the reflecting surface. Some
systems of this "type are in an advanced re-
search stage; however, they are not availa-
ble commercially as far as the authors are
aware. Details concerning this method are
found in Koerner, et al. (1978).
DESCRIPTION OF FIELD SITES
The first site'was in a nearly ideal
dry sandy soil in an open field free of
man-made interference (Lord, et al. 1982b).
This site provided .an excellent starting
point for our experimental work. Three
methods (ER, SR and CWM) were eliminated
from further consideration based on the
field data at this first site. The second
site was much more formidable. (Koerner,
et al., 1982) Here a saturated silty clay
soil overlying shallow shale rock was used.
Recognizing that containers are sometimes
dumped directly into water (Weston, 1981)
and that the salinity of the water can
range from fresh to brine, the third study
was directed at drums under water (Lord,
et al., 1984).
Important is the extent to which
ground water salinity influences the detec-
tion capability of these NOT methods. To
this end, studies were made at a fourth
site with steel containers buried in a soil
of varying electrical conductivity.
(Koerner, et al., 1984) The ocean was used.
as an electrical conductivity extreme and
the conductivity decreased substantially
as the survey moved inland. Site 5 was
the same location as Site 4 but, in this
case, plastic containers were used instead
of steel. (Lord and Koerner, -1986b)
RESULTS
Overall View
Table 1 is an overall view of the
ability of all four NOT methods to detect
steel and plastic containers in a wide
variety of subsurface situations. It
should be considered as a guide to the
practitioner when considering the use of
NOT methods for a particular soil. (More
detail can be found in last year's Con-
ference Proceedings- Lord and Koerner,
1986a and the Final Report on the Project-
Lord and Koerner, 1987).
Elimination of Certain NOT Methods
Figure 1 shows the results for a
seismic refraction (SR) survey over the
four steel drum pattern indicated on the
figure. The drums were placed horizontally
and with about four feet of soil cover at
Site 1. There is essentially no indica-
tion of the drums in the survey. This ex-
perimental result, taken together with
the fact that no work has been reported
using seismic refraction (SR) in.very short
range container searching eliminated the
technique from further consideration.
Figure 2 shows the results of an elec-
trical resistivity (ER) survey over the
same four steel drum pattern. Again there
is no indication of the buried drums, even
though the electrode spacing was choosen
to be quite sensitive to the scale of the
burial. The-large increase at the right
end of the survey corresponds to a gulley
where water runs during rain storms. Thus
for the same reason as with SR, this method
was eliminated from further consideration.
Figure 3 shows the results of a con-
tinuous microwave (CWM) scan over the same
four steel drum pattern. There seems to-be
an indication of the drums. However the
calculated depths are too low by a factor
of 3-4. Thus the method appears to be of
quite limited value in searching for buried
drums. As there were also no published
reports of its use in this type applica-
tion, the CWM method was eliminated from
'future consideration. A prime reason for
the elimination of the three methods de-
scribed above in this section is that there
appeared to be much more attractive methods
available, namely, MD, EMI, MAG, and GPR.
Location of a. "Trash Dump"
At Site 1 a pit approximately 12 ft x
12 ft x 5 ft deep was filled with six
steel drums, two plastic drums of various
sizes and a few steel plates of various
sizes. This site was called the "Trash
Dump". Figure 4 shows the disposition of
the buried objects.. The responses of the
-226-
-------
four NOT methods over the "Trash Dump" are
shown in Figures 5-8.
Figure 5 gives the MD results in terms
of the strong response region. MD certain-
ly detects the "Trash Dump" quite well.
Figure 6 gives the EMI results in
terms of equal response amplitude contour
lines. Note the position of the "posts"
outlining the burial. Again, EMI locates
the "Trash Dump" very well.
The MAG response is shown in Figure 7
again in terms of response contour lines.
The center of the contour appears somewhat
east (about 4-6 feet) of the center'of the
steel drum distribution. This characteris-
tic of the MAG response is quite typical,
but would not seem to be a serious draw-
back in the use of the method, for the
"Trash Dump" is still located quite well.
Figure 8 shows the GPR response over
the "Trash Dump". Here the side walls of
the excavations are clearly seen, however
the individual objects in the "dump" can-
not be delineated. This is an extremely
important use of the GPR method, i.e., to
locate the side walls of a burial site.
Lateral Scan Sensitivity and Drum Orienta-
tion Effects
It is important to know how close the
scan line must be to the actual buried con-
tainer before detection if possible.
Figure 9 shows the results of EMI scans at
various distances laterally displaced from
the line centered on the four 55-gallon
steel drum sequence. The drums were buried
at Site 1 with about four feet of soil
cover. It is seen that the method is quite
sensitive until the offset is greater than
8 feet for the antenna perpendicular to
the center line and greater than 4 feet "
offset for the antenna parallel to the cen-
ter line of the drums.
Figure 10 indicates the lateral scan
sensitivity of the MAG method for a single
30-gallon steel drum buried at Site 1, with
about three feet of soil cover, and at dif-
ferent orientations. The MAG method is
sensitive to about 8 feet of lateral off-
set for single drums.
Figure 11 shows that the GPR method
loses location ability if the scan is dis-
placed by two feet from the center of
drum. The MD is a less sensitive version
of EMI and hence has, as was shown in our
experiments, very little lateral scan
sensitivity.
Of interest also is the effect of
steel drum orientation on the response of
the various methods. The MAG results are
shown in Figure 10. EMI results are shown
in Figure 12. -Figure 13 shows the effect
of drum orientation on the GPR response.
The MD does not have the requisite sensi-
tivity to be very sensitive to drum orien-
tation.
In summary EMI and MAG are moderately
dependent on the orientation of the drum
whereas GPR is extremely sensitive to
orientation with the horizontal burial
giving, by far the best response. • .
Water Table Determination
Figure 14 shows the detection of the.
water table at Site 4. The varying depth
water table here was of quite high elec-
trical conductivity, so contrast was quite
high and the table could be seen very
easily to sixteen feet. The soil was a"
sand of very little stratigraphy, so the
GPR method worked very well. This is, one
of the deepest water tables that has been
reported using GPR, to the authors' knowl-
edge.
Detection of Plastic Drum with EMI
Figure 15 shows the results of EMI
scans over 40 gallon empty plastic drums
at Site 1. The drum at one and three feet
of cover are easily detected, even with
many feet of lateral off-set'of the instru-
ment. In essence the EMI is acting like
an air-void detector here. The excellent
resolving power here is undoubtedly due
to the fact that the sand is very dry, very
uniform and of quite low electrical con-
ductivity;
SUMMARY
The basic results of our work have
been presented In detail elsewhere (Lord
and Koerner 1986, 1987)... The summary is
given in Table 1. Here is presented a
"pot pourri" of interesting and important
results which due to space limitation
could not be fitted into the paper at last
year's conference. (Lord and Koerner,
1986). •
-227-
-------
In particular, detail is given as to
experimental results which allowed us to
eliminate certain NOT methods from future
considerations. Considerable detail is
given as to the ability of the various NOT
methods to "pick up" an artificial "Trash
Dump". The methods used were metal detec-
tion (MO), electromagnetic induction (EMI),
magnetometer (MAG) and ground penetrating
radar (GPR). The ability of GPR to pick
up trench boundaries cannot be over em-
phasized.
Results concerning the sensitivity of
the NOT methods when passing on a scan not
directly over the center line of the drums
are presented. The GPR response suffers
greatly when "off-center". The effect of
drum orientation is catastrophic only for
GPR; only the horizontal burials are easily
picked up. Water table determination with
GPR to 16 feet is demonstrated as well as
plastic drum detection with EMI. The
latter case is an example of void detction.
ACKNOWLEDGEMENTS
We thank the U. S. Environmental Pro-
tection Agency for funding of this work
through Cooperative Agreement No. CR
807777020. Special thanks are due to Dr.
John E. Brugger, the project officer on
this Cooperative Agreement, for his con-
stant interest, advice and encouragement.
We also wish to thank Ira Wilder, Frank
Freestone and Jack Farlow of EPA for their
support. A large number of fine Drexel
University graduate students deserve credit
for their cooperation and enthusiasm in
this work.
REFERENCES
Bowders, J. 0., Jr., Koerner, R. M. and
Lord, A. E., Jr. (1982). "Buried Container
Detection Using Ground Penetrating Radar,"
Journal of Hazardous Materials, Vol. 7,
pp. 1-17.
Dobrin, M. B., (1976). Introduction to
Geophysical Prospecting, McGra'w Hill, New
York.
Koerner, R. M., Lord, A. E., Jr.,
Okrasinski, T. A.'and Reif, J. S., (1978).
"Detection of Seepage and Subsurface Flow
of Liquids by Microwave Interference
Methods," Proc. Conf. on Control of Hazard-
our Materials, April 11-18, 1978, Miami
Beach, Florida, pp. 287-292.
Koerner, R. M., et al., U982). "Use of
NOT Methods to Detect Buried Containers in
Saturated Clayey. Silt Soil," Proceedings
of Management of Uncontrolled Waste Sites,
Nov. 29 - Dec. I , 1982, Washington, DC,
HMCRI, pp. 12-16.
Koerner, R. M. and Lord, A. E., Jr. (1984).
"NOT Location of Containers Buried in
Saline Contaminated Soils," Proceedings of
Management of Uncontrolled Waste Sites,
Nov. 7-9, 1984, Washington, DC, HMCRI Pub!.,
Silver Spring, MD, 20910.
Lord, A. E., Jr., Koerner, R. M. and
Freestone, F. J. (1982a), "The Identifica-
tion and Location of Buried Containers Via
Non-Destructive Testing Methods," Journal
of Hazardous Materials. Vol. 5, pp. 221-
233.
Lord, A. E., Jr.^ et al. (1982b). "Use of
NOT Methods to Detect and Locate Buried
Containers Verified by Ground Truth Mea-
surement," Proceedings Hazardous Materials
Spill Conference, April 19-22, Milwaukee,
WI, HMCRI, pp. 12-16.
Lord, A. E., Jr., Koerner, R. M. and
Arland, F. J. (1984). "The Detection of
Containers Located Beneath Water Surfaces
Using NOT (Remote) Sensing Techniques,"
Proceedings of Hazardous Waste and Environ-
mental Emergencies, March 12-14, 1984,
Houston, HMCRI, pp. 392-395.
Lord, A. E., Jr. and Koerner, R. M. (1986a).
"Nondestructive Testing (NOT) for the Lo-
cation of Containers Buried in So.il,"
Land Disposal, Remedial Action, Incinera-
tion and Treatment of Hazardous Waste,
Proceedings of the Twelfth Annual Research
Symposium at Cincinnati, Ohio, April 21-
23, 1986. U. S. Environmental Protection
Agency.
Lord, A. E., Jr. and Koerner, R. M. (1986b).
"Nondestructive Testing (NOT) for the Lo-
cation of Plastic Containers Buried in
Soil," Proc. 1986 Hazardous Material Spills
Conference, St. Louis, May 5-8, 1986,
Govt. Inst. Inc.
Lord, A. E., Jr. and Koerner, R. M. (1987).
"Nondestructive Testing (NOT) Techniques
to Detect Contained Hazardous Waste," Final
Report on Cooperative Agreement No.
CR807777 Hazardous Waste Engineering Re-
search Laboratory, Releases Control Branch,
U. S. Environmental Protection Agency,
-228-
-------
Edison, New Jersey, February 1987.
McNeil], J. .0. (1982). "Electromagnetic
Resistivity Mapping of Contaminant Plumes,"
Proceedings of Management of Uncontrolled
Hazardous Waste Sites, Nov. 29-Dec. 1,.
1982, Wash., DC, Hazardous Materials Con-
trol Research Institute, Silver Spring,
MD, pp. 1-6.
Tyagi, S., Lord, A. E., Jr. and Koerner,
R. M. (1983). "Use of Proton Precession
Magnetometer to Detect Buried Drums in
Sandy Soil," Journal of Hazardous Materials,
Vol. 8, pp. 11-23.
Weston Consultants (1981). "Ground Pene-
trating Radar Survey Elizabeth River, New
Jersey," Final Report to U. S. Coast
Guard, New York-, 30 June 1981, 14 pgs.
so
a
£!
10
0
FIG. 1
10 20 30
SPACING (FEET)
10
2O
3O
50
SPACING (FEET)
Seismic refraction survey re-
sponse; (a) no drums present,
(b) four drum pattern present.
;WENNER ARRAY
POSITION OF c& ELECTRODE: (FEED
FIG. 2 Electrical resistivity survey re-
sponse over the four drum
pattern.
16
14
12
10
8
6
4
2
0
FIG. 3
8 12 16 2O 24 28 32 35
POSITION OF ANTENNAE (FEET)
Continuous wave microwave results
for the four drum pattern.
-229-
-------
REFERENCE POSTS
FIG. 4 Distribution of objects in
"trash dump". Key: metal drums:
1: horizontal, 30 gal; 2 horizon-
tal, 55 gal; 3: horizontal, 5
gal; 4: vertical, 5 gal; 5: verti-
cal, 30 gal; 6: 45° angle, 5 gal;
plastic drum: 7: horizontal, 30
gal.
FIG. 6 Electrical conductivity contours
in the vicinity of the "trash
dump" determined by the EMI
technique.
METAL DETECTOR
STRONG RESPONSE REGION
'TRASH DUMP' OUTLINE
FIG. 5 MD response in the vicinity of
the "trash dump".
FIG. 7 Magnetic field contour map of the
"trash dump" using the MAG. The
actual magnetic field is 55 x 103
Y plus the contour magnetic field
label units of Y. The "trash
dump" boundary is indicated by the
dashed lines.
-230-
-------
-
-
-10*-
k.
U.
3:
K.
a
Q
FIG. 9
30 40
DISTANCE (It)
9 (b)
Results of EMI scans over the
four 55-gallon steel pattern with
various lateral offset (a)
antenna \_ to scan direction; (b)
antenna II to scan direction.
FIG. 8 GPR survey over the "trash dump"
showing the excavation boundaries
of the burial pit. •'( Arrows in-
dicate trench boundaries).
^ O.40—
to
g O.30—
5
Ul O.20—
CC
O.IO—
•
4 — • *- 2' OFFSET 1 ft
1 fl
T j
\ 6'OFFSET \
\
' 8'OFFSET . V
• 4
1 | 1 1
i
30 10
DISTANCE (ft)
9(0)
S 300
in
•in
I
i
1 200
LJ
cc
100
20 40
DISTANCE (FT)
60
FIG. 10 MAGi scans over single steel drums
(With about three feet of soil
cover)of differing orientations
as shown on Figure. Scans at
various lateral offsets.
-231-
-------
T
20 30 40
DISTANCE (ft)
12 (b)
•FIG. 12 EMI scans for 55-gallon steel
drums buried with about three
feet of soil cover. Drums of
various orientations as shown on
Figure. Scans at various
lateral offsets.
FIG. 11 GPR printouts for a single 55-
gallon steel drum buried hori-
zontally with three feet of soil
cover, (a) Antenna directly
over the drum (b) Antenna dis-
placed two feet from drum center-
line.
aso-
•| 040—
Iai°:
§ 0 30-
* 0,10—
10
DISTANCE (ft)
30
12 (a)
GROUND SURFACE
O1
-... _....,.__.,lft8ifas^WlEUH
FIG. 13 GPR printout showing survey of
30-gallon steel drums at
various orientations buried 3
ft. beneath the surface, (a)
Drum standing on end; (b) Drum
at 45° angle to ground surface;
(c) Drum buried horizontal to
ground surface.
-232-
-------
Figure 14- GPR scans showing the water table, starting at a depth of about 4 feet in
the left and going to a depth of about 16 feet at the right (Site 4). ;
J h K " h
o-ll I
K0.60—
0
. | 0.50—
i 0.40—
H 0.30 — i
§ -
£o.20-
_
0.10—
O- — •^\^./-^^^ysST^^st:
-• — •— »• 2' OFFSET
V S'OFFSET
" ' "8'OFFSET
1 T ' i i T i i ' i* ' M i
10 20 30 10 50 60 70 80 90 IOO
DISTANCE (ft)
FIG. 15 Results for EMI over empty 40-
gallon plastic drums buried
with the indicated soil
covers. (Site 1).
-233-
-------
CJ C
i- CU
S- l/l
3 CU
CO J_
o.
T3
CO C
NT-
00 "§
t— >
CO CO
O't-
CU4->
4-> O.
CO CU
(JO
5 c
~J *—
o
O'>-
•P4J
CO
W S"
•04J
O CO
r- c
4-> CO
CO O_
§1 £
C/) CO '1-
•»— c
S--O C
CO CO O
^"co t—
ovr- <|
ui co
=3 C
1. & ^_^
O 4-* CO
•r- CU
>>"CJ CO
4-> C CJ
•r- Q -C
r-0 4->
T- C
.Q S- CU
CO CU S-
JJ-O CO
0.1= 0-
CU 3-
U
owe
< S- -r-
CJ
r- e co
cO-r- CO
S. CO i-
CU 4-> T3
C C 3
CJ O +->
C3 O CO
1
«-«
r—
CO
CO
JO
(U — .
o =c
"cu —
C7>
CO
CJ)
~
^
ct
o rctS
C T3 0-
> co cr
3
3
O
.
>
O
J
j c:
c o
cn-i— •—
CO 4-5 i — i
= o s;
3 3 LL
J C
[}
Jj
S-
3 Si
^- CU
q_
CU'i- ?
^-O n
C
o
•1—
4->
CO
3
CO
co
co ^~
Ot—
CO CO
4- •!—
S- S-
3 CO
CO 4->
CO Q.
S^_
0^0
CU CO CO
'cu'cu co
o o o
XXX
CO CU CO
•^^
0^
dis
c c -
CU CO CM
CU CO S-
o o o
X X O
co cu a.
•^"^
0^_^
excellent
average. (
not go'od
*— *»^-^
U3 CM
> — x^_-«
C CT3
cu cu o
i — i — O
i — i — en
CO CU
0 0 4->
X X 0
CO CO C
CO
"cO
-o
, CO
S
.e s- c
c/1 CO CO
CU 40 CO
S- C O
<4- •!- O
O O 0
CM in o
1 1 1
o o o
•—i LO
S-
CO
"3T3
c c .
CO CO
S- V>
c
<
i
s-
o
o
Q.
c~~
S.
*
CO
JJ
CO
S-
O)
TD
_
2..
moderate*
**™s
^t-
*— " *
A;
cu
CO
S-
co
T3
FT
^=
00
CO
S-
6?
CJ
C3
1— 1
'
o
in
CO
CU CO
S-H.
co ro co
LJ Tl i> i|_>
EEC
CO CO CO
'co'cu co
o o o
XXX
CO CU CO
^
• —
E 13 -O
CO O O
i— O O
i — CJ) en
CO
O 4-> +J
X 0 0
CO C C
_
CO
E -O -O
CO O O
i— O O
i — CJ) CJ)
CO
O 4-> 4-»
X O O
CO C E
**"^s
co ,
' '
cu o
i — O
i — en
CO S-
O O 4J
X O O
CU O. E
cu
CO
•o
CO
E
U) CO CO
CO 4-> CO
s_ c: o
M-T- 0
o o o
000
t-H t~ H i~H
S-
cu
4J
CO
itainers
&-
8
-J
•r
t/
(C
SI
-a -a
0 2
O C5
o o
c c
1
I *^"^* C
8"
• a-* 5
c/l O •
jj 4-d CU <4— C Cn
CO CO •!— *~ U) -r-
^ "E t> ' E o
CO O 3 S- CO 3 S-
0 o -a .r- +J "a o
x 45 a
•§ o
0 0
en en
•o "2
0 . 0
0 0
en en
? g
E ^
cu
CO
•o
CO
t C
<1) «3
4-> CD
o o
rH
I 1
S 1
CO
"3
E
CO
CD
> E
co en
" E
•a •!—
s-
CO
CU E
f__ ip—
"cu >>
o E
X CO
*
-234-
-------
THERMODYNAMIC ANALYSIS OF POST-FLAME REACTIONS
APPLIED TO WASTE COMBUSTION
Daniel P.Y. Chang, Ph.D.
Visiting Research Engineer to EPA/HWERL from the
Department of Civil Engineering
University of California
Davis, California 95616
and
Robert E. Mournighan and George L. Huffman
Thermal Processes Research Staff
Thermal Destruction Branch
Alternative Technologies Division
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The equilibrium compositions of product gases resulting from the combustion of a few
model waste/fuel mixtures have been calculated. These include some chlorinated hydrocar-
bons (CHCs) and some high nitrogen-containing species. The calculations were carried out
with the aid of an interactive, PC-compatible version of a powerful equilibrium solver,
STANJAN. Examples are drawn from the more interesting results in order to demonstrate
how the theoretical calculations can be interpreted and used to provide insight into the
occurrence of products of incomplete combustion (PICs) in incinerator effluents. Practi-
cal applications and extensions of the method are also discussed.
INTRODUCTION
It has long been recognized that
reaction kinetics, heat and mass trans-
fer are limiting processes to the
attainment of chemical equilibrium in
combustion processes. Nevertheless,
because of the high temperatures and
long residence times commonly found in
hazardous waste incinerators, there may
be situations in which the state of
chemical equilibrium is approached, at
least on a local basis. In such cases,
thermodynamic equilibrium analyses can
provide a useful semi-quantitative
description of the final system compo-
sition. Even when kinetic limitations
prevent the attainment of equilibrium,
the equilibrium composition of a system
is indicative of the state toward which
a reacting system is tending. Perhaps
most importantly, thermodynamic equi-
librium computations provide an over-
view of system behavior which is not
readily apparent in detailed chemical
kinetic computations.
•There are reasons to believe that
the penetration of principal organic
hazardous constituents (POHCs) and
products of incomplete combustion
(PICs) through incinerators has often
been associated with some type of non-
stoichiometric, fuel-rich condition
having occurred within the combustion
device. Although the precise reasons
leading to a locally fuel-rich stoichio-
metry have not been fully determined,
the occurrence of non-stoichiometric
"pockets", "plugs" or "puffs", even
with an average condition of excess
air, has been suggested (1,2,9,10). For
-235-
-------
this exploratory study of the post-
flame combustion environment (500 K <
T < 1500 K), it was assumed that con-
ditions ranged from stoichiometric
to fuel-rich. Chemical species
examined were primarily hydrocarbons
(HC), chlorinated hydrocarbons (CMC),
and nitrogen species for which thermo-
dynamic data were readily available.
Because of space limitations, only
highlights of the investigation are
presented in this paper. A preliminary
paper containing initial results for
CHCs should soon be available (2) and a
more complete presentation of results
will be available in a future EPA
report (3).
METHODS
A powerful, "menu-driven", personal
computer (PC) implementation of a code
for solving chemical equilibrium prob-
lems, STANJAN V3.6, was employed in
this research (5). The standard PC
version of STANJAN permits the inclu-
sion of up to 20 compounds in an equi-
librium computation. By judicious
selection, many more than 20 species
can be studied under the same specifi-
cation of temperature and pressure by
repeatedly solving the equilibrium
problem with new "minor" species. The
errors incurred in doing so can be
expected to be smaller than the accuracy
embodied in the thermodynamic data.
The "major" species present at equi-
librium were determined by conducting
trial computations, including all
available compounds, and systematically
retaining only those that appeared at
high concentrations. It quickly became
evident which species needed to be
retained for the final computations as
whole classes of compounds could be
ignored for the cases studied, e.g.,
alkanes.
Thermochemical data on entropy and
enthalpy functions were prepared in a
form suitable for use with STANJAN.
The chemical species used in the calcu-
lations are shown in Table 1 along with
the source of the data. The data were
used without further critical evalua-
tion. In some cases the thermodynamic
functions were only available to 1000 K
and a smoothing algorithm in the STANJAN
package was used to extend the data to
6000 K as required by STANJAN. Since
only the lower temperature range (T <
1500 K) was of interest, extrapolation
errors probably do not nave a signifi-
cant effect upon the predicted equi-
librium concentrations, as many com-
pounds occur at negligible concentra-
tions in the temperature range above
1000 K.
For all the cases examined, equiva-
lence ratio (E.R. = [fuel/air] actual/
[fuel/air] stoichiometric) was ,
varied from 1.0 (stoichiometric) to
2.0 (highly fuel-rich). For the CMC
computations, hydrogen-to-chlorine
ratio was also varied from 1:3 to 3:1.
Although the lower H:C1 ratio would be
unrealistic in an "average" sense, it
is conceivable that incomplete blending
of the waste/fuel mixture could lead to
introduction of short-term fluctuations
in composition that could approach or
even be less than 1:3, e.g., tetrachloro-
ethene (specific gravity = 1.6) that
may have stratified in a tank. Further-
more, in some cases the fuel and the
waste are introduced into the combus-
tion unit as separate streams, again
resulting in the possibility of low
H:C1, on a local basis, caused by
inadequate mixing. As will become
evident from the computational results,
it would not require large amounts of
an inadequately blended feedstock or
poorly mixed gases to result in emission
levels comparable to those seen in
field tests.
The predicted equilibrium con-
centrations are only meaningful in the
sense that the "correct" chemical
species have been included in the
computation. Thus, if a major specie
which would form at equilibrium, such
as CO, were to be left out of a compu-
tation, the results would be in serious
error. Lack of thermodynamic data
prevented a full study of all chlorin-
ated ethenes and benzenes. Thus some
species that have been observed in
source tests and which may have measur-
able concentrations, e.g., a mole
fraction greater than about 1 E-10 (0.1
ppb), have been ignored. The figures
presented in the results show concen-
trations only of those species of
particular interest.
Some compounds were predicted to
reach significant concentrations, but
-236-
-------
only at the lower extreme of the
temperature range (T < 750 K), e.g.,
graphitic carbon, C(s). Under such
conditions, kinetic limitations exist
which prevent that condition from
becoming an accessible thermodynamic
state. In such cases, the computations
were repeated, removing the "offending
species" to determine the "next"
thermodynamically favored compound.
The validity of such manipulations
is questionable, nevertheless, insight
regarding alternative reactions of
possible importance can be obtained
using such a procedure, e.g., benzene,
chlorinated benzenes and dioxins became
significant species in some cases when
formation of C(s) was eliminated.
It is well known that molecular
nitrogen, N£, present in the combustion
air is not particularly reactive.
Therefore, only the nitrogen that was
assumed to be present in the waste
molecules was allowed to be "reactive".
A pseudo-species having the thermal
properties of molecular nitrogen was
included in the equilibrium calcula-
tions to achieve the composition of
reaction products that would correspond
to combustion with air. The system
compositions were varied in accordance
with the type of waste assumed, e.g.,
hydrogen cyanide, acetonitrile or
acrylonitrile. Equivalence ratios were
varied as described above. No attempt
was made to undertake a systematic
survey of nitrogen-chlorine inter-
actions, e.g., nitrosyl chloride
formation.
SELECTED RESULTS AND DISCUSSION
The most striking results of the
equilibrium calculations occurred for
simultaneous conditions of fuel-rich
stoichiometry, E.R. > 1.0, and an H:C1
ratio £ 1.0. An example of the results
is shown in Figures la, Ib and Ic for a
system composition 1C:1H:3C1 (H:C1 =
0.33) and equivalence ratios of 1.0
(stoichiometric), 1.01 (1% oxygen
deficiency), 1.05 (5% oxygen deficien-
cy). Note that at lower temperatures,
under even the slightest fuel-rich
conditions, CC14 and COC12 became
major thermodynamically stable species.
The temperature region over which their
concentration increased rapidly began
at about 1000 K (about 700°C). This is
probably a sufficiently high temperature
for the system to tend toward the
equilibrium state given typical post-
flame residence times of about two
seconds, and the relatively low bond
strength of Clg molecules (239 kJ/mole).
However, detailed kinetic calculations
or experiments would be needed to con-
firm these predictions. Certainly for
temperatures below 500 K, reactions
would have slowed dramatically and the
predicted equilibrium states would not
be attained.
Several additional observations
were noteworthy in the CMC calculations.
1. There was a competition among
Cl and 0 ("oxidizing" species) for the
C and H (more-or-less chemically
"reducing" species in the systems
examined). Formation of C02, ^0 and
HC1 was essentially quantitative under
stoichiometric or fuel-lean conditions.
2. Under fuel-rich conditions, as
the H:C1 ratio decreased, increasingly
chlorinated methanes were observed as
stable products in the lower tempera-
ture range < 1000 K, i.e., chloromethane
—> tetrachloromethane.
3. Under slightly fuel-rich
conditions (1.0 < E.R. < 1.05), solid
carbon, benzene, chlorinated benzenes
and chlorinated dioxins were each
predicted to form at 1ow temperatures.
Solid carbon and benzene were thermo-
dynamically favored products with
an excess of hydrogen present, while
solid carbon, chlorinated dioxins and
chlorinated benzenes were favored
products for an H:C1 ratio of 1.0. ,
When the H:C1 ratio was less than
1.0, none of the above species were
predicted to form because the Cl would
directly oxidize the C atoms to produce
primarily chlorinated methanes
(especially CC14).
4. Recognizing that kinetic limi-
tations would prevent the attainment of
the equilibrium predictions described
in (3) above, the occurrence of solid
carbon, benzene, chlorinated benzenes
and chlorinated dioxins was successive-
ly eliminated. For H:C1 > 1.0, when
solid carbon was disallowed, the ten-
dency was to form benzene, then ethyl-
ene and chloromethane (Figures 2a,2b,
-237-
-------
2c). For H:C1 = 1.0, and a fuel-rich
stoichiometry, E.R. = 2.0 the corres-
ponding tendency was to form solid
carbon, chlorinated dioxins, chlorinated
benzenes, benzene, chlorinated methanes
(Figures 3a,3b,3c,3d, 3e,3f).
One inference that can be drawn
from the tendencies described above is
that it is quite likely that trace
amounts of the above species could be
formed in the cooling combustion gases
if the composition of the system corres-
ponded to those assumed. One can also
speculate that chemically simpler
species, e.g., chlorinated methanes and
ethenes, would be more likely to be
formed at detectable concentrations
than those which require a longer
sequence of elementary reaction steps,
i.e., chlorinated methanes > benzene,
ethylene, or chlorinated ethenes and
benzenes > chlorinated dioxins.
For the most part, it is the authors'
belief that this is consonant with
observed test results obtained from
incinerators that most closely corres-
pond to the equilibrium assumptions,
e.g., mineral kilns and circulating bed
combustors. In these types of devices,
strong transverse turbulent mixing may
actually impede longitudinal mixing,
leading to a condition approximating
"plug-flow". As a practical application
of these insights, it can be suggested
that it would be prudent to insure
adequate blending of waste streams with
an excess of hydrogen (in the fuel) and
adequate mixing of the combustion
gases in a temporal sense, i.e., the
mixing process should produce both good
microscale mixing and avoidance of
non-stoichiometric "puffs" or plugs".
Otherwise, high temperature alone will
not guarantee low emission levels
because some CHCs can simply re-
form upon cooling.
The initial exploration of the
nitrogen compounds under fuel-rich
conditions has not led to any remark-
able insights, other than that molecular
nitrogen in the combustion air should
not be included as a reactant in the
system. If a diluent gas having the
properties of N£ was not substituted
for the N£ in the combustion air, then
unrealistically large amounts of HCN
(hydrogen cyanide), C^HsN (acetonitrile)
and CsHsN (acrylonitrile) were predicted
to be present as stable combustion
products. Otherwise, the principal
nitrogen-containing species observed
was ammonia with only trace levels of
HCN, C2H3N, and C3H3N. Some typical
results for the "pseudo-nitrogen" compu-
tations are shown in Figures 4a, 4b and
4c.
CONCLUSIONS
This exploratory study has shown
that some valuable insights regarding
combustion processes can be obtained
from thermodynamic equilibrium analyses.
A wealth of thermodynamic data exists
for use with a very convenient compu-
tational tool, STANJAN. Besides the
CHCs and nitrogen compounds, data are
available for fluorine, sulfur and
metals. A systematic study of trace
species of environmental concern may
prove to be valuable, not only for
hazardous waste incineration but for
other types of thermal processes as
well.
A summary of specific findings is
given below:
1. CHCs are thermodynamically
stable compounds under fuel -rich con-
ditions ,
____-
dition to be particularly avoided.
Trace levels of such species can per-
sist in the combustion effluent even if
only a small fraction of the waste
encounters the conditions described
above. Such conditions can arise from
inadequate blending of fuel /waste
mixtures and incomplete mixing with
combustion air before leaving the high
temperature zone of an incinerator.
2. In real systems CO would be
expected under the fuel -rich conditions.
There are reasons why CO levels might
increase without a concommitant in-
crease in CHC emissions. However, it
appears unlikely that high levels of
CHC emissions are possible without at
least a moderate increase in CO levels.
3. Thermodynamic analyses may be
useful in identifying species for which
sampling methods need to be developed.
As an example, CUClg (phosgene) is
predicted to be present at trace levels
when compounds such as CC14 (carbon
tetrachloride) are present in the
-238-
-------
combustion effluent. Because COC12
is a gas at ambient temperatures and
readily hydrolyzed, its presence may
have gone undetected using current
sampling methods.
4. The initial screening of
nitrogen-containing wastes did not
identify any major potential problems
with residual emissions under fuel-rich
conditions, under the assumption that
N2 in air was not available for reaction.
5. Additional kinetic studies
are needed to determine the range of
applicability of the thermod.ynamic
predictions. At present, it can only be
said that the predictions are in
qualitative agreement with field and
laboratory tests where the equilibrium
assumptions were most closely satisfied.
REFERENCES
1. Chang, D.P.Y, N. Sorbo, G. Mur-
chison, R. Adrian and D. Simeroth,
June, 1985. "Evaluation of a
Pilot-Scale Circulating Bed Combus-
tor as a Potential Hazardous Waste
Incinerator." Paper No.85-776,
Presented at the 78th Annual Air
Pollution Control Association Meet-
ing, Detroit, MI. Submitted for
publication to the Journal of the
Air Pollution Control Association.
2. Chang, D.P.Y., R.E. Mournighan and
G.L. Huffman, 1987. "An Equilib-
rium Analysis of the Combustion of
Chlorinated Hydrocarbons." Submitted
to the Journal of the Air Pollution
Control Association.
3. Chang, D.P.Y., 1987. "Spray Combus-
tion Studies of Hazardous Waste
Incineration." A report to the U.S.
EPA Hazardous Waste Engineering
Research Laboratory for Cooperative
Research Agreement CR813333-OT-0, to
be submitted September, 1987.
4. Mallard, G., 1986. Chemical Ki-
Netics Division, National Bureau of
Standards. Through personal communi-
cation, supplied thermochemical
data contained in the NASA-Lewis
chemical equilibrium code.
5. Reynolds, W.C., January, 1986. "The
Element Potential Method for Chemi-
cal Equilibrium Analysis: Implemen-
tation in the Interactive Program
STANJAN Version 3." Department of
Mechanical Engineering, Stanford
University, Stanford, CA 94305.
6. Shaub, W.M., 1982. "Estimated
Thermodynamic Functions for Some
Chlorinated Benzenes, Phenols and
Dioxins." Thermochemica Acta.
58:11-44.
7. Stull, D.R. and H. Prophet, 1971.
JANAF Thermochemical Tables. 2nd
tdition. NSRDS-NBS 37. RaTional
Bureau of Standards, Washington,
D.C.
8. Stull, D.R., E.F. Westrum Jr. and
G.C. Sinke, 1969. The Chemical
Thermodynamics of Organic Compounds.
John Wiley and Sons, Inc., New
York. „.
9. Tsang, W., 1986. "Fundamental
Aspects of Key Issues In Hazardous
Waste Incineration." Presented at
the ASME Winter Annual Meeting,
Anaheim, California, December,
1986.
10. Wendt, J.L., W. P. Linak, J. D.
Kilgroe, T. A. McSorley and J. E.
Drum, September, 1986. "Parametric
Studies Delineating the Occurrence
of Transient Puffs in a Rotary Kiln
Simulator." Presented at the Third'
International Symposium on Operating
European Hazardous Waste Management
Facilities, Odense, Denmark, Sep-
tember, 1986.
Erratum:
In the set of calculations for H:C1 = 3:1 (fuel equivalent to CH3C1), methane
(CH4) was inadvertently left out under fuel-rich conditions. As a result C(s) was
predicted to be the major thermodynamically stable species at low temperatures (Figure
2a). When CH4 is included, it becomes the major species and C(s) disappears as a
stable phase. Figures 2a, b and c therefore represent the predicted system composi-
tions if CH4 (Figure 2a), C(s) (Figure 2b), and C6H6 (Figure 2c) are successively
removed from the system. The major conclusions of the study remain unchanged.
-239-
-------
TABLE 1. LISTING OF CHEMICAL SPECIES STUDIED
Composition
Reference
Organic;:
methane
acetylene
ethylene
benzene
toluene
styrene
naphthalene
dibenzo-p-dioxin
blphenyl
anthracene
Chlorine Containing Species:
tetrachloromethane (carbon tetrachloride)
trlchloromethane (chloroforn)
dichloromethane (methylene chloride)
chloromethane
carbonyl dichloride (phosgene)
dichloroacetylene
tetrachloroethene (perchloroethylene)
hexachloroethane
trlchloroethene (trichloroethylene)
pentachloroethane
1,1-dichloroethene (dichloroethylene)
l,2-d1chloroethene (dichloroethylene)
chl oroethene (vinyl chloride)
1,1,2-tricMoroethane
1,1-dichloroethane
1,2-di chloroethane
hexachlorobenzene
1,2,3-tri chlorobenzene
meta-dichlorobenzene
chl orobenzene
monochlorodibenzo-p-dioxin (HCDD)
tetrachlorodi benzo-p-di oxi n (TCDD)
octachlorodibenzo-p-dioxin (OCDD)
chlorine, atomic
chlorine, molecular
hydrogen chloride
Nitrogen Containing Species:
methyl ami ne
acetonitrile
ethylenimine
ethyl ami ne
acrylonitrile
pyrrolidine
pyridine
aniline
benzonitrile
hydrocyanic acid (cyanide)
nitroxyl hydride (nitrosyl hydride)
nitrous acid
nitric acid
nitrogen, atomic
ammonia
nitrogen monoxide (nitric oxide)
nitrogen dioxide
nitrogen trioxide
nitrogen, molecular
dinitrogen monoxide (nitrous oxide)
Additional Inorganic Species:
carbon monoxide
carbon dioxide
dihydrogen monoxide (water)
oxygen, atomic
oxygen, molecular
CH4
C2H2
C2H4
C7H8
C8H8
C12HiQ
C14H10
CC14
CHC13
CH2Cl2
CH3C1
COC12
C2C12
C2C16
C2HC13
C2HC15
1,1-C2H2C12
1,2-C2H2C12
C2H3C1
1.1.2.C2H3C13
1,1-C2H4C12
1,2-C2H4C12
Cede
1,2,3-C6H3C13
2,4-C6H4Cl2
C12H702C1
Cl2H4°2c14
C1202C18
Cl
Cl2
HC1
CH5N
C2H3N
C2H5N
C3H3N
C$3
C7H5N
HCN
HNO
HN02
HN03
N
NH3
NO
N02
N03
N2
N20
CO
C02
H20
0
02
7
7
7
7
7
7
7
7
8
8
8
8
8
8
8
8
8
6
8
8
6
6
6
6
7
7
8
8
8
8
8
8
8
8
8
7
7
7
7
7
7
7
7
7
7
7
-240-
-------
Log(Concentration) VS 1/T
(a) Fuel-Equivalent to CHCI3, E.R
-pptn
f -
3
¥
0.0004 O.OOOS 0.001? 0.0016 O.OO2O C.OO24 O.OO28 O.OO32
Ibl Fuel Equivalent to CHCI3, E.R = 1.O1
5'-
— 1
-2
-3 -
—5 -
— 6 -
-7 -
-8 -
—9 -
-1O -
-11 -
12 -
3 -
14 -
•15 -
•16 -
17 -
18 -
19 -
2O
2500K
1000K
500K
294K
0.0004 00008. 0.0012 O.OO1S O.OO2O O.OO24 O.OO28 O.O032
MFuel Equivalent to CHCI3. E.R = 1 .05
ppm
ppt
—3 -
— 4 -
—5 -
— 6 -
—7 -
-8 -
—9 -
— 1O -
— 1 1 -
-12 -
— 13 -
—14 -
— 15 -
-16 -
— 17-
-,18 -
—19 -
-2O
2500K
1000K
500K
294K
O.OOOS 0.0012 0.0015 O.OO2O O.OO24 O.OO2S o!oO32
ppm
ppt
Figure 1. Composition versus inverse temperature for a system equivalent to
CHClafuel (chloroform) burned with air under (a) stoichiometric con-
ditions, (b) 1% oxygen deficiency (c) 5% oxygen deficiency. Symbols
spaced every 50K.
-241-
-------
Log (Concentration) VS 1 /T
(») Fuel Equivalent to CH3CI. E.R. ' —
-ppm
I
s
O OOO4- O OOOS O.OO12 O.OO16 O.OO2 O.OO24 O.OO28
VCCS) XC6H&
(b)Fuel Equivalent to CH3CI. E.R. — 1 .OS
-ppm
O.OOOA O.OOOS O.OO12 O.OO1S O.OO2O O.OO24 O.OO2S
XC6H6
MFuel Equivalent to CH3CI, E.R. = 1.05
-ppm
S -
O.OOO+ O O008 O.OO12 O.OO16 D.OO2 O.OO2+ O.OO2S O.O032
1/T — [1 /K]
D CO 4- COCI2 O CH2CI2 A CH3C! X C2H4
Figure 2. Composition versus inverse temperature for a system equivalent to a
CHsCl fuel (chloromethane) burned with air at E.R. = 1.05 (slightly
fuel-rich): (a) allowing C(s) as a reaction product (b) disallowing
C(s) (c) disallowing both C(s) and C^Q as reaction products. In
reality, the same levels of CHsCl are predicted, order of ppb, since
reaction kinetics slow dramatically below 750K.
-242-
-------
Log (Concentration-) VS 1/T
la) Fuel Equivalent to CH2CI2, E.R. = 2.OO
,—7 -
—8 -
-a
— 1O
-1 1
-12
-13
-14
-15
-16
—17
-19
-2O
O.OOO5 O.OCOV O.CCO9 O.OO1 1 O.OO12 O.OO15 O.OO17 C.OO19
ppm
ppt
1 CHCI3
V C
-------
Log (Concentration-) VS 1/T
-ppm
o» _
—20
O.OOO5 O.OOO7 O.OOO9 O.OO1 1 O.OO13 O.OO15 O.OO17 O.OO19
1/T - D/K]
CO + CH3CI « CH2CI2 a CHCI3 X CC14-
2QQOK
(a) Fuel Equivalent to CH2CI2, E.R. = 2.OO
I100QK
-ppm
I
1
O.OOO5 O.OOO7 O.OOO9 O.OO11 O.OO13 O.OO15 O.OO17 O.OO19
1/T — Q1/K1
+ C6H5C1 o C6H4-CI2 A C6H3CI3 X C6CI6
20_flQ_K.
(f)Fuol Equivalent to CH2CI2, E.R. — 2.OO
11 OOP K
500 K
-ppm
DQODD-B a a a'D-D—a a—B—3 B s B
O.OOO5 O.OOO7 O.OOO9 O.OO1 1 O.OO13 O.OO15 O.OO17 O.OO19
1/T"— [1/K]
CO + COCI2 o MCDD A TCDD x OCDD
Figure 3. Continued - Composition versus inverse temperature for a system equiva-
lent to CH?Cl2 fuel (dichloromethane) burned with air under a fuel-rich
condition (E.R. = 2.0): (d), (e), (f) present the same system, but
excluding C(s).
-244-
-------
Log (Concentration) VS 1/T
(alFuel Equivalent to HCN, E.R. = 2.OO
o -i
-ppm
- ppm
O.OO1 O.OO14 O.OO1S O.OO22 O.OO26 O.OCX2
(b)Fuel Equivalent to C2H3N, E.R. = 2.OO
1/T
- [1/K]
A C3H3
Figure 4. Composition versus inverse temperature for nitrogen based fuels,
E.R. = 2.0: (a) fuel equivalent to HCN (hydrogen cyanide) (b) fuel
equivalent to Cz^ (acetonitrile) (c) fuel equivalent to
(acrylonitrile).
-245-
-------
INFLUENCE OF ATOMIZATION PARAMETERS ON
DROPLET STREAM TRAJECTORY AND INCINERATION
J. A. Mulhoi land
U.S. Environmental Protection Agency
Air and Energy Engineering Research Laboratory
Research Triangle Park, N.C. 27711
R. K. Srivastava
Acurex Corporation
Durham, N.C. 27713
ABSTRACT
In the incineration of liquid hazardous wastes, atomization quality may limit
destruction efficiency. Large, non-mean droplets in a fuel spray can pass through the
flame zone prior to complete evaporation, and may subsequently fail to burn completely
due to insufficient temperature and/or flame radicals. A study is ongoing to develop a
predictive understanding of individual droplet trajectories in turbulent diffusion
flames. Experiments in a cold quiescent flow environment, a laminar flow flat-flame
burner, and a 100 kW swirling, turbulent combustor have been conducted to calibrate a
model to predict three-dimensional trajectories of single monodisperse droplet streams.
Escape from the flame zone of large (> 200 ym diameter) fuel oil/xylene droplets
has been observed as a function of initial droplet size, velocity, spacing, and injection
angle. Incomplete incineration of these droplets was found to be related strongly to
droplet penetration of the flame zone. Minimum model requirements to successfully
predict droplet trajectories in turbulent diffusion flames include: droplet spacing
effects on drag; droplet/droplet interaction effects on evaporation; evaporation effects
on drag; and turbulence effects on droplet ballistics.
INTRODUCTION
Of the hazardous organic wastes pro-
duced in the U.S., about 75 percent are
liquids or dissolved in liquids. With land-
fill disposal of these wastes becoming
Increasingly unacceptable due to growing
public concern and increased levels of
hazardous waste production, thermal destruc-
tion is being considered as an alternative
disposal method. Successful incineration
requires that flame radicals and high tem-
peratures be combined in such a way that
the principal organic hazardous components
(POHCs) of the feed waste are destroyed
and that the formation of products of
incomplete combustion (PICs), which may
be as hazardous as, or even more hazardous
than, the original POHCs, are minimized.
Proper spray atomization provides the
necessary dispersion of liquid fuel and
waste into the oxidizer to avoid inciner-
ator failure by inadequate mixing. Fuel
spray nozzles degrade with use and must
be replaced periodically. Therefore,
there is a need to understand and quantify
how atomization parameters limit liquid
waste incineration so that a sound ra-
tionale for selecting and replacing spray
nozzles can be defined.
Four group combustion modes of a fuel
droplet cloud have been identified [1],
with single droplet combustion being
applicable in practice to only a very
limited number of special situations. Such
-246-
-------
a special situation, however, can arise
during the incineration of liquid hazardous
wastes, where droplets with large diameters
(as much as an order of magnitude larger
than the mean) congregate at the outer edge
of fuel spray cones [2], One or more of
these large, errant droplets may individu-
ally pass through, or bypass, the main
flame zone and lead to a failure mode in
the incinerator. For example, bypassing of
as few as 1 drop out of 10 million drops
can lead to failure to meet a destruction
removal efficiency (ORE) in excess of 99.99
percent, as is required by law.
Motivation for the current study lay,
therefore, in the need to predict single,
non-mean droplet trajectories in a turbu-
lent flame zone. To this end, experiments
have been conducted (1) to determine mini-
mum requirements of a model that success-
fully predicts measured trajectories of
single monodisperse droplet streams in
three-dimensional turbulent diffusion
flames, and (2) to relate droplet penetra-
tion of the flame zone to potential
incinerator failure modes.
THEORY
A semi-empirical numerical model is
being developed for predicting the ballis-
tics of burning droplet streams in tur-
bulent diffusion flames. An important
input to this model is a proper represent-
ation of the drag coefficient on a non-i,
evaporating droplet in a stream.
Drag Coefficient
While much information is available on
the relationship of the drag coefficient, -
CQ, of an isolated sphere to Reynolds
number, Re, little is available on the
dependence of CQ on Re and droplet spacing,
non-dimensionalized by droplet diameter, D,
as L/D. As droplet spacing is reduced,
drag is reduced due to wake effects. For
large droplet spacings, when droplet^ cease
to interact, the drag coefficient, CQ, for
each one is that of an isolated sphere, and
is only a function of Re. For the slightly
distorted droplets formed by a vibrating
capillary droplet generator, the relation
of Lambiris and Combs [3] is recommended:
In the other, extreme, as droplet spacing
approaches the droplet diameter, the drag
coefficient may be assumed to approach the
friction drag coefficient of a long rod,
which can be calculated from the theory
of Glauert and Lighthill [4]. A general
form for the asymptotic form for the drag
coefficient, CQ, of a sphere as L/D
approaches unity is obtained by a Taylor
expansion about L/D = 1; thus, CQ is
hypothesized to be:
CQ(Re,L/D) = 0.755/Re + aReb [L/D-1] (2)
where a and b are parameters to be deter-
mined from experiment. The factor [aReb]
multiplying [L/D-1] is hypothesized to be
a function of Reynolds number, since the
extent of droplet interaction as L/D
approaches 1 will depend on Re, the latter
quantity determining the length of wake
behind each droplet [5], An effective
and universal technique to correlate data,
obtained between regions of validity for
the asymptotic forms for CQ as L/D -•-> 1
and L/D -->«, is to use the asymptotic
expansion formula developed by Churchill
[61:
[CD(Re,L/D)]-n =
[C5(Re,L/D).]-n + [C0(Re)]-n
(3)
= 2
Re > 1(T
where n is a parameter to be obtained by
experiment.
Trajectory Model • ,
A simple, numerical model has been
developed for predicting the ballistics
of an isplated burning droplet [7]. The
model is being refined to include effects
of droplet interaction on drag and evapor-
ation rate, evaporation effects on drag,
and turbulence effects .on droplet ballis-
tics. The model is used to solve the
uncoupled equations of droplet motion in a
Lagrangian framework. A three-dimensional
grid structure is established for specify-
ing the background gas velocity (vg), tem-
perature (Tg), and chemical specialion.
Calculations are terminated when the drop-
let exits the computational domain.
Drag and gravitational forces are
assumed to be the only external forces on
the droplet. The drag force, FQ, has two
vector components (FQI and F02). The drag
coefficient with droplet spacing effects
(CQ) is applied to the drag component in
the droplet stream direction; the drag
-247-
-------
coefficient without spacing effects (CQ) is
applied to the other drag component. FQI
is defined to be aligned with the droplet
velocity vector, v^; FQZ is then perpen-
dicular to the plane containing vg and vj
(Fig. 1). The magnitudes of the force
vectors are given by:
=
CD* "n
C£*^vr
(4)
where vri is the component of relative
velocity (vr = vj - vg) aligned with
Vd and vr2 = vr - vri. The evaporating
droplet drag coefficient (* denotes mass
transfer) has been determined by Eisenklam,
et al. [8]. Evaporation has been found to
decrease drag by a factor ne (ie = CD*/CD)>
given by:
B)
where B is the mass transfer number,
defined by equation 8c. Force balance
yields:
SFX * max = FDlx + FD2x
ZFy s may *
ZFZ s maz =
where Fgrav is the gravitational force
(= -mg).
(5)
(6)
Heat and mass transfer are modeled in
two steps: droplet heating to the liquid
boiling point, followed by droplet evapora-
tion. During droplet heating, its diameter
is assumed to be constant (i.e., no eva-
poration), and its temperature (T) is
assumed to be uniform (i.e., low Biot
number). Here, the time rate of change of
the droplet temperature is given by:
Nu k
(7).
where the Nusselt number (Nu) is calculated
using an empirical relation. After the
droplet temperature reaches the boiling
point, evaporation begins. In this stage,
the time rate of change of the droplet
diameter is given by:
dD 1 OR
_ = _CR — [1 + 0.23 Re0-5] ns
dt 20
where the evaporation constant (CB) is
given by:
(8a)
8 k
CB =
in (1+B)
(8b)
and where the mass transfer number (B) is
given by:
R =
(Be)
Combustion.-* O Droplet
Gas ^-^ O stream „
O y v-o—V.
O J—x =$> v
z
Figure 1. Vector diagram in three dimen-
sions. Velocities shown with
dotted lines; forces shown with
solid lines.
Here, kg, p&, and cpjj, are properties of the
liquid and background gas, Q is the heat of
combustion of the droplet, L is the latent
heat of vaporization, Y0
-------
.determined statistically, with the number
being sufficient to describe a distribution
of trajectory endpoints.
EXPERIMENTAL METHODS
The objectives of this study are: (1)
to measure and predict the three-dimen-
sional trajectories of single monodisperse
droplet streams in turbulent diffusion
flames, and (2) to study the relation of
these trajectories with droplet inciner-
ation effectiveness.
The 100 kW combustor, modified for
simulation of flame aerodynamics at the
expense of time/temperature simulation,
and monodisperse droplet generator (Fig. 2)
have been described previously [11]. Fir-
ing natural gas, two flame shapes with very
different mixing patterns were established:
Types C and A. The short, swirling Type C
flame is stabilized by strong internal
recireulation, whereas the long, axial Type
A flame is stabilized by external recireu-
lation. Radial profiles of temperature,
velocity (three components), and species
were measured. Gas temperatures were
^Natural
Gas
Air
Figure 2. Turbulent diffusion flame modu-
lar combustion facility. A 50
cm internal.diameter, 90 cm long
near-burner combustion module is
water-cooled and insulated with
a 4 cm thick refractory lining.
An International Flame Research
Foundation burner has a movable-
block air swirl generator and
interchangeable fuel nozzles. A
monodisperse droplet generator
is equipped with electronics for
droplet separation. It consists
of a vibrator (1), liquid supply
(2), orifice (3), pulsed charg-
ing ring (4), deflection plates
(5), and trap (6). A water-
cooled coil allows" for quenching
the combustion gases at variable
axial locations.
measured by suction pyrometry; axial,
radial, and tangential velocity components
were measured by using a specially designed
five-hole pitot probe (water-cooled with
0.8 cm uncooled tip, calibrated). "In-
stantaneous" velocities of the turbulent
flow were recorded, with means and standard
deviations calculated by microcomputer.
Measured radial profiles of axial velocity
were consistent with overall mass balance.
The droplet generator is a vibrating
orifice device with ancillary electronics
to facilitate droplet spacing variation.
Initial droplet diameter and spacing,were
measured via strobe photography. Initial
droplet velocity was.calculated from the
vibrator frequency setting. The fluid
tested, chosen on the basis of conductivity
(for electrostatic charging and deflection)
and viscosity (for droplet formation),
was a mixture of 80 percent (by volume)
Shell Oil Company fuel additive ASA-3,
comprised mostly of xylene (CgHio), and 20
percent distillate fuel oil.
Droplet trajectories were measured
with the aid of high speed photography.
Droplet destruction efficiency was deter-
mined by measuring the increased levels of
exhaust unburned hydrocarbons (UHC, or
surrogate POHC) and carbon monoxide (CO, or
PIC) due to droplet injection. These-
emission measurements were taken from the
stack where the combustion gases are well
mixed. Flame ionizatiori was used to detect
UHC, and an infrared instrument was used
for CO measurement. A removable water-
cooled coil for quenching the combustion
gases was inserted, resulting in a bulk gas
residence time of 0.6 s before quenching.
To determine droplet spacing effects
on the non-evaporating droplet drag coeffi-
cient, trajectory experiments were con-
ducted in the cold flow, quiescent environ-
ment within an observation tunnel. Empiri-
cal parameters a, b, and n (equations 2 and
3) were evaluated. Two-dimensional laminar
flow experiments in a bench-scale flat-
flame burner were then conducted to cali-
brate the model, allowing minimum model
requirements to be determined without the
complexities of three dimensions and turbu-
lence. In the 100 kW turbulent diffusion
flame experiments, baseline measurements
(i.e., without droplet injection) of com-
bustion gas temperature and velocity were
taken. Droplet trajectories were observed
for a variety of initial droplet stream
-249-
-------
conditions, including variation in droplet
size, velocity, spacing, and injection
angle. Finally, droplet incineration was
measured with the quench coil inserted.
COLD FLOW TRAJECTORIES
Parameters a, b, and n were estimated
using data from 10 full trajectories. Non-
linear regression for parameter estimation
was accomplished using a modified Marquardt
[12] algorithm, which is based on a multi-
dimensional search in parameter space for
the minimum value of a sum of squares
functional measuring deviation from the
data. The least squares parameter estimates
were: a = 34.8, b = -1.009, and n = 0.7072.
Measured and predicted trajectory end-
points are shown in Fig. 3 for the 10 tra-
jectories in regression and for 39 addi-
tional trajectories. More detailed results
from the experiments will be published in a
separate paper.
LAMINAR FLOW EXPERIMENTS
Trajectories of non-burning and burn-
ing droplets in a flat-flame burner were
measured to test the model in this aero-
dynamically simple environment. These
results, shown in Fig. 4, demonstrate that
droplet interaction effects on evaporation
rate and evaporation effects on drag are
important model requirements. Droplet
interaction can reduce the evaporation rate
by up to 25 percent (ns = 0.75); evapora-
tion can reduce drag by as much as 70
percent (ne = 0.30).
TURBULENT DIFFUSION FLAME TRAJECTORIES
In baseline tests without droplet
injection the combustion gas flow field was
characterized in terms of temperature and
velocity. A peak temperature of 1750 K was
measured at the combustor centerline,
decreasing to 1100 K near the wall. Axial,
radial, and tangential velocity components
were measured, with a peak total velocity
of 26 m/s measured near the burner quarl
exit. The standard deviation of velocity
measurement often exceeded 50 percent of
the mean velocity. Axial symmetry was
observed, both in temperature and velocity.
(See reference 11 for more details.) These
results were then used as inputs to the
80
60
40
20
• Re'
V Re'
O Re
A Re'
C>Re
= 210, h • 11.7 cm
» 292, h * 14.0 cm
- 136, h = 7.3 cm
= 110, h - 7.3 cm
» 89.6, h = 8.6 cm
01 23 45 678
In (L°/D)
Figure 3. Cold flow trajectory endpoints.
The range represents the hori-
zontal distance travelled by
the droplet stream while fall-
ing a 'height h. Points repre-
sent measured trajectory end-
points; lines represent model
results. The solid points and
dotted line are data and model
results from the best-fit non-
linear regression.
trajectory model.
The nominal droplet test condition
was: 225 urn initial diameter, 6.5 m/s
initial velocity, 130 diameters initial
spacing, and 30 degree injection angle.
The domain of droplet trajectories observed
for this nominal test condition is shown in
the two planes of observation in Fig. 5.
Superimposed on the cross-sectional view
along the centerline axis are lines of con-
stant mean axial velocity of the combustion
gases. The droplets were observed to
follow an approximately common path until
they entered the high shear layer of the
reacting gas. The burning times were very
short, as indicated by the burning droplet
"flecks" stretching only a few centimeters.
The large shaded area in Fig. 5 denotes the
distribution of burning droplets, Igniting
-250-
-------
30
25
Horizontal Distance (cm)
20 15 10
-25
Flat Plane
1 1
L°/D° =2.54
260 1030
80%
S
CCHCN/20% fuel oil
b b
= 234 urn
= 7 m/s
Figure 4. Laminar flow results. The top
plot shows trajectories of
evaporating, nonburning drop-
let streams; the bottom plot
shows trajectories of burning
droplet streams.
0 20 40 60
Axial Distance (era)
Figure 5. Trajectories of the nominal drop-
let stream test condition. The
shaded area represents the domain
of droplet trajectory endpoints.
An axial velocity profile is
superimposed in the top view.
and burning out at various locations in
this area. Thus, no single trajectory
describes a droplet stream test condition.
The existence of a domain of droplet stream
burnout locations, rather than a single
burnout point, is further evidence that the
gas velocity fluctuates with time (i.e.,
turbulence).
Measured droplet axial penetration
distances (i.e., axial distance droplet
travels prior to complete evaporation)
are listed in Table 1. Closely spaced
droplets penetrated farther than isolated
droplets. Little effect of spacing was
observed for values greater than 20 dia-
meters. Axial penetration distance roughly
doubled with a doubling of initial droplet
velocity. Droplet penetration increased as
droplet diameter increased, as well.
Little change in axial penetration distance
was observed for isolated droplet injection
at 0, 30, and 45 degrees. Finally, a
change in the combustion gas flow field
from a Type C to a Type A flame resulted in
a slight reduction in droplet penetration.
Model development to include turbulence
effects is ongoing.
TABLE 1. MEASURED TRAJECTORIES IN TURBULENT
DIFFUSION FLAME EXPERIMENTS
Input Test
Variable
Diameter
Velocity
Spacing
Angle
Flame
Condition:
Value
225 pm
297
378
3.8 m/s
6.5
9.8
2.5 dia.
4.7
8.8
16.7
32.3
130
0 deg.
30
45
Type C
Type A
Axial
Mean
30
45
55
20
30
45
55
45
35
30
30
30
30
30
35
30
25
Penetration:
Min.
15
35
45
15
15
30
35
30
10
10
25
15
20
15
25
15
10
Max
45
60
70
25
45
60
75
65
55
45
40
45
45
45
50
45
35
*
cm
cm
cm
cm
cm
-251-
-------
DESTRUCTION EFFICIENCY
Tests were conducted with the water-
cooled coil inserted downstream of the
Type C flame to quench the combustion gases
and measure droplet incineration. Combus-
tion gases were extracted from the stack,
with emissions analyzed with and without
droplet injection. Droplet destruction
efficiency (DE) was calculated as the mass
of carbon emitted as CO and UHC divided by
the mass of carbon injected.
Droplet incineration results are shown
in Fig. 6. The previously measured mean
axial penetration is also shown, with dashed
lines representing the range of droplet
trajectories. These data indicate that
droplet destruction efficiency is related
inversely to droplet penetration, which has
been shown to depend on droplet atomization
properties. Thus, predicting droplet ballis-
tics may be one tool for anticipating inciner-
ator failure modes due to poor atomization.
CONCLUSION
Large droplet penetration of the flame
zone has been observed as a function of drop-
let atomization parameters in tests with
single monodisperse droplet streams injected
into turbulent diffusion flames. The incom-
plete incineration of these hydrocarbon
droplets has been approximated by measuring
CO and UHC emissions, and a strong correla-
tion with droplet penetration has been
observed. The short burning distances and
relatively long trajectories prior to igni-
tion exhibited suggest that droplet aerody-
namics prior to ignition is of primary impor-
tance in predicting incinerator ORE.
60
so
$. 40
% 30
&
£ 20
3
O10
0
60
-I 01
o ~s
(0 cr
100
90 !
70
200 gg
' O
x -a
100 sl
—3-
•a *<
100 200 300 400
Initial Diameter (urn)
(a) Size Variation
500
In (L°/D°)
(c) Spacing variation
Initial Velocity (m/s)
(b) Velocity variation
DUI
I'0
J 40
1 30
O)
a.
.|H 20
O 10
0
**
•
***
• - - - -"*
1UU ,_,
1
90 S-5
— ^
80 ||
70
i i
200 =0
x -o
100 ^1
0-
0 "i~"
'0 15 30 45 60
Injection Angle (degrees)
(d) Angle variation
75
Figure 6. Droplet destruction efficiency and axial penetration results.
-252-
-------
It was found that in order to predict
droplet penetration the model must include:
o Droplet spacing effects on drag.
o Droplet/droplet interaction effects on
evaporation rate.
o Evaporation effects on drag.
o Turbulence effects on droplet ballistics.
Cold flow experiments have shown that
decreased droplet spacing reduces drag.
Laminar flow flat-flame burner experiments
have demonstrated that (1) droplet inter-
action reduces evaporation rate, and (2)
mass transfer reduces drag; both effects
increase penetration. Turbulent diffusion
flame experiments have shown the need for
statistically representing droplet pene-
tration. The model is currently being
developed to include turbulence effects.
This study represents a first step
toward developing a predictive understanding
of potential incinerator failure modes due
to poor atomization. While it is recognized
that further study with full sprays is ne-
cessary to establish nozzle operating guide-
lines based on atomization parameters for
incinerators, this ongoing work with mono-
disperse droplet streams provides a funda-
mental basis for establishing such criteria.
ACKNOWLEDGMENTS
All financial support was provided by
the U.S. Environmental Protection Agency's
Air and Energy Engineering Research Labor-
atory. The authors are grateful to J.V.
Ryan and J. VanRoy, of Acurex Corporation,
J.O.L. Wendt, Professor of Chemical Engi-
neering at the University of Arizona, and
S.B. Robinson and W.S. Lanier, of Energy
and Environmental Research Corporation.
REFERENCES
1. Chiu, H.H., Kim, H.Y., and Croke, E.J.,
"Internal Group Combustion of Liquid
Droplets," Nineteenth Symposium (Inter-
national) on Combustion, The Combustion
Institute, 1982, pp. 971-980.
2. Kramlich, J.C., Seeker, W.R., and Samu-
elsen, G.S., "Influence of Atomization
Quality on the Destruction of Hazardous
Waste Compounds," Twenty-First Sympo-
sium (International) on Combustion, The
Combustion Institute, in press.
3. Lambiris, S., and Combs, L.P., "Steady-
State Combustion Measurements in a LOX-
RP-1 Rocket Chamber and Related Spray
Burning Analysis," Detonation and Two-
Phase Flow, Vol. 6, 1962, p. 283.
4. Glauert, M.B., and Lighthill, M.J.,
"The Axisymmetric Boundary Layer on a
Long Thin Cylinder," Proceedings of the
'Royal Society of London, A, Vol. 230,
1955, pp. 188-203.
5. Panton, R.L., "Incompressible Flow,"
1st Edition, John Wiley, New York,
1984, pp. 384-401.
6. Churchill, S.W., "The Interpretation
and Use of Rate Data: The Rate Con-
cept," Ch. 10, McGraw-Hill, New York,
1974, pp. 290-296.
7. Ayers, W.H., Boysan, F., and Swithen-
bank, J., "Droplet Trajectories in
Three-Dimensional Gas Turbine Flow
Fields," Air Force Report AFOSR-TR-81-
0543, September 1980.
8. Eisenklam, P., Arunachalam, S.A.3 and
Weston, J.A., "Evaporation Rates and
Drag Resistance of Burning Drops,"
Eleventh Symposium (International) on
Combustion, The Combustion Institute,
1967, pp. 715-728.
9. Labowsky, M., "Calculation of the Burn-
ing Rates of Interacting Fuel Droplets,"
Combustion Science and Technology, Vol.
22, 1980, pp. 217-226.
10. Dukowicz, J.K., "A Particle-Fluid
Numerical Model for Liquid Sprays,"
Journal of Computational Physics, Vol.
35, 1980, pp. 229-253.
11. Mulholland, J.A., Srivastava, R.K., and
Ryan, J.V., "The Role of Rogue Droplet
Combustion in Hazardous Waste Incinera-
tion," In: Proceedings: Twelfth Annual
Research Symposium on Incineration and
Treatment of Hazardous Wastes, EPA-600/
9-86-022 (NTIS PB87-119491), 1986, pp.
413-420.
12. Marquardt, D.L.,,"An Algorithm for
Least-Squares Estimation of Non-Linear
Parameters," J. "Soc. Indust. Appl.
Math, Vol. 2, 1963, pp. 431-441.
-253-
-------
DISTRIBUTION OF VOLATILE TRACE
ELEMENTS IN EMISSIONS AND RESIDUALS
FROM PILOT-SCALE LIQUID INJECTION INCINERATION
Johannes W. Lee, Robert W. Ross, II,
Ralph H. Vocque, Jerry W. Lewis,
and Larry R. Water!and
Acurex Corporation
Environmental Systems Division
Combustion Research Facility
Jefferson, Arkansas 72079
ABSTRACT
The EPA is currently developing regulations on trace element emissions from hazardous
waste incineration. However, the data base to support these regulations is very sparse.
Data on the effects of waste composition and incinerator operation on trace element
emissions are particularly lacking. In response to these data needs, EPA is conducting
several test series at the Combustion Research Facility (CRF) Jefferson, AR. The first
series of tests was performed to investigate the fate of volatile elements in liquid
Injection incineration. In these tests, trace amounts of arsenic in the form of arsenic
trloxlde (As203) and antimony in the form of antimony trichloride (SbCl3) were fired in a
nvethanol base containing varying amounts of chlorobenzene and carbon tetrachloride. Test
variables included incinerator temperature, excess air level, and feed chlorine content.
As usually occurs in tests of this type, the data show a general inability to obtain mass
balance closure for the trace elements. Both elements are found in the vapor phase at
high temperatures, but they condense to particulate at scrubber exit temperatures. Other
conclusions await further data reduction and evaluation.
INTRODUCTION
In support of the EPA Office of Solid
Waste (OSW) regulation development, the CRF
conducted a series of incineration tests to
determine the fate of two Appendix VIII
metals, antimony and arsenic. These
metal-emission tests are important because
risk assessments show that metal emissions
from otherwise properly operated
Incinerators can be the largest component
of risk to human health and the
environment. For these tests, the CRF
liquid incinerator system (LIS) fired
mixtures of methanol, chlorobenzene, and
carbon tetrachloride, spiked with SbCl3 and
As203- The objectives of the test program
were:
• To determine the amount and form of
arsenic and antimony at various points
in the incinerator system
* To provide air pollution control device
collection efficiency information
• To measure the distribution of trace
elements between flue gas and scrubber
blowdown water
» To quantify the relationship between
trace element emissions and incinerator
operating conditions
» To evaluate whether the presence of
trace elements affects other waste
emission characteristics
-254-
-------
APPROACH
The LIS at CRF incinerates "clean,"
pumpable and atomizable liquid wastes.
Figure 1 shows a simplified schematic of
the incinerator system. Combustion takes
place in two refractory-lined chambers.
The system cleans the combustion product
gases in a packed-column scrubber, an
ionizing wet scrubber, an activated carbon
bed adsorber, and a high-efficiency
particulate (HEPA) filter. An induced-
draft (ID) fan downstream of the HEPA
filter draws a slight vacuum throughout the
incinerator/scrubber system and vents the
combustion products via a stack.
The test series included variations of
three operating parameters: feed chlorine
content, incinerator temperature, and
combustion excess air. A Box-Wilson
factorial experimental design specified 18
test conditions, which included five levels
of each parameter. Table 1 lists the
planned and achieved conditions. Arsenic
and antimony concentrations in the feed
were held constant at 12 and 40 ppm, re-
spectively, for all tests. These concen-
trations in the feed material ensure that
the worst-case concentrations in the stack
gas will never exceed the threshold limit
values (TLVs) of 0.2 mg As/m3 and 0.5 mg
Sb/m3. Furthermore, if all the trace
metals leave the incinerator via the scrub-
ber blowdown water, their concentrations
would not exceed the EP toxicity limits
(5 mg As/L, no limit established for Sb).
Blends of methanol, carbon tetrachlor-
ide, and chlorobenzene in the feed produced
the various inlet chlorine concentrations
(0 to 28 percent). The liquid feed entered
the main combustion chamber via a steam-
atomized nozzle at about 45 kg/hr (100
Ib/hr). Auxiliary propane maintained the
incinerator at the specified temperature
that ranged from 1,035° to 1,450°C (1,900°
to 2,640°F). Combustion air entered via
swirl registers in the burner to produce
from 6- to 11-percent excess oxygen at the
incinerator exit.
Figure 2 identifies the sampling points
and protocols. Continuous monitors, (CMs)
measured 02, CO, and C02; volatile organic
sampling trains (VOSTs) collected samples
at the afterburner and the scrubber exits.
Thermal desorption purge and trap GC/ECD
was used to evaluate 21 volatile organic
compounds.
EPA reference Method 5 was used to
collect samples for particulate load and
arsenic and antimony analyses. The Method
5 train was modified to ensure collection
of antimony and arsenic that passed through
the particulate filter. It consisted of a
probe and a glass cyclone, followed by a
filter and five impingers. The first
impinger contained 0.1 N NaOH. The second,
third, and fourth impingers contained 0.2 M
(NH4)2S208 + 0.02 M AgN03. The last im-
pinger contained silica gel. Following
collection and SW-846 digestion, furnace
atomic adsorption (AA) methods were used to
analyze for arsenic and antimony.
RESULTS
Table 1 shows that 12 out of 18 planned
tests were completed. The remaining six
test conditions were generally unattainable
due to flame stability problems at low
flame temperatures. Of the 12 tests for '
which stable conditions were achieved, data
from 7 are currently available. These
results are discussed below. A final test
report will be published to present all -
results and conclusions.
Trace Element Discharges - • •'•'•
Mass balance for antimony and arsenic
could not be established for any of these
tests. Table 2 lists the fractions of
inlet antimony found at the afterburner
exit, the scrubber exit, and the.scrubber
blowdown water. For each of the tests
where the feed had 28-percent chlorine and
one of the 18-percent chlorine tests, the
samples drawn at the afterburner exit 'could
account for no more than 2 percent of the •'<'•
inlet antimony. Three tests with 18-
percent chlorine showed 53 to 146 percent
of inlet antimony in the afterburner exit
gas. These three are in the range of
acceptable mass balance closure based on
past experience. '•
At the scrubber exit, between 7 and 36
percent of the inlet antimony can be
accounted for. The higher recovery levels
occurred during lower-excess-oxygen tests.
Higher combustor temperature also appeared •
to coincide with with higher recovery at
the scrubber exit.
-255-
-------
Stack
HEPA I— Carbon bed
filter / filter
Building wall
Afterburner chamber
(unflred)
Ha In chamber
Aux. propane,
Liquid feed
Atomizing steam,
Combustion air.
.—Ionizing wet
/ scrubber
Sampling port
Quench chamber,
VentuH scrubber—/
Figure 1. Liquid injection incinerator system.
TABLE 1. SUMMARY OF TEST CONDITIONS
Feed Cl content
(percent)
Primary combustor
temperature (°C (°F))
Afterburner exit
02 (percent)
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Target
18
18
18
28
28
8
28
8
8
28
8
18
18
36
0
18
18
18
Actual
18
18
28
28
8
28
•.»
8
—
....
w»
18
35
0
18
....
17
Target
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
,250
,250
,250
,365
,130
,365
,365
,130
,365
,130
,130
,250
,250
,250
,250
,450
,045
,250
(2,280)
(2,280)
(2,280)
(2,490)
(2,070)
(2,490)
(2,490)
(2,070)
(2,490)
(2,070)
(2,070)
(2,280)
(2,280)
(2,280)
(2,280)
(2,640)
(1,915)
(2,280)
1
1
1
1
1
1
1
1
1
1
1
1
Actual
— —
,336
,228
,307
,137
,290
,399
—
,425
—
—
—
,203
,293
,265
,450
—
,176
__
(2,437)
(2,243)
(2,384)
(2,079)
(2,354)
(2,551)
—
(2,597)
—
—
--
(2,198)
(2,359)
(2,310)
(2,642)
—
(2,148)
Target
8.4
8.4
8.4
10.1
10.1
10.1
6.7
10.1
6.7
6.7
6.7
5.5
11.3
8.4
8.4
8.4
1 8.4
11.3
Actual
—
8.8
8.4
10.6
11.0
10.1
6.7
--
5.9
— —
— -
--
11.7
9.2
8.1
7.5
—
10.8
Note
a
b
b
a
b
a
a
a
b
b
a
_T— j^^jjrj-
aUnable to attain target condition due to unstable flame.
bSample analyses not complete.
-256-
-------
Stack
Parameter
Auxiliary CMs H5 VOST
Sampling Haste Waste fuel Air Scrubber (02, CO, C02, Volume (particulate, (volatile
point feedrate feed feedrate feedrate blowdown NOX) flow _ metals) organics)
1 X X X X
2
3 X
4
5
XXX
X
X X
X X Xa
X
X
aParticulate on1.y.
Figure 2. Sampling protocol.
TABLE 2. ANTIMONY DISCHARGE DISTRIBUTIONS
Antimony discharge distribution
(percent of feedrate)
Test
no.
4
5
7
13
18
16
2
Feed Cl
content
(percent)
28
28
28
18
17
18
18
Primary
combustor Afterburner
temperature
(°C (°F))
1,307 (2,384)
1,137 (2,079)
1,399 (2,551)
1,203 (2,198)
1,176 (2,148)
1,450 (2,642)
1,336 (2,437)
exit 02
(percent)
10.5
10.8
6.1
11.8
10.7
7.4
9.2
Afterburner
exit
flue gas
0.18
0.24
1.8
oai
53.
146.
104.
Scrubber
exit
flue gas
14
13
30
6.6
14
36
20
Scrubber
bl owdown
water
22
16
12
87
124
8.3
12
Scrubber blowdown water contained
between 8 and 124 percent of the inlet
antimony. No clear trend is apparent for
these blowdown recovery rates.
Table 3 shows that, at the afterburner
exit, where temperatures were above 760°C
(1,400°F), the vapor portions of the ex-
tracted samples contained an average of 55
percent (range 24 to 90) of the collected
antimony. At the scrubber exit where the
temperature was 74°C (165°F), the vapor
accounted for an average of 8 percent
(range 0 to 38) of the collected antimony.
Tables 4 and 5 list the arsenic mass
balance and phase distribution data.
Except for one test, arsenic recovery was
poor. Analyses of collected afterburner
exit gas, scrubber exit gas, and blowdown
-257-
-------
TABLE 3. ANTIMONY FLUE GAS DISTRIBUTIONS
Antimony distribution between particulate
and vapor phase in the flue gas
(percent)
Primary
Afterburner exit
Scrubber exit
Test
no.
4
5
7
13
18
16
2
Feed C1
content
combustor
temperature
(percent) (°C (°F))
28
28
28
18
17
18
18
1
1
1
1
1
1
1
,307
,137
,399
,203
,176
,450
,336
(2,384)
(2,079)
(2,551)
(2,198)
(2,148)
(2,642)
(2,437)
Afterburner
exit 02
(percent)
10
10
6
11
10
7
9
.5
.8
.1
.8
.7
.4
.2
Parti cul ate
62
49
53
42
21
76
10
Vapor
38
51
47
58
79
24
90
Particulate
99
62
99
99
100
99.9
82
Vapor
1
38
1
1
0
0
18
.1
TABLE 4. ARSENIC DISCHARGE DISTRIBUTIONS
Arsenic discharge distribution
(percent of feedrate)
Test
no.
4
5
7
13
18
16
2
Feed C1
content
(percent)
28
28
28
18
17
18
18
Primary
combustor
temperature
(°C (°F))
1,307 (2,384)
1,137 (2,079)
1,399 (2,551)
1,203 (2,198)
1,176 (2,148)
1,450 (2,642)
1,336 (2,437)
Afterburner
exit 02
(percent)
10.5
10.8
6.1
11.8
10.7
7.4
9.2
Afterburner
exit
flue gas
ND
ND
1.7
ND
4.0
77
10.2
Scrubber
exit
flue gas
9.0
5.1
58
ND
2.2
9.4
4.3
Scrubber
bl owdown
3.8
2.3
6.4
2.8
3.1
1.9
0.51
ND -- Not detected.
-258-
-------
Table 5. ARSENIC FLUE GAS DISTRIBUTIONS
Arsenic distribution between particulate
and vapor phase in the flue gas
(percent)
Test
no.
4
5
7
13
18
16
2
Feed Cl
content
(percent)
28
28
' 28 '
18
17
18
18
Primary
combustor
temperature
. 1
1
1
1
1
1
1
(°C
,307
,137
,399
,203
,176
,450
,336
(F
(2
(2
(2
(2
(2
(2
(2
Afterburner exit
Afterburner
Scrubber
exit
exit 0?
°)). (percent)
,384)
,079)
,551)
,198)
,148)
,642)
,437)
10
10
6
11
10
7
9
.5
.8
.1
.8
.7
.4
.2
Particulate Vapor
ND
ND
43
.. ND
32
76
83
ND
ND
57
ND
68
24
17
Particulate
100
83
100
ND
100
100
94
Vapor
0
17
0
ND
0
0
6
ND -- Not detected.
water samples could account for no more
than 10 to 15 percent of the incoming
arsenic. The general low recovery rates
rendered it difficult to identify any
trends. However, it was apparent that
most, if not all, of the collected arsenic
was found on the particulate portion of the
sample train. This is in agreement with
the antimony results.
Destruction and. Removal Efficiencies , (DREs)
The tests demonstrated high DREs for
both carbon tetrachloride and chlorobenzene
(Table 6). Chlorobenzene DREs were higher
than those for carbon tetrachloride.
Carbon tetrachloride DREs were greater than
99.99 percent at the afterburner exit.
These increased to greater than 99.999
percent at the scrubber exit. Similarly,
chlorobenzene DREs ranged from 99.999
percent at the afterburner exit to greater
than 99.9999 percent at the scrubber exit.
The available data do not suggest any
discernible effect of temperature, feed
composition nor excess air. This is in
general agreement with previous CRF data,
which has consistently shown high DREs for
concentrated feed materials.
Products of Incomplete Combustion (PICs)
Incineration produces low levels of PICs.
The CRF routinely analyzes VOST samples for
21 organic compounds with GC/ECD. Tables 7
and 8 summarize the higher concentration
compounds found during these tests. At the
afterburner exit, carbon tetrachloride
(POHC) was highest among the 21 compounds
(>100 ug/dscm). Other chlorinated alkanes
and alkenes were present in the 10- to 100-
ug/dscm range. Chlorobenzene (POHC)
concentrations were similar to those of the
common chlorinated'PICs.
At the scrubber exit, PIC concentra-
tions are about one-tenth of those at the
scrubber inlet; i.e., afterburner exit.
Carbon tetrachloride (a POHC) ranged from 4
to 23 jig/dscm. The highest concentration
occurred at a low excess air condition.
Chlorobenzene (the other POHC) ranged from
1 to 4 ug/dscm and did not appear to
correlate with temperature or excess air.
Other PICs include methylene chloride,
which is ever-present at high levels '(15 to
67 ug/dscm). Chloroform was also present
and ranged from 9 to 74 ug/dscm. Other
.chlorinated alkanes and alkenes, hexane,
benzene, and toluene were below 10
ug/dscm.
DISCUSSION
The data evaluated to., date point to the
lack of mass balance closure for the trace
volatile elements. The exact cause for the
low recovery rates for the two elements at
-259-
-------
TABLE 6. POHC DREs
Carbon
tetrachloride
Chlorobenzene
ORE (percent)
Test
no.
4
5
13
18
16
2
Afterburner
exit 02
(percent)
10.5
10.8
11.8
10.7
7.4
9.2
Corabustor
Temperature
(°C ("F))
1,307 (2,384)
1,137 (2,079)
1,203 (2,198)
1,176 (2,148)
1,450 (2,642)
1,336 (2,437)
Feed
concentration
(percent)
22
24
11
11
11
11
Afterburner
exit
99.9988
99.9982
Scrubber
exit
..a
99.999945
99.99983
99.99973
99.99953
99.99985
Feed
concentration
(percent)
24
22
24
23
24 '
23
DRE
Afterburner
exit
99.9996
99.999995
(percent)
Scrubber
exit
99.999984
99.999987
99.999981
99.999981
99.999985
'Data not available; sampling train contamination occurred.
TABLE 7. AFTERBURNER EXIT PIC DATA
Test no. 42
Afterburner exit 02 (percent) 10.5 9.2
Primary combustor temperature (°C (°F)) 1,307 (2,384) 1,336 (2,437)
POHC concentration (tig/dscm)
Carbon tetrachloride
Chlorobenzene
PIC concentration (yg/dscm)
Methylene chloride
1,1-dichloroethylene
Chloroform
1,2-dichloroethylene
Tri chloroethylene
Benzene
1,1,2-trichloroethane
Hexane
Toluene
110
37
18
38
70
38
21
13
61
15
12
150
11
77
6.
12
58
14
3
23
10
11
-260-
-------
TABLE 8. SCRUBBER EXIT PIC DATA
Test no.
Afterburner exit 02 (percent)
Primary combustor temperature (°C (°F))
5
10.8
1,137
(2,079)
13
11.8
1,203
(2,198)
18
10.7
1,176
(2,148)
16
7.4
1,450
(2,642)
2
9.2
1,336
(2,437)
POHC concentration (ug/dscm)
Carbon tetrachloride
Chlorobenzene
PIC concentration (ug/dscm)
Methylene chloride
1,1-di chloroethylene
Chloroform
1,2-dichloroethylene
1,1,1-t ri chloroethane
Tri chloroethylene
Benzene
1,1,2-trich!oroethane
Hexane
Toluene
4.4
1.6
7.5
1.4
12
1.6
23
1.8
12
4.0
15
0.6
74
8.7
0.2
1.3
2.6
9.5
0.9
1.6
34:
ND
8.9
NDa *
0.8
1.9
1.3
1.9
1.9
1.9
42
ND
8.8
ND
0.6
3.3
1.6
0.5
5.5
8.0
67
ND
18
2.4
1.0
2.1
1.0
5.2
2.0
2.7
18
ND
25
3.1
1.3
1.6
6.3
ND
2.0
4.0
ND — Not detected.
the afterburner are not known. It may be
due to inconsistent capture efficiencies or
the sampling train's ability to capture
these volatile metals. Other unknown
factors may play a role in the sampling
train's performance.
The information obtained to date tends
to suggest the following:
• The presence of trace amounts of
antimony and arsenic does not affect
POHC ORE.
• Both arsenic and antimony appear to
show up in considerably greater amounts
as solids, especially at lower
temperatures.
The data reported here are complex.
Trends identified earlier are weak at best.
Upon completion of the sample analysis of
the remaining test samples, a clearer
picture may emerge. These will be reported
in the test final report, to be submitted
at the conclusion of the test program.
ACKNOWLEDGEMENTS
The work reported in this paper was
performed under the CRF operations and
research contract with EPA/HWERL. This
support and the guidance provided by the
EPA Project Officer, R. E. Mournighan, is
gratefully acknowledged.
-261-
-------
ASSESSMENT OF RESIDUES FROM INCINERATION
OF RCRA WASTES
Joan V. Boegel
Metcalf & Eddy, Inc.
Wakefield, Massachusetts 01880
ABSTRACT
Incineration is generally recognized as a well-demonstrated technology
for the treatment of organic hazardous wastes including spent solvent
wastes. Most studies of incineration have been concerned with the
effectiveness of the process to destroy key organic constituents of a waste
(destruction and removal efficiency, DRE) as measured by the relative quantity
of those organics in the incinerator off-gas. In contrast, this paper focuses
on characterization of the solid and liquid residues generated by incineration
of RCRA wastes.
Two incineration systems are evaluated - one at a commercial treatment,
storage and disposal facility (TSDF) accepting organic wastes from a variety
of industrial generators (Facility A) and the other operated on-site at a
chemical industry manufacturing plant (Facility B). Both systems generate two
types of residue - ash and scrubber wastewater. Ash from both facilities is
currently landfilled. Treatment of the scrubber wastewater at Facility A
results in a metal sulfide sludge, which is also landfilled. At Facility B,
scrubber wastewater is neutralized and injected into a deepwell on site.
All ash, sludge and wastewater samples collected at these facilities were
analyzed for priority pollutant organics and metals. The ash and sludge
samples were also subjected to the Toxicity Characteristic Leaching Procedure
(TCLP). Ash from Facility A exhibited unacceptably high TCLP extract
concentrations of two volatile organics - methylene chloride and
tetrachloroethylene, indicating incomplete combustion of solvent wastes. Ash
from facility B passed the TCLP for both metals and organics, but both ash and
extract levels of three non-TCLP metals - Copper, Nickel, and Zinc - were
high. Scrubber wastewater from both facilities had no significant
concentrations of toxic organics. However, copper, lead, nickel and zinc were
found at concentration greater than 50 mg/1 in the scrubber wastewater from
Facility B. This paper presents and evaluates quantitative data describing the
wastes incinerated and the resulting residues at both facilities.
-262-
-------
INTRODUCTION
The RCRA Hazardous and Solid Waste
Amendments of 1984 call for a ban on the
land disposal of hazardous wastes. EPA is
required to evaluate technologies for each
waste category and to establish treatment
performance levels for wastes based on the
lowest levels achieved by a "best demon-
strated available technology". Treatment
levels from incineration have been
selected for F001-F005 spent solvents.
Incineration is also likely to be an
important technology for most other listed
organic wastes. Therefore, the quantity,
chemical characteristics, and fate of
incineration residuals are important.
Incineration is a controlled high
temperature oxidation process by which
organic wastes are destroyed. The com-
bustion reactions of incineration should
convert organic wastes to gaseous C02 and
water vapor. If the waste contains cnlori-
nated organics, then HC1 and free chlorine
gas are also generated as incineration by-
products. Incineration of wastes contain-
ing organic sulfur or nitrogen results in
formation of SOx and NOx, respectively.
Air pollution control devices - often wet
scrubbers - are used to control emissions
from hazardous waste incinerators.
The residues from incineration are
the off-gas, ash and scrubber wastewater.
The effectiveness of an incinerator
and its associated air pollution system is
normally measured in terms of destruction
and removal efficiency (ORE), as defined
by the following formula:
10055
Win
Win =
mass feed rate of the principal
hazardous constituent(s) to the
incinerator
Wout = mass emission rate of the
principal organic hazardous
constituent(s) to the
atmosphere as measured in the
stack prior to discharge.
This formula considers only off-gas
emissions to the atmosphere, ignoring the
other residuals of incineration - ash and
scrubber wastewater.
Samples of wastes, ash and scrubber
wastewater were collected from two full-
scale incineration facilities as part of a
program to evaluate alternatives to land
disposal. This paper summarizes the
operations at both facilities and presents
and evaluates quantitative data describing
the wastes incinerated and the resulting
residues.
FACILITY A
Facility Description
Facility A is a commercial hazardous
waste treatment facility accepting drummed
organic, hazardous wastes from a variety of
industrial generators. Typical wastes
accepted for incineration at Facility A
include chlorinated and nonchlorinated
solvent wastes (F001, F002, F003, F004,
F005), paint formulating wastes, furniture
strippers, printing inks and dyes, polymer
wastes, agricultural products,
pharmaceutical production wastes, lab
packs and contaminated soils. The facility
does not accept PCBs, dioxins, compressed
gases, radioactive waste, or organic
wastes containing significant
concentrations of potassium, sodium,
lithium, mercury or lead.
Figure 1 shows a process flow diagram
for Facility A. Waste organic liquids are
pumped from drums into one of four 6000
gallon steel blend tanks. Waste liquids
are selected such that the final blend has
greater than 8000 BTU/lb and less than 40$
chlorine.
Drums containing primarily solids,
nonpumpable materials or residues
remaining from liquid pumping are
transported to the drum processing
building. Operators use an air-driven
cutting tool to remove the tops of steel
drums. Drums are then upended and dumped
-263-
-------
QBf,
£io
C 0)
II
II
-264-
-------
onto a coarse-mesh screen above a steel
tank. Any free liquids drain into the tank
and are subsequently pumped to a blend
tank. Solids remaining on the screen are
shoveled into 30 gallon fiber packs. Empty
drums are steam-cleaned, and the cleaning
solutions are added to a blend tank.
The incinerator at Facility A is a
dual chamber unit with a thermal input
rating of 19.8 million BTU/hr. All wastes
are introduced into the lower chamber.
Liquid waste from the blend tank is
injected into this chamber by atomization
with steam. Solids in fiber packs are
loaded into the incinerator by a ram
feeder. Solids residence time in the lower
chamber is estimated at 15 to 30 minutes.
The upper chamber is an afterburner used
to complete combustion. Fuel oil is burned
in this chamber to maintain required
temperature.
Dry ash residue is conveyed from the
lower chamber out of the incinerator
building and is dumped into a roll-off
container. The ash is transported off-site
to a permitted hazardous waste landfill.
A caustic scrubber is employed for
removing air pollutants, particularly HC1,
from the stack gas. This scrubber permits
burning wastes with up to 40 percent total
chlorine content.
The approximate feed and effluent
flowrates during the sampling period were
as follows:
Feed
Solid waste = 48.1 ft3/hr
Liquid waste = 28.3 ft^/hr
Residuals
Bottom Ash = 5.25 ft^/hr
Scrubber Wastewater = 722 ft^/hr
Scrubber wastewater treatment
includes pH adjustment, hydroxide and
sulfide precipitation of heavy metals,
clarification, pressure chamber
filtration, and sludge dewatering with a
recessed plate filter press (Figure 2).
Scrubber water collects in a 400
gallon steel tank and is pumped to the
scrubber wastewater treatment building.
The wastewater flow is typically 90 gpm.
The wastewater pH is adjusted to 9.0
with 25 percent caustic in a 7,000-gallon
mixed tank. Sodium sulfide is added to
precipitate heavy metal sulfides which
remain in solution after hydroxide
precipitation' at pH 9.0. Ferric chloride
is added to this tank as a conditioning
agent to improve clarification and
filtration. High cationic synthetic
polyelectrolyte polymer is added before an
in-line static mixer as the wastewater is
pumped to the clarifier.
Wastewater is clarified in a 20,000
gallon steel tank which overflows to a
second 20,000 gallon steel storage tank.
Clarified effluent is then pumped through
a seven element filter. The elements are
operated in a parallel and each element
contains a 200 mesh polypropylene filter
bag. The filter is backwashed once each
shift, and the backwash water is returned
to the clarifier.
A portion of the treated, filtered
wastewater is recycled to the scrubber, A
blowdown stream is stored in a third
20,000 gallon steel tank. Blowdown
wastewater from this storage tank is
trucked periodically to the municipal
Wastewater treatment plant.
Sludge from the clarifier is pumped
to a 7000 gallon steel conditioning tank.
Synthetic polyelectrolyte is added to this
mixed tank as a filter aid. The
conditioned sludge is pumped to a recessed
plate filter press once each shift. The
filtrate is returned to the clarifier, and
the dewatered sludge is collected for
landfill disposal.
-265-
-------
g
I
I
-266-
-------
Waste Characterization
During a week-long sampling program,
samples of five discrete solid wastes and
five discrete liquid wastes were obtained.
In addition, a sample of the liquid waste
from the blend tank was taken. As Table 1
shows, the major volatile organic
components of these wastes were the common
solvent compounds acetone, benzene,
chloroform, methylene chloride, methyl
ethyl ketone, tetrachloroethylene,
toluene, trichloroethylene and xylene.
These nine components account for about
54 percent by weight of the liquid waste
blend. Surprisingly, the liquid waste
contained 28.7 weight percent water and
its fuel value was only 6,824 BTU/lb,
below the plant's stated goal of
_>8000 BTU/lb. The liquid blend's organic
chloride content of 27.99 percent was
below the plant's stated upper limit of
40 percent.
The five solid wastes represent
diverse industrial sources. Information
supplied to Facility A by the waste
generators indicates that the five solid
wastes were (1) epoxy resins, (2) used
nylon mesh filter bags, (3) waste
tetrachloroethylene and wax mixture,
(4) waste methylene chloride and polymer,
and (5) eye shadow with tetrachloroethylene.
Measureable levels of chromium,
copper, lead, nickel and zinc were found
in all waste samples. The highest metal
concentration was 759 mg/L copper in the
liquid waste blend.
Residue Characterization
The wastes described in the previous
section were incinerated during a 36-hour
period. Beginning 4 hours after the first
fiberpack of solid waste was fed to the
incinerator, grab samples of ash were
collected from the ash chute every two
hours until the incineration run was
complete. The individual grab samples
were composited to form a single ash
residue sample representative of the
burn. Tables 2 and 3 summarize the
analytical data for this sample.
Compositional analysis of the ash
shows significant concentrations of
several volatile and semivolatile organic
compounds. The tokicity characteristic
leaching procedure was run on the ash
sample. Extract levels of two volatile
organics - methylene chloride and
tetrachloroethylene - exceed the
regulatory levels for these compounds,
meaning that the ash must be classified as
a hazardous waste.
Metals appear to have concentrated in
the ash as expected, with most metals
present in the ash at levels greater than
their concentrations in any of the waste
samples. Arsenic exhibits the greatest
concentration factor. Its ash
concentration - 42 nig/kg - is 52 times the
highest waste sample arsenic
concentration. Copper at 13,800 mg/kg is
the major metal component of the ash.
Both TCLP and EP extracts were analyzed
for toxic metals. None of the eight
listed EP Toxicity metals were found above
regulatory levels in either extract.
Comparison of extract and compositional
metals levels indicates that the ash is
fairly resistant to leaching of metals.
The ash had a total solids content of
about 60 percent by weight and a specific
gravity of 1.28.
Data on the scrubber wastewater and
its treatment products are summarized in
Table 4. With the exception of acetone
which was also found in the laboratory
blank, none of the organic compounds
measured in the wastes or ash were
detected in the scrubber wastewater or the
filtered effluent.
Seven toxic metals were reported
above detection levels in the scrubber
wastewater at concentrations ranging from
0.6 mg/L for arsenic to 11.0 mg/L for
copper. Metal concentrations in the
-267-
-------
8
•H
J3
n)
N
LI
Q)
o
co
LI
cd
6
X>
CO
n<
3
4J
O
£
T^
a
X
n
ir\
•a
•H
•g
CO
o
0 0
10 - -2 SJ
nt X! £ C
x: x> o co
CO "H (fl i-H
cu S -P J?
W EH CD
a> »a
rr
•a
.-I
w
m
2
i-H
o
CO
C Li
Q) tt) CO CU
jj i-i -a E
CO r*»*lH P*>
«! X: L, ,-H
3 -u ; o o
Q> rH Cu
0
o m
1?
L, m
JJ
O) 0)
H §
0) i-H
co x:
a) -u
3
on
ooooooooo
ooooooooo
too C^OOt—.=3*vO CMCO
COCMCMT~OOOvOC— CM
o =f <^ en on oo co
.=r \o CM en o «-
o
— . (DO)
CO TD O (U . 'a) a.
S 0 trf XJ M —
OS O >> O -P J
0 guSb1 § g
Ct] O C W •-! t« H
^J 0)0) X1QJO 2-
E-* OQ)£->j>*rtl> O
> «3lCQOS.Ec-|c-|c-|X t—
O'— inCMCMOOOOOOO
• - CM • • O • • • • '
j3~ ^^ T— ^3 ^- OO ^3 LO LO Cj ^O
V • ^" f1"
Q
0<-^OCMt-OCOOO
*— O«— OOOOOCM>J3
o
O'— ocMinc— ooo
v-O'-OOOCM'-CM
0
OCOOCMOOOOOO
«— ooO'— in^-ooo
v v o »-
O»— OCMLO"— OOO
.— O'— OO -CMOOO
v
CM CO
o o
t- 0
o in
=r o
«- o
t—
«- o
O CM
. T— <— . -co • ir> • •— •
<— • «— O O • O C— •=!• 'CO
VO v^"— CM OOO
4J
c
CO
1 — 1
«J
cd "cd
X -p
a> o
g i-|
c1 o -3 E 3 3
O -H § "—I 3 >(H "'H Li
>>
Li i-H
•iHco-H>>Eooo.'ap^
4i CQ Li Li TJ Lj tj M* C
S^f^t^cjooOb.
*~
1 *•« *-*
1 ^ S
-268-
-------
c
c
a
is
£.
O
ctf
t.
a.
X
u
0)
03
n)
3
cfi
4-5
•iH
|jj
O
cd
• — •
lu
§
•rH
§
0
0)
rH
a)
E~*
O
in o o
=*= T3 rH (U
•a xi 4-5 o (u
•rH CO -H s
CO >> 0) J->
ca H ,.H ;»
•H «) XI £-. rH
rH 3 4J 0 0
O 0) rH 0,
co 2: x;
o
i
o
S- X
O (B
" — 1 3
00 X!
=* o -o
nJ c
"0 £H Cfl
•H 4-3
rH 0) tU
O E-< C
CO *,
03 S
3 4-5 60
•HI — i i — i
NX *— .
O =f O O
c- o ro in
o
0 CM O O
c— o •*- co
NX NX
VO Cvl O CM
• » • •
«— O «— «—
NX NX VO
CO
ta
CO
^
11 §
•rH t, -H
C (1) rH 03
0) > rH o Cd
rH rH OJ C 3=
co co e-< t>a o
^3- t— CO «- ^- O «-
VO O OO OMn O CM
• • • • CTi CO O
ON c— «— CM in - -
CM • CM =r
t- vo
VO
o .=r in c— in o o>
t- ^r CM ^r o o *-
«- «- CM • CO C—
»- CM
in «- in «- in o o
cr. in «- 3- on o in
• • • • «- co t~
o cr>-r- CM ro - >
co in • o in
CM
ON O CO C— in O )— i
co .=r co m^r o s
• • • • C- CM Q
=r co CM o oo -
oo • <-
^— CO
in
o co crv «- in o o
vo o =r ro «— o in
. . . . CM <— co
co en o o o - -
vo • in in
' — -=r <~
c—
in CM t— cr- =f o 3-
CNJ O • CTvCO O CM
• -CO • O O CO
CM t— CM t— • - -
ro CM «— vo vo
in
o 03
C 60 -rH >,T3
O O £- C 4-> -H
XI t. o
al >>
-------
Table 2. Facility A Ash Analytical Data* - Organics
PARAMETER
Volatile Organics
Methylene chloride
Acetone
Chloroform
2-Butanone
1,1,1 trichloroethane
1,2 dichloropropane
trichloroethylene
benzene
l»-methyl-2-pentanone
tetrachloroethylene
toluene
chlorobenzene
ethylbenzene
styrene
xylenes
raethanol
Semivolatile Organics
Phenol
Nitrobenzene
2,4 dimethyl phenol
naphthalene
2-nitroaniline
dimethyl phthalate
diethyl phthalate
di-n-butyl phthalate
Compositional
(ug/kg)
38,000
20,000
46
2,000
12
32
120
42
2,300
1,200,000
2,500
27
380
320
1,900
410,000
40,000
29,000
23,000
24 , 000
180,000
55,000
120,000
160,000
TCLP Extract
(jig/L)
-
8,800
<3,300
<1,700
<3,300
< 1 , 700
<1,700
< 1,700
< 1 , 700
<3,300
48,000
11,000
<1,700
< 1 , 700
<1,700
< 1 , 700
M.M.
< 1,400
<200
< 1 , 000
310
1,300
370
410
<200
Regulatory
Level
(yg/L)
8,600
None
70
7 , 200
30,000
None
70
70
None
100
14,400
1 ,400
None
None
None
None
14,400
130
None
None
None
None
None
None
*Analysis of a single composite sample. , -
composite were collected every two hours during the incineration run.
-270-
-------
Table 3. Facility A Ash Analytical Data* - Metals
PARAMETER
Compositional
(mg/kg)
TCLP Extract
(mg/L)
EP Toxicity
Extract
(mg/L)
Regulatory
Level
(mg/L)
Toxic Metals
Antimony
Arsenic
Barium
Beryllium
Cadmium
Hexavalent Chromium
Total Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
Other Analyses
Total solids (mg/kg)
Specific gravity (g/mL)
Paint Filter Test
8,0
42,0
150
<0.2
2.0
0.083
71.0
13,800
30.0
0.2
190
<1.0
0.4
2.0
280
599,300
1.2809
PASS
0.094
0.062
0.026
<0,005
0.02
NM
0.01
0.729
<0.05
0.00025
1.14
<0.001
<0.005
<0.001
1.15
0.2
<1.0
<0.2
<0.5
NM
<0.3
4.0
2.0
<1.0
<0.2
<1.0
0.3
None
5.0
100.0
None
1.0
None
5.0
None
5.0
0.2
None
1.0
5.0
None
None
*Analysis of a single composite sample.Aliquots of ash making up the composite
were collected every two hours during the incineration run.
-271-
-------
40
f
C
0)
E
JO
03
,
"c
n
In
rr
a
j;
H
40
O
n)
) LI
jj
X
w
•° ;>,
JJ
' a
i X
1 O
1
a a.
w
(!)
00
T)
3
CO
•a
0)
L,
0)
jj
^
0)
o
JJ
c
d)
3
d
8
CO V N/ ^
a
V.
UC
5 0> § 0) C 0) • "^ "Sdjltf
Hl-S §£§— ^ rH^rti. i i-* "
I- f\ "^ c~* L .O I O fl) flj *"• C ^} ttJ "^ •*-* •*-*
A tj t=S>>?Sm (-VI51 NK1 -H (UjSC'HD.Xl'H
° n% eSSSn 10 CC 40 N4O(UC O.>>
2 ^mcT-2 ^4^ fem rH 0) e>* 5 S § -S •=>• a C B _o> ^
o ^^g-g^T^T^^-j.^S'S'cijwx'E co euist\rcc^-a-OT3
-272-
-------
n
4-
cd
Q
c
(1
a
CU
i.
I*
fr
a.
s
a
V
3
J_^
a;
XI
X
c
CO
•=&
4-5
rH
•H
o
cd
73
(U
4J
c
o
o
.j!
0)
i-H
X!
cd
o
cu cd
60 't,
73 4^
3 £4
rH Cd
CO
'S.H?
s- o
CO -H
4-5 X
cd o
3 E-H
cu
a a,
Cd
CO
60
73
3
CO
73
CO
£4
0)
4-5
cd
3
CU
Q
c
CO
3
clj
Cu
Cd
73
0)
cu
rH
•H
SL,
s
tf
>2
D ^-,
n c
cd cd
3 E-<
5 a.
XI — '
XI
3
3
CO
RAMETER
•=fi
Cu
O «, O CVJ LO CO Q O < — CO O C\J O i.-.
••- o''- OOSO^'-OO^O'-Q ' ' ' ' ' '
\X NX NX
*— -=3" co •oomoo^oo • • • o o s s cvi s co
co <— <— o .=t- • co vo CM •— o co *- .=r cr\ co
NX o CO *- CM NX CO - CM -
NX 0 • C-
in '-co
co.
°. incjrvj.u?'ocy?oco^;t-0. C^OCD £2. jg! S £?; ™ ij
*~~ f~i * — ^^ ^^ r""t ^^ i o - C^ * — ^^ T— . * •* *. ^* ^^ ^*
V V XX NX V V^VVV IgLg _'
o vo °. ^ c- [n ^- °. o ^ o °. ^ °. o S S ° S "^ §
^ °' v S °' P "~ ^ "^ S =•• v S v c°' vo" oT^° °- ^
CO CO »-
uJ uJ
60 60 CO
E E 0
cd
to co o
T3 73
•rH -H to
^ E -— rH rH Cd
i-J 3 J O O
^ 'E ' ^ 6O
•— J-.E IO---Q)CI)-HE
x: 3 CD > 73 >
K> O-H tOCOrHCCd----
•-J E > 73 O CO £-.
Cd _-*iP rH-rHtOCXO>,
4-» ECt. CdrHtOt04J
^ £>,, ^-^-C EE C O-H3O -H
s o-HErni1"1 ^5 3 "* cooco-rH c
CJ EC3rH—f>rH(l) 30>CCl)rH Ll tHrHrH-rH ,H
•H •rHO).H>,EcdCdQ.73O^l!(U>rHO 0) CtfCdCdO Cd
O CLCdCDCdCUOOCUCJ -H. CP-HXI'rH 4J OOOCU^CrH
E~<
3
CO
cd
1 •
•H C
X! 0
•H
C 4->
•H cd
c
73 -H
73 C E
0) 3 cd
Si O 4->
to o
cd co a
CU 4J
E ^s -ZyS
rH C
4-5 Cd Cd
O C -H
C =C XI
II II
S CO
-273-
-------
filtered effluent are less than those of
the scrubber wastewater.
No information is available on
Facility A's discharge permit to the local
POTW. However, strictly for purposes of
comparison, Table 5 presents filtered
effluent data for five metals along with
the daily maximum concentrations of metals
permitted to be discharged to publicly
owned treatment works (POTWs) by metal
finishing operations (Effluent Guidelines
for Metal Finishing. Amended September,
1981). Facility A's effluent exceeds the
metal finishing effluent guidelines
concentrations for copper, lead and zinc.
The dewatered sludge from scrubber
wastewater treatment was about 45 percent
total solids by weight. Metals
concentrations in the sludge were greater
than 1000 ppm for copper, lead, nickel and
zinc. An EP Toxicity test was done on
this sludge. None of the metals exceeded
regulatory levels in the extract.
FACILITY B
Facility Description
Facility B is a chemical
manufacturing plant owned and operated by
a major chemical company. Both solid and
liquid hazardous organic wastes generated
at Facility B and at several of the
company's other plants are incinerated at
Facility B.
Solid wastes and sludges are received
in drums and typically include paint
sludges, chlorinated hydrocarbons, coke
solids, vacuum filter solids, waste filter
elements and polymeric tar.
Liquid wastes are transported by
tanker truck and are transferred on site
to one of four 20,000 gallon waste storage
tanks. One is used as a feed tank, another
as a waste blending tank, and the
remaining two as holding tanks. Liquid
wastes commonly include waste organic
solutions, waste solvents, tank farm
nitriles and chloroprene catalyst sludge.
The incineration system (Figure 3) at
Facility B includes both a rotary kiln
incinerator and a single-stage liquid
injection incinerator. The total system
thermal- input rating is 40 million BTU/hr.
Both the incinerators use natural gas as
auxiliary fuel. The liquid injection
incinerator burns only liquid wastes,
which are predominantly generated from the
manufacturing operations on site. Ash is
removed about four times per year from the
liquid injection incinerator. The rotary
kiln incinerator will typically burn the
same liquid wastes, in addition to a
variety of solid wastes.
Solid wastes are all contained in
drums. The drums are fed to the kiln by an
automated feed system. First the drum top
is removed, and the drum is lifted
vertically up an elevator. It is then
transported horizontally across a
conveyor. The drum then enters an airlock
before being inverted and the contents
dumped into the rotary kiln. Empty drums
are returned down the elevator and placed
in storage. If the drums must be cleaned
they are washed with toluene or water and
set out to dry. The toluene and water
wastes are also incinerated.
The rotary kiln is followed by a
natural gas-fired afterburner designed to
further burn gases exiting the kiln.
Normal operating temperatures for the
rotary kiln are 1200-1400°C. Afterburner
temperatures for the rotary kiln are 100-
150°C higher than the kiln. Average solid
residence times in the kiln are between 40
and 60 minutes, and the kiln rotation is
between 0.10 and 0.20 revolutions per
minute. Ash is discharged continuously
from the kiln to an ash sluice.
Liquid wastes, as described
previously, are stored in four (20,000
gallon) tanks. These tanks are
recirculated by pumps to mix the tank
contents. It normally takes 2 to 3 days to
burn the contents of one tank.
-274-
-------
'es
2
•3
o
O)
2
a.
CO
cfi
£
-275-
-------
TABLE 5. FACILITY A EFFLUENT METALS
Metal
Filtered
Effluent (mg/L)
Metal Finishing
Effluent Guidelines (mg/L)
Copper
Nickel
Lead
Chromium
Zinc
5.0
1.7
1.3
<0.3
5.0
3.38
3.98
0.69
2.77
2.61
-276-
-------
The liquid incinerator is air
atomized and care must be taken to prevent
plugging of the feed lines. Tanks and
feed lines are flushed with toluene
between burning incompatible liquid
wastes. Normal operating temperatures for
the liquid incinerator ranges from 1400 to
1600°C.
Each of the two incinerators is
followed by a water quench and a cyclone.
The quench serves to cool effluent gases.
The cyclone acts to remove entrained
particulates and droplets prior to
scrubbing the exhaust gases.
Gases exiting the two cyclones
combine to form a single exhaust stream.
The combined gas stream then passes
through a three-stage packed tower
absorption (scrubber) system for removal
of particulates and HC1.
There are two waste streams generated
by this incineration facility: scrubber
wastewater and ash residue. A portion of
the scrubber blowdown (about 10-12,000
gal/hr) is neutralized and processed at an
on-site wastewater treatment facility
prior to deep well injection. Ash residue
is typically allowed to accumulate in the
ash sluice and slurry tank for about four
weeks. The rotary kiln is shut down
approximately once a month for ash
removal. Ash is landfilled on-site.
Waste Characterization
During a week-long sampling period,
Facility B incinerated the contents of one
20,000 gallon liquid blend tank along with-
183 drums representing five different
types of solid waste. The solid waste
types were: (1) chlorinated hydrocarbon
coke from dlchlorobutene synthesis,
(2) polymeric tar, (3) filtered organo-
metallic waste solids from adiponitrile
area, (4) adiponitrile spent filter
cartridges, and (5) scrap paint solids.
Analytical data summarizing the chemical
and physical characteristics of each waste
are presented in Table 6.
Toluene is the major volatile organic
component"of the waste. Several
semivolatile compounds were reported in
one or more of the waste samples. Copper
and, nickel are the two metals present at
highest concentrations in the liquid waste
and they are reported at significant
levels in each of the solid wastes as
well. Solid waste #5, the scrap paint
solids, exhibits the highest metals
concentration with high levels of lead,
chromium and zinc, three common pigment
metals.
The liquid waste blend had a water
content of about 17 percent,:an organic
halide content of 14.24 percent and a fuel
value of 11,926 BTU/lb. The fuel value of
the solid wastes ranged from 5,866 BTU/lb
for the polymeric tar to 12,000 BTU/lb for
the scrap paint solids.
Residue Characterization
No volatile or semivolatile organic
compounds were found in the ash from
Facility B. This ash had a total solids
content of about 81 weight percent and a
specific gravity o'f about 1.94. The ash
contained nearly 18 weight percent
silica. Measurable concentrations of
eight toxic metals were reported, with
copper (4,600 mg/kg), nickel (4,200 mg/kg)
and zinc (1,160 mg/kg) being present at
the highest concentrations. Although none
of the regulated TCLP metals were found
above regulatory level in the TCLP
extract, extract levels of copper, nickel
and zinc were quite high.
Facility B's scrubber wastewater • is
very acidic, wi-th pH <1. Four toxic
metals - copper, lead, nickel and zinc -
are present at concentrations greater than
25 mg/L. No volatile or semivolatile
organic compounds were found in the
scrubber wastewater.
-277-
-------
*
§
*H
03
a
t_
CO
JJ
O
n)
i.
o ro c— ON o
C— C— O O T— C— O
fO ;=r O CO CO =»• «-
O «~ NX
o"
ro <-^
Q
a
XI CU
Q C CO
Q O O
"^ (fl Z
CO JJ > J
r | o CO CO i~H Q) C t ^ ^>
J t. C rH S: >>rH E S
O,i u
ooooooooo
ooooooooo
t— t— C— Ot—C— t— CMC—
CM
anaanaooci
S2SSS3OSS
o
*.
3;
OO
""
ooooooooo
ooooooooo
OOOOOC— OOO
oc— c— oooooo
ON in CT» CT* O C — CT* ON CT*
»— m^r ^r ON »— «— ^—
NX *• NX NX NX
CM
ooooooooo
ooooooooo
ooooooooo
ooooooooo
VO VO VO VO VO VO VO VO vO
ooooooooo
ooooooooo
C^- ^~ C"^~ C~~ C^- C"~ C— t— C^-
ooooooooo
ooooooooo
ooooooooo
Ot— OOOOOOOO
ovoooinininininin
^"^.^
«— *- OO
CO CO
C 4J
•rH 0) Cd
Q) E CO 4J rH
C CO C a) cd — -
CU rH CO rH X! E
N >, rH Cd 4J Q
c c cd s: ^: a
rH CO CO X! 4J O. -— '
o xi x: 4J si
£ O Q. S2 CO Q-rH1 CO
OOT3Cn)CO rHN <=£
rHrHOCUCX: r»>C EH
>x: 3 xi S
rH-rH L.X:XI CUrHXI
>,-a 40 4-> -P cd o i rH o
l^fffll Z t g
03 «— C C CM cd CXT3 XI E-"
CM *- CT> CM Ln in
«— 'CM • -CO
0 000
^ \/ \s •
o
<- «- in CM in in
o a- o o o
NX NX NX •
o
vo »— »— CM on i— i
«— CM CM • -CO
,- . o «-
O NX
a- •r- T- CM in vo
i NX • 'CM
o o o •
*—*—<— CM in ON
NX • NX • • O
o o o •
\S V V
T— *— t— c\j in oo
\s • \x • • OJ
o o o •
J_J
c
CU
t-H
cd
^
cd
» S O
C i-i cd o cd .C
cc
-------
*
c
o
Jj
M
_J
*r|
rti
• ^
o
2
r*
O
40
Cd
UJ
*^
C3
.5
"tj
"'":
•.-1
O
Cti
Ei-t
^^
•a
QJ
?•
C
*""^
r*
0
^
*^^
.
^0
0)
*o
£
in
•a
•H
rH
o
CO
^f
s»fc
•a
•rH
rH
£
m
:**-
•a
•rH
rH
0
co
CM
=«(=
•a
. -H
rH
O
CO
,_
=«=
•a
•rH
rH
3
T3
3
cr
•H
,_]
40
C
•H
cd co
a, -a
•rH
O.-H
cd o
t. CO
CO
CO CO
40 CO
rH 60
•H -a
[*Tj •!— (
£-.
C £-
O 0)
C 40
fl) CO
bO cd
£-3
O vo men -vo
CTv O • in • vO • 1 • " — VO CO
C~-J3- O • O «— O «— O~> ON • C— 1 in • CM -
vo t— on co in ^ ^* ^ *~~ CM vo CM i *— *^ ^^ in
CM CM CO a-
m <— CM «— o in a- m -a-
CO • CM • O • t — ^" 1 -=3~ CM CT> CTv
• in >— o • *— o >— • cn-'i ••in-
o
CO O
o o <- =r vo co
O ' — CM O ' — CTN C — in —a~
•r-t-- • -co • vo • i c— o in CM
. -.— o <— «— o <— • co-cni ..vo-
CM in .=3- o in
vO O vO
vo vO ^3- <— C— O CM
<— CM c^vo «— CM O in o cr.
CM C— • t— -CO ..Z3-...CT.--
•CO • O CM »— O »- • a- • t— 3- O -OO-—
CM vO < — N/ CM N/" "^ v -=t~ In CO ' — * — v O t — « —
^— ^
bO
, "O
1 J C 60 --H CB 40 .-I
•=E OOL.CO-rHr-1
E E E 2Xis-cocd'H>o
3 >>33 S Cd S- •«-< 2- Xt
O O. *O CJ *v^ CO ^ ' — i O Cd Cd *^
£-,cxcd£-OrH.— icdc X «e.^s.^a.^s.s«. -403
X:OCOCO'HCO'rHX:-H E-i 4040404040Q.OE-1
CJO-JSZCO'/lf-cJ O 33333COE-ICQ
•
co
CO
rH
Q.
E
cd
CO
X!
, fd
60
CO
£
x:
40
0
CO
60
cd
^,
C8
CO
x:
40
CO
40
CO CO C
O rH CO
c a. co
CO E CO
s-. cd t.
CO CO Q.
OH CO
2_j 4_J j_,
CO C
40 CO CO
C -r-l 3
PH O rH
— 1 Cd
CO Ci_i >
'3. 3 x:
E 03 0
cd c cd
CO r-H Cd
II II
1— I CO
CO — *
-279-
-------
Table 7. Facility B Ash Analytical Data * - Metals
PARAMATER
TOXIC METALS
Antimony
Arsenic
Barium
Beryllium
Cadmium
Hexavalent Chromium
Total Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
OTHER ANALYSES
Silica (WT56)
Total Solids (mg/kg)
Specific Gravity
Compositional
(mg/kg)
14.5
<0.1
75
<0.2
-------
Table 8. Facility B Scrubber Wastewater Data*
PARAMETER
Scrubber Wastewater
Toxic Metals (mg/kg)
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium-hex
Chromium-total
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
Other Analyses
Total Solids (mg/kg)
Total Dissolved Solids (mg/kg)
Total Suspended Solids (mg/kg)
Total Organic Halide (wtf,)
Total Chlorine (wt$)
Specific Gravity (g/mL)
PH
<0
<0
0
1
241
106
27
<0.2
363
3867
3500
67
0.08
1.59
0.9936
0.7
*Each value represents the average of 3 grab samples.
-231-
-------
CONCLUSIONS
Because of the limited sampling
periods, the data presented in this paper
represents only a snapshot of operations
at two hazardous waste incineration
facilities. It is not known whether this
data is typical of normal operation at
each of these facilities or of
incineration in general.
However, evaluation of the limited
data does indicate that significant levels
of toxic metals are present in waste
solvents and organic process wastes.
Since metals are not destroyed by
incineration, they tend to concentrate in
the treatment residues - ash and scrubber
wastewater. Properly designed and
operated incineration systems should
achieve virtually complete destruction of
organic compounds. This appears to be the
case at Facility B, where no detectable
levels of volatile or semivolatile
organics were found in either the ash or
scrubber wastewater. Facility A's ash did
contain significant concentrations of
organic compounds. This may be indicative
of incomplete combustion due to inadequate
detention time, mixing, or temperature in
the incinerator.
-282-
-------
WASTE CHARACTERIZATION AND THE GENERATION OF
TRANSIENT PUFFS IN A ROTARY KILN INCINERATOR SIMULATOR
William P. Linak, Joseph A. McSorley
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Jost 0. L. Wendt
University of Arizona '
Tucson, AZ 85721
James E. Dunn
University of Arkansas
Fayetteville, AR 72701
ABSTRACT
The batch introduction of waste-filled drums or containers into rotary kiln incin-
erators can lead to transient overcharging conditions, which are denoted as "puffs." This
paper describes results of an in-house investigation at the U.S. EPA into the waste prop-
erties and kiln parameters that determine both the intensity and the magnitude of transi-
ent puffs leaving the kiln. The experimental apparatus utilized was a 73 kW (250,000
Btu/hr) laboratory rotary kiln simulator. Surrogate solid wastes in the form of plastic
rods and surrogate liquid wastes on corncob sorbent in cardboard containers were investi-
gated. Parametric studies were used to determine the extent to which waste and kiln
variables (such as charge mass, charge surface area, charge composition, kiln temperature,
and kiln rotation speed) affected the intensity (peak hydrocarbon emission) and magnitude
(time-integrated hydrocarbon emission) of puffs.
Results demonstrate the relative ease with which failure conditions are achieved,
even at high excess air values and high kiln temperatures. Chemical analysis indicates
that puffs arising from even innocuous surrogate wastes can contain numerous hazardous
compounds even though adequate DREs (>99.99%) are achieved. Increasing kiln temperature
and rotation speed can adversely affect puff intensity, due to increased devolatilization
and liquid evaporation rates. There are large effects of waste composition and, for
solid wastes, waste surface area is a critical variable.
Stoichiometric oxygen requirement is an important variable distinguishing the trans-
ient behavior of different kinds of wastes. Thermogravimetric analyses may be useful in
characterizing the propensity of solids to generate transient puffs, while liquid wastes
may be best characterized by their normal boiling points and latent heats.
INTRODUCTION
Rotary kiln incinerator systems have
been described elsewhere (1,2) and may
comprise one of the most versatile waste
disposal systems available. Available
data (1,3) indicate that these systems
are generally safe and effective. How-
ever, most data on destruction and removal
efficiencies (DREs) of many principal or-
ganic hazardous constituents (POHCs) have
been obtained under steady-state operating
conditions. Little information is avail-
able on potential problems during transi-
ent conditions that occur when drummed
liquids or solid materials are batch fed
into the combustion chamber. Overcharging
conditions can occur if waste parameters
-283-
-------
are not properly matched to the incinera-
tor operating conditions. They can lead
to heavy loadings of products of incom-
plete combustion (PICs) in the afterburner
and gas cleanup systems. For the sake of
brevity, we denote this transient condi-
tion as a "puff."
The overall objectives of the current
research are to gain engineering insight
Into potential problem areas related to
rotary kiln incineration, with a view to
developing predictive methods that corre-
late possible failure modes to waste
characteristics and kiln operating parame-
ters using continuous on-line instrumenta-
tion. The specific problem addressed in
this paper is that of transient puffs
originating from the batch introduction
of either solids or liquids on sorbent in
containers.
The problem is approached through
parametric experimentation on a laboratory-
scale rotary kiln incinerator simulator.
The value 1n utilizing a prototype kiln,
rather than an actual permitted field
device, lies in its ability to allow para-
metric variation, in a controlled fashion,
without loss of the overall complicated
salient features of practical kiln opera-
tion.
It is hypothesized that a waste param-
eter of direct importance in the genera-
tion of transient puffs is the volatility,
or the rate of volatile release, of the
waste. When devolatilization occurs
rapidly, local oxygen concentrations in
the flue gas are almost totally depleted
or displaced by the volatilized waste
species. A transient puff then moves as
a plug through the system. When volatiles
are released slowly, however, they are
transferred into the flue gas stream with-
out totally depleting the local oxygen
levels, allowing more complete oxidation
to occur and puffs to be minimized.
Therefore, the research results presented
here focus primarily on the effects caused
by differences in waste volatility. Spe-
cifically, we compare results from the
batch introduction of surrogate solid
wastes, in the form of plastic rods, to
those resulting from contained surrogate
liquid/sorbent wastes.
LABORATORY ROTARY KILN SIMULATOR
The EPA laboratory-scale rotary kiln
simulator was designed to contain the
salient features of full-scale kilns, but
to remain sufficiently versatile to allow
parametric experimentation by varying one
parameter at a time, or by controlling a
set of parameters independently. A sche-
matic view of the simulator is shown on
Figure 1. Its overall characteristics,
compared to those of full-scale units,
and a detailed description of the hard-
ware specifications have been previously
presented (4). The simulator consists of
five refractory-lined sections, and is
rated at 73 kW (250,000 Btu/hr). The sim-
ulator matches the volumetric heat release
and gas-phase residence times of full-
scale units, but is rated at less than 5
percent of the gross heat input of full-
scale units (5).
Surrogate waste.materials are batch
charged through a sliding gate/ram assem-
bly located on the transition/afterburner
section (Figure 1). Quantification of a
system upset condition (puff) requires the
real-time measurement of system variables.
Peak responses of fixed gas and hydrocar-
bon analyzers indicate extremes. Other
useful variables include time-integrated
responses; in particular, the time-
integrated response of the total volatile
hydrocarbon analyzer. This method permits
determination of emitted mass. By holding
constant sampling temperature, pressure,
and flow rate, a constant volumetric flow
is delivered to the analyzer's flame ioni-
zation detector (FID). Provided the
incinerator flue gas flow rate is constant
and the gases well mixed, the integrated
mass measured is then proportional to the
total mass of volatile hydrocarbons pass-
ing up the stack during a puff. The peak
response, on the other hand, represents
the maximum concentration in a puff. The
FID is very sensitive to hydrocarbon
species but less sensitive to chlorinated
compounds.
This paper is concerned only with
phenomena occurring in the kiln (primary
combustion chamber), so samples withdrawn
represent effluent leaving the kiln and
entering the afterburner (secondary com-
bustion chamber). Throughout this work,
there was neither supplemental heating
nor afterburning downstream of the kiln
exit. Samples were withdrawn at sample
-284-
-------
port 4 (Figures 1 and 2). Linak et al.
(4) have discussed the extractive sampling
and analysis procedures. The results
allow determination of the inlet transient
conditions with which an afterburner must
be capable of dealing, and how these de-
pend on kiln and waste parameters.
TO SLOWER AND STACK
BACKFIRE EXHAUST
• •THERMOCOUPLE
) ^-flAMHOO
CHARGING BASKET
ROTARY LEAF
WRINOSIAI.
ITRANSmON/AFTEflaURNER)
ft 0 1 2 3
Figure 1. EPA rotary kiln incinerator
simulator.
Figure 2 shows the axial time/
temperature profile through the kiln and
the subsequent control temperature tower
(without additional heat addition) under
the low and high firing conditions uti-
lized in the parametric experiments
reported in later sections. Gas phase
kiln temperatures range from 1258 K
(1804 °F) to 1056 K (1440 °F) after 3.6
seconds, and from 1383 K (2030 °F) to
1327 K (1928 °F) after 3.1 seconds under
low and high firing conditions, respec-
tively. These gaseous time/temperature
profiles simulate full-scale units well.
6 8
RESIDENCE TIME.*
Figure 2. Temperature profile vs resi-
dence time, low and high fire
test conditions
PARAMETRIC STUDIES
Statistical Design of Experiments
Experimentation was divided into two
phases: Phase 1 was concerned with well
defined prototype solid plastic wastes;
and Phase 2 dealt with liquids absorbed
on corncob sorbent in cardboard con-
tainers. In both experimental phases,
the general methodology employed was that
of response surface experimentation (6).
This methodology allows one to determine,
on the basis of one experiment involving
a minimal number of trials, an empirical
relationship between the response and the
controlled parameters in the experimental
region.
For both phases of experimentation,
low, intermediate, and high settings of
the variables to be investigated were
determined. For Phase 1, dealing with
prototype solid plastic wastes, these
variables were charge mass, charge sur-
face area, and kiln temperature. For
Phase 2, which was treated separately and
concerned with contained liquid on sorbent
wastes, these variables were charge mass,
kiln temperature, and kiln rotation speed.
Each phase involved four waste com-
positions and consisted of at least 24
trials. This allowed quadratic models to
be formulated as well as full degree of
-285-
-------
freedom tests of significance of all main
effects and all two-factor interactions
(6). Each trial involved between 5 and
15 replicates, depending on the response
variance and the availability of waste
material. Replicate values of the re-
sponse variables, including FID measure-
ments of time-resolved total hydrocarbon
peak area, and peak height, were averaged.
A quadratic response surface model was
fitted to these mean values using weighted
least squares.
The quadratic model included linear
and quadratic dependencies on each inde-
pendent variable, as well as all the
cross-product terms denoting interactions
between pairs of variables. The prepro-
grammed procedures in SAS (Statistical
Analysis System), which are relevant to
response surface methodology, were used
further to interpret the data. Similari-
ties in model coefficients between differ-
ent wastes (within one phase of experimen-
tation) were statistically tested for
evidence that dependencies of the response
on various parameters altered from waste
to waste. Further tests denoted which
linear, quadratic, or interaction depen-
dencies were statistically significant.
Phase 1; Prototype Solid Plastic Wastes
Plastic rods composed of low density
polyethylene (LDPE), high density poly-
ethylene (HOPE), polystyrene (PS), and
polyvinylchloride (PVC) were chosen as
prototype plastic wastes. These rods had
well defined initial shapes and allowed,
within a certain range, for surface area
and mass to be varied independently by
varying rod diameters. The plastics span
a range of molecular structures, molecular
weights, and elemental compositions that
are of intrinsic interest to the inciner-
ation community because they are widely
used and disposed of. In addition to com-
position, the other independent variables
included kiln temperature, charge mass,
and charge surface area. The representa-
tive kiln temperature used as a control
parameter was that measured by a bare
type K thermocouple at the kiln exit,
0.97 m (38 in.) from the front burner
wall. Low and high fire conditions were
used and correspond to the two temperature
distributions in Figure 2. The following
represent the minimum and maximum values
for the parametric test matrix:
(a) Kiln exit temperature, 1105 K (1530°F)
minimum, 1339 K (1950 °F) maximum.
High fire conditions were obtained by
making use of the thermal inertia of
the kiln to heat the flue gas to a
temperature greater than that occur-
ring naturally at a stoichiometric
ratio (SR) of 2.0, as shown in Figure
2.
(b) Charge surface area, 0.050 m2 (78 -
in.2) minimum, 0.095 m2 (147 in.2)
maximum.
(c) Charge mass, 75-100 g (0.17-0.22 Ib)
(minimum), 500-750 g (1.10-1.65 Ib)
(maximum). Charge mass was deter-
mined by the mass of 6.35, 12.7, and
25.4 mm (0.25, 0.5, and 1.0 in.)
diameter rods which produced the
proper charge surface area, and was
dependent on waste density.
The following kiln operating parame-
ters were kept constant for all runs
reported here:
(a) Auxiliary gas flow rate, 5.66 m3/hr
(200 cfh).
(b) Air flow rate, 109.0 m3/hr (3850 cfh)
(leading to constant SR of 2.0 or
excess air of 100 percent).
(c) Rotation speed, 1 rpm.
(d) Burner position, -0.34 m (-13.5 in.)
inside front burner wall.
(e) Kiln pressure, -37.4 Pa (-0.15 in.
H20).
(f) Sample port, at position 4 (See
Figures 1 and 2).
The area under the transient curve
measured by flame ionization detection
(FID) is, under constant sampling condi-
tions, proportional to the total quantity
of volatile hydrocarbon in a puff. Peak
height, on the other hand, represents the
instantaneous volatile hydrocarbon concen-
trations in the puff. The volatile hydro-
carbons are defined as those sampled at
420 K (300 °F) and analyzed by continuous
FID. Figure 3 shows typical traces for
two different plastics at similar test
conditions. Clearly there are significant
quantitative and qualitative variations
in the mechanisms generating transient
puffs from solid wastes of differing com-
positions.
Quadratic models were fitted to the
data, using weighted least squares where
the weights were equal to the number of
-286-
-------
1000
800
600
400
200
I I
(a) LOPE
307.6 g (0.678 Ib)
0.0547 m2 (84.8 in.2)
1360 K( 1988 °F)
50
40
30
20
10
I I
(b) PVC
- 442.2 g (0.975 Ib)
0.0538m2 (83.4 in.2)
1339 K (1950 <>F)
0.5
1.0
1.5
2.0 0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
TIME, min
Figure 3.
Comparison of LDPE and PVC transient hydrocarbon peak shapes
at similar trial conditions.
replicates. These models, obtained using
the preprogrammed procedures in SAS,
should not be considered unique, since
other models might attain fits essentially
indentical to those reported here. Once
determined, these models can be described
pictorially through response surface
plots. These contour plots have been pre-
sented previously (4) and are not included
here.
The final models for peak area ac-
counted for 95.9 percent of the variation
in the reponse caused by changes in the
controlled variable. The data indicate
that, for LDPE, the largest puffs (mass
basis) occur at large surface areas and
high charge masses. There exists a thres-
hold charge mass, above which puffs rapid-
ly appear, and this threshold is not sen-
sitive to kiln temperature. For LDPE,
charge surface area is important and,
within a certain charge mass range, can
also determine the presence and magnitude
of a puff. Increasing kiln temperature
decreases the total volatile hydrocarbon
species contained in a puff. This sig-
nificant result was also true for the
HOPE, PS, and PVC wastes. PS and PVC
wastes, however, showed little effect of
surface area on the integrated traces.
Data for maximum peak heights, represent-
ing the maximum instantaneous intensity
of the puff, were analyzed statistically
using the same procedures. The resulting
quadratic model accounted for 84.3 per-
cent of the variability of the 37 averaged
data points used and is significantly dif-
ferent from the model for peak area. This
demonstrates the importance of distinguish-
ing between total mass of the puff and
the peak transient concentration. The
resulting contour plots for LDPE yield
the important result that increasing the
kiln temperature increases the maximum •
instantaneous concentration of volatile
hydrocarbons emitted, which is contrary
to the school of thought that destruction
efficiencies necessarily increase with
temperature. Charge surface area is very
significant at lower kiln temperature,
but less so at higher kiln temperatures.
PS arid PVC peak heights showed strong
surface area effects, but the role of kiln
temperature was more ambiguous. For the
PVC waste, peak heights were very much
smaller, and peak durations very much
longer than those for the HOPE and LDPE
wastes at similar test conditions (see
Figure 3).
-287-
-------
It 1s likely that the formation of
puffs from solid plastic wastes is gov-
erned by a number of related chemical and
physical processes. The stoichiometric
oxygen requirement of the plastic waste
being incinerated is the most important
parameter distinguishing the behavior of
different waste compositions. High kiln
temperatures enhance rapid devolatiliza-
tion of the solids, thus increasing the
rate of formation of volatile gases to
displace or deplete the excess oxygen
from the main burner. This may explain
the significant effect of kiln temperature
on the maximum instantaneous concentration
of hydrocarbons in the puff (peak height).
It is interesting to note, however, that
Increasing kiln temperature decreased the
total mass of volatile hydrocarbons emitted
(peak area). This points to the benefi-
cial (and not unexpected) effects of high
temperatures to enhance oxidation mecha-
nisms. Clearly, definition of an optimum
condition is not trivial. In practice,
afterburners must be able to handle large
Instantaneous pulses of wastes from the
kiln, and so the peak height results are
of practical significance.
That surface area appears to be an
important parameter, under several condi-
tions, indicates the importance of physi-
cal processes, such as heat transfer to
the solid and devolatilization from the
surface. Decreasing surface area of the
solid usually decreased both the intensity
of the puff and the total volatile hydro-
carbon mass emitted. The exceptions to
this occurred at high kiln temperature
when heat transfer and subsequent devol-
atilization are both more rapid.
Heat transfer is clearly not the only
important mechanism, since there were
vast differences in behavior between waste
types. PVC rods had the slowest rate of
volatile release, for all surface areas,
and therefore always formed the least
intense, but longest, puffs. It is ex-
pected that their effluent could be
handled best by an afterburner. The PE
wastes had the most rapid rate of vola-
tile release, and thus are most likely to
cause difficulties in the downstream
afterburner. This was evident also from
thermogravimetric analyses of the solids.
These "devolatilization profiles" indicate
that, while PVC begins to lose mass at a
lower temperature as compared to both LDPE
and HOPE, the mass loss rate is slower.
Programmed at a temperature ramp of 30
K/min (56 °F/min) in a nitrogen atmos-
phere, the PVC begins to lose mass at
approximately 525 K (486 °F) and is still
losing mass at 1175 K (1656 °F). Both
LDPE and HOPE, under the same conditions,
begin losing mass rapidly at approximately
675 K (756 °F), and essentially all mass
is lost by 750 K (891 °F). Clearly, vol-
atile release rates are significant in
characterizing the propensity of solid
wastes to form transient puffs. The PE
wastes had the most rapid rate of vola-
tile release, and thus are most likely
to cause difficulties in the downstream
afterburner.
Phase 2: Liquid/Sorbent Wastes
In practice, when liquids are incin-
erated in a batch mode, they are usually
held on sorbents in containers or drums;
In the research reported here, the con-
tainers consisted of 0.947 1 (1 qt)
closed, cylindrical cardboard vessels,
with 135 g (0.298 Ib) of shredded agri-
cultural corncob as the sorbent. Proto-
type liquid wastes included toluene, No.
5 fuel oil, carbon tetrachloride, and
methylene chloride. These compounds were
chosen because they spanned a range of
volatility (the fuel oil, for example,
required heating before being absorbed on
the sorbent), chemical structure, stoi-
chiometric oxygen requirement, and chlo-
rine substitution.
The experimental design was identical
to that used for the prototype solid
wastes (Phase 1), except that the variable
denoting the solid plastic surface areas
was replaced by kiln rotation speed. Ro-
tation speed was chosen because, like
surface area for a solid, it directly
influences contacting between waste/
sorbent and the surroundings. The follow-
ing represent the minimum and maximum
values for the parametric test matrix:
(a) Kiln exit temperature, 1105 K (1530°F)
minimum, 1339 K (1950 °F) maximum.
(b) Kiln rotation speed, 0.5 rpm minimum,
2.0 rpm maximum.
(c) Charge mass, 50 g (0.11 Ib) minimum,
200 g (0.44 Ib) maximum, not includ-
ing the container and sorbent tare
weight of 180 g (0.40 Ib).
The other kiln operating parameters were
kept constant as described in Phase 1.
-288-
-------
Reproducibility was much better for
the liquids reported here than for the
solids reported in Phase 1. Figure 4
shows typical FID traces for two different
liquids at similar test conditions, and
again demonstrates the significant quanti-
tative and qualitative variations result-
ing from charge composition. Nonlinearity
of the FID response was observed during
the most intense puffs (greater than
10,000 ppm of THC).
Again quadratic models were fitted
to the data, using weighted least squares,
where the weights were equal to the number
of replicates divided by the sample vari-
ance between replicates. The final re-
sponse models for peak area accounted for
99.4 percent variation in the response
caused by changes in parameters. Quad-
ratic models for peak heights were very
similar to those for peak area, and
accounted for 98.9 percent of the varia-
tion in the response. In all cases,
increasing the kiln rotation speed in-
creased the peak area of the FID hydrocar-
bon trace and the peak height. This was
also true for the two chlorinated liquids.
Furthermore, for toluene and fuel oil,
both peak area (puff magnitude) and peak
height (puff intensity) increased with
increasing temperature. This result dif-
fers slightly from the solid plastic
results, where only puff intensity fol-
lowed this trend, and it conflicts with
the more traditional viewpoint held with-
in the incineration community that high
incineration temperatures are always bene-
ficial. This apparent perverse effect of
kiln temperature was also observed for
methylene chloride, but not for carbon
tetrachloride,.where results were less
definitive.
Contour plots for toluene peak area
and peak height exhibited saddle contours
when concentrations of volatile hydrocar-
bons were very high (in excess of 10,000
ppm)(7). These saddle contours are anom-
alous, and cannot be explained by nonlin-
ear response of the FID. FID response
was found to be monotonic with THC concen-
tration up to 40,000 ppm. A possible
explanation for the presence of saddles
in these plots (7) is that soot formation
is disproportionately accelerated at high
hydrocarbon loadings and that, therefore,
the remaining volatile hydrocarbons meas-
ured by the FID are no longer quantita-
tively representative of total puff size.
§20-
°10-
{•) TOLUENE
200.0 g IQ.441 Ib)
0.5 rpm
1329K|1933°FI
(b) CARBON TETRACHLORI06
200.0 g(0.441 Ib)
i 0.5 rpm ;
I 1323 Kf1922°F|
Figure 4. Comparison of toluene and carbon
tetrachloride transient hydro-
carbon peak shapes at similar
trial conditions.
Saddles did not appear for the less
volatile fuel oil where hydrocarbon load-
ings were significantly lower. Therefore,
we believe that, for the less volatile
wastes, such as the plastics described
.above, the FID peak area and peak height
contour plots can represent the magnitude
and intensity of the puff, provided the
hydrocarbon concentrations are not too
high.
The similarity between contours of
integrated measurements (peak area) and
those of instantaneous values (peak
height) for the liquid wastes (7) suggests
that other integrated measurements, such
as that of particulate matter trapped on
filters during a puff, should also repre-
sent trends in both puff magnitude and
intensity. Therefore, a statistical anal-
ysis was performed on the filter weight
recorded for each test. The ensuing model
(r2 = 0.999) contained similar terms as
the peak area model. Contour plots for
toluene and fuel oil, Figure 5, show con-
clusively that, as kiln temperature is
increased, the total soot loading from
the kiln to the afterburner is increased;
i.e., a much larger puff is formed. This
is significant, and once again points to
potential failure if the kiln is operated
at too high a temperature.
Filter residue mass cpntour plots for
methylene chloride were qualitatively
similar to those shown on Figure 5, but
-289-
-------
(a)
0.5'
1.0
1.5
1.0
1.5
2.0 0.5
KILN SPEED, rpm >
Figure 5. Effect of charge mass, charge surface area, and kiln temperature
on (a) toluene and (b) No. 5 fuel oil particulate mass. (Contour
values denote filter loadings in milligrams).
2.0
indicated only a slight dependence on
kiln temperature. Carbon tetrachloride
contour plots for both filter residue and
peak area differed sharply from those
arising from all other liquids, and did
not lend themselves to physical interpre-
tation. Furthermore, the carbon tetra-
chloride filter residues were not black
and did not resemble soot while those of
the others did. This suggests that, for
wastes containing only carbon and chlo-
rine, neither the magnitude nor the
intensity of a puff can be determined from
measurements of either volatile (FID
measureable) hydrocarbons or integrated
filter residue mass. Therefore, we
explored the value of using CO peak height
as an indication of the presence of puffs,
since this has been investigated by
others (8-15) as a measure of incinerator
performance. Statistical analysis of the
data describing maximum peak values of CO
allowed a quadratic model, similar to
that for FID hydrocarbon peak area, to be
formulated with an r2- of 0.988. The
resulting CO contour plots for this, and
indeed for all other liquids except
toluene, were similar to the FID hydro-
carbon peak height plots for fuel oil.
This suggests that, for fully chlorinated
liquid waste compounds, CO may be a
useful puff indicator. However, when
copious quantities of soot are formed, as
for toluene at high kiln temperatures,
this indicator ceases to be valid and
should not be used alone. It would
appear, therefore, that no single indi-
cator of puffs exists for all types of
wastes, and that it is desirable always
to monitor at least all three indicators
used in this work; namely, FID measurable
hydrocarbons, CO, and integrated filter
residue. For the wastes considered here,
however, it was always possible to choose
one or more of these indicators such that
physically interpretable contour plots
resulted.
-290-
-------
TRANSIENT PUFF COMPOSITION
Phase 1: Prototype Solid Plastic Wastes
The compositions of the transient
puffs from simple polyethylene (PE) and
polyvinyl chloride (PVC) pipe pieces were
determined using VOST and SASS stack
sampling trains (16,17), in conjunction
with GC-MS analysis (18). These screening
tests were used for qualitative identifi-
cation only since it was not within the
scope of this research to perform exhaus-
tive quantitative chemical analyses.
Results are presented for a solvent blank
and for the test conditions:
(a) A baseline in which the exact proce-
dure for charging was followed, but
no charge was entered'(i.e., the gate
was opened and closed).
(b) One test involving PE pipe with nine
charges totalling 2.3 kg.
(c) Two replicate tests (c-1 and c-2),
each involving 50 percent PE and 50
percent PVC .pipe, (by weight), each
test consisting of 10 charges total-
ing 2.7 kg.
(d) One test involving 10 charges of PVC
pipe totalling 2.9 kg.
In all cases, the kiln was at high
fire conditions (1339 K, 1950 °F) and
100 percent excess air, and the charges.
were introduced at a rate of one every 10
minutes. The results are shown on Table
1 and were significant in that they
showed that hazardous constituents are
contained in the puff, and that more
diverse organochlorides were detected when
the waste consisted of PE/P'VC mixtures
but not in tests with PVC alone. This is
consistent with the transient results
described in the preceding sections. PVC
alone is.devolatilized slowly, and forms
small puffs of low intensity and magni-
tude. PE alone, on the other hand, forms
intense puffs of great magnitude, but
with no chlorine content. Therefore, the
greatest likelihood of hazardous chlori- '
nated hydrocarbons being formed occurs
when PE and PVC are mixed. These tests
also indicated the presence of trace quan-
tities of penta- through octachloro
paradibenzo dioxins (PCDDs) and furans
(PCDFs) in several of the samples. These
data are presented in Table 2. GC/MS with
selected ion monitoring was employed on
concentrated samples. No PCDDs or PCDFs
were detected from the baseline test (a),
the PE test (b), or the solvent blank.
Both PE/P.VC tests (c-1 and e-2) indicated
the presence of several of these compounds
but only the second test (c-2) was quanti-
fied. The PVC test (d) indicated the
presence of octa-CDF. Again, these
results are consistant with proposed mech-
anisms regarding puff formation.
Phase 2: Liquid/Sorbent
As in Phase 1, compositions of the
transient puffs from the liquid wastes
were determined using VOST and SASS stack
sampling traias .in conjunction with GC/MS
analysis. The ability of the kiln alone
, to produce PICs and, additionally, its
ability adequately to destroy the POHC
species, were examined. Results are
presented for a solvent blank and for the
following test conditions:
(a) A baseline in which the exact proce-
dure for charging was followed, but
no charge was entered (i.e., the gate
was opened and closed).
(b) A sorbent/container blank (10 charges
totaling 1.8 kg).
(c) Two replicate tests (c-1 and c-2),
each involving 10 charges of toluene
alone totaling 3.8 kg (including 1.8
kg of sorbent and container).
(d) Two replicate tests (d-1 and d-2), of
50 percent toluene and 50 percent
carbon tetrachloride (by weight),
each test consisting of 10 charges
totaling 3.8 kg (including 1.8 kg
of sorbent and container).
(e) One test involving 10 charges of
carbon tetrachloride totaling 3.8
kg (including 1.8 kg of sorbent and
container).
As in Phase 1, the kiln was operated
at high fire conditions (1339 K, 1950°F),
100 percent excess air, and the charges1
were introduced at a rate of 'one every 10
minutes. The VOST results show benzene '
and toluene to be present for all of the
tests including the baseline test (a),
and carbon tetrachloride to be present in
both the toluene/carbon tetrachloride
(d-1 and d-2) and carbon tetrachloride
(e) tests. ORE calculations, however,
indicate adequate destruction of the POHC
species in excess of 99.99 percent.
Particulate loadings, however, determined
by SASS filter weights averaged over the
entire 100 minute test, exceeded the RCRA
regulated limit of 180 mg/dscm (0.08
-291-
-------
Taole 1.
Selected RCKA Appendix VIII compounds identified in test burns charging
.polyethylene and polyvinyl chloride pipe.1
Solvent Basel1ne2 PE PE/PVC3 PE/PVC3 PVC
Blank (a) (b) (c-1) (c-2) (d)
(2-Ethylhexyl Jphthalate
Naphthalene
Dichlorobenzene
1 ,2 ,4-Trl chl orobenzene
Hexachlorobenzene
Pentachlorophenol
Fluoranthene
X XX
X XX
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1 In all cases, the kiln was at high fire conditions (1339 K, 1950 °F),
100 percent excess air (SR=2.0), 1 rpm kiln rotation, and 9 to 10 charges,
averaging 270 g each, were charged one every 10 minutes.
2 Baseline test sampled auxiliary fuel combustion gas products; no waste
material charged.
3 PE/PVC mixed charges were 50 percent each by weight.
Table 2. PCOD and PCDF flue gas concentra-
tions determined from test burns
charging polyethylene and poly-
vinyl chloride pipe.
Average flue gas concentration, ppt (volume).
Solvent Baseline PE
Blank (a) (b)
Mono-CDO
Di-CDD
Trl-COD
Tetr«-CDD
Penta-CDO
Hexa-COD
Hepta-CDD
Octa-COO
Hono-COF
01 -CDF
TM-COF
Tetra-CDF
Penta-COF
Kexa-COF
Hepta-CDF
Octa-COF
H
0
H
E
D
E
T
E
C
T
E
D
N
0
H
E
D
E
T
E
C
T
E
D
H
0
N
E
D
E
T
E
C
T
E
D
PE/PVC
(c-1)
D N
E 0
T T
E
c 0
T U
E A
D N
T
I
F
I
E
D
PE/PVC PVC
(c-Z) (d)
-1
-
-
5.4
3.4
-
18.4
13.4
-
-
-
20.1
-
58.1
62.8
27.2 10.9
1- Hot detected
gr/dscf, corrected to 7 percent 03) (19)
for all samples except the baseline (a),
sorbent/container blank (b), and carbon
tetrachloride (e) experiments. The
toluene (c-1 and c-2) and toluene/carbon
tetrachloride (d-1 and d-2) tests exceeded
the particulate limit by as much as a fac-
tor of nine.
Extracted and concentrated samples
from all parts of the SASS train were
combined. Table 3 presents a selected
list of major peaks identified by GC/MS
analysis. The data show contaminant
phthalate ester and chlorocyclohexanol
species present in all the samples, in-
cluding the baseline (a) and sorbent/
container blank (b). Of interest is the
detection of several polycylic aromatic
hydrocarbon (PAH) compounds (acenaphtha-
lene, anthracene, fluoranthene, and
pyrene) in the toluene samples (c-1 and
c-2), but not in the toluene/carbon
tetrachloride (d-1 and d-2) or carbon
tetrachloride (e) samples. In addition,
low concentrations of tetra- and penta-
chlorobenzene were detected in the
toluene/carbon tetrachloride samples (d-1
and d-2), but not in any of the other
samples. The carbon tetrachloride (e)
sample was found to be free of both the
chlorinated benzene and PAH compounds.
Again these results are consistent with
the transient results presented in the
previous sections of this paper and with
the analytical results from Phase 1.
Samples were also analyzed for PCDDs
and PCDFs. Again, GC/MS with selected
ion monitoring was employed on concentra-
ted samples. Table 4 summarizes the in-
tegrated concentrations of PCDD and PCDF
species identified in these flue gas
samples. No PCDDs or PCDFs were detected
in the solvent blank, baseline (a),
sorbent/container blank (b), or toluene
(c-1 and c-2) samples. Both the toluene/
carbon tetrachloride mix (d-1 and d-2)
and the carbon tetrachloride (e) samples
contain parts per trillion concentrations
of tetra- through octachloro dioxins and
furans. The carbon tetrachloride sample
(e) contained substantially less PCDDs
-292-
-------
Taole 3.
Selected compounas identified in test burns charging toluene --jnd
carbon tetracnloride.l
Solvent Baseline Sorbent/ Toluene Toluene Toluene/ Toluene/ Carbon
Blank (a)3 Container (c-1) (c-2) Carbon Carbon Tetrachloride
Blank Tetrachloride Tetrachloride (e)
(b) (d-l)4 (d-2)4
Chlorocyclohexanol X X X
Phthalate ester XXX
Tetrachl orobenzene^
Pentachl orobenzene^
Naphthalene
Acenaphthalene
Anthracene
Fluoranthene
Pyrene
X
X
X
X
X
X
X
X X
X
X
X
X
X
X
X
X
X X
X X
X
X
1 In all cases, the kiln was at high fire conditions (1339 K, 1950 °F), 100 percent excess air (SR=2.0), 0.5 rpm kiln
rotation, and 10 charges of 380 g (0.84 lb) each (Including 180 g, 0.40 Ib, centalner/sorbent), were charge one every
10 minutes. Trace quantities of varying chlorinated compounds were identifed in all samples Including baseline,
sorbent/container blank, and toluene samples.
2 Present in low concentrations in comparison to others listed, but not identified in any other sample.
3 Baseline test sampled auxiliary fuel combustion gas products; no waste material charged.
4 Toluene/carbon tetrachloride charges were 50 percent each by weight.
and PCDFs than did the toluene/carbon
tetrachloride mix samples (d-1 arid d-2)
in comparison. However, it must be
restated that all these samples represent
effluent from the kiln prior to the
afterburner, and therefore describe what
must be destroyed by the afterburner or
removed downstream.
DISCUSSION
A significant distinguishing property
of both solid and liquid waste, regarding
the generation of puffs, is its stoichio-
metric oxygen requirement. However,
other fundamental mechanisms governing
the generation of puffs for plastic
wastes are different than those for
liquid wastes on sorbents. For the
former, the volatile release rate de-
pended on the rate of pyrolysis of the
plastic, and this could be correlated
with differential thermogravimetric
analyses. For liquid wastes on sorbents,
thennogravimetric analysis showed no
interesting distinguishing features be-
tween wastes and therefore could not be
related to tne contour plots describing
the transients. Rather, the volatile
release rate is most probably controlled
by heat transfer to (and simple evapora-
tion from) the sorbent, and the surface
exposure of the sorbent to the surround-
ings. If this hypothesis is correct,
then the important waste parameters for
liquid wastes would be their normal boil-
ing points and their latent heats'of
vaporization. One might further hypothe-
size that surface exposure of the sorbent
is increased as kiln rotation speed is
increased. This hypothesis is consistent
with all the data presented here, which
indicate that both puff magnitude and in-
tensity increase witn increasing kiln
rotational speed. Work is in progress to
test these hypotheses using phenomenolog-
ical models.
Time resolved measurement of puffs
is not trivial. Volatile hydrocarbons
give a fair indication when levels are
not too hign (e.g., at Tow to moderate
volatile release rates), as for fuel oil
or plastics. Filter residue mass was a
better indicator for toluene, when black
particulate matter appeared to be the
prime constituent of a puff. Neither of
the above was suitable for carbon tetra-
chloride waste, for which CO appeared to
be the better on-line indicator. However,
CO was a poor indicator for toluene.
Therefore, we conclude that no single
on-line measurement is suitable for all
liquid wastes, and that the omission of
any one of the indicators used here might
lead to error. However, siiiultaneous
-293-
-------
Table 4 PCUO and PCDF flue gas concentrations determined from test burns
charging toluene and carbon tetrachlonde.
Average flue gas concentration, ppt (volume)
Kono-CDD
D1-CDD
Trl-CDO
Tetra-CDO
Penta-CDD
Kexa-CDD
Hepta-COD
Octa-CDD
Mono-CDF
D1-CDF
Trl-CDF
Tetra-COF
Penta-CDF
Hexa-COF
Itepta-CDF
Octa-COF
Solvent
Blank
N
0
N
E
D
E
T
E
C
r
E
D
Baseline Sorbent/ Toluene
(a) Container (c-1)
Blank
(b)
H
0
N
E
D
E
T
E
C
T
E
D
N
0
N
E
D
E
T
E
C
T
E
D
N
0
N
E
D
E
T
E
C
T
E
D
Toluene
(c-2)
N
0
N
E
D
E
T
E
C
T
E
D
Toluene/ Toluene/ Carbon
Carbon Carbon Tetrachlorlde
Tetrachlorlde Tetrachlorlde (e)
(d-1) (d-2)
.1
~
1.3
1.3
0.8
-
-
14.1
16.9
7.7
1.2
-
"
1.8
2.5
17
./
0.7
0.2
-
~
8.4
21.5
12.9
2.5
0.2
-
0.9
0.6
.
0.9
0.7
0.7
0.8
1.8
1- Hot detected
measurements of volatile hydrocarbon
concentration, filter residue mass, CO,
C02» 02. and NOX as puff indicators appear
to span the waste conditions encountered
in this study.
Puffs consist of toxic secondary
combustion products (PICs). Chlorinated
PIC compounds are more likely to be
formed when mixtures of dissimilar mater-
ials, such as PE and PVC or toluene and
carbon tetrachlonde, are burned, than
when PVC or carbon tetrachloride are
burned alone. PVC or carbon tetrachloride
form relatively small puffs. Excess oxy-
gen is neither totally depleted nor dis-
placed, and even though large quantities
of chlorine are present, few chlorinated
PICs are formed. PE or toluene alone, on
the other hand (toluene is a known soot
precursor), form intense puffs of great
magnitude. Excess oxygen is much more
extensively depleted or displaced, and
although no chlorine is present to allow
the formation of chlorinated PICs, there
1s a propensity to form large quantities
of soot and associated PAH compounds.
PE/PVC or toluene/carbon tetrachloride
mixtures provide both the necessary
chlorine and pyrolytic conditions to pro-
mote chlorinated PICs. Therefore, carbon
tetrachlonae alone may not be the most
suitable surrogate compound to investigate
the formation of PICs, at least in the
transient mode.
Finally, the thermudynamic stability
of certain chlorinated compounds, includ-
ing TCDD [thermochemistry from Tsang and
Shaub (20)], was investigated for a wide
range of local stoichiometries likely to
be valid in the puff. Multicomponent
equilibrium calculations, involving 33
species, indicated that very fuel-rich
conditions possibly encountered in the
puff could create thermodynamically
stable trace quantities of CHgCle
and COC1 but that predicted TCDD mole
fractions were always less than 10"19.
The levels of TCDD compounds measured in
this work are above equilibrium concentra-
tions for all possible conditions in the
puff; therefore, their formation is
likely to be kinetically, rather than
equilibrium, controlled.
CONCLUSIONS
A laboratory rotary kiln simulator
can yield useful insight into factors
influencing the formation of transient
puffs caused by the batch introduction of
both solid and liquid wastes. The data
suggest that the volatility of the waste,
or rather, the rate of volatile matter
release, is of paramount importance in
determining the occurrence of puffs.
There are therefore large differences
between solid and liquid wastes, and more
volatile substances yield larger puffs
more readily. These transients can be
minimized at lower kiln temperatures,
-294-
-------
which is contrary to nonia'1 practice.
Devolatilization of the waste, however,
was visually observed to depend on many
physical events occurring in the kiln,
including, but not limited to, the way in
which the waste contacted both the kiln
walls and the flue gas from the main
burner.
Transient puffs can be easily gener-
ated in the kiln and contain hazardous
substances even from quite innocuous com-
pounds such as polyethylene and polyvinyl-
chloride. DREs calculated for liquid
wastes were greater than 99.99 percent.
Mixtures of wastes can lead to more di-
verse hazardous substances than wastes
incinerated individually.
The most significant' implication for
practical systems is that the afterburner
plays an essential role in preventing
transient puffs from both solid and
liquid wastes from entering the environ-
ment. Great care must £>e taken in the
afterburner design to ensure that the puff
components are properly oxidized and de- '
stroyed. Future work could be directed
at ensuring that the kiln performs more
efficiently as a thermal oxidizer rather
than merely as a pyrolysis furnace. This
work suggests that this is not to ba
achieved simply by operating at high
excess air, at high temperatures, and
with good contacting. In fact, for the.
latter two variables, the converse may be
true.
REFERENCES
1. H.W. Fabian, P. Reher, M. Schoen,
"How Bayer Incinerates Waste,"
Hydrocarbon Processing, 4, 183-192
(1979).
2. Tanner, R.K., "Incineration of Indus-
trial Waste," Progress in Energy and
ConiDustion Science, 5, 245-251 (1979).
3. T.A. Bonner, C. L. Cornett, 8.0. Desai,
J.M. Fullenkamp, T.W. Hughes, M.L.
Johnson, E.D. Kennedy, R.J. McCormick,
J.A. Peters, D.L. Zanders, "Engineer-
ing Handbook for Hazardous Waste
Incineration," Monsanto Research
Corp., EPA-SW-889 (NTIS PB 81-248
163), U.S. EPA, Industrial Environmen-
tal Research Laboratory, Cincinnati,
OH (1981).
4. W.P. Liriak, J.D. Kilgroe, J.A.
McSorley, J.O.L. Wendt, J.E. Dunn,
"On the Occurrence of Transient Puffs
in a Rotary Kiln Incinerator Simu-
lator: I. Prototype Solid Plastic
wastes," J. Air Pollut. Control Assoc.
37(1):54 (1987).
5. I. Frankel, N. Sanders, 6. Vogel,
"Survey of the Incinerator Manu-
facturing Industry," Chemical Engi-
neering Progress, 3, 44-55 (1983).
6. C.R. Hicks, "Fundamental Concepts in
the Design of Experiments," 2nd
Edition, Holt, Rinehart and Winston
Inc., New York, NY (1973).
7. W.P. Linak, J.A. McSorley, J.O.L.
Wendt, J.E. Dunn, "On the Occurrence
of Transient Puffs in a Rotary Kiln
Incinerator Simulator: II. Contained
Liquid Wastes on Sorbent," Submitted
to J. Air Pollut. Control Assoc.
(1987).
8. J.C. Kramlich, M.P. Heap, W.R. Seeker,
S.S. Samuelsen, "Flame-mode Destruc-
tion of Hazardous Waste Compounds,"
20th Symposium (International) on
Combustion, The Combustion Institute
(1984).
9. W.R. Seeker, J.C. Kramlich, M.P. Heap,
"Laboratory-scale Flame Mode Study Of
Hazardous Waste Incineration." 9th
Annual Research Symposium-Incineration
And Treatment of Hazardous Waste,
EPA-600/9-84-015 (NTIS PB84-234525),
U.S. EPA, Industrial Environmental
Research Laboratory, Cincinnati, OH
(May 1984).
10. S.L. Daniels, D.R. Martin, R.A.
Johnson, A.D. Potoff, J.A. Jackson,
R.H. Locke, "Experience In Continuous
Monitoring Of A Large Rotary Kiln
Incinerator for CO, C02 and 02," 79th
Annual Meeting - Air Pollution Control
Association (June 1985).
11. S.L. Daniels, R.A. Johnson, J.D.
Wilson, "Significance Of Major Gaseous
Species In Combustion And Destruction
Of Hazardous Waste Constituents,"
1985 Annual Meeting - American Instir
tute of Chemical Enginers (November
1985).
-295-
-------
12. R.K. LaFond, J.C. Kramlich, W.R.
Seeker, G.S. Samuelsen, "Evaluation
Of Continuous Performance Monitoring
Techniques For Hazardous Waste Incin-
erators," J. Air Pollut. Control
Assoc. 35(6):653 (1985).
13. L.J. Staley, "Carbon Monoxide and ORE:
How Well Do They Correlate?" llth
Annual Research Symposium - Incinera-
tion And Treatment Of Hazardous
Wastes, EPA-600/9-85-028 (NTIS PB86-
199403), U.S. EPA Hazardous Waste
Engineering Research Laboratory, Cin-
cinnati, OH (April 1985).
14. V.A. Cundy, J.S. Morse, D.W. Senser,
"Practical Incinerator Implications
From A Fundamental Flat Flame Study
Of Dichloromethane Combustion," J.
Air Pollut. Control Assoc. 36(7):824
(1986).
16. G.L. Huffman, L.O. Staley, "The Forma-
tion Of Products Of Incomplete Combus-
tion In Research Combustors," 12th
Annual Research Symposium - Land Dis-
posal, Remedial Action, Incineration
And Treatment Of Hazardous Waste,
EPA-600/9-86-022 (NTIS PB87-119491),
U.S. tPA Hazardous Waste Engineering
Researcn Laboratory, Cincinnati, OH
(April 198b).
16. L.D. Johnson, R.6. Merrill, "Stack
Sampling for Organic Emissions,"
Toxicological and Environmental Chem-
istry, 6, 109-126 (1983).
17. L.U. Johnson, "Detecting Waste Com-
oustion Emissions," Environmental
Science and Technology, 20:3, 223-227
(1986).
18. ft.H. James, R.E. Adams, J.M. Finkel,
H.C. Miller, L.O. Johnson, "Evalua-
tion of Analytical Methods for the
Determination of POHC in Combustion
Products," J. Air Pollut. Control
Assoc., 35(9):959 (1985).
19. 1984 Hazardous and Solid Waste Amend-
ments to the 1976 Resource Conservation
and Recovery Act.
20. W. Tsang, W. Shaub, "Environmental
Consequences Arising From The Combus-
tion of Municipal Solid Waste," Pro-
ceedings of Resource Recovery From
Solid Wastes, S. Sengupta, K.V. Wong,
eds., Miami Beach, FL (May 1982),
Pergamon Press, New York, NY.
-296-
-------
ON-LINE MONITORIMG OF ORGANIC EMISSIONS WITH A MOBILE LABORATORY
Sharon L. No!en
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Jeffrey V. Ryan and Richard Bridge, Jr.
Acurex Corporation
P. 0. Box 13109
Research Triangle Park, NC 27709
ABSTRACT
EPA's Hazardous Air Pollutants Mobile Laboratory (HAPML) was designed as an inte-
1 ana Packa96 f°r "*1 -time monitoring of comEion sSurcel
" and i
an
*»t,u-r recently Participated in a total mass emissions test at a full scale
KS7rkll-n ir*inerator' The complete field test was conducted by EPA's Hazardous
Waste Engineering Research Laboratory under the direction of Robert C. Thirnau and will
be reported separately. The HAPML collected continuous emission monitor
™™l raton
normal operation.
/Pfn °PSet3 2hich simulated conditions which might occur
GC/FID was used for on-line analysis of light hydrocarbons
train (VOST) was used ^ collect samplls for ISalysis by
compounds in the stack 9as- co-
?he dfta C011ect*l during the field test and other capabil-
nclude usina MS 1?^ Delude exploring those other capabilities whichP
include using the MS as a single ion monitor and testing the HC1 monitor.
INTRODUCTION
Because of the interest in determin-
ining the performance of an incinerator
on a real-time basis and the need to
characterize the emissions from a variety
of incinerators, EPA initiated a project
to build a mobile laboratory for continuous
monitoring of emissions and operating
parameters of hazardous waste incinerators.
The objective of this project was to
provide an integrated sampling and analy-
tical package for evaluation and develop-
ment of continuous and semicontinuous
monitors for hazardous waste incinerators.
The result of the project is the Hazardous
Air Pollutants Mobile Laboratory (HAPML),
designed to be easily transported for
use by research projects at a variety of
incinerators.
This paper willdiscuss the develop-
ment of the HAPML, its capabilities, data
obtained while monitoring a hazardous
waste incinerator, and future plans.
DEVELOPMENT OF THE HAZARDOUS AIR POLLUTANTS
MOBILE LABORATORY
The HAPML was constructed as an
-297-
-------
Integrated sampling and analytical package
for use in hazardous waste incineration
research. The HAPML utilizes extractive
sampling technology, and all components
are housed in a standard 8 m long,
self-propelled van.
The HAPML contains continuous moni-
tors for real-time analysis of CO, C02,
NOX, 02, S02, and HC1. Organic analytical
instrumentation includes a gas chromato-
graph (GO equipped with a flame ioniza-
tion detector (FID), and a mass spectro-
meter (MS). The layout inside the van is
shown in Figure 1.
The sampling system delivers a
portion of the gaseous effluent in the
stack to the HAPML while maintaining
sample integrity. Major considerations
in on-line analysis include the interfer-
ences caused by particulates and moisture
and minimizing reactions and interactions
in the sample line.
All particulates must be removed
from the sample stream because of problems
they may cause in the analytical instru-
mentation: plugging of the lines, damage
to the optical system, or light scatter
resulting in erroneous data. The HAPML
uses a particulate removal system located
3 m from the stack sampling port. The
particulate removal system is housed in a
heated aluminum filter box and consists
of a high surface area, low pressure
drop, spun glass-wool filter. Valves
located in this box allow calibration
gases to be sent from the van, through
the sampling system, back to the van,
and into the appropriate analytical instru-
ment for system checks.
Water, a natural combustion product,
may condense in the analyzers or can
result in optical interferences. The
problems may be handled in different ways
depending on the species of interest.
The stack gas emissions flowing to the
CO, C02, 02, and NOX monitors are cooled
and dried using a condenser and a
Perma-Pure dryer. N02 is soluble in
water. However, since NOX is typically
about 95% NO, the loss is negligible.
Since S02 and HC1 are also soluble in
water, these gases would be removed along
with the water. Therefore, these problems
must be resolved with a different proc-
edure. A sample conditioner is used to
dilute these gas streams with dry, clean
air to minimize the interference effects
of water vapor.
Considerations in temperature and
material are required to ensure that reac-
tions in the sample line do not occur. A
Teflon line, heated to approximately 150°C,
is used to transport the sample from the
stack into the HAPML. This limits the
on-line organic analysis to those compounds
with boiling points less than the tempera-
ture of the sample line. Many species of
interest lie outside this range and may not
be monitored in this fashion. Care must be
taken to eliminate any cool spots in the
line, such as at fittings. If any chlorine
is in the system, stainless steel connec-
tions can react with the sample and must be
removed. Teflon fittings or stainless
steel fittings lined with Teflon, which
eliminates contact between the stainless
steel and the sample, must be used.
The organic analytical system was
designed to provide information in a variety
of ways. The sampling system may be used
to deliver sample gas: directly to the
GC/MS or the GC/FID for semicontinuous
monitoring, to the MS for on-line analysis
of selected organic compounds, or to the
FID for total hydrocarbon (THC) concentra-
tion determination. The HAPML may also be
used as an on-site laboratory for volatile
organic analysis. It is equipped with a
Volatile Organic Sampling Train (VOST) for
collection and concentration of the gas
effluent if the compounds of interest are
present below the detection limits of the
analytical instrumentation (1). These sam-
ples may be analyzed on the GC/MS or the
GC/FID using a NuTech purge and trap system
with a clamshell oven for thermal desorption.
The purge and trap unit is equipped with a
Tekmar sparger tube assembly. A quick
analysis may be invaluable in determining
the amount of stack gas to be concentrated
to provide the proper amount of sample for
more extensive analytical procedures.
Also, orgam'cs collected on a sorbent
material may be suspected of reacting,
partially desorbing, or becoming contam-
inated if they are not stored properly or
if they are not analyzed quickly enough.
Therefore, analyzing the samples on-site
improves the reliability of the results.
A two column option has recently been
added to the GC to allow simultaneous
-298-
-------
operation of the GC/MS and the GC/FID.
VOST samples may be analyzed on the GC/MS
while the GC/FID is used for semicontin-
uous monitoring of selected compounds or
vice versa. The only limitation is the
temperature of the oven. It may be
operated isothermally or, if a temperature
program is used, the runs for the GC/FID
and the GC/MS must be started simultan-
eously.
Standard gases traceable to the
National Bureau of Standards (NBS) are
used to calibrate the continuous emission
monitors (CEMs) for quantitative analysis.
Onboard storage is available for gas
cylinders. A zero calibration (typically
nitrogen gas) and standards in two concen-
trations are used each day of sampling to
calibrate the instruments. The zero and
the standard closer to the concentration
seen while sampling are checked during
the middle and at the end of the sampling
period. If analyzed concentrations of
these gases fall outside specified limits,
the instruments are recalibrated. The
values from the calibration checks are
used to determine the precision and
accuracy of the instruments.
The organic system is calibrated
similarly, although the standards vary.
When VOST analyses are being done, organic
liquids are injected onto VOST cartridges,
desorbed, and analyzed according to the
VOST protocol. Standard gases for common
organic classes are sometimes used, as
well as liquids injected directly into
the GC. The HAPML is also equipped with
a permeation chamber for calibration of
the organic system and the HC1 monitor.
Most analysis done with the organic
system is qualitative and semi-quanti-
tative. Precision and accuracy of the
instruments are determined similarly to
the CEMs.
All MS and CEM data are acquired by
two onboard computers (a DEC POP 11 and a
Compaq Deskpro 286). FID data are ac-
quired using an HP 3393A integrator.
Backup of data collected is provided by
strip chart recorders for the CEMs and a
tape backup in the Compaq computer.
TOTAL MASS EMISSIONS (TME) TEST
Field tests conducted by the
Hazardous Waste Engineering Research
Laboratory (HWERL) indicate that most
facilities achieve high DREs (2). How-
ever, both the studies and the EPA's
Science Advisory Board's later review
raised additional questions about the
performance of hazardous waste inciner-
ators. A major concern is the unident-
ified emissions, both compounds referred
to in Appendix VIII of the Solid Waste
Methods 846 (SW846) as well as those not
listed that are formed in the combustion
process. The concerns expressed are not
limited to simply stack emissions, but
all other posssible effluents. These
include organics, trace metals, and other
chemicals associated with incinerator
ash, spent water, and particulates.
Because of this concern, HWERL initiated
a project to quantify the total mass
emissions (TME) from a commercial incin-
erator. HWERL had complete responsibil-
ity for the TME test, under the direction
of Robert C. Thurnau. Services of the
HAPML were requested to provide CEM and
on-line GC/FID data. The opportunity
also allowed for further evaluation of
the capabilities of the HAPML in field
conditions.
The TME project was conducted at a
full scale, industrial rotary kiln incin-
erator. Two sets of triplicate 2 hour
tests were scheduled. One set was per-
formed under steady state conditions
similar to those used for a trial burn.
The second set of tests was conducted
under transient conditions which included
variations in the process similar to
those expected during routine incinerator
operations.
The process tested is a multiple
stream hazardous waste incinerator with a
horizontal rotary kiln, secondary combus-
tion chamber, and wet scrubber. Waste
feed streams consist of the solid waste
to the kiln and a liquid mixture of
organic and aqueous waste to the second^
ary chamber and the kiln.
During this study, three types of
solid waste were fed to the kiln: alter-
nate drums containing hydroxypropylmethyl
cellulose, polyethylene wax or powder,
and chlorinated pyridine tar were fed
into the kiln at a rate of 20 bbl/hr
(3200 1/hr). The waste drums began
-299-
-------
feeding Into the kiln about 30 min prior
to the beginning of each test period to
compensate for the 30 to 60 min kiln
residence time.
A single liquid organic tank was
prepared with a 5% spike of carbon tetra-
chloride. The tank was continuously
agitated to preserve homogeneity. The
plant fired other wastes until the samp-
ling crews were ready and then switched
to the spiked waste. Sampling took place
during this period. During the steady
state tests, the liquid waste was fed
into the kiln* and the secondary chamber
at about 2 gpm (7.6 1pm). This feed rate
was increased by fully opening the control
valve for 7 sec approximately every 30
rain to simulate transient conditions.
HAPHL THE RESULTS
CEH data were acquired every 5 sec
and recorded in 30 sec averages. Table 1
lists the average concentrations for the
six tests for QZ, C02, and CO. Tests 1
through 3 are steady state tests, while 4
through 6 are transient tests. The
triplicate steady state runs show very
good reproducibility. The QZ varied from
8.1 to 8.9%, the COz from 6.3 to 6.4%,
and the CO from 0 to 1 ppm.
Although an effort was made to
achieve reproducibility for the transient
tests as well, the results show a fairly
wide range of values resulting from the
three tests. In all cases, the peak
shape for the CO concentration resulting
from the induced upset and corresponding
increase in CO concentration was similar.
The peak maximum occurred about 1 min
after the initial increase in CO concen-
tration and did not decrease to the
normal level until about 3 min after
first rising above the standard concen-
tration. The minimum level of 02 and the
maximum level of COe occurred simultan-
eously during the transient condition.
The C0£ and QZ appeared consistently as
mirror images of the other. There is
little difference between the two types
of tests in the 02 and C02, and the
transient conditions are not obvious.
The relationship between the time of
occurrence for the extremes of the 02 and
C02 concentrations and the maximum in the
CO concentration varied.
The GC/FID was used for on-line
monitoring of low molecular weight hydro-
carbons. Only methane and ethylene were
detected. Table 2 gives the results.
Averages of all readings for the steady
state tests are given. Two averages are
given for each of the transient tests.
The data which occurred between the
transient conditions are given under
steady state, and the data acquired
during the transient conditions (when CO
maxima were apparent) are listed under
transient. Ethylene was detected only
during the transient tests and, except in
one instance, occurred simultaneously
with the transient condition.
Test 4 shows an obvious difference
between the methane concentrations occurr-
ing between and at transient conditions,
as expected. Tests 5 and 6 do not exhibit
this behavior. A Mann-Whitney Two-Sample
Test was performed on the methane data to
determine if the two sets of numbers could
have occurred purely by chance; i.e.,
that there was no real difference between
the two. A Kendall's Rank Correlation
was performed to determine if there was a
correlation between the methane data and
the CO level. Table 3 summarizes the
results. The larger numbers indicate a
greater probability that any difference
between the methane values or any corre-
lation with CO is purely chance. The low
numbers for Test 4 indicate that there is
a real difference between the two sets of
methane data in that test and that there
is a good correlation between the methane
and CO concentrations. The unexpected
results of Tests 5 and 6 could possibly
be explained by some other upset in the
system that damped out the effect of the
induced upset.
VOST samples were taken to determine
if any volatile organics were present
below the detection limit of the instru-
ment. The major peak detected could not
be identified. The next two peaks in
order of intensity were benzene and
toluene. Benzene averaged 13 ppb for the
steady state tests. Figure 2 compares
the benzene concentration of the transient
tests, after subtracting the background
or steady state level, with the peak
maximum of the CO concentration which
occurred during the VOST run. Although
data are limited, Figure 2 suggests that,
while CO may not be a good indicator of
ORE (3), the CO level may be related to
-300-
-------
the concentrations of at least some PICs.
Each data point represents only one VOST
analysis. No change was seen in the
toluene concentration between the steady
state and transient tests. It remained
at 2 or 3 ppb for all tests.
CONCLUSIONS
The results of the field test indi-
cate that the HAPML can be successfully
operated in the field and produce quality
data. The HAPML was audited during the
field trip discussed in this paper and
was found to be acceptable with minor
recommendations. All recommendations
have since been implemented.
Several problems in the MS have
since been identified, and steps have
been taken to correct them. These
include an update of all software, im-
provement of the library search capabili-
ties, and making the single ion monitoring
capability operational. It is hopeful
that the MS, in the single ion monitoring
mode, can be used as a continuous monitor
for PICs. There is, without question, a
need for real-time measurement capability
for hazardous waste incinerators. The
HAPML provides a means of testing a
number of incinerators and can be used to
characterize the volatile organics and
permanent gas emissions.
REFERENCES
1. Hansen, E. M. "Protocol for the
Collection and Analysis of
Volatile POHCs Using VOST,"
EPA-600/8-84-007, NTIS PB84-
170042, March 1984.
2. Trenholm, A. R. and C. C. Lee.
"Analysis of PIC and Total Mass
Emissions from an Incinerator."
In Proceedings of the Twelfth
Annual Research Symposium on
Incineration and Treatment of
Hazardous Waste, April 21-23,
1986, pp. 376-381. EPA-600/
9-86-022, NTIS PB87-119491, July
1986.
3. Staley, L. "Carbon Monoxide
and ORE: How Well Do They
Correlate?" In Proceedings of the
Eleventh Annual Research
Symposium on Incineration and
Treatment of Hazardous Waste,
April 29 - May 1, 1985, pp. 23-25.
EPA-600/9-85-028, NTIS PB 86-199403,
September 1985.
-301-
-------
TABLE 1. CEM RESULTS
Test No.
1
2
3
4
5
6
02
(*)
8.9
8.1
8.6
8.2
8.7
10.3
8!
6.3
6.4
6.3
6.2
6.3
5.2
CO
(ppm)
1.0
0.0
0.1
9.0
11.7
14.3
TABLE 2. RESULTS OF ON-LINE 6C/FID MONITORING
Test No.
1
2
3
4
5
6
Steady State
Methane tthylene
(ppm) (ppm)
1 .7 ND
1.1 ND
1.0 ND
1.9 0.8
93.2 ND
48.3 ND
Transient
Methane
(ppm)
NA
NA
NA
11.0
92.9
52.6
ttny i ene
(ppm)
NA
NA
NA
1.7
1.3
0.3
NO = Not detected
NA = Not applicable
TABLE 3. SUMMARY OF STATISTICAL ANALYSIS
Test No.
Probability
4
5
6
Mann-Whitney
0.036
0.421
0.238
Kendall
0.005
0.364
0.168
-302-
-------
Cylinder
Storage
Bench
Instrument Racks
Storage
Bench
Computer
Console
1
Inc
2
rganic
3
Anal)
4
rsis
5
^^_ Organic
Analysis
Door
Air and Vacuun
Controls
Operator
Compartment
Figure 1. HAPML floor plan.
4A
JQ
O.
0)
0)
0)
m
40-
36-
32-
28-
24-
20-
16-
12
100
200
500
600
300 400
CO (ppm)
Figure 2. Comparison of benzene concentration and
maximum CO concentration.
-303-
-------
TOTAL MASS EMISSIONS FROM A HAZARDOUS WASTE INCINERATOR
Andrew R. Trenholm
MIDWEST RESEARCH INSTITUTE
Kansas City, Missouri
Robert Thurnau
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio
ABSTRACT
Past studies of hazardous waste incinerators by the Hazardous Waste Engineering Re-
search Laboratory have primarily examined the performance of combustion systems relative
to the destruction and removal efficiency (ORE) for Resource Conservation and Recovery
Act (RCRA) Appendix VIII compounds in the waste feed. These earlier studies demonstrated
that In general most facilities performed quite well relative to the ORE However, sub-
sequent review by the Environmental Protection Agency's (EPA) Science Advisory Board
raised questions about additional Appendix VIII or non-Appendix VIII constituents that
were not identified in the earlier tests and might be emitted from hazardous waste com-
bustion. This paper presents results of a characterization of incinerator effluents to
the extent that the emitted compounds can be identified and quantified Measurements
were made of both Appendix VIII and non-Appendix VIII compounds in all effluents (stack,
ash water, etc.) from a full scale incinerator. A broad array of sampling and analysis
techniques were used. Sampling methods included Modified Method 5, volatile organic sam-
pling train (VOST), and specific techniques for compounds such as formaldehyde. Analy-
sis techniques included gas chromatography (GC) and gas chromatography/mass spectrometry
(GC/HS). Continuous measurements were also made for a variety of compounds including
total hydrocarbons by flame ionization detector (FID).
INTRODUCTION
A common theme in literature and re-
ligion dealt with the idea that the tri-
als and tribulations experienced in life
are a form of punishment resulting from
our ancestor's sins. The problems of
toxic and hazardous wastes can be com-
pared with this concept in that we are
now faced with the cleanup of hundreds of
old disposal sites that now threaten the
health and well-being of thousands of
citizens. Coupled with the cleanup ef-
fort is the idea that toxic and hazardous
wastes generated today will now be dis-
posed of in an environmentally safe
manner so as not to be a problem to fu-
ture generations. As a public expression
of commitment to cleaning up existing
hazardous waste sites, the Comprehensive
Environmental Response Compensation and
Liability Act (CERCLA) legislation and
its reauthorization Superfund Amendments
Reauthorization Act (SARA) were enacted.
To handle the present day problems of
toxic and hazardous waste disposal, the
RCRA was enacted in 1976 and amended in
1984 by the Hazardous and Solid Waste
Amendments (HSWA). Commensurate with
these statutes, the EPA regards incinera-
tion as one of the principal technology
candidates for the ultimate safe disposal
-304-
-------
of wastes and promulgated the following
standards in the Federal Register. Vol-
ume 46, No. 15, on January 23, 1981.
1. An incinerator must achieve a ORE of
99.99 percent for each principal or-
ganic hazardous constituent (POHC)
designated for each waste feed.
2. An incinerator burning hazardous
waste must not emit more than
1.8 kilogram/hour (kg/hr) of hydro-
gen chloride (HC1) or must remove
99 percent of the hydrogen chloride
from the exhaust gas.
3. An incinerator burning hazardous
waste must not emit particulate mat-
ter exceeding 180 milligrams per dry
standard cubic meter (mg/dscm).
The above standards only address the
POHC residues at the stack and fail to
address other possible effluents such as
products of incomplete combustion (PICs)
associated with stack gases, and POHC
residues, trace metals, and other chemi-
cals associated with incinerator ash,
spent water, and particulates. Because
these effluents may be equally or more
hazardous than POHCs themselves, research
is needed to qualitatively and quantita-
tively study the characteristics of all
possible effluents and to provide engi-
neering data for regulatory support.
To meet this need, Midwest Research
Institute (MRI) was contracted by the
Hazardous Waste Engineering Research Lab-
oratory (HWERL), EPA, Cincinnati, Ohio,
to conduct an inventory of a hazardous
waste incinerator in which the total mass
emissions (TME) were measured and quanti-
tated to the best extent possible. This
paper presents preliminary results from
this study.
APPROACH
This study addressed two primary ob-
jectives. First, a wide array of sam-
pling and analysis (S&A) techniques were
used to identify and quantify constitu-
ents in all waste streams to the extent
possible. Second, the emissions were
measured under two operating conditions,
steady state and with a transient combus-
tion upset. The steady state operation
was similar to conditions during a trial
burn and the transient upsets were in-
tended to simulate conditions that may
occur at times during normal operation.
The test site for this study was se-
lected after a number of sites had been
surveyed. It was desirable to select a
site that would be most representative of
hazardous waste incinerators in the United
States. Though it is impossible to be
completely representative at any one site,
the Dow Chemical Company incinerator in
Plaquemine, Louisiana, was chosen because
it had many of the equipment components
common to hazardous waste incinerators
and burned a wide variety of types of
wastes. The incinerator consisted of a
rotary kiln with liquid waste injection
and drummed solid feed, secondary combus-
tion chamber, quench, and particulate and
HC1 control devices. The wastes fed to
the incinerator during the tests included
solids and liquids as follows:
o Solids - Substituted cellulose
- Polyethylene wax
- Chlorinated pyridine tar
o Organic liquids - Isopar (2,2,4-tri-
methylpentane)
- Carbon tetrachlo-
ride
o Aqueous liquids - Runoff from diked
area
Each type of solid waste was drummed
separately and the drums were automati-
cally weighed and fed through a ram
feeder to the kiln every 4 minutes (min).
Drums of the different types of waste
were fed alternately. A uniform supply
of the liquid organic waste sufficient
for about 100 hours (hr) of operation was
accumulated in a tank and spiked with
about 10 percent carbon tetrachloride be-
fore the test. Liquid organic waste was
fed continuously to both the kiln and
secondary chamber, and the aqueous liquid
waste was fed only to the kiln. Natural
gas was fired in both the kiln and secon-
dary chamber as needed to maintain tem-
peratures. Combustion parameters (tem-
perature, airflows, etc.) and all feed-
rates are controlled by computer from a
central control room.
Essentially all input and effluent
streams were sampled and analyzed using,
-305-
-------
a wide variety of techniques. Table 1
shows a summary of the sampling and anal-
ysis parameters and methods used. EPA's
mobile monitoring van assisted MRI with
the sampling and Dow Chemical conducted
parallel sampling and analysis for many
of the measurements (VOST, waste feed,
scrubber water outlet).
A sufficient number of the solid
waste drums were set aside, numbered, and
sampled during the week prior to the
test. An equal number of drums contain-
ing substituted cello!use, polyethylene
wax, and chlorinated pyridine tars were
used. The solid waste samples were ana-
lyzed for volatile and semivolatile
POHCs, chlorides, ash, and heating value.
Samples of the liquid organic wastes were
taken from a tap at 15-min intervals, as
were the aqueous wastes. Both these sam-
ple types were analyzed for volatile and
seraivolatile POHCs, chlorides, ash, and
heating value. Viscosity of the liquid
organic waste was also measured. The
scrubber waters were sampled from a tap
at approximately 30-min intervals, and
analyzed for POHCs and other major con-
stituents. The bottom ash hopper was
emptied at the start of each run and a
sample was collected after each test run.
The ash samples were analyzed for semi-
volatile compounds.
The stack sampling included two
Modified Method 5 (MM5) trains, a Vola-
tile Organic Sampling train, a midget im-
pinger train for aldehydes, an ORSAT sam-
pler, continuous gas analyzers and three
GCs. One MM5 train was analyzed for par-
ti culates and anions and the other was
analyzed for semivolatile organics. The
VOST samples were analyzed for volatile
organic constituents. The ORSAT sample
was analyzed for 02 and C02. The con-
tinuous analyzers included 02, CO, C02,
NO and THC. The grab sampling frequency
for the GCs depended on the column being
used and the actual run conditions. The
G! and C3 hydrocarbons were detected and
quantified using a Porapak QS column and
FID. The aromatic compounds were de-
tected and quantified using a Megabore
624 column and a photoionization detector
(PID). The halogenated compounds were
detected and quantified using a Carbopack
624 column and a Hall detector. The
GC techniques used were designed for
research measurements and provided
intermittent, instantaneous values at se-
lected times.
DISCUSSION OF RESULTS
Process Data and Waste Characteristics
Key process data for each run are
presented in Table 2. The process data
shown consists of kiln and secondary com-
bustion chamber temperatures, percent 02,
stack flow rate, natural gas heat input,
and total heat input. The total heat in-
put is the sum of heat input from the
natural gas and waste feed. Ranges for
selected process data are shown in paren-
theses.
Runs 1 through 3 were conducted un-
der steady state operating conditions.
However, runs 4 through 6 were altered
so that the incinerator operated with a
transient combustion upset. Several
techniques were attempted to accomplish
this transient condition by causing a
sudden increase in the feed of volatile
waste to the incinerator. Attempts with
waste fed to the rotary kiln, spiking
10 gallons (gal.) of volatile hydrocar-
bons in drums of solid waste or suddenly
increasing the liquid waste feed, failed
to produce a transient condition. The CO
level did not change. A sudden increase
in liquid organic waste feed to the sec-
ondary chamber did produce sharp in-
creases in CO and THC levels, and that
technique was chosen to cause the transi-
ent condition.
The liquid organic waste feedrate to
the secondary combustion chamber was sud-
denly increased from a normal 2 gallons
per minute (gal/min) to 6 gal/min for ap-
proximately 7 seconds (sec) and then re-
turned to the original setting. This
fluctuation in the liquid organic waste
feedrate occurred 15 min into the run and
approximately every 30 min throughout the
run. A review of the data in Table 2
shows that the transient conditions did
not cause any major change in the process
parameters shown on the table.
Table 3 presents the average waste
characteristics for the steady state and
transient test runs. The characteristics
compared are feedrates, heating values,
percent chlorine, and percent ash. The
waste characteristics did not change
-306-
-------
significantly between the two sets of
test runs.
Carbon Monoxide (CO) and Total Hydro-
carbon (THC) Emissions
One way to characterize the opera-
tion .of the incinerator under steady
state and transient conditions is to com-
pare CO and THC levels. Table 4 presents
average CO levels and ranges of CO for
each run. CO levels for the'steady state
conditions (runs 1 through 3) averaged
near zero while the average for the tran-
sient conditions (runs 4 through 6)
ranged from 10 to 15 parts per million
(ppm). During transient conditions, a CO
spike was registered approximately every
30 min with the levels increasing rapidly
to about 700 ppm (1-min average) then de-
creasing quickly.
Table 4 also contains the average
and range of THC values for each run.
The THC values are calculated as ppm of
methane. MRI's runs 1 through 3 showed a
steady THC level with averages of 6 to
8 ppm. Dow's average THC level for
runs 1 through, 3 was less than the detec-
tion limit (1 ppm). MRI run 4 had a
slightly elevated THC level of 9 ppm arid
peaks to 160 ppm during the transient up-
set. MRI runs 5 and 6 had significantly
higher THC levels. Run 5 produced an
average THC of 150 ppm with peaks up to
220 ppm, while run 6 produced an average
of 110 ppm and peaks to 190 ppm. Dow's
THC analyzer malfunctioned during run 4.
Dow1 s run 5 produced an average THC of
99 ppm with peaks up to 150 ppm. Run 6
showed an average THC of 60 ppm with
peaks up to 63 ppm. These high THC lev-
els were largely accounted for by meth-
ane. The difference between run 4 and
ruris 5 and 6 is under investigation.
Volatile Compound Emissions
Tables 5 and 6 present the stack
concentrations of the volatile organic
compounds found during the steady state
and transient operating conditions, re-
spectively. Results from both MRI sam-
pling and parallel sampling by Dow Chemi-
cal are shown. MRI data for run 2 were
not available for presentation in this
paper. Comparison of the two sets of
data provides some perspective on the
variability encountered in measurement
of these low levels of volatile organic
compounds in incinerator stacks. Table 7
shows the average results for the two op-
erating conditions and also presents
blank values for MRI's data. The sample
values in these tables are not blank cor-
rected; therefore, the values should be
viewed relative to the blank values. The
acetone values, for example, are uncer-
tain when the blank values are consid-
ered. Also, the chlorpmethane data were
obtained from GC analysis of periodic
grab samples and may represent a somewhat
different average than the other com-
pounds collected in the VOST. Chloro-
methane has a low boiling point and is
not retained well on the Tenax used in
the VOST. There were indications in the
data that some portion of the chloro-
methane may also be attributable to blank
background levels.
During steady state conditions in
runs 1 and 3, methane, chloromethane, and
chloroform were emitted at the greatest
rates, based on the MRI data. Methane
accounted for 27 percent of the total
volatile emissions and chloromethane and
chloroform accounted for .another 7 per-
cent. All other compounds were emitted
at much lower levels.
The largest change in concentrations
between the steady state conditions and
the transient conditions during runs 4
through 6 was a substantial increase in
methane emissions. Dichloromethane and
benzene also increased significantly, by
factors of 30 and 12, respectively. The
higher values for these two compounds,
however, were still below 100 ppb. No
other volatile compounds showed the large
increase in concentration as those three
compounds. Runs 5 and 6 showed much
higher concentrations of methane and di-
chloromethane than run 4 (see Table 6).
This is not explained at this time but
is similar to the trend in THC discussed
earlier.
To compare the volatile compound
emissions to the total THC emissions, all
results were converted to an equivalent
concentration in parts per billion (ppb)
as methane. The conversion factors from
ppb to ppb as methane are based on the
effective carbon number commonly used in
estimating FID response factors and on
measured literature values. Table 8
-307-
-------
presents a summary of the volatile com-
pound concentrations in ppb as methane.
This comparison is useful to assess the
degree that total emissions have been
identified, but it does distort the rela-
tive mass emissions between compounds.
Nonchlorinated compounds have a higher
response on a FID detector than chlori-
nated compounds. For example, compare
the relative levels of chloroform and
benzene on Table 7 versus Table 8.
Table 9 compares the total hydro-
carbon concentrations with methane, total
volatile and other compounds, based on
the data in Table 8. Further analyses
are being completed to identify some of
the other category. A comparison of the
average concentrations of the two sets of
runs shows a significant increase in the
total hydrocarbon (THC) concentration.
Methane concentration increases and also
becomes a much larger proportion of the
emissions during the transient runs. The
volatile compounds increase slightly in
concentration but their percent contribu-
tion to the total hydrocarbon emissions
decreases.
In addition to measurement of emis-
sion levels of volatile compounds, the
destruction and removal efficiency (ORE)
was measured for carbon tetrachloride,
the principal volatile compound in the
waste. The DREs calculated from the MRI
data for the steady state test runs
ranged from 99.998 to 99.9995. DREs dur-
ing the transient test runs were slightly
lower, ranging from 99.994 to 99.998.
The DREs calculated from the Dow data
ranged from 99.9992 to 99.9998 for the
steady state runs and from 99.9995 to
99.9997 for the transient runs.
Particulate Hatter and HC1 Emissions
Table 10 presents the particulate
and HC1 emissions, and HC1 efficiency for
each run. The range of particulate emis-
sions was 9.0 to 35 milligrams/cubic
meter (mg/m3). Ash loading to the incin-
erator was very low compared to typical
operation of this incinerator and to typ-
ical operation of other rotary kilns that
burn hazardous waste. The range of HC1
emissions was 0.066 to 0.11 kg/hr. HC1
efficiencies averaged 99.95 percent.
These rates are all very low compared to
the regulatory limits and to typical re-
sults from hazardous waste incinerator
tests.
CONCLUSIONS
Final conclusions for this project
will not be available until all of the
data have been analyzed. However, some
preliminary conclusions can be reached,
based on the data presented in this
paper. These conclusions are:
o Methane was a substantial portion of
the total hydrocarbons emitted from
the hazardous waste incinerator
tested. The percent contribution
from methane was higher for the
transient upset conditions than dur-
ing steady state operation.
o MRI's preliminary data show a large
portion of the total hydrocarbons
emitted during steady state opera-
tion was not accounted for with the
volatile organic compounds identi-
fied. Further analysis for semi-
volatile compounds is expected to
account for some of the unidentified
portion.
o The change in concentration of vola-
tile compounds between the steady
state and transient upset conditions
was not uniform across compounds.
It appeared to be compound specific.
-308-
-------
TABLE 1. SUMMARY OF SAMPLING AND ANALYSIS PARAMETERS AND METHODS
Sample
Liquid organic waste
Sampling
method
Tap (S004)
Sampling
frequency
for each run
One grab sample
every 15 nrin
composited into
one sample for
each run
Analytical
parameters
SV organics
Chlorides
Heating value
Ash
Viscosity
Analytical method
GC.MS
Organic halide (04327-84
or D808-81)
Calorimeter (D240-73)
Ignition (0482-80)
Viscometer (D-88-81)
Aqueous waste
Solid waste
Scrubber feed and
effluent water
Ash
Stack gas
VOA vial1
filled from
composite
Tap (3004)
Tap (S004)
Scoop (S007)
Dipper (S002)
VOA vial
filled from
grab sample
Scoop (S007)
MM52
One at end of
run
One grab sample
every 15 rain
composited into
one sample for
each run
One VOA vial
every 15 min
One grab sample
per solid charge,
composited at end
or test
One grab sample
every 30 min
composited into
one sample each
run
One VOA vial
every 30 min
One grab sample
per run
2-hr composite
per run
V organics3
SV organics
Chlorides
Heating value
Ash
V organics
V organics
SV organics
Chlorides
Heating value
Ash
SV organics
V organics
SV organics
Particulate
HC1
Moisture
Temperature
Velocity
GC/MS
GC/MS
Organic halide (D4327-84
or 0808-81)
Calorimeter (0240-73)
Ignition (D482-80)
GC/MS
GC/MS
GC/MS
Organic halide (D432-84)
Calorimeter (02015-77)
Ignition (D482-80)
GC/MS
GC/MS
GC/MS
Gravimetric (EPA RMS)
Organic halide (D4327-84)
Gravimetric
Thermocouple
Pitot tube
-309-
-------
r
TABLE 1 (Continued)
Sutple Sampling
method
MM5
VOST (S012)4
EPA Reference
Method 3
A132
Continuous
Gas sampling valve
Gas sampling valve
Gas sampling valve
or syringe
Sampling
frequency
for each run
2-hr composite
per run
Three traps
pairs at
40 min per
pair per run
One composite
sample per run
One composite
sample per run
1 min averages
~ once/30 min
~ once/30 nrin5
~ once/30 min5
Analytical
parameters
SV organ ics
Moisture
Temperature
Velocity
Method 624
Compounds
Oxygen,
carbon
dioxide
Aldehydes
CO, C02, 02
NO^, THC
Cj. to C3
hydrocarbons
Aromatics
Halogenated
organics
Analytical method
GC/MS
Gravimetric
Thermocouple
Pi tot tube
GC/MS
Orsat
HPLC
NOIR, NDIR, paramagnetic,
chemi luminescent, FID
GC/FID
GC/PIO
GC/Hall
or PID
NOTE: Saopling method numbers (e.g., S004) refer to methods published in "Sampling and Analysis Methods for
Hazardous Waste Combustion," December 1983; analytical methods beginning with prefix 0 refer to ASTM
oettiods.
1 Volatile organic analysis vial.
* HM5 - Modified Method 5.
3 Volatile organic constituents.
4 VOST = Volatile organic sampling train.
s Maxinua rate permitted by analysis time.
-310-
-------
^c
J"—
^c
a
to
V)
UJ
o
a;
a.
CM
UJ
m
1—
•P
s
•f—
U)
c
CO
f.
r-
ID
C
3
OS
m
3
OS
^"
C
ce
CU
IB
•P
U>
>,
CO
cu
•P
to
CO
c
3
a;
CM
C
3
^v*
rH
C
3
3£
ID O
«*• in
rH in
rH
1
O
o
rH
•* ID
r- o
rH 00
rH
i
t~.
rH
tn oo
oo cn
i-H 00
rH
|
CM
00
i— 1
O
•P Ll_
.a
E cu
o i-
U 3
+J
£>£
to cu
•a D.
C E
O O)
O 4->
cu
to
CTl ^"*s
• CO
cn .
rH
I-H
1
cn
^
ID ^-N
• in
o .
rH T-T
I-H
1
'PP
00
CM /••>>'••
• I-H
rH .
rH CM
rH
1
co
' 0'
rH
in /— s
. CM
rH' .
rH CM
rH
CM
O
. rH
O ^"N
rHr-i
i-H
CO
6V
x^
rH^-x
. CO
Sr-1
rH
|
O
cn
v*-^
^
^0
s*^ t
N
0
0 00 0
CD • •
rH CM 00
rH in
in
CM
o oo oo
CD
CM •* CO
I-H ' . m'. : ",
in
CM
o co in
o
CO CO rH
- rH in
CO
C\J
O CM CO
O . • • ;
co CD cn
rT
CM
o o cn
0 rH
CO , . CO t
M ••«•"•
CM
CD o — av .. .
.CD •••'.'
CM cn CM
in
CM
CM
'E'
U
co
0) i- 31-
•P J= Q.J=
CO V, c "x.
S- *H 3 »r- 3
U) -P CO O
-z. \—
f
' -a
cu
CO
4->
U)
cu
U)
.g
-Q
3
GO
O
CD
rH
^"
CM
t-
O
CU
3
r—
CO
>
C
j V
CO
cu
t/>
CO
O)
r—
CO
3
CO
iH
ui
co
D)
S-
3
•P
CO
c~
E
Q
t[
-P
' 3
a.
c:
•P'
CO
cu
U)
a>
-o
3
(J
.^
4J
3
a.
c-
4->
CO
CU
jr
CO
-p
O
(_
IM
-311-
-------
TABLE 3. AVERAGE WASTE CHARACTERISTICS
Solid
Chlorinated
Polyethylene
Substituted
Liquid organic
To kiln
To SCC
Liquid aqueous
Feedrates
(Ib/hr)
steady state/
transient
pyridine 1,300/1,350
wax 510/470
cellulose 1,110/1,230
520/490
820/820
5,010/4,940
HHV C1 (%)
(Btu/lb) steady state/
steady state/ transient
transient
1,970/1,700 26.40/20.80
15,240/17,100 0.016/0.010
6,980/6,250 0.39/0.61
18,000/17,980 10.70/10.80
18,000/17,980 10.70/10.80
15/0 0.0044/0.0023
Ash (%)
steady state/
transient
35.10/37.60
0.16/0.087
4.10/3.00
0.033/0.026
0.033/0.026
0.031/0.034
TABLE 4. PRELIMINARY AVERAGE CO AND THC LEVELS
,
Run CO (ppm)
EPA van
1 ~ 1
(0-86)
2 NA
(2
3 ~ 1
(0-2.6) (2
4 10
(0-720)
5 11
(0.58-710) (1
6 15
(0.03-710) (6
THC (ppm)
Dow MRI Dow
3.5 7,6 0
(3-4) (7.4-9.0)
3.5 6.8 0
.6-4.7) (5.8-7.9)
2.7 6.2 0
.5-3.3) (4.9-10.0)
20 8. 8 NA
(2-691) (4. 1-160)
7.9 150 99
.6-538) (100-220) (79-150)
33 110 60
.2-819) (79-190) (55-63)
-312-
-------
TABLE 5. PRELIMINARY STACK CONCENTRATIONS OF VOLATILE COMPOUNDS FOR
STEADY-STATE RUNS (ppb) (NOT BLANK CORRECTED)
Compound
Methane
Chi oromethane
Dimethyl ether
Methyl bromide
Vinyl chloride
Methyl ene chloride
Acetone
Trichlorofluoromethane
1,1-Dichloroethylene
Chloroform
1,2-Dichloroethane
1 , 1 , 1-Tri chl oroethane
Carbon tetrachloride
Di chl orobromomethane
Benzene
Chlorodibromomethane
2-Chloroethyl vinyl ether
Bromoform
1,1,2, 2-Tetrachl oroethyl ene
CgHis
Toluene
Chlorobenzene
Ethyl benzene
MRI run 1
l.yOO1
2201
19
0.051
0.86
3.7
8.7
4.2
0.99
63
2.7
0.25
3.9
14
4.9
. 2.3
2.0
0.14
1.2
1.2
7.9
0.11
1.0
Dow run 1
NA4
30
NQ3
ND
2.1
0.85
NQ
NQ
ND
16
1.2
0.16
2.0 ..
4.4
8.0
1.3
ND
1.2
0.36
ND
7.3
0.060
0.66
Dow run 2
NA
3.7
NQ
ND
ND
0.68
NQ
NQ
0.020
31
1.3
1.5
0.83
5.6
11
0.91
ND
0.083
0.33
ND
2.4
0.12
0.22
MRI run 3
l.SOO1
1001
0.63
0.15
1.2
1.7
7.9
0.18
0.050
65
0.22
1.2
1.3
13
2.1
1.7
0.25
0.029
0.41
0.76
0.94
0.086
0.14
Dow run 3
NA
ND2
NQ
X
ND
ND
0.77
NQ
"X
NQ
§1X
0.017
26
0.16
0.83
0.61
5.7
3.4
0.80
ND
0.044
0.30
ND
4.7
0.086
0.061
Total
2,100
NA
NA
1,500
NA
1 Data from GC analysis of grab samples; other data from GC/MS analysis of VOST
samples.
2 ND = Not detected.
3 NQ = Not quantitated. Compound detection was qualitative by GC/MS. Quantisation
will be made available for final report.
4 NA = Not available.
-313-
-------
TABLE 6. PRELIMINARY.STACK CONCENTRATIONS OF VOLATILE COMPOUNDS FOR
TRANSIENT RUNS (ppb) (NOT BLANK CORRECTED) -,
Compound
MRI
run 4
Dow
run 4
MRI
run 5
Dow
run 5
MRI
run 6
Dow
run 6
Methane 4.4001
Chloromethane ' 200
Dimethyl ether - 1.4
Methyl bromide 2.3
Vinyl chloride 0.74
Methylene chloride 5.4
Acetone 13
Trichlorofluoromethane 0.26
1,1-Dichloroethylene 0.12
Chloroform 73
1,2-Dichloroethane , 0.54
1,1,1-Trichloroethane 0.74
Carbon tetrachloride 7.4
Dichlorobromomethane 16
Benzene 62
Chlorodibromomethane 4.6
2-Chloroethyl vinyl ether 1.4
Bromoform 6.4
1,1,2,2-Tetrachloroethylene 0.25
C8H18 44
Toluene 1.7
Chlorobenzene 0.56
Ethyl benzene . .. . 0.80
Total 4,800
NA4
2.3
NQ2
ND3
ND
.1.7
NQ
NQ
ND
28
0.49
0.40
0.69
6.8
25
0.52
ND
11
0.16
NQ
1.4
0.19
0.041
93,000*
200 1
1.5
3.8
1.8
150
12
0.39
0.27
> 24
0.097
0.40
11
1.2
19
0.54
2.2
0.64
0.16
3.6
0.92
0.58
0.18
NA
140
NQ
ND
ND
40
NQ
NQ
0.070
3.8
0.056
0.20
1.1
0.18
11
0.15
ND
0.47
0.039
NQ
11
0.36
0.034
51.0001
3101
1.5
, 4.3
2.8
91
12
0.75
0.25
18
0.46
0.22
3.3
2.9
56
0.35
0.66
0.11
0.087
< 0.0085
0.71
0.60
0.11
NA
81
NQ
3.0
0.70
37
NQ
NQ
0.12
9.9
0.34
0.25
1.1
2.1
7.4
0.16
ND
0.12
0.0087
NQ
2.0
0.54
0.045
NA 94,000
NA
52,000
NA
1 Data from GC analysis of grab samples; other data from GC/MS analysis of VOST
samples.
2 NQ = Not quantitated. Compound detection was qualitative by GC/MS. Quantitation will
be made avail able,for final report. .
3 NO = Not detected.
4 NA = Not available.
-314-
-------
^
.a
a.
a.
**•''
to
a
CD
a.
g
u
UJ
— J
I-H
:ȣ
z
•=C
OO
z
^
OO
z
o
HH
Ej
) —
UJ
o
z
o
o
•f
t-
a
UJ UJ
0 I—
. o; •
S>
T3
.3
O
O.
E
O
O
CO
c
s
s-
Co in
to c:
!- 3
01 s-
en 00 00 OO ID ID <3*
Z ZZZ ZZZ .... .Z..Z*.. Z
r- , CM rH rH
? r^ CM, r^ in f*^ oo vo
<4-mCX> «^-CM CO^-<^-00 OD^«^-rH rHincO
^ O rH CO rH CM CM O O 00 O O f^ to ID rH rH CM O ID rH CO O O
O ^J* 00 rH >CO «^ rH O
0 CM o
2! • ID CO OOrH
, CM
ID ID
O CM o OOrHcncnoO
O ID rH ' • UD • t— 1 •' O
in rH OO
* . . . ^*
O)
c
t- a) 4J
tOOJ CO)O}O) fl)f-.
a) -uai aiJT-i-nj nj"?, s-
•a 0)1— t=4J5-j= xrc: o
&- ox: jc o i — o) ai> XT
i- a) a) o t-+j +> £. x: E E o
ojoj'O'Oi — OQJ a>ouo o i — to 01
^$-EtI"5 ^2 °x:Eo ox? HJ Sc
4-> S-i — OJ Oi — S-r— -r- O) Xt Xt 0) E r— C N flj
fl^E>» LI O) OJ O L) M— O 1 — ft- 0) TD t- O CM O) XI 0) O
10 S- 4-> >,i— >,OX:e3S-CJrHOr— 0)5-!— OCM OOQ)S-r—
x: o a> x: >,x: <-> u i o i -xi X:NOX:E »' <-• a o >> •
•<->•— E4-> C-l-> O)T-rHi— CMrH S- O C i— CJ O rH 3: i— r— X?
oix:*!— o»*i— o»ot- «x: •• ~toM-o»x:i &. *>ooox:-fj
.
o>
Q.
Q
U)
i
tn
>»
"io
c
10
t/)
o
o
g
g
*•
10
+J
Ol
x:
0
. ^_
^!
t
tn
^^
IO
s.
D)
*|
O
tn
V)
nj
|
to
a
rH
Q)
i
'S
o
z
II
g
(M
t-
(U
1
»r—
>
«j .
at
•S
m
XI
,-.
•r-
3:
c '
o
4-*
.**
jJ
C
Itf
3
cy
m
to
^s
0
^»
^J •
9) i .
>
•*•>
5'
15
13
O~
U)
c
0
LI
"
Q)
XI
•o
~
O
o (
JJ
,
I
n>
c .
to 4->
3 S-
cr o
Q.
4J Q)
O !-
Z
II *g
z"
z
II
a
z
-315-
-------
V)
o
8
31
§£
UJ
u-s:
in a.
z: Q-
:*;=>
o o-
0) C
co
c
Qi
c
Of
o
a.
o
o
CM co H CM co IB r-- t~- oo «o in o CM
r-cM
55
o
o
CM O CT> CO •* CO CO rH LO COrHr-lrHOr-linCn
O 00 CM rH
co
CO
co
o
00
V r».
CM UD CO O in CM in CM ^* *T in
o
in
in rH CD
CO CO r-^
rH
OinOrHCM^rHO^CMrHrHCMCMrHCOOOCOCOO^ |
Ol-HrHin rHi-H CO CO 00 inOOO«d-C\)lf)inrH
OOi-HdcMrHinddindrHdrHCOrHOOOUSUSOrH O
O> r-4 in i-H rH ^ J>j
*^"
CM CM rH in i-H CO in UD
V ^*^
S- OJ
a) i—
•^ >>
a>
c:
re ai
j= c
•p ai
« 73
JC -r-
+J S-
>
0)
O
E +»
a E > Q
r- S- S-
<(- O E O
O i— S- r—
S- J= O -C
0) O O t- U
_ C I— •!- O T-
r— o .c: ea s- a
_; +J o i o i
U
•!- O>
S- 4->
rH o
--Q
I-H S-
** re
tH O
•P
(U
O
I
JD
O
t- QJ
O C
<— CO
J= N
U C
•r- 01
O CO
CO
O i
U
i— re
>, s-
^ +J
+J d)
a> e K-
o £ i
S. O CM
O <*-
CD
c a)
c
CD -0 O>
COX)
;:cg*iss&s;
• O O rH 3C i— i— -C
: i s- -co o j= 4->
I CM CO rH O I— O Ul
-316-
-------
TABLE 9. PRELIMINARY MRI CONCENTRATIONS AS A PERCENT
OF TOTAL HYDROCARBONS
THC Methane Total
Run 1
Run 3
Run 4
Run 5
Run 6
Averages:
Runs ,1&3
Runs 4-6
* All ppm
(ppm)* (ppm)* 5
7.6 1.7
6.2 1.3
8.8 4.3
145 93
106 51
6.9 1.5
87 49
units are as methane.
£ Total volatiles
(ppm)* % Total
22 0.58 8
21 0.23 4
49 1.1 13
64 0.58 0.4
48 0.84 0.8
22 0.41 6
56 0.85 1
Other
% Total
70
75
38
36
51
72
43
TABLE 10. PRELIMINARY PARTICULATE AND
. HC1 EMISSIONS
.Run Parti cul ate
(mg/m3 )
1 15.9
2 14.2
3 9.0
4 11. 1
5 23.6
6 35.5
HC1 HC1
emissions1 efficiency1
(kg/hr)
0.066 0.99981
0.074 0.99948
0.074 0.99956
0.11 0.99918
0.081 0.99963
0.11 0.99955
1 Average of two values.
-317-
-------
INCINERATION OF CLEANUP RESIDUES
FROM THE BRIDGEPORT RENTAL
AND OIL SERVICES SUPERFUND SITE
Larry R. Waterland, Johannes W. Lee,
Robert W. Ross, II, Jerry W. Lewis,
and Carlo Castaldini
Acurex Corporation
Environmental Systems Division
Combustion Research Facility
Jefferson, Arkansas 72079
ABSTRACT
Several PCB-contaminated wastes will be generated through remedial actions at the
Bridgeport Rental and Oil Services (BROS) Superfund site in Bridgeport, New Jersey. Among
these are a lagoon surface oil, an underlying sludge, and contaminated soil.
Inclnerability testing of these three wastes plus a combination of the soil and sludge was
performed at the Environmental Protection Agency's (EPA) Combustion Research Facility
(CRF) to determine whether thermal treatment via incineration was a viable treatment
option for these wastes. Tests under three incinerator operating conditions were
performed in the CRF rotary kiln incineration system for each waste. Test variables
included rotary kiln temperature and rotation speed (solids residence time) and
afterburner temperature. All wastes contained between 100 and 300 ppm polychlorinated
biphenyls (PCBs) as Arochlor 1254. PCB destruction efficiency (DE) was in the 99.99 to
99.999 percent range for all tests. All scrubber blowdown samples had nondetectable PCB
levels (<1 ppb) and hazardous constituent trace element concentrations well below
extraction procedure (EP) toxicity thresholds. Kiln ash samples for the soil, sludge, and
soil/sludge wastes were not PCB contaminated, having nondetectable PCB levels (<0.4 ppm).
The composite kiln ash for the lagoon surface oil tests contained 2.6 ppm PCB. EP
leachates of all kiln ash samples had hazardous constituent trace element concentrations
well below EP toxicity thresholds.
INTRODUCTION
One of the primary functions of the
Combustion Research Facility (CRF) is to
support the Environmental Protection Agency
(EPA) Regional Offices in evaluations of
incineration as a disposal option for
wastes generated through remedial actions
taken at Superfund sites. This report
summarizes test results obtained during the
Incineration of hazardous wastes from the
Bridgeport Rental and Oil Services (BROS)
Superfund site located in Bridgeport, New
Jersey. Several PCB-contaminated wastes
that will be generated from the BROS site
are being considered for thermal treatment
using incineration. Among these wastes are
PCB-contaminated lagoon surface oil, lagoon
sludge, and contaminated soil. The primary
objective of the CRF tests was to determine
whether these PCB-contaminated wastes could
be treated via conventional incineration to
render them uncontaminated, while meeting a
PCB destruction efficiency (DE) of
99.9999 percent, as mandated under the
Toxics Substances Control Act (TSCA).
Emissions of products of incomplete
combustion (PICs) in the flue gas and the
EP toxicity trace element content of
incinerator residual streams were also
investigated.
-318-
-------
APPROACH
The CRF rotary kiln system, depicted
in Figure 1, was used to incinerate the
wastes during these tests. The primary
rotary kiln chamber has a maximum design
temperature of about 900°C (1,650°F) with a
nominal bulk gas residence time of about
1.7 sec. Further thermal treatment takes
place in the afterburner chamber which is
designed for a maximum temperature of
approximately 1,200°C (2,200°F) and a bulk
residence time of about 1.0 sec at a
typical excess oxygen level of 8 percent.
The air pollution control system consists
of a venturi and a packed column scrubber,
followed by a carbon adsorber and a high
efficiency particulate filter.
Four waste materials were tested
during this program. The chemical
composition and PCB and metals
concentrations of each waste material are
summarized in Table 1. The lagoon surface
oil, a dark brown viscous fluid containing
some debris, was pumped in the kiln via a
progressive cavity pump. The sludge, a
black aqueous gel also containing much
debris, was also fed into the kiln via the
progressive cavity pump. The contaminated
soil, a clumped clay mud containing the
most debris (rocks, grass, roots, twigs)
was fed into the kiln using 5.7L (1.5 gal)
fiberpacks via the available ram feeder.
The mixture of sludge and soil was also fed
in fiberpacks via the ram feed system.
PCB contamination in these test waste
materials ranged from about 100 mg/kg for
the soil and soil/sludge mixture to nearly
300 mg/kg for the lagoon surface oil.
Highest concentrations of hazardous
constituent trace metals were measured for
lead (46 to 2,900 mg/kg). Barium and
chromium were also detected in each test
material in the range of 12 to 1,030 mg/kg.
Leachate analyses of each test material
showed trace elements concentrations well
below the EP toxicity limits. Thus, these
test materials would not be considered
hazardous wastes under RCRA but, because of
PC? contamination, they are considered
hazardous under TSCA provisions.
Table 2 summarizes the average
incinerator conditions during each test. A-
Venturi
inlet tluct
Burner
No. 2
er~
Propane -»•* •• I r
Transfer I
duct
flecirculation
pumji
RecircuUtion Rlpwilown Rlowrtown
tank tank tant.
. No. 1 No. ?
Figure 1. Rotary kiln incinerator system.
-319-
-------
•*• "* 1
§o!
O —
• - > e
. <*._
-------
TABLE 1. BROS WASTE CHARACTERIZATION — COMPOSITION
Ultimate
analysis
(% by weight as fed)
C
H
0
N
S
Cl
Total
High heating value,
MJ/kg (Btu/lb)
Total PCBs
(mg/kg as Arochlor
1254)
Metals (mg/kg):
Arsenic, As
Barium, Ba
Cadmium, Cd
Chromium, Cr
Lead, Pb
Mercury, Hg
Selenium, Se
Silver, Ag
EP toxicity leachate
Arsenic, As (5.0)
Barium, Ba (100)
Cadmium, Cd (1.0)
Chromium, Cr (5.0)
Lead, Pb (5.0)
Mercury, Hg (0.2)
Selenium, Se (1.0)
Silver, Ag (5.0)
Area 1
soil
11.4
4.6
25.0
0.1
0.4
0.44
41.94
0
67.3-167
(110)a
<1
744
<1
55
756
<1
<1
<5
(mg/L):
<0.1
0.12
<0.1
<0.1
0.46
<0.1
<0.1
<0.1
Lagoon
surface
oil
54.4
10.9
29.9
0.1
0.7
0.1
96.1
8.62
(3,716)
270-300
(286)b
2
1,035
<10
46
2,888
<1
<1
<10
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
<0.1
Lagoon
sludge
1.0
11.1
81.8
1.0
1.0
0.01
95.91
10.0
(4,348)
250
<1
23
<5
12
46
<1
<1
<5
0.19
<0.1
<0.1
<0.1
<0,1
<0.1
<0.1
<0.1
Soil and
lagoon
sludge
13.1
4.7
32.2
0.4
0.01
0.06
50.86
2.43
(1,048)
78.6-170
(123)b
11
823
4
65
1,034
<1
<1
<5
0.1
0.30
<0.1
<0.1
0.12
<0.1
<0.1
<0.1
^Average value of 9 analyses on 3 composite samples
"Average value of 3 analyses on 1 composite sample
-321-
-------
TABLE 2. AVERAGE TEST CONDITIONS
Test
Lagoon surface oil
Test 1
Test 2
Test 3
Soil
Test 1
Test 2
Test 3
Soil plus sludge
Test 1
Test 2
Test 3
Sludge
Test 1
Test 2
Test 3
Feedrate
Feed method kg/hr (Ib/hr)
Progressive
cavity
pump
Flberpack
drum
ram-feed
Flberpack
drum
ram-feed
Progressive
cavity
pump
18 (39)
20 (45)
24 (53)
44 (98)
44 (96)
45 (100)
42 (92)
42 (92)
43 (95)
40 (88)
35 (78)
33 (72)
Kiln
Kiln
Afterburner
rotation
speed Temperature Exit 02 Temperature Exit 03
(rpm) °C (°F) (percent dry) °C (°F) (percent dry)
0.5
0.5
0.5
0.2
0.2
0.2
0.2
0.2
0.4.
0.4
0.4
0.4
667 (1,250)
677 (1,250)
893 (1,640)
927 (1,700)
700 (1,290)
888 (1,630)
715 (1,320)
800 (1,650)
893 (1,640)
654 (1,210)
882 (1,620)
654 (1,210)
12.6
10.4
10.3
8.2
10.3
8.8
13.1
NA
9.8
NA
NA
NA
1,140 (2,080)
1,210 (2,210)
1,130 (2,070)
1,130 (2,060)
1,130 (2,060)
1,120 (2,050)
1,120 (2,050)
1,120 (2,050)
1,120 (2,050)
1.120 (2,040)
1,130 (2,050)
1,210 (2,210)
. 5.6
6.0
6.5
8.0
7.2
6.7
8.5
7.5
6.8
10.1
6.2
6.2
total of 12 tests were performed, 3 for
each of the waste materials, with variable
kiln and afterburner temperatures, excess
oxygen levels, and kiln rotational speed.
The effect of maximum afterburner
temperature in combination with low kiln
temperatures was investigated during two
tests, one with the lagoon surface oil and
the other with the sludge as waste feed
materials. All other tests were performed
with afterburner temperatures at
approximately 1,120°C (2,050°F). Propane
was used as the primary fuel in both the
kiln and afterburner chambers to maintain
test temperatures. An increase in excess
oxygen at the afterburner exit from 5 to
10 percent corresponds to an approximate
45 percent increase in gas flowrate at
constant temperature. This increase in gas
flowrate translates to a corresponding
reduction in bulk gas residence time.
Figure 2 illustrates the sampling
protocol used during these tests. Organic
compound emissions were measured at the
packed scrubber outlet location and
upstream of the carbon bed absorber, using
standard EPA methods (2,4). Particulate
mass and HC1 emissions were measured at
three locations, namely afterburner exit,
packed scrubber outlet, and stack, using
the EPA Method 5 modified for acid gas
(HC1) collection in the impinger section of
the sampling train (5). The kiln ash and
scrubber blowdown were also sampled for
priority pollutant metals and organic
contents. Analytical methods conformed to
those prescribed-in References 1 and 3.
TEST RESULTS
PCB destruction measured during these
tests generally ranged between 99.99 and
99.999 percent as illustrated in Figure 3.
None of the test conditions investigated
resulted in a DE of 99.9999 percent
required under TSCA provisions for PCB
incineration. Attempts to correlate DE
results with individual process parameters
did not reveal any statistically
significant correlations. However, test
data illustrated in Figure 4 may indicate
a tentative effect of reduced gas residence
time at temperature on measured DE. In
this figure, the DE is plotted versus a
mean gas temperature, defined by the
arithmetic average of kiln and afterburner
temperatures, normalized by the gas
flowrate through the afterburner. The
linear regression line shows a relatively
good fit (correlation coefficient = 0.82)
with increasing DE as temperature and gas
residence time increase. Extrapolation of
the linear fit to 99.9999 percent DE would
indicate a gas flowrate requirement of
about 1.2 m^/sec for a mean temperature of
-322-
-------
5 -
Test 1
UCOON ML J771
Test 2
17771 SOCL+SUJDGE
Figure 3. PCB DEs.
S9.9999
99.999-
99.99-
99.9-
99-
90-
Linear regression
coefficient •= 0.8Z
0.5 ,
0.7
0.9
(Thousands)
1.1
Figure 4.
(Tkiln * TAB)/acm/sec
PCB DE vs. temperature/flowrate
(°F/acm/sec).
970°C (1,780°F). This gas flowrate
corresponds to a residence time in the
afterburner of 2.0 sec or .about 2.3 times
that recorded during these tests.
Table 3 summarizes the concentrations
of most common volatile organic compounds
detected in the waste feed materials.and
the incineration, effluents. Flue gas
concentrations of individual compounds were
generally below 10 ug/dscm (approximately
1 ppb). Particulate mass and HC1 emissions
were measured at levels below those
mandated by RCRA regulations (180 mg/dscm
corrected to 7 p,ercent 02 and, 1.8 kg/hr,
respectively) for all tests as shown in
Table 4. Highest particulate levels were
recorded during incineration of
contaminated soil. "
Table 5 summarizes the concentrations
of most common trace elements detected in
the kiln ash and its leachate and in the
scrubber blowdown solids and liquids. The
kiln ash was found to contain barium,
chromium, and lead in concentrations
ranging from about 100 to 2,160 mg/kg.
Lead concentrations in the kiln ash and
blowdown solids for the soil and soil plus
sludge tests clearly reveal partitioning of
this metal to the flyash rather than the
kiln ash. The partitioning is not evident
for the other metals. Kiln leachate
analyses and blowdown liquids indicate
concentrations typically less than
1 mg/L or well below the EP toxicity
levels.
ACKNOWLEDGMENTS
This research was performed under the
Combustion Research Facility operations and
research contract with the U.S.
Environmental Protection Agency's Hazardous
Waste Engineering Research Laboratory. The
invaluable assistance and guidance provided
by the EPA Project Officer, Robert ,
Mournighan are gratefully acknowledged.
The authors would also like to thank Ronald
Borsellins of EPA Region II, and Henry
Munoz of the U.S. Army Corps of Engineers
for their significant contributions to this
effort.
REFERENCES
1. EPA, 1984. Test Methods for
Evaluating Solid Wastes: Physical
Chemical Methods. EPA SW-846, 2nd,
ed., Revised, U.S. EPA, OSW,
Washington, DC 20460, April 1984.
2. Hansen, Earl M., 1984. Protocol for
the Collection and Analysis of Vo 1 atfi 1 e
POHCs Using VOST. EPA-600/8^84Po6T;
U.S. EPA, AEERL, Research Triangle
Park, North Carolina 27711, March
1984.
3. Harris, Judith C., Deborah J. Larsen,
Carl E. Rechsteiner, and Kathleen E.
Thrun, 1984. Sampling and Analysis
Methods for Hazardous Waste~
Incineration. EPA-60~0/8-84~-002,
U.S. EPA, AEERL, Research Triangle
Park, North Carolina 27711, February
1984.
-323-
-------
TABLE 3. MOST COMMON VOLATILE ORGANICS
Compound
1,1-dichloroethylene
Chloroform
1,2-dichloroethane
Carbon tetrachloride
1,2-dichloropropane
Trichloroethylene
Benzene
Hexane
Tetrachloroethylene H
Tetrachloroethane
Toluene
Feed concentration (ppm)
Scrubber exit flue gas concentration
(pg/dscm)
Lagoon
surface
oil
4.7
27
16
68
4
21
3.7
23
ND
Sludge
ND
ND
14
63
25
12
2.1
6.0
16
Soil +
sludge
15
ND
59
57
ND
5
ND
ND
ND
Lagoon
surface
oil
ND
ND-5
5-31
6-76
ND
ND
ND-1,730
ND-6
2-7.5
Soil
ND-2.4
2.4-2.7
ND-16
4.4-17
ND
ND
0.7-1.4
0.6-3
7.4-17
Sludge
ND
5.8-9.2
ND
ND-5.1
0.7-1.4
2.1-3.5
6.5-8.4
2.2-9.3
ND-8.8
Soil +
sludge
ND
3.2-5.3
ND
14-18
ND
ND
1.2-2.0
1-1-0
2.0-6.6
44
4.7
5.1
2-10
7-12 2.9-5.1 3.4-5.2
TABLE 4. SUMMARY OF PARTICULATE AND HC1 EMISSIONS
Afterburner exit Scrubber discharge Stack
(Location 4) (Location 6) (Location 7)
Waste feedstock
Lagoon surface oil
Soil
Sludge
Soil plus sludge
Particulate
mg/dscma
8.30-17.9
(13.1)b
48.8-279
(126)
18.3-31.9
(27.1)
15.8-224C
(112)
HC1
mg/dscm
9.60-17.3
(13.2)
16.8-30.9
(21.9)
4.90-7.20
(5.90)
17.8-23.2
(20.6)
Particulate
mg/dscma
<0.2-0.5
(0.33)
9.70-16.9
(13.1)
4.70-8.00
(6.60)
9.30-124
(47.6)
Particulate
mg/dscma
NA
13.7-26.1
(19.4)
12.5-45.7
(23.6)
39.6-182
(129)
HC1
mg/dscm
NA
<9.6
• <9.8
<10.0
aParticulate matter concentration corrected to 7 percent plus 02
^Number in parenthesis is the arithmetic average of three tests
C0ata suspect due to unusually high proportion of particulate catch in the probe rins
NA -- Not available. Samples not taken.
-324-
-------
TABLE 5. TRACE ELEMENT CONCENTRATIONS
Test /sample
Lagoon Surface Oil
Composite
Composite kiln ash leachate (mg/L)
Average blowdown liquids (mg/L)a
Soil
Average kiln ash (ppm)a
Average kiln ash leachate (mg/L)a
Average blowdown solids (ppm)b
Average blowdown liquid (mg/L)a
Sludge
Composite kiln ash (ppm)
Composite kiln ash leachate (mg/L)
Average blowdown liquid (mg/L)a
Soil plus Sludge
Average kiln ash (ppm)a
Average kiln ash leachate (mg/L)a
Average blowdown solids (ppm)
Average blowdown liquid (mg/L)a
Arsenic
<2
0.33
<2
<60
<2
<2
<24
Barium
120
0.33
0.54
550
0.45
980
0.33
680
<0.1
0.41
740
0.32
820
0.37
Chromium
1,090
<0.1
0.16
130
100*
0.29
110
<0.1
0.17
87
<0.1
74
0.31
Lead
2,160
0.23
0.70
<
910
2,400*
1.3
800
0.12
<0.1
450
<0.1
5,010
0.12
aAverage over three tests.
^Average over Tests 2 and 3 only. Test 1 blowdown contained no solids.
Note: EP toxicity limits are 5.0 mg/L for arsenic, chromium and lead, and
100 mg/L for barium.
4. Schlickenreider, Lynn M., Jeffrey W.
Adams, and Kathleen E. Thrun, 1975.
Modified Method 5 Train and Source
Assessment Sampling System Operators
Manual. EPA-60078-85-003, U.S. EPA,
AEERL, Research Triangle Park, North
Carolina 27711, February 1985.
5. 40 CFR Part 60, Appendix A.
-325-
-------
PILOT-SCALE TESTING OF NONSTEADY BOILER WASTE COFIRING
Howard B. Mason, Julie A. Nicholson, Carlo Castaldini,
Robert J. DeRosier, and C. Dean Wolbach
Acurex Corporation
Mountain View, California 94043
and
Ivars J. Licis
Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Waste destruction efficiencies were measured for volatile and semivolatile
chlorinated organic compounds cofired with gas, oil, and coal in a pilot-scale boiler
simulator with a maximum capacity of 3 million Btu/hr. The tests were run to help
Interpret waste destruction data from 14 prior boiler cofiring field test programs.
Specific issues addressed in the pilot-scale tests were: what is the background level of
waste emissions due to residual deposition on boiler surfaces?; what operating conditions
fall to yield acceptable destruction efficiency?; what waste products of incomplete
combustion are formed, and in what quantities?; and what is the form and fate of trace
metals contained in the waste? Operating parameters varied were excess 02; atomizatlon
patterns; residence time; firing rate; wall cooling; and waste or fuel flow transients.
INTRODUCTION
Since 1981, the Environmental
Protection Agency (EPA) has tested 14
Industrial boilers for performance in the
thermal destruction of hazardous wastes
cofired with conventional fuels (1). The
test facility specifications and operating
conditions were selected to obtain a
reasonable coverage of the diverse range of
Industrial boiler designs and operating
practices. Field test data were obtained
for the following range of conditions:
• Flretube, package and field
erected watertube designs; 0.3 to
>2 seconds residence time in main
f1rebox
• 2,500 to 350,000 Ib/hr steam
capaci ty
• Gas, oil, coal, and wood fuels
• Volatile and semivolatile wastes
with heating value from 0 to
18,000 Btu/lb
• 0 to 100 percent heat input from
wastes
The results of these 14 tests showed an
overall mass weighted waste destruction
efficiency of 99.998 percent. This high
level of destruction was generally
exhibited across the broad range of
designs, waste characteristics, and
operating conditions tested, and prevailed
even when the boilers were intentionally
operated in off-design upset conditions
such as substoichiometric air,
malfunctioning atomizers and waste flow
transients. In fact, the destruction
efficiencies were uniformly so high that it
was not possible within the constraints of
plant operation to identify the operational
boundary between acceptable and
-326-
-------
unacceptable destruction efficiencies.
This remains an open issue.
Within the context of these high
destruction efficiencies, several patterns
in the data recurred at most of the sites.
Although the average destruction efficiency
was nearly five nines, the data did exhibit
variability of up to an order of magnitude
which was apparently not due to artifacts
in the sampling and analysis scheme or to
contamination. Additionally, the
variations did not correlate with any '
boiler operational setting or with boiler
performance indicators such as CO or NOX
emissions, or smoke. It was also observed
at several sites that the stack
concentrations of waste species continued
for several hours after waste firing was
curtailed. This hysteresis effect may be
attributable to deposition of trace
concentrations of waste species on boiler
surfaces, particularly on areas with
carbonaceous deposits from oil or coal
firing. •
Soot blowing, the regular steam
cleaning of the boiler heat transfer "
surfaces, is a normal part of boiler '
operation. Limited data were obtained
during the previous tests indicating that
organics were present in higher- " '
concentrations during the soot blow cycle.
Only the gaseous portion'of the flue gas
was analyzed due to the characteristics of
the sampling systems. Species and
concentration of organics adsorbed on the
carbonaceous and solid materials being
expelled during soot blowing cycles were
not determined.
Other significant issues remained at
the conclusion of the 14 full-scale boiler
tests.
A key issue is the identification of
the form arid fate of heavy metals.' Very
little data in this area has been generated
by the previous testing for which the •
emphasis was destruction efficiency (ORE).
The effectiveness of air pollution control
devices is also not well defined.
A second/issue is the more complete
identification of products of incomplete
combustion (PICs). Several tests showed '
PIC emissions in boilers a"s a higher
fraction of the organic constituents in the
flue gas than those in incinerators. The
types and concentrations of PICs resulting
from fuels in comparison to cofired waste
have not been identified in sufficient
detail.
A third issue is the interpretation of
data from field tests in the presence of
background variability and how this
variability relates to the condition of
boiler heat exchange surfaces and operating
conditions. Some understanding of this
issue is needed to gauge the confidence
level of conclusions to be reached
regarding the relative insensitivity of
waste destruction to boiler operating
conditions.
The present pilot-scale tests
addressed the above issues. The
pilot-scale approach was selected over
full-scale testing for a number of reasons.
The range of parametric variation and
degree of off-specification operation
envisioned would be difficult to achieve at
full-scale field installations.
Additionally, the experiments could be more
closely controlled at a test facility..
There would also be some cost saving's
realized in conducting pilot-scale testing
at a home facility as compared to
maintaining a test team at a field
facility. /
FACILITY '' . •" ' ' r " •' •"
The pilot-scale facility is a
refractory-lined furnace and convective
section designed with geometric similarity'
and time-temperature simulation of a
tangentially or wall-fi.red watertube boiler
(Figure 1). For the present tests, the
unit was fired in the. wall-fired mode,'
typically with gas, .oil, or coal,fired
through three of the burners and a
simulated waste through'the remaining.'
burner. The fuel feed system, cooliKg
system, and postcombustiori cleanup system
are shown schematically in Figure 2. For
most tests, the unit was fired at. a firing
rate of 1.5 million Btu/hr which gives a
residence time in the firebox prior to
quenching in the convective section of
approximately 2 seconds. The wall-fired
burners were variable swirl "research
burners equipped with a single fuel
injector tube. Combustion air is preheated
to 400°F upstream of the burner by an
electric heater. For oil firing, 'steam
atomization is typically used, although air
atomization is available. The convective
section consists of approximately 20 tube
-327-
-------
Convective section
Hot sampling ports
Heat exchange-*-^
tube bundles V\
Cold sampling
station
To baghouse-
Ash pit
Figure 1* Pilot-Scale Experimental Furnace
-328-
-------
Ol
+J
01
o
t-
ia
0)
oj
O)
u
.•a:
a.
Ul
CM
0)
3
-329-
-------
bundle heat exchange drawers that are
cooled to about 300°F with Dowtherm.
For oil firing, a 2.2-percent sulfur
residual oil with heating value of
18,300 Btu/lb was used for the three
burners fired on conventional fuel. The
oil was heated to about 220°F by drum
heaters and heat traced recirculation
loops. For coal firing, an Illinois No. 6
coal with a 3.6-percent sulfur content was
pneumatically conveyed to the burners.
The synthetic waste fired through the
remaining burner was a blended mixture of
80 percent by weight No. 2 distillate oil,
and 5 percent each carbon tetrachloride,
trichloroethylene, monochlorobenzene, and
trichlorobenzene. The mixture was
recirculated continually with a pump to
retard stratification, and fed t'o the steam
atomized burner at room temperature. For
baseline testing in the absence of waste,
all four burners were fired on conventional
fuel.
Thermocouples, sampling ports, and
viewports are positioned along the furnace
and convective section. Temperatures,
flowrates, and emission measurements are
recorded and processed on a microcomputer
or data logger. Continuous monitoring was
done for 02» CO (two scales), C02,
hydrocarbons, SQz, and opacity.
Volatile waste constituents and PICs
were sampled in the flue gas with the
volatile organic sampling train (VOST).
Two methods of using the VOST were used.
For periodic comprehensive quantification,
the full -VOST sampling protocol was used
with GC/MS analysis following thermal
desorption. To reduce the cost and
turnaround time for multiple samples, a
reduced protocol termed "mini VOST" was
used. The sampling was done on a single
VOST trap for 10 minutes and the trap was
taken to a nearby gas chromatograph and
desorbed to a parallel flame ionization
detector and a Hall detector to quantitate
hydrocarbons and chlorinated species,
respectively. This procedure can provide
near real time feedback on volatile organic
waste concentrations.
Semivolatile wastes and products of
incomplete combustion were sampled by a
Modified Method 5 train fitted with an
organics module containing 65 grams of XAD
resin. The train was typically run for
approximately 4 hours to a .total sample
volume of 4.3 dscm (150 dscf). The
expended XAD was Soxhlet extracted along
with other catches from the Method 5 and
analyzed with the GC/MS.
TEST PROGRAM
To address the issues identified in
the Introduction, seven test series were
run as follows:
I Baseline — Fuel Only
II Baseline — Cofiring
III Data variability, replication,
and hysteresis
IV Evaluation of high destruction
efficiency operational window
V Form and fate of trace metals
VI Evaluation of PIC production
VII Confirmatory tests
The Series I and II baselines were run to
obtain reference levels of waste and PIC
emissions under well controlled operation
and for a clean test facility. Later runs
as heat exchanger deposition increases will
be compared to the baselines to determine
possible background and hysteresis effects.
This effect will be explored in Series III,
where a sequence of volatile and
semivolatile measurements will be compared
to determine the background levels of waste
and PICs and the reproducibility of the
waste destruction efficiencies. With the
data variability established and some
calibration made on the confidence level,
the remainder of the testing will be on
obtaining a data base on waste emissions
and PICs and the influence of operating
conditions. Series IV will quantify the
effect of off-design and upset operation on
the high destruction efficiency window.
Factors to be tested include low combustion
air to the level of smoking and CO
excursions; atomizer upsets; and flow or
load transients. Additionally, wall
cooling and lower residence times via
capacity increases will be evaluated. In
Series V, trace metals will be injected
into the flame and the concentrations will
be determined upstream and downstream of
the baghouse as a function of particle
size. Series VI will test the influence of
-330-
-------
waste identity on the concentration and
identity of PICs. Inorganic chlorides will
be used in place of the waste to determine
if waste identity is dominant in PIC
production. Finally, Series VII will test
specific conditions found from the Series I
to VI results to have a strong effect on
waste or PIC emissions.
REFERENCES
1. Castaldini, C., et. al. "Engineering
Assessment Report: Hazardous Waste
Cofiring in Industrial Boilers," Final
Report TR-84-159/EE on EPA
Contract 68-02-3188, Acurex
Corporation, Mountain View, California,
June 1984.
-331-
-------
TECHNICAL/ECONOMIC ASSESSMENT OF SELECTED PCB DECONTAMINATION PROCESSES
Ben H. Carpenter
Research Triangle Institute
Research Triangle Park, North Carolina 27709
and
Donald L. Wilson
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
ABSTRACT
Eleven emerging alternative treatments for Polychlorinated Biphenyl (PCB) contaminated
sediments have been compared and ranked using technical performance status of develop-
ment; test and evaluation data needs, and cost as factors. In ranking the processes,
weights were assigned the factors to emphasize the extent of decontamination, the esti-
mated cost of treatment, and the versatility of the process. The emerging treatment proc-
esses are based on five different technologies: one on low-temperature oxidation, two on
chlorine removal, one on pyrolysis, four on removing and concentrating, and three on
fdcroorqanlsms. Types of technologies not developed are chlorinolysis, stabilization, and
enzymes: On the basis of the comparisons made, the treatment processes were ranked in the
following order from highest to lowest: KPEG, LARC, Acurex, Bio-Clean, Supercritical
Water, Advanced Electric Reactor, Vitrification, OHM Extraction, Soilex, Composting, and
Svbron B1-Chem 1006. The first eight processes show potential for reduction of PCB con-
centrations to the desired background levels (1 to 5 ppm) or less, with minimum environ-
mental impacts and low to moderate cost. All the technologies except,the advanced elec-
tric reactor required further development and testing.
INTRODUCTION
The PCB contamination problems in the
Hudson River and New Bedford,
Massachusetts are reported to be among the
worst in the United States in terms of
concentration and total quantity of PCB's.
It 1s estimated that 290,000 kg of PCB's
are contaminating 382,000 m3 (500,000 yd*)
of sediments of the Hudson River. During
the 70s, approximately 907,000 kg of PCBs
were used in the New Bedford area annual-
ly, of which an estimated 45,500 kg were
improperly disposed. The PCB contamina-
tion problems pose threats to both drink-
Ing water and the fishing industry. There
are also numerous industrial lagoons con-
taminated with large quantities,of PCBs.
The only available proven technology
is dredging and expensive incineration.
Land disposal of the sediments untreated
has legal restrictions. Biodegradation is
a possibility, but sufficient information
does not exist to design and operate such
a system. There is little experience in
the application of encapsulation tech-
nology to PCB-contaminated sediments.
This study was undertaken to identify
the most technically feasible processes
that have been proposed by research
concerns for the removal of PCBs from
sediments; to identify their extent of
development, effectiveness, limitations
and probable costs; and to determine needs
for further development. The study in-
volved four phases: data acquisition,
-332-
-------
screening and'selection of the most tech-
nically feasible processes, development of
criteria for process assessment, and proc-
ess assessment.
DATA ACQUISITION
Three major sources of data were:
EPA's file of proposals and correspondence
concerning problems of PCB contamination
.and possible approaches to alternative
solutions; the open literature; and direct
contacts with proponents of treatment
technologies.
A bibliography (171 references) was
prepared, which included treatment
feasibility study reports, process test
and evaluation reports, process develop-
ment proposals, and patents. As processes
were Identified, direct contacts were made
with the investigators for details of
their process studies.
SCREENING AND SELECTION OF MOST
TECHNICALLY FEASIBLE PROCESSES
Alternati ve destructi on/detoxi f1ca-
tion/removal (DDR) processes were subject-
ed to screening to identify those to be
assessed further. The processes were
categorized according to their generic
technology so that their potential perfor-
mance could be judged appropriately.
Processes with undesirable aspects were
rejected from further assessment. For
example, lack of tolerance for water by a
process is undesirable because extensive
sediment drying is required. Processes
showing insufficient tolerance for water
were therefore rejected from further con-
sideration as a primary treatment process
in favor of more tolerant alternatives.
Table 1 lists the processes screened,
identifies those selected for further ,
assessment and gives the reasons for
rejection of the rest. Some of the tech-
nologies (e.g. nucleophilic substitution)
have provided several processes. Some
(e.g. enzymes) have not yet provided any
processes. A process evaluated as "1" in
Table 1 was selected for further assess-
ment. Other evaluation numbers assigned
to the rest of the screened processes
refer to footnotes that identify the
reason for rejection of the process for
further assessment. References cited are
identified fully in the bibliography.
DEVELOPMENT OF CRITERIA FOR PROCESS
ASSESSMENT
The PCB contamination problem in the
Hudson River is representative of the type
of PCB destruction/detoxification problems
focused on in this study. It is expected
that the contaminated sediments will have
to be dredged from all sites and that the
dredged sediments will have high water
content.
Criteria for assessment of alternative
treatments were chosen which relate to a
broad range of principles of operation of
diverse applied technologies, yet can be
used effectively in comparing one treat-
ment process with another. Additional
factors, specific to a technology, were
included to help portray the inherent
strengths and limitations of a process.
Table 2 lists the seven criteria used in
comparative process evaluation and three
additional factors relating to the needs
for further process development and evalu-
ation. The table also includes an overall
description of the findings for the proc-
esses evaluated.
The goal set for process performance
is to reduce the PCB concentration in
treated sediments to levels of 1 to 5 ppm.
Several of the processes were found to
meet this goal. Those that showed reduc-
tion to less than 2 ppm were assigned a
rating of "6". Those that attained a
.level between 2 and 10 ppm were assigned a
"4". Those with residual concentrations
greater than 10 ppm were rated "2".
Available capacity was found not to
exist for any of the processes. However,
several were developed sufficiently to
permit projections of the time required to
build a facility for application of the
treatment. Those for which such projec-
tions could not be made were rated "2".
Those requiring 24 or more months were
rated "4". Those requiring 12 to 16
months were rated "6".
Conditions/limitations that were rated
were tolerance for water, required proc-
essing time, and controllability of proc-
ess conditions. Those treatments that
could tolerate water up to about 40 per-
cent would not require a drying step with
its attendant fines' control problems.
Those requiring only 1 day for treatment
-333-
-------
TASLE 1. SCREENING OF PCS TREATMENT PROCESSES
Generic
technology
References
Process
Evaluation3
CHEMICAL
General
Low-tepperature oxidation
Hat air oxidation
Centofanti 1971; Chan 1982; Childs 1382;
Craddock 1982; Edwards etal. 1982;
Envircnrant Canada 1933; Homig 1984;
tossey and Walsh 1985; Rogers'and Komel
1985; Rogers 1983; Rogers 1985.
Baillcdetal. 1978; Miller and
Sevientonievski (n.d.); Miller and Fox
1982.
Supercritical water oxidation Model 1 et al. 1982.
Chemical oxidants FMC Corporation (n.d.).
Ozcnation
Chlorine recoval
Oehydrochlorination
Arisman et al. 1981; Lacy and Rice
Deschlaeger 1976; Prengle and Mauk 1978.
U.S.P. 4,469,661
Ctu and Vick 1985; Lapiere et al. 1977.
Reducing agents
Chu and Vick 1985; Sworzen aid Ackerman 1982.
Nucleophilic substitution
Brovn et al. 1985a; Srunelle and Singleton
1985; terch 1968; New York University 1984;
Rozz et al. 1985; Smith and Gurbachsn 1981;
Sunohio (n.d.); Sweeny and Fischer 1970;
United States Patent Office 1984b; Weitzman
1984; Weitzman 1984; Weitzman 1985.
Uncatalyzed, general 2
Zimpro Process, Santa Maria, 4,13
CA Waste Site
Catalyzed
Dow Chemical Co. Patent 3,984,311 2
IT Environmental Science' 2
Modar 1
Potassium permanganate plus Chromic
Acid and Nitric Acid 6
Chloroiodides 4,7
Ruthenium tetroxide 3,4,8
G£ UV/ozonation process 2
Molten aluminum/distillation 14
Catalytic: 2,3
Nickel on kieselguhr 2,3
Pd on charcoal 2,3
Lithium aluminum hydride 2,3
Butyl lithium 2,3
Raney Nickel 2,3
Sodium in liquid ammonia • 7,9
Nickel-catalyzed zinc reduction 7,9
Hydrazine 7,9
UV light plus hydrogen 2
Mildly acidic zinc powder,
Sweeney and Fisher (1970) 2,14
Sodium-based processes:
Goodyear, sodium in naphthalene (1980) 10
Acurex, proprietary solvent 10
PCBX/Sun Ohio ' 10
PPM '- 10
Ontario Hydro Power 10
(continued)
-334-
-------
TABLE 1 (continued)
Generic
technology
Nucleophilic substitution
(continued)
Radiant energy
References
Bailin and Hertzler 1977; Sailin and
Hertzler 1378; Bailin et al. 1978;
Craft etal. 1975; Oevetal. 1985;
• Kate et al. 1981; Meuser and Weimer 1982;
Plimer 1978; Rogers and Komel 1985;
Rogers 1985; Trump et al. 1979; West et al.
1983. .,
Process
Potassium poly (ethylene glycolate)
based:
EPA In-house KPE8
KPB3 Terraclean-CI
SEKOH-PES
New York University KPEG
UV/photolysis
Syntax photolytic
Thermal corona glow
Microwave plasma
RF insitu heating
Gama radiation (Craft et al. 1975)
LARC
Evaluation9
19
1
11
12
3
3,4
5
9,17
18
9
1
Electromechanical reduction Massey and Walsh 1985.
Cjilorinolysis Sworzen and Ackerman 1982.
Pyrolysis
PHYSICAL
Removing and concentrating
Boyd 1985; New York State Department of
Environmental Conservation 1985a; New
York State Oepartnsit of Environmental
Conservation 1985b.
Angiola and Soden 1982; Caron 1'985; Gilmer
and freestone 1978; Githens 1984; Hancher
et al. 1984; Hawthorne 1982; Lee et al.
'1979; Saunders 1985; Schwirn et al. 1984;
Versar, Inc. 1984.
Electromechanical research process 14
Hoechst process . 9
Gocdyear catalytic hydrogenolysis 9
, Exhaustive chlorination 9
Advanced Electric Reactor 1
Wright-Malta alkaline catalyst fuel-gas
process u
Heated Air Stripping
Extraction
Adsorption
Vitrification
TiirnErraan 1985.
American Toxics Disposal, Inc. i«
Critical Fluid Systems, COj 14
Furfural 15
Acurex solvent wash 1
0. H. H. extraction 1
Soilex process \
Carbon adsorption, general 13
Neoprene rubber .adsorption 15
Sattelle vitrification process 1
(continued)
-335-
-------
TABLE 1 (continued)
Generic
technology
Stabilizing
References
Ghassemi and Haro 1985; Law Engineering
Testing Coipany 1982; Stroud et al. 1978;
SutnacBnian and Rahalingam 1977; Tittlebaun
et al. 1985.
Process
Asphalt with lime pretreatment
Z-Lnpremix
Sulfur-asphalt blends (K-20)
Ground freezing
Evaluation3
16
15
16
13
Bottora
BIOLOGICAL
HicrooroanisES
Erayces
Carich and Toffleraire 1983; Hand and Ford
1978; Murakami and Takeishi 1978; U.S. Array
Corps of Engineers teter Resources Support
Carter 1983; Ziimrie and Tofflanire 1978.
Bedard et al. 1985; Bumpus et al. 1985;
Clark et al. 1979; Oawes and Sutherland
1976; Furakawa 1982; Isbister et al. 1984;
Kong and Sayler 1983; HcCoraick 1985;
New York State Department of Environmental
Conservation 1985a; New York State Depart-
ment of Environmental Conservation 1985b;
Rhee et al. 1985b; Rhee et al. 1985;
Untenran et al. 1985.
Catelani et al. 1971; Rochkind et al.
Merman et al. 1985.
Dredging
Bio-Clean
Sybron Bi-Chsm 1006 P6
Composting
Bio-Surf
Ecolotrol, Inc.
Wormes Biochemical's Phenobac
Rhee anaerobic degradation
No processes found.
13
1
1
1
4,13
4,13
11,13
14
^Explanation of process rating:
1. Identified emerging ssdieent treatment process.
2. Destruction efficiency appears to be too low to meet environmental goals.
3. Processing tine appears to be extremely long for practical timely cleanup.
4. Data availabla for dioxin, other chlorinated compounds, or other contaminants, hut not PCB's.
5. Process has been show to destroy PCB's in gas streams only. It may be feasible for sediments, but has not been show to
be.
6. PCB's with 5-7 chlorine atcois per molecule are not destroyed.
7. Products of partial degradation may be toxic.
8. Reagent fe very costly/toxic or both.
9. Process costs appear to be excessively high compared with other emerging treatment processes.
-336-
-------
10. Water destroys the reagent or interferes with its action, thus the process would require excessive drying of sediitents and.
probably, extraction in pretreatments. The process would therefore haw application only as a subordinate final step to
several extraction and concentration operations.
11. This particular process was not evaluated because data were not available for assessment.
12. This process is an alternative to another process using the sane generic technology, but it is in very early stages of
development, and data were not available for assessment.
13. This technique is basically applicable to preliminary operations prior to treatment or to treatment of wastestreans (e g
vastewaters) from chemical or physical treatments.
U. This process is in the concept stage and data are insufficient to assess it for PC8-«ntanrinated sediments.
15. This process has been found to be ineffective.
16. This technology provides only for encapsulation of the PCS-caitaminated sediments.
17. This process supports incineration of PCS's.
18. The process does not appear to be feasible for submerged sediments.
19. Basic data support to identify emerging treatment process.
-337-
-------
UTBRIA AND TECHNICAL FACTORS USED IN PROCESS ASSESSMENT
tj-i
o
CS3
Cd
trt
S3
£-«
Description
£4
Q
^J
1*4
(0
^*
cd
*^i
CD
4J
•*-{
f-j
0
C
•pHj
(1)
0 >>
•p
rt
o*
P CU U
C X Ll -H
o co p bo
0 • 0
co
p
•H
B
•H f
^ —
tfi
a
o
P
73
o
u
cesses required more than 3 weeks. • -
o
o
o
CO
o
4J
O
.§•.
The PCB concentration of the sediments treated ranged from u
ppm. Some processes had limits inherent in the technology.
73
CU
pH
73
a
cs
;•
cu •
bo bo
RI a
JJ -H ,
CO P
CO
P CU
a p
cu .
o
CD
Q
CO
Ed 73
CU
73 CU
d 23
Rt
Rl
P P
CO 19
0, 0
1
CO -H
O P 73
•H CO CU
^ O CO
3 O
O 73
CO CU
Li 3 P
(U rH Rl
o< a cu
P - p
coco
0 6 •, o
The estimated costs of treatments were made in terms of the
meter of dry sediment treated, assuming a density of 1.68 Mg
of associated operations-dredging, transportation, handling
ments, as required. All costs are stated in 1985 dollars.
.p
CO
o
0
73
CD
P
id
•H
p
10
Ed
73
-------
•5
3
e
•H
P
s
o
o
"~*
•
CV1
w
1
e .
0
•^
a
•H
LI
o
CO
cu
a
LI
o
.u
o
(8
to
(8
•»-l
Li
CU
•H
Ll
u
e
o
.P
(8
CU N
P] •*•)
•M t<
CD
- P
CO U
p ce
C Li
CU (8
M J3
C8 U
CU 0)
> -p
•rt CU
•P i-H
u a
(8 g
o
CD O
3,
e
to a
•a c
3 0
r-l -H
U P
13 0
•rt 3
- -P
T3 CO
CU CD
.0 TJ
•H
LI oa
0 0
CO Ou
T3 4~*
o
CO
C8 co
- . x e
to •
>> -rt CO
So e c
o ce o
i-H >13 !rt
O O 4J
C 01 18
J3 S L,
O CD
cu -a a
P C O
C8
co P
CO CO -rt
cu cu e
O -H S
o a,
Ll «rt .-1
ft 0 rt
C C8
CD -rt
e-i a o
CO
13
O
•H
L!
o1
Li ^
0 -H
o B
CQ
CO
p
e
cu
s
p
C8
CU
Li
•P
CU
si
p
g
0
L>
>
^ C
cu cu
•H *H
f£3 O
O -rt
(8
CD
CO — -
CO 06
cu •^
u o
o ^
ft "^^
CU -H
• X! 18
p >
o
i— 1 CD
< tl
u
c
cu
•rt
o
•rt
Ed
^£
\
Q
\
Q
•d
cu
•rt
+J
00
-339-
-------
could generally show a faster rate of
cleanup than those requiring 3 days. Some
biological processes required more than 3
weeks. The treatments generally provided
control of the processing conditions;
however, a few (e.g., composting) would
not necessarily do so. The three condi-
tions/limitations were ranked as follows:
Conditions/limitations Rank
Tolerates to 40 percent water and
treats in 1 day 6
Sediment needs to be dried 5
Tolerates to 40 percent water and
treats in 3 days 4
Tolerates water and treats in >3
weeks 3
Sediment needs to be dried, treats
in >3 weeks 2
Processing conditions uncontrollable 1
Concentration range handled in data
developed for the processes ranged from
unknown to 3,000 ppm. Ratings were
assigned based on the upper limit of feed
concentration. The ratings were as fol-
lows:
PCB concentration treated, ppm Rank
;>3,000 6
2,000 to 3,000 5
1,500 to 2,000 4
500 3
250 to 350 2
Unknown 1
Status of development ratings were "1"
for no data, "2" for laboratory-scale
tests completed, "3" for bench-scale tests
completed, "4" for pilot-scale tests com-
pleted, "5" for field tests completed; and
"6" for commercial system designed and
ready for construction.
Test and evaluation data needs could
be rated differently, depending upon the
purpose. For indicating the extent to
which a treatment process is readied for
use, the more data that are available the
better. For indicating the need to sup-
port a very promising technology that
lacks sufficient progress, the potential
and the data needs should be rated in
combination. The ratings used here are
for the former purpose and are as follows:
Test and evaluation data needs Rank
None except permits and checkout 6
Field tests 5
Pilot tests and costs 4
Laboratory and bench tests 3
Conceptual treatment process design 2
D/D/R data, residual PCB data,
RCRA waste data 1
The application of any treatment proc-
ess can involve the need for one or more
of the following unit operations:
dredging, transport, storage, landfill
disposal, land treatment disposal, incin-
eration, and/or alternative treatment.
Estimates were developed for all of these
so that, in any given process evaluation,
the proper elements could be added to
obtain an estimate of the cost of applica-
tion. The estimates were made in terms of
the cost per cubic meter of sediment
treated. The sediment was assumed to have
a density of 1.68 Mg/m3.
Dredging costs for those treatments
requiring removal of the sediment before
treatment are estimated at $20/m3 based on
the recent experience of the U.S. Army
Corps of Engineers in contracting for
dredging in the New York State area
(Wheeler, 1986).
Transport costs are given as a range.
The Corps' experience is $13/m3 for short
hauling distances (Wheeler, 1986). A cost
of $126/m3 was used for long hauling dis--
tances, which represents an assumed 483-km
average transport distance to RCRA land-
fills capable of accepting PCB-
contaminated wastes (Industrial Economics,
Inc., 1985).
Storage cost will sometimes be in-
curred to hold the dredged sediments pend-,
ing treatment; e.g., where dredging rates
exceed the rates at which the treatment
can be applied. These have been set
arbitrarily at $10/m3.
Land treatment was used in one of the
processes to degrade residual solvent left
in the soil after treatment. This in-
volves the controlled application of
-340-
-------
wastes to the surface of the soil. At
land-treatment facilities, wastes are
either spread on or injected into the
soil, followed by tilling into the soil
with farm equipment. The physical and
chemical properties of the soil, in unison
with the biological component of the soil
and sunlight work together to immobilize,
degrade, and transform portions of the
wastes. The application and tilling proc-
ess can be repeated many times on the same
plot, making land treatment a dynamic
system designed to reduce and ultimately
eliminate a portion of the waste, as
opposed to permanent storage such as land-
fills.
The American Petroleum Institute
(1983) has reported that there were 213
land-treatment facilities in operation
handling waste from 16 different industry
sectors. The most extensive use of land
treatment is for petroleum refinery
wastes, with 105 land-treatment facili-
ties, many of which are located on the
same site as the refinery. More recently,
EPA verified the existence of 114 land-
treatment facilities and obtained informa-
tion on operating parameters at some of
these sites (Thorneloe, 1986).
Wastes are typically mixed to a depth
of 0.5 to 1.0 feet, where biochemical
reactions take place. Application fre-
quencies can range from daily to yearly,
with tilling occurring as frequently as
daily.
The average cost of controlled,
managed land treatment cited by the
American Petroleum Institute, $60/ton,
equates to $lll/m3 of sediments. For
short-term land treatment of readily-
degradable solvents remaining in treated
sediments free of PCBs after they are
washed or dried, the cost is estimated at
$33/m3 (Cap!an, 1986).
Redeposition costs of decontaminated
sediments were also estimated at $33/m3.
Slightly lower costs might be expected in
special cases.
Because the regulations permit the use
of incineration or chemical waste landfill
and the application costs of these two
methods are available from firms engaging
in their practice, these costs were used
as lower and upper limits with which to
compare the costs of applying new alterna-
tive technology.
Landfill disposal costs, incurred when
the sediments must be placed in authorized
chemical waste landfills, are estimated as
ranging from $260/m3 for the Michigan area
(EPA Regional Office) to $490/m3, based on
the highest prices charged for hazardous
wastes by commercial facilities (Indus-
trial Economics, Inc., 1985). This range
includes an intermediate value of $420/m3
reported by the Corps of Engineers.
Costs for incineration techniques
capable of achieving 99.9999 percent
destruction and removal efficiencies for
PCBs are difficult to predict. Even more
difficult is prediction of the price com-
mercial facilities will charge to accept
the responsibility of handling such a
sensitive waste. Surveys made to deter-
mine the likely charges to incinerate
dioxin-containing wastes resulted in a
reported price on the order of $l,000/Mg
(Pope-Reid Associates 1985). This trans-
lates to $l,680/m3, the value adopted for
this evaluation, and the cost of disposal
of residue from incineration is included.
The total cost of use of incineration
including dredging at $20/m3 and transport
at $13 to $126/m3 is $1713 to $1826/m3.
When available, alternative treatment
costs were obtained from the proponent of
the process. Otherwise, they were esti-
mated based on the types of unit processes
involved and the environmental controls
required, or they were determined not to
be estimable considering the status of
development of the process.
While all costs are in 1985 dollars,
the treatment costs are not all neces-
sarily based upon the same labor rates,
corporate fixed charges, or profit. These
costs vary from one firm to another. The
cited estimates are costs of purchasing
the treatments. Further cost analyses
will be needed to provide a basis for
comparison of processes on the basis of
individual cost elements.
Table 3 shows the unit cost estimates
used to develop cost ranges for the
emerging treatments.
Estimated costs were rated by compar-
ing the range of the cost estimates
-341-
-------
r
TABLE 3. UNIT COST ESTIMATES FOR STEPS INVOLVED IN TREATMENT
AND DISPOSAL OF PCB-CONTAMINATED SEDIMENTS
Operation
Cost,
Dredging
Transport
Storage
Landfill and Disposal
Landfarming
Restricted Land Disposal
Incineration
20
13 to 126
10
260 to 490
33
111
1680
-342-
-------
obtained with the cost of placing theni
into a chemical waste landfill. Treatment
processes showing the-lowest estimated •••
cost range were rated '"6"; those showing a
probable cost Tower than landfill were
rated "4"; those showing an estimated cost
equal to landfill were rated "2"; and
those showing an estimated cost range
greater than landfill were rated "1".
Overall ranking was accomplished
through the use of weighting factors
assigned to each rated factor. The
weighted average rank was then obtained by
summing the products of the weighting
factors and the ratings and dividing by
the sum of the weighting factors. The
weighting factors were:
Factor Weigti
Residual PCB concentration 5
Capacity 2
Conditions/limitations 3
Concentration range handled 2
Status of development 2
Test and evaluation data needs 1
Estimated costs 4
The weightings tend to give greatest
emphasis to the ability of the treatment
to reduce the PCBs and to the probable
cost of the treatment. Much less emphasis
is placed on the status of development.
Thus, an almost fully developed process
with an extremely high cost would be rank-
ed lower by application of the weighting
process than a less developed process with
a much lower potential cost. Test and
evaluation data needs have not been
heavily weighted because nearly all the
alternative treatment processes that show
low potential cost require more data to be
proven.
Under this procedure, the perfect
process for treating PCB-decontaminated
sediments would show the following levels
for each ranking factor and would receive,
using the ratings given, a weighted rating
of 6.0:
Factor level
Rating, R Wt R x Wt
• 1. Residual PCB, - ':- •- "
treated sediment
" less than 1" ppm: ' ., 6
) - ,' , ,
2. Capacity adequate1'
for site, cleanup, :
available in
12-16 mo. ' . " '' '- --.; 6
3. Tolerates to 40 '"' '
"percent water' and '"
treats in 1 day
(24 hr) 6
4. Handles concentra-
tions greater
than 3,000 ppm 6
5. Commercial system
designed and ready
for construction 6
6. No test and evalu-
ation data needs
except permits and
checkout 6
7. Lowest estimate cost
range among
alternative emerging
technologies 6
Total R x Wt E R x Wt
Weighted rating (E R x Wt)/EWt
PROCESS ASSESSMENT
30
18
12
12
24
114
6
The processes were assessed by charac-
terization and ranking. Characterization
provided for objective comparison of the
processes. Ranking provided a subjective
comparison of the processes based on the
seven criteria.
CHARACTERIZATION
Table 4 summarizes five characteris-
tics of the processes: unit operations,
available capacity, conditions/limita-
tions, concentration handled, and any
generated RCRA wastes. The unit opera-
tions employed are given, and each is
identified by a number. Generally, a
greater number of unit operations will
mean a greater effect on treatment costs.
-343-
-------
TA8LE4. TR&WENT PROCESS ASSESSMENT
Process
Unit
operations
Available
capacity
(or time
to provide)
Candidas
and limits
Concentration
handled
RCRA
waste
generated
Cheaical
Supercritical wt«r oxidation 1,4,10
KPEG, Terraclean-a
1,3,4,7 (24 no)
20-40% solids; 374 °C, >3000 ppm
23.3 HPa organic
content >5% or supple-
mental fuel
150 °C, 0.5-2 h
500 ppm or
greater
Here
w.w.tr.
act.
carbon
KPS3, m
KP6B, EPA in-house
URC
Advanced electric
reactor (l.H. Huber)
Physical
0. H. teterials
nathanol extraction
'Soil ex* keroseneAater
Acurex solvent wash
Vitrification
1.2,3.4,5,
6,7,9
Basic process data
1,2,5,15,
7,8,12.13
14
2,7.8,14
15
1,2,5,15
2,4,5.6,
10.11
8.12,14
(24 nx>) tolerates 25% water. 480 ppm
(16 no) 2204 "C, 2.400 WSVm3 >3000 ppm
heeds predryer
predry to <1% moisture >4flO ppra
251 of kerosene sol- to 350 ppm
vent retained in soil; tested
3 d per batch
3-12 washes, tolerates up to 1,983
<40% water. ppm
Electrical power usage 500 ppm
increases with soil
moisture; submerged
sediments dredged
and treated
Ncne
toe
PCfi-'oaoed
carton from-
solvent
cleanup
Concentrat-
ed PCS from
still to
incinera-
tion
Concentrat-
ed PCB's to
KPES
None
(continued)
-344-
-------
TABLE 4 (continued)
Process
Biological
Composting
Available
capacity
Unit (or tire
operations to provide)
15,16. (16 mo)
Conditions Concentration
and limits handled
Seasonal effects, 1,530 ppi
reaction time must be
>4 weeks
. RCRA
waste
generated
Treated
material
is still a
8io-Clean
Sybrcn 8i-Chan 1006
1,2,17
15,17
27 nfyd avail- Proved for PCP, labor- i3DO ppm
able, 12 so for atory confirmed for
full-size PCB's
Unknown
Unknow
fiCRA waste
None
Unkrni
NOTE-Unit operations key:
1. Liquid/solids separation
2. Extraction/solubilization (liquid-solids)
3. Liquid/liquid extraction
4. Chemical reactor
5. Stripping still
6. Solvent recovery still
7. Adsorption
8. Dryer (solids)
9. Dryer (liquids) ,•
10. Filtration
11. Steam cleaning
12. Therm! reactor
13. Grinding
14. Air pollution controls
15. Landfara
16. Irmoculation/digestion
17. UV light reactor
-345-
-------
None of the processes has currently
available capacity approaching that
required for major cleanups. Therefore,
the time required to build capacity is '•
listed. Construction time ranges from 12 ..
to 24 months.
Certain conditions that typify the
process or limit*its versatility'are given'
in column 4'of TabVe'4. fable 4 also
identifies any RCRA waste streams gener-
ated by the process.
The data from studies of the processes
were examined for ranges of PCB concentra-
tions handled to date. Generally, the
values are not limitations on the process,
but only on the data acquired. The value
£300 ppm for the Bio-Clean process may,
however,, be a limitation requiring process
adjustment to control.
Table 5 lists .five additional charac-
teristics of the processes and 'the rating
developed in the ranking process. The
characteristics shown here relate to the
needs for further process development and
evaluation. The process status is given
in terms of stages of development com-
pleted. t. The processes range in stages
completed from concept to pilot plant.
Both PCB destruction and residual PCB
concentration in treated sediments are
given to the extent available. Certain
processes may require extension of the
unit operations, employed (e.g., more
stages of extraction) to attain the re-
quired performance levels.
Test and evaluation data needs are
indicated for each process. Needs vary
from none (AER process) to complete site-
specific evaluation!
The estimated costs of applying the ~
process are listed in $/m3. Although cost
estimates lack the necessary accuracy at .,
this stage of development of the alterna-
tive processes to serve as the sole
criterion of potential, they nevertheless
indicate that seven of the processes may
cost no more than landfill and five could
cost less. (Cost estimates could not be
made for the Sybron process and compost-
ing.)
The processes varied in complexity as
evidenced by the number of unit operations
employed. Supercritical water oxidation,
Bio-Clean and vitrification each employed
three unit operations; KPEG employed
eight. -Operator training requirements
. were,not• .evaluated comparatively due to
insufficient data. However, for the
sealed-up treatment processes, the operat-
ing labor requirements are expected to be
quite similar.
RANKING OF TREATMENT PROCESSES
In contrast to process characteriza-
tion which involves all factors listed in
Tables 4 and 5, ranking is subjective and
is based on the seven criteria previously
described. An attempt was made to define
and determine a single number that could
represent the overall position of each
process relative to an arbitrarily defined
' perfect process.
Based on the weighted ratings, the
processes rank as follows from highest to
lowest: KPEG, LARC, Acurex, Bio-Clean,
Modar-Supercritical Water, Advanced
Electric Reactor, Vitrification, OHM
Extraction, Soil ex, Composting, and Sybron
Bi-Chem 1006 PB/Hudson River Isolates.
CONCLUSIONS
Emerging treatment processes for
decontamination of sediments containing
PCBs that''show potential as alternatives
to incineration and chemical waste land-
fill have been identified. Eleven alter-
native treatments were compared and ranked
using technical performance,'status of
development, test and evaluation data
needs, and cost as factors. The first
eight processes show potential for reduc-
tion of PCB concentrations to the desired
-"background levels (1-5 ppm) or less, with
minimum environmental impacts and low to
moderate cost. The sediments must be
dredged for application of these treat-
, ments...
Of the eleven processes assessed, all
but the Advanced Electric Reactor (AER)
are in various stages of development from
laboratory-scale through field tests. The
AER is a permitted treatment under TSCA in
EPA Region VI, based on completed trial
burns. There is no immediately available
capacity for any of the treatment proc-
esses. Further data are needed in most
cases to define the final system designs
for the processes.
-346-
-------
TABUS. TREATMENT PROCESS ASSESSMENT
Process
Estimated
0/0/R
Status3 efficiency, *b
Estimated
residual Test and evaluation Estiasted
PCS. ppm data heeds cats. I/*3 toting
Chemical/physical
Supercritical wter Field test with
oxidation, Modar PCS liquids
>99.9995 <0.1 ppb
1.2,3,4,5,6,1, 250-733
4.58
KPEG Terraolean-CL Pilot tests
>98
1.6
208-375
5.42
LARC
Lab tests
>90
38-50
2,3,4,5,6,7
223-336
S.»
Advanced electric
reactor
Pilot tests
>99.9999
<1ppb
830-943
Physical
0. H. Materials, Field tests under
methanol extraction way
97
<25ppn
2.3,6.7
401-514
4.16
Soilex
Pilot tests
6-9 ppn
5.6.7
856-913 3.26
(Jcurex solvent wash Pilot-scale e
(field tests
planned)
In-situ vitrification Pilot test of soil 99.9
Sattelle Pacific
NWforEPRI
- <2ppm
None in vitri-
fied block, 0.7
ppm in adjacent
soil
Identity of 196-569
mixed solvent,
6,7
6 255-548
5.21
4.S3
See footnotes at end of table.
(continued)
-347-
-------
Process
TABLE 5 (continued)
Estimated Estimated
0/D/R residual
Status9 efficiency, %b PCS, ppm
Test and evaluation Estimated
data needs costs, S/m3 Rating6
Biological
Cotpcsting, aerobic Lab-scale
anaerobic Lab-scale
62
18-47
504-908
825-1268
4,5,6
4,5,6
— 2.47
— 2.47
Bio-Clean, aerobic Bench-scale
99.99
Sybren Bi-Chsa 1006 Lab-scale and concept 50
25ppb
3,5,6,7
3,4,5,6,7
191-370
4.84
1.46
NOTE—Data needs key:
1. 0/D/R data
2. Residual PCS data
3. Unit operations data
4. Eench-scale data
5. Pilot-scale data
6. Field test data
7. Cost data
8. RCRA waste
Status is drfined in terras of the types of studies completed.
ba/Tj/R = destrocticn/detoxificaticn/reraoval.
cite rating was obtained as shew by the example, under Characterization.
dAER is fully permitted under TSCA in EPA Region IV for destruction of PCS.
<%eateent is continued until a residual of <2 ppm PCB's is obtained.
-348-
-------
At this stage, estimated costs of
application of these eleven processes are
less than or within the range of costs of
chemical waste landfill, .except for the
AER estimated cost, which exceeds that of
landfill, but is less than incineration.
These costs are planning estimates only.
In most cases, further research is needed
to provide data suitable for more definite
cost estimates.
The emerging treatment processes are
based on six types of generic tech-
nologies: low-temperature oxidation,
chlorine removal, pyrolysis, removal and
concentration, vitrification (melting),
and microorganisms. Types of generic
technologies not yielding competitive
emerging processes are: chlorinolysis,
stabilization, and enzymes. A search of
these technologies yielded no suitable
candidate processes at this time.
On the basis of the comparisons made,
the treatment processes were ranked in
order, from highest to lowest, as shown in
Table 6. The estimated cost range (1985
dollars) per cubic meter of sediment
treated is also shown for each process.
Costs of chemical waste landfill and in-
cineration are given.for comparison.
NOTICE
Although the research described in
this paper has been funded wholly or in
part by the United States Environmental
Protection Agency through contract number
68-02-3992 to the Research Triangle
Institute, it has not been subjected to
Agency review and therefore does not
necessarily reflect the views of the
Agency and no official endorsement should
be inferred.
BIBLIOGRAPHY
Adams, G.P., and R.L. Peterson. 1985.
Non-Sodium Process for Removal of PCBs
from Contaminated Transformer Oil, PCB
Seminar, EPRI, Seattle, Washington,
October 22-25.
Addis, G., and J. Marks, eds. 1982.
Proceedings: 1981 PCB Seminar, Dallas,
Texas, December 1-3, 1981. EPRI EL-
2572, Electric Power Research Institute,
Palo Alto, California. 331 pp.
Agnew, R.W. 1984. Removal and Treatment
. of Contaminated River Bottom Muds:
Field Demonstration. NTIS Publication
No. PB84-129022, EPA-600/52-84-006,
U.S. Environmental Protection Agency,
_Cincinnati, Ohio. 70 pp.
American Petroleum Institute. 1983. Land
Treatment—Safe and Efficient Disposal
of Petroleum Waste. Washington, D.C.,
21 pp.
Angiola, A.J., and J.M. Soden. 1982.
Predicting and Controlling Downwind
Concentrations of PCB from Surface
Impoundments. In: Proceedings of the
Annual Meeting of the Air Pollution
Control Association. 75(4):82.
Arisman, R.K., R.C. Musick, J.D. Zeff, and
T.C. Crase. 1981. Experience in Opera-
tion of a Ultraviolet-Ozone (Ultrox )
Pilot Plant for Destroying Polychlori-
nated Biphenyls in Industrial Waste
Influent. In: Proceedings of the 35th
Industrial Waste Conference, Purdue
University, West Lafayette, Indiana,
May 11-13, Ann Arbor Science, Ann Arbor,
Michigan, pp. 802-808.
Bailin, L.J., and B.L. Hertzler. 1977.
Development of Microwave Plasma Detoxi-
fication Process for Hazardous Wastes-
Phase I. NTIS PB-268526, EPA-600/2-77-
030, U.S. Environmental Protection
. Agency, Cincinnati, Ohio. 82 pp.
Bailin, L.J., and B.L. Hertzler. 1978.
Development of Microwave Plasma Detoxi-
fication Process for Hazardous Wastes-
Part I. Environ. Sci. Technol.
12(6):673-679.
Bailin, L.J., B.L. Hertzler, and D.A.
Oberacker. 1978. Detoxification of
Pesticides and Hazardous Waste by the
Microwave. Plasma Process. ACS Symposium
Series, No. 73, American Chemical
Society, Washington, DC, p. 49.
Baillod, R., et al. 1978. "Wet Oxidation
of. Toxic Organic Substances," presented
at the Purdue Industrial Waste Confer-
ence, W. Lafayette, Indiana.
Ball, J., F. Prizmar, and P. Peterman.
1978. Investigation of Chlorinated and
Nonchlorinated Compounds in the Lower
Fox River Watershed. NTIS PB-292-818/2,
-349-
-------
TABLE 6. SUBJECTIVE RANKING OF TREATMENT PROCESSES ON OVERALL
SUITABILITY, AND ESTIMATED COST OF APPLICATION
Process
Cost of application,
$/n)3 treated
KPEG
LARC
Acurex Solvent Wash
Bio-Clean
Hodar Supercritical Water
Advanced Electric Reactor
Vitrification
OHM Methanol Extraction
Soil ex Solvent Extraction
Composting
Sybron Bi-Chem 1006
Chemical Waste Landfill
Incineration
$211-378
$223-336
$196-569
$191-370
$250-733
$830-942
$255-548
$400-514
$856-913
Unable to estimate cost
Unable to estimate cost
$260-490
$1713-1826
-350-
-------
EPA-905/3-78-004, U.S. Environmental
Protection Agency, Great Lakes National
Program Office, Chicago, Illinois. 235
PP- . .,-'..;
Bartos, M.J. 1978. NYC's Plan to Meet
the Water Quality Challenge. Civ. Eng.
(Am. Soc. Civ. Eng.) 48(11):80-84.
Bedard, D.L., et al. 1985. Rapid Screen-
ing Assay for PCB-Degradative Ability,
Bio. Sci. Br., General Electric Company,
Corporate Research and Development,
Schenectady, New York 12301.
Berg, O.W., P.L. Diosady, and G.A.V. Rees.
1972. PCB Contamination of the Aquatic
Environment in Ontario. Paper presented
before the Division of Water, Air, and
Waste Chemistry at the 173rd Meeting of
the American Chemical Society, August
28-September 1, 12(2):59.
Bergh, A.K., and R.S. Peoples. 1977.
Distribution of Polychlorinated Bi-
phenyls in a Municipal Wastewater
Treatment Plant and Environs. Sci. Total
Environ. 8(3):197-204.
Billings, W.N;, T.F. Bidleman, and W.B.
Vernberg. 1978. Movement of PCB from a
Contaminated Reservoir into a Drinking
Water Supply. Bull. Environ. Contam.
Toxicol. 19(2):215-222.
Boyd, J.W., 1985. Thermal Treatment of
PCB's and PCB-contaminated Soils, EPRI.
PCB Seminar Oct. 21-25, Seattle,
Washington.
Brown, J.F., M.E. Lynch, J.C. Carnahan,
and J. Singleton. 1981. Chemical
Destruction of PCB's in Transformer Oil.
Preprint Extended Abstract. Presented
before the Division of Environmental
Chemistry at the 182nd Meeting of the
American Chemical Society, New York, New
York, August 23-28, 21(2):90-92..
Brown, M.P. 1985a. Characterization of
Hudson River Sediment Provided to
American Toxics Disposal, Inc. by the
New York State Department of Environ-
mental Conservation. Office of Special
Projects New York State Department of
Environmental Conservation, Albany, New
York. 7 pp.
Brown, M.P. 1985b.
Rogers, May 24.
Letter to C.J.
Brown, M, P. 1986; Letter pf.April 16 to
Donald L. Wilson'.
BruneTle, D.J., and D.A. Singleton. 1985.
Chemical Reaction of Polychlorinated
Biphenyls on Soils with PolyCEthylene
Glycol)/KOH. Chemosphere . 14(2):173-
181. \f '
Bumpus, J.A., et al. 1985. >;.Oxidation of
Persistent Environmental 'Pollutants by a
White Rat Fungus, Science 228, 1434-
1436, .June 20, 1985. , ,
Business Week. • 19781 : The Lagging Cleanup
of Great Lakes Pollution. Bus. Wk.,
May 29, p. 76. ' ' '
Caplan, J. 1986., Personal Communication
to Ben H. Carpenter', February 6.
Carcich, I.G., and J. Toff1emire. 1983.
PCB Cleanup Activities on the Upper
Hudson River.-'In: Management of Bottom
Sediments Containing Toxic Substances:
, Proceedings of, the 7th U.S./Japanese
Experts Meeting, National Technical
Information Service, Springfield,
Virginia, pp. 168-180.
Carcich, I.G., and T.J. Tofflemire. 1982.
Distribution and Concentration of PCB in
the Hudson River and Associated Manage-
ment Problems. Environ. Intl. 7 (2):73-
85. --•
Caron, R. 1985. Superfund Cleanup of PCB
Contaminated Soil, Minden, West
Virginia. Personal Communication with
Ben H. Carpenter, December.
Catelani, D., C. Sorlini, and V. Treccani.
1971. The metabolism of biphenyl by
Pseudomaas Putida. Experimentia
27:1173-1174.
Centofanti, L.F. 1984. Analysis of PCB
Chemical Destruction Effluents. Pre-
print Extended Abstract. Presented
before the Division of Environmental
Chemistry at the 185th Meeting of the
American Chemical Society, St. Louis,
Missouri, April 8-13, 24(1):11-12.
Chang, V.S., and D.A. Fast. 1983. PCBs
Spill at Federal Pioneer Limited's
Regina Plant. In: Proceedings of the
Environment Canada 1st Technical • Chemi-
cal Spills Seminar, Toronto, Canada,
October 25-27, p. 225.
-351-
-------
Chen, K. 1982. The Reclamation of Trans-
former Oils Containing PCBs by the
Sunohio PCBX Process, Final Report.
Prepared for the Tennessee Valley
Authority Office of Power, Division of
Energy Demonstrations and Technology,
Chattanooga, Tennessee. TVA/OP/EDT-
83/3, DE83 901683. 112 pp.
ChUds, K. 1982. PCB Treatment and
Destruction Technologies. In: Proceed-
ings of the 28th Ontario Industrial
Waste Conference, Ottawa, Ontario,
Canada, June 13-16, Ontario Ministry of
the Environment, Ontario, Canada, pp.
121-148.
Chu, N.S., and S.C Vick. 1985. Chemical
Destruction of PCBs, EPRI PCB Seminar.
Seattle, Washington. October 27.
Clark, R.R., E.S.K. Chian, and R.A.
GHffin. 1979. Degradation of Poly-
chlorinated Biphenyls by Mixed Microbial
Cultures. Appl. Environ. Mlcrobiol.
37(4):680-685.
Counsel, G. 1982. Wrench-Lok : Ground-
ing Connections Simplified. Paper pre-
sented before the Electrical Equipment
Committee at the Pennsylvania Electric
Association Fall Conference, Lancaster,
Pennsylvania, September 14.
Craddock, J.H. 1982. Polychlorinated
Biphenyls (PCBs) Disposal and Treatment
Technologies: An Update. Paper pre-
sented at the Fertilizer Institute
Environmental Symposium, San Antonio,
Texas, March 8-10, p. 161.
Craft, T.F., R.D. Kimbrough, and C.T.
Brown. 1975. Radiation Treatment of
High Strength Chlorinated Hydrocarbon
Wastes. NTIS PB-244388, EPA-660/ 2-75-
017, U.S. Environmental Protection
Agency, Athens, Georgia, 33 pp.
Dallalre, G. 1979. Toxics in the N.J.
Environment: Microcosm of U.S. Ills.
Civ. Eng. (Am. Soc. C1v. Eng.)
49 (9)-.74-80.
Dawes, I.W., and I.W. Sutherland. 1976.
Microbial Physiology. John Wiley and
Son, New York.
Dev, H., J.E. Bridges, G.C. Stesty, and
C.J. Rogers. 1985. In-Situ Decontami-
nation of Spills and Landfills by Radio
Frequency Heating. Paper presented at
the American Chemical Society National
Meeting, Chicago, Illinois, September.
Edwards, B.H., J.N. Paullin, and K.
Coghlan-Jordan. 1982. Emerging Tech-
nologies for the Control of Hazardous
Wastes. EPA-600/2-82-011, NTIS PB 82-
236993, U.S. Environmental Protection
Agency, Cincinnati, Ohio. 145 pp.
Electrical World. 1978. One Solution to
the PCB-Disposal Dilemma. Elect. World
190(9):52-53.
Environment Canada. 1983. Destruction
Technologies for Polychlorinated Bi-
phenyls (PCBs). Economic and Technical
Review Report EPS 3-EC-83-1, Environmen-
tal Protection Service, Ottawa, Ontario,
July (reprint). 81 pp.
Environment Canada. 1983. Proceedings of
the Technical Seminar on Chemical
Spills. Toronto, Canada, October 25-27,
291 pp.
Erler, T.J., J. Dragun, and D.R. Weider.
1983. Two Case Studies of Cost- Effec-
tive Remedial Actions for PCB Contami-
nated Soils. In: Proceedings of the
38th Industrial Waste Conference, Purdue
University, West Lafayette, Indiana, May
11-13, Ann Arbor Science, Ann Arbor,
Michigan, pp. 369-375.
Exner, J.H., ed. 1982. Detoxification of
Hazardous Waste. Ann Arbor Science
Publishers, Ann Arbor, Michigan. 362
pp.
Fedorko, J.W. 1982. Protection of the
Hosensak 230 kV and Three Mile Island
500 kV Capacitor Banks. Paper presented
at the Pennsylvania Electric Association
Fall Conference, Lancaster,
Pennsylvania, September 14.
FMC Corporation, (n.d.) 'Industrial Waste
Treatment with Hydrogen Peroxide,
Industrial Chemical Group, 2000 Market
Street, Philadelphia, Pennsylvania.
Fox, L. L., and N.J. Merrick. 1983.
Managing Polychlorinated Biphenyls in
the Industrial Environment—A Case His-
tory. In: Toxic and Hazardous Waste:
Proceedings of the 15th Mid-Atlantic
-352-
-------
Industrial Waste Conference, Lewisburg,
Pennsylvania, June 26-28, Butterworth
Publishers, Boston, Massachusetts, pp.
336-344.
Fradkin, L., and S. Barisas. 1982. Waste
Management Options for PCBs. In:
Industrial Waste: Proceedings of the
14th Mid-Atlantic Industrial Waste Con-
ference, College Park, Maryland, June
27-29, Ann Arbor Science, Ann Arbor,
Michigan, pp. 398-407.
Furakawa, K. 1982. Microbial degradation
of PCBs, pp. 33-57. A.M. Chakrabarty
(Ed.), Biodegradation and Detoxification
of Environmental Pollutants, CRC Press,
Inc. Boca Raton, Florida.
Ghassemi, M., and M. Haro. 1985. Hazard-
ous Waste Surface Impoundment Technolo-
gy, J. Envir. Eng. Ill, 5, October.
Gilmer, H., and F.J. Freestone. 1978.
Cleanup of an Oil and Mixed Chemical
Spill at Dittmer, Missouri, April-May
1977. In: Proceedings of the 1978
National Conference on Control of Haz-
ardous Material Spills, Miami Beach,
Florida, April 11-13, Information Trans-
fer, Rockville, Maryland, pp. 131-134.
Githens, G.D. 1984. Carbon Adsorption
Onsite for Remedial Actions. Pollut.
Eng. 16(l):22-25.
Gruenfeld, M., F. Freestone, and I.
Wilder. 1978. EPA's Mobile Lab and
Treatment System Responds to Hazardous
Spills. Ind. Water Eng. 15(5):19-23.
Hancher, C.W., J.M. Napier, and F.E.
Kosinski. 1984. Removal of PCB from
Oils and Soils. Preprint prepared for
submission to: 5th Department of Energy
Environmental Protection Conference,
November 6-8, Albuquerque, New Mexico,
DE85 002619. 7 pp.
Hand, T.D., and A.W. Ford. 1978. The
Feasibility of Dredging for Bottom
Recovery of Spills of Dense, Hazardous
Chemicals. In: Proceedings of the 1978
National Conference on Control of Haz-
ardous Material Spills, April 11-13,
Miami Beach, Florida, pp. 315-324.
Hawthorne, H.S. 1982. Solvent Decontami-
nation of PCB Electrical Equipment.
In: IEEE Conference Record of the
Industrial and Commercial Power Systems
Technical Conference, Philadelphia,
Pennsylvania, May 10-13, Institute of
Electrical and Electronics Engineers,
Inc., New York, New York, pp. 74-78.
Hedley, W.H., S.C. Cheng, B.O. Desai, C.S.
Smith, and H.D. Toy. 1983. Alternate
Treatment of Organic Solvents and
Sludges from Metal Finishing Operations:
Final Report. NTIS Publication No.
PB84-102151, EPA-600/52-83-094, U.S.
Environmental Protection Agency,
Cincinnati, Ohio. 363 pp.
Hennings, T.J., P.A. Painter, L.L. Scinto,
and A.M. Takata. 1982. Preliminary
Operations Plan and Guidelines for the
At-Sea Incineration of Liquid PCB
Wastes, NTIS Publication No. PB83-
181834, EPA-600/52-82-068, U.S. Environ-
mental Protection Agency, Research
Triangle Park, North Carolina. 121 pp.
Herbst, E., et al. 1977. Fate of PCBs-
14C in Sewage Treatment—laboratory
experiments with activated sludge.
Chemosphere 6:725-730.
Hetling, L.J., T.J. Tofflemire, E.G. Horn,
R. Thomas, and R. Mt. Pleasant. 1978.
The Hudson River PCB Problem: Manage-
ment Alternatives. Paper presented at
the New York Academy.of Sciences Confer-
ence on the Health Effects of Halo-
genated Aromatic Hydrocarbons, New York,
New York, June 24-27, p. 630.
Hornig, A.W. 1984. Destruction of PCB-
Contaminated Soils with a High-Tempera-
ture Fluid-Wall (HTFW) Reactor. In:
Proceedings of the 1984 Hazardous
Material Spills Conference: Prevention,
Behavior, Control, and Cleanup of Spills
and Waste Sites, Nashville, Tennessee,
April 9-12, Government Institute,
Rockville, Maryland, pp. 73-79.
Hornig, A.W. 1984. Destruction of PCB-
Contaminated Soils with a High-
Temperature Fluid-Wall (HTFW) Reactor..
NTIS PB84-168798, EPA-600/D-84-072,
U.S. Environmental Protection Agency,
Cincinnati, Ohio. 23 pp.
Industrial Economics, Inc. 1985. Regula-
tory Analysis of Proposed Restrictions
on Land Disposal of Certain Dioxin-
-353-
-------
Containing Wastes, Studies and Methods
Branch, EPA, Washington, D.C.
Isbister, J.D., G.L. Anspach, and J.F.
Kitchens. 1984. Composting for Degra-
dation of PCBs in Soil. In: Proceed-
ings of the 1984 Hazardous Materials
Spills Conference: Prevention,
Behavior, Control, and Cleanup of Spills
and Waste Sites, Nashville, Tennessee,
April 9-12, Government Institute,
Rockville, Maryland, pp. 104-109.
Jenkins, T.F., D.C. Leggett, L.V. Parker,
J.L. Oliphant, C.J. Martel, B.T. Foley,
and C.J. Deiner. 1983. Assessment of
the Treatability of Toxic Organics by
Overland Flow. CREL 83-3, U.S. Army
Corps of Engineers, Cold Regions
Research and Engineering Laboratory,
Hanover, New Hampshire.
Kalmaz, E.V., R.B. Craig, and G.W.
Zimmerman. 1981. Kinetics Model and
Simulation of the Concentration Varia-
tions of the Species of Polychlori-
nated Biphenyls (PCBs) Involved in
Photochemical Transformation. Preprint
Extended Abstract. Presented before the
Division of Environmental Chemistry at
the 182nd Meeting of the American Chemi-
cal Society, New York, New York, August
23-28, 21(2):100-103.
Kitchens, J.F., G.L. Anspach, L.B.
Mongoba, and E.A. Kobylinski. 1984.
Cleanup of Spilled Chlorinated Organics
with the LARC Process. In: Proceed-
ings of the 1984 Hazardous Materials
Spills Conference: Prevention,
Behavior, Control, and Cleanup of Waste
Sites, Nashville, Tennessee, April 9-12,
Government Institute, Rockville,
Maryland, pp. 110-115.
Kitchens, J.F., W.E. Jones, G.L. Anspach,
and D.C. Schubert. 1982. Light- acti-
vated Reduction of Chemicals for
Destruction of Polychlorinated Biphenyls
in Oil and Soil. In: Symposium on the
Detoxification of Hazardous Waste, Ann
Arbor Science, Ann Arbor, Michigan, pp.
215-226.
Kluge, F.R., P.M. Balma, W. Turiansky, and
R.V. Snow. 1982. PJM Joint Reactive
Project Installation of 230 kV Shunt
Capacitor Banks on the Public Service
Electric and Gas Company System. Paper
presented at the Pennsylvania Electric
Association Fall Conference, Lancaster,
Pennsylvania, September 14.
Kokoszka, L., and J. Flood. 1985. A
Guide to EPA-Approved PCB Disposal
Methods. Chem. Eng. pp. 41-43, July 8^
Komai, R.V. 1982. Current PCB Research.
Presented before the Edison Electric
Institute/Envirosphere Company 5th
Annual Conference on Environmental
Licensing and Regulatory Requirements
Affecting the Electric Utility Industry,
October 27-29, Washington, D.C. 13 pp.
Komai, R.Y., D.C. Van der Meer, J. Figler,
M.A. Bender, R.P. Kale, M.S. Makar, and
B.E. Pyatt. 1982. Current PCB
Research. Paper presented at the 5th
Conference of the Edison Electric
Institute, Envirosphere Environmental
. Licensing, and the Electric Industry,
Washington, DC, October 27-29.
Kong, H.L., and G.S. Sayler. 1983.
Degradation and total mineralization of
monohalogenated biphenyls in natural
sediment and mixed bacterial cultures,
Appl . Environ. Microbiol. 46:666-672.
Kopeoky, Anne L. 1985. PCB Biodegrada-
tion Using Sybron Bi-Chem lOOb PB/Hudson
River Isolates, Status Report. Sybron
Chemicals, Birmingham, N.J.
Lacy, W.J., and R.G. Rice. 1977. The
Status and Future of Ozone for Water -and
Wastewater Treatment. Ind. Water Eng.
Lafornara, J.P. 1978. Cleanup After
Spills of Toxic Substances. Water
Pollut. Control Fed. J. 50(4) :617-627.
Lahey, W., and M. Connor. 1983. The Case
for Ocean Waste Disposal. Techno!. Rev.
86(6):61-68.
Lapiere, R.B., et al . 1977. Catalytic
Hydrochlori nation of Polychlorinated
Pesticides and Related Substances.
Worcester Polytechnic Institute. Pre-
pared for Municipal Environmental
Research Lab. NTIS PB-262804.
Lauber, J.D. 1982. Burning Chemical
Wastes as Fuels in Cement Kilns. J. Air
Pollut. Control Assoc. 32(7) -.771-777.
-354-
-------
Law Engineering Testing Company. 1982.
Unsolicited Proposal for the Applica-
tion of Ground Freezing Technology to
Subsurface Contamination Control and
Clean-up. Law Engineering Proposal No.
2032.91.
Lee, M.C., R.A. Griffin, M.L. Miller, and
E.S.K. Chian. 1979. Adsorption of
Water-Soluble Polychlorinated Biphenyl
Aroclor 1242 and Used Capacitor Fluid by
Soil Materials and Coal Chars. J.
Environ. Sci. Health A14(5):415- 442.
Ludwigson, J., ed. 1984. Proceedings of
the 1984 Hazardous Material Spills Con-
ference: Prevention, Behavior, Control
and Cleanup of Spills and Waste Sites,
Nashville, Tennessee, April 9-12. 445
PP.
MaCallum, A. D. 1948. A Dry Synthesis of
Armoatic Sulfides: Phenylane Sulfide
Resins. J. Org. Chem. 13:154-160.
March, J. 1968. Advanced Organic Chemis-
try. McGraw-Hill, New York.
Massey, M.J., and F.M. Walsh. 1985. An
Electrochemical Process for Decontami-
nating PCB-Contaminated Transformer
Coolants, EPRI, PCB Seminar. Seattle,
Washington, October 22-25.
McCormick, D. 1985. One Bag's Meat.
Biotechnology 3(May):429-435.
McGranaghan, M.F., and R.F. Gustin. 1982.
Transient Switching Studies for EHV
Shunt Capacitor Applications on the PJM
System. Paper presented at the
Pennsylvania Electric Association Fall
Conference, Lancaster, Pennsylvania,
September 14.
McGrath, M.F. 1980. The Fourth Coast.
EPA Journal. 6(5):22-24. Washington,
DC
Mclntyre, A.E., R. Perry, and J.N. Lester.
1981. The Behavior of Polychlorinated
Biphenyls and Organochlorine Insecti-
cides in Primary Mechanical Wastewater
Treatment. Environ. Pollut. (Series B)
21(3):223-233.
McManamon, K. 1982. Sunohio's PCBX Proc-
ess. In: Minutes of the Meeting of the
Pennsylvania Electric Association,
Engineering Section Electrical Equipment
Committee, Fall 1982, Lancaster,
Pennsylvania, September 14-15,
Pennsylvania Electric Association,
Harrisburg, Pennsylvania. 9 pp.
Mercer, B.W., G.W. Dawson, J.A. McNeese,
and E.G. Baker. 1984. Methods/ Materi-
als Matrix of Ultimate Disposal Tech-
niques for Spilled Hazardous Materials.
NTIS Publication No. PB85-116853, EPA-
600/2-84-170, U.S. Environmental Protec-
tion Agency, Cincinnati, Ohio. 130 pp.
Merrick, N.J., L.L. Fox and D.M. Coker.
1983. Polychlorinated Biphenyl Removal
by a Combined Industrial Sanitary Treat-
ment Plant. In: Toxic and Hazardous
Waste: Proceedings of the 15th Mid-
Atlantic Industrial Waste Conference,
Lewisburg, Pennsylvania, June 26-28,
Butterworth Publishers, Boston,
Massachusetts, pp. 402-415.
Meuser, J.M., and W.C. Weimer. 1982.
Amine-Enhanced Photodegradation of Poly-
chlorinated Biphenyls, Final Report.
EPRI-CS-2513. Batelle Pacific Northwest
Labs, Rich!and, Washington.
Millan, R.C., and N.A.-Ostenso. 1984. A
Case Study of Storage and Disposal of
PCB Contaminated Soils. In: Proceed-
ings of the Conference on Hazardous
Wastes and Environmental Emergencies:
Management, Prevention, Cleanup, and
Control. Hazardous Materials Control
Research Institute, Silver Spring,
Maryland, pp. 250-254.
Miller, R.A., and M.D. Sevientoniewski.
(n.d) The Destruction of Various Organ-
ic Substances by a Catalyzed Wet Oxida-
tion Process, Report for EPA, Contract
68-03-2568. Work Directive T-7016.
Miller, R.A., and R.D. Fox. 1982.
Catalyzed Wet Oxidation of Hazardous
Wastes from Detoxification of Hazardous
Wastes. J.H. Exner, ed., Ann Arbor
Science.
Modell, M., G.G. Gandet, M. Simson, G.T.
Hong, and K. Blemen, 1982. Super-
critical Water, Testing Reveals New
Process Hold Promise, Solid Waste
Management, August.
-355-
-------
Honlck, B., and A. Blake. 1983. New
Industrial Wastewater Treatment Method
for Removal of Multiple Contaminants.
In: Proceedings of the 70th American
Electroplaters Society Annual Technical
Conference, Indianapolis, Indiana, June
27-30, American Electroplaters Society,
Winter Park, Florida, 13 pp.
Muller, B.W., A.R. Brodd, and J.P. Leo.
1983. Hazardous Waste Remedial Action-
Picillo Farm, Coventry, Rhode Island;'An
Overview. J. Hazard. Mater. 7(2):113-
129.
Murakami, K., and K. Takeishi. 1976.
Behavior of Heavy Metals and PCB's in
the Removal and Treatment Operations of
Bottom Deposits. Paper presented at the
Toxic Substances 2nd U.S./Japanese
Experts Meeting, October 25-29.
National Institute for Occupational Safety
and Health. 1977. NIOSH Criteria for a
Recommended Standard Occupational Expo-
sure to Polychlorinated Biphenyls
(PCBs). DHEW (NIOSH) Publication No.
77-225, U.S. Department of Health, Edu-
cation, and Welfare, Washington, DC.
224 pp.
New England River Basins Commission.
1980. Housatonic River Basin Overview.
NTIS Publication No. PB82-107152, Water
Resources Council, Washington, D.C. 209
pp.
New England River Basins Commission.
1980. Housatonic River Basin Overview.
NTIS PB82-107152. Prepared for the
Water Resources Council, Washington,
D.C. 199 pp.
New York State Department of Environmental
Conservation. 1985a. Status Report:
PCB Biodegradation Using Sybron Bi-Chem
1006 PB/Hudson River Isolates. Office
of Special Projects, Albany, New York,
May 21. 2 pp.
New York State Department of Environmental
Conservation. 1985b. Status Report:
Wright-Malta Steam Gasification Process.
Office of Special Projects, Albany, New
York, May 17. 2 pp.
New York University. 1984. Development
and Evaluation of a Low Energy Process
Technology for Extraction and Chemical
Destruction of Polychlorinated Biphenyls
(PCB's) from Contaminated Soils and
Sludges. -NYU/DASP 84-02. New York,
New York. 30 pp.
Oeschlaeger, H.F. 1976. Reactions of
Ozone with Organic Compounds, Proc.
Ozone Conference, Cincinnati, Ohio.
Pavlov, S.P., and W. Horn. 1978. PCB
Removal from the Duwamish River Estuary:
Implications to the Management Alterna-
tive for the Hudson River PCB Cleanup.
Paper presented at the New York Academy
of Sciences Conference on the Health
Effects of Halogenated Aromatic Hydro-
carbons, New York, New York, June 24-27,
p. 651.
Peterson, R.C., A.S. Batolomeo, M.H.
Corbin, and F. Roy. 1983. Lehigh
Electric Site Superfund PCB Cleanup:
Case History. In: 1983 National Con-
ference on Environmental Engineering:
Proceedings of the ASCE Specialty Con-
ference, Boulder, Colorado, July 6-8,
American Society of Chemical Engineers,
New York, New York, pp. 766-774.
Peterson, S.A., and K.K. Randolph, eds.
1977. Management of Bottom Sediments
Containing Toxic Substances: Proceed-
ings of the Second U.S.-Japan Experts
Meeting, October 1976, Tokyo, Japan.
NTIS PB-272684, EPA-600/3-77-083, U.S.
Environmental Protection Agency,
Con/all is, Oregon. 303 pp.
Peterson, R., and E. Milicic. 1985.
Chemical Destruction/Detoxification of
Chlorinated Dioxins in Contaminated
Soils, Summary Report (10-84—5-85,
sponsored by EPA and Air Force E.56),
Gal son Research Corporation, E.
Syracuse, New York.
Plimmer, J.R. 1978. Approaches to Decon-
tamination or Disposal of Pesticides:
Photodecomposition. ACS Symposium
Series, No. 73, American Chemical
Society, Washington, DC, p. 13.
Prengle, H.W., and C.E. Mauk. 1978. New
technology: Ozone/UV chemical oxidation
wastewater process for metal complexes,
organics, and disinfection, AIChE Symp.
Series 7:228-243.
-356-
-------
Rhee, G-Y. 1985a. Anaerobic Biodegrada-
tion of PCBs in Hudson River Sediments
and Dredged Sediment Disposal Sites.
Proposal submitted to the New York State
Department of Environmental Conserva-
tion, Albany, New York. 13 pp.
Rhee, G-Y. 1985b. Letter to R. Mt.
Pleasant, May 16.
Rhee, G-Y, M. Chen, and B. Bush. 1985.
Anaerobic Biodegradation of PCBs.
Proposal submitted to the New York State
Department of Environmental Conserva-
tion, Albany, New York. 17 pp.
Rochkind, M.L., G.S. Saylor, and J.W.
Blackburn. 1985. Microbial Decomposi-
tion of Chlorinated Hydrocarbons, EPA
Contract 68-03-3074-3.
Rogers, C. J., and A. Kernel. 1985.
Laboratory and Field Tests of Chemical
Reagents for In Situ Dextoxification of
Chlorinated Dioxins fn Soils. Paper
presented at the American Chemical
Society National Meeting, Chicago,
Illinois, September.
Rogers, C.J. 1983. Chemical Treatment of
PCBs.in the Environment In: Proceedings
of the 8th Annual Research Symposium.
Fort Mitchell, Kentucky, March 8-10,
1982, EPA-600/9-83-003, U.S. Environ-
mental Protection Agency, pp. 197-201.
Rogers, C.J. 1985. Letter to J.B. White,
July 19.
Ruzz, B., M. Iqbal, W. Brenner, and C.J.
Rogers. 1985. A Low Energy Process for
the Extraction and Chemical Destruction
of PCBs from Contaminated Soils and
Sludges. Paper presented at the
American Chemical Society National Meet-
ing, Chicago, Illinois, September.
Saunders, M.B. 1985. Pilot Studies for
Solvent Extraction of PCB from Soil,
Proceedings of EPRI PCB Seminar,
Seattle, Washington. October 22-25.
Schact, R.A. 1974. Pesticides in the
Illinois Waters of Lake Michigan.- NTIS
PB-245150, EPA 600/3-74-002, U.S..
Environmental Protection Agency.
Washington, D.C. 55 pp.
Schiaris, M.P. , and G.S. Saylor. 1982.
Biotransformation of PCB by Natural
Assemblages of Freshwater Microorgan-
isms. Environ. Sci . & Tech., 16:367-
369.
Scholz, R., and J. Milanowski. 1984. EPA
Project Summary: Mobile System for
Extracting Spilled Hazardous Materials
from Excavated Soils. J. Hazard. Mater.
91(2):241-252.
Schwinn, D.E., D.F. Storrier, R.J. Moore,
and W.S. Carter. 1984. PCB Removal by
Carbon Adsorption. Pollut. Eng.
Shen, T.T. 1982. Estimation of Organic
Compound Emissions from Waste Lagoons.
J. Air Pollut. Control Assoc. 32(1): 79-
82.
Sloan, K.J., et al . 1983. Temporal
Trends Toward Stability of Hudson River
PCB Decontamination, Bull. Envir.
Contam. Toxicol., 31:377-385.
Smith, J.G., and L.B. Gurbacham. 1981.
The Use of Sodium Naphthalenide to
Chemically Destroy Polychlorinated Bi-
phenyl by Dechlori nation. Preprint
Extended Abstract. Presented before the
Division of Environmental Chemistry at
the 182nd Meeting of the American Chemi-
cal Society, New York, New York, August
23-28, 21(2):88-89.
Snell Environmental Group, Inc. 1982.
Rate of Biodegradation of Toxic Organic
Compounds While in Contact with Organics
Which Are Actively Composting: 1982
Final Report: NTIS Publication No.
PB84-193150, National Science Founda-
tion, Washington, DC.
Sonksen, M.K., and J.A. Lease. 1983.
Evaluation of Cement Dust Stabilization
of Polychlorinated Biphenyl -Contaminated
Sludges. In: Proceedings of the 37th
Industrial Waste Conference, Purdue
University, West Lafayette, Indiana, May
11-13, Ann Arbor Science, Ann Arbor,
Michigan, pp. 405-412.
Spittler, T.M. 1984. Field Measurement
of Polychlorinated Biphenyls in Soil and
Sediment Using a Portable Gas Chromato-
graph. ACS Symposium Series, American
Chemical Society, Washington, DC,
267:37-42.
-357-
-------
Star, R. 1977. The Development of Pollu-
tion Control in Japan. IV. American and
Japanese Controls of Polychlorinated
Blphenyls (PCBs). Harv. Environ. Law
Rev. 1:561-567.
Strek, H.J. 1980. Factors Affecting the
Bloavailability of Polychlorinated Bi-
phenyls (PCBs) in Soils. NTIS PB81-
209223, Prepared for the Office of Water
Research and Technology, Washington,
D.C. 109 pp.
Strek, H.J., J.B. Weber, P.J. Shea, E.
Mrozek, and M.R. Overcash. 1981.
Reduction of Polychlorinated Biphenyl
Toxicity and Uptake of Carbon-14 Activi-
ty by Plants Through the Use of Acti-
vated Carbon. J. Agric. Food Chem.
29(2):288-293.
Stretz, L.A., L.C. Borduin, W.E. Draper,
R.A. Koenig, and J.S. Vavruska. 1982.
Controlled Air Incineration of Hazardous
Chemical Waste at the Los Alamos Nation-
al Laboratory. In: Proceedings of the
1982 Symposium on Waste Management:
Waste Isolation in the U.S. and Else-
where, Technical Programs and Public
Communications. Vol. 1: General,
Tucson, Arizona, March 8-11, Arizona
Board of Regents, Tucson, Arizona, pp.
281-299.
Stroud, F.B., R.T. Wilkerson, and A.
Smith. 1978. Treatment and Stabiliza-
tion of PCB Contaminated Water and
Waste Oil: A Case Study. In: Proceed-
ings of the 1978 National Conference on
Control of Hazardous Material Spills,
Miami Beach, Florida, April 11-13, In-
formation Transfer, Rockville, Maryland,
pp. 135-144.
Subnamanian, R.V., and R. Mahalingam.
1977. Immobilization of Hazardous
Residuals by Encapsulation (Semi-Annual
Technical Report). NTIS PB-271410.
National Science Foundation, Washington,
D.C. 93 pp.
Sunohio. (n.d.) PCBX: Chemical Destruc-
tion of PCB's. Brochure. Sunohio, 1700
Gateway Boulevard, SE, Canton, Ohio
47707.
Sweeny, K.H., and J.R. Fischer. 1970.
Investigation of Means for Controlled
Self-Destruction of Pesticides. NTIS
Publication No. PB-198224. U.S.
Environmental Protection Agency,
Washington, DC. 131 pp.
Sworzen, E.M., and D.G. Ackerman. 1982.
Interim Guidelines for the Disposal/
Destruction of PCBs and PCB Items by
Non-thermal Methods. NTIS Publication
No. PB82-217498, EPA-600/2-82-069, U.S.
Environmental Protection Agency,
Washington, DC.
Taylor, D. 1984. Managing Organics in
Sludge Reuse Programs. BioCycle
25(6):20-22.
Thomason, T.B. and M. Model 1, 1984.
Supercritical Water Destruction of Aque-
ous Wastes, Hazardous Waste 1(4):453-
467.
Thorneloe, S. 1986. Land Treatment Data
Base, Memorandum to J. Durham, January
31, Environmental Protection Agency,
Office of Air Quality. Planning and
Standards.
Timmerman, C.L. 1985. In Situ Vitrifica-
tion of PCB-Contaminated Soils, Proceed-
ings of EPRI PCB Seminar, Seattle,
Washington, October 22-25.
Tittlebaum, M.E., et al. 1985. "State of
the Art on Stabilization of Hazardous
Organic Liquid Wastes and Sludges." CRC
Critical Reviews in Environmental
Control, Vol. 15, Issue 2, pp. 179-211.
Trump, J.G., K.A. Wright, A.J. Sinskey,
D.N. Shan, and R. Fernald. 1979.
Disinfection of Municipal Sludge and
Wastewater by Energized Electrons.
Report No. INIS-MF-6718, In: Proceed-
ings of the International Seminar on
Radioisotopes and Radiation Applied to
Environmental Protection, Sao Paulo,
Brazil, September 17.
Tucker, E.S., et alI. 1975. Activated
sludge primary biodegradation of PCB's.
Bull. Environ. Contam. Toxicol. 14:705-
713.
Tulp, M.T.M., R. Schmitz, and 0.
Hutzinger. 1978. The Bacterial
Metabolism of 4,4'-Dichlorobiphenyl and
Its Suppression by Alternative Carbon
Sources. Chemosphere 7(1):103-108.
-358-
-------
U.S. Army Corps of Engineers, Water
Resources Support Center. 1983.
Management of Bottom Sediments Contain-
ing Toxic Substances: Proceedings of
the 7th U.S./Japan Experts Meeting,
November 2-4, 1981, New York, New York.
Cont. Accession No, AI36740. 421 pp.
U.S. Environmental Protection Agency.
1980. Identification and Listing of
Hazardous Waste. 40 CFR 261, Final
regulation, Federal Register.
U.S. Environmental Protection Agency.
1980. Polychlorinated Biphenyls (PCBs)
in Manufacturing, Processing, Distribu-
tion in Commerce, and Use Prohibition.
40 CFR 761, Final regulation, Federal
Register.
U.S. Environmental Protection Agency.
1985. Assessment of Incinerators as a
Treatment Method for Liquid Organic
Hazardous Wastes: Summary and Conclu-
sions. Washington, DC, March. 5 pp.
U.S. Environmental Protection Agency, and
Oil Spill Control Association of
America. 1978. Control of Hazardous
Material Spills: Proceedings of the
1978 National Conference in Control of
Hazardous Material Spills, Miami Beach,
Florida, April 11-13. 458pp.
United States Patent Office. 1983.
Method and Apparatus for Treating Poly-
chlorinated Biphenyl Contaminated
Sludge. U.S. Patent No. 4,402,274,
Issued to W.C. Meenan and G.D. Sullivan,
September 6.
United States Patent Office. 1984a.
Destruction of Polychlorinated Biphenyls
and Other Hazardous Halogenated Hydro-
carbons, U.S. Patent No. 4,469,661,
Issued to C.E. Shaltz, September 4.
United States Patent Office. 1984b.
Methods for Decontaminating Soil. U.S.
Patent No. 4,447,541, Issued to R.L.
Peterson, May 8.
Unterman, R., et al. 1985. Microbial
Degradation of Polychlorinated Bi-
phenyls, Proceedings of Tenth.Conf. on
New Frontiers for Hazardous Waste
Management, Pittsburgh, Pennsylvania. .
September 15-18. EPA/600-9-85-025.
v. Low, E. 1983. Vorkommen und
Mikrobieller Umund Abbau von
aromatischen Polyzyklen im Boden und in
Siedlungsabfallen. Forum Stadte-Hygiene
34(5):263-267.
V. Meyer, W. C. 1986. A Study of Thiona-
tion Reactions for Use in Destruction of
Toxic Wastes, Initial Tests of Thiona-
tion Reaction Conditions with Priority
Pollutants. Fairview Industries, Inc.,
Middleton, Wisconsin July. • EPA Contract
No. 68-02-3992.
Versar, Inc. 1976. Assessment of Waste-
water Management, Treatment, Technology,
and Associated Costs for Abatement of
PCB Concentrations in Industrial Influ-
ents.
Versar, Inc. 1984. Work Plan for PCB
Analyses. Versar Proposal No. .84-989.
Submitted to.the New York State Depart-
ment of Environmental Conservation,
Albany, New York, June 28.
Walker, K.H.. 1977. The Great Lakes
Cleanup. Water: Sewage Works
124(11):85.
Weaver, G. 1982. PCB Pollution in the •
New Bedford, Massachusetts Area: A
Status Report. Massachusetts Office of
Coastal Zone Management, Report sub-
mitted to the Massachusetts New Bedford
PCB Task Force, June. 62 pp. .
Webber, M.D., H.D. Monteith, and D.G.M.
Corneau. 1981. Assessment of Heavy
Metals and PCB's at Selected Sludge
Application Sites in Ontario. Research
Report No. 109, Research Program for the
Abatement of Municipal Pollution Under
the Provisions of the Canada-Ontario
Agreement on Great Lakes Water Quality,
Environment Canada, Environmental Pro-
tection Service, Ottawa, Ontario. 27-
PP.
Weitzman, L. 1981. Treatment and
Destruction of PCB and PCB Contaminated
Materials. Preprint Extended Abstract.
Presented before the Division of
Environmental Chemistry at the 182nd
Meeting of the American Chemical
Society, New York, New York, August 23-
28, 21:42-44.
-359-
-------
Weltzman, L. 1984. Cleaning of PCB Con-
taminated Soils, Proceedings of 1983 PCB
Seminar, EPRI, Atlanta, Georgia.
December 6-8.
Weltzman, L. 1985. Solvent Cleaning of
PCB Contaminated Soils, Proceedings of
PCB Seminar, EPRI, Seattle, Washington,
October 22-25.
West, P.R., S.K. Chaudhary, and R.H.
Mitchell. 1983. Photodechlorination of
Polychlorinated Biphenyls Induced by
Hydroquinone in Basic Media. Presented
before the Division of Environmental
Chemistry at the 186th Meeting of the
American Chemical Society, Washington,
DC, August 28-September 2, 23(2):384-
387.
West, R.H., and P.G. Hatcher. 1980.
Polychlorinated Biphenyls in Sewage
Sludge and Sediments of the New York
Bight. Marine Pollut. B. (11):126-
129.
Wheeler, J. 1986. U.S. Army Corps of
Engineers, Buffalo District. Contact
Summary, Research Triangle Institute,
January 15.
WhHmore, F.C., and J.D. Barden. 1983. A
Study of PCB Destruction Efficiency and
Performance for a Coal-Fired Utility
Boiler. Vol. 1 and 2. NTIS Publication
Nos. PB84-110147 and PB84-110154, EPA-
600/52-83-101a and b, U.S. Environmental
Protection Agency, Research Triangle
Park, North Carolina. 79 pp.
Young, D.R., and T.C. Heeson. 1977.
Polychlorinated Biphenyls in the Near-
shore Marine Ecosystem off San Diego,
California. Report No. SCCWRP-109,
Prepared by the Southern California
Coastal Water Research Project for the
California Regional Water Quality Board,
San Diego, California. NTIS PB-283090,
21 pp.
Young, D.R., D. McDermott-Erlich, and T.C.
Heesoon. 1977. Sediments as Sources of
DDT and PCB. Marine Pollut. Bull.
8(ll):254-257.
Zimmie, T.F., and T.J. Tofflemire. 1978.
Maintenance Dredging and Toxic Sub-
stances. In: Proceedings of the 2nd
Conference on International Waterborne
Transport. ASCE Urban Transportation
Division Specialty Conference, New York,
New York, October 5-7, 1977, American
Society of Chemical Engineers, New York,
New York, pp. 704-719.
-360-
-------
MOBILE KPEG DESTRUCTION UNIT FOR PCBs, DIOXINS
AND FURANS IN CONTAMINATED WASTE
Charles J. Rogers, Alfred Kernel
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
Robert L. Peterson
Gal son Research Corporation
E. Syracuse, New York 13057
ABSTRACT
The presence of highly toxic and persistent chemicals.in liquids, soils, sediments,
and sludges in abandoned waste sites poses a threat to both public health and the environ-
ment. Incineration is frequently used to destroy highly hazardous wastes, however, when
operated under less than optimum combustion conditions acutely hazardous products includ-
ing polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs)
can be formed and emitted in the combustion products. Various biological, chemical, and
physical methods have been tested and have been demonstrated to be effective to varying
degrees in destroying halo-organics.
The U. S. Environmental Protection Agency (EPA) has supported research intramurally
and extramurally since 1980, to develop an alternative method for in-situ or on-site de-
struction of halogenated pollutants. Chemical reagents prepared from polyethylene glycols
and potassium .hydroxide fKPEGs) have been demonstrated under mild conditions (25°-140°C)
to dehalogenate PCDDs and PCDFs to less than 1 ppb of starting materials. The reaction
mechanism is nucleophilic substitution at an aromatic carbon.
Toxicological tests have established that arylpolyglycol by-products from KPEG reac-
tions are non-toxic. In July and August, 1986, a 2700 gallon KPEG reactor was used in
Butte, Montana and Kent, Washington to successfully destroy PCDDs and PCDFs (120 ppb - 200
ppm) in 17,000 gallons of liquid waste to non-detectable levels. A new 2 cubic yard KPEG
reactor designed to treat both liquids and soils will be field tested in 1987.
INTRODUCTION
The accumulation pf polychlorinated
biphenyls (PCBs) and polychlorinated
dibenzo-p-dioxins (PCDDs, "dioxins") in
soil, sediment, and living tissue is a
serious problem that has received consider-
able public attention in recent years. As
an example, there existed in 1982 an esti-
mated 415 commercial wood-preserving plants
in the U.S. generating daily a total of
nearly 5.5 million gallons of process waste
which contained toxic materials (1), The
pentachlorophenol (PCP) products purchased
by industry to treat wood normally contain
100-200 ppm heptachlorodibenzo-p-dioxin
(HpCDD) and 1000 to 25000 ppm octachloro-
dibenzo-p-dioxin (OCDD). Also, pre-
liminary investigations of dioxins showed
that in the use of PCP with these high
concentrations of HpCDD and OCDD, and with
the high temperatures and pressures in-
volved, there existed the possibility that
the more toxic tetrachlorinated dioxins
and furans might be formed. Of the dis-
posal options available, the majority of
the wood preserving plants practiced stor-
ing the waste water on site, relying on
evaporation to reduce liquid waste to
sludge.
-361-
-------
Conventionally, the "clean up" of such
contaminated sites usually involves land-
filling and is not really a permanent
solution but rather a transfer of a toxic
waste from one region to another. Land-
filling will be curtailed in 1988 under the
1984 RCRA amendments.
The chemical stability of PCDD, PCBs
and other haloorganics precludes their
destruction by,conventional refuse incin-
eration methods. Most municipal inciner-
ators cannot achieve the high temperatures
necessary to destroy these chemicals.
(•Jobile incinerator technology has been
developed for on-site treatment of waste,
however, early estimates are that on-site
incineration of toxic waste at a wood
preserving site will be costly.
CHEMICAL PROCESS
Chemical decontamination is an alter-
native to thermal processing or landfilling
of soils contaminated with PCDDs or other
aromatic halides such as chlorobenzenes or
PCBs. Chemical decontamination, like in-
cineration, involves changes to the chemi-
cal structure of the dioxin molecule.
While chlorinated dioxins are thermally
stable, they readily dechlorinate to water
soluble compounds under relatively mild
conditions of temperature and pressure.
For example, chlorinated dioxins in oil are
readily reduced to the parts per trillion
level within 15 minutes at 80°C by reacting
them with a compound which is not oil
soluble. In soils processing, the PCBs and
PCDDs are dechlorinated to a water soluble
form which will remain with the soil or be
contained in the reagent that is recovered
for reuse. Dechlorination also affects the
toxicity of the dioxin, with dioxins con-
taining fewer than three chlorine atoms
generally showing low toxicity (2).
The proposed mechanism for these reac-
tions is shown in the following example
using 2,3,7,8-tetrachlorodibenzo-p-dioxin:
PROCESS CHEMISTRY
ROH + KOH
Cl
glycol 400 (PEG 400) to form an alkoxide.
The alkoxide (R0~) reacts with one of the
chlorine atoms on the chlorinated dioxin to
produce an ether and the alkali metal salt.
This dechlorination may proceed to complete
dechlorination, although replacement of a
single chlorine is sufficient to make the
reaction products water soluble.
TOXICITY CONSIDERATIONS
A major concern in this type of pro-
cessing involves the toxicity of any re-
agents and/or reaction products which may
be left in the decontaminated matrix after
treatment. Some toxicity data on reagents
used in the process are shown in Table 1,
along with comparison values for sodium
chloride and 2,3,7,8-TCDD. The reagents
used in this process are some five times
less toxic than table salt, and roughly
six orders of magnitude less toxic than
2,3,7,8-TCDD, the dioxin isomer of major
concern. Polyethylene glycol 400 is an
FDA approved material for use in foods and
cosmetics.
Toxicity testing of the reacted aro-
matic halides were conducted in EPA's
Research Triangle Park and Duluth labora-
tories. Structural assessment of the theo-
retical toxicity of the reaction products
is favorable, i.e., the known reaction pro-
ducts would not be expected to show signif-
icant toxicity. Results of the Ames test
for mutagenicity are negative, i.e., the
reaction products do not demonstrate car-
cinogenic potential. Bioaccumulation tests
also produced negative results, which is
not surprising given the water solubility
of these reaction products (3,4).
In earlier LV$Q studies conducted by
the EPA Duluth laboratory and others, it
was demonstrated that 2,3,7,8-TCDD is
extremely toxic to fish, producing
mortality in carp at water concentrations
TABLE 1. TOXICITY OF REAGENTS AND
COMPARISON MATERIALS
Material
J-Dsn. Oral-rat(3)
An alkali metal hydroxide, usually potas-
sium hydroxide (KOH) is reacted with an
alcohol or glycol such as polyethylene
polyethylene glycol 400
sodium chloride
(comparison value)
2,3,7,8-TCDD
(comparison value)
27,500 mg/kg
3,000 mg/kg
0.022 mg/kg
-362-
-------
as low as 60 ppq (5,6,7,8). By analyzing
tissue residue levels of TCDD in these carp
and expressing the results as ug accumula-
tion dose/kg body weight, it was possible
to make a rough comparison between.these
results and mammalian studies investigating
TCDD. These comparisons indicate that carp
are approximately as sensitive to TCDD as
guinea pigs. (LD50 600 ng/kg body weight.)
In 1050 tests 5.8 mg of by-products
from the KPEG-TCDD reaction was injected
into a 40 liter aquarium containing 15
carp. This resulted in a nominal water
concentration of-2501 ppt of
reacted TCDD (3). In a previous study
conducted under similar conditions except
at a measured water concentration of 60 ppq
TCDD, carp showed gross pathological damage.
These damages included cranial deformation,
lateral line lesions, fin darkening, hemor-
rhages and ulceration of the ventral body
wall (5,6). The absence of mortality and
gross pathological damage in exposures to
KPEG-TCDD adducts concentrations more than
4000-fold greater than those in the previ-
ous study lead to the conclusion that
treatment of TCDD with KPEG changes it to
a non-toxic form or at least reduces its
toxicity by several orders of magnitude
(3). The non-toxic properties of KPEG
reaction products improves the prospects of
on-site treatment, delisting and disposal
of haloorganic contaminated materials.
FIELD TESTS
Studies were initiated in January 1986
to determine if KPEG could be used to treat
PCDD and PCDF contaminated oil at a wood
preserving industrial site near Butte,
Montana. The wood preserving site con-
tained approximately 9000 gallons of light
petroleum oil collected previously from
groundwater over a period of two years.
The oil contained 3.5% pentachlorophenol,
PCDD and PCDF homologs ranging from 422 ppb
of tetrachlorinated isomers to 83,923 ppb
of octachlorinated isomers. Because of the
presence of these highly toxic chlorinated
dioxins and furans, the oil could not be
transported off-site for incineration.
Bringing i.n and operating a mobile inciner-
ator for on-site destruction of contami-
nated oil was rejected because of high
costs.
In-April 1986, U.S. EPA Region 8
agreed, after review of laboratory data,
that the chemical process, based upon a
KPEG reagent, could be used to decontami-
nate on-site the PCDD/PCDF contaminated
oil.
Initial treatment experiments with
some of the PCP waste-contaminated samples
obtained from Butte, Montana were accom-
plished on January 9, 1986. For these
tests, two 500-ml samples were obtained,
one an oil waste and the other a contam-
inated soil. The procedures utilized for
the treatment of the oil samples with the
KPEG reagents are described in the
following:
0 Two aliquots of oil samples (24.94 grams)
were removed from the container and
accurately weighed into new, pre-cleaned
125-ml sample bottles fitted with Teflon-
lined caps. A reagent blank bottle was
simultaneously prepared.
0 Twenty-five (25) grams of the KPEG re-
agent were added, at ambient temperature,
to each of the three sample bottles
mentioned above.
0 The bottles were placed in sand baths
maintained at temperatures of 70°C and
100°C for a period of 8 hours and the
contents of each bottle were stirred with
a motor-driven stirring rod during this
entire period.
0 Accurately weighed aliquots (typically
1-5 grams) of the sample were removed
from the reactor at intervals of 15
minutes and placed in 125-ml flint glass
bottles. The destruction reaction which
had been occurring was quenched by adding
a sufficient quantity of 50% 1^04 to
each bottle to adjust the pH to about
pH 7, as determined by indicator paper.
Following quenching of the treated
samples, the bottles were set aside until
test portions could be successfully extrac-
ted and analyzed (9).
The procedures for extraction and
analysis of PCDD and PCDF were based on
adding appropriate internal standards to
each aliquot taken from the reaction. At a
minimum, 1-2 ng_each of 2,3,7,8-[13Cs]-
TCDD, 2,3,7,8-[':i7Cl4]-TCDD and 2,3,7,8-TCDF
were added to each 125-ml flint glass
bottle containing 1-5 grams of the treated
oil sample. Sample extracts were prepared
and were analyzed using DB-5 capillary GC
columns to obtain data on the concentra-
-363-
-------
tions of PCDDs and PCDFs. Perkin-Elmer
Sigma II or Varian 3740 GCs were used for
the analysis. The Kratos MS-30 Mass
Spectrometer was used in the low resolution
mode to detect PCDDs and PCDFs. If PCDDs
and PCDFs were detected, a second portion
of the sample extract was analyzed using
high resolution MS (9).
RESULTS
The oil samples obtained from the
Butte, Montana site were treated as de-
scribed with KPEG and were analyzed by GC
MS to determine if PCDDs and PCDFs had been
destroyed, as required by the regulation,
to less than 1 ppb. Table 2 shows that
many of the congeners of PCDD and PCDF were
destroyed effectively by the KPEG with a
temperature as low as 70°C for 15 minutes.
All PCDD and PCDF congeners were completely
destroyed by KPEG when the sample was
heated to 100°C for 30 minutes. As a
result of the successful laboratory tests,
arrangements were made to lease equipment
to treat the remaining oil stored on the
site (10).
The field equipment used to process
the oil consisted of a 2700-gallon reactor
mounted on a 45-foot trailer equipped with
a boiler, cooling system, a laboratory and
a control room. Heating of the oil and re-
agent was achieved by recirculating the oil
and KPEG via a pump and high shear mixer,
through a shell and tube heat exchanger.
The heat transfer fluid on the shell side
of the heat exchanger was heated using a
boiler or cooled through a series of radi-
ator type air coolers. The PCDD- and PCDF-
contaminated oil was treated in five
batches, each consisting of 1400 to 2000
gallons, requiring 600 gallons of KPEG
reagent per batch. The mixture was heated
to 150°C and allowed to react for one and
one half hours before cooling. The treated
oil was pumped out of the reactor and into
a holding tank to confirm by subsequent
analysis that PCDDs and PCDFs in the oil
had been destroyed below detection limits.
The data from the sampling and analy-
sis of each tank revealed that all PCDDs
and PCDFs were destroyed to below the 1 ppb
detection limit. The treatment cost was
only 10% of the projected incineration
cost.
In August 1986 the equipment was
transported to Kent, Washington at the
request of U.S. EPA Region 10 and was
used successfully to treat approximately
8000 gallons of 2,3,7,8,-TCDD-contaminated
liquid waste discovered on an industrial
hazardous waste disposal site.
Currently, a new, improved 1 to 2-
cubic yard reactor, designed to treat both
liquids and solids, is being constructed.
In early 1987, field verification studies
will be conducted with this new KPEG re-
actor on four selected PCDD/PCDF- and PCB-
contaminated sites. In 1987, it is
expected that the KPEG process can be
demonstrated to be cost-effective when used
at wood preserving and Superfund sites to
TABLE 2. LABORATORY-SCALE TREATMENT OF OIL
Contaminants Concentration in
CDD/CDF Untreated Oil (ppb)
TCDD (2,3,7,8-)
TCDD (total)
PeCDD
HxCDD
TCDF (2,3,7,8-)
TCDF (total)
PeCDF
HxCDF
HpCDF
OCDF
28.2
422
822
2982
23.1
147
504
3918
5404
6230
Concentration in
Treated Residue (ppb) *MDC
70°C, 15 min. 100°C, 30 min.
-
"
—
12.1
33.3
4.91
5.84
0.65
0.37
0 71
0 I -3
0.28
0.35
0 36
0.76
1.06
2.62
*Minimum detectable concentration in parts per billion.
-364-
-------
destroy toxic halogenated pollutants in
liquids, sediments, sludges and soils.
ACKNOWLEDGEMENTS
Dr. Thomas 0. Tiernan and Staff, Brehm
Laboratory, Wright State University, in
cooperation with EPA HWERL, treated the
oil samples and provided the analytical
services under a U.S. EPA contract (Viar
Co. SAS NO 1724-X).
The testing of KPEG to destroy PCDDs
and PCDFs was co-funded by the Department
of the U.S. Air Force, Headquarters Air
Force Engineering and Services Center
Tyndall Air Force Base, FL. Captain Edward
Heyse was the U.S.A.F. Project Officer.
REFERENCES
1. Wilson, D., Summary Status of The Wood
Preserving Industry, EPA Food and Wood
Products Branch Internal Report, 1982.
2. Esposito, et al., EPA 600/2-80-197
p. 187.
3. Cook, P. et al. Bioaccumulation Tests
of Detoxified By-products. Internal
Report, Dec. 1986.
4. DeMarini, D. et al. Bioassay Testing
of Detoxified By-products. Internal
Report, Dec. 1986.
5. Cook, P. M., D. W. Kuehl and A. R.
Batterman. Presented at SETAC Annual
Meeting, November 2-5, 1986. Radisson
Mark Plaza Hotel, Alexandria, VA.
6. Johnson, R. D., D. B. Lothenbach and
A. R. Batterman. Presented at SETAC
Annual Meeting, November 2-5, 1986.
Radisson Mark Plaza Hotel, Alexandria
VA. ,
7. Mehrle, P. M., D. R. Buckler, E. E.
Little, J. D. Petty, D. L. Stalling,
G. M. DeGreeve, J. M. Coyle and W. J.
Adams. Presented at SETAC Annual
Meeting, November 2-5, 1986. Radisson
Mark Plaza Hotel, Alexandria, VA.
8. Kuehl, D. W., et al. Bioavailability
of Polychlorinated Dibenzo-p-Dioxins
and Dibenzofurans from Contaminated
Wisconsin River Sediment to Carp.
Submitted to Chemosphere.
9. Tiernan, T., et al., Wright State
University, Dayton, Ohio, Interim.
Report on the Assessment of Chemical
Reagents for Destruction of Higher
PCDD/PCDFs. Internal Report prepared
for U.S. EPA, April 15, 1986.
10. Peterson, R., Potassium Polyethylene
Gycol Treatment of PCDD/PCDF - Con-
taminated Oil In Butte, Montana. IT
Corp./Galson Research Corp., Proiect
#86-706, July 1986. J
-365-
-------
SUPERCRITICAL SOLVENT EXTRACTION
Charles A. Eckert, Gregory W. Leman,
Joan F. Brennecke, and Steven R. Alferi
Department of Chemical Engineering
University of Illinois, Urbana, IL, 61801
ABSTRACT
technology to problems in environmental control.
This paper first discusses the technical advantages of supercriticalfluid
significantly less than that for traditional methods.
INTRODUCTION
Supercritical fluids (SCF's) are the
medium for an exciting new separation
technology which has the potential to
provide a much safer and less expensive
alternative for the detoxification of
materials at waste sites and the delist-
ing of industrial adsorbants. In recent
years there have been a wide variety of
successful new applications of SCF tech-
nology which demonstrate the technical
feasibility as well as the cost-
effectiveness of such processes. In
this paper we discuss the theoretical
basis of SCF processing and report on
design and economic studies made at the
University of Illinois aimed at the
development of mobile units for waste
treatment using SCF separations, in a
manner to optimize safety, economics,
and public acceptability.
A SCF is a fluid that has been
heated and compressed beyond its criti-
cal temperature and pressure, existing
as a single phase with some unique prop-
erties. It is quite dense, almost as
dense as a liquid, which gives a large
capacity for solutes; it has a high
molecular diffusivity and low viscosity,
which make it an ideal medium for effi-
cient mass transfer; and finally it has
and unusually high compressibility,
which permits large density changes with
very small pressure changes, thus yield-
ing extraordinary selectivity character-
istics.
The solubility of heavy organic
solutes in SCF's is frequently many
orders of magnitude greater than the
solubility in an ideal gas. The ratio
-356-
-------
of these two solubilities is defined as
thg enhancement factor and values of
10 '-10 are quite common. In addition,
we have found that a one phase mixture
of a pure SCF and a properly-chosen
chemical "entrainer" greatly increases
the solubility of solid solutes and/or
enhances the separation of two solutes.
An "entrainer" is a small percentage
(1-5%) of an additional solvent added to
the SCF, such as acetone, alcohol, or
water added to O>2. This increase in
solubility and separation is due to spe-
cific chemical interactions between the
entrainer and the solute (e.g. Lewis
acid/base interaction or hydrogen bond-
ing) . As an example of the effect that
is possible, the increase in selectivity
of the removal of acridine from an acri-
dine/anthracene mixture when 1%
entrainer is added to supercritical CO
is shown in Figure I. 2
Different entrainers in different
supercritical solvents will exhibit var-
ying degrees of solubility and separa-
tion ability. A careful choice of the
solvent and entrainer permits the
exploitation of the chemical interac-
tions with the solutes to be separated
to create an optimum solvent mixture or
"tailor" a solvent for a process. For
example, the removal of trichlorophenol
from contaiminated soil might be greatly
increased by the addition of a small
amount of amine as an entrainer. The
specific chemical interaction between
the acidic trichlorophenol and basic
amine is likely to greatly increase the
solubility. Since the equilibrium or
maximum concentration of the solute in
the SCF will ultimately determine the
economic feasibility of SC extraction
processes we recognize the importance of
tuning solvent properties and composi-
tions to maximize solubilities.
SCF technology has been widely used
in many fields. In the oil industry
"miscible flooding" with supercritical
C02 has Ion9 been a highly successful
technique in tertiary oil recovery.
Also the Kerr-McGee prizewinning ROSE
process uses supercritical pentane to
permit the processing of heavier and
dirtier crudestocks. In the food busi-
ness supercritical CO has been used for
a variety of processes, including coffee
decaffeination, flavor recovery, removal
of cholesterol from eggs, defatting of
foods, and others. There are many other
applications now in practice, such as
the extraction of thermally labile mate-
rials from natural products, separation
of ethanol from fermentation beers, coal
desulfurization and liquefaction, ana-
lytical techniques such as SCF chroma-
tography, and more. Thus there exists
copious proof of the technical feasibil-
ity and economic advantages of SCF pro-
cessing.
Many of the above applications were
originally developed empirically, and
were limited due to the paucity of
experimental data and scientific under-
standing of the SCF state that existed a
few years ago. However recently a num-
ber of studies have been carried out
which have given us not only a better
physical understanding of the SCF state,
but the data to develop valid and
reliable mathematical models from which
one may do economic feasibility studies
as well as designs and scaleups [1-10].
In the work reported here we have relied
heavily upon these recent research
results, and they were essential to the
design process.
There are two major ways in which
waste materials are detoxified by SCF
extraction. The first is the direct
contact, or single step method, in which
contaminated material, solid or liquid,
is contacted directly with the SCF,
which extracts the toxic substances in
highly concentrated form for subsequent
treatment. This method is generally
used with solids and with liquid streams
with relatively high contaminant concen-
trations. For more dilute liquid
streams, a two-step method is generally
superior. In this the contaminant is
removed at ambient pressure and tempera-
ture by an adsorbant, for example granu-
lar activated carbon (GAG), and the
adsorbant is subsequently regenerated by
SCF extraction. Studies [11,12] have
demonstrated such regeneration to be far
superior to conventional thermal regen-
eration.
The types of environmental problems
that seem best solved by SCF technology
involve the removal of organics from the
soils and sludges found in the 546
National Priorities List waste sites,
the detoxification of leachates and
groundwaters, plus the regeneration of
-367-
-------
the GAG used in thousands of industrial
pollution control operations. In the
past much material has been disposed of
by burial, preceded by transportation to
a chosen site. Burial will not be a
viable alternative in the future, and
the transportation itself constitutes a
hazard. We and others have designed and
shown economic feasibility for station-
ary SCF toxic waste removal plants, but
these do not address both parts of the
problem. Thus, we have developed
designs and economic analyses for mobile
SCF separation units, capable of trans-
portation by highway trailer to waste
sites for rapid setup and safe and eco-
nomical operation on site.
CURRENT TECHNOLOGY
The methods in current use for
hazardous waste treatment—particularly
for the National Priority List Sites
—represent an iiranediate response to
very recently defined problems, but not
viable long-range solutions. Up to the
present there has been much emphasis on
hazard containment and volume reduction,
including the transportation of
abandoned wastes to approved containment
sites. Such procedures, however, are
not long-term solutions; not only does
this in no way lessen the total amount
of material which must be detoxified,
but substantial extra hazard, and
public-perception of hazard, is
introduced by the transportation
process. The current direction of
research therefore must be containment
and on-site treatment, developing
methods that can be applied economically
to a variety of wastes and which produce
treated materials suitable for reuse in
industry or for return to the
environment.
One of the methods which has been
applied to detoxification of contami-
nated solids is incineration. This
technology has been scaled for rela-
tively mobile operation (6-8 semitrail-
ers and several weeks for setup are
required), and it does produce residues
which can be buried on-site. The great-
est limitation, however, is the cost,
which has been variously estimated at
$50O-1000/yd .
The cost of incineration of GAG is
comparable. Thermal regeneration per-
mits reuse of GAC, but is expensive
itself, and involves a substantial loss
of capacity on each cycle. Thus the
treatment of groundwaters and leachates
with GAC is expensive by current tech-
nology. Various reports of GAC costs
run from $0.35-1.50/lb.
There also exists a class of prob-
lems involving sludges, with relatively
high concentrations of metals and
organic materials. The best disposal
method seems to be solidification for
burial. Vendor and literature informa-
tion compiled recently by the the EPA
[13] indicates that alcohols, aldehydes,
ketones, chlorinated hydrocarbons,
organic acids, acid chlorides, phenols
and other soluble organics tend to
increase set times for some mixes, and
cause consistently reduced durability in
the solidified product. Extraction of
the organic components cited is listed
in the EPA report as a removal method
which also has the advantages of volume
reduction and conversion to nonhazardous
material.
Thus current methods are improving
and can accomplish detoxification of
most types of waste, and mobile scale
design and testing is in progress for
some applications to allow on-site pro-
cessing. However, even where the tech-
nology is adequate, the costs for the
large-scale operations needed to impact
even a portion of the sites on the
National Priorities List alone is pro-
hibitive.
RESULTS OF ECONOMIC FEASIBILITY STUDIES
The scope of these investigations
encompasses the design and evaluation of
mobile SCF units for the treatment of
contaminated soils, organic removal from
sludges, and the regeneration of spent
GAC. Common features of the designs
include modularization for ease of
transportation, and include the produc-
tion of a concentrated toxin discharge
stream for chemical detoxification (such
as by dehydrohalogenation). Each system
consists of two essentially independent
segments: a continuous SCF flow loop to
control fluid density by step changes in
pressure, for continuous removal of
extracted organics, and a semi-batch or
continuous contactor.
-368-
-------
In the design and cost estimations
presented here, we have drawn on the
available thermodynaraic data and mathe-
matical models, as well as on the very
limited rate information available, and
where uncertainty existed, made rather
conservative contingency assumptions.
Thus we consider that the unit costs for
processing put forth here may have an
uncertainty as large as a factor of two.
Nonetheless, the SCF technology appears
to be so economically attractive that
even a doubling of the estimated costs
would represent very substantial savings
over current technology.
Soil Systems
Soil detoxification is the most com-
plicated of the three types of waste
treatment problems studied. It may
require site excavation to 10 feet or
more to achieve complete removal of con-
taminants; pretreatment is required to
screen large rocks and grind wood; and
varying moisture levels necessitate pro-
cess flexibility. Each factor leads to
more peripheral equipment, which exacer-
bates both the processing and mobility
problems.
Constraints for this design include
the limitation to a total of six semi-
trailers, including pretreatment, labo-
ratory, and control room facilities.
Other design bases include semi-batch
operation of the contactor units and a
set ratio of 8 Ibs. SCF/lb. soil treated
in the detoxification cycle {for pure
COg solvent). The latter condition
(possibly too conservative) is based on
the preliminary results reported by
Knopf et al. at Louisiana State Univer-
sity [14]. Under these conditions high-
way weight limits result in treatment
capacity of 50 yds /day (2.1 yds /hr.).
If necessary permits could be obtained
(as seems likely) for 2095 overweight on
one trailer, the capacity could be
expanded by 78% (3.7 yds /hr, or 89
yds /day). The design results are
summarized in Table I and Figure 2.
Sludge Systems
Organics removal from high-metal
sludges has the potential to be a vastly
simpler process if the sludge is pump-
able, so that completely continuous
countercurrent processing can be used.
Data are limited at this point, but it
appears that capital costs might be com-
parable to those for SCF soil process-
ing, and operating costs could be a bit
more favorable in a continuous process.
GAG Systems
The handling of leachates and
groundwaters is really just an applica-
tion of the GAG technology. These gen-
erally represent a situation where the
contaminant is dilute, and a two-step
separation is appropriate. Thus one
would envision an adsorption on GAC,
followed by SCF regeneration.
The regeneration of spent GAC is
essentially just a simplified version of
the soil treatment process because any
screening or grinding pretreatment oper-
ations are unnecessary. The solids are
more homogenous so less variations in
processing are required. They are also
more flowable, which facilitates easier
semi-batch or even continuous operation
of the separator modules. Finally the
volume to be treated is normally much
less. We have developed a design for a
mobile unit carried on just two trailers
with a regeration capacity of 500
Ibs./hr. (dry basis), and we have used
the same flow ratio that we have used
previously in stationary plants, 2O Ibs.
SCF/lb.dry GAC. Capital and operating
costs have been scaled from the designs
for soil detoxification, and the eco-
nomic summary is given in Table II.
To calculate return on investment we
have made the rather conservative
assumption that the total cost of
thermal regeneration and make-up GAC to
offset capacity reduction are at least
35c/lb. The minimum net gain is then at
least 20C/lb, since GAC regeneration by
SCF processing is estimated at 15$/lb.
CONCLUSIONS
From these economic studies, it is
easily concluded that supercritical
extraction offers an attractive alterna-
tive to several current hazardous mate-
rials treatment methods. However,
before construction of a full-scale
operational mobile detoxification unit
based on SCF technology, an intermediate
bench scale or pilot unit should be con-
-369-
-------
structed and operated on a batch or
semi-batch basis.
The following points should be taken
into consideration. The results of the
above research should be incorporated
into any design and the best system to
use for the initial study would be a
unit for regeneration of GAG. The unit
should be sized so that only one scal-
e-up step is needed, which means a
bench-scale unit processing 50-100
Ibs/day of GAC (dry basis). The unit
should be operated over a wide range of
operating conditions and the possibility
of a continuous contactor for flowable
GAC in contact with a supercritical
fluid should be investigated.
REFERENCES
1. K. P. Johnston and C. A. Eckert,
AIChE Journal 27, 418, 1981.
2. K. P. Johnston, D. H. Ziger, and C.
A. Eckert, Ind. Enq. Chem. Fundamen-
tals 21 191 (1982).
3. D. H. Ziger and C. A. Eckert, Ind.
Enq. Chem. Process Des. Develop. 22,
582 (1983).
4. C. A. Eckert, D. H. Ziger, K. P.
Johnston, and T. K. Ellison, Fluid
Phase Equilibria 14, 167 (1983).
5. C. A. Eckert, Materials Research
Society Svmp. Proc. 22, II, 81
(1984).
6. C. A. Eckert, J. G. Van Alsten, and
T. Stoicos, Env. Sci. Tech. 20, 319
(1986).
7. C. A. Eckert, D. H. Ziger, K. P.
Johnston, and S. Kim, J. Phvs. Chem.
86, 2738 (1986).
8. S. W. Gilbert and C. A. Eckert,
Fluid Phase Equilibria 30. 41
(1986).
9. C. A. Eckert, P. C. Hansen, and T.
K. Ellison, Fluid Phase Equilibria,
in press.
10. C. A. Eckert, B. S. Hess, and J. G.
Van Alsten, Fluid Phase Equilibria,
in press.
11. C. P. Eppig, R. P. deFilippi and R.
J. Robey, Project Sumnary,
EPA-6OO/2-82-O67, Washington, D. C.,
1981.
12. R. P. deFilippi and R. J. Robey,
Project Summary, EPA-600/52-83-038,
Washington, D. C., 1983.
13. JACA Corp., Technical Report, EPA
Contract No. 68-03-3186, Cincinnati,
OH, 1986.
14. F.. C. Knopf, B. Brady, and F. R.
Groves, CRC Crit. Rev. Environ.
Control 15, (3), 237 (1985).
15.0
12.5
10.0
01
V)
7.5
5.0
2.5
SOLVENT'
C02+I%CH3OH
SOLVENT* PURE C02
I
I
I
50 150 250 350
PRESSURE (BAR)
Fig. 1 Entrainer effect on
selectivity: The
supercritical fluid
extraction of acridinfe
from phenanthiene.
-370-
-------
o
•H
<*-! TJ
!«! 3 T3
O .-I W O
O -H 4J
§
-371-
-------
TABLE I
SCF EXTRACTION OF CONTAMINATED SOIL
Design Basis: 2.1 yd3/hr., 250 days/yr.
CAPITAL COSTS
EQUIPMENT
ENGINEERING AND SUPERVISION
CONSTRUCTION
CONTINGENCY
WORKING CAPITAL
TOTAL CAPITAL INVESTMENT
OPERATING COSTS
DIRECT COSTS:
Solvent
Labor
Direct Supervision
Utilities
Maintenance and Repairs
Operating Supplies
Laboratory Costs
FIXED CHARGES:
Depreciation
Taxes and Insurance
Plant Overhead
Administrative Costs
Financing
CONTINGENCY FOR TRANSPORTATION:
TOTAL PRODUCT COST:
UNIT COST = S98.74/YD3
1-45
0.15
O.15
0.20
0.25
2.20
MS/YR
10.8
225.0
30.0
170.4
110.0
11.0
30.0
587.2
1234.2
COST FOR MOBILE INCINERATOR $5OO/yd
ANNUAL PRETAX PROFIT S5.02MM
RETURN ON INVESTMENT 228*
PAYOUT TIME 5.3 MONTHS
TABLE II
SCF REGENERATION OF CONTAMINATED GAG
Design Basis: 500 Ibs/hr GAG, dry basis
10,000 Ibs/hr CO. recirculation
250 days/year operation
16 ft3)
CAPITAL COSTS:
Vessels (4
Main SCF Loop
Recovery System
Cooling Tower
Material Handling, Conveyors
Process Trailer
Analytical/Office Trailer
Direct Capital Investment
Engineering and Supervision
Construction Expense
Contingency
FIXED CAPITAL INVESTMENT
Working Capital
TOTAL CAPITAL INVESTMENT
216.0
213.4
65.4
6.0
15.8
15.0
100.0
, 634.6
50.0
50.0
50.0
784.6
100.Q
$884.6
OPERATING COSTS: M$
DIRECT COSTS:
Makeup CO 2.5
Labor (4 @ 25,000) 100.0
Supervision (% @ 30,000) 15.0
Utilities 39.5
Maintenance and Repairs 44.2
Operating Supplies 4.4
Laboratory Costs 30.0
235.6
FIXED CHARGES:
Depreciation 63.5
Taxes and Insurance 15.7
, Plant Overhead 31.8
Administrative 10.0
Financing 88.5
TOTAL PRODUCT COST 445.1
UNIT COST: 14.80/lb dry GAG
CURRENT TECHNOLOGY: Thermal regen-
eration costs at a minimum 35
-------
SUPERCRITICAL FLUID EXTRACTION AND CATALYTIC
OXIDATION OF TOXIC ORGANICS FROM SOILS
-Kerry M. Dooley1, Robert Gambrell2 and F. Carl Knopf1
Department of Chemical Engineering1
: , Department of Marine Science2
Louisiana State University
Baton Rouge, Louisiana 70803
**
ABSTRACT
Supercritical fluid (SCF) extraction is a promising new technique for the cleanup of
soils, sediments, and sludges that are contaminated with hazardous wastes. In this
investigation, supercritical carbon dioxide (SC-CO ) has been used to extract PCBs, DDT
and toxaphene from contaminated topsoils and subsoils. An attractive feature of this '
process is that the CO , being virtually inert, leaves no solvent residue on the
processed soil.
In our initial extraction studies, supercritical CO at 100 atm and 40°C was
continuously passed through a fixed bed of 10 g of soil. Approximately 70% of the DDT
and 75% of the toxaphene could be leached from a topsoil (12.6% organic matter)
contaminated with 1000 ppm DDT and 400 ppm toxaphene in under ten minutes using SC-CO at
a rate of 0.7 g/s. The extraction of contaminated (with 1000 ppm Aroclor 1254) subsoil "
(0.74% organic matter) proved to be even more promising, because more than 90% of the
PCBs could be extracted in under one minute at the same CO rate.
Recently SC-C02 with a single entrainer*, either methanol or toluene, was compared
to pure C02; comparison was made on the basis of extraction rate and the removal
efficiency for DDT or PCBs from contaminated topsoils. The supercritical mixtures at 100
atm and 40°C were continuously passed through a fixed bed'of 10 g of soil. The most
effective solvent system, SC-C02 with 5 wt% methanol at a flowrate of 0.7 g/s, was able
to leach 95% of the DDT from the soil in under 5 minutes, as compared to either pure CO
or C02 with 5 wt% toluene at the same conditions, which resulted in only 70% extraction2
in 10 minutes. This same extraction mixture (SC-CO with 5 wt% methanol) was also
applied to a highly contaminated spill site topsoil containing ca. 3500 ppm Aroclor 1260
and 2100 ppm Aroclor 1242.
minutes.
Over 98% extraction of the contaminants was realized in 10
With the demonstrated ability to extract contaminants from soils using supercritical
C02 with an entrainer, a logical further treatment would be the destruction of the wastes
while they are in the supercritical phase. An evaluation of catalysts for the low
temperature (below 350°C) oxidation of polycyclic aromatic hydrocarbons and chlorinated
hydrocarbon wastes in the SCF phase will be presented.
An entrainer is a volatile organic compound which, when added in low levels to super-
critical C02, dramatically increases the solubilities of certain nonvolatile orcanics in
the SCF.
JUA.
To whom correspondence should be addressed.
-373-
-------
INTRODUCTION
frustrated some applications.
Supercritical fluid (SCF) extrac-
tion has received much attention as a
technique for separating relatively non-
volatile materials (Paulaitis et al.,
1982). Typically in SCF extraction a
solvent gas such as carbon dioxide, at
high pressure and moderate temperature,
is contacted with a solid or liquid phase.
Slight changes in the system temperature
or pressure can cause large changes in
the solvent density and consequently in
its ability to solubilize relatively
nonvolatile components. For example, at
200 atm and 35°C the density of C02
approaches 0.8 g/cc and at these condi-
tions a solute such as naphthalene would
have a solubility some 10,000 times that
predicted if C0_ behaved as an ideal gas.
By taking advantage of these facts a
process can be envisioned whereby manipu-
lation of the system pressure effects
extraction of a nonvolatile material. A
pressure letdown, to a pressure below the
system critical conditions, can cause
near complete precipitation of the
relatively nonvolatile material from the
solvent. In addition to the liquid-like
densities of a typical SCF, viscosities
and molecular diffusivities of SCFs are
intermediate to typical liquid and gas
values for these properties. For these
reasons the extraction efficiencies of
SCFs are usually higher than those of
liquids.
The many advantageous properties of
SCFs have opened up new technologies in
environmental control; the subject has
been recently reviewed by Groves, Brady
and Knopf (1985). Successful efforts
have included the use of SCFs in: the
regeneration of adsorbents contaminated
with volatile organics using pure SC-C02
(Eppig et al., 1981); oxidation of
organic contaminants in waste streams
using SC-water (Modell, 1982); liquid-
liquid extraction of waste streams again
using pure CO- as the extraction medium
(Ringhand andTCopf ler, 1983); the extrac-
tion/reaction comprising the generation
of low sulfur chars from coals utilizing
either SC-toluene or SC-alcohols
(Vasilakos et al., 1985); and finally the
regeneration with SC-CO of activated
carbon used in the cleanup of liquid
waste streams (de Filippi et al., 1980a).
The regeneration of activated carbon was
not universally practical, because the
buildup of irreversibly bound organics
We are currently examining the
capabilities of supercritical fluids to
extract toxic chemicals such as DDT and
PCBs from soils, thereby providing a tool
for cleaning up hazardous waste sites.
Such removal offers the obvious advantage
of the creation of a much smaller volume
if further treatment such as combustion,
biological degradation or other disposal
method is desired.
In conventional extraction technology
the pressure would be reduced subsequent
to extraction, precipitating a solid or
liquid phase rich in organic contaminants
from the CO.. In order to recycle the
C0_, the contaminant level in the CO- must
often be further reduced by distillation
(deFilippi et al, 1980b). In addition,
since the pressure of the CO- has been
substantially lowered, recompression is
necessary.
In conjunction with our work with SC-
CO plus entrainers, we are investigating
an alternative to the above separation
scheme. We are attempting to totally
oxidize the extracted organic contaminants
by passing the high pressure mixture plus
air over total oxidation catalysts such as
CuO. Temperatures necessary for total
oxidation should be in the 420-600 K
range, this being characteristic of free-
radical assisted autoxidation processes
(Schuit and Gates, 1980). After drying
to remove the water formed in the
reactions, and neutralization, both of
which can be accomplished at high
pressure, the CO- can be recycled with
small recompression costs.
In order to explain the extraction
process, the thermodynamics involved, and
oxidation catalysis in a supercritical
fluid medium, this paper is divided into
two sections. The first discusses back-
ground, including early SCF extraction
results and the phase equilibria expected
for the ternary system of CO -methane1-DDT
(here methanol is an entrainer), and
briefly introduces both reactions in SCFs
and catalytic autoxidation of priority
pollutants. The second section presents
recent data on the effectiveness of CO-
with entrainers for the extraction of DDT
and PCBs from topsoils of high organic
content. Some preliminary data showing
the solid-catalyzed partial oxidation of a
pollutant (toluene)...in a SCF are also
-374-
-------
presented.
PAST RESULTS
In previous work (Brady et al.,
1987) we investigated the use of pure SC-
CO- to extract PCB's, DDT and toxaphene
from a contaminated topsoil (12.6%
organic matter) and subsoil (0.74%
organic matter). Experiments with SC-
CO- continuously flowing through a fixed
bea of 10 grams of contaminated soils
showed that over 90% of the PCBs (original
contamination level 1000 ppm) could be
extracted from the subsoil in under one
minute using a CO. flowrate of 0.7 g/s.
However, only 70% of the DDT and 75% of
the toxaphene could be leached from the
topsoil contaminated with 1000 ppm DDT
and 400 ppm toxaphene in 10 minutes at
the same C02 rate. These results are
not unexpected in that pure SC-CO-,
being a relatively nonpolar solvent,
would show only non-specific van der
Waals interactions with adsorbed
compounds that may be strongly bound to
polar sites on the topsoil. The
presence of these sites is consistent
with the high organic content of the
topsoil. Mixed solvents, for example CO-
and a few wt% of a polar entrainer, coula
result in specific chemical attraction of
adsorbate and extraction medium, and thus
enhance contaminant extraction from
solids containing polar adsorption sites,
such as activated carbon or topsoils of
high organic content; ^
PHASE DIAGRAMS
• The "entrainer" solvents used in
this work to date are toluene and
methanol. By using the Peng-Robinson
cubic equation of state (Peng and
Robinson, 1976), the triangular phase
diagram for the ternary system CO--DDT-
'nethanol at 40°C, 100 atm can be
constructed as shown in Figure 1.
It is assumed the solubility of the
solvent (CO with methanol) in solid DDT
(component 2) is negligible. The solid
phase is considered pure DDT, and its
fugacity f- in phase equilibria
calculations was evaluated from
assuming the Peng-Robinson equation is
valid for calculating the-fugacity of the
pure subcooled liquid (f° ). The critical
points are calculated according to the
algorithm proposed by Heideman and Khalil
(1980).
At 40°C and 100 atm, the addition of
a small amount of toluene or methanol to
.CO increases the DDT solubility. The
solubility of DDT in a mixed solvent of 95
wt% CO. and 5 wt% toluene is estimated as
2.5 x 10 (~3.6 times the value in pure
C02). If toluene is replaced by methanol
the estimated DDT solubility from Figure 1
is 0.01 (~14 times that in pure CO-).
Therefore a mixed solvent with 95 wt% CO-
and 5 wt% methanol should be more
effective in extracting DDT from soil than
a mixed solvent with 95 wt% C02 and 5 wt%
toluene, although the latter is still a
more effective solvent than pure CO-. The
approximate supercritical phase composi-
tions for the ternary system at the
operating conditions are summarized in
Table I. The interaction parameter k..
(k. . = 0.07) was empirically determine*! by
regressing VLB data for the binary system
of CO- plus methanol (Semenova et al.,
1979). The phases present ("conditions")
are denoted by either SFE (solid and SCF at
equilibrium), or SLVE (solid-liquid-vapor
phases at equilibrium).
Table I. Couposttions of Supercritical Fluid Phases fo;
C02-Entrainer-DDT-Syste«s
Systei
C02(l) •)• DDT(2) 0.07
DDT(2) + Methanol 0
C02 + DDT + Hethanol
y2 = 0.7 X 10~3 .-
y2 = 0.3 x io"3
X2 = 0.055 X3 = 0.132
y2 = 0.014 y3 = 0.079
SFE
SFE
SLVE
ritical pt
Vo1 - v°s
2 V2
CD
REACTIONS IN SUPERCRITICAL SOLVENTS
. Supercritical fluids have potential
advantages as solvents in coal lique-
faction, high temperature thermal.
condensations and alkylations, catalyzed
hydrocarbon isomerizations, and oxidation
reactions. The application of SGF
solvents improves the final product
distributions of these reactions,
sometimes by diminishing heat and mass
transport limitations, and sometimes by
allowing separation of desired products
prior to further reaction in a solid or
liquid phase. Currently, the most
-375-
-------
promising application of SCF as reaction
media is coal liquefaction/extraction,
where the condensed extract is a fuel or
chemical feedstock comparable to
hydrogen-intensive processes. These
findings have prompted a growing interest
in reactions using supercritical fluid
media, the subject has been recently
reviewed (Subramaniam and McHugh, 1986).
•^>
The utility of SCF,as reaction
media, apart from their utility in coal
processing, is documented. SCF have
proven feasible as reaction media for high
temperature deep oxidations (Modell,
1982), where SC water homogenizes heavy
organic-air-salt mixtures and thereby
enhances combustion efficiencies.
SCF have also proven feasible as
reaction media for hydrocarbon isomeriza-
tion processes, either with a homogeneous
or heterogeneous catalyst (Kramer and
Leder, 1975, Tiltscher et al., 1981). The
details of the isomerization processes
are sketchy but interesting. In the
homogeneous process, the SC paraffin
reactant solubilizes not only the most
efficient catalyst, AlBr3, but also H_,
which controls the selectivity of this
reaction by suppressing cracking rates *
Isoraerization-to-crackirig ratios of
better than 50 were obtained at optimum
reaction conditions.
Finally, Metzger et al. (1983) have
found that many compounds containing
carbonyls or double bonds will undergo
condensation and alkylation reactions at
temperatures not much above their
critical temperatures, in the 300-600K
range. At these conditions they
prepared, for example, the cyclohexane
addition products to acrylonitrile,
heptene-1, and methyl acrylate in high
yields. The intermediates in these
reactions were remarkable stable with
respect to polymerization.
In summary, it has been observed
that SCF provide favorable reaction media
in the following cases:
(1) Where homogenization of the
reaction mixture removes dif-
fusion limitations for a key
reactant, catalyst, or promoter;
(2) Where the SCF separate unstable
products of reaction from a
solid or liquid phase.
(3) Where the SCF play a direct
role in reaction, say, in the
stabilization .of free radicals
or other intermediates, or
assisting in the transfer of
hydrogen.
(4) Where the SCF aid in the
solubilization of reaction
products that could result in
catalyst deactivation.
TOTAL CATALYTIC OXIDATION OF PRIORITY
POLLUTANTS
The chemical behavior with respect to
catalytic total oxidation of the many
compounds listed as priority pollutants
(U.S. EPA, 1980b; Robertson et al., 1980)
is so diverse that not all compound
classes could be studied. This is true
even if we restrict ourselves to single
model compounds or well-characterized
mixtures of homologous compounds. Our
attention has been focused on three
classes of difficult-to-oxidize compounds:
(1) Single-ring aromatics with
deactivating side groups, such as
chlorobenzenes, chlorobenzidines, and
nitrophenols; (2) Fused-ring polycyclic
aromatic hydrocarbons (PAHs) such as
anthracene and benzanthrene; (3) Chlori-
nated biphenyls and naphthenic compounds
such as DDT, PCBs, and dieldrin. Our
catalytic oxidation Work is not performed
independently of the entrainer studies;
feeds to be oxidized include the
entrainer, which in some cases will be
oxidized as well. Entrainer oxidation is
often desirable, since in free radical
processes small molecules that can form
radicals easily (acetone, for example)
help initiate the more difficult free
radical oxidations of the refractory
compounds. Such oxidizable initiators are
known as co-oxidants and represent a more
economical solution to the initiation
problem than do such common initiators as
peroxides.
We have examined total oxidation by
free-radical-assisted redox processes,
which are well known for the partial
oxidations of alkylaromatic and naphthenic
hydrocarbons in aqueous or organic
solutions; the reaction mechanisms and
applicable catalysts are exhaustively
surveyed by Kaeding et al. (1970), Sheldon
and Kochi (1974), Schuit and Gates (1980),
and Parshall (1980). The catalysts are
-376-
-------
organometallic r|dox Couples containing,
for example, Co /Co .Peroxides and
hydroperoxides are the primary products,
which decompose to more stable products
at temperatures above about 80-100°C.
Typical pathways for a typical reaction,
the oxidation of toluene to benzaldehyde/
benzoic acid, are given in Fig. 2. In
the presence of acids, condensation takes
place concurrently with oxidation,
resulting in the production of phenols
and acetates. A typical pathway for
toluene Oxidation is given below:
species present in the feed, if the
amount of catalyst employed is large
enough. One way in which olefinic,
aromatic, and other double-bond
containing molecules can be totally
oxidized is through free-radical
epoxidation on a surface or even by
desorbed radicals in the fluid-phase
(Haber, 1985):
R0_- + R'C = CR" -> RO.
I
C - C - R" -» JO- + K' - C - C - S"
I o
H*
<—?(
The processes and catalysts common
to solution can be adapted to a dense ,gas
(supercritical) phase and certain solid
heterogeneous catalysts. We have
demonstrated this in a recently completed
study of the oxidation of toluene to *
benzaldehyde with feeds composed mostly
of supercritical CO and some air (Dooley
and Knopf, 1986). Not, all redox-
containing catalysts are effective; in
fact, in catalytic reaction experiments
employing a series of Co and Co-Mo oxides
supported on Al 0 2+onlv+one catalyst
with a specific Co /Co initial ratio
was both active for partial oxidation and
inactive for the undesired condensation
reactions to polycyclic products. For
this work the temperatures were in the
180-220°C range and the catalysts were
not modified with electron transfer
promoters (as they are in solution) to
enhance their activities.
At slightly higher temperatures and
with the inclusion of electron transfer
promoters in the solid catalyst,
peroxides and hydrogen peroxides are more
easily produced and then ruptured at the
0-0 bond. The rupture results in a net
increase in the number of free radicals
and therefore in a dramatic increase in
oxidation rates. Schuit and Gates (1980)
have shown that at these conditions the
total concentration of peroxides and
hydroperoxides is proportional to the
exponential of the product of the rate
of formation of all products times the
reactor residence time. Therefore once
initiated this process must result in
complete oxidation of all oxidizable
The reason why total oxidation is not
more prevalent in solution is that the
reaction conditions are carefully chosen
to inhibit this generally undesired
reaction: the solutions are 0- starved
due to the low solubilities of low-
pressure air in most organics, and the
solutions often contain polar compounds
(acetic acid, for example) which act as
free radical traps. Both of these limits
on total oxidation will be removed in the
proposed process. The upper limit on
free-radical-assisted total oxidation is
about 300-350°C; above these temperatures
the reverse rates of hydroperoxide
formation reactions are large enough that
oxidation must take place by other routes.
Simple oxides may not catalyze the
total oxidation of refractory aromatic and
naphthenic compounds. The ideal catalyst
must be able to adsorb the hydrocarbon by
electron' transfer to or from a metal
cation site, to insert electrophilic
oxygen to form aldehydes or epoxides, and
finally to decompose these to carbon
oxides.
EXPERIMENTAL METHODS FOR SCF SOIL
EXTRACTIONS
Two contaminated topsoils were
obtained for our extraction experiments.
The first topsoil was obtained from a DDT
spill site near Lake Providence,
Louisiana. The second topsoil was
obtained from a PCB spill site. The soils
were air-dried (35°C) and the initial DDT
or PCB levels checked via EPA-approved
test procedures (U.S. EPA, 1980; 1982).
A schematic diagram of the SC-CO.
extraction apparatus is given in Figure 3.
Liquid CO^ at ambient temperature is fed
-377-
-------
to a diaphragm compressor and compressed.
to a pressure between 200 and 350 atm.
The compressed C02 is stored in surge
tanks to dampen any pressure fluctua-
tions. From the surge tanks the C02
flows at ~0.7 g/s to a vertical tube
fixed bed containing the contaminated
soil; the pressure here is controlled to
±5 psi. Upstream of the vertical tube,
approximately 5 wt% of either toluene or
mcthanol are combined with the super-
critical C0_ by the use of a Ruska high-
pressure metering pump. In the tube the
fixed bed of contaminated soil is
contacted by the mixed solvent. The
extraction pressure is monitored by a
Heise digital pressure gauge. Both the
feed line and the fixed bed are immersed
in a constant temperature bath maintained
at 40 ± 1°C. Downstream of the bed two
micrometering valves are used to control
the solvent flow rate and reduce the
pressure to atmospheric. The extract
mixture is first passed through two dry
ice/acetone cold traps to collect the
precipitated contaminants, followed by an
activated carbon trap, and finally by a
dry test meter to totalize the solute-
free CO-. All wetted parts of the
apparatus are stainless steel, teflon, or
viton.
The residual concentration of DDT or
PCBs in the soil is measured according to
EPA-approved test procedures (U.S. EPA,
1980; 1982). Typically 5 g of soil are
weighed to ± 0.01 g into cellulose
extraction thimbles, which are placed in
a Soxhlet apparatus and extracted for 8
hours with a 60/40 mixture of acetone/
hexane. The acetone/hexane extracts are
then washed with water to remove acetone,
and the hexane portion is subjected to
florisil cleanup and finally diluted with
hexane as necessary for analysis by gas
chromatography.
SCF EXTRACTION OF DDT CONTAMINANTED
TOPSOIL - RESULTS AND DISCUSSION
The initial DDT level of the
contaminated topsoil was 1271 mg/kg.
Extractions of this test soil showed that
approximately 60-70% of the DDT could be
removed in approximately 10-20 minutes as
shown in Figure 4. Longer-time
extractions using pure SC-C02 did not
show any improvement over this reduction.
These data imply that a portion of the
DDT is strongly bound to the soil and
that SC-CO- at these conditions cannot
extract this strongly-bound DDT.
Using the same extraction conditions
as with pure SC-CO- (40°C, 100 atm, flow
rate ~ 315 cm3/s at 25°C, 1 atm), the two
mixed solvent systems (CO^-toluene and
CO -methanol) were investigated. These
data are plotted in Figure 4 along with
the extraction data obtained using pure
SC-CO, solvent. Little if any improvement
in extraction efficiency is observed upon
comparing the data for SC-CO with 5 wt%
toluene to that for pure SC-CO,. A
maximum of 60-70% of the DDT can be
removed in 10-20 minutes with either
solvent system. Extraction with SC-C02
and 5 wt% methanol, however, is more
successful. Approximately 95% of the
contaminants can be removed in 5 minutes
or less (Figure 4). These results imply
that for maximum extraction efficiency a
mixed solvent system should be used, where
the entrainer or co-solvent is tailored to
the soil type and the solute being
extracted.
To determine the rate-limiting-step
of the extraction process, we also carried
out extractions of DDT-contaminated
topsoils at various SC-CO_ flowrates,
while maintaining the metfianol entrainer
composition at 5 wt%. These results are
presented in Table II. The data indicate
that the process is apparently not
external-transfer limited (in the fluid
phase).
Table II. Variation of Soil Concentration in the Extraction of DDT-
Contaminatcd. Spill-Site Topsoil using SOCO, with 5 vtt,
Hethanol
Conditions: AO°C, 100 atm, 470 and 47 cmVsec
Extraction Time
(min)
470 craVsec
@ 25°C, 1 atm
47 cmVsec
8 25°C, atm
0
1
1
2
2
5
5
10
10
1271
*
221, 207
222, 236
82, 105
9S, 94
66, 58
58, 61
52, 37
59, 48
1271
"
60
20
40
60
120
86
34
51
25
Multiple entries in concentration column are replicate analyses.
-378-
-------
SCF EXTRACTION OF PCB CONTAMINATED
TOPSOIL -. RESULTS AND DISCUSSION
To again demonstrate the effective-
ness of SC-CO with 5 wt% methanol, this
solvent system was used to extract a
spill site topsoil containing ca. 3,500
ppm Aroclor 1260 and 2,100 ppm Aroclor
1242. The results, plotted in Figure 5,
show over 98% removal of the contaminant
is possible within a 10 minute extraction
period.
EXPERIMENTAL METHODS FOR SCF OXIDATION
STUDIES.
We are currently investigating the
oxidation of toluene with redox
catalysts, using the system shown in Fig.
6. CO- is compressed to slightly above
critical conditions (T = 31°C and P -
72.8 atm), and at a flow rate of 1 g/s is
mixed with toluene delivered by a
metering pump and air from a cylinder.
All flow rates are controlled by metering
valves. The reaction mixture is ca. 1.5
wt% toluene, 1.5% oxygen, 5% N and 92%
C02, and is controlled at 30-3DO°C ±1°C
by encasing the reactor (V1 o.d. tube
with a fritted disk) in an insulated heat
source (4" o.d. heated alumnium block). '
For the initial experiments the pressure
at the reactor was 80 atm although the
reactor is capable of pressures in excess
of 300 atm.
Before the reactor, the air/CO.
ratio and toluene/CO- ratio are checked
by on-line sample injection into a gas
chromatograph with a thermal conductivity
detector; after the reactor, product
samples are injected into an on-line gas
chromatograph interfaced to a mass
spectrometer (Extrel Model ELQ 400) for
detailed analysis. The products are then
cooled by expansion through a heated
micro-metering valve. The uncondensed
CO is passed, through a totalizing meter.
It was possible to dissolve 1.5 wt%
toluene and 6.5 wt% air in supercritical
C02 at 80 atm and 293-493K. Using this
mixture the toluene could be oxidized at
low rates and conversions to benzalde-
hyde, benzyl alcohol, the cresol isomers,
and a lesser amount of condensation
products and carbon oxides. A 5% CoO/
A1.03 catalyst, calcined at 200°C in
order to preclude oxidation to Co_0, and
the cobalt "aluminates", proved to be an
active and selective (for partial oxida-
tion) catalyst. Its turnover number for
partial oxidation was about 10 /s at
473K.
SUMMARY
We have reported the use of SC-CO-,
with either toluene or methanol as an
entrainer, for the extraction of DDT and
PCBs from contaminated topsoils. The
extraction efficiency was shown to be a
strong function of the entrainer selected.
A supercritical mixture of C02 plus 5 wt%
toluene showed no improvement over CO.
alone, with only ~ 75% removal of the DDT
possible; the residual DDT was strongly
adsorbed on the soil. However, near
complete removal of DDT was possible using
SC-C02 with 5 wt% methanol at 40°C and a
flowrate of 0.7 g/s. The efficiency of
the C02-methanol system remained unchanged
with a large variation in flowrate, from 1
g/s to 0.1 g/s, demonstrating that no
external transfer resistances are present
at these conditions. Near complete
removal of a mixture of Aroclor 1260 and
Aroclor 1242 was also possible using
SC-CO2 with 5 wt% methanol at 40°C and a
flowrate of 0.7 g/s. Once these
contaminants have been dissolved in the
SCF phase, partial oxidation is feasible
if catalyzed by certain supported metal
oxides. Further work on total oxidation
by supported mixed-metal oxides is in
progress.
ACKNOWLEDGMENT
This study was supported in part by
Grant CR-809714 from the US Environmental
Protection Agency. This support does not
signify that the contents necessarily
reflect the views and policy of the.
Agency; no mention of trade names or
commerical products constitute endorsement
or recommendation for use. We acknowledge
the experimental assistant o'f DeAnn Leach
and James Torres.
NOMENCLATURE
F ' = ' fluid phase
AH1 =
fugacity, atm
heat of fusion, cal/mol
-379-
-------
k
L
P
R
S
T
V
interaction, parameter
liquid phase
pressure, atro
gas constant, 1.987 cal/mol-°K
or 82.06 atm-cm3/mol-°K or soil
particle radius, cm
solid phase
temp e rature, °K
vapor phase or molar volume,
cm3/mol
SUBSCRIPTS
m = melting point
o = initial value
1 = supercritical solvent
2 = solute
SUPERSCRIPT
o& = pure subcooled liquid
os = pure solid
REFERENCES
Brady, B.O., Kao, C.-P., Gambrell, R.P.,
Dooley, K.M., and Knopf, F.C.,
Ind. Eng. Chem. Research, 26, 261
(1987).
deFilippi, R.P., Krukonis, V.J., Robey,
R.J., and Modell, M., EPA Report -
600/2-80-054, National Technical
Information Service, Washington,
D.C., 1980a.
deFilippi, R.P., Krukonis, V.J., Robey,
R.J., and Modell, M., United States
EPA Report No. 600/2-80-054; EPA
Office of Research and Development:
Research Triangle Park, North
Carolina, 1980b.
Dooley, K.M., and Knopf, F.C., 1986,
submitted to Ind. Eng. Chem.
Research.
Eppig, C.P., deFilippi, R.P., and Murphy,
R.A. United States EPA Report No.
600/2-82-067; EPA Office of Research
and Development: Research Triangle
Park, North Carolina, 1981.
Gates, B.C., Katzer, J.R., and Schuit,
G.C.A., Ch. 4 of "Chemistry of
Catalytic Processes," McGraw-Hill,
1979.
Groves, F.R., Brady, B.O., and Knopf,
F.C., CRC Reviews in Env. Control
15, 237 (1985).
Haber, J., ACS Symp. Ser. 279, 3 (1985).
Heideman, R.A., Khalil, A.M., AIChE J 26,
769 (1980).
Kaeding, W.W., Lindblom, R.O., Temple,
R.G., and Mahon, H.I., Ind. Eng.
Chem. Proc. Des. Dev. 4, 97 (1965).
Kramer, G.M. and Leder, F., U.S. Patent
3,880,945 (1975).
Metzger, J_0., Hartmanns, J., Malwitz, D.,
and Koll, P. in "Chemical Engineering
at Supercritical Fluid Conditions,"
Paulaitis, M.E., Penninger, J.M.,
Gray, R.D., and Davidson, P. (eds.),
p. 515 (1983).
Modell, M., Gaudet, G.G., Simson, M.,
Hong, G.T., and Biemann, K., Eighth
Annual Research Symp. Land Disposal,
Incineration and Treatment of
Hazardous Waste, Ft. Mitchell,
March 8 to 10, 1982.
Parshall, G.W., Chs. 7 and 10 of
"Homogeneous Catalysis," Wiley, 1980.
Paulaitis, M.E., Krukonis, V.J., Kurnik,
R.T., and Reid, R.C., Rev. Chem. Eng.
I, 1791 (1982).
Peng, D.Y; Robinson, D.B., Ind. Eng. Chem.
Fundam. 15, 59 (1976).
Ringhand, P.H. and Kopfler, F.C., 186th
National Meeting of the American
Chemical Society, Washington, D.C.,
August 28 to September 3, 1983.
Robertson, J.H., Cowen, W.F., and
Longfield, J.Y., Chem. Eng., June 30,
1980, p. 102.
Schuit, G.C.A. and Gates, B.C., pp. 461-
475 of "Chemistry and Chemical
-300-
-------
Engineering of Catalytic Processes,'
R. Prins and G.C.A. Schuit, Eds.,
Sijthoff and Nordhoff, 1980.
Semenova, A.I., Emel'yanova, E.A.,
Tsimmerman, S.S., and Tsiklis, D.S.,
p'"-- J. Phys. Chem. 53, 1428
Russ.
(1979).
Sheldon, R.A. and Kochi, J.K., Adv.
Catal. 25, 272 (1974).
Subramaniam, B. and McHugh, M.A., Ind.
Eng. Chem. Proc. Des. Dev. 25, 1
(1986).
Tiltscher, H., Wolf, H., and Schelschorn,
J., Angew. Cfaem. Int. Ed. Engl. 20,
892 (1981).
U.S. EPA; "Interim Methods for the
Sampling and Analysis of Priority
Pollutants in Sediments and Fish
Tissue"; U.S. EPA, Environmental
Monitoring and Support Laboratory:
Cincinnati, Ohio, 1980.
U.S. EPA, "Priority Pollutant Frequency
listing Tabulations and Descriptive
Statistics", U.S. EPA, Effluent
Guidelines Division, 1980b.
U.S. EPA 600/4-82-057; "Qrganochlorine
Presticides and PCB's-Method 608";
U.S. EPA Environmental Montioring
and Support Laboratory: Cincinnati,
Ohio, 1982.
Vasilakos, N.P., Dobbs, J.M., and Parisi,
A.S. Ind. Eng. Chem. Process Des.
Dev. 24. 121 (1985).
-381-
-------
2(DDT)
35 C
40*C
I(COz)
3(CH3OH)
Figure 1. Phase behavior of the ternary system of CO~-DDT-methanol
at 35°C and 40°C, 100 atm.
-332-
-------
01
o
-IN
a
o
•H
4J
KJ
•o
•H
01
a
0)
fi
u
>.
ra
at
VI ,
S"
00
-383-
-------
0>
M
ai
o
•rl
4J
O
«t
HI
H
w
3
4J
ID
M
n
o.
a,
eo
u
•H
4->
ra
I
u
VO
01
3
00
o>
c
o>
•3
^
r~ ~Z
_ 0) O
CD Q. Q_ OL H- O
till)
m a. o cr J-
D. a.
-334-
-------
w
UIO
01 Z
Is
o> S
o o
O O
«, «,
= 3
Q>
1
O
"5. o>
O
o>
o
o
S S
o>
ol (£ t-
i i i
o
OL Q. H-
° ~ S
|S -2
6*0
o>
X
O
•H
J-)
O
CO
S-l
4J
•H
o
tn
a
o
•H
4J
•ri
ta
ta
S
2
a)
o
•H
u
CO
^
00
•H
(4
£D X
-335-
-------
0.6 r
9Q" 1000 fj.q DDT/g soil
O - Pure SC-C02
• - SC-C02with 5 wt% toluene
* - SC-C02 with 5wt% methanol
O
O
10
20 3O 40
TIME (Minutes)
50
60
Figure 4. Temporal variation of relative soil concentration in the
extraction of a DDT-contaminated, spill-site topsoil,
showing the effect of entrainers.
-386-
-------
3300
2000
500
o>
40O
or
i-
UJ
o
o
CO
o
a.
300
2OO
IOO
SC-CO EXTRACTION w 5 wt% METHANOL
@ 40°C, 1400 psi AND 0.7 g/sec
DRY SPILL SITE TOPSOIL
90 = 3331 fj.q Aroclor 1260 /g Soil
= 2052 ^.g Aroclor 1242/g Soil
A Aroclor 1260 Concentration/g Soil
• Aroclor 1242 Concentration/g Soil
1 1
•
A
t
IO 15 2O
EXTRACTION TIME (min.)
25
30
Figure 5. Temporal variation of relative soil concentration in the
extractxon of a PCB-contaminated, spill site topsoil.
-387-
-------
MICROBIAL DEGRADATION OF SYNTHETIC CHLORINATED COMPOUNDS
R. A. Haugland, U. M. X. Sangodkar, A. M. Chakrabarty
Department of Microbiology and Immunology
University of Illinois at Chicago
Chicago, Illinois 60612
P. H. Tomasek
Pesticide Degradation Laboratory
U. S. Department of Agriculture
Beltsville, Maryland 20705
P. R. Sferra
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Pseudomonas cepacia strain AC1100 is a novel organism from the standpoint of
being the product of a facilitated evolution process that has resulted in the formation of
a unique metabolic pathway for the utilization of the herbicide 2,4,5-trichlorophenoxy-
acetic acid. A review is presented of past research pertaining to this organism as well
as features it possesses that make it a highly desirable subject for further investiga-
tion. Recent results of an on-going research program designed to isolate the genetic
determinants responsible for 2,4,5-T metabolism by this organism are also presented.
These include the isolation of a series of spontaneous mutants affected at several differ-
ent 2,4,5-T degradation-specific loci, the construction of a genomic library of AC1100 DNA
sequences in Escherichia coli and the use of this library in the complementation of a pre-
sumed transposon Tn5-induced mutant.
INTRODUCTION
With growing public awareness of the
health threats posed by synthetic chlorin-
ated compounds in the environment there is
an increasing demand to develop effective
methods for their elimination. Microbial
degradation is known to contribute to the
destruction of many of these compounds in
the environment. A major constraint to
this process occurs, however, when approp-
riate dissimilatory pathways have not yet
evolved in natural organisms. In this
regard, it has been demonstrated that
deliberate genetic selection and genetic
manipulation procedures in the laboratory
may lead to the construction of strains
having wider biodegradative capabilities
than their natural counterparts (1-5).
One of the more notable products of
this form of deliberately facilitated evol-
ution is Pseudomonas cepacia strain AC1100
(6). AC1100 has acquired the ability to
completely degrade the herbicide 2,4,5-
trichlorophenoxyacetic acid (2,4,5-T) and
in the process utilize the compound as a
source of carbon and energy for growth. As
such, it presently provides a rather unique
opportunity to study mechanisms by which
genetic information may be assembled and
evolve towards the formation of novel bio-
degradative pathways in a single organism.
It has also been shown that AC1100 has the
capability to degrade or dechlorinate a
fairly wide variety of chlorophenol congen-
ers although available evidence suggests
that the majority of these compounds are
not utilized for growth (7). Another
-388-
-------
potential use of this organism may, there-
fore, be as a source of genetic information
in future attempts to assemble complete
metabolic pathways for these compounds and
other priority pollutants. Finally, treat-
ment of contaminated soils in the labora-
tory with AC1100 has been found to result
in significant reductions of 2,4,5-T (8,9).
This organism can therefore be considered
as one of the better candidates presently
available for use in evaluating the effects
(both in terms of treatment efficacy and
environmental impact) of deliberately
introducing a microorganism into the
environment for the purpose of soil decon-
tamination.
In order to fully realize both these
and other potential applications of AC1100,
it is clear that a good understanding of
the genetics and physiology of this organ-
ism (particularly as they relate to its
biodegradative capabilities) will be bene-
ficial. The purpose of this report will be
both to summarize past information that has
been obtained in this regard as well as to
describe ongoing studies that are being
conducted in a collaborative research
effort by the Gulf Breeze and Cincinnati
EPA laboratories and the University of
Illinois at Chicago.
PHYSIOLOGICAL STUDIES ON 2,4,5-T AND
CHLOROPHENOL DEGRADATION BY AC1100
Studies have indicated that 2,4,5-
trichlorophenol (2,4,5-TCP) is an early
intermediate in the 2,4,5-T breakdown
pathway of AC1100 and that the enzyme(s)
responsible for carrying out this conver-
sion are constitutive (10). The enzymes
involved in the further breakdown of 2,4,5-
TCP are as yet unknown. Proposed pathways
for the degradation of 2,4-dichlorophenoxy-
acetic acid through a catechol intermediate
(11) and pentachlorophenol through a hydro-
quinone (12) currently provide two alterna-
tive models for potentially describing
portions of this activity (Figure 1). It
also appears that one or more of the later
enzymes involved in 2,4,5-T degradation by
AC1100 is or are inducible (10). In this
regard, the available evidence suggests
that 2,4,5-TCP or a subsequent breakdown
intermediate may be active as the inducer.
Considering the nature of the selec-
tive pressure employed in generating AC1000
(i.e., for growth with 2,4,5-T as a sole
source of carbon and energy) it is not sur-
prising to observe that this organism has
evolved a set of enzymes that efficiently
metabolize the related compound 2,4,5-TCP
and are perhaps synthesized in response to
this compound. The ability of AC1100 to
metabolize various other chlorophenol con-
geners, however, may be impaired in several
ways. Studies of oxygen uptake and chlor-
ide release in the presence of various
chlorophenol congeners by 2,4,5-T-grown
resting cells of AC1100 (Table 1) have
shown that each of these compounds is de-
chlorinated to some degree and in many
instances to an extent comparable to that
of 2,4,5-TCP. In contrast, the corres-
ponding oxygen uptake values observed for
these compounds were in each case well be-
low that observed for 2,4,5-TCP and also
below the levels that might have been pre-
dicted based on.their individual degrees of
dechlorination. In general the more highly
chlorinated congeners showed the highest
ratios of dechlorination to oxygen uptake.
These results suggest that to varying de-
grees, the dissimilation of these compounds
may be blocked or impaired at steps in the
process that occur after at least some
dechlorination has taken place. Such a
phenomenon might be expected to occur if
specific intermediates formed in the ini-
tial degradation of these compounds are
poorly recognized as substrates by sub-
sequent AC1100 degradative enzymes. Par-
ticularly in the case of the more highly
chlorinated phenols, it is also tempting to
speculate that dechlorination may at least
to some extent be mediated by hydrolytic
and/or reductive mechanisms similar to
those demonstrated by Steiert and Crawford
(12) for pentachlorophenol (PCP) degrada-
tion by a Flavobacterium species (Figure 1).
Degradation of PCP by AC1100 has also
been proposed to be constrained at the
level of gene regulation. In this regard
studies have revealed that, unlike 2,4,5-
TCP, PCP is unable to induce its own de-
chlorination in resting cells of AC1100
(10). At present, the ability of other
chlorophenol congeners to induce their own
degradation is unknown.
Besides requiring induction, there is
also some evidence to indicate that the
2,4,5-T degrading activity of AC1100 may be
repressed or inhibited by a number of al-
ternative carbon sources such as glucose,
lactate and succinate (10). Under certain
conditions this effect can be shown to be
at least partially mediated by inhibited
-389-
-------
9CH2COOH
,-CI
COOH
COOH
COOH
:OOH
CO, + Cl-
Figure 1. Pathways for 2,4-dichlorophenoxyacetic acid (2,4-D) and pentachlorophenol (PCP)
biodegradation. Part A: Proposed pathway for the degradation of 2,4-D by A. eutrophus
OMP134 (after Don et^a]_., 11). The metabolites involved are: 2,4-dichloropTfenoxyacetic
acid, 1; 2,4-dichlorophenol, 2; 3,5-dichlorocatechol, 3; 2,4-dichloromuconic acid, 4;
trans-2-chlorodiene-1actone, 5; cis-2-chlorodiene lactone, 6; 2-chloromaleylacetic acid, 7
and 3-oxoadipic acid, 8. The enzymes involved are: 2,4-D monooxygenase, a; 2,4-dichloro-
phenol hydroxylase, b; chlorocatechol 1,2-dioxygenase, c; chloromuconate cycloisomerase,
d; 4-carboxymethylenebut-2-en-4-olide (dienelactone) hydrolase, e and trans-chlorodiene
lactone isomerase, f. Part B: Proposed pathway for the degradation of PCP by a Flavo-
bacterium sp. (after Steiert and Crawford, 12). The metabolites involved are: penta-
chlorophenol, 1; tetrachloro-£-benzoquinone, 2; tetrachloro-p_-hydroquinone, 3; 2,3,6-tri-
chlorohydroquinone, 4 and 2,6-dichlorohydroquinone, 5.
uptake of an inducer compound (e.g., 2,4,5-
TCP). Under other conditions this effect
also appears to be exerted at a step or
steps in the pathway following the conver-
sion of 2,4,5-T to 2,4,5-TCP and has been
suggested to create a peculiar problem for
AC1100 cells grown in the presence of
2,4,5-T and one of these alternative carbon
sources. Culturing AC1100 under these con-
ditions has been found to result in prema-
ture termination of growth, considerable
loss of viability and relatively high
levels of 2,4,5-TCP accumulation in the
medium. The explanation proposed for these
phenomena is that the inhibitory effect of
the alternative substrates leads to the
accumulation of 2,4,5-TCP to levels that
may become toxic to the cells.
In addition to evolving the genetic
information required for metabolizing
2,4,5-T, there is evidence that AC1100 has
also evolved the capability to produce a
2,4,5-T emulsifying agent (13). This emul-
sifying agent, which the available data in-
dicate is a high molecular weight lipid
derivative, has been shown to be specific
for 2,4,5-T and to some extent other
-390-
-------
TABLE 1. CHLOROPHENOL DEGRADATION BY
2,4,5-T-GROWN AC1100 RESTING
CELLS (FROM KARNS et al ., 7)
Substrate*
Dichloro-
phenols
2,3-
2,4- , ,
2,5- '
2,6-
3,4-
3,5- -
Trichloro-
phenols
2,4,5-
2,3,5-
2,3,6-
2,4,6-
2,3,4-
3,4,5-
Tetrachloro-
phenols
2,3,4,6-
2,3,4,5-
2,3,5,6-
Pentachloro-
phenol
Chloride
Release**
65
71
84
,36
60
- : 38
90
29
29
56
70
27
80
60
94
92
Oxygen
Uptake***
16.0
16.5
35.9
10.1
27.0
8.1
100.0
17.8
2.6
ND
ND
ND
16.1
ND
ND
9.3
*Each substrate was used at a concentration
of 0.1 mM. Incubation period was 3 hr.
**Percentage of the theoretical maximum.
***Expressed as a percentage of the net ,
respiration rate observed for 2,4,5-TCP.
ND: Not determined.
related chlorophenols and is not produced
by other P. cepacia strains. While a
definite role for this compound .has not as
yet been established, the production of
such agents is known to be important in
facilitating the uptake and utilization of
hydrophobic compounds by microorganisms.
MOLECULAR AND GENETIC STUDIES OF AC1100
Studies of various organisms showing
the capability to degrade different syn-
thetic compounds have revealed that the
genetic information for these activities is
often located,on plasmids. Populations of
AC1100 cells have been analyzed for plasmid
content and have revealed some heterogene-
ity in this regard (14). It was observed
that the predominant fraction of the cells
(80% or more) contained a single .plasmid of
approximately 170 kb. A smaller fraction
(about 10%) contained a plasmid of approxi-
mately 40 kb. Other plasmids ranging in
size from 2 kb to 30 kb were also observed.
Direct evidence for the involvement of
plasmids in the 2,4,5-T degrading activity
of AC1100 has not as yet been forthcoming.
Efforts to cure this strain of plasmids as
a means of making this determination using
conventional mitomycin C treatment have not
succeeded (14). The strongest indirect
evidence for such involvement has resulted
from the demonstration that hybridization
occurs between a specific region of AC1100
plasmid DNA and a portion of the plasmid
pJP4 (15). The hybridizing region of pJP4
harbors structural genes coding for both 3-
chlorobenzoic acid (3CBA) and 2,4-dictiloro-
phenoxyacetic acid (2,4-D) degradative
enzymes (11). It was also rep.orted that
AC1100 plasmid DNA did not hybridize to the
chlorocatechol (clc) gene cluster of
another plasmid TpAC27) under stringent
conditions (15). These clc genes (coding
for enzymes c,d;e; Figure 1A) are both
isofunctional and homologous to the corres-
ponding genes of pJP4. This observation
has. lead to the conclusion that the
sequence homology between AC1100 plasmid
DNA and pJP4 may be limited to genes in-
volved in the degradation of chlorophenoxy
acetates (e.g., enzyme a, Figure 1A).
Further preliminary evidence concern-
ing the location of determinants for
2,4,5-T metabolism in AC1100 has resulted
from the generation of random transppson
Tn5 mutants of this organism (15). Plasmid
and total DNAs isolated from a number of
Tn5-containing AC1100,derivatives that had
also lost their 2,4,5-T degradative ability
were hybridized with a labeled Tn5 DNA.
probe and in each case only the total DNA ;
was found to show homology. A rigorous
demonstration that the Tn5 insertions in
these mutants are directly responsible for
their loss of 2,4,5-T degrading activity
has not yet been made. Efforts to make
such a determination with one of these
mutants, PT88, were pursued by the con-
struction of a gene library of its total
genomic DNA and the .isolation of clones
containing the kanamycin resistance deter-
minant of Tn5 (15). Subsequent use of a
flanking AC1100 DNA fragment from one side
of one of these clones as a hybridization.
probe to filter blots of total AC1100 DNA
has revealed that a sequence is present
within this DNA fragment that is repeated.
numerous times in the AC1100 genome. These
-391-
-------
results have lead to the interesting specu-
lation that repeated sequences may be
associated with determinants for 2,4,5-T
degradation in AC1100 and hence may have
had a role in the evolution of this capa-
bility. In this regard, recent studies
with P. cepacia strain 249 have not only
identified a number of repeated insertion
sequences within this organism but have
also suggested a potential role for these
sequences in creating genomic plasticity
by processes such as insertional activation
of genes as well as the recruitment of
foreign plasmid-born genes by replicon
fusion (16-18).
Efforts to isolate and identify the
2,4,5-T degradative genes of AC1100 are
presently being continued in our labora-
tories. Toward this end a new genomic
library of AC1100 DNA sequences has re-
cently been constructed. To do this, total
DNA from AC1100 was isolated and partially
digested with restriction endonuclease Bam
HI under conditions that maximized the for-
mation of DNA fragments in the size range
of 20 to 30 kb. DNA of the 23 kb broad
host range cosraid vector pCPIS (19) was
also digested with Bam HI (in this case to
completion), dephosphorylated with calf in-
testine alkaline phosphatase and then
ligated with the AC1100 DNA fragments. The
ligation products were packaged into lambda
phage heads in vitro and transfected into
E. coli strain" AC80. A library containing
approximately 10,000 tetracycline resistant
clones was generated in this manner.
The library was subsequently mobilized
en masse using a modification of a previ-
ously described triparental conjugative
mating system (20) into a series of 2,4,5-T
defective, transposon Tn5 insertion mutants
of AC1100 (15, see above). Of 12 such mut-
ants screened in this manner, only one
(PT88) was initially observed to produce
tetracycline-resistant transconjugants that
appeared to be phenotypically complemented
based on their ability to grow on media
containing 2,4,5-T as a sole carbon source.
Plasmid DNA from a 2,4,5-T metabolizing
transconjugant was isolated and transformed
back into _E_. coli. This plasmid, desig-
nated pUSl, was confirmed to complement the
PT88 mutation by producing nearly 100%
2,4,5-T metabolizing transconjugants when
remobilized into PT88 from E. coli. The
recombinant plasmid pUSl has also subse-
quently been demonstrated to complement two
additional mutants (PT8 and PT9) in a
similar manner. Further studies are cur-
rently in progress to characterize the in-
sert DNA of this plasmid with respect to
the location, number and organization of
2,4,5-T degradative genes that it contains.
INSTABILITY OF 2,4,5-T DEGRADATIVE
CAPABILITY IN AC1100
It has been documented that AC1100
cells can spontaneously lose 2,4,5-T
degradative capability when grown in the
absence of this compound (6). These ex-
periments indicated that the frequency of
such events may vary with culture age but
could approach a rate of approximately 6 X
10"* per cell per generation or higher. As
previously mentioned, AC1100 has been shown
to be effective in decontaminating 2,4,5-T
treated soils in a laboratory setting. An
additional observation in these studies was
that cells with 2,4,5-T degradative capa-
bility rapidly disappeared from the soil
samples once the 2,4,5-T levels were dimin-
ished. While not proven, it is reasonable
to suggest that the instability of the
2,4,5-T degradative genes in AC1100 may
have contributed significantly to this
disappearance.
The underlying cause behind the insta-
bility of ACllOO's 2,4,5-T degradative
phenotype has not been established although
several potential mechanisms can be en-
visaged. One obvious possibility is that
2,4,5-T determinants are located on a
plasmid that is readily lost through segre-
gation. Other possibilities include legit-
imate or illegitimate recombination events
mediated by the presence of repeated
sequences or transposable elements. As
previously mentioned, numerous sequences
of this type have recently been identified -
in P. cepacia strain 249 and it is poten-
tially significant to note that this organ-
ism also appears to undergo spontaneous
mutations at a high frequency.
The mechanism(s) underlying the spon-
taneous loss of 2,4,5-T degradative ability
by AC1100 are currently under further in-
vestigation in our laboratories. A number
of spontaneous mutants have been identified
by replica plating AC1100 colonies grown on
a nonselective medium (containing basal
salts and 0.1% yeast extract, 6) onto
medium containing 2,4,5-T as sole carbon
source. It has been observed that these
mutants can be phenotypically differenti-
ated on the basis of their growth charac-
-392-
-------
TABLE 2. GROWTH CHARACTERISTICS OF SPONTANEOUS ACUOO MUTANTS
Mutant Class
2,4.5-T
Growth Substrate
Glucose
2,4,5-T + Glucose
no growth
comparable to
wild-type
comparable to wild-type
but produces a dark red
pi gment
B no growth
C no growth
D no growth
comparable to
wild -type
comparable to
wild-type
comparable to
wild -type but
colonies are
more mucoid
very little or no growth
comparable to wild-type
comparable to wild-type
but colonies are more
mucoid
teristics in basal salts medium containing
2,4,5-T plus glucose. These differences
which are summarized in Table 2 suggest
that multiple genetic loci may be inde-
pendently susceptible to spontaneous muta-
tional events in AC1100. ,In view of this
observation it appears unlikely that tne
loss of a plasmid can be the sole explan-
ation for loss of 2,4,5-T degrative capa-
bility in this organism.
Studies are currently in progress to.
both identify the steps in the 2,4,5-T
degradative pathway at which these mutants
are blocked and to isolate the affected
genes by means of complementation with our
AC1100 genomic library. As the cloned DNA
sequences corresponding to these mutated
regions are isolated, it should be possi-
ble to use them as hybridization probes in
determining the physical basis for the
mutational events.
ACKNOWEDGEMENTS
This investigation was supported by a
cooperative program grant from the U. S.
EPA (CR809666) and in part by a Public
Health Service grant (ES04050) from the
National Institute of Environmental Health
Sciences.
REFERENCES
1. Reinke, W. and H.-J. Knackmuss. 1979.
Construction of haloaromatics utiliz-
ing bacteria.
385-386.
Nature (London) 277:
5.
6.
Chatterjee, D. K. and A. M.
Chakrabarty. 1982. Genetic rear-
rangements in.plasmids specifying
total degradation of chlorinated
benzoic acids. Mot. Gen. Genet.
188:279-285.
Sjchweinj U. and E. Schmidt. 1982.
Improved degradation of monochloro-
phenols by a constructed strain.
Appl, Environ. Microbiol. 44:33-39.
Reinke, W., S. W. Wessels, M. A.
Rubio, J. Latorre, U. Schwein, E.
Schmidt, M. Schlomann and H.-J.
Knackmuss. 1982. Degradation of
monochlorinated aromatics following
transfer of genes encoding chloro-
catechol catabolism. FEMS Microbiol.
Lett. 14:291-294.
Lehrbach, P. R., J. Zeyer, W. Reinke,
,H.-J. Knackmuss and K. N. Timmis.
1984. Enzyme .recruitment in vitro:
Use of cloned genes to extend the
range of haloaromatics degraded by
Pseudomonas sp. strain 813. J.
Bacteriol. 158: 1025-1032.
Kilbane, J. J., D. K. Chatterjee,
J. S. Karns, S. T. Kellogg and A. M.
Chakrabarty. 1982. Biodegradation of
2,4,5-trichlorophenoxyacetic acid by a
-393-
-------
9.
10.
11.
12.
13.
pure culture of Pseudomonas cepacia.
Appl. Environ. Microbiol. 44:72-78.
Karns, J. S., J. J. Kilbane, S.
Duttagupta and A. M. Chakrabarty.
1983. Metabolism of halophenols by
2,4,5-trichlorophenoxyacetic acid-
degrading Pseudomonas^ cep_acijK Appl.
Environ. Microbiol. 46:1176-1181.
Chatterjee, D. K., J. J. Kilbane and
A. M. Chakrabarty. 1982. Biodegrada-
tion of 2,4,5-trichlorophenoxyacetic
acid in soil by a pure culture of
Pseudomonas cepacia. Appl. Environ.
Microbiol. 44:514-516.
Kilbane, 0. J., D. K. Chatterjee and
A. M. Chakrabarty. 1983. Detoxifica-
tion of 2,4,5-trichlorophenoxyacetic
acid from contaminated soil by
Pseudomonas cepacia. Appl. Environ.
Microbiol. 45:1697-1700.
Karns, J. S., S. Duttagupta and A. M.
Chakrabarty. 1983. Regulation of
2,4,5-trichlorophenoxyacetic acid and
chlorophenol metabolism in Pseudomonas
cepacia AC1100. Appl. Environ.
Microbiol. 46:1182-1186.
Don, R. H., A. J. Weightman, H.-J.
Knackmuss and K. N. Timmis. 1985.
Transposon mutagenesis and cloning
analysis of the pathway for degrada-
tion of 2,4-dichlorophenoxyacetic acid
and 3-chlorobenzoate in Alcaligenes
eutrophus JMP134(pJP4). J. Bacteriol.
161:85-90.
Steiert, J. G. and R. L. Crawford.
1986. Catabolism of pentachloro-
phenol by a Flavobacterium sp.
Biochem. Biophys. Res. Comm. 141:
825-830.
Banerjee, S., S. Duttagupta and A. M.
Chakrabarty. 1983. Production of
emulsifying agent during growth of
Pseudomonas cepacia with 2,4,5-
trichlorophenoxyacetic acid.
Microbiol. 135:110-114.
Arch.
14. Ghosal, D., I.-S. You, D. K.
Chatterjee and A. M. Chakrabarty.
1985. Plasmids in the degradation
of chlorinated compounds. Plasmids
in Bacteria, D. Helinski, S. N. Cohen,
UT Clewell, D. Jackson and A.
Hollaender (eds.) Plenum Press, New
York, pp 667-686.
15. Tomasek, P. H. and A. M. Chakrabarty.
1985. Bacterial degradation of chlor-
inated compounds. Proceedings of the
llth Annual Hazardous Waste Research
Symposium. U. S. Environmental
Protection Agency, Cincinnati, Ohio
pp 127-134.
16. Gaffney, T. D. and T. G. Lessie.
1987. Insertion sequence-dependent
rearrangements of Pseudomonas cepaci a
plasmid pTGLl. J. Bacteriol. 169-
224-230.
17. Barsomian, G. and T. G. Lessie. 1986.
Replicon fusions promoted by insertion
sequences on Pseudomonas cepaci a
plasmid pTGL6. Mol. Gen. Genet. 204:
273-280.
18. Scordilis, G. E., H. Ree and T. G.
Lessie. 1987. Identification of
transposible elements which activate
gene expression in Pseudomonas
cepacia. J. Bacteriol. 169:8-13.
19. Darzins, A. and A. M. Chakrabarty.
1984. Cloning of genes controlling
alginate biosynthesis from a mucoid
cystic fibrosis isolate of Pseudomonas
aeruginosa; J. Bacteriol. 159:9-18.
20. Ditta, G., S. Stanfield, D. Corbin and
D. R. Helinski. 1980. Broad host
range DNA cloning system for gram
negative bacteria: Construction of a
gene bank of Rhi zobi urn meliloti.
Proc. Nat'l. Acad. Sci. USA. 77:7347-
7351.
-394-
-------
BACTERIAL OXIDATION OF POLYCHLORINATED BIPHENYLS
Louise M. Nadim, Mark J. Schocken, Frank K. Higson and David T. Gibson
Center for Applied Microbiology
The University of Texas at Austin
Austin, Texas 78712
and
Donna L. Bedard, Lawrence H. Bopp and Frank J. Monde!lo
General Electric Company
Corporate Research and Development
Schenectady, New York 12301
ABSTRACT
The present studies represent a summary of the results obtained on the degradation of
polychlor.nated biphenyls by two strains of bacteria. The organisms, Alcaligenes eutro-
P"us H850 and Pseudomonas put?da LB400, are capable of metabolizing a wide range oFPCf
congeners. The initial reactions involved in the oxidation of 2,5,2',5'-tetrach1oro-
biphenyl by both organisms appears to involve oxidation at the unsubstituted 3,4-posi-
tions. The properties of the enzymes involved in these reactions and the regulation of
their activities are being studied by modern molecular biological techniques. It is
anticipated that the results obtained will lead to the construction of improved strains
of bacteria that can efficiently degrade a wide range of PCB congeners.
INTRODUCTION
Polychlorinated biphenyls (PCBs) are
manufactured by the controlled chlorine
substitution of the biphenyl molecule.
Mixtures of PCB isomers and congeners of a
differing degree of chlorine substitution
were first marketed in the USA under the
name "Aroclor" by the Monsanto Corporation
almost 50 years ago.
PCBs are extremely stable compounds,
their excellent heat resistance and dielec-
tric properties have led to wide industri-
al use in capacitors, transformers,
dielectric fluids, fire retardants and
plasticizers (10,12).
The stability of PCBs also makes them
quite resistant to biodegradation. This
is particularly true for congeners contain-
ing 5 or more chlorine substituents (7,12)
which tend to accumulate in the environ-
ment. The lipophilic nature of PCBs
causes them to partition into the fatty
tissue of higher organisms, thus concen-
trating them in the food chain by several
orders of magnitude at each succeeding
trophic level. This is of concern to
human health since the long term effects
of PCBs are not fully understood.
Two microorganisms, A lealigenes eu-
trophus strain H850 and Pseudomonas~puti-
da strain LB400 were used in this study!"
Both organisms were isolated from PCB con-
taminated soil and tested for their abili-
ty to degrade a wide range of PCBs (1,2,A)
including congeners that have no unsubsti-
tuted 2,3-positions (the favored sites of
initial oxidative attack), such as
2,5.2",5'-tetrachlorobiphenyl (2,5,2',5'-
CB). Congener depletion assays indicate
that chlorine substitution at the 2,5-pos-
itions may facilitate degradation of cer-
tain PCBs (2,3). The pathway used for the
degradation of a 2,5-dichlorinated aroma-
tic ring has not been established.
-395-
-------
However, It seems possible that alterna-
tive pathways to that shown in Figure 1
for blphenyl may exist. The general fea-
tures of the pathway have often been extra-
polated to account for the biodegradation
Of certain PCB congeners (7,12). For exam-
ple, it is generally assumed that the ini-
tial oxidative reaction occurs by the in-
corporation of both atoms of molecular oxy-
gen Into the aromatic nucleus that con-
tains the least number of chlorine sub-
stltuents. The resulting dihydrodiol is
then oxidized by a pyridine nucleotide-
dependent dehydrogenase to form the appro-
priate ring-fission substrate. Subsequent
steps involving hydrolysis, hydration and
aldol cleavage are thought to account for
the formation of small molecules that can
enter the tricarboxylic acid cycle. It is
difficult to see how 2,5,2',5'-CB and simi-
lar PCBs can be degraded by this pathway.
The results presented, show for the first
time that bacteria can initiate oxidation
of the biphenyl molecule at the 3,4-posi-
tions.
COOH
Acid
4. 2,3-Dlhydroxybiphenyl Benzole acid
Blphenyl els.-Blphenyl dihydrodiol Ring fission product
(J) Biphenyl dioxygenase
(5) els-Biphenyl dihydrodiol dehydrogenase
(f) 2,3-Dlhydroxyblphenyl dioxygenase
© Hydrolyase
Figure 1. Initial reactions in the oxidation of biphenyl by A_. eutrophus H850 and £.
put!da LBAOO
MATERIALS AND METHODS
Microorganisms and growth conditions
Alcaligenes eutrophus strain H850 and
Pseudomonas put Ida strain LBkOO were
isolated as described previously (1,4).
Transformation studies were usually con-
ducted with biphenyl-Snduced cells. Metab-
olites were isolated by high pressure li-
quid chromatography (HPLC) and identified
from the information provided by conven-
tional chemical techniques. These includ-
ed proton magnetic resonance spectrometry
(PMR), mass spectrometry (MS) and gas
chromatography (GC). In certain instances
crude mixtures of metabolites were sepa-
rated and identified by GC/MS.
Mutant Isolation
Mutagenesis with N-methyl-N'-nitro-N-
nitrosoguanidine was carried out as des-
cribed previously (5). Transposon mutagen-
esis was carried out using either E_. col i
NEC0100 (pRKTVH) as described by Finette
(6) or £. coli C600 (pBEE-132) as des-
cribed by Kuner et al (11). Presumptive
biphenyl strains were selected on the
basis of their failure to produce large
colonies when grown on mineral salts agar
containing 0.2% succinate and 40 microgram
(y g/ml) of triphenyltetrazolium chloride.
Biphenyl crystals were placed in the lid
of each petri dish.
Rapid screening techniques were util-
ized to identify mutants defective in
structural genes of the biphenyl catabolic
pathway. Biphenyl dioxygenase activity
-396-
-------
was detected by the ability of strains to
clear zones of bipheny] from the surface
of agar plates that had been sprayed with
an ether solution of biphenyl.
Mutants defective in cis-2,3-dihydroxy-
2,3-dihydrobiphenyl dehydrogenase and
2,3-dihydroxybiphenyl dioxygenase were
detected by their ability to clear zones
of biphenyl from the surface of agar
plates as described above. Dehydrogenase
mutants were distinguished from dioxygen-
ase mutants by their ability to form a
yellow ring fission product when colonies
on agar plates were sprayed with an ether
solution of 2,3-dihydroxybiphenyl. No
color change was observed when 2,3-dihy-
droxyb iphenyl dioxygenase mutants were
exposed to 2,3-dihyroxybiphenyl. Mutants
defective in the ring fission enzyme,
2-hydroxy-6-oxo-pheny1hexa-2,4-d i eno i c
acid hydrolase, were detected by the
accumulation of the yellow ring fission
product when colonies were exposed to
biphenyl, c|s-2,3-dihydroxy-2,3-dihydro-
fa iphenyl and 2,3-dihydroxybiphenyl.
RESULTS
Whole cell studies
Both A lealigenes eutrophus H850 and P.
P"tida LB400 were able to utilize biphen^l
as the sole source of carbon and energy
for growth. In order to establish the
initial reactions involved in biphenyl
degradation mutant strains were isolated
that contained defects in the structural
genes for the first four enzymes of the
pathway shown in Figure 1. The properties
of these mutant strains are shown in Table
1. Strains FM200 and FM202 do not produce
clear zones on agar plates that have been
sprayed with an ether solution of biphen-
yl. However, crossfeeding experiments
with Beijerinckia B8/36, a mutant that
accumulates cis-biphenyl dihydrodiol (8),
results in growth of both FM200 and
FM202. These strains may have defects in
the structural gene(s) for biphenyl dioxy-
genase or they may be unable to transport
biphenyl. Strains FM203, 204, and 408
each oxidize bipheny] to cis-biphenyl
TABLE 1. PRODUCTS FORMED FROM BIPHENYL BY MUTANT STRAINS
OF P. PUTIDA LB400
Strain Designation
Source
Products formed from Biphenyl
FM200
FM202
FM203
FM204
FM408
FM903
FM905
FM205
FM206
Tn5 Kmr2
Tn5 Kmr
Tn5 Kmr
Tn5,Kmr
NTG-5
Tn5-132 tet™
Tn5-132 tetr
Tn5 Km
Tn5 Kmr
None
None
cis-B iphenyl dihydrodiol
cis-B iphenyl dihydrodiol
cis-Biphenyl dihydrodiol
2, 3-D ihydroxyb iphenyl (2,3-DB)
2, 3-D ihydroxyb iphenyl (2,3-DB)
2,3-DB Ring fission product
2,3-DB Ring fission product
Strains were isolated as described in Materials and Methods.
Kanamycin resistant.
N-methyl-N'-nitro-N-nitrosoguanidine.
Tetracycline resistant.
-397-
-------
dthydrodiol which has an absorbtlon maxi-
mum at 303 nm (8). Biphenyl and cis-bi-
phenyl dihydrodiol are both oxidized by
strains FM903 and 905 to a compound that
shows an absorption maximum at 2^7 nm. In
addition, cell free supernatant solutions
containing this metabolite are rapidly
oxidized to a yellow ring fission product
by blphenyl-induced cells of the parent
strain of P. put?da LB400. These observa-
tions strongly indicate that strains FM903
and 905 contain a defective gene for
2,3-dihydroxybiphenyl dioxygenase and that
the product formed from biphenyl and cis-
biphenyl dihydrodiol is 2,3-dihydroxybi-
phenyl. When biphenyl, cJ£-biphenyl dihy-
drodiol and 2,3-dihydroxybiphenyl are in-
cubated with strains FM205 and 206 a
bright yellow ring fission product which
has an absorbance maximum at k3k nm accu-
mulates from each substrate. These re-
sults show that neither strain contains
an active enzyme for the metabolism of the
ring-fission product formed from 2,3-dihy-
droxybiphenyl. The results are consistent
with the pathway shown in Figure 1. Polar-
ographic studies .indicate that A. eutro-
phus H850 also utilizes this pathway (data
not shown).
chlorophenyl ring improves degradabi1ity,
even for congeners that contain 1-3 chlor-
ine substituents on the other ring (Table
2D).
Oxidation of PCB congeners by biphenyl-
Induced cells of A. eutrophus HS50 and P.
put Ida LB500~~
Previous studies have shown that
congener depletion assays can provide
valuable information on the ability of
different bacterial strains to oxidize
individual PCBs (1,2,3,4). The results of
these Investigations are summarized in
Table 2. Quantitative analysis of the
data is difficult due to the fact that
percentage depletion was only measured at
one time point (2kh) and that the data
represents a summation of several differ-
ent experiments. Nevertheless, the re-
sults clearly indicate that chlorine sub-
stitution at the Jf,4'-positions renders a
molecule less susceptible to degradation
by these strains and that P. putida LB400
Is more effective than .A. eutrophus H850
In oxidizing congeners with this substi-
tution pattern (Table 2B). The effect of
chlorine substitution at the 2,6-positions
Is shown in Table 2C. This substitution
pattern seems particularly effective in
preventing significant PCB degradation by
both organisms. In contrast, a 2,5~di-
Oxidation of 2,5.2'.S'-tetrachlorobiphenyl
(2.5,2'.5'-CBT'
Alcaligenes eutrophus H850 and Pseudo-
monas putida LB400 were unable to utilize
2,5,2',5'-tetrachlorobiphenyl as a source
of carbon and energy for growth. However,
biphenyl-grown cells of both organisms
rapidly transformed this tetrachlorobi-
phenyl to polar products.
The first product formed from
2,5,2',5'-CB was isolated and identified
by conventional chemical techniques as
3,k-dihydroxy-3,4-d ihydro-2,5,2',5'-tetr'a-
ch1orob i pheny1 (3,4-di hydrod i ol). S ubse-
quent metabolism of this dihydrodiol led
to the accumulation of a more polar prod-
uct which appeared to be resistant to fur-
ther degradation. This metabolite was iso-
lated and identified as 3,4,3'.^'-tetrahy-
droxy-3,4,31 ,V-tetrahydro-2,5,2' ,5'-tetra-
chlorobiphenyl (b_i_s-3,A-di hydrod i ol).
There were no indications of prior dechlor-
ination reactions which would leave 2,3-
positions aval Table for oxidation. How-
ever, the identification by GC/MS of 2,51-
dichloroacetophenone as a minor metabolite
confirms the studies of Bedard et al (3),
and may indicate the presence of a novel
pathway for the degradation of biphenyIs
that contain certain chlorine substitution
patterns.
The second reaction in the bacterial
degradation of biphenyl is catalyzed by a
pyridine nucleotide-dependent dehydrogen-
ase which oxidizes cis-2,3-dihydroxy-2,3-
dihydrobipheny1 to 2,3-dihydroxybiphenyl.
This enzyme was very active in cell ex-
tracts prepared from bipheny1-grown cells
of A. eutrophus H850 and P_. putida LB400
(dala not shown). The same cell extracts
were not active against the 3,4-dihydro-
diol or the bis-3,4-dihydrodio1.
DISCUSSION
£. eutrophus H850 and IP. putida LBAOO
can Use biphenyl as a sole source of
carbon and energy for growth. Preliminary
studies with the mutants listed in Table 1
indicate that £*' putida LB400 oxidizes
-398-
-------
TABLE 2. DEGRADATION OF PCB CONGENERS BY £. EUTROPHUS H850
AND jP. PUT I DA LB4001
PCB
Congener
2,3-
2,5-
2,6-
2,2'-
2,4'-
4,4'-
2,4,4'-
2,5,2'-
2,5,4'-
2,4,2 ',4 '-
2, 4,3', 4'-
2,4.6,4'-
2, 5,2', 3'-
2, 5,3', 4'-
2,4, 6.2', 4'-
2,3, 4.2', 5'-
PCB
Congener
2,6-
2,4.6-
2,4,6,4'-
.2, 6,2', 6'-
2, 4, 6,2', V-
2, 4, 6,3', 4'-
2, 4, 6,3'. 5'-
2,3,6,2',3',6'-
2, 4, 6,2', 4', 6'-
A
Percent
H850
100 -
100
30
100
100
40
75
100
100
55
20
0
100 ,
90
0
60
C
Percent
RSso
30
20
0
0
0
0
0
20
0
degraded
LB400
100
100
40
100
100
50
100
100
100
100
65
0
100
100
0
100
degraded
LB400
40
25
0
0
0
0
0
35
0
PCB
Congener
4,4'-
2.4,4'-
2. 4,2'. 4'-
2. 4,3', 4'-
2,4,6,4'-
3, 4,3', 4'-
2, 4, 6,2', 4'-
2,4,5,2',4',5'-
PCB
Congener
2,5-
2,5,2'-
2,5.4'-
2.5.2'. 3'-
2,5, 3', 4'-
2,5,2',6'-
2, 5,2', 5'-
2, 5,2', 4', 6'-
2, 5.2', 4', 5'-
2, 5,2', 3'. 4'-
B
Percent
H850
40
75
55
20
0
0
0
15
D
Percent
H850
95
100
100
100
90
90
100
40
70
60
degraded
LB400
50
100
100
65
o
0
0
70
degraded
LB400
100
100
100
100
100
100
100
75
100
100
Data obtained from defined congener assays as described in (2) and reported
in part in references 3 and 4. Each mixture contained 10-11 chlorinated
biphenyls in which each congener was present at a concentration of 5.0 y M.
-399-
-------
blphenyl at the 2,3-position to form cis-
2,3-dIhydroxy-2,3~dIhydrobIpheny1. S ub-
sequent oxidation of this metabolite fol-
lows the reaction sequence shown in Figure
1. A mutant strain of A. eutrophus H850
(Strain FM803) also oxidizes biphenyl to
cis-2,3-dihydroxy-2,3-dihydrobiphenyl and
poTarographic studies with bipheny1-grown
cells of the parent organism (data not
shown) indicates that the pathway shown in
Figure 1 also represents the major reac-
tion sequence for the oxidation of bi-
phenyl by £. eutrophus H850.
The results in Table 2 show that both
A. eutrophus H850 and _P. putida LB400 have
"the ability to oxidize a wide range of PCB
congeners. Strain LB400 appears to be
more effective than strain H850 in terms
H
of the extent of oxidation of individual
PCBs. However, this aspect has not been
investigated in detail. In general terms,
the results shown in Table 2 indicate that
a 2,5-chlorine substitution pattern en-
hances degradation and this feature has
been reported in some detail by Bedard et
al (3). In contrast a 4,4'- or 2,6-chlor-
ine substitution pattern inhibits oxida-
tion even if the molecule has free 2,3-pos-
itions available for oxygenation.
Since both organisms oxidized
2,5,2',5'-tetrachlorobiphenyl (2,5,2',5'-
CB) this substrate was chosen for further
study. The results obtained are shown In
Figure 2. The major reactions involve
sequential dihydroxylation at both of the
open 3,4-positions in 2,5,2',5'~CB.
H OH
Figure 2.
Cl
Major reactions involved in the transformation of 2,5,2',5'-CB by biphenyl-
grown cells of £. eutrophus H850 and £. putida LB400. The formation of the
minor metabolite, 2,5-dichloroacetophenone is not shown.
The relative stereochemistry of the hydrox-
yl groups In both products has not been
firmly established. The cis orienta-
tlon(s) shown in Figure 2 is based on the
results of previous studies which have
shown that bacteria initiate the oxidation
of biphenyl and several other aromatic
hydrocarbons by enzymatically incorpor-
ating both atoms of oxygen into the aro-
matic nucleus to form dihydrodiols^in
which the hydroxyl groups have a cis-rel-
ative stereochemistry (9).
There are several possible explana-
tions for the results obtained in this
study. At the enzyme level, different
chlorine substitution patterns cause
steric effects and these could permit or
retard access of different PCBs to the
active site of the dioxygenase induced by
biphenyl. Thus the oxidation of 2,5,2',5'
-CB at the 3,4-positions may be catalyzed
by the 2,3-dihydroxybiphenyl dioxygenase
that is induced by growth in the presence
of biphenyl. Alternatively, a different
biphenyl oxygenase may be responsible for
the observed results. Different chlorine
substitution patterns may be responsible
for effects at the transcriptional level.
For example, a 2,5-substitution pattern
-400-
-------
could induce an enzyme with different pro-
perties and substrate specificity to the
enzyme induced by biphenyl. The difficul-
ties associated with the interpretation of
the results 'are due to the low levels of
substrates used and the length of time
required to demonstrate significant degra-
dation of individual PCB congeners. Stud-
ies currently being conducted are directed
towards the elucidation of these problems.
As described in the present study mutants
have been isolated that are defective in
the structural genes for the first four
enzymes of the biphenyl degradative path-
way. The use of these organisms in bio-
transformation experiments should provide
valuable information on the pathways util-
ized for the oxidation of different PCB
congeners.
Our future studies will be directed
towards cloning the genes for biphenyl di-
oxygenase into high expression vectors.
The availability of such strains will fa-
cilitate the isolation and characteriza-
tion of the initial dioxygenase and permit
unambiguous determination of the substrate
specificity of the enzyme for individual
PCB isomers and homologs. In addition,
such studies will provide valuable informa-
tion on the regulation of the enzyme(s) in-
volved in PCB metabolism. We anticipate
that this approach will lead to the con-
struction of bacterial strains that can ef-
ficiently biodegrade complex mixtures of
PCBs.
ACKNOWLEDGMENTS
This work was supported in part by
Grant CR812/27 from the Office of Research
and Development, the Environmental Protec-
tion Agency. We thank Dr. P.R. Sferra,
EPA Project Officer for his interest, sup-
port and suggestions. Initial studies on
the oxidation of 2,5,2',5'-tetrach1orobi-
phenyl by A^ eutrophus H850 at the Univei—
sity of Texas were supported by grant
A02-A00579000 from General Electric Com-
pany. We thank Catherine Potter for her
patience and skill in organizing and typ-
ing the manuscript.
REFERENCES
1. Bedard, Donna L., Michael J. Brennan,
and Ronald Unterman, 1984. Bacterial
degradation of PCBs: Evidence of
distinct pathways in Corynebacterium
sp. MB1 and A lealigenes eutrophus
HSJO. In Proceedings of" the 1983 PCB
seminar. Electric Power Research
Institute (G. Addis and R. Komai.
eds).Palo Alto, CA., pp4-101 to
4-118.
2. Bedard, Donna L., Ronald Unterman,
Lawrence H. Bopp, Michael J. Brennan,
Marie L. Haberl, and Carl Johnson,
1986. Rapid assay for screening and
characterizing microorganisms for the
ability to degrade polychlorinated
biphenyls. Appl. Environ. Microbio.
51, PP761-76S:
3. Bedard, Donna L., Marie C. Haberl,
Ralph J. May, and Michael J. Brennan,
1987. Evidence for novel mechanisms
of PCB metabolism in A lealigenes eu-
trophus H850, Appl. Environ. MicroT,
in press.
4. Bopp, Lawrence H., 1986. Degradation
of highly chlorinated PCBs by Pseudo-
monas strain LB400. J. Ind. Microbio-
logy 1, pp23-29.
5. Finette, Barry A., Venkiteswaran
Subramanian, and David T. Gibson,
1984. Isolation and characterization
of Pseudomonas put?da PpF1 mutants
defective in the,toluene dioxygenase
enzyme system. J_. Bacter iol. 160,
pp1003-1009.
6. Finette, Barry A., 1984. Molecular
characterization of the toluene (tod)
operon from Pseudotnonas put!da PpFI,
Ph.D. dissertation. The University of
Texas at Austin, Austin, Texas.
7. Furukawa, Kensuke, 1982. Microbial
degradation of polychlorinated
biphenyls (PCBs), In Biodegradation
and Detoxification of Environmental
Pollutants (A.M. Chakrabarty, ed.).
CRC Press, Inc., Boca Raton, Florida,
PP33-57.
8. Gibson, David T., Rowena L. Roberts,
Martha C. Wells, and Val M. Kobal,
1973. Oxidation of biphenyl by a
Beijerinckia sp., Biochem. Biophys.
Res. Commun. 50, pp211-219.
-401-
-------
9. Gibson, David T., and Venklteswaran
Subramanian, 1984. Microbial
degradation of aromatic hydrocarbons.
In Microbial Degradation of Organic
Compounds (D.T. Gibson, ed.), Marcel
Dekker, Inc., New York, ppl8l-252.
10. Hutzinger, Otto, Stephen H. Safe, and
Vladimir Zitko, 1974. Commercial PCB
preparations, properties and composi-
tion. In The_CjTem]^try_of_PCBs_, C.R.C.
Press, Cleveland, p8.
11. Kuner, Jerry M., Leon Avery, Douglas
E. Berg, and Dale Kaiser, 1981. Uses
of transposon Tng in the genetic
analysis of Myxococcus xanthus. In
Microbiology (D. Schlessinger, ed).
ASM Publications, Washington D.C.,
pp128-132.
12. Safe, Stephen H., 1984. Microbial
degradation of polychlorinated
biphenyls.In Microbial Degradation of
Organic Compounds (David T. Gibson,
ed).Marcel Dekker, Inc., New York,
PP361-370.
-402-
-------
ENGINEERING P450 GENES IN YEAST
C. Chena, C.R. Deya, V.F. Kalfaa, D. Sanglarda, T.R. Sutterb, T. Turia
and J.C. Lopera,b
aDepartment of Microbiology and Molecular Genetics
bDepartment °f Environmental Health
University of Cincinnati College of Medicine
Cincinnati, OH 45267-0524
ABSTRACT
Cytochrome P450 systems catalyze the monooxygenation of a broad range of xenobiotic
compounds. These systems are most extensively characterized in mammals, where for a given
species a single form of NADPH-cytochrome P450 oxidoreductase (P450R) donates reducing power
to any of 30-to-100 unique members of a P450 protein superfamily. We are interested in gene
engineering P450 systems in yeast for the oxidative detoxication and biodegradation of
environmentally stable organic pollutants. Two organisms, Saccharomyces cerevisiae (baker's yeast)
and Candida tropicalis ATCC750, a yeast capable of n-alkane assimilation, have been used as
models. S. cerevisiae is particularly useful for gene manipulation; C. tropicaZis is of interest for its
possible advantages for genetically engineering the uptake and catabolism of hydrophobic toxicants.
The major proteins of interest in these model yeasts are: P450R; the P450 lanosterol 14a-
demethylase, involved in sterol biosynthesis; and P450 n-alkane oi-hydroxylase.
Genes for these proteins have been isolated and determination of their DNA sequence has been
completed or is in progress. Ammo acid sequences deduced from these DNAs were compared to
sequences reported for mammalian P450 system proteins. Our results indicate that the P450R
protein is highly conserved among yeast and mammals. The yeast P450 proteins share patterns
observed for mammalian P450s, with sequence similarity among enzymes of similar function and
with wide sequence diversity between P450s of different substrate specificity. The yeast P450s for
lanosterol demethylation and alkane n-hydroxylation belong to two new families in the P450
superfamily. Characterization of these genes forms a basis for the gene engineering of P450
expression in yeasts.
INTRODUCTION
Viewed broadly in terms of the natural
cycles in the biosphere it is clear that nearly all
compounds are degraded. These biological
degradations are primarily effected by bacteria,
yeasts, molds and other microorganisms.
Included in the biodegradable category are a
variety of halogenated aromatic compounds.
Thus when environmental persistence of various
polychlorinated aromatic hydrocarbons is
encountered, it is reasonable .to consider the
isolation or modification of such organisms in
order to catalyze the entry of these persistent
compounds into the pre-existing degradative
pathways. Numerous national workshops and
symposia have proposed the development of
microbial biodegradation technologies for their
potential as low cost, environmentally sound
tools in the treatment of high priority pollutant
targets in land fills, waste sites, and industrial
effluents. (13,9).
Because of the highly substituted structure
of the persistent polychlorinated aromatics, the
P450 monooxygenases are likely candidates as
catalysts for their specific oxidation or
reductive dechlorination. Fungi and yeasts are
eukaryotie microorganisms which are known to
catalyze a broad range of these reactions using
-403-
-------
P450 enzymes (10) and mammalian P450s have
been implicated in the degradation of such
compounds as 2,3,7,8-tetrachlorodibenzo-p-
dioxin (8,4), 2,3,7,8-tetrachlorodibenzofuran (7),
and hexachlorobenzene (12). We are interested
In gene engineering yeasts to provide organisms
combining both the P450-catalyzed degradation
of persistent toxic compounds and the capacity
to survive or to survive and grow in
contaminated environments.
PURPOSE AND APPROACH
Fig. 1 shows a diagram of a cytoplasmic
membrane-bound P450 monooxygenase system.
In mammals as many as 30-to-100 different
P450s make up the P450 protein superfamily (5).
Each P450 receives electrons for its reaction
from a single NADPH-P450 oxidoreductase
(P450R), or in some cases from P450R plus a
second donor, cytochrome b5. The individual
P450 enzymes provide the specificity for a
broad range of oxidative reactions, shown here
as the hydroxylation of the substrate with the
concomitant formation of a water molecule.
Until recently the molecular genetics of
eukaryotic P450s was based entirely upon
studies of mammalian systems. Although major
P450 enzymes had been described for yeasts (2),
none of the relevant genes or the signals
controlling expression of those genes were
characterized. Since an understanding of these
genetic elements would provide a basis for
engineering P450s in yeasts, the goal of this
phase of our work has been the isolation and
molecular characterization of genes coding for
yeast P450 systems. Two organisms were used:
baker's yeast, Saccharomyces cerevisiae; and
Candida tropicalis ATCC750. S. cerevisiae is
highly characterized genetically and has become
the eukaryotic microorganism of choice for the
application of molecular genetic techniques. C.
tropicalis ATCC750 is representative of a group
of yeasts capable of growth upon n-alkanes in
petroleum as a carbon and energy source; such
yeasts already express lipophilic properties
which may be useful in gene engineering cells
for the uptake and metabolism of hydrophobic
hazardous compounds.
The major proteins of interest in both of
these yeasts are P450R and the P450 enzyme
lanosterol 14o-demethylase (14DM), an enzyme
well studied for its role in the biosynthesis of
ergosterol. Additionally C. tropicalis employs
the P450 enzyme u-hydroxylase (P450alk) as
catalyst of the initial step in n-alkane
catabolism. In the recently recommended
nomenclature for the P450 gene superfamily,
14DM and P450alk are designated as LIA1 and
LIB1 respectively (6).
RESULTS
Progress in our isolation and DNA
sequence characterization of these genes has
been described (3,1,11, and manuscripts
submitted for publication) and is summarized in
Table I. Access to the DNA sequence of a gene
makes possible the description of the deduced
amino acid sequence of the encoded protein.
Thus this summary Table I also indicates that we
have full or partial data for the deduced amino
NADPH.H +
NADP-
Fig. 1. A schematic of the membrane bound P450 monooxygenase system.
-404-
-------
Table 1. PROGRESS IN THE CHARACTERIZATION OF YEAST P450 GENES
Organism
Genes
Isolated Sequenced
References
S. cerevisiae
P45014DM
LIA1*
P450R
C. tropicalis
P450alk
LIB1*
P45014DM
P450R
X
X
X
in progress
X
in progress
in progress
Kalb et al., 1986
Kalb et al., manuscript submitted
Kalb and Loper, manuscript submitted
Sutter et al., unpublished
Sanglard et al. 1987
Sanglard et al., manuscript in
preparation
Chen et al., unpublished
Sutter et al., unpublished
*Nomenclature of Nebert et al., see ref. 6.
acid sequence of each of these proteins. We
have now begun to ask how these lower
eukaryotic P450R and P450 proteins compare to
the P450 system proteins in the P450
superfamily. Results of such comparisons are
presented here.
Our first test was to compare the deduced
ammo acid sequence of one mammalian P450
with all available protein sequences. Bovine
adrenal cortex P450C21 was arbitrarily chosen
for the mammalian P450 sequence. The pool of
protein sequences contained all of the several
thousand sequences in the National Biomedical
Research Foundation (NBRF) Protein Sequence
Data Bank Release 10.0, our data for two yeast
P450 proteins, and sequence data for three
mammalian P450s not yet included in the NBRF
compilation. The yeast P450 sequences were for
S. cerevisiae 14DM and C. tropicalis P450alk.
The algorithm used was that of Wilbur and
Lipman (14) implemented at Bionet (The
parameters used were: WORD-LENGTH, 1;
GAP-PENALTY, 2; WINDOW, 30; DENSITY,
Less; and FAST, No).
These comparisons appear as relatedness
scores in Fig. 2. The majority of proteins of
course show only random sequence
relatedness; for the total population the mean
score was 4.6 +/- 0.9 s.d. The horizontal bar
covers the location of all 22 of the eukaryotic
P450s in this distribution and illustrates the
wide diversity of protein sequence structure
among the P450 superfamily. Five entries are
plotted with values of >60; one of these is for
the P450C21 protein itself. The relatedness
scores for the remaining 17 P450s in this
population range from 60 down to 19, a score
four times that of random relatedness. The
relatedness scores of S. cerevisiae 14DM and C.
tropicalis P450alk, indicated as S. c. and C. t.
in Fig. 2, are 20 and 29 respectively. Although
relatively low, these scores are similar to the
score seen with bovine P450scc (score = 19).
Thus these two P450s from the lower eukaryotes
lie within the diversity range of the mammalian
P450s.
Homology and diversity among yeast P450s. A
relatedness score as discussed above gives an
overall average for two protein sequences.
Homology matrix comparisons between pairs of
-405-
-------
H
U
H
B
E
R
0
F
S
E
Q
O
E
H
C
E
S
1400-
1000-
500-
100-
50-
10-
5™*
«.
- *
i
0
* S.C.
* **
1 1
10 20
C.t.
** * **
1
30
*
*
** ** * ** *
1 1 1
40 50 60 >60
SCORE
Fig. 2. Relatedness scores of P450C21 with the sequences in the National Biomedical
Research Foundation protein database.
-406-
-------
P450s allow the examination of the localization
of similarities and differences across the length
of the amino acid sequences. Such comparisons
were obtained using the Pustell dot matrix
program (International Biotechnologies, Inc.; the
parameters used were: RANGE, 6; SCALE, 0.9;
HASH LEVEL, 1; JUMP LEVEL, 1; STEP, 1;
MINIMUM VALUE, 40; COMPRESSION, 5.)
Regions of homology between paired sequences
are indicated by the presence of a letter which
encodes a score proportional to the degree of
homology: A = 100%, B = 98-99%, Z = 50-51%, a
= 48-49%, b = 46-47% etc. Thus in this program
a self-comparison for any amino acid sequence
would generate a diagonal plot consisting of a
continuous string of A's.
Our data of deduced amino acid sequence
for the 14DM genes from both S. cerevisiae and
C. tropicaZis were used to examine the
relatedness of this P450 in the two yeast genera.
Since the present DNA sequence
characterization of the C. tropicalis 14DM gene
contains a few ambiguous bases, the homology
matrix comparison which resulted is only a close
approximation. However it is clear from the
pattern obtained that these two proteins have
extensive homology throughout their deduced
amino acid sequences, see Fig. 3A.
This homology for the 14DM proteins is in
marked contrast to the minimal relatedness seen
between the two C. tropicalis P450s, 14DM and
P450alk, see Fig. 3B. Again the homology
matrix comparison of these two proteins is only
a close approximation, pending resolution of a
few ambiguous nucleotide, residues in our
present DNA sequence for the C. tropicalis
14DM gene. This lack of pattern structure
reveals a remarkable sequence diversity for
these two P450s from the same yeast.
2. P450R proteins
Determination of the DNA sequence.for
the P450R gene from C. tropicalis is nearly
complete and sequencing of the S. cerevisiae
P450R gene is in progress. Preliminary data
indicate that these P450R proteins are quite
similar in protein sequence and also show high
homology to mammalian P450R proteins.
DISCUSSION
Isolation and DNA sequence determination
of the P450 genes for 14DM of S. cerevisiae and
for P450alk of C. tropicalis has allowed the first
comparisons of deduced protein sequences for
P450s of lower and higher eukaryotes. We have
shown that each of these P450s is a member of a
distinct P450 family (Kalb et al. manuscript
submitted for publication 1987, Sanglard et al.,
unpublished). Together with the preliminary
homology comparisons presented here involving
the C. tropicalis 14DM sequence, the data
strongly support the conclusion that P450
structure in yeasts resembles mammalian P450
structure. The two yeast 14DM P450s show
extensive homology, and in results not presented
here we have demonstrated that the 14DM gene
of C. tropicalis is functionally expressed in S.
cerevisiae and can compensate for 14DM-
def icient mutants in that yeast (C. Chen et al.,
unpublished). On the other hand, comparisons of
P450alk with deduced sequences of 14DM of
either yeast show great diversity, an observation
which is consistent with the diversity between
these P450s regarding substrate and reaction
specificity. That the structure for P450R
protein is highly conserved across yeasts and
mammals is of particular interest, since
molecular genetic tools for a detailed structure
function analysis can be applied to the yeast
enzyme as a model of all eukaryotic P450R,
including human.
Access to this DNA sequence data makes
possible the molecular genetic manipulation of
these genes. For example, based upon DNA
sequence information T. Turi has inactivated the
14DM gene in S. cerevisiae by gene disruption,
an experiment which showed that this gene is
essential for normal growth (Kalb et al.,
manuscript submitted for publication, 1987).
The availability of such genetically inactivated
strains provides an ideal background for tests in
which restoration of activity can be examined
under genetically altered levels of expression.
Also, the normal P450 gene can be mutated in
vitro and the altered construct put into the
tester strain and its level of function
determined. Access to the DNA sequence of
the C. tropicaZis gene for P450alk will facilitate
our testing of its expression as a foreign gene in
S. cerevisiae. Experiments using these methods
form important steps toward gene engineering
other specific P450s in yeasts for the
degradation of hazardous wastes.
ACKNOWLEDGEMENTS
This work was supported in part by
Cooperative Agreements CR810605 and^
CR813366 to JCL from the U.S. Environmental
Protection Agency, Office of Research' and
Development, Hazardous Waste Engineering
-407-
-------
S. cerevisiae 14DM AMINP ACID RESIDUE NUMBER
100
200
300
400
500
100
H
Q
H
200
n 30°
•H
H
§
•H
O 400
e
a
\
"Y
Yv
»H.
\
e
d
d
\
a
zb
Q
\
"x
i
Yx
SK
d
Ij
ED£
Ilvz
d
e
d
\.
"u
c
T
°X
?
t"
y.
FH
Hjs
\
a
Fig. 3A. Homology matrix comparisons of P45014DM from two yeasts.
-408-
-------
t. tropicalis 14DM AMINO ACID RESIDUE NUMBER
100
H
CO
200
nJ
o
in
to
•H
rH
(0
O
300
£ 400
O
500
100
200
300
400
d e
Fig. 3B. Homology matrix comparisons of two P450s from C. tropicalts ATCC750.
-409-
-------
Research Laboratory, Cincinnati, P.R. Sferra,
Project Officer. DS was supported in part by a
research fellowship from the Swiss National
Foundation. DS and VFK are postdoctoral
fellows and CC, TRS and TGT are Ph.D.
graduate students supported by their respective
Departments in the University.
References
1. Kalb, V.F., Loper, J.C., Dey, C.R., Woods,
C.W. and Sutter, T.R. 1986. Isolation of a
cytochrome P450 structural gene from
Saccharomyces cerevisiae. Gene 45,237-
245.
2. KSppeli, O. 1986. Cytochrome P450 of
yeasts. Microbiological Reviews 50,244-
258.
3.
4.
5.
6.
Loper, J.C., Chen, C. and Dey, C.R. 1985.
Gene engineering in yeast for
biodegradation: Immunological cross-
' reactivity among cytochrome P450 system
proteins of a Saccharomyces cerevisiae
and Candida tropicalis. Hazard. Wastes
and Hazard. Mat. 2,131-141.
Neal, R., Gasiewicz, T., Geiger, L., Olson,
J., and Sawahata, T. Metabolism of
2,3,7,8-tetrachlorodibenzo-p-dioxin in
mammalian systems, in Biological
Mechanisms of Dioxin Action, Bambury
Report 18, A. Poland and R. Dikinbrough,
eds., Cold Spring Harbor Laboratory, 1985.
Nebeft, D.W. and Gonzalez, F.J. 1985.
Cytochrome P450 gene expression and
regulation. TIPS 6,160-164.
Nebert, D.W., Adesnik,
T?ofoHi*r»(-*lr "D IA7 Cinnw.
Sato, R., & Waterman, M.R. 1987.
P450 gene family, recomemmended
nomenclature. DNA 6,1-11.
Poiger, H., Buser, H.R. and Schlatter, C.H.
1984. The metabolism of 2,3,7,8-
tetrachlorodibenzofuran in the rat.
Chemosphere 13.351-357.
Poiger, H. and Buser, H. The metabolism
of TCDD in the dog and rat, in Biological
Mechanisms of Dioxin Action, Bambury
Report 18, A. Poland and R. Dikinbrough,
eds., Cold Spring Harbor Laboratory, 1985.
9. Proceedings of USEPA Workshop on
Biotechnology and Pollution Control,
March, 1986.
10. Rosazza, J.P. and Smith, R.V. 1979.
Microbial models of drug metabolism,
Advances in Applied Microbiology 25,169-
208.
11. Sanglard, D., Chen, C. and Loper, J.C.
1987. Isolation of the alkane inducible
cytoehrome P450 (P450alk) gene from the
yeast Candida tropicalis. Biochem.
Biophys. Res. Comm. in press.
12. Takazawa, R.S. and Strobel, H.W. 1986.
Cytochrome P450 mediated reductive .
dehalogenation of the perhalogenated
aromatic compound hexachlorobenzene.
Biochemistry 25,4804-4809.
13. USEPA, Report of the Research Planning
.» Workshop on Bioavailability of Dioxins,
EPA/600/9-86/004 Washington, DC. 56 pp.
' 1986.
14. Wilbur, W.J. and Lipman, D.J. 1983.
Rapid similarity searcher of nucleic acid
protein data banks. Proc. Natl. Acad. Sci.
USA 80,726-730.
-410-
-------
BIODEGRADATION OF ORGANOPOLLUTANTS BY
PHANEROCHAETE CHRYSOSPORIUM! PRACTICAL CONSIDERATIONS
John A. Bumpus, Tudor Fernando, Gerald J. Mileski, Steven D. Aust
Department of Biochemistry
Michigan State University
East Lansing, MI 48824
ABSTRACT
We previously reported that a wide variety of structurally diverse organopollutants
are mineralized by the lignin degrading system of the white rot fungus Phanerochaete
chrysosporium (Science 228, 1434, 1985). Current research is directed towards applica-
tion of this technology for the biodegradation of environmental pollutants and hazardous
wastes. The system is effective in both liquid and solid matrices. Bulking agents such
as wood chips or corn cobs can also serve as a carbon source for the fungus. Degradation
of chemicals is supported by a carbon source for the fungus but readily available carbon
sources such as glucose do not support sustained rates of degradation. Sustained rates
were obtained with complex carbohydrates including natural sources. Rates of degradation
increased with respect to the concentration of chemical. Degradation of mixtures often
preceded faster than the rate of degradation of pure chemicals. For example, the mineral-
ization of pure 2,4,5,2',4',5'-hexachlorobiphenyl preceded much slower (1.1? in 30 days)
than did the mineralization of Aroclor 1254 and 1242 (14.3? and 20.3?, respectively, in
30 days). Two percent of pure 1^C-naphthalene was mineralized in 30 days but 32? was
mineralized when the ]^C-naphthalene was present in coal tar contaminated soil. Toxicity
of chemicals to the fungus was rare but could be circumvented. The fungus would grow in
the presence of used motor oil or coal tar contaminated soil.' The toxicity of the fungi-
cide pentachlorophenol (PCP) was reduced by starting with mature mycelia instead of
fungal spores. Under these conditions, the fungus continued to mineralize substantial
amounts of PCP at concentrations up to 100 ppm. For example, when the initial concentra-
tion of PCP was 100 ppm, greater than 20? of the PCP initially present was mineralized in
30 days. Furthermore, disappearance studies showed that approximately 95? of the PCP
initially present had been metabolized. '
INTRODUCTION
The wood rotting fungus Phanerochaete
chrysosporium is a member of the Basidiomy-
cotina. Recent studies in our laboratory
(2-9) and by others (1,11,12, 20) have
shown that under nutrient nitrogen limit-
ing conditions, this microorganism is able
to degrade a wide variety of structurally
diverse and environmentally persistent
organopollutants to carbon dioxide. The
ability to degrade such a wide variety of
chemicals has been found to be due, at
least in part, to the lignin degrading
system of this funugs (1-9,11,12,20).
Lignin is a complex heteropolymer and is
possibly the most difficult-to-degrade,
naturally occurring organic compound (10).
When nutrient nitrogen is depleted in '
cultures of P_. chrysosporium, this fungus
secretes a family of peroxldases that are
able to catalyze the oxidative
depolymerization of lignin as well as a
number of other reactions which occur
during the biodegradation of lignin
(21,22). The peroxidases, which are
commonly referred to as ligninases, are
also able to catalyze the initial
oxidation of many hard-to-degrade.
organopollutants (5,12,13,20). This is of
-411-
-------
great significance because the initial
oxidation of a chemical is often the most
difficult step in its biodegradation.
Because of its broad biodegradative
abilities, we have proposed that it may be
possible to develop waste treatment
systems based on the use of this microorga-
nism in axenic culture or in conjunction
with other microorganisms (2). Also, the
use of ligninases as additives may prove
to be a viable option in some waste treat-
ment processes.
Although the use of P_. chrysosporium
in the treatment of organochemical wastes
holds great potential, in practice there
are a number of factors which must be
addressed before this microorganism or its
extracellular ligninases can be used in
such systems. The present study addresses
some of these concerns.
METHODS AND MATERIALS
Phanerochaete chrysosporium
(BKM-F-1767) was obtained from the United
States Department of Agriculture, Forest
Products Laboratory (Madison, WI). This
fungus was maintained on malt agar slant
cultures and was stored at room tempera-
ture until used. Subcultures were made
routinely every'30-60 days.
Radiochemicals
Carbon-14 labeled Aroclor 1242 (33
mCi/mmol), Aroclor 1254 (32 mCi/mmol) and
DDT (1,1,1-trichloro-2,2-bis(4-chloro-
phenyl)ethane) (75 mCi/mmol) were obtained
from the Amersham Co. (Arlington Heights,
IL) while carbon-14 labeled POP (penta-
chlorophenol) (10.57 mCi/mmol), DDE (1,1-
dichloro-2,2-bis(4-chlorophenyl)ethene),
2,4,5,2',4',5'-HCB (2,4,5,2',4',5'-hexa-
chlorobiphenyl (13.09 mCi/mmol) and
glucose (1.88 mCi/mmol) were obtained from
Pathfinder Laboratories Inc. (St. Louis,
MO). Carbon-14 labeled dicofol (2,2,2,-
trichloro-1,1-bis(4-chlorophenyl)ethanol)
(9.78 mCi/mmol) was a gift from the Rohm
and Haas Co. (Springhouse, PA). The
purity of radiolabeled chemicals was
monitored via thin layer chromatography
(TLC) or by high performance liquid
chromatography (HPLC) using solvent
systems recommended by the vendors. When
necessary, these compounds were repurified
by TLC. The purity of radiolabeled com-
pounds used in these experiments was 98%
or greater. For the special case of the
polychlorinated biphenyl mixtures (Aroclor
1254 and Aroclor 1242), the radiolabeled
mixtures were compared with authentic
unlabeled Aroclor 1242 and Aroclor 1254 by
gas liquid chromatography. These studies
showed that, although some differences
were apparent, the radiolabeled Aroclors
appeared to be very similar to their
unlabeled counterparts.
Culture Conditions and Mineralization
Studies
For most experiments, P_. chrysospori-
um was incubated at 37°-39°C'in 10 ml of
nutrient nitrogen deficient culture medium
as described by Kirk et al. (16). This
medium consists of 56 mM glucose; 1.2 mM
ammonium tartrate, trace elements and
thiamine (1 mg/L) in 20 mM dimethylsucci-
nate buffer, pH 4.2. Cultures were estab-
lished by inoculating this medium with
spores as described (16). In most experi-
ments, 1^C-labeled chemicals, in a minimal
volume of acetone, were added at this time.
During the first three days of incubation,
cultures were grown under ambient atmos-
phere in 250 ml Wheaton Bottles equipped
with a gas exchange manifold. After three
days and at three day intervals thereaf-
ter, cultures were flushed with oxygen.
The atmosphere from each culture was
forced through 10 ml of an ethanolamine
containing scintillation cocktail which
served as a carbon dioxide trap.
Carbon-14 labeled carbon dioxide was quan-
titated'by liquid scintillation spectrome-
try. Details of this mineralization assay
have been previously described (2,16).
High Performance Liquid Chromatography
The amount of PCP remaining following
its incubation in ligninolytic cultures of
£. chrysosporium was determined using a
Beckman HPLC system equipped with an
Alltech R-Sil C-18 reverse phase column
(4.6 x 250 mm). 'Isocratic elution was
performed using acetonitrilezHgOtglacial
acetic acid (75:25:0.125) at a flow rate
of 2 ml/min. Elution'of PCP was monitored
at 238 run. Quantitation of PCP was accom'-
plished using a Hewlett-Packard model
3390A integrator.
-412-
-------
In a typical experiment, a culture
that had been incubated with 1 or 100 ppm
of PCP was poured into a 30 ml tissue
grinder. The culture bottle was then
rinsed with 5 ml of ethanol and then with
5 ml of water. Both the ethanol and water
rinses were added to the tissue grinder.
The fungal mats were then homogenized with
five strokes using a motor driven teflon
pestle and the final volume was measured. '
One ml of NaCl saturated water was mixed
with each aliquot prior to extraction with
hexane. When the initial concentration
of PCP in culture was 1 ppm, 4-5 ml of
homogenate was extracted with 1 ml of
hexane by mixing on a vortex for 30 s.
The aqueous and organic phases were then
separated by centrifugation. Typically,
50-100 ul allquots of hexane were used for
PCP determinations. Cultures in which the
initial concentration of PCP was 100 ppm
were extracted in a similar manner except
that 0.5 ml of homogenate was extracted
with 2 ml of hexane.
RESULTS AND DISCUSSION
Biodegradative Abilities of P.
chrysosporlum
The development of waste treatment
systems based on the use of £. chrysospori-
um is an attractive concept for many
reasons. For example, a wide-variety of
structurally diverse organopollutants,
including some of the most recalcitrant
environmental pollutants known, are
degraded to carbon dioxide (i.e., mineral-
ized) by this fungus (1-9,11,12,20). The
fact that these compounds are degraded to
carbon dioxide is an important concept
because it demonstrates that a metabolic
pathway exists for the complete biodegrada-
tion of each of these compounds. It also
implies that intermediates formed during
biodegradation of the parent compound are
also metabolized. It is critical that
this occur because metabolites of some
organopollutants may be as toxic, or even
more toxic than the parent compound. They
may also be as difficult to degrade.
Table 1 shows that dicofol (2,2,2-trichloro
1,1-bis(4-chlorophenyl)ethanol) a metabo-
lite formed by £. chrysosporium during DDT
degradation is itself readily mineralized.
Table 1 also demonstrates that DDE is
mineralized by this fungus. Although DDE
is not a substantial DDT metabolite in
this system, it is a major bacterial (17)
and mammalian (19) metabolite of DDT. It
may also be formed from DDT in the environ-
ment and it is at least as toxic and as
environmentally persistent as DDT (15,17,
18). Also sites contaminated with DDT
would be expected to be contaminated with
DDE, especially when these sites have been
contaminated with technical preparations
of DDT which contain substantial amounts
of DDE initially.
The observation that metabolites of
organopollutants are themselves metabo-
lized in consistent with the report by
Sanglard et al. (20) in which it was shown
that water soluble metabolites of benzo(a)-
pyrene were formed following the formation
of organic soluble metabolites in nutrient
nitrogen deficient cultures of this fungus.
The organic soluble benzo(a)pyrene metabo-
lites were subsequently identified as the
benzo(a)pyrene 1,6-, 3,6- and 3,12-quin-
ones (12). It is interesting to note
that, unlike the reaction products of the
cytochrome P-450 monooxygenases of mammali-
an systems, the highly carcinogenic 7,8
-diol- 9,10-epoxide of benzo(a)pyrene was
not reported to be a benzo(a)pyrene metabo-
lite in this fungus. Hammel et al. (13)
have also presented data which indicate
that the mutagenic intermediates of pyrene
(pyrene-1,6-dione and pyrene-1,8-dione)
are also further metabolized by P_.
ohrysosporlum. It is also important to
note that mutagenic or otherwise toxic
metabolites formed by the ligninases of j?.
chrysosporlum would be formed extracellu-'
larly thus limiting their interaction with
intracellular components such as DMA.
Enhancement of Biodegradation by P.
chrysosporium
The initial report (2) from our labo-
ratory was designed to show that a number
of environmentally persistent organo-
halides, as well as the polycyclic aromat-
ic hydrocarbon, benzo(a)pyrene, could be
degraded by £. chrysosporium in nutrient
nitrogen deficient cultures. Although
mineralization did occur, the rate of
mineralization in some cases was relative-
ly slow. For example, only 9.3? of the
1^C-benzo(a)pyrene initially present was
mineralized during 30 days of incubation.
Because of the desirability of developing
systems which rapidly degrade organopollu-
tants, we have initiated studies which
focus on ways to increase the rate of bio-
degradation of organopollutants using this
microorganism. However, it should be
-413-
-------
TABLE 1. MINERALIZATION OF DDE AND DICOFOL BY P_. CHRYSOSPORIUMa
Initial
Concentration
(nmoles/culture)
Amount
Mineralized13
(pinoles/culture)
Mineralized13
DDE
Dicofol
15.0
5:0
714 ± 180
800 ± 232
it. 8 ± 1.2
16.0 ±4.6
a. Each culture was grown in 10 ml of the nutrient nitrogen limiting media described by
Kirk et al. (17).
b. These values represent the amount or percent of 1l*C-labeled substrate that was
converted to 1l|C02 in 30 days (mean +_ standard deviation, N = 4).
noted that if waste treatment systems can
be developed which produce sustained
levels of biodegradation at the low rates
already demonstrated, even these systems
would be sufficient to ensure the timely
destruction of these chemicals.
Efforts to increase the rate of bio-
degradation of organopollutants by P_.
ohrysosporium have employed a number of
strategies. For example, we have shown
that increasing the concentration of
growth substrate (glucose) from 26 mM to
112 mM resulted in a 2.5 fold increase in
the amount of 1l*C-DDT mineralized during a
30 day incubation period (3). An interest-
ing facet of these studies is the fact
that the initial rate of mineralization
(i.e. the rate observed between day 3 and
day 18) was the same for all of the glu-
cose concentrations studied. The differ-
ence in the amount of 114C-DDT mineralized
was due to the fact that, in cultures con-
taining low (28 mM and 56 mM) glucose con-
centrations, the rate of mineralization
rapidly declined after 18 days of incuba-
tion whereas this decline was not as pro-
nounced in cultures containing higher con-
centrations of glucose. These results
suggest that the ability to mineralize
organopollutants is dependent upon the
availability of a carbon source such as
glucose which can serve as a growth sub-
strate. This hypothesis was confirmed in
studies in which supplemental glucose was
shown to increase the rate and extent of
I^C-DDT mineralization when added to cul-
tures in which the glucose concentration
had been depleted (2,3).
The ability of supplemental glucose
to increase the rate and extent of bio-
degradation may be due to at least two
phenomena. First, the addition of supple-
mental glucose may simply increase the
overall rate of fungal metabolism and, as
a consequence, organopollutants or their
metabolites are mineralized at a faster
rate. Secondly, since hydrogen peroxide
is a required cosubstrate for the lignin-
ases which are responsible for the initial
oxidation of many organopollutants, it is
possible that the addition of supplemental
glucose allows the fungus to produce the
required hydrogen peroxide as a co-product
of the glucose oxidase system.
The suitability of using more complex
carbohydrates to support growth and miner-
alization of organopollutants has also
been studied. Figure 1 demonstrates that
1? cellulose and 1$ starch are both sub-
stantially better than 1$ glucose in
supporting mineralization of 1^C-DDT.
Possibly the most important aspect of this
experiment is the fact that, although the
initial rate of mineralization with all
three substrates was nearly identical, the
rate of mineralization when starch or
cellulose served as growth substrate did
not decline as rapidly as when glucose was
used. Furthermore, in the cultures con-
taining cellulose or starch, substantial
amounts of 1l*C-DDT continued to be miner-
alized for the duration of the 90 day
incubation period.
In addition to cellulose and starch,
a number of inexpensive and readily avail-
able cellulosic materials were tested for
their ability to support growth and
-414-
-------
g
8 80
0 12 24 36 4« 60 ' 72 84
TINE (DAYS)
Figure 1. Mineralization of ^C-DDT by J?.
ohrysosporium using 1% starch (closed ~
squares), cellulose (closed circles) or
glucose (closed triangles) as growth
substrate. The initial concentration of.
14C-DDT was 365 pmoles/culture.
mineralization. Figure 2 shows
that 5% (w/v) wheat straw, used newspaper
and corn cobs all support growth of the
fungus and mineralization of 1l)C-DDT.
Although these materials were less effi-
cient that starch or cellulose in the
support of 1^C-DDT mineralization, they
did support substantial amounts of miner-
alization (13.2? ± 5.8, 8.1? ± 5.0, and
8.2? ± 3.3, respectively). This, coupled
with the fact these materials are inexpen-
sive and readily available suggests that
they may be suitable for use as growth
substrates in a practical waste treatment
system.
Biodegradation of Complex Mixtures
The lignin (and organopollutant)
degrading system of P_. chrysosporium is
relatively non-specific and non-stereo-
selective. This property is very impor-
tant because, unlike other microorganisms,
one would expect that this fungus might be
able to degrade complex mixtures of xeno-
biotics. Indeed, Table 2 demonstrates
that 1^C-Aroclor 125*1 (a mixture of
tetra-, penta- and hexa-chlorobiphenyls)
and 1^C-Aroclor 1242 (a mixture of mono-,
di-, tri- and tetra-chlorobiphenyls) under-
go substantial mineralization in nutrient
nitrogen deficient cultures of P. chryso-
1«
f 30
B
§
12 II
TME (MYS)
Figure 2. Mineralization of ,1l*C-DDT by P_.
chrysosporium grown on 5% wheat straw .~-
(closed circles), newspaper (closed
squares) or corn cobs (closed triangles).
The initial concentration of 11(C-DDT was
365 pmoles/culture.
sporlum. These studies are in good agree-
ment with the work of Eaton (11) who first
showed that 1l*C-Aroclor 1254 was degraded
by £. chrysosporium. ,It is interesting-to
note that mineralization of the Aroclor ... .
mixtures proceeded more rapidly than minerr
alization of pure 1^C-2,4,5,2',4',5'-hexa-
chlorobiphenyl. The reason for this is,
at present, unknown. It may be that this
highly chlorinated biphenyl congener is
simply more resistent to degradation than
other congeners in these PCB mixtures.: On
the other hand, It is interesting to specu-
late that certain individual components of
complex mixtures may exert synergistic. .
effects on the oxidative biodegradation of.
other components in the mixture. There is
precedent for such interactions. For exam-
ple, addition of veratryl alcohol,has,been
shown to increase the rate and extent,of.
degradation of benzo(a)pyrene (12) and
4-methoxymandelic acid (4). We have .
observed a similar phenomena. Carbon-14
labeled naphthalene is mineralized slowly
(255/30 days) in nutrient nitrogen defi-
cient cultures of P. ohrysosporium. How-
ever, when 1''c-naphthalene was added to •
cultures containing coal tar contaminated
soil, Tween 20 and veratryl alcohol, 32?
of the 1^C-naphthalene was mineralized
during the same incubation period (.Bumpus,,
and Aust, unpublished observation).
-415-
-------
TABLE 2. MINERALIZATION OF BIPHENYL AND POLYCHLORINATED BIPHENYLS BY £. CHRYSOSPORIUMa
Initial
Concentration
(nmoles/oulture)
Amount
Mineralized13
(pmoles/culture)
% Mineralized13
Biphenyl
Aroclor 1212
Aroclor 1254
2,H,5,2',4',5'-HCBC
1.25
1.-25
1:25
5^0
367 ± 91
254 ± 19
179 ± 47
'57 ± 20
29. 4 ± 7.3
20:3 ± 0:2
14.3 ± 3:8
1.1 ± 0:4
a. Culture conditions were the same as described in Table 1.
b. These values represent the amount or percent of 1l*C-labeled substrate that was
converted to 1l|C02 in 30 days (mean ± S.D., N = 4 for Biphenyl, Aroclor 1254 and
2,11,5,2',I1,5'-HCB, N = 3 for Aroclor 1242).
o. Because of the low amounts of 2,4,5,2',4I,5'-HCB that were mineralized, this
experiment was continued for a total of 115 days. At the end of this extended
incubation period 126.0 ± 32.0 pmoles (2;5 ± 0.6?) of the 2,4,5,2',4I,5'-HCB initially
present had been mineralized. During this period, supplemental glucose (56 mM) was
added at 30 day intervals.
Toxloity Problem
A prerequisite for any microorganism
used in waste treatment systems is the
ability to survive in the presence of the
organopollutants that it is intended to
degrade. Our toxicity studies have
focused of the ability of £. chrysosporium
to mineralize the fungicide; pentachloro-
phenol. We have shown that, like a number
of other organopollutants, '^c-PCP mineral-
ization is promoted in nutrient nitrogen
deficient cultures of £. chrysosporium
whereas mineralization is suppressed in
nutrient nitrogen sufficient cultures,
thus suggesting that mineralization is
due, at least in part, to the lignin
degrading system of this funugs (9).
Furthermore, our studies demonstrated that
the initial rate and the extent of mineral-
ization increased with respect to I^C-PCP
concentration over the range of 16.6 ppb
to 333 ppb (Bumpus and Aust, unpublished
observation). Attempts to study 1J*C-PCP
mineralization at higher concentrations
(i.e. in the parts per million range) were
hindered because PCP concentrations higher
than 4 ppm were lethal to the fungus when
cultures were intiated with spores. The
problem of lethality was circumvented by
allowing cultures of P_. chrysosporium to
grow for six days before adding PCP. In
these studies (Table 3), in which toxicity
was measured by the ability of PCP to
inhibit respiration (i.e. conversion of
1llC-glucose to 1^002),'it was shown that
although PCP concentrations between 10 and
100 ppm caused substantial inhibition of
respiration, they were not lethal to the
fungus. And, significantly, substantial
amounts of 14c-PCP were still degraded as
shown in disappearance and mineralization
studies.
In other studies we have shown that
this microorganism was able to survive in
the presence of a number of relatively
adverse conditions. For example, we have
shown this fungus was able to grow in the
presence of high concentrations (300 ppm)
of DDT, used motor oil (20% w/w) and coal
tar contaminated soils.
Taken together, these results suggest
that conditons may be developed which will
allow this fungus to grow and degrade
organopollutants in a number of potential-
ly toxic environments. The suitability of
using this microorganism in a given situa-
tion will, of course, require individual-
ized study of the site in question.
-416-
-------
TABLE 3- Mineralization and Disappearance of PCP by P. chrysosporiuma
Initial Concentration
of PCP
nmoles/culture
ppm
Amount Mineralized
nmoles/culture
% Mineralized
% Disappearance
37.5
375:0
3750.0
1
10b
100b
18.8 ±1.9
'158 ± 45
841 ±615
50.1 ± 5.1
42.1 ± 12.0
22.- 4 ± 16.4
100°
N.D.d
95
a. Each culture (10 ml) was allowed to grow for 6 days at which time PCP was added in a
minimal volume of acetone (< 20 ul). Other culture conditions were the same as those
described in Table 1 . Cultures were incubated for 30 days.
b. PCP caused 57.1 ± 37% and 74.3 ± 12.5? inhibition of the initial rate of respiration
in cultures containing 10 and 100 ppm PCP, respectively. Respiration was measured as
the amount of ^C02 evolved from 1^C-glucose per day.
c. Detection limit = 50 ppb
d. Not determined.
ACKNOWLEDGEMENTS
This research was supported by Cooper-
ative Agreement CR813369, U.S. Environment-
al Protection Agency, Office of Research
and Development, Hazardous Waste Engineer-
ing Research Laboratory, Cincinnati, OH,
P.R. Sferra, Project Officer. The authors
also wish to thank Teresa Vollmer for her
expert secretarial assistance.
REFERENCES
Arjmand, M. and H. Sandermann, 1985.
Mineralization of Chloroaniline/Lig-
nin Conjugates and of Free Chloroani-
lines by the White Rot Fungus Phanero-
chaete chrysosporium. J. Agric. Food
Chem. 33:1055-1060. ' '
Bumpus, J.A., M. Tien, D. Wright, and
S.D. Aust, 1985a. Oxidation of Per-
sistent Environmental Pollutants by a
White Rot Fungus. Science 228:1434-
1436.
Bumpus, J.A. and S.D. Aust, 1985.
Studies on the Biodegradation of
Organopollutants by a White Rot
Fungus. Proceedings of the Interna-
tional'Conference on New Frontiers
for Hazardous Waste Management.
United States Environmental
Protection Agency, EPA/600/9-85/025,
Pittsburgh, PA, pp404-4!0.
Bumpus, J.A., M. Tien, D. Wright, and
S.D. Aust; t985b. Biodegradation of
Environmental Pollutants by the White
Rot Fungus Phanerochaete
chrysospor ium. Proceedings of the
Eleventh Annual Research Symposium on
Incineration and Treatment of
Hazardous Waste. United States
Environmental Protection Agency,
EPA/600/9-85/028, Cincinnati, OH,
pp120-126.
Bumpus, J.A. and S.D. Aust, 1986a.
Biological Oxidations by Enzymes from
a White Rot Fungus. American
Institute of Chemical Engineers 1986
Summer Meeting, Boston, MA, paper
86c.
Bumpus, J.A. and S.D. Aust, 1986b.
Biodegradation of Chlorinated Organic
Compounds by Phanerochaete
chrysosporium, A Wood Rotting Fungus.
Solving Hazardous Waste Problems (ACS
Symposium Series), (J.H. Exner, ed.),
ACS Books, Washington, D.C. (In
Press). ' '
-417-
-------
7. Bumpus, J.A. and S.D. Aust, 1987a.
Biodegradation of Environmental Pollu-
tants by the White Rot Fungus Phanero-
chaete chrysospor ium: Involvement of
the Lignin Degrading System.
BioEssays 6:166-170.
8. Bumpus, J.A. and S.D. Aust, 1987b.
Biodegradation of DDT (1,1,1-tri-'
chloro-2,2-bis (1-chlorophenyl) ethane
by the White Rot Fungus Phanerochaete
chrysosporium. (Submitted).
9. Bumpus, J.A. and S.D. Aust, 1987o.
Mineralization of Recalcitrant Envi-
ronmental Pollutants by a White Rot
Fungus. Proceedings of the National
Conference on Hazardous Wastes and
Hazardous Materials, Hazardous Materi-
als Control Research Institute,
Silver Spring, MD. Library of
Congress Catalog No. 87-80469, pp.
116-151.
10. Crawford, R, 1981. Lignin Biodegrada-
tion and Transformation. John Wiley
and Sons, Inc., NY, p15l.
11. Eaton, D.C., 1985. Mineralization of
Polychlorinated Biphenyls by Phanero-
chaete chrysosporium; A Ligninolytic
Fungus. Enzyme Microb. Technol.
7:191-196.
12. Haemmerli, S.D., M.S.A. Leisola, D.
Sanglard, and A. Fiechter, 1986.
Oxidation of Benzo[a]pyrene'by Extra-
cellular Ligninases of Phanerochaete
chrysosporium; Veratryl alcohol and
Stability of Ligninases. J. Biol.
Chem. 261:6900-6903.
13. Hamrael, K.E., Kalyanaraman and T.K.
Kirk, 1986.' Oxidation of Polycyclic
Aromatic Hydrocarbons and Dibenzo[pJ-
dioxins by Phanerochaete chrysospori-
um Ligninase. J. Biol. Chem.
261:16918-16952.'
11. Harvey, R.J., H.E. Schoemaker, and
J.M. Palmer,' 1986. Veratryl Alcohol
as a Mediator'and'the Role of Radical
Cations in Lignin Biodegradation by
Phanerochaete chrysosporium. FEES
Lett. 195:212-216.
15. Johnson, R.E, 1976. DDT Metabolism
in Microbial Systems. Residue
Reviews: Residues of Pesticides and
Other Contaminants in the Total
Environment. 61:1-28.
16. Kirk, T.K., E. Schultz, W.J. Connors,
' L.F. Lorenz, and J.G. Zeikus, 1978.
Influence of Culture'Parameters on
Lignin Metabolism by Phanerochaete
chrysospor ium. Arch. Microbiol.
117:277-285. '
17. Lai, R. and D.M. Saxena, 1982. Accum-
' ulation, Metabolism and Effects of
Organochlorine Insecticides on Micro-
organisms. Microbiological Reviews
16:95-127.'
18. Muir, D.C.G., A.L. Yarechewski, R.L.
' ' Corbet, G:R;B. Webster, and A.E. ' •
Smith, 1985: 'Laboratory and Field
Studies'on the Fate of 1,3,6,8-Tetra-
chlorodibenzo-p-dioxin in Soil and
Sediments. J. Agric. Food Chem.
33:518-523. '
19. Peterson, J.G. and W.H. Roblson, 1961.
' ' Metabolic Products of p,p'-DDT in'the'
Rat. Toxicology and Applied Pharma-
cology 6:321-327.
20. Sanglard, D., M.S.A. Leisola, and
' A.Fiechter, 1986. Role of Extracellu-
lar Ligninases in Biodegradation of
Benzo[a]pyrene by Phanerochaete
chrysosporium. Enzyme Microbiol.
Teohnol. 8:209-212.
21. Tien, M. and T.K. Kirk, 1983. Lignin
' ' Degrading Enzyme from the Hymenomy-
cete Phanerochaete ohrysosporium
Burds. Science 221:661-663.
22. Tien, M. and T.K. Kirk, 1985. Lignin-
Degrading Enzyme'from Phaneroohaete
chrysospor ium; Purification, Charac-
terization, and Catalytic Properties
of a Unique H202~Requiring Oxygenase.
Proc. Natl. Acad. Sci. USA 81:2280- '
2281; '
-418-
-------
GROWTH OF THE WHITE-ROT FUNGUS PHANEROCHAETE CHRYSOSPORIUM IN SOIL
Richard T. Lamar1, Michael J. Larsen1, T. Kent Kirk1 and John A. Glaser2
1USDA Forest Products Laboratory, Madison, Wisconsin 53705
2Hazardous Waste and Engineering Research Laboratory,
Cincinnati, Ohio 45268
ABSTRACT
Phanerochaete chrysosporium is a white-rot fungus with a demonstrated ability to degrade chlorinated organ-
ics in pure liquid culture to carbon dioxide. This ability suggests that the fungus may have potential as an in situ
hazardous waste degrader. However, no data exist regarding the ability of P. chrysosporium to survive and
grow in soil. That information is required for an effective evaluation of the ability of the fungus to degrade orga-
no-pollutants in situ. The objective of this study was to investigate the influence of soil biotfc and abiotic factors
on survival and growth of the organism. This paper will summarize our research results to date on the effects of
soil type, temperature, water potential and acidity on growth of the fungus in sterile soils.
INTRODUCTION
The fungus Phanerochaete chrysosporium Burds.
(Burdsall and Eslyn 1974) (Basidiomycotina, Corticia-
ceae), has been shown to oxidize several halogenated
aromatic pollutants to carbon dioxide and has been
proposed as an agent for biological treatment of recal-
citrant organo-halides (Eaton 1985; Bumpus et al.
1985; Huynh et al. 1985). On-site processing or in
situ hazardous waste treatment methods that employ
P. chrysosporium will require application of the fun-
gus to soil for treatment of contaminants. Phanero-
chaete chrysosporium is a white-rot wood decay fun-
gus that is often found in wood chip storage piles
(Burdsall and Eslyn 1974). Its growth in soils has not
been reported. Therefore, the ability of the fungus to
grow in soils, as well as abiotic and biotic soil factors
affecting fungal survival, growth and degradative abil-
ity must be determined before the possibility of devel-
oping a biological treatment method can be evaluated.
The investigations reported in this paper were de-
signed to assess the effects of various soil types, wa-
ter potentials and temperatures on growth of P. chry-
sosporium in sterile soils. Hazardous waste sites oc-
cur on a great variety of soil types. Soil water potential
and temperature are major factors influencing both
fungal growth and decomposition processes (Cooke
and Rayner 1984). These parameters were chosen for
the initial studies described here, because an assess-
ment of their influence should yield information criti-
cal to future investigations into the degradative ability
of P. chrysosporium in soil, and ultimately to the de-
velopment of a biological treatment method.
Materials and Methods
Inoculum Preparation
Phanerochaete chrysosporium Burds. (BKM F-
1767; ATCC no. 24725) was maintained at room tem-
perature (25°C) on 2% malt agar slants (Kirk et al.
1978). Soil inoculum consisted of aspen (Populus
tremuloides Michx.) pulpwood chips (1.5 x 0.5 x
0.25 cm3) thoroughly grown through with P. chrysos-
porium . Chips were inoculated with mycelial sus-
pensions of the fungus as follows: The fungus was
cultured in 125-ml Erlenmeyer flasks containing 10 ml
of basal medium (Kirk etal. 1978). Inoculum for
these starter cultures consisted of conidial suspensions
taken from two-week-old slants (Kirk et aL 1978).
After incubating at 39°C for 1.5 d in Erlenmeyer
flasks, mycelial mats were collected, washed with
sterile deionized distilled water and fragmented in a
-419-
-------
blender. The resulting slurry was resuspended in the
volume of sterile deionized distilled water needed to
adjust the moisture content of the aspen chips to 60%
(dry weight basis). Aspen chips had been sterilized by
autoclaving in aluminum foil-covered 2-L Erlenmeyer
flasks at 121'C and 103.5 kPa., for 1 h. Mycelia
were added to sterile chips at a rate of 0.02% myce-
lia:chips (w/w, dry weight basis). Inoculated chips
were incubated at 39*C for approximately 3 wk.
Soils
Soil samples from the A horizons of Marsham
sandy loam (Fine-silty, mixed, mesic, Mollic-
Hapludalf); Xurich sandy loam (Fine-silty over sandy,
mixed, mesic, Typic-Haplaquoll) and the B2t horizon
of Batavia silty clay loam (Fine-silty, mixed, mesic,
Typic-Hapludalf) were used in the present study. Soils
were air-dried, passed through a 2 mm sieve and
stored in 4-mil plastic bags at 4°C. Samples of each
soil were analyzed for physical and chemical proper-
ties by Dr. M. F. Jurgensen of the Michigan Techno-
logical University, School of Forestry and Wood
Products, Houghton, Michigan. These soils were
chosen to represent ranges of texture, acidity and or-
ganic matter and nutrient contents. Results of soil
chemical analyses are presented in Table 1.
Experimental Design
An experiment designed to test the influence of
soil type, water potential and temperature on growth of
P. chrysosporium in sterile soils consisted of a 3 x 3
x 3 factorial in a completely randomized design; the
experiment was repeated once.
Moisture content of each soil was adjusted to lev-
els corresponding to soil water potentials of -0.03, -
0.15 and -1.5 MPa. Petti plates (20 x 90mm diam.)
were filled with soil; within a soil type, an equal
weight of soil on a dry weight basis was added to each
plate. The weight of each filled plate was recorded and
soils and plates sterilized by autoclaving for 0.5 h at
121'C and 103.5 kPa. Autoclaving was performed on
each of three consecutive days to assure sterility; no
contamination problems were encountered.
Soils were inoculated by aseptically placing an in-
oculum chip into the soil in the center of a plate. Con-
trols were prepared by inoculating plates with inocu-
lum chips which had been re-sterilized by autoclaving
for 1 h at 121'C and 103.5 kPa. Plates were then
placed in incubators held at 25°, 30' and 39°C. Soil
water potentials were maintained daily by weighing
each plate and adjusting, if necessary, with the appro-
priate amount of sterile deionized distilled water. Each
run of the experiment was 2 weeks in duration.
In a separate experiment the effect of acidity on
growth of P. chrysosporium in the Batavia soil was
investigated. Appropriate amounts of CaCO^ were
added to Batavia soil to raise the pH 1 and 2 pH units
above the original (pH 4.8). Studies showed that 1 g
of CaCO3 per kg of Batavia soil and a 2-week incuba-
tion period were required to raise the pH value of the
soil 1 unit. After autoclaving, pH values were 4.8,5.8
and 6.8. Soil water potential was adjusted to -0.03
MPa. Sterilization and inoculation procedures were as
previously described. The experiment was 2 weeks
long.
Growth Assessment
At the end of 2 wk, growth of P. chrysosporium
was assessed using a qualitative rating system. This
system was based on a visual assessment of the
amount of growth by the fungus over the soil surface.
The rating system was as follows:
Growth rating-description
0-no growth of fungus from chip into soil
1-some hyphae growing from chip into soil
2-some scattered hyphal growth in the vicinity of the
chip plus some conidiation
3-medium coverage of soil surface by hyphae plus
conidiation
4-medium-dense coverage of soil surface plus dense
conidiation
5-dense coverage of soil surface plus dense conidia
tion
A rank transformation approach was used to ana-
lyze the data. The entire set of observations (i.e.
growth ratings) were ranked, with the smallest obser-
vation having rank 1, the second smallest rank 2, and
so on. Average ranks were assigned ties (Conover and
Iman 1981). The transformed data were subjected to
analysis of variance and reported as ranks of the
growth assessment data. Soil type means were com-
pared using Tukey's HSD method. All other treatment
means were compared using single degree of freedom
contrasts (Chew 1977).
Results and Discussion
Soil Type
Soil type greatly influenced growth of P. chrysos-
porium (p = .0001) . Growth of the fungus was
greatest in the Marsham, intermediate in the Xurich
and least in the Batavia soil (Table 2). It is interesting
to consider which factor(s) were responsible for the
growth differences of P. chrysosporium among the
three soils.
Soil acidity is one possibility. Cooke and Rayner
-420-
-------
Table 1. Selected soil chemical properties of the three studied soils.
_Mg_
JL
_E Mn zn
Soil type CEC1 BS% Acidity2 OM% N%3 Ca
-E A_ ppm
pH
Batavia 22.9 29.5 5.06 4.65 0.5 0.05 1950 850 145 75 0.6 12.5 1.9 9.7
Marsham 50.7 66.4 7.13 6.59 12.0 0.46 4900 -1650 90 17 1.3 9.0 12.2 11.3
Xurich 16.5 24.0 7.74 7.12 39.0 0.18 1675 640 80 17 0.8 77.5 6.4 3.9
1 Determined using the Ammonium Acetate Method (Thomas 1982)
2B= before autoclaving; A= after autoclaving
3Determined using the Kjeldahl prodedure for Total N (Bremmer and Mulvaney 1982)
Table 2. Mean growth ranking of Phanerochaete chry-
sosporium in three soils.
Soil Tvoe
Batavia
Marsham
Xurich
Growth Ranking
27c*
80a
57b
*Means followed by a different letter are significantly
different (Tukey's HSD, a = .05)
(1984) observed, however, that the influence of soil
acidity on fungal growth is difficult to assess because
of the abili ty of fungi to radically alter the pH value of
theu: environment, and because the effects of acidity
are modified by so many other environmental factors.
It was not surprising, therefore, that growth of P.
chrysosporium, which has an optimum of pH 5.5 in
liquid culture (Kirk et al. 1978), was greater in the
neutral to slightly acid Marsham and Xurich soils than
in the strongly acid Batavia soil. If acidity were re-
sponsible for the poor growth in the Batavia soil, re-
ducing soil acidity should have been beneficial. How-
ever, this was not the case (p = .4523). The failure of
decreasing soil acidity in the Batavia soil to benefit
fungal growth, and the observation that optimum
growth was found in the slightly acid (i.e. pH 6.59)
Marsham soil when the optimum far P. chrysospori-
um in liquid culture is pH 5.5, suggests that soil acid-
ity does not play a significant role in mediating the
growth of this fungus in soil. Evidently this organ-
ism, too, alters the pH value in the vicinity of its hy-
phae to create a favorable environment.
A second factor likely to influence growth is soil
organic matter. The three studied soils differed greatly
in this respect (Table 1). The ability of P. chrysospori-
um to assimilate the organic matter in the three soils is
unknown. Given the magnitude of the differences, if
organic matter content were the primary factor in con-
trolling growth of P. chrysosporium, greatest growth
would have been expected in the Xurich soil. Because
this was not the case (Table 2), some factor other than
organic matter content was controlling growth.
A third factor likely to influence growth is soil ni-
trogen content. In natural environments, this is the
nutrient which is most likely to be in such a limited
supply as to affect mycelial growth (Cooke and Rayn-
er 1984). Nitrogen content varied greatly between the
three soils studied. Regression of fungal growth on
soil nitrogen content revealed a strong positive rela-
tionship (R2 = .924) (Figure 1). The strength of this
relationship suggests that soil nitrogen content played
a significant role in mediating growth of P. chrysos-
porium .
Growth habit of the fungus was also affected by
soil type. In the Marsham and Xurich soils, hyphae
were observed to be growing primarily across the soil
surface, although at the highest levels of growth it was
evident from observing the bottom of the plates that
the fungus had penetrated through the soil. In the Ba-
-421-
-------
1001
80-
60-
40-
20
0.0 0.1 0.2 0.3 0.4 0.5 0.6
SOIL NITROGEN (%)
Figure 1. Regression of growth ranking on soil nitro-
gen content for Phanerochaete chrysosporium grown
in three soils (y = 123.6x + 26.1, R2 = .924).
tavia soil, hyphae were not observed growing across
the soil surface, but appeared to grow through the soil,
surfacing at various locations.
There was evidence for an interaction between soil
type and experimental run (p = .0003). This interac-
tion was probably the result of better growth in the
Marsham and Xurich soils in Run 2 than in Run 1
(Figure 2). During Run 1, relative humidity in the in-
cubators was maintained at levels which resulted in ap-
preciable daily soil moisture losses, particularly at 39
"C. During the second run, humidity levels in the in-
SoilType
• Balavia
Marsham
Xurich
1 2
Experimental Run
Figure 2. Growth ranking of Phanerochaete chrysospori-
um in three soils by experimental run.
cubators were increased and soil moisture losses great-
ly reduced. Phanerochaete chrysosporium might have
been able to benefit from the increased incubator hu-
midity levels during Rim 2 by obtaining moisture di-
rectly from the atmosphere, or loss of hyphal vigor
due to partial desiccation could have been prevented.
Ether might explain the significantly greater growth of
the fungus in the Marsham and Xurich soils during
Run 2. Because P. chrysosporium did not grow on
the soil surface in the Batavia soil it would not be ex-
pected to benefit as much from an increase in atmo-
spheric moisture. Indeed, growth of the fungus in the
Batavia soil during Run 2 was not significantly differ-
ent from that in Run 1 (Figure 2). In light of the
above, the data from each run were analyzed separate-
ly to assess the effect of soil water potential on growth
of P. chrysosporium.
Soil Water Potential
A significant (p = 0.0002) linear relationship be-
tween growth of P. chrysosporium and soil water po-
tential over all soils was observed in Run 1 (Figure 3).
In contrast, soil water potential had less influence on
growth in Run 2 (p = .1399). As indicated above,
however, the fungus growing in the Marsham and Xu-
rich soils might have benefited from increased levels
of atmospheric moisture which effectively masked the
influence of soil water potential on growth. In the Ba-
tavia soil where surface growth of the fungus was
sparse, there was a significant (p = 0.0017) linear rela-
tionship between growth of the fungus and soil water
potential (Fig. 4).
The linear increase in growth of P. chrysosporium
in response to increasing water potential over the range
401
30-
20-
10
0.0 0.5 1.0 1.5 2.0
SOIL WATER POTENTIAL (-MPa)
Figure 3. Regression of growth ranking on soil water
potential far Phanerochaete chrysosporium grown in
three soils-Run 1 (y = 1.119x + 33.767, R2 = .233).
-422-
-------
30-
20-
§
o
10-
0
Table 3. Effect of soil temperature on growth ranking
of Phanerochaete chrysosporium in three soils.
0.0 0.5 1.0 1.5 2.0
SOIL WATER POTENTIAL (-MPa)
Figure 4. Regression of growth ranking on soil water
potential for Phanerochaete chrysosporium grown in
the Batavia soil-Run 2 (y = 1.299x + 23.551, R2
=.47)
-1.5 to -0.03 MPa is consistent with what other
workers have observed for both wood decay and soil
fungi in general. Griffin (1977) proposed that growth
of wood decay fungi occurs at a decreasing rate from
0 to -4.0 MPa, and presented data indicating that
maximum growth of five wood decay basidiomycetes
occurred at -0.15 MPa osmotic potential. Wood decay
fungi have been placed in a group of organisms,
which includes some soil basidiomycetes and gram- •.
negative bacteria, that is extremely sensitive to water
potential, with an optimum growth response ca. -0.1
MPa osmotic potential, and little growth below -2.0
MPa (Griffin 1981). The optimum soil water potential
range for microbial growth and metabolic activity is
generally considered to be between -0.01 and -0.03
MPa (Sommers et al. 1981). The soil water potential
growth optimum far P. chrysosporium might be
greater than -0.03 MPa.
Sommers et al. (1981) suggested that water po-
tential influences decomposition of soil organic mate-
rials in a 2-phase process whereby an initial rapid de-
crease in decomposition in the -0.03 to -1.5 MPa mat-
ric potential range is followed by a second phase in
which decomposition decreases linearly with decreas-
ing water potential. Therefore, soil water potential op-
tima for growth of, and degradatiye activity by, P.
chrysosporium may be very similar.
Temperature
Evidence was obtained for better growth of P.
chrysosporium at 30° and 39°C compared to growth
of the fungus at 25°C (Table 3). The magnitude of
Soil
Soil Temperature ("O
39°
30°
25°
p*
Batavia
Marsham
Xurich
28
86
64
31
89
59
21
66
47
.1198
.0170
.0531
* Probability of a larger F value for the single degree of
freedom contrast of growth at 30° and 39°C versus
growth at 25°C.
this difference appeared to be directly related to the rel-
ative growth among soil types. Most soil fungi are
mesophiles with temperature optima between 25° and
35°C, but with an ability to grow from ca 15° to 45°C
(Cooke and Rayner 1984). Phanerochaete chrysospori-
um is also mesophilic and has a temperature optimum
of ca. 39°C on 2% malt agar (Burdsall and Eslyn
1974). Ability of the fungus to grow as well at 30°C as
it did at 39°C was not expected. It is possible that at
39°C some other factor (e.g. nutrient availability or wa-
ter stress) offsets the benefits of the more optimal tem-
perature. However, P. chrysosporium did not appear
to exploit the total soil volume in any of the plates, indi-
cating that nutrient pools were also not fully exploited.
Studies are underway to determine the minimum tem-
perature at which P. chrysosporium will grow in the
soil.
Conclusions
All parameters investigated, except acidity, in- «'
fluenced growth.of P. chrysosporium in the three soil
types. Soil type had a significant effect on growth and
growth habit. Nitrogen content appeared to play a ma-
jor role in mediating growth of the fungus in the soils, ,
and could well have been responsible for the differenc-
es in fungal growth among the soil types. Future stud-
ies to elucidate the extent to which available nitrogen in-
fluences growth of P. chrysosporium in soil are war-
ranted. Increasing soil water potential from -1.5 MPa
to -0.03 MPa resulted in greatly increased growth of P.
chrysosporium. The data indicate that growth of the
fungus might benefit from soil water potentials above -
0.03 MPa. Growth of the fungus was significantly
greater at 30° and 39°C than at 25°C.
-423-
-------
Acknowledgments
The authors wish to thank C B. Davey for critical
review of the manuscript and D. M. Dietrich for techni-
cal assistance.
Literature Cited
1. Bremmer, J.M. and C.S. Mulvaney. 1982. Nitro
gen-Total. In: Methods and Principles of Soil Anal
ysis, Part 2. Chemical and Microbiological Proper
ties-Agronomy Monograph no. 9 (2nd Edition).
2. Bumpus, J.A., M. Tien, D. Wright and S.D. Aust.
1985. Oxidation of persistent environmental
pollutants by a white rot fungus. Science
228:1434-1436.
3. Burdsall, H.H., Jr. and W.E. Eslyn. 1974. A new
Phanerochaete vntha. Chrysosporium
imperfect state. Mycotaxon 1:123-133.
4. Chew, V. 1977. Comparisons among treatment
means in analysis of variance. Agric. Res. Ser.
H-6. 60pp.
5. Conover, WJ. and R. L. Iman. 1981. Rank trans
formations as a bridge between parametric
and non-parametric statistics. American Statistician
35(3):124-129.
6. Cooke, R.C. and A.D.M. Rayner. 1984. Ecology
of saprophytic fungi. New York:Longman
415pp.
7. Eaton, D.C. 1985. Mineralization of polychlorinat
ed biphenyls by Phanerochaete chrysosporium: a
ligninolytic fungus. Enzyme Microb. Technol.
7:194-196.
8. Griffin, D.M. 1981. Water and microbial stress.
Adv. Microbial Ecol. 5:91-136.
9. Griffin, D.M. 1977. Water potential and wood de
cay fungi. Ann. Rev. Phytopathology
15:319-329.
10. Huynh, V.-B., H.-M. Chang, T.W. Joyce and
T.K. Kirk. 1985. Dechlorination of
chloro-organics by a white-rot fungus. Tappi J.
68:98-102.
11. Kirk, T.K., E. Schultz, W.J. Connors, L.F. Lo
renz and J.G. Zeikus. 1978. Influence of culture
parameters on lignin metabolism by Phanero
chaste chrysosporium. Arch. Microbiol. 117:277-
285.
12. Seitz, L. M., D. B. Sauer, R. Burroughs, H. E.
Mohr and J. D. Hubbard. 1979. Ergosterol as
a measure of fungal growth. Phytopathology
69:12024203.
13. Sommers, L.E., C.M. Gilmour, R.E. Wilding and
S.M.Beck. 1981. The effect of water potential on
decomposition processes in soils, pp. 97-118. In:
Water potential relations in soil microbiology.
Soil Sci. Soc. Am. Spec. Pub. No. 9.
14. Thomas, G.W. 1982. Exchangeable cations. In:
Methods of Soil Analysis, Part 2. Chemical and
Microbiological Properties-Agronomy Monograph
no. 9 (2nd Edition).
-424-
-------
BIOLOGICAL TREATMENT OF SELECTED AQUEOUS
ORGANIC HAZARDOUS WASTES
Richard J.Lesiecki, Margaret K. Koczwara, James E. Park
University of Cincinnati
Dept. of Civil and Environmental Engineering
Cincinnati, Ohio 45221
i
Douglas W. Grosse
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
ABSTRACT
This paper describes tests performed in order to evaluate the fate of aqueous organic
hazardous waste compounds in the activated sludge process. Gas, liquid and waste solids
samples were taken from acclimated activated sludge systems to determine amounts that were
volatilized, biodegraded and associated with the wasted solids. Results discussed here
include two compounds, methyl ethyl ketone and 1,1,1 trichloroethane.
INTRODUCTION
With the reauthorization of RCRA and
the concurrent restrictions on land dis-
posal of hazardous wastes, the EPA is
assessing waste treatment technique's that
can be substitutes for, or precursors to,
land disposal. Concentrations of
hazardous, organic compounds .in aqueous,
RCRA listed wastes must be greatly reduced
in order to achieve levels acceptable for
land disposal. For this type of waste in
concentrations up to 10,000 ppm, the
activated sludge process is a potentially
applicable treatment process. However, a
lack of data on the treatment of solvents
at high concentrations exists.
To fulfill this need, two bench-scale
activated sludge systems were constructed.
The systems were operated at a theoretical
10-day solids retention time (SRT) with a
24-hour hydraulic retention time (HRT).
Each system was fed a synthetic wastewater
spiked with the test compound. Compounds
tested in this study were methyl ethyl
ketone (MEK) and 1,1,1 trichloroethane
(TCA). Each compound was tested at two
concentrations.
The purpose of this program was to
study the changes in fate of selected
organic priority pollutants in continuous
bench-scale activated sludge systems over a
broad concentration range. This informa-
ti&n will be helpful id determining whether
or not this treatment process introduces
toxic compounds to the atmosphere, concen-
trates them in the wasted sludge, or
destroys them via biodegradation.
MATERIALS & METHODS
Two completely mixed, continuous flow,
bench-scale activated sludge systems were
utilized. Each system consisted of a pyrex
glass column 15.2 cm in diameter and 91.4
cm in height. The total liquid volume was
11 liters with a depth of 58.4 cm. This
allows for an air volume above the liquid
level of 5.7 liters. A metal plate with
inlet and outlet holes covers the top of
the reactor. Ports are located along the
side of the reactor for inlet/outlet lines.
An airstone located just above the bottom
of the reactor diffuses air into the system
for mixing and aerating purposes. Air flow
rate and pressure were monitored. Air flow
rates varied between approximately
-425-
-------
800 ml/min to 1800 ml/min depending on the
dissolved oxygen (D.O.) concentration and
mixed liquor solids concentration of the
reactor.
The primary and secondary clarifiers
were also made of pyrex glass and were the
same size. A glass dome, with inlet and
outlet ports for air venting and sampling
purposes, covered each clarifier. Each had
a liquid volume of 1.5 liters with an air
head space volume of 0.8 liters. Liquid
flowed by gravity in sequence from the
primary clarifier to the reactor and on to
the secondary clarifier. A recycle pump
connected to a timer was used to periodi-
cally transfer the settled secondary solids
back to the reactor. Solids were wasted
directly from the reactor by a pump
connected to a timer. Synthetic feed was
peristaltically pumped from a refrigerated
stock solution to the primary clarifier.
The organic spike compound was introduced
to the feed lines by means of a syringe
pump. A static, in-line mixer was used
immediately in front of the primary
clarifier to blend the spike with the feed.
Off-gases from the reactor and clarifiers
were vented.
The activated sludge systems were used
to treat a synthetic wastewater consisting
of a "base mix" plus the selected priority
pollutant of interest. The synthetic feed
was prepared daily and stored at 4°C. The
base mix constituents are listed in Table 1.
Table 1
Synthetic Feed Constituents
Ethylene glycol
Ethyl alcohol
Acetic acid
Phosphoric acid
Glutamic acid
Glucose
Phenol
Ammonium sulfate
Magnesium sulfate
Manganese sulfate
Calcium chloride
Ferric chloride
The base mix was diluted so that the
wastewater exerted a 5-day Soluble
Biochemical Oxygen Demand (SBODs) of
approximately 500 mg/1, with an average
Soluble Organic Carbon (SOC) of 420 mg/1.
This synthetic wastewater is the formula
used by Kincannon, et. al. (1983) in
studying priority pollutants in activated
sludge treatment systems. A synthetic feed
was chosen over an industrial or municipal
wastewater for two basic reasons: A
synthetic feed is more consistent in terms
of SBODs and SOC concentrations than an
industrial or municipal wastewater; and
secondly, it facilitates comparisons
between spiked compounds.
Much of the existing data on the fate
of compounds in activated sludge are at
concentrations up to 200 mg/1 (e.g.
Kincannon et. al., 1983). Based on this,
initial concentrations not exceeding
200 mg/1 were chosen. The first measured
concentrations tested were 55 mg/1 for MEK
and 141 mg/1 for TCA. The second
concentrations tested were 430 mg/1 for MEK
and 174 mg/1 for TCA. The addition of MEK
increased the influent SBODs to 620 mg/1
and 1340 mg/1 for each concentration
respectively. The addition of TCA had a
negligible affect on the influent SBOD5 at
either concentration.
Return activated sludge, for the
purpose of initial seeding, was obtained
from a large-scale system at the EPA Test
and Evaluation (T&E) Facility. A 100 ml
sample pf primary effluent from the same
large-scale system was added daily to each
systems' reactor throughout the study.
This was done to provide a continued supply
of organisms found in wastewater that are
not available in a synthetic feed. The two
individual systems were acclimated to the
synthetic wastewater for at least one month
before the spike was introduced. Both sys-
tems were allowed to acclimate to the first
spike concentration for at least a month.
A theoretical solids retention time (SRT)
of 10 days was targeted, controlled by
wasting 10% of the reactor volume over a 24
hour period. Average SRT's were about 7
days for both systems, due to a loss of
solids in the effluent. The hydraulic
retention time (HRT) was 24 hours. Routine
process performance monitoring measurements
and their respective frequencies are listed
in Table 2.
When the systems assumed steady state
conditions following acclimation to the_
spike compound, a two-week sampling period
ensued. During the sampling period, sam-
ples for gas chromatograph analyses were
taken three times a week for the two weeks.
Samples taken during this period consisted
of 24 hour composite liquid influent and
effluent samples (pumped to teflon bags to
minimize volatilization), a 24-hour
composite wasted mixed liquor sample, a
grab primary effluent liquid sample and
grab off-gas samples from the primary and
-426-
-------
Table 2 Routine Process Monitoring
Measurement Frequency Method
SB005
SOC
TSS/VSS
D.O.
D.Q. Uptake Rate
USEPA (1983)
3 x per wk
5 x per wk
1 x per day
6 x per day
1 x per day
EPA 405.1*
EPA 415.2*
EPA 160.1*
EPA 360.1*
Std. Meth.
213A
secondary clarifiers, as well as the
reactor. All liquid composite samples were
kept at 4°C. Liquid samples were trans-
ferred to clean amber glass bottles with
Teflon caps, sealed with no headspace.
Off-gases were sampled using 0.63 cm O.D.
borosilicate glass sorbent tubes containing
0.25 g Tenax. Two tubes were used in
series for each sample. Air sampling pumps
were used to provide a constant air flow
rate, which was further controlled by
needle valves and measured by rotameters.
Purified air at atmospheric pressure was
used to replace the off-gas removed from
the clarifiers.
Methyl ethyl ketone measurements were
obtained by EPA Method 8015 (SW-846, 3rd
ed., 1986). EPA Method 601 for Purgeable
Halocarbons was used to measure 1,1,1
trichloroethane (Federal Register, 40 CFR
Part 136, Part VIII, 10/26/84).
Table 3
Process Monitoring Data for MEK System
(all values in mg/1)
MEK Concentration
55 mg/1
Before Testing
During Testing
Mean Std. Dev. Mean Std. Dev.
MLVSS 779 121 1093
EFF. VSS 101 62 44
EFFL. SBOD5 — 16
EFFL. SOC 38 21 17
336
22
8
6
MEK Concentration
430 mg/1
Before Testing
During Testing
Mean Std. Dev. Mean Std. Dev.
2430
55
33
10
298
30
56
6
2189
84
23
4
182
57
19
4
Table 4
Process Monitoring Data for TCA System
(all values in mg/1)
TCA Concentration
• 141 mg/1
Before Testing
During Testing
Mean Std. Dev. Mean Std. Dev.
MLVSS 958 129 1335
EFFL. VSS 134 112 91
EFFL. SBOD5 17 16 63
EFFL. SOC 20 12 27
178
65
53
14
TCA Concentration
174 mg/1
Before Testing
Mean Std. Dev.
682
327
6
13
221
58
4
During Testing
Mean Std. Dev.
467
345
10
29
75
3
1
-427-
-------
RESULTS
Two organic compounds, each at two
concentrations, have been studied in the
completely-mixed, continuous flow,
activated sludge systems. Table 3 and
Table 4 summarize basic process monitoring
data used to assess the systems' perfor-
mance. Mean values with standard
deviations for a two-week period prior to
testing, and during the test period itself,
are presented. All test period values are
14-day means, except TCA concentration #2,
Which is a 3-day mean. Because a five-week
period separated sample days two and three
thru seven for the first MEK concentration,
the data for the two weeks prior to testing
are prior to the last five sample days.
For the first concentration of each
compound seven data points were collected.
For the second concentration, six data
points for MEK and three data points for
TCA were collected. The average mass
fluxes with the mean percentages of the
influent mass distributed to each fraction,
as calculated from an average mass balance
around the primary clarifier and both the
reactor and secondary clarifier, are shown
in Tables 5 and 6. Data from the second
TCA concentration were excluded since this
concentration was inhibitory to the system.
Table 5
Fates of MEK in Activated Sludge
MEK Concentration
55 mg/1
MEK Concentration
430 mg/1
Influent
Primary Effluent
Primary Volatilization
Unknown Primary
Loss (+)/Gain (-)
Flux
(mg/min)
0.418
0.404
0.046
-0.030
Percent
PRIMARY MASS BALANCE
96.7
11.0
-7.7
Flux
(mg/min)
3.269
3.358
0.084
-0.170
SECONDARY MASS BALANCE
Percent
102.7
2.6
-5.2
Reactor Influent
Apparent Biodegradation
Reactor Stripping
Secondary Volatilization
Waste Sludge
Effluent
0.404
0.382
0.012
0.0002
0.0004
0.009
94.6
3.0
0.1
0.1
2.3
3.358
3.012
0.342
0.003
0.000
0.001
89.7
10.2
0.1
0.0
0.0
-428-
-------
Table 6
Fates of TCA in Activated Sludge
Influent
Primary Effluent
Primary Volatilization
Unknown Primary
Loss(+)/Gain(-)
Reactor Influent
Apparent Biodegradation
Reactor Stripping
Secondary Volatilization
Waste Sludge
Effluent
TCA Concentration
. 141 mg/1
Flux (mg/min)
PRIMARY MASS BALANCE
1.072
0.741
0.221
0.110
SECONDARY MASS BALANCE
0.741
0.161
0.565
0.004
0.001
0.010
Percent
69.0
20.6
10.4
21.8
76.2
0.5
0.1
1.4
Table 7
Removal Efficiencies of Organic Pollutants
Compound
Influent (mg/1)
Mean Std. Dev.
Effluent (mg/1)
Mean Std. Dev.
Percent Overall
Methyl ethyl ketone
1,1,1 Trichloroethane
55
430
141
174
9
59
37
32
0.5
0.9
3.8
3.5
1.2
0.3
6.5
0.3
>99
>99
97
98
DISCUSSION
Table 7 summarizes the overall organic
pollutant removal efficiencies of both
systems at each concentration. It is seen
that greater than 89% of the MEK was bio-
degraded at both concentrations. The large
increase in MLVSS concentrations after the
increase in MEK concentration from 55 mg/1
to 430 mg/1 would seem to support the
notion that MEK is very easily biodegraded.
Operating the activated sludge system
while it was spiked with MEK was somewhat
difficult. The floe formed in the system
was fine and light, hence, not settling
well. Bridging at the bottom of the secon-
dary clarifier developed, having the effect
of lowering the concentration of solids in
the return activated sludge (RAS).
Therefore, solids were washed out of the '
system in the secondary effluent. As the
MEK concentration increased, secondary
effluent solids concentration increased as
more solids were being formed in the
reactor, compounding the settling problems.
-429-
-------
Another factor influencing the
performance -of the MEK system may have been
nutrient deficiency. As the concentration
of MEK was increased, the influent SBODs
increased, but the base mix remained
constant. This resulted in a decrease of
the ratio of nutrients to SBOD5. The poor
floe formation observed at the higher MEK
concentration may have been partially
caused by this deficiency.
As is shown in Table 6, most of the
TCA was removed via stripping in the
reactor. This is to be expected since TCA
is highly volatile. Primary volatilization
and biodeg'radation also accounted for
significant percentages of the overall
removal.
As with MEK, poor system operation
occurred when spiked with TCA. For each
concentration tested, TCA seemed inhibitory
to biological growth. This is evident by
the fact that initial acclimation took four
months; one month at 10 mg/1 and three
months at HI mg/1. Effluent solids
concentrations at 141 mg/1 TCA averaged
approximately 100 mg/1 VSS. The solids
were fine and light and did not settle
well. When the influent concentration was
further increased to a measured concen-
tration of 174 mg/1, the system exhibited
signs of shock and inhibition. The primary
clarifier contents, usually cloudy, turned
crystal clear. Little growth occurred in
the reactor. Solids did not settle in the
secondary clarifier. Effluent solids
concentrations were almost the same as
MLVSS concentrations as shown in Table 4.
One major problem encountered when
working with TCA was in uniformly dis-
persing it into the feed solution. For the
first test phase, an amount equal to
'295 mg/1 was injected into the feed line,
yet only 141 mg/1 was measured in the feed.
For the second test concentration, an
amount equal to 705 mg/1 was injected,
whereby only 174 mg/1 was measured in the
feed, and 251 mg/1 measured in the primary
effluent. Many factors may account for
this inconsistency including: the density
of TCA (1.339 g/cc), low solubility in
water, volatility, as well as inadequate
sampling techniques. Thus, effects on the
system may have been.caused by concentra-
tions of TCA higher than reported.
SUMMARY
This study has shown that methyl ethyl
ketone and 1,1,1 trichloroethane can be
removed by the activated sludge process,
but both chemicals at the concentrations
tested are disruptive to the basic opera-
tion of the unmodified process. For MEK,
biodegradation was the most important
removal mechanism, while most TCA was
removed via stripping.
ACKNOWLEDGEMENTS
This work was carried out at the
USEPA's Test and Evaluation Facility in
Cincinnati, Ohio, as a part of the
Hazardous Waste Engineering Research
Laboratory (HWERL), Technology Treatment
Staff's Program. The Technical Project
Monitor was Douglas W. Grosse. The gas
chromatography was performed by EER under
the direction of Sam Hayes. The authors
would like to express their gratitude to
Terry Harris, Paul Mishurda, Daniel DiCarlo
and Michael Morelock for their invaluable
work.
REFERENCES
Kincannon, D.F. et.'al, "Removal Mechanisms
for Toxic Priority Pollutants", JWPCF Vol.
55, No. 2, 1983, pp 157-163
Standard Methods for the Examination of
Water and Wastewater, 16th Ed.,1985,
APHA-AWWA-WPCF
USEPA "Methods for Organic Chemical
Analysis of Municipal and Industrial
Wastewater" 40 CFR Part 136, Oct. 1984
USEPA "Methods for Chemical Analysis
of Water and Wastes" EPA 600/4-79-020,
March 1983
-430-
-------
ASSESSMENT OF ALTERNATIVE TECHNOLOGIES FOR TREATING
SPENT ELECTROPLATING SOLUTIONS AND SLUDGES
Katherine Driscoll and Barry Kaplan
Metcalf & Eddy, Inc.
Wakefield, MA 01880
ABSTRACT
Off-site commercial hazardous waste treatment facilities were evaluated
to gene'rate support data for the Environmental Protection Agency's land
disposal ban. Establishing treatment standards for electroplating wastewater
and sludges is a high priority task with respect to the land disposal ban.
One facility treated electroplating solutions with cyanide oxidation,
hexavalent chromium reduction, a combination of lime and sulfide
precipitation, and vacuum filtration. Electroplating sludges were stabilized
with calcium hypochlorite, ferric sulfate or lime. Of particular interest is
the use of waste streams as treatment reagents. This report summarizes data
used to evaluate these treatment technologies for electroplating solutions and
sludges.
INTRODUCTION
The RCRA Hazardous and Solid
Waste Amendments of. 1984 mandate
that a ban on the land disposal of
hazardous wastes be implemented.
EPA will choose the "best demon-
strated available technology" (BOAT)
for each waste category banned and
will determine treatment performance
levels for each waste category based
on the BOAT. Hazardous constituents
of treated wastes must be reduced
below specified levels. To meet the
provisions of the land disposal ban,
EPA has prioritized waste categories
to be addressed. Electroplating
wastes and other wastes containing
metals and cyanide are a high
priority.
In support of the efforts of
the Office of Solid Waste (OSW) to
implement the provisions of the RCRA
land disposal ban, the Hazardous
Waste Engineering Research
Laboratory (HWERL) has initiated
several programs to evaluate
alternatives to land disposal. A
major component of these programs
has been the field evaluation of
existing full-scale commercial
treatment facilities.
FACILITY DESCRIPTION
One facility evaluated under
this program treated wastes
generated by the electroplating
industry, as well as other metal
finishing operations. The facility
treats liquid wastes by the follow-
ing unit operations: cyanide
oxidation, hexavalent chromium
reduction, metals precipitation and
vacuum filtration. -Total treatment
capacity for liquid wastes is ten
million gallons per year. Solid
wastes contaminated with cyanide,
hexavalent chromium, and various
metals are stabilized with calcium
hypochlorite, ferric sulfate
crystals, and lime, respectively.
This process is designed to delist
the solid wastes. The facility
operates 24 hours per day for six
days a week.
-431-
-------
PROCESS DESCRIPTIONS - LIQUID WASTE
Incoming Waste
Liquid waste streams, defined
as having less than 20 weight
percent total suspended solids, are
screened to confirm their manifest
composition. A treatment simulation
is conducted to determine the
required amount of treatment
reagents. Wastes are segregated by
type: acid, neutral, alkali,
cyanide content, and hexavalent
chromium content. If a waste
contains greater than 1 part per
million (ppm) of cyanide, it is
segregated for pretreatment. Wastes
containing greater than 25,000 ppm
of cyanide are not accepted. As
many as five separate sources may be
incorporated into one batch for
treatment.
Cyanide Oxidation
Cyanide wastes are treated by
alkaline chlorination. A lime
slurry is used for pH adjustment to
10 and a waste sodium hypochlorite
solution is used as the oxidizing
agent. The cyanide is oxidized to
cyanate in a single-stage batch
reaction. The amount of
hypochlorite added to each batch is
predetermined by the plant chemist
and controlled by the operator. The
level of excess hypochlorite is
monitored every 15 minutes by an
operator using potassium iodide
starch test paper. Additional
hypochlorite is added, if needed, to
maintain an excess concentration of
at least 100 ppm. One hour after
the initial hypochlorite addition a
filtered sample is analyzed for
amenable cyanide. Treatment
continues until cyanide is not
detectable in a filtered sample.
The cyanide reaction is optimum at a
pH of 11 to 12. Since cyanide forms
strong complexes with ferrous iron,
it is important to oxidize the
cyanide before reducing hexavalent
chromium. Further treatment of the
waste is conducted in the same
reactor tank.
Hexavalent Chromium Reduction
Hexavalent chromium is reduced
to trivalent chromium in a single-
stage batch reaction, using waste
iron acid as the reducing agent.
The treatment tank is first charged
with lime slurry to minimize
corrosion of the stainless steel
mixing system. Enough lime slurry
is added to neutralize all waste
acids that would be treated for
hexavalent chromium reduction and
subsequent metals precipitation.
Maintaining an alkaline mixture also
helps to eliminate emissions of by-
product gases (e.g. M02, C12). The
stoichiometric weight ratio of iron
acid to hexavalent chromium is 3.2
to 1. A 10 percent excess of iron
is maintained to ensure complete
reduction. Temperature and pH are
monitored periodically throughout
the reaction. Treatment is
continued until a spot test shows no
detectable hexavalent chromium.
Additional treatment is conducted in
the same reactor vessel.
Metals Precipitation
Metals are first precipitated
with a lime slurry. A predetermined
volume of lime slurry is used which
includes 200 to 300 gallons in
excess of the required volume. At
the optimum pH range of 8 to 10,
metal cations will precipitate as
insoluble hydroxides. The process
is dependent on the solubilities of
the metal hydroxides, which are a
function of pH and water quality.
Because each metal hydroxide has a
minimum' solubility at a different
pH, it is difficult to obtain high
removal efficiencies for every
-432-
-------
metal. An agitator keeps the
suspended floe in suspension. After
ten minutes of agitation the pH is
checked. When the pH does not
change with time, the reaction with
lime is complete.
Because of stringent discharge
limits, precipitation with a waste
sulfide stream is used as a
polishing step to remove the
remaining dissolved metals. In
general, metal sulfides have lower
. solubilities than metal hydroxides
at the same pH. After sulfide
addition, a filtered sample is
analyzed for chromium, copper,
nickel and zinc. The sample filter
cake is subjected to an abbreviated
EP Toxicity test. The facility has
determined that if a waste can pass
their abbreviated test, it will also
pass the 21-hour test. When the
sample results are acceptable, the
contents of the tank are pumped to a
precoat vacuum filter.
Vacuum Filtration
The treated wastewater is
subjected to vacuum filtration to
separate liquids and solids.
Perlite, a volcanic ash, is
currently used as a precoat although
diatomaceous earth has been used in
the past. The dewatered solids fall
onto a solids conveying system and
are stored in a bulk container. A
sample is subjected to the 24 hour
EP toxicity test to ensure the waste
is nonhazardous.
Ultimate Disposal
The dewatered sludge is sent to
the facility's lined, nonhazardous
landfill. The filtrate is
discharged to the local publicly
owned treatment works (POTW). The
facility continuously monitors the
sewer discharge to assure compliance
with local standards. The liquid
waste treatment process is
illustrated in Figure 1.
Air Emissions
The facility uses a two-stage
scrubber to capture fumes from the
treatment tanks, caustic storage and
sulfide storage tanks.
A one-stage scrubber is used to
capture fumes released from the acid
storage tanks. Both scrubbers have
an induced flow of 1000 standard
cubic feet per minute (scfm). The
pH in each scrubber is maintained
between 12 and 14 by the addition of
sodium hydroxide. The overflow
spillage from each scrubber is
collected in its respective sump and
pumped to a treatment tank when the
sump is full.
PROCESS DESCRIPTION - SOLID WASTE
Incoming Waste
Prior to being unloaded at the
treatment facility, each load of
solid waste is weighed and tested to
verify its contents. A treatment
simulation is conducted oh a sample
to determine the required amounts of
treatment chemicals and to make sure
the waste can pass the EP Toxicity
test. Incoming loads are segregated
on the basis of their water content.
Treatment
The solids handling/treatment
process is designed to delist solids
that may be contaminated with
cyanide, hexavalent chromium, and/or
various other metals. The untreated
solids are lifted by an overhead
traveling crane into a mix chamber
that is agitated by a modified
pugmill operation. After mixing,
bags of reagents are manually added
into the mix cell. Calcium
-433-
-------
LL.O.
U.O
LUl-
UI U.
QLU
LU>
CEO
LUCC
Ho.
wo
OH
O
o
LLJ
_1
D.
85
LULL.
cc
D-
ujO
e5
z*1-
x
°0
111
LuXUJ
1
zuu
O
Q
CC
LLJ
D
-434-
-------
hypochlorite, ferric sulfate
crystals and lime are added for
sludges contaminated with cyanide,
hexavalent chromium and various
metals, respectively. Water is also
added if the mix is too dry. After
thirty minutes of treatment, the mix
cell contents are dropped into a
roll-off container. All treated
solids are then subjected to the
24 hour EP Toxicity test to confirm
they have been rendered
nonhazardous. Figure 2 depicts the
solids handling treatment operation.
Ultimate Disposal
The delisted sludge is disposed
of in the facility's lined,
nonhazardous landfill.
SAMPLING RESULTS
The objective of this sampling
program was to provide HWERL with
information on existing hazardous
waste treatment technologies which
provide alternatives to land
disposal. Specific objectives
included the following:
identify industries whose
wastes are treated by
alternative technologies.
determine treatment
capacity, waste acceptance
criteria and other
pertinent information about
each facility.
evaluate the effectiveness
of each unit operation.
discuss environmental
impacts of alternative
technologies (e.g. air
emissions, sludge disposal,
effluent discharge).
Fourteen complete batches of
waste were processed during the
sampling episode, including eleven
liquid and three solid batches. Of
the eleven batches of liquid waste,
three required pretreatment for
cyanide oxidation, six required
hexavalent chromium reduction and
all eleven were treated by chemical
precipitation and vacuum filtration.
Liquid Waste Treatment System
Influent Wastes. Discrete wastes
treated during the sampling episode
were primarily from fabricated metal
products industries. Each batch
treated contained an average of 46
percent by volume electroplating
waste. Table 1 summarizes the types
of waste streams including their SIC
code and RCRA waste code.
These discrete waste streams
have widely varying concentrations
of metals. The wastes had high
concentrations of solids, ranging
from 4,500 to 259,200 ppm. Ninety
six percent of the solids were
dissolved. The pH ranged from 1 to
13. In the three batches pretreated
by cyanide oxidation, the total
cyanide concentration in discrete
wastes averaged 7.6 ppm. Free
cyanide was present at less than
0.10 ppm.
The most concentrated waste was
a chromic acid plating bath solution
with a pH of 1, containing the
following metals: antimony
(68 ppm), hexavalent chromium
(78,400 ppm), total chromium
(103,800 ppm), copper (3,500 ppm),
lead (60 ppm), nickel (40 ppm) and
zinc (100 ppm).
Treatment Tank Composite. As many
as five discrete wastes were
composited for batch treatment.
Treatment tank composite samples
were analyzed for volatile organics,
total organic halide (TOX) and total
organic carbon (TOO, in addition to
-435-
-------
C/l
Q
8.
Q =:
o
O
w
erf
CO
§
O
CO
O
-436-
-------
TABLE 1. INDUSTRIAL SOURCES OF LIQUID WASTES
Industry
Category
Industry
(SIC Code)
Process
RCRA waste
code
Fabricated
metal products
Fabricated
metal products
Fabricated
metal products
Fabricated
metal products
Chemical manu-
facturing and
processing
Fabricated metal
products
Fabricated
metal products
Fabricated
metal products
Fabricated
metal products
Refurbishing of used
field artillery (3471)
Electroplating (3471)
Manufacture of toasters,
toaster ovens, and
coffee makers (3634)
Reclamation of tin
from scrap metal (3341)
Manufacture of glass
perfume bottles (3621)
Manufacture of aluminum
windows, doors, sashes
and mouldings (3442)
Manufacture of semi-
conductors and electrical
connectors (3678, 3643)
Manufacture of
aluminum bottle
caps (3466)
Surface finishing of
sheet steel-anodizing,
painting and etching
(3316)
Chrome & brass F006
electroplating
Chrome D002
electroplating
Nickel and chrome F006
electroplating
Etching,. D002
electroplating
Etching D002
Conversion coating F019
of aluminum
Nickel, chrome F006
and gold electro-
plating
Cleaning D002
Cleaning, K062
D002 - Corrosive waste that has a pH less than or equal to 2 or greater than
or equal to 12.5.
F006 - Wastewater treatment sludges from electroplating operations except from
the following processes: (1) sulfuric acid anodizing of aluminum; (2)
tin plating on carbon steel; (3) zinc plating (segregated basis) on
carbon steel; (4) aluminum or zinc - aluminum plating on carbon steel;
(5) cleaning/stripping associated with tin, zinc and aluminum plating
on carbon steel; and (6) chemical etching and milling of aluminum.
F019 - Wastewater treatment sludges from the chemical conversion coating of
aluminum.
K062 - Spent pickle liquor from steel finishing operations.
-437-
-------
toxic metals and other parameters.
Toluene, 1,1-dichloroethane,
ethylbenzene and 1,1,1
trichloroethane were detected in
several treatment tank composite
samples at levels below 300 ppb.
The pH ranged from 7 to 10. TOX
varied from 0.00 to 0.36 percent
with a mean of 0.11 percent. TOO
varied from 0 to 5,900 ppm with a
mean of 1,821 ppm. Total and free
cyanide were not detected above the
analytical detection limit of 2 ppm.
Waste Treatment Chemicals. Waste
streams were used as treatment
reagents for cyanide oxidation,
hexavalent chromium reduction and
sulfide precipitation.
A waste hypochlorite stream is
used as the oxidizing agent in
cyanide treatment. The waste used
during the sampling episode was a
by-product scrubber solution from
the production of chlorinated
chemicals (RCRA Waste Code D003).
Analytical data showed low
concentrations of barium (43 ppm)
and zinc (4 ppm). The total solids
content ranged from 250,000 ppm to
330,000 ppm with the majority of
solids being dissolved. The pH of
the wastestream was 13.
A waste iron-bearing stream is
used to reduce hexavalent
chromium. The waste used during the
sampling episode was a waste
pickling acid from the metal
processing industry (RCRA Waste Code
K062). The following metals were
detected in this waste stream:
arsenic (3 ppm), total chromium
(3500 ppm), copper (532 ppm), nickel
(1553 ppm), and zinc (7 ppm). The
total solids content ranged from
70,700 ppm to 148,600 ppm with the
majority of solids being
dissolved. The pH of the
wastestream was 1.
A waste stream containing
sulfide is used in a polishing step
following precipitation with lime.
In all batches the waste stream was
a by-product scrubber solution from
the production of a sulfurized ester
(RCRA Waste Code D003).
Insignificant concentrations of
total chromium, lead and zinc were
detected in this waste stream. The
total solids content was 94,000 ppm
with most solids being dissolved.
The pH of the wastestream was 13.
Vacuum Filter Effluent. Soluble
metals data for the vacuum filter
effluent were consistently below
detection limits. This indicates
that the precipitation process was
effective in removing soluble
metals. Total metals data for the
filtrate were comparable to the
soluble metals data, which indicates
insignificant levels of insoluble
metals. The filtration system was
effective in removing insoluble
metal hydroxides and metal sulfides.
Metal removal efficiencies by
precipitation and filtration are
summarized in Table 2. The removal
efficiency for hexavalent chromium
is a result of chromium reduction as
well as precipitation and
filtration. Metals not listed in
Table 2 were not detected during the
sampling episode.
In Table 3, data for the vacuum
filter effluent are compared to the
facility's discharge limits.
Analytical detection limits for
mercury and cyanide were higher than
the discharge limits. The
hexavalent chromium concentration
exceeded the discharge limit three
times for waste treated by chromium
reduction and once for waste treated
only by precipitation and
filtration. The zinc concentration
exceeded the discharge limit once
during the sampling episode. All
-438-
-------
TABLE 2. METAL REMOVAL EFFICIENCY
Metal
Percent Removal Range
Chromium, Total
Chromium, Hexavalent
Copper
Lead (3 batches only)
Nickel
Zinc
99.97 - 100.00
99.98 - 100.00
99.78 - 100.00
99.94- - 100.00
99.90 - 100.00
70.88 - 99.93
TABLE 3. COMPARISON OF FACILITY DISCHARGE TO CITY DISCHARGE LIMITS
Pollutant
Vacuum Filter Effluent
(ppm, range)
Discharge Limit
(ppm)
Toxic Metals;
Chromium (total)
Chromium (hexavalent)
Copper
Lead
Mercury
Nickel
Silver
Zinc
Other Parameters:
Cyanide
PH
0.10
0.011
0.07
<0.01
0.20
0.190
0.21
0.01
0.31 - 0.40
<0.2
0.06 -- 1.62
<2
7
- 10
0.6
0.05
0.8
0.5
0.003
1.0
1.0
1.0
0.15
5.5 - 10.5
NS = Not Sampled
-439-
-------
other parameters listed in Table 3
were detected below the discharge
limits.
The vacuum filter effluent
samples contained an average solids
content of 5.2 weight percent with
90 percent of the solids being
dissolved. Volatile organic
compounds were not present above
analytical detection limits.
Vacuum Filter Cake. The filter cake
samples were dry with a solids
content of 35 weight percent. All
samples passed the paint filter
test. The total organic carbon
content averaged 0.57 percent.
Alkalinity was 220,000 mg/1 as
CaCOo. The filter cake samples were
subject to the EP toxicity and the
recently promulgated Toxicity
Characteristic Leaching Procedure
(TCLP) test. Table 4 summarizes the
analytical results and the EPA
treatment standards. All metals
were found at levels below the
regulatory limits. Volatile organic
compounds were detected in TCLP
extracts at levels below regulatory
limits.
Solid Waste Treatment System
Influent Wastes. Discrete solid
wastes treated during the sampling
episode were primarily from
fabricated metal products
industries. Table 5 summarizes the
types of waste streams, including
their SIC code and RCRA waste
code. All waste streams were
dewatered residuals from one or more
of the following treatment
processes: cyanide oxidation,
hexavalent chromium reduction and
lime neutralization. Discrete
wastes were composited prior to
sampling.
The untreated solids composite
had a solids content of 41 weight
percent and a total organic carbon
content of 0.48 percent. All
samples passed the paint filter
test. Acidity ranged from 0 to
1,400 mg/1 as CaCOo and alkalinity
ranged from 0 to 35,000 mg/1 as
CaCOo. The metals content was
highly variable. Total chromium,
lead, nickel and zinc were detected
at levels ranging from 100 ppm to
7,000 ppm. Barium, cadmium,
hexavalent chromium and copper were
detected at levels below 100 ppm.
Low concentrations of volatile
organic compounds were present.
Stabilized Solids. Metals results
for stabilized solids are summarized
in Table 6. All metals were found
at levels below the regulatory
limits in the sample extracts.
Styrene (200 ppb) and total xylenes
(78 ppb) were detected as well as
other volatile organics (<30 ppb).
Volatile organics in TCLP extracts
were detected at levels below the
regulatory limits. The stabilized
solids had a solids content of 46
weight percent and a total organic
carbon content of 0.57 percent. All
stabilized solids samples passed the
paint filter test. Alkalinity
averaged 135,000 mg/1 as CaCOo.
CONCLUSION
The facility discussed in this
paper was successful in treating
electroplating and other metal
finishing waste. The cyanide
oxidation operation could not be
evaluated due to high analytical
detection limits. Metal removal
efficiency was consistently higher
than 99.8 percent except for one
batch where zinc removal was 70.9
percent. In most cases, the
facility's effluent was acceptable
with respect to discharge limits.
Hexavalent chromium and zinc
exceeded discharge limits at least
once during the sampling episode.
-440-
-------
The filter cake from liquid waste
treatment and stablili2ed solids can
be delisted and disposed of in a
nonhazardous landfill. Toxic metal
and volatile organic analyses of
TCLP extracts showed compliance with
regulatory limits. In general, the
facility was well maintained with
proper attention to controlling air
emissions, to monitoring effluent
quality and to producing non-
hazardous solids.
TABLE 4. METALS RESULTS FOR VACUUM FILTER CAKE
Compositional
(ppm)
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium (hexavalent)
Chromium (total)
Copper
Le'ad
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
<10
5
40
<2
53
1.77
16,300
775
2,800
<1
4,700
<10
14
<10
2,100
EP Toxiclty
Extract
(mg/1)
0.01
<0.05
0.03
0.01
0.031
<0.005
<0.01
0.03
0.18
<0.002
0.09
<0.01
, 0.06
0.06
0.10
TCLP
Extract
(mg/1)
__
0.016
0.28
—
<0.020
—
<0.05
—
<0.10
< 0.0002
—
<0.040
' <0.020
—
—
TCLP (a)
Regulatory
Limit
(mg/1)
._
5.0
100.0
--
1.0
—
5.0
—
5.0
0.2
—
1.0
5.0
—
—
(a) Federal Register, Volume,5,1, Number 144, June 13, 1986, p. 21675.
-441-
-------
TABLE 5. INDUSTRIAL SOURCES OF SOLID WASTES
Industry
Category
Metal processing
and refining
Fabricated
metal products
Fabricated
metal products
Metal processing
Industry
(SIC Code)
»
Primary iron and
steel production (3312)
Manufacture of
semiconductors (3674)
Manufacture of
chains (3471)
Primary steel production
Process
Pickling
Etching and nickel
electroplating
Nickel and chrome
electroplating
Pickling, chrome
RCRA waste
RCRA waste
code
DOO?
F006
F006
D007
and refining
Electrical and
electronic
equipment
Fabricated
metal products
Fabricated
metal products
with secondary plating
of sheet steel (3312)
Secondary battery
manufacture (3691)
Manufacture of motor
vehicles and passenger
car bodies (3711)
Manufacture of iron
pipe fittings (3312)
and zinc electro-
plating
Lead-acid battery D008
production
Conversion coating F019
of aluminum
Cleaning K062
D007 - A solid waste that exhibits the characteristic of EP toxicity due to
the concentration of the toxic contaminant chromium.
D008 - A solid waste that exhibits the characteristics of EP toxicity due to
the concentration of the toxic contaminant lead.
F006 - Wastewater treatment sludges from electroplating operations except
from the following processes: (1) sulfuric acid anodizing of aluminum;
(2) tin plating on carbon steel; (3) zinc plating (segregated basis)
on carbon steel; (4) aluminum or zinc - aluminum plating on carbon
steel; (5) cleaning/stripping associated with tin, zinc and aluminum
plating or carbon steel; and (6) .chemical etching and milling of
aluminum.
F019 - Wastewater treatment sludges from the chemical conversion coating of
aluminum.
K062 - Spent pickle liquor from steel finishing operations.
-442-
-------
TABLE 6. METALS RESULTS FOR STABILIZED SOLIDS
Toxic Metal
Antimony
Arsenic
Barium
Beryllium
Cadmium
Chromium (hexavalent)
Chromium (total)
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
Compositional
(ppm)
30
5
• ,60
<2
21
0.30
7,500
775
.1,500
<1
4,900
<10
8
<10
11,000
EP Toxicity
Extract
(mg/1)
<0.01
<0.05
0.20
<0.01
0.014
<0.005
0.04
0.11
0.14
<0.002
0.22
<0.01
0.01
<0.01
0.04
TCLp(a)
TCLP Regulatory
Extract Limit
(mg/1) (mg/1)
_.
0.015
0.19
—
<0.05
—
0.15
—
<0.10
< 0.0002
—
<0.20
<0.02
—
—
._
5.0
100.0
—
1.0
__
5.0
. • ..
5.0
0.2
—
1.0
5.0
__
—
Federal Register, Volume 51, Number 144, June 13, 1986, p. 21675.
-443-
-------
SOLVENT RECOVERY TECHNOLOGIES
By
Robert A. Olexsey
Benjamin L. Blaney
Ronald J. Turner
Treatment Technology Staff
Thermal Destruction Branch
Alternative Technologies Division
Hazardous Waste Engineering Research Laboratory
Cincinnati. Ohio 45268
ABSTRACT
The increasing cost for disposal of hazardous wastes presents a favorable climate
for recovery of materials and energy from hazardous wastes. In the case of waste sol-
vents, the land disposal restrictions imposed by EPA on those materials on November 7,
1986, will make disposal much more difficult and costly.
This paper describes approaches to recovery of solvent wastes: fuel blending,
distillation, and steam stripping. The technologies are described and data are presented
from EPA programs to evaluate these technologies.
INTRODUCTION
Annual waste solvent generation in the
U.S. is approximately 3.2 billion gallons
(1). Prior to the imposition of the cong-
ressionally mandated land disposal re-
strictions on November 7, 1986, about 1.2
billion gallons of hazardous waste sol-
vents were disposed of to the land, pri-
marily to surface impoundments. About 300
million gallons went to deep well injec-
tion and 195 million gallons were inciner-
ated. About 545 million gallons were re-
cycled and the remainder went to storage.
Of the amount that was recycled, 80 per-
cent was recycled at the source of gener-
ation (onsite) (2).
New regulations promulgated by EPA
under the Hazardous and Solid Waste Amend-
ments of 1984 (HSWA) require substantial
treatment of the waste prior to land dis-
posal and also put severe limitations on
waste storage. HSWA also requires that
EPA evaluate deep well injection to deter-
mine its suitability for long term waste
disposal. It is obvious that acceptable
management techniques must be found to
accommodate the discrepancy between cur-
rent waste solvent generation rates and
treatment capacity.
Certain conclusions are obvious.
First, more solvent wastes will be incin-
erated. In particular, commercial incin-
eration will increase. Wastes which are
burned onsite are most likely burned as
fuel in boilers. Waste that will tend to
show up at commercial incineration facil-
ities will have low value as fuel.
Second, waste minimization will get
increased attention as costs increase and
disposal options decrease. Minimization
practices will more than likely result in
lower volumes of total solvent waste en-
tering the waste management system but
these solvent wastes will be more dilute
and of lower quality as a fuel, therefore
making the wastes less amenable to incin-
eration.
Therefore, there will exist a greater
need for application and improvements of
-444-
-------
concentration techniques. Once the waste
solvent is concentrated, it will be a
function of purity, degree of separation,
and fuel value that determines whether the
solvent is reused as a solvent or burned
as a fuel.
Solvent Recovery Technologies
Solvent recovery technologies consist
of three general types of processes: con-
centration, refinement, and use as fuel.
Examples of concentration processes in-
clude drying, evaporation, steam strip-
ping, and simple distillation. The most
prominent example of refinement is frac-
tional distillation. Solvents are burned
as fuel in industrial boilers and indus-
trial furnaces such as cement kilns, lime
kilns, and steel making furnaces. The
remainder of this paper will provide de-
scriptions of the principal solvent recla-
mation processes, including thin film
evaporation, steam stripping, and frac-
tional distillation. Data are provided on
the performance of these technologies in
spent solvent processing operations.
Steam Stripping
Figure 1 illustrates a typical steam
stripping process. Waste enters near the
top of the column and then flows by grav-
ity countercurrent to steam (3). As the
waste passes down through the column it
contacts vapors rising from the bottom of
the column that contains progressively
less volatile organic compounds. The
concentration of volatile compounds in the
waste reaches a minimum at the bottom of
the column where it is discharged. The
overhead vapor is condensed as it exits
the column and the condensate is then
decanted to achieve solvent/water separa-
tion. Reflux may or may not be used,
depending on the desired composition of
the overhead stream.
Steam stripping is most applicable to
low concentrations of organic materials in
water. Volatile organic concentrations in
excess of 10 percent may be more cost
effectively treated with distillation.
Also, organic compounds with boiling
points in excess of 150°C cannot be
treated effectively with steam stripping.
Recently, EPA conducted bench scale
steam stripping of a petrochemical process
wastewater stream containing halogenated
organics. The waste stream contained a
number of constituents found in waste sol-
vents. Table 1 exhibits data on compound
removal efficiency for the 5.08 cm dia-
meter, 367 cm high, counterflow column.
Recoveries of volatile organics averaged
85 percent with steam flows of 18 to 24
percent of the feed (4).
The concentrated solvent product from
the steam stripper could be used in an
application where mixed constituents are
acceptable (e.g. certain degreasing oper-
ations). It could also be processed fur-
ther to separate constituents (through
distillation) or burned as a fuel.
Thin Film Evaporation
Agitated thin film evaporation (ATFE)
is the most effective high volume evap-
oration technique for separating low and
medium boiling point waste components,
while concentrating high boiling and/or
solid constituents. Its ability to
handle a wide range of waste viscosities
(1 to 1,000,000 cps) makes it the most
common treatment device at large commer-
cial solvent recycling companies.
Liquid waste is fed to the top of
ATFEs where longitudinal blades mounted
on a motor driver rotor centrifugally
force the waste against the heat transfer
surface which is the inside wall of the
cylindrical vessel. This surface is en-
closed in a heating jacket which employs
steam or hot oil as the heating medium.
Temperatures of the working fluid can
reach 650°F. The agitation and liquid
film are maintained by the blades as they
move along the heat transfer surface. The
blade tips typically travel 30 to 40 feet
per second at a clearance of 0.007 to 0.10
inches which creates high turbulence.
This facilitates efficient heat and mass
transfer, shortens required waste resi-
dence time and creates a degree of mixing
which maintains solvents in a manageable
suspension without fouling the heat trans-
fer surface. Viscosity is the critical
waste property that affects thin film
evaporator performance.
A schematic of an ATFE and associated
pretreatment and post-treatment options
are shown in Figure 2 (5). The pretreat-
ment techniques most often undergoing ATFE
are a previous solvent recovery process,
oil or suspended solids removal, or a
-445-
-------
CO
CO
o
o
cc
a
o
Z
a
a
cc
ui
»-
CO
AQUEOUS PHASE)
c
td
U
Cd
Q
^s
X
Cu
Id
as
OS
NON-CONTACT
DOLING WATER
u
-"*
§
cd at
2~~ O
cvi
^ d
a
r-) c_i
1- td Z
O H
td O 1
04 H H
W CO
S» Cd O
82 * '
Id 2 04
PS eu o
t
»
0 C3
otf •< cu £
Cd U Pu ^
H H (-" iJ
Z en oi o
3 H cj
ej __
A
«
td f H
< C W ==
S > M H
a f §d
BU (. S S
[RECT STEAM
ADDITION
a
en td
td •<
O M
o-o
•< en
^, n S
52
td en
oi
-------
POST-TREATM
OPTIONS
§
o£2
z <
>-< z tn
SS*
H-&3
DEC
DE
-
-
HI
u.
H
<
cc
o
LL
5'5
< 1-1
WOE-
s
§3
PS
W
ea bi i-i o- u o cr
z ea cn z M >H
w o j
3 Z O
fc< M to
O >• >« '
oi E o,
•« o c- o
II
O 32 t>
^ H W
UJ
u
CC
LU
OS
Z3
CD
P z
< O
£jj rt
a- H
6- Cu
WO
M o
H >-H
< H
en
Q W
O
•J Z
o o o
Z en >-•
I-. -^ f-
t- e -
s
. . <
O C' C.
UJ i-. LU
O J CO
I
t> O
>J bJ
O W
en CL.
CU
co
fl
en
-447-
-------
Post treatment is almost always required.
These operations basically include further
refinement of the overhead produce through
dewatering or fractionation, further sol-
vent recovery, or disposal of bottom
products.
The USEPA recently evaluated a thin
film evaporator at a recycler which was
using the process for the treatment of
paint and lacquer thinner. The thinner
contains up to 20 percent dissolved paint
resins plus suspended pigments and non-
halogenated solvents. The overhead prod-
uct from the ATFE is sold back to the
customers without high boiling components,
such as xylene. The bottoms product is
sold as a waste fuel.
Table 2 displays performance data for
the ATFE (6). The unit has 20 square feet
of heated surface area with a steam heated
jacket. The blades operate at 300 revolu-
tions per minute (rpm) and the unit oper-
ates under up to 28 in Hg of vacuum.
During the test, the unit processed 1530
gallons of feed material over a 6.75
hour period, generating 830 gallons of
distillate product and 688 gallons of
still bottoms. Steam pressure was 150
psig and steam temperature was 365°F.
Overhead product temperature was 125°F
and bottom temperature was 140°F.
Viscosity is the critical waste
property that affects thin film evaporator
performance. ATFEs can operate in the
range of from 1 to 1,000,000 cps.
Fractional Distillation
Distillation is a separation tech-
nique that operates on the principle of
differential volatility. More volatile
constituents can be enriched or separated
from less volatile constituents by heat-
ing. Distillation can be simple, in
which the objective is to concentrate a
volatile component, or it can be frac-
tional, in which the goal is to separate
two or more volatile components.
Fractionation is a multi-stage proc-
ess used for separating solvent mixtures
when the value of the pure component pro-
duct justifies the additional processing.
As with simple distillation, indirect heat
from steam or oil provides the thermal
driving force. While simple distillation
most often occurs in a pot or kettle,
fractionation takes place in a column.
Figure 3 is an illustration of a continu-
ous process. In this operation feed is
constantly charged to the column at a
point which provides the specified top and
bottoms product. The section of the tower
above the feed point is the rectifying or
enriching section and the section below
the feed point is the stripping section.
A reboiler is connected to the bottom
of the fractionation tower to provide the
heat needed for additional reflux (conden-
sate recycle) and better fractionation of
complex mixture (6).
EPA recently conducted a field test
of a full scale batch distillation unit
that was used to process a liquid feed
containing three major components:
methanol, methyl ethyl ketone (MEK), and
methylene chloride. The water content of
the feed was 3.59 percent. The atmos-
pheric distillation unit was equipped
with a 6 foot high, 10 inch diameter
packed column and had a batch capacity
of 500 gallons.
Table 3 provides data on the perform-
ance of the fractionation unit on the
three principal constituents. The data
in Table 3 shows that the lowest boiling
component (methylene chloride) boiled off
early, appearing at highest concentrations
in the first product cut and decreasing
concentrations in subsequent cuts.
Methanol concentration in the products was
relatively constant. The least volatile
of the three major components (MEK) was
present in low concentration in the first
two cuts and increased in concentration as
distillation continued. From Table 3 it
can be seen that distillation provides ex-
cellent separation for the low boiling
point compounds. Further refinement could
be achieved through additional product
cuts (higher column) or through further
distillation of the individual product
cuts (7).
Waste Solvent End Use
Waste solvents can be either reused
as solvents or burned as fuel. Histori-
cally, the extent of solvent recovery has
been dictated by equipment processing
capability and local marketing situation.
However, regulatory constraints imposed on
the bottoms product will increasingly be-
come the decisive factor in routing sol-
vents to recovery through treatment.
-448-
-------
VAPOH
*- STCAU
CONOCHSAIf
BOTTOUS PRODUCT
FIGURE 3.CONTINUOUS FRACTIONATION
-449-
-------
TABLE 1. REMOVAL EFFICIENCY DATA FOR STEAM STRIPPER
Feed Composition
(mg/L)
Average Removal
Efficiency (Percent)
1,2-Dichloroethylene
Ethyl ene Di chloride
Dichloromethane
1,1,1,2-Tetrachloroethane
Chloroform
1583.3
1593.0
800.9
572.8
140.3
85.9
91.3
76.7
99.9
89.9
TABLE 2. PERFORMANCE OF THIN FILM EVAPORATION
(Al1 values in mg/kg)
Constituent
Waste Feed
Still Bottoms
Distillate Product
Xylene
Acetone
Ethyl Acetate
Ethyl Benzene
Methyl Isobutyl
Ketone
n-Butyl Alcohol
Toluene
Methyl Ethyl
Ketone
Isopropanol
66,000
190,000
11,000
22,000
14,000
11,000
220,000
180,000
76,000
210,000
5,200
12,000
48,000
5,600
5,500
81,000
57,000
5,500
84,000
180,000
12,000
28,000
16,000
13,000
240,000
220,000
84,000
TABLE 3. FRACTIONAL DISTILLATION PERFORMANCE
Compound
Boiling
Point (°C)
Concent rat i on in g/1
—StTTl
Product Cuts
Influent Bottoms
2
3
Methanol 64.5
Methylene Chloride 40.0
Methyl Ethyl Ketone 79.6
300
460
240
230
<0.5
370
160
1100
42
190
710
65
180 190
590 58
280 340
-450-
-------
Whether the end use is for solvents
or for fuel, the treatment concentration
requirements are identical. The decision
to further concentrate and/or fractionate
will be a function of the relative value
of-the recovered solvent versus the in-
creased cost of treatment for recovery.
In any case, solvent bottoms will
find increasing use as fuels. In 1984,
at least 62 percent of solvent bottoms
from commercial recovery facilities were
used as supplemental fuel.
Commercial Facilities
The commercial solvent recycling in-
dustry currently consists of at least 135
firms with 243 facilities in the U.S.
Large firms tend to serve regional markets
and are capable of producing high-purity
solvents through the use of distillation,
fractionation, and other specialized re-
covery techniques. The list below de-
scribes solvent recovery technology utili-
zation in the commercial solvent recycling
industry (3):
Process
Simple distillation
Fractionation
Thin-film evaporation
Steam stripping
Drying
Solvent extraction
Use as fuel
No. of Facilities
44
22
43
18
17
6
48
Of course, many facilities use combi-
nations of equipment in series to produce
a finished product. With the imposition
of the land disposal restrictions and the
possible constraints to be imposed on deep
well injection, both the number of facil-
ities and the adoption of advanced tech-
nologies should increase.
Conclusion
Solvent recovery is an important
waste management practice today. A
greater percentage of solvents are re-
covered today than any other hazardous
waste streams. The degree of purity and
selection of end use are not limited by
technology but are a function of relative
costs of treatment versus product or fuel
value. Treatment and disposal costs will
rise as land disposal alternatives are
eliminated. Therefore, more solvent re-
covery will be practical in the future.'
References
1. 40 CRF Parts 260-271, Hazardous
Waste Management System, Land Dis-
posal Restrictions, Proposed Rule,
January 14, 1986.
2. Engineering Science, "Supplemental
.Report on the Technical Assessment
of Treatment Alternatives for Waste
Solvents," Report to USEPA, 1985.
3. Breton, M., et a!., "Technical
Resource Document: Treatment Tech-
nologies for Solvent Containing
Wastes," Report to USEPA, HWERL,
August 1986.
4. Coco, J. H., et al., "Development of
Treatment and Control Technology for
Refractory Petrochemical Wastes,"
Report to USEPA by Gulf South
Research Institute, EPA-600/2-79-080,
April 1979.
5. Koppenberger, P. F., et al., "Thin-
Film Technologies Environmental
Protection," Chemical Age of India,
V. 36 (1), January 1985.
6. Allen, C. C., et al., "Field Eval-
uation of Hazardous Waste Pretreat-
ment as an Air Pollution Control
Technique," Report to USEPA, HWERL,
by Research Triangle Institute, EPA
600/2-86/048, January 1986.
7. Metcalf and Eddy, Inc., Facility
Report Test Report for Environmental
Waste Enterprises," Report to USEPA,
HWERL, February 1986.
-451-
-------
EVALUATION OF HAZARDOUS WASTE RECYCLING PROCESSES
IN THE PRINTED CIRCUIT BOARD INDUSTRY
Thomas J. Nunno
Stephen Palmer
Mark Arienti
Alliance Technologies Corporation
Bedford, MA 01730
ABSTRACT
In response to the 1984 RCRA Amendments, EPA1s Hazardous Waste Engineering Research
Laboratory (HWERL) initiated a program to develop case studies demonstrating waste
minimization and recycling options for hazardous waste management. The program focused
on solvent and metal waste streams from the semiconductor and printed circuit board
industries, specifically: 1) waste solvents from resist stripping and developing
operations; and 2) metal wastes from plating operations. Two case studies involved the
use of solvent distillation units which achieved over 95 percent recovery of spent
halogenated solvents. The results suggest that solvent recovery can be widely applied
to printed circuit board manufacturing facilities. The other four case studies focused
on technologies to reduce metal-plating wastes. Two of these, evaluating the use of
sodium borohydride reduction as a substitute for lime/ferrous sulfate precipitation,
found that the technology was a viable substitute in one case and was marginally
acceptable in another. Another case study, involving carbon adsorption removal of
organic contaminants from plating bath wastes, found that this technology significantly
reduced both disposal costs and waste volume. A final case study of electrolytic
recovery indicated that while acid copper electroplating rinses are amenable to
electrolytic recovery, other metal-bearing rinses, such as those from solder (tin/lead)
plating or etching are less appropriate.
BACKGROUND
With the enactment of the Hazardous
and Solid Waste Amendments (HSWA) in
November 1984, Congress set forth a
schedule for evaluating the land disposal
restriction of various classes of hazardous
wastes including: (I) solvents; (2) metals
and cyanides; (3) halogenated organics;
(4) corrosives; and (5) dioxin wastes. A
key issue identified in the evaluation of
the waste bans is the availability of
commercial treatment capacity to handle
the wastes proposed for banning.
Therefore,Congress also asked EPA to
evaluate the potential for onsite waste
minimization to reduce the quantity or
toxicity of wastes being considered under
the ban.
In an effort to identify successful
waste minimization technologies, EPA's
Office of Solid Waste (OSW) and Office of
Research and Development (ORD) Hazardous
Waste Engineering Research Laboratory
(HWERL) set forth on research efforts
aimed at assessing the viability of waste
minimization as a means of reducing the
quantities of land disposed hazardous
waste. OSW's research focused on an
exhaustive literature review identifying a
broad spectrum of waste minimization
technologies and their various
applications. The primary emphasis of
HWERL's work was on demonstrating the
effectiveness of specific minimization
technologies through case studies and
process sampling.
-452-
-------
INTRODUCTION
The purpose of this project was to
evaluate the effectiveness of various
waste minimization practices or
technologies in the printed circuit board
and semiconductor manufacturing industries.
The most significant waste streams in
these industries are waste halogenated
solvents from photoresist stripping and
developing operations (RCRA Haste Code
F001-F003), and metal-bearing sludges
(RCRA Waste Code F006) from the treatment
of metal plating and etching rinsewaters.
This paper presents the findings of case
studies conducted at five printed circuit
board manufacturing facilities and one
commercial .treatment/recovery facility.
Each facility investigated employs some
practice that requires offsite disposal.
Two of the case studies focus on the
recovery of spent halogenated solvents,
and the remaining four discuss the recovery
or reduction of metal plating and etching
process wastes. Table 1 summarizes
characteristics of facilities investigated
which range from small job shops to large
integrated facilities.
The common objectives of each of the
technologies evaluated are:
(1) minimization of metals sludges
generated; (2) compliance with effluent
guidelines or local discharge limitations;
and (3) reduction in operating costs over
other conventional alternatives. The
following discussion briefly summarizes
each case study, the nature of the
minimization technology, the measurements
data collected and the results obtained.
METAL PLATING BATH'WASTE MINIMIZATION CASE
STUDY RESULTS
Metal plating wastes'generated from
plating bath dumps, rinses, etching
machines and scrubbing operations generate
copper-, nickel-, tin-, and lead-
contaminated wastes. Four of the six case
studies investigated under this research
project focus on the minimization of
sludges generated primarily by copper
plating and etchant baths and copper and
tin/lead rinsewaters.
Facility A Case Study
Description--
Facility A is an offsite Treatment,
Storage, and Disposal (TSDF) facility
which processes concentrated dumps from
the metal plating and printed circuit
board industries, including alkaline
etchants, acid plating baths, nitric acid
rack strip baths, and electroless plating
cyanide baths. The average total metals
concentration in the incoming waste was
reportedly 12 g/L (12,000 ppm). Initially,
the facility was designed to operate using
lime and ferrous sulfate precipitation of
TABLE 1. SUMMARY OF FACILITIES TESTED UNDER WASTE MINIMIZATION CASE STUDY PROGRAM
Fac i1ity name
Description
Wastes treated/reduced
Technology
Facility A Treatment storage disposal
facility handling electro-
plating baths, waste
etchante, spills., etc.
Capacity: 1}000 gph
(24,000 gpd).
Facility B Contract PC board
manufacturing shop
Employees: 77
Production: 500,000 ft2/yr
Sales: $7 MiLlion/yr
Nickel plating baths
Copper plating baths
Cyanide
Cupric chloride etchant
Electroless plating rinses
Electroplating rinses
Sodium hydroxide precipitat ioi
Sodium borohydride reduction
Alkaline chlbrihation
Sodium borohydride reduction
Memtek ultrafiltration system
Sludge product
Sludge product
Facility C
Facility D
Facility E
Facility F
Computer manufacturer.
Employees: 10,000
Electronic equipment rafgr.
PC board manufacturing using
the subtractive technique in
the HacDermid process.
' Employees : 260
Computer manufacturer.
PC board manufacturing using
additive techniques.
Employees: 600
Production: 600,000 ft2/yr
PC Board manufacturer.
2-sided single layer circuit
' boards.
Production: 480,000 ft^/yr
- Methyl chloroform resist
developer
- Freon resist developer
~ 1, 1, 1-trichloroethane
resist developer
— If 1, 1-trichloroethane
still bottoms
- Acid copper plating bath
- Acid copper plating
rinsewaters
- Tin/lead plating
rinsewaters
- Solvent distillation/
fractionatlon recovery of
resist developers.
- 2-stage solvent distillation
- (1) DuPont RISTOH SRS-120
solvent recovery still
- (2) Recyclene Products, Inc.
RX-35 still
- Activated carbon regeneration
of spent plating baths.
- Agtnct Equipment Corp.
- electrolytic recovery units.
Still bottoms
Still bottoms
Spent activate'
carbon
Metal foil
-453-
-------
metals as the primary means of waste
treatment. When the high cost of land
disposal of the lime sludges was
considered, alternate means of treating
and disposing of the waste were evaluated.
The unit processes selected to
detoxify the wastes and recover metals at
Plant A currently include sodium
hypochlorite oxidation of cyanides
(alkaline chlorination), sodium hydroxide
precipitation, pH adjustment, sodium
borohydride reduction (with sodium
tnetabisulfite stabilization),
sedimentation, plate and frame filter
press (for sludge dewatering), rapid sand
filtration, and ion exchange columns for
effluent polishing.
Results—
The primary purpose of the Facility A
case study was to evaluate sodium
borohydride as a viable waste treatment
alternative for reducing RCRA Hazardous
Waste Code F006 spent electroplating
baths. The evaluation criteria were the
ability of sodium borohydride (SBH) to
effectively meet local compliance
standards and produce a high density,
low-volume sludge. The test program
evaluation relies mainly on the trace
metals results to evaluate system
performance.
The SBH reactor was sampled for eight
selected metals on the influent, effluent,
and sludge streams. Analytical results
demonstrated individual metal reduction
efficiencies which ranged from 16.1 to
99.8 percent. The observed range, in
efficiency data was attributed to
variations in concentration and chemical
potential (quantity of free energy
required for an ionic species to obtain
equilibrium) of each of the metallic ions
contained in the solution. Overall, SBH
was able to reduce 6.91 kg of the initial
influent metals loading of 7.25 kg. These
results represent a greater than
95 percent reduction in total metals for a
complex waste stream. The remainder of
the metals influent loading (0.337 kg)
consisted of over 70 percent calcium.
An additional objective of this
program was to evaluate the ability of
Facility A to consistently meet local
pretreatment requirements. The resultant
data for two separate batch runs showed
discharges in excess of effluent limits,
apparently due to incomplete polishing
caused by cation exchange column
breakthrough. Since the test program was
completed, Facility A has instituted the
use of a. quality control holding tank and
further waste processing optimization to
remedy these problems. Follow-up
discussions with the local sewer authority
revealed that Facility A's effluent
quality has improved considerably and is
now consistently meeting compliance
guidelines.
In addition to assessing wastewater
effluent characteristics, the testing
program was designed to evaluate
uncontrolled process air emissions. The
results were obtained by Draeger tube
analysis of grab and integrated samples of
exhaust gases taken from the process
reactor exhaust ducts. The emission
results showed a frequent presence of
hydrochloric acid and hydrogen gas
accompanied by occasional presence of
ammonia and sulfur dioxide. One of the
hydrogen emissions grab sample results
(6.0 percent) is significant since this
value is greater than the lower flammable
limit for hydrogen (4.0 percent). Grab
sample concentrations for ammonia and
sulfur dioxide also exceeded adopted
short-term exposure limits (STEL) for
these substances.
Due to SBH sludge filter press
operational difficulties, realistic data
as to SBH's ability to produce a high
density, low volume sludge was
unobtainable. However, EP toxicity
analyses of the sludge produced indicated
that for Facility A influent metals
concentrations, the SBH sludge produced is
fairly stable in that its leachate
characteristics are below EP Toxicity
limits for all metals. However, note that
the waste is still classified as F006
hazardous waste.
An additional objective of the
Facility A case study was to evaluate the
ability of sodium borohydride to
economically reduce F006 waste streams.
At the time of testing, Facility A
reduction chemistry was very inefficient
at $19.8/lb of copper reduced. However,
through process optimization, chemical
costs have reportedly decreased over
63 percent,! bringing process economics
within acceptable limits. The case study
follow-up for Facility A has indicated
-454-
-------
that the cost of copper reduction has been
lowered to $7.27/lb of copper. ' •
Facility B Case Study
Description--
Facility B is a captive printed
circuit board manufacturing facility.
Production at Facility B uses a special
hybrid process, employing elements of both
additive arid semi-additive printed circuit
production techniques. Facility B uses a
slow acting etchant {sodium chloride,
sodium chlorate, and muriatic acid) which
etches copper from the board, and yields
cupric chloride in the waste stream.
Facility B uses a unique end-of-pipe
treatment system employing sodium
borohydride treatment and ultrafiltration
(Memtek) technology for solids
separation. In this process, incoming
plating and etching wastes are adjusted to
pH 7-11 by addition of sodium hydroxide or
sulfuric acid. Sodium borohydride is
added to obtain an oxidation reduction
potential (ORP) of approximately -250 mv
or less. The reacted waste then feeds
from the concentration tank to a Memtek
ultrafiltration unit from which the
permeate is discharged to municipal
treatment, and the concentrate is returned
to the concentration tank. A small plate
and frame sludge filter press dewaters the
sludge which is drawn from the bottom of
the concentration tank.
Points of interest in evaluating the
Facility B waste treatment system for this
case study were: (1) compliance of the
ultrafiltration permeate (wastewater
discharge) with local and Federal
discharge standards; (2) the volume and EP
toxicity of the sludge filter cake; arid
(3) economic, evaluation against comparable
technology Clime and ferrous sulfate
treatment).
Results—
Analysis of the influent and effluent
streams metals characteristics, showed
that copper was reduced most efficiently
(99.82 percent), while nickel reduction
was the least efficient at (45.5 percent).
Differences in removal efficiencies were
attributed to variations in concentration
(higher removals for higher
concentrations), but the chemical
potential may also have been a factor.
Approximately 144.7 Ibs of combined metals
were reduced to elemental.form by the SBH
reaction system, representing a combined
reaction efficiency of 99.8 percent.
Despite deviations from design operating
conditions, the SBH/ultrafiltration "system
performed very well. EP Toxicity leachate
test results for Facility B filter press ,
sludge clearly sbbw\that the sodium
borohydride sludge produced is fairly
stable with leachate characteristics below
EP Toxicity limits for all metals.
An economic comparison pf the use of
sodium borohydride versus lime-ferrous
sulfate chemistries was conducted. The
results demonstrate that in this,
application, sodium borohydride w^uld be
superior to lime-ferrous sulfate for the
following reasons: (1) sludge disposal '
costs and volumes would be reduced by
93.5 percent; (2) overall operating
expenses would be 48 percent lower; and
(3) sludge generated by the SBH reduction
process was 78 percent copper and suitable
for reclamation (due to the high copper "
content). Note that the use of the sodium
borohydride and ultrafiltration treatment
at Facility B is favored by the use of the
sodium chloride etch process in lieu of
the more commonly preferred ammonium
chloride etch process. The ammonium-based
etchants create borohydride sludge
reduction problems which require tighter-
treatment process control and the use of
primary reductants such as sodium
metabisulfite. Additional factors which
favor the economics of sodium borohydride
treatment at Facility B include: 1) high
copper concentrations and low organic
loadings; and 2) low effluent limitations'
required by the sanitation district.
Facility E Case Study
Description— • '
In Facility E plating operations,
addition agent and photoresist breakdown
products incrementally accumulate and
contaminate electrolytic (charge carrying)
plating baths. In the absence of a bath
regeneration system, Facility E would"
typically be forced to either discharge'
the spent plating bath to the wastewater
treatment plant or send it offsite for
disposal. In either case, large quantities
of metals containing sludge (RCRA Waste
Code F006) would be .generated and
subsequently land disposed. instead,
-455-
-------
-------
the performance of these units, samples of
the plating bath, dragput, and rinse bath
were analyzed.
The following conclusions were drawn
based on the resultant data. Recovery
rate of copper from the acid copper
solution was approximately 4 to 5 grams/
hour/unit, representing an efficiency of
nearly 90 percent. The investigators also
noted that concentrations of tin and lead
in the dragout were not significantly less
than those in the plating bath indicating
a poor level of recovery. Thus, use of
in-line electrolytic recovery was not able
to reduce metal concentrations sufficiently
to enable this facility to meet
pretreatment standards.
Electrolytic recovery would
significantly reduce the amount of sludge •
generated if a lime precipitation system
were utilized to remove metals from the
final plant effluent. For this facility,
a reduction and subsequent waste
minimization of 32 tons'/year would be
realized. At a sludge disposal cost of
$200/ton, the annual cost of electrolytic
recovery would exceed the savings.
However, if sludge disposal costs
increased to $300/ton, fch6 savings (at
least for copper recovery) would exceed
the processing costs.
RESIST DEVELOPING SOLVENT RECOVERY CASE
STUDIES
Two case studies evaluated under this
program focused upon the minimization of
developer solvent wastes and sludges which
might require either land disposal or
incineration. In the case of Facility C,
the technology involves the separation of
a two-solvent system with subsequent
recovery and reuse of each solvent. In
the case of Facility D, the technology
evaluated further recovers the solvent
bottoms product of the initial recovery
unit.
Facility C Case Study
Description—
The spent solvents from photoresist
stripping and developing are contaminated
with photoresist solids at up to 1 percent,
and the solvents used for surface cleaning
are contaminated by dust, dirt or grease.
Waste solvents are recovered at Plant C by
distillation or evaporation and returned
to the process in which they were used.
Several types of equipment are used
including flash evaporators to recover
methyl chloroform, and a distillation
column to recover Freon.
There are two identical flash
evaporators at the facility, each with a
capacity to recover 600 gallons of methyl
chloroform (MCF) per hour. The flash
chamber operates at a vacuum of 20 in. Hg,
allowing the MCF to vaporize at 100 to
110°F. The units are operated one to
two shifts/day -depending on the quantity
of waste solvent being generated.
A packed distillation column is used
to recover pure Freon from a waste solvent
stream containing approximately 90 percent
freon and 10 percent methyl chloroform.
Waste is continuously fed to a reboiler
where it is vaporized and rises up the
packed column. Vaporized freon passes
through the column, is condensed and
recovered at a rate of 33 gal/hour. MCF
condenses on the packing and falls back
into the reboiler. The distillation
bottoms are removed when the concentration
of methyl chloroform reaches 80 percent
(approximately i to 2 weeks).
Results—
Sampling and analysis was conducted
on process streams associated with the
solvent recovery processes. One of these
processes was the flash evaporator used
for recovery of methyl chloroform
(1,1,1-trichloroethane), and the other was
the distillation column used to recover
Freon from a Freon/methyl chloroform
mixture. The conclusions drawn from the
sampling and testing program were: 1) at
least 95 percent of the solids are removed
from the solvent waste influent; 2) the
recovered product is at least as clean as
the virgin material; and 3) the still
bottoms from recovery of contaminated
solvent still contain a high fraction
(90 percent) of solvent.
In recovering spent solvent, the
company saves over $10 million annually,
compared to offsite recovery. The savings
per pound of methyl chloroform, methylene
chloride, and Freon recovered is $0.18,
$0.18, and $0.61, respectively. The high
cost savings are primarily due to the fact
that the solvents recovered are reused
-457-
-------
onsite, thus reducing the quantity (by
greater than 95 percent) of new or virgin
solvent that must be purchased. Because
the rate of generation of spent solvent is
so high, the initial expense of purchasing
recovery equipment is quickly returned.
Incentives other than economic
reasons for onsite recovery include:
1} reduction in the risk of a spill of
solvent in transporting the waste to a
TSDF; and 2) reduced liability related to
an accident at the XSDF resulting in the
release of spent solvent.
Facility D Case Study
Description—
Printed circuit boards are produced
at Facility D using the subtractive
technique and solvent-based photoresists.
Methylene chloride resist stripper and
1,1,1-trichloroethane (TCE) developer are
continuously recycled in closed-loop
stills. The TCE developer wastes (Haste
Code F002) are recovered in a DuPont
Riston SRS-120 solvent recovery still
(referred to as the primary still) and
returned to the developer line. Until
recently, all still bottoms from the
primary still were drummed and shipped
offsite for reclamation at a solvent
recycling facility. Facility D purchased
a Recyclene Industries RX-35 solvent
recovery system (referred to as the
secondary still) in October 1985, to
recover additional TCE from the SRS-120
still bottoms.
The Recyclene Industries RX-35 solvent
recovery system is a batch distillation
system with a 30 gallon capacity, silicone
oil immersion heated stainless steel
boiler, a noncontact, water-cooled
condenser, and a 10 gallon temporary
storage tank. The boiler is equipped with
a vinyl liner inside a Teflon bag. The
Teflon bag provides temperature resistance
and the vinyl bag collects solid residue,
eliminating boiler clean-out and minimizing
sludge generation after distillation. Two
thermostats control the temperature of the
boiler and the vapor, automatically
shutting down the boiler when all the
solvent has evaporated. The maximum
operating temperature of the still is
370°F, so recovery of solvents with higher
boiling points would not be practical.
Recovery of a 20 to 25 gallon batch of
Still bottoms requires approximately
90 minutes at Facility D, and four to six
batches are completed each day.
Results—
Evaluation of the system consisted of
the analysis of the contaminated feed,
overhead product, and distillation bottoms.
Based on a mass balance and analytical
data, the following conclusions were made:
1) purity of recovered solvent was
99.99 percent; 2) total solvent recovery
was 99.78 percent; 3) Still bottoms
contained 7.5 weight percent
1,1,1-trichloroethane; and 4) reduction in
waste generation was 97.5 percent..
Annual cost savings ($43,000) and
waste reduction (10,602 gal) were
calculated for Plant D, based on the first
year of RX-35 operation. In addition, the
investment payback period for the RX-35
was calculated considering credit for
reclaimed solvent and reductions in waste
transportation and disposal costs. The
estimated payback period was 7.3 months,
given the current level of solvent
reclamation. Thus, the low capital cost
of the unit and the relatively high costs
of virgin solvent ($4.50/gal) favor the
second-stage recovery of TCE developer
still bottoms.
Potential drawbacks to the
implementation for a RX-35 back still
include 1) while this technology
significantly reduces the volume and
toxicity of the solvent still bottoms, it
continues to generate a hazardous waste
product, 2) there is a large potential for
the accumulation of contaminants and/or
breakdown products.
SUMMARY OF FINDINGS
The findings of the waste minimization
case studies evaluated under this program
are presented in Table 2, which includes
data collected by the facilities and
verified by sampling and laboratory
results. These results indicate that a
good variety of technologies exist to
minimize metals-containing and solvent
wastes produced by the printed circuit
board and semiconductor industries. The
technologies discussed range from simple
changes in treatment system reagents with
nominal capital costs to large onsite
solvent reclamation facilities with
significantly higher capital costs.
-458-
-------
TABLE 2. SUMMARY OF FINDINGS OF WASTE REDUCTION/RECYCLING CASE STUDIES
Facility
name
Facility A
Facility B
Facility C
Technology
Sodium borohydride reduction
Sodium borohydride reduction
Solvent batch distillation
Waste reduction
Metals sludge
Metals sludge
Methylene chloride
Methyl chloroform
Freon
Annual waste
reduction
achieved
a
962 tons
6,152,000 gal
Capital costs
(J)
Nominal
Nominal
709, AGO
'•
Projected
annual cost
savings
(4)
__b
115,870
16,000,000
Facility D 2-Stage solvent distillation 1 , 1, 1-Trichloroethane
Resist developer
still bottoms
10,625 gal
26,150
"Not quantifiable, but a significant waste reduction was realized.
^Not demonstrated during testing.
c( ) indicates negative value.
43,105
Facility E
Facility F
Carbon adsorption
plating bath reclamation
Agraet electrolytic
recovery unit
Platini
(metal!
Metals
; bath wastes
! sludge)
sludge
10,600 gal
32 tons
9,200
30,350
57,267
(10,685)c
Four of the case studies investigated
under this program focused on technologies
to reduce metal-plating rinsewater sludges.
The use of sodium borohydride as a
substitute for lime/ferrous sulfate was
found to be viable in one case and •
appeared to be marginally acceptable in
another. The case study on carbon
adsorption recovery of plating bath wastes
found that this technology significantly
reduced both disposal costs and waste
volume. The case study of electrolytic
recovery indicated that this technology is
highly waste stream specific. An acid
copper electroplating rinse is an ideal
waste stream for electrolytic recovery.
However, other metal-bearing rinses, such
as those from solder (tin/lead) plating or
etching, are not appropriate for use of
electrolytic recovery. Electrolytic
recovery units are, however, generally
inexpensive to purchase and can be used in
many cases to supplement an end-of-pipe
treatment process.
Two of the case studies presented in
this paper involve the recovery of spent
halogenated solvents using solvent
distillation units. Both of these case
studies indicate that onsite solvent
recovery is successful -from a technical
and an economic standpoint. In both
cases, over 95 percent of the waste
solvent was recovered and reused onsite.
Splvent recovery appears to be a technology
that could be applied to a number of
printed circuit board manufacturing
facilities.
The results of this project indicate
that waste reduction can be achieved
through the use of appropriate technology,
and it^can be achieved with significant
reductions in cost. The case studies also
indicate that the success of waste
reduction is in many cases waste stream
specific. The technologies will not
necessarily be successful in all cases. A
slight variation between one waste stream
and another may make waste reduction
either technically or economically
impractical. Therefore, successful waste
reduction is dependent on a thorough
knowledge of waste quantities and
characteristics.
REFERENCES
1. Rosenbaum, W., ETTCAM-RI Inc.
Warwick, R.I. Personal communication
with Thomas J. Nunno, Alliance
Technologies Corporation,
December 12, 1985.
-459-
-------
THE CALIFORNIA INNOVATIVE ALTERNATIVE TREATMENT AND
RECYCLING DEMONSTRATION PROJECTS PROGRAM
Jan Radimsky, P.E. and Robert Ludwig
California Department of Health Services
Sacramento, CA 95814
ABSTRACT
.
Alternative waste management strategies being studied include, in
order of preference, (1) source reduction, (2) recycling, and
(3) treatment.
This paper discusses state and federal statutory requirements and
problems encountered in the process of implementation of th e ^ents
of California's Waste Reduction Program. Specific studies described
incSSe (1? waste stream -information collection, (2) assessment of
eSSSt of use of waste treatment in California, (3) waste audit
SSgSams, *4? waste management information transfer, and (5) discus-
sion of preliminary results from demonstration projects.
INTRODUCTION
The California Department of
Health Services' Waste Reduction
Program was initiated in July,
1984. The program has been devel-
oped with input from environmental
groups and industry. In addition,
the Department has reviewed the
efforts of waste reduction programs
from six states and incorporated
their regulations which were appli-
cable to California. The various
program elements are designed to
overcome different technical and
financial barriers, to help facili-
tate permitting, and to meet regu-
latory requirements. These key
elements include (1) technical
assistance, (2) information and
technology transfer, (3) economic
incentives, and (4) regulatory
incentives.
The major waste reduction
strategies of the California Waste
Reduction Program that industry can
apply to reduce the volume of un-
treated hazardous wastes going to
land disposal in order of prefer-
ence are: (1) Source reduction—the
elimination or reduction of the
generation of hazardous wastes; (2)
Recycling and reuse—reprocessing
of a waste material to a point that
it can be used again for the origi-
nal or different purposes; (3)
Treatment—including incineration
to eliminate or reduce the hazar-
dous characteristics of wastes; and
(4) Limit land disposal to treated
wastes residues only.
Source reduction is the best
solution as it eliminates the
-450-
-------
problems associated with hazardous
waste generation, storage, trans-
portation, treatment, and residual
disposal. This strategy reduces
significantly the volume of hazar-
dous wastes and can usually be
accomplished by improvements in the
manufacturing process, waste stream
segregation, and handling. Its
implementation beyond a certain
point may require major technolog-
ical changes and may become too
costly to implement. Also, treat-
ment and disposal of the reduced
volume of the waste will still be
required.
Recycling feasibility for
different wastes is affected by
costs of raw materials, cost of
disposal alternatives, and specific
process demands. The availability
of low-cost raw materials in com-
parison with the cost of recycled
wastes may limit this option.
Treatment has the potential to
detoxify most wastes although it
may shift the public health and
environmental risks from land to
air, ground and surface waters.
Treatment usually results in a
residue less toxic than the
original wastes which can then be
disposed in a non-hazardous waste '
landfill. However, there are some
instances where a smaller volume of
highly concentrated waste is pro-
duced which will require disposal
in a hazardous waste landfill.
A major part of the waste
reduction program has involved the
EPA in a Cooperative Agreement with
the Department of Health Services
(DHS). This three year multi-task
project, started in 1985, and under
the supervision of Mr. Harry
Freeman of EPA's Hazardous Waste
Engineering Research Laboratory,
has provided the Department with
staff and contractual support.
Projects related to the program
include: (1) Waste Reduction
(Minimization), (2) Technical
Information Dissemination, (3)
Evaluation of Potential Economic
Incentives, (4) Land Restrictions
Impact on Management of Restricted
Wastes, and (5) On-site Small
Treatment Demonstration Program.
_This paper will discuss the
specific studies made possible by
the EPA Cooperative Agreement as
well as state and federal statu-
tory requirements and problems
encountered in the process of
implementing the elements of
California's Waste Management
Program.
REGULATORY INCENTIVES AND STATUTORY
MANDATES
Regulatory Incentives
The regulatory element of the
Waste Reduction Program provides a
major driving force for moving
industry towards alternatives to
land disposal. In addition, it
provides some of the necessary
tools for the Department to effec-
tively implement the technical
assistance and information transfer
components of the program. Given
this, the Department has set the
following objectives: (l) to pro-
vide a clear message that hazardous
waste management in California must
move towards waste reduction and
away from land disposal; (2) to
ensure utilization of available
technology for hazardous waste
treatment; (3) to require that
industry be aware of and consider
waste reduction whenever possible,
and (4) to expand the California
land disposal restriction program
to coincide with the federal land
disposal restrictions program.
California Land Disposal
Restrictions Program
California's Land Disposal
Restriction Program was initiated
in December 1982, when the Depart-
ment implemented a regulatory pro-
gram to phase out land disposal of
certain hazardous wastes. This
program provided a schedule of land
disposal restrictions for specific
hazardous wastes, the California
List, to be implemented, contingent
upon the availability of alterna-
tive treatment and/or recycling
capacity in the State. The
restricted liquid wastes and
implementation schedule dates
-4C1-
-------
include: (1) free cyanides greater
than 1000 mg/L (6-1-83); (2) dis-
solved metals, compounds, elements
containing arsenic, cadmium, chrom-
ium, lead, mercury, nickel, sele-
nium, or thallium (1-1-84); (3)
acid wastes with a pH less than 2.0
(1-1-84); (4) PCB's greater than 50
mg/L (1-1-84); and (5) wastes con-
taining halogenated organic com-
pounds greater than 1,000 mg/L
(1-8-85).
An additional land disposal
restriction for solid hazardous
wastes containing halogenated
compounds in total concentration
greater than 1,000 mg/L is
scheduled to go into effect 7-8-87.
Another requirement of Senate Bill
509 (Carpenter, 1985), restricts
hazardous wastes having a heating
value of 3,000 BTU's per pound, to
incineration or treatment as the
only means of disposal, will go
into effect on 1-1-88. Other
regula- tions restricting land
disposal of hazardous waste
" containing volatile organics above
a range of one to eight percent (to
be specified) are to be developed
by the Department on or before
1-1-90.
Federal Land Disposal Restrictions
Section 3004 of the Resource
Conservation & Recovery Act (RCRA),
as amended by the Hazardous and
Solid Waste Amendments of 1984
(HSWA), prohibits the continued
placement of RCRA-regulated
hazardous wastes in or on the land,
including placement in landfills,
land treatment areas, waste_piles,
and surface impoundments (with
certain exceptions used for the
treatment of hazardous wastes).
The amendments specify dates by
which these prohibitions are to
take effect for specific hazardous
wastes with May 8, 1990, being the
deadline for the last third of all
RCRA listed wastes. The wastes
restricted and schedule dates are
as follows: (1) bulk liquids:
5_8_85; (2) solvent waste: 11-8-86;
(3) California List: 7-8-87; (4)
first third of all RCRA listed
wastes, decision on underground
injection, clean up wastes subject
to retrictions: 8-8-87; (5) dioxin
containing wastes (catagories F020
to F028): 11-8-88; (6) second third
of all RCRA listed wastes: 6-8-89;
and (7) last third of all RCRA
listed wastes: 1-1-90.
Implementation of these state
and federal mandates has been im- .
peded by a number of problems. The
first problem area is inadequate
waste composition and volume infor-
mation. Without this information
it is extremely difficult for the
state and industry to assess the
potential for waste reduction,
facility needs, and compliance with
land disposal restrictions. The
second problem area is the lack of
performance data from different
demonstration treatment technology
projects. Industries, regulatory
agencies, and the public are
seeking information related to the
evaluation of technologies before
extensive funds and time are
invested and before the technology
becomes a permanent fixture in an
area.
The third area of concern
limiting the implementation of
waste reduction programs is the
limited number of off-site treat-
ment facilities in the state. The
siting and regulatory requirements
and the difficulty in obtaining
environmental and liability insur-
ance have severely limited the
number of on-line treatment facil-
ities in California. Today there
is a lack of incineration treatment
capacity and only one off-site
operating hazardous waste treatment
facility in the state, General
Portland.
The Waste Reduction Program is
addressing these implementation
problems with the following pro-
jects: (1) Waste stream information
accumulation and analysis; (2)
Determination of extent and use of
on-site treatment; (3) Waste audit
program; (4) Waste management
information transfer; and (5)
Demonstration studies. These are
presented below.
-462-
-------
Waste Stream Information
Accumulation and Analysis
Hazardous waste generators and
operators of treatment, storage,
and disposal facilities in Cali-
fornia must now submit reports to
the Department that identify the
type and quantity of hazardous
wastes shipped off-site or wastes
they have handled. In addition,
the Department has also requested
more specific information pertain-
ing to the constituents and concen-
trations of materials in their
waste streams'. These reporting
requirements should improve the
Department's and industry's ability
to assess the potential for waste
reduction, facility needs, and
compliance with the land disposal
restrictions.
As part of this program, the
Department of Chemical Engineering
at the University of California,
Davis, is assisting the Department
to characterize California's waste
streams in terms of their suitabi-
lity to various treatment, recy-
cling, and reduction processes.
Over 11,000 individual waste stream
description reports, containing
information on constituents and
concentrations of different wastes,
will be analyzed. The principal
tasks are to: (l) prepare protocol
for analyzing the data according to
specific criteria for a particular
treatment, recycling, or reduction
process; (2) develop an easy access
data base format; (3) prepare an
analyses of this year's data, and
(4) prepare recommendations for
future waste stream description
forms. A final report will be
available in Fall, 1987.
Determination Of On-site Treatment
Use
The Department has reviewed
the 1985 Biennial Facility Reports
submitted on EPA's forms. Although
useful, some of the major problems
with the EPA biennial facility
reports were identified as follows:
(1) no correlation existed between
the amount of hazardous wastes
received and the ultimate disposal;
(2) discrepancies in report form's
disposal codes and disposal codes
in 40 CFR; (3) no code for recy-
cling of wastes with subsequent
reporting recycling as storage; (4)
misunderstanding in reporting due
to similarities in appearance of
facility reporting forms and gener-
ator's report forms; and (5) double
counting of hazardous wastes
created by generators reporting on-
site storage as disposal and then
having TSD facilities claiming
these wastes as well. Staff has
redesigned the form and the Depart-
ment will be using this in 1986
data gathering via the Annual
Facility Report due.March l, 1987.
After review and discussion of
the EPA reports, California repor-
ting requirements have been modi-
fied to be consistent with federal
law and also improved to obtain
more complete and accurate infor-
mation. Hazardous waste generators
must now submit a biennial report
to the Department that covers the
type and quantity of hazardous
waste shipped off site. Operators
of treatment, storage, and disposal
facilities must submit similar
reports on the wastes they have
handled. The Department has, in
the forms provided to the genera-
tors and operators, requested addi-
tional information on the consti-
tuents and concentrations of mater-
ials in their waste streams. These
reporting requirements will improve
the Department's ability to assess
potential for waste reduction,
facility needs, and compliance with
the land disposal restrictions.
Another reporting provision
which came into effect with the
passage of Assembly Bill 685 (Farr,
1985) requires that generators des-
cribe, in their biennial, reports,
waste reduction efforts and changes
in waste generation from the prev-
ious year. These reports will
assist the Department in developing '
information on effective waste
reduction projects and an assess-
ment of waste reduction potential
for specific industries.
-463-
-------
New reporting requirements for ;
generators concern changes to the
Uniform Hazardous Waste Manifest.
Generators shipping waste off site
must now sign a certification form
which states that a program to
reduce the volume and toxicity of
waste generated to the degree
determined to be economically
feasible is in place and the method
of treatment, storage, or disposal
currently available which minimizes
the present and future threat to
humans and the environment has been
selected.
After the above information is
collected and analyzed, the Depart-
ment will be able to predict how
much hazardous waste is produced in
California, how much is being dis-
posed on land, and what remaining
disposal capacity exists in the
state. Projections can then be
made as to the types and volumes of
wastes which will be impacted and
what types and scale of storage,
recycling, and treatment facilities
will be needed.
WASTE AUDIT PROGRAM
The Department has been admin-
istering a program to provide tech-
nical assistance for California
industries as part of its coordina-
tion, research, and development
efforts to promote reduction or
recycling of hazardous wastes.
This program includes the Waste
Audit Studies for hazardous waste
generators targeting various
industries. Each waste audit study
will focus on a specific industry
and include an on-site evaluation
of three to six firms within that
industry. For each firm, the manu-
facturing process and/or operating
systems will be audited, and the
current waste management practices
will be appraised to identify
options that could be utilized to
recycle, treat, or reduce the
generation of hazardous wastes.
The first set of waste audit
study contracts for approximately
$25,000 each were awarded in March,
1986, for the following: (1) pesti-
cide formulators, (2) paint indus-
try, (3) circuit board manufact-
urers, (4) automotive repairs,and
(5) automotive paint shops. Final
reports of the individual waste
audit studies will be available in
Summer, 1987.
A second set of waste audit
studies will target the following
industries: (1) photographic pro-
cessing laboratories, photographic
finishing, and motion picture film
processing; (2) commercial print-
ing, letterpress and screen,
engraving and etching; (3) educa-
tional, scientific, and research
institutions; (4) general medicine
and surgical hospitals excluding
infectious wastes; and (5) fiber-
glass and reinforced molded or
rigid plastic products including
watercraft, electrical panels,
figures, surfboards, and building
panels. Contracts will be executed
in Spring, 1987, with completion of
the final reports in Fall, 1987.
WASTE MANAGEMENT INFORMATION
TRANSFER
The objectives of this program
are (1) to make industry more aware
of economic, technical, and
environmental advantages of waste
reduction; (2) to establish credi-
bility with industry that the State
is available to assist them in the
area of land disposal capacity and
restrictions on land disposal_ of
hazardous waste; and (3) to dis-
seminate information and facts
resulting from other components of
the waste management program.
The Department has been
accomplishing these objectives
through seminars, fact sheets,
reports, and by operation of the
California Waste Exchange. The
-464-
-------
Waste Exchange utilizes three major
avenues to encourage recycling. A
Directory Of Industrial Recyclers
is published annually to inform
generators about companies presen-
tly available to recycle their
wastes. A Newsletter/Catalog is
published on a triannual basis.
More details about this activity
are provided under the technology
transfer component of the program
which publicizes waste reduction
opportunities or successful waste
reduction projects. Staff also work
directly with individual companies
to assist them in recycling their
waste.
Seminars provide a unique
opportunity for industry, regula-
tory agencies, and the public to
meet and discuss common goals,
problems, and solutions. Seminars
on specific waste management have
been a successful strategy to
expand awareness of the regulated
community on waste reduction.
The first seminar series,
fully funded and sponsored by the
Department, involved Solvent Waste
Management Alternatives. Two semi-
nars were held in October, 1986;
one in Los Angeles and the second
in San Francisco. A final report
entitled "Guide To Solvent Waste
Reduction Alternatives" was pre-
pared by ICF Consulting Associates,
Inc. and presented to all 500
attendees. The focus of the study
was on practical waste management
alternatives to land disposal that
have potential for reducing the
amount and/or toxicity of solvent
waste generated. Some of the major
tasks performed included: (1) iden-
tification of major solvent-user
industries in California; (2) a
review of current waste management
methods and technologies; and (3)
characterizations of source
reduction alternatives for minimi-
zing solvent wastes, on-site and
off-site solvent recycling treat-
ment alternatives. A conference
proceedings entitled "Solvent Waste
Reduction Alternatives Symposia"
was also prepared and is available.
A second symposium series
designed to educate California's
oil waste generators about alter-
natives to land disposal is
scheduled for March, 1988, in Los
Angeles and San Francisco. The
study on which the symposium will
be based, will be limited to oil
wastes resulting from oil usage and
can include lubricants, coolants,
cutting and machining oils, and
hydraulic fluids. A final report
summarizing the results of the
study and identifying the evalua-
tion methods of managing oil wastes
will be presented to each attendee.
A conference proceedings will like-
wise be available after the sympo-
sium.
DEMONSTRATION STUDIES
EPA Funded Projects
The following projects are
receiving direct funding through
the EPA/DHS Cooperative Agreement.
This element of the project will
evaluate commercially available
innovative alternative technologies
for the treatment and on-site
recycling of hazardous wastes.
Each technology will be monitored
and evaluated for treatment effici-
encies and economic feasibility.
Extent of funding ranges from
$30,000 to $50,000. The projects
are presented below.
California Agricultural
Research will demonstrate the
efficacy and economic feasibility
of commercially available Aerobic
Composting for the treatment of
pesticide rinsewaters. The demon-
stration project will take place at
the Chemical Waste Management
Treatment Facility in Kettleman
Hills, California, during the
Spring of 1987.
-465-
-------
The Department of Environ-
mental Toxicology of the University
of California, Davis will evaluate
the treatment efficiency of an
Oxidation/Ultraviolet Light System
in conjunction with a biological
treatment system for the degrada-
tion of pesticide rinseates. The
project will occur at the U.C.,
Davis campus during the Spring and
Summer of 1987.
Wesley M. Toy, P.E. will
evaluate the efficacy and economic
feasibility of an Evaporative Waste
Concentrator for automotive repair
wastes. The project will be per-
formed in conjunction with Safeway
Chemical in Saratoga, California in
the Spring of 1987.
Roy F. Weston, Inc. will
demontrate the treatment effi-
ciency of a Low Temperature Thermal
Treatment System for soils contam-
inated with volatile organic com-
pounds. This process involves the
stripping of volatile organic com-
pounds (VOC's) from soils utilizing
a recycling, low temperature oil
system (3OOF). The demonstration
project will take place at a site
to be determined in the
Spring/Summer of 1987.
Woodward-Clyde Consultants
will evaluate the treatment effic-
iency of Ambersorb XE-340 as an
alternative to granular activated
carbon for the adsorption of chlor-
inated solvents from industrial
waste stream. This project will be
conducted in Cupertino, California
in the Spring of 1987.
Ogden Environmental Services,
Inc., who recently purchased GA
Technologies circular combustion
division, will monitor and evaluate
a Circular Fluidized Bed Combustion
Process to determine treatment
efficiencies of cyanide and
fluoride removal from aluminum
spent potlining wastes. The pro-
ject will occur in San Diego, CA in
the Spring of 1987.
Hazardous Waste Reduction Grant
Program
The California Waste Reduction
Grant Program, established under
Assembly Bill 685 (Farr) to promote
innovative processes for managing
hazardous wastes, is entering its
second year. The Department seeks
cost effective, practical strate-
gies and technologies which would
reduce the volume, mobility or
toxicity of hazardous wastes. The
Department awarded a total of
$920,000 to 26 out of 97 applicants
for fiscal year 85-86.
The major components of the
projects funded were: (1) waste
types which included waste oils,
low heating value liquids and
solids, pesticide rinsewaters,
metals, chlorinated still bottoms,
plating wastes and foundry sands;
(2) technologies which included
biological, chemical, & thermal
treatment, recycling, freeze
crystallization, mobile treat-
ment, and oxidation; and (3)
industries which included oil and
gas refining, oil recycling,
semiconductor manufacturing,
pesticide application, aerospace,
cast metals manufacturing, plating
and electric utilities. Fifteen
feasibility studies, including six
for project designs, three for con-
struction and demonstration, and
two with public agency or univer-
sity applicants were among the
funded projects.
Based on the interest gener-
ated, the Department believes the
program will be highly successful
in demonstrating and evaluating
waste reduction activities. The
Department is now reviewing 106
grant proposals for the second year
funding of $1,000,000.
-466-
-------
CONCLUSIONS
Better knowledge in the areas
of waste composition and volume,
extent of treatment, and source
reduction methods should make the
California's hazardous waste
management program more effective
in the implementation of its
ultimate goal, reduction of the
volume of untreated waste being
land disposed. New facilities and
expansions of existing facilites
will be needed to fully implement
the ambitious goals of the federal
and state statutes. Major efforts
to overcome siting problems are
needed to allow for the use of
tested treatment alternatives. The
State's hazardous waste management
program is being continuously
reviewed to assure that all its
elements contribute as effec-
tively as possible to the meeting
of California's waste reduction
goals.
-467-
-------
FIELD ASSESSMENT OF STEAM STRIPPING VOLATILE ORGANICS
FROM AQUEOUS WASTE STREAMS
Marvin Branscome, Clark Allen, Scott Harkins, and Keith Leese
Research Triangle Institute
Research Triangle Park, North Carolina
and
Dr. Benjamin L. Blaney
U.S. Environmental Protection Agency
Cincinnati, Ohio
ABSTRACT
This paper discusses the removal of volatile organics (VO) from aqueous waste streams
by steam stripping and summarizes the effectiveness of VO removal from the.waste, the air
emissions from the process, and the cost of the treatment process. Tests were conducted
at two chemical plants that used continuous steam strippers to remove VO from the waste-
water. The operation at Plant H, which produces ethylene dichloride and vinyl chloride
monomer, treated about 852 liters per minute (L/min) or 225 gallons per minute (gal/min)
of aqueous waste containing about 6 grams per liter (g/L) of VO. The .operation at Plant
I, which produces one-carbon chlorinated solvents, was smaller and treated 42 L/min (11
gal/min) of aqueous waste containing about 6 g/L of VO.
The test program evaluated the removal of VO from the water, which was about 99.8 to
99.999 percent at the two plants. At Plant H, the concentration of VO in the stripper
bottoms ranged from 0.34 to 36 parts per million (ppm) with an average of 9.7 ppm. This
wide range was caused by variations in the concentration of chloroform (the major consti-
tuent in the bottoms), which was apparently related to column fouling. This stripper
processes wastewater containing about 1.4 g/L of filterable solids. At Plant I, the con-
centration of VO in the bottoms ranged from less than 0.005 to 0.13 ppm. Solids and an
organic layer are removed in decanters at Plant I prior to steam stripping to provide a
feed stream containing.about 0.01 g/L of filterable solids. Emissions of VO from the
decanter and storage tank vents at Plant I were estimated as 46 megagrams per year
(Mg/yr). Significant vent rates of VO were also measured from the condensers at both
sites. The condenser vent.rate at Plant H averaged about 20 Mg/yr compared to 11 Mg/yr at
Plant I. The condenser efficiency at Plant H ranged from an average of 6 percent for
vinyl chloride to 99.5 percent for ethylene dichloride. At Plant I, the condenser effi-
ciency ranged from 89 percent for chloromethane to 94 percent for chloroform.
INTRODUCTION
The EPA Office of Air Quality Planning
and Standards (OAQPS) is currently devel-
oping regulations under the Resource
Conservation and Recovery Act (RCRA) and
its 1984 amendments to control air emis-
sions from hazardous waste treatment,
storage, and disposal facilities (TSDFs).
In support of this regulatory development
effort, EPA's Hazardous Waste Engineering
Research Laboratory (HWERL) is conducting
-468-
-------
field assessments of hazardous waste
treatment processes that could be used to
remove volatile organics (VO) from wastes.
Volatile organics are defined as those
organic compounds detected and quantified
by EPA procedures, which include both
purgeable and extractable compounds.
Aqueous wastes represent a high percentage
of the total volume of hazardous wastes.
These aqueous wastes are often stored,
treated, or disposed of in open area .
sources, such as surface impoundments and
open wastewater treatment units, which are
sometimes aerated. Steam stripping of the
waste prior to placement in these open
units is one promising technique that may-
reduce VO emissions from these sources.
APPROACH
This project focused on the use of
continuous steam strippers. There are
potential cost and cost-effectiveness
advantages of continuous stripping,
particularly for high volume aqueous
wastes. Batch processes have been evalu-
ated previously. The goals of the testing
effort included measuring the effective-
ness of the process for VO removal from
the waste, measuring any air emissions,
and assessing the cost of treatment.
Samples of the feed stream to the steam
strippers were taken for analysis by gas
chromatography/mass spectroscopy (GC/MS)
to identify purgeable and extractable
organic compounds. No significant quanti-
ties of extractable organics were found,at
either plant; consequently, the sampling
and analysis for each test focused on the
purgeable organic compounds that were
identified.
The sampling at each site was conduct-
ed over a 2-day period with 5 samples
taken each day at about 2-hour intervals
from each sampling point. The primary
sampling points included the steam strip-
per's feed stream, bottoms, and condensate
as shown in Figures 1 and 2. These liquid
samples were collected in duplicate in 40
milliliter (mL) vials with no headspace.
At Plant H, these samples were analyzed by
EPA Method 624, which is a purge-and-trap
procedure with analysis by GC/MS. The
vapor flow rate and composition from the
steam stripper's condenser system were
also measured. At Plant H, the con-
denser/decanter was vented within a closed
system to an incinerator. The system was
maintained at a pressure of 1.3 atmos-
pheres. The vapor flow rate in this
closed system was measured by a tracer gas
dilution technique. Propane was metered
into the vapor inlet to the condenser at a
known rate and the concentration of pro-
pane was measured downstream to calculate
the vapor flow rate.
Samples of the wastewater from Plant I
were screened by GC/MS to identify the
organic compounds. Because only purgeable
chlorinated compounds were found, EPA
Method 601 was used for analysis of the
samples from Plant I. At Plant I, the
vapor flow rate from the condenser system
was measured directly with a wet gas meter
because the system was vented to the
atmosphere. Vapor samples were collected
in electropolished stainless steel canis-
ters that had been previously cleaned and
evacuated. The vapor flow rates and con-
centration measurements were used to
calculate condenser efficiency. Addition-
al liquid samples at each plant were taken
for analysis of solids content, metals,
pH, and VO in the headspace. Process data
were collected for the steam stripping
system and included flow rates and temper-
atures of the various streams entering and
leaving the stripper.
The treatment system at Plant I also
included removal of solids and any separ-
ate organic layer in decanters prior to
steam stripping. Samples were taken from
the streams entering and leaving the de-
canter for analysis of VO, solids, and
metals to assess the effectiveness of the
treatment and to determine the character-
istics of residuals. In addition, vapor
samples were taken from the headspace of
the decanter and from the feed tank for
the steam stripper to estimate air emis-
sions.
PROCESS DESCRIPTION
The steam stripper at Plant H is used
to treat wastewater from the production of
ethylene dichloride and vinyl chloride
monomer. At Plant I, the steam stripper
is used to treat wastewater generated from
the production of methylene chloride,
carbon tetrachloride, and chloroform. The
characteristics of the two waste streams
are summarized in Table 1. The primary
constituents at Plant H were ethylene
dichloride (5,630 ppm) and chloroform (271
-469-;
-------
ppm). A total of 12 other compounds,
mostly chlorinated organics, were also
detected at Plant H at average levels that
ranged from 0.3 ppm (for benzene) to 11
ppm (for 1,1-dichloroethane). At Plant I,
methylene chloride (4,490 ppm) and chloro-
form (1,270 ppm) were the major constitu-
ents. A total of four additional chlori-
nated compounds were detected at average
levels that ranged from 5.3 ppm (for
1,1,2-trichloroethane) to 55 ppm (for
carbon tetrachloride). Total VO at both
plants averaged about 6,000 ppm (6 g/L).
The stream stripper at Plant H used a
tray column to treat water at a rate of
about 852 L/m1n (225 gal/min). Solids at
this plant were not removed prior to steam
stripping and were processed through the
steam stripper at a level of 1.4 g/L for
filterable solids. Fouling of the heat
exchanger and column from the accumulation
of solids requires that this system be
backflushed or cleaned periodically. The
company prefers to process the solids
through the steam stripper rather than
install a system for solids removal.
Removal of solids prior to steam stripping
would generate a sludge that could be a
hazardous waste requiring treatment before
disposal. The vapors from the stripper
pass through a primary condenser cooled
with cooling tower water followed by a
secondary condenser cooled with refriger-
ated glycol. Noncondensibles vented from
the secondary condenser are routed to an
Incinerator. The condensate, which con-
tains both an aqueous and organic phase,
is recycled to the production process.
The steam stripper at Plant I was a
packed column used to treat about 42 L/min
(11 gal/min) of wastewater. The waste-
water is treated for removal of solids and
any separate organic phase in a decanter
prior to stripping. The treatment in-
cludes pH adjustment, addition of floccu-
lant, mixing, and settling for phase
separation of each batch in the decanter
over a 24-hour period. The aqueous phase
1s decanted and stored In the feed tank
prior to steam stripping. Vapors from the
steam stripper pass through a primary
condenser and a secondary condenser, both
cooled with cooling tower water. The
condensate 1s separated in a decanter and
the aqueous layer is returned to the
column. The heavier organic layer is
removed periodically and returned to the
production process. The two solids
decanters are also vented to the secondary
condenser. The feed storage tank is vent-
ed to the atmosphere through a conserva-
tion vent.
After steam stripping, the wastewater
from Plant H is sent to the wastewater
treatment process, which includes solids
removal and biological treatment. At
Plant I, no additional treatment usually
is needed (other than occasional pH ad-
justment) prior to discharge to the Y-iver.
VO REMOVAL FROM WATER
At Plant H, the removal of the major
component (ethylene dichloride) was gener-
ally on the order of 99.999 percent with a
feed concentration of 5,630 ppm reduced to
0.097 ppm in the stripper bottoms (see
Table 1). The removal of chloroform aver-
aged 99.6 percent for 6 of the 10 runs and
averaged only 92.4 percent for the other 4
runs. A feed concentration of 271 ppm
chloroform was reduced to an average in
the bottoms of 9.6 ppm with a range of
0.13 to 36 ppm. The variations in chloro-
form removal appeared to be related to
column fouling problems because the lowest
values in the stripper bottoms were found
after backflushing the steam stripper, and
the highest levels were found before back-
flushing when the column pressure drop was
increasing. None of the other 12 volatile
compounds found in the feed at 0.3 to 11
ppm was detected in the stripper bottoms
at a detection limit of 0.01 ppm. The
percent removal for these compounds gener-
ally exceeded 99 percent. Total VO at
this plant was reduced from an average
feed concentration of about 6,000 ppm to
an average of 9.8 ppm in the effluent, or
about 99.8 percent removal of total VO.
The major constituent at Plant I
(methylene chloride) was reduced from an
average feed concentration of 4,490 ppm to
0.011 ppm (99.999 percent removal).
Chloroform was reduced from 1,270 ppm to
0.006 ppm and carbon tetrachloride was
reduced from 55 ppm to <0.005 ppm. The
other chlorinated compounds present in the
feed (chloromethane, trichloroethylene,
and 1,1,2-trichloroethane) were not de-
tected in the bottoms at a detection limit
of 0.005 ppm. The total VO at this plant
was reduced from about 6,000 ppm to <0.037
ppm with about 99.999 percent removal.
-470-
-------
The removal efficiency of this steam
stripper over the 2-day test was more
consistent than that observed at Plant H.
During the first test day, the levels of
chloroform in the bottoms ranged from 7 to
9 ppb compared to all values <5 ppb on the
second test day. None of the volatile
compounds in the stripper bottoms exceeded
0.023 ppm for any of the samples.
CONDENSER EFFICIENCY AND AIR EMISSIONS .
The condenser efficiency (Table 2) was
evaluated at both plants from the quantity
of VO entering the overhead system and the
quantity leaving with the noncondensible
gases. The condenser system at Plant H
included two condensers in series that
used cooling tower water followed by
refrigerated glycol at 2 *C. A total of 7
of the 14 compounds detected in the feed
were also detected and quantified in the
vented vapors. The condenser removed 99.5
percent of the major constituent (ethylene
dichloride) and about 96 percent of the
chloroform. The removal efficiency for
compounds present in lower concentrations
was much lower. Vinyl chloride removal in
the condenser averaged only 6 percent and
indicated that this compound passed
through the condenser in the vapor phase
and was sent to the incinerator. The
average flow rate from the condenser was
3.1 L/s and the total VO from the vent was
about 20 Mg/yr (0.62 g/s). Although the
condenser removed 99.5 percent of the
ethylene dichloride from the vapor,
ethylene dichloride was the major compo-
nent of the vented vapors and comprised ..
about one-half of the total VO vented to
the incinerator.
The condenser system at Plant I also .
included two condensers in series; how-
ever, both were cooled with cooling tower
water. Because of no significant differ-
ences in vapor phase concentrations and
temperatures from samples taken after the
two condensers, the small secondary con-
denser apparently was not removing any
additional constituents. All of the sig-
nificant condensation was provided by the
primary condenser on the steam stripper.
Vapor phase concentrations of VO after the
primary condenser were high and averaged
44 percent VO by volume in the vapor phase
or a mass concentration of 1.6 g/L at 25
*C. The vent rate was measured as 0.2
L/s. The condenser removed about 90
percent of the major constituent
(methylene chloride) and about 94 percent
of the chloroform from the vapors. The
overall removal efficiency for total VO
was approximately 91 percent. Emissions
from the secondary condenser vent due to
steam stripping were measured as 11 Mg/yr
(0.34 g/s).
The vapor space in the solids decanter
contained primarily methylene chloride (28
to 32 volume percent), chloroform (6.4 to
7.6 percent), carbon tetrachloride (1.5 to
3.8 percent), and chloromethane (0.6 to
1.0 percent). The vapor space in the feed
tank also contained methylene chloride
(9.7 to 12 percent), chloroform (2.5 to,
3.1 percent), carbon tetrachloride (0.7 to
0.8 percent), and chloromethane (0.2 per-
cent). Emissions from these two tanks
were estimated based on the measured vapor
phase concentrations and the working
losses from the tanks based on-a water
transfer rate of 11 gal/min. Total VO
emissions from the solids decanter were
estimated as 35 Mg/yr or 1.7 grams per
liter (g/L) of water treated. Total VO
emissions from the feed tank were esti-r,
mated as .11 Mg/yr or 0.5 g/L treated. ' The
total VO emissions from the,three major
sources (condenser vent, solids decanter
vent, and feed tank vent) were estimated
as 57 .Mg/yr or about 2.7 g/L treated. The
estimate of annual emissions is based on
an average water treatment rate of 11
gal/min for 50 weeks during a year.
Samples taken from the-solids decanter
showed that the treatment process could
reduce filterable solids from 1,100 ppm to
50 ppm. Samples from the storage tank,
which represented water decanted before
our test, showed filterable solids levels
of 11 ppm. Reductions in chromium, cop-
per, nickel, lead, and zinc were, also
observed after treatment. Solids removal
prior to steam stripping decreases column
cleaning requirements and probably
improves the consistency of the operation.
However, the process generates about
190,000 L/yr of sludge, which contains the
chlorinated compounds at levels of 20 to
30 percent. The solids removal process
also increases the capital and operating
cost of the treatment system. However,
the bottoms from the steam stripper at
this plant do not require any additional
wastewater treatment (other than occasion-
al pH adjustment) before discharge.
-471-
-------
COST
Cost data were not available for the
small steam stripper at Plant I. However,
cost data were available for the steam
stripper at Plant H and for a similar
operation at Plant K, which treated a
wastewater stream similar to that at Plant
H; The steam stripper at Plant K was not
evaluated 1n a full-scale test; however,
process data, cost data, and samples were
obtained during a one-day plant visit.
The basic process equipment for the
steam stripping operation at Plant H
includes a feed storage and surge tank,
heat exchanger, the column and trays, two
condensers 1n series, a decanter, 8 pumps,
Instrumentation, piping, and insulation.
The total installed capital cost was esti-
mated as $950,000 (1986 dollars). The
major annual operating cost components
Include utilities (primarily steam),
operating and maintenance labor, and
laboratory support for analyses. A credit
is Included for the recovery of ethylene
dichloride that is recycled to the produc-
tion process. The annual operating cost
was estimated as $250,000/yr. The total
annualized cost, which Includes capital
recovery based on an interest rate of 10
percent and a lifetime of 10 years, was
estimated as $405,000/yr or $0.89/1,000 L
treated.
The data from the steam stripper at
Plant H are compared in Table 3 with data
obtained from a similar stripper at Plant
K during a one-day plant visit. One dif-
ference between the two types is that the
stripper at Plant H used trays for vapor/
liquid contact whereas the Plant K opera-
tion uses a packed column. The basic feed
constituents are similar; however, Plant K
has a higher concentration of 1,2-di-
chloroethane 1n the feed. The difference
in annual operating cost is probably
attributable to the higher rate of steam
usage at Plant K where approximately 75
percent of the annual operating cost is
for steam. Both of these steam strippers
achieve similar effluent (or bottoms)
concentrations of VO in the range of 1 to
2 ppm. Although the steam usage for Plant
K appears to be higher than that at Plant
H, the steam rates in terms of VO removed
are very similar (6.2 and 6.7 kg steam/kg
VO removed). The small difference in cost
effectiveness is probably not significant
and can likely be attributed to the higher
feed concentrations observed at Plant K.
Steam usage contributes significantly
to the annual operating cost of a steam
stripper. The steam usage at Plants I and
K averaged about 0.1 kg/kg water treated
compared to 0.036 kg/kg treated for Plant
H. (Each of these plants used heat ex-
changers to preheat the stripper feed with
the hot bottoms stream from the steam
stripper.) These values compare favorably
with some published design information on
steam usage. Typical values of 0.07 to
0.24 (1), 0.1 to 0.3 (3), and 0.31 (2) kg
steam/kg water treated have been reported.
CONCLUSIONS
Steam stripping can remove 99.8 to
over 99.999 percent of the purgeable
organic compounds found in the two waste
streams. The presence of solids in the
wastewater can lead to fouling problems
and variations in performance; however,
wastewater containing 1.4 g/L of filter-
able solids can be processed in a tray
column steam stripper. When the condenser
and tanks are Vented to the atmosphere,
emissions of 20 to 57 Mg/yr can result.
The cost effectiveness of stripping a
saturated wastewater stream in systems
designed for 680 to 820 L/min was on the
order of $120 to $220/Mg VO removed. The
water treatment cost ranged from $0.89 to
$1.57 per Mg or $3.38 to $5.96 per 1,000
gallons of water.
REFERENCES
1. Ehrenfeld, J., and J. Bass. Handbook
for Evaluating Remedial Action
Technology Plans.EPA 600/2-87-076.
August 1983.
2. Nathan, M. F. Choosing a Process for
Chloride Removal. Chemical
Engineering. January 1978. p. 93-
100.
3. Shukla, H. M., et al. Process Design
Manual for Stripping of Organics. EPA
600/2-84-139.August 19847
-472-
-------
TABLE 1. SUMMARY OF AVERAGE STRIPPER FEED (IN) AND BOTTOMS (OUT)
CONCENTRATIONS (ppm)a
Constituent
Ethyl ene di chloride
Chloroform
Benzene
Carbon tetrachloride
Chlorobenzene
1 , 1-Di chl oroethane
1 , 1-Di chl oroethene
1 , 2-Di chl oroethene
Methyl ene chloride
Tetrachl oroethene
1,1, 2-Tri chl oroethane
Tri chl oroethene
Vinyl chloride
Total
Plant H
, In •
5,630
271
0.27
1.7
0.38
11
4.7
8.9
1.2
1.4
7.5
4.8
8.4
5,950
Out
0.097
9.6
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<9.8
Filterable solids (g/L)
1.4
0.93
Plant I
Constituent
Methyl ene chloride
Chloroform
Carbon tetrachloride
Chloromethane
Trichloroethylene
1,1, 2-Tri chl oroethane
Total
Filterable solids (g/L)
•-In
4,490
1,270
55
33
5.6
5.3
5,860
0.011
Out
0.011
0.006
<0.005
<0.005
<0.005
<0.005
<0.037
0.009
aAverages of 10 samples taken over a 2-day test.
-473-
-------
TABLE 2. SUMMARY OF AVERAGE CONDENSER VENT RATES AND EFFICIENCIES
Plant H
Constituent
Vinyl chloride
Chloroethane
1 , 1-Di chl oroethene
1 , 1-Di chl oroethane
1 , 2-Di chl oroethene
Chloroform
Ethyl ene di chloride
Condenser
loading
(g/s)
0.089
0.081
0.036
0.11
0.006
2.9
63
Vent rate
(g/s)
0.084
0.043
0.031
0.013
0.001
0.11
0.34
Condenser
efficiency
(percent)
6
47
15
88
84
96
99.5
Total
66
0.62
99
Plant I
Constituent
Condenser
loading
(g/s)
Vent rate
(g/s)
Condenser
efficiency
(percent)
Chloromethane 0.021
Methylene chloride 2.9
Chloroform 0.81
Carbon tetrachloride 0.038
Total 3.8
0.0024
0.29
0.045
0.0039
0.34
89
90
94
90
91
-474-
-------
TABLE 3. COST COMPARISON
Item
Capital cost ($)
Operating cost ($/yr)
Total annual! zed cost ($/yr)a
Average feed rate (L/min)
Steam rate (kg/h)
Feed constituents (ppm)
1 , 2-Di chl oroethane
Chloroform
Other VO
Bottoms constituents (ppm)
1, 2-Di chl oroethane
Chloroform
Other VO
Steam usage
kg/kg water
kg/ kg VO removed
Cost-effectiveness
$/Mg VO removed
$/l,000 L treatedf
Plant H
950,000
250,000
405,000
820
1,790
5,600
270
59
0.16
0.8C
<0.01
0.036
6.2
220d
0.89
Plant K
700 000
450,000
564,000
680
4,090
15,000b
175
31b
.037b
1.3b
1.4b
0.10
6.7
120e
1.57
aBased on a 10-year lifetime at 10% (Capital recovery factor = 0.163).
bBased on a single sample analysis from presurvey trip.
cBased on 6 of 10 runs.
^Based. on 1,820 Mg/yr recovered.
eEstimated from single analysis and 329 days/year operation.
equals $/Mg treated.
-475-
-------
a.
a>
_c
"a.
CO
w
II
OO
JO
a.
g
o
"a.
CO
v>
•n
c
to
i_
a>
a.
.£•
CD
to
O
O
+5
CO
a>
.e
u
CO
I
at
-476-
-------
I
oo
ta
.1
C SJ
8*
I
o
CQ
a
i
8'
o"
55
K
O
O £
_co
CL
a a>
C Q.
'5 E
aa
°> l_
c o
It
<0 O
O O)
J= C
"1.
•S i
cs
1
O)
I
1 I
is
S3
« g
S1
o
N —
ta
T
8
-477- •
-------
FIELD ASSESSMENT OF THE FATE OF VOLATILE ORGANICS IN
AERATED WASTE TREATMENT SYSTEMS
David Green
Research Triangle Institute
Research Triangle Park, North Carolina 27709
and
Bart Eklund
Radian Corporation
Austin, Texas 78766
ABSTRACT
Aeration of wastewater containing volatile organic compounds in activated sludge sys-
tems effectively removes many of these compounds from the wastewater prior to discharge.
Studies were conducted at a full-scale treatment systems to determine the relative extent
to which various compounds were destroyed biologically and stripped into the air. Direct
measurements of air emissions were made through sampling and chemical analysis of off-
gases from the aeration tank of an activated sludge unit. Indirect measurements were made
by comparing compound specific biological oxidation rates obtained in closed bottles to
total disappearances across the treatment units. Additional measurements were made to
determine potential removal of organics in waste sludge streams. This paper describes
these measurement techniques and results of the studies.
INTRODUCTION
Industrial wastewater is often treated
1n aerated biological oxidation units to
remove dissolved organic chemicals prior
to discharge or reuse. For relatively
dilute streams containing biodegradable
compounds, treatment processes such as
activated sludge are typically effective
and economical. Organic material is re-
moved from wastewater in these processes
by aerobic organisms which use the materi-
al as a food source. Oxygen is supplied
to the organisms either by bubbling air
through the wastewater, agitating the
surface of the wastewater or some combina-
tion of these mechanisms. In addition to
supplying oxygen to the organisms respon-
sible for biodegradation, the aeration 1s
necessary to keep the biomass suspended
and to provide adequate mixing.
When volatile and semi volatile com-
pounds are present in the wastewater which
is aerated, the possibility exists that
some or all of the removal of these com-
pounds results from mass transfer to the
air rather than biological decomposition.
In order to determine the relative
importance of these mechanisms, the total
removal in the unit must be determined and
a direct or indirect measurement of air
emissions must be made. An indirect
measurement of air emissions can be made
by determining the rate of biooxidation in
a closed system where mass transfer to the
air is not significant, and comparing this
rate to the total.disappearance measured
across the actual treatment unit. Direct
air emissions measurements can be made by
sampling and analyzing air over the
aeration tank in a way that permits an
accurate calculation of the flux from the
-478-
-------
aeration tank. The accuracy of either
direct or indirect methods depends on the
capability of the technique to monitor the
physical or biochemical process without
significantly affecting the rate. The
accuracy of chemical analyses of gas or
liquid samples is also crucial to the
accuracy of the overall measurement.
This study was done as part of a
project to support efforts by the EPA
Office of Air Quality, Planning, and
Standards to estimate air emissions from
specific treatment processes. An
understanding of the potential magnitude
of emissions of this type is required for
development of national emissions models
in support of regulations under the
Resources Conservation and Recovery Act.
The major objectives of this field test
were to obtain direct measurements of air
emissions from a submerged aeration
activated sludge tank using an isolation
mass flux chamber at various locations on
the tank surface, and to also measure
compound specific biodegradation rates and
total removals in the activated sludge
tank so that a calculation of air
emissions could be made by difference. In
addition, process operating conditions
which affect both biochemical and physical
removal were monitored.
At the outset of the project, it was
hoped that a comparison between the direct
and indirect measurement techniques could
be made on a compound-specific basis. A
confirmation of the indirect measurements
by the direct measurement technique would
support use of the indirect technique at
surface aerated facilities where direct
air measurements are more difficult.
Analytical limitations precluded this
comparison at this site except for the
compound, 1,1,1-trichloroethane. The
phenolic compounds determined by EPA
priority pollutant methods in liquid
samples were not quantifiable with the gas
phase analytical system used in this
study.
Minor objectives of the field test
included determination of overall removals
of specific organic compounds from the
wastewater treatment system as a whole and
determination of experimentally measured
air/water and water/biomass solids
partition coefficients for use in generic
evaluations of similar systems. The test
was conducted by Research Triangle
Institute (RTI) and Radian Corporation.
RTI-had primary responsibility for liquid
sampling and analysis, total removal
calculations, air/water and water/biomass
solids partition coefficients, and
compound specific biodegradation rate
determinations. Radian had primary
responsibility for air sampling and
analysis and direct air emissions
calculations. Sampling was conducted
between September 22 and September 26,
1986.
TEST DESCRIPTION
At the facility tested, wastewater is
collected at various points in the
manufacturing area of the plant and pumped
intermittently to a sump in the wastewater
treatment area. Wastewater is pumped
intermittently from this sump to an
equalization tank with a residence time of
approximately 90 hours. The equalization
tank is not completely mixed and is
operated primarily to accommodate •
hydraulic surges. '
Wastewater is then pumped to a split-
ter box where it is mixed with recycled
sludge and divided between two identical
parallel above-ground concrete aeration
tanks providing approximately six days
residence time. Air is supplied through
static mixers in each tank. Approximately
two inches of foam was present on the
surface of the tanks except in the areas
directly above the mixers. The aeration
tanks contained 2,500 mg/L of mixed liquor
suspended solids during the test. 'The
water level is maintained by an overflow
weir.
The wastewater from the two tanks
overflows to a splitter box where it is
recombined and then divided evenly between
two clarifiers. Sludge is returned to the
aeration tanks at the influent splitter
box in an amount sufficient to maintain
the desired volatile suspended solids
content of the mixed liquor. '
AIR SAMPLING
The tank was divided into twenty-seven
2.44 m x 2.44 m grids. An enclosure > -
device, the isolation emission flux •
chamber (see Figure 1),-was used to ••'?••
measure the off-gas flow rate from the
-479-
-------
different parts of a grid. A slipstream
of the sample gas was collected for
hydrocarbon analysis.
The first day and a half of sampling
was used to assess the spatial variability
of a grid in terms of VOC emissions and
air flow rate. To accomplish this, flow
rate measurements and THC concentration
measurements were made at various grid
points and locations within the grids.
Only the on-site total hydrocarbon
analysis was performed for samples
collected on the first days of sampling.
Based upon this preliminary work, the tank
was divided into three zones, directly
above the aerators (A), foam covered (C),
and non-aerated (NA). These zones are
shown schematically in Figure 2.
The order of sampling was randomly
determined, with 18 grid points (six per
zone) being sampled over the following
two-day period. At each location, a gas
sample was collected for on-site methane,
and total non-methane hydrocarbons (TNMHC)
and screening-level compound analysis, and
a second gas sample was collected for
detailed speciation at Radian's Austin
laboratories. A total of 20 stainless
steel canister samples (including
duplicates) and a canister blank were
collected over the two-day sampling
period. Two liquid samples were collected
1n VOA vials at each of the eighteen
sampling points. The total air flow rate
to the tank was determined by using the
fan performance curve (amperes vs. air
flow rate) and amperage data continuously
recorded on a stripchart recorder during
the site investigation.
The air sampling approach chosen used
an enclosure device, referred to as an
emission Isolation flux chamber, to sample
gaseous emissions from a 0.13 m2 surface
area. A pump was used to withdraw sample
gas from the flux chamber at the same rate
at which it entered. This was confirmed
by monitoring the pressure in the chamber.
The volumetric flow rate of air through
the, chamber was recorded and the
concentration of the species of interest
was measured at the exit of the chamber.
The emission rate was calculated as:
E.R.-i
where: E.R.j = emission rate of species,
i (/
-------
electrode. Volatile suspended solids
analyses of aeration tank effluent, re-
cycle sludge, and influent splitter box
samples corresponding to biodegradation
rate determination samples were conducted
by Standard Method 209E. Volatile sus-
pended solids data were supplemented with
data supplied by plant personnel deter-
mined using the same method. Chemical
oxygen demand data were obtained from
plant records.
BIODEGRADATION RATE TESTING
In order to distinguish.between re-
moval of organics from mixed liquor due to
biodegradation and removal due to mass
transfer into the air, experiments were
conducted which permitted biodegradation
to take place while limiting air strip-
ping. Samples of a mixture of aeration
tank feed and recycled sludge were dipped
from the influent splitter box at the
upstream end of the aeration tank. :
Each sample was divided using a two
liter Nalgene graduated cylinder as fol-
lows: up to seven, one liter bottles were
partially filled with 500 ml of mixture,
one one-liter bottle was completely filled
with mixture, and one specially prepared
500 ml bottle was partially filled with
250 mL of mixture. The filled bottle was
designated for volatile suspended solids
analysis and immediately stored on ice.
One of the partially filled one liter
bottles was immediately preserved with 10
ml of saturated copper sulfate solution
and agitated gently to assure that the
copper sulfate solution was distributed.
This bottle was then used to fill two 40
mL septum vials. The one liter bottle and
the two 40 mL bottles were stored on ice
immediately thereafter for shipment to a
laboratory for organic compound analysis.
The specially prepared 500 mL bottle
had a plastic tubing stub fitted into and
protruding through the cap. Tygon tubing
was connected to the stub leading to a
plastic T-connector. One side of the T-
connector was attached to a short length
of tubing filled with lithium hydroxide.
The other side of the T-connector was con-
nected to a mercury manometer. This
bottle was used to monitor oxygen uptake
over time.
The partially filled one liter bottles
and the partially filled 500 mL bottle
were then mounted on a wrist action shaker
and continuously agitated. Over a period
of about 19 hours, bottles were removed
from the shaker, one by one and preserved
with copper sulfate using the same
procedure as for the initial sample.
Similarly, 40 mL vials were filled for
purgeable organics analysis.
Biodegradation rate test samples were
analyzed for purgeable organics, acid
extractable priority pollutants, and meth-
anol. A total of four tests were conduct-
ed, with two tests run simultaneously from
the same sample of aeration tank feed/
recycle sludge mixture.
The total oxygen uptake within the
specially prepared bottle was monitored to
insure that the oxygen within the
headspace was not exhausted. When the
oxygen in the headspace was approximately
50% of that initially present, the bottles
remaining on the shaker were vented and
fresh air was allowed to replace that
initially present. This was done once in
the course of the experiment.
RESULTS
Direct Air Emissions Measurements
Compound specific emissions rates are
given in Table 1. Methane emissions were
larger than those of any class of
compound. The TNMHC emitted from the
lagoon was much less than 1 kg/day.
Emissions were highest for Zone C (foam
covered areas) in all cases except in the
case of the insignificant amount of sulfur
species and unidentified VOs emitted.
Zone A (directly above the aerators)
emissions are generally less than Zone C
reflecting the smaller total area of
Zone A. It is probable that sufficient
air is present in Zone A to strip out all
VOs and that any excess air merely acts to
dilute the off-gas.
These values have been corrected for
the blank sample values and for a sampling
bias at low flow rate. The total air flow
rate based on the fan curve was 15 m3/min.
The spatial variabilities observed
were not expected and indicate that the
subsurface flow patterns or the emission
-431-
-------
processes are not fully understood. The
large methane emissions from the aerated
zone Imply either a non-biological origin
or production outside the aerated zone.
Indirect Measurements
Typical biodegradation test data are
shown 1n Figure 3. The slope of the
linear regression line through the data
points represents the best estimate of the
compound specific biodegradation rate.
Concentrations would be expected to
decline monotonically in the absence of
chemical analysis errors. This slope was
then normalized for the biomass concentra-
tion. Selected biodegradation rate con-
stants are given in Table 3. Multiple
rates for the same compound reflect data
obtained during different tests. Taking
the rate constant for phenol, as an ex-
ample, as 0.25 /jg/min g biomass, would
Imply that a tank with a six day residence
time, operated with mixed liquor volatile
suspended solids of 2500 mg/L could effec-
tively blodegrade 5400 /jg/L of phenol.
The actual difference between phenol in
the influent and the effluent of the aera-
tion tank during the study period averaged
6200 /ig/L (based on a weighted average of
aeration tank feed concentration and
recycled sludge vs. aeration tank efflu-
ent, the effluent and recycle streams
were below the detection limit of 250
/jg/L). Based on this rate, 86% of the
phenol removed in the system was biode-
graded and 14% was emitted to the air.
The uncertainty in the rate calculation is
great enough that all of the phenol
removed may be due to biodegradation, as
the coefficient of variation of the bio-
degradation rate varied between 90% and
200%. The rate for 2,4,6-trichlorophenol,
0.037 * 47% /jg/min-gram biomass, indicates
that approximately 800 /jg/L of this com-
pound could be biodegraded in this system.
The average aeration tank influent concen-
tration (feed plus recycled sludge) was
approximately 1900 /jg/L; aeration tank
effluent concentration was consistently
below the detection limit of 250 /tg/L.
Potential air emissions of 2,4,6-trichlo-
rophenol may be as great as 58% of the
Influent.
The biodegradation rate for 1,1,1-
trichloroethane was essentially zero.
This compound was detected in the direct
air samples taken with the flux chamber at
1,330 /tg/min, but this was much less than
the mass in the aeration tank influent
(10,000 /*g/L). The difference is either
due to failure to detect the bio-
degradation of this compound (perhaps by
not continuing the test long enough) or
else to incomplete capture in the flux
chamber.
A similar calculation for 1,1,1-tri-
chloroethane indicates that most of the
difference in concentration between influ-
ent and effluent could be due to volatili-
zation.
Water/hiomass solids partition coeffi-
cients are given in Table 3.
CONCLUSIONS
The results of this field test did not
confirm the suitability of the indirect
air emissions estimation approach. The
major reason for this was that chemical
limitations prevented, for the most part,
analysis of the same compounds in both the
liquid and gaseous streams. Direct air
emissions measurements of a number of
compounds were obtained and can be cor-
related with influent concentrations.
Similarly, biodegradation rate estimates
and water/biomass solids partition coeffi-
cients were experimentally obtained for
use in the absence of more accurate site-
specific data.
-482-
-------
TABLE 1. TOTAL EMISSIONS FOR SELECTED INDIVIDUAL COMPOUNDS
Total emissions (^g/min)
Compound
Methane
C-2 VOC
Cyclopentane
Isobutene + 1-Butene
t-4-Methyl-2-Pentene •
Toluene
Methyl ene chloride
1,1, 1-Tri chl oroethane
Acetaldehyde
Acetone
Dimethyl sul fide
Total Air Flux (mVmin)
Aerated
Zone A
50,000 * 22,600
287 4 236
696 4 555
62.3 4 89.6
105 * 93.2
114 * 1,120
161 * 149
385 ±166
5,480 * 1,930
7.71 * 298
127 * 70.6
18.6
Foam covered
Zone C
263,000 4 50,700
1,800 * 571
1,040 ± 1,520
187 * 220
80.0 * 85.3
6,100 * 10,300
153 * 269
934 * 795
5,350 4 2,320
159 4 874
112 * 78.7
25.4
Non-aerated
Zone NA
2,080 4 11,600
-55.6 * 81.2*
26.8 ± 77.3
-15.5 4 28.3*
•17.7 4 62.4
-630 4 1,780*
-73.3 4 29.9*
10.2 4 44.5
-213 * 262*
-278 4 51.9*
0.00 4 0.0
3.85
Total
Tank
316,000
2,040
1,770
234
202
5,570
241
1,330
10,600
-111*
239 4 124
47.9
*Negative emissions result from correction for blank.
-483-
-------
TABLE 2. BIODEGRADATION RATE CONSTANTS
Compound
Rate Constant
/jg/(min-g biomass) ± 1 S.D.
Hethanol
Phenol
2,4,6-Tri chlorophenol
Styrene
Oxi rane
1,1,1-Trlchloroethane
12.8 ± 38 percent
5.7 ± 54 percent
0.087 ± 150 percent
0.25 * 91 percent
0.29 ± 200 percent
0.037 * 47 percent
0.0011 * 123 percent
0.38 ± 56 percent
0.59 ± 47 percent
TABLE 3. SLUDGE/WATER PARTITION COEFFICIENTS
Recycle
Compound [(/ig/kg)/(/*g/L)]
Methyl ene chloride
Oxi rane
2-Propanone
2-Propanol
1 , 3-Di oxo 1 ane-2-methanol
Dimethyl disulfide
140
170
160
170
170
150
Effluent
[(/KJ/kg)/(/KJ/L)]
NM
69
57
130
140
NM
Not meaningful (samples below detection level).
-484-
-------
.-485-
-------
Nonaerated Zone
7
1" = 4'
Zone C (Foam-covered Area)
Zone A (Above Aeration Jet)
Figure 2. Aeration tank zone designations.
10
8.
3 '
,§ fi-
ll '
o -
2-
S
800
Last bottle vented
at this point
200
400
600
1000
Time (minutes)
Figure 3. Concentration of phenol in biodegradation rate test samples.
-486-
-------
PILOT-SCALE EVALUATION OF A THIN-FILM EVAPORATOR FOR VOLATILE ORGANIC REMOVAL
FROM LAND TREATMENT SLUDGES
Coleen M. Northeim, C. Clark Allen, and Scott M. Harkins
• Research Triangle Institute
P. 0. Box 12194 ' . ,• .
Research Triangle Park, North Carolina 27709
ABSTRACT ,• • •,••_- •- , •
_ The U.S. Environmental Protection Agency Office of Air Quality Planning and Standards
is currently developing regulations to control air emissions from waste treatment, stor-
age, and disposal facilities. In support of this regulatory development effort, the
Research Triangle Institute has conducted a study of thin-film evaporators (TFE) for
removing volatile organics,(V,0) from refinery wastes. Jhin-film evaporators were studied
to evaluate their use to remove and recover VO from waste petroleum sludges prior to land
treatment. This would reduce the amount of VO available for release to the atmosphere
during land treatment of the sludges. '
The treatment of two refinery sludges was investigated in a pilot-scale agitated TFE.
The fraction of feed removed by the TFE ranged from 11 to 95.7 percent. At the greatest
overhead fraction, more than 99.9 percent of the VO and 75 percent of the semi volatile
compounds were removed from the sludge. At the lowest overhead fraction, greater than
98.5 percent of the VO and 10 to 43 percent of the semivolatiles were removed from the
sludge. The sludge processed with the lowest overhead fraction contained water and
maintained suitable handling characteristics for land treatment.
INTRODUCTION
The TFE test was conducted at Luwa
Corporation in Charlotte, NC, during the
week of September 8-12, 1986. Due to
regulatory constraints, the wastes that
were tested were npnhazardous (as defined
by RCRA) refinery wastes.. These wastes
were selected based on'their similarity to
hazardous refinery, wastes, such as API
separator sludge (hazardous waste code
K051), that are currently land-treated:
The use of TFEs was Investigated to
determine if VO can be removed and
recovered from waste petroleum sludges
prior to land treatment of Jthe/sludge.
This would reduce the amount of VO avail-
able for release to the atmosphere during
land treatment, of the sludges. The VO
would be recovered as an organic conden-
sate and recycled to petroleum refineries
as product. In addition to VO removal,
the process can also be operated to remove
water and low boiling oils from sludges,
reducing s.ludge volume while recovering
oil from sludges prior to disposal. These
benefits are not limited to sludges dis-
posed of by land treatment.
-487-
-------
THIN-FILM EVAPORATION
Agitated TFEs are designed to spread a
thin layer of viscous liquids or sludges
on one side of a metallic surface with
heat supplied to the other side. This
promotes the transfer of heat to the
material while simultaneously exposing a
large surface for evaporation of volatile
compounds. Heat can be supplied by either
steam or heated oil; heated oils are used
to heat the waste to temperatures higher
than can be achieved with saturated steam.
Volatile constituents separate from the
feed liquid or sludge, producing a vapor
stream of volatiles and leaving a treated
waste which flows out of the bottom of the
evaporator. Vapors can be condensed and
recovered.
The unique feature of this equipment
is not the thin film itself (falling- and
rising-film evaporators use thin liquid
layers), but rather the mechanical agi-
tator device for producing and agitating
the film. This mechanical agitator per-
mits the processing of high-viscosity
liquids and sludges with suspended solids.
The agitation at the heat transfer surface
not only promotes heat transfer but also
maintains precipitated or crystallized
solids in manageable suspension without
fouling the heat transfer surface.
With typical tip speeds of 900 to
1,200 cm/s (30 to 40 ft/s), centrifugal
forces distribute the feed as a thin film
on the heated cylinder wall, and the wave
action produced by the rotating blades
provides rapid mixing and frequent surface
regeneration of the thin liquid layer on
the transfer surface. A typical vertical
thin-film evaporator is illustrated in
Figure 1.
PILOT FACILITY AND TEST
The tests were conducted at the pilot
facility of Luwa Corporation, Charlotte,
NC. This facility contains a variety of
evaporators produced by the company and is
used to test potential applications of
their equipment for clients. The equip-
ment used for the tests was judged most
suitable for our applications. Samples of
tank bottoms sludges were obtained from an
oil refinery, tested in the equipment, and
then returned to the refinery for normal
disposal.
Drive
System
Rotor
Heating Medium (=
Modular Heating t_
Bodies
Product Outlet
Figure 1. Luwa vertical thin-film evaporator.
Figure 2 shows the equipment Used in
the pilot-scale tests. The 100-gal feed
tank was agitated with both an axial ntixer
and continuous recirculation of the feed
liquid through a centrifugal pump. A
positive displacement pump was used to
pump the feed sludge through the preheater
and into the top of the TFE. There, the
sludge was continuously spread over the
heated surface of the TFE. Unevaporated
material flowed down through the TFE to a
collection pot at the base. Materials
evaporating in the TFE passed through an
empty demister and were condensed in a
condenser that was cooled with cooling
tower water. Condensate flowed from the
condenser and was collected directly into
liquid sample jars or a flask used for
measuring condensate volumes. Any uncon-
densed vapors flowed from the condenser
through a wet testmeter for flow measure-
ment.
-488-
-------
s 'fiitt
«> S o o §
« U. 00 O>
U. CO
OJ
3
.5*
iff
I
£
3
O)
-489-
-------
The TFE was heated by hot oil,
although steam could be used for lower
temperatures. Entrainment separators are
used frequently with TFEs to remove en-
trained liquids from the vapors flowing to
the condenser. The entrainment separator
was empty (but heated) during the testing,
and very little material condensed there.
Five feed lines from the preheater to the
TFE were heated with low-pressure steam,
as were the vapor lines from the TFE
through the entrainment separator and to
the condenser.
A total of 22 runs were performed with
the TFE, using two different wastes, three
temperatures, three flow rates, and under
both atmospheric and vacuum operation.
The feed rates and temperatures were
chosen to operate the TFE over its normal
range of operation and to demonstrate the
removal of VO compounds from the feed
sludges. Tests 1 through 18 used five
55-gal drums of an emulsion tank sludge
while tests 19 through 22 were performed
on oily tank bottoms. The process test-
ing, sampling, and analysis concentrated
on the tests using the first waste while
the second waste was used to gather addi-
tional process data and demonstrate proc-
ess operation on a second waste sludge.
Matrices indicating operating condi-
tions and run numbers for the tests are
shown in Table 1. This test plan studied
the two major variables affecting TFE
performance, the temperature of the heat-
Ing jacket and the feed rate. The indi-
cated flows and temperatures were the
nominal process parameters during the
tests and the actual measured parameters
varied somewhat from these values. Runs
5, 6, and 7 were a series of tests at a
constant heating jacket temperature
(150 *C) at three different feed rates.
Runs 8, 9, and 10 were conducted at simi-
lar flow rates to Runs 5, 6, and 7 but
were at a much higher heating jacket tem-
perature (310 •C). Runs 14, 15, and 16
were conducted at an intermediate tempera-
ture (230 *C) and were limited to the two
lower flow rates. Runs 1-4, 12, and 13
were used as shakedown runs. These runs
allowed practice samples to be taken and
potential problems to be solved before the
samples to be analyzed were taken.
Three vacuum runs were performed wi th
the first waste sludge during Runs 11, 17,
, 18. The purpose of these runs was to
examine the effect of vacuum operation on
the removal of VO compounds from the feed
waste. The final four runs (19, 20, 21,
and 22) were performed with the second
waste. These tests were principally to
demonstrate the operation of the TFE with
a second waste sample and were not exten-
sively sampled and analyzed during the
project.
Four of the tests (5, 7, 8, and 10)
were selected for extensive sampling and
analysis of process streams. These four
runs allowed the process to.be examined
with both high and low feed rates and at
both high and low heating temperatures.
They represent the range of reasonable
operating conditions for the TFE process-
ing waste sludges for the removal of vola-
tiles, water, and oils from the sludge.
SAMPLING AND ANALYSIS
Samples were taken to characterize the
wastes treated during the pilot studies
and to determine the efficiency of the TFE
process. Four process streams were ,
sampled: feed, bottoms, condensate, and
condenser vent gas. The procedures for
obtaining these samples are outlined in a
specific test and quality assurance plan.
In all cases, special precautions were
taken to obtain representative samples and
to prevent the loss of VO from the samples
prior to analyses.
Sample analysis was performed onsite
by RTI and offsite by contract labora-
tories. The onsite measurements performed
by RTI were: (1) the analysis of head-
space concentrations of VO from feed
sludge samples and bottoms samples and
(2) the measurement of vent gas flow rates
and overall VO concentrations in the vent
gas and bottoms collection pot. Two types
of analyses for headspace concentrations
of VO were employed. The first used
syringes to transfer gas samples from
half-filled sample bottles and a portable
GC to measure the concentrations of VO in
air above the samples. The second method
of measuring the headspace concentrations
of VO used a calibrated total hydrocarbon
analyzer. This instrument was a Bacharach
TLV Sniffer that pulls a continuous sample
that is continuously oxidized by a
catalyst-coated resistance element. The
resistance of this element varies with
-490-
-------
TABLE 1. THIN-FILM EVAPORATOR TEST PLAN
Feed No. 1, Emulsion Tank Bottoms
Flow Rate (Ib/hr)
Temperature (*C)
150 230 310
70
100
150
Run 5
Run 6
Run 7
Run 14
Run 15
Run 16
Run 8
Run 9
Run 10
Feed No. 2, 01ly Tank Bottoms
Flow Rate (Ib/hr)
150
Temperature (*C)
310
45
80
Run 21
Run 22
Run 19
Run 20
-491-
-------
temperature, which is in turn proportional
to the hydrocarbon concentration of the
analyzed gas. These measurements were
Intended to be a rough analysis of head-
space concentrations and to confirm the
results from GC analysis. The Bacharach
TLV Sniffer had a maximum measurable con-
centration of 10,000 ppm. This limit was
exceeded in the headspace of all of the
feed samples.
The Bacharach TLV was also used to
measure concentrations of organics in the
vent gas, feed tank headspace, and vapors
above the bottoms when the sample pot was
removed from the TFE. Measured organic
concentration of the vent gas and feed
tank headspace also exceeded the range of
the instrument.
Samples of the TFE feed, bottoms, and
condensate from selected tests were ana-
lyzed by a contract laboratory for vola-
tile and semivolatile organics; percent
oil, solids, and water; and metals using
the EPA Contract Laboratory Program (CLP).
The purpose of these analyses was to
evaluate the process effectiveness on an
Individual component basis and to obtain
Information to calculate a material bal-
ance around the TFE.
RESULTS AND DISCUSSION
The feed rate and temperature of the
TFE were varied over its normal range with
no observed operational problems when
operated at atmospheric pressure with the
tested sludges. There were difficulties
when it was operated under vacuum at 320
*C, as some carryover of feed into the
condensate was observed. The condensate
from the vacuum runs was a milky-white
emulsion, which would require additional
treatment to separate the oils.
When the TFE was operated at the 150
*C, some of the water in the feed was
evaporated along with most of the VO. As
illustrated in Table 2, the VO removals
ranged between 99.5 and 99.8 percent at
the low feed rate and decreased slightly
when the feed rate was increased (98.6 to
99.8 percent at a feed rate of 154 Ib/h).
The bottoms temperatures for these runs
were 98 to 102 *C, indicating that water
was still boiling from the bottoms as it
exited the base of the TFE.
The removal efficiency for volatiles
was greater when the TFE was operated at
higher temperatures. The VO removals when
the TFE was operated at 320 *C were 99.88
to 99.99 percent, with no clear trends
relative to changes in feed rate. These
runs removed essentially all of the water
and VO from the feed sludge, along with
much of the higher boiling oils. The
amount of bottoms sludge produced ranged
between 10 and 13 percent of the feed
rate, substantially reducing the amount of
material requiring disposal. This bottoms
product was a relatively viscous, high
solids content sludge, which was still
pumpable.
Several vacuum runs were performed
with the TFE. These runs produced a
milky-white emulsion as condensate, which
contrasted substantially with the cleanly
separating organic/aqueous condensate of
the atmospheric pressure runs. A high-
temperature vacuum run (320 °C) produced a
bottoms product that was only 4.3 percent
of the feed sludge. This indicates that a
two-stage process (first-stage removal of
water and volatiles at atmospheric pres-
sure, second-stage removal of heavier oils
under vacuum operation at high tempera-
ture) could be employed to reduce substan-
tially the amount of sludge material
requiring disposal.
Metals in the feed sludge appeared to
remain in the bottoms products. Only
minor amounts of metals were found in the
organic condensate.
The organic condensates produced dur-
ing the atmospheric pressure tests could
easily be recycled as raw products to the
refinery operation. This would reduce the
actual operating costs of the TFE while
removing organics from the wastes prior to
disposal. The aqueous condensate was
water saturated with the recovered organ-
ics and could be sent to existing waste-
water treatment facilities at refineries.
If the process is applied for the removal
of volatiles only, the aqueous condensate
could be recycled to the process feed, so
that all water would exit with the process
bottoms. (This would not be practical if
large quantities of water were condensed,
as in high-temperature operation) The
bottoms sludge produced by the process
could be treated by existing methods (land
treatment) or perhaps incinerated. The
-492-
-------
TABLE 2. TFE VO REMOVAL FOR SELECTED COMPOUNDS
Operating Conditions
Test
No.
5
7
8
10
Temperature
Cc)
150
150
310
310
Flow
rate
Ob/hr)
71.6
153.7
68.5
143.4
Reduction in concentrations from feed (%)a
Benzene Toluene Ethyl benzene
99.58 99.61 99.48
99.73 99.78 98.83
99.72 99.84 99.68
99.76 99.90 99.78
m-Xyl ene
99.54
98.64
99.67
99.75
aBased on GC/MS analyses.
-493-
-------
ultimate treatment for any bottoms product
will require additional testing of pos-
sible disposal methods.
The results from this test were used
to verify a model that can be used to
predict the effectiveness of TFE treatment
on different waste sludges. Generally, a
TFE 1s modeled as a one theoretical stage
separation device. Countercurrent steam
purging is thought to improve the separa-
tion to approximately one and one half
theoretical stages. The number of
theoretical stages, together with the
partition coefficient (K) of the volatile
organic component at the TFE operating
temperatures, and the flow rates can be
used to predict the percent removal of
volatile organics from the sludge in a
• treatment device. For a complete discus-
sion of the TFE model and the pilot test,
the reader is referred to Harkins, Allen,
and Northeim (1987) (1).
CONCLUSIONS
The TFE was found to have very high
removal efficiencies of VO compounds from
the waste sludges tested. In each of
three methods used to assess the reduction
of volatiles (two headspace analyses, one
analysis of VO compounds in feed and bot-
toms), the removal efficiencies for VO
compounds were greater than 99 percent.
REFERENCE
1. Harkins, S. M., C. C. Allen, and C. M.
Northeim, 1987. Pilot-Scale
Evaluation of a Thin-Film Evaporator
?or Volatile Organic Removal from Land
Treatment Sludges. U.S. Environmental
Protection Agency, Hazardous Waste
Engineering Research Laboratory,
Alternatives Technology Division.
(Draft. Prepared by the Research
Triangle Institute under EPA Contract
No. 68-02-3253, Work Assignment 1-6.)
-494-
-------
U.S. ENVIRONMENTAL PROTECTION AGENCY
HAZARDOUS WASTE CONTROL TECHNOLOGY DATA BASE
C. S. Fore
DOE Hazardous Waste Remedial Actions Program
Oak Ridge National Laboratory
Oak Ridge, TN 37831
P. H. Dalfonso
Automated Sciences Group, Inc.
Oak Ridge, TN 37830
C. C. Lee
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, OH 45268
ABSTRACT
The U.S. Environmental Protection
Agency (EPA) Hazardous Waste Control
Technology Data Base was developed by
the U.S. Department of Energy's
Hazardous Waste Remedial Actions Program
through a joint interagency agreement
with EPA's Hazardous Waste Engineering
Research Laboratory. The data base
functions as an information resource on
thermal treatment technology for
handling hazardous wastes. It serves as
a multifunctional information tool to
support permit writers, researchers,
private industry, and decision makers in
managing, analyzing, and comparing
similar waste components and
technologies. The features of the data
base incorporate (1) engineering data on
permit applications for existing, new,
and research development and
demonstration facilities; (2) trial burn
and design data; (3) on-line report
generation capabilities; and (4) methods
for conducting similarity analyses.
Interactive menu-driven retrieval
options have been designed to generate
summary reports through selection of key
parameters. The data base can be used
to conduct similarity analyses of the
data for (1) re-evaluating the
technology data to compare with actual
industry performance and operating
conditions, (2) providing a reference
guide to meet new and existing
regulatory standards, and (3) providing
a means of calculating the theoretical
performance of trial burns. Overall,
the data base functions as a means of
tracking the status of permit
applications, assists the decision
makers in determining future, research
strategies, provides support data for
public hearings on permit decisions, and
supports EPA's regulatory standards and
procedures. The scope of the data base
is being expanded to incorporate
technical data on chemical, physical,
and biological treatment technologies.
-495-
-------
ANALYSIS OF SAMPLES FROM THE GATEWAY NATIONAL RECREATION
AREA AT JAMAICA BAY, NY
David G. Olson
NUS Corporation
Edison, NJ 08837
Uwe Frank and Michael Gruenfeld
USEPA Hazardous Waste Engineering Research Laboratory
Edison, NJ 08837
John Tanacredi
National Park Service
Brooklyn, NY
The National Park Service (NPS)
provided the Releases Control Branch
(RCB) of the Hazardous Waste
Engineering Research Laboratory with
funding to perform a preliminary study
of chemical pollutants entering the
Jamaica Bay ecosystem at several
designated sites near municipal
landfills in Brooklyn, NY. The NPS
needs this information to make an
initial assessment of whether these
landfills are potential contributing
sources of pollution. The landfills
will eventually be reclaimed by NPS as
part of the Gateway National
Recreation Area park system.
Bay samples were screened for the
presence of priority pollutants.
These samples were collected by NPS
personnel at several sites near the
landfills along the north shore of
Jamaica Bay, NY and analyzed according
to the requirements in 40 CFR Part 136
(Guidelines Establishing Test
Procedures for the Analysis of
Pollutants under the Clean Water Act:
Final Rule and Interim Final Rule and
Proposed Rule). The resulting data
will be presented in this poster. In
addition, the poster presents
designated sampling sites on the bay
and demonstrates the impact of
pollution on the surrounding
communities.
In the first part of the study (March
1985), appreciable concentrations of
pollutants were found at the following
levels: base/neutral and acid
extractables (BNAs) 8-42 ppb (water),
300-9800 ppb (soil); polynuclear
aromatic hydrocarbons (PAHs) 1.74 ppb
(one water sample), 55.9-3515 ppb
(soil); polychlorinated biphenyl
(PCBs) aroclor 1260 5.45-43.2 ppb
(soil); metals 0.11-6.41 ppm (water),
0.42-125 ppm (soil).
In the second part of the study
(May-June 1986), appreciable amounts
of base/neutral and acid extractables
(BNAs) and metals were found in the
landfill sediment samples, with the
following ranges of maximum values:
495-6,740 ppb BNAs, 2.4-910 ppm
metals. Aroclor 1260 was the only
polychlorinated biphenyl (PCB) found
in any of the samples (88.4 ppb in bay
water and 128 ppb in sediment).
Polynuclear aromatic hydrocarbons
(PAHs) were found in both the bay
water and sediment samples (5.16 ppb
maximum); landfill samples were not
analyzed for PAHs. In order to
evaluate and document the quality of
the analytical data generated, an
extensive internal QA/QC program was
followed, which will also be described
in this poster.
-496-
-------
CASE EVALUATIONS OF RD&D PERMIT APPLICATIONS
By
W. Clark, P.M. Maly, W.R. Seeker
Energy and Environmental Research Corporation
18 Mason
Irvine, California 92718
And
C.C. Lee
U.S. Environmental Protection Agency
26 West Saint Clair Street
Cincinnati, Ohio 45268
The Environmental Protection Agency
(EPA) is charged with regulating the in-
cineration of hazardous waste under the
Resources Conservation and Recovery Act
(RCRA). Under current procedures the
regional permit writer must make a number
of engineering assessments during the
course of the permit application concern-
ing the adequacy of the design data, the
consistency of the trial burn data, the
appropriate limits to be set on operating
conditions, and the parameters to be moni-
tored to ensure continual compliance. A
computer model has been developed to
assessments, allowing them to perform fast
and accurate energy and mass balance cal-
culations for hazardous waste incinera-
tors based on sound engineering principles.
The mass balance is based on simple
stoichiometric calculations assuming com-
plete combustion. The energy balance
solves for gas temperature, wall temper-
ature, and shell temperature considering
chemical heat, heat of vaporization, sen-
sible heat, radiation, convection, and
conduction, this procedure can be useful
to incinerator designers and operators
and especially to permit writers. Permit
writers can use the procedure to evaluate
the feasibility of incinerator designs
and concepts, to issue research, develop-
ment and demonstration (RD&D) or con-
struction permits and to evaluate the
consistency of trial burn measurements
and set appropriate operating limits on
the basis of those measurements for op-
erating permist of hazardous waste incin-
erators. The procedure has been used to
evaluate an actual RD&D permit application
by Waste-Tech Services, Inc. for a fluid-
ized bed incineration system. The in-
cinerator is predicted to be capable of
achieving the times, temperatures, and
velocities claimed in the permit applica-
tion, and incinerator performance is
predicted to be most sensitive to changes
in air flow; waste feed, waste heating
value, and auxiliary fuel flow.
-497-
-------
OVERVIEW: EPA's MOBILE INCINERATION SYSTEM AND
TRANSPORTABLE INCINERATION
A. C. Gangadharan, H. Mortensen, and A. Sherman
Enviresponse, Inc.
Livingston, NO 07039
F. Freestone and J. Yezzi
USEPA Hazardous Waste Engineering Research Laboratory
Edison, NJ 08837
Over the past five years, the EPA
mobile incineration system has been
successfully tested using
PCB-contarainated liquids and
dioxin-contaminated liquids and
solids. A field demonstration on a
variety of dioxin-contaminated liquids
and solids was conducted between July
1985 and February 1986. A total of
2,000,000 Ib of solids and 180,000 Ib
of liquids were successfully
decontaminated during that time.
Subsequently, four design
modifications have been made to
improve throughput capacity and
availability.
The modifications include changes to
the ram feed system, the addition of
an oxygen enrichment system in the
kiln, installation of a cyclone
between the kiln, and secondary
combustion chamber (SCO, and
installation of a wet electrostatic
precipitator downstream of the SCC.
Key design features of these
modifications and their potential
benefit to performance and operating
costs will be discussed. Updated
operational data will be presented as
available.
This presentation will also discuss
conceptual configurations of a
transportable incineration system that
has a thermal duty of 75 million
Btu/hr. Cost estimates of a
truck-mounted transportable
incineration system will be described.
-498-
-------
BOILER COFIRING OF CHLORINATED HYDROCARBONS
John H. Wasser
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
ABSTRACT
An in-house project to study the destruc-
tion of liquid hazardous waste by cofiring
in an industrial package boiler has recently
been completed. The test facility
consisted of a 2.5 million Btu per hour
North American Scotch Marine boiler, a
surrogate hazardous liquid feed system, a
mini-VOST sampling train, and a continuous
emissions monitoring system (CEMS).
Hazardous wastes were simulated by mixing
NO. 2 fuel oil with carbon tetrachloride,
monochlorobenzene, and perch!oroethylene.
Each of these principal organic hazardous
compound (POHC) surrogates constituted 2%
by weight of the mixture. This mixture
was cofired with natural gas in the
boiler. The overall program objective
was to define the envelope of operating
conditions where a 99.99% or greater
destruction and removal efficiency (ORE)
could be achieved for the POHC. Waste
surrogate and fuel (natural gas) were
cofi red under both steady state and
transient conditions. Boiler load,
stoichiometry, waste heat percent, and
degree of atomization were varied under
steady state conditions, while load ramp
up/down, stoichiometric fluctuations, and
waste feed rate fluctuations were the
parameters evaluated under transient
conditions. Stack emissions of NOX, CO,
HC, and oxygen were measured by CEMS,
while chlorinated POHCs were sampled with
the mini-VOST train and analyzed using a
GC (Hall detector). Samples of the waste
mixture and internal boiler deposits were
also collected. Indications from the
results are that less than 99.99% DRE was
experienced with a boiler stoichiometry
near 1.0, a high firing rate (80%), and
50% waste heat input under both good and
bad atomization Conditions. A DRE less
than 99.99% resulted under transient con-
ditions of load ramp up/down, stoichio-
metry fluctuation, and waste feed rate
fluctuation. Monochlorobenzene is the
most difficult of the tested compounds to
destroy, indicating that aromatic chloro-
hydrocarbons could be the major concern
in cofiring. Steady, moderate boiler
operating conditions give the best
assurance of maintaining 99.99% DRE.
-499-
-------
DEMONSTRATION, TESTING, AND EVALUATION OF COMMERICAL
TECHNOLOGIES UNDER THE SITE PROGRAM
Seymour Rosenthal
Enviresponse, Incorporated
Livingston, NJ 07039
M. Stinson
USEPA Hazardous Waste Engineering Research Laboratory
Edison, NJ 08837
The Superfund Innovative Technology
Evaluation (SITE) Program provides a
mechanism by which the US EPA can
cooperate with the private sector in
maximizing the use of innovative
remediation technologies.
The SITE Program provides certain
support and financial assistance for
the development and demonstration of
promising technologies, thus enhancing
their establishment and availability.
The-poster describes some technologies
currently being examined under the
program. They include: the
International Waste Technologies
solidification process, which will be
demonstrated on PCB-contaminated soil
in Hialeah, Florida; the Westinghouse
plasma arc process that will be tested
at Love Canal; and the Shirco
incinerator planned for a
demonstration at Tampa, Florida.
-500-
-------
CONDITIONS WHICH ENHANCE BIODEGRADATION OF
ORGANIC COMPOUNDS BY WHITE ROT FUNGI
T. Fernando, J. A. Bumpus, S. D. Aust
Department of Biochemistry
Michigan State University
East Lansing, Michigan 48824
ABSTRACT
The white rot fungus Phanerochaete
chrysosporium is able to degrade a broad
spectrum of typically hard-to-degrade
environmental pollutants. The present
study focused on culture conditions that
enhance the mineralization of environmental
pollutants. Special attention was given to
the use of inexpensive carbohydrate sources
to develop a useful and economically prac-
tical waste treatment system. The initial
rate of mineralization of ^^C-DDT, which
was used as a model organopollutant, was
Unear (i.e., first order) with respect to
C-DDT concentration. Cultures containing
one percent glucose, starch, or cellulose
mineralized 14.5%, 25.1%, and 32.1%,
respectively, of the 14C-DDT during a 90
day incubation period. No significant
difference in the initial rate of 14C-DDT
mineralization was observed. However,
mineralization was minimal after 18 days of
growth on glucose, but continued at sub-
stantial rates in cultures in which cellu-
lose or starch served as the carbohydrate
source. Cultures grown on used newspaper,
wheat straw, or corn cobs mineralized 8.1%,
13.3%, and 8.2%, respectively, of the 14C-
DDT during 30 days of incubation. In other
studies it was shown that 42% of the 14C-
DDT initially present was mineralized in 60
days in culture in a medium composed of
(1:7.5 w/v) horse manure liquor, 8% cellu-
lose and ,2.5% glucose. The rate and extent
of 14C-DDT mineralization was also found to
be affected by the concentration of trace
minerals. For example, a twenty-fold in-
crease (from 0.004 uM to 0.08 uM) in the
concentration of FeS04 resulted in a two-
fold increase in the amount of 14C-DDT
mineralized. Similar effects were observed
when the concentrations of CuS04, ZnS04,
and MnSOd. were varied. These studies dem-
onstrated that it is possible to substan-
tially increase the rate and extent of
xenobiotic degradation by P^_ chrysosporium.
They further suggest that it may be possible
to develop an efficient and economical
waste treatment system based on the use of
this microorganism. (Supported by USEPA
Cooperative Agreement CR811464).
-501-
-------
DEMONSTRATION AND EVALUATION OF THE EPA
MOBILE CARBON REGENERATOR
Patricia Brown and H. Mortensen
Enviresponse, Inc.
Livingston, NJ 07039
R. Traver
USEPA Hazardous Waste Engineering Research Laboratory
Edison, NJ 08837
This poster presents information on
the EPA Mobile Carbon Regenerator,
which was developed as part of EPA's
overall mission to actively encourage
the use of cost-effective, advanced
technologies during cleanup
operations. Specifically, the
regenerator is intended to provide
on-slte reactivation of granular
activated carbon (GAC) used to adsorb
contaminants at Superfund sites. The
regenerator consists of a rotary kiln
with afterburner, scrubber, and
auxiliary equipment mounted on a
serai-trailer, and is capable of
regenerating up to 100 Ib/hr (dry
basis) of carbon.
Automated feed- and product- handling
systems were added during the current
phase of system development.
Most recently, the regenerator was.run
in a demonstration operation using GAC
from the Stringfellow hazardous waste
site, Riverside County, CA. Results
of this operation are presented,
together with the results of a
previous operation using GAC spiked
with tetrachloroethylene and with
ortho-dichlorobenzene.
-502-
-------
PRETREATMENT OF LAND-TREATED WASTES
William E. Gallagher* PE
Thomas C. Ponder, Jr., PE, CCE
Joseph B. Murray
PEI Associates, Inc.
Arlington, Texas 76012
ABSTRACT
The purpose of this project is to
investigate hazardous waste dewatering
at petroleum refineries to evaluate
air emissions associated with this
pretreatment operation. Refineries
routinely dewater.API sludges and dis-
solved air flotation (DAF) float prior
to disposing of this material. The
sludge from these dewatering opera-
tions is a listed hazardous waste and
is typically either land-treated or
disposed of in a hazardous waste dis-
posal facility.
Specifically, this project deter-
mined the fate of both semivolatile
and volatile organic carbon during the
dewatering operation. By reviewing
existing data, the different types of
dewatering equipment used in this
operation were identified. Sev-
eral refineries were visited and a re-
finery using a belt filter press was
selected for testing. The refinery
processed API and DAF sludge separate-
ly (most refineries mix these sludges
prior to dewatering). The facility
had a forced draft ventilation system
which made it possible to test air
emissions from the belt filter through
an exhaust stack. .-.-•,
Two, days of testing.were conducted,
;one day,while API sludge was being,
processed and a second day while DAT
float was being dewatered. Each test
ran for six hours. During the test,
inlet and outlet streams to,the filter
were measured and sampled hourly along
with testing to quantify volatile arid
semi-volatile organics.in the air", ,
:vented from the filter, building.
The samples were analyzed, for 13
specific organic compounds, as well as
purgible and non-purgible organic,car-
bon. Solid, oil and water content of
each stream was also analyzed. Data
taken during the test were used to
quantify the flows in and out of the
filter, and a material balance was done
as a cross-check on the accuracy of our
data. The results of the sampling and
analysis effort were used to determine
the fate of the organics and determine
the level of air emissions generated
from this type of operation.
-503-
-------
GEOTECHNICAL ANALYSIS FOR REVIEW OF DIKE STABILITY
R.M. McCandless, A. Bodocsi, P. Cluxton and M.S. Meyers
Department of Civil and Enrivonmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
The structure and capabilities of a
user-friendly, interactive computer pro-
gram developed for the stability analysis
of dikes (CARDS) are described. The CARDS
program is designed to guide the geotechni-
cal user through the customary steps of
earth dike analysis. The significant
difference between CARDS and other stabil-
ity programs commonly available is the
opportunity to perform all of the follow-
ing analyses in a single package: automa-
tic search to determine the critical fail-
ure surface for both rotational (slip-cir-
cle) and trans!ational (wedge) stability
analyses; automatic search to locate zones
of greatest liquefaction potential and to
compute total and differential settlements
of foundation soils; finite element hydrau-
lic analysis to determine the steady state
piezometric surface through the section
(including the case of an impermeable
barrier such as a clay liner); evaluation
of excess pore pressure conditions pro-
duced by confined steady state flow and
potential slope stability and uplift con-
ditions resulting therefrom; determine the
maximum exit gradient and the potential
for piping failure.
The program was developed under the
the sponsorship of the U.S. Environmental
Protection Agency and therefore emphasize^
hazardous waste applications although it
is suitable for general use. The GARDS
package is designed for use on the IBM/PC/
XT microcomputer. Documentation consists
of a technical users maunal which presents
basic theory, program operational proce-
dures, and example computer solutions.
-504-
-------
LAND BAN DATA NEEDS
Ronald J* Turner
TDB, ATDj HWERL
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
All of the RCRA hazardous wastes
listed as of November 8, 1984 were
ranked by their intrinsic hazards and
their volumes, and a time frame was
established for conditional land dis-
posal restrictions. The EPA will
obtain datel on the operation and per-
formance of demonstrated, available
treatment technologies for these
hazardous wastes. The HWERL sampling
and analysis program is described.
-505-
-------
DEMONSTRATE COMPUTER ASSISTED ENGINEERING (CAE)
TECHNIQUES. FOR REMEDIAL ACTION ASSESSMENT
Phillip R. Cluxton
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Computer Assisted Engineering (CAE)
refers to a broad range of powerful soft-
ware tool packages which aid engineers
primarily in the fields of machine design,
microelectronic circuit design and struc-
tural design. Since powerful personal
computers are becoming widely available, ,
CAE systems using these computers are
being developed to assist with many other
types of engineering problems.
A CAE system customized for remedial
action assessment will be developed by
integrating several existing software
packages, including a Computer Aided
Design/Drafting (CADD) package, a Geo-
graphic Information System (GIS), and a
groundwater modeling package. The result-
ing system will be demonstrated by applying
it to evaluation of remedial action
alternatives at a Superfund site. ,
-506-
-------
HAZARDOUS WASTE RESIDUALS CHARACTERIZATION
H. Paul Warner
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
The Office of Solid Waste and •
Emergency Response is establishing
regulations for the disposal of waste
or residue into the land. In the
development of these regulations, an
evaluation of several hazardous waste
treatment techniques was made at se-
lected field sites. This presentation
summarizes the results of these eval-
uations with respect to the character-
ization of treatment residuals and
their suitability for disposal to the
land. Included in this presentation,
where data is available, is a summary
of the results from the application of
the Toxicity Characteristic Leaching
Procedures (TCLP) on treatment resid-
ual s. Also discussed will be the
Paint Filter Test and Liquid Release
Test, however, primary emphasis will
be placed on a discussion of the TCLP.
The TCLP, the Paint Filter Test, and
the Liquid Release Test are three
criteria proposed to be used by the
U.S. Enviromental Protection Agency
to regulate disposal of hazardous
waste residuals to the land on a
pass/fail basis.
-507-
-------
COST ENGINEERING MODELS FOR REMEDIAL RESPONSE TECHNOLOGIES
William Kemner, John Abraham, Jay Palmisano
PEI Associates, Inc.
Cincinnati, Ohio 45246
ABSTRACT
The purpose of this project was
to develop a cost estimation model
for remediation technologies at
hazardous waste sites. The work was
conducted for the U.S. Environmental
Protection Agency (EPA), Hazardous
Waste Engineering Research Laboratory
in cooperation with the Office of
Emergency and Remedial Response. The
model is designed for operation on a
personal computer and can provide
levels of detail in the estimate
commensurate with the detail of input
parameters available, i.e., the, stage
of design.
The cost estimation model takes
into account variations in site
characteristics, waste quantities,
worker safety requirements, and
regional costs. Output formats and
types of cost calculations are con-
sistent with the EPA Remedial Action
Costing Procedures Manual. Default
values for interest rates, discount
rates, planned life, contingency,
allowances, etc. are consistent with
the costing procedures manual.
The model can be used by Federal
or State government and contractors
(REM/FIT Contractors, cleanup con-
tractors, etc.) for quick and effec-
tive calculations of costs for evaluat-
ing and comparing a variety of different
cleanup or remediation alternatives.
-508-
-------
TRIAL BURN MEASUREMENT GUIDANCE
Roy Neulicht
Andrew Trenholm
Midwest Research Institute
Kansas City, Missouri 64110
ABSTRACT
The Resource Conservation and Recovery
Act (RCRA) requires the U.S. Environmental
Protection Agency (EPA) to develop, promul-
gate, and implement regulations which
control the generation transportation,
treatment, storage, and disposal (TSD) of
hazardous wastes. An integral part of
these activities is reviewing and issuing
permits to hazardous waste incineration
(HWI) facilities.
The performance of HWI facilities is
evaluated by conducting a "Trial Burn"
during which measurements are made on
the various input and output streams to
the incinerator. In this presentation,
the measurement aspects of a trial burn
are reviewed. The review is oriented
towards how the measurements are made.
The major elements of incineration measure-
ments related to the process monitoring,
sampling, and analysis aspects of trial
burns are discussed. Measurements reviewed
will include volatile and semivolatile
Principal Organic Hazardous Constituents
(POHCs) and Carbon Monoxide (CO). A
Volatile Organic Sampling Train (VOST) will
be demonstrated.
-509-
-------
MICROSCOPIC AND MICROCHEMICAL ANALYSES OF SOLIDIFIED
INORGANIC WASTES CONTAINING INTERFERENCE COMPOUNDS
H. C. Eaton, M.. E. Tlttlebaum and F. K. Cartledge
Hazardous Waste Research Center
Louisiana State University
Baton Rouge, Louisiana 70803
Solidification/stabilization processes are thought to be affected
by the presence of certain interference compounds. The mechanism of
interference is unknown but is believed to involve the microscopic
phases which are normally formed during hydration of the pozzolanic
matrix materials. Recent studies, for example, have characterized the
interaction between selected organic wastes and portland cement and have
shown that set retardation results from an alteration of the chemical
and morphological structure of the calcium silicate hydrates that nor-
mally impart structural integrity to the hardened paste.
The present study is an attempt to determine if inorganic wastes
(interferences) have the same detrimental effect upon similarly solidi-
fied inorganic sludges. X-ray powder diffraction is used to identify
the crystalline phases which differ in chemical composition or amount
from those contained in sludge solidified in the absence of interference
compounds. Scanning electron microscopy is used to characterize the
morphology and to identify the form of the important phases. The exper-
imental results are described and their importance to the design of
effective solidification schemes is discussed.
r510-
-------
VACUUM-ASSISTED IN-SITU STEAM STRIPPING
TO REMOVE POLLUTANTS FROM CONTAMINATED SOIL
Arthur E. Lord, Jr., Robert M. Koerner and Vincent P. Murphy
Geosynthetics Research Institute
Drexel University
Philadelphia, Pennsylvania 19104
and
John E. Brugger
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Edison, New Jersey 08837
ABSTRACT
A long term project has recently been
initiated (Fall 1986) by Drexel University,
under the sponsorship of the U. S. Environ-
mental Protection Agency, to develope the
means by which to steam-strip (distill)
pollutants from contaminated, soils.
The overall project is divided into
four phases:
Phase I - Perform literature and"infor-
mation search and design of
small scale laboratory experi-
ments.
Phase II - Run small scale laboratory
experiments to determine the
feasibility of steam stripping
of a wide range of chemicals
for pertinent soil types and
their different conditions.
Phase III - Design and perform small pilot
scale steam stripping and
vacuum entrapment experiments
on selected chemical/soil
combinations.
Phase IV - Perform final design of field
unit. Possible supervision
of the construction of a field
deployable unit working close-
ly with EPA-Edtson personnel.
Although a great deal of chemistry
and chemical engineering, literature is
available as regards steam stripping and
steam distillation, there .is virtually no
literature concerning the interaction of
steam with soils. Thus the preliminary
work has involved simple experiments to
develope basic knowledge in this area.
The volume expansion of a variety of
soil types exposed to steam (at about
100°C and one atmospheric pressure) in an
autoclave is discussed. Also the behavior
of a steam front (injected at 100°C and
5 psi) in a soil-filled, thin transparent
box will be presented. From these steam
front movements, effective permeabilities
for steam movement are derived and com-
pared with water permeabilities.
Containment of the stripped waste is
an absolutely essential aspect of the work.
Our geosynthetic containment system is of
a unique design, and some preliminary
results with this containment system are
given.
-511-
-------
USE OF MODIFIED CLAYS FOR ADSORPTION AND CATALYTIC
DESTRUCTION OF CONTAMINANTS
S.A. Boyd and M.M. Mortland
Department of Crop and Soil Sciences
and
T.J. Pinnavaia
Department of Chemistry
Michigan State University
East Lansing, Michigan 48824
ABSTRACT
The unique properties of smectite clays
that are important in designing effective
adsorbents and catalysts are (1) high surface
area (2) high cation exchange capacity (3)
swelling properties of the clay platelets and
(it) high surface acidity and reactivity. Such
properties make these materials inherently
reactive and allows tremendous synthetic
versatility to produce structures having
specific properties for detoxication. Our
goal is to develop a process for treating
liquid wastes in which the contaminant is
first immobilized on the clay material and
then detoxified via catalysis on the clay
surface. The clay-based materials being
developed for this purpose are (1) transition
metal saturated clays (2) organo-clays (3)
pillared clays and (4) delaminated clays. The
metal clays are versatile catalysts which can
be used to form reactive radical cations of
organic toxicants such as dioxins and
chlorophenols. These reactive radical
species are subject to a variety of potential
reactions to form less toxic products.
Important reactions such as polymerization
and dechlorination have been achieved.
Organic cations can also be placed on the
surface of smectite clays to impart
hydrophobic properties important for
adsorption of trace organic contaminants
from aqueous streams. Pillared and
delaminated clays are microporous to
macroporous derivatives formed by the
reaction of natural clays with robust
polyoxocations. The relatively large pore
sizes of these clays make them ideal
adsorbents for large chlorinated
hydrocarbons. Also, metal catalysts can be
introduced into their structures to facilitate
subsequent oxidation of the adsorbed
toxicant.
-512-
-------
STRINGFELLOW LEACHATE TREATMENT WITH A ROTATING BIOLOGICAL CONTACTOR'
Edward J. Opatken, Hinton K. Howard, and James J. Bond
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
A study was conducted with a rotating
biological contactor (RBC) for treatment
of leachate from the Stringfellow hazard-
ous waste site in Riverside County, Cali-
fornia. The leachate was transported from
California to Cincinnati, where a pilot
sized RBC was installed at the U.S. EPA's
Testing and Evaluation (T&E) Facility.
A series of kinetic runs were made
with primary effluent from the City of
Cincinnati's Mill Creek Sewage Treatment
Plant to develop the biomass on disks and
to obtain a standard kinetic removal rate.
These runs were then followed with String-
fellow leachate experiments that included
0 Operations at various ratios of
leachate to primary effluent
0 Operations at 100% leachate
0 Operations to increase the percentage
removal of dissolved organics.
This poster presentation reports on
the results from these experiments and the
effectiveness of an RBC to adequately treat
leachate from the Stringfellow hazardous
waste site.
-513-
-------
SEPARATION AND RECOVERY OF HAZARDOUS WASTES
Kenneth E. Noll
Charles N. Haas
James W. Patterson
Pritzker Department of Environmental Engineering
Illinois Institute of Technology
Chicago, IL
ABSTRACT
Two projects at the Illinois Institute
of Technology that are currently supporting
the basic research activities of EPA in-
volve multi-component adsorption mechanisms
and metals speciation and precipitation
processes. These two pieces of work are
Intended to extend the basic understanding
of separation and recovery technologies
which are vital to the current legislation
controlling hazardous waste generation,
treatment and disposal.
The first project, dealing with organ-
ic pollutant separation and recovery, has
concentrated on the dynamics of multi-
component adsorption/desorption systems.
The kinetic data and isotherm parameters
obtained from a series of experiments
with a quartz spring apparatus have been
used to develop computer models of multi-
component adsorption in air and in water.
Preliminary tests in long column adsorption
experiments are in good agreement with both
models.
The second project, also investigating
a separation process, has focused on metal
speciation and precipitation phenomena.
The soluble phase kinetics of metal reac-
tions in both single ligand and competitive
ligand systems have been evaluated and
reaction orders and rate constants have
been proposed for several systems. In
addition, the effects of reaction stoichio-
metry and reactor hydraulic regime on the
onset of nucleation and precipitation have
been studied. The resulting precipitates
have been characterized with respect to
their composition, morphology, and surface
charge.
-514-
-------
TREATMENT OF AQUEOUS METAL AND CYANIDE BEARING
HAZARDOUS WASTES
Sardar Q. Hassan, James E. Park, Margaret K. Koczwara
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, Ohio 45221
Douglas W. Grosse
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio 45268
ABSTRACT
With the amendment of the Resource
Conservation and Recovery Act (RCRA) by the
Hazardous and Solids Waste Amendments
(HSWA) and.the resulting restrictions on
land disposal of hazardous wastes, the U$ •
Environmental Protection Agency is
assessing technologies that can be sub/-
stitutes for, or precursors to land
disposal. This paper describes the
research work being conducted at the EPA's
Test and Evaluation Facility in Cincinnati,
Ohio, involving treatment of electroplating
and metal finishing hazardous wastewater.
, A series of treatment units has been
designed and fabricated to determine the
optimum combination of the units for any
given waste. These units are: lime
precipitation, flocculation, clarification,
mixed-media filtration, sulfide precipita-
tion, activated carbon adsorption, ion
exchange and alkaline chlorination of
cyanide. All of these units except
alkaline chlorination of cyanide have been
tested with industrial metal bearing wastes
(EPA RCRA codes D006/D008) .in the first
experimental phase. A process flow rate of
2.5 gallons per minute.was used during
these test runs and a total of 450 gallons
of waste was treated in each run. Ion
exchange was found to be a promising
polishing technology.
In the second experimental phase, the
alkaline chlorination of cyanide will be
tested after the necessary safety features
are incorporated into the system. The
system will be modified to operate at a ,
lower flow rate of 1.0 gallons per minute
to reduce the amount of hazardous waste
that has to be imported and handled during
each run. Plans for utilizing additional
treatment units (chromium reduction,
stabilization/fixation of precipitated
sludges) will also be presented.
-515-
-------
AN EXPERIMENTAL INVESTIGATION OF SINGLE DROPLET
COMBUSTION OF CHLORINATED HYDROCARBONS
N.W. Sorbo, D.P.Y. Chang, C.K. Law
Departments of Civil and Mechanical Engineering
Davis, California 95616
R.R. Steeper
Combustion Research Facility, Scandia National Laboratories,
Livermore, California, 94550
ABSTRACT
Experiments have been conducted with
chlorinated hydrocarbons to investigate
the combustion characteristics of liquid
hazardous wastes. In this study, the
combustion and vaporization of single
droplets of pure chlorinated hydrocar-
bons were examined and the effect of
adding hydrocarbons to the chlorinated
hydrocarbon to facilitate burning was
investigated. Single droplets were pro-
duced by a piezoelectric generator,
injected into a hot chamber, and sized
by microphotography. In addition, the
liquid phase history of the droplet was
determined using phase discriminated
sampling and GC analysis. The size
history of the droplets was used to
generate D2-Law plots from which burning
rate parameters were extracted. Evidence
of droplet extinction appeared on D2-Law
plots for certain mixtures of chlorinated
and non-chlorinated hydrocarbons. From
this study, the effects of chlorine
loading, volatility differentials in mix-
tures, and mixture percentage in single
droplet combustion and vaporization are .
presented. The implications of these
effects on the behavior of full scale
hazardous waste incinerators are dis-
cussed.
-516-
-------
CATALYTIC DESTRUCTION OF HALOGENATED
HAZARDOUS WASTE
Howard L. Greene
The University of Akron, Akron, Ohio 44325
Edward Katz
U.S.E.P.A., Cincinnati, Ohio 45268
ABSTRACT
Disposal of halogenated wastes using
methods of catalytic oxidation is being
evaluated from both technology and eco-
nomic standpoints. Vapor phase oxidation
of halogenated compounds over supported
transition metal oxides shows substantial
reactivity at temperatures well below
those necessary for homogeneous reaction.
The spectrum of oxidation products
obtained depends strongly on the choice
of catalyst and the molecular makeup of
the halogenated species along with any
other molecules present in the vapor
phase.
Catalyst deactivation and corrosion
of the support system are important
design considerations based on the quan-
tities of halogen-acids formed during
the oxidation process. Screening of
both catalysts and supports has yielded
several systems capable of withstanding
reactor conditions.
Although considerable research re-
mains to be done, the underlying flexi-
bility of, the catalytic process makes
it a potentially viable method for
treatment of halogenated wastes.
-517-
-------
EXPERT SYSTEM SCREENING OF REMEDIAL ACTION
TECHNOLOGIES FOR CERCLA SITES
Lewis A. Rossman
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
A prototype computerized expert
system called TECHSCRN has been
developed for identifying hazardous
waste cleanup, control, and treatment
technologies that would be feasible
to apply at Superfund remedial action
sites. The system matches character-
istics of the site and of the materials
being released against the conditions
under which a particular type of tech-
nology is applicable. This initial
screening produces a more manageable
11st of action alternatives that can
then be subjected to more detailed
study. The knowledge base for this
system is being acquired primarily
from existing literature sources.
The prototype system has been
coded in the PROLOG language and
operates on a personal computer.
The system can function as both
a site-specific screening tool
and as a general purpose data
base system on remedial action
technology characteristics. In-
formation on 30 different types
of containment and removal tech-
nologies is currently available.
This knowledge base is being
expanded to include treatment
technologies as well. Future
plans call for adding a design
capability to the system that
combines individual technologies
into overall system alternatives
as well as a cost estimating
capability to provide a better
basis for screening of alter-
natives. A hands-on demonstra-
tion of TECHSCRN will be provided.
-518-
-------
ACTIVITIES AT LOUISIANA STATE UNIVERSITY'S HAZARDOUS WASTE RESEARCH CENTER
Louis J. Thibodeaux, Director
Louisiana State University
Baton Rouge, Louisiana 70803
ABSTRACT
The Hazardous Waste Research Center
at Louisiana State University (LSU) is
conducting fundamental and exploratory
research in these general areas:
environmental media/waste interaction,
incineration, and alternative methods of
treatment/destruction. Although admin-
istered through LSU's College of Engineer-
ing, individual research projects are
being conducted by multi-disciplinary
groups representing 'a number of academic
departments. Ongoing research projects for
1987 are:
. A Study of Immobilization
Mechanisms In Solidification/Stabilization
Using Cement/Silicate Fixing Agents:
. Incineration of Liquid Hazardous
Waste Droplets.
. An Indepth Investigation of Rotary
Kiln Incineration Performance.
. Supercritical Extraction and
Catalytic Oxidation of Toxic Organics,
from Soil. >
. Short Range Air Dispersion of
Toxics from Area Sources.
Detoxification of Dioxin contami-
nated Sludges Using Combined Microbio-
logical and Photolytic Degradative
Approaches.
Quickening regulatory requirements
imposed on the hazardous waste industry
have resulted in> the Center's responding
to industry needs by implementing in
1985 the Industry Associates Program.
Through this program Center resources
and research finding are being applied
to current industrial problems in
hazardous waste treatment and disposal.
One of the first projects conducted under
this program is "Fate and Transport of
Hazardous Materials After Deep Well
Disposal". The purpose of this
research is to determine whether
interactions occur between specific
hazardous waste materials and
disposal formation rock matrices or
impurities within a rock matrix.
-519-
-------
OXIDATION OF PERSISTENT AROMATIC POLLUTANTS
BY LIGNIN-DEGRADING ENZYMES
Kenneth E. HammeT
State University of New York
College of Environmental Science and Forestry
Syracuse, New York 13210
B. Kalyanaraman
National Biomedical ESR Center
Medical College of Wisconsin
Milwaukee, Wisconsin 53226
T. Kent Kirk
Forest Products Laboratory
U. S. Department of Agriculture
Madison, Wisconsin 53705
John A. Glaser
Hazardous Waste Engineering Research Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Ifgninase of the wood-rotting fungus, Phanerochaete chrysosporium, catalyzes the
oxidation of a variety of lignin-related compounds. This enzyme also catalyzes the oxida-
tion of certain aromatic pollutants and compounds related to them. Polycyclic aromatic
hydrocarbons with ionization potentials less than ca. 7.55 eV, are oxidized to quinones,
with the source of the quinone oxygens being water. A related enzyme, horseradish peroxi-
dase, is known to oxidize only those polycyclic aromatic hydrocarbons that have ionization
potentials less than ca. 7.35 eV, j_.e. ligninase is the more oxidizing of the two peroxid-
ases. Ligninase also oxidizes chlorophenols and certain halogenated am"soles. In the
case of 2,4,6-trichlorophenol, the reaction proceeds with dechlorination at the 4-position
to yield 2,6-dichloro-p-benzoquinone as the product, as determined by gas chromatography/
mass spectrometry analysis. Certain dibenzo(p)dioxins are also substrates for ligninase,
and the visible and electron spin resonance spectra of reactions in progress show that
substrate cation radicals are intermediates in catalysis.
-522-
-------
LABORATORY STUDY OF THE THERMAL DECOMPOSITION OF SULFUR HEXAFLUORIDE
Philip H. Taylor
University of Dayton Research Institute
Dayton, Ohio 45469
John F. Chadbourne
General Portland, Incorporated
Dallas, Texas 75221
ABSTRACT
This report presents the results of a
laboratory evaluation of the gas-phase thermal
decomposition of sulfur hexafluoride (SFg).
Thermal decomposition profiles were gener-
ated for pure SFg and SFg doped in a mix-
ture containing carbon tetrachloride (CC14),
tetrachloroethylene (C£Cl4), 1,2,4-
trichlorobenzene (TCB), and toluene (CyHs).
The data were generated in both oxidative
and near-pyrolytic atmospheres with the
residence time for reaction held at 2.0
seconds.
Thermal decomposition profiles for pure
SFg as a function of reaction atmosphere
indicate that SFg is an extremely stable
material with an extrapolated temperature
for 99.99% destruction much greater than
1050°C. The data also indicate no substan-
tial effect of SFg stability on oxygen con-
centration. Thermal decompositon profiles
for the SF5 mixture components evaluated
under oxidative reaction conditions indi-
cate that although SFg begins to degrade at
a temperature of 500°C, its stability
decreases slowly at higher temperatures.
The remaining mixture components exhibit
rapid oxidative degradation above tempera-
tures of 500°C. As a result, SFg is the
most stable component in the mixture by far
with an extrapolated temperature for. 99.99%
destruction much greater than 1050°C.
Similar profiles for the SFg mixture con-
stituents evaluated under near-pyrolytic
reactor conditions indicate that SFg is the
most stable component in the mixture by far
with an extrapolated temperature for 99.99%
destruction much greater than 1050°C.
Considerations of the molecular struc-
ture (S-F bond dissociation energy ~90
kcal/mole) are consistent with the experi-
mental results. Measurements of SFg-OH
radical reactivity in this laboratory
indicate very low reaction cross-sections.
This finding is consistent with SFg's
approximate zeroeth-order decomposition
behavior with respect to oxygen concentra-
tion. In addition, comparison of SFg
unimolecular decomposition as calculated
via RRKM theory is in good agreement with
the pure compound experimental data.
A major application of this study is
the use of SFg as a surrogate to monitor
hazardous .waste incinerator performance as
defined by current regulations for princi-
pal organic hazardous constituent (POHC)
destructability. Two major requirements of
an ideal surrogate are high thermal stabil-
ity (relative to hazardous constituents in
the feed), and the ability of the surrogate
to "track" hazardous consitiuent destruc-
tion. With respect to these requirements,
this laboratory evaluation has shown that
measurement of effluent SFg concentration
may represent an accurate measure of incin-
erator performance. Field studies have
confirmed this hypothesis as 99.9% ORE of
SFg has resulted in 99.999% (destruction
limit 99.9999%) ORE of CC14, C-^k' TCB> °*
any known hazardous waste constituent.
Areas of additional study include
evaluation of the relationship between SFg
and products of incomplete combustion (PICs)
thermal stability. This is important
because previous studies of POHC degrada-
tion indicate much higher stability for
thermal reaction products, especially under
near-pyrolytic conditions.
-523-
-------
THE U.S. EPA COMBUSTION RESEARCH FACILITY
R.W. Ross, R.H. Vocque, and L.R. Waterland
Acrurex Corporation
Jefferson, Arkansas
In the past year, three
major incineration test programs
have been completed at EPA's
Combustion Research Facility
(CRF). The first consisted of
incinerating a synthetic waste
comprised of distilled fuel oil
containing 30 percent PCB-1260 in
the CRF liquid injection
incineration system (LIS). The
second test program consisted of
incinerability testing of four
wastes being generated through
remediation of the Bridgeport
Rental and Oil Services (BROS)
Superfund site in Bridgeport, New
Jersey. The third program
investigates the fate of trace
elements fed to an incinerator as
a function of feed composition
and the incinerator temperature
and excess air level. Results
from these tests are abstracted
in this poster (the BROS test and
the HIS trace element tests are
discussed in other papers in the
symposium).
-524-
-------
CONSTRUCTION AND SHAKE DOWN OF AN ENVIRONMENTAL TESTING
CHAMBER FOR SOIL REAGENT RESEARCH
Michael Black
United States Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
26 W. St. Clair Street
Cincinnati, Ohio 45268
Chemical and biochemical means to
detoxify soil are becoming more promi-
nent in the arsenal of treatment techno-
logies for the detoxification of soil
contaminated with hazardous waste. Due
to permitting restrictions and the lack
of controlled field conditions, the
application of these new technologies
to actual field sites continues to be
throttled. The Environmental Testing
Chamber (ETC) has been assembled to
alleviate some of these difficulties.
At this stage of investigation, the
chamber is currently built and awaiting
more auxiliary equipment in order to
begin shake down and testing to deter-
mine design performance. It is designed
as a generic device, that can study
chemical, biological, physical and
engineering design aspects of hazardous
waste treatment technologies. This
testing chamber will give a better
assessment of the treatability of a
technology at this intermediate scale.
The loss of contamination is best
partitioned into the various environ-
mental and treatment process parts
affecting the hazardous waste site. The
Environmental Testing Chamber will permit
the allocation of contaminant losses to
the various major contributors to the
overall loss.
The main objectives of the ETC
research are: 1) close simulation of
field conditions; 2) improved control of
treatment technology; 3) engineer treat-
ment optimization; 4) enhance scale-up
assessment; 5) provide cost-effectiveness
treatment design; 6) access to all con-
taminated media in the chamber; 7) com-
plete controls of contaminant losses.
The planned future developments of the
chamber include additional aspects to
more completely simulate the real world
such as microclimatology and weathering.
Also, to develop additional prototype
chambers, as necessary, to study other
treatment processes and environmental
settings. Other possible chamber design
changes may include: 1) rainfall simu-
lation; 2) photolytic decomposition
reactions; and 3) surface and subsurface
waterflow as it affects the treatment
process.
-525-
-------
EARTHEN LINERS: PROTOTYPE PHASE OF A FIELD STUDY OF TRANSIT TIME
Karen A. Albrecht, Beverly L. Herzog, Robert A. Griffin, Wen-June Su, Ivan Krapac
and Keros Gartwright
Illinois State Geological Survey
Champaign, IL 61820
ABSTRACT
In the first phase of a project to
determine transit times of water and solute
through a partially-saturated field-scale
earthen liner, a prototype clay liner
(10x30x3 feet) was constructed. The main
goal of the prototype was to determine if
desired moisture, density, and hydraulic .
conductivity relations could be achieved in
the field using the selected clay material
(Batestown Till) and full-size compaction
equipment. Specifically, the prototype was
necessary to determine if the U.S. EPA
hydraulic conductivity criterion of less
than IxlO"7 cm/s for earthen liners could
be met at the study s.ite. Additional goals
were to test instrument installation tech-
niques and provide undisturbed samples of
liner materials.
In-situ prototype infiltration rates
were measured using two large (5T inner
ring diaia.) double-ring infiltrometers.
Apparent steady-state flux of .1.5x10-7 cm/s
was achieved after about 3 weeks. Similar
rates (2xlO~7 cm/s) were obtained from
three smaller infiltrometers (8" inner ring
diam.). Results suggest that-the prototype
met the low hydraulic conductivity require-
ment .
Rhodaraine and fluorescein were placed
in the inner and outer rings, respectively,
of one of the large infiltrometers -to indi-
cate possible preferential flow pathways.
Dye patterns observed during excavation
indicated that some lateral flow occurred
between lift interfaces. Although the
performance criterion was met, the dye
experiments show a need for better lift
bonding.
Morphological study of two prototype
profiles revealed variation in the degree
of compaction within lifts. Thinner lifts
or longer compactor feet are expected to
improve compaction uniformity and lift
bonding.
A combination of vertically and
horizontally, installed tensiometers, gyp-
sum blocks, neutron access tubes, and ly-
simeters were used to monitor moisture and
dye movement. Instruments worked in
either orientation (except lateral neutron
access tubes which were damaged during
construction). However, the horizontal
instruments were necessarily installed at
lift interfaces which may be preferential
flow paths. In addition, horizontal
instruments may be destroyed if the lift
thickness is decreased. Vertical instal-
lations 'showed no evidence of channeling
and were deemed more satisfactory than
horizontal ones.
The next phase of the project is to
build a large clay liner (30x60x3 feet)
for long-term monitoring of ponded water
and tracers.
-526-
•
ft US. GOVERNMENT PRINTING OFFICE: 1087— 7 "» 8- 1 2 II 67008
-------
|